TlAN Xue-liang,LlU Jia-jia,LlU Quan-chengXlA Xin-yao,PENG Yong,Alejandra l.HUERTA,YAN Jian-bing,Ll Hui,LlU Wen-de

1 State Key Laboratory for Biology of Plant Diseases and Insect Pests,Institute of Plant Protection,Chinese Academy of Agricultural Sciences,Beijing 100193,P.R.China

2 Henan Engineering Research Center of Biological Pesticide &Fertilizer Development and Synergistic Application,Henan Institute of Science and Technology,Xinxiang 453003,P.R.China

3 School of Biological Science and Technology,University of Jinan,Jinan 250022,P.R.China

4 National Key Laboratory of Crop Genetic Improvement,Huazhong Agricultural University,Wuhan 430070,P.R.China

5 Department of Entomology and Plant Pathology,North Carolina State University,Raleigh 27606,USA

Abstract The diversity of prokaryotic communities in soil is shaped by both biotic and abiotic factors. However,little is known about the major factors shaping soil prokaryotic communities at a large scale in agroecosystems. To this end,we undertook a study to investigate the impact of maize production cropping systems,soil properties and geographic location (latitude and longitude) on soil prokaryotic communities using metagenomic techniques,across four distinct maize production regions in China. Across all study sites,the dominant prokaryotes in soil were Alphaproteobacteria,Gammaproteobacteria,Betaproteobacteria,Gemmatimonadetes,Acidobacteria,and Actinobacteria. Non-metric multidimensional scaling revealed that prokaryotic communities clustered into the respective maize cropping systems in which they resided. Redundancy analysis (RDA) showed that soil properties especially pH,geographic location and cropping system jointly determined the diversity of the prokaryotic communities. The functional genes of soil prokaryotes from these samples were chiefly influenced by latitude,soil pH and cropping system,as revealed by RDA analysis. The abundance of genes in some metabolic pathways,such as genes involved in microbe-microbe interactions,degradation of aromatic compounds,carbon fixation pathways in prokaryotes and microbial metabolism were markedly different across the four maize production regions. Our study indicated that the combination of soil pH,cropping system andgeographic location significantly influenced the prokaryotic community and the functional genes of these microbes. This work contributes to a deeper understanding of the composition and function of the soil prokaryotic community across large-scale production systems such as maize.

Keywords:metagenome,cropping system,maize,soil prokaryotes

1.lntroduction

Soil prokaryotes have an important ecological role in nature,they recycle nutrients and sustain plant growth.In agricultural ecosystems they can increase nutrient availability to crops and suppress soil-borne diseases (Mülleret al.2016). Agricultural production systems are more homogeneous at large spatial scales than natural ecosystems. This is largely due to lower plant genetic diversity especially in monocropping systems and frequent human intervention,for example through fertilization and agrochemical treatments (Kennedy and Smith 1995). In these homogeneous ecosystems,crops interact with multiple species of metabolically diverse microbes to form complex and dynamic interactive networks (Berendsenet al.2012;Busbyet al.2017;de Vries and Wallenstein 2017).

The prokaryotes in soils of agricultural ecosystems are influenced by agricultural practices,such as farm operation,crop type,pesticide application,fertilization programs,plant host genotype,and natural factors including physical and chemical properties of soil,and other prokaryotic and eukaryotic organisms in the soil (Pennanenet al.1999;Oline 2006;Joneset al.2009;Jacobsen and Hjelmsø 2014;McDanielet al.2014;Dinget al.2018;Daset al.2019;Huanget al.2019). These factors contribute to a feedback loop where they influence soil prokaryote community composition which in turn interacts with the plant host. The plant host influences the soil by modifying the soil physicochemical characteristics through litter input (Wanget al.2016),and the excretion of exudates (Acosta-Martínezet al.2008;Haicharet al.2008;DeAngeliset al.2009;Bakkeret al.2012). Two studies on structure and function of the soil microbiome at a global scale have provided a deep understanding of the multiple factors influencing soil bacterial communities (Bahramet al.2018;Manuelet al.2018),but these studies mainly focus on soil microbes from natural ecosystems rather than from agricultural ecosystems. It is well known that soil microbes play an important role in maintaining plant health and productivity (Chenet al.2020;Shiet al.2020). Understanding the biotic and abiotic factors that shape the soil prokaryotic community at large agricultural scales will improve our ability to develop synthetic communities that maximize prokaryotic community functional capabilities in agricultural ecosystems,such as in maize.

