Philippa Kaur

Field samplingLake Masoko fish were chased into fixed gill nets and SCUBA by a team of professional divers at different target depths determined by diver depth gauge (12× male benthic, 12× male littoral). Riverine fish (11× Mbaka River and 1× Itupi river) were collected by local fishermen. On collection, all fish were euthanized using clove oil. Collection of wild fish was done in accordance with local regulations and permits in 2015, 2016, 2018 and 2019. On collection, fish were immediately photographed with color and metric scales, and tissues were dissected and stored in RNAlater (Sigma-Aldrich); some samples were first stored in ethanol. Only male specimens (showing bright nuptial coloration) were used in this study for the practical reason of avoiding any misassignment of individuals to the wrong population (only male individuals show clear differences in phenotypes and could therefore be reliably assigned to a population). Furthermore, we assumed that any epigenetic divergence relevant to speciation should be contributing to between-population differences in traits possessed by both sexes (habitat occupancy, diet). To investigate the role of epigenetics in phenotypic diversification and adaptation to different diets, homogenized liver tissue – a largely homogenous and key organ involved in dietary metabolism, hormone production and hematopoiesis – was used for all RNA-seq and WGBS experiments.Common-garden experimentCommon-garden fish were bred from wild-caught fish specimens, collected and imported at the same time by a team of professional aquarium fish collectors according to approved veterinary regulations of the University of Bangor, UK. Wild-caught fish were acclimatized to laboratory tanks and reared to produce first-generation (G1) common-garden fish, which were reared under the same controlled laboratory conditions in separate tanks (light–dark cycles, diet: algae flakes daily, 2–3 times weekly frozen diet) for approximately 6 months (post hatching). G1 adult males showing bright nuptial colors were culled at the same biological stages (6 months post hatching) using MS222 in accordance with the veterinary regulations of the University of Bangor, UK. Immediately on culling, fish were photographed and tissues collected and snap-frozen in tubes.Stable isotopesTo assess dietary/nutritional profiles in the three ecomorph populations, carbon (δ13C) and nitrogen (δ15N) isotope analysis of muscle samples (for the same individuals as RRBS; 12, 12 and 9 samples for benthic, littoral and riverine populations, respectively) was undertaken by elemental analyzer isotope ratio mass spectrometry by Iso-Analytical Limited. It is important to note that stable isotope analysis does not depend on the use of the same tissue as the ones used for the RRBS/WGBS samples45. Normality tests (Shapiro–Wilk, using the R package rstatix v.0.7.0), robust for small sample sizes, were performed to assess sample deviation from a Gaussian distribution. Levene’s test for homogeneity of variance was then performed (R package carData v.3.0-5) to test for homogeneity of variance across groups. Finally, Welch’s ANOVA was performed followed by Games–Howell all-pairs comparison tests with adjusted P value using Tukey’s method (rstatix v.0.7.0). Mean differences in isotope measurements and 95% CI mean differences were calculated using Dabestr v.0.3.0 with 5,000 bootstrapped resampling.Throughout this manuscript, all box plots are defined as follows: centre line, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers.RNA-seqNext-generation sequencing library preparationTotal RNA from liver tissues stored in RNAlater was extracted using a phenol/chloroform approach (TRIzol reagent; Sigma-Aldrich). Of note, when tissues for bisulphite sequencing samples were not available, additional wild-caught samples were used (Supplementary Table 3). The quality and quantity of RNA extraction were assessed using TapeStation (Agilent Technologies), Qubit and NanoDrop (Thermo Fisher Scientific). Next-generation sequencing (NGS) libraries were prepared using poly(A) tail-isolated RNA fraction and sequenced on a NovaSeq system (S4; paired-end 100/150 bp; Supplementary Table 3), yielding on average 32.9 ± 3.9 Mio reads.Read alignment and differential gene expression analysisAdaptor sequence in reads, low-quality bases (Phred score  More

