Field location and soil sampling
The soil used in this experiment was collected from an agricultural field cultivated with maize at the “Instituto Tecnológico Superior del Oriente del Estado de Hidalgo” (ITESA) located in Apan, State of Hidalgo, Mexico (19° 73′ N, 98° 46′ W). The 0–20 cm top soil layer of three 400 m2 plots was sampled 20 times. The soil from each plot was pooled separately so that three soil samples (n = 3) were obtained. This field based replication was maintained in the greenhouse experiment so as to avoid pseudo-replication. The soil samples were passed separately through a 5 mm sieve and characterized.
The soil is classified as a Phaeozem according to “World Reference Soil (WRS) system”, with pH 6.6, electrolytic conductivity (EC) 0.22 dS m−1 and water holding capacity (WHC) 515 g kg−1. The sandy clay loam soil with clay content 240 g kg−1, sand content 530 g kg−1 and silt content 230 g kg−1, had an ammonium content 8.16 mg kg−1 dry soil, nitrate 1.91 mg kg−1 dry soil and nitrite 0.01 mg kg−1 dry soil. The maize seeds were the hybrid variety 215 W obtained from Eagle® Sinaloa (Mexico).
Characteristics of the biofertilizer
Although a biofertilizer can be described in different ways we use the definition as given by38. Vessey defined (2003) a biofertilizer as “a substance which contains living micro-organisms which, when applied to seeds, plant surfaces, or soil, colonize the rhizosphere or the interior of the plant and promotes growth by increasing the supply or availability of primary nutrients to the host plant”. As the consortium used in this study fits the definition of a biofertilizer as given by Vessey38 we will refer to the consortium as the biofertilizer or when sterilized to the sterilized biofertilizer throughout the manuscript.
The “biofertilizer” used in this study was a mixture of bacteria and leachate from compost of cow manure and was obtained from a local farmer in Hidalgo (Mexico) and characterized chemically and microbiologically. The cow manure was composted on a cement floor with a small inclination so that leachate could be collected easily. The farmer adds the leachate to the mixture of the bacteria to guarantee their survival and as an additional plant nutrient source. The farmer applies this solution regularly to fertilize his fields cultivated with maize. A same application protocol and procedure was used in this study to mimic the field experiment. Half of the biofertilizer obtained from the local farmer was sterilized by autoclaving at 121 °C for 20 min on three consecutive days so as to determine the effect of the microorganisms in the biofertilizer on the maize plants and the bacterial community structure, and the effect of the nutrients added with the biofertilizer.
Experimental design and a greenhouse experiment
The research was done in a greenhouse at Cinvestav-Zacatenco situated to the north of Mexico City (Mexico). The experiment used a completely randomized block design with six treatments. The treatments combined as a first factor soil cultivated with maize or left uncultivated. A second factor included soil amended with the biofertilizer, sterilized biofertilizer or not fertilized. The daily temperature in the greenhouse ranged from 15 °C as minimum and reached a maximum 35 °C from April to August of 2017.
As the experimental protocol was complex, a diagram of the different treatments and sampling is given in Supplementary Fig. S11 online. A total of 162 PVC columns with diameter 17 cm and height 60 cm were used in the experiment. Each pot was filled at the bottom with 0.5 kg tezontle, a highly porous volcanic rock, and 10 kg soil was added on top. The 162 columns included 6 treatments (uncultivated unamended soil, uncultivated soil amended with biofertilizer, uncultivated soil amended with sterile biofertilizer, maize cultivated unamended soil, maize cultivated soil amended with biofertilizer, maize cultivated soil amended with sterile biofertilizer; n = 6), 3 sampling times (day 44, day 89 and day 130; n = 3), three different soil samples (n = 3), with three columns planted with a maize plant per soil sample (n = 3). Three columns of each soil sample were planted with a maize plant to account for plants that might die so that at least one mature plant was obtained per treatment, sampling time and soil sample. The soil in the 162 PVC columns was adjusted to 40% WHC with distilled water and conditioned in the greenhouse for a week. Additionally, three PVC columns were filled with soil from each soil sample (n = 3), adjusted to 40% WHC with distilled water and conditioned for a week. These three soil samples were used to extract DNA as described below and defined the bacterial community at the onset of the experiment, i.e. time 0.