Maize (ZeamaysL.subsp.mays) is a globally important crop used for human food,animal feed,and industrial products. It is planted in the tropics,temperate and frigid zones of the world and thrives under humid,semi-humid,and arid conditions. The maize plant secretes a wide variety of amino acids,sugars,and organic acids,all of which modify the soil (Vílchezet al.2000). In China,more than 35 445 million hectares are in maize production and the total yield in 2017 was estimated at 21 589 million tons (NBSC 2017). Maize was introduced to China in the 14th century and has since been planted across a large range of geographical locations and under different cropping systems (Li 2009). To date,the commonalities and differences between prokaryotic communities and their functional potential in soil under different maize cropping systems across large spatial scales in China remain unexplored. Considering that maize has been planted in China for centuries under different cropping systems,we hypothesized that the prokaryotic communities and functional genes in maize soil are shaped by a combination of soil properties,cropping systems,and geographic location. Here,we present a metagenomic analysis on the soil prokaryotic communities under four distinct maize cropping systems in China,including continuous cropping of maize,and wheat-maize,oil seed-maize and potato-maize rotations.

2.Materials and methods

2.1.Soil sample collection

Soil samples were collected from 22 locations across 20 provinces in China that span four maize production regions with different cropping systems (Appendix A). Information about the sampling locations and soil chemical properties are listed in Appendix B. The production regions and cropping systems sampled were the North China spring maize region under continuous maize cropping (M),the Huanghuai Plain summer maize region under a continuous wheat-maize rotation (WM);the southern China hill maize region under an oil seedmaize rotation (OM);and the Northwest China mountain maize and Southwest China maize regions both under continuous potato-maize rotations (PM). For each maize production region,the representative maize fields were chosen for sample collection. Five soil samples were taken randomly across the selected field and pooled together,giving one representative sample for each location. Soil samples were collected after removing the first 5-cm of top soil and taking a soil core 5-15 cm deep in locations without roots. Each composite soil sample was divided into two sub-samples:one for molecular analyses,the second for chemical analyses. The soil pH,and soil organic matter (SOM),total nitrogen (N),total phosphorus (P) and total potassium (K) concentrations were determined using standard laboratory methods (Analysis and Test Center,Institute of Environment and Sustainable Development in Agriculture,Chinese Academy of Agricultural Sciences) as follows. The soil pH was measured using a pH-meter (UB-10;DENVER Instrument,Denver,USA) with a soil-liquid ratio of 1:2.5;SOM was determined by potassium dichromate (K2Cr2O7) oxidation followed by ammonium ferrous sulfate titration;and N,P and K concentrations were determined by semimicro Kjeldahl distillation method,sodium hydroxide melting-molybdenum sulfate anti-colorimetric method,and sodium hydroxide melting-flame photometry method,respectively. The differences in soil properties among the cropping systems were determined using one-way analysis of variance with SPSS Software.

2.2.DNA extraction and metagenomic sequencing

Total DNA was extracted from 10 g of soil per sample using a Power Soil DNA Kit (MOBIO,USA). Purified DNA was stored at -20°C until further analysis. DNA concentrations in extracts were quantified with a NanoDrop 2000 spectrophotometer (Thermo Scientific,USA),and extracts were diluted to a final concentration of approximately 25 ng μL-1DNA with distilled water and stored at -20°C. Each composite DNA sample was sheared into fragments of approximately 300 bp using an M220 Focused-ultrasonicator (Covaris Inc.,Woburn,MA,USA) to build a metagenomic DNA library according to the manufacturer’s recommendations (Illumina,San Diego,USA). Samples were paired-end sequenced on the Illumina HiSeq2500 platform by using the PE150 mode (Illumina,San Diego,USA).