Bacterial strains and mediaA collection of 185 genome-sequenced bacterial isolates, described previously14, was utilized to assemble the synthetic community used in this work. These isolates were obtained from surface-sterilized Brassicaceae roots, primarily Arabidopsis thaliana, grown in two soils from North Carolina, USA35. This isolate collection includes strains V. paradoxus CL14, Arthrobacter CL28, Acinetobacter CL69 and Acinetobacter CL71, which are also used in this work in individual strain contexts. V. paradoxus CL14 ΔHS33, which has a clean deletion of genes with gene ID 2643613677 through 2643613653 was constructed previously14 and used here. Additional strains were obtained from the American Type Culture Collection (ATCC): E. soli LF7 (ATCC BAA-2102), R. pomeroyi (ATCC 700808) and B. japonicum (ATCC 10324). P. phytofirmans PsJN (DSMZ 17436) was obtained from the DSMZ-German Collection of Microorganisms and Cell Cultures. P. putida strain 1290 was generously provided by Johan Leveau (University of California Davis). Pseudomonas strain Root 562 was generously provided by Paul Schulze-Lefert (Max-Planck-Gesellschaft). All bacteria, with exceptions noted below, were routinely grown on LB agar plates (10 g l−1 tryptone, 5 g l−1 yeast extract, 10 g l−1 NaCl, 1.5% (w/v) agar) and in 2xYT liquid medium (16 g l−1 tryptone, 10 g l−1 yeast extract, 5 g l−1 NaCl) at 28 °C. The 175-member (185-member minus 10 Variovorax strains) synthetic community (SC185-10V) was grown on KB medium as was done previously to culture this synthetic community14. B. japonicum (ATCC 10324) was routinely grown on liquid and solidified YM medium (1 g l−1 yeast extract, 10 g l−1 mannitol, 0.5 g l−1 dipotassium phosphate, 0.2 g l−1 magnesium sulfate, 0.1 g l−1 NaCl, 1 g l−1 CaCO4, pH 6.8, solidified with 1.5% agar as necessary) at 28 °C. R. pomeroyi (ATCC 700808) was routinely grown on liquid and solidified LB medium supplemented with 2% sea salt (Millipore Sigma S9883) and solidified with 1.5% (w/v) agar as necessary. M9 base medium was formulated using 1x M9 minimal salts medium (Sigma M6030) supplemented with 2 mM MgSO4, 0.1 mM CaCl2 and 10 µM FeSO4. A carbon source or sources were added to this M9 base medium to support bacterial growth. Unique strains constructed in this study are available upon request.Bacterial 16S rRNA sequencingBacterial colonization of Arabidopsis roots was assessed using a method similar to the previous study14. Roots from 8–10 plants were collected into sterilized 2 ml tubes containing three 4 mm glass beads and root fresh weight in each tube was obtained. Five such samples were collected for each bacterial treatment. The roots were washed three times with sterile distilled water and stored at −80 °C until further processing. The roots were then lyophilised for 48 h using a Labconco freeze-dry system and pulverized using an MPBio tissue homogenizer. DNA was extracted from the root samples and bacterial cell pellets saved from the bacteria for input into the experiment using the DNeasy PowerSoil HTP 96 kit (Qiagen) according to manufacturer instructions. The V3-V4 region of the bacterial 16S rRNA gene was amplified and sequenced as previously described14.16S amplicon sequence data processingThe 16S sequencing data from synthetic community experiments were processed as previously described14. Briefly, usable read output from MT-Toolbox36 (reads with 100% primer sequences that successfully merged with their pair) were filtered for quality with Sickle37 by not allowing any window with Q score under 20. The resulting sequences were globally aligned to a 16S rRNA gene sequence reference dataset from genome assemblies of the synthetic community members. For strains that do not have an intact 16S rRNA sequence in their assembly, Sanger sequencing was used to obtain the 16S rRNA gene sequence of the strains for inclusion in the reference dataset. The reference dataset also included sequences from Arabidopsis organellar sequences and known bacterial contaminants. Sequence alignment was performed with USEARCH v.7.109038 using the optional usearch_global at a 98% identity threshold. On average, 85% of read sequences matched an expected isolate. The 185 isolates of our 185-member synthetic community could not all be distinguished from one another on the basis of the V3-V4 sequence. They were thus classified into 97 unique sequences encompassing a set of identical (clustered at 100%) V3-V4 sequences coming from a single or multiple isolate strains. An isolate abundance table was created from the sequence mapping results.We estimated 16S rRNA absolute abundance using a plasmid spike-in method39. Synthetic DNA was spiked at known quantities into samples before DNA extraction and the ratio of added to recovered synthetic DNA served as a conversion factor by which the total number of 16S rRNA molecules in a given sample was estimated. We designed a plasmid which included 16S V3-V4 primer binding sequences flanking a randomly generated DNA sequence matching the most frequent length and Guanine + Cytosine (GC) content of amplicons generated using the same primer sequences from wild soil. These sequences were synthesized by Geneart (Invitrogen) and supplied cloned in plasmid pMA-T. The plasmid was transformed into E. coli and isolated using a miniprep spin kit (Qiagen). Specific volumes of this isolated plasmid were then added to individual samples before DNA extraction to spike-in approximately 20% of the predicted 16S copies occurring within the sample. We performed colony-forming units (c.f.u.) counting using similarly treated plant samples (that is, growth on SynCom-inoculated agar plates) to obtain an estimate of the 16S copy number per mg fresh weight of plant roots. We plated serial dilutions of plant root samples ground in MgCl2 on LB to perform c.f.u. counts. The c.f.u. count multiplied by a given sample’s fresh weight were used to calculate sample-specific predicted 16S copy numbers.Plant growth conditions and root growth inhibition assayA. thaliana ecotype Col-0 seeds were sterilized in 70% household bleach, 0.2% Tween-20 for 10 min with vigorous agitation and then rinsed 10 times with sterile distilled water. Seeds were stratified at 4 °C in sterile distilled water for 1–2 d. Plants were germinated for 7 d on 0.5x MS agar medium (2.22 g l−1 PhytoTech Labs M-404: Murashige & Skoog modified basal medium with Gamborg vitamins, 0.5 g l−1 MES hydrate, pH adjusted to 5.7, solidified with 1% (w/v) agar) supplemented with 0.5% (w/v) sucrose in vertical 12 ×12 cm square plates under long-day conditions (21 °C/18 °C, 16 h light/8 h dark, day/night cycle). Then 8 to 10 plants were aseptically transferred to 12 ×12 cm plates containing 0.5x MS agar medium without sucrose where the medium surface was spread with the bacterial inoculum. For assays with IAA addition, 100 nM IAA was added to the medium before pouring the plates. The plant root tip location was marked on plates after transfer to record the initial root tip position. The plates containing the plants and bacteria were incubated vertically under short-day conditions (22 °C/18 °C, 9 h light/15 h dark, day/night cycle) for an additional 11 d. Plates were imaged on a document scanner and primary root elongation was determined using imageJ to quantify the change in root tip position from the initial to the final position.Bacterial inoculation of plantsIndividual bacterial strains were grown on agar plates of the media types specified above at 28 °C. Before plant inoculation, a single colony was picked into the appropriate liquid medium and grown at 28 °C to late exponential or early stationary phase. To remove the medium from the bacteria before inoculation, strains were washed three times in sterile 10 mM MgCl2. The optical density at 600 nm (OD600) was measured for each washed strain and normalized to OD600 of 0.01 in 10 mM MgCl2. For plant experiments with mono-association of an individual strain, 100 µl of OD600 = 0.01 washed bacteria was spread on the 12 ×12 cm plate before plant transfer. For experiments in duo-association with Arthrobacter CL28, 100 µl of OD600 = 0.01 washed Arthrobacter CL28 was spread along with 100 µl of OD600 = 0.01 of the second strain.The 175-member synthetic community (SC185-10V) was prepared as described for the 185-member synthetic community used previously14 by leaving out the 10 isolates from the genus Variovorax. Briefly, 7 d before plant transfer, strains were inoculated individually into 600 µl KB medium in a 96-well plate and grown at 28 °C for 5 d. At 2 d before plant transfer, 20 ul from these 5-day-old cultures were transferred to 380 ul fresh KB medium in a new set of 96-well plates and both sets of plates were returned to the incubator for 2 d. This resulted in two cultures of each strain, one 7 d old and the other 2 d old, which were combined. The OD600 of the strains in each well was measured and the strains were combined while normalizing the OD600 of each strain in the pool. This pool was washed twice with 10 mM MgCl2 and diluted to OD600 = 0.2. For experiments with the SC185-10V SynCom, 100 µl of this OD600 = 0.2 washed pool was spread on 12 ×12 cm plates. For treatments where an additional strain was added to the SC185-10V SynCom, the individual strain was washed as described above, diluted to OD600 = 0.0034 in 10 mM MgCl2, and 100 µl of this dilution was spread on the plates with the SC185-10V SynCom. This addition of the individual strain corresponded to an OD600 three times that of a single strain in the SC185-10V SynCom (0.0034 = (0.2/175) × 3). For the addition of the 10 Variovorax strains to the SC185-10V SynCom experiment, the 10 Variovorax strains were grown individually in 2xYT medium from colonies grown on plates. The OD600 values of the 10 cultures were measured and the 10 strains were pooled while normalizing the OD600 of each strain to the same value. This mixture of the 10 Variovorax strains was then treated as the individual strains for washing and addition of 100 µl of OD600 = 0.0034 to the SC185-10V SynCom on plates.Construction of vectors with Variovorax CL14 iad gene insertsPortions of the V. paradoxus CL14 IAA degradation locus were subcloned into broad host range vector pBBR1MCS-232. Primers JMC579 through JMC604 (Supplementary Table 8) were used to amplify 3–5 kb segments of the locus by PCR using Q5 DNA polymerase (New England Biolabs). These primers were designed to amplify sections beginning and ending at gene start codons and with appropriate overlapping sequences for Gibson assembly either into the pBBR1MCS-2 backbone or to the adjacent section to make larger vector inserts, as appropriate. The pBBR1MCS-2 vector backbone was prepared for Gibson assembly by amplifying the vector by PCR using primers JMC577 and JMC578 (Supplementary Table 8) and subsequently treating with DpnI to remove circular vector template. PCR fragments were cleaned up as necessary using the QIAquick PCR purification kit (Qiagen). Appropriate fragments were mixed to construct the vectors by Gibson assembly using HiFi DNA Assembly Mastermix (New England Biolabs) according to manufacturer instructions. Gibson assembly products were transformed into NEB 10beta chemically competent E. coli (New England Biolabs) and selected on LB plates supplemented with 50 µg ml−1 kanamycin. Vectors were miniprepped using either the ZR plasmid miniprep classic kit or Zymo BAC DNA miniprep kit (Zymo Research) and confirmed via restriction mapping with PstI-HF (New England Biolabs) and Sanger sequencing (Genewiz).To construct vectors that are derivatives of pBBR1::70–66, the Q5 site-directed mutagenesis kit (New England Biolabs) was used for gene deletion. Briefly, vector pBBR1::70–66 was used as a PCR template and portions of this vector were amplified by PCR using primers JMC641 through JMC650 (Supplementary Table 8) and Q5 DNA polymerase (New England Biolabs). PCR products were cleaned up and circularized using KLD Mastermix (New England Biolabs). The product was transformed into NEB 10beta chemically competent E. coli (New England Biolabs) and selected on LB plates supplemented with 50 µg ml−1 kanamycin. Vectors were miniprepped and Sanger sequenced as described above to confirm the construction of the correct vectors.Conjugation of vectors to V. paradoxus CL14 ΔHS33Vectors were conjugated into V. paradoxus CL14 ΔHS33 using tri-parental mating. The helper E. coli strain carrying plasmid pRK201340 and donor NEB 10beta E. coli strains containing the pBBR1MCS-2-based vectors with Variovorax IAA degradation locus gene inserts were cultured in LB media containing 50 µg ml−1 kanamycin at 37 °C. V. paradoxus CL14 ΔHS33 was grown in 2xYT medium containing 100 µg ml−1 ampicillin at 28 °C. V. paradoxus CL14 wild type and derivative strains such as ΔHS33 are naturally resistant to ampicillin and this ampicillin selection allows for recovery of only Variovorax from the conjugation reaction. To prepare for conjugation, all bacteria were pelleted by centrifugation at 5,000 × g for 5 min and washed 3 times in 2xYT medium without antibiotics. For each conjugation reaction, equal volumes (100–300 µl) of each of the three washed bacteria: recipient V. paradoxus CL14 ΔHS33, donor NEB 10beta E. coli containing a pBBR1MCS-2-based vector, and helper E. coli pRK2013 were mixed. Control conjugation mixtures of each pair of strains and individual strains alone were performed in parallel to ensure successful selection of exconjugants only from mixtures of all three strains together. Conjugation mixtures were pelleted by centrifugation at 5,000 × g for 5 min, resuspended in 50 µl 2xYT media, transferred to LB media plates without antibiotics and allowed to dry in a laminar flow hood. These conjugation plates were incubated overnight at 28 °C. After 18–24 h, exconjugants were selected by streaking from the pooled conjugation mixtures on the LB plate without antibiotics to LB plates containing 50 µg ml−1 kanamycin and 100 µg ml−1 ampicillin. This selects for only V. paradoxus CL14 ΔHS33 (ampicillin resistant) containing the pBBR1MCS-2-based vector (kanamycin resistant). Individual colonies were picked into and subsequently cultured in 2xYT medium containing 50 µg ml−1 kanamycin and 100 µg ml−1 ampicillin at 28 °C.Construction of V. paradoxus CL14 gene deletionsUnmarked gene deletions in V. paradoxus CL14 were constructed as described previously14 using the suicide vector backbone pMo130 originally developed for gene knockouts in Burkholderia spp.41. Primers JMC203 and JMC204 (Supplementary Table 8) were used to amplify the pMO130 vector backbone by PCR. This product was subsequently treated with DpnI (New England Biolabs) to digest circular template DNA. Primers JMC605 through JMC612 and JMC671 through JMC677 (Supplementary Table 8) were used to amplify flanking regions for the gene deletion targets from V. paradoxus CL14 genomic DNA. All PCR was performed using Q5 DNA polymerase (New England Biolabs) and products were cleaned up, as appropriate, with the QIAquick PCR purification kit (Qiagen). These PCR products were assembled into suicide vectors using HiFi Gibson Assembly Mastermix (New England Biolabs), transformed into chemically competent NEB 5alpha E. coli (New England Biolabs), and selected on LB plates with 50 µg ml−1 kanamycin. Vectors were miniprepped using the ZR plasmid miniprep classic kit (Zymo Research) and confirmed by Sanger sequencing (Genewiz). Confirmed vectors were transformed into the chemically competent bi-parental mating strain E. coli WM3064. Transformants were selected at 37 °C on LB media supplemented with 50 µg ml−1 kanamycin and 0.3 mM diaminopimelic acid (DAP), and single colonies picked into LB medium also with 50 µg ml−1 kanamycin and 0.3 mM DAP.Bi-parental mating was performed by growing E. coli WM3064 containing the appropriate suicide vector as described above, and V. paradoxus CL14 was grown in 2xYT medium containing 100 µg ml−1 ampicillin at 28 °C. Both E. coli and Variovorax were washed separately three times using 2xYT medium, then mixed in a 1:1 ratio and pelleted. All centrifugation steps were performed at 5,000 × g for 5 min. The pelleted conjugation mixtures were resuspended in 1/10 the volume of 2xYT, plated on LB agar with 0.3 mM DAP and grown at 28 °C overnight. Exconjugants from these plates were streaked out and grown on LB agar with 100 µg ml−1 ampicillin, 50 µg ml−1 kanamycin, and no DAP at 28 °C. These strains were purified by streaking and growing on plates of the same medium once more. These strains with suicide vector integration were then grown once in liquid LB containing 100 µg ml−1 ampicillin and 1 mM isopropyl 1-thio-d-galactopyranoside (IPTG) at 28 °C and then streaked on plates containing media with 10 g l−1 tryptone, 5 g l−1 yeast extract, 100 g l−1 sucrose, 1.5% agar, 100 µg ml−1 ampicillin and 1 mM IPTG. Colonies from these plates were picked and grown in the same liquid media. These strains were then assessed for gene deletion by PCR using primers JMC657 through JMC660 and JMC697 through JMC699 (Supplementary Table 8). The Quick-DNA miniprep kit (Zymo Research) was used to isolate all genomic DNA for PCR screening. To purify the knockout strains, they were streaked and grown out three times on LB plates containing 100 µg ml−1 ampicillin before a final PCR confirmation. To check the purity of the final strains, PCR was performed with one primer outside the deletion region and one inside the deleted gene to ensure no product is produced for the knockout strain. The sequences for the primers used for this PCR reaction (JMC691, JMC717, JMC718, JMC693 and JMC694) can be found in Supplementary Table 8.Measurement of bacterial growth and IAA degradationIndividual strains were grown in 5 ml cultures in various media types supplemented with IAA at 28 °C and 250 r.p.m. To screen the V. paradoxus CL14 ΔHS33 pBBR1 vector complemented mutants, 2xYT medium supplemented with 0.1 mg ml−1 IAA was used. For comparison of other V. paradoxus CL14 mutants, M9 medium with 15 mM succinate and 0.1 mg ml−1 (0.57 mM) IAA was used. For comparison of IAA-degrading strains from diverse genera, M9 medium with 0.1% (w/v) casamino acids (Bacto) and 0.1 mg ml−1 IAA was used. For R. pomeroyi, 2% (w/v) sea salts were added to this M9 medium with casamino acids and IAA. The pBBR1 vector library in E. coli NEB 10beta was screened in LB medium supplemented with 0.04 mg ml−1 IAA and grown at 37 °C and 250 r.p.m. For all media types, IAA was solubilized in 100% ethanol at 20 mg ml−1 and diluted to 0.1 mg ml−1 in the media, resulting in 0.5% (v/v) ethanol in the media.To measure growth, a 200 µl sample was taken from the growing cultures and OD600 was determined on an Infinite M200 Pro plate reader (Tecan). Subsequently, cells were pelleted by centrifugation at 4,200 × g for 15 min and 50 µl of supernatant was transferred to a new 96-well plate and