More stories

  • in

    Performance-based criteria for safe and circular digestate use in agriculture

    AbstractAnaerobic digestion converts organic waste into renewable energy (biogas) and recyclable nutrients (digestate), generating over one billion tons of digestate annually. While this represents a major resource, its safe reuse remains a bottleneck for nutrient circularity, particularly for closing global nitrogen loops. We analyzed digestates from 23 full-scale digesters in Sweden, Norway, and Denmark across whole, liquid, and solid fractions using germination index (GI) assays and chemical profiling. Three parameters predicted phytotoxicity: total ammonia nitrogen (TAN ≥ 1,122 mg N L− 1), potassium (K ≥ 39.6 × 103 mg kg− 1), and boron (B ≥ 22.5 mg kg− 1). When all thresholds were exceeded, germination indices dropped below 50% in every case. Based on these findings, we propose a decision-ready framework linking TAN-K-B thresholds to germination outcomes, guiding mitigation through acidification, stripping, blending, or source control. This outcome-based screening reduces monitoring complexity while maintaining compliance with EU and US pollutant ceilings. Its implementation strengthens nitrogen use efficiency, curbs NH3 and N2O emissions, and secures crop establishment. By shifting from origin-based restrictions to performance-based thresholds, our framework provides transparent certification, builds farmer confidence, and positions digestate reuse as a global lever for climate mitigation, nutrient circularity, and food system resilience.

    Similar content being viewed by others

    ​ Uncovering associations between DUS test traits and biochemical composition in safflower germplasm​​

    Article
    Open access
    05 December 2025

    Addressing nitrogenous gases from croplands toward low-emission agriculture

    Article
    Open access
    02 June 2022

    Biochar addition influences C and N dynamics during biochar co-composting and the nutrient content of the biochar co-compost

    Article
    Open access
    10 October 2024

    Data availability

    All data underlying the figures and analyses are provided as Source Data files. Additional processed tables and metadata are available in the Supplementary Material.
    ReferencesRichardson, K. et al. Earth beyond six of nine planetary boundaries. Sci. Adv. 9, eadh2458 (2023).
    Google Scholar 
    Fu, Z. et al. A comprehensive review on the Preparation of Biochar from digestate sources and its application in environmental pollution remediation. Sci. Total Environ. 912, 168822 (2024).
    Google Scholar 
    Ciurli, A. et al. Dried anaerobic digestate from slaughterhouse by-products: emerging cues for a bio-based fertilization. Waste Biomass Valorization. 16, 927–943 (2025).
    Google Scholar 
    European Biogas Association. Exploring Digestate’s Contribution to Healthy Soils. www.europeanbiogas.eu. (2024).Kadam, R. et al. A review on the anaerobic Co-Digestion of livestock manures in the context of sustainable waste management. Energies (Basel). 17, 546 (2024).
    Google Scholar 
    Alberici, S., Wouter, G. & Toop, G. Feasibility of REPowerEU 2030 targets, production potentials in the Member States and outlook to 2050. (2022).Samoraj, M. et al. The challenges and perspectives for anaerobic digestion of animal waste and fertilizer application of the digestate. Chemosphere 295, 133799 (2022).
    Google Scholar 
    Surendra, K. C., Takara, D., Hashimoto, A. G. & Khanal, S. K. Biogas as a sustainable energy source for developing countries: opportunities and challenges. Renew. Sustain. Energy Rev. 31, 846–859 (2014).
    Google Scholar 
    IEA, I. E. A. Outlook for Biogas and Biomethane: Prospects for Organic Growth. (2020). https://iea.blob.core.windows.net/assets/03aeb10c-c38c-4d10-bcec-de92e9ab815f/Outlook_for_biogas_and_biomethane.pdfPecorini, I. et al. Evaluation of MSW compost and digestate mixtures for a circular economy application. Sustainability 12, 3042 (2020).
    Google Scholar 
    Carraro, G., Tonderski, K. & Enrich-Prast, A. Solid-liquid separation of digestate from biogas plants: A systematic review of the techniques’ performance. J. Environ. Manage. 356, 120585 (2024).
    Google Scholar 
    Tallaksen, J., Bauer, F., Hulteberg, C., Reese, M. & Ahlgren, S. Nitrogen fertilizers manufactured using wind power: greenhouse gas and energy balance of community-scale ammonia production. J. Clean. Prod. 107, 626–635 (2015).
    Google Scholar 
    Shi, L. & Zhu, H. En route to a circular nitrogen economy. Nat. Sustain. 7, 1221–1222 (2024).
    Google Scholar 
    AHDB. Nutrient Management Guide (RB209). (2023).IPCC. Working Group III Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change Mitigation of Climate Change. (2022).Daramola, D. A. & Hatzell, M. C. Energy demand of nitrogen and phosphorus based fertilizers and approaches to circularity. ACS Energy Lett. 8, 1493–1501 (2023).
    Google Scholar 
    Giwa, A. S. et al. Advancing resource recovery from sewage sludge with IoT-based bioleaching and anaerobic digestion techniques. J. Environ. Chem. Eng. 13, 116293 (2025).
    Google Scholar 
    Lazzari, A. et al. Evaluating urban sewage sludge distribution on agricultural land using interpolation and machine learning techniques. Agriculture 15, 202 (2025).
    Google Scholar 
    Rocha, F., Ratola, N. & Homem, V. Heavy metal(loid)s and nutrients in sewage sludge in Portugal – Suitability for use in agricultural soils and assessment of potential risks. Sci. Total Environ. 964, 178595 (2025).
    Google Scholar 
    European Parliament and Council of the European Union. Regulation (EU) 2019/1009 of the European Parliament and of the Council of 5 June 2019. Off. J. Eur. Union L 170, 1–114 (2019).CONAMA. Resolução no 498, de 19 de agosto de 2020 [Resolution No. 498, of August 19, 2020]. (2020).Zhang, B., Zhou, X., Ren, X., Hu, X. & Ji, B. Recent research on municipal sludge as soil fertilizer in china: a review. Water Air Soil. Pollut. 234, 119 (2023).
    Google Scholar 
    Logan, M. & Visvanathan, C. Management strategies for anaerobic digestate of organic fraction of municipal solid waste: current status and future prospects. Waste Manage. Res. 37, 27–39 (2019).
    Google Scholar 
    Lencioni, G., Imperiale, D., Cavirani, N., Marmiroli, N. & Marmiroli, M. Environmental application and phytotoxicity of anaerobic digestate from pig farming by in vitro and in vivo trials. Int. J. Environ. Sci. Technol. 13, 2549–2560 (2016).
    Google Scholar 
    Hatamleh, A. A., Danish, M., Al-Dosary, M. A., El-Zaidy, M. & Ali, S. Physiological and oxidative stress responses of solanum lycopersicum (L.) (tomato) when exposed to different chemical pesticides. RSC Adv. 12, 7237–7252 (2022).
    Google Scholar 
    Fast, B. J. et al. Aminopyralid soil residues affect rotational vegetable crops in Florida. Pest Manag Sci. 67, 825–830 (2011).
    Google Scholar 
    Soukupová, M. & Koudela, M. Impacts of aminopyralid on tomato seedlings. Horticulturae 9, 456 (2023).
    Google Scholar 
    Abdourahime, H. et al. Modification of the existing maximum residue levels for aminopyralid in certain cereals. EFSA J. 17, 1–30 (2019).
    Google Scholar 
    van Midden, C., Harris, J., Shaw, L., Sizmur, T. & Pawlett, M. The impact of anaerobic digestate on soil life: A review. Appl. Soil. Ecol. 191, 105066 (2023).
    Google Scholar 
    Kabata-Pendias, A. & Szteke, B. Trace Elements in Abiotic and Biotic Environments (CRC, 2015). https://doi.org/10.1201/b18198Quina, M. J. et al. Studies on the chemical stabilisation of digestate from mechanically recovered organic fraction of municipal solid waste. Waste Biomass Valorization. 6, 711–721 (2015).
    Google Scholar 
    Li, S. et al. A critical review of plant adaptation to environmental Boron stress: Uptake, utilization, and interplay with other abiotic and biotic factors. Chemosphere 338, 139474 (2023).
    Google Scholar 
    Broadley, M. R., White, P. J., Hammond, J. P., Zelko, I. & Lux, A. Zinc in plants. New Phytol. 173, 677–702 (2007).
    Google Scholar 
    Sarker, S., Lamb, J. J., Hjelme, D. R. & Lien, K. M. A review of the role of critical parameters in the design and operation of biogas production plants. Appl. Sci. 9, 1915 (2019).
    Google Scholar 
    Rivero-Marcos, M. et al. Plant ammonium sensitivity is associated with external pH adaptation, repertoire of nitrogen transporters, and nitrogen requirement. J. Exp. Bot. 75, 3557–3578 (2024).
    Google Scholar 
    Britto, D. T. & Kronzucker, H. J. NH4 + toxicity in higher plants: a critical review. J. Plant. Physiol. 159, 567–584 (2002).
    Google Scholar 
    Atta, K. et al. Impacts of salinity stress on crop plants: improving salt tolerance through genetic and molecular dissection. Front. Plant. Sci. 14, 1241736 (2023).
    Google Scholar 
    An, X. et al. Nutrient dynamics during the growth period of epimedium pubescens and its impact on growth and Icariin-Flavonoids composition. Ind. Crops Prod. 225, 120520 (2025).
    Google Scholar 
    Pandey, A., Khan, M. K., Hakki, E. E., Gezgin, S. & Hamurcu, M. Combined Boron toxicity and salinity stress—An insight into its interaction in plants. Plants 8, 364 (2019).
    Google Scholar 
    Bony, L., Dhar, A., Wilkinson, S. R. & Naeth, M. A. Assessing electrical conductivity and sodium adsorption ratio as soil salinity indicators in reclaimed well sites. Land. (Basel). 14, 2125 (2025).
    Google Scholar 
    Lilay, G. H. et al. Linking the key physiological functions of essential micronutrients to their deficiency symptoms in plants. New. Phytol. 242, 881–902 (2024).
    Google Scholar 
    Gjata, I., van Drimmelen, C. K. E., Tommasi, F., Paciolla, C. & Heise, S. Impact of rare Earth elements in sediments on the growth and photosynthetic efficiency of the benthic plant myriophyllum aquaticum. J. Soils Sediments. 24, 3814–3823 (2024).
    Google Scholar 
    Tyler, G. Rare Earth elements in soil and plant systems – A review. Plant. Soil. 267, 191–206 (2004).
    Google Scholar 
    Council of the European Communities. Council Directive 91/676/EEC of 12 December 1991 Concerning the Protection of Waters against Pollution Caused by Nitrates from Agricultural Sources. (1991). https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:31991L0676Wang, R. et al. Electrochemical ammonia recovery and co-production of chemicals from manure wastewater. Nat. Sustain. 7, 179–190 (2024).
    Google Scholar 
    Kanter, D. R., Chodos, O., Nordland, O., Rutigliano, M. & Winiwarter, W. Gaps and opportunities in nitrogen pollution policies around the world. Nat. Sustain. 3, 956–963 (2020).
    Google Scholar 
    Jwaideh, M. A. A. & Dalin, C. The multi-dimensional environmental impact of global crop commodities. Nat. Sustain. 8, 396–410 (2025).
    Google Scholar 
    Brueck, C. L., Nason, S. L., Multra, M. G. & Prasse, C. Assessing the fate of antibiotics and agrochemicals during anaerobic digestion of animal manure. Sci. Total Environ. 856, 159156 (2023).
    Google Scholar 
    Coggins, S. et al. Data-driven strategies to improve nitrogen use efficiency of rice farming in South Asia. Nat. Sustain. 8, 22–33 (2025).
    Google Scholar 
    American Public Health Association. Standard Methods for the Examination of Water and Wastewater. APHA vol. 21. (American Public Health Association, Washington, DC, 2005).Download referencesAcknowledgementsThis study was partly financed by the Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ) through PhD and post doctoral scholarship for TMA. HRO thanks the Brazilian National Council for Scientific and Technological Development (CNPq) for the PhD scholarship. AE-P and AB gratefully acknowledge financial support from the funding agency Formas [Grant number: 2021-02429] and from the Swedish Energy Agency [Grant number: 35624-2] at the Biogas Research Solutions Center hosted by Linköping University, Sweden, respectively.FundingOpen access funding provided by Linköping University. This research was funded by Formas [Grant number: 2021–02429] and from the Swedish Energy Agency [Grant number: 35624-2] at the Biogas Research Solutions Center hosted by Linköping University, Sweden.Author informationAuthors and AffiliationsPrograma de Pós-Graduação em Biotecnologia Vegetal e Bioprocessos, Universidade Federal do Rio de Janeiro, Rio de Janeiro, BrazilThuane Mendes AnacletoUnidade Multiusuário de Análises Ambientais, Instituto de Biologia, Universidade Federal do Rio de Janeiro, Rio de Janeiro, BrazilThuane Mendes Anacleto, Helena Rodrigues Oliveira & Alex Enrich-PrastBiogas Solutions Research Center, Linköping University, Linköping, SwedenThuane Mendes Anacleto, Helena Rodrigues Oliveira, Giacomo Carraro, Luka Šafarič, Sepehr Yekta Shakeri, Annika Björn & Alex Enrich-PrastCentro Federal de Educação Tecnológica Celso Suckow da Fonseca (CEFET/RJ), Rio de Janeiro, BrazilHelena Rodrigues OliveiraDepartment of Thematic Studies – Environmental Change, Linköping University, Linköping, SwedenGiacomo Carraro, Luka Šafarič, Sepehr Yekta Shakeri, Annika Björn & Alex Enrich-PrastDepartment of Ecology and Environmental Protection Technologies, Sumy State University, Sumy, UkrainePolina SkvortsovaInstituto de Agronomia, Universidade Federal Rural do Rio de Janeiro, Rio de Janeiro, BrazilÉrika Flávia Machado PinheirosInstitute of Marine Science, Federal University of São Paulo (IMar/UNIFESP), Santos, BrazilAlex Enrich-PrastAuthorsThuane Mendes AnacletoView author publicationsSearch author on:PubMed Google ScholarHelena Rodrigues OliveiraView author publicationsSearch author on:PubMed Google ScholarGiacomo CarraroView author publicationsSearch author on:PubMed Google ScholarPolina SkvortsovaView author publicationsSearch author on:PubMed Google ScholarLuka ŠafaričView author publicationsSearch author on:PubMed Google ScholarSepehr Yekta ShakeriView author publicationsSearch author on:PubMed Google ScholarAnnika BjörnView author publicationsSearch author on:PubMed Google ScholarÉrika Flávia Machado PinheirosView author publicationsSearch author on:PubMed Google ScholarAlex Enrich-PrastView author publicationsSearch author on:PubMed Google ScholarContributionsT.M.A.: Conceptualization, methodology, experimental analysis, formal analysis, writing – original draft. H.R.O.: Writing – review & editing, experimental analysis, formal analysis. G.C.: Writing – review & editing experimental analysis. P.S.: Experimental analysis. L.Š.: Writing – review & editing. S.Y.S.: Writing – review & editing. A.B.: Writing – review & editing. E.F.M.P.: Writing – review & editing. A.E.-P.: Supervision, writing – review & editing. All authors discussed the results and approved the final manuscript.Corresponding authorCorrespondence to
    Alex Enrich-Prast.Ethics declarations

    Competing interests
    The authors declare no competing interests.

    Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Supplementary Material 2Rights and permissions
    Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
    Reprints and permissionsAbout this articleCite this articleAnacleto, T.M., Oliveira, H.R., Carraro, G. et al. Performance-based criteria for safe and circular digestate use in agriculture.
    Sci Rep (2025). https://doi.org/10.1038/s41598-025-33314-xDownload citationReceived: 14 October 2025Accepted: 17 December 2025Published: 24 December 2025DOI: https://doi.org/10.1038/s41598-025-33314-xShare this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative
    KeywordsBiofertilizerPhytotoxicityCircular economyNutrient recoverySolid-liquid separation More

  • in

    Effect of leguminous green manures on alleviating continuous cropping obstacles in muskmelon cultivation

    AbstractContinuous cropping obstacles in greenhouse muskmelon cultivation pose a significant threat to sustainable production. While leguminous green manures are known to mitigate soil degradation in other crops, their efficacy and micro-ecological mechanisms in muskmelon systems remain unexplored. Here, we demonstrate for the first time that winter planting of two leguminous green manures, common vetch (Vicia sativa L.) and smooth vetch (Vicia villosa Roth var. glabresens Koch), during fallow periods alleviates continuous cropping obstacles by reshaping soil micro-ecology. Field trials revealed that both green manures significantly increased muskmelon yield (13.71% and 10.68%, respectively), elevated soil pH and organic matter, and reduced salinity (EC by 51.1% and total salt by 35.7%). High-throughput sequencing uncovered enriched microbial diversity, with beneficial taxa (Pirellula, Gemmata, Myceliophthora, Talaromyces) positively correlated with yield, while suppressing pathogenic fungi (Fusarium). Redundancy analysis highlighted soil pH and organic matter as key drivers of beneficial microbial recruitment, whereas salinity promoted harmful taxa. This study establishes a green manure-driven micro-ecological remediation framework, providing a cost-effective strategy for sustainable muskmelon cultivation in southern China.

    Similar content being viewed by others

    Spinach (Spinacia oleracea) as green manure modifies the soil nutrients and microbiota structure for enhanced pepper productivity

    Article
    Open access
    13 March 2023

    Colored plastic mulch impacts on soil properties, weed density and vegetable crop productivity: A meta-analysis

    Article
    Open access
    29 August 2025

    Green manure-induced shifts in nematode communities associated with soil bacterial and fungal biomes

    Article
    Open access
    13 December 2025

    IntroductionCucumis melon, a global economic fruit crop, is primarily classified into two types ingcluding muskmelon (thick-skin) and pellicle melon (thin-skin). China holds a global leadership position in terms of both cultivation area and production output for cucumis melon. These melons are highly favored across national and international markets for their distinctive sweetness and juicy texture. In southern China, successful cultivation of the muskmelon necessitates greenhouse farming to protect the crops from rain, ensuring abundant yields. Such greenhouses typically conduct biannual muskmelon cultivation. However, this practice of long-term continuous cropping, combined with intensive fertilizer use and absence of natural leaching due to rainfall, leads to several soil-related challenges, including acidification, salinization, nutrient and microbial community imbalance, accumulation of pathogens, collectively known as continuous cropping obstacles that resulting in a decrease in crop yield1. Some scientists have attempted to use biocontrol agents to solve the problem of continuous cropping obstacles in melons, and found that Bacillus subtilis C3 alleviated the continuous cropping obstacles of melon by eliminating phenolic acids and inhibiting the growth of Fusarium (pathogen) and root-knot nematodes, as well as improving the composition and structure of the rhizosphere microbial community2. Trichoderma. viride T23 relieved the continuous cropping limitation in muskmelon by improving soil physicochemical properties, elevating the biomass and diversity of soil microbial communities, and stimulating the production of soil active substances1. However, the use of biocontrol agents is limited by their singular strain, short storage life, and susceptibility to climatic conditions in the field, leading to inconsistent effectiveness and hindering large-scale application. In addition, grafting is another common method to alleviate the continuous cropping obstacles in cucurbit crops, offering the advantages of high efficiency and environmental friendliness. It not only enhances the disease resistance and stress tolerance of plants but also improves nutrient uptake3. Nevertheless, in southern China, there are no rootstock varieties suitable for muskmelon, making it impossible to employ grafting as a solution to the continuous cropping obstacles in muskmelon cultivation.Some studies have shown that planting green manure, particularly leguminous green manure, can enhance soil fertility, improve soil physicochemical properties, and effectively alleviate the continuous cropping obstacles of some crops such as wheat, potatoes, and cotton, thereby increasing crop yields4,5,6,7. However, there are relatively few green manure varieties suitable for cultivation in southern China. Whether planting leguminous green manure can alleviate the continuous cropping obstacles in muskmelon cultivation, as well as the mechanisms involved, remains unclear.Common vetch (Vicia sativa L.) and smooth vetch (Vicia villosa Roth var. glabresens Koch) are two leguminous green manure varieties that are relatively well-suited for cultivation in southern China. Studies have shown that using common vetch as green manure can improve soil fertility, moisture retention, and microbial abundance, thereby increasing potato yields8. It has also been demonstrated to effectively enhance the yields of corn and wheat4,9. Planting common vetch in vineyards reduces soil erosion and improves the soil micro-ecological environment10,11. Additionally, cultivating common vetch as green manure in tobacco fields effectively suppresses weed growth and boosts tobacco yields12. In contrast, research on smooth vetch remains limited.Currently, there are no research reports on the application of common vetch and smooth vetch in muskmelon cultivation. To clarify the effects and micro-ecological mechanisms of these two leguminous green manures in alleviating continuous cropping obstacles in muskmelon, this study involves planting and incorporating common vetch and smooth vetch during the winter fallow period of muskmelon cultivation. The research aims to analyze the impacts of these green manures on muskmelon yield, soil chemical properties (pH, salinity, nutrient content), and soil microbial diversity. By doing so, the study seeks to elucidate the efficacy and micro-ecological mechanisms of these green manures in mitigating continuous cropping obstacles in muskmelon. The research results can provide a theoretical foundation for soil improvement and the sustainable, efficient development of muskmelon in southern China.Materials and methods Field site descriptionThe study was conducted from November 16, 2023, to June 12, 2024, at the Wumao Farm, Wuming District, Nanning City, Guangxi, located at 23° 38′ N, 108° 22′ E, and 111 m above sea level. This region experiences a subtropical monsoon climate with yearly average temperatures ranging from 20 °C to 27 °C. The experimental greenhouse, measuring 30 m in length, 6 m in span, with sidewall height of 1.8 m and a central height of 3.0 m, had been consistently used for cultivating muskmelon over nine years, accounting for 18 growing cycles since autumn 2014. The cultivation soil was enriched with various decomposed organic materials, including tree bark, sugarcane bagasse, cassava residue, and poultry manure. Soil analysis conducted before the experiment revealed a pH of 5.96, an electrical conductivity (EC) of 1.75 ms/cm (indicating water-soluble salt concentration), organic matter at 219.2 g/kg, total salts at 3.66 g/kg, total nitrogen at 11.1 g/kg, total phosphorus at 13.0 g/kg, total potassium at 7.89 g/kg, available nitrogen at 1.16 g/kg, available phosphorus at 0.14 g/kg, and available potassium at 2.2 g/kg.Experimental designThe experiment employed a randomized block design, and the data statistical model utilized analysis of variance (ANOVA). The design included two treatments: common vetch (T1), and smooth vetch (T2), and a control group (CK) with no green manure. The common vetch and smooth vetch used in this experiment are cultivated varieties. The Agricultural Resources and Environmental Research Institute, Guangxi Academy of Agricultural Sciences provided the green manure seeds and approved the experimental protocols. Each treatment was replicated across three blocks, each being a single-span greenhouse measuring 30 m by 6 m, totaling an area of 180 m² per block. Post the autumn muskmelon harvest in 2023, green manures were sown on October 20 at a rate of 7.5 g/m². By December 28, the green manures, having reached about 70 cm, were harvested and chopped into 15–20 cm segments. These were evenly distributed over the soil surface, then plowed under and irrigated to maintain approximately 80% soil moisture. The soil was then covered with a white polyethylene film with a thickness of 0.02 mm for 68 days to aid in the fermentation and decomposition of the green manures. The muskmelon variety ‘Huang Meng Cui’ was sown on February 10, 2024, and transplanted to the experiment greenhouses on March 8, 2024. Each block consisted of 300 plants. Identical irrigation and fertilization regimes were maintained across all treatments, culminating in the harvest of the muskmelon on June 12, 2024.Measurement of muskmelon yield and soluble solids contentOn harvest day, the total yield from each treatment plot was weighed to determine the yield per hectare. Additionally, the soluble solids content at the center of the muskmelon was tested using a digital refractometer. For this analysis, 15 muskmelons were randomly selected from each plot to measure and calculate the average soluble solids content.Soil samplingPost-harvest, soil from each treatment area was thoroughly mixed and leveled. Samples were collected from the top 20 cm of soil using a five-point sampling method. Each plot’s sample weighted 1 kg, and was divided into two portions: one (0.95 kg) sealed in plastic bags for chemical property analysis (including pH, salinity, and nutrient content) and the other stored in 50 mL centrifuge tubes, instantly frozen in liquid nitrogen, then preserved at −80 °C for microbial diversity analysis (focusing on fungi and bacteria).Assessment of soil chemical propertiesThe chemical properties of the soil used for muskmelon cultivation were evaluated using methods adapted from Ding13 and Shen14. The soil pH was determined potentiometrically. EC, reflecting the concentration of water-soluble salts in the soil, was measured using conductivity testing, and the total salt content was quantified using the gravimetric method. To assess soil fertility, organic matter content was determined by the potassium dichromate volumetric method in an oil bath. Nutrient levels were also thoroughly examined: total nitrogen content and available nitrogen content were analyzed using the Kjeldahl method and alkaline diffusion method, respectively, while total phosphorus and available phosphorus were measured via sodium hydroxide fusion and hydrochloric acid-sulfuric acid extraction methods. Lastly, total potassium and available potassium were quantified through acid dissolution and flame photometry, respectively.Extraction of total microbial DNA from soil and high-throughput sequencingTotal microbial DNA was extracted from the cultivated soil samples using the HiPure Soil DNA Extraction Kit (Magen, Guangzhou, China). DNA concentration and purity were assessed with a NanoDrop 2000 UV-Vis spectrophotometer (Thermo Scientific, Wilmington, USA), and DNA integrity was verified through 1% agarose gel electrophoresis. For the amplification of bacterial DNA, primers 341 F (5’-CCTACGGGNGGCWGCAG-3’) and 806R (5’-GGACTACHVGGGTWTCTAAT-3’) were used to target the V3-V4 hypervariable regions15, and the fungal internal transcribed spacer (ITS) ITS2 region was PCR-amplified using primers ITS3F (GATGAAGAACGYAGYRAA) and ITS4R (TCCTCCGCTTATTGATATGC)16. The PCR protocol included an initial denaturation at 95 °C for 5 min, followed by 30 or 35 cycles (for bacteria and fungi, respectively) of denaturation at 95 °C for 1 min, annealing at 60 °C for 1 min, and extension at 72 °C for 1 min, ending with a final extension at 72 °C for 7 min. The PCR products were purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) and quantified with the ABI StepOnePlus Real-Time PCR System (Life Technologies, Foster City, USA). Sequencing libraries were prepared and subjected to paired-end sequencing on the Illumina HiSeq 2500 PE 250 platform.Data analysisAnalysis of variance (ANOVA)Differences between data sets were statistically analyzed using SPSS v.20. A one-way ANOVA was performed at a significance level of P = 0.05 to assess the variability.High-throughput sequencing data analysis for microbial 16 S rRNA and ITS gene ampliconsThe raw sequencing data obtained from the Illumina platform were first processed using FASTP (version 0.18.0) to filter out low-quality reads. The resulting high-quality reads, known as ‘clean reads,’ were then merged into longer sequences or ‘tags’ using the FLASH software (version 1.2.11). These tags were subjected to further quality control by removing low-quality segments, thus obtaining ‘clean tags.’ The clean tags were clustered into Operational Taxonomic Units (OTUs) at a similarity threshold of 97% using the UPARSE algorithm (version 9.2.64). To ensure data accuracy, the UCHIME algorithm was employed to remove any chimeric sequences from these tags, resulting in what are termed ‘effective tags.’ For species identification, the representative sequences of bacterial OTUs were compared with the SILVA database (version 132), and fungal OTUs were compared with the UNITE database (version 8.0). The classification of these sequences was then refined using the RDP classifier (version 2.2) with the Naive Bayesian model, with a confidence threshold set between 0.8 and 1, to provide detailed annotations of the microbial species present in each sample.In this study, a comprehensive suite of advanced statistical and bio-informatics tools was employed to evaluate soil microbial diversity and community composition, thereby providing an in-depth understanding crucial for sustainable agricultural practices. Microbial diversity indices, including Sobs, Chao1, Simpson, and Shannon, were calculated using QIIME software (version 1.9.1)17. These indices quantitatively assess the microbial richness and diversity within the soil samples. Principal Component Analysis (PCA) was conducted using the ‘vegan’ package (version 2.5.3; Oksanen et al. 2010) in R language to analyze and visually represent the compositional differences in microbial communities based on Operational Taxonomic Units (OTUs). The ‘VennDiagram’ package in R language18 was used to construct Venn diagrams, elucidating common or unique OTUs across treatments. For the analysis of microbial species composition, the ‘ggplot2’ package (version 2.2.1)19 in R was utilized to create abundance stack plots. Biomarker analysis involved using the ‘vegan’ package in R for Tukey’s HSD test and Kruskal-Wallis H test to evaluate species abundance differences between treatments, followed by Linear discriminant analysis Effect Size (LEfSe) software (version 1.0) for analyzing biomarkers (LDA > 2; P < 0.05) and generating phylogenetic trees13. Finally, Redundancy Analysis (RDA) was performed using the ‘vegan’ package (version 2.5.3)20in R to determine the impact of environmental factors on the microbial community composition.ResultsImpact of different treatments on yield and soluble solids content in muskmelonThe impact of planting and incorporating different green manures on the yield and soluble solids content in the center of muskmelon is summarized in Table 1. The treatments involving the incorporation of common vetch (T1) and smooth vetch (T2) significantly increased the yield of muskmelon compared with the control group (CK). Specifically, the yields were 43,785 kg/ha for T1 and 43,110 kg/ha for T2, representing increases of 13.71% and 10.68% over the control, respectively. Additionally, the soluble solids content in the center of muskmelon was slightly higher in T1 and T2 treatments compared with that in CK treatment, although these differences were not statistically significant.Table 1 Impact of different green manure treatments on yield and soluble solids content in muskmelon.Full size tableImpact of different treatments on soil chemical propertiesAs indicated in Table 2, the treatments with common vetch (T1) and smooth vetch (T2) significantly enhanced the pH and organic matter content of the cultivated soil. Specifically, T1 and T2 treatments increased soil pH by 7.9% and 6.4%, respectively, and raised organic matter content by 11.3% and 10.6%, respectively. Furthermore, there was a notable reduction in the soil’s EC values and total salt content; EC values decreased by 51.1% in T1 and 48.9% in T2, while total salt content decreased by 34.1% and 35.7%, respectively. The T1 treatment also resulted in increases in total nitrogen, available nitrogen, total phosphorus, and available phosphorus levels in the soil, although these increases did not reach a level of statistical significance. However, there was a significant reduction in the content of total potassium and available potassium, with the available potassium content decreasing by 35.0%. T2 treatment showed similar trends, increasing total nitrogen and phosphorus while decreasing total and available potassium, although these changes were not significantly different from the control group. Between the two green manure treatments, no significant differences were observed in parameters other than available potassium.Table 2 Impact of different green manures on soil chemical properties.Full size tableSoil microbial genome sequence informationHigh-throughput sequencing of the 16 S rRNA and ITS gene sequences from nine soil samples across the three treatments, following chimera filtering and quality control, yielded 734,250 high-quality 16 S rRNA gene sequences and 711,827 ITS gene sequences (effective tags), which were used for further community analysis. Clustering at 97% similarity resulted in the identification of 29,207 bacterial and 2,844 fungal OTUs. The average Good’s coverage for bacteria was 98.1% and for fungi 99.9%, indicating adequate sequencing depth that likely covered the majority of species present in the samples. Importantly, both the bacterial and fungal rarefaction curves gradually flattened, suggesting that most microbial species in the samples were captured and the sequencing depth was sufficient.Analysis of microbial diversity indices under different treatmentsAlpha (α) diversity analysis provides insights into the abundance and diversity of microbial communities. The Sobs and Chao1 indices, indicative of community abundance, suggested greater species abundance in T1 and T2 treatments compared to those in CK treatment. Simpson and Shannon indices, reflecting community diversity, indicated increased diversity in these treatments. Table 3 shows that both T1 and T2 treatments significantly enhanced the abundance of bacteria and fungi in the cultivated soil. There was a notable increase in bacterial community diversity and a slight but non-significant increase in fungal diversity under these treatments. The abundance and diversity of both bacteria and fungi in T1 treatment were slightly higher than those in T2 treatment, although these differences were not statistically significant.Table 3 Analysis of microbial diversity indices under different treatments.Full size tablePCA of microbial communities under different treatmentsPCA was employed to elucidate the compositional similarities and disparities among the microbial communities under different treatments based on the abundance of OTUs. This analysis provides an insightful depiction of the spatial relationships between the samples in a multidimensional space, where proximity indicates compositional similarity. The PCA revealed distinct variations in the bacterial and fungal community structures across the treatments. Specifically, both T1 and T2 treatments exhibited bacterial community compositions that were significantly divergent from CK treatment. However, the difference between T1 and T2 bacterial communities was relatively minor (Fig. 1a). In terms of fungal community compositions, a pronounced difference was observed between T1 and T2 treatments, while the differences between T1 and CK, and T2 and CK, were comparatively marginal (Fig. 1b).Fig. 1PCA of microbial communities under different treatments (a: Bacteria; b: Fungi).Full size imageAnalysis of shared and unique OTUs across different treatmentsThe distribution of shared and unique OTUs among the treatments was analyzed using Venn diagrams, offering a visual representation of the overlap and exclusivity of microbial communities. The bacterial community analysis (Fig. 2a) indicated a larger number of unique OTUs in T1 treatment, amounting to 868 OTUs, with 2237 OTUs shared with the control group (CK). Conversely, T2 treatment manifested a lower count of unique bacterial OTUs (499), but a larger number of shared OTUs with CK (2338). This pattern suggests that T1 treatment exerts a more substantial influence on the bacterial community than T2 treatment. Fungal community analysis (Fig. 2b) revealed that the T1 treatment harbored a larger number of unique fungal OTUs (207), with 113 OTUs shared with CK. In contrast, T2 treatment had fewer unique fungal OTUs (109), though it shared more OTUs with CK (136), indicating a more pronounced impact of T1 treatment on the fungal community as well.Fig. 2Number of OTUs in different treatments (a: Bacteria; b: Fungi).Full size imageComparative analysis of microbial community composition across treatmentsThe relative abundance of bacteria across treatments was evaluated at various taxonomic levels—phylum, class, and genus (Fig. 3a–c). At the phylum level, the top ten bacterial groups, in descending order of abundance, were Chloroflexi, Proteobacteria, Planctomycetes, Gemmatimonadetes, Acidobacteria, Actinobacteria, Patescibacteria, Rokubacteria, Bacteroidetes, and Verrucomicrobia, constituting 96.65–97.65% of the total bacterial sequences in each treatment. Notably, T1 and T2 treatments enhanced the relative abundance of Planctomycetes, Acidobacteria, Patescibacteria, and Rokubacteria, and reduced that of Chloroflexi, Proteobacteria, Gemmatimonadetes, and Actinobacteria (Fig. 3a). At the class level, the top ten bacterial taxa, arranged from highest to lowest abundance, were Gemmatimonadetes, Planctomycetacia, Alphaproteobacteria, Subgroup_6, KD4-96, Phycisphaerae, Gammaproteobacteria, Gitt-GS-136, Acidimicrobiia, and Chloroflexia. Compared with CK treatment, T1 and T2 treatments significantly favored the abundance of Planctomycetacia, Subgroup_6, and Phycisphaerae, with T1 additionally promoting Gammaproteobacteria and T2 fostering Acidimicrobiia. Notably, CK showed higher relative abundance of Gemmatimonadetes and Alphaproteobacteria (Fig. 3b). At the genus level, Gaiella, Sphingomonas, Pirellula, Gemmata, and RB41 ranked top five, with each having a relative abundance of over 1%. T1 and T2 treatments notably increased the abundance of Pirellula, Gemmata, and SH-PL14, with T1 also enhancing RB41. CK exhibited higher relative abundance of Gaiella and Sphingomonas (Fig. 3c).The relative abundance of fungus in each treatment was assessed at the phylum, class, and genus levels (Fig. 3d–f). Fungal communities at the phylum level were dominated by Ascomycota, Mortierellomycota, Basidiomycota, and Chytridiomycota. Ascomycota was the most abundant, constituting 72.84–81.38% of the total fungal sequences. As indicated in Fig. 3d, T1 and T2 treatments significantly increased the abundance of Mortierellomycota and Basidiomycota. Furthermore, T2 treatment also significantly enhanced the relative abundance of Chytridiomycota. At the class level, Sordariomycetes, Eurotiomycetes, Pezizomycetes, Dothideomycetes, Agaricomycetes, and Saccharomycetes showed relative abundance of over 1%. Sordariomycetes had the highest relative abundance, accounting for 46.74–48.80% of the total sequences across treatments. As depicted in Fig. 3e, compared with CK treatment, T1 treatment enhanced the relative abundance of Dothideomycetes Agaricomycetes, and Saccharomycetes, and T2 also favored Agaricomycetes. The relative abundance of Sordariomycetes, Eurotiomycetes, and Pezizomycetes decreased in T1 and T2 treatments. The genus-level analysis revealed Scopulariopsis, Penicillium, Aspergillus, Chaetomium, Myceliophthora, Fusarium, Talaromyces, Mycothermus, and Scedosporium as dominant, each exceeding 1% relative abundance. Compared with CK treatment, T1 treatment enhanced Scopulariopsis, Penicillium, Chaetomium, Myceliophthora, Talaromyces, and Scedosporium, while T2 favored Myceliophthora, Talaromyces, and Scedosporium. Notably, the relative abundance of pathogenic fungi like Aspergillus and Fusarium displayed a significant decrease in both the T1 and T2 treatments (Fig. 3f).Fig. 3Relative abundance of microorganisms in different treatments (a: Bacterial abundance at the phylum level; b: Bacterial abundance at the class level; c: Bacterial abundance at the genus level; d: Fungal abundance at the phylum level; e: Fungal abundance at the class level; f: Fungal abundance at the genus level).Full size imageComparative assessment of microbial biomarkersLEfSe analysis identified unique biomarker species that were significantly different among treatments. For bacteria, five species, including 67_14, AKYG587, and Sandaracinaceae were predominantly associated with T1 treatment, whereas Defluviicoccus and Rhodopirillaceae were more abundant in CK treatment (Fig. 4a).In the case of fungi, Mycosphaerella, Mycosphaerella etlingerae, Trichoderma, Exobasidiaceae, Exobasidiales and Exobasidium were significantly more abundant in T1 treatment (Fig. 4b).Fig. 4Cladogram plotted from LEfSe comparison analysis (a: Bacteria; b: Fungi).Full size imageImpact of environmental factors on soil microbial community compositionRDA was employed to elucidate the influence of environmental variables on the composition of microbial communities within the soil. This investigation was centered on the top ten most abundant bacterial and fungal genera, in relation to their correlation with specific environmental factors. The analysis depicted in Fig. 5a revealed significant relationships between these bacteria and various soil parameters. The RDA plot for bacterial relative abundance (Fig. 5a) demonstrated that the first and second axes accounted for 53.79% and 28.87% of the total variance, respectively. Soil pH, organic matter content, and muskmelon yield were aligned in the same direction, showing significant positive correlations with Pirellula, Gemmata, and SH-PL14, that primarily found in T1 and T2 treatments. This correlation indicates that an increase in soil pH and organic matter content is associated with a higher relative abundance of these bacterial genera, which, in turn, positively influences muskmelon yield. Conversely, EC, total salt, and available potassium content were positively correlated with Marmoricola, Aeromicrobium, and Gaiella in CK treatment, inversely impacting the yield. In addition, the available nitrogen and phosphorus content showed a significant positive correlation with the genera MND1 and Sphingomonas.The fungal community composition, as depicted in the RDA plot (Fig. 5b), showed that the first and second axes explained 38.62% and 26.33% of the variation, respectively. Soil pH, organic matter content, and yield were aligned in the same direction, displayed a positive correlation with Myceliophthora and Talaromyces in T1 and T2 treatments. This finding indicates that higher levels of soil pH and organic matter are conducive to the proliferation of these fungal genera, which are beneficial for enhancing muskmelon yield. In contrast, EC, total salt, and available potassium content showed a significant positive correlation with Fusarium, Scedosporium, Mycothermus, and Chaetomium in CK treatment. These fungi, negatively correlated with yield, suggest that increased soil salinity and potassium levels may favor fungal species detrimental to muskmelon yield. Additionally, a significant positive correlation was observed between EC and Aspergillus in CK, negatively impacting muskmelon yield, thereby implicating Aspergillus as a potential yield-reducing factor. Moreover, available nitrogen and phosphorus were found to be significantly positively correlated with Penicillium and Cladosporium.Fig. 5RDA analysis of microorganisms and environmental factors (a: Bacteria; b: Fungi; Gai: Gaiella; Sph: Sphingomonas; Pir: Pirellula; Gem: Gemmata; SH: SH-PL14; Aer: Aeromicrobium; Mar: Marmoricola; Gemm: Gemmatimonas; Sco: Scopulariopsis; Pen: Penicillium; Asp: Aspergillus; Cha: Chaetomium; Myce: Myceliophthora; Fus: Fusarium; Tal: Talaromyces; Myco: Mycothermus; Sce: Scedosporium; Cla: Cladosporium; OM: Organic matter; TS: Total salt; AN: Available N; AP: Available P; AK: Available K).Full size imageDiscussionThe increasing severity of continuous cropping obstacles in muskmelon cultivation necessitates sustainable solutions. While leguminous green manures have proven effective in alleviating such obstacles for crops like wheat and potato, their efficacy in muskmelon systems, particularly in southern China, remained unexplored. Our study provides the first mechanistic evidence that leguminous green manures alleviate continuous cropping obstacles in muskmelon by synergistically improving soil chemical properties and restructuring microbial communities. The observed yield increases align with earlier findings in wheat and potato systems4,5, but extend these benefits to a greenhouse muskmelon context where salinity and pathogen pressure are exacerbated by intensive cultivation. Critically, both vetches elevated soil pH and organic matter—key drivers of microbial recruitment—while suppressing salinity (EC, total salt), contrasting with biocontrol approaches like Trichoderma viride T23, which primarily targets microbial restructuring without addressing soil chemistry1. This dual action underscores the advantage of green manures in tackling multi-faceted degradation.Both green manures improved soil pH and organic matter while reducing salinity (EC, total salts), aligning with findings in wheat and potato systems4,5. Notably, the pH elevation counteracts soil acidification common in southern China, while increased organic matter enhances nutrient retention. Though nitrogen and phosphorus levels showed non-significant gains, repeated green manure applications may amplify these effects, as observed in multi-year studies5,21. The significant potassium reduction, particularly under common vetch, warrants further investigation but likely reflects crop-specific nutrient competition.Both green manures can significantly increase the abundance of bacterial and fungal species and the diversity of bacterial communities in the soil. The results of microbial community composition analysis indicate that both green manures can recruit bacteria such as Pirellula, Gemata, SH-PL14, and Subgroup-6. In addition, common vetch can also recruit RB41. Among these bacteria, Pirellula, Gemmata, and SH-PL14 belong to the Planctomycetes phylum, which is a common and important bacterium for both the environment and biotechnology. They are key participants in the global carbon and nitrogen cycle22. Subgroup-6 belongs to the Bathyrachaeota phylum and plays important roles in phototroph, autotrophy, as well as nitrogen and sulfur cycling23. RB41 belongs to Acidobacteria phylum that can enhance carbon and nitrogen cycling and phosphorus absorption in the rhizosphere environment. It is significantly correlated with soil pH and promoted crop growth24. Furthermore, both green manures can recruit some fungi such as Myceliophthora, Talaromyces, and Scedosporium. Specifically, common vetch additionally recruited fungal genera including Scopulariopsis, Chaetomium, and Penicillium. Among these fungi, four genera—Myceliophthora, Talaromyces, Chaetomium, and Penicillium—are recognized as beneficial fungi. Myceliophthora, a thermophilic fungus, has been extensively studied and applied due to its ability to produce thermostable enzymes such as xylanases, heat-stable proteases, amylases, chitinases, and cellulolytic enzymes25,26. Talaromyces, another thermophilic fungal genus, synthesizes nematode-antagonistic macrocyclic lactones27 and produces bioactive compounds with efficacy against human pathogens including Plasmodium (malaria parasite), tumor cells, and Staphylococcus aureus28,29. Chaetomium enhances plant disease resistance through antimicrobial metabolite production and has been widely utilized in phytopathology research for plant protection30,31. Penicillium plays critical ecological roles in environmental processes such as nutrient cycling and pollutant degradation32. Thus it can be seen that both green manures can recruit beneficial bacteria and fungi that may be increase the yield of muskmelon.On the contrary, both green manure treatments substantially diminished the relative abundance of fungi Fusarium. This fungi is a genus implicated in significant soil-borne diseases affecting muskmelon, including wilt and root rot33,34. This reduction suggests a potential mitigation of soil-borne disease risks in muskmelon cultivation.Our RDA at the genus level assessed the impact of environmental factors on the bacterial and fungal communities ranking in the top ten for abundance. The interplay between soil parameters and microbial taxa highlights a dual mechanism: green manures improve soil chemistry, fostering beneficial microbes while suppressing pathogens. This aligns with Ding13, who emphasized green manure-induced microbial restructuring. However, our study uniquely identifies Pirellula, Gemmata, Myceliophthora and Talaromyces as key yield-linked taxa in muskmelon systems, advancing the understanding of crop-specific microbial drivers.ConclusionBoth leguminous green manures can alleviate continuous cropping obstacles and increase muskmelon yield. The micro-ecological mechanism is to enhance soil pH, organic matter content, and reduce salinity indicators (EC value, total salt content), thereby indirectly increasing the microbial diversity and abundance of beneficial microorganisms, and diminishing the relative abundance of harmful microorganisms.

