LONG-TERM IMPACT OF MAIZE AGROECOLOGICAL MANAGEMENT ON BACTERIAL COMMUNITIES AND SOIL HEALTH IN THE ARID NORTH OF MEXICO

Background. Soil management practices modify the microbial communities and the carbon stocks (organic, inorganic, and total). The increase in microbiological communities’ diversity improves the production of plants; thus, it is essential to understand the predominant bacterial taxa in the soil. Objective. The objective of the present study was to establish the bacterial communities’ alteration by agroecological management in maize crops in arid northern Mexico. Methodology. Bacterial diversity and composition were determined in soils from Coahuila, Mexico, under three different scenarios: i ) Agroecological management (AM), ii ) Conventional management (CM), and iii ) Control (T, with no vegetation). In addition, pH, electrical conductivity (EC), and soil organic matter (SOM) were analyzed using standard methods. Bacterial DNA was extracted from the soil, amplifying the V3-V4 region of the 16S rRNA gene and sequencing with Illumina. The gene sequences were analyzed in QIIME. Results. A total of 20 bacterial phyla and 631 genera were registered. For AM, CM, and T, respectively, the most abundant genera were Tepidisphaera (7.02, 9.29, and 9.93 %), Sphingomonas (6.55, 5.15, and 4.06 %), Microvirga (2.64, 2.39, and 3.63 %), and Blastococcus (2.91, 3.10, and 3.37 %). A significant difference was observed among groups (p = 0.004), where AM was different, which suggests that the type of substrate determines the diversity and abundance of the bacterial community. Significant differences were found for pH and EC, with higher pH in CM (7.87) and T (7.86) soils. The EC was higher in AM (446 μ Scm -1 ) and T (419 μ Scm -1 ). On the other hand, AM showed the best result in SOM content (21.80 ± 1.10 gC kg -1 ). Implication. Therefore, AM in maize crops has the potential


INTRODUCTION
Modern agriculture faces new challenges as it integrates ecological and molecular approaches to achieve higher crop yields while minimizing environmental impacts.Farmers have generally improved their productivity by applying synthetic fertilizers and pesticides to crops, especially in arid and semi-arid areas; however, their uncontrolled use over the years has led to major environmental consequences, including air and water pollution as well as loss of soil fertility (Kosty et al., 2020).In addition, the structure and function of soil microbial communities have been affected by irrational agricultural practices (Semenov, 2021;Shen et al., 2021).Thus, there is a growing interest in the agroecological management of the agroecosystem, integrating the ecology of soil food systems, and the economic and social aspects of sustainable agriculture (Altieri, 2020;Francis et al., 2003;Gliessman, 2013).Agroecology offers an integrated set of solutions that reconciles three central challenges facing agriculture today: (i) Feeding a growing population, (ii) Conserving natural resources, providing sustainable lives and livelihoods for farmers in the food system for agricultural workers, and (iii) For the people who consume their products (Gliessman, 2020).To improve production without the use of fertilizers of synthetic origin is necessary to focus on the beneficial intervention of microorganisms in the soil due to their potential to promote plant growth and reduce the use and abuse of chemically synthesized contaminants (Moreno-Reséndez et al., 2018).Moreover, agroecology tends to achieve system self-regulation by biodiversification in first place (Chavarria et al., 2018).Furthermore, the key role of microorganisms in nutrient cycles and organic matter mineralization, as well as their interactions with plants could be used as indicators to evaluate the effect of different agricultural management systems on soil´s quality (Burton et al., 2010).Therefore, analyzing the microbiome is necessary to improve plant growth, which in turn will benefit sustainable food production.
For a long time, research on soil microbial composition was limited by isolation procedure in Petri plates.Moreover, in the mid-1980s, cultivable microbial colonies were estimated to constitute a small proportion (0.1-5%) of the total soil microbial diversity (Torsvik et al., 1990;Torsvik and Ovreas, 2002).In addition, the high variability of the physicochemical properties of the soil (the content of organic carbon, nitrogen, pH, humidity, and porosity, among others) influences the diversity of its microbial communities (Semenov, 2021).According to Marchesi and Ravel (2015), new molecular technologies and bioinformatics have transformed how the soil microbial community is interpreted.Therefore, microbiome analysis is considered the predominant research approach in soil microbiology and provides a starting point for new research directions and hypotheses (Nesme et al., 2016).Knowing the dominant bacterial taxa of the soil could improve the ability to actively manage communities and promote their functional capacities to increase the production of plants of agricultural importance (Delgado-Baquerizo et al., 2018).Due to this, the objective of the present study was to establish the bacterial communities' alteration by agroecological management in maize crops in arid northern Mexico comparing three scenarios i) Agroecological management (AM), ii) Conventional management (CM), and iii) Control (T, with no vegetation, that represents degraded soil), and considering changes in the physicochemical properties, as well as on carbon sequestration in a semi-arid region of Northern Mexico.