frozen at −80 °C until further analysis. IAA degradation was determined by thawing the plates containing 50 µl aliquots of culture supernatant and combining this with 100 µl of Salkowski reagent (10 mM ferric chloride and 35% perchloric acid)42. This was performed alongside mixing 50 µl of IAA standards with 100 µl of Salkowski reagent in the same 96-well plate format. Colour development was allowed to proceed for 1 h and absorbance was read at 530 nm on the Infinite M200 Pro plate reader (Tecan). The absorbances measured were converted to IAA concentration on the basis of the absorbances measured for the IAA standards.Liquid Chromatography Dual Mass Spectroscopy (LC–MS/MS) metabolomics on Variovorax IAA degradationV. paradoxus CL14 was grown in 5 ml cultures of M9 minimal medium supplemented with either 0.1 mg ml−1 IAA, 0.1 mg ml−1 13C6-IAA (with the 6 carbons of the benzene ring of the indole labelled, Cambridge Isotope Laboratories CLM-1896-PK), and/or 15 mM succinate. Cultures and parallel media controls were incubated at 28 °C with shaking at 250 r.p.m. Cultures and media controls were centrifuged (4,200 × g for 15 min at 4 °C) to pellet cells; supernatants were transferred to new tubes and both pellets (intracellular fraction) and supernatants (extracellular fraction) were stored frozen at −80 °C until extraction. All subsequent work was performed over dry ice or in chilled cold blocks. Frozen pellets from the intracellular fraction were thawed for 3 h at 4 °C, then 800 µl of cold LCMS-grade water was added to the pellets with repeated pipetting to break up the pellet until visually homogeneous. Samples were then quickly returned to −80 °C to freeze the suspension. Frozen pellet suspensions and extracellular solutions were lyophilised until dry. The cells from the dried pellet suspensions were lysed and homogenized with a bead mill (BioSpec Mini-Beadbeater-96) using one sterile 3.2 mm steel ball in each tube for 3 rounds of 5 s each with 10 s breaks in between to reduce heat production. Dried extracellular samples were concentrated by resuspension in 100 µl LCMS- grade methanol, vortexed 3 times for 10 s each, water bath sonicated for 20 min, incubated at 4 °C overnight, centrifuged (1,000 × g, 4 °C, 5 min), and the methanol supernatant was dried using a speed vacuum concentrator. On the day of LC–MS/MS analysis, homogenized dry material was suspended in LCMS-grade methanol with internal standard mix (100 µM U-13C/15N-labelled amino acids, SIGMA 767964). Intracellular samples were suspended at 11.1 µl mg−1 of original sample cell pellet wet weight; extracellular samples were suspended at 38.9 µl mg−1 of corresponding cell pellet wet weight from the culture. The solutions were vortexed 3 times for 10 s each, bath sonicated in ice water for 10 min, chilled at −20 °C for 10 min, then centrifuged (10,000 × g, 5 min, 10 °C) to pellet insoluble material. Supernatants containing the methanol extracts were filtered through 0.22 µm PVDF microcentrifuge filtration tubes (10,000 × g, 5 min, 10 °C); filtrates were transferred to glass vials and immediately capped. Filtrates were then analysed by LC–MS/MS using an Agilent 1290 UHPLC system connected to a Thermo Q Exactive Hybrid Quadrupole-Orbitrap mass spectrometer equipped with a heated electrospray ionization (HESI-II) source probe. Extracts were chromatographically separated on a ZORBAX RRHD Eclipse Plus C18, 95 Å, 2.1 × 50 mm, 1.8 µm column (Agilent) for non-polar metabolomics. Separation, ionization, fragmentation and data acquisition parameters are specified in Supplementary Table 7. Briefly, metabolites were separated by gradient elution followed by MS1 and data-dependent (top 2 most abundant MS1 ions not previously fragmented in last 7 s) MS2 collection; targeted data analysis was performed by comparison of sample peaks to a library of analytical standards analysed under the same conditions or by searching the raw data files for predicted m/z values based on structural information of compounds of interest. Three parameters were compared: matching m/z, retention time and fragmentation spectra using Metabolite Atlas (https://github.com/biorack/metatlas)43,44. Identification and standard reference comparison details are provided in Supplementary Table 6. Raw and processed data are available for download at the JGI Joint Genome Portal under ID 1340427. Statistical comparisons were performed using R version 3.6.2, using package agricolae 1.3–5 and stats 3.6.245; boxplots were generated with base R graphics using the boxplot function.Phylogenomic analysisTo guide the delineation of the IAA degradation operons across the bacterial tree of life, we constructed two Hidden Markov Model (HMM) profiles of the genes iacC and iacD by subsetting all homologous genes from the previously validated operons (Extended Data Fig. 4). In parallel, we downloaded the assembly files for all available complete genomes deposited in the NCBI RefSeq 202 repository46. For the 220,000 assembly files downloaded, we performed open reading frame (ORF) prediction using prodigal. We then used the two HMM profiles described above to query the predicted ORFs. Utilizing ad hoc scripts, we constructed a table of HMM hits along the genomes scanned and subset genomic loci where both iacC and iacD genes appeared adjacent to one another. The logic of using the iacC and iacD genes as anchor genes for our search is that the adjacent physical location of both iacC and iacD homologues is a conserved feature across all previously experimentally validated IAA-degrading operons (Extended Data Fig. 4). Next, for each region containing the adjacent iacC and iacD homologue genes, we extracted the gene neighbourhood adjacent to the anchor hit by extracting the amino acid sequence of ORFs +10 kb and −10kb with respect to the anchor hit. Using hmmscan from the Hmmer v3.1.b2 suite47, we performed HMM profiling in all ORFs extracted via our neighbourhood delineation against the COG database version 2003. Finally, we used the COG profiles across the neighbourhoods to create a matrix describing the prevalence of COGs across the regions (candidate regions) with the adjacent iacC and iacD homologue genes.For each genome containing at least one candidate region, we performed taxonomic classification using the GTDB database48. Due to the size of our estimated genomic matrix and to reduce potential biases due to over-representation of certain lineages within RefSeq, we performed principal coordinate analysis (PCoA) using a reduced matrix containing one representative candidate region per species. Species labelling was obtained from the GTDB taxonomic classification described above. PCoA was performed using the oh.pco function from the ohchibi package49, taking as input a binary version of the gene matrix described above. We classified candidate reads into the two types of IAA-degrading operon (iac-like and iad-like), utilizing a majority count-based approach using marker COGs conserved between the previously experimentally validated IAA-degrading operons (Extended Data Fig. 4). Specifically, for each potential operon, we determined the prevalence of COGs that a priori (Extended Data Fig. 4) showed differential prevalence across the two degrading operons (for example, iacA, iacB and iacI are markers of the iac operon, while iorB/iadB and iotA/iadA are exclusive markers of the iad-like operon). Hybrid gene clusters were defined as operons that exhibited the hallmark COGs of both operons.In parallel, we performed phylogenetic inference over all the genomes belonging to genera with at least one representative strain harbouring any of the two types of IAA-degrading operon. This phylogenetic tree was constructed using a super-matrix-based approach as previously described35. Finally, for each genus with at least one assembly harbouring a positive IAA-degrading operon, we estimated the prevalence of the trait across the genus by dividing the total number of isolates with detectable IAA degradation locus by the total number of isolates belonging to that genus in the dataset. In addition, to see the phylogenetic evenness of the distribution of the IAA degradation trait across each genus, we calculated the phylogenetic ratio by calculating the ratio between the average phylogenetic distance (computed via the cophenetic.phylo function from the ape R package50) of isolates with a detectable IAA degradation locus and the total average phylogenetic distance of all isolates within that genus. We constructed the MarR phylogeny using the MarR sequences from candidate regions with 100% markers of one of the two types of IAA-degrading operon. Amino acid sequences of the MarR homologues were aligned using MAFFT51 and phylogenetic inference was performed using FastTree 252.RNA-seq on Variovorax strainsV. paradoxus CL14 was grown in 5 ml cultures of M9 minimal medium supplemented with 15 mM succinate and 0.5% (v/v) ethanol alone or containing IAA. IAA was at a final concentration of 0.1 mg ml−1 in the medium to which it was added. Cultures were prepared at a starting OD600 of 0.02 and incubated at 28 °C, shaking at 250 r.p.m. Cells from all samples were collected for RNA-seq at 18 h to ensure IAA was still present in the cultures of strains that degraded IAA most rapidly. Cells were pelleted by centrifuging the culture at 4,200 × g for 15 min and removing the supernatant. Cell pellets were frozen at −80 °C before RNA extraction. To extract RNA, cells were lysed in TRIzol reagent (Invitrogen) according to manufacturer instructions for lysis and phase separation. After these steps, RNA was purified from the aqueous phase using the RNeasy mini kit (Qiagen) including the optional on-column DNase digestion with RNase-free DNase set (Qiagen). Total RNA was quantified using the Qubit 2.0 fluorometer (Invitrogen) and RNA-seq libraries were prepared using the Universal Prokaryotic RNA-Seq Prokaryotic AnyDeplete kit (Tecan) according to manufacturer instructions. The resulting libraries were pooled and sequenced on the Illumina HiSeq4000 to generate 50 bp single-end reads.RNA-seq data analysisThe V. paradoxus CL14 RNA-seq sequence data were analysed as described previously14. Briefly, the raw reads were mapped to the V. paradoxus CL14 genome (fasta file available at https://github.com/isaisg/variovoraxRGI/blob/master/rawdata/2643221508.fna) using bowtie253 with the ‘very sensitive’ flag. Hits to each individual coding sequence were counted and annotated using the function featureCounts from the R package Rsubread54, inputting the V. paradoxus CL14 gff file (available at https://github.com/isaisg/variovoraxRGI/blob/master/rawdata/2643221508.gff) and using the default parameters with the flag allowMultiOverlap = FALSE. Finally, DESeq255 was used to estimate Differentially Expressed Genes (DEGs) between treatments, with the corresponding fold-change estimates and False Discovery Rate (FDR) adjusted P values. For visualization purposes, we performed z-score standardization of each gene across samples and we visualized this standardized expression values utilizing a heat map constructed using ggplot256. These data can be found in Supplementary Table 9.MarR protein expression and purificationThe coding sequence for each gene can be found in Supplementary Table 10. MarR expression plasmids were synthesized as codon-optimized genes for E. coli expression by BioBasic in the pLIC-His N-term vector (pMCSG7) and transformed into E. coli BL21 (DE3) Gold cells for expression. Cells were grown in the presence of ampicillin in LB medium with shaking at 225 r.p.m. at 37 °C to an OD600 of 0.5, at which point the temperature was reduced to 18 °C. At an OD600 of 0.8, protein expression was induced by the addition of 0.1 mM IPTG and incubation continued overnight. Cells were collected by centrifugation at 4,500 × g for 20 min at 4 °C in a Sorvall (model RC-3B) swinging bucket centrifuge. Cell pellets were resuspended in buffer A (20 mM potassium phosphate, pH 7.4, 50 mM imidazole, 500 mM NaCl), DNase, lysozyme and a Roche Complete EDTA-free protease inhibitor tablet. Resuspended cells were sonicated and clarified via centrifugation at 17,000 × g for 60 min in a Sorvall (model RC-5B) swinging bucket centrifuge. The lysate was applied to a nickel-nitrilotriacetic acid HP column (GE Healthcare) on an Aktaxpress Fast Performance Liquid Chromatography (FPLC) system (Amersham Bioscience) and washed with buffer A. Protein was eluted with buffer B (20 mM potassium phosphate, pH 7.4, 500 mM imidazole, 500 mM NaCl). Fractions containing the protein of interest were combined and passed over a HiLoadTM 16/60 SuperdexTM 200 gel filtration column. Proteins were eluted in S200 buffer (20 mM HEPES, pH 7.4, 300 mM NaCl). Fractions were combined and concentrated for long-term storage at −80 °C.MarR mutant proteins were created by site-directed mutagenesis using primers from Integrated DNA Technologies. The mutant plasmids were sequenced to confirm the mutations. The mutants were produced and purified using E. coli BL21 (DE3) Gold as described above.Ligand binding studies by isothermal titration calorimetry (ITC)All ITC measurements were performed at 25 °C using an Auto-ITC200 microcalorimeter (MicroCal/GE Healthcare). The buffer employed was 20 mM HEPES, pH 7.4, 50 mM NaCl and 0.5% dimethly sulfoxide (DMSO) for protein/ligand binding and 20 mM HEPES, pH 7.4 and 300 mM NaCl for DNA/protein binding experiments. For ligand binding experiments, the calorimetry cell (volume 200 ml) was loaded with MarR wild-type, mutant or homologue protein at a concentration of 50 μM. The syringe was loaded with a ligand concentration of 0.5 or 2 mM. For DNA binding experiments, wild-type MarR_73 did not bind any of the DNA oligos examined; however, we hypothesized that this arose from the ability of this native receptor to remain bound to ligands retained from its recombinant expression in E. coli. Thus, we employed the MarR_73 S28A protein with reduced ligand binding capacity. Here, the calorimetry cell was loaded with duplex oligo at a concentration of 25 μM and the syringe was loaded with MarR S28A mutant protein, which was necessary to prevent ligand binding during expression and purification, at a concentration of 0.5 mM. A typical injection protocol included a single 0.2 μl first injection followed by 20 1.5 μl injections of the syringe sample into the calorimetry cell. The spacing between injections was kept at 180 s and the reference power at 8 μcal s−1. The data were analysed using Origin for ITC version 7.0 software supplied by the manufacturer and fit well to a one-site binding model. Two independent ITC measurements were performed for each condition. A non-integer N value (for example, 0.73 in Fig. 2a) indicates that some protein monomers may not be in an active conformation, and thus do not bind ligand. Additionally, small measurement errors in assessing the protein or ligand concentrations may also contribute to non-integer N values in ITC. To confirm that 300 mM NaCl did not negatively impact DNA binding, MarR_73 S28A was examined by ITC in 150 mM NaCl. In this condition, the KD for the 22 bp duplex was 0.428 ± 0.002 μM (N = 1.75 ± 0.014), while the KD for 24 bp duplex was 0.151 ± 0.025 μM (N = 2.51 ± 0.26).Protein crystallographyV. paradoxus MarR_73 was crystallized using the sitting drop vapour diffusion method at 20 °C in conditions outlined in Supplementary Table 4. Crystallization drops were set up using the Oryx4 protein crystallization robot (Douglas Instruments) and contained 0.15 μl protein and 0.15 μl well solution. For all V. paradoxus MarR_73 wild-type conditions, ligands were added at 10-fold molar excess before crystallization trials and crystals appeared within 2–5 d. V. paradoxus MarR_73 with the S28A and R46A mutations was crystallized in similar conditions as the wild-type protein. Similarly, P. putida MarR_iacR, B. japonicum MarR_Bj1, A. baumannii MarR_Ab and E. soli MarR_Es were crystallized using vapour diffusion methods in sitting drop trays at 20 °C and crystals appeared within 3–5 d. All crystallization conditions are outlined in Supplementary Table 4. Crystal specimens were cryoprotected with the well solution supplemented with glycerol to 20% (v/v) (Supplementary Table 4). X-ray diffraction data were collected at the Advanced Photon Source beamline 23-ID-D (Supplementary Table 3). Diffraction images were reduced using either XDS or Denzo and scaled with either Aimless or Scalepack57,58,59. The V. paradoxus MarR_73 structure in complex with IAA was determined by molecular replacement using the structure of 3CDH as a search model in Phaser60. All subsequent structures of V. paradoxus MarR_73 were determined using the V. paradoxus MarR_73 IAA complex structure (PDB: 7KFO) as a search model. The P. putida MarR_iacR and B. japonicum MarR_Bj1 structures were determined by molecular replacement using the structure of 3CJN as the search model. P. putida MarR_iacR (PDB: 7KUA) was subsequently used as the search model for molecular replacement to solve A. baumannii MarR_Ab and E. soli MarR_Es. A nickel ion was placed in the model of MarR_Ab. The following ions or molecules were examined and refined in this location in the MarR_Ab structure: water, Na, Mg, K, Ca, Mn, Fe, Co, Ni, Cu, Zn and Ba. Water, Na, Mg, K, Ca and Ba were deemed unacceptable in this site due to poor difference density. Of the remaining ions considered, there were no sources of Mn, Fe, Co, Cu or Zn in the protein expression media, protein purification buffers, protein storage buffer, crystallization condition or cryoprotectant solutions. Thus, we concluded that the ion present in this structure is Ni due to the use of a nickel-affinity column during the protein’s purification. It is unclear why this ion remained bound to MarR_Ab even after the subsequent size exclusion chromatography purification step, or why such an ion is only observed in this structure of the proteins examined. All structures were refined with either Phenix.refine or Refmac using iterative model building in Coot to the final parameters outlined in Supplementary Table 361,62. MarR_73 is a dimer with one protein monomer in the asymmetric unit and the dimer generated by crystallographic symmetry. PDB accession codes and associated crystallographic data are reported in Supplementary Table 3.Statistics and reproducibilityNo statistical method was used to predetermine sample size, but our sample sizes are similar to those reported in previous publications14,63,64. No data were excluded from the analyses. The experiments were randomized. The investigators were not blinded to allocation during experiments and outcome assessment. Where not stated, data distribution was assumed to be normal, but this was not formally tested.Reporting summaryFurther information on research design is available in the Nature Research Reporting Summary linked to this article. More