    Data availability

    The datasets generated and/or analysed during the current study are available in the Genome Sequence Archive (GSA: CRA030214) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa.
    ReferencesZhang, Z. et al. The effects of Trichoderma viride T23 on rhizosphere soil microbial communities and the metabolomics of muskmelon under continuous cropping. Agronomy 13 (4), e1092. https://doi.org/10.3390/agronomy13041092 (2023).Article 
    CAS 

    Google Scholar 
    Ku, Y. et al. Biological control of melon continuous cropping obstacles: weakening the negative effects of the vicious cycle in continuous cropping soil. Microbiol. Spectr. 10, e01776–e01722. https://doi.org/10.1128/spectrum.01776-22 (2022).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Colla, G., Rouphael, Y., Cardarelli, M., Salerno, A. & Rea, E. The effectiveness of grafting to improve alkalinity tolerance in watermelon. Environ. Exp. Bot. 68 (3), 283–291. https://doi.org/10.1016/j.envexpbot.2009.12.005 (2010).Article 
    CAS 

    Google Scholar 
    Badaruddin, M. & Meyer, D. W. Green-manure legume effects on soil nitrogen, grain yield, and nitrogen nutrition of wheat. Crop Sci. 30 (4), 819–825. https://doi.org/10.2135/cropsci1990.0011183X003000040011x (1990).Article 
    CAS 

    Google Scholar 
    Sincik, M., Turan, Z. M. & Göksoy, A. T. Responses of potato (Solanum tuberosum L.) to green manure cover crops and nitrogen fertilization rates. Am. J. Potato Res. 85, 390–391. https://doi.org/10.1007/s12230-008-9011-9 (2008).Article 

    Google Scholar 
    Rochester, I. & Peoples, M. Growing Vetches (Vicia villosa Roth) in irrigated cotton systems: inputs of fixed N, N fertiliser savings and cotton productivity. Plant. Soil. 271, 251–264. https://doi.org/10.1007/s11104-004-2621-1 (2005).Article 
    CAS 

    Google Scholar 
    Lahti, T. & Kuikman, P. The effect of delaying autumn incorporation of green manure crop on N mineralization and spring wheat (Triticum aestivum L.) performance. Nutr. Cycl. Agroecosys. 65, 265–280. https://doi.org/10.1023/A:1022617104296 (2003).Article 
    CAS 

    Google Scholar 
    Che, T. et al. Common Vetch intercropping with reduced irrigation ensures potato production by optimizing microbial interactions. Field Crop Res. 307, e109267. https://doi.org/10.1016/j.fcr.2024.109267 (2024).Article 

    Google Scholar 
    Pál, V. & Zsombik, L. Effect of common Vetch (Vicia sativa L.) green manure on the yield of corn in crop rotation system. Agronomy 14 (1), e19. https://doi.org/10.3390/agronomy14010019 (2024).Article 
    CAS 

    Google Scholar 
    Rodrigo-Comino, J., Terol, E., Mora, G., Giménez-Morera, A. & Cerdà, A. Vicia sativa Roth. Can reduce soil and water losses in recently planted vineyards (Vitis vinifera L). Earth Syst. Environ. 4, 827–842. https://doi.org/10.1007/s41748-020-00191-5 (2020).Article 
    ADS 

    Google Scholar 
    Longa, C. M. O. et al. Soil microbiota respond to green manure in organic vineyards. J. Appl. Microbiol. 123 (6), 1547–1560. https://doi.org/10.1111/jam.13606 (2017).Article 
    CAS 
    PubMed 

    Google Scholar 
    Travlos, I. S. et al. Green manure and pendimethalin impact on Oriental sun-cured tobacco. Agron. J. 106 (4), 1225–1230. https://doi.org/10.2134/agronj13.0557 (2014).Article 
    CAS 

    Google Scholar 
    Ding, T., Yan, Z., Zhang, W. & Duan, T. Green manure crops affected soil chemical properties and fungal diversity and community of Apple orchard in the loess plateau of China. J. Plant. Nutr. Soil. Sc. 21, 1089–1102. https://doi.org/10.1007/s42729-021-00424-0 (2021).Article 
    CAS 

    Google Scholar 
    Shen, Z. et al. Induced soil microbial suppression of banana fusarium wilt disease using compost and biofertilizers to improve yield and quality. Eur. J. Soil. Biol. 57, 1–8. https://doi.org/10.1016/j.ejsobi.2013.03.006 (2013).Article 

    Google Scholar 
    Ali, A., Ghani, M. I., Li, Y., Ding, H. & Meng, H. Microbial community, diversity structure, and predictive functional profiling in continuous cucumber planted soil affected by diverse cropping systems in an intensive greenhouse region of Northern China. Int. J. Mol. Sci. 20, 2619–2640. https://doi.org/10.3390/ijms20112619 (2019).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Toju, H., Tanabe, A. S., Yamamoto, S. & Sato, H. High-coverage ITS primers for the DNA-based identification of ascomycetes and basidiomycetes in environmental samples. PloS One. 7 (7), e40863. https://doi.org/10.1371/journal.pone.0040863 (2012).Article 
    CAS 
    PubMed 
    PubMed Central 
    ADS 

    Google Scholar 
    Caporaso, J. G. et al. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods. 7, 335–336. https://doi.org/10.1038/nmeth.f.303 (2010).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Chen, H. & Boutros, P. C. VennDiagram: a package for the generation of highly-customizable Venn and Euler diagrams in R. BMC Bioinform. 12 (e35). https://doi.org/10.1186/1471-2105-12-35 (2011).Wickham, H. ggplot2. WIREs Comput. Stat. 3(2), 180–185. (2011). https://doi.org/10.1002/wics.147Oksanen, J. et al. Vegan: community ecology package. Software http://CRAN.R-project.org/package=vegan (2010).Zhang, D. et al. Improving soil aggregation, aggregate-associated C and N, and enzyme activities by green manure crops in the loess plateau of China. Eur. J. Soil. Sci. 70 (6), 1267–1279. https://doi.org/10.1111/ejss.12843 (2019).Article 
    CAS 

    Google Scholar 
    Wiegand, S., Jogler, M. & Jogler, C. On the maverick planctomycetes. FEMS Microbiol. Rev. 42 (6), 739–760. https://doi.org/10.1093/femsre/fuy029 (2018).Article 
    CAS 
    PubMed 

    Google Scholar 
    Pan, J. et al. Genomic and transcriptomic evidence of light-sensing, porphyrin biosynthesis, Calvin-Benson-Bassham cycle, and Urea production in bathyarchaeota. Microbiome 8, e43. https://doi.org/10.1186/s40168-020-00820-1 (2020).Article 
    CAS 

    Google Scholar 
    Zhang, H. et al. Microbial communities in the rhizosphere soil of Ambrosia Artemisiifolia facilitate its growth. Plant. Soil. 492, 353–365. https://doi.org/10.1007/s11104-023-06181-6 (2023).Article 
    CAS 

    Google Scholar 
    Van den Brink, J., Samson, R. A., Hagen, F., Boekhout, T. & de Vries R. P. Phylogeny of the industrial relevant, thermophilic genera Myceliophthora and Corynascus. Fungal Divers. 52, 197–207. https://doi.org/10.1007/s13225-011-0107-z (2012).Article 

    Google Scholar 
    Maheshwari, R., Bharadwaj, G. & Bhat, M. K. Thermophilic fungi: their physiology and enzymes. Microbiol. Mol. Biol. R. 64 (3), 461–488. https://doi.org/10.1128/mmbr.64.3.461-488.2000 (2000).Article 
    CAS 

    Google Scholar 
    Guo, J. P. et al. Thermolides, potent nematocidal PKS-NRPS hybrid metabolites from thermophilic fungus Talaromyces thermophilus. J. Am. Chem. Soc. 134 (50), 20306–20309. https://doi.org/10.1021/ja3104044 (2012).Article 
    CAS 
    PubMed 
    ADS 

    Google Scholar 
    Zang, Y. et al. Unexpected talaroenamine derivatives and an undescribed polyester from the fungus talaromyces stipitatus ATCC10500. Phytochemistry 119, 70–75. https://doi.org/10.1016/j.phytochem.2015.09.002 (2015).Article 
    CAS 
    PubMed 

    Google Scholar 
    Zang, Y. et al. Antimicrobial Oligophenalenone dimers from the soil fungus Talaromyces stipitatus. J. Nat. Prod. 79 (12), 2991–2996. https://doi.org/10.1021/acs.jnatprod.6b00458 (2016).Article 
    CAS 
    PubMed 

    Google Scholar 
    Shanthiyaa, V. et al. Use of Chaetomium globosum for biocontrol of potato late blight disease. Crop Prot. 52, 33–38. https://doi.org/10.1016/j.cropro.2013.05.006 (2013).Article 

    Google Scholar 
    McLean, K. L. & Stewart, A. Application strategies for control of onion white rot by fungal antagonists. New. Zeal J. Crop Hort. 28 (2), 115–122. https://doi.org/10.1080/01140671.2000.9514131 (2000).Article 

    Google Scholar 
    Park, M. S., Oh, S. Y., Fong, J., Houbraken, J. & Lim, Y. W. The diversity and ecological roles of Penicillium in intertidal zones. Sci. Rep. 9, e13540. https://doi.org/10.1038/s41598-019-49966-5 (2019).Article 
    CAS 
    ADS 

    Google Scholar 
    Imazaki, I. & Kadota, I. Control of fusarium wilt of melon by combined treatment with biocontrol, plant-activating, and soil-alkalizing agents. J. Gen. Plant. Pathol. 85, 128–141. https://doi.org/10.1007/s10327-018-00833-7 (2019).Article 
    CAS 

    Google Scholar 
    Abdul-Hasan, F. & Hussein, Z. H. Genetic diversity of Fusarium Solani f. sp. cucurbitae, the causal root and crown rot of cucurbits (melon) by using molecular markers and control. J. P S. 7 (15), 2151–2172. https://doi.org/10.4236/ajps.2016.715191 (2016).Article 
    CAS 

    Google Scholar 
    Download referencesAcknowledgementsWe would like to thank Dr. Tieguang He from the Agricultural Resources and Environmental Research Institute, Guangxi Academy of Agricultural Sciences for providing leguminous green manure seeds.FundingThis work was supported by the Guangxi Natural Science Foundation (2023GXNSFAA026437), the Guangxi Key Research and Development Project (Guikenong AB2506910004), the National Natural Science Foundation of China (32260677), the China Earmarked Fund for Modern Agroindustry Technology Research System (CARS-25), the Earmarked Fund for CARSGI(nycytxgxcxtd-2024-17-02), and the Special Project of Basic Scientific Research of Guangxi Academy of Agricultural Sciences (Guinongke 2021YT045 and Guinongke 2024ZX26).Author informationAuthor notesYunfeng Ye and Guifen Li: Co-first author.Authors and AffiliationsHorticultural Research Institute, Guangxi Academy of Agricultural Sciences, Nanning, 530007, People’s Republic of ChinaYunfeng Ye, Guifen Li, Huayun Xie, Sihua Qin, Jinyan Huang, Demei Zhang, Yi He, Tangjing Liu & Rixin HongGuangxi Key Laboratory of Biology for Crop Diseases and Insect Pests/Plant Protection Research Institute, Guangxi Academy of Agricultural Sciences, Nanning, 530007, People’s Republic of ChinaChanjuan Du & Gang FuAuthorsYunfeng YeView author publicationsSearch author on:PubMed Google ScholarGuifen LiView author publicationsSearch author on:PubMed Google ScholarHuayun XieView author publicationsSearch author on:PubMed Google ScholarSihua QinView author publicationsSearch author on:PubMed Google ScholarJinyan HuangView author publicationsSearch author on:PubMed Google ScholarChanjuan DuView author publicationsSearch author on:PubMed Google ScholarDemei ZhangView author publicationsSearch author on:PubMed Google ScholarYi HeView author publicationsSearch author on:PubMed Google ScholarTangjing LiuView author publicationsSearch author on:PubMed Google ScholarRixin HongView author publicationsSearch author on:PubMed Google ScholarGang FuView author publicationsSearch author on:PubMed Google ScholarContributionsY. Y. and G. L. wrote the main manuscript text, prepared figures and tables, and participated in field experiments. H. X., S.Q., J. H., C.D., D. Z., Y. H. and T. L. participated in field and indoor experiments. R. H. and G. F.guided experiments and reviewed the manuscript. All authors reviewed the manuscript.Corresponding authorsCorrespondence to
    Rixin Hong or Gang Fu.Ethics declarations

    Competing interests
    The authors declare no competing interests.

    Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Rights and permissions
    Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
    Reprints and permissionsAbout this articleCite this articleYe, Y., Li, G., Xie, H. et al. Effect of leguminous green manures on alleviating continuous cropping obstacles in muskmelon cultivation.
    Sci Rep 15, 44507 (2025). https://doi.org/10.1038/s41598-025-28033-2Download citationReceived: 12 September 2025Accepted: 07 November 2025Published: 24 December 2025Version of record: 24 December 2025DOI: https://doi.org/10.1038/s41598-025-28033-2Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative
    KeywordsMuskmelonContinuous cropping obstaclesMicrobial diversityCommon vetchSmooth vetch More

  • in

    Cesium accumulation in nodules is involved in mitigating cesium transfer to shoot

    AbstractRadiocesium (137Cs) transfer from soil to crops is largely regulated by soil potassium (K) levels owing to the chemical similarity between K and cesium (Cs). However, the mitigation of Cs translocation in soybean through soil K is lower than in other crops, highlighting the importance of clarifying soybean-specific Cs translocation mechanisms. Although root nodule symbiosis has been proposed to alter nutrient transport systems, its impact on Cs dynamics remains unclear. We hypothesized that Cs translocation mechanisms are altered under root nodule symbiosis. To elucidate these mechanisms, we conducted field experiments using three soybean genotypes with different nodulation abilities and analyzed their elemental distribution patterns. Additionally, hydroponic experiments using inoculated soybeans were conducted to investigate 137Cs distribution. We found that Cs concentrations were consistently higher in nodules than in other organs. Radioisotope imaging also showed predominant 137Cs accumulation in nodules. Covariance analysis revealed that Cs translocation to shoot was lower in genotypes with higher nodule formation under the same soil exchangeable K conditions. Furthermore, increased nodule formation, especially nodule number, was associated with reduced Cs translocation to shoot. These results suggest that nodules contribute to suppressing Cs translocation to shoot and provide new insights into Cs dynamics under root nodule symbiosis.

    Similar content being viewed by others

    Nodulation number tempers the relative importance of stochastic processes in the assembly of soybean root-associated communities

    Article
    Open access
    28 August 2023

    Genetically optimizing soybean nodulation improves yield and protein content

    Article

    09 May 2024

    Integrated single-nucleus and spatial transcriptomics captures transitional states in soybean nodule maturation

    Article

    13 April 2023

    IntroductionThe hydrogen explosions at TEPCO’s Fukushima Daiichi Nuclear Power Plant in 2011 resulted in the release and deposition of radioactive substances, including radiocesium (137Cs), onto agricultural fields. Crops cultivated in decontaminated fields remain at risk of radioactive substances transfer to edible plant parts1. The transfer of 137Cs from soil to plants has become a significant environmental and agricultural concern due to its long half-life (ca. 30.2 years). Potassium (K) fertilization has been shown to effectively reduce cesium (Cs) uptake in various crops, including soybean, wheat, and rice2,3,4. For rice and soybean cultivation, maintaining exchangeable K concentrations in soils (over 25 mg K2O per 100 g soil) through cultivation has been recommended5,6. This mitigation strategy is based on the chemical similarity between Cs+ and K+, thereby elevating soil K concentrations competitively inhibiting Cs uptake. It is well known that Cs is primarily taken up and translocated via K transport systems in plants. The high-affinity K⁺ transporter family KUP/HAK/KT has been reported to be involved in Cs⁺ transport7. In the case of Arabidopsis thaliana, AtHAK5 has been identified as a key transporter contributing to increased Cs uptake under K-deficient conditions8. In rice, OsHAK1 has also been associated with Cs transport9,10. While the role of the soybean homolog GmHAK5 in K uptake and distribution in root was evaluated11, the function of this transporter for Cs remains poorly understood. Regarding other Cs transport pathways, the Arabidopsis inward-rectifying K+ channel AKT1 has been suggested to be involved in Cs transport, although its role appears to be limited12,13. Cs is also absorbed and distributed primarily in voltage-insensitive cation channels under K-sufficient conditions14.Under equivalent soil K concentrations, soybean exhibits high Cs transfer to seeds compared with other crops, such as wheat and maize15,16. Even among leguminous species, soybean tends to accumulate more 137Cs than peanut17. It is considered that the higher Cs accumulation in soybean seed is due to high K demand. Additionally, the structural characteristics of soybean seeds facilitate Cs translocation into the edible portions18. Soybean continues to absorb K from soil until the full seed stage, and it has been reported that the ability of soybean to discriminate between K and 137Cs declines from the flowering stage to seed development, resulting in increased 137Cs uptake during this period4. K fertilization is highly effective to reduce Cs uptake and improve the production in soybean; however, due to the economic cost of fertilizers, continuous K application may not be a sustainable strategy to mitigate 137Cs transfer from soil to plant. Therefore, other mitigation strategies in addition to K fertilization should be explored for effective risk management of Cs transfer in soybean.To better understand Cs transfer, several studies have investigated the role of nonexchangeable K19, Cs adsorption to soil mineral nutrients20, and the effects of cattle manure application21. While these studies focus on soil-based countermeasures, it is also necessary to explore plant-specific strategies for reducing Cs transfer. One potential approach in soybean is the use of symbiotic microorganisms. Soybean establishes symbiotic relationships with rhizobia and arbuscular mycorrhizal (AM) fungi, which primarily enhance nitrogen and phosphorus acquisition. The role of AM fungi in Cs transfer from soil to shoot remains inconsistent, with some studies showing Cs delivery to host plants from soil22, while others report limiting its movement to shoot due to its retention within fungal hyphae23,24. Nodules, which share evolutionarily conserved common signaling pathways with AM fungi symbiosis25, are reportedly more sensitive to salinity and heavy metal stress than roots or whole plants26. In terms of the role of K in nodules, it is presumed to contribute to intercellular potential regulation, cell expansion, and energy supply27,28. In rhizobia, a positive relationship between K levels and nitrogenase activity has been reported29. These limited studies suggest that K plays a critical role in both nodule tissues and rhizobia. However, while the mechanisms of K uptake and distribution in nodules are not yet fully understood, the involvement of Cs in nodules remains even more obscure. Furthermore, both nodule formation and Cs transfer are inherently influenced by soil nutrient conditions. In particular, soil nitrogen and phosphorus conditions have been suggested to play a role in nodule formation and Cs dynamics30,31,32,33. Thus, it is necessary to investigate how nodules affect Cs transfer under various soil conditions.This study aimed to reveal the effects of nodules on the mechanism of Cs transfer to shoot. To achieve this goal, we conducted field experiments in 2021 and 2024 using three soybean genotypes with different nodulation abilities, under various fertilization treatments, to investigate the dynamics of monovalent cations in each plant organ and soil. We also attempted to conduct a spatial analysis of 137Cs in root and nodule tissue to investigate the 137Cs distribution pattern. We found that Cs levels were higher in nodules than in other plant organs and increased nodule formation was potentially associated with suppressing Cs translocation to shoot. Our findings suggest that nodules may play a key role in mitigating Cs transfer in addition to K fertilization, providing a potential strategy for 137Cs risk management in leguminous crops.Materials and methodsPlant materials and growth conditionsThree soybean (Glycine max (L.) Merrill) genotypes with different nodulation abilities [a normal nodulating genotype (Enrei), a hyper-nodulating genotype (En-b0-1), and a non-nodulating genotype (En1282), all derived from Enrei as described by Hamaguchi et al.34] were cultivated in a field at Hokkaido University in 2021 and 2024. These fields were established in 1914, with four fertilizer treatments, namely, complete fertilization (+ NPK), without nitrogen (− N), without phosphorus (− P), and without potassium (− K), applied continuously for over 100 years35. Soil type is categorized as gray lowland soil. In the − N, − P, − K, and + NPK treatments, each plot size was 5.25 × 15.80 m. Moreover, the N, P, and K fertilizers were applied as ammonium sulfate, superphosphate, and potassium sulfate, respectively (100 kg N, P2O5, K2O ha−1). Three seeds were sown at 70 cm × 20 cm intervals and then thinned to two plants per hill. Plant organs (leaf, petiole, stem, pod, seed, and root) and bulk soil (the 0–15 cm soil in the inter-row space) were sampled at the flowering and maturity stages with three replications. Plant samples at the flowering stage were divided into leaf, petiole, stem, root, and nodule. Plant samples were then dried at 80 °C for 1 week, and their dry weights were measured. These dried plant samples were ground for mineral analysis. Soil samples were air-dried and passed through a sieve with a 2 mm diameter for mineral analysis.Measurement of plant mineralPlant samples were digested in 2 mL of 61% (w/v) HNO3 (EL grade; Kanto Chemical, Tokyo, Japan) at 107.5 °C in a DigiPREP apparatus (SCP Science, Canada). After approximately 1.5 h, about 0.5 mL of H2O2 (EL grade; Kanto Chemical, Tokyo, Japan) was added, following by further digestion for another 15 min. The digested solution was cooled and filled with 10 mL with 2% HNO3 in Milli-Q water. Cs, K, sodium (Na), and rubidium (Rb) were analyzed using inductively coupled plasma mass spectrometry (ICP-MS) (ELAN DRC-e; Perkin Elmer, Waltham, MA, USA). To determine δ15N in leaves, plant samples were weighed into tin capsules and analyzed using an isotope ratio mass spectrometer (Integra2, sercon, UK). The δ15N values were reported relative to atmospheric N2.Measurement of soil mineralFor the soil exchangeable cation, soil was extracted with 1 M ammonium acetate at a soil/solution ratio of 1:20, shaken for 1 h and filtered using an Advantec® quantitative filter paper No. 5C (Toyo Roshi Co., Ltd., Tokyo, Japan). The filtrates were measured using ICP-MS (ELAN DRC-e; PerkinElmer, Waltham, MA, USA) to determine the concentration of soil exchangeable cation. Total nitrogen concentration in soil samples was determined by Kjeldahl digestion. Available phosphorus concentrations were measured using the Bray II method36. Soil pH (soil/Milli-Q water = 1:5, w/v) was measured with a pH meter (HORIBA Ltd., Kyoto, Japan).Spatial imaging of 137Cs in root and nodule tissuesTo study the distribution of 137Cs in the nodules and roots, germinated soybean (Glycine max cv. Enrei) seeds were cultivated hydroponically in 3-L plastic containers while aerating and inoculated with Bradyrhizobium japonicum USDA110 (obtained from the National Agriculture and Food Research Organization, NARO, Japan) at a concentration of 1 × 107 colony-forming units mL−1. The solution was an N-free half-strength modified Hoagland’s nutrient solution (3 mM K, 2 mM Ca, 1 mM P, 0.5 mM S, 0.5 mM Mg). The nutrient solution inoculated with rhizobia was replaced every 3 days. At 18 days after the start of hydroponic cultivation, 3.7 kBq mL−1 137Cs was added to the nutrient solution immediately after 18 days of hydroponic cultivation without any pre-treatment, and the plants and rhizobia were allowed to absorb it for 24 h. After washing the root with pure water, the root with attached nodules was cut into approximately 1 cm pieces and immediately frozen by submerging in liquid nitrogen. Thereafter, the roots and sliced sections were prepared according to a previously described method37,38. Each sample was embedded in an embedding medium, covered with an adhesive film (Cryo-Film transfer kit, Finetec, Tokyo, Japan), and sliced with a cryostat into serial sections of 5 µm thickness at − 20 °C. A BAS IP TR imaging plate (GE Healthcare) was stuck to the frozen section on the side directly opposite the attached film. The imaging plate (IP) with the samples was kept in the freezer at − 80 °C for 6 days. The image on the IP was scanned with an FLA5000 image analyzer (Fujifilm, Tokyo, Japan) at a resolution of 10 µm.Statistical analysesTo compare dry weight and elemental distribution of plant organs, a three-way analysis of variance (ANOVA) was first performed to evaluate the effects of organ, fertilizer treatment, and genotype in R (version 4.4.1). However, strong effects of treatment and genotype masked organ-specific differences. To address this, a one-way ANOVA was subsequently conducted for organs within each treatment–genotype combination, followed by Tukey’s multiple comparison analysis. To assess effects of treatment and genotype by year, a two-way ANOVA was conducted, with Tukey’s multiple comparison analysis applied when interactions were significant. Regarding correlation analysis, we used the Spearman’s rank correlation to accommodate potential nonlinearity and non-normality. An analysis of covariance (ANCOVA) was conducted in Python (version 3.9.18) using the statsmodels library (version 0.14.0). Least‐squares means (LS‐means) for each genotype were calculated at the overall mean of covariates, and pairwise t-tests were then performed on those LS-means. False discovery rate correction was applied using the two-stage Benjamini–Hochberg method (BH).ResultsSoil chemical properties under different fertilization treatmentsTo clarify the effect of nodule formation on cation transfer from soil to soybean, we investigated three soybean genotypes with different nodulation abilities (normal nodulating genotype: Enrei; hyper-nodulating genotype: En-b0-1; and non-nodulating genotype: En1282) grown under various fertilization treatments. Compared with the + NPK treatment, soil exchangeable K concentrations were significantly reduced under − K treatment consistently across both years and growth stages (Supplementary Table S1). Meanwhile, soil exchangeable Cs was highest under − K treatment. Soil exchangeable K was higher and Cs was lower in the − N and − P treatments than in the + NPK treatment. Other exchangeable monovalent cations in soil showed similar trends to Cs. Available phosphorus in soil was lowest under − P treatment, and soil total nitrogen tended to be lowest under − N treatment. Soil pH was significantly increased under − N treatment.Growth of genotypes with different nodulation abilities under different soil nutrient treatmentsIn the normal nodulating genotype, shoot (combined leaf, petiole, and stem) dry weight was highest in the + NPK treatment in both 2021 and 2024, and decreased in the order of − N, − P, and − K treatment (Table 1). In contrast, both the hyper-nodulating and non-nodulating genotypes reduced shoot biomass in − N treatment. Across all genotypes and treatments, leaf dry weight was consistently the highest among organs, while nodule dry weight was the lowest (Fig. 1a,b). Under − K treatments, growth was consistently suppressed in all genotypes, with the number and dry weight of nodules being the lowest across treatments in both 2021 and 2024 (Table 1). Nodule formation showed no significant difference between the − P and + NPK treatments but was significantly increased under the − N treatment. Seed weight at maturity showed a similar tendency to shoot dry weight at the flowering stage, with a significant reduction in − K treatment observed particularly in the normal nodulating genotype. While overall biomass and yield were higher in 2021 compared to those in 2024, the effects of fertilizer treatments and genotypes were largely consistent across years.Table 1 Plant growth at flowering stage and maturity stage.Full size tableFig. 1Plant growth per organ at flowering stage. (a, b) Each organ dry weight of soybean genotypes at flowering stage in 2021 (a) and 2024 (b). For each genotype and treatment, statistical analyses were performed as one-way ANOVA with Tukey’s multiple comparisons test for each plant organ. Dots represent distinct biological replicates for each treatment (n = 3), and bars indicate mean ± SE. Different letters indicate statistically significant differences (P < 0.05) as determined by Tukey’s multiple comparisons test. “n.s.” indicates no significant differences among plant organs (P > 0.05).Full size imageElemental distribution within plant in field experimentA two-way ANOVA was conducted for shoot Cs and K concentrations to evaluate the effects of genotype and treatment on the elemental distribution to shoot. Significant effects on shoot Cs concentration were observed only in 2021. Cs levels in the hyper-nodulating genotype tended to be lower compared with those in the normal and non-nodulating genotypes, particularly in − K and + NPK treatment (Table 2). Across all genotypes, shoot Cs concentration was highest in − K and lowest in − N treatment. Shoot K concentrations consistently decreased in the − K treatment in both years across all genotypes. To investigate elemental distribution among organs, soybean samples at the flowering stage were separated into leaves, petioles, stems, roots, and nodules. In the normal nodulating genotype grown in 2021, Cs concentrations were highest in nodules across all treatments compared with those in other organs, with particularly pronounced accumulation in − K and + NPK treatment (Fig. 2a). A similar tendency was observed in the hyper-nodulating genotype, which also exhibited significantly increased concentrations of Cs in nodules under these treatments. Although one-way ANOVA detected no significant differences under –N, –P, and + NPK treatment in 2024, Cs concentrations tended to be higher in nodules (Fig. 2b). In 2021, nodules had significantly higher K concentrations than other organs in the − K treatment in both the normal and hyper-nodulating genotypes (Fig. 2c). However, this trend was not observed in 2024, showing no consistent distribution pattern (Fig. 2d). For Cs concentrations in shoot, leaves were the primary organ of Cs allocation. In contrast, K concentrations were significantly higher in petioles than in leaves in all treatments except − K in both years. At maturity, seed Cs concentrations increased in the − K treatment, while seed K concentrations decreased (Table 2). For other monovalent cations of organs at the flowering stage, Na significantly accumulated in roots in all genotypes in both 2021 and 2024 (Supplementary Fig. S1). In the − K treatment, Na concentration increased in nodules but remained lower than that in roots. Rb tended to be more concentrated in nodules than in other organs in the normal nodulating and hyper-nodulating genotypes in 2021. A similar tendency was also observed in the normal nodulating genotype under the − K treatment in 2024.Table 2 Elemental concentration in shoot and seed.Full size tableFig. 2Cs and K concentrations in each soybean organ. (a, b) Concentration of Cs in each organ at flowering stage in 2021 (a) and 2024 (b). (c, d) Concentration of K in each organ at flowering stage in 2021 (c) and 2024 (d). Statistical analyses were conducted using one-way ANOVA followed by Tukey’s multiple comparisons test to evaluate differences among plant organs within each genotype and treatment. Dots represent distinct biological replicates for each treatment (n = 3), and bars indicate mean ± SE. Different letters indicate statistically significant differences (P < 0.05) as determined by Tukey’s multiple comparisons test. “n.s.” indicates no significant differences among plant organs (P > 0.05).Full size imageDistribution ratios among organs were calculated to evaluate the effect of nodules on the elemental distribution within plant. The proportion of Cs distributed to nodules was highest in the − N treatment in both 2021 and 2024 (Supplementary Fig. S2). Compared with the normal nodulating genotype, the hyper-nodulating genotype showed higher Cs proportions allocated to nodules across all treatments. K exhibited similar distribution patterns to Cs in both years. Spearman’s rank correlation coefficient (r) between the proportion of Cs allocated to nodules and that allocated to shoot was significantly negative (r =  − 0.77, P < 0.005) (Supplementary Fig. S3). Similar patterns were observed for K (r =  − 0.66, P < 0.005). In contrast, Na was predominantly distributed to roots across all treatments, and increased allocation to nodules did not result in the change of shoot Na level (Supplementary Figs. S2 and S3). Rb showed a distribution pattern similar to those of Cs and K (Supplementary Fig. S2), and a negative correlation was observed between nodule and shoot allocation ratios (r =  − 0.70, P < 0.005) (Supplementary Fig. S3).Spatial localization of 137Cs in root systemAlthough high concentrations of stable Cs in nodules were observed in field experiments, it is also necessary to confirm the distribution pattern of 137Cs. A transverse section (Fig. 3a) and its corresponding autoradiograph (Fig. 3b) were superimposed to generate Fig. 3c. According to the overlaid image, a strong signal was observed in the nodule, particularly in the infection zone. In contrast, the signal in the main and lateral roots was very weak. These observations suggest that 137Cs is taken up and accumulated in the nodule at 24 h after the start of absorption.Fig. 3137Cs autoradiography in a soybean root and nodule after 24 h of 137Cs absorption. (a) Transverse section of the soybean root including the root nodule, (b) Autoradiograph of (a) showing the localization of 137Cs. (c) Superimposed image of (a) and (b) after the black color in (b) was converted to white. Bar, 1 mm.Full size imageRelationship between soil minerals and plant mineralsAlthough the Cs distribution pattern within plant was clarified, a comprehensive understanding of the mineral nutrient dynamics requires not only a distribution analysis but also an investigation of nutrient uptake from the soil. We then investigated the relationship between mineral nutrients in the soil and those in soybean. A consistently significant positive correlation was observed between soil exchangeable Cs concentrations and Cs concentrations of both shoots and seeds in both 2021 and 2024 (Supplementary Fig. S4). In contrast, a significant negative correlation was found between soil exchangeable K concentrations at the flowering stage and shoot Cs concentrations in 2021 (r =  − 0.74, P < 0.005), with a similar tendency observed in 2024 (r =  − 0.79, P < 0.005) (Fig. 4a,b). In all three genotypes, soil exchangeable K was negatively correlated with shoot Cs. Given this strong inverse relationship, soil exchangeable K was included as a covariate when comparing shoot Cs among genotypes. To compare the effects of genotypes on shoot Cs after adjusting for soil exchangeable K, an ANCOVA (shoot Cs ~ exchangeable K + genotype) was performed for the hyper-nodulating, normal, and non-nodulating genotypes, followed by pairwise t-tests of the resulting LS-means. The exchangeable K concentrations adjusted shoot Cs concentrations of the hyper-nodulating genotype were significantly lower than that of both the normal and non-nodulating genotypes (FDR-corrected two-stage BH P < 0.05), whereas the exchangeable K concentrations adjusted shoot Cs concentrations of the normal and non-nodulating genotypes were not significantly different. At the maturity stage, significant negative correlations were found between soil exchangeable K concentrations and seed Cs concentrations in both 2021 (r =  − 0.84, P < 0.005) and 2024 (r =  − 0.88, P < 0.005) (Fig. 4c,d). However, ANCOVA did not show significant differences among genotypes in this relationship.Fig. 4Relationship between soybean Cs concentration and soil exchangeable K concentration. (a, b) Shoot Cs concentration at flowering stage in 2021 (a) and 2024 (b). (c, d) Seed Cs concentration at maturity stage in 2021 (c) and 2024 (d). Line types represent soybean genotypes: Enrei (solid), En-b0-1 (dotted), and En1282 (dashed). Power-law equations (y = axb) were estimated separately for each genotype. Spearman’s rank correlation coefficient (r) and corresponding two-sided P values are indicated (†P < 0.1; *P < 0.05; **P < 0.01; ***P < 0.001). When pooled across genotypes: (a) r =  − 0.74, P < 0.001; (b) r =  − 0.79, P < 0.001; (c) r =  − 0.84, P < 0.001; (d) r =  − 0.88, P < 0.001.Full size imageRelationship between nodulation and Cs translocationThe relationship between soil exchangeable K concentration and shoot Cs concentration varied among nodulation genotypes, indicating that nodule formation may influence Cs translocation to shoot. We then further investigated the relationship between nodulation traits and shoot Cs concentration at the flowering stage. Shoot Cs concentration decreased as the nodule dry weight increased (r =  − 0.55, P < 0.005), and a similar negative relationship was found with nodule number (r =  − 0.66, P < 0.005) (Fig. 5a,b). Shoot Cs concentration was negatively correlated with soil exchangeable K concentration (Fig. 4a,b), and both nodule number and nodule dry weight were significantly reduced under K-deficient conditions (Table 1). The relationship between nodule formation and shoot Cs concentration may be potentially confounded by soil K availability. To address this, partial correlation analysis was performed using soil exchangeable K concentration at the flowering stage as a confounding factor. While no significant relationship was found between shoot Cs concentration and nodule dry weight, a significant negative partial correlation was retained with nodule number (r =  − 0.37, P < 0.01). δ15N in leaves, which is used as an indirect indicator of nitrogen fixation, showed no significant correlation with shoot Cs concentration (Supplementary Fig. S5). In contrast, a significant correlation was observed between nodule Cs concentration and leaf δ15N, and this relationship remained significant after accounting for soil exchangeable K (Supplementary Fig. S6). Similar correlation analyses were conducted for other monovalent cations in shoot. K concentration showed significantly positive correlations with both nodule dry weight and nodule number (Supplementary Fig. S7). Meanwhile, Na concentration was not significantly correlated with either trait. Rb showed a pattern similar to Cs. It was not significantly correlated with nodule dry weight but was negatively correlated with nodule number (r =  − 0.49, P < 0.005). Regarding nitrogen fixation, no significant correlations were found between δ15N and other monovalent cation concentrations in leaves (Supplementary Fig. S5). For nodule monovalent cation concentrations, significant correlations were observed for Na and Rb, but not for K (Supplementary Fig. S6).Fig. 5Relationship between shoot Cs concentration and nodule formation. (a) Nodule dry weight. (b) Number of nodules. Spearman’s rank correlation coefficient (r) and corresponding two-sided P values are indicated (†P < 0.1; *P < 0.05; **P < 0.01; ***P < 0.001). Partial correlation coefficient (partial r) was calculated using soil exchangeable K concentration at flowering stage as a confounding factor (†P < 0.1; *P < 0.05; **P < 0.01; ***P < 0.001).Full size imageDiscussionRoot nodule symbiosis affects the uptake and transport systems not only for nitrogen but also for other mineral nutrients, such as K39,40. K has been suggested to play a role in maintaining ion homeostasis and nodule development and nitrogenase activity41. Our study revealed that Cs consistently accumulated at higher concentrations in nodules at the flowering stage in both 2021 and 2024, whereas K levels in nodules exhibited an inconsistent trend (Fig. 2). Moreover, 137Cs was mainly accumulated in nodules under root nodule symbiotic conditions (Fig. 3). As reported by Shinano et al.42, K is preferentially translocated to the shoot rather than Cs, possibly limiting Cs mobility and leading to Cs accumulation in the root system, particularly in nodules. Supporting this notion, Cs accumulation in nodules was significantly high under K-deficient conditions, indicating that limited K availability enhances Cs accumulation in nodules. Furthermore, 137Cs was particularly enriched in the infected cells of nodule. K is essential for rhizobia to maintain osmotic balance43. This suggests that Cs may also accumulate preferentially in bacteroids. However, Cs localization within nodules remains inconclusive due to the short-term observation and the overall K requirement of nodule tissues. To clarify this, further studies, such as merging Cs imaging with rhizobia staining or conducting experiments using artificial nodules, will be necessary. Additionally, because nutrient uptake and distribution mechanisms may differ between symbiotic and non-symbiotic conditions, results such as 137Cs imaging should be regarded as specific to the root nodule symbiotic condition.Increasing exchangeable K concentration in soil has been widely regarded as the most effective strategy for reducing Cs transfer, regardless of species. We also observed a significant negative correlation between soil exchangeable K and shoot Cs concentrations in soybean (Fig. 4). However, ANCOVA showed the different relationships of different nodulation genotypes, suggesting that nodule formation influences Cs uptake at the flowering stage. In nodules, K supports ATP supply and nitrogenase activity and is positively associated with carbon and nitrogen transport44,45,46. In the hyper-nodulating genotype, increased total photosynthate allocation to nodules may result from enhanced nodulation in lateral root and reduced root elongation47. Changes in the root system may have increased carbon costs in roots and altered nutrient uptake mechanisms, potentially affecting translocation and consequently reducing Cs translocation to shoot. This suggestion is supported by the different elemental distributions in roots and nodules observed in the hyper-nodulating genotype compared with those in other genotypes (Supplementary Fig. S2). In contrast, there were no significant differences among genotypes at the maturity stage, indicating that seed Cs levels remain homeostatic despite genotypic differences in Cs uptake at the flowering stage.Furthermore, our results suggest that nodules may contribute to suppressing Cs translocation to shoot (Fig. 5). The high Cs concentrations observed in nodules (Fig. 2a,b) indicate that nodules may function as “buffer zones”. While heavy metals and alkali metals generally differ in terms of chemical properties and translocation mechanisms, nodules have been proposed to serve as buffer zones to mitigate heavy metals48,49,50. We also found that increased nodule number was associated with increased shoot K concentrations (Supplementary Fig. S7), suggesting that preferential K translocation to the shoot may have indirectly promoted Cs accumulation in nodules. Notably, δ15N, one of nitrogen fixation activity indicators, suggested that nitrogen fixation may not contribute to reduced Cs translocation to shoot (Supplementary Fig. S5). In contrast, Cs accumulation in nodules did not appear to inhibit nitrogen fixation activity (Supplementary Fig. S6). Further studies should use more quantitative indicators of nitrogen fixation activity, such as acetylene reduction assays, to examine its relationship with Cs dynamics.Elemental uptake and distribution in nodules primarily occur via two pathways, i.e. direct transport through the nodule epidermis and cortex51 and translocation from roots via vascular tissues into nodules52. Several K transporters, including HAK5, GORK, SKOR and KUP7, are expressed in the epidermal and vascular tissues53,54,55,56. Interestingly, our study found that the suppression of Cs translocation was more strongly associated with nodule number than with nodule dry weight (Fig. 5). In soybean nodules, the vascular bundles extend radially around the infected zone57,58. Therefore, the increase in the nodule number implies an expansion of the total epidermal and vascular tissue area. Treating transporter density per unit membrane area as approximately constant, this expansion broadens the expression domain of K transporters at these sites and, in turn, altering Cs translocation. Verification of this hypothesis requires functional analyses of transporters, including assessments of Cs selectivity, and spatial localization of expression in the relevant regions. As another hypothesis, our findings suggest that a single nodule does not accumulate Cs indefinitely. Soybean forms determinate nodules, where nodule development is driven by cell expansion rather than division59. To facilitate cell expansion, intercellular potential balance needs to be maintained, potentially absorbing Cs in nodules during nodule development. However, in mature nodules, an elevated intracellular potential may limit further passive Cs transport. K transporters such as GmHAK5 and LjKUP, which are expressed in nodules, have been identified, but their involvement in active Cs sequestration within nodules remains unclear60,61. To clarify these physiological mechanisms, it is necessary to investigate the subcellular localization of elements and gene expression within nodules spatiotemporally, in accordance with nodule developmental stages.For other cations, Cs+, K+, Na+ and Rb+ are chemically similar and generally exhibit competitive uptake due to overlapping parts of transport pathways7. In this study, exchangeable Na and Rb concentrations in the soil increased under K-deficient conditions, like Cs (Table S1). Regarding Rb distribution within the plant, Rb concentrations in nodules tended to increase, similar to Cs, and nodule number appeared to be associated with reduced Rb translocation to shoot (Figs. 2, S1, S2, S3 and S7). These similar dynamics of Cs and Rb suggest the involvement of a common transport mechanism. Further studies are needed to identify nodule-localized transporters responsible for Cs and Rb uptake and to clarify their ion selectivity. In contrast, Na was predominantly distributed to the roots, and no significant relationship was observed between nodule traits and Na translocation patterns (Figs. 2, S1, S2, S3 and S7). In legumes, Na taken up from the soil is retained prior to crossing the endodermis, with reduced levels in xylem vessels, thereby limiting its upward movement62. Although recent work in rice suggested that Na indirectly influences Cs transport through Na transporters63, Cs and Rb transport is thought to rely mainly on K transport systems7. This indicates that Na is likely distributed through a distinct mechanism from Cs and Rb. To elucidate differences in these distribution mechanisms, future studies of rhizobia-inoculated soybean should employ techniques such as isotope tracing and elemental mapping. Importantly, because exchangeable Na concentrations in soil differ substantially from those of Rb and Cs, potential ionic interactions and competition among these monovalent cations should be carefully considered.ConclusionOur study revealed the influence of nodules on the mechanism of Cs translocation to shoot. Through field experiments conducted using soybean genotypes with different nodulation abilities, comparison of elemental concentrations in each organ revealed that Cs was highly accumulated in nodules. Similarly, spatial mapping of 137Cs in the root system revealed that 137Cs was predominantly distributed in nodules. These results highlight the importance of nodules as key sites for Cs sequestration. This study also demonstrated a significant negative correlation between nodule number and shoot Cs concentration, suggesting that increased nodulation may suppress Cs translocation to shoot. These findings provide fundamental insights into the potential use of root nodule symbiosis as a strategy for mitigating Cs transfer to shoots. To further elucidate the mechanisms underlying Cs accumulation in nodules and its suppressed translocation to shoot, spatiotemporal analyses of elemental distribution and gene expression within nodules are needed.

    Data availability

    Raw data were generated at Hokkaido University. Derived data supporting the findings of this study are available from Kazuki Murashima on request.
    ReferencesShinano, T. Mitigation of radioactive contamination from farmland environment and agricultural products. MESE 2, 454–461 (2016).Article 

    Google Scholar 
    Zhu, Y. & Smolders, E. Plant uptake of radiocaesium: A review of mechanisms, regulation and application. J. Exp. Bot. 51, 1635–1645 (2000).Article 
    CAS 
    PubMed 

    Google Scholar 
    Fujimura, S. et al. Theoretical model of the effect of potassium on the uptake of radiocesium by rice. J. Environ. Radioact. 138, 122–131 (2014).Article 
    CAS 
    PubMed 

    Google Scholar 
    Matsunami, H., Uchida, T., Kobayashi, H., Ota, T. & Shinano, T. Comparative dynamics of potassium and radiocesium in soybean with different potassium application levels. J. Environ. Radioact. 233, 106609. https://doi.org/10.1016/j.jenvrad.2021.106609 (2021).Article 
    CAS 
    PubMed 

    Google Scholar 
    MAFF, NARO & NIAES. Factors and Countermeasure of Soybean with High Radioactive Cesium Concentration. Investigation of Factors and Result of Trial Examination. https://www.maff.go.jp/j/kanbo/joho/saigai/pdf/youin_daizu_3.pdf (2015).Kato, N. et al. Potassium fertilizer and other materials as countermeasures to reduce radiocesium levels in rice: Results of urgent experiments in 2011 responding to the Fukushima Daiichi Nuclear Power Plant accident. Soil Sci. Plant Nutr. 61, 179–190 (2015).Article 
    CAS 

    Google Scholar 
    White, P. J. & Broadley, M. R. Tansley Review No. 113: Mechanisms of caesium uptake by plants. New Phytol. 147, 241–256 (2000).Article 
    CAS 

    Google Scholar 
    Qi, Z. et al. The high affinity K+ transporter AtHAK5 plays a physiological role in planta at very low K+ concentrations and provides a caesium uptake pathway in Arabidopsis. J. Exp. Bot. 59, 595–607 (2008).Article 
    CAS 
    PubMed 

    Google Scholar 
    Rai, H. et al. Cesium uptake by rice roots largely depends upon a single gene, HAK1, which encodes a potassium transporter. Plant Cell Physiol. 58, 1486–1493 (2017).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Nieves-Cordones, M. et al. Production of low-Cs+ rice plants by inactivation of the K+ transporter OsHAK1 with the CRISPR-Cas system. Plant J. 92, 43–56 (2017).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Chao, M. et al. Structural features and expression regulation analysis of potassium transporter gene GmHAK5 in soybean (Glycine max L.). Plant Growth Regul. 102, 471–483 (2024).Article 
    CAS 

    Google Scholar 
    Broadley, M. R., Escobar-Gutierrez, A. J., Bowen, H. C., Willey, N. J. & White, P. J. Influx and accumulation of Cs+ by the akt1 mutant of Arabidopsis thaliana (L.) Heynh. lacking a dominant K+ transport system. J. Exp. Bot. 52, 839–844 (2001).Article 
    CAS 
    PubMed 

    Google Scholar 
    Adams, E., Miyazaki, T., Saito, S., Uozumi, N. & Shin, R. Cesium inhibits plant growth primarily through reduction of potassium influx and accumulation in Arabidopsis. Plant Cell Physiol. 60, 63–76 (2019).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Kanter, U., Hauser, A., Michalke, B., Dräxl, S. & Schäffner, A. R. Caesium and strontium accumulation in shoots of Arabidopsis thaliana: genetic and physiological aspects. J. Exp. Bot. 61, 3995–4009 (2010).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Hirayama, T., Takeuchi, M., Nakayama, H. & Nihei, N. Effects of decreasing radiocesium transfer from the soil to soybean plants and changing the seed nutrient composition by the increased application of potassium fertilizer. Bull. Fukushima Agric. Technol. Centre (Jpn. Engl. Abstr.) 9, 1–10 (2018).
    Google Scholar 
    IAEA. Environmental Transfer of Radionuclides in Japan Following the Accident at the Fukushima Daiichi Nuclear Power Plant: Report of Working Group 4 Transfer Processes and Data for Radiological Impact Assessment Subgroup 2 on Fukushima Data (IAEA, 2020).
    Google Scholar 
    Kubo, K. et al. Comparative study of radioactive cesium transfer from soil to peanut and soybean. Soil Sci. Plant Nutr. 67, 707–715 (2021).Article 
    CAS 

    Google Scholar 
    Nihei, N. et al. The concentration distributions of Cs in soybean seeds. Radioisotopes 66, 235–242 (2017).Article 
    CAS 

    Google Scholar 
    Kurokawa, K. et al. Advanced approach for screening soil with a low radiocesium transfer to brown rice in Fukushima based on exchangeable and nonexchangeable potassium. Sci. Total Environ. 743, 140458. https://doi.org/10.1016/j.scitotenv.2020.140458 (2020).Article 
    CAS 
    PubMed 

    Google Scholar 
    Ogasawara, S. et al. Phytoavailability of 137Cs and stable Cs in soils from different parent materials in Fukushima, Japan. J. Environ. Radioact. 198, 117–125 (2019).Article 
    CAS 
    PubMed 

    Google Scholar 
    Suzuki, M. et al. Effects of cattle manure compost application on crop growth and soil-to-crop transfer of cesium in a physically radionuclide-decontaminated field. Sci. Total Environ. 908, 167939. https://doi.org/10.1016/j.scitotenv.2023.167939 (2024).Article 
    CAS 
    PubMed 

    Google Scholar 
    Dupré De Boulois, H., Voets, L., Delvaux, B., Jakobsen, I. & Declerck, S. Transport of radiocaesium by arbuscular mycorrhizal fungi to Medicago truncatula under in vitro conditions. Environ. Microbiol. 8, 1926–1934 (2006).Article 

    Google Scholar 
    Joner, E. J. et al. No significant contribution of arbuscular mycorrhizal fungi to transfer of radiocesium from soil to plants. Appl. Environ. Microbiol. 70, 6512–6517 (2004).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Dupré De Boulois, H. et al. Role and influence of mycorrhizal fungi on radiocesium accumulation by plants. J. Environ. Radioact. 99, 785–800 (2008).Article 
    PubMed 

    Google Scholar 
    Ivanov, S. et al. Rhizobium–legume symbiosis shares an exocytotic pathway required for arbuscule formation. Proc. Natl. Acad. Sci. U.S.A. 109, 8316–8321 (2012).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Fedorova, E. E. et al. Potassium content diminishes in infected cells of Medicago truncatula nodules due to the mislocation of channels MtAKT1 and MtSKOR/GORK. J. Exp. Bot. 72, 1336–1348 (2020).Article 
    PubMed Central 

    Google Scholar 
    Cakmak, I., Hengeler, C. & Marschner, H. Partitioning of shoot and root dry matter and carbohydrates in bean plants suffering from phosphorus, potassium and magnesium deficiency. J. Exp. Bot. 45, 1245–1250 (1994).Article 
    CAS 

    Google Scholar 
    Ragel, P., Raddatz, N., Leidi, E. O., Quintero, F. J. & Pardo, J. M. Regulation of K+ nutrition in plants. Front. Plant Sci. https://doi.org/10.3389/fpls.2019.00281 (2019).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Gober, J. W. & Kashket, E. R. K+ regulates bacteroid-associated functions of Bradyrhizobium. Proc. Natl. Acad. Sci. U.S.A. 84, 4650–4654 (1987).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Gyuricza, V., Dupré de Boulois, H. & Declerck, S. Effect of potassium and phosphorus on the transport of radiocesium by arbuscular mycorrhizal fungi. J. Environ. Radioact. 101, 482–487 (2010).Article 
    CAS 
    PubMed 

    Google Scholar 
    ten Hoopen, F. et al. Competition between uptake of ammonium and potassium in barley and Arabidopsis roots: molecular mechanisms and physiological consequences. J. Exp. Bot. 61, 2303–2315 (2010).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Xia, X., Ma, C., Dong, S., Xu, Y. & Gong, Z. Effects of nitrogen concentrations on nodulation and nitrogenase activity in dual root systems of soybean plants. Soil Sci. Plant Nutr. 63, 470–482 (2017).Article 
    CAS 

    Google Scholar 
    Mirriam, A. et al. Role of phosphorus and inoculation with Bradyrhizobium in enhancing soybean production. Adv. Agric. 2023, 3231623. https://doi.org/10.1155/2023/3231623 (2023).Article 

    Google Scholar 
    Hamaguchi, H. et al. Nitrogen fertilization affects yields and storage compound contents in seeds of field-grown soybeans cv Enrei (Glycine max L) and its super-nodulating mutant En-b0–1 through changing N2 fixation activity of the plants. Soil Soil Sci. Plant Nutr. 66, 299–307 (2020).Article 
    CAS 

    Google Scholar 
    Watanabe, T., Urayama, M., Shinano, T., Okada, R. & Osaki, M. Application of ionomics to plant and soil in fields under long-term fertilizer trials. SpringerPlus 4, 781. https://doi.org/10.1186/s40064-015-1562-x (2015).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Bray, R. H. & Kurtz, L. T. determination of total, organic, and available forms of phosphorus in soils. Soil Sci. 59, 39–46 (1945).Article 
    ADS 
    CAS 

    Google Scholar 
    Kawamoto, T. Use of a new adhesive film for the preparation of multi-purpose fresh-frozen sections from hard tissues, whole-animals, insects and plants. Arch. Histol. Cytol. 66, 123–143 (2003).Article 
    PubMed 

    Google Scholar 
    Kobayashi, N. I., Tanoi, K., Hirose, A. & Nakanishi, T. M. Characterization of rapid intervascular transport of cadmium in rice stem by radioisotope imaging. J. Exp. Bot. 64, 507–517 (2013).Article 
    CAS 
    PubMed 

    Google Scholar 
    Udvardi, M. & Poole, P. S. Transport and metabolism in legume-rhizobia symbioses. Annu. Rev. Plant Biol. 64, 781–805 (2013).Article 
    CAS 
    PubMed 

    Google Scholar 
    Ma, Y., Xiao, C., Liu, J. & Ren, G. Nutrient-dependent regulation of symbiotic nitrogen fixation in legumes. Hortic. Res. 12, uhae321. https://doi.org/10.1093/hr/uhae321 (2025).Article 
    CAS 
    PubMed 

    Google Scholar 
    Li, Y. et al. Metal nutrition and transport in the process of symbiotic nitrogen fixation. Plant Commun. 5, 100829. https://doi.org/10.1016/j.xplc.2024.100829 (2024).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Shinano, T. et al. Varietal difference in radiocesium uptake and transfer from radiocesium deposited soils in the genus Amaranthus. Soil Sci. Plant Nutr. 60, 809–817 (2014).Article 
    CAS 

    Google Scholar 
    Domínguez-Ferreras, A., Muñoz, S., Olivares, J., Soto, M. J. & Sanjuán, J. Role of potassium uptake systems in Sinorhizobium meliloti osmoadaptation and symbiotic performance. J. Bacteriol. 191, 2133–2143 (2009).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Mengel, K., Haghparast, M.-R. & Koch, K. The effect of potassium on the fixation of molecular nitrogen by root nodules of Vicia faba. Plant Physiol. 54, 535–538 (1974).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Lynd, J. Q., Odell, G. V. Jr. & McNew, R. W. Soil potassium effects on nitrogenase activity with associated nodule components of hairy vetch at anthesis. J. Plant Nutr. 4, 303–318 (1981).Article 
    CAS 

    Google Scholar 
    Cao, H.-R. et al. Carbon-nitrogen trading in symbiotic nodules depends on magnesium import. Curr. Biol. 32, 4337–4349 (2022).Article 
    CAS 
    PubMed 

    Google Scholar 
    Ito, S. et al. Real-time ananlysis of translocation of photosynthates to nodules in soybean plants using 11CO2 with a positron-emitting tracer imaging system (PETIS). Radioisotope (Jpn. Engl. Abstr.) 59, 145–154 (2010).CAS 

    Google Scholar 
    Chen, W.-M., Wu, C.-H., James, E. K. & Chang, J.-S. Metal biosorption capability of Cupriavidus taiwanensis and its effects on heavy metal removal by nodulated Mimosa pudica. J. Hazard. Mater. 151, 364–371 (2008).Article 
    CAS 
    PubMed 

    Google Scholar 
    Wani, P. A., Khan, Md. S. & Zaidi, A. Effect of metal-tolerant plant growth-promoting Rhizobium on the performance of pea grown in metal-amended soil. Arch. Environ. Contam. Toxicol. 55, 33–42 (2008).Article 
    CAS 
    PubMed 

    Google Scholar 
    Guo, J. & Chi, J. Effect of Cd-tolerant plant growth-promoting rhizobium on plant growth and Cd uptake by Lolium multiflorum Lam. and Glycine max (L.) Merr in Cd-contaminated soil. Plant Soil 375, 205–214 (2014).Article 
    CAS 

    Google Scholar 
    Slatni, T., Krouma, A., Gouia, H. & Abdelly, C. Importance of ferric chelate reductase activity and acidification capacity in root nodules of N2-fixing common bean (Phaseolus vulgaris L.) subjected to iron deficiency. Symbiosis 47, 35–42 (2009).Article 
    CAS 

    Google Scholar 
    Roy, S. et al. Celebrating 20 years of genetic discoveries in legume nodulation and symbiotic nitrogen fixation. Plant Cell 32, 15–41 (2020).Article 
    CAS 
    PubMed 

    Google Scholar 
    Gaymard, F. et al. Identification and disruption of a plant shaker-like outward channel involved in K+ release into the xylem sap. Cell 94, 647–655 (1998).Article 
    CAS 
    PubMed 

    Google Scholar 
    Han, M., Wu, W., Wu, W.-H. & Wang, Y. Potassium transporter KUP7 is involved in K+ acquisition and translocation in Arabidopsis root under K+-limited conditions. Mol. Plant 9, 437–446 (2016).Article 
    CAS 
    PubMed 

    Google Scholar 
    Qin, Y.-J., Wu, W.-H. & Wang, Y. ZmHAK5 and ZmHAK1 function in K+ uptake and distribution in maize under low K+ conditions. J. Integr. Plant Biol. 61, 691–705 (2019).Article 
    CAS 
    PubMed 

    Google Scholar 
    Drain, A. et al. Functional characterization and physiological roles of the single Shaker outward K+ channel in Medicago truncatula. Plant J. 102, 1249–1265 (2020).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Guinel, F. C. Getting around the legume nodule: I. The structure of the peripheral zone in four nodule types. Botany 87, 1117–1138 (2009).Article 

    Google Scholar 
    Livingston, D., Tuong, T., Nogueira, M. & Sinclair, T. Three-dimensional reconstruction of soybean nodules provides an update on vascular structure. Am. J. Bot. 106, 507–513 (2019).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Crespi, M. & Gálvez, S. Molecular mechanisms in root nodule development. J. Plant Growth Regul. 19, 155–166 (2000).Article 
    CAS 
    PubMed 

    Google Scholar 
    Desbrosses, G., Kopka, C., Ott, T. & Udvardi, M. K. Lotus japonicus LjKUP Is induced late during nodule development and encodes a potassium transporter of the plasma membrane. MPMI 17, 789–797 (2004).Article 
    CAS 
    PubMed 

    Google Scholar 
    Liu, J. et al. Investigate the effect of potassium on nodule symbiosis and uncover an HAK/KUP/KT member, GmHAK5, strongly responsive to root nodulation in soybean. J. Plant Biol. 65, 459–471 (2022).Article 
    CAS 

    Google Scholar 
    Iqbal, M. S., Clode, P. L., Malik, A. I., Erskine, W. & Kotula, L. Salt tolerance in mungbean is associated with controlling Na and Cl transport across roots, regulating Na and Cl accumulation in chloroplasts and maintaining high K in root and leaf mesophyll cells. Plant Cell Environ. 47, 3638–3653 (2024).Article 
    CAS 
    PubMed 

    Google Scholar 
    Kanno, S. et al. Rice Na+ absorption mediated by OsHKT2;1 affected Cs+ translocation from root to shoot under low K+ environments. Front. Plant Sci. https://doi.org/10.3389/fpls.2024.1477223 (2024).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Download referencesAcknowledgementsThis work was financially supported by the Japan Science and Technology Agency (JST), project no. JPMJSP2119 “Support for Pioneering Research Initiated by the Next Generation (SPRING)”. Additional support was provided by JSPS KAKENHI Grant-in-Aid for Scientific Research B (23K21251); JSPS KAKENHI Grant-in-Aid for Scientific Research C (26511006); Kuribayashi Scholarship Academic Foundation. Bradyrhizobium japonicum USDA110 and soybean seed (En1282 and En-b0-1) was kindly provided by the National Agriculture and Food Research Organization (NARO; formerly NIAS), Japan.Author informationAuthors and AffiliationsGraduate School of Agriculture, Hokkaido University, Sapporo, 060-8589, JapanKazuki Murashima, Hayato Maruyama, Toshihiro Watanabe & Takuro ShinanoFaculty of Food and Agricultural Sciences, Fukushima University, Fukushima, 960-1248, JapanNaoto NiheiRIKEN CSRS, Tsukuba, Ibaraki, 305-0074, JapanNao OkumaAuthorsKazuki MurashimaView author publicationsSearch author on:PubMed Google ScholarNaoto NiheiView author publicationsSearch author on:PubMed Google ScholarNao OkumaView author publicationsSearch author on:PubMed Google ScholarHayato MaruyamaView author publicationsSearch author on:PubMed Google ScholarToshihiro WatanabeView author publicationsSearch author on:PubMed Google ScholarTakuro ShinanoView author publicationsSearch author on:PubMed Google ScholarContributionsK.M.: Writing—original draft, visualization, methodology, investigation, statistical analysis. N.O.: Statistical analysis, writing—review & editing. N.N.: Investigation, writing—review & editing. H.M., T.W., & T.S.: Investigation, supervision, resources, writing—review & editing. All authors have read and agreed to the published version of the manuscript.Corresponding authorCorrespondence to
    Takuro Shinano.Ethics declarations

    Competing interests
    The authors declare no competing interests.

    Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationSupplementary Information.Rights and permissions
    Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
    Reprints and permissionsAbout this articleCite this articleMurashima, K., Nihei, N., Okuma, N. et al. Cesium accumulation in nodules is involved in mitigating cesium transfer to shoot.
    Sci Rep 15, 44449 (2025). https://doi.org/10.1038/s41598-025-28137-9Download citationReceived: 04 July 2025Accepted: 07 November 2025Published: 24 December 2025Version of record: 24 December 2025DOI: https://doi.org/10.1038/s41598-025-28137-9Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative
    KeywordsCesiumNodulePotassiumRadioisotope imagingRoot nodule symbiosis More

  • in

    Continent-wide view of genomic diversity and divergence in the wolves of Asia

    AbstractGrey wolves (Canis lupus) in Asia hold most of the species’ genetic diversity and many endangered populations. However, a clear understanding of the evolutionary history of wolves in Asia is lacking, hindering their conservation. We used 98 whole genomes of wolves across Eurasia to better resolve their evolutionary history and conservation status. The strongest barriers to gene flow coincided with boundaries separating the three major wolf lineages – Indian, Tibetan, and Holarctic. Wolves in the central Asian mountain ranges belonged to the Holarctic lineage and share little ancestry with the nearby Tibetan lineage. In contrast, wolves from eastern Asia share population-wide ancestry with the Tibetan lineage, which may reflect an unsampled lineage similar to the Tibetan lineage. Wolves from southwestern Asia share population-wide ancestry with the Indian lineage, likely due to old (>6 kya) admixture events. Long-term declines and recent inbreeding have left Indian and Tibetan wolves with some of the lowest levels of genetic diversity and highest realized genetic loads. In contrast, adjacent populations have some of the highest genetic diversity, due in part to admixture along contact zones. Our study highlights southern regions of Asia as hotspots of wolf diversity and the need to conserve these remaining populations.

    Data availability

    All raw reads are publicly available on the National Center for Biotechnology Information Sequence Read Archive under Project ID PRJNA1285574, with accession numbers SRR35174885- SRR34347608. Results and tables of the analyses can be found on GitHub: https://github.com/hennelly/Asia_wide_wolf_genomics.
    Code availability

    The bioinformatic scripts can be found on GitHub: https://github.com/hennelly/Asia_wide_wolf_genomics.
    ReferencesHogg, C. J. Translating genomic advances into biodiversity conservation. Nat. Rev. Genet. 25, 362–373 (2024).
    Google Scholar 
    Fredrickson, R. J., Siminski, P., Woof, M. & Hedrick, P. W. Genetic rescue and inbreeding depression in Mexican wolves. Proc. R. Soc. 274, 2365–2371 (2007).
    Google Scholar 
    Battilani, D. et al. Beyond population size: whole-genome data reveal bottleneck legacies in the peninsular Italian wolf. J. Hered. 116, 10–23 (2025).
    Google Scholar 
    Schweizer, R. M. et al. Genetic subdivision and candidate genes under selection in North American gray wolves. Mol. Ecol. 25, 380–402 (2015).
    Google Scholar 
    Kardos, M. et al. Genomic consequences of intensive inbreeding in an isolated wolf population. Nat. Ecol. Evol. 2, 124–131 (2018).
    Google Scholar 
    Robinson, J. A. et al. Genomic signatures of extensive inbreeding in Isle Royale wolves, a population on the threshold of extinction. Sci. Adv. 5, eaau0757 (2018).
    Google Scholar 
    Taron, U. H. et al. A sliver of the past: the decimation of the genetic diversity of the Mexican wolf. Mol. Ecol. 30, 6340–6354 (2021).
    Google Scholar 
    Lobo, D., Lopez-Bao, J. V. & Godinho, R. The population bottleneck of the Iberian wolf impacted genetic diversity but not admixture with domestic dogs: a temporal genomic approach. Mol. Ecol. 32, 5986–5999 (2023).
    Google Scholar 
    Werhahn, G., Senn, H., Macdonald, D. W. & Sillero-Zubiri, C. The diversity of genus Canis challenges conservation biology: a review of available data on Asian wolves. Front. Ecol. Evol. 10, 10.3389 (2022). 2022.
    Google Scholar 
    Wang, M. S. et al. Ancient hybridization with an unknown population facilitated high-altitude adaptation of canids. Mol. Biol. Evol. 37, 2616–2629 (2020).
    Google Scholar 
    Hennelly, L. M. et al. Ancient divergence of Indian and Tibetan wolves revealed by recombination-aware phylogenomics. Mol. Ecol. 30, 6687–6700 (2021).
    Google Scholar 
    Wang, M. S. et al. Genome sequencing of a gray wolf from Peninsular India provides new insights into the evolution and hybridization of gray wolves. Genome Biol. Evol. 2, evac012 (2022).Bergstrom, A. et al. Grey wolf genomic history reveals a dual ancestry of dogs. Nature 607, 313–320 (2022).
    Google Scholar 
    Gopalakrishnan, S. et al. Interspecific gene flow shaped the evolution of the Genus Canis. Curr. Biol. 28, P3441–P3449 (2018).
    Google Scholar 
    Hennelly, L. M. et al. Genomic analysis of wolves from Pakistan clarifies boundaries among three divergent wolf lineages. J. Heredit. 115, 339–348 (2023).Mallil, K. et al. Population genetics of the African wolf (Canis lupaster) across its range: first evidence of hybridization with domestic dogs in Africa. Mamm. Biol. 100, 645–658 (2020).
    Google Scholar 
    Niemann, J. et al. Extended survival of Pleistocene Siberian wolves into the early 20th century on the island of Honshu. iScience 24, 101904 (2020).
    Google Scholar 
    Ramos-Madrigal, J. et al. Genomes of Pleistocene Siberian wolves uncover multiple extinct wolf lineages. Curr. Biol. 31, 198–206 (2021).
    Google Scholar 
    Segawa, T. et al. Paleogenomics reveals independent and hybrid origins of two morphologically distinct wolf lineages endemic to Japan. Curr. Biol. 32, 2494–2504 (2022).
    Google Scholar 
    Marcus, J., Ha, W., Barber, R. F. & Novembre, J. Fast and flexible estimation of effective migration surfaces. eLife 10, e61927 (2021).
    Google Scholar 
    Green, R. E. et al. A draft sequence of the Neandertal genome. Science 328, 710–722 (2010).
    Google Scholar 
    Zhang, C., Rabiee, M., Sayarri, E. & Mirarab, S. ASTRAL-III: polynomial time species tree reconstruction from partially resolved gene trees. BMC Bioinform. 19, 153 (2018).
    Google Scholar 
    Burbrink, F. T., DeBaun, D., Foley, N. M. & Murphy, W.J. Recombination-aware phylogenomics. Trends Ecol. Evol. 40, 900–912 (2025).Pease, J. B. & Hahn, M. W. More accurate phylogenies inferred from low-recombination regions in the presence of incomplete lineage sorting. Evolution 67, 2376–2384 (2013).
    Google Scholar 
    Schumer, M. et al. Natural selection interacts with recombination to shape the evolution of hybrid genomes. Science 360, 565–660 (2018).
    Google Scholar 
    Martin, S., Davey, H., Salazar, C. & Jiggins, C. D. Recombination rate variation shapes barriers to introgression across butterfly genomes. PLoS Biol. 17, e2006288 (2019).
    Google Scholar 
    Schaffner, S. T. The X chromosome in population genetics. Nat. Rev. Genet. 5, 43–51 (2004).
    Google Scholar 
    Webster, T. H. & Sayres, M. A. W. Genomic signatures of sex-biased demography: progress and prospects. Curr. Opin. Genet. Dev. 31, 62–71 (2016).
    Google Scholar 
    de Jong, M. J. et al. Range-wide whole-genome resequencing of the brown bear reveals drivers of intraspecies divergence. Commun. Biol. 6, 153 (2023).
    Google Scholar 
    de Jong, M. J. et al. Red deer resequencing reveals the importance of sex chromosomes for reconstructing Late Quaternary events. Mol. Biol. Evol. 42, 1–17 (2025).
    Google Scholar 
    Nachman, M. W. & Payseur, B. A. Recombination rate variation and speciation: theoretical predictions and empirical results from rabbits and mice. Philos. Trans. R. Soc. B 367, 409–421 (2012).
    Google Scholar 
    Li, G., Figueiro, H. V., Eizirik, E. & Murphy, W. J. Recombination-aware phylogenomics reveals the structured genomic landscape of hybridizing cat species. Mol. Biol. Evol. 36, 2111–2126 (2019).
    Google Scholar 
    vonHoldt, B. M. et al. Persistence and expansion of cryptic endangered red wolf genomic ancestry along the American Gulf coast. Mol. Ecol. 31, 5440–5454 (2021).
    Google Scholar 
    Feng, C., Wang, J., Liston, A. & Kang, M. Recombination variation shapes phylogeny and introgression in wild diploid strawberries. Mol. Biol. Evol. 40, msad049 (2023).
    Google Scholar 
    Jiang, Z. et al. Gene flow and an anomaly zone complicate phylogenomic inference in a rapidly radiated avian family (Prunellidae). BMC Biol. 22, 49 (2024).
    Google Scholar 
    Monthey, J. D. & Spellman, G. M. Recombination rate variation shapes genomic variability of phylogeographic structure in a widespread North American songbird (Aves: Certhia americana). Mol. Phylogenet. Evol. 196, 108088 (2024).
    Google Scholar 
    Excoffier, L., Dupanloup, I., Huerta-Sánchez, E., Sousa, V. C. & Foll, M. fastsimcoal2: Demographic inference under complex evolutionary scenarios. Bioinformatics 37, 4842–4849 (2021).
    Google Scholar 
    Li, H. & Durbin, R. Inference of human population history from individual whole-genome sequences. Nature 475, 493–496 (2011).
    Google Scholar 
    Mazet, O., Rodriguez, W., Grusea, S., Boitard, S. & Chikhi, L. On the importance of being structured: instantaneous coalescence rates and human evolution—lessons from ancestral population size inference?. Heredity 116, 362–371 (2015).
    Google Scholar 
    Hanghoj, K., Molke, I., Andersen, P. A., Manica, A. & Korneliussen, T. S. Fast and accurate relatedness estimation from high-throughput sequencing data in the presence of inbreeding. GigaScience 8, 1–9 (2019).Narasimhan, V. et al. BCFtools/RoH: a hidden Markov model approach for detecting autozygosity from next-generation sequencing data. Bioinformatics 32, 1749–2751 (2016).
    Google Scholar 
    Thompson, E. A. Variation in meiosis, across genomes, and in populations. Genetics 194, 301–326 (2013).
    Google Scholar 
    Gomez-Sanchez, D. et al. On the path to extinction: inbreeding and admixture in a declining grey wolf population. Mol. Ecol. 27, 3599–3612 (2018).
    Google Scholar 
    Bertorelle, G. et al. Genetic load: genomic estimates and applications in non-model animals. Nat. Rev. Genet. 23, 492–503 (2022).
    Google Scholar 
    Cingolani, P. et al. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff. Fly 6, 80–92 (2012).Heffelfinger, J. R., Nowak, R. M. & Paetkau, D. Clarifying historical range to aid recovery of the Mexican wolf. J. Wildl. Manag. 81, 766–777 (2017).
    Google Scholar 
    Mathur, S. & DeWoody, J. A. Genetic load has potential in large populations but is realized in small inbred populations. Evol. Appl. 14, 1540–1557 (2021).
    Google Scholar 
    Dussex, N., Morales, H. E., Grossen, C., Dalen, L. & Oosterhout, C. Purging and accumulation of genetic load in conservation. Trends Ecol. Evol. 38, P961–P969 (2023).
    Google Scholar 
    Smeds, L. & Ellegren, H. From high masked to high realized genetic load in inbred Scandinavian wolves. Mol. Ecol. 32, 1567–1580 (2022).
    Google Scholar 
    Iurino, D. A. et al. A Middle Pleistocene wolf from central Italy provides insights on the first occurrence of Canis lupus in Europe. Sci. Rep. 12, 2882 (2022).Sotnikova, M. & Rook, L. Dispersal of the Canini (Mammalia, Canidae: Caninae) across Eurasia during the Late Miocene to Early Pleistocene. Q. Int.212, 86–97 (2010).Gelabert, P. et al. Genome-scale sequencing and analysis of human, wolf, and bison DNA from 25,000-year-old sediment. Curr. Biol. 31, 3564–3574 (2021).
    Google Scholar 
    Hibbins, M. S. & Hahn, M. W. Phylogenomic approaches to detecting and characterizing introgression. Genetics 220, iyab173 (2021).
    Google Scholar 
    Wang, G. D. et al. Genomic approaches reveal an endemic subpopulation of gray wolves in Southern China. iScience 20, 110–118 (2019).
    Google Scholar 
    Wang, L. et al. The geographical distribution of grey wolves (Canis lupus) in China: a systematic review. Zool. Res. 37, 315–326 (2016).
    Google Scholar 
    Pinxian, W. & Xiangjun, S. Last glacial maximum in China: comparison between land and sea. Catena 23, 341–535 (1994).
    Google Scholar 
    Dayan, T., Simberloff, D., Tchernov, E. & Yom-Tov, Y. Canine car-nassial: character displacement in the wolves, jackals, and foxes ofIsrael. Biol. J. Linn. Soc. 45, 313–331 (1992).
    Google Scholar 
    Kurten, B. The carnivora of the Palestine caves. Acta Zool. Fenn. 107, 74 (1965).
    Google Scholar 
    Mashkour, M. et al. Carnivores and their prey in the Wezmeh Cave (Kermanshah, Iran): a Late Pleistocene refuge in the Zagros. Int. J. Osteoarchaeol. 19, 678–694 (2008).
    Google Scholar 
    Plessis, S. J., Blaxter, M., Koepfli, K. P., Chadwick, E. A. & Hailer, F. Genomics reveals complex population history and unexpected diversity of Eurasian otters (Lutra lutra) in Britain relative to genetic methods. Mol. Biol. Evol. 40, msad207 (2023).
    Google Scholar 
    Statham, M. J. et al. Range-wide multilocus phylogeography of the red fox reveals ancient continental divergence, minimal genomic exchange, and distinct demographic histories. Mol. Ecol. 23, 4813–4830 (2014).
    Google Scholar 
    Alvares, F. et al. Old World Canis spp. with taxonomic ambiguity: workshop conclusions and recommendations. Canid News 21 http://hdl.handle.net/10138/327703 (2019).Krofel, M., Hatlauf, J., Bogdanowicz, W., Cambell, L. A. D. & Godinho, R. Towards resolving taxonomic lineages in wolf, dog, and jackal of Africa, Eurasia, and Australasia. J. Zool. 316, 155–168 (2022).
    Google Scholar 
    Sillero-Zubiri, C. Family Canidae. In D.E. Wilson and R.A. Mittermeier (Eds.), The Handbook of the Mammals of the World (Lynx Edicions in association with Conservation International and IUCN, 2009).Sharma, D. K., Maldonado, J. E., Jhala, Y. V. & Fleischer, R. C. Ancient wolf lineages in India. Proc. R. Soc. B 271, S1–S4 (2004).
    Google Scholar 
    Werhahn, G. et al. Himalayan wolf distribution and admixture based on multiple genetic markers. J. Biogeogr. 47, 1272–1285 (2020).
    Google Scholar 
    Jhala, Y. V., Saini, S., Kumar, S. & Qureshi, Q. Distribution, status, and conservation of the Indian Peninsula wolf. Front. Ecol. Evol. 10, 814966 (2022).
    Google Scholar 
    Sheikh, K. M. & Molur, S. Status and Red List of Pakistan’s Mammals. Based on the Conservation Assessment and Management Plan 312 (IUCN Pakistan, 2003).Hennelly, L. M. et al. Canis lupus ssp. pallipes (The IUCN Red List of Threatened Species, 2025).Werhahn, G. et al. Himalayan Wolf: Canis lupus spp. chanco (The IUCN Red List of Threatened Species, 2023).Kardos, M. et al. The crucial role of genome-wide genetic variation in conservation. Proc. Natl. Acad. Sci. USA 118, e2104642118 (2021).
    Google Scholar 
    Sowerby, A. C. China’s Natural History: A Guide to the Shanghai Museum (Royal Asiatic Society of Great Britain and Ireland, North China Branch, 1936).Fellowes, J. R., Chan, B. P. L., Lau, C. M. N., Sai-Chit, N. & Siu, G. L. P. 2003. Report of Rapid Biodiversity Assessment at Shiwandashan National Nature Reserve and National Forest Park, Southwest Guanxi, China, 2000–2001 (Kadoorie Farm and Botanic Garden in collaboration with Guangxi Forestry Department, Guangxi Institute of Botany, South China Normal University, 2003).Bonsen, G. T. et al. Navigating complex geopolitical landscapes: challenges in conserving the endangered Arabian wolf. Biol. Conserv. 296, 110655 (2024).
    Google Scholar 
    Caroe, C. et al. Single-tube library preparation for degraded DNA. Methods Ecol. Evol. 9, 410–419 (2017).
    Google Scholar 
    Mak, S. S. T. et al. Comparative performance of the BGISEQ-500 vs Illumina HiSeq2500 sequencing platform for paleogenomic sequencing. GigaScience 6, 1013 (2017).
    Google Scholar 
    Gilbert, M. T. P. et al. Whole-genome shotgun sequencing of mitochondria from ancient hair shafts. Science 317, 1927–1930 (2007).
    Google Scholar 
    Dabney, J. et al. Complete mitochondrial genome sequence of a Middle Pleistocene cave bear reconstructed from ultrashort DNA fragments. Proc. Natl. Acad. Sci. USA 110, 15758–15763 (2013).
    Google Scholar 
    Allentoft, M. E. et al. Population genomics of Bronze Age Eurasia. Nature 522, 167–172 (2015).
    Google Scholar 
    Schubert, M. et al. Characterization of ancient and modern genomes by SNP detection and phylogenomic and metagenomic analysis using PALEOMIX. Nat. Protoc. 9, 1056–1082 (2014).
    Google Scholar 
    Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics 25, 1754–1760 (2009).
    Google Scholar 
    McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 20, 1297–1303 (2010).
    Google Scholar 
    Meisner, J. & Albrechtsen, A. Inferring population structure and admixture proportions in low-depth NGS data. Genetics 210, 719–731 (2018).
    Google Scholar 
    Skotte, L., Korneliussen, T. S. & Albrechtsen, A. Estimating individual admixture proportions from Next Generation sequencing data. Genetics 3, 693–702 (2013).
    Google Scholar 
    Korneliussen, T. S., Albrechtsen, A. & Nielsen, R. ANGSD: analysis of next generation sequencing data. BMC Bioinform. 15, 356 (2014).
    Google Scholar 
    Danecek, P. et al. The variant call format and VCFtool. Bioinformatics 27, 2156–2158 (2011).
    Google Scholar 
    Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
    Google Scholar 
    Kalyaanamoorthy, S., Minh, B. Q., Wong, T. K. F., Haeseler, A. & Jermiin, L. S. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat. Methods 14, 587–589 (2017).
    Google Scholar 
    Nguyen, L. T., Schmidt, H. A., Haeseler, A. & Minh, B. Q. IQTREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol. Biol. Evol. 32, 268–274 (2014).
    Google Scholar 
    Auton, A. et al. Genetic recombination is targeted towards gene promoter regions in dogs. PLoS Genet. 9, e1003984 (2013).
    Google Scholar 
    Smeds, L. et al. Whole-genome analysis provide no evidence for dog introgression in Fennoscandian wolf populations. Evol. Appl. 14, 721–734 (2021).
    Google Scholar 
    Patterson, N. et al. Ancient admixture in human history. Genetics 192, 1065–1093 (2012).
    Google Scholar 
    Martin, S. H. & Belleghem, S. M. V. Exploring evolutionary relationships across the genome using topology weighting. Genetics 1, 429–438 (2017).
    Google Scholar 
    Delaneau, O., Howie, B., Cox, A. J., Zagury, J. F. & Marchini, J. Haplotype estimation using sequencing reads. AJGH 93, P687–P696 (2013).
    Google Scholar 
    Danecek, P. et al. Twelve years of SAMtools and BCFTools. GigaScience 10, giab008 (2021).
    Google Scholar 
    Hilgers, L. et al. Avoidable false PSMC population size peaks occur across numerous studies. Curr. Biol. 35, 927–930 (2025).
    Google Scholar 
    Koch, E. M. et al. De novo mutation rate estimation in wolves of known pedigree. Mol. Biol. Evol. 36, 2536–2547 (2019).
    Google Scholar 
    Mech, D. L. & Barber-Meyer, S. Use of erroneous wolf generation time in assessments of domestic dog and human evolution. Sci. Lett. http://science.sciencemag.org/content/352/6290/1228/tab-e-letters (2017).Download referencesAcknowledgementsL.M.H. thanks the National Science Foundation Postdoctoral Research Fellowship (award number 2208950) for funding and support. The Norwegian Environment Agency (project 18088069) provided funding and support for sequencing efforts on newly sequenced wolf genomes. Ç.H.Ș. thanks to Fondation Segré and the Sigrid Rausing Trust for providing the majority of the funding for this project, H. Batubay Özkan and Barbara Watkins for their support of the Biodiversity and Conservation Ecology Lab at the University of Utah, and Bilge Bahar, Seha İşmen, Ömer Koç, Ömer Külahçıoğlu, Burak Över, Emin Özgür, Suna Reyent, and Ceren Sağlamer for supporting this project. Türkiye’s Department of Nature Conservation and National Parks and the Ministry of Agriculture and Forestry granted the permit for Türkiye (No. 72784983-488.04-114100). P.H. thanks the Technology Agency of the Czech Republic (project SS07010447) for support. We thank the Museum of the Institute of Plant and Animal Ecology UB RAS for access to their collections.Author informationAuthors and AffiliationsSection for Evolutionary Hologenomics, Globe Institute, University of Copenhagen, Copenhagen, DenmarkLauren M. Hennelly, Bárbara R. Parreira, Camilla H. Scharff-Olsen, Xin Sun, Nuno Filipes Gomes Martins, M. Thomas P. Gilbert & Shyam GopalakrishnanCenter for Conservation Genomics, Smithsonian’s National Zoo and Conservation Biology Institute, Washington, DC, USALauren M. HennellyMammalian Ecology and Conservation Unit, Veterinary Genetics Laboratory, Department of Population Health and Reproduction, School of Veterinary Medicine, University of California, Davis, CA, USALauren M. Hennelly & Benjamin N. SacksDepartment of BioSciences, Rice University, Houston, TX, USALauren M. HennellyOrganisms and Environment, School of Biosciences, Cardiff University, Cardiff, UKAsh Noble & Frank HailerDepartment of Molecular Biology and Genetics, Koç University, Istanbul, TürkiyeM. Çisel Kemahlı Aytekin & Çağan H. ŞekercioğluSchool of Biological Sciences, University of Utah, Salt Lake City, UT, USAÇağan H. ŞekercioğluKuzey Doğa Society, Kars, TürkiyeÇağan H. ŞekercioğluUral Federal University, Yekaterinburg, RussiaPavel KosintsevInstitute of Plant and Animal Ecology, Ural Branch of Russian Academy of Science, Yekaterinburg, RussiaPavel KosintsevTechnical University in Zvolen, Zvolen, SlovakiaLadislav PauleDepartment of Zoology, Charles University, Prague, CzechiaPavel HulvaDepartment of Biology and Ecology, University of Ostrava, Ostrava, CzechiaPavel HulvaNTNU University Museum, Norwegian University of Science and Technology, Trondheim, NorwayHans K. Stenøien & M. Thomas P. GilbertWildlife Institute of India, Dehradun, IndiaBilal HabibUniversity of Education – Attock Campus, Attock, PakistanHira FatimaUniversity of the Punjab, Lahore, PakistanGhulam SarwarLebanese Wildlife, Jitta, El Metn, LebanonSamara P. El-Haddad & Alexandra YoussefCardiff University-Institute of Zoology Joint Laboratory for Biocomplexity Research (CIBR), Beijing, ChinaFrank HailerDepartment of Biology, University of Copenhagen, Copenhagen, DenmarkMikkel-Holger S. SindingAuthorsLauren M. HennellyView author publicationsSearch author on:PubMed Google ScholarBárbara R. ParreiraView author publicationsSearch author on:PubMed Google ScholarAsh NobleView author publicationsSearch author on:PubMed Google ScholarCamilla H. Scharff-OlsenView author publicationsSearch author on:PubMed Google ScholarM. Çisel Kemahlı AytekinView author publicationsSearch author on:PubMed Google ScholarÇağan H. ŞekercioğluView author publicationsSearch author on:PubMed Google ScholarPavel KosintsevView author publicationsSearch author on:PubMed Google ScholarLadislav PauleView author publicationsSearch author on:PubMed Google ScholarPavel HulvaView author publicationsSearch author on:PubMed Google ScholarHans K. StenøienView author publicationsSearch author on:PubMed Google ScholarBilal HabibView author publicationsSearch author on:PubMed Google ScholarHira FatimaView author publicationsSearch author on:PubMed Google ScholarGhulam SarwarView author publicationsSearch author on:PubMed Google ScholarSamara P. El-HaddadView author publicationsSearch author on:PubMed Google ScholarAlexandra YoussefView author publicationsSearch author on:PubMed Google ScholarFrank HailerView author publicationsSearch author on:PubMed Google ScholarXin SunView author publicationsSearch author on:PubMed Google ScholarNuno Filipes Gomes MartinsView author publicationsSearch author on:PubMed Google ScholarM. Thomas P. GilbertView author publicationsSearch author on:PubMed Google ScholarBenjamin N. SacksView author publicationsSearch author on:PubMed Google ScholarMikkel-Holger S. SindingView author publicationsSearch author on:PubMed Google ScholarShyam GopalakrishnanView author publicationsSearch author on:PubMed Google ScholarContributionsL.M.H. conceived and designed the project with guidance from S.G., M.H.S.S., and B.N.S. C.S.O., M.H.S.S., and N.F.G.M. conducted laboratory work; H.F., G.S., B.H., F.H., S.P.E.H., C.K., Ç.H.Ş., P.K., H.S., M.H.S.S., L.P., P.H., A.Y., and M.T.P.G. provided logistics, field work, wolf samples, and sequencing efforts; L.M.H. led data analysis with assistance from B.R.P., A.N., X.S., N.F.G.M., and L.M.H. wrote the manuscript with input from all co-authors.Corresponding authorCorrespondence to
    Lauren M. Hennelly.Ethics declarations

    Competing interests
    The authors declare no competing interests.