Study area and sampling
The study was conducted in the Universidad Autónoma Agraria Antonio Narro, Unidad Regional Laguna, Torreón, Coahuila, Mexico, on December 2020.During summer the air temperature ranges from 22 to 35 °C, and in winter from 8 to 22 °C; the annual total rainfall average is 225 mm; the climate is mainly very dry semi-warm (BWH); there are sandy loam soils.Soil samples (texture sandy clay crumb) were collected at three contrasting sites: i) Agroecological management (AM-1 to AM-3; coordinates 25.553589N, -103.372848W;adding corn crop and weeds residues from the last year; no fertilizers were applied), ii) Conventional management (CM-4 to CM-6; 25.555654N, -103.372333W;adding NPK fertilizers (250 N, 70 P2O5, both in kg ha -1 ), artificial irrigation, mechanized land tilling systems, and machinery-assisted weed control measures), and iii) Control (T-7 to T-9; 25.565618N, -103.372189;without crops or plants to reflect the degradation degree of analyzed soils).Three random samples from each treatment were taken at a depth of 10 cm, using a California Type auger (SP06505 Model J), which was sterilized with alcohol 96º and fired for 1 min between samples.Likewise, a conventional auger was used to collect other three soil samples from each treatment (mixed to get a combined sample) to determine the physicochemical parameters.

DNA extraction, amplification, and sequencing
An approximate portion of 100 mg was taken from the soil samples and placed in BashingBead™ lysis tubes; then 750 μl of Xpedition™ Zymo Research™ lysing buffer/stabilizer was added.Each tube was placed in a cell disruptor (TerraLyzer™) for 30 sec.DNA from the samples was extracted using the Zymo Research™ DNA Zymobiomics MiniPrep kit according to the manufacturer's instructions.The amount of DNA per sample was measured in a Qubit® brand fluorometer.The amplification of the V3-V4 region of the 16S rRNA gene was performed using the primers suggested by Klindworth et al. (2013): S-D-Bact-0341-b-S-17, 5´-CCTACGGGNGGCWGCAG-3´ and S-D-Bact-0785a-A-21, 5´-GACTACHVGGGTATCTAATCC-3´, which produces an amplicon of ~460 bp.The PCR was made following Illumina protocol for 16S metagenomics, as well as quantification, normalization (equimolarity), library pooling, and massive next-generation sequencing (MiSeq Illumina® 2 × 250 paired-end reads) (Illumina, 2017a).

Bioinformatic analysis
The analysis of the sequences was carried out in an Oracle VM VirtualBox 5.1.14virtual machine in MGLinux using the bioinformatics software Quantitative Insights into Microbial Ecology (QIIME) v.1.9.0 (Caporaso et al., 2010).The process began by assembling the forward and reverse sequences using the PEAR program (Zhang et al., 2014), with quality criteria Q30.Chimeras were eliminated with USEARCH (Edgar, 2010) and the selection of OTUs was performed with the UCLUST method (Edgar, 2010) at 97% similarity; a representative sequence was obtained for each OTU and the taxonomy was assigned taking as reference the updated 2020 EzBioCloud database (Yoon et al., 2017).The absolute abundance of OTUs was obtained and the number of sequences was plotted by the number of taxa at the genus level to observe the coverage depth (asymptote trend curves); PAST Ver.3.15 software was used (Hammer et al., 2001).A simple random rarefaction process was carried out (Weiss et al., 2017) to match the samples to the same number of sequences.Beta diversity was calculated using the Bray-Curtis Index (Beals, 1984); the obtained matrix was used to perform a PERMANOVA test (p < 0.05) to observe significant difference in the microbiota among groups.Also, relative abundances for phylum and genera were obtained.The genera, whose relative abundance was greater than 0.01%, were represented in a heatmap; the hierarchical cluster method with Euclidean measurement was used for the dendrogram of the samples; this visualization was made using Morpheus software (https://software.broadinstitute.org/GENE-E/).

Physicochemical properties of the soil
The pH and electrical conductivity (EC) were determined in a soil water suspension of 1:2 and 1:2.5 (w/v) ratio, respectively (Luján-Soto et al., 2021).On the other hand, the soil organic matter (SMC, gC kg -1 ) was measured by Walkley and Black (1934), and labile carbon (L) and non-labile (NL) carbon fractions due to the oxidation of C by KMnO4 (Blair et al., 1995).

Statistical analysis
First, the data normality and homogeneity of variance were examined in IBM SPSS Statistics program.After both assumptions were met, one-way ANOVA was applied to analyze differences among treatments.The taxa were enumerated to visualize the groups of samples using a principal component analysis (PCA) to observe associations between the ten main phyla and the physicochemical variables of the soils using RStudio (ver.2021.09.2), and thus, determine the variables that have greater relationship to the phyla.The patterns in AM, CM, and T physicochemical properties were analyzed by multivariate redundancy.A redundancy analysis (RDA) was performed on RStudio (Di Felice et al., 2012).Finally, a LEfSe analysis was made to statistically and biologically determine the key biomarkers, which contributed to the differences among groups.The selected clades were those less than 0.05 in the alpha value of the Kruskal-Wallis factorial test >4.0 in the logarithmic LDA score.This analysis was performed on the website http://huttenhower.sph.harvard.edu/lefse/.

Analysis of the bacterial microbiota
The mean obtained from the total number of sequences in AM before performing the assembly was 135,976; for the CM was 123,261, and for T was 106,820.Averages of the assembled sequences were 50,133; 48,016, and 48,759, respectively.After the elimination of the singletons, the respective averages were 15,471; 17,549, and 14,799 (Table 1).An acceptable coverage depth was observed since the curves tended to the asymptote (Fig. 1).
When performing the PERMANOVA test a significant difference was observed among groups (p = 0.004).The AM, CM, and T groups were visualized using principal coordinate analysis (PCoA), showing the separation among them (Fig. 4).

Taxa biomarkers
A bar graph was made for a discriminant analysis of biomarkers (Fig. 5A), and a cladogram to represent the results obtained through the LEfSe analysis to make a comparison of high-dimensional classes with a focus on taxa biomarkers (Fig. 5B).The discriminant analysis and cladogram showed an evident separation among the AM, CM, and T samples, in addition to identifying the predominant taxa by population.

Physicochemical properties of the soil
The results of the physicochemical characterization of the soils are shown in Table 2.The pH of soils was affected by the agroecological management (AM), showing the lowest mean (7.62 ± 0.05).Also, electrical conductivity showed significant difference among treatments, being higher for AM (446.0 ± 25.3 μScm -1 ), and T (419.3 ± 14.86 μScm -1 ).
Regarding SOM, significant differences were found among treatments, where its content was higher in the AM soil (21.80 ± 1.10 gC kg -1 ).The AM soil also displayed a higher TC than the CM soil, but not concerning to T (control).Within the labile carbon (L) and non-labile (NL) carbon fractions due to the oxidation of C by KMnO4, both showed significant differences among treatments, AM soil showing the highest fraction of L (17.82 ± 0.61 gC kg -1 ), and T registering the highest fraction of carbon NL (5.94 ± 0.43 gC kg -1 ).Principal component analysis (PCA) indicated differences between the analyzed soils with a decentralized distribution and an aggregate one (Fig. 6), explaining 92.7% of the variance with 2 components.From the results, it was possible to observe how the addition of organic amendments is related to a higher content of SOM and SOC, especially labile C. Similarly, CM influences a higher pH, while soils without vegetation seem to be related to non-labile C fractions (Table 1).Therefore, the environmental variables that were selected for their importance on the analyzed soils were SOM, SOC, L, NL, and pH.

Correlations between environmental variables and bacterial community
The canonical redundancy analysis of the samples is found in Figure 7.All the phyla found to the left of the "Y" axis are related to the NL carbon and the T (control) soil (Actinobacteria, Firmicutes, Deinococcus).On the other hand, all the bacterial phyla found to the right will be related to the labile carbon fraction and the SOM (Armatimonadales, Gemmatimonadales, Proteobacteria (AM); Acidobacteria, Saccharibacteria_TM7, and Cyanobacteria (CM)).It is highlighted that the control samples were located on the left side of the axis; while the CM and AM samples are on the right.In this way, all the bacterial phyla that are found up the "X" axis are more abundant in soils with a higher pH (Armatimonadales, Gemmatimonadales, and Proteobacteria), such as soil with CM; at the same time phyla below "X" axis will be in soils with higher EC, and higher concentration of TC and SOM, (Deinococcus, Thermus, Planctomyces, phylum Cyanobacteria, and Chloroflexi).

DISCUSSION
The pH is one of the most important chemical properties of the soil and it is a good indicator of the balance among nutrients (Smita-Tale and Ingole, 2015) due to its significant effects on the concentration and absorption of solutes (Akpoveta et al., 2010).In The Comarca Lagunera, calcisols (from the Latin calx, lime) are soils with a substantial accumulation of calcareous material, so their pH ranges from 7.2 to 8.3 (moderately alkaline), with the presence of carbonates (Yescas-Coronado et al., 2018), such as the pH in the soils of the present study.The lowest pH was for the site with agroecological management (7.62), which also had the highest content of SOM and SOC with 1.27% and 2.19%, respectively.The decrease in pH can be attributed to SOM rise (Martínez et al., 2008).The pH decreased by organic matter incorporation favors processes such as nitrification and mineralization, and a certain acidity generated by the various edaphic microbial groups.This coupled to the increase of CO2 (due to the combination of atmospheric CO2 and soil water that results in carbonic acid), which when dissociated, generates H + and causes the pH reduction (Carrasco, 1992;cited by Martínez et al., 2008).
Soil organic matter increased under agroecological management (Table 2).Additionally, degraded soil (T) presented the lowest SOM content.The soil organic carbon (SOC) is directly associated with soil organic matter content (SOM), thus the addition of an exogen organic amendment source normally increases its concentration.Once in soil, organic matter plays an important role in soil biological activity, due to it gives energetical resources to organisms.SOM also provides colloids with a high cation exchange capacity, which effect on physical properties is manifested by modifying the structure and distribution of the pore space of the soil.The amount of SOM and SOC not only depends on local environmental conditions but is strongly affected by its management (Martínez et al., 2008;Yescas-Coronado et al., 2018).On the other hand, electrical conductivity measures ions present in a solution and varies with depth (Dutta and Ram, 2012;Ingole, 2015).It is a determination that correlates with soil properties that affects its texture, the capacity of cation exchange, drainage conditions, organic matter level, salinity, and subsoil characteristics (Solanki and Chavda, 2012).If the electrical conductivity is less than 1000 (µS cm-1), it is considered a normal soil (Deshmukh, 2012), thus the results of this research correspond to this indicator (265-419 µS cm-1).The electrical conductivity of soils varies according to the amount of moisture retained by soil particles and is useful for monitoring the mineralization of its organic matter (De Neve et al., 2000), which could indicate the reason for the differences in the EC between treatments.In that sense, the use of bovine manure seems to be a cause of a higher content of salts and increasing the EC in the AM, thus the type of manure should be considered.
Identification of labile SOC fractions let to understand the turnover of SOC and the enhancement of soil quality by management.The results showed an increase of LC under AM, where the incorporation of crop residues seems to increase this C fraction as other agroecological management were reported (Ma et al., 2021;Xue et al., 2018;Zhang et al., 2020).The increase of labile KmO4-C fraction could be related with a low C/N ratio in added amendments, favoring their decomposition (Mi et al., 2016).However, other parameters such as soil microbiota could also affect the lability of SOC.The benefits from residue management in soil quality and crop yields have been reported by several authors (Piccoli et al., 2020;Shafi et al., 2007;Singh et al., 2018), results that agree with this research.Nevertheless, their potential increases when are combined with other agroecological management such as legumes crop rotation, amendments incorporation or conservation agriculture (Singh et al., 2018;Yescas-Coronado et al., 2018).
Metasequencing has been used to study different soil environments (Alteio et al., 2020;Cabrera-Rodríguez et al., 2020;Clark et al., 2021).Moreover, metagenomic analysis allows analyzing the composition and function of soil microbial communities (Navarrete et al., 2015).There are two fundamental attributes of organisms that affect their distribution and species diversity: their niches and metabolism.The metabolic capacities of an organism must, at least in part, determine its niche, and its niche imposes restrictions on its distribution.Thus, niche distribution, mediated by metabolic requirements, constrains the size and composition of the subset of organisms with niches that match local conditions; thereby limiting local species richness (α diversity) and change in species composition among sites (β diversity) (Okie et al., 2015).In the present study, significant statistical differences were found for β diversity, as indicated by the PERMANOVA results (p<0.004), and the distribution of groups by principal component results (Fig. 6).
According with PCA, higher pH was associated with conventional management, as well as the genus Cyanobacteria (Figure 5).Cyanobacteria proliferation in soils is related with several physicochemical properties, as well as other issues determined by biodiversity in the ecosystem.This genus participates in different soil processes such as nitrogen fixation, soil genesis and conservation, SOM decomposition, soil phosphorus cycling, biocontrol, soil aggregation and aeration, among others (Alvarez et al., 2021;Crouzet et al., 2019).Moreover, Cyanobacteria could protect plants from adverse environmental conditions due to their physiological characteristics, including the high viscosity of their protoplasmic gel and sheaths with high molecular-weight heteropolysaccharides and proteins (Bertocchi et al., 1990;Gantar et al., 1995;S. Singh, 2014;Zulpa et al., 2008).In addition, their exopolysaccharides interact with soil particles to stabilize them (Sepehr et al., 2019), and at the same time, they are a source of SOC for soil microorganisms also rising microbial activity (Crouzet et al., 2019;Nisha et al., 2018).Those mechanisms to resist extreme environmental conditions could be the reason why Cyanobacteria was a biomarker for conventional management in maize crops; Knapen et al., (2007) reported that Cyanobacteria and algae increase cropped topsoil resistance to degradation.
The agroecological management was linked to higher SOC, SOM and L content, as the time that Proteobacteria and Alphaproteobacteria are their more representative biomarkers; Aquidulcibacter was the only biomarker genus.Alphaproteobacteria can conform to several environmental conditions, as well as they can interact with plants by symbiotic and nonsymbiotically.Besides, this phylum is an important group to measure anthropogenic and environmental impact in a soil ecosystem functional diversity (Singh et al., 2022).For example, Matsushita et al. (2015) reported more abundance and diversity of Alphaproteobacteria in organic apple Orchards than in conventional ones; also, Gazdag et al. (2018) had similar results in organic and conventional crop managements of maize, wheat, rye and sunflower production systems.Those results agree with the obtained data in the present study.On the other hand, Proteobacteria are copiotrophic bacteria which prefer living in nutrient-sufficient environmental, which could be the reason of their proliferation in the analyzed agroecosystem with manure addition.Mei et al. (2021) also reported a significant increase in Proteobacteria abundance in maize agroecosystems where manure was added to soils.Moreover, Zhang et al. (2019) found that the abundance of Proteobacteria is directly related with SOC and its labil fractions, such as our results.Likewise, Aquidulcibacter population was also increased by agroecological management.This genus was reported recently isolated from cyanobacterial aggregates in the eutrophic Lake Taihu in China as part of the family Caulobacteraceae (Cai et al., 2017).Aquidulcibacter are Gram-negative aerobic bacteria, which growth at a pH range between 5.5 and 8.5 (optimum pH 7.0) and at 20-40°C (optimum 30°C).They are catalase and oxidase positive and are capable of nitrate reduction.
Control soil with no vegetation, was related with higher no labile fraction probably due to a higher content of soil inorganic C (SIC) with respect of the other sites, which represent an important carbon fraction in arid and semiarid areas (Gao et al., 2018).
Barren spots areas are also an indicator of high salt content in soils (Reddy et al., 2017).The content of soil carbon (SIC and SOC) varies with several environmental variables, including temperature, precipitation, nitrogen availability, among others (Shi et al., 2012).Additionally, three unidentified phylotypes were the highest biomarkers in T samples: PAC002290_o, FJ479147_g and FJ479147_f, all of them part of Actinobacteria phylum.The phylum Actinobacteria, considered one of the most widely distributed in soils, is recognized for its ability to degrade plant residues in vitro, encoding 16% of the total enzymes with activity on carbohydrates.According to Bao et al. (2021), their taxonomic and functional compositions were relatively stable during straw decomposition.In our study, it was the dominant phylum (38%), which demonstrates its importance in the degradation process of the incorporation of crop residues, especially in less fertile soils, since they have genes to fix nitrogen and produce antibiotics that favor them to compete and acquire carbon sources and protect against environmental disturbances (Swarnalakshmi et al., 2016;van Bergeijk et al., 2020;Bao et al., 2021).Moreover, members of this phylum live under extreme conditions, due to this taxon size and diversity, its geographical expansion and ecological relevance (Bao et al., 2021).FJ479147_g, for example, was reported as one of the most abundant actinobacteria members in Atacama Desert in Chile, the most extreme non-polar biome on Earth (Idris et al., 2017).In a similar way, Thomson et al. (2010) reported that Acidobacteria increased their abundance in bare soils, as occurs in this study.Thus, microorganisms exhibit an extraordinary phylogenetic and functional diversity and support biogeochemical cycles (Okie et al., 2015), this explains the differences in biomakers among sites (Fig. 4), with pH being a determining factor at a global level (Delgado-Baquerizo et al., 2018;Lauber et al., 2009;Zhou et al., 2020).

CONCLUSION
The agroecological management of the soil is the way to conserve and expand microbial diversity and favor the development of bacterial genera that efficiently degrade the organic matter that is incorporated, which will favor the health of the soil.Genomic technologies allow us to have an approach to understanding the importance of microbial groups in soils and their biotechnological implications for sustainable and resilient agricultural development in the medium and long term, especially of the main predominant genera Tepisiphaera, Sphingomonas, Microvirga, Geodermatophilus and Blastococcus, among others.

Figure 4 .
Figure 4. Principal coordinate analysis (PCoA) of the dissimilarity among the soil samples: PCoA plotted against the PC1 vs. PC2 vs. PC3 axes.The percentages indicate the relative contribution of the three principal coordinates (PC1-PC2-PC3).

Figure 5 .
Figure 5. LEfSe analysis of bacterial microbiome from Agroecological management (AM), Conventional management (CM), and Control (T) soil samples.(A) Bar graph shows LDA scores which indicate the taxonomic key for differentiation among treatments.(B) The cladogram generated by LEfSe indicates the main biomarkers among treatments.Each successive circle represents one phylogenetic level.Red-colored regions indicate taxa enriched in AM, green-colored region represents taxa enriched in CM, and blue-colored regions indicate taxa enriched in T.

Figure 6 .
Figure 6.Principal component analysis (PCA) of soil properties under different managements.Vectors show the physicochemical soil properties which were related to each treatment.Vectors that pointed in the same direction are associated variables since samples closer to each other exhibited similarity.AM -Agroecological management; CM -Conventional management; T -Control; SOM -Soil organic matter; EC -Electrical conductivity; pH -Potential of hydrogen; TC -Total carbon; L -Labile carbon fraction; NL -No labile carbon fraction.

Figure 7 .
Figure 7. Redundancy analysis showing the effect of soil properties on the relative abundance of the most abundant bacterial taxa in soil samples under different treatments.AM -Agroecological management; CM -Conventional management; T -Control; SOM -Soil organic matter; pH -Potential of hydrogen; L -Labile carbon fraction; NL -No labile carbon fraction.