Genomic insights into the versatility of the plant growth-promoting bacterium Azospirillum amazonense

Background The species Azospirillum amazonense belongs to a well-known genus of plant growth-promoting bacteria. This bacterium is found in association with several crops of economic importance; however, there is a lack of information on its physiology. In this work, we present a comprehensive analysis of the genomic features of this species. Results Genes of A. amazonense related to nitrogen/carbon metabolism, energy production, phytohormone production, transport, quorum sensing, antibiotic resistance, chemotaxis/motility and bacteriophytochrome biosynthesis were identified. Noteworthy genes were the nitrogen fixation genes and the nitrilase gene, which could be directly implicated in plant growth promotion, and the carbon fixation genes, which had previously been poorly investigated in this genus. One important finding was that some A. amazonense genes, like the nitrogenase genes and RubisCO genes, were closer phylogenetically to Rhizobiales members than to species of its own order. Conclusion The species A. amazonense presents a versatile repertoire of genes crucial for its plant-associated lifestyle.


Background
The genus Azospirillum (α-proteobacteria class) encompasses free-living bacteria that can improve the growth of many economically important plants, mainly cereals (for an extensive review, see [1]). Therefore, these microorganisms are considered as plant growth-promoting rhizobacteria (PGPR). Species of this genus are widely distributed in nature, living in soils of tropical, subtropical and temperate regions all over the world. Several aspects of their physiology seem to be related to a plant stimulatory effect, notably their ability to synthesize phytohormones. Although these microorganisms are able to fix atmospheric nitrogen, the exact contribution of this process to plant growth is still disputable [1][2][3].
So far, fifteen species of the Azospirillum genus have been described ( [4] and references therein). However, most research efforts have been dedicated to the species Azospirillum brasilense, neglecting the potential offered by the biological diversity of this genus.
The bacterium A. amazonense, the focus species of this study, was initially isolated from forage grasses grown in the Amazon region. Further studies revealed its broad ecological distribution, as it is also found in association with the roots of gramineous plants like rice, maize, sugarcane and sorghum [5,6]. This species is phylogenetically closer to Azospirillum irakense and Rhodospirillum centenum (also known as Rhodocista centenaria) than to A. brasilense. Unlike the latter, A. amazonense can use sucrose as a sole carbon source and is better adapted to acid environments [6].
In order to access the valuable information that genomic sequences can provide on the physiology of azospirilla, there have been independent efforts by research groups in sequencing their genomes. Currently, the genomes of three members of the Azospirillum-R. centenum group are available: Azospirillum sp. B510 [7], A. brasilense Sp245 [8] and R. centenum SW [9]. Although many years have passed since the discovery of A. amazonense, there is scarce information about this species. Nevertheless, a recent study under greenhouse conditions showed that especially A. amazonense Y2 (wildtype strain) contributed to the growth of rice plants by means of biological nitrogen fixation [10], showing its potential for use as an agricultural inoculant. Therefore, the objective of our study was to sequence the A. amazonense Y2 genome and to analyze specific regions that could exert fundamental roles in its survival in the soil and in its ability to promote plant growth.

Results and Discussion
General features of the A. amazonense draft genome database The A. amazonense Y2 presents four replicons with the following estimated sizes: 2.7 Mbp, 2.2 Mbp, 1.7 Mbp and 0.75 Mbp [11] The genomic G+C content of A. amazonense Y2 is 66.89%. The draft genome sequence consists of 7.044.835 bp divided in 1617 contigs. The average gene length is 1080. Currently, there are 3319 predicted CDS, where 2299 have sequence similarity to known genes, 501 are homologs to genes of unknown function and 519 are hypothetical genes exclusive to A. amazonense.