    Peer review

    Peer review information
    Communications Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Yi-Jyun Luo and Michele Repetto.

    Additional informationPublisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary informationSupplementary InformationDescription of Additional Supplementary MaterialsSupplementary Data 1Reporting SummaryRights and permissions
    Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
    Reprints and permissionsAbout this articleCite this articleHennelly, L.M., Parreira, B.R., Noble, A. et al. Continent-wide view of genomic diversity and divergence in the wolves of Asia.
    Commun Biol (2025). https://doi.org/10.1038/s42003-025-09379-9Download citationReceived: 12 February 2025Accepted: 04 December 2025Published: 24 December 2025DOI: https://doi.org/10.1038/s42003-025-09379-9Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative More

  • in

    Coupling models to assess impacts of land and carbon changes on sustainable ecological safety networks of Gui’an New Area, China

    AbstractGui’an New Area, the largest state-level new area in western China, serves as a hub for economic growth and a demonstration area for ecological civilization in China. Its sustainable development heavily relies on ecological protection. In this study, we explored the impacts on the regional ecosystem by analyzing the land use change, carbon storage (CS) dynamics, and ecological safety network pattern of Gui’an New Area during the period from 2010 to 2060. The study utilizes the FLUS model to predict land use in 2030 and 2060, applies the InVEST model to assess CS, and combines the MSPA and MCR models to construct an ecological safety network. The results show that policies drive land use types and landscape spatial changes in Gui’an New Area. The land use types of Gui’an New Area shifted significantly from 2010 to 2060, especially the decrease of cropland and the growth of buildings, as well as the fluctuating changes around woodland, grassland, and water; the core area of the landscape also showed a decreasing trend in the area share between 2010 and 2060. Gui’an New Area’s CS displayed an overall decline from 2014 to 2060, despite an initial increase until 2030; this trend showed significant spatial heterogeneity, with woodland and building areas undergoing the most substantial changes due to variations in ecological space area and carbon density. The analysis of the ecological safety network shows that the number and area of ecological source land in Gui’an New Area fluctuated and decreased between 2010 and 2060; the number of ecological corridors declined as a whole; the spatial distribution was uneven; and the ecological space in the eastern part of the area was compressed by the influence of economic development and the growth of population density, which limited the formation and development of the ecological corridors. The study emphasizes that regional CS and ecosystem services can be enhanced through rational planning and ecological restoration. It is recommended that ecological space protection and long-term management be strengthened to achieve a win-win situation between ecological protection and economic development in Gui’an New Area and to promote sustainable development.

    Similar content being viewed by others

    Multi-scenario simulation of land use change based on the objectives of cultivated land, ecological protection, and economic development in Yunnan Province, China

    Article
    Open access
    27 October 2025

    Multi-scenario simulation of carbon stock and landscape ecological risk changes in Jinpu new area and analysis of spatial conflict relationships

    Article
    Open access
    01 July 2025

    Prediction of land use for the next 30 years using the PLUS model’s multi-scenario simulation in Guizhou Province, China

    Article
    Open access
    07 June 2024

    IntroductionThe management and assessment of territorial spatial changes and carbon storage(CS) are essential for alleviating the greenhouse effect and attaining carbon neutrality1,2. Urbanization, an inevitable trend in societal progression3, not only transforms land use and cover types but also significantly impacts carbon reserves and the ecological safety framework4,5. This process exacerbates landscape fragmentation, resulting in habitat area reduction, diminished landscape connectivity, lower biodiversity, and compromised ecosystem health and service functions3,5,6,7. Land use transitions can influence ecosystem service functions, particularly key aspects such as vegetation change and carbon sequestration8,9,10. Consequently, examining land use alterations and terrestrial ecosystem CS assessments is vital for developing ecological safety patterns that enhance regional carbon sinks, thereby positively affecting natural ecosystems.Advanced modeling techniques such as the Future Land Use Simulation (FLUS), Patch Generating Land Use Simulation (PLUS), Integrated Valuation of Ecosystem Services and Tradeoffs (InVEST), Circuit Theory (CT), and Minimum Cumulative Resistance (MCR) have been crucial in comprehending the intricate interactions among land use types, CS, and ecological safety across various landscapes in China3,11,12,13,14. These models have not only enhanced our predictive capabilities but have also improved our understanding of ecological risks and the optimization of ecological networks. For instance, the application of ecological carrying capacity assessments in the Southwest Guangxi Karst – Beibu Gulf region has been instrumental in constructing ecological security patterns11, while in Xinjiang, the InVEST-Conefor-MCR model has further strengthened these patterns with robust scientific backing12. These studies have collectively improved our predictive capabilities and understanding of ecological risks and the optimization of ecological networks. The synergistic application of FLUS and InVEST models has been particularly effective for an in-depth evaluation of land use changes and CS dynamics, owing to their efficiency and practical utility. For example, research utilizing the Markov-PLUS-InVEST model in the Sichuan and Chongqing urban agglomeration under complex terrain conditions has illustrated how urban expansion into undeveloped mountainous areas can decelerate CS growth and enhance spatial heterogeneity3. Urban growth patterns in Hangzhou have similarly been associated with reduced habitat quality(HQ)15. The FLUS-InVEST model has been particularly enlightening in the Chengdu-Chongqing urban agglomeration, revealing the impact of land use change on CS13, and similar advancements have been noted in the Pearl River Delta, where ecological protection scenarios have significantly improved HQ and CS16. Even with these contributions, a cohesive framework is still needed to integrate the temporal and spatial variability of land use change and CS assessments within ecological safety networks. This gap is apparent in the use of models on complex terrains like the Hohhot-Baotou-Ordos-Yulin urban area17, in studies along China’s swiftly urbanizing coastal areas18, and in the Qinghai Lake Basin19, and Guilin20, highlighting the need for a more unified approach. Efforts to optimize ecological protection in southern China’s hilly regions21 and to simulate future land use scenarios by coupling ecological security patterns with multiple scenarios22 have advanced the field. However, constructing a comprehensive ecological safety network, as highlighted by research focusing on Shenzhen’s urban ecological security pattern23, which proposes various development zone patterns based on HQ assessment and the MCR model, remains a challenge. A synthesis of these studies reveals a research gap: while individual models have been applied with success, an integrated approach that harnesses their collective power to construct a predictive framework for ecological safety networks, accounting for the intricate interplay between land use changes, CS, and ecological dynamics, is still required. However, while individual models have contributed valuable insights, there is an urgent need for an integrated approach that harnesses the collective power of these models.To address this research gap, this study proposes an integrated framework coupling the FLUS, InVEST, and MCR models. This synergistic approach leverages the predictive power of FLUS to generate future land use scenarios, providing essential inputs for InVEST to quantify the resulting spatial and temporal dynamics of CS, thereby explicitly linking land use transitions to impacts on this critical ecosystem service. Crucially, the MCR model utilizes these future-oriented land use patterns and associated ecosystem service assessments (e.g., HQ/CS from InVEST) to construct ecological safety networks that inherently account for anticipated changes21. Although both Li et al.16 and Zhang et al.3 have verified the efficiency and universal potential of FLUS-InVEST in land use simulation and CS assessment, their research frameworks have not been systematically coupled with the construction of ecological safety networks, leaving a gap in relevant integrated studies. Consequently, this sequential FLUS-InVEST-MCR workflow forms a holistic, closed-loop framework that predicts land use drivers, evaluates ecological service functions (e.g., CS), and designs adaptive spatial solutions (e.g., eco-network). This extends existing integrated approaches by explicitly linking projected CS to dynamic eco-network design. It uniquely enables the construction of forward-looking ecological safety networks grounded in projected ecosystem service status and responsive to anticipated land use transformations, directly addressing the identified need for a cohesive framework integrating land use change, CS, and ecological network dynamics.Applying this integrated framework, this study focuses on the Gui’an New Area, as the eighth state-level new area approved by the State Council in 2014, assigned the important roles of economic growth pole in the western region, new inland open economy highway, and ecological civilization demonstration area24. However, under the special karst natural environment of the new area, its construction and development face serious challenges in terms of ecological safety. Therefore, in the context of the “dual-carbon” goal and the strategy of ecological civilization construction, we utilize the coupled FLUS-InVEST-MCR approach to conduct a comprehensive assessment of the land use change and CS in Gui’an New Area and to construct an ecological safety network. The study addresses the following three scientific questions: (1) Analyze the trend of land use change in Gui’an New Area from 2010 to 2022 and predict the change characteristics of land use in 2030 versus 2060 under the ecological protection scenario. (2) Evaluate the spatial and temporal differentiation of CS in Gui’an New Area using the InVEST model, as well as the differences in CS among different land use types. (3) Construct and analyze the ecological safety network pattern of Gui’an New Area at different development stages.Materials and methodsStudy areaGui’an New Area is China’s eighth state-level new area, established in 2014, located in the combined area of Guiyang City and Anshun City (Fig. 1), Guizhou Province, with a planned area of 1,901 square kilometers and a current resident population density of about 404,000 people. Since its conception in 2011, it has developed rapidly with the support of the national and provincial governments and is positioned as an economic growth pole in the west, a new inland open economy highland, and an ecological civilization demonstration area24. Meanwhile, it is strongly supported by policies, and Gui’an New Area has been approved as a number of national experimental demonstration zones, which provides a strong guarantee for development25. In terms of the economy, the GDP of the direct administration area increased from 5.23 to 22.913 billion yuan between 2014 and 2023, with an annual average growth rate of 18.2%. The new zone focuses on modern industrial clusters such as big data, high-end electronic information manufacturing, high-end specialty equipment manufacturing, cultural tourism, and health and high-end service industries and has attracted large-scale data centers, including the three major carriers, Huawei, Tencent, and Apple. Since the establishment of the new area, a series of public service facilities have been built, the integration of urban and rural water supply has achieved full coverage, and the urbanization rate has reached 76%.In terms of the ecological environment, the area is characterized by its karst landscape and exhibits a diverse vegetation cover. The predominant natural vegetation types include subtropical evergreen broad-leaved woodland, mixed evergreen and deciduous broad-leaved forests, and secondary shrubland and grassland resulting from historical land use changes. Key native tree species contributing significantly to the regional carbon pool include Cyclobalanopsis glauca, Castanopsis spp, Pinus massoniana, and Cunninghamia lanceolata, particularly in forest areas. Additionally, significant areas are covered by bamboo forests (Phyllostachys spp.) and various fruit orchards26. The notable 42% forest coverage provides a substantial foundation for carbon sequestration. Therefore, in the plan, it is expected that the urban population density will reach 2 million by 2030 and the land for urban construction will be controlled at 220 square kilometers, further promoting regional economic and social development and ecological civilization.Fig. 1Study area (This figure was obtained by using ArcGIS 10.8 through the open-access data process.).Full size imageData sources and processingThe data used in this study include land use data, climate and environmental data, and socioeconomic data (Table 1). Among them, the land use data were obtained from Wuhan University 30-meter-resolution Chinese surface cover data27. The data were merged into six primary categories: cropland, woodland, grassland, water, building land, and unused land. This establishes a unified benchmark dataset for characterizing the pattern of land cover changes and supporting CS assessment based on the InVEST model. Climate and environmental data include elevation, slope, NDVI, mean annual precipitation, mean annual temperature data, soil moisture28, and rocky desertification index. Among them, we coupled the carbonate content with 30 m DEM and NDVI data, calculated the rock exposure rate, and finally obtained the pixel scale rocky desertification index. Socio-economic data included roads, water system, population density, GDP, nighttime lights index29, and so on. Highway, main road, secondary road, and water system data for each period were extracted from the road network as the main research object. The annual average temperature data is based on the monthly average temperature raster data with 1 km by 1 km resolution and the annual average temperature raster obtained by calculating the average of the 12 months of the year’s month-by-month average temperature raster. The monthly average precipitation is based on the monthly average precipitation raster data at 0.1° × 0.1° resolution and is obtained by calculating the average of the 12-month average precipitation using the raster calculation tool.Table 1 Sources of data.Full size tableMethodsThe objective of this study is to investigate the impact of land-use changes at various temporal and spatial scales on landscape patterns, CS functions, and ecological safety networks in the Gui’an New Area. To this end, we analyzed land-use changes in the Gui’an New Area before and after its construction (in the years 2010, 2014, 2018, and 2022) and predicted the land-use trends for the next two periods (2030 and 2060). Concurrently, we conducted a quantitative analysis of the landscape patterns and CS during these periods. Based on these analyses, we constructed the corresponding ecological safety network patterns, aiming to provide a scientific basis for future ecological restoration and protection strategies in the Gui’an New Area. Figure 2 illustrates the specific structure of this technical framework.Fig. 2Research Technology Roadmap (Created by ArcGIS 10.8 and Microsoft office PowerPoint.).Full size imageFLUS modelWith cellular automaton as its theoretical core, the FLUS model can capture the nonlinear coupling relationship between multi-source driving factors and land use types2,30. The Gui’an New Area has witnessed rapid urban expansion and significant land changes, and the use of data from adjacent years can enable the model to more truly reproduce the evolution trajectory. The simulation process is divided into three steps: (1) The occurrence probability module based on Artificial Neural Network (ANN). Ten categories of driving factors such as elevation, slope, annual average precipitation, annual average temperature, distance to water, population density, GDP, distance to highway, distance to main road, and distance to secondary road were selected, and a transfer cost matrix was defined based on existing studies2 (Table 2). (2) Introduce an adaptive inertia competition mechanism, coupling the conversion rules with neighborhood weights. Neighborhood weight settings: Cropland, 0.2; Woodland, 1.0; Grassland, 0.9; Water, 0.6; Other land, 0.1; Building, 0.4; so as to depict the expansion/shrinkage behaviors of different land use types13. (3) Use the FLUS model to predict the future land use pattern. The relevant calculation formulas are as described in the studies by Li et al.2 and Zhang et al.30.This study takes the land use data of Gui’an New Area in 2018 as the benchmark to simulate the land use pattern in 2022, and verifies it by comparing with the actual data31. After ensuring the overall accuracy, the land demand of 2030 output by the System Dynamics (SD) model is embedded into the FLUS framework. Starting from 2022, it further predicts the changes in spatial distribution in 2030 and 2060 under the principle of ecological priority.Table 2 Transfer cost matrix for eco-priority scenarios.Full size tableLandscape pattern analysisBased on the raster data of land use types from 2010 to 2060, woodland, grassland, and water were extracted as foreground elements of the MSPA, and other land as background. Based on the Guidos Toolbox 2.8 software platform, seven landscape elements such as core area, bridge, islet, loop, edge, branch, and perforation were identified, and finally, the core area of the landscape types, which is important for maintaining connectivity, was extracted as the landscape element for the later connectivity analysis (Fig. 2).Assessment of CS changesThe estimation of CS in the InVEST model is based on the alternative method of pools. In this paper, the average carbon densities of above-ground carbon pools, below-ground carbon pools, soil carbon pools, and dead organic matter carbon pools of various land classes were counted according to the land use situation, and then the carbon densities were multiplied by the area of each land class to obtain the CS of four major pools, and the total value of CS in the whole study area could be obtained by accumulating the CS of the four major pools3. The calculation formula is as follows:$$:C_total=C_above+C_below+C_soil+C_dead$$
    (1)
    Eq. C_total represents the total CS. C_above is the aboveground CS. C_below is the subsurface CS. C_soil is the soil CS. C_dead is the dead organic CS.The carbon density parameters of different land covers were mainly referred to the relevant studies in Guizhou or the Southwest Karst region3,32, which are shown in Table 3.Table 3 Reference values of land use carbon density in the study area (Mg/hm2).Full size tableEcological safety network construction(1) Ecological source identification: In this study, important habitat patches in the study area were identified as potential ecological sources using the MSPA model and landscape connectivity analysis method21,23. Three landscape indices of overall connectivity, possible connectivity, and patch importance were selected, and the threshold of patch connectivity distance was set to 40,000 m and the probability of connectivity was set to 0.5 to evaluate the landscape connectivity of the core area. The patches with the area of the core area larger than 2.5 km2 and the value of dPC larger than 2.5 were considered ecological sources to avoid the subjectivity of ecological source selection. The specific calculation formula is as follows:$$:IIC=frac{sum:_{i=1}^{n}sum:_{j=1}^{n}frac{{a}_{i}:{:a}_{j}}{1+n{l}_{ij}}}{{A}_{L}^{2}}$$
    (2)
    $$:PC=frac{sum:_{i=1}^{n}sum:_{j=1}^{n}{p}_{ij:}^{*}:{a}_{i}:{a}_{j}}{{A}_{L}^{2}}$$
    (3)
    $$:dI=frac{I-{I}_{remove}}{I}times:100text{%}$$
    (4)
    Ea. denotes the total number of patches in the landscape. (:{a}_{i:})and (:{a}_{j}) denote the area of patch i and patch j, respectively. (:{nl}_{ij}:)denotes the connection between patch i and patch j. (:{A}_{L}) is the total area of the landscape, and (:{P}_{ij}^{text{*}}) is the maximum likelihood of direct dispersal of species in patches i and j. (:I) is the value of the connectivity index for a given landscape, which in this paper refers to the index of overall connectivity (IIC) and the index of possible connectivity (PC); the (:{I}_{remove}) is the value of the connectivity index of the landscape after removing patch i from this landscape.(2) Resistance surface construction: based on the karst characteristics of Gui’an New Area, seven factors were selected by synthesizing the natural environment, ecological resources, and economic society, as shown in Table 4. Referring to the existing research33, the corresponding resistance values were assigned to different land use types, and the resistance coefficients of cropland, woodland, grassland, water, other land, and buildings were 30, 1, 5, 10, 25, and 100, respectively. The AHP method was then used to calculate the weights of the factors (λmax = 7.52209, CI = 0.0870153, CR = 0.0639819 < 0.1, passed the consistency test) to get the ecological resistance factor weights of Gui’an New Area (Table 4). Completed by spatial superposition analysis resistance surface construction (Fig. 3).Table 4 Classification and weighting of ecological resistance factors.Full size tableFig. 3Characteristics of the spatial distribution of normalized ecological resistance factors. (a): land use; (b): elevation; (c): slope; (d): soil moisture; (e): nighttime lights index; (f): NDVI; (g): rocky desertification index; (h): combined resistance surface. (This figure was obtained by using ArcGIS 10.8 through the open-access data process.)Full size image(3) Ecological corridor identification: this paper adopts the Minimum Cumulative Resistance (MCR) model to extract the ecological corridors in Gui’an New Area. The basic formula is as follows:$$:MCR=fminsum:_{j=n}^{i=m}{D}_{ij}times:{R}_{i}$$
    (5)
    Where: MCR is the minimum cumulative resistance value;(:{D}_{ij}) is the spatial distance of the species from source j to landscape unit i; (:{R}_{i:})is the coefficient of resistance of landscape unit i to the movement of a species; f denotes the positive correlation between the minimum cumulative resistance and the ecological process. And based on the gravity model, the interaction matrix between source sites was constructed to quantitatively evaluate the interaction strength between habitat patches to scientifically determine the relative importance of potential ecological corridors. Based on the results of the matrix and combining them with the actual situation of the study area, the corridors with interaction strengths greater than 3000 were extracted as important corridors, and the others were treated as general corridors, and the ecological network map of the study area was obtained.The gravity model can calculate the interaction matrix between ecological source sites, and the higher the interaction force between two source sites, the more important the ecological corridor between two source sites is in the ecological service system of the study area. The calculation formula is as follows:$$:F=frac{{N}_{i}{N}_{j}}{{D}_{ij}^{2}}=frac{{L}_{max}^{2}{ln}left({S}_{i}right){ln}left({S}_{j}right)}{{L}_{ij}^{2}{P}_{i}{P}_{j}}$$
    (6)
    Where, (:{F}_{ab}) is the interaction force between the source sites a & b. (:{N}_{i}), (:{N}_{j}) is the weight value of i, j; (:{D}_{ij}) is the resistance value of the potential corridor between source sites i and j; (:{L}_{max}^{:}) is the maximum resistance value of the potential corridor in the study area; (:{S}_{i}), (:{S}_{j})are the areas of i and j, respectively; (:{L}_{ij}) is the cumulative resistance value of the potential corridor between i and j; (:{P}_{i}{P}_{j}) is the average resistance value of i and j.ResultsLand use change and landscape pattern characterizationThis study reveals the land use change characteristics of Gui’an New Area from 2010 to 2020 and predicts the development characteristics of the district in 2030 and 2060 (Figs. 4 and 5a-f). Land-use simulations for Gui’an New Area in 2030 and 2060, conducted with the FLUS model, achieved an overall accuracy of 0.93 and a Kappa coefficient of 0.85, indicating strong agreement between the projected outcomes and actual land-use patterns and underscoring the model’s robust predictive reliability. Between 2010 and 2022, the area of woodland and cropland in Gui’an New Area decreased, with cropland decreasing by 5.2% to 1221.62 km2. Woodland, grassland, and buildings grow by 4.3%, 40.49% and 87.36%, respectively. Cropland is projected to continue to decrease by 5.5% by 2030, while woodland, grassland and buildings will grow by 5.6%, 15.5% and 39.8% respectively. During 2010–2030, built-up areas expanded annually while cropland decreased and woodland increased post-2014. Water grew marginally. By 2060, cropland is projected to rise 14% from 2030 levels with built-up areas doubling. Woodland and grassland will decline, whereas other land use types increase.Eight landscape types were identified in this study: core, bridging, loop, perforation, islet, branch, edge, and background. The background has the largest percentage of area, followed by the core and edge. The share of the core in the total area of the ecological landscape between 2010 and 2060 decreased with proportions of 18.35%, 16.67%, 17.70%, 19.59%, 17.78% and 10.37%, respectively (Fig. 6). It is mainly distributed in the northwestern and central parts of the study area with good spatial connectivity, while it is less distributed and poorly connected in the eastern part. The marginal zone is large and growing steadily, and the perforation and loop zones account for a relatively small proportion of the area.Fig. 4Characteristics of land use types changes in Gui’an New Area, 2010–2060 (km2).Full size imageFig. 5Characteristics of the spatial distribution of land use types from 2010 to 2060. (This figure was obtained by using ArcGIS 10.8 through the open-access data process).Full size imageFig. 6Characteristics of the spatial distribution of land use types and landscape types from 2010 to 2060. (This figure was obtained by using the Guidos Toolbox 2.8 and ArcGIS 10.8 through the open-access data process).Full size imageSpatial and Temporal variations in CSIn this study, the CS of Gui’an New Area from 2010 to 2060 was calculated at six time points, and the results showed that the CS slightly decreased in 2014, but the overall trend of growth was observed between 2014 and 2030, reaching 28,977,800 T in 2030, with a growth rate of 4.6%. By 2060, CS had decreased by 19.7%. Among the land use types, the CS of cropland, woodland, grassland, buildings, and other land decreases in order. From 2010 to 2060, the static CS of buildings land increased by a total of 1,147,600 T, and the CS of woodland increased by a total of 1,339,200 T, with a growth rate of 10.6%. Cropland, grassland, and other land CS fluctuated, but water CS was not accounted for due to model limitations (Fig. 7). CS in Gui’an New Area show spatial heterogeneity, mainly concentrated in the woodland and grassland areas in the northwest, forming a high-density CS area (Fig. 8). In contrast, CS is lower in the northern and eastern regions, which are densely populated and have more watersheds. The CS potential of Gui’an New Area is influenced by the area of ecological space and carbon density. Future projections show that changes in land use and landscape types will maintain the “large concentration and small dispersion” pattern of CS in Gui’an New Area.Fig. 7CS in different land use types in Gui’an New Area, 2010–2060 (104T).Full size imageFig. 8Characteristics of spatial differentiation of CS, 2010–2060. (This figure was obtained by using ArcGIS 10.8 through the open-access data process.)Full size imageCharacterization of ecological safety network patternsThe number and area of ecological source sites in the Gui’an New Area experienced fluctuations between 2010 and 2030 (Fig. 9a-e). There were 14 ecological source sites in 2010, while they decreased to 12 in 2014, 2030, and 2060, and 13 in 2018 and 2022. The total area of ecological source sites also varied from year to year, ranging from 161.94 km2 to 196.44 km2, and the intersection area between the periods was 114.74 km². The number and type of ecological corridors changed significantly over time, with 91 corridors in 2010, including 17 important corridors; in 2014, the number decreased to 66 and 11 important; in 2018 and 2022, there were 78 corridors each, with 11 and 19 important corridors, respectively; and in 2030, there were 63 corridors and 15 important. By 2060, the number of primary ecological corridors will decrease to 66, with no important corridors (Fig. 9f). The overall trend shows a decrease in the number of ecological corridors.In terms of the spatial distribution of the ecological safety network, the ecological sources in Gui’an New Area are mainly concentrated in watersheds, woodland, and grassland and are denser in the northwestern area and the south-central part of the area and less in the southwestern part of the area and the eastern part of the area. The ecological corridors are densely distributed in 2010, 2018, and 2022, and the important ecological corridors in 2010 connect all the major ecological sources. The important ecological corridors in the central and western parts of the area are decreasing in 2014, and general ecological corridors are increasing. and general ecological corridors increase. 2018 important ecological corridors in the west further decrease, general ecological corridors continue to increase, and the distribution pattern is like that of 2030. By 2060, only first-class ecological corridors will remain to be interspersed. It is evident that ecological corridors in the eastern region are restricted in their formation and development due to the compression of ecological space.Fig. 9Characteristics of the ecological safety network, 2010–2060. (This figure was obtained by using ArcGIS 10.8 through the open-access data process).Full size imageDiscussionPolicies drive Spatial changes in the landscape of Gui’an New AreaSince the establishment of Gui’an New Area in 2014, there has been a continuous expansion of land for construction, a decrease in cropland, and a significant increase in grassland area (Fig. 5). This is a key role of policy orientation in driving spatial changes in regional landscapes. Rapid economic growth and sustained population growth, coupled with accelerated urbanization, have together led to a significant increase in the demand for construction land, which in turn has triggered a significant shift in the type of land use3,18,34. The land use policy and urban planning of Gui’an New Area (Tables 5 and 6), similar to that of Shenzhen SAR and Xiong’an New Area, focuses on promoting regional economic development and urbanization35,36, with special support for the development of big data and high-end manufacturing industries, which further increases the land resource demand and puts pressure on ecological space37.Table 5 Characteristics of land use type changes in Gui’an New Area, 2010–2030 (km2).Full size tableTable 6 Carbon stocks in different land use types in Gui’an New Area, 2010–2030 (104T).Full size tableUrban expansion has not only reshaped the original natural landscape but also reduced the area of the core area of the landscape and its ecological connectivity in Gui’an New Area. Although some greening measures have been taken during urban development and the area of woodland has increased in 2022 and is expected to continue to grow under the impetus of ecological preservation policies, ecological preservation in the core area is still insufficient, and part of the ecological space has been replaced by economic activities38. Population growth is driving the expansion of residential and commercial space, especially in the more economically developed eastern regions, which further exacerbates the pressure on ecological space39. This rapid urbanization-induced ecological space compression exhibits convergent evolutionary patterns in Shenzhen35 and the Yangtze River Delta (Shanghai)40. Land use projections to 2060 indicate that planning policies will have a significant impact on long-term trends in land use types. Building land is expected to continue to grow, while cropland will undergo a process of decreasing and then increasing. This suggests that policy-oriented urban expansion is a direct reflection of the increased demand for non-agricultural land, especially the growth of building land41,42. While policy decisions and land management may be more inclined to support economic development, the karst ecosystem of Gui’an New Area also maintains its inherent woodland and water. This change reflects a general trend in the urbanization process and highlights future challenges in land management. This study confirms that the changes in land use in Gui’an New Area are the result of changes in socio-economic conditions and natural factors brought about by policy-oriented urbanization.Land use types dominate CS patternsOur study shows that the type of land use types in Gui’an New Area is a key factor that dominates the regional CS pattern (Figs. 7 and 8). The study shows that despite a decline in the region’s CS in 2014, the overall trend of CS growth from 2014 to 2030 is 1,277,700 T, a growth rate of 4.6%. However, this growth trend did not continue until 2060, when the CS declined. This suggests that land use changes, especially urban expansion, have a significant impact on CS in ecosystems3,18. Woodland and grassland, as important carbon pools in Gui’an New Area, contribute more to the total CS. In the process of urbanization, cropland and natural vegetation are converted into building land, which not only reduces carbon sinks but also increases carbon emissions. However, sound land planning and ecological restoration projects can increase carbon sequestration capacity, especially in woodland and grassland areas. From 2010 to 2060, the CS of built-up land increased by 1,147,600 T, and this increase may be related to the increase in urban greening and NDVI. The karst topography, geomorphology, climate, and soil conditions in Gui’an New Area have a special impact on carbon sequestration, making it different from other regions9,43. For example, urbanization expansion in the Sichuan-Chongqing urban agglomeration has led to a decline in CS and HQ, which underscores the urgency of sustainable land management practices3. In addition, the land-use change of ” Grain for Green” also enhances the carbon sequestration capacity of ecosystems44, suggesting that land-use changes, such as vegetation restoration, can increase regional CS.CS in Gui’an New Area exhibits distinct spatial heterogeneity, predominantly concentrated in northwestern woodland and grassland forming a high-density carbon zone. From 2010 to 2030, CS is influenced by a variety of factors, including land management practices, vegetation restoration projects, and natural carbon cycle changes. Urbanization has led to localized declines in CS, especially in the northern and eastern regions with high population density and watersheds, which have been affected by reduced NDVI due to urbanization activities and watershed development. Among all land use types, cropland has the largest CS, but with a decreasing trend, while the CS of woodland has experienced a decrease and then an increase and is expected to grow to 13,919,900 T by 2030. This emphasizes the importance of protecting and restoring woodland and optimizing land use structure and habitat to maintain and enhance regional CS45. Therefore, to protect and enhance the regional CS capacity, promote sustainable development, and conserve biodiversity, Gui’an New Area needs to adopt active ecological protection measures, strengthen ecological spatial planning and management, raise public awareness of ecological protection, and conduct in-depth studies on the impacts of changes in landscape patterns. These measures are essential to achieving long-term stability and growth in regional CS.Ecological safety network patternThis study reveals a significant decrease in the number of ecological source sites and ecological corridors in Gui’an New Area from 2010 to 2060, as well as the resulting decrease in ecological network connectivity. In 2010, Gui’an New Area had 14 ecological source sites, whereas by 2060, this number was reduced to 12, suggesting a clear downward trend. At the same time, the total number of ecological corridors decreases from 91 to 66, with the number of important ecological corridors decreasing from 17 to zero, a change that reflects the significant impacts of land-use changes and disturbances in landscape patterns on the ecological safety network18. The overall decline in the number of ecological corridors between 2010 and 2030 is of particular concern, especially the reduction of important ecological corridors, while the increase in the number of common ecological corridors indicates a shift from natural landscapes to built-up land in the process of urbanization, which not only reduces biodiversity but also weakens the connectivity of the ecological network44. Although some ecological corridors have been preserved, their connectivity, function, and importance have declined, posing an impact on the overall health of the ecosystem. For example, in the coastal cities of eastern China18, urbanization has also led to the reduction of ecological source sites and fragmentation of ecological corridors, suggesting that insufficient ecological protection and irrational planning are the main reasons. The analysis shows socio-economic pressures in eastern Gui’an New Area compress ecological space, limiting the formation of ecological corridors. However, compared with the successful case of ecological conservation in Shenzhen35, the changes in the ecological network of Gui’an New Area appear to be more complex. The ecological source areas in Gui’an New Area vary significantly in scale (with the largest being 85.76 km² and the smallest 2.82 km² by 2030), highlighting patch fragmentation. Scenario projections for 2060 indicate that the network is showing a degenerative trend, with the number of corridors and connectivity decreasing simultaneously, which is highly consistent with the findings of existing studies in rapidly urbanizing regions such as Shanghai40, Nanchang33, and the Chengdu-Chongqing13 area.Despite the continuous changes over time, the ecological source areas in Gui’an New Area are still mainly concentrated in watersheds, woodland, and grassland, especially denser in the northwestern part of the area and in the south-central part of the area, and less in the southwestern part of the area and in the eastern part of the area, showing the heterogeneity of the spatial distribution. This emphasizes the need to prioritize the protection and restoration of the ecological functions of the core zone as a key part of the ecological safety network. Therefore, we emphasize that the Gui’an New Area should strictly adhere to the ecological red line to lock in core areas and corridors; adopt a compact city model to curb urban sprawl while incorporating green corridors; carry out targeted restoration in key fragile areas; and establish a collaborative assessment mechanism for ecology and development.Limitations and perspectivesThis study adopts two major models of land use prediction and CS estimation to quantitatively analyze the characteristics of land use change and CS change in Gui’an New Area from 2010 to 2060 and construct the corresponding ecological safety network pattern. This study can provide a scientific basis for ecological protection planning, environmental policymaking, and sustainable development, and at the same time improve the understanding of the value of ecosystem services and biodiversity conservation. Nevertheless, this study still has certainshortcomings and foresight for future research. The CS model only considers the static CS and ignores the carbon cycle and the dynamic transformation between different carbon pools, which does not differ from the actual situation, thus affecting the accuracy of CS calculation. Additionally, the present work relies on a single “ecological-priority” scenario; the projected 2030 peak and the 19.7% net loss of carbon stock by 2060 are therefore scenario-dependent. Whether an “economic-growth-priority” trajectory (e.g., faster urban expansion, higher industrial land demand) would erase the 2030 peak or even accelerate the decline remains untested. In future research, we will couple the FLUS model with scenario drivers such as GDP, population and policy constraints to generate a spectrum of development pathways. These steps will refine targeted strategies for green and low-carbon development as well as for the ecological security pattern of Gui’an New Area.ConclusionsGui’an New Area, as China’s 8th state-level new area’s spatial changes in land use, CS characteristics, and ecological safety network are crucial to the green and sustainable development of the region. In this study, the following conclusions are drawn by analyzing the land use changes in Gui’an New Area from 2010 to 2060, predicting the land use in 2030 and 2060 using the FLUS model, conducting the spatial-temporal dynamic analysis of CS using the InVEST model, as well as constructing the ecological safety network pattern by combining the MSPA and MCR models:(1) Policies have driven changes in land use types and landscape spatial patterns in Gui’an New Area. For example, in 2014, the building land of Gui’an New Area promoted the reduction of woodland and cropland and the significant increase of building land. As Gui’an New Area becomes more urbanized, the landscape pattern continues to change. The area of woodland and grassland is predicted to decline by 2060, while built-up land expands year by year. The area of the landscape core area also shows a decreasing trend in terms of percentage decline between 2010 and 2060.(2) The CS in Gui’an New Area rises to a peak in 2030, then declines toward 2060; the net change from 2010 to 2060 is 19.7%. From 2014 to 2030, the CS of cropland, woodland, and grassland remains the highest, while that of building land, though lower, increases significantly during the study period. By 2060, land use changes in Gui’an New Area exert a pronounced negative impact on CS, driving the post-2030 decline.(3) Between 2010 and 2060, the number and area of ecological source sites in Gui’an New Area fluctuated, and the number of ecological corridors generally decreased. Ecological source areas and ecological corridors are mainly concentrated in watersheds, woodland, and grassland, especially in the northwestern area and the central and southern parts. In the eastern part, due to economic development and population growth, ecological space is compressed, and corridor formation is limited, reflecting the challenge of economic development to ecological protection.Overall, future land use planning for Gui’an New Area should focus on ecological protection and restoration, optimizing the land use structure, and enhancing ecological service functions. In the eastern part of the area, where there are fewer ecological corridors, the construction and maintenance of ecological networks should be strengthened to enhance ecosystem connectivity and biodiversity. At the same time, the land-use change, CS capacity, and ecological safety network pattern of Gui’an New Area will be continuously monitored and assessed so that problems can be identified in a timely manner and appropriate management measures can be taken to promote sustainable development.

    Data availability

    The Land use dataset, which includes the land use type, is available at a 30 m resolution for the period 2010-2022 and can be accessed at https://zenodo.org/ (accessed on 15 April 2024). The Terrain dataset comprises elevation, slope, and slope direction. These data are available at a 30 m resolution, and the source can be accessed at http://www.gscloud.cn/ (accessed on 1 April 2024). The climate dataset consists of average monthly temperature and average monthly precipitation for the period 2010-2022. The monthly average precipitation from the Copernicus Climate Data Store (https://cds.climate.copernicus.eu/ accessed on 2 May 2024). The monthly average temperature from the National Tibetan Plateau Data Centre (China), (https://data.tpdc.ac.cn/ accessed on 18 December 2023). Environmental data: the NDVI dataset is sourced from the National Ecological Data Centre (China), (http://www.nesdc.org.cn accessed on 16 June 2024). The soil moisture data from the A Fine-Resolution Soil Moisture Dataset for China in 2002–2018, (https://doi.org/10.5281/zenodo.4738556 accessed on 18 January 2024). The desertification index from the Food and Agriculture Organization of the United Nations, (https://www.fao.org/ accessed on 5 January 2024). Socioeconomic factors such as GDP and population (POP) are included in the dataset at a 1 km resolution for the period 2010-2022. The data can be accessed at https://www.resdc.cn/ (accessed on 10 April 2024). Accessibility factors such as the distance to railway, highway, primary road, secondary road, and tertiary roads are derived from vector data analyzed using ArcGIS Euclidean distance. The data for these factors are available at a 30 m resolution for the period 2010-2022. The specific sources include Open Street Map for roads, which can be accessed at https://www.openstreetmap.org/ (accessed on 10 April 2024).
    ReferencesMo, L. et al. Integrated global assessment of the natural forest carbon potential. Nature 624, 92–101 (2023).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Li, L., Huang, X. & Yang, H. Optimizing land use patterns to improve the contribution of land use planning to carbon neutrality target. Land. Use Policy. 135, 106959 (2023).Article 

    Google Scholar 
    Zhang, H., Li, X., Luo, Y., Chen, L. & Wang, M. Spatial heterogeneity and driving mechanisms of carbon storage in the urban agglomeration within complex terrain: Multi-scale analyses under localized SSP-RCP narratives. Sustain. Cities Soc. 109, 105520 (2024).Article 

    Google Scholar 
    Pan, Y. et al. A large and persistent carbon sink in the world’s forests. Science 333, 988–993 (2011).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Weiskopf, S. R. et al. Biodiversity loss reduces global terrestrial carbon storage. Nat. Commun. 15, 4354 (2024).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Zhang, C. et al. Impacts of urbanization on carbon balance in terrestrial ecosystems of the Southern united States. Environ. Pollut. 164, 89–101 (2012).Article 
    CAS 
    PubMed 

    Google Scholar 
    Roebroek, C. T. J., Duveiller, G., Seneviratne, S. I., Davin, E. L. & Cescatti, A. Releasing global forests from human management: how much more carbon could be stored? Science https://doi.org/10.1126/science.add5878 (2023).Article 
    PubMed 

    Google Scholar 
    Wu, Y. et al. Low carbon storage of Woody debris in a karst forest in Southwestern China. Acta Geochim. 38, 576–586 (2019).Article 
    CAS 

    Google Scholar 
    Wu, Y. et al. NDVI-Based vegetation dynamics and their responses to climate change and human activities from 2000 to 2020 in Miaoling karst mountain Area, SW China. Land 12, 1267 (2023).Article 
    CAS 

    Google Scholar 
    Hasan, S. S., Zhen, L., Miah, M. G., Ahamed, T. & Samie, A. Impact of land use change on ecosystem services: A review. Environ. Dev. 34, 100527 (2020).Article 

    Google Scholar 
    Zhang, Z., Hu, B., Jiang, W. & Qiu, H. Construction of ecological security pattern based on ecological carrying capacity assessment 1990–2040: A case study of the Southwest Guangxi Karst – Beibu Gulf. Ecol. Model. 479, 110322 (2023).Article 

    Google Scholar 
    Cao, C., Luo, Y., Xu, L., Xi, Y. & Zhou, Y. Construction of ecological security pattern based on InVEST-Conefor-MCRM: A case study of Xinjiang, China. Ecol. Indic. 159, 111647 (2024).Article 

    Google Scholar 
    Shao, Z. et al. Impact of Land Use Change on Carbon Storage Based on FLUS-InVEST Model: A Case Study of Chengdu–Chongqing Urban Agglomeration, China. Land 12, 1531 (2023).Wu, J., Luo, J., Zhang, H., Qin, S. & Yu, M. Projections of land use change and habitat quality assessment by coupling climate change and development patterns. Sci. Total Environ. 847, 157491 (2022).Article 
    CAS 
    PubMed 

    Google Scholar 
    Ding, Y., Gui, F., Tian, S. & Zhao, S. Temporal and spatial changes of habitat quality in the area around Hangzhou Bay based on InVEST model. in International Conference on Artificial Intelligence, Virtual Reality, and Visualization (AIVRV 2021) (eds Li, T., Chen, S., Wu, D. & Gao, G.) 6SPIE, https://doi.org/10.1117/12.2626517 (Sanya, China, 2021).Li, B. et al. Prediction and valuation of ecosystem service based on land use/land cover change: A case study of the Pearl river delta. Ecol. Eng. 179, 106612 (2022).Article 

    Google Scholar 
    Wang, C. et al. Land use change and its impact on carbon storage in Northwest China based on FLUS-In VEST – A case study of Hubao-Eyu urban agglomeration. J. Ecol. Environ. 31, 1667–1679 (2022). [Chinese].CAS 

    Google Scholar 
    Zhu, L., Song, R., Sun, S., Li, Y. & Hu, K. Land use/land cover change and its impact on ecosystem carbon storage in coastal areas of China from 1980 to 2050. Ecol. Indic. 142, 109178 (2022).Article 
    CAS 

    Google Scholar 
    Li, J., Gong, J., Guldmann, J. M., Li, S. & Zhu, J. Carbon dynamics in the Northeastern Qinghai–Tibetan plateau from 1990 to 2030 using landsat land Use/Cover change data. Remote Sens. 12, 528 (2020).Article 
    ADS 

    Google Scholar 
    He, Y., Ma, J., Zhang, C. & Yang, H. Spatio-Temporal evolution and prediction of carbon storage in Guilin based on FLUS and invest models. Remote Sens. 15, 1445 (2023).Article 
    ADS 

    Google Scholar 
    Yi, L. I. et al. Optimization of ecological red line in the hilly region of Southern China based on invest and MCR model. J. Nat. Resour. 36, 2980–2994 (2021).
    Google Scholar 
    Nie, W. et al. Simulating future land use by coupling ecological security patterns and multiple scenarios. Sci. Total Environ. 859, 160262 (2023).Article 
    CAS 
    PubMed 

    Google Scholar 
    Zhang, Y. Z. et al. Construction and optimization of an urban ecological security pattern based on habitat quality assessment and the minimum cumulative resistance model in Shenzhen city, China. Forests 12, 847 (2021).Xi, X. et al. The State Council approved the establishment of Gui’an New Area. Contemporary Guizhou 13,03 (2014). [Chinese].Yi, P. Gui’an New Area is the new fulcrum of Guizhou’s leapfrog development. Contemporary Guizhou 66,04 (2014). [Chinese].Qi, Y. et al. Exploring the development of the sponge City program (SCP): the case of gui’an new District, Southwest China. Front. Water. 3, 676965 (2021).Article 

    Google Scholar 
    Yang, J. & Huang, X. The 30 m annual land cover dataset and its dynamics in China from 1990 to 2019. Earth Syst. Sci. Data 13, 3907–3925 (2021).Meng, X. et al. A fine-resolution soil moisture dataset for China in 2002–2018. (2020). https://doi.org/10.5194/essd-2020-292Zhong, X. et al. Long Time Series Nighttime Light Dataset of China (2000–2020). https://doi.org/10.3974/geodb.2022.06.01.V1Zhang, K., Fang, B., Zhang, Z., Liu, T. & Liu, K. Exploring future ecosystem service changes and key contributing factors from a past-future-action perspective: A case study of the yellow river basin. Sci. Total Environ. 926, 171630 (2024).Article 
    CAS 
    PubMed 

    Google Scholar 
    Li, X. et al. A cellular automata downscaling based 1 Km global land use datasets (2010–2100). Sci. Bull. 61, 1651–1661 (2016).Article 
    CAS 

    Google Scholar 
    Yang, J. A study of land cover types and carbon storage changes in Guiyang City, 1980–2018. J. Southwest. Forestry Univ. (Natural Science). 40, 115–121 (2020). [Chinese].
    Google Scholar 
    Wang, C. et al. Can the establishment of ecological security patterns improve ecological protection? An example of Nanchang, China. Sci. Total Environ. 740, 140051 (2020).Article 
    CAS 
    PubMed 

    Google Scholar 
    Zhang, W., Chang, W. J., Zhu, Z. C. & Hui, Z. Landscape ecological risk assessment of Chinese coastal cities based on land use change. Appl. Geogr. 117, 102174 (2020).Article 

    Google Scholar 
    Liu, X., Su, Y., Li, Z. & Zhang, S. Constructing ecological security patterns based on ecosystem services trade-offs and ecological sensitivity: A case study of Shenzhen metropolitan area, China. Ecol. Indic. 154, 110626 (2023).Article 

    Google Scholar 
    Ye, L. Solidly promoting the planning and construction of Xiongan new area. Econ. Manage. 31, 6–12 (2017). [Chinese].
    Google Scholar 
    Kang, P., Chen, W., Hou, Y. & Li, Y. Spatial-temporal risk assessment of urbanization impacts on ecosystem services based on pressure-status – response framework. Sci. Rep. 9, 16806 (2019).Article 
    ADS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Wu, W., Xu, L., Zheng, H. & Zhang, X. How much carbon storage will the ecological space leave in a rapid urbanization area? Scenario analysis from Beijing-Tianjin-Hebei urban agglomeration. Resour. Conserv. Recycl. 189, 106774 (2023).Article 

    Google Scholar 
    Guo, H. et al. Higher water ecological service values have better network connectivity in the middle yellow river basin. Ecol. Indic. 160, 111797 (2024).Article 

    Google Scholar 
    Wu, Y. et al. Decoding carbon pathways of Shanghai megacity through historical land use patterns and urban ecosystem transitions. Sci. Rep. 15, 6326 (2025).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Wei, Y., Zhou, P., Zhang, L. & Zhang, Y. Spatio-temporal evolution analysis of land use change and landscape ecological risks in rapidly urbanizing areas based on Multi-Situation simulation – a case study of Chengdu plain. Ecol. Indic. 166, 112245 (2024).Article 

    Google Scholar 
    Wu, K., Wang, D., Lu, H. & Liu, G. Temporal and Spatial heterogeneity of land use, urbanization, and ecosystem service value in china: A national-scale analysis. J. Clean. Prod. 418, 137911 (2023).Article 

    Google Scholar 
    Ding, Y. et al. Estimating land use/land cover change impacts on vegetation response to drought under ‘Grain for green’ in the loess plateau. Land. Degrad. Dev. 32, 5083–5098 (2021).Article 

    Google Scholar 
    Nolan, C. J., Field, C. B. & Mach, K. J. Constraints and enablers for increasing carbon storage in the terrestrial biosphere. Nat. Rev. Earth Environ. 2, 436–446 (2021).Article 
    ADS 

    Google Scholar 
    Niu, L., Zhang, Z., Liang, Y. & Huang, Y. Assessing the impact of urbanization and Eco-Environmental quality on regional carbon storage: A multiscale Spatio-Temporal analysis framework. Remote Sens. 14, 4007 (2022).Article 
    ADS 

    Google Scholar 
    Download referencesAcknowledgementsWe thank the anonymous reviewers for their valuable comments. We gratefully acknowledge the design of Siliang Li and the contributions of the co-authors. We appreciate Zhenghua Shi, Lei Gu, Liqing Wu, Yue Fu, Songchi Xie and Shasha Li’s suggestions for paper revision. We thank Ruixue Fan’s contribution to the English revision of the manuscript.FundingThis work was supported by Guizhou Provincial Key Project of Philosophy and Social Science Planning (24GZZD61); Guizhou Provincial Science and Technology Projects (QKHZC [2023] YB228); Guizhou Provincial Science and Technology Projects (QKHPT KXJZ [2024] 032); Guizhou Provincial Science and Technology Projects (Qian Ke He Cheng Guo [2023] Zhong Da 006); National Natural Science Foundation of China (U24A20579); Guizhou Provincial Digital Rural Innovation Team in Higher Education (QJJ [2023] 076).Author informationAuthor notesYangyang Wu and Huangting Luo contributed equally to this work.Authors and AffiliationsSchool of Geography and Resources, Guizhou Education University, Guiyang, 550018, ChinaYangyang Wu, Huangting Luo & Guangjie LuoSchool of Earth System Science, Tianjin University, Tianjin, 300072, ChinaYangyang Wu, Silang Li & Chunzi GuoGuizhou Provincial Key Laboratory of Geographic State Monitoring of Watershed, Guizhou Education University, Guiyang, 550018, ChinaHuangting Luo, Jinli Yang, Haobiao Wu, Panli Yuan & Guangjie LuoCollege of Ecology and Environment, Xinjiang University, Urumqi, 830017, ChinaJinli Yang, Xiaodong Yang, Haobiao Wu & Panli YuanAdministration of Ecology and Environment of Haihe River Basin and Beihai Sea Area, Ministry of Ecology and Environment of People’s Republic of China, Tianjin, 300061, ChinaChunzi GuoState Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences, Guiyang, 550081, ChinaYue XuSchool of Environmental and Life Sciences, Nanning Normal University, Nanning, 530100, ChinaZhonghua ZhangDepartment of Geography and Spatial Information Techniques, Ningbo University, Ningbo, 315211, ChinaXiaodong YangState Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Science, Nanjing, 210008, ChinaSongyu YangAuthorsYangyang WuView author publicationsSearch author on:PubMed Google ScholarHuangting LuoView author publicationsSearch author on:PubMed Google ScholarSilang LiView author publicationsSearch author on:PubMed Google ScholarJinli YangView author publicationsSearch author on:PubMed Google ScholarChunzi GuoView author publicationsSearch author on:PubMed Google ScholarYue XuView author publicationsSearch author on:PubMed Google ScholarZhonghua ZhangView author publicationsSearch author on:PubMed Google ScholarXiaodong YangView author publicationsSearch author on:PubMed Google ScholarSongyu YangView author publicationsSearch author on:PubMed Google ScholarHaobiao WuView author publicationsSearch author on:PubMed Google ScholarPanli YuanView author publicationsSearch author on:PubMed Google ScholarGuangjie LuoView author publicationsSearch author on:PubMed Google ScholarContributionsConceptualization, Y.W. and S.L.; methodology, X.Y. and G.L.; software, H.L. and J.Y.; validation, S.Y. and Z.Z.; formal analysis, C.G. and Y.X.; investigation, H.W. and P.Y.; data curation, S.Y., Z.Z. and Y.X.; writing—original draft preparation, Y.W. and H.L.; writing—review and editing, S.L. and J.Y.; visualization, J.Y. and H.L.; supervision, C.G.; project administration, G.L. and X.Y.; funding acquisition, Y.W. All authors have read and agreed to the published version of the manuscript.Corresponding authorsCorrespondence to
    Xiaodong Yang or Guangjie Luo.Ethics declarations

    Competing interests
    The authors declare no competing interests.

    Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Rights and permissions
    Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
    Reprints and permissionsAbout this articleCite this articleWu, Y., Luo, H., Li, S. et al. Coupling models to assess impacts of land and carbon changes on sustainable ecological safety networks of Gui’an New Area, China.
    Sci Rep 15, 44580 (2025). https://doi.org/10.1038/s41598-025-28233-wDownload citationReceived: 23 September 2024Accepted: 10 November 2025Published: 24 December 2025Version of record: 24 December 2025DOI: https://doi.org/10.1038/s41598-025-28233-wShare this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative More

  • in

    Impact of particle size and surface modifications on the neurotoxic potential of copper oxide nanoparticles

    AbstractCopper oxide (CuO) nanoparticles (NPs) have widespread applications in electronics, energy storage, and healthcare domains owing to their high surface-to-volume ratio, catalytic activity, and anti-bacterial and anti-microbial properties. However, the health hazard of direct CuO exposure to humans has raised safety concerns. CuO NPs can cross the blood–brain barrier, access the central nervous system, and trigger neurotoxicity. Previous studies have investigated the neurotoxicity of CuO NPs. However, the effects of different sizes and comparable size NPs with and without surface coating have not been previously reported. In this study, two differentially sized NPs (CuO-25 and CuO-48 NPs) and one polyvinylpyrrolidone-coated NP (CuO-P NPs; 46 nm) were synthesized and characterized. The neurotoxic potential of these NPs was examined in vitro using PC-12 cells. CuO NPs significantly decreased cell viability at concentrations of ≥ 1 μg/mL by inducing oxidative and nitrosative stress in a time-dependent and concentration-dependent manner. Additionally, CuO NPs altered mitochondrial membrane potential, upregulated Il6 and Tnf levels, induced apoptosis by upregulating Casp3 activity, and inhibited acetylcholinesterase activity. Furthermore, CuO NPs upregulated the expression of Maoa and Snca, which are associated with dopamine metabolism and the pathogenesis of neurodegenerative disorders. The three NPs exerted differential effects. The cytotoxic effects of CuO-25 NPs were higher than those of CuO-48 NPs. Additionally, the cytotoxic effects of coated NPs (CuO-P) were lower than those of uncoated NPs. Cu2+ ions released from NPs mediate the neurotoxic effects of NPs.

    IntroductionCopper (Cu), a vital trace element, is involved in several physiological functions, such as neurotransmission, energy metabolism, and antioxidant defense1. The health hazard of Cu toxicity in the general population is minimal as its homeostasis is regulated at physiological levels. In the last two decades, CuO nanoparticles (NPs) have been widely applied in various fields, such as electronics, energy, and healthcare owing to their high surface area-to-volume ratio, catalytic activity, and anti-bacterial capabilities2. The potential adverse effects of these NPs are serious health concerns, necessitating thorough toxicological evaluations. Previous studies have revealed that CuO NPs may enter the human body via ingestion, inhalation, or dermal route and can potentially accumulate in various organs3. The potential targets for CuO NP toxicity include the lungs, kidneys, and liver4. CuO NPs are reported to cross the blood–brain barrier (BBB). This excessive Cu exposure can lead to neurotoxicity characterized by cognitive impairment, motor dysfunction, and neurodegenerative disorders (NDs)5,6,7.Previous in vitro and cell studies have demonstrated that CuO NPs exert toxic effects on primary brain cells8,9. Analysis of adhesion kinetics, growth, proliferation, and DNA damage revealed that CuO NPs exert cytotoxic and genotoxic effects on the primary cultures of brain microvascular endothelial cells and astrocytes10. CuO NPs exhibit anti-proliferative properties and can induce cell death in glioma and neural cells11,12. One study demonstrated that CuO NPs promote apoptosis in glial and neuronal cell lines13. At subtoxic doses, CuO NPs activate the NF-κB signaling pathway and enhance amyloid precursor protein expression in neural cells, suggesting a correlation between NPs and NDs14.In vivo studies have also reported the neurotoxic effects of CuO NPs. For example, mice intranasally exposed to CuO NPs exhibited nerve cell damage and dysfunction in the cerebellum, cerebral cortex, hippocampus, and striatum15. Additionally, CuO NPs altered the metabolic and antioxidant characteristics of brain tissues, acetylcholinesterase (AChE) activities, glutathione levels, and lipid peroxidation and downregulated the expression of the cytochrome P-450 enzyme system16,17,18. In higher organisms, AChE is involved in neurotransmission, cognition, and memory. The principal function of AChE in cholinergic synapses is to hydrolyze the neurotransmitter acetylcholine. However, ingested NPs may bind to AChE and alter its activity19 .NPs promote brain tissue damage by inducing apoptosis, inflammation, protein aggregation, and oxidative stress. These mechanisms can lead to the onset and progression of degenerative diseases. Limited studies have examined the correlation between CuO NPs and specific disorders. Therefore, there is a need to evaluate the health risks of CuO NPs, especially ND risk.This study aimed to examine the neurotoxicity of CuO NPs and assess the effects of their size, surface coating, and released Cu2+ ions on the neurotoxic effects. CuO NPs of varying sizes, as well as particles of similar dimensions with or without surface coating, were synthesized. This approach enables the comprehensive evaluation of the effect of particle size and surface coating on neurotoxicity. The polymer polyvinylpyrrolidone (PVP) was selected to examine the effect of coating on the neurotoxicity of CuO NPs, ensuring comparability with other research endeavors. PC-12 cells, which are routinely used to examine neurotoxic effects and can differentiate into neuron-like cells, were used as the in vitro model.Additionally, this study examined the impact of CuO NPs on the expression of genes associated with the dopaminergic system (Th, Maoa, and Comt) and neurodegeneration initiation (Snca, Prkn, and Gpr37) and AChE enzyme activity.Materials and methodsNP Synthesis and characterizationCuO NPs were synthesized using the precipitation method with cupric nitrate and copper sulfate as precursors. First, 0.5 M aqueous solutions of cupric nitrate and copper sulfate were prepared. The aqueous cupric nitrate and copper sulfate solutions were incubated with 0.5 M aqueous sodium hydroxide solution (added dropwise) with constant vigorous stirring at room temperature to obtain black residue (pH 10). The precipitate was centrifuged, washed with deionized water (to neutralize pH), and divided into two portions. The first portion was subjected to calcination at 80 °C, while the second portion was annealed at 400 °C for 2 h to obtain the black powder, which was finely ground and stored in a vacuum desiccator.PVP-coated CuO NPs were synthesized by combining the capping agent (PVP) and precursor salts (cupric nitrate) in a 1:1 ratio with the precursor salt (copper acetate monohydrate). The CuO NPs were coated separately with PVP during synthesis.The particle shape and size were analyzed using a field-emission scanning electron microscope (FEI Nova NanoSEM 450). The mean particle diameter was estimated by analyzing 100 particles (ImageJ, National Institutes of Health). The elements were screened using an energy-dispersive detector (Bruker XFlash 6I30). The CuO NPs were ultrasonicated for 5 min in deionized water at 1 mg/mL and analyzed using a dynamic light scattering instrument (DLS, Sympatec Nanophox) to determine the hydrodynamic diameter. The zeta potential was calculated using Beckman Coulter (Delsa™ Nano). Fourier-transform infrared (FTIR) spectroscopy was used to determine the properties of the functional groups at wavelengths of 400–4000 cm−1 using an FTIR spectrometer (Bruker Tensor-27).A solid-state ultraviolet–visible (UV–Vis) spectrophotometer (Jasco) was used to analyze the optical characteristics of CuO in the 200–800 nm region. The absorption edge of Tauc plot was extrapolated from the UV–vis spectra to determine the optical bandgap of NPs.Cell culture and NP treatmentPC-12 (rat pheochromocytoma) cells, which were obtained from the National Centre for Cell Science (Pune, India), were maintained in F-12 Ham, Kaighn’s modification medium supplemented with 10% heat-inactivated horse serum, 5% heat-inactivated fetal bovine serum, and an antibiotic solution containing 10,000U of penicillin and 10 mg of streptomycin at 37 °C in a humidified 5% CO2 incubator until 80%–90% confluency.Before NP treatment, an optimization study was performed to establish the appropriate cell density (2000 to 5 × 104 cells per well). This optimization study aimed to mitigate cell death from nutrient depletion in the culture medium during the 96-h treatment. The optimal cell density was determined to be 1 × 104 cells per well, which minimized the potential confounding factors associated with cell stress and nutrient depletion during prolonged NP exposure.All NPs were sterilized through autoclaving before treatment and tested for contaminants. NPs were dispersed in the complete growth media and subjected to probe sonication for 15 min before each assay. Adhered cells were incubated with various concentrations (0.1, 1, 10, 50, and 100 µg/mL) of NPs for 24–96 h.Cell viability assayThe viability of cells treated with CuO NPs (0.1–100 µg/ml) for 24–96 h was examined using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. The absorbance of the samples at 570 nm was examined using a microplate reader (Biotech Instruments, USA). Cells not treated with NPs served as a control with 100% cell viability. To eliminate NPs interference with the MTT reagent, a blank control was prepared with each concentration of NPs in complete growth media.The cells were subjected to neutral red uptake (NRU) assay, which was performed using protocols similar to those of the MTT assay, after a specified exposure time, following the manufacturer’s instructions (Neutral Red Cell Assay Kit, Himedia).Quantification of intracellular Cu2+ IonsThe intracellular Cu2+ ions (including both internalized nanoparticles and dissociated ions) were quantified by exposing 5 × 104 cells to CuO NPs. Cytotoxicity was first observed at 10 μg/mL after 24 h, and higher concentrations produced a clear dose-dependent increase in toxicity, as shown in Fig. 6. For quantification, cells were exposed to 10 μg/mL CuO NPs for 96 h, washed with phosphate-buffered saline (PBS), trypsinized, and subjected to ultracentrifugation. The resulting cell pellet was digested in nitric acid, and samples were analyzed using atomic absorption spectroscopy (AAS, Shimadzu AA-7000). The AAS values were normalized to the total protein content of each sample, determined by the Bradford assay, to correct for variations in cell number and biomass. Data are expressed as μg Cu per mg protein.Evaluation of cell membrane integrityThe culture supernatant of cells treated with CuO NPs for different durations (24–96 h) was analyzed using the lactate dehydrogenase (LDH) cell assay kit (Himedia), following the manufacturer’s instructions. The fluorescence intensity was assessed at excitation and emission wavelengths of 560 and 590 nm, respectively.Evaluation of mitochondrial membrane potential (MMP)The treated cells were incubated with 1 mM JC-10 dye in the dark for 60 min, following the manufacturer’s instructions. The MMP was determined by quantifying the red-to-green fluorescence intensity ratio.Evaluation of cellular reactive oxygen species (ROS) and reactive nitrogen species (RNS) productionPC-12 cells were treated with different concentrations of NPs, washed with 100 μL of PBS, and incubated with 100 μL of 2′,7′-dichlorodihydrofluorescein diacetate (DCFDA) in the dark for 45 min. The fluorescence intensity was measured using a microplate reader (Ex/Em = 485/535 nm) to determine the ROS levels.The Griess assay was used to quantify RNS. The culture supernatant of treated cells was subjected to the RNS detection assay with a nitric oxide estimation kit (Himedia), following the manufacturer’s instructions.Casp3 activation assayPC-12 cells were incubated with CuO NPs (0.1–100 μg/mL) in 96-well plates for 24 h. The fluorescence intensity (λex = 470, λem = 520 nm) was measured to quantify Casp3 activity using the Caspase-3 assay kit (DEVD-R110 Fluorometric Assay kit, Biotium), following the manufacturer’s instructions.Determination of pro-inflammatory cytokines levelsThe Il6 and Tnf levels were quantified using the enzyme-linked immunosorbent assay (ELISA) kits (Invitrogen, USA), following the manufacturer’s instructions. Briefly, the PC-12 cells were incubated with 0.1–100 μg/mL of CuO NPs for 24 h. ELISA was performed using the cell culture supernatant. The absorbance of the sample at 450 nm was measured using a multi-plate reader.AChE activity inhibition assayThe effect of CuO NPs on AChE activity was assessed using the acetylcholinesterase inhibitor screening kit (Sigma Aldrich), following the protocol based on the modified Ellman assay. Donepezil (half-maximal inhibitory concentration = 40 nM) served as the positive control.RNA isolation and quantitative real-time polymerase chain reaction (qRT-PCR)RNA was isolated from cells (3 × 106 cells) using Trizol (Invitrogen) and Qiagen TM RNeasy Plus (Qiagen, Valencia, CA), following the manufacturer’s instructions. The isolated RNA was quantified using NanoDrop (NanoDrop Technologies). RNA quality was determined using Bioanalyzer (Agilent Technologies).Total RNA was isolated from the differentiated PC-12 cells using the RNeasy® Plus Mini kit, following the manufacturer’s instructions. The RNA was reverse-transcribed using random primers (Table S1). qRT-PCR analysis was performed using SYBR green with QuantStudio3 (Applied Biosystems, USA). The expression levels of target genes were normalized to those of Gapdh (internal control).Effect of released Cu2+ ions on cellular toxicityThe concentration of Cu2+ions released from the CuO NPs was quantified. The NP sample (10 μg/mL) was incubated in a cell growth medium at 37 °C for 24–96 h. The dispersion was centrifuged at 15,000 rpm for 30 min at 4 °C. Only the supernatant with dissolved Cu2+ was used for downstream tests, and not the pellet. The supernatant was collected for quantifying the released ions via AAS. Experiments were performed in triplicate.Under the same experimental settings, the collected supernatant was subjected to in vitro assays (MTT, DCFDA, IL-6, Casp3 activity, and AChE inhibition) to assess the role of released ions in toxicity and directly compare with the neurotoxicity of CuO NPs.Statistical analysisThe data are presented as mean ± standard error of mean. The results of the non-treated control group were considered as the control values. The NP characterization data were analyzed using Origin Pro 8.0. (Origin Lab Corporation, MA, USA). Means were compared using two-way analysis of variance, followed Tukey’s multiple comparisons analysis using GraphPad Prism 9.4.1 (USA). All experiments were performed at least thrice.ResultsCharacterization of CuO NPsCuO NPs aggregated in the form of nanoclusters. At high magnification, CuO NPs exhibited a spherical shape (Fig. 1a–c). The mean particle diameter was measured using ImageJ (National Institutes of Health) and represented as histograms (Fig. 1d–f) along with hydrodynamic diameter (Fig. 1g–i). Elemental composition analysis determined using EDX spectroscopy confirmed the presence of Cu and O (Fig. S1). Table 1 shows the particle size, hydrodynamic diameter, and zeta potential of NPs.Fig. 1Morphological analysis of synthesized CuO NPs via FE-SEM (a–c). Histograms depicting mean particle diameter distribution (d–f). Hydrodynamic diameter determination by DLS in deionized water (g–i).Full size imageTable 1 The results for size analysis, hydrodynamic diameter and zeta potential of synthesized CuO NPs.Full size tableThe XRD data from JCPDS file 45-–0937 reveals that CuO NPs have a monoclinic crystal structure (Fig. 2a). This is corroborated by the observed diffraction peak positions and relative intensities, which are consistent with the known crystal structure of CuO. Some of the significant peak positions and their corresponding 2θ angles are 35.561° (002), 38.686° (111), 48.840° (202), 53.825° (020), 57.520° (202) and 61.257° (113). In the case of CuO, the (111) plane is the most intense peak in the XRD pattern. The average crystallite sizes for CuO-25, CuO-48, and CuO-P were determined to be 19, 28, and 25 nm, respectively. The variance between the grains and crystallites as well as the particle size determined by FE-SEM were, was in good agreement.Fig. 2Crystallographic analysis of synthesized CuO NPs by XRD (a). The result for CuO NP’s optical behavior via UV–visible absorption (b).Full size imageAs Cu2+ ions are in their crystal lattice, CuO NPs exhibit an absorption peak in the UV region. Cu2+ ions in the NP form a shoulder peak in the 550–600 nm region (Fig. 2b). The strength of this peak may increase with a decrease in particle size. According to a recent theoretical paradigm, the correlation between the cellular redox potential and the band gap of metal oxides can explain the oxidative stress-inducing and toxic capabilities of these materials20. The energy difference between the valence band and the conduction band, which is known as the optical band gap (Eg), was determined by extending the linear region of a graph depicting the square of the photon energy (hv)2 versus energy. The band gap for CuO-25, CuO-48, and CuO-P, which was calculated using the Tauc plot, was 1.84 eV, 1.25 eV, and 1.24 eV, respectively (Fig. S2).While FTIR does not provide comprehensive elemental composition or extensive impurity profiles for CuO nanoparticles, it remains an important instrument for determining their surface chemistry. Functional groups and surface-bound species such as hydroxyl, carboxyl, or other organic moieties, which may originate from synthesis precursors or stabilizing agents, can be identified by FTIR, which primarily detects molecular vibrations. FTIR is essential for understanding how surface chemistry affects interactions between nanoparticles and cells in the context of in- vitro investigations21.CuO NP composition and vibration modes were analyzed using FTIR spectroscopy (400–4000 cm−1) (Fig. 3). The vibrations at 480, 530, and 580 cm–1 are attributed to Cu–O, confirming the purity of CuO NPs. The following two peaks were observed due to the presence of moisture in CuO NPs: 3418 cm−1 (O–H stretching) and 1622 cm−1 (O–H bending). The asymmetric C-O in the CuO structure was connected to the two peaks at 1381 and 1060 cm−1. The PVP coating was the predominant element in the FTIR spectrum of CuO-P NPs. The PVP molecule comprises several functional groups, including carbonyl (C = O) and amide (C-N), exhibiting distinctive absorption peaks (Fig. S3). The broad peak around 3200–3500 cm−1 indicates the O–H stretching of hydroxyl groups in the PVP. The intensity peak around 1638 cm−1 indicates the N–H bending and C-N stretching of amide groups in PVP22. Multiple peaks around 1000–1400 cm−1 correspond to the C-O stretching of the PVP ether groups and the Cu–O stretching of the CuO NPs.Fig. 3The result for functional group analysis of CuO NPs by FTIR analysis.Full size imageEffect of CuO NPs on cell viabilityNPs time-dependently and concentration-dependently decreased cell viability (Fig. 4). In particular, CuO-25 and CuO-48 NPs significantly decreased cell viability at 1 μg/mL with the effect being evident after 72 h of exposure.Fig. 4The results of MTT assay indicating cell viability post CuO NPs exposure at different time intervals with reference to control (100% viability). Data is presented as the mean ± SEM (n = 3). *P < 0.05; **P < 0.01, ***P < 0.001, ns not significant.Full size imageTreatment with 10 μg/mL CuO-25 NPs decreased the viability of PC-12 cells to 65% (P < 0.05), which further decreased to 35% (P < 0.05) at 96 h. At a concentration of 50 μg/mL, the viability of CuO-25 NP-treated cells was < 30% (P < 0.05) and 12% (P < 0.01) at 24 and 96 h, respectively. Meanwhile, the viability of cells treated with 100 μg/mL CuO-25 NPs was 17% (P < 0.01) and 6% (P < 0.01) at 24 and 96 h, respectively. The viability of cells treated with 10 μg/mL CuO-48 NPs was 75% (P < 0.05) and 51% (P < 0.05) at 24 and 96 h, respectively. Additionally, the viability of cells treated with 50 μg/mL CuO-48 NPs was 40% (P < 0.05) and 19% (P < 0.01) at 24 and 96 h, respectively, while those treated with 100 μg/mL CuO-48 NPs was 29% (P < 0.05) and 12% (P < 0.01), respectively.The results of the NRU assay suggested that at 96 h, CuO-25 and CuO-48 NPs decreased the viability of cells to 63% (P < 0.05) and 74% (P < 0.05), respectively (Fig. 5). Treatment with 10 μg/mL CuO-25 NPs significantly decreased cell viability to 60% (P < 0.05) and 25% (P < 0.01) at 24 and 96 h, respectively. The viability of cells treated with 50 μg/mL CuO-25 NPs was 38% (P < 0.05) and 12% (P < 0.01) at 24 and 96 h, respectively, while that of cells treated with 100 μg/mL CuO-25 NPs was 25% (P < 0.01) and 8% (P < 0.01), respectively. At 24 h, the viability of cells treated with 10, 50, and 100 μg/mL CuO-48 NPs was 68% (P < 0.05), 49% (P < 0.05), and 37% (P < 0.05), respectively. Additionally, the viability of cells treated with 10, 50, and 100 μg/mL CuO-48 NPs for 96 h was 34% (P < 0.05), 20% (P < 0.01), and 14% (P < 0.01), respectively.Fig. 5The results of NRU assay indicating cell viability post CuO NPs exposure at different time intervals with reference to control (100% viability). Data is presented as the mean ± SEM (n = 3). *P < 0.05; **P < 0.01, ***P < 0.001, ns: not significant.Full size imageComparative analysis of the MTT and NRU assay results suggested that the cytotoxicity of CuO-25 NPs against PC-12 cells was significantly higher than that of CuO-48 NPs at concentrations ≥ 1 μg/mL.In the CuO-P NP-treated groups, the viability of cells treated with 10 μg/mL CuO-P NPs for 96 h was 55% and 68% in the NRU and MTT assays, respectively. At 100 μg/mL, CuO-P NPs decreased the cell viability to below 30% in both assays. Additionally, the cytotoxicity of PVP-coated NPs was significantly lower than that of uncoated NPs of similar sizes.Effect of CuO NPs on Intracellular Cu2+ ionsThe AAS analysis demonstrated that the intracellular uptake of CuO NPs exhibited a dose-dependent pattern (Fig. 6). CuO-25 NP-treated groups exhibited significantly higher (P < 0.05) uptake per mg of cellular protein than CuO-48 NP-treated (P < 0.05) and CuO-P NP-treated groups, indicating a significant influence of size on the cellular uptake of NPs.Fig. 6Cellular uptake of CuO by flame atomic absorption spectroscopy.Full size imageEffect of CuO NPs on cell membrane integrityDamaged cell membranes promote cell death, which manifests as enhanced extracellular release of LDH enzyme. CuO NPs enhanced the LDH levels in the PC-12 cell culture supernatant (Fig. 7). At 1 μg/mL, the LDH release in the CuO-25 NP-treated and CuO-48 NP-treated groups was 40% (P < 0.05) and 31% (P < 0.05), respectively, at 96 h.Fig. 7The results of LDH assay indicating cell membrane integrity post CuO NPs exposure at different time intervals with reference to untreated cells (control). Data is presented as the mean ± SEM (n = 3). *P < 0.05; **P < 0.01, ns not significant.Full size imageTreatment with 10 μg/mL CuO-25 NPs increased the LDH release rate to 43% (P < 0.05) and 71% (P < 0.05) at 24 and 96 h, respectively. The LDH release rate in cells treated with different concentrations of CuO-25 NPs at 24 and 96 h was as follows: 50 μg/mL: 60% (P < 0.05) and 91% (P < 0.01), respectively; 100 μg/mL: 75% (P < 0.01) and 91% (P < 0.01), respectively. Meanwhile, the LDH release rate at 24 and 96 h in cells treated with 50 μg/mL CuO-48 NPs was 46% (P < 0.05) and 78% (P < 0.01), respectively, while that in cells treated with 100 μg/mL CuO-48 NPs was 64% (P < 0.05) and 86% (P < 0.01), respectively. Comparative analysis revealed that the LDH release rate in the CuO-25 NP-treated group was significantly higher than that in the CuO-48 NP-treated group.The LDH release rate in the CuO-P NP-treated group was significantly lower than that in the CuO-48 NP-treated group. At the maximum treatment concentration (100 μg/mL), the LDH release in the CuO-P NP-treated group was 53.55% (P < 0.05) and 77% (P < 0.01) at 24 and 96 h, respectively, which was significantly lower than that in the uncoated NP-treated groups. This indicated that PVP coating mitigated the cytotoxicity of CuO NPs.Effect of CuO NPs on MMPTreatment with 1 μg/mL CuO-25 and CuO-48 NPs decreased the MMP to 63% (P < 0.05) and 73% (P < 0.05), respectively, at 96 h (Fig. 8). Compared with that in the control group, the MMP decreased to 62% (P < 0.05) and 39% (P < 0.01) in the 10 μg/mL CuO-25 NP-treated group at 24 and 96 h, respectively. Meanwhile, the MMP decreased to 37% (P < 0.05) and 17% (P < 0.01) in the 50 μg/mL CuO-25 NP-treated group at 24 and 96 h, respectively. In the 100 μg/mL CuO-25 NP-treated group, the MMP decreased to 24% (P < 0.01) and 10% (P < 0.01) at 24 and 96 h, respectively. The MMP decreased to 54% (P < 0.05), 25% (P < 0.05), and 15% (P < 0.01) in the groups treated with 10, 50, and 100 μg/mL CuO48 NPs, respectively, at 96 h. The MMP decline was mitigated in the CuO-P NP-treated groups. In particular, the MMP decreased to 74% (P < 0.05), 62% (P < 0.05), and 58% (P < 0.05) in the groups treated with 10, 50, and 100 μg/mL CuO-P NPs, respectively, which was significantly lower than that in the groups treated with uncoated NPs.Fig. 8The results for analysis of mitochondrial membrane potential (ΔΨm) via JC-10 assay post exposure to CuO NPs at different time intervals with reference to reference to control (100%). Data is presented as the mean ± SEM (n = 3). *P < 0.05; **P < 0.01, ***P < 0.001, ns not significant.Full size imageEffect of CuO NPs on intracellular ROS and RNS productionThe effects of 1 μg/mL CuO NP treatment for 72 h were evident in the cell viability assays. However, the ROS levels increased by 1.72-fold (P < 0.05) at 48 h only in the CuO-25 NP-treated group and remained steady till 96 h (Fig. 9). Treatment with 1 μg/mL CuO-48 NPs increased the ROS levels by 1.43-fold and 1.42-fold at 72 and 96 h, respectively. In contrast, CuO-P NPs did not increase the ROS levels.Fig. 9The result for quantification of reactive oxygen species levels via DCFDA assay post CuO NPs exposure at different time intervals with reference to reference to untreated control. Data is presented as the mean ± SEM (n = 3). *P < 0.05; **P < 0.01, ns not significant.Full size imageAt 10 μg/mL, CuO-25 NPs increased the ROS levels by 2.04-fold (P < 0.05) and 2.54-fold (P < 0.05) relative to the control at 24 and 96 h, respectively. Meanwhile, treatment with 50 and 100 μg/mL CuO-25 NPs increased the ROS levels by 3.32-fold (P < 0.01) and 4.07-fold (P < 0.01), respectively, at 96 h. The ROS levels increased by 2.44-fold (P < 0.01), 2.91-fold (P < 0.01), and 3.54-fold (P < 0.01) upon treatment with 10, 50, and 100 μg/mL CuO-48 NPs for 96 h, respectively, while those increased by 1.64-fold (P < 0.01), 2.35-fold (P < 0.01), and 3.02-fold (P < 0.01) upon treatment with 10, 50, and 100 μg/mL CuO-P NPs for 96 h, respectively.Treatment with 1 μg/mL CuO-25 NPs and CuO-48 NPs increased the intracellular RNS levels at 48 and 72 h, respectively. However, treatment with 1 μg/mL CuO-P NPs did not increase the RNS levels (Fig. 10). The RNS levels increased by 1.76-fold and 2.01-fold at 24 and 96 h, respectively, upon treatment with 10 μg/mL CuO-25 NPs. Treatment with 50 and 100 μg/mL CuO-25 NPs increased the RNS levels by 2.43-fold and 2.88-fold, respectively, at 96 h. The RNS levels increased by 1.7-fold (P < 0.05), 2.17-fold (P < 0.01), and 2.5-fold (P < 0.01) after treatment with 10, 50, and 100 µg/mL CuO-48 NPs, respectively, for 96 h. Treatment with 10, 50, and 100 µg/mL CuO-P NPs for 96 h increased the RNS levels by 1.46-fold (P < 0.05), 1.58-fold (P < 0.05), and 2.14-fold (P < 0.01), respectively. Thus, at concentrations of ≥ 1 μg/mL, CuO-25 NPs upregulated the ROS and RNS levels when compared with CuO-48 and CuO-P NPs.Fig. 10The result for the evaluation of extracellular reactive nitrogen species levels post CuO NPs exposure at different time intervals. Data is presented as the mean ± SEM (n = 3). *P < 0.05; **P < 0.01, ***P < 0.001, ns not significant.Full size imageEffect of CuO NPs on Il6 and Tnf levelsTreatment with 1, 10, 50 and 100 μg/mL for 24 h CuO-25 NPs upregulated the Il6 levels by 3.38-fold (P < 0.05), 6.38-fold (P < 0.05), 11.66-fold (P < 0.01), and 14.41-fold (P < 0.01), respectively (Fig. 11a). Meanwhile, the Il6 levels were upregulated by 2.52-fold (P < 0.05), 5.04-fold (P < 0.05), 10.09-fold (P < 0.01), and 12.57-fold (P < 0.01) upon treatment with 1, 10, 50, and 100 μg/mL for 24 h. The exposure of CuO-P NPs did not affect the Il6 levels at 1 μg/mL. However, the Il6 levels were upregulated by 3.85-fold (P < 0.05), 7.80-fold (P < 0.01), and 10.14-fold (P < 0.01) upon treatment with 10, 50, and 100 μg/mL CuO-P NPs, respectively.Fig. 11The assessment of IL-6 via ELISA after 24 hours of exposure to CuO nanoparticles at a concentration of 10 µg/ml (a) evaluation of TNF-α levels via ELISA after 24-h CuO NPs exposure at 10 µg/ml (b). The results for evaluation of caspase-3 activity after a 24-h exposure to CuO NPs at 10 µg/ml (c). The assessment of AChE inhibition after 24-h exposure to CuO NPs at 10 µg/ml (d). Data is presented as the mean ± SEM (n = 3). Significance levels are indicated as follows *P < 0.05; **P < 0.01, ***P < 0.001, ns not significant.Full size imageTreatment with 1 μg/mL CuO NPs did not affect the Tnf levels (Fig. 11b). The Tnf levels increased by 2.32-fold (P < 0.05), 3.60-fold (P < 0.05), and 4.47-fold (P < 0.001) upon treatment with 10, 50, and 100 μg/mL CuO-25 NPs, respectively. Meanwhile, treatment with 10, 50, and 100 μg/mL CuO-48 NPs upregulated the Tnf levels by 1.89-fold (P < 0.05), 2.91-fold (P < 0.05), and 3.80-fold (P < 0.05), respectively. CuO-P NPs did not affect the Tnf levels at 10 μg/mL. However, the Tnf levels increased by 2.43-fold (P < 0.05) and 3.29-fold (P < 0.05) upon treatment with 50 and 100 μg/mL CuO-P NPs, respectively.Effect of CuO NPs on cellular apoptosisCompared with negative control, treatment with 0, 50, and 100 μg/mL CuO-25 NPs for 24 h significantly upregulated Casp3 activity by 2.1-fold (P < 0.05), 2.56-fold (P < 0.001), and 2.93-fold (P < 0.001), respectively (Fig. 11c). Meanwhile, treatment with 10, 50, and 100 μg/mL CuO-25 NPs for 24 h increased Casp3 activity by 1.82-fold (P < 0.05), 2.32-fold (P < 0.05), and 2.68-fold (P < 0.001), respectively, which was significantly lower than that in the CuO-25 NP-treated group.At 10 μg/mL, CuO-P NPs did not significantly affect Casp3 activity when compared with CuO-48 NPs. However, treatment with 50 and 100 μg/mL CuO-P NPs upregulated the Casp3 activity by 1.86-fold (P < 0.05) and 2.4-fold (P < 0.05), respectively, which was significantly lower than that in the CuO-48 NP-treated group.Effect of CuO NPs on AChE activityCuO NPs concentration-dependently inhibited AChE activity (Fig. 11d). Treatment with 10, 50, and 100 μg/mL CuO-25 NPs significantly inhibited AChE activity to 63% (P < 0.05), 45% (P < 0.05), and 24% (P < 0.01), respectively. Meanwhile, treatment with 10, 50, and 100 μg/mL CuO-48 NPs suppressed AChE activity to 71% (P < 0.05), 57% (P < 0.05), and 35% (P < 0.05), respectively. CuO-P NPs did not inhibit AChE activity at 10 μg/mL. However, treatment with 50 and 100 μg/mL CuO-P NPs inhibited AChE activity to 73% (P < 0.05) and 59% (P < 0.05), respectively.Effect of CuO NPs on dopaminergic gene expressionTreatment with CuO-25, CuO-48, and CuO-P NPs for 24 h significantly upregulated the expression levels of Mao-A gene by 2.31-fold (P < 0.05), 2.21-fold, and 1.65-fold (P < 0.05), respectively (Fig. 12a). Additionally, treatment with CuO-25, CuO-48, and CuO-P NPs upregulated the expression of Th gene by 1.72-fold, 1.64-fold, and 1.66-fold, respectively. However, the Th gene expression levels were not significantly different between the CuO-25 NP-treated, CuO-48 NP-treated, and CuO-P NP-treated groups.Fig. 12The results for effect of CuO NPs exposure post 24 h at 10 µg/ml on gene expression of Mao-A, Th, and Comt (a). The results for effect of CuO NPs exposure on gene expression of α- synuclein, Gpr37, and Parkin (b) after 24 h at 10 µg/ml. Data is presented as the mean ± SEM (n = 3). *P < 0.05, ns not significant.Full size imageNext, three genes (Snca, Prkn, and Gpr37) associated with the etiology of NDs were analyzed. Treatment with CuO-25, CuO-48, and CuO-P NPs significantly upregulated α-synuclein expression by 3.53-fold (P < 0.05), 2.96-fold (P < 0.05), and twofold (P < 0.05), respectively (Fig. 12b). The Snca levels in the CuO-25 NP-treated group were significantly higher than those in the CuO-48 NP-treated group. However, unlike the Mao-A gene, a significant difference was observed between CuO-25 and CuO-48 NPs, with smaller NPs inducing significantly more up-regulation in gene expression compared to bigger counterparts.Effect of released Cu2+ ions on cellular toxicityThe behavior of CuO NPs in the cell culture media was influenced by their size, exposure time, and surface coating. The disintegration of these NPs varied depending on these factors. CuO-25 NPs were associated with enhanced release of Cu2+ ions when compared with CuO-48 NPs (Fig. S4). Furthermore, CuO-P NPs exhibited the least Cu2+ ion release.The released Cu2+ ions exerted neurotoxic effects at 10 μg/mL. At 96 h, cell viability decreased to 53% (P < 0.05) and 63% (P < 0.05) in the cells treated with culture media from the CuO-25 NP-treated and CuO-48 NP-treated groups, respectively (Fig. S5). Additionally, exposure to culture media from the CuO-25 NP-treated and CuO-48 NP-treated groups containing Cu2+ ions increased the ROS levels by 1.7-fold (P < 0.05) and 1.5-fold (P < 0.05), respectively (Fig. S6). The Il6 levels in the cells treated with culture media from the CuO-25 NP-treated and CuO-48 NP-treated groups increased by 1.85-fold (P < 0.05) and 1.65-fold (P < 0.05), respectively (Fig. S7). The Tnf levels in the cells treated with culture media from the CuO-25 NP-treated and CuO-48 NP-treated groups were upregulated by 3.1-fold (P < 0.05) and 2.42-fold (P < 0.05), respectively (Fig. S8). Furthermore, Casp3 activity was upregulated by 1.65-fold (P < 0.05) and 1.46-fold (P < 0.05) in the cells treated with culture media from the CuO-25 NP-treated and CuO-48 NP-treated groups, respectively (Fig. S9). The AChE activity decreased to 76% (P < 0.05) in the cells treated with culture media from the CuO-25 NP-treated group but was not affected in the cells treated with culture media from the CuO-48 NP-treated group (Fig. S10). Cell culture media containing the Cu2+ ions released from CuO-P NPs did not exert cytotoxic effects.DiscussionCuO NPs have various industrial and medicinal applications owing to their distinct physicochemical features. However, the adverse effects of CuO NPs on human health and the environment, especially their capacity to induce oxidative stress and damage biological components, have raised safety concerns23,24. The accumulation level of CuO NPs in the brain tissue is unclear. Particles with a large size are generally considered safe as they cannot enter the brain owing to the BBB. However, efforts are ongoing to determine the maximal size of NPs that can traverse the BBB.Similar to other metal NPs, CuO NPs are reported to exert neurotoxic effects. CuO NPs induce oxidative stress, which can damage macromolecules, such as lipids, proteins, and DNA25,26, adversely affecting metabolic activity, cell viability, and neuronal structure. However, conflicting findings have been reported on the cytotoxicity of CuO NPs. One study reported that CuO NPs (53 nm) do not decrease the viability of MCF-7 cells (breast carcinoma cells) below 50% even after treatment at a high concentration of 1600 μg/mL for 24 h27. Similar outcomes were observed in the N2A mouse neuroblastoma cell line exposed to CuO (102 ± 34 nm). The viability of neuroblastoma cells did not decrease even at 400 mg/L after 24 h of exposure28. This can be explained by the agglomeration of NPs in cells. Previous studies did not perform DLS or zeta potential analysis. The sonication of NPs before treatment is essential and can significantly affect the outcomes29 However, one study reported decreased viability and Cu accumulation in C6 glioma cells30. The findings of this study are consistent with those reported in the previous study. CuO NPs decreased cell viability in a time-dependent, concentration-dependent, size-dependent, and coating-dependent manner. Treatment with 1 μg/mL CuO-25 NPs for 72 h decreased cell viability by 25–30%, exerting the maximum cytotoxicity. Meanwhile, CuO-48 NPs exerted the second highest cytotoxic effects. The MTT, NRU, and LDH assay results revealed that PVP coating mitigated the cytotoxicity of NPs.Neuronal function is dependent on MMP. The primary source of energy for the neurons is the mitochondria in which energy is generated through oxidative phosphorylation31. MMP is critical for neuronal function and survival. Neurons require a steady supply of adenosine triphosphate (ATP) to maintain membrane potential and release neurotransmitters32. Mitochondrial dysfunction can lead to impaired ATP synthesis, oxidative stress, and cell death, as well as decreased MMP33. In this study, CuO NPs differentially impaired MMP. The MMP decreased to 63% and 75% upon treatment with 1 μg/mL CuO-25 NPs and CuO-48 NPs for 96 h, respectively. However, the MMP in the CuO-P NP-treated group was significantly higher than that in the uncoated NP-treated group.MMP decline is a major inducer of oxidative stress, disrupting the electron transport chain and increasing the formation of ROS, such as superoxide anions and hydroxyl radicals34. ROS can damage cellular macromolecules, such as lipids, proteins, and DNA, and activate various signaling pathways that result in cell death35. Additionally, nitric oxide (NO) is the major RNS produced by NO synthases. NO can combine with other molecules to form other RNS, such as peroxynitrite (ONOO-), nitrogen dioxide (NO2), and nitrous oxide (N2O3). RNS is involved in several physiological and pathological processes in brain cells, including neurotransmission, synaptic plasticity, and inflammation36. Excess RNS production can induce oxidative and nitrosative stress, which can promote cellular damage and contribute to the development of various neurological disorders37. This study demonstrated the correlation between MMP decline and elevated ROS in the effect of NPs, which was concentration-dependent and time-dependent. CuO NPs significantly upregulated the ROS and RNS levels in PC-12 cells. The ROS and RNS levels were the highest in the CuO-25 NP-treated group and the lowest in the CuO-P NP-treated group. These findings were consistent with those of previous studies, which reported that CuO NPs upregulated a specific pro-apoptotic gene (Bax) and downregulated an anti-apoptotic gene (Bcl2) in the mouse hippocampus HT-22 cell line. Additionally, CuO NPs downregulated the activity of various detoxification enzymes (glutathione S-transferase and superoxide dismutase)38. An in vivo study reported increased oxidative stress in the murine brain post-intra-nasal exposure to 23.5 nm CuO NPs39. The increased oxidative stress in the brain of rats orally administered with CuO NPs was due to the downregulation of cytochrome P-450 enzymes40.Pro-inflammatory cytokines, such as TNF-α and IL-6 have critical roles in immune responses. Additionally, pro-inflammatory cytokines are involved in various physiological and pathological processes in the brain, including learning and memory41. Under physiological conditions, TNF-α is produced in small quantities to regulate neurogenesis, synaptic plasticity, and neuroprotective mechanisms42. TNF-α and IL-6 activate various signaling pathways, including the NF-κB and MAPK pathways, promoting the production of ROS, the activation of adhesion molecules, and the migration of immune cells to the brain43. Furthermore, TNF-α and IL-6 may regulate synaptic plasticity, cognition, and neurotransmitter systems, such as glutamate and gamma-aminobutyric acid41. The dysregulation of TNF-α and IL-6 can lead to the development of various neurological disorders, such as Alzheimer’s and Parkinson’s disease44. RNS and ROS synergistically promote the release of pro-inflammatory mediators. Consistently, this study demonstrated that treatment with 10 μg/mL CuO NPs significantly upregulated the levels of the pro-inflammatory cytokines Il6 and Tnf in PC-12 cells. The upregulation levels of Tnf and Il6 were dependent on the size and coating of the NPs. In particular, the levels of Tnf and Il6 were the highest in the CuO-25 NP-treated group, followed by the CuO-48 NP-treated and CuO-P NP-treated groups. These results were supported by the findings of a previous study, which reported that prostaglandin E2, Tnf, and Il1b were upregulated in rat brain microvessel endothelial cells exposed to CuO NPs (40 and 60 nm) for 8 h45.The activation of caspase-3, which is involved in programmed cell death or apoptosis, is one of the downstream effects of the upregulation of pro-inflammatory cytokines. Under physiological conditions, caspase-3 exists as an inactive proenzyme. However, caspase-3 is cleaved and activated in response to specific signals46. Caspase-3 activation leads to DNA fragmentation and the breakdown of cellular components. This mechanism is tightly regulated to prevent excessive cell death, which can cause tissue damage and disease. CuO NPs upregulated caspase-3 activity in SH-SY5Y, H4, and PC-12 cells12. The findings of this study are consistent with these findings. Casp3 activity was the highest in the CuO-25 NP-treated group, followed by the CuO-48 NP-treated group (at NP concentrations of ≥ 10 μg/mL). In contrast, CuO-P NPs did not upregulate Casp3 activity at 10 μg/mL but upregulated its activity at 50 and 100 μg/mL.CuO NPs concentration-dependently inhibited AChE activity, which may be the potential mechanism underlying their neurotoxicity. Treatment with CuO-25 NPs at concentrations ≥ 10 μg/mL inhibited AChE activity. In contrast, CuO-48 and CuO-P NPs inhibited AChE activity to a lesser degree when compared with CuO-25 NPs, suggesting a significant size-dependent effect. This difference in AChE inhibitory activities can be attributed to differential NP size and surface characteristics, which can affect their interaction with the enzyme47. AChE inhibition is a biomarker of neurotoxicity as it prevents the breakdown of acetylcholine, a critical neurotransmitter for synaptic communication48. The prolonged presence of acetylcholine in the synaptic cleft can overstimulate neuronal circuits, contributing to nicotinic and muscarinic toxicity49. CuO-P NPs did not affect AChE activity at 10 μg/mL but significantly inhibited AChE activity at 50 and 100 μg/mL. This indicates that PVP coating may partially prevent AChE inhibition.CuO NP exposure upregulated the expression of genes involved in dopamine metabolism and NDs. In particular, CuO-25 and CuO-48 NPs upregulated the expression of MaoA gene. The upregulation of MaoA gene in the CuO-P NP-treated group was lower than that in the CuO-25 NP-treated and CuO-48 NP-treated groups. This suggests that surface coating moderately mitigates MaoA gene upregulation. The upregulation of MaoA gene alters dopamine metabolism, which may have implications for neurotransmitter modulation and potentially lead to neurotoxicity50. This is consistent with the findings of previous studies, which reported that metal ions, such as Cu, alter dopamine homeostasis and contribute to neurotoxicity7,51.Furthermore, CuO NP exposure upregulated α-synuclein expression. α-synuclein is associated with the etiology of neurodegenerative diseases, including Parkinson’s disease52. CuO-25 and CuO-48 NPs differentially upregulated α-synuclein expression. In particular, the upregulation of α-synuclein induced by CuO-25 NPs was higher than that induced by CuO-48 NPs, indicating increased neurotoxicity of particles with small sizes. Thus, CuO NPs may contribute to the development of neurodegenerative diseases through the upregulation of α-synuclein. This finding is consistent with the emerging evidence suggesting a role for metal NPs, including Cu, in the aggregation and toxicity of α-synuclein. However, further studies are needed to elucidate the underlying mechanisms and implications for neurodegenerative processes53,54.Excessive Cu2+ ion levels in the brain can lead to NDs, including Wilson’s disease. Cu2+ ions released from NPs can exert toxic effects by inducing oxidative stress, disrupting the neurotransmitter systems, and promoting protein aggregation. CuO NPs release Cu2+ ions when they come in contact with biological fluids or cells55. The free Cu2+ ions can interact with various biological components, exerting cytotoxic effects and inducing oxidative stress56. One study reported that the release of Cu2+ ions from NPs was higher than that from micrometer-sized particles. However, the cytotoxic effects of Cu2+ ions released from NPs were lower than those of NPs57. In this study, AAS analysis revealed that the release of Cu2+ ions in the cell culture media was upregulated in the CuO NP-treated group. The released Cu2+ ions mediated the neurotoxic effects of CuO NPs. The exposure of PC-12 cells to culture media (containing released Cu2+ ions) of CuO-25 NP-treated and CuO-48 NP-treated groups decreased cell viability and increased ROS, Il6, and Tnf levels. Additionally, the Cu2+ ions in the culture media upregulated the activity of caspase-3, a marker of apoptosis, confirming the negative effects of the ions on neural PC-12 cells. Thus, excess Cu2+ ions can lead to neurotoxicity and other NDs. However, further studies are needed to examine the effects of excess or deficient Cu on ND etiology58,59. Cu2+ ions released from CuO-25 NPs significantly inhibited AChE activity, indicating potential cholinergic signaling disruption, which is associated with various neurodegenerative disorders (NDs).The mechanisms underlying CuO NP-induced neurotoxicity are complex and involve various pathological processes, including oxidative stress, inflammation, mitochondrial dysfunction, and protein misfolding60,61. Oxidative stress is one of the key mechanisms through which CuO NPs exert neurotoxic effects. CuO NPs can promote protein oxidation, lipid peroxidation, and DNA damage through the Fenton and Haber–Weiss reactions62. ROS can destabilize MMP and trigger cytochrome-c release, leading to apoptosis induction. In this study, CuO NP-induced neurotoxicity was dependent on particle concentration, physical diameter, and surface coating. However, further studies are needed to elucidate the mechanism underlying CuO NP-induced neurotoxicity. These mechanisms have been implicated in the pathophysiology of a number of neurodegenerative illnesses, including as Parkinson’s, Alzheimer’s, and prion disorders, which are all marked by the build-up of aggregated or misfolded proteins63.It is important to acknowledge the limitations of the 2D cell culture model used in this study. While PC-12 cells provide valuable insights into nanoparticle-induced cytotoxicity and neurotoxic mechanisms, 2D systems lack the three-dimensional architecture, extracellular matrix interactions, and physiological gradients present in-vivo. These factors may influence the extent of CuO nanoparticle uptake, mitochondrial stress, and downstream neurotoxic outcomes observed in our study. For example, the dose-dependent decline in mitochondrial membrane potential and increase in intracellular Cu accumulation reported here may differ in magnitude or kinetics in more physiologically relevant systems. Future work using 3D neural cultures or in vivo models will be essential to validate and extend these findings.ConclusionsCuO NP-induced neurotoxic effects can be attributed to several factors, including concentration, particle size, surface coating, and the release of Cu2+ ions. These properties influence the cellular uptake of NPs and their interaction with cellular components as well as their propensity to induce oxidative stress. Although CuO NPs have shown neurotoxic effects in both in- vitro and in- vivo models, it is crucial to remember that the doses used in in- vitro studies frequently surpass physiologically relevant levels, especially given the limited ability of nanoparticles to translocate across the BBB after systemic exposure. According to the literature, due to the barrier’s limiting properties and intricate biodistribution dynamics, there is usually little real accumulation of nanoparticles in brain tissue after crossing the BBB. As a result, care must be taken when extrapolating concentration–response results from in- vitro experiments to in- vivo exposure situations. We acknowledge this limitation and emphasize that our findings represent a first step in understanding nanoparticle -neuron interactions, primarily providing mechanistic insights under carefully controlled experimental conditions.