Maize seeds variety 215 W Eagle hybrid seeds® were obtained from the farmer that provided us with the biofertilizer. Three washed maize seeds were planted at 3 cm depth in 81 columns, while the remaining columns were left uncultivated. Seven days after emergence, the most vigorous plantlet was kept and the other two discarded. After 44 days, the biofertilizer or the sterilized biofertilizer was diluted with water and applied with an atomizer (10 ml m−2 or similar to 100 l applied ha−1 by the farmer) so that it was added as fine spray evenly on soil of each pot when the seeds were planted. A similar volume of water was applied in the same way to the unfertilized treatment. Five more applications of the biofertilizer, sterilized biofertilizer or water by aspersion were done during the cultivation of the maize plants. As such, the uncultivated or maize plant cultivated soil was applied with the biofertilizer, sterile biofertilizer or water on 13th April, 28th May, 5th June, 13th July, 2nd August and 12th August 2017.
Soil and plant sampling
After 44 (27th May 2017), 89 (11th July 2017) and 130 days (21st August 2017), three columns from each treatment (n = 6) and soil sample (n = 3) were selected at random. Soil was removed from each column. The cultivated and uncultivated soil was sampled, characterized, and extracted for DNA as described below. The non-rhizosphere soil was separated from the rhizosphere soil by shaken the plants gently. The soil adhered to the roots was considered the rhizosphere soil. A 20 g sub-sample of the uncultivated, non-rhizosphere and rhizosphere soil was stored at − 20 °C pending extraction of DNA, while the pH and mineral N was determined in the remaining soil. Roots and shoots were separated, weighted and their length measured. The roots and shoots were dried in an oven at 60 °C for 24 h and weighed.
Soil physicochemical characterization
The moisture content of the soil was determined by weight loss after samples were dried at 60 °C in an oven for 24 h. The WHC was determined by saturating 50 g dry soil with distilled water, left to drain overnight and measuring the amount of water retained. The EC was measured in a soil paste (200 g soil/110 ml distilled H2O) with an HI 2300 microprocessor (HANNA Instruments, Woonsocket, RI, USA), while the particle size distribution was determined with the hydrometer method as described by Gee and Bauder39. The pH was determined in a 10 g soil–25 ml distilled water mixture with a calibrated pH meter (Denver Instrument, Bohemia, NY, USA) fitted with a glass electrode (3007281 pH/ATC Termofisher Scientific, Waltham, MA, USA).
Mineral nitrogen (NO3−, NO2− and NH4+) was measured in the soil and biofertilizer. A 20 g soil sub-sample was extracted with 100 ml 0.5 M K2SO4 and filtered through Whatman filter paper® while mineral N was measured with a SKALAR automatic analyser system (Breda, the Netherlands)40. A 20 g biofertilizer sub-sample was mixed with 80 ml 0.5 M K2SO4, filtered through Whatman filter paper® and mineral N measured as described previously.
DNA extraction and PCR amplification
A 5 ml sub-sample of the sterilized and unsterilized biofertilizer was centrifuged at 3500 rpm for 15 min and the supernatant removed. A 0.5 g sub-sample of soil was washed with 10 ml 0.15 mol l−1 sodium pyrophosphate to eliminate the humic and fulvic acids, centrifuged at 3500 rpm for 15 min and this process was repeated until the supernatant was clear41. The excess pyrophosphate was eliminated with 10 ml 0.15 mol l−1 phosphate buffer pH 8. Three different methods were used to extract DNA from the soil and the sterilized and unsterilized biofertilizer samples. The first technique was based on the method described by Green and Sambrook42. In the second method, cells were lysed with two lysis solutions and a thermal shock as described by Valenzuela-Encinas et al.43. The third method consisted of a mechanical disruption and detergent solution for cell lysis44. Each method was used to extract three times 0.5 g soil or 5 ml sterilized and unsterilized biofertilizer (a total of 1.5 g soil or 15 ml sterilized and unsterilized biofertilizer). The extracts from the soil and sterile or unsterilized biofertilizer were pooled separately.
The 16S rRNA gene (V3–V4 region of bacteria) was amplified using the primers 341F (5′-CCTACGGGNGGCWGCAG-3′) and 805R (5′-ACHVGGGTATCTAATCC-3′45. The PCR conditions were 94 °C for 5 min, followed by 25 cycles of 60 s at 94 °C, 45 s at 53 °C, and 60 s at 72 °C, with a final extension of 10 min at 72 °C. The PCR was repeated three times for each sample. After PCR amplification, the obtained products were cleaned using the FastGen Gel/PCR extraction Kit (Nippon Genetics Duren, Germany) and quantified using a Nanodrop 3300 fluorospectrometer (TermoFisher, Wilmington, DE, USA) with PicoGreen dsDNA. The samples were mixed in equimolar amounts and sequenced using MiSeq 300-pb paired-end runs (Illumina, CA, USA) at Macrogen Inc. (Seoul, Korea).
16S rDNA sequences analysis
The raw sequences were analysed with “Quantitative insights into microbial ecology pipeline” (QIIME) software (version 1.9.1)46. The barcode reads were demultiplexing removed from the sequences using the script extract_barcodes.py. The chimeric sequences were identified using “identify_chimeric_seqs.py” with the usearch61 method and removed47. The taxonomic assignment was done using the Ribosomal Data Project (rdp)48, against the Greengenes 16S rRNA database with a 0.8 confidence49. The sequences were clustered as operational taxonomic units (OTU) at 97% similarity level with the UCLUST algorithm47. Sequences were aligned against the Greengenes reference database using PyNAST version 1.2.250. The obtained 16S dataset was filtered, all OTUs assigned to Archaea were discarded and the dataset normalized. Alpha diversity indices (Chao1, Shannon and Simpson) were calculated from 478000 rarefied sequences with QIIME.
Statistical analysis
All statistical analyses were done in R (R 4.0.2 GUI 1.72 Catalina build51). The characteristics of the maize plants (n = 3) obtained per plot (n = 3) were averaged and the sequences obtained from the replicate rhizosphere or non-rhizosphere soil were summed (n = 3) per plot before the statistical analysis. A non-parametric test was used to determine the effect of biofertilizer application and time on the plant and soil characteristics with the non-parametric t1way test of the WRS2 package (A collection of robust statistical methods)52. A non-parametric test was used to determine the effect of biofertilizer application or cultivation of maize on the bacterial alpha diversity with the non-parametric t1way test of the WRS2 package52. Heatmaps of the relative abundances of the bacterial groups were constructed with the pheatmap package53. Ordination [principal component analysis (PCA)], multivariate comparison (perMANOVA) and differential abundance (ALDEx2) was done with converted sequence data using the centred log-ratio transform test returned by the aldex.clr argument (ALDEx2 package54). The PCA was done with the vegan package55. Effect of biofertilizer application and cultivation of maize on the bacterial groups was determined using a compositional approach, i.e. analysis of differential abundance taking sample variation into account (aldex.kw argument, ALDEx2 package). A permutational multivariate analysis of variance (perMANOVA) analysis was also done with sequence counts converted using the centred log-ratio transform, i.e. aldex.clr argument (ALDEx2 package (aldex.clr(counts, mc.samples = 128, denom = ”all”, verbose = FALSE, useMC = FALSE)). The adonis2 argument (Vegan package) was used for the perMANOVA analysis to test the effect of cultivation of maize, time and its interaction, biofertilizer application, time and their interaction, and cultivation of maize, biofertilizer application and their interaction on the bacterial community structure (#adonis2(clrcounts ~ maize*biofertilizer, data = code, permutations = 999, method = ”euclidean”). Raw counts were used as input and Monte Carlo Dirichlet instances of the clr transformation values were generated with the function ‘aldex.clr’ of ALDEx2 (v.1.23.2) R package54. Distance pairwise matrices were calculated using the Aitchison distance and the principal coordinate analysis (PCoA) was calculated on the distance matrices with vegan R package55.
Informed consent
Permission was obtained from the farmer to use the maize seeds he provided.
Ethical approval
The experiment in the greenhouse complied with and was conducted as stipulated by national regulations.
Source: Ecology - nature.com