Taxonomic features of A. amazonense
In this study, a phylogenetic tree was constructed using the 16S rDNA sequences from microorganisms belonging to the orders Rhodospirillales and Rhizobiales (Figure 1). The resulting phylogenetic tree clearly shows a split between these orders. The outermost clade containing all the Azospirillum species divides in two main subclades: one containing A. amazonense, A. irakense, Rhodocista pekingensis and R. centenum, and another containing the other Azospirillum species. This result is in agreement with previous studies, showing the close evolutionary relationship between A. amazonense and R. centenum [12][13][14], and is also supported by the greater number of A. amazonense genes (22%) with best BLAST hits (KEGG Database) to R. centenum genes.
Another relevant feature of the 16S rDNA phylogenetic reconstruction is that R. centenum does not cluster with other Rhodospirillum species, suggesting that the genus nomenclature of R. centenum is not appropriate, as has been pointed out by previous publications [13,15].

Carbon metabolism
Azospirilla display versatile carbon metabolism in order to support their lives in the soil. A. amazonense is capable of growing on various disaccharides, hexoses and pentoses and a previous study suggested that A. amazonense is able to catabolize carbohydrates exclusively through the Entner-Doudoroff pathway (ED pathway) [16]. The genes encoding the key enzymes of this pathway, 6-phosphogluconate dehydratase and 2-dehydro-3deoxy-phosphogluconate aldolase (KDPG aldolase) (Additional file 1), were found in the A. amazonense genome and seem to be organized as an operon. This same study also suggested that the glycolysis pathway (Embden-Meyerhof-Parnas pathway) was inoperative in A. amazonense, because no activity of 6-phosphofructokinase and fructose bisphosphate aldolase was detected in crude extracts [16]. However, predicted genes encoding those enzymes were found in the A. amazonense genome (Additional file 1). Therefore, although the genomic approach indicates that most probably A. amazonense is also able to consume carbohydrates via glycolysis, this catabolic feature should be experimentally retested.
As stated previously, one of the main differences between A. amazonense and A. brasilense is that A. amazonense is capable of consuming sucrose as the sole carbon source [6]. In the genome of A. amazonense, a predicted gene that codes for a putative α-glucosidase was identified (Additional file 1). This enzyme converts sucrose to glucose and fructose, substrates that can be promptly consumed by catabolic pathways.
The A. amazonense genome also harbors homologs of the genes salB and salA of A. irakense (Additional file 1). These genes encode β-glucosidases, enzymes implicated in the acquisition of glucose by means of the hydrolysis of aryl-β-glucosides, such as salicin [17].
Bacteria of the genus Azospirillum produce high levels of poly-β-hydroxybutyrate (PHB), the energy and carbon storage source utilized under nutritional stress conditions [2,18]. The essential genes for PHB biosynthesis are present in the A. amazonense genome: phbA (βketothiolase), phbB (aceto acetyl coenzyme A reductase) and phbC (PHB synthase) (Additional file 1). Furthermore, the phaZ gene that encodes a PHB depolymerase (Additional file 1), the first enzyme of the PHB degradation pathway, was also found in its genome.
One of the most surprising features of the A. amazonense genome is the presence of a gene cluster implicated in carbon fixation (the Calvin-Benson-Basham cycle) (Figure 2 and Additional file 1). The main genes of this cluster are the genes cbbL and cbbS, and they encode, respectively, the large and small subunits of ribulose-1,5-bisphosphate carboxylase (RubisCO). This enzyme is responsible for the incorporation of carbon dioxide in a molecule of ribulose-1,5-bisphosphate, generating two molecules of 3-phosphoglycerate, which can subsequently be used in biosynthetic pathways. A phylogenetic analysis of the concatenated RubisCO small and large subunits of A. amazonense revealed that they belong to the Form IC of RubisCOs ( Figure  3). This type of enzyme is commonly found in α-Proteobacteria and it is adapted to environments with medium to high CO 2 and the presence of O 2 (in general, RubisCOs also have affinity to O 2 and high levels of this molecule can inhibit CO 2 fixation) [19]. So far, there have been no reports showing that A.
amazonense has autotrophic behavior. However, from the Azospirillum group, at least R. centenum and A. lipoferum are known to be capable of growing autotrophically by means of RubisCO [9,20], unlike Azospirillum sp. B510 and A. brasilense Sp245, which do not Figure 1 The evolutionary history of 16S rDNA sequences from members of the orders Rhodospirillales and Rhizobiales. The tree was constructed using the Neighbor-Joining method with 1000 bootstrap replicates (percentages greater than 50% are shown next to the branches). The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. There were a total of 1254 positions in the final dataset. The species Azospirillum amazonense is indicated by an arrow. The gray boxes indicate members of the order Rhodospirillales.
contain Form I or II of RubisCOs ("true" RubisCOs) encoded in their genomes.
The RubisCO phylogenetic reconstruction also indicated the close relationship of the A. amazonense enzyme with those from members of the family Bradyrhizobiaceae (order Rhizobiales) (Figure 3), namely Rhodopseudomonas palustris and Bradyrhizobium spp. In fact, the genetic organization of the carbon-fixation cluster of A. amazonense resembles that found in Bradyrhizobium sp. BTAi1, contrasting with the genetic organization of R. centenum. These incongruities, i.e. the genetic organization and phylogenetic relationship closer to Bradyrhizobiaceae members than to R. centenum, suggests that horizontal gene transfer may be an important driving force in the evolution and dispersion of RubisCOs in Proteobacteria.