2.3.Gene prediction,functional annotation and taxonomic assignment

Raw data were first processed with Trimmomatic Software,which removed and filtered reads with ambiguous N bases,sequencing adapters,and reads with average base quality score of less than 15 (Bolgeret al.2014).The BWA Software (http://bio-bwa.sourceforge.net/) was used to filter host plant DNA contamination. After quality control,we used the SOAPdenovo2 (v2.0) Software to performdenovometagenome (scaffold) assembly with valid reads (Luoet al.2015). Statistical tests were carried out on scaffolds with a length of 500 bp or more. Each scaffold was broken down at gap regions to extract new contig sequences. Only final contigs with lengths greater than 500 bp were retained and used for the gene predictions and annotations (Luet al.2018).

Open reading frames (ORFs) were predicted and only those with lengths over 100 bp were retained. The ORFs were translated into amino acid sequences with GeneMark.hmm (v3.26) in GeneMark (http://exon.gatech.edu/meta_gmhmmp.cgi) using a Hidden Markov Model (HMM) (Besemer and Borodovsky 2005). Subsequently,a non-redundant gene catalog was constructed using CDHIT (v4.5.4) Software with a 95% identity parameter and a 90% coverage parameter (Fuet al.2012). The longest gene was selected as a representative sequence of each gene set. The valid reads from each sample were aligned against the non-redundant gene catalog (95% identity) by Bowtie2 (Langdon 2015) to estimate the abundance of genes in each sample. The representative amino acid sequence for each gene was annotated using the NCBI Non-Redundant (NR) Proteins database with a cutoff e-value of 10-5. Taxonomic assignment of soil microbes in our samples was performed by aligning gene sequences in the NR Library,and the best hits to the NCBI microbial taxonomy database were selected.

2.4.Alpha-and beta-diversity of soil prokaryotic communities

The abundance of prokaryotes in each sample was calculated using the corresponding abundance of predicted genes. Next,we constructed an abundance profile of Domain,Kingdom,Phylum,Class,Order,Family,Genus,and Species,and abundance statistics were calculated at each taxonomic level. To quantify the alpha-diversity of prokaryotic communities in each soil sample,Chao1 and Shannon indicies were calculated using the R package vegan (Oksanenet al.2012). The beta-diversities of the prokaryotic communities under different cropping systems were determined using non-metric multidimensional scaling (NMDS) analysis and Adonis analysis in the R program (https://www.r-project.org/). Networks of the Pearson correlation coefficients between cropping systems and soil properties were calculated and plotted. Each factor was designated as a node and the correlation coefficients between two factors was represented by the line,which plotted by Cytoscape (https://cytoscape.org/). To detect the effects of soil properties,geographic location (latitude and longitude) and cropping system on the soil prokaryotic community we conducted RDA using the R package vegan (Oksanenet al.2012).

2.5.ldentification of dominant and indicator taxa under four cropping systems

We defined the dominant taxa across soil samples under different cropping systems according to the following criteria:1) the dominant taxa were found in all of the samples;and 2) only the top 10% of the most abundant taxa across all samples were retained (Manuelet al.2018). We then calculated the percentage or relative abundance of these dominant taxa compared to total taxa at the genus level,as well as the relative abundance of dominant genera under each cropping system and within individual sampling sites. Identification of indicator taxa among the four cropping systems was performed using the linear discriminant analysis effect size (LEfSe) algorithm (Shayanet al.2015),by implementing a Kruskal-Wallis test (P≤0.05) followed by linear discriminant analysis (LDA=3.0,P≤0.05) to screen significantly enriched indicator taxa in different cropping systems.

2.6.Functional gene categories of soil prokaryotic communities

To reveal the functions of soil prokaryotic communities,a search was conducted for each representative amino acid sequence for each gene set using BLAST and comparing sequences against the Kyoto Encyclopedia of Genes and Genomes database (Kyoto Encyclopedia of Genes and Genomes) with a set cutoff e-value of 10-5. We used a principal component analysis (PCA) to identify the functional genes of soil prokaryotes using the R program (https://www.r-project.org/). The effect of soil properties,latitude,longitude,and cropping systems on functional genes was determined by RDA using the R package vegan (Oksanenet al.2012). Significant differences in abundance of functional gene between cropping systems were determined with one-way ANOVA statistical analysis using the STAMP program (Parkset al.2013).

3.Results

3.1.Taxonomic features of the prokaryotic communities

Over the 22 locations,soil pH varied from 5.2 in cropping system M to 8.8 in cropping system PM with an average soil pH of 7.1. Soil pH from the M and OM systems were markedly lower than that from the PM and WM systems (Appendix B).

The total reads of the raw data were 225 Gb and the average sample read was 10.2 Gb. Sequencing data were uploaded and are available in the NCBI sequence read archive (SRA). The BioSample IDs are SAMN12406927-SAMN12406948 and the BioProject ID is PRJNA557509. After filtering,the metagenomic DNA library consisted of approximately 34 364 851 clean,highquality sequence reads per soil sample and the total length of reads in each sample averaged 5 098.32 Mb (Appendix C). The average number of assembled scaffolds in the metadata was 904 288 and there were a total of 1 253 426 ORFs identified which ranged between 8 666 and 219 494 per soil sample. Among the unique genes annotated in our study,an average of 987 525 came from bacteria (relative abundance of 95.96%); 40 128 were from archaea (relative abundance 3.90%),1 418 were from virus (relative abundance 0.14%),and 8.91 were from fungi (relative abundance 0.0007%).

Of the prokaryotic community in the soil samples,Alphaproteobacteria,Gammaproteobacteria,and Betaproteobacteria were the dominated classes,followed by Gemmatimonadetes,Acidobacteria,and Actinobacteria (Fig.1-A). The relative abundance of the classes varied under the different cropping systems. The Shannon index and the Chao1 index of the prokaryotic community under the M and OM cropping systems were higher than that under the WM and PM systems (Fig.1-B and C),showing that the soil prokaryotic community under the M and OM systems had higher species richness and diversity than the WM and PM systems.

Fig.1 Relatice abundance (A),Shannon index (B) and Chao1 index (C) of soil prokaryotic communities across four maize cropping systems. M,continuous maize cropping;OM,oil seed-maize rotation;PM,potato-maize rotation;WM,wheat-maize rotation. No rank class contains classes not in the top 10.

3.2.Beta-diversity of the prokaryotic communities

The NMDS analysis revealed that the prokaryotic communities were classified into four groups consistent with the four cropping systems (Fig.2-A). The communities under the M and OM systems clustered to the left of zero (NMDS1＜0),whereas the communities under the WM and PM systems mostly clustered to the right (NMDS1＞0). Moreover,the Adonis test based on the Bray-Curtis betadiversity analysis showed that the communities were significantly different (P＜0.01) across the four cropping systems. The communities under the same cropping system were more similar to each other forming distinct clusters in the ordination space. This finding was also supported by the correlation heatmap that showed the communities from one group were generally located in one big branch of the phylogenetic tree (Fig.2-B).

3.3.lnfluence of soil properties,geographic location and cropping system on the prokaryotic communities

We found that soil properties differed among the four cropping systems,and we therefore analyzed the relationships between cropping systems and soil properties (pH,SOM,N,P and K). Cropping system had a positive correlation with soil pH (r=0.71) and negative correlation with N,P and K concentrations (Fig.2-C). The RDA demonstrated that pH had the greatest influence on the composition of prokaryotic communities (Fig.2-D),followed by geographic location (longitude and latitude),then cropping system and finally soil N concentration. Other soil properties like P,K and SOM levels were less important regulators of soil communities and only explained small amount of variation in community composition. The above results indicated that the combined effect of soilpH,cropping system and geographic location played an important role in shaping the communities.

Fig.2 The influence of soil properties,geographic location and cropping systems on the soil prokaryotic community. A,non-metric multidimensional scaling (NMDS) analysis of soil prokaryotic communities under four cropping systems. B,correlation heatmap denotes similarity of the communities under one cropping system. C,network of correlation between cropping systems and soil properties. Line thickness represents magnitude of correlation coefficients. D,redundancy analysis of the effect of soil properties,latitude,longitude and cropping system on the communities. M,continuous maize cropping;OM,oil seed-maize rotation cropping system;PM,potato-maize rotation;WM,wheat-maize rotation;SOM,soil organic matter.