    Data availability

    All data supporting the findings of this study are available within the paper and it’s Supplementary Information.
    ReferencesScheiber, I. F., Mercer, J. F. B. & Dringen, R. Metabolism and functions of copper in brain. Prog. Neurobiol. 116, 33–57 (2014).Article 
    CAS 
    PubMed 

    Google Scholar 
    Verma, N. & Kumar, N. Synthesis and biomedical applications of copper oxide nanoparticles: An expanding horizon. ACS Biomater. Sci. Eng. 5, 1170–1188 (2019).Article 
    CAS 
    PubMed 

    Google Scholar 
    Naz, S., Gul, A. & Zia, M. Toxicity of copper oxide nanoparticles: a review study. IET Nanobiotechnol. 14, 1–13 (2020).Article 
    PubMed 

    Google Scholar 
    Dey, A. et al. Biodistribution and toxickinetic variances of chemical and green Copper oxide nanoparticles in vitro and in vivo. J. Trace Elem. Med. Biol. 55, 154–169 (2019).Article 
    CAS 
    PubMed 

    Google Scholar 
    Joshi, A., Farber, K. & Scheiber, I. F. Neurotoxicity of copper and copper nanoparticles. in Advances in Neurotoxicology 5, 115–157 (Elsevier Inc., 2021).An emerging discipline. Bencsik, A., Lestaevel, P. & Guseva Canu, I. Nano- and neurotoxicology. Prog. Neurobiol. 160, 45–63 (2018).
    Google Scholar 
    Lutsenko, S., Washington-Hughes, C., Ralle, M. & Schmidt, K. Copper and the brain noradrenergic system. JBIC J. Biol. Inorg. Chem. 24, 1179–1188 (2019).Article 
    CAS 
    PubMed 

    Google Scholar 
    Prabhu, B. M., Ali, S. F., Murdock, R. C., Hussain, S. M. & Srivatsan, M. Copper nanoparticles exert size and concentration dependent toxicity on somatosensory neurons of rat. Nanotoxicology 4, 150–160 (2010).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Bulcke, F., Thiel, K. & Dringen, R. Uptake and toxicity of copper oxide nanoparticles in cultured primary brain astrocytes. Nanotoxicology 8, 775–785 (2014).CAS 
    PubMed 

    Google Scholar 
    Lian, D., Chonghua, Z., Wen, G., Hongwei, Z. & Xuetao, B. Label-free and dynamic monitoring of cytotoxicity to the blood–brain barrier cells treated with nanometre copper oxide. IET Nanobiotechnol. 11, 948–956 (2017).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Dobrucka, R., Kaczmarek, M., Łagiedo, M., Kielan, A. & Dlugaszewska, J. Evaluation of biologically synthesized Au-CuO and CuO-ZnO nanoparticles against glioma cells and microorganisms. Saudi Pharm. J. 27, 373–383 (2019).Article 
    PubMed 

    Google Scholar 
    Shi, Y., Pilozzi, A. R. & Huang, X. Exposure of CuO Nanoparticles Contributes to Cellular Apoptosis, Redox Stress, and Alzheimer’s Aβ Amyloidosis. Int. J. Environ. Res. Public Health 17, 1005 (2020).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Kukia, N. R., Abbasi, A. & Froushani, S. M. A. Copper oxide nanoparticles stimulate cytotoxicity and apoptosis in glial cancer cell line. Dhaka Univ. J. Pharm. Sci. 17, 105–111 (2018).Article 
    CAS 

    Google Scholar 
    Mou, X. et al. Exposure to CuO nanoparticles mediates NFκB activation and enhances amyloid precursor protein expression. Biomedicines 8, (2020).Bai, R. et al. Integrated analytical techniques with high sensitivity for studying brain translocation and potential impairment induced by intranasally instilled copper nanoparticles. Toxicol. Lett. 226, 70–80 (2014).Article 
    CAS 
    PubMed 

    Google Scholar 
    Bugata, L. S. P. et al. Acute and subacute oral toxicity of copper oxide nanoparticles in female albino Wistar rats. J. Appl. Toxicol. 39, 702–716 (2019).Article 
    CAS 
    PubMed 

    Google Scholar 
    Wang, X. et al. Observation of acetylcholinesterase in stress-induced depression phenotypes by two-photon fluorescence imaging in the mouse brain. J. Am. Chem. Soc. 141, 2061–2068 (2019).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Minigalieva, I. A. et al. In vivo toxicity of copper oxide, lead oxide and zinc oxide nanoparticles acting in different combinations and its attenuation with a complex of innocuous bio-protectors. Toxicology 380, 72–93 (2017).Article 
    CAS 
    PubMed 

    Google Scholar 
    Wang, Z., Zhao, J., Li, F., Gao, D. & Xing, B. Adsorption and inhibition of acetylcholinesterase by different nanoparticles. Chemosphere 77, 67–73 (2009).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Burello, E. & Worth, A. P. A theoretical framework for predicting the oxidative stress potential of oxide nanoparticles. Nanotoxicology 5, 228–235 (2011).Article 
    CAS 
    PubMed 

    Google Scholar 
    Mudunkotuwa, I. A., Minshid, A. Al & Grassian, V. H. ATR-FTIR spectroscopy as a tool to probe surface adsorption on nanoparticles at the liquid-solid interface in environmentally and biologically relevant media. Analyst 139, 870–881 (2014).Rahma, A. et al. Intermolecular interactions and the release pattern of electrospun Curcumin-Polyvinyl(pyrrolidone) Fiber. Biol. Pharm. Bull. 39, 163–173 (2016).Article 
    CAS 
    PubMed 

    Google Scholar 
    Pohanka, M. Copper and copper nanoparticles toxicity and their impact on basic functions in the body. Bratislava Med. J. 120, 397–409 (2019).Article 
    CAS 

    Google Scholar 
    Hejazy, M., Koohi, M. K., Pour, A. B. M. & Najafi, D. Toxicity of manufactured copper nanoparticles – A review. Nanomedicine Res. Journal 3, 1–9 (2018).CAS 

    Google Scholar 
    He, H. et al. Copper oxide nanoparticles induce oxidative dna damage and cell death via copper ion-mediated P38 MAPK activation in vascular endothelial cells. Int. J. Nanomedicine 15, 3291–3302 (2020).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Manke, A., Wang, L. & Rojanasakul, Y. Mechanisms of nanoparticle-induced oxidative stress and toxicity. Biomed Res. Int. 2013, (2013).Alishah, H., Pourseyedi, S., Ebrahimipour, S. Y., Mahani, S. E. & Rafiei, N. Green synthesis of starch-mediated CuO nanoparticles: preparation, characterization, antimicrobial activities and in vitro MTT assay against MCF-7 cell line. Rend. Lincei 28, 65–71 (2017).Article 

    Google Scholar 
    Perreault, F. et al. Genotoxic effects of copper oxide nanoparticles in Neuro 2A cell cultures. Sci. Total Environ. 441, 117–124 (2012).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Cronholm, P. et al. Effect of sonication and serum proteins on copper release from copper nanoparticles and the toxicity towards lung epithelial cells. Nanotoxicology 5, 269–281 (2011).Article 
    CAS 
    PubMed 

    Google Scholar 
    Joshi, A. et al. Uptake and toxicity of copper oxide nanoparticles in C6 glioma cells. Neurochem. Res. 41, 3004–3019 (2016).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Sakamuru, S., Zhao, J., Attene-Ramos, M. S. & Xia, M. Mitochondrial membrane potential assay. Methods Mol. Biol. 2474, 11–19 (2022).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Fields, R. D. Nonsynaptic and nonvesicular ATP release from neurons and relevance to neuron–glia signaling. Semin. Cell Dev. Biol. 22, 214–219 (2011).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Satoh, T., Enokido, Y., Aoshima, H., Uchiyama, Y. & Hatanaka, H. Changes in mitochondrial membrane potential during oxidative stress-induced apoptosis in PC12 cells. J. Neurosci. Res. 50, 413–420 (1997).3.0.CO;2-L” data-track-item_id=”10.1002/(SICI)1097-4547(19971101)50:3<413::AID-JNR7>3.0.CO;2-L” data-track-value=”article reference” data-track-action=”article reference” href=”https://doi.org/10.1002%2F%28SICI%291097-4547%2819971101%2950%3A3%3C413%3A%3AAID-JNR7%3E3.0.CO%3B2-L” aria-label=”Article reference 33″ data-doi=”10.1002/(SICI)1097-4547(19971101)50:3<413::AID-JNR7>3.0.CO;2-L”>Article 
    CAS 
    PubMed 

    Google Scholar 
    Maharjan, S., Oku, M., Tsuda, M., Hoseki, J. & Sakai, Y. Mitochondrial impairment triggers cytosolic oxidative stress and cell death following proteasome inhibition. Sci. Rep. 4, 1–11 (2014).Article 

    Google Scholar 
    Schieber, M. & Chandel, N. S. ROS Function in Redox Signaling and Oxidative Stress. Curr. Biol. 24, R453–R462 (2014).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Picón-Pagès, P., Garcia-Buendia, J. & Muñoz, F. J. Functions and dysfunctions of nitric oxide in brain. Biochim. Biophys. Acta – Mol. Basis Dis. 1865, 1949–1967 (2019).Yokoyama, H., Kuroiwa, H., Yano, R. & Araki, T. Targeting reactive oxygen species, reactive nitrogen species and inflammation in MPTP neurotoxicity and Parkinson’s disease. Neurol. Sci. 29, 293–301 (2008).Article 
    PubMed 

    Google Scholar 
    Niska, K., Santos-Martinez, M. J., Radomski, M. W. & Inkielewicz-Stepniak, I. CuO nanoparticles induce apoptosis by impairing the antioxidant defense and detoxification systems in the mouse hippocampal HT22 cell line: Protective effect of crocetin. Toxicol. Vitr. 29, 663–671 (2015).Article 
    CAS 

    Google Scholar 
    Liu, Y. et al. Oxidative stress and acute changes in murine brain tissues after nasal instillation of copper particles with different sizes. J. Nanosci. Nanotechnol. 14, 4534–4540 (2014).Article 
    CAS 
    PubMed 

    Google Scholar 
    Wang, Y. et al. Effect of copper nanoparticles on brain cytochrome-P450 enzymes in rats. Mol. Med. Rep. 20, 771–778 (2019).CAS 
    PubMed 

    Google Scholar 
    Bourgognon, J.-M. & Cavanagh, J. The role of cytokines in modulating learning and memory and brain plasticity. Brain Neurosci. Adv. 4, 239821282097980 (2020).Article 

    Google Scholar 
    Khairova, R. A., Machado-Vieira, R., Du, J. & Manji, H. K. A potential role for pro-inflammatory cytokines in regulating synaptic plasticity in major depressive disorder. Int. J. Neuropsychopharmacol. 12, 561 (2009).Article 
    CAS 
    PubMed 

    Google Scholar 
    Morgan, M. J. & Liu, Z. G. Crosstalk of reactive oxygen species and NF-κB signaling. Cell Res. 21, 103–115 (2011).Article 
    CAS 
    PubMed 

    Google Scholar 
    Smith, J. A., Das, A., Ray, S. K. & Banik, N. L. Role of pro-inflammatory cytokines released from microglia in neurodegenerative diseases. Brain Res. Bull. 87, 10–20 (2012).Article 
    CAS 
    PubMed 

    Google Scholar 
    Trickler, W. J. et al. Effects of copper nanoparticles on rat cerebral microvessel endothelial cells. Nanomedicine 7, 835–846 (2012).Article 
    CAS 
    PubMed 

    Google Scholar 
    Wang, X., Chen, S., Ma, G., Ye, M. & Lu, G. Involvement of proinflammatory factors, apoptosis, caspase-3 activation and Ca2+ disturbance in microglia activation-mediated dopaminergic cell degeneration. Mech. Ageing Dev. 126, 1241–1254 (2005).Article 
    CAS 
    PubMed 

    Google Scholar 
    Laban, B. B., Lazarević‐Pašti, T., Veljović, D., Marković, M. & Klekotka, U. Methionine Capped Nanoparticles as Acetylcholinesterase Inhibitors. Eur. J. Inorg. Chem. 26, (2023).Kalafatakis, K. et al. Acetylcholinesterase activity as a neurotoxicity marker within the context of experimentally-simulated hyperprolinaemia: An in vitro approach. J. Nat. Sci. Biol. Med. 6, S98–S101 (2015).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Binukumar, B. K. & Gill, K. D. Cellular and molecular mechanisms of dichlorvos neurotoxicity: cholinergic, nonchlolinergic, cell signaling, gene expression and therapeutic aspects. Indian J. Exp. Biol. 48, 697–709 (2010).CAS 
    PubMed 

    Google Scholar 
    Shih, J. C., Chen, K. & Ridd, M. J. MONOAMINE OXIDASE: From Genes to Behavior. Annu. Rev. Neurosci. 22, 197–217 (1999).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Bisaglia, M. & Bubacco, L. Copper ions and Parkinson’s disease: Why is homeostasis so relevant? Biomolecules 10, (2020).Stefanis, L. α-Synuclein in Parkinson’s disease. Cold Spring Harb. Perspect. Med. 2, 1–23 (2012).Article 

    Google Scholar 
    Valensin, D., Dell’Acqua, S., Kozlowski, H. & Casella, L. Coordination and redox properties of copper interaction with α-synuclein. J. Inorg. Biochem. 163, 292–300 (2016).Article 
    CAS 
    PubMed 

    Google Scholar 
    Okita, Y. et al. Metallothionein, copper and alpha-synuclein in alpha-synucleinopathies. Front. Neurosci. 11, 1–9 (2017).Article 

    Google Scholar 
    Jeong, J. et al. Differential contribution of constituent metal ions to the cytotoxic effects of fast-dissolving Metal-Oxide Nanoparticles. Front. Pharmacol. 9, 1–10 (2018).Article 

    Google Scholar 
    Moriwaki, H., Osborne, M. R. & Phillips, D. H. Effects of mixing metal ions on oxidative DNA damage mediated by a Fenton-type reduction. Toxicol. Vitr. 22, 36–44 (2008).Article 
    CAS 

    Google Scholar 
    Shi, M., Kwon, H. S., Peng, Z., Elder, A. & Yang, H. Effects of surface chemistry on the generation of reactive oxygen species by copper nanoparticles. ACS Nano 6, 2157–2164 (2012).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Desai, V. & Kaler, S. G. Role of copper in human neurological disorders. Am. J. Clin. Nutr. 88, 855S-858S (2008).Article 
    CAS 
    PubMed 

    Google Scholar 
    Bagheri, S., Squitti, R., Haertlé, T., Siotto, M. & Saboury, A. A. Role of copper in the onset of Alzheimer’s disease compared to other metals. Front. Aging Neurosci. 9, 1–15 (2018).Article 

    Google Scholar 
    Gupta, G. et al. /Zn superoxide dismutase 1 (SOD1)Copper oxide nanoparticles trigger macrophage cell death with misfolding of Cu. Part. Fibre Toxicol. 19, 1–27 (2022).Article 
    MathSciNet 

    Google Scholar 
    Chojnacka-Puchta, L., Sawicka, D., Zapor, L. & Miranowicz-Dzierzawska, K. Assessing cytotoxicity and endoplasmic reticulum stress in human blood–brain barrier cells due to silver and copper oxide nanoparticles. J. Appl. Genet. 66, 87–103 (2025).Article 
    CAS 
    PubMed 

    Google Scholar 
    Angelé-Martínez, C., Nguyen, K. V. T., Ameer, F. S., Anker, J. N. & Brumaghim, J. L. Reactive oxygen species generation by copper(II) oxide nanoparticles determined by DNA damage assays and EPR spectroscopy. Nanotoxicology 11, 278–288 (2017).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Koszła, O. & Sołek, P. Misfolding and aggregation in neurodegenerative diseases: protein quality control machinery as potential therapeutic clearance pathways. Cell Commun. Signal. 22, (2024).Download referencesAcknowledgementsThe authors thank Symbiosis International (Deemed University) for its facility, funding, and support. The authors thank the Director and Deputy Director of Symbiosis School of Biological Sciences for the discussions. The authors thank National Chemical Laboratory Venture Centre at Pune for FT-IR and AAS analysis. Authors than Pune University for FE-SEM studies.FundingOpen access funding provided by Symbiosis International (Deemed University). This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.Author informationAuthors and AffiliationsSymbiosis School of Biological Sciences, Faculty of Medical and Health Sciences, Symbiosis International (Deemed University), Lavale, Pune, 412115, IndiaJitendra Kumar Suthar, Anuradha Vaidya & Selvan RavindranSymbiosis Centre for Stem Cell Research, Symbiosis International (Deemed) University, Pune, IndiaAnuradha VaidyaAuthorsJitendra Kumar SutharView author publicationsSearch author on:PubMed Google ScholarAnuradha VaidyaView author publicationsSearch author on:PubMed Google ScholarSelvan RavindranView author publicationsSearch author on:PubMed Google ScholarContributionsJitendra Kumar Suthar: Conceptualization, Methodology, interpretation of data, writing an original draft. Anuradha Vaidya: Review, edit, and proofread. Selvan Ravindran: Writing, review & editing, interpretation of data, proofreading.Corresponding authorCorrespondence to
    Selvan Ravindran.Ethics declarations

    Competing interests
    The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

    Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
    Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
    Reprints and permissionsAbout this articleCite this articleSuthar, J.K., Vaidya, A. & Ravindran, S. Impact of particle size and surface modifications on the neurotoxic potential of copper oxide nanoparticles.
    Sci Rep 15, 44532 (2025). https://doi.org/10.1038/s41598-025-28114-2Download citationReceived: 17 April 2025Accepted: 07 November 2025Published: 24 December 2025Version of record: 24 December 2025DOI: https://doi.org/10.1038/s41598-025-28114-2Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative
    KeywordsCopperNanoparticlesNeurotoxicityPC-12ApoptosisOxidative stress More

  • in

    Population dynamics of seed and seedlings of Albizia procera (Roxb.) in Mizoram, India

    AbstractSeed production, dispersal, germination, and seedling establishment are critical life phases of a tree species. Understanding these processes is crucial to recognize species composition and directional change for ecosystem restoration. This study aimed to estimate seed production, dispersal, and fate of the seed population of A. procera (Roxb.) and evaluate its seedling growth performance in relation to microclimates under natural conditions. Seed production was estimated from 15 sampled trees for three years, while seed dispersal using circular sample plots and seed traps under mother trees. The mean seed production per tree was 145,352, 43,607, and 41,490 during year 2022, 2023, and 2024 respectively, and it significantly differed between years (F = 12.09, P < 0.0001) and among individual trees (F = 4.63, P < 0.0001) while correlated positively with tree traits. Additionally, the seed density decreases with increased distance from the mother trees. A majority of the seeds (55.02% in 2022, 54.25% in 2023 and 52.92 in 2024) fell under the mother tree, while seeds disappeared due to predation and other losses reached 56.60%, 48.00%, and 49.80%, respectively. Germination rate in natural conditions were moderate (39.00%, 47.90%, and 45.40% in 2022, 2023, and 2024, respectively), and less than half (46.07%) of the germinated seedlings survived after 14 months. Further, relative seedling growth rate was strongly influenced by soil temperature, moisture and relative humidity indicating their crucial role in successful establishment. The findings provide essential insights into the population dynamics of A. procera and can inform strategies for monitoring growth and restoring degraded lands.