Sarwar, N. et al. Phytoremediation strategies for soils contaminated with heavy metals: Modifications and future perspectives. Chemosphere 171, 710–721 (2017).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Lin, H. M. et al. Cadmium-stress mitigation through gene expression of rice and silicon addition. Plant Growth Regul.: Int. J. Nat. Synthetic Regul. 81(1), 91–101 (2017).Article 
CAS 

Google Scholar 
Pan, F. S. et al. Enhanced Cd extraction of oilseed rape (Brassica napus) by plant growth-promoting bacteria isolated from Cd hyperaccumulator Sedum alfredii Hance. Int. J. Phytorem. 19(1/6), 281–289 (2017).Article 
CAS 

Google Scholar 
Puangprasert, S. & Prueksasit, T. Health risk assessment of airborne Cd, Cu, Ni and Pb for electronic waste dismantling workers in Buriram Province, Thailand. J. Environ. Manag. 252, 109601 (2019).Article 
CAS 

Google Scholar 
Tipu, M. I. et al. Growth and physiology of maize (Zea mays L.) in a nickel-contaminated soil and phytoremediation efficiency using EDTA. J. Plant Growth Regul. 40(2), 774–786 (2021).Article 
CAS 

Google Scholar 
Chaturvedi, N., Dhal, N. K. & Patra, H. K. EDTA and citric acid-mediated phytoextraction of heavy metals from iron ore tailings using Andrographis paniculata: A comparative study. Int. J. Min. Reclam. Environ. 29(1), 33–46 (2015).Article 
CAS 

Google Scholar 
Wang, G. Y. et al. Heavy metal removal by GLDA washing: Optimization, redistribution, recycling, and changes in soil fertility. Sci. Total Environ. 569–570, 557–568 (2016).Article 
ADS 
PubMed 

Google Scholar 
Kołodyńska, D. Cu(II), Zn(II), Co(II) and Pb(II) removal in the presence of the complexing agent of a new generation. Desalination 267(2–3), 175–183 (2011).Article 

Google Scholar 
Guo, X. F. et al. Mixed chelators of EDTA, GLDA, and citric acid as washing agent effectively remove Cd, Zn, Pb, and Cu from soils. J. Soils Sediments 18(2), 835–844 (2017).
Google Scholar 
Wang, X. et al. Subcellular distribution and chemical forms of cadmiun in Bechmeria nivea L. Gaud. Environ. Exp. Bot. 62(3), 389–395 (2008).Article 
CAS 

Google Scholar 
Gallego, S. M. et al. Unravelling cadmium toxicity and tolerance in plants: Insight into regulatory mechanisms. Environ. Exp. Bot. 83, 33–46 (2012).Article 
CAS 

Google Scholar 
Clemens, S., Aarts, M. G. M., Thomine, S. & Verbruggen, N. Plant science: The key to preventing slow cadmium poisoning. Trends Plant Sci. 18(2), 92–99 (2013).Article 
CAS 
PubMed 

Google Scholar 
Zhou, J. T. et al. Integration of cadmium accumulation, subcellular distribution, and physiological responses to understand cadmium tolerance in apple rootstocks. Front. Plant Sci. 8, 966 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Yang, L. P., Zhu, J., Wang, P., Lyu, D. G. & Li, H. F. Effect of Cd on growth, physiological response, Cd subcellular distribution and chemical forms of Koelreuteria paniculata. Ecotoxicol. Environ. Saf. 160, 10–18 (2018).Article 
CAS 
PubMed 

Google Scholar 
Wang, W. J., Zhang, M. Z. & Liu, J. N. Subcellular distribution and chemical forms of Cd in Bougainvillea spectabilis Willd. as an ornamental phytostabilizer: An integrated consideration. Int. J. Phytorem. 20(11), 1087–1095 (2017).Article 

Google Scholar 
Weigel, H. J. & Jäger, H. J. Subcellular distribution and chemical form of cadmium in bean plants. Plant Physiol. 65(3), 480–482 (1980).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Khanna, K., Kohli, S. K., Ohri, P., Bhardwaj, R. & Ahmad, P. Agroecotoxicological aspect of Cd in soil–plant system: Uptake, translocation and amelioration strategies. Environ. Sci. Pollut. Res. 29, 30908–30934 (2022).Article 
CAS 

Google Scholar 
Wei, Z. B., Chen, X. H., Wu, Q. T. & Tan, M. Biodegradable chelator GLDA induced remediation of heavy metal contaminated soil in Southeast Jingtian. Environ. Sci. 36(5), 1864–1869 (2015).CAS 

Google Scholar 
Wang, K., Liu, Y. H., Song, Z. G., Wang, D. & Qiu, W. W. Chelator complexes enhanced Amaranthus hypochondriacus L. phytoremediation efficiency in Cd-contaminated soils. Chemosphere 237, 124480 (2019).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Meng, N., Wang, M., Chen, L., Zheng, H. & Chen, S. B. Remediation effects of different herbaceous plants intercropping on Cd-contaminated soil. China Environ. Sci. 38(7), 2618–2624 (2018).CAS 

Google Scholar 
Jones, D. & Willett, V. Experimental evaluation of methods to quantify dissolved organic nitrogen (don) and dissolved organic carbon (doc) in soil. Soil Biol. Biochem. 38(5), 991–999 (2006).Article 
CAS 

Google Scholar 
Su, F. L. et al. The distribution and enrichment characteristics of copper in soil and Phragmites australis of Liao River estuary wetland. Environ. Monit. Assess.: Int. J. 190(6), 1–9 (2018).Article 
CAS 

Google Scholar 
Shahid, M., Dumat, C. & Khalid, S. Reviews of Environmental Contamination and Toxicology Vol. 241, 3–137 (Springer, 2016).
Google Scholar 
Yuliya, V. et al. Comparison of soil-to-root transfer and translocation coefficients of trace elements in vines of Chardonnay and Muscat white grown in the same vineyard. Sci. Hortic. 192, 89–96 (2015).Article 

Google Scholar 
Liu, Q. Q., Chen, Y. H., Shen, Z. G. & Zheng, L. Q. Roles of cell wall in plant heavy metal tolerance. Plant Physiol. J. 50(5), 605–611 (2014).
Google Scholar 
Zhen, S. et al. Foliar application of Zn reduces Cd accumulation in grains of late rice by regulating the antioxidant system, enhancing Cd chelation onto cell wall of leaves, and inhibiting Cd translocation in rice. Sci. Total Environ. 770, 145302 (2021).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Shi, Y. X. et al. Simulation of the absorption, migration and accumulation process of heavy metal elements in soil-crop system. Environ. Sci. 37(10), 3996–4003 (2016).
Google Scholar 
Yan, X. X. et al. Effect of foliar application of different manganese fertilizers on cadmium accumulation and subcellular distribution in pak choi. J. Agro Environ. Sci. 38(8), 1872–1881 (2019).
Google Scholar 
He, S., Wu, Q. & He, Z. Effect of DA-6 and EDTA alone or in combination on uptake, subcellular distribution and chemical form of Pb in Lolium perenne. Chemosphere 93(11), 2782–2788 (2013).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Li, C. C. et al. Integration of metal chemical forms and subcellular partitioning to understand metal toxicity in two lettuce (Lactuca sativa L.) cultivars. Plant Soil 384(1/2), 201–212 (2014).Article 
CAS 

Google Scholar 
Li, D., He, T., Saleem, M. & He, G. Metalloprotein-specific or critical amino acid residues: Perspectives on plant-precise detoxification and recognition mechanisms under cadmium stress. Int. J. Mol. Sci. 23(3), 1734 (2022).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Perriguey, J., Sterckeman, T. & Morel, J. L. Effect of rhizosphere and plantrelated factors on the cadmium uptake by maize(Zea mays L.). Environ. Exp. Bot. 63(1/3), 333–341 (2008).Article 
CAS 

Google Scholar 
Dai, S. et al. Effects of biochar amendments on speciation and bioavailability of heavy metals in coal-mine-contaminated soil. Hum. Ecol. Risk Assess. Int. J. 24(7), 1887–1900 (2018).Article 
CAS 

Google Scholar 
Hou, S., Zheng, N., Tang, L., Ji, X. F. & Li, Y. Y. Effect of soil pH and organic matter content on heavy metals availability in maize (Zea mays L.) rhizospheric soil of non-ferrous metals smelting area. Environ. Monit. Assess. 191(10), 634 (2019).Article 
PubMed 

Google Scholar 
Wu, H. J. et al. Effects of Astragalus smicuson cadmium effectiveness in paddy soil and cadmium accumulation in rice plant. Chin. Agric. Sci. Bull. 33(16), 105–111 (2017).ADS 

Google Scholar 
Jin, P. K., Liu, K. J. & Wang, X. B. Conversion and utilization of slowly biodegradable organic matter. Chin. J. Environ. Eng. 10(5), 2168–2174 (2016).CAS 

Google Scholar 
Kopáček, J. et al. Factors affecting the leaching of dissolved organic carbon after tree dieback in an unmanaged European mountain forest. Environ. Sci. Technol. 52(11), 6291–6299 (2018).Article 
ADS 
PubMed 

Google Scholar 
Anwar, S. et al. Impact of chelator-induced phytoextraction of cadmium on yield and ionic uptake of maize. Int. J. Phytorem. 19(6), 505–513 (2017).Article 
CAS 

Google Scholar 
Wu, J. M., Xi, M. & Kong, F. L. Review of researches on the factors influencing the dynamics of dissolved organic carbon in soils. Geol. Rev. 59(5), 953–961 (2013).CAS 