3.4.ldentification of dominant and indicator taxa

Among all 1497 genera identified in our soil samples (Appendix D),only 748 were prevalent in all samples (Fig.3-A). Moreover,we identified 141 out of 748 genera as the most abundant taxa across all 22 soil samples (Fig.3-B). These 141 abundant genera accounted for 64% of the relative abundance of all samples (Fig.3-B). The abundant genera were classified intoSphingomonas,Rhodanobacter,Bradyrhizobium,Pseudolabrys,Candidatus_Koribacter,Candidatus_Solibacter,Pyrinomonas,Acidobacterium,GemmatirosaandGemmatinonas(Fig.3-C). In addition,we identified the indicator taxa from each cropping system using the LEfSe method. At the genus level,Sphingomonasand Candidatus_Koribacterwere indicators under the M cropping system,Candidatus_Nitrosotaleawas an indicator under the OM system,Nitrososphaeraand Candidatus_Nitrosocosmicuswere indicators under the PM system,andPyrinomonas,NitrospiraandLysobacterwere indicators under the WM system (Fig.4-A and B).

Fig.3 Dominant taxa of prokaryotic communities under four maize cropping systems. A,the prevalent and unique genera across all soil samples. The central circle shows the number of genera common to all samples under the four cropping systems,while each petal shows the number of genera unique to each soil sample. B,percentage of the number of genera and relative abundance of 141 taxa. C,phylum composition and relative abundances of the most dominant genera of bacteria. M,continuous maize cropping;OM,oil seed-maize rotation;PM,potato-maize rotation;WM,wheat-maize rotation.

Fig.4 Indicator taxa across cropping systems revealed using the linear discriminant analysis effect size (LEfSe) method. A,cladogram representation of the phylogenetic relationships of indicator taxa in each cropping system. The circles from inner to outer stand for phylum,class,order,family and genus. B,LEfSe analysis (P＜0.05,logarithmic LDA score threshold≥3.0) of the four cropping systems.M,continuous maize cropping;OM,oil seed-maize rotation;PM,potato-maize rotation;WM,wheat-maize rotation.

3.5.Functional potential of soil prokaryotes

Functional annotations for approximately 251 604 unigenes were achieved by using BLAST and the KEGG database. A total of 15 610 KOs were assigned to the annotated unigenes. The KEGG classification of functional groups includes“metabolic pathways,biosynthesis of secondary metabolites,microbial metabolism in diverse environments,biosynthesis of antibiotics,carbon metabolism,biosynthesis of amino acids,two-component systems,purine metabolism,ABC transporters,and pyruvate metabolism”(Appendix E).The distribution of functional genes of soil prokaryotes across the four cropping systems was revealed by PCA and the results showed clusters of M and OM mainly in the right lower quadrant,while PM clustered primarily at the upper left and WM was mainly clustered at the lower left (Fig.5-A). RDA showed that latitude primarily shaped the distribution of functional genes of the soil prokaryotes,followed by soil pH,cropping system and soil N concentration (Fig.5-B). Longitude and other soil properties (P,K and SOM) explained only small variations in the distribution of functional genes.

We found that the genes in some pathways were different in abundance under different cropping systems (Fig.5-C). For example,genes involved in microbemicrobe interactions,such as ‘polyketide sugar unit biosynthesis’ (P＜0.01),‘terpenoid backbone biosynthesis’ (P＜0.01),‘biosynthesis of antibiotics’ (P=0.012) and ‘limonene and pinene degradation’ (P=0.043) under the OM cropping system and ‘biosynthesis of type II polyketide backbone’ (P=0.048) under the M cropping system,had higher abundance than under the PM and WM systems. The WM system possessed a higher proportion of genes related to ‘degradation of aromatic compounds’ (P=0.011) than other cropping systems.The abundance of genes belonged to ‘carbon fixation pathways in prokaryotes’ (P＜0.01) and ‘microbial metabolism in diverse environments’ (P＜0.01) were higher under PM than under other cropping systems.