Nitrogen metabolism
Azospirillum species are able to utilize distinct nitrogen sources, including ammonia, nitrate, nitrite, dinitrogen and amino acids [3,21]. The A. amazonense species has several genes implicated in nitrogen metabolism, which encode transporters, enzymes and regulatory proteins (Additional file 1).
Ammonia is the central compound of nitrogen metabolism and the preferred nitrogen source of many microorganisms. In general, nitrogen sources other than ammonia are converted into it to be assimilated [22]. The A. amazonense genome contains genes that are implicated in this conversion of alternative nitrogen sources, like nitrate/nitrite, urea and dinitrogen (Additional file 1). Once available, ammonia can be incorporated into the metabolism by the glutamine synthetase (GS)/glutamine:oxoglutarate aminotransferase (GOGAT) pathway, the genes for which are also encoded in the A. amazonense genome (Additional file 1).  Overall, the conversion of nitrogen compounds to ammonia expends some energy and, therefore, the metabolic pathways implicated in this process are strictly regulated to minimize energy waste. The central regulators of nitrogen metabolism are the PII proteins [22,23]. Three PII homolog genes (glnB, glnK and glnK2) were found in the A. amazonense genome (Additional file 1). The glnK gene and the glnB gene have ortholog counterparts in Azospirillum sp. B510, A. brasilense Sp245 and R. centenum. The glnK gene is upstream of the aat gene (aminotransferase) and the glnB is upstream of the glnA gene (glutamine synthetase) [24]. The third gene, glnK2, which is located downstream of the amtB gene, is absent in A. brasilense Sp245, Azospirillum sp. B510 and R. centenum, although this genetic association is frequently found in diverse prokaryotes [25].
The PII protein interactions with transporters, transcription factors and regulatory enzymes are well-established in the literature (for a review, see [22,23]) and the potential interaction targets found in the A. amazonense genome will be briefly discussed. One putative target is the glnD gene that codes for an uridilyl-transferase, an enzyme that uridylylates the PII proteins under nitrogen-limiting conditions [23]. Other potential targets found in the A. amazonense genome are two ammonium transporters, encoded by the amtB genes, which in the presence of high nitrogen levels are inhibited by PII proteins [26]. The adenylyltransferase enzyme (encoded by the glnE gene, Additional file 1) which regulates glutamine synthetase via covalent modifications [27] could also interact with PII proteins [28].
The analysis of the A. amazonense genome revealed the presence of the NtrBC and NifA systems (Additional file 1), which are PII-regulated signal transduction systems responsible for the coordination of genes implicated in the scavenging of alternative nitrogen sources [23]. Both NtrC and NifA rely on the presence of the sigma N factor (also known as RpoN, or sigma 54) to activate the transcription of specific genes [29], which is also present in the A. amazonense genome (Additional file 1).
One of the main characteristics of the Azospirillum species is that they are able to fix nitrogen, i.e. convert N 2 to ammonia, by means of the nitrogenase enzyme complex. The main genes implicated in this process are known as nif genes, and they are highly conserved among nitrogen-fixing proteobacteria [30].
A preliminary BLAST analysis showed that the A. amazonense nif genes exhibit high similarity with genes of some species of the order Rhizobiales (Additional file 1). Since these observations were unexpected, the phylogenetic history of nifH was reconstructed utilizing sequences from species of the orders Rhodospirillales and Rhizobiales. The resulting nifH tree (Figure 4) was clearly incongruent to the 16S rDNA tree (Figure 1): in the nifH tree, A. amazonense and A. irakense grouped with Bradyrhizobium sp. BTAi1, Xanthobacter diazotrophicus and Azorhizobium caulinodans, instead of grouping with other Azospirillum species.
The genetic organization of the nif genes between the Rhodospirillales and Rhizobiales bacteria is somewhat homogeneous ( Figure 5). As demonstrated in Figure 5, some features of the nif cluster of A. amazonense are exclusively similar to the homolog cluster of the species Bradyrhizobium sp. BTAi1, such as the presence of three conserved hypothetical genes that are indicated by the numbers 2, 5 and 6. On the other hand, Azospirillum sp. B510 and A. brasilense have the aerC gene between the nifHDK and nifENX operons and the draG and draT genes in the upstream region of the nifH gene, features not shared with the nif cluster of A. amazonense. The genes draG and draT code for a posttranslational control system of the nitrogenase, which are not present in A. amazonense and R. centenum [9,31].
The domain composition of the deduced NifE and NifU proteins is also evidence that the nitrogenase complex of A. amazonense is more distantly related to Azospirillum spp. (with the exception of A. irakense) than to some Rhizobiales species (i.e. those from the genus Bradyrhizobium, Xanthobacter and Azorhizobium). The NifE protein of A. amazonense (like the Bradyrhizobium, Xanthobacter and Azorhizobium proteins NifE) has a bacterioferritin-associated ferredoxin [2Fe-2S] binding domain (BFD), not present in the NifE protein from A. brasilense Sp245 and Azospirillum sp. B510, and R. palustris (Rhizobiales) (Additional file 2). Furthermore, its NifU protein is smaller than those of A. brasilense Sp245, Azospirillum sp. B510 and R. palustris CGA009 because it does not contain the BFD and an N-terminal NifU domain, which are present in the NifU protein of the latter species (Additional file 2). Therefore, taken together, these results indicate that complex events occurred in the evolution of the nitrogenase system in these bacteria, for instance, horizontal gene transfers and/or gene duplication followed by differential gene loss, culminating in the current distribution of the nif genes among the members of these taxonomic orders.
Nitrogen fixation is a very energy-demanding process and it is not surprising that the nitrogenase system is elaborately regulated. In all diazotrophic species of the Proteobacteria examined so far, the transcriptional activator NifA and the sigma N alternative RNA polymerase-associated factor are the master regulators of nitrogen fixation genes [32]. Sequence motifs similar to the consensus region of sigma N and NifA binding sites are present upstream of the nifH gene, the homolog of "orf2" (indicated by the number 7 in Figure 5) of the orf2nifUSVorf4 cluster from A. brasilense [33] and the nifB gene (Additional file 3).

Energy production and conversion for nitrogen fixation
Nitrogen fixation demands the systematic action of different genes. The fix genes are essential for this process, and they encompass different functional categories. These genes were found in the A. amazonense genome divided into three main clusters, namely fixABCX, fixLJ-fixK and fixNOQP-fixGHIS.
The fixABCX genes from A. amazonense, responsible for electron transfer to nitrogenase, are located downstream of the nifW gene ( Figure 5). As seen in Figure 5, this genetic cluster is highly conserved among the analyzed species, and it is tightly associated with the nif cluster. Putative NifA and sigma N binding sites were found upstream of the operon fixABCX from A. amazonense (Additional file 3), indicating that these transcription factors could be key elements for the expression of this operon. This evidence is corroborated by reports showing that the operon fixABCX is regulated by the NifA protein in A. brasilense and Rhizobium spp. [34][35][36].
The A. amazonense genome also possesses a gene cluster including fixLJ and fixK. In symbiotic diazotrophs, the transcription of fix genes involves the oxygen-responsive FixLJ two-component system. The FixL protein, in the absence of oxygen, autophosphorylates and transfers the phosphate group to FixJ. Finally, the phosphorylated FixJ activates the expression of FixK, which activates the transcription of genes required for microaerobic growth [32].
Molecular nitrogen reduction requires high levels of energy under microaerobic conditions. The fixNOQP and fixGHIS genes encode membrane-bound cytochrome c oxidase and the redox process-coupled cation pump, respectively, which are intimately implicated in respiration under microaerobic conditions, supplying energy for nitrogen fixation [37,38]. These clusters were found in tandem in the A. amazonense genome (Additional file 1), and they show identical organization in many diazotrophic α-Proteobacteria, like R. centenum, A. brasilense Sp245, R. palustris and Bradyrhizobium spp.
Nitrogen fixation forms molecular hydrogen (H 2 ) as a byproduct. Therefore, diazotrophic bacteria have several hydrogenase systems that are responsible for oxidizing molecular hydrogen to recover part of the energy expended during nitrogenase activity. Genes encoding for an uptake NiFe hydrogenase (hupSL) were identified in A. amazonense ( Figure 6). Furthermore, the accessory proteins required for maturation of the subunits [39,40], encoded by the hup and hyp genes, are situated downstream of the genes encoding hupSL ( Figure 6). This organization resembles that found in members of the order Rhizobiales, where, in general, the hup and hyp genes are clustered (Figure 6), although their ordering is quite heterogeneous among the species. In contrast with this observation, comparisons with closely-related species showed that Azospirillum sp. B510 and A. brasilense Sp245 have their hup and hyp genes scattered across the genome. Moreover, the bacteria R. centenum does not have the hydrogenase gene cluster.
An integrated model relating the components discussed in this section with nitrogen metabolism is depicted in Figure 7, taken into consideration the similarity of A. amazonense genes to those from other wellknown bacterial systems.

Phytohormone production
Species of the Azospirillum genus can positively influence plant growth and crop yield by the biosynthesis and secretion of indole-3-acetic acid (IAA) [2,41]. However, although A. amazonense strains are able to synthesize IAA [10], very little is known about the molecular mechanisms responsible for this process.
In A. brasilense, at least three pathways for IAA biosynthesis exist, two tryptophan-dependent pathways (indole-3-acetamide pathway (IAM) and indole-3-pyruvate pathway (IPyA)) and one tryptophan-independent pathway [3,42]. Similarly, the genome of Azospirillum sp. B510 contains genes responsible for the IAM pathway [7]. However, the iaaM, iaaH and ipdC genes, related to the IAM or IPyA pathways, were not located in the A. amazonense genome. Further analysis of the   genome sequence of A. amazonense revealed a gene encoding a protein with about 70% similarity to nitrilases from plant species, like Arabidopsis thaliana and Zea mays, which catalyze the conversion of indole 3acetonitrile to IAA [43,44]. Future studies may verify if this gene is implicated in IAA biosynthesis in A. amazonense.

Quorum sensing and biofilm formation
Quorum sensing is an intercellular signaling process implicated in the regulation of several traits of bacteria, notably antibiotic biosynthesis and biofilm formation. The archetype for quorum sensing regulation is the LuxIR system, which involves an acyl-homoserine lactone (AHL) synthase (LuxI homolog) and an AHLdependent transcriptional regulator (LuxR homolog) [45]. The quorum sensing phenomenon of Azospirillum species is strain-specific and seems to regulate functions linked to rhizosphere competence and adaptation to plant roots [46].
The acyl-homoserine lactone (AHL) biosynthesis ability of forty Azospirillum strains (including A. amazonense Y2) was previously investigated, and only four A. lipoferum strains seemed of being capable of synthesizing these compounds [47]. However, the genome analysis of A. amazonense revealed the presence of genes encoding for LuxI and LuxR homologs proteins (Additional file 1). Therefore, these results indicate that A. amazonense Y2 could synthesize AHLs and respond to its presence in the environment. The genome of A. amazonense also presents a Klebsiella pneumoniae ahlK homolog [48], a predicted gene that codes for a putative homoserine lactonase (Additional file 1) implicated in AHL degradation. Since bacterial plant pathogens rely on quorum sensing mechanisms to infect plants [45], a study of A. amazonense homoserine lactonase activity on the deleterious activities of these pathogens would be relevant.
Extracellular polysaccharides are loosely bound to the cell surface and play an important role in bacteriumplant interactions through the firm and irreversible anchoring of cells to the plant roots [3]. In the A. amazonense genome, two genes, noeJ (mannose-6-phosphate isomerase, Additional file 1) and noeL (GDP-mannose 4,6-dehydratase, Additional file 1), which are related to extracellular polysaccharide biosynthesis and biofilm formation, were also found [49].

Chemotaxis/Motility
Different species of Azospirillum attach to and colonize plant root surfaces and these processes depend on chemotaxis. Azospirilla exhibit chemotaxis towards sugars, amino acids, organic acids and root exudates [50]. This ability offers the bacteria the advantage of moving towards favorable nutrient conditions. Genes encoding for the central signal transduction pathway for chemotaxis (che) are present in nearly all motile bacteria. This signal transduction system is composed of the conserved cheAWYBR genes and a group of transmembrane chemoreceptors (known as MCPs or methyl-accepting proteins) that perceive environmental signals. Homologs of the cheAWYBR and MCP genes were identified in the A. amazonense genome (Additional file 1), and some che genes display similar organization to those found in the model organism Escherichia coli [51].
The A. amazonense Che1 gene cluster (cheAWYBR) revealed a conserved organization with the major chemotaxis gene cluster from A. brasilense Sp245, which modulates cell length and clumping behavior [52]. Apparently, this gene cluster also affects the production of exopolysaccharide and flocculation of A. brasilense. Other che-like genes (Che2 and Che3, Additional file 1) are present in the A. amazonense genome, probably encoding parallel signal transduction pathways that could have distinct functions, similar to those found in R. centenum [53,54].
It is also worth noting that genes encoding MCP domain proteins are spread throughout the A. amazonense genome. Some of them were classified accordingly to their similarities to E. coli chemoreceptors. All five types of E. coli MCP receptors were found in the A. amazonense genome (Additional file 1), and it is probable that some of them are related to cell motility by regulating the histidine kinase CheA that phosphorylates a response regulator, which in turn controls the rotational direction of the flagellar motor [55].
The flagellum is a key structure for the chemotactic response. In the A. amazonense genome, 39 flagellar genes were identified, and the majority of them are distributed among different gene clusters (Additional file 1). The reduced flagellar gene number in A. amazonense compared to those of the closely-related A. brasilense (79 annotated genes) and R. centenum (72 genes) species is in agreement with previous evidence indicating that A. amazonense synthesizes only the polar flagellum for swimming motility [56].
Previous studies demonstrated that the transcription factor FlbD is related to the biosynthesis of lateral flagella in A. brasilense and the polar flagella in R.
centenum [9,57]. The presence of the flbD gene in the A. amazonense genome suggests that it could have a similar regulatory function as that found in R. centenum.

Transport, antibiotics resistance and lantibiotic production
Transport systems allow the uptake of nutrients and ions, excretion of end products of metabolism and communication between cells and the environment.
Several components of the cationic efflux pump, the ATP-Binding Cassette (ABC) transporter superfamily, the Major Facilitator Superfamily (MFS) and the TonBdependent transporters were identified in the A. amazonense genome.
TonB-dependent transport is a mechanism of active uptake across the outer membrane normally related to iron uptake, signal transduction and environmental perception [58,59]. A. amazonense has a high number of genes encoding TonB receptors (Additional file 1) when compared with other diazotrophs such as Azospirillum sp. B510 (9 annotated genes) and R. centenum SW (32 annotated genes). Homology analyses suggest that a set of putative TonB receptors for specific substrates like ferrioxamine, cobalamin (B12 vitamin) and heme are present in this bacterium.
Among the different families of transporters, only two occur ubiquitously in all kingdoms of life: the Major Facilitator Superfamily (MFS) and the ATP-Binding Cassette (ABC) superfamily, representing the largest and most distributed families of transmembrane proteins. MFS proteins are single polypeptide secondary carriers that utilize uniport, symport or antiport mechanisms to transport various small substrates [60]. The ABC transporter proteins utilize energy from adenosine triphosphate (ATP) hydrolysis to carry out the uptake of essential nutrients and/or the extrusion of toxic substances [61]. In the A. amazonense genome, several genes encoding for putative MFS and ABC transporters were identified, and they could be implicated in the transport of a wide range of putative substrates (Additional file 1).
Multidrug resistance (MDR) transporters increase drug excretion through an efflux pump, which expels a wide variety of toxic products from the cell, playing a central role in bacterial drug resistance. The MDR transporters belong to various transporter families [62]. In E. coli, the transport of diverse substrates out of the cell by the AcrAB-TolC efflux transporter can confer broad resistance to antibiotics [63]. The acrA and acrB genes normally form an operon whose transcription is regulated by the acrR gene product, and are found in the A. amazonense genome.
Putative drug resistance transporters of the QacA subfamily were found in the A. amazonense genome (Additional file 1), which could confer resistance to monovalent and bivalent cationic lipophilic antiseptics and disinfectants such as quaternary ammonium compounds [64].
Further analysis revealed that, in addition to the MDR transporters, a set of genes whose products could be related to specific antimicrobial resistance are present in the A. amazonense genome. Genes that code for penicillin, glyoxalase/bleomycin and tetracycline resistance are also present in the A. amazonense genome (Additional file 1). These findings corroborate the experimental data that shows that A. amazonense is tolerant to tetracycline and resistant to penicillin [6].
Lantibiotics are peptide-derived antibacterial substances produced by some bacteria, and are characterized by the presence of unusual amino acids like lanthionines and dehydrated amino acids [65]. Lantibiotic biosynthesis is frequently coregulated as part of a stress response when cells enter the late-log or stationary phase [66]. Most lantibiotics exert their antibiotic effect by either forming pores in the target cell membrane or by inhibiting cell wall synthesis, and many lantibiotics are bactericidal against a variety of Grampositive bacteria [67]. Genes related to lantibiotic synthesis were found in A. amazonense genome (Additional file 1).
The genes implicated in antibiotic resistance and in lantibiotic production are probably essential for successful establishment of this microorganism in the soil due to constant contact with niche competitors like fungi and other bacteria.

Bacteriophytochrome
The A. amazonense genome also harbors a bacteriophytochrome gene (Additional file 1). Similarly, Azospirillum sp. B510 has two genes that code for distinct types of bacteriophytochromes [7], and Bradyrhizobium sp. possesses three bacteriophytochrome genes [68]. In plants, phytochromes regulate the metabolic response to the light environment, but a variety of functions is found in other organisms [69]. The bacteriophytochromes in R. palustris regulate the biosynthesis of the photosynthetic apparatus [70], while in Deinococcus radiodurans and R. centenum, they regulate pigment biosynthesis [71]. Subsequent studies must be carried out to understand the role of the bacteriophytochrome in A. amazonense physiology.

Conclusion
In order to thrive, bacteria must adapt readily to environmental shifts by means of a wide variety of genotypic and phenotypic accommodations [72]. The rhizosphere is a good example of a dynamic environment, where fluctuations in its biological and chemical activities demand an appropriate response from its inhabitants. The species A. amazonense is a free-living plant growth-promoting rhizobacterium that is found in association with plants of agricultural importance. In this study, we identified a series of A. amazonense genes that could be essential for adaptation to the competitive environment of the rhizosphere. Its wide genetic repertoire confers a versatile metabolism (e.g. the ability to use different carbon and nitrogen sources), as well as different mechanisms of perceiving and exploring its surroundings. These characteristics could directly influence plant growth, for instance, by providing nitrogen and stimulatory compounds to plants. Another important finding was the greater similarity of some genes, e.g. nitrogenase and RubisCO genes, to genes of members of the order Rhizobiales than to genes from other Azospirillum species. This evidence illustrates the genetic plasticity of this species and indicates that evolutionary phenomena like horizontal gene transfer could be fundamental for adaptation to its environment.
The major impact of this work will be to guide subsequent studies for a better understanding of the potential of A. amazonense.

Methods
Bacterial strain, culture conditions, and DNA isolation A. amazonense Y2 (ATCC 35120) was cultured in M79 medium (10 g/L of sucrose as a carbon source, 0.1 g/L of K 2 HPO 4 , 0.4 g/L of KH 2 PO 4 , 0.2 g/L of MgCl 2 .7H 2 O, 0.1 g/L of NaCl, 0.4 g/L of yeast extract, pH 6.5) [73] with shaking at 150 rpm and 35°C for 18 hours. Genomic DNA was isolated as described by Wilson [74]. The quality of the isolated genomic DNA was assessed by agarose gel electrophoresis.

Genome sequencing, assembly, draft annotation
Total genomic DNA was sequenced using the Roche 454 pyrosequencing platform following the manufacturer's instructions. The contigs were assembled using Newbler software version 2.3 with the default parameters. The estimated coverage of the genome was 35x. Some gaps present in the genes of interest were filled in by sequencing PCR fragments obtained from genomic DNA.
The annotation and analysis of the sequences were carried out using the System for Automated Bacterial Integrated Annotation (SABIA) [75]. This Whole Genome Shotgun project has been deposited at DDBJ/ EMBL/GenBank under the accession AFBX00000000. The version described in this paper is the first version, AFBX01000000.

Phylogenetic analysis
Gene sequences were retrieved from GenBank. Most of the 16S rDNA sequences were retrieved from the Ribosomal Database Project [http://rdp.cme.msu.edu/] [76,77]. The accession numbers of the sequences utilized in the phylogenetic reconstructions are listed in Additional file 4.
Multiple sequence alignments were performed using MUSCLE version 3.8 [78] and CLUSTALW (built into the MEGA 4 software) [79]. Phylogenetic trees were inferred using the neighbor-joining method (1000 bootstrap replicates) by the MEGA 4 software [80]. The evolutionary distances were computed using the Maximum Composite Likelihood method for the nucleotide sequences and the Jones-Taylor-Thornton (JTT) matrixbased method for the amino acid sequences. All positions containing gaps and missing data were eliminated from the datasets (complete deletion option).

Additional material
Additional file 1: Supplementary table 1. Genes of Azospirillum amazonense described throughout the study.

Acknowledgements and Funding
We especially thank Dr. Augusto Schrank for the valuable effort that enabled the realization of this work. We also thank the Empresa Brasileira de Pesquisa Agropecuária -Centro Nacional de Pesquisa de Agrobiologia (EMBRAPA-CNPAB) for providing the A. amazonense Y2 strain. This work was supported by grants from the Brazilian National Research Council (CNPq) and the Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul (FAPERGS). FHS, RC, LAR and FMS received scholarships from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).