Google Scholar 
AkzoNobel. Dissolvine GL® Technichal Brochure 1–5 (AkzoNobel Amsterdam, 2010).
Google Scholar 
Beygi, M. & Jalali, M. Assessment of trace elements (Cd, Cu, Ni, Zn) fractionation and bioavailability in vineyard soils from the Hamedan, Iran. Geoderma 337, 1009–1020 (2019).Article 
ADS 
CAS 

Google Scholar 
Gul, I. et al. Comparative effectiveness of organic and inorganic amendments on cadmium bioavailability and uptake by Pelargonium hortorum. J. Soils Sediments 19(5), 2346–2356 (2019).Article 
CAS 

Google Scholar 
Wang, H., Sun, L. N., Li, H. B. & Sun, T. Y. Effect of different chelators application on Cd accumulation in metal polluted soils by Beta vulgaris var. cicla L. Ecol. Environ. 17(6), 2249–2252 (2008).
Google Scholar 
Zhang, G. X. et al. Effects of biochars on the availability of heavy metals to ryegrass in an alkaline contaminated soil. Environ. Pollut. 218, 513–522 (2016).Article 
CAS 
PubMed 

Google Scholar 
Gu, M. H. et al. Effects of manganese application on the formation of manganese oxides and cadmium fixation in soil. Ecol. Environ. Sci. 229(2), 360–368 (2020).
Google Scholar 
Bradl, H. B. Adsorption of heavy metal ions on soils and soils constituents. J. Colloid Interface Sci. 277(1), 1–18 (2004).Article 
ADS 
CAS 
PubMed 

Google Scholar  More

Anderson, M. The evolution of eusociality. Annu. Rev. Ecol. Syst. 15, 165–189 (1984).Article 

Google Scholar 
Wilkinson, G. S. Reciprocal food sharing in the vampire bat. Nature 308, 181–184 (1984).Article 

Google Scholar 
Feistner, A. & Mcgrew, W. Food-sharing in primates: a critical review. Perspect. Primate Biol 3, (1989).Hoelzel, A. R. Killer whale predation on marine mammals at Punta Norte, Argentina; food sharing, provisioning and foraging strategy. Behav. Ecol. Sociobiol. 29, 197–204 (1991).Article 

Google Scholar 
Behmer, S. T. Animal behaviour: feeding the superorganism. Curr. Biol. 19, R366–R368 (2009).Article 
CAS 
PubMed 

Google Scholar 
Cassill, D. L. & Tschinkel, W. R. Information flow during social feeding in ant societies. in Information Processing in Social Insects (eds. Detrain, C., Deneubourg, J. L. & Pasteels, J. M.) 69–81 (Birkhäuser, 1999). https://doi.org/10.1007/978-3-0348-8739-7_4.Hunt, J. H. Trophallaxis and the evolution of eusocial Hymenoptera. in The Biology of Social Insects (CRC Press, 1982).Sorensen, A. A., Busch, T. M. & Vinson, S. B. Trophallaxis by temporal subcastes in the fire ant, Solenopsis invicta, in response to honey. Physiol. Entomol. 10, 105–111 (1985).Article 

Google Scholar 
Meurville, M.-P. & LeBoeuf, A. C. Trophallaxis: the functions and evolution of social fluid exchange in ant colonies (Hymenoptera: Formicidae). Myrmecol. News 31, 1–30 (2021).Bodner, L. et al. Nutrient utilization during male maturation and protein digestion in the Oriental hornet. Biology 11, 241 (2022).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Sorensen, A. A., Kamas, R. S. & Vinson, S. B. The influence of oral secretions from larvae on levels of proteinases in colony members of Solenopsis invicta Buren (Hymenoptera: Formicidae). J. Insect Physiol. 29, 163–168 (1983).Article 
CAS 

Google Scholar 
Erthal, M., Peres Silva, C. & Ian Samuels, R. Digestive enzymes in larvae of the leaf cutting ant, Acromyrmex subterraneus (Hymenoptera: Formicidae: Attini). J. Insect Physiol. 53, 1101–1111 (2007).Article 
CAS 
PubMed 

Google Scholar 
Went, F. W., Wheeler, J. & Wheeler, G. C. Feeding and digestion in some ants (Veromessor and Manica). BioScience 22, 82–88 (1972).Article 

Google Scholar 
Ishay, J. & Ikan, R. Food exchange between adults and larvae in Vespa orientalis F. Anim. Behav. 16, 298–303 (1968).Article 
CAS 
PubMed 

Google Scholar 
Hunt, J. H. The evolution of social wasps. (Oxford University Press, USA, 2007).Abe, T., Yoshiya, T., Hiromitsu, M. & Kawasaki, Y. Y. Comparative study of the composition of hornet larval saliva, its effect on behaviour and role of trophallaxis. Comp. Biochem. Physiol. Part C: Comp. Pharmacol. 99, 79–84 (1991).Article 

Google Scholar 
Ishay, J. & Ikan, R. Gluconeogenesis in the Oriental hornet Vespa orientalis F. Ecology 49, 169–171 (1968).Article 

Google Scholar 
Brock, R. E., Cini, A. & Sumner, S. Ecosystem services provided by aculeate wasps. Biol. Rev. 96, 1645–1675 (2021).Article 
PubMed 

Google Scholar 
Ueno, T. Flower-visiting by the invasive hornet Vespa velutina nigrithorax (Hymenoptera: Vespidae). Int. J. Chem., Environ. Biol. Sci. 3, 444–448 (2015).
Google Scholar 
Käfer, H., Kovac, H. & Stabentheiner, A. Respiration patterns of resting wasps (Vespula sp.). J. Insect Physiol. 59, 475–486 (2013).Article 
PubMed 
PubMed Central 

Google Scholar 
Bodner, L., Bouchebti, S. & Levin, E. Allocation and metabolism of naturally occurring dietary amino acids in the Oriental hornet. Insect Biochem. Mol. Biol. 139, 103675 (2021).Article 
CAS 
PubMed 

Google Scholar 
Levin, E., Lopez-Martinez, G., Fane, B. & Davidowitz, G. Hawkmoths use nectar sugar to reduce oxidative damage from flight. Science 355, 733–735 (2017).Article 
CAS 
PubMed 

Google Scholar 
Hunt, J. H., Baker, I. & Baker, H. G. Similarity of amino acids in nectar and larval saliva: the nutritional basis for trophallaxis in social wasps. Evolution 36, 1318–1322 (1982).Article 
CAS 
PubMed 

Google Scholar 
Hunt, J. H., Jeanne, R. L., Baker, I. & Grogan, D. E. Nutrient dynamics of a swarm-founding social wasp species, Polybia occidentalis (Hymenoptera: Vespidae). Ethology 75, 291–305 (1987).Article 

Google Scholar 
Cassill, D. L. & Tschinkel, W. R. Allocation of liquid food to larvae via trophallaxis in colonies of the fire ant, Solenopsis invicta. Anim. Behav. 50, 801–813 (1995).Article 

Google Scholar 
Buffin, A., Denis, D., Simaeys, G. V., Goldman, S. & Deneubourg, J.-L. Feeding and stocking up: radio-labelled food reveals exchange patterns in ants. PLOS ONE 4, e5919 (2009).Article 
PubMed 
PubMed Central 

Google Scholar 
Quque, M. et al. Hierarchical networks of food exchange in the black garden ant Lasius niger. Insect Sci. 28, 825–838 (2021).Article 
PubMed 

Google Scholar 
Cassill, D. L. & Tschinkel, W. R. A duration constant for worker-to-larva trophallaxis in fire ants. Ins. Soc. 43, 149–166 (1996).Article 

Google Scholar 
Cassill, D. L. & Tschinkel, W. R. Regulation of diet in the fire ant, Solenopsis invicta. J. Insect Behav. 12, 307–328 (1999).Article 

Google Scholar 
Wilson, E. O. & Eisner, T. Quantitative studies of liquid food transmission in ants. Ins. Soc. 4, 157–166 (1957).Article 

Google Scholar 
Markin, G. P. Food distribution within laboratory colonies of the argentine ant, Tridomyrmex humilis (Mayr). Ins. Soc. 17, 127–157 (1970).Article 

Google Scholar 
Howard, D. F. & Tschinkel, W. R. The flow of food in colonies of the fire ant, Solenopsis invicta: a multifactorial study. Physiol. Entomol. 6, 297–306 (1981).Article 

Google Scholar 
Suryanarayanan, S. & Jeanne, R. L. Antennal drumming, trophallaxis, and colony development in the social wasp Polistes fuscatus (Hymenoptera: Vespidae). Ethology 114, 1201–1209 (2008).Article 

Google Scholar 
Greenwald, E., Segre, E. & Feinerman, O. Ant trophallactic networks: simultaneous measurement of interaction patterns and food dissemination. Sci. Rep. 5, 12496 (2015).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Baltiansky, L., Sarafian-Tamam, E., Greenwald, E. & Feinerman, O. Dual-fluorescence imaging and automated trophallaxis detection for studying multi-nutrient regulation in superorganisms. Methods Ecol. Evol. 12, 1441–1457 (2021).Article 

Google Scholar 
Feldhaar, H. et al. Stable isotopes: past and future in exposing secrets of ant nutrition (Hymenoptera: Formicidae). Myrmecol. N. 13, 3–13 (2010).
Google Scholar 
Bouchebti, S., Bodner, L., Bergman, M., Magory Cohen, T. & Levin, E. The effects of dietary proline, β-alanine, and γ-aminobutyric acid (GABA) on the nest construction behavior in the Oriental hornet (Vespa orientalis). Sci. Rep. 12, 7449 (2022).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Motro, M., Motro, U., Ishay, J. S. & Kugler, J. Some social and dietary prerequisites of oocyte development in Vespa orientalis L. workers. Ins. Soc. 26, 155–164 (1979).Article 

Google Scholar 
Levin, E., McCue, M. D. & Davidowitz, G. More than just sugar: allocation of nectar amino acids and fatty acids in a Lepidopteran. Proc. R. Soc. B: Biol. Sci. 284, 20162126 (2017).Article 

Google Scholar 
Wright, G. A., Nicolson, S. W. & Shafir, S. Nutritional physiology and ecology of honey bees. Annu. Rev. Entomol. 63, 327–344 (2018).Article 
CAS 
PubMed 

Google Scholar 
Helm, B. R. et al. The geometric framework for nutrition reveals interactions between protein and carbohydrate during larval growth in honey bees. Biol. Open 6, 872–880 (2017).CAS 
PubMed 
PubMed Central 

Google Scholar 
Paoli, P. P. et al. Nutritional balance of essential amino acids and carbohydrates of the adult worker honeybee depends on age. Amino Acids 46, 1449–1458 (2014).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Stabler, D., Paoli, P. P., Nicolson, S. W. & Wright, G. A. Nutrient balancing of the adult worker bumblebee (Bombus terrestris) depends on the dietary source of essential amino acids. J. Exp. Biol. 218, 793–802 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Arganda, S. et al. Parsing the life-shortening effects of dietary protein: effects of individual amino acids. Proc. R. Soc. B 284, 20162052 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Csata, E. & Dussutour, A. Nutrient regulation in ants (Hymenoptera: Formicidae): a review. Myrmecol. N. 29, 111–124 (2019).
Google Scholar 
Gottsberger, G., Schrauwen, J. & Linskens, H. F. Amino acids and sugars in nectar, and their putative evolutionary significance. Pl. Syst. Evol. 145, 55–77 (1984).Article 
CAS 

Google Scholar 
Ozimek, L. et al. Nutritive value of protein extracted from honey bees. J. Food Sci. 50, 1327–1329 (1985).Article 
CAS 

Google Scholar 
Nicolson, S. W. & Thornburg, R. W. Nectar chemistry. in Nectaries and Nectar (eds. Nicolson, S. W., Nepi, M. & Pacini, E.) 215–264 (Springer Netherlands, 2007). https://doi.org/10.1007/978-1-4020-5937-7_5.Contrera, F. A. L., Imperatriz-Fonseca, V. L. & Koedam, D. Trophallaxis and reproductive conflicts in social bees. Insect Soc. 57, 125–132 (2010).Article 

Google Scholar 
Carter, G. G. & Wilkinson, G. S. Food sharing in vampire bats: reciprocal help predicts donations more than relatedness or harassment. Proc. R. Soc. B: Biol. Sci. 280, 20122573 (2013).Article 

Google Scholar 
Nalepa, C. A. Origin of termite eusociality: trophallaxis integrates the social, nutritional, and microbial environments. Ecol. Entomol. 40, 323–335 (2015).Article 

Google Scholar 
Werenkraut, V., Arbetman, M. P. & Fergnani, P. N. The Oriental hornet (Vespa orientalis L.): a threat to the Americas? Neotrop. Entomol. 51, 330–338 (2022).Article 
PubMed 

Google Scholar 
Darchen, R. Biologie de Vespa orientalis. Les premiers stades de développement. Ins. Soc. 11, 141–157 (1964).Article 

Google Scholar 
Van Itterbeeck, J. et al. Rearing techniques for hornets with emphasis on Vespa velutina (Hymenoptera: Vespidae): A review. J. Asia-Pac. Entomol. 24, 103–117 (2021).Article 

Google Scholar 
R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. (2020).Bouchebti, S., Bodner,L. & Levin, E. Continuous exchange of nectar nutrients in an Oriental hornet colony- Dataset [Data set]. Zenodo. https://doi.org/10.5281/zenodo.7135100 (2022). More

Study areaThe present study was conducted in Uttarkashi district, Uttarakhand, located between 38° 28′ to 31°28′ N latitude and 77°49′ to 79°25′ E longitude with an area of about 8016 km2, covering primarily hilly terrain with an altitudinal range of 715–6717 m (Fig. 3). The terrain is mountainous, consisting of undulating hill ranges and narrow valleys with temperate climatic conditions. The district lies in the upper catchment of two major rivers of India, viz., the Ganges (Bhagirathi towards upstream) and the Yamuna. The major vegetation types of the study area are Himalayan moist temperate forest, sub-alpine forest and alpine scrub59. The Uttarkashi district forests are managed under three Forest Divisions viz., (i) Uttarkashi Forest Division (ii) Upper Yamuna Badkot Forest Division and (iii) Tons Forest Division) with two Protected Areas (PAs) (i) Gangotri National Park and (ii) Govind Pashu Vihar National Park. The forested habitats of the study landscape are home to top conservation priority species, including Asiatic Black bear (Ursus thibetanus), Musk deer (Moschus spp.), Common leopard (Panthera pardus), Himalayan brown bear (Ursus arctos isabellinus) and Western Tragopan (Tragopan melanocephalus), Himalayan monal (Lophophorus impejanus). The study was conducted after a study permit issued by the Chief Wildlife Warden, Forest Department, Uttarakhand government, vide letter no. 848/5-6 dated 31/08/2019, we have not handled the species for doing research. Instead, remote camera traps have been used for collecting the data with the permission of the Chief Wildlife Warden, Government of Uttarakhand. Further, informed consent was taken before interviewing the local communities. The data was collected according to the institutional guidelines and approved by the Research Advisory and Monitoring Committee of the Zoological Survey of India.Figure 3Map of the study area Uttarkashi, Uttarakhand. ArcGIS 10.6 (ESRI, Redlands, CA) was used to create the map. (Map created using ArcGIS 10.6; http://www.esri.com).Full size imageSampling protocolThe basic sampling protocol and assumptions for multi-species occupancy modelling are identical to the single-species case7. Briefly, a set of 62 intensive sites, were randomly selected, and each site i was surveyed j times. During each survey, detection/non-detection of S focal species was recorded. Additionally, direct or indirect evidences of species presence from the different areas were also recorded.Data collectionThe complete study area was divided into 10 × 10 km grids, consisting of n = 60 grids. Based on the reconnaissance survey, out of these 60 grids, we selected 25 girds that were accessible to conduct the survey and have the species presence. Further, these grids were divided into 2 × 2 km grids to maximize our effort so that all logistically accessible grids could be covered, and we conducted intensive sampling in N = 62 grids after excluding the grids with human settlements. T The field surveys were conducted during 2018–2019, and a team of researchers systematically visited selected grids to collect data on the detection/non-detection of these ungulates. A total of 62 camera traps were deployed in selected grids, and 650 km were traversed, accounting for N = 54 trails in these sampled grids. These camera traps were visited once in every fifteen days for replacing the batteries as well as documenting the presence of the species through the sign surveys. The ultra-compact SPYPOINT FORCE-11D trail camera (SPYPOINT, GG Telecom, Canada, QC) and Browning trail camera (Defender 850, 20 MP, Prometheus Group, LLC Birmingham, Alabama, https://browningtrailcameras.com) camera traps were used to detect the presence/absence of ungulate species. The cameras were mounted 40–60 cm above ground on natural trails without lures.Data explorationWhile deploying camera traps, we also noted habitat variables through on-site observation such as distance to the village and human disturbance. We tested site covariates for collinearity and discarded one of a pair if the Pearson’s correlation was greater than 0.760. Hence, we assumed each of the site covariates could influence the occupancy and detectability of these ungulates.CovariatesWe hypothesized that habitat variables may influence these ungulates’ occupancy and detection probability. A total of 21 variables were extracted either from the field or using the ArcGIS v. 10.6 software (ESRI, Redlands, CA), and only 14 were retained after collinearity testing60 (Table 3). These covariates were classified into the following categories (Topographic variables, Habitat variables and anthropogenic variables). The topographic variables (elevation, slope and aspect) were generated using 30× resolution SRTM (Shuttle Radar Topography Mission) image downloaded from EarthExplorer (https://earthexplorer.usgs.gov/). The habitat/ land cover classification was carried out using Landsat 8 satellite imagery (Spatial resolution = 30 m) downloaded from Global Land Cover Facility by following the methodology suggested by61 using the ArcGIS v. 10.6 software (ESRI, Redlands, CA). The study area was classified into nine Land use/land cover (LULC) classes viz., West Himalayan Sub-alpine birch/fir Forest (FT 188), West Himalayan upper oak/fir forest (FT 162), West Himalayan Dry juniper forest (FT 180), Ban oak forest (FT 152), Moist Deodar Forest (FT 155), Western mixed coniferous forest (FT 156), Moist temperate Deciduous Forest (FT 157) which were used for further analysis considering their importance to species ecology and behavior60. The values for all the covariates were extracted at 30 m resolution, and a single value per site was obtained by averaging all the pixel values within each sampling site (camera trap locations).Table 3 Habitat variables used for multi species occupancy analysis of three ungulate species in Uttarkashi, Uttarakhand.Full size tableOccupancy modelling frameworkWe used multi-species occupancy modelling62 of barking deer, goral and sambar to estimate the probability of the species (s) occurred within the area (i) sampled during our survey period (j), for accounting the imperfect detection of the species8. Distinguishing the true presence/absence of a species from detection/non-detection (i.e., species present and captured or species present but not captured) requires spatially or temporally replicated data. We used camera stations to record the presence/absence of species along with sign survey in all the studied grids. The camera traps were placed along the trail/transects in the studied grids hence each grid needs to be visited once in every fifteen days to check the camera traps as well as to document the presence of the studied species. Therefore, we treated 15 trap nights as one sampling occasion at a particular camera station resulting in ~ 7 sampling occasions per camera station.Our aim was to record the presence/ absence of the species at a particular gird hence we incorporated sign survey data if the species was not detected in camera station but recorded through sign survey. We pooled the presence/absence data in a single sheet of each species following6 and fitted occupancy and detectability models using programme Mark63,64. We model the species (s) presence (ysij = 1) and absence (ysij = 0) at site i during survey j, and the sampling protocol was identical to single species case65, where the Bernoulli random variable was conditional on the presence of species s (Zs = 1) following6$${text{y}}_{sij} sim {text{ Bernoulli}}left( {{text{p}}_{sij} {text{z}}_{si} } right),$$
where Psij represents the probability of detecting species S during replicate survey j at site i and Zsi = presence or absence of species s at site i.Furthermore, we model the latent occupancy state of species s at site i as a multivariate Bernoulli random variable:$${text{Z}}_{i} sim {text{MVB}}left( {uppsi _{i} } right)$$
where Zi = {Z1i, Z2i….., ZSi} is an S-dimensional vector of 1’s and 0’s denoting the latent occupancy state of all S species and (ψi) is a 2S-dimensional vector denoting the probability of all possible sequences of 1’s and 0’s Zi can attain such that ∑ ψi = 1 with corresponding probability mass function (PMF) adopted from6,64.$$fleft( {{text{Z}}_{i} } right) = {text{ exp}}left( {left( {{text{Z}}_{i} {text{log}}(uppsi_{{text{i}}} {1}/uppsi_{{text{i}}} 0} right) , + {text{ log}}left( {uppsi_{{text{i}}} 0} right)} right).$$The quantity f = log (ψi1/ψi0), is the log odds species S occupies a site often referred to as a ‘natural parameter’.Since we are modeling three ungulate species (S = 3), 2S = 23 the possible encounter histories included in the dataset were eight, if neither of the two species were detected the value of ‘00’ was assigned; similarly ‘01’ indicates detection of species 1; ‘02’ indicates detection of species 2; ‘03’ indicates detection of both the species; ‘04’ indicates detection of species 3; ‘05’ indicates detection of species 1 and species 3; ‘06’ indicates detection of species 2 and species 3 and ‘07’ indicates detection of all the three species. We modelled constant occupancy and detection probability for each of the three species. Hence, we specified 6 f and p parameters, an intercept (β) for each of one-way f parameter and detection parameter p following64.$$f_{{1}}=upbeta_{{{1},}} ;;{text{p}}=upbeta_{{4}}$$$$f_{{2}} = upbeta_{{{2},}} ;{text{p }} = , upbeta 5$$$$f_{{3}} = , upbeta_{{{3},}}; {text{p }} = , upbeta_{{6}}$$We fit a set of models including the detection probability as a constant, p(.), and variable function to occupancy ψ(covariate) for site-specific covariates and models include occupancy as constant ψ(.) and variable function of the detection p(covariates) for the respective site covariates.As we have assumed the independence among all three species, the model shows marginal occupancy probabilities of species 1, species 2 and species 3 varies as a function of environmental variables. We incorporated site-level characteristics affecting species-specific occurrence (f1: occupancy of species 1, f2: occupancy of species 2, & f3: occupancy of species 3) and detection probabilities using a generalized linear modelling approach42. This requires 9 parameters: an intercept (β1, β3, β5) and slope (β2, β4, β6) coefficient for each 1-way f parameter f1, f2, f3 and an intercept parameter for each detection parameter (β7, β8, β9). Below mentioned is the model for 1-way f parameters.$$f_{{1}} = , upbeta_{{{1 } + }} upbeta_{{2}} left( {{text{Covariate}}} right),;;{text{ p }} = , upbeta_{{7}}$$$$f_{{2}} = , upbeta_{{{3 } + }} upbeta_{{4}} left( {{text{Covariate}}} right),;;{text{ p}} = , upbeta_{{8}}$$$$f_{{3}} = , upbeta_{{5}} + , upbeta_{{6}} left( {{text{Covariate}}} right),;;{text{ p }} = , upbeta_{{9}} .$$All covariates were standardized before model fitting. We fitted the most complex model to each species and considered all possible combinations of covariates using the logit link function. Our rationale for including these variables in the occupancy and detectability component of the model was that we expected these variables to influence the occupancy and detectability of the study species.Since multi-species occupancy simultaneously model environmental variables, & interspecific interactions. Further it also allows to understand the influence of environmental variables on one species occupancy, in the presence or absence of other sympatric species64. Hence, we also modeled two species occur together as a function of covariates. We examined how the variables of each camera site influenced the pair-wise interaction of the three ungulate species. This model assumes that the conditional probability of one species varies in the presence or absence of other species. We assumed f123: co-occurrence of species 1, species 2 & species 3 = 0, hence we did not include higher-order interactions in any of our models, we assumed the conditional probability of 3 species occurred together was purely a function of species-specific (f1, f2, f3) and pair-wise interaction (f12: co-occurrence of species1 & species 2, f13: co-occurrence of species 1 & species 3, f23: co-occurrence of species 2 & species 3) parameters. We modeled pair-wise interaction of species varies as a function of environmental variables keeping detection probability constant. Hence, we specified 15 f and p parameters, an intercept and slope coefficient for each of the one-way (f1, f2, f3) and the two-way f parameters (f12, f13, and f23); as well as an intercept parameter for each of the detection models. The model equation below implies for 2-way f parameters:$$f_{{{12}}} = , upbeta_{{{7 } + }} upbeta_{{8}} left( {{text{Covariate}}} right),;;{text{ p }} = , upbeta_{{{13}}}$$$$f_{{{13}}} = , upbeta_{{{9 } + }} upbeta_{{{1}0}} left( {{text{Covariate}}} right),;;{text{ p }} = , upbeta_{{{14}}}$$$$f_{{{23}}} = , upbeta_{{{11 } + }} upbeta_{{{12}}} left( {{text{Covariate}}} right),;;{text{ p }} = , upbeta_{{{15}}} .$$We also fitted models including co-occurrence and detection probability of a species varies as a function of environmental variables. Hence, we specified 18 f and p parameters, an intercept and slope coefficient for each of one-way (f1, f2, f3) and two-way f parameters (f12, f13, f23); and an intercept as well as the slope parameters for each of the detection models. The model equation below implies for 2-way f parameters:$$f_{{{12}}} = , upbeta_{{{7 } + }} upbeta_{{8}} left( {{text{Covariate}}} right),{text{ p }} = , upbeta_{{{13 } + }} upbeta_{{{14}}} left( {{text{covariate}}} right)$$$$f_{{{13}}} = , upbeta_{{{9 } + }} upbeta_{{{1}0}} left( {{text{Covariate}}} right),{text{ p }} = , upbeta_{{{15}}} + , upbeta_{{{16}}} left( {{text{covariate}}} right)$$$$f_{{{23}}} = , upbeta_{{{11 } + }} upbeta_{{{12}}} left( {{text{Covariate}}} right),{text{ p }} = , upbeta_{{{17}}} + , upbeta_{{{18}}} left( {{text{covariate}}} right)$$A total of 38 models were run to test the influence of environmental variables on occupancy and detection probability of species-specific (f1, f2, f3) and pair-wise interaction of the three ungulate species. The best-supported model was identified by selecting the model with the lowest AICc value and highest model weights66, where higher model weights indicate a better fit of the model to the data. Second-Order Information Criterion (AICc)67 values were used to rank the occupancy models, and all the models whose ΔAICc  More

We hypothesized that functionally distinct forest types can be mapped at moderate spatial resolutions, using a combination of canopy foliar traits and canopy structure information. Our analysis of LiDAR and imaging spectroscopy data at spatial resolutions ranging from 4 to 200 m (16 m2–40,000 m2), with an emphasis on the 30 m (900 m2) spaceborne hyperspectral spatial resolution, reveals that few remotely sensed canopy properties are needed to successfully identify ecologically distinct forest types at two diverse tropical forest sites in Malaysian Borneo. In testing our second hypothesis that mapped forest types exhibit distinct ecosystem function, we found that forest types identified using remotely sensed leaf P, LMA, Max H, and canopy cover at 20 m height (Cover20) closely align with forest types defined from field-based floristic surveys29,30,31,32,33 and inventory plot-based measurements of growth and mortality rates (Fig. 4b). Our approach, however, enables mapping of their entire spatial extent (Fig. 1) and reveals important structural and functional variation within areas characterized as a single forest type in previous studies (Fig. 3). Current and forthcoming satellite hyperspectral platforms, including PRISMA (30 m), CHIME (20–30 m), and SBG (30 m), have or will have comparable spectral resolution, higher temporal revisits, and much greater geographic coverage. The ability to conduct this type of analysis using remote sensing measurements at 30 m resolution suggests that our method can be applied to these emerging spaceborne imaging spectroscopy data to reveal important differences in structure and function across the world’s tropical forests.Nested functional forest types revealedTo test our first hypothesis, rather than making an a priori decision about the number of k-means clusters (k), we explored the capacity of remotely sensed data to reveal ecologically relevant variation in forest types. Baldeck and Asner took a similar unsupervised approach to estimating beta diversity in South Africa34. Because the choice of k directly influences analysis outcomes, careful selection of k is required. Different approaches for identifying the number of clusters, using the Gapk and Wk elbow metrics35, yielded varying optimal numbers of clusters for the Sepilok and Danum landscapes (Fig. 1, Supplementary Figs. 4 and 5). However, at both sites, a comparison of results based on different values of k revealed ecologically meaningful structural and functional differences and graduated transitions between forest types (Fig. 2, Supplementary Figs. 7 and 8), indicating that the exploration of traits that aggregate or separate forest types as k changes is a valuable exercise. Overlap between the remotely sensed forest type boundaries and inventory plots within distinct forest types indicate that the series of clustered forests align closely with forest types defined based on in situ data on species composition and ecosystem structure. In part, this type of analysis requires careful selection of the number of clusters. Additionally, however, we gained valuable insights via the exploration of varying numbers of clusters as it relates to biologically meaningful categorization of forest types. Extending this method to other parts of the tropics will require similar decision-making, which will either require user input, or the development of robust automated algorithms for selecting k.Forest types capture differences in ecosystem dynamicsWe further evaluated the canopy traits and structural attributes that were most critical for mapping distinct forest types, hypothesizing that mapped forest types exhibit distinct ecosystem function. Forest types revealed by the cluster analyses were distributed along the leaf economic spectrum, where the leaf economic spectrum characterizes a tradeoff in plant growth strategies36. LMA, which can covary strongly with leaf N and P, is a key indicator of plant growth strategies along the spectrum37. At the slow-return end of the leaf economics spectrum, plants in nutrient-poor conditions with low leaf nutrient concentrations invest in leaf structure and defense, expressed as high LMA, strategizing longer-lived, tougher leaves with slower decomposition rates. This strategy comes at the cost of slower growth. At the quick-return end of the spectrum, plants in nutrient-rich environments with higher leaf nutrient concentrations invest less in structure and defense, enabling faster growth and more rapid leaf turnover, i.e., shorter leaf lifespans. This quick-return growth strategy supports higher photosynthetic rates and more rapid carbon gain36.In this study, the principal components and clustering results yielded forest types that are indicative of community level differences associated with leaf economic spectrum differences. The nutrient rich sites (Danum1 and Danum2, Supplementary Fig. 8) show high canopy N and P and low LMA compared to the nutrient poor and acidic sites (Sandstone and Kerangas), which contributes to lower leaf photosynthetic capacity (Vcmax) and growth (Fig. 4b). Foliar N:P also increased with site fertility, confirming that tropical forests are primarily limited by phosphorus, and not nitrogen38,39, with large implications for carbon sequestration in these forests. Orthogonal differences in canopy structure and architecture between Danum forest types and Sepilok Sandstone and Alluvial forests could be indicative of ecosystem scale differences in the sensitivity of these forests to endogenous disturbance processes40.The significant differences in aboveground carbon stocks and growth and mortality rates between forest types further suggests strong differences in ecosystem dynamics. In general, growth rates varied inversely to aboveground carbon, and higher aboveground carbon corresponded to lower mortality rates. As an example, the Sepilok sandstone forests, which are largely comprised of slow-growing dipterocarp species29,33, had the highest median aboveground carbon (236 Mg C ha−1), with higher canopy P and N, and lower LMA. The taller canopy and low canopy leaf nutrient concentrations are consistent with the low growth and mortality rates found in the sandstone forest, indicating a slow-growth strategy yielding larger trees and higher aboveground carbon stocks. In contrast, alluvial forests exhibit high turnover with mortality and growth rates higher relative to Sandstone forests corresponding to lower aboveground carbon on average. Kerangas forests exhibited low aboveground carbon despite an intermediate plot-level growth rate, and mortality rates that were significantly lower than the Danum or alluvial forest types. Kerangas forests, which were characterized by the highest LMA, lowest foliar P and N (Fig. 2a), and the lowest plot-level aboveground carbon density (186 Mg C ha−1; Fig. 4a), are known to have higher stem densities, lower canopy heights, and long-lived leaves5,32,41, suggesting well-developed strategies for nutrient retention42. Interestingly, despite significantly different aboveground carbon and demography, the kerangas and sandstone forests did not differ in LAI or canopy architecture (P:H); although maximum height, Cover20, and Hpeak LAI were significantly higher in the sandstone forest, highlighting the need to account for differences beyond LAI when scaling processes from leaves to ecosystems.In addition, when three forest types were distinguished at Sepilok, the alluvial inventory plot had significantly higher aboveground carbon than the remote sensing-derived alluvial forest extent (Fig. 4a, p  More

Bates, H. W. XXXII. Contributions to an insect fauna of the Amazon Valley. Lepidoptera: Heliconidæ. Trans. Linn. Soc. Lond 23, 495–566 (1862).Article 

Google Scholar 
Müller, F. Ituna and thyridia: A remarkable case of mimicry in butterflies. Trans. Entomol. Soc. Lond. 1879, 20–29 (1879).
Google Scholar 
Baxter, S. W. et al. Convergent evolution in the genetic basis of Müllerian mimicry in Heliconius butterflies. Genetics 180, 1567–1577 (2008).Article 
CAS 

Google Scholar 
Sheppard, P. M., Turner, J. R. G., Brown, K., Benson, W. & Singer, M. Genetics and the evolution of Muellerian mimicry in Heliconius butterflies. Philos. Trans R. Soc. Lond. B, Biol. Sci. 308, 433–610 (1985).Article 
ADS 

Google Scholar 
Mallet, J. & Gilbert, L. E. Jr. Why are there so many mimicry rings? Correlations between habitat, behaviour and mimicry in Heliconius butterflies. Biol. J. Lin. Soc. 55, 159–180 (1995).
Google Scholar 
Brower, A. V. Parallel race formation and the evolution of mimicry in Heliconius butterflies: A phylogenetic hypothesis from mitochondrial DNA sequences. Evolution 50, 195–221 (1996).Article 
CAS 

Google Scholar 
Wilson, J. S. et al. North American velvet ants form one of the world’s largest known Müllerian mimicry complexes. Curr. Biol. 25, R704–R706. https://doi.org/10.1016/j.cub.2015.06.053 (2015).Article 
CAS 
PubMed 

Google Scholar 
Wilson, J. S., Williams, K. A., Forister, M. L., Von Dohlen, C. D. & Pitts, J. P. Repeated evolution in overlapping mimicry rings among North American velvet ants. Nat. Commun. 3, 1272 (2012).Article 
ADS 

Google Scholar 
Wilson, J. S., Pan, A. D., Limb, E. S. & Williams, K. A. Comparison of African and North American velvet ant mimicry complexes: Another example of Africa as the ‘odd man out’. PLoS ONE 13, e0189482. https://doi.org/10.1371/journal.pone.0189482 (2018).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Plowright, R. & Owen, R. E. The evolutionary significance of bumble bee color patterns: A mimetic interpretation. Evolution 34, 622–637 (1980).Article 
CAS 

Google Scholar 
Williams, P. The distribution of bumblebee colour patterns worldwide: Possible significance for thermoregulation, crypsis, and warning mimicry. Biol. J. Lin. Soc. 92, 97–118 (2007).Article 

Google Scholar 
Hines, H. M. & Williams, P. H. Mimetic colour pattern evolution in the highly polymorphic Bombus trifasciatus (Hymenoptera: Apidae) species complex and its comimics. Zool. J. Linn. Soc. 166, 805–826 (2012).Article 

Google Scholar 
Koch, J. B., Rodriguez, J., Pitts, J. P. & Strange, J. P. Phylogeny and population genetic analyses reveals cryptic speciation in the Bombus fervidus species complex (Hymenoptera: Apidae). PLoS ONE 13, e0207080 (2018).Article 

Google Scholar 
Ezray, B. D., Wham, D. C., Hill, C. E. & Hines, H. M. Unsupervised machine learning reveals mimicry complexes in bumblebees occur along a perceptual continuum. Proc. R. Soc. B 286, 20191501 (2019).Article 

Google Scholar 
Bateson, W. The alleged “Aggressive Mimicry” of volucellæ. Nature 46, 585 (1892).Article 
ADS 

Google Scholar 
Poulton, E. B. The volucellœ as alleged examples of variation “almost unique among animals”. Nature 47, 126 (1892).Article 
ADS 

Google Scholar 
Cockerell, T. D. New social bees. Psyche A J. Entomol. 24, 120–128 (1917).Article 

Google Scholar 
Koch, J., Strange, J. & Williams, P. In: Bumble bees of the western United States (US Forest Service, San Francisco California, 2012).
Google Scholar 
Williams, P. H., Thorp, R. W., Richardson, L. L. & Colla, S. R. In: Bumble bees of North America: An identification guide Vol. 87 (Princeton University Press, Princeton, 2014).
Google Scholar 
Ruxton, G. D., Franks, D. W., Balogh, A. C. & Leimar, O. Evolutionary implications of the form of predator generalization for aposematic signals and mimicry in prey. Evol Int. J. Org. Evol. 62, 2913–2921 (2008).Article 