Fig.5 The distribution and effect factors of functional genes and signigicantly different metabolic pathways of soil prokaryotes. A,principle component analysis of the functional genes under four maize cropping systems. B,redundancy analysis of distribution of the functional genes by soil properties,latitude,longitude and maize cropping system. C,metabolic pathways of soil prokaryotes under four maize cropping systems. M,continuous maize cropping;OM,oil seed-maize rotation;PM,potato-maize rotation;WM,wheat-maize rotation.

4.Discussion

4.1.Effect of soil properties,geographic location and cropping system on prokaryotic communities

In our study,we found that the alpha-diversity of prokaryotes in the different cropping systems were significantly different,as was the beta-diversity across cropping systems,suggesting that cropping systems contribute to shaping the composition of prokaryote communities in this study. In general,in one cropping system,similar cropping and planting practices,which influence the alpha-diversity of communities,are used (Hartmannet al.2015). The M cropping system is characterized by low-input farming practices,such as low fertilizer and pesticide application,and soils possess high organic matter content (Duanet al.2016;Puet al.2019). These factors promote prokaryotic abundance and diversity,leading to higher Shannon and Chao1 indicies of prokaryotic communities under the M cropping system. The OM system characteristically undergoes less tillage and other agricultural operations,including low fertilizer application,than other rotation systems (Chaiet al.2014;Donget al.2018),which also results in large populations of prokaryotes. In contrast,the WM and PM cropping systems are intensively managed at planting,and exposed to high chemical fertilizer and pesticide application rates (Maet al.2017;Umairet al.2017),which disturbs the soil prokaryotic communities and decreases their diversity.

Soil samples analyzed in this study were collected from 22 distinct geographic locations,across different climate zones and under different cropping systems,leading us to hypothesize that the soil environment and geographic location would result in different microbial communities.The RDA analysis of cropping system,geographic location and soil chemical properties showed that soil pH was the primary factor driving differences in species diversity between sites. This is consistent with previous findings that soil pH determines microbial community diversity at larger spatial scales (Lombardet al.2011;Fiereret al.2012;Manuelet al.2018). Geographic location has been recognized as an important driver of soil geochemistry,and thus fundamentally impacts microbial communities in soils (Bissettet al.2010;Reithet al.2012;Wanget al.2015). In our study we confirmed the importance of geographic location as it was ranked as the second most important factor shaping the community structure.

Cropping system was the third most important factor,among those analyzed,that influenced prokaryote community composition,which aligned with previous studies showing that long-term cropping systems can partly shape the soil microbiome (Hartmanet al.2018).Although the cropping system effects were not as significant as soil pH and geographic location,the soil prokaryotic communities showed a ‘cropping system’ pattern,indicative of a correlation between cropping system and soil pH (Fig.2-C). Two previous studies found similar results,where soil pH was influenced by cropping system (Xu and Coventry 2003;Anandaet al.2006;Sofiet al.2016). It is important to note that long-term cropping practices may change soil physicochemical properties (Sofiet al.2016). We found that soil chemical properties,geographic location,and cropping system formed a tripartite interaction which impacted on the prokaryotic communities in our study sites. This point is consistent with the report that contemporary disturbances (soil N,P,K and organic manure) and historical contingencies (distinct geographic sampling locations) together affect the soil bacterial community (Geet al.2008).

4.2.Dominant and indicator prokaryotes in different cropping systems

The dominant generaStenotrophomonas,Herbaspirillum,Pseudomonas,andChryseobacteriumin our study were also observed by Niuet al.(2017),suggesting that some dominant taxa are globally distributed.Koribacterwas indicator species under the M cropping system with low pH soil,reflecting that they are suited to an acidic environment (Wanget al.2018).Sphingomonaswas an indicator under the M system,probably due to high SOM providing more carbon sources,which is consistent with the results of Murakamiet al.(2010). CandidatusNitrosotaleais adapted to nitrifying acid soil (Lehtovirta-Morleyet al.2011) and was specific to the OM system possibly due to high N levels and low pH in OM soil.Nitrososphaeraand CandidatusNitrosocosmicushave a strong tolerance to high ammonia concentrations and have wide environmental adaptation (Kerouet al.2016;Lehtovirta-Morleyet al.2016),which would benefit them in adapting to high soil N levels and dry conditions under the PM cropping system.Pyrinomonasis a moderately acidophilic bacterium (Leeet al.2015),and was an indicator in mildly alkaline soil in the WM system.Nitrospirais a common nitrite-oxidizing bacteria in alkaline soil (Zhanget al.2017),which explains why it was an indicator species in the mildly alkaline WM soil.Lysobacterwas indicator species in the WM cropping system,which was probably related to abundantFusariumpathogens of wheat in this system,asLysobacteract as biocontrol bacteria against wheatFusariumhead blight (Zhaoet al.2019),