    IntroductionAlbizia procera (Roxb.) (Family Fabaceae) is a fast growing, medium sized tree with an open canopy that can reach up to a height of 30 m1,2. It is native to India, America, Pakistan and Australia3 and found in tropical countries such as Myanmar, Thailand, Bangladesh, Malaysia, Laos, Cambodia, Vietnam4, Nepal, Indonesia, China, Andaman, Kenya, South Africa and Uganda1,5. Due to its varied adaptability and economic importance, A. procera is being planted in agroforestry systems including tea gardens, reforestation, afforestation and social forestry programmes for restoration of degraded lands6,7. Besides, the species is reported to have high medicinal properties8,9. Additionally, it is extensively used for other purposes, such as its leaves are consumed as a vegetable2 and its bark used as a fish poison5,10. These extensive uses can impact the species status and reduce its distribution in its native area.Seed dynamics and seedling establishment are critical life phases of a tree species11,12. Understanding these processes is crucial to recognize species recruitment and directional change in ecosystem restoration11. In contrast, seed production is considered a critical bottleneck of tree life cycle13 and its variation may affect regeneration and numerous biological phenomena including interaction between plants and animals, vegetation dynamics and nutrient cycle14. It may be limited by various external factors such as adverse climate, pollination failure, predation of flowers, fruits and leaves, wind speed etc15, and internal factors such as genetic condition, age and size of the tree16,17. These factors could lead to seed production reduction and significantly impact seed dynamics and seedling recruitment of a given species.Seed dispersal is another important link in the reproductive cycle of a tree from the end of the adult plant to the beginning of the new plant. The distance to which the seeds get dispersed has important ecological significance. For example, seed dispersal can avoid high density-dependent mortality in close proximity to the parent plants18, while it can facilitate maintaining species diversity and may induce spatial accumulation of seeds and seedlings in pioneer trees19. Soil seed banks, on the other hand, are influenced by seed production and dispersal and various governing factors that affect seed germination12,20. The fate of the seeds after dispersal from the parent plants can be influenced by several microclimatic and biotic factors21 and in turn primarily determine the future population structure. It is therefore of paramount importance to have a complete understanding of the population dynamics of a species from seed production to seedling recruitment. However, there are very limited studies in trees that deal with population dynamics of seeds and seedlings.A review of literature reveals that studies on seeds dynamics of A. procera is lacking or very limited. A. procera is a pioneer fast growing legume tree having hard seed coat and this property of seeds results in reduced germination in wild22,23 which may cause seedling population mosaic or heterogeneity in natural conditions24,25. The results have shown that seed pre-treatments of A. procera6,23,26, can bring seedling population homogeneity in nursery and plantations7. We hypothesized that seed production of A. procera is influenced by both intrinsic traits (e.g. tree size, and crown structure) and extrinsic environmental factors, and seedling success is modulated by microclimatic variation. Accordingly, the study aimed to quantify interannual variation in seed production and its relationship with tree characteristics, assess seed dispersal patterns and post-dispersal fate (germination, disappearance, and seed bank formation), and evaluate seedling growth and survival in relation to soil temperature, moisture, and humidity. To achieve these objectives, we conducted field experiments for A. procera species to address the following questions: (a) is there variation in seed production between years and among individual trees?, (b) how do the tree traits affect its seed production?, (c) do the seed dispersal distance influence seed germination?, and (d) which microclimatic variables affect the seedling growth the most?. The findings provide essential insights into population dynamics of the A. procera and can inform strategies for monitoring growth and restoring degraded forests.Materials and methodsStudy areaThe study was carried out at Mizoram University (MZU) campus (23°39’52”-23°48’43″N and 92°39’49”-. 92°46’39″E) (Fig. 1), located 15 km away from Aizawl, the capital city of Mizoram having elevations ranging from 300 m to 880 m asl. MZU campus encompasses roughly 980 acres of land, and harbors tropical wet evergreen forests, small biodiversity park and protected forested water catchment reserve in the north. Several streams flow through the campus27,28,29. The forest included diversified plants species, where Wapongnungsang et al.29 reported 384 plants belonging to 290 genera and 107 families and many grass species27. The area receives an average annual rainfall of approximately 1,850 mm, influenced heavily by the southwest monsoon, with most precipitation occurring between May and September. The mean annual temperature is about 21.6 °C, with summer highs reaching 30 °C and winter lows ranging between 10 and 12 °C (Meteorological data of Mizoram, Aizawl, 2023).Data collectionSeed productionSeed production of A. procera was estimated from randomly selected 15 individual trees between 2022 and 2024. However, for each tree, the number of branches (B), secondary branches (sub-branches (SB)), tertiary branches (sub-sub-branches (SSB)), and inflorescences (INF) per sub-sub-branches were recorded to calculate total seed produced per plant following formula (1). The seed were estimated during November and December of each year before seed maturation and fall16. Additionally, in each tree, the mean number of inflorescences per SSB was determined from random 5-SSB, while the average number of fruits per inflorescences was calculated from a random selection of 10 inflorescences. Conversely, the mean number of seeds per fruit were estimated from fifteen fruits (Fig. 2A) and the number of seeds per kilogram was estimated using the average weight of 25 seeds (Fig. 2B) following the formula (2).Seed dispersalTo study the seed dispersal, five individual fruiting trees were marked in the forest stand following the method suggested by Khan et al.30. The five trees selected for dispersal studies were a subset of initial 15 trees, chosen based on their health, canopy structure, accessibility, and adequate spacing (> 100 m) to avoid seed overlap. This smaller number was selected to allow intensive monitoring of seed-fall and dispersal pattern around each individual tree. Each selected individual tree base was considered as a center, of which concentric circles of 2.5 m circular increments were marked on the ground around the mother tree, extending outside the crown radius as deliberated by Sahoo and Lalfakawma17. The first circle had a radius of a minimum five meters and maximum was 25 m. In the meantime, the seeds that fell under the tree crown were not considered as dispersed seeds. However, each selected individual tree was visited at three-days interval over 8-weeks during seed-fall season (Fig. 2C). All seeds collected within each marked circle were counted separately and the seeds damaged during dispersal were excluded from the analysis.Fig. 1Study area: (a) India, (b) Mizoram state, (c) Aizawl district and Mizoram University location, and (d) sampled trees. The map was created using ArcGIS Pro and can be accessed via (https://pro.arcgis.com/en/pro-app/latest/get-started/download-arcgis-pro.htm).Full size imageFate of seed populations in the soilThe fate of seed population in the soil was assessed, where the number of seeds disappearing (fraction) during the seed fall period was studied. Five seed traps sizing (1 m x 1 m with 30 cm depth) were placed randomly on the ground under the mother tree crown at the beginning of seeds fall (Fig. 3A). The seed traps were visited every five-days until completed seed shedding following the method described by Sahoo and Lalfakawma17 and de Sá Dechoum, et al.31. Simultaneously, during each visit, all seeds in the traps were counted, and seeds damaged (due to insects and rodents) were separated from undamaged one. However, the difference between total produced seed and undamaged seeds that fell under the individual mother tree was estimated as the fraction of seeds loss during the seed fall period17.To estimate the germination of healthy seeds after fall and/or dispersal, the fate of undamaged seeds after dispersal was studied by sowing 20 seeds in plot sizing (1 m x 1 m) under the five sampled individual trees (Fig. 3B). Additionally, seed fate was assessed in relation to distance from the mother tree. At each distance e.g. (5, 15, 25, and 35 from the mother tree), plots sizing (1 m x 1 m each) were established (n = 3 replicates per distance). Twenty seeds were placed per plot (total 60 seeds per distance), and seed fate was recorded weekly for three months. The number of seeds that germinated, disappeared (e.g. translocated or consumed by dispersal agents) or rotted was recorded, and germination was defined as emergence of radicle visible through the seed coat.Soil seed bankThe study on soil seed soil bank was conducted at the end of rainy season following the method described by Souza et al.32. Four sites were selected using stratified sampling20, and from each site, five soil samples from area of (5 × 5 cm and 5 cm depth) were collected and bulked to estimate buried seed density. From each of the bulked sample, 100 g of soil was weighed and washed gently using a jet of water (Fig. 2D) following the procedure and method described by Padonou et al.20 to recover the seeds. The number of seeds found in the sample was then extrapolated to 1 m x 1 m area.Seedling growth performance in natural conditions and its relation to microclimate variablesTo evaluate seedling growth performance and its relation to microclimatic variables, 10 quadrats (1 m ×1 m) were laid randomly in the forest near A. procera species stand. All seedlings recruited from the dispersed seeds within quadrates were monitored for 14 months. The seedlings height and stem collar diameter (SCD) were measured at two-weeks interval; seedlings height was measured using 30 cm ruler, while SCD measured using digital Vernier caliper (150 mm). Monthly seedling growth increments were calculated along with relative growth. Seedling mortality and microclimatic variables such as soil temperature, soil moisture, and humidity were recorded monthly throughout the study period.In contrast, microclimate data were collected in each study plot, where soil temperature was measured using a stainless-steel dial thermometer (Model ST-9283B, 0.1 °C accuracy), soil moisture using digital soil moisture meter (Kelway HB-2), while relative humidity using portable digital hygrometer (HTC-1). One set of sensors was installed per plot (10 total). Sensors were positioned in the center of the (1 m x 1 m) seedling quadrats at 5 cm soil depth. Moreover, measurements were taken three days in a week during morning and afternoon, and averaged to monthly means for correlation with seedling growth increments.Population flux of A. procera
    The life cycle of A. procera species is computed from the average of seed production, seed dispersal and fate of seeds during seed-fall and post-seed fall, and soil seed bank over three years (2022–2024), as well as seedling growth, survival, and mortality for the period of 14 months. The mean seed production from 15 trees was estimated, along with seed dispersal and seed damage. Consequently, the percentage of seed disappeared was calculated from the fate of undamaged seed at varying distances from the mother trees, which also included seed germinated and contribution to the soil seed bank. Moreover, seedling survival and mortality were monitored over 14 months to determine overall seedling survival rates.Fig. 2A. procera seed at different stages from seed development to soil bank formation: (A) Fruits pod, (B) Mature seeds, (C) Seeds in the fruit pod after fall from the mother tree, and (D) excavation of buried seeds using wet sieving technique.Full size imageFig. 3Experiments layout: (A) Laying seed trap to collect the dispersed seed, (B) Monitoring the fate of undamaged seed population (towards germination, disappearance and rotted seeds) from plots under the mother tree and at varying distance from the mother tree.Full size imageData analysisAll field data collected on seed production, dispersal, fate of seed population, and seedling growth were compiled and organized into meaningful table for analysis. Data were summarized and expressed as mean and standard deviation (SD). The mean seed production for each year was computed using formula (1), and seed germinated and disappeared was calculated using formulas (3 and 4 respectively). While the spatial distribution of seeds (under the mother tree and at varying distances) was first determined by calculating the percentage of seeds in each location, this percentage was then used to allocate the annual mean seed production across the same spatial categories, yielding the estimated number of seeds at each distance as studied by researcher33. The relationship between seed production and tree characteristics such as DBH, number of branches, number of inflorescences, crown height, crown diameter was examined. In contrast, post-hoc Fisher LSD, Tukey, and One-way ANOVA test (Sig. 0.05) were used to examine the variation of seed production between years and among individual trees. Conversely, seedlings’ growth (height and SCD) were converted to monthly basis, and growth increment was calculated. Seedling growth rates were correlated with microclimate variables such as soil moisture, temperature, and humidity. In the meantime, post-hoc Tukey (Sig. 0.05) was used to examine the variation in monthly seedling growth increments during study period. One-way ANOVA (Sig. 0.05) was used to assess the variation in rainfall, temperature, and humidity between the study years. The seed population flux was computed as accumulative mean of seed production, dispersal, germination, disappearance, soil seed bank, and survival rates. All data analysis were performed using SPSS, Version 22.0, Jamovi (Version 2.5.6)34, OriginPro 2025 and Microsoft Excel (Version 365).Formulas and equationsAll abbreviations in formulas are defined as full form in (Table 1).$${text{Total seed production}}/{text{tree }} = {text{ B }} times {text{ SB }} times {text{ SSB }} times {text{ INF }} times {text{ FI }} times {text{ SF}}$$
    (1)
    $$Number{text{ }}of{text{ }}seeds{text{ }}in{text{ }}kg{text{ }} = frac{{1000~}}{{Mean~seeds~weight}}~$$
    (2)
    $${text{Germination }}left( % right){text{ }} = frac{{Number~of~seeds~germinated~}}{{total~~seeds~sown}} times 100$$
    (3)
    $${text{Seeds disappeared }}left( % right){text{ }} = frac{{Number~of~missing~or~predated~seeds~}}{{total~~seeds~sown}} times 100$$
    (4)
    Table 1 Summary of parameters and definitions.Full size tableResultsSeed production, seed dispersal, and soil seed bankThe mean seed production of A. procera significantly differed among individual trees (F = 4.63, P < 0.0001) (Fig. 4) and between years (F = 12.09, P < 0.0001) (Fig. 5). The mean seed production per individual tree was 145,352 seed/tree in 2022, dropping sharply to 43,607 in 2023, and 41,490 in 2024 (Table 2). The variation in seed production across the years can be driven from variation in number of inflorescences (F = 12.16, P < 0.0001) and number of fruits per inflorescences (F = 8.92, P < 0.000) (Table 2). Additionally, these declines correspond with notable interannual (Table 2) and monthly (Fig. 6) variation in rainfall, mean temperature, and humidity. In contrast, the result suggests that the pronounced decline in seed output was likely driven by climatic stress during the flowering and fruit set. The mean seed weight was 0.039 g resulting estimated 25,654 seed/kg (Table 2). A strong correlation was observed between mean seed production/tree and tree traits such as DBH, number of branches, inflorescences, fruits, crown height, and crown diameter (Table 3). The seed soil bank was relatively low across the years, without showing discernable variation among sample plots and years (Table 2).The number of seeds per unit area decreased as the dispersal distance from the mother tree increased showing a negative relation (reverse J-shaped curve) between seed density and dispersal distance across the years (Fig. 7). The distance to which the seeds got dispersed and the dispersed seed density varied significantly between years (F = 13.46, P < 0.0001, Table 4). A significant proportion of the total seeds produced (55.38% in 2022, 54.25% in 2023 and 52.92 in 2024) fell directly under the mother tree (Fig. 8) and the dispersed seed density gradually decreased with increased dispersal distance and this was true for all the years (Fig. 8). Further, the maximum distance to which the seeds were found dispersed was 22.50 m from the mother trees (Fig. 8).Table 2 Inter-annual variation in seed production, soil seed bank formation, and climatic factors.Full size tableFig. 4Mean seed production of A. procera among individual trees over three years (2022 to 2024). The error bar shows the variation of seed production between years within studied tree. The different alphabetic letters show significant variation of mean seed production between individual trees (where T1 to T15 are sampled trees).Full size imageFig. 5Variation in mean seed production across study years of A. procera. Post-hoc Tukey test (Sig. 0.05).Full size imageFig. 6Interannual variation in climatic factors: (a) monthly rainfall, (b) temperature, and (c) relative humidity.Full size imageTable 3 Correlation between mean seed production and tree characteristics.Full size tableTable 4 Seed density under the mother tree (UMT) and with different distances from the mother tree.Full size tableFig. 7Post-hoc Tukey test (Sig. 0.001) shows seed density from the mother tree (m) and variation in seed dispersal across the years. The error bar (standard deviation) showing the variation in number of seeds dispersed (m) within the sampled trees, while different letter shows significant difference in number of seeds within the year and between years across various distance from the mother tree.Full size imageFig. 8The percentage of seeds fall under the mother tree (UMT) and dispersed with distance (m).Full size imageFate of seed populations in the soilDuring the seed-fall period, more than 97% of the seeds that fell under the mother tree was found undamaged and viable when observed year-wise variation (Table 5). However, during post-seed-fall, the fraction of seed disappeared was 56.6% in 2022 while it decreased to 48% in 2023, and 49.80% in 2024. Similarly, the seed germination rates of the sown seeds were quite moderate (39.00%, 47.90% and 45.40% in 2022, 2023, and 2024 respectively) and seeds those got rotten were minimal (Table 5).Seed disappearance and germination in relation to distance from the mother treeThe seed disappearance due to predation was significantly (F = 9.61, P < 0.001) higher around the mother trees and it decreased gradually as the dispersal distance increased from the mother trees, a trend commonly observed in all the studied years. Our results showed that a very high proportion (82.33%) of the sown seeds disappeared within a 5 m circle while the disappeared seeds drastically reduced (67.33%) at a distance of 35 m from the mother tree (Fig. 9).Seedlings dynamics in the forest and its relation to microclimatic variablesDuring 14-months period, the seedlings of A. procera attained an average height and SCD of 22.50 cm and 3.25 mm respectively (Fig. 10). The species showed significant change in its growth increment with time (F = 3.29, P < 0.001 for seedling height, and F = 8.36, P < 0.001 for seedling collar diameter) (Fig. 11). The soil temperature and soil moisture showed seasonal variation; the highest soil temperature was recorded in June (26.86 °C), and soil moisture in July (70.30%) (Fig. 12). Among the microclimatic variables, soil moisture showed the strongest positive correlation with height growth and collar diameter (Table 6). The other parameters (relative humidity and soil temperature) too showed positive relationship with the seedling growth increment (height and collar diameter), however, soil pH was negatively related to these seedling attributes (Table 6).Population flux of A. procera
    The population flux of A. procera integrates seed production, seed dispersal, seed soil bank, and seedlings recruitment over three years (2022–2024). Based on 15-sampled trees, the mean seed production of A. procera over three years was 76,816 seeds/tree. Of these 97.63% seeds fell beneath the mother tree and dispersed at varying distances. However, during post-seed-fall, we found that 73.44%, 25.50% and 1.06% of the felled seeds got disappeared, germinated and stored in soil bank respectively. Among the germinated seeds, only 46.07% of seeds developed into seedlings that survived till the end of the experiment (Fig. 13). In the meantime, the seedling continues with vigorous growth (height and diameter) during the rainy season while dropping the leaves and stop shoot growth during dry season (Fig. 14).Table 5 Fate of seed population during seed-fall period and post seed-fall period, expressed as fractions (%) of the total seeds trapped (Mean ± SD).Full size tableFig. 9Germination and disappearance of seeds as influenced by distance from the mother tree.Full size imageFig. 10Mean seedling growth performance in the natural forest conditions over study period: (a) mean monthly seedling height (cm) and (b) mean monthly stem diameter (mm). Note: the box plot size shows the variation in seedling growth (N = 50 seedling in 10 quadrats), ANOVA test (Sig. 0.05).Full size imageFig. 11Post-hoc Tukey test (Sig. 0.05) for seedling growth increment: (a) mean seedling height and (b) stem diameter increment. The different alphabetic letters show significant variation in seedling growth increment between months during study period, while error bar (standard deviation) shows the variation in seedling growth (N = 50 seedling in 10 quadrats).Full size imageFig. 12Soil temperature and soil moisture in forest conditions during the study period.Full size imageFig. 13A. procera population flux: average of 3-years seed dynamics to seedling survival.Full size imageTable 6 Correlation between seedlings parameters and microclimate variables.Full size tableFig. 14A. procera seedling growth performance in natural forest conditions: (A) seedling during rainy season of the first year, (B) seedling response to dry winter season, and (C) seedling growth at the onset of rainy season.Full size imageDiscussionSeed productionSeed production of A. procera showed marked interannual variation, with a pronounced peak in 2022 followed by significant declines in 2023–2024. Similar variation in seed production across years have been reported in other trees species11,14,16,17,35,36,37. However, such variability is typical of tropical legumes and can be linked to fluctuations in rainfall and temperature that affect flowering and fruiting set. In contrast, the seed production variation was obviously related to the fruit loading of the species and the number of fruits bearing species in a given year. Several authors have suggested that seed production in a species will be influenced by several intrinsic and extrinsic factors during the flowering and seed setting period. For example, Iralu et al.12 reported that the annual rainfall acts as a limited factor for seed production and seedling survival while16,17,38 explained that seed production of tree species can be influenced by a variety of factors such as availability of resources, pollination failure, predation on flower, fruits, climatic condition, plant age and size. We observed a strong correlation between mean seed production and tree characteristics such as DBH, number of branches, number of inflorescences, and crown cover. On the other hand, larger trees with wider canopies tended to produce more seeds; however, this relationship must be interpreted cautiously because estimated production was driven from tree traits. Direct seed count in a validation subset are recommended to refine predictive models, and similarly used by several researchers11,13,14,16,36. In the meantime, Khan et al.30 observed that dominant trees species with large crowns, which receive a lot of light, tend to produce an optimum number of seeds. Conversely, the higher seed production occurring in warmer year, meanwhile increased temperature negatively affects seedling establishment39.Seed dispersalSeed dispersal is the movement of seed from the mother tree by help of different dispersal agents. For successful tree regeneration, it is important that the seeds should disperse to a safe location where they can germinate, survive and translate into mature plants. Thus, it determines the seeds distribution and trees during natural regeneration process which can be influenced by several factors (e.g. biological and environmental). In A. procera, seed dispersal occurred through a combination of wind dispersal and gravity. We found seed dispersal declined exponentially with distance, confirming that most seeds remain beneath or near the parent crown. A significant variation was observed in seed dispersal across the study years, this variation is mainly driven by seed production. We found the maximum seed dispersed up to 22.50 m from the mother tree, which indicated that the species dispersed its seed via explosion/gravity, while a small fraction of seeds was transported to distance places by wind. The limited seed dispersal pattern may increase density-dependent mortality and seed predation under the mother trees, a similar pattern reported for other Fabaceae species40,41. These findings suggest that the seeds with capsules tend to be exposed to secondary dispersal where wind moves it to another location. Seed dispersal, however, was highly restricted in this species, and for successful regeneration, the seeds need to be transported to far off places. The restricted seed dispersal by gravity nevertheless possesses some challenges for this species as it leads to overcrowding and competition for resources among the closely spaced individuals upon germination. To avoid for this challenge, the seeds of this species have wings/appendages which help them to be carried out and disperse to new locations. Several studies observed that seeds from same mother tree with varying seed mass are influenced by dispersal distance and differ significantly among tree species sharing the same dispersal mode41,42,43,44. Similar findings of decline the density of seed with increased dispersal distance from the mother trees have been observed for other trees species17,45. In contrast, Nathan et al.46 argued that long-dispersal distance is more common in open terrestrial landscapes and driven by migratory animals and wind. While Chen et al.35 and Kasi and Ramasubbu47 indicated that tree species that dispersed their seed by gravity are aggregated around the parent tree.Soil seed bankDespite high seed production, only a small proportion (≈ 1%) contributed to the persistent soil seed bank. Most seeds either germinated shortly after dispersal or disappeared due to predation and decay. Conversely, soil seed-bank contribution suggests that A. procera relies primarily on current-year’s recruitment rather than long-term soil storage. However, soil seed bank nevertheless plays a crucial role in regeneration of tree species, and its size is determined by seed dispersal and seed characteristics, and its ability to remain viable during the unfavourable conditions. Berihun et al.48 reported that seed bank can serve as a “memory” of past plant communities by containing seed from previous years and enhancing future plant communities. Meanwhile, the prevailing microclimate such as low temperature and moisture can have a bearing on soil seed bank by limiting germination49. The relatively low rates of seed germination under natural conditions in this species reveal that a larger fraction of the seeds remains viable-dormant in soil for a longer period, contributing to the resilience and persistent in its natural habitats. The size of the soil seed bank of a species in a given time is related to seed inputs (through seed rain) and seed outputs (through germination, predation and other losses) which are influenced by several factors including environmental and anthropogenics16,17,33,37. Though predation is reported to reduce the seed soil bank in forest floor and may affect survival and mortality12,50, the small viable seed bank can ensure the species’ survival and persistent in the nature.Seed disappearance and germination in relation to distance from the mother treeSeed disappearance in the present study was closely related to distance from the mother tree. The seeds that escaped or dispersed far off from the mother trees were found to be less predated than those which fell near the mother trees. Higher rate of predation and low survival near the mother tree were obviously due to density-dependent competition and predator preference, as also has been reported by several other workers16,17,33. Additionally, our results revealed that with increased distance from the mother tree the germination increased while seed disappearance decreased, in conformity with Souza et al.51. We observed that low seed germination compared to the fractions of the seeds that disappeared or got predated after seed fall in their natural habitats at varying distances from the mother trees. Majority of seeds that still remained in capsule could not transform into successful seedlings due to unfavourable environmental conditions (Fig. 15), which significantly reduce the seed germination and seedling recruitment in the forest. In contract, it reported that the ability of plant propagules to reach microhabitats with the adequate conditions for seed germination and establishment of sapling will have direct effect on the plant’s fitness52. The results clearly demonstrate that the seeds will have a better chance of survival if they are dispersed far away from the parent plants and get favorable germination conditions.Fig. 15Unfavorable environmental conditions limit the successful seed germination and lead to mortality.Full size imageSeedling dynamics and microclimate effectsSeedling growth and survival are often influenced by a combination of environmental factors such as soil temperature, moisture, light availability, and relative humidity. In the present study, A. procera seedlings reached an average height of 22.5 cm and SCD of 3.25 mm over study period and survival after germination was below 50% indicating the sensitivity of the seedling during early growth to various environmental stresses. This find support as the seedling growth of the species was positively correlated with soil moisture, soil temperature, and relative humidity, indicating that water availability is a key driver of recruitment success. Seasonal declines in soil moisture during the dry period sharply reduced the seedling survival, underscoring the vulnerability of young seedlings to drought. Similar findings are reported by Musa and Sahoo26, who reported that moisture availability and temperature significantly affect seedling performance of tropical species while Bebre et al.53 stated that multiple environmental factors in forests influence the seedling growth such as light, temperature, soil moisture, litter depth, intra and interspecific competition for various resources. Consequently, Greenwood et al.54 and Wieser et al.55 observed that higher soil temperature and moderate temperatures can improve physiological processes such as photosynthesis and nutrient uptake, leading to improved seedling establishment and growth. We found that seedling mortality during early growth was higher compared to their survival, a result that find support from the studies reported by Johnson et al.56. These findings further underscore the importance of the favorable microclimates during this critical stage of seedling development and can drastically affect survival rate among species57. In the meantime, the presence of canopy gaps, for instance, has been shown to provide improved conditions for seedling recruitment and survival due to increased light availability and moderated competition, and similar findings have been observed by several researchers58,59. While Awal60 and Kharuk et al.61 reported that soil temperature not only affects plant growth directly but also regulates microbial activity and nutrient cycling, which indirectly supports seedling vigor. Our study results confirm that microclimatic conditions particularly the soil temperature and moisture at the forest floor significantly influenced the seedling dynamics of A. procera.Study limitations and future directionsThis study excludes determining the viable-dormant seed fraction and association of various types of seed dormancy in the soil seed bank which would have provided better clues in understanding the role of the small-sized soil seed bank in regeneration of this fast-growing species in natural habitats. The reportedly under-estimation of dispersed seed density might have occurred due to limited seed trap coverage and secondary removal by predators. Incorporating camera monitoring or automatic traps could have enhanced better accuracy in estimation of various fractions of seeds during post seed-fall. Extending this study across climatic gradients and disturbance regimes would provide a more comprehensive understanding of A. procera regeneration ecology.ConclusionsSeed production and limited dispersal significantly constrain the natural regeneration of A. procera. Although the species produces abundant seeds, most fell beneath the mother tree canopies, where predation and low moisture reduced germination and survival. The low persistent soil seed bank ensure the regeneration of this species in its natural landscape, however, high mortality of seedling during early growth stage highlights a strong bottleneck in its seedling survival. To enhance the regeneration and restoration success, management should focus on assisted seed dispersal, moisture retention, and partial shade maintenance to improve seedling establishment.

    Data availability

    All data used/analyzed in this paper are available from corresponding author upon request.
    ReferencesHossen, K. & Kato-Noguchi, H. Evaluation of the allelopathic activity of albizia procera (Roxb.) Benth. As a potential source of bioherbicide to control weeds. Inter J. Plant. Biol. 13, 523–534. https://doi.org/10.3390/ijpb13040042 (2022).Article 
    CAS 

    Google Scholar 
    Kumar, S. V., Panwar, V. S., Deep, S., Verma, S. & P., & A brief review on phytopharmacological reports on albizia procera. Asian J. Pharm. Sci. 6, 144–149 (2020).
    Google Scholar 
    Pachuau, L., Lalhlenmawia, H. & Mazumder, B. Characteristics and composition of albizia procera (Roxb.) Benth gum. Ind. Crops Prod. 40, 90–95. https://doi.org/10.1016/j.indcrop.2012.03.003 (2012).Article 
    CAS 

    Google Scholar 
    Shah, S. S. et al. Preparation and characterization of manganese oxide nanoparticles-coated albizia procera derived carbon for electrochemical water oxidation. J. Mater. Sci. Mater. Electron. 30, 16087–16098. https://doi.org/10.1007/s10854-019-01979-6 (2019).Article 
    CAS 

    Google Scholar 
    Rai, E. & Ansari, S. A. A protocol for recovery of whole plants through in vitro shoots regeneration from leaflet explants of mature trees of Albizia procera (Roxb.) Benth. Indian For. 143, 487–498 (2017).
    Google Scholar 
    Azad, M. S., Biswas, R. K. & Matin, M. A. Seed germination of albizia procera (Roxb.) Benth. In bangladesh: a basis for seed source variation and pre-sowing treatment effect. Forestry Stud. China. 14, 124–130. https://doi.org/10.1007/s11632-012-0209-z (2012).Article 

    Google Scholar 
    Khurana, E. Influence of seed size on seedling growth of albizia procera under different soil water levels. Ann. Bot. 86, 1185–1192. https://doi.org/10.1006/anbo.2000.1288 (2000).Article 

    Google Scholar 
    Bunluepuech, K. & Tewtrakul, S. Anti-HIV-1 integrase activity of Thai medicinal plants in longevity preparations. Sonklanakarin J. Sci. Tech. 33, 693 (2011).CAS 

    Google Scholar 
    Khatoon, M. M., Khatun, M. H., Islam, M. E. & Parvin, M. S. Analgesic, antibacterial and central nervous system depressant activities of albizia procera leaves. Asian Pac. J. Trop. Biomed. 4, 279–284. https://doi.org/10.12980/APJTB.4.2014C348 (2014).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Rai, E., Bouddha, S. & Ansari, S. A. Biochemical investigations during in vitro adventitious shoot regeneration in leaflet explants from nodal segments of a mature albizia procera tree. J. Res. 27, 699–705. https://doi.org/10.1007/s11676-015-0196-8 (2015).Article 
    CAS 

    Google Scholar 
    Abiyu, A., Teketay, D., Glatzel, G. & Gratzer, G. Seed production, seed dispersal and seedling establishment of two Afromontane tree species in and around a church forest: implications for forest restoration. Ecosystem 3 https://doi.org/10.1186/s40663-016-0076-5 (2016).Iralu, V., Barbhuyan, H. S. A. & Upadhaya, K. Ecology of seed germination in threatened trees: A review. Ener Ecol. Environ. 4, 189–210. https://doi.org/10.1007/s40974-019-00121-w (2019).Article 

    Google Scholar 
    Mattana, E., Fenu, G. & Bacchetta, G. Seed production Andin situgermination of Lamyropsis microcephala (Asteraceae), a threatened mediterranean mountain species. Arct. Antarct. Alp. Res. 44, 343–349. https://doi.org/10.1657/1938-4246-44.3.343 (2012).Article 
    ADS 

    Google Scholar 
    Wesołowski, T., Rowiński, P. & Maziarz, M. Interannual variation in tree seed production in a primeval temperate forest: does masting prevail? Eur. J. Res. 134, 99–112. https://doi.org/10.1007/s10342-014-0836-0 (2014).Article 

    Google Scholar 
    Kameswara, Rao, N., Dulloo, M. E. & Engels, J. M. A review of factors that influence the production of quality seed for long-term conservation in genebanks. Genet. Resourc Crop Evol. 64, 1061–1074. https://doi.org/10.1007/s10722-016-0425-9 (2017).Article 
    CAS 

    Google Scholar 
    Barik, S. K., Tripathi, R. S., Pandey, H. N. & Rao, P. Tree regeneration in a subtropical humid forest: effect of cultural disturbance on seed production, dispersal and germination. J. Appl. Ecol. 33, 1551. https://doi.org/10.2307/2404793 (1996).Article 

    Google Scholar 
    Sahoo, U. K. & Lalfakawma, L. Population dynamics of Castanopsis tribuloides A. (DC) in an undisturbed and disturbed tropical forest stands of North-east India. J. Exp. Bio Agri Sci. 1, 454–463 (2013).
    Google Scholar 
    Herrero-Jáuregui, C., García-Fernández, C., Sist, P. L. J. & Casado, M. A. Recruitment dynamics of two low-density Neotropical multiple-use tree species. Plant. Ecol. 212, 1501–1512. https://doi.org/10.1007/s11258-011-9924-0 (2011).Article 

    Google Scholar 
    Howe, H. F. & Miriti, M. N. When seed dispersal matters. BioScience 54, 651. https://doi.org/10.1641/0006-3568(2004)054 (2004).Article 

    Google Scholar 
    Padonou, E. A., Akakpo, B. A., Tchigossou, B. & Djossa, B. Methods of soil seed bank estimation: a literature review proposing further work in Africa. iForest-Biogeosciences Forestry. 15, 121. https://doi.org/10.3832/ifor3850-015 (2022).Article 

    Google Scholar 
    Von Arx, G., Pannatier, E. G., Thimonier, A. & Rebetez, M. Microclimate in forests with varying leaf area index and soil moisture: potential implications for seedling establishment in a changing climate. J. Ecol. 101, 1201–1213. https://doi.org/10.1111/1365-2745.12121 (2013).Article 

    Google Scholar 
    Nongrum, A. & Kharlukhi, L. Effect of seed treatment for laboratory germination of albizia chinensis. J. Res. 24, 709–713. https://doi.org/10.1007/s11676-013-0408-z (2013).Article 
    CAS 

    Google Scholar 
    Sahoo, U. K. Effect of pretreatments on seed germination and seedling vigour of four different species of albizzia. Seed Res. 35, 124–128 (2007).
    Google Scholar 
    Daibes, L. F. et al. Fire and legume germination in a tropical savanna: ecological and historical factors. Ann. Bot. 123, 1219–1229. https://doi.org/10.1093/aob/mcz028 (2019).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Koutouan-Kontchoi, M. N., Phartyal, S. S., Rosbakh, S., Kouassi, E. K. & Poschlod, P. Seed dormancy and dormancy-breaking conditions of 12 West African Woody species with high reforestation potential in the forest-savanna ecotone of Côte d’ivoire. Seed Sci. Tech. 48, 101–116. https://doi.org/10.15258/sst.2020.48.1.12 (2020).Article 

    Google Scholar 
    Musa, F. I. & Sahoo, U. K. Impact of pre-treatments on Albizia procera and Albizia chinensis seed germination and early growth performance in Nursery, Mizoram, India. Sci. Tech. 21, 138–148. https://doi.org/10.1080/21580103.2025.2465604 (2025).Article 

    Google Scholar 
    Sailo, L., Solanki, G. S., Lalhruaizela, C. & Mizoram, A. I. Z. A. W. L. Avian Diversity in Mizoram University Campus, Sci. Tech. J. 7, 54–68 https://doi.org/10.22232/stj.2019.07.01.08 (2019).Article 

    Google Scholar 
    Zothanpuii, J. H., Gouda, S., Parida, A. & Solanki, G. S. A study on diversity of mammalian species using camera traps and associated vegetation in Mizoram university Campus, Aizawl, Mizoram. J. Threat Taxa. 12, 17330–17339. https://doi.org/10.11609/jott.6465.12.17.17330-17339 (2020).Article 

    Google Scholar 
    Wapongnungsang, N., Ovung, E. Y. & Tripathi, S. Assessment of tree diversity in tropical moist deciduous forest of Mizoram University, Northeast India. J. Appl & Nat. Sci. 13, 95–100. https://doi.org/10.31018/jans.v13i1.2436 (2021).Article 

    Google Scholar 
    Khan, M. L., Bhuyan, P. & Tripathi, R. S. Effects of forest disturbance on fruit set, seed dispersal and predation of Rudraksh (Elaeocarpus Ganitrus Roxb.) in Northeast India. Curr. Sci. 88, 133–142 (2005).
    Google Scholar 
    De Sá Dechoum, M., Rejmánek, M., Castellani, T. T., Zalba, S. & Martin Limited seed dispersal May explain differences in forest colonization by the Japanese raisin tree (Hovenia dulcis thunb.), an invasive alien tree in Southern Brazil. Trop. Con Sci. 8, 610–622. https://doi.org/10.1177/194008291500800303 (2015).Article 

    Google Scholar 
    Souza, M. et al. Weed emergence in a soil with cover crops in an agroecological no-tillage system. Planta Daninha. 36 https://doi.org/10.1590/S0100-8358201836010065 (2018).Musa, F. I. et al. Regeneration ecology of Lagerstroemia speciosa (L.) Pers in the Eastern Himalayan region of India. J. Environ. Sci. 41, 169–181. https://doi.org/10.7747/JFES.2025.41.2.169 (2025).Article 

    Google Scholar 
    Fox, J., Weisberg, S. & car Companion to Applied Regression. [R package]. (2023). Retrieved from https://cran.r-project.org/package=carChen, L. et al. Seed dispersal and seedling recruitment of trees at different successional stages in a temperate forest in Northeastern China. J. Plant. Ecol. 7, 337–346. https://doi.org/10.1093/jpe/rtt024 (2014).Article 

    Google Scholar 
    Calama, R. & Montero, G. Cone and seed production from stone pine (Pinus Pinea L.) stands in central range (Spain). Eur. J. Res. 126, 23–35. https://doi.org/10.1007/s10342-005-0100-8 (2005).Article 

    Google Scholar 
    Khan, M. L., Bhuyan, P., Shankar, U. & Todaria, N. P. Seed germination and seedling fitness in Mesua Ferrea L. in relation to fruit size and seed number per fruit. Acta Oecol. 20, 599–606 (1999).Article 
    ADS 

    Google Scholar 
    Owens, J. N. Constraints to seed production: temperate and tropical forest trees. Tree Physiol. 15, 477–484. https://doi.org/10.1093/treephys/15.7-8.477 (1995).Article 
    PubMed 

    Google Scholar 
    Ibáñez, I., Katz, D. S. W. & Lee, B. R. The contrasting effects of short-term climate change on the early recruitment of tree species. Oecologia 184, 701–713. https://doi.org/10.1007/s00442-017-3889-1 (2017).Article 
    ADS 
    PubMed 

    Google Scholar 
    Howe, H. F. & Smallwood, J. Ecology of seed dispersal. Ann. Rev. Ecol. Syst. 13, 201–228. https://doi.org/10.1146/annurev.es.13.110182.001221 (1982).Article 

    Google Scholar 
    Musa, F. I., Sahoo, U. K., Mohammed, E. M. I., Ramtharmawi & Swain, S. K. Seed population dynamics and early seedling growth of albizia chinensis (Osbeck) Merr. In the forests of humid subtropics of India. Int. J. Res. 2025(1), 3199982. https://doi.org/10.1155/ijfr/3199982 (2025).Article 

    Google Scholar 
    Jordano, P. What is long-distance dispersal? And a taxonomy of dispersal events. J. Ecol. 105, 75–84. https://doi.org/10.1111/1365-2745.12690 (2016).Article 

    Google Scholar 
    Muller-Landau, H. C., Wright, S. J., Calderón, O., Condit, R. & Hubbell, S. P. Interspecific variation in primary seed dispersal in a tropical forest. J. Ecol. 96, 653–667 (2008).Article 

    Google Scholar 
    Vittoz, P. & Engler, R. Seed dispersal distances: a typology based on dispersal modes and plant traits. Bot. Helv. 117, 109–124. https://doi.org/10.1007/s00035-007-0797-8 (2007).Article 

    Google Scholar 
    Miranda, A. et al. Traits of perch trees promote seed dispersal of endemic fleshy-fruit species in degraded areas of endangered mediterranean ecosystems. J. Arid Environ. 170, 103995. https://doi.org/10.1016/j.jaridenv.2019.103995 (2019).Article 
    ADS 

    Google Scholar 
    Nathan, R., Schurr, F. M., Spiegel, O., Steinitz, O. & Trakhtenbrot, A. Tsoar, A. Mechanisms of long-distance seed dispersal. Trends Ecol. Evol. 23 (11), 638–647 (2008).Article 
    PubMed 

    Google Scholar 
    Kasi, K. S. & Ramasubbu, R. Phenological patterns, fruit predation, and seed dispersal in two endangered trees (Elaeocarpus spp.) of Southern Western Ghats, India. J. Asia-Pac Biodivers. 14, 275–282. https://doi.org/10.1016/j.japb.2021.02.002 (2021).Article 

    Google Scholar 
    Berihun, T., Bekele, T., Lulekal, E. & Asfaw, Z. Study of the soil seed bank composition in Arjo-Diga humid Afromontane forest under different land use types and its implications for the restoration of degraded lands in Western Ethiopia. Int. J. Ecol. 2024(1), 5119487. https://doi.org/10.1155/2024/5119487 (2024).Article 

    Google Scholar 
    Saatkamp, A., Affre, L., Dutoit, T. & Poschlod, P. Germination traits explain soil seed persistence across species: the case of mediterranean annual plants in cereal fields. Ann. Bot. 107, 415–426. https://doi.org/10.1093/aob/mcq255 (2011).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Fenner, M. Ecology of seed banks. In: Routledge eBooks. 507–528. https://doi.org/10.1201/9780203740071-19 (2017). Souza, M. L., Silva, D. R. P. & Fantecelle, L. B. De Lemos Filho, J. P. Key factors affecting seed germination of Copaifera langsdorffii, a Neotropical tree. Acta Bot. Brasil. 29, 473–477. https://doi.org/10.1590/0102-33062015abb0084 (2015).Article 

    Google Scholar 
    Lomascolok, S. B. Indirect assessment of seed dispersal effectiveness for Solatium riparium (Solanaceae) based on habitat use and rate of fruitdisappearance. Ecol. Austral. 26 (1), 64–71. https://doi.org/10.25260/EA.16.26.1.0.101 (2016).Article 

    Google Scholar 
    Bebre, I., Riebl, H. & Annighöfer, P. Seedling growth and biomass production under different light availability levels and competition types. Forests 12 (1376). https://doi.org/10.3390/f12101376 (2021).Greenwood, S., Chen, J., Chen, C. & Jump, A. S. Temperature and sheltering determine patterns of seedling establishment in an advancing subtropical treeline. J. Veg. Sci. 26, 711–721. https://doi.org/10.1111/jvs.12269 (2015).Article 

    Google Scholar 
    Wieser, G., Oberhuber, W., Walder, L., Spieler, D. & Gruber, A. Photosynthetic temperature adaptation of Pinus cembra within the timberline ecotone of the central Austrian alps. Ann. Sci. 67, 201. https://doi.org/10.1051/forest/2009094 (2010).Article 
    CAS 

    Google Scholar 
    Johnson, D. M., McCulloh, K. A. & Reinhardt, K. Tree physiology/Tree physiology. (Dordrecht) 65–87. https://doi.org/10.1007/978-94-007-1242-3_3 (2011).Brown, N. The implications of climate and gap microclimate for seedling growth conditions in a Bornean lowland rain forest. J. Trop. Ecol. 9, 153–168. https://doi.org/10.1017/s0266467400007136 (1993).Article 
    CAS 

    Google Scholar 
    Charru, M., Seynave, I., Hervé, J. C. & Bontemps, J. D. Spatial patterns of historical growth changes in Norway Spruce across Western European mountains and the key effect of climate warming. Trees 28, 205–221. https://doi.org/10.1007/s00468-013-0943-4 (2014).Article 

    Google Scholar 
    Moore, C. E. et al. The effect of increasing temperature on crop photosynthesis: from enzymes to ecosystems. J. Exp. Bot. 72 (8), 2822–2844. https://doi.org/10.1093/jxb/erab090 (2021).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Awal, M. Effects of changes in soil temperature on seedling emergence and phenological development in field-grown stands of peanut (Arachis hypogaea). Environ. Exp. Bot. 47, 101–113. https://doi.org/10.1016/s0098-8472(01)00113-7 (2002).Article 

    Google Scholar 
    Kharuk, V. I., Ranson, K. J., Im, S. T. & Vdovin, A. S. Spatial distribution and Temporal dynamics of high-elevation forest stands in Southern Siberia. Glob Ecol. Biogeogr. 19, 822–830. https://doi.org/10.1111/j.1466-8238.2010.00555.x (2010).Article 

    Google Scholar 
    Download referencesAcknowledgementsThe first author (F.I.M.) thanks Indian Council for Cultural Relations (ICCR) for PhD scholarship and their valuable support and University of Blue Nile for granting study leave to carry out this research.FundingThe first author (F. I. M) grateful acknowledge the support of Indian Council for Cultural Relations (ICCR) under India-Africa Maitri Scholarship Scheme (Formerly called Africa Scholarship Scheme) through a PhD scholarship.Author informationAuthors and AffiliationsDepartment of Forestry, School of Earth Sciences and Natural Resource Management, Mizoram University, Aizawl, 796004, Mizoram, IndiaFaisal Ismail Musa, Uttam Kumar Sahoo, Uttam Thangjam & Mamta ChettriDepartment of Forestry, Faculty of Agriculture and Natural Resources, University of Blue Nile, Ad-Damzin, 26611, SudanFaisal Ismail MusaDepartment of Environmental Sciences, School of Earth Sciences and Natural Resource Management, Mizoram University, Aizawl, 796004, Mizoram, IndiaAhmed Abdallah Adam MohamedAuthorsFaisal Ismail MusaView author publicationsSearch author on:PubMed Google ScholarUttam Kumar SahooView author publicationsSearch author on:PubMed Google ScholarAhmed Abdallah Adam MohamedView author publicationsSearch author on:PubMed Google ScholarUttam ThangjamView author publicationsSearch author on:PubMed Google ScholarMamta ChettriView author publicationsSearch author on:PubMed Google ScholarContributionsFaisal Ismail Musa: Conceptualization, data curation, formal analysis, investigation, methodology, software, validation, visualization, writing—original draft; Uttam Kumar Sahoo: Conceptualization, methodology, supervision, writing—review & editing; Ahmed Abdallah Adam Mohamed: Data curation, formal analysis, validation, visualization, writing—review & editing; Uttam Thangjam: Methodology, Review & editing; Mamta Chettri writing—Review & editing. All authors read and approved the final version for publication.Corresponding authorCorrespondence to
    Faisal Ismail Musa.Ethics declarations

    Competing interests
    The authors declare no competing interests.

    Ethical approval
    The seeds used in this experiment were collected from the natural forests following the standard guidelines, and this research was carried out as per the local legislation and approval from the research ethical committee of Mizoram University, India.

    Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Rights and permissions
    Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
    Reprints and permissionsAbout this articleCite this articleMusa, F.I., Sahoo, U.K., Mohamed, A.A.A. et al. Population dynamics of seed and seedlings of Albizia procera (Roxb.) in Mizoram, India.
    Sci Rep 15, 44613 (2025). https://doi.org/10.1038/s41598-025-28651-wDownload citationReceived: 22 September 2025Accepted: 11 November 2025Published: 24 December 2025Version of record: 24 December 2025DOI: https://doi.org/10.1038/s41598-025-28651-wShare this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative
    Keywords
    Albzia procera
    Seed production and dispersalDispersal distanceGermination More

  • in

    Social familiarity shapes collective decision-making in response to looming stimuli in Medaka fish

    AbstractSocial familiarity within groups promotes behavioural synchrony and facilitates information transfer. Whether it shapes collective decision-making under predator threat is unknown. Here groups of six medaka (Oryzias latipes) familiarised for one month were used to test whether familiarisation promotes instantaneous collective decision-making in response to a looming stimulus (LS) mimicking a predator attack. First, we analysed behavioural transitions, defined as changes among three behavioural states: high-speed, normal and freezing-like before, during, and after LS in groups of six individuals. Individuals showing high-speed state in response to LS typically tended to shift to freezing-like state afterwards, whereas non-responders were more likely to maintain normal state. Group-level analysis revealed a bimodal distribution in the number of individuals exhibiting freezing-like state, with peaks at zero and six individuals, corresponding to ‘all non-freezing’ and ‘all-freezing’. Clustering analysis further identified three consistent group profiles: ‘freezing-dominant’, ‘non-freezing-dominant’, and ‘mixed-type’ based on behavioural tendencies across 10 trials. In contrast, in unfamiliar groups assembled immediately before testing, the ‘freezing-dominant’ profile was absent, and the distribution in the number of individuals exhibiting freezing-like state shifted to unimodal. In these groups, even at the individual level, responses more often showed a transition from high-speed state to normal state rather than freezing-like state. The results indicate that social familiarity promotes synchronous freezing-like state and consensus decisions under looming threat. Our study presents a behavioural assay for predator-evoked collective decision-making in a genetic model fish, providing a framework for future efforts to link behavioural ethology with neuroscience.

    Similar content being viewed by others

    Optimization, purification, and characterization of xylanase production by a newly isolated Trichoderma harzianum strain by a two-step statistical experimental design strategy

    Article
    Open access
    22 October 2022

    IoT and ML approach for ornamental fish behaviour analysis

    Article
    Open access
    05 December 2023

    Comparison of anxiety-like and social behaviour in medaka and zebrafish

    Article
    Open access
    28 June 2022

    IntroductionAnimals living in groups often exhibit synchronous behaviour in contexts such as migration, foraging, and predator avoidance. The emergence of coordinated behavioural patterns through interactions among individuals is termed collective behaviour, where group-level order and synchrony are thought to arise from local rules at the individual level, including attraction, alignment, and repulsion1. Collective behaviour includes collective decision-making, in which a group selects a single option from among multiple alternatives, and this phenomenon is observed in various contexts such as movement2,3, foraging4, and predator evasion5,6,7.In the context of predator avoidance, consensus formation during collective decision-making has been studied across various species. For instance, in sticklebacks, once the number of individuals escaping in a particular direction exceeds a threshold, the remaining group members tend to follow6. Similarly, simulations involving humans have shown that when a critical number of escape responses are observed, the group tends to adopt avoidance behaviour5. In elephants, the oldest female has been reported to be particularly sensitive to predator vocalisations and to influence the group’s decision to flee7. However, few studies have quantitatively investigated collective decision-making as an instantaneous group response under emergency conditions. In particular, how dynamic group-level responses to a rapidly approaching predator emerge remains poorly understood. While mechanisms underlying rapid individual decision-making in response to visual looming stimuli have been demonstrated8, systematic analyses of dynamic group-level responses remain scarce.Social familiarity has also been shown to affect individual recognition and interaction patterns, thereby influencing behaviour and information transmission9. At the dyadic level, social familiarisation enhances responsiveness to predators in predatory mites and reduces encounter frequency10. In sticklebacks, familiarisation reduced leadership tendencies in bold individuals, leading to more balanced coordination11. In cichlids, social familiarity promotes exploratory behaviour and reduces fear responses to novel stimuli12, while in zebrafish, the transmission of social fear is enhanced among familiar individuals13. At the group level, wild female guppies tend to associate with familiar individuals14, avoidance frequency in response to predator odour increases in fathead minnows15, and the latency to initiate avoidance of a predator model is reduced in brown trout16. In tropical damselfish, both responsiveness to fear stimuli and inter-individual information transmission are enhanced through familiarisation17. Overall, the literature indicates that social familiarity enhances alignment coordination and information transmission; however, how these factors influence collective decision-making remains unclear.Furthermore, most empirical research in this field has relied on observations in natural environments or on wild individuals6,15, yet relatively few studies have been conducted in controlled experimental settings. Although collective decision-making regarding movement direction has been demonstrated in zebrafish18, reports of such decisions in response to predators are lacking. Moreover, integrative frameworks that link decision-making mechanisms at both the individual and group levels with molecular and neural analysis, particularly in genetically tractable model organisms, are still lacking.To address these gaps, we focused on medaka Oryzias latipes, a well-established model organism in molecular genetics. Medaka are known to exhibit coordinated behaviour with conspecifics19 and to improve foraging efficiency through visual social learning20. Preliminary observations revealed that small groups of medaka responded synchronously to a human approach, either by showing ‘freezing after escape’ or by maintaining continuous movement without escape. Motivated by these findings, we aimed to develop a behavioural assay capable of quantitatively assessing instantaneous collective decision-making using a looming stimulus (LS) mimicking the sudden approach of a predator. LS has been used to elicit individual avoidance responses in mice21, zebrafish22, and fruit flies23. However, most previous studies have focused on individual-level decision-making8.Even in group contexts, the focus has remained on how individual responses are influenced by conspecifics23,24, with little attention paid to whether the group as a whole converges on a single collective choice.In this study, we optimised LS parameters and established a quantitative behavioural system capable of replicating the distinct response patterns during preliminary observations. We then tested whether medaka groups exhibit instantaneous collective decision-making in response to LS and examined the effects of social familiarity on the decision-making patterns. Our findings establish an experimental model for examining collective decisions under acute threat. They also lay the foundation for elucidating how familiarisation modulates synchrony and group-level decision-making processes in a genetically accessible model organism.MethodsEthics statementAll the methods in this study were carried out in accordance with relevant guidelines and regulations. The work in this paper was conducted using protocols specifically approved by the Animal Care and Use Committee of Tohoku University (permit number: 2022LsA-003). All efforts were made to minimise suffering following the NIH Guide for the Care and Use of Laboratory Animals. Fish and breeding conditions are described above. The study was carried out in compliance with the ARRIVE guidelines (https://arriveguidelines.org/arrive-guidelines).AnimalsMedaka (Oryzias latipes, fading strain) were obtained from Dr Tetsuro Takeuchi (Fig. 1a)25. This strain gradually loses body pigmentation in different body parts and at different timing among individuals, which initially appeared suitable for individual identification based on body colour. However, we later found that distinguishing individuals within a group of six remained difficult. During group observations, we consistently noted characteristic and reproducible collective behavioural patterns. As this strain has been maintained as a closed colony with limited genetic variation, we considered it an appropriate model for investigating the mechanisms underlying such stable group-level behavioural patterns. All individuals were hatched and bred in our laboratory. Medaka fish were maintained in groups of six individuals in plastic aquariums (22.6 cm× 14.6 cm× 14.5 cm, Sanko) or custom-made acrylic aquariums(22 cm × 14.5 cm × 14.5 cm) under controlled temperature (26 ± 1 °C) and light (14 h: 10 h light: dark) conditions. Every day, fish were fed brine shrimp between 12:00 and 13:00 and solid bait Otohime β−2 (Marubeni Nissin Feed, Tokyo, Japan) at least twice around 10:00 and 17:00 on weekdays. This study used individuals aged 2–9 months post hatch.Fig. 1Experimental design, state transition analysis, and statistical evaluation of freezing-like behaviour in medaka after looming stimulus. (a) Medaka fish (fading strain) (b) Groups of medaka were transferred, with their tanks, from the circulating water system to the apparatus after feeding. One hour later, the looming stimulus (LS) was presented five times at 30-minute intervals. The experiment was conducted over two consecutive days, giving ten LS. For analysis, three 10-second periods (before, during, and after each LS) were used, totalling 30 s per trial. (c) A state transition diagram visualises individual-level states (FS: freezing-like state, NS: normal state, and HS: high-speed state) across three intervals: before, during, and after LS. Node size represents the proportion of individuals, and numbers within nodes indicate counts. Node colours are red for HS, light blue for NS, and grey for FS. Numbers on edges indicate the transition probabilities, and edge thickness corresponds to the number of individuals. Edge colours indicate the originating state: red for transitions from HS, blue from NS, and grey from FS. (d) A bar plot showing the frequencies of transition patterns across the three intervals. Each pattern is categorised according to the sequence of states. Red bars indicate transitions to HS during LS followed by FS after LS. Blue bars indicate individuals that remained NS during and after LS. Grey bars represent all other patterns. To assess statistical significance, a binomial test with false discovery rate (FDR) correction was applied. Patterns with q < 0.001 are marked with ***. (e) The X-axis shows the number of individuals in FS after LS, and the Y-axis shows its frequency. The blue and red lines represent the observed data (17 groups) and the simulated data (17 groups × 1000 trials, seed = 1, …, 1000), respectively. A chi-square test result is shown in the graph (χ² (6) = 147, p < 0.001).Full size imageBehavioural experiments under familiar conditionsMedaka (Oryzias latipes) aged either one month or nine months were randomly selected to form groups. Ten groups of six individuals each were formed from one-month-old fish (N = 60), and seven groups were formed from nine-month-old fish (N = 42). These groups were formed without regard to sex, because the sex ratio of one-month-old fish could not be reliably determined. In contrast, the nine-month-old groups were adjusted to achieve a 1:1 sex ratio. Each group of six individuals was maintained separately for a month. In total, 17 groups of sexually mature individuals aged either two months or ten months that had undergone familiarisation, were used for experiments.Behavioural experiments under unfamiliar conditionsMedaka aged 4 or 9 months, reared in a recirculating aquaculture system, were randomly selected. From the 4-month-old fish (N = 36), six groups were formed, and from the 9-month-old fish (N = 36), another six groups, each consisting of six individuals, yielded a total of 12 groups. The sex ratio in each group was adjusted to 1:1. Group formation took place in the morning, and experiments were conducted 1 to 2 h after the afternoon feeding. Subsequent procedures were conducted in accordance with those described in the familiarisation experiments.Looming stimulus (LS)The looming stimulus (LS) was a visual stimulus mimicking an approaching predator, created using the “Zoom” animation function in Microsoft PowerPoint (Figure S1). The stimulus expanded to a width of 14.5 cm over 5.5 s, gradually darkened over 10 s, and remained black for 3 min (Figure S1). It was presented on an LCD monitor (EXLDH271DB, I-ODATA) mounted on the side of the aquarium (Figure S2). Both plastic and acrylic aquaria, also used as breeding tanks, were employed. Approximately 1–2 h after daytime feeding (12:00–13:00), each group was transferred in its breeding tank directly to the behavioural testing apparatus and the water level was adjusted to 6 cm. The LS was presented five times at 30-minute intervals for two consecutive days, beginning 1 h after transferring the aquarium from the rearing system to the testing apparatus (Fig. 1b).Behavioural recording and trackingBehaviour was recorded from above using an action camera (M80 Air, Apexcam, or HERO8, GoPro) at a resolution and frame rate of 4 K (30 fps) or 2.7 K (50 fps). Recordings lasted 5 min, beginning 2 min before the LS and ending 3 min after. For analysis, a 30-s segment was extracted for each trial: 10 s before, during, and after LS (Fig. 1b). Video files were extracted using QuickTime Player, converted to JPEG format using FFMPEG (v4.4.1), and subsequently converted to MP4 format at 5 fps. Tracking was performed using UMATracker26, and coordinate data were obtained with the UMATracker-Tracking tool, applying either the Pochi-Pochi (manual positioning) or Group Tracker GMM algorithm. Tracking errors, such as identity swaps, were manually corrected using UMATracker-TrackingCorrector.To convert pixel values to centimetres, the number of pixels along the centre of the long side of the aquarium was measured in ImageJ, which was based on the actual inner length (20.0–20.5 cm). Velocity (cm s⁻¹) was calculated from coordinate data (5 fps). A velocity matrix (6 individuals × 10 trials × 17 groups; 1020 × 150 frames) was compiled, and a moving average was applied (window size = 5) using pandas v1.4.0 to smooth short-term fluctuations.Definition of behavioural types and statesTo capture how individual fish responded to LS (looming stimulus) in terms of state transitions, we expressed behavioural responses as transition patterns across three intervals (before, during, and after LS). In total, 17 groups of six individuals each (N = 102) were tested. For each group, fish were transferred to the experimental arena and habituated for one hour, after which the looming stimulus was presented five times at 30-minute intervals over two consecutive days (Fig. 1b). This protocol yielded a total of 6 individuals × 10 trials × 17 groups of individual-level datasets for subsequent analyses. For this purpose, we first defined behavioural types for each interval based on velocity data. Using histograms and kernel density estimation (KDE) curves of velocity, we categorised behaviour into three types: ‘freezing-like behaviour’, ‘normal swimming’, and ‘high-speed swimming’. The rationale for these definitions is as follows. Some individuals exhibited freezing-like behaviour after LS. The velocity histogram for the post-LS interval showed a bimodal distribution with a trough at approximately 0.2 cm/s (Figure S3a). Therefore, frames with speeds below 0.2 cm/s were defined as ‘freezing-like behaviour’. During LS, escape-like responses characterised by high-speed swimming were observed. Such behaviours were rarely seen before the LS onset. Comparison of KDE curves for the pre-LS and LS intervals revealed minimal overlap above 6 cm/s (Figure S3b). Thus, frames with speeds of 6 cm/s or higher were defined as ‘high-speed swimming’. Frames with velocities between 0.2 cm/s and 6 cm/s were categorised as ‘normal swimming’. All histograms, KDE curves, and heat maps were generated using Python v3.8 and matplotlib v3.7.5.Based on speed-based behavioural types, we then defined the behavioural state for each 10-second interval (before LS, during LS, and after LS). An interval was categorised as a ‘freezing-like state (FS)’ if freezing-like behaviour persisted for ≥ 8 s, and as a ‘normal state (NS)’ if freezing-like behaviour lasted for < 2 s. During LS, escape behaviour occurred rapidly. Therefore, if high-speed swimming (≥ 6 cm/s) was sustained for 0.2 s (equivalent to one frame at 5 fps), we defined this as a ‘high-speed state (HS)’. This threshold reflects the minimum temporal resolution required to identify continuous motion.Statistical analysisCharacterisation of state transition patterns at the individual levelTo calculate the state transition probabilities for behavioural transitions before, during, and after LS, we constructed state transition matrices by counting the number of transitions between states and normalising each row, following established methods using Markov chain analysis27,28. To visualise the transition dynamics, we created state transition diagrams using python-graphviz v0.20.3, where each state (NS, FS, and HS) was represented by a node, with edges indicating transition probabilities.Furthermore, we used a binomial test to compare whether there were significantly more specific state transition patterns in the series of flows from before LS to after LS. In the binomial test, we set the null hypothesis that ‘the 27 behavioural patterns occur with equal probability (1/27)’ and performed a one-sided test. To control for type I errors due to multiple comparisons, we applied FDR correction to the binomial test results.Analysis of group-level freezing-like states after LSBased on the analysis of individual-level behavioural patterns, we next examined group-level freezing-like states (FS) after LS. For each trial, the number of individuals in the FS after LS was counted. To test whether synchronous FS occurred, virtual datasets were generated by randomly shuffling the states of each trial among groups (17 groups × 1,000 trials; seeds = 1, 2, …, 1,000). The proportions of individuals in each state across all trials were compared between the virtual datasets and the observational data using a chi-square test (scipy v1.10.1). The null hypothesis was defined as: “The presence or absence of the FS for each individual is independent, and synchronous FS for the entire group occur at random.”Classification of group response profilesTo classify these characteristics, we performed a principal component analysis (PCA) on 27 individual-level behavioural patterns from before to after the LS intervention. We then calculated the cumulative contribution rate and reduced the number of dimensions to the minimum required to explain > 95% of the variance. To visualise the cluster structure, we further projected the PCA-reduced data using UMAP29 (umap-learn v0.5.7). Classification was performed using spectral clustering (scikit-learn v1.2.2), and the optimal number of clusters was determined based on the silhouette coefficient. This coefficient approaches 1 when intra-cluster cohesion and inter-cluster separation are high; therefore, the number of clusters yielding the highest silhouette coefficient was selected.Comparison of state transition patterns at the individual level between clustersDifferences in the frequency of state transition patterns between clusters were evaluated using binomial tests with FDR correction, as described above.Analysis of group-level freezing-like statesTo examine differences in group-level freezing-like states (FS) across clusters and between familiar and unfamiliar groups after LS, we applied a generalized linear mixed model (GLMM)30. The dependent variable was the number of individuals exhibiting the FS (0–6) within each group. Cluster identity and the presence or absence of familiarisation were included as fixed effects. Experimental group identity, trial number, and group identity were incorporated as random effects to account for repeated measurements and inter-group variability. The model assumed a binomial distribution with a logit link function, which is appropriate for categorical or count data with hierarchical structure. Analyses were performed in Python v3.8.12. We used pyper v1.1.2 to call R v4.1.2, and the lme4 and multcomp packages for model fitting and post hoc tests. Tukey’s method was applied for multiple comparison correction.ResultsDetection of individual-level state transition characteristicsTo examine how individuals responded to the looming stimulus (LS), we analysed behavioural transitions across three intervals: before, during, and after LS. Fish behaviour was first classified into three types based on swimming velocity: freezing-like (< 0.2 cm/s), normal (0.2–6 cm/s), and high-speed (≥ 6 cm/s). Each 10-second interval was then categorized into one of three behavioural states—freezing-like, normal, or high-speed—according to duration thresholds (≥ 8 s freezing, < 2 s freezing, and ≥ 0.2 s high-speed). These definitions enabled consistent identification of state transitions across trials. Using these thresholds, each of the three temporal intervals was classified into one of the three behavioural states: ‘freezing-like state (FS)’, ‘normal state (NS)’, or ‘high-speed state (HS)’ (Figure S4).To examine how the behavioural states of individuals transitioned from before LS to during LS, and from during LS to after LS, we calculated state transition probabilities using a Markov chain and visualised them as a state transition diagram (Fig. 1c). Between the pre-LS and LS intervals, 39% of individuals transitioned from NS to HS, whereas 60% remained in NS. Among individuals that entered HS during LS, 70% transitioned to FS after LS. In contrast, individuals that remained in the NS during LS had a 76% probability of continuing in that state after LS. These findings suggest that individuals showing escape-like behaviour during LS tended to transition into FS after LS, whereas those unresponsive to LS generally maintained NS.To statistically evaluate trends in state transition patterns, we extracted behavioural state sequences across the three intervals (before, during, and after LS). Each sequence was expressed as a combination of the three defined behavioural states: HS, NS, and FS, resulting in 27 possible transition patterns. We quantified the frequency of each pattern across trials (Fig. 1d).The most frequent transition pattern was the maintenance of NS throughout the three intervals (NS→NS→NS; blue; q < 0.001). The second most frequent pattern was NS→HS→FS (red; q < 0.001), and the third was FS→HS→FS (red; q < 0.001), both involving a transition to HS during LS followed by FS: NS→HS→FS (red; q < 0.001) and FS→HS→FS (red; q < 0.001). Additional significantly overrepresented patterns were NS→HS→NS (grey; q < 0.001), NS→HS→FS (grey; q < 0.001), and FS→NS→NS (blue; q < 0.001), all of which exceeded the expected frequency under a uniform distribution (1/27).Among these six prominent transition patterns (Fig. 1d), NS→NS→NS and FS→NS→NS represent non-reactive behaviours where individuals maintained or returned to the NS during and after LS (blue). In contrast, NS→HS→FS and FS→HS→FS represent reactive responses, characterised by HS during LS followed by FS (red). These two reactive patterns accounted for 41% and 30% of all observations, respectively, and thus constitute the typical individual-level responses to LS stimulation. Notably, escape without subsequent FS (NS→HS→NS; approx. 9%) and no initial response followed by FS after LS (NS→NS→FS; approx. 8%) were also observed at appreciable frequencies.Population polarisation into synchronous freezing-like and non-freezing-like states after LSTo investigate whether medaka groups exhibited synchronous responses (either FS or non-FS) after LS, we counted the number of individuals exhibiting FS in each trial. The distribution was bimodal, with peaks at 0 and 6 individuals (Fig. 1e, blue line), suggesting that entire groups tended to respond uniformly.To determine whether this distribution could be explained by chance, we generated a virtual dataset by randomly shuffling the individual-level FS and non-FS classifications within groups of the same size (Fig. 1e, red line). This reconstructed the expected distribution under the assumption that individuals responded independently of one another. A chi-square test comparing the observed and expected distributions revealed a significant difference (χ² (6) = 147, p < 0.001). This result suggests that the strong bias towards either ‘all-freezing’ or ‘all non-freezing’ within groups after LS is unlikely to have occurred by chance alone. Instead, it indicates that individuals within a group reacted in a synchronous manner through social interaction.Group response profiles to LS classified into three typesDuring the behavioural experiments, we observed groups in which all individuals synchronously exhibited FS, as well as groups in which all individuals remained unresponsive and continued swimming. Moreover, the same groups tended to display similar response tendencies across repeated trials. Based on these preliminary observations, we hypothesised that groups exhibit consistent and characteristic behavioural tendencies, which we define as group response profiles. To evaluate this hypothesis, we classified groups according to their individual-level behavioural patterns. Specifically, behavioural data from 10 trials per group were aggregated, dimensionality reduction was performed using principal component analysis (PCA) followed by UMAP, and spectral clustering was applied to classify the groups.PCA indicated that 16 dimensions were required to exceed a cumulative variance contribution of 95%, and this was adopted as the optimal dimensionality (Figure S5). The reduced data were then projected into two dimensions using UMAP, revealing a clear distinct group-level structures (Fig. 2g). Spectral clustering, guided by the silhouette coefficient identified three as the optimal number of clusters (Figure S6). Accordingly, groups were classified into three clusters (Figure S7).Fig. 2State transition diagrams and frequencies of behavioural state transition patterns for each cluster. (a–c) State transition diagrams for each cluster were generated to represent individual-level behavioural states (FS: freezing-like state, NS: normal state, and HS: high-speed state) before, during, and after LS. All the visual elements are consistent with those in Fig. 1c. (a) State transition diagram for Cluster 0. b) State transition diagram for Cluster (1) (c) State transition diagram for Cluster (2) (d–f) Bar graphs showing the frequency of occurrence for each behavioural state transition pattern in each cluster. Details of colour coding and statistical tests are as described in Fig. 1d. (d) Distribution of transition pattern frequencies in Cluster 0. (e) Distribution of transition pattern frequencies in Cluster (1) (f) Distribution of transition pattern frequencies in Cluster (2) (g) The X-axis represents the first UMAP component and the Y-axis the second. Each point shows the group centroid, obtained by reducing the original 23 dimensions to 16 dimensions using PCA and further to two dimensions using UMAP. Colours indicate classification results based on spectral clustering. (h) The X-axis represents the number of individuals in the freezing-like state after LS, and the Y-axis represents the frequency of these counts across all trials. The lines correspond to the IDs of the three clusters. Tukey’s post hoc test based on a GLMM was performed, and the results of the cluster comparisons are shown within the graph.Full size imageTo analyse state transitions of individuals across the pre-, during-, and post-LS intervals in each cluster, we calculated state transition probabilities using a Markov chain and visualised them as state transition diagrams (Fig. 2a-c).In Cluster 0 (Fig. 2a), the probability of transitioning from NS to HS from before LS to during LS was 72.5%, while the transition probability from FS to HS was 76%. Individuals that exhibited HS during LS had an 89% probability of subsequently transitioning to the FS. In addition, even individuals that remained in NS during LS had an 80.4% probability of transitioning to FS afterwards. These results suggest that in Cluster 0, both responsive (HS) and unresponsive (NS) individuals tended to synchronise into a FS after LS. The most frequent pattern was NS→HS→FS (red; q < 0.001), and the second was FS→HS→FS (red; q < 0.001), both involving a transition to HS during LS followed by FS (Fig. 2d). These findings suggest that Cluster 0 corresponds to groups in which all individuals synchronise and exhibit FS after LS.In Cluster 1 (Fig. 2b), the probability of maintaining the NS from before LS to during LS was 86.2%, and individuals that remained unresponsive during LS continued NS after LS with a probability of 94.2%. The dominant patterns were those where individuals maintained NS throughout (NS→NS→NS and FS→NS→NS, q < 0.001; Fig. 2e, blue). These results suggest that Cluster 1 corresponds to groups in which all individuals synchronise and maintain NS after LS.In Cluster 2 (Fig. 2c, f), NS→NS→NS remained the most frequent pattern (q < 0.001). However, various other transitions were also observed, including NS→HS→NS (q < 0.001; grey), NS→HS→FS (q < 0.001; red), and NS→NS→FS (q < 0.001; grey). This diversity of transitions indicates that Cluster 2 represents a heterogeneous group response profile, reflecting a mixture of multiple individual-level response types rather than a single dominant pattern.Synchronisation of freezing-like and non-freezing-like states across the group profilesIndividual-level state transition analysis revealed that in Cluster 0, individuals frequently responded to the looming stimulus (LS) and then entered FS, whereas in Cluster 1, transitions in which individuals did not respond to LS and continued NS were predominant. We next examined whether all individuals in Cluster 0 synchronised to exhibit FS, and whether all individuals in Cluster 1 synchronised to continue NS. To this end, we counted the number of individuals in FS after LS for each group and compared these counts across clusters. A significant difference was observed in the number of individuals exhibiting FS (Fig. 2h, p < 0.001). In Cluster 0, the most frequent outcome was that all six individuals showed FS, whereas in Cluster 1, the most likely outcome was that no individual showed FS. In Cluster 2, the number of individuals exhibiting FS ranged mostly from zero to three, yielding a distribution distinct from both Clusters 0 and 1. These results indicate that in Cluster 0, individuals tended to synchronise to FS, whereas in Cluster 1 they synchronised to non-FS (continued NS). In contrast, Cluster 2 showed no clear synchronisation, with only a subset of individuals exhibiting FS after LS.Individual-level differences in behavioural transition patterns between familiar and unfamiliar groupsWe determined the behavioural patterns of individuals in the unfamiliar group (Figure S7-8), constructed a state transition diagram (Fig. 3a), and classified individual-level transition patterns into 27 categories, comparing their frequencies of occurrence (Fig. 3b). As in the familiar groups, the most frequent pattern was the non-reactive type (NS→NS→NS, blue; q < 0.001). The pattern (NS→HS→NS, grey) in which fish transitioned to HS during LS and returned to NS afterwards also appeared at a significantly high frequency (q < 0.001). In addition, the pattern in which fish transitioned to HS during LS and then entered FS after LS (NS→HS→FS, red) occurred significantly more often (q < 0.05). However, the patterns in which individuals entered FS after LS (NS→HS→FS, red; NS→NS→FS, grey), which were significantly enriched in the familiar groups, did not reach significance in the unfamiliar group. These findings suggest that individual-level transition patterns differed between the two conditions.Fig. 3State transition diagram of individual-level behaviour (Unfamiliar group). (a) For individual-level behaviours classified as FS: freezing-like state, NS: normal state, or HS: high-speed state, the state transition probabilities were shown from before to during LS and from during to after LS using a Markov chain. Details are as described in Fig. 1c. (b) Bar graphs showing transition patterns and their frequencies from before LS to during and after LS for FS, NS, and HS. Statistical significance is denoted as follows: ***q < 0.001, **q < 0.01, *q < 0.05, and no notation for q > 0.05. Details are as described in Fig. 1d.Full size imageAbsence of group-level synchronous freezing-like state in the unfamiliar groupTo test whether individuals exhibited synchronous responses after LS, we counted the number of individuals in FS per trial for each group and compared the observed data with control data generated by virtual shuffling, as in the familiar groups (Figure S9). In the observed data, the number of freezing-like individuals peaked at zero, and there was a significant difference in both the number and frequency of freezing-like individuals between the observed and virtual data (χ² (6) = 22.1, p < 0.01) (Figure S9). These results indicate that the peak at zero was not coincidental, but rather that all individuals within the unfamiliar group tended to exhibit synchronous non-FS, continuing to NS after LS.Differences in synchronous freezing-like states between familiar and unfamiliar groupsTo examine whether the occurrence of FS at the group level after LS differed depending on familiarisation, the number of individuals exhibiting FS per trial was counted for each group. The aggregated group-level data were then compared between the familiar and unfamiliar groups using GLMM (Figure S10). The results showed that, following LS, the familiar groups tended to have a higher number of individuals exhibiting FS (Figure S10, β = 2.02, p < 0.05) and displayed a bimodal distribution. In contrast, the unfamiliar group showed fewer freezing-like individuals and exhibited a unimodal distribution. This indicates that, unlike the familiar groups, the unfamiliar group lacked the peak where all six individuals exhibited FS after LS.Disappearance of the collective freezing in the unfamiliar groupTo clarify similarities and differences in collective behavioural patterns between familiar and unfamiliar groups, we integrated and analysed data from both conditions. Specifically, we performed dimensionality reduction and clustering based on 27 individual-level behavioural transition patterns. Principal component analysis (PCA) revealed that 17 dimensions were required to explain 95% of the variance, which was therefore set as the optimal number (Figure S11). The silhouette coefficient indicated that the optimal number of clusters was three (Figure S12). The clustering results were visualised using a two-dimensional UMAP embedding derived from the 17 principal components and classified into three clusters by spectral clustering (Fig. 4a). Groups in the familiar condition were distributed across all three clusters, whereas the unfamiliar groups were absent from Cluster 0, indicating a clear bias (Fig. 4b). We next verified that in Cluster 0, all individuals tended to exhibit synchronous FS, while in Cluster 1 they tended to exhibit synchronous non-FS (continued NS). In Cluster 2, synchrony was absent, with only a subset of individuals showing FS after LS. Importantly, the classification showed that the unfamiliar groups were not represented in Cluster 0, showing that the ‘freezing-dominant’ cluster was absent from their collective behavioural profiles (Fig. 4c). To evaluate whether this disappearance of the freezing-dominant response could be attributed to the immediate formation of unfamiliar groups, we compared the distributions of the number of freezing-like individuals between groups tested on the first and the following day. Although a significant difference was detected (χ² (6) = 14.2, p = 0.028), this was mainly due to a slight increase in groups with two or four freezing-like individuals, whereas the frequency of ‘all-freezing’ remained almost unchanged (Figure S13). These results suggest that handling or grouping stress immediately after formation had little effect, and that the disappearance of freezing-dominant is a robust feature of unfamiliar groups.Fig. 4Visualisation of dimensionality reduction using PCA and UMAP, and comparison of freezing-like states across clusters. (a–b) The X-axis represents the first UMAP component and the Y-axis the second. Each point corresponds to the group centroid. (a) Colours indicate cluster IDs obtained by spectral clustering. (b) Colours indicate familiarisation status: familiar groups (blue) and unfamiliar groups (orange). (c) Distribution of freezing-like states (FS) after LS exposure in the integrated dataset combining familiar and unfamiliar groups. Details are as described in Fig. 2h.Full size imageDiscussionIn this study, we established a quantitative behavioural assay to analyse collective decision-making in medaka (Oryzias latipes) in response to a looming stimulus (LS). In this system, small groups of medaka were presented with an LS that mimicked an approaching predator, and their collective behavioural choices were examined. Two dichotomous collective response patterns consistently emerged at the group level: ‘all-freezing’ and ‘all non-freezing’. Furthermore, the distribution of the number of FS individuals per trial was bimodal, with clear peaks at either zero or six individuals. These results demonstrate the presence of a dichotomous collective behavioural choice in medaka and validate as a robust tool for investigating collective decision-making under controlled laboratory conditions. Moreover, this assay will provide a platform for elucidating the genetic and neural bases of collective decision-making in vertebrates.Previous studies of collective decision-making under laboratory conditions have primarily used small fish species, such as sticklebacks and golden shiners2,4,6, often focusing on wild populations in ecological contexts. By contrast, our study employed medaka, a well-established genetic model organism, thereby enabling experimental systems in which genetic and environmental factors can be controlled. This approach enables the establishment of highly reproducible behavioural assays and allows for long-term monitoring of behavioural development, from the individual to the group level.Most previous studies of collective decision-making have focused on gradual responses to predators6,7. In contrast, our findings revealed a novel phenomenon: rapid collective responses to sudden visual threats. In coral reef fishes, escape responses of individuals can be predicted from the expansion rate of looming stimuli or the behaviour of neighbours24. However, how these individual responses converge into synchronous group-level behaviour remains unclear. Our results demonstrate that under time-constrained predatory threat, rapid collective decision-making can emerge, thereby complementing existing models of gradual escape behaviour.Our findings further suggest that a certain period of familiarisation is required for collective behavioural choices in response to the LS. In particular, in familiar groups, many individuals transitioned from HS during LS to FS afterwards, and entire groups tended to enter the FS. Such consistent behavioural synchrony was mainly observed in groups that had undergone sufficient familiarisation, suggesting that social familiarity may contribute to coordinated collective decisions. In our experiments, social familiarity increased the proportion of individuals that exhibited a FS after escape-like HS, indicating a change in behavioural regularity at the individual level. However, these individual-level changes alone cannot fully explain the emergence of dichotomous collective outcomes (all-freezing versus all non-freezing). It remains unclear whether familiarisation (1) enhanced each individual’s social sensitivity, making them more likely to be influenced by others, or (2) homogenised behavioural traits within groups. Distinguishing between these two possibilities was beyond the scope of this study. Future approaches incorporating longitudinal tracking of individually identified fish and quantitative measures of behavioural synchrony will be necessary to address this question.If explanation (1) is correct, repeated interactions during familiarisation may allow individuals to recognise and predict the behaviour of conspecifics, thereby strengthening group-level properties such as polarisation and alignment. Previous studies have reported various effects of social familiarity in fish. For example, in female guppies, 12 days of familiarisation led to preferential associations with familiar conspecifics31. Social familiarity has also been shown to promote group cohesion and alignment in guppies14 and to enhance information transfer under social threat in damselfish17. In addition, familiarisation may also induce social fear contagion. In zebrafish, individuals are known to switch from high-speed swimming to freezing when exposed to alarm cues from conspecific skin extracts13,32, suggesting that this behavioural pattern may be widespread among fishes. Moreover, zebrafish exhibit similar freezing-like responses when observing familiar conspecifics or groups displaying fear responses13,33.On the other hand, explanation (2), that familiarisation homogenises behavioural traits within groups, cannot be excluded. Previous studies have shown that bold individuals tend to maintain stable behavioural traits, whereas shy individuals are more plastic and influenced by social context. For instance, in guppies, bold individuals rely on their own information and explore independently, whereas shy individuals adjust their behaviour according to social information34. In sticklebacks, bold individuals also show stable exploratory behaviour, while shy individuals display behavioural plasticity and can change over time35. However, to our knowledge, no studies have directly demonstrated long-term homogenisation of behavioural traits caused by familiarisation in any animal species. Thus, we consider explanation (1) to be the more plausible mechanism underlying our findings.Our study therefore extends previous fish familiarity research, which has reported average increases in shoal cohesion, by showing that long-term social familiarity also structures the variability of collective decisions. Familiar groups not only became cohesive; they also differentiated into groups that reached full consensus (freezing-dominant or non-freezing-dominant) and groups that failed to do so (mixed-type), revealing that familiarity regulates the probability—rather than the inevitability—of consensus formation under threat.In summary, our results indicate that social familiarity promotes the dichotomisation of collective behavioural choices in medaka, and that factors such as social familiarity or changes in social sensitivity may contribute to this process. Although further investigation will be required to directly verify these mechanisms, our study provides a foundation for exploring how social experience shapes collective decision-making under time-constrained predatory threats in vertebrates.

    Data availability

    All data generated or analysed during this study are available from the corresponding author on reasonable request.
    ReferencesKrause, J. & Ruxton, G. D. in Living in Groups. (eds Krause, J. & Ruxton, G. D.) (Oxford University Press, 2002). https://doi.org/10.1093/oso/9780198508175.002.0001Couzin, I. et al. Uninformed individuals promote democratic consensus in animal groups. Science 334, 1578–1580 (2011).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Strandburg-Peshkin, A., Farine, D. R., Couzin, I. D. & Crofoot, M. C. Shared decision-making drives collective movement in wild baboons. Science 348, 1358–1361 (2015).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Ward, A. J., Krause, J. & Sumpter, D. J. Quorum decision-making in foraging fish shoals. PloS One. 7, e32411 (2012).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Clément, R. J. G., Wolf, M., Snijders, L., Krause, J. & Kurvers, R. H. J. M. Information transmission via movement behaviour improves decision accuracy in human groups. Anim. Behav. 105, 85–93 (2015).Article 

    Google Scholar 
    Ward, A. J. W., Sumpter, D. J. T., Couzin, I. D., Hart, P. J. B. & Krause, J. Quorum decision-making facilitates information transfer in fish shoals. Proc. Natl. Acad. Sci. 105, 6948–6953 (2008).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    McComb, K. et al. Leadership in elephants: the adaptive value of age. Proc. R Soc. B Biol. Sci. 278, 3270–3276 (2011).Article 

    Google Scholar 
    Shang, C. et al. Divergent midbrain circuits orchestrate escape and freezing responses to looming stimuli in mice. Nat. Commun. 9, 1232 (2018).Article 
    ADS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Ward, A. J. W. & Hart, P. J. B. The effects of kin and familiarity on interactions between fish. Fish. Fish. 4, 348–358 (2003).Article 

    Google Scholar 
    Strodl, M. A. & Schausberger, P. Social familiarity reduces reaction times and enhances survival of group-living predatory mites under the risk of predation. PLOS ONE. 7, e43590 (2012).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Riley, R. J., Kwon, Y. M., Manica, A. & Savage, J. L. Familiarity dampens the effect of boldness on coordination in three-spined sticklebacks. Behaviour 162, 191–206 (2025).Article 

    Google Scholar 
    Galhardo, L., Vitorino, A. & Oliveira, R. F. Social familiarity modulates personality trait in a cichlid fish. Biol. Lett. 8, 936–938 (2012).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Fernandes Silva, P., Garcia de Leaniz, C. & Luchiari, A. C. Fear contagion in zebrafish: a behaviour affected by familiarity. Anim. Behav. 153, 95–103 (2019).Article 

    Google Scholar 
    Davis, S., Lukeman, R., Schaerf, T. M. & Ward, A. J. W. Familiarity affects collective motion in shoals of guppies (Poecilia reticulata). R Soc. Open. Sci. 4, 170312 (2017).Article 
    ADS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Chivers, D., Brown, G. & Smith, J. Familiarity and shoal cohesion in Fathead minnows (Pimephales promelas): implications for antipredator behaviour. Can. J. Zool. 73, 955–960 (1995).Article 
    ADS 

    Google Scholar 
    Griffiths, S., Brockmark, S., Höjesjö, J. & Johnsson, J. Coping with divided attention: the advantage of familiarity. Proc. Biol. Sci. 271, 695–699 (2004).Article 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Nadler, L. E., McCormick, M. I., Johansen, J. L. & Domenici, P. Social familiarity improves fast-start escape performance in schooling fish. Commun. Biol. 4, 897 (2021).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Kadak, K. & Miller, N. Follow the straggler: zebrafish use a simple heuristic for collective decision-making. Proc. R. Soc. B Biol. Sci. 287:20202690 (2020).Imada, H. et al. Coordinated and cohesive movement of two small conspecific fish induced by eliciting a simultaneous optomotor response. PLoS One. 5, e11248 (2010).Article 
    ADS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Ochiai, T., Suehiro, Y., Nishinari, K., Kubo, T. & Takeuchi, H. A new data-mining method to search for behavioral properties that induce alignment and their involvement in social learning in Medaka fish (Oryzias latipes). PLoS One. 8, e71685 (2013).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Yilmaz, M. & Meister, M. Rapid innate defensive responses of mice to looming visual stimuli. Curr. Biol. 23, 2011–2015 (2013).Article 
    CAS 
    PubMed 

    Google Scholar 
    Temizer, I., Donovan, J. C., Baier, H. & Semmelhack, J. L. A visual pathway for looming-evoked escape in larval zebrafish. Curr. Biol. 25, 1823–1834 (2015).Article 
    CAS 
    PubMed 

    Google Scholar 
    Ferreira, C. H. & Moita, M. A. Behavioral and neuronal underpinnings of safety in numbers in fruit flies. Nat. Commun. 11, 4182 (2020).Article 
    ADS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Hein, A. M., Gil, M. A., Twomey, C. R., Couzin, I. D. & Levin, S. A. Conserved behavioral circuits govern high-speed decision-making in wild fish shoals. Proc. Natl. Acad. Sci. 115, 12224–12228 (2018).Article 
    ADS 
    CAS 
    PubMed 
    PubMed Central 

    Google Scholar 
    Takeuchi, T. & Manabe, E. Genetical study on the new mutant of the fading medaka, Oryzias latipes. Res. Bull. Shujitsu Women’s Coll. Shujitsu Jr Coll. 14, 1–18 (1984).
    Google Scholar 
    Yamanaka, O. & Takeuchi, R. UMATracker: an intuitive image-based tracking platform. J. Exp. Biol. 221, jeb182469 (2018).Article 
    PubMed 

    Google Scholar 
    Tuqan, M. & Porfiri, M. Mathematical modeling of zebrafish social behavior in response to acute caffeine administration. Front. Appl. Math. Stat. 7, 751351. https://doi.org/10.3389/fams.2021.751351 (2021).Article 
    PubMed 
    PubMed Central 

    Google Scholar 
    Syme, J., Kiszka, J. J. & Parra, G. J. Behavioural variation facilitates coexistence and explains the functions of mixed-species groups of sympatric delphinids. Anim. Behav. 210, 395–408 (2024).Article 

    Google Scholar 
    McInnes, L., Healy, J. & Melville, J. U. M. A. P. Uniform Manifold Approximation and Projection for dimension reduction. arXiv:1802.03426v3 (2020).Bolker, B. M. et al. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24, 127–135 (2009).Article 
    PubMed 

    Google Scholar 
    Griffiths, S. W. & Magurran, A. E. Familiarity in schooling fish: how long does it take to acquire? Anim. Behav. 53, 945–949 (1997).Article 

    Google Scholar 
    Masuda, M. et al. Identification of olfactory alarm substances in zebrafish. Curr. Biol. 34, 1377–1389e7 (2024).Article 
    CAS 
    PubMed 

    Google Scholar 
    Akinrinade, I. et al. Evolutionarily conserved role of Oxytocin in social fear contagion in zebrafish. Science 379, 1232–1237 (2023).Article 
    ADS 
    CAS 
    PubMed 

    Google Scholar 
    Trompf, L. & Brown, C. Personality affects learning and trade-offs between private and social information in guppies, Poecilia reticulata. Anim. Behav. 88, 99–106 (2013).Article 

    Google Scholar 
    Jolles, J., Briggs, H., Araya, Y. & Boogert, N. Personality, plasticity and predictability in sticklebacks: bold fish are less plastic and more predictable than shy fish. Anim. Behav. 154, 193–202 (2019).Article 

    Google Scholar 
    Download referencesAcknowledgementsWe thank Dr. Tetsuro Takeuchi for sharing the fading strain. We thank Drs. Masahiro Daimon and Masayuki Koganezawa for their advice on the development of the behavioural assay. We thank Drs. Ken-Ichiro Tsutsui, Hiromu Tanimoto, Jamie M. Kass and Towako Hiraki-Kajiyama for comments on the manuscript.FundingThis work was supported by the National Institute for Basic Biology Priority Collaborative Research Project 10–104 (to H.T.), 19–347 (to H.T.), and 21–335 (to H.T.); a grant for Joint Research (#01111904) by the National Institutes of Natural Sciences (to H.T.); Japan Society for the Promotion of Science (JSPS) KAKENHI Grants 21H04773 (to H.T.), 20H04925 (to H.T.), 18H02479 (to H.T.), 22H05483 (to H.T.), 23K27205 (to H.T.), 24H01216 (to H.T.) and 24K21957 (to H.T.). Takeda Science Foundation (to H.T.), and the natural science grant of the Mitsubishi Foundation (to H.T.); Japan Science and Technology Agency (JST) SPRING, Grant Number JPMJSP2114(to R.N.).Author informationAuthors and AffiliationsMolecular Ethology Laboratory, Graduate School of Life Science, Tohoku University, Sendai, 980-8577, JapanRyohei Nakahata & Hideaki TakeuchiDepartment of Cardiac Regeneration Biology, National Cerebral and Cardiovascular Centre, Osaka, 564-8565, JapanRyohei NakahataAuthorsRyohei NakahataView author publicationsSearch author on:PubMed Google ScholarHideaki TakeuchiView author publicationsSearch author on:PubMed Google ScholarContributionsR.N. and H.T. designed experiments. R.N. conducted experiments, wrote code and analysed data. R.N. and H.T. co-wrote and edited the paper and supervised the project.Corresponding authorsCorrespondence to
    Ryohei Nakahata or Hideaki Takeuchi.Ethics declarations

    Competing interests
    The authors declare no competing interests.

    Additional informationPublisher’s noteSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.Supplementary InformationBelow is the link to the electronic supplementary material.Supplementary Material 1Rights and permissions
    Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
    Reprints and permissionsAbout this articleCite this articleNakahata, R., Takeuchi, H. Social familiarity shapes collective decision-making in response to looming stimuli in Medaka fish.
    Sci Rep 15, 43650 (2025). https://doi.org/10.1038/s41598-025-30656-4Download citationReceived: 14 September 2025Accepted: 26 November 2025Published: 23 December 2025Version of record: 23 December 2025DOI: https://doi.org/10.1038/s41598-025-30656-4Share this articleAnyone you share the following link with will be able to read this content:Get shareable linkSorry, a shareable link is not currently available for this article.Copy shareable link to clipboard
    Provided by the Springer Nature SharedIt content-sharing initiative More