Google Scholar 
Rowe, C., Lindström, L. & Lyytinen, A. The importance of pattern similarity between Müllerian mimics in predator avoidance learning. Proc. R. Soc. Lond. Ser. B Biol. Sci. 271, 407–413 (2004).Article 

Google Scholar 
Beatty, C. D., Beirinckx, K. & Sherratt, T. N. The evolution of Müllerian mimicry in multispecies communities. Nature 431, 63 (2004).Article 
ADS 
CAS 

Google Scholar 
Chittka, L. & Osorio, D. Cognitive dimensions of predator responses to imperfect mimicry. PLoS Biol. 5, e339 (2007).Article 

Google Scholar 
Dittrigh, W., Gilbert, F., Green, P., McGregor, P. & Grewcock, D. Imperfect mimicry: A pigeon’s perspective. Proc. R. Soc. Lond. Ser. B Biol. Sci. 251, 195–200 (1993).Article 
ADS 

Google Scholar 
Sherratt, T. N., Whissell, E., Webster, R. & Kikuchi, D. W. Hierarchical overshadowing of stimuli and its role in mimicry evolution. Anim. Behav. 108, 73–79 (2015).Article 

Google Scholar 
Beatty, C. D., Bain, R. S. & Sherratt, T. N. The evolution of aggregation in profitable and unprofitable prey. Anim. Behav. 70, 199–208 (2005).Article 

Google Scholar 
Kazemi, B., Gamberale-Stille, G., Tullberg, B. S. & Leimar, O. Stimulus salience as an explanation for imperfect mimicry. Curr. Biol. 24, 965–969 (2014).Article 
CAS 

Google Scholar 
Kikuchi, D. W., Dornhaus, A., Gopeechund, V. & Sherratt, T. N. Signal categorization by foraging animals depends on ecological diversity. Elife. 8, e43965 (2019).Article 

Google Scholar 
Rapti, Z., Duennes, M. A. & Cameron, S. A. Defining the colour pattern phenotype in bumble bees (Bombus): A new model for evo devo. Biol. J. Lin. Soc. 113, 384–404 (2014).Article 

Google Scholar 
Wilson, J. S., Sidwell, J. S., Forister, M. L., Williams, K. A. & Pitts, J. P. Thistledown velvet ants in the desert mimicry ring and the evolution of white coloration: Müllerian mimicry, camouflage and thermal ecology. Biol. Lett. 16, 20200242 (2020).Article 

Google Scholar 
Ascher, J. & Pickering, J. Discover Life bee species guide and world checklist (Hymenoptera: Apoidea: Anthophila) (2019).iNaturalist. Available from https://www.inaturalist.org. Accessed [2022].Bombus Latreille, 1802 in GBIF Secretariat (2021). GBIF Backbone Taxonomy. Checklist dataset https://doi.org/10.15468/39omei accessed via GBIF.org on 2021-12-03. More

Sakschewski, B. et al. Leaf and stem economics spectra drive diversity of functional plant traits in a dynamic global vegetation model. Glob. Change Biol. 21, 2711–2725 (2015).Article 

Google Scholar 
Berzaghi, F. et al. Towards a new generation of trait-flexible vegetation models. Trends Ecol. Evol. 35, 191–205 (2020).Article 
PubMed 

Google Scholar 
Bruelheide, H. et al. Global trait–environment relationships of plant communities. Nat. Ecol. Evol. 2, 1906–1917 (2018).Article 
PubMed 

Google Scholar 
Joswig, J. S. et al. Climatic and soil factors explain the two-dimensional spectrum of global plant trait variation. Nat. Ecol. Evol. 6, 36–50 (2021).Article 
PubMed 
PubMed Central 

Google Scholar 
van Bodegom, P. M., Douma, J. C. & Verheijen, L. M. A fully traits-based approach to modeling global vegetation distribution. Proc. Natl Acad. Sci. USA 111, 13733–13738 (2014).PubMed Central 

Google Scholar 
Moreno Martínez, A. et al. A methodology to derive global maps of leaf traits using remote sensing and climate data. Remote Sens. Environ. 218, 69–88 (2018).Article 

Google Scholar 
Pérez-Harguindeguy, N. et al. New handbook for standardized measurment of plant functional traits worldwide. Aust. J. Bot. 23, 167–234 (2013).Article 

Google Scholar 
Kattge, J. et al. TRY—a global database of plant traits. Glob. Change Biol. 17, 2905–2935 (2011).Article 

Google Scholar 
Kattge, J. et al. TRY plant trait database-enhanced coverage and open access. Glob. Change Biol. 26, 119–188 (2020).Article 

Google Scholar 
Jetz, W. et al. Monitoring plant functional diversity from space. Nat. Plants 2, 16024 (2016).Article 
PubMed 

Google Scholar 
Butler, E. E. et al. Mapping local and global variability in plant trait distributions. Proc. Natl Acad. Sci. USA 114, E10937–E10946 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Boonman, C. C. et al. Assessing the reliability of predicted plant trait distributions at the global scale. Glob. Ecol. Biogeogr. 29, 1034–1051 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Madani, N. et al. Future global productivity will be affected by plant trait response to climate. Sci. Rep. 8, 2870 (2018).Vallicrosa, H. et al. Global distribution and drivers of forest biome foliar nitrogen to phosphorus ratios (N:P). Glob. Ecol. Biogeogr. 31, 861–871 (2022).Article 

Google Scholar 
Meyer, H. & Pebesma, E. Predicting into unknown space? Estimating the area of applicability of spatial prediction models. Methods Ecol. Evol. 12, 1620–1633 (2021).Article 

Google Scholar 
Schiller, C. et al. Deep learning and citizen science enable automated plant trait predictions from photographs. Sci. Rep. 11, 16395 (2021).Aguirre-Gutiérrez, J. et al. Pantropical modelling of canopy functional traits using sentinel-2 remote sensing data. Remote Sens. Environ. 252, 112–122 (2021).Article 

Google Scholar 
Homolova, L. et al. Review of optical-based remote sensing for plant trait mapping. Ecol. Complex. 15, 1–16 (2013).Article 

Google Scholar 
Van Cleemput, E. et al. The functional characterization of grass-and-shrubland ecosystems using hyperspectral remote sensing: trends, accuracy and moderating variables. Remote Sens. Environ. 209, 747–763 (2018).Article 

Google Scholar 
Kattenborn, T., Fassnacht, F. E. & Schmidtlein, S. Differentiating plant functional types using reflectance: which traits make the difference? Remote Sens. Ecol. Conserv. 5, 5–19 (2019).Article 

Google Scholar 
Hauser, L. T. et al. Explaining discrepancies between spectral and in-situ plant diversity in multispectral satellite earth observation. Remote Sens. Environ. 265, 112684 (2021).Article 

Google Scholar 
Wäldchen, J. & Mäder, P. Plant species identification using computer vision techniques: a systematic literature review. Arch. Comput. Methods Eng. 25, 507–543 (2018).Article 
PubMed 

Google Scholar 
Jones, H. G. What plant is that? Tests of automated image recognition apps for plant identification on plants from the British flora. AoB Plants 12, plaa052 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Hampton, S. E. et al. Big data and the future of ecology. Front. Ecol. Environ. 11, 156–162 (2013).Article 

Google Scholar 
WÜest, R. O. et al. Macroecology in the age of big data—where to go from here? J. Biogeogr. 47, 1–12 (2020).Article 

Google Scholar 
Mäder, P. et al. The Flora Incognita app—interactive plant species identification. Methods Ecol. Evol. 12, 1335–1342 (2021).Article 

Google Scholar 
Di Cecco, G. J. et al. Observing the observers: how participants contribute data to iNaturalist and implications for biodiversity science. BioScience 71, 1179–1188 (2021).Article 

Google Scholar 
Mahecha, M. D. et al. Crowd-sourced plant occurrence data provide a reliable description of macroecological gradients. Ecography 44, 1131–1142 (2021).Article 

Google Scholar 
Botella, C. et al. Jointly estimating spatial sampling effort and habitat suitability for multiple species from opportunistic presence-only data. Methods Ecol. Evol. 12, 933–945 (2021).Article 

Google Scholar 
iNaturalist Research-Grade Observations (GBIF, accessed 5 January 2022); https://www.gbif.org/dataset/50c9509d-22c7-4a22-a47d-8c48425ef4a7Callaghan, C. T. et al. Three frontiers for the future of biodiversity research using citizen science data. BioScience 71, 55–63 (2020).
Google Scholar 
Dickinson, J. L., Zuckerberg, B. & Bonter, D. N. Citizen science as an ecological research tool: challenges and benefits. Ann. Rev. Ecol. Evol. Syst. 41, 149–172 (2010).Article 

Google Scholar 
Kosmala, M. et al. Assessing data quality in citizen science. Front. Ecol. Environ. 14, 551–560 (2016).Article 

Google Scholar 
Boakes, E. H. et al. Patterns of contribution to citizen science biodiversity projects increase understanding of volunteers’ recording behaviour. Sci. Rep. 6, 33051 (2016).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Bowler, D.E. et al. Temporal trends in the spatial bias of species occurrence records. Ecography 2022, e06219 (2022). https://doi.org/10.1111/ecog.06219GBIF Occurrence Download (GBIF, 4 January 2022); https://doi.org/10.15468/dl.34tjreBruelheide, H. et al. sPlot—a new tool for global vegetation analyses. journal of vegetation science. J. Veg. Sci. 30, 161–186 (2019).Article 

Google Scholar 
Sabatini, F. et al. sPlotOpen—an environmentally balanced, open access, global dataset of vegetation plots. Glob. Ecol. Biogeogr. 30, 1740–1764 (2021).Article 

Google Scholar 
Whittaker, R.H. et al. Communities and Ecosystems (Macmillan/Collier Macmillan, 1970).Olson, D. M. et al. Terrestrial ecoregions of the world: a new map of life on Earth: a new global map of terrestrial ecoregions provides an innovative tool for conserving biodiversity. BioScience 51, 933–938 (2001).Article 

Google Scholar 
Joswig, J., Wirth, C. & Schuman, M. Climatic and soil factors explain the two-dimensional spectrum of global plant trait variation. Nat. Ecol. Evol. 6, 36–50 (2022).Article 
PubMed 

Google Scholar 
Díaz, S. et al. The global spectrum of plant form and function. Nature 529, 167–171 (2016).Article 
PubMed 

Google Scholar 
Ploton, P. et al. Spatial validation reveals poor predictive performance of large-scale ecological mapping models. Nat. Commun. 11, 4540 (2020).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Meyer, H. & Pebesma, E. Spatial validation reveals poor predictive performance of large-scale ecological mapping models. Methods Ecol. Evol. 12, 1620–1633 (2021).Article 

Google Scholar 
Schrodt, F. et al. Bhpmf—a hierarchical Bayesian approach to gap filling and trait prediction for macroecology and functional biogeography. Glob. Ecol. Biogeogr. 24, 1510–1521 (2015).Article 

Google Scholar 
Kuppler, J. et al. Global gradients in intraspecific variation in vegetative and floral traits are partially associated with climate and species richness. Glob. Ecol. Biogeogr. 29, 992–1007 (2020).Article 

Google Scholar 
Scheiter, S., Langan, L. & Higgins, S. I. Next-generation dynamic global vegetation models: learning from community ecology. New Phytol. 198, 957–969 (2013).Article 
PubMed 

Google Scholar 
Taubert, F. et al. Confronting an individual-based simulation model with empirical community patterns of grasslands. PLoS ONE 15, e0236546 (2020).Roger, E. & Klistorner, S. (2016) Bioblitzes help science communicators engage local communities in environmental research. J. Sci. Commun. https://doi.org/10.22323/2.15030206 (2016).Legendre, P. & Legendre, L. Numerical Ecology 3rd edn (Elsevier, 2012).Warton, D. I. et al. Smatr 3—an R package for estimation and inference about allometric lines. Methods Ecol Evol 3, 257–259 (2012).Article 

Google Scholar 
Wolf, S. et al. iNaturalist_traits: iNaturalist trait maps version 1 (January 5, 2022) Zenodo https://doi.org/10.5281/zenodo.6671891 (2022). More