4.3.Effect of soil properties,cropping system and geographic location on the functional genes of soil prokaryotes

A latitudinal diversity gradient has been observed for many plant and animal species,and is also documented in microbial global distribution patterns (Green and Bohannan 2006). We found that the functional genes of soil prokaryotes were primarily regulated by latitude,which agrees with studies that latitude as well as temperature influence the kinetics of biological processes such as reproduction,dispersal,and species interactions (Rohde 1992;Allenet al.2002). Moreover,soil bacteria natural product biosynthetic domain genes most consistently correlate to latitude (Lemetreet al.2017). However,global topsoil bacterial genes have been found to be more strongly correlated with mean annual precipitation than latitude (Bahramet al.2018). This discrimination may be explained by the fact that the topsoil in that study was obtained from natural ecosystems,whereas our soils were collected from agricultural ecosystems with manual irrigation,thus mean annual precipitation was no longer a major factor shaping the functional genes of the soil prokaryotes. Soil pH also drives the functional genes of soil prokaryotes,which is consistent with the result that pH has been shown to play an important and similar role in shaping the composition and functional profile of soil bacteria (Kaiseret al.2016). Here,we also show that cropping systems play important roles in shaping the functional genes of soil prokaryotes. Similar results have been observed by Maulet al.(2019) who demonstrated that the abundance of soil denitrification genes of soil microbes in agroecosystems was driven by cropping systems in combination with soil properties (Maulet al.2019).

More genes were related to microbe-microbe interaction pathways under the OM cropping system than other systems,which may be due to nutrient competition (Zuñigaet al.2017) among the more complex and diverse communities within that system,as well as a suitable environment,including high temperature and humidity,for microbes in OM soil. Genes in the biosynthesis of the type II polyketide backbone pathway were more abundant under the M cropping system than any other system and are probably involved in bacteria-fungi interaction (Liuet al.2019). This is consistent with the presence of heavy soil-borne fungal diseases of maize in the M system because of continuous cropping and straw amendments (Wanget al.2006). The enrichment of pathways involved in ‘degradation of aromatic compounds’ were observed in soil samples from the WM system,reflecting excess aromatic compounds coming from pesticide application and industrial emissions (Sunet al.2017). Low OM content in soil from the PM system probably stimulated more autotrophic bacterial activity and encoded more genes that took part in ‘carbon fixation pathways in prokaryotes’ in PM soils than in soils from other cropping systems. In addition,the climates in PM cropping systems in Northwest China have large temperature and moisture fluctuations (Chenet al.2015),which makes soil prokaryotes encode more genes that regulate the ‘microbial metabolism in diverse environments’ pathway.

5.Conclusion

Overall,this work demonstrated that soil properties,geographic location and cropping systems shaped the prokaryotic communities and functional genes in soil of maize cropping systems over a large geographical area of China. The metagenome analysis provided a rich knowledge base for the better understanding of the soil prokaryotic community at a large scale under different maize cropping systems.

Acknowledgements

This research was supported by the National Program for Support of Top-notch Young Professionals,China,the Open Research Fund of State Key Laboratory for Biology of Plant Diseases and Insect Pests,Institute of Plant Protection,Chinese Academy of Agricultural Sciences (SKLQF201508) and the Project of Plant Protection Key Discipline of Henan Province,China. We thank Dr.Liao Yingyu from Department of Entomology and Plant Pathology,North Carolina State University,for manuscript review and feedback.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on http://www.ChinaAgriSci.com/V2/En/appendix.htm