Stenotrophomonas comparative genomics reveals genes and functions that differentiate beneficial and pathogenic bacteria

Background In recent years, the number of human infections caused by opportunistic pathogens has increased dramatically. Plant rhizospheres are one of the most typical natural reservoirs for these pathogens but they also represent a great source for beneficial microbes with potential for biotechnological applications. However, understanding the natural variation and possible differences between pathogens and beneficials is the main challenge in furthering these possibilities. The genus Stenotrophomonas contains representatives found to be associated with human and plant host. Results We used comparative genomics as well as transcriptomic and physiological approaches to detect significant borders between the Stenotrophomonas strains: the multi-drug resistant pathogenic S. maltophilia and the plant-associated strains S. maltophilia R551-3 and S. rhizophila DSM14405T (both are biocontrol agents). We found an overall high degree of sequence similarity between the genomes of all three strains. Despite the notable similarity in potential factors responsible for host invasion and antibiotic resistance, other factors including several crucial virulence factors and heat shock proteins were absent in the plant-associated DSM14405T. Instead, S. rhizophila DSM14405T possessed unique genes for the synthesis and transport of the plant-protective spermidine, plant cell-wall degrading enzymes, and high salinity tolerance. Moreover, the presence or absence of bacterial growth at 37°C was identified as a very simple method in differentiating between pathogenic and non-pathogenic isolates. DSM14405T is not able to grow at this human-relevant temperature, most likely in great part due to the absence of heat shock genes and perhaps also because of the up-regulation at increased temperatures of several genes involved in a suicide mechanism. Conclusions While this study is important for understanding the mechanisms behind the emerging pattern of infectious diseases, it is, to our knowledge, the first of its kind to assess the risk of beneficial strains for biotechnological applications. We identified certain traits typical of pathogens such as growth at the human body temperature together with the production of heat shock proteins as opposed to a temperature-regulated suicide system that is harnessed by beneficials. Electronic supplementary material The online version of this article (doi:10.1186/1471-2164-15-482) contains supplementary material, which is available to authorized users.


Background
During the last years, the number of human infections caused by opportunistic pathogens has increased dramatically. One natural reservoir of opportunistic pathogens is the rhizosphere, the zone around roots that is influenced by the plant [1,2]. Due to a high content of nutrients, this habitat is a 'microbial hot-spot' , where bacterial abundances including those with strong antagonistic traits are enhanced [3]. Various bacterial genera, including Burkholderia, Enterobacter, Herbaspirillum, Ochrobactrum, Pseudomonas, Ralstonia, Staphylococcus and Stenotrophomonas, contain root-associated strains that can encounter bivalent interactions with both plant and human hosts [2]. Mechanisms responsible for colonization of the rhizosphere and antagonistic activity against plant pathogens are similar to those responsible for colonization of human organs and tissues, and pathogenicity [2]. Multiple resistances against antibiotics are not only found with clinical strains but also with strains isolated from the rhizosphere [4]. High competition, the occurrence of diverse antibiotics and secondary antimicrobial plant metabolites, and enhanced horizontal gene transfer and mutation rates in this microenvironment contribute to the high levels of natural resistances [5]. On the other hand, rhizosphere inhabitants have an enormous potential as biocontrol or stress protecting agents or as fertilizers for sustainable agriculture [6,7].
Stenotrophomonas maltophilia is an emerging global pathogen and already one of the most common opportunistic pathogens in hospitals [8,9]. A recent study shows that approx. 5% of the Gram-negative infections were caused by S. maltophilia in intensive care units in the United States. The two most common diseases caused by S. maltophilia are bacteremia and pneumonia, which are often associated with high mortality rates. S. maltophilia strains are characterized by multi-resistance to many antibiotics [10]. For a long time it was not possible to differentiate between the clinical and environmental S. maltophilia strains [4]. Using a polyphasic approach, Wolf et al. [11] were able to describe a new plant-associated species within the S. maltophilia complex. Interestingly, no human-pathogenic potential has ever been observed in this phylogenetically and ecologically closely related species [11]. Moreover, both species can be easily distinguished with regard to the production of the osmoprotective substance glucosylglycerol (only present in S. rhizophila) and the occurrence of specific multidrug-efflux pumps (only present in S. maltophilia) [12]. S. rhizophila is both rhizosphere-and phylloplane-competent and shows pronounced salt tolerance, and is hence a model bacterium among the plant growth-promoting rhizobacteria (PGPR) [13]. Plant growth promotion by S. rhizophila strain DSM14405 T (syn. strain e-p10) was observed under greenhouse conditions [14] and in the highly salinated soils of Uzbekistan at levels up to 180% [15]. However, S. maltophilia was a typical rhizosphere bacterium used as an efficient biocontrol agent, and until the 1980s, no capacity to cause infection had ever been reported. Now, the theory is established that the ancestors of virulent bacteria including Stenotrophomonas, as well as the origin of virulence and resistance determinants, lay most likely in the environmental microbiota [5]. It is now one of the main challenges to predict any risk for human health [16]. Currently, these potential risk factors are a main obstacle in registration procedures, especially in the European Union [17]. Next generation sequencing and the corresponding bioinformatic analyses have an enormous impact on our understanding of microbial communities and the host-microbe interactions [18,19]. However, is it possible to use these techniques to solve this problem?
The objective of our study was to find out if there is a borderline based on the distinguishing features between the beneficials and pathogens within the genus Stenotrophomonas. Moreover, using genomics, transcriptomics and physiological assays, we try to predict the potential risk of the stress protecting agent S. rhizophila strain DSM14405 T for humans by studying its genetic potentials and comparing these with two Stenotrophomonas model strains, the human-pathogenic S. maltophilia K279a [20] and the plant-associated S. maltophilia R551-3 [21].

Comparisons of plant and human-associated Stenotrophomonas genomes
General genomic features of S. rhizophila DSM14405 T were compared to the plant-associated S. maltophilia R551-3 and the human pathogenic S. maltophilia K279a ( Table 1). The genome of S. rhizophila DSM14405 T consists of 4,648,976 nucleotides with a GC content of 67.26%, and has been predicted to code for 4,033 CDSs. Compared with S. maltophilia R551-3 and S. maltophilia K279a with each 4,573,969 n (4,039 CDSs) and 4,851,126 n (4,386 CDSs), respectively, the size of the genome and the predicted number of CDSs are slightly smaller. There is no plasmid present in S. rhizophila DSM14405 T . Figure 1 shows genomic comparisons between the genome of S. rhizophila DSM14405 T , S. maltophilia R551-3 [21] and S. maltophilia K279a [20]. Overall, there is a very high degree of sequence similarity between the genome of the plant growth-promoting S. rhizophila and both S. maltophilia R551-3 and the humanpathogenic K279a. The homology boxes are however separated by non-homologous regions. The whole genome sequence comparison between the three genomes was also performed using the Artemis Comparison Tool (ACT) which revealed the same result as in Figure 1 and is presented in Additional file 1: Figure S1. Figure 2 represents circular genomic map of S. rhizophila DSM14405 T , wherein DNA coordinates are presented and coding sequences (CDSs) on both the leading and lagging strands are assigned colors with respect to functional gene groups. In addition, the figure depicts the position of tRNA and rRNA genes, and other characteristics of the S. rhizophila DSM14405 T genome such as GC content as well as the excess of C over G throughout the whole genome (GC skew) with the regions above and below average shown in different colors. Furthermore, gene orthology analyses were performed, in which S. rhizophila genes were compared with those of S. maltophilia R551-3 and K279a to detect orthologous genes shared among strains ( Figure 3). In addition to orthologous genes, Figure 3 depicts the position and color-coded function of the S. rhizophila DSM14405 T -specific genes. The numbers of orthologous and strain-specific unique genes are shown in the Venn diagrams ( Figure 4a). As presented in the figure, the number of CDSs shared between S. rhizophila and S. maltophilia R551-3 or S. maltophilia K279a is 3171 (862 specific to S. rhizophila) and 3149 (884 specific to S. rhizophila), respectively. Moreover, the orthology analysis including all three strains revealed that 762 CDSs are unique to S. rhizophila DSM14405 T , as these are absent from both S. maltophilia R551-3 and S. maltophilia K279a (Figure 4a). The percentage distribution of the 762 S. rhizophila-unique CDSs with regard to functional gene groups revealed that genes involved in carbohydrate transport and metabolism (4.07%), and the biogenesis and transport of cell wall, outer membrane or cytoplasmic membrane (3.02%) are of relatively high abundance (Figure 4b).
Specific gene characteristics of S. rhizophila DSM14405 T within the Stenotrophomonas/Xanthomonas group Quorum sensing Similar to various xanthomonads, S. rhizophila DSM14405 T does not possess a homoserine lactone-based quorum sensing system, but instead uses the rpf/DSF system for quorum sensing and cell-cell communication. The rpf (regulation of pathogenicity factors) gene cluster is responsible for the synthesis and perception of the DSF molecule which is a quorum sensing regulatory molecule of fatty acid nature with similarity to enoyl-CoA hydratase, and was first detected in Xanthomonas [22,23]. The rpfF gene product, known as DSF synthase, is essential for the synthesis of DSF [24,25]. Other members of the rpf gene locus (rpfC,  rpfG and rpfB) have been revealed to each fulfill a particular function, with the RpfC/RpfG two-component system consisting of a sensory (RpfC) and regulatory (RpfG) component that are responsible for DSF perception and signal transduction, respectively [22,26]. Similar to Stenotrophomonas strains K279a and R551-3, the core of the rpf gene locus of S. rhizophila DSM14405 T consists of four genes: rpfB, rpfF, rpfC and rpfG. Fouhy et al. [27] described the positions of these in the humanpathogenic S. maltophilia K279a, which we found to be similar to those in the plant-associated S. maltophilia R551-3. Moreover, in both K279a and R551-3, rpfB and rpfF are located on the lagging strand while rpfC and rpfG are located on the leading strand. In S. rhizophila DSM14405 T , however, the rpfB and rpfF genes are located on the leading strand while rpfC and rpfG are on the lagging strand. In addition, in the genome of S. rhizophila, there is a 228-nucleotide gene of unknown function on the lagging strand between rpfF and rpfB, which extends from 2469447 to 2469674 and was annotated as Sr14405 DX03_10710. In addition, Sr14405 DX03_10710 is transcribed in the cell, as we detected the corresponding mRNA in a whole genome expression analysis approach (data not shown). There is no homologue to Sr14405 DX03_10710 in either S. maltophilia R551-3 or K279a, and its function in S. rhizophila DSM14405 T remains to be elucidated.

Flagella, fimbriae
Flagella and fimbriae-driven motility is crucial for biofilm formation and host-plant colonization by bacteria [28][29][30]. S. rhizophila DSM14405 T possesses several genes responsible for motility. A gene block encoding a flagellar apparatus was detected that includes 22 genes, and extends over the genome from Sr14405 DX03_10335 to DX03_10430 with most genes located on the leading strand. Another flagella-encoding gene block was also detected which includes 26 genes and is located between Sr14405  4. tRNA genes; 5. rRNA genes; 6. GC plot with regions above and below average in gray and black, respectively; 7. GC skew showing regions above and below average in dark yellow and magenta, respectively (window size: 10,000 bp). The circular genome map was constructed using DNAPlotter.
Chitinase, extracellular proteases, antibiotic and salinity resistance S. rhizophila DSM14405 T is a biocontrol agent capable of synthesizing extracellular enzymes with anti-pathogenic activity such as chitinase and extracellular proteases, and is antagonistic against important fungal pathogens such as Verticillium dahliae and Rhizoctonia solani [11,31]. In addition to its direct effect, S. rhizophila DSM14405 T is thought to also indirectly promote plant growth through biological control [14]. In the genome, Sr14405 DX03_17135 codes for a putative extracellular chitinase gene, and Sr14405s DX03_03415, DX03_05635, DX03_16295, DX03_17120 are predicted to code for extracellular proteases.
In general, Stenotrophomonas species are known to show resistance against a broad range of antibiotics [8,31]. There are numerous resistance genes against various antibiotics in the genome of S. rhizophila DSM14405 T , some code for general resistance, while others provide resistance against particular classes of antibiotics. The gene cluster extending from Sr14405 DX03_01540 to DX03_01550 was predicted to code for a multidrug export system. Another two multidrug resistance gene clusters were detected from Sr14405 DX03_02075 to DX03_02085 and from DX03_13460 to DX03_13470. A number of single multidrug resistance genes, such as mdtN, mdtA and Sr14405 DX03_03380 are scattered throughout the genome as well. Moreover, macA and macB code for the macrolide-specific efflux protein and a macrolide export ATP-binding/permease, respectively. Other identified genes include: Sr14405 DX03_06420 and ampH that code for β-lactamases, Sr14405 DX03_08150 that codes for an aminoglycoside efflux pump, and a transposon tetracycline resistance gene (tetX).
S. rhizophila DSM14405 T possesses both ggpS and ycaD, which are essential for the synthesis and transport of the important osmolyte glucosylglycerol, which provides tolerance against salinity and salt stress [32]. Both ggpS and ycaD are absent in S. maltophilia R551-3 and K279a. Figure 3 Gene orthology analyses between S. rhizophila DSM14405 T , S. maltophilia R551-3 and the clinical S. maltophilia K279a. Circles show from the outermost to the innermost: 1. DNA coordinates; 2, 3. Function-based color-coded mapping of the CDSs predicted on the forward and reverse strands of the S. rhizophila DSM14405 T genome, respectively. 4. Orthologous CDSs shared between S. rhizophila DSM14405 T and S. maltophilia R551-3. 5. S. rhizophila-specific CDSs, compared with S. maltophilia R551-3. 6. Orthologous CDSs shared between S. rhizophila and S. maltophilia K279a. 7. S. rhizophila-specific CDSs, compared with S. maltophilia K279a. 8. GC plot depicting regions above and below average in gray and black, respectively; 9. GC skew showing regions above and below average in yellow green and magenta, respectively (window size: 10,000 bp). The assessment of orthologous CDSs was carried out using the reciprocal best BLASTp hit approach with an identity threshold of 30% and evalue of 10 −6 .

Surface polysaccharides
Homologs to xanA, xanB, and rmlAC were detected in S. rhizophila DSM14405 T . These genes are involved in the biosynthesis of the Xanthomonas well-known surface polysaccharide xanthan, in biofilm formation [33] and the biosynthesis of lipopolysaccharides. It is noteworthy that S. rhizophila DSM14405 T , similar to other members of the Stenotrophomonas species known so far, does not have a gum gene cluster and therefore cannot produce xanthan.
The bacterial capsule is an extracellular structure usually composed of polysaccharides which is considered an important virulence factor in surface adherence, antibiotic resistance, and preventing phagocytosis [34,35]. Reckseidler et al. [36] demonstrated that the ability to synthesize capsule is crucial for virulence in the human pathogenic Burkholderia. In S. rhizophila DSM14405 T , a gene block from Sr14405 DX03_11185 to DX03_11265 is homologous to a capsule biosynthesis gene cluster of Pseudomonas pseudomallei, described by [37]. This gene block includes genes that code for proteins of various functions such as signal transduction, transport, and biosynthesis of capsule polysaccharide components. None of the genes present in the S. rhizophila capsule gene block were detected in S. maltophilia R551-3 and K279a. Alginate, an exopolysaccharide, is involved in the development and architecture of biofilms and protects bacteria from antibiotics and other antibacterial mechanisms [38][39][40]. Alginate biosynthesis genes algI and algJ code for the poly (beta-D-mannuronate) O-acetylase and the alginate biosynthesis protein, respectively. While both were detected in S. rhizophila DSM14405 T , neither the plant-associated strain S. maltophilia R551-3 nor the human pathogenic S. maltophilia K279a contained either of these genes. algI is preceded by four genes which are also absent from both S. maltophilia R551-3 and K279a, with one of these being homologous to a gene coding for a cell morphology protein from the biocontrol agent P. fluorescens SBW25.

Secretion systems
While type II and V secretion system genes were identified in S. rhizophila DSM14405 T , there is no type III secretion system present, as this is typical of Stenotrophomonas. Although there are several genes belonging to the type IV secretion system, a complete gene set was not detected in S. rhizophila DSM14405 T . Furthermore, a gene block extending from Sr14405 DX03_08870 to DX03_09120 was identified in S. rhizophila, which includes numerous genes of the type VI secretion system (T6SS) including icmF, impA, genes belonging to the Hcp1 family, and genes coding for proteins with a T6SS Rhs element. With the exception of Sr14405 DX03_09050, DX03_09095, and DX03_09115, there were no homologs in S. maltophilia K279a and R551-3 to any of the genes of the S. rhizophila type VI secretion system block.
One genus, two entirely different habitats and life styles? : Genome comparison between S. rhizophila DSM14405 T and S. maltophilia K279a All genes of the plant growth-promoting environmental S. rhizophila DSM14405 T and the clinical human pathogenic S. maltophilia K279a were compared. While absent from S. maltophilia K279a, numerous S. rhizophila-specific genes play a role in host-plant colonization. Some of these genes, as described earlier, are crucial for surface attachment, biofilm formation, secretion systemsdriven molecular mechanisms, and tolerance of environmental stress such as high soil salinity. In addition, another S. rhizophila-specific gene was predicted to code for spermidine synthase (speE). Spermidine is a plant growth regulator and has been recently shown to strongly promote the growth of arugula plants [41]. There are also S. rhizophila-specific genes that are predicted to be involved in the biodegradation of bacterial and plant cell wall. mltD, located closely to the S. rhizophila type VI secretion system gene block, codes for muramidase that plays an important role in the bacterial cell wall breakdown. Furthermore, a gene block stretching from Sr14405 DX03_09870 to DX03_09895 was predicted to code for several genes involved in the breakdown of plant cell walls.
As a next step, the S. rhizophila DSM14405 T having a counterpart in the plant-associated S. maltophilia R551-3, but with no homologous genes in the clinical S. maltophilia K279a were studied. Of these 88 genes, several help with the need to adapt to the plant and rhizosphere as the natural habitat. For instance, the endo-1,4-betaxylanase B gene (xynB) is involved in plant cell wall biodegradation. Other genes conserved in S. rhizophila DSM14405 T and S. maltophilia R551-3 are the ferrichrome receptor genes, fcuA and fhuA which code for siderophore receptors, and the outer membrane adhesin-like gene (Sr14405 DX03_14745). Table 2 presents a list of selected S. rhizophila DSM14405 T -specific genes with no homologs in the human pathogenic S. maltophilia K279a, together with their biological role. The complete list of the 884 S. rhizophila DSM14405 Tspecific genes which are not present in S. maltophilia K279a is provided in Additional file 2: Table S1.
Next, the genomes of S. rhizophila DSM14405 T and S. maltophilia K279a were compared using the latter as the reference, which revealed that 1230 genes are specific to the human pathogenic K279a with no homologs in the plant-growth promoting S. rhizophila. While many of these genes are hypothetical or have unknown protein function, others play a specific role. Of the genes with a known or predicted function, many are involved in pathogenicity and virulence. For example, the gene block extending from Smlt 2997 to 3005 codes for proteins of the type IV secretion system, which is known to have a dual role of both horizontal gene transfer and pathogenicity. Other S. maltophilia K279a-specific virulence genes are Smlt 3048, 3683, and 4452, which were predicted to code for an outer membrane-located adhesin, hemolysin, and hemagluttinin, respectively. Furthermore, Smlt 4391, and afaD code for putative exopolysaccharide and adhesin, respectively. In addition, a K279a-specific putative fimbriae gene block (Smlt 706-709) codes for fimbrial adhesin proteins and their chaperones. Moreover, there are also several S. maltophilia K279a heat shock and chaperone-encoding genes that have no homologs in S. rhizophila DSM14405 T , such as Smlt 1818 and Smlt 4629-4631, as both code for heat shock chaperone proteins. Synthesis of chaperones to cope with temperaturecaused stress is fundamental for the natural habitat of S. maltophilia K279a as a human pathogenic strain. There are various S. maltophilia K279a-specific genes coding for antibiotic resistance and multidrug efflux pumps that are absent from S. rhizophila DSM14405 T . For instance, smeABC (Smlt 4474-4476), which code for a multidrug efflux pump typical of S. maltophilia strains, is not present in DSM14405 T . Nevertheless in total, six of the nine K279a genes that code for multidrug efflux pumps involved in antibiotic resistance [20] have a homolog in S. rhizophila DSM14405 T . Other S. maltophilia K279aspecific genes with respect to antibiotic resistance include β-lactamase genes, Smlt 4159 and 4211, and Smlt 1071 (qnrB), which codes for a fluoroquinolone resistance protein. Moreover, several heavy metal resistance genes for the transport of arsenate, mercuric, and copper are another genomic characteristic of the human pathogenic S. maltophilia K279a. Table 3 shows various S. maltophilia K279a genes with no homologs in the plant growthpromoting S. rhizophila DSM14405 T together with their products and biological roles. The complete list of the 1230 S. maltophilia K279a-specific genes is presented in additional file 3: Table S2. Figure 5 represents a model that summarizes crucial specific mechanisms harnessed by S. rhizophila DSM14405 T and S. maltophilia K279a to best adapt to their particular habitats together with the corresponding genes, and the important mechanisms that are shared between both species.
The temperature limit: the transcriptional response of S. rhizophila DSM14405 T to 35°C In contrast to other Stenotrophomonas strains including S. maltophilia K279a and R551-3, S. rhizophila DSM14405 T is not able to grow at 37°C, which is a critical temperature   for strains to colonize the human body. To understand why this strain and other S. rhizophila strains cannot grow at this temperature and to reveal the underlying molecular mechanisms behind, the impact of high temperature (35°C) on S. rhizophila DSM14405 T was studied using a transcriptomic approach. In comparison with growth at 30°C, 328 and 49 genes are significantly up and downregulated, respectively, as the result of a severe heat shock of 35°C ( Figure 6). As presented, genes responsible for translation and post-translational modification along with other functional gene groups including those involved in the metabolism and transport of amino acids, nucleotides, lipids, and the production and conversion of energy are up-regulated at 35°C. Moreover, those genes whose functions are only generally known are also mostly up-regulated suggesting that, in S. rhizophila DSM14405 T , a significant number of the genes involved in the response to severe heat shock code for proteins with unknown function. In contrast, genes responsible for cell motility, secretion and intracellular trafficking show a pronounced down-regulation at 35°C, making up 24.49% of all down-regulated genes. Of the genes with a significant transcription fold change as the result of the shift from 30 to 35°C, some code for products with known biological function ( Table 4). As shown in the table, genes coding for chaperones involved in both general stress and heat shock specific response including dnaJ-K, and groS-L and htpG, respectively, are up-regulated at 35°C. Other up-regulated genes include rpoH which codes for the heat shock specific sigma factor, and numerous genes coding for proteases that are responsible for the degradation of stress-denatured proteins. Furthermore, genes responