Applied Soil Ecology 143 (2019)26-34 Pro sents high metabolic versatility and includes a large group of ba t of the also f Is,mainly in the the th studies( ng et a Whe ed os bet d pH i nay ha ely pr D rding 2017)that reported root exu s and soil chemie cal properties lies have shown that the host genotype,to a extent,shape 2017:Mendes t al it is stil uncle arwhether 即ng On the other hand the iner sed abundance of Acidohe Finally,we use tiemtrepositioninth il.The these g oups of bacteria tinely.ou le.both ering and bulk soil.Als th bed as oligc probable c tition bet tha e,the g wi nalysis om wer prefere nt in m important in the net s in soils.b ized cial role b initiating the degra tion of c For thi phere of both the t et al.,2012).In th linke e comple occus-Ther antly found in maia 2013)and pr ogens this phylu the phere of ma ntly found i ia found in s nts and,althou t transformation (Li et al.2014).In cov ea.Mic ell c oting b the rhia Taken ether the c alysis revealed that th sof the ommu that the group Mi d o)Th 5.Conclusion m has small ge and a pr ed symbiotic or pa udied.this found in the hi ,nds taL,20184d 2014 can present similar.or nd diversity than hizosphe ng the s then et al.20 05).Chaudh et a erial community and iatro and did not find can infl ce Acknowledgments ed by This study was funded by on 30506920181T thar to the Centro de C e matica(CeGenBio Proteobacteria presents high metabolic versatility and includes a large group of bacteria found in soils, mainly in the rhizosphere (Janssen, 2006). Proteobacteria was also found abundantly in the rhizosphere of maize in previous studies (Chauhan et al., 2011; Yang et al., 2017). On the other hand, cowpea presents the ability to change soil pH in the rhizosphere (Rao et al., 2002), which may have contributed to decrease the abundance of Proteobacteria. These results agree with Zhou et al. (2017) that reported root exudates and soil chemical properties as factors driving the soil microbial community, especially pH that is the primary chemical driver of the bacterial community in soil (Lauber et al., 2009). On the other hand, the increased abundance of Acidobacteria, Armatimonadetes, and OP11 in cowpea senescence suggests depletion of nutrients in the soil, since cowpea grew after maize and there was not nutrient reposition in the soil. Thus, these groups of bacteria were in- fluenced by the presence of oligotrophic soils as discussed previously. The increase of these groups in the senescence stage can also be linked to the decrease of rhizospheric effect. For example, both Acidobacteria and Armatimonadetes are described as oligotrophic groups that grow in environments with low input of nutrients (Tamaki et al., 2011; Kielak et al., 2016). The rhizospheric effect, i.e. the rhizodeposition of nu￾trients by plant roots, occurs up to the plant flowering, decreasing with the plant developmental stage. The differential abundance of some groups between flowering and senescence is also related to the type of exudate. Bacteria present in the rhizosphere of young plants have a preference for simple amino acids, while the groups present in mature plants prefer more complex carbohydrates (Houlden et al., 2008). This change in abundance of specific groups is an indicative of a change in the quality of plant root exudates. For example, Firmicutes plays a crucial role by initiating the degradation of complex substrates such as plant cells walls, starch particles and mucin (Flint et al., 2012). In this context, the increasing of these groups in the senescence of maize can be linked to the ability to use complex carbohydrates and respond under competition for nutrients. Interestingly, Deinococcus-Thermus was abundantly found in maize rhizosphere as compared with cowpea rhizosphere and it agrees with previous studies that found this phylum in the rhizosphere of maize (Chauhan et al., 2011; Correa-Galeote et al., 2016). Deinococcus com￾prises a group of bacteria found in several environments and, although this phylum is not well characterized, there is information that they can act as plant growth-promoting bacteria in the rhizosphere (Lai et al., 2006). Specifically, Deinococcus-Thermus was found acting as biolo￾gical control of soil-borne pathogens in the rhizosphere of cotton (Zhang et al., 2011). It is also worthy to note that the group OP11 presented a significant enrichment only in cowpea senescence. This bacterial group was recently named Microgenomates and proposed to belong to the superphylum Patescibacteria (Rinke et al., 2013). This superphylum has small genomes and a presumed symbiotic or parasitic lifestyle (Sanchez-Osuna et al., 2016). Although Microgenomates is still poorly studied, this group was already found in the rhizosphere of maize, tomato, and soybean (Liang et al., 2014; Correa-Galeote et al., 2016; Sanchez-Osuna et al., 2016). The richness and diversity did not vary between bulk soil and maize rhizosphere, while decreased in cowpea rhizosphere. Usually, bulk soil can present similar, or higher richness and diversity than rhizosphere since plant rhizosphere selects a subset of bacterial groups from the soil, then decreasing the diversity (Weisskopf et al., 2005). Chaudhry et al. (2012) evaluated the microbial diversity in the rhizosphere of switch￾grass and jatropha and did not find differences between bulk soil and rhizosphere soils. Comparing plant species, we found a difference in richness and diversity between legumes and grass, a similar result re￾ported by Zhou et al. (2017), which indicated that bacterial community indices associated with legume were different from that associated with grass. When the rhizosphere samples were compared to the bulk soil, we found that 11% of the genera detected in the rhizosphere were not found, in a detectable ratio, in the bulk soil. On the other hand, most of the genera (67%) are shared between bulk soil and rhizosphere. In this sense, the rhizospheric community is a subset of the bulk soil, which is the main source of microbial species colonizing the rhizosphere (Mendes et al., 2014). When compared the rhizosphere groups between the two different plants, we detect a small proportion of genera ex￾clusively present in each plant, reinforcing our observations of differ￾ential community assembly according to the plant species. Recent stu￾dies have shown that the host genotype, to a minor extent, shapes the root microbiota profiles (Bulgarelli et al., 2012; Pérez-Jaramillo et al., 2017; Mendes et al., 2018a); however, it is still unclear whether mi￾crobiome divergence is more significant in host species belonging to different plant families (Abbo et al., 2014). Finally, we used network analysis to understand the microbial community dynamics between bulk soil and rhizosphere and the dif￾ferent developmental stages of the plants. Interestingly, our data re￾vealed that the rhizosphere community in the senescence was more complex than the flowering and bulk soil. Also, in the senescence period, there was an increase in negative correlations. These results suggest a probable competition between microorganisms since that, during the senescence, the rhizosphere activity and root exudation decrease (Miranda et al., 2018). Therefore, these characteristics of rhizosphere during the senescence may contribute to increase the complexity of the bacterial community. According to the network analysis, some keystone species of bacteria were identified in each treatment. In the bulk soil, we found that the genus DA101 was the most important in the network structure. This genus belongs to Ver￾rucomicrobia phylum and was recently described as abundant and ubiquitous in soils, being characterized as an aerobic heterotroph with many putative amino acid and vitamin auxotrophies (Brewer et al., 2016). For rhizosphere of both maize and cowpea, the genus Bacillus was detected as keystone species in the network (specifically for MF, MS and CS treatment). The genus Bacillus has long been considered dominant in the rhizosphere of plants being recognized for their func￾tions in several important biological processes related to plant growth (Mendes et al., 2013) and protection against pathogens (Mendes et al., 2018b). In maize, we also detected the genus Mycobacterium as a key group in the community network. This genus was abundantly found in the rhizosphere of maize and harbors diverse functional genes for nu￾trient transformation (Li et al., 2014). In cowpea, Microlunatus was detected as key species, and this group is related to phosphorus meta￾bolism in soils (Kawakoshi et al., 2012). Taken together, the co-oc￾currence network analysis revealed that the dynamics of the commu￾nities in the rhizosphere change according to the plant developmental stage, and the key groups are involved in beneficial traits related to plant nutrition and health. 5. Conclusion In this study, the response of the bacterial community was different between bulk and rhizospheric soils, confirming the power of the rhi￾zosphere environment to shape the microbial community. Also, this study showed that the structure and diversity of bacterial community vary significantly according to plant species and, in a minor extent, their developmental stage. Interestingly, the complexity of bacterial community increases during the senescence as compared with the flowering. It confirms that the variation in rhizosphere activity during the plant growth can drive the responses of the bacterial community, which in turn, can influence plant growth and health. Acknowledgments This study was funded by Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq/Brazil (grant 305069/2018-1). The authors thank to the Centro de Genetica e Bioinformatica (CeGenBio) from the Unit of Research (NPDM/UFC). Ademir Sergio Ferreira de Araujo, Vania Maria Maciel Melo, and Marcia do Vale Barreto A.S.F. de Araujo, et al. Applied Soil Ecology 143 (2019) 26–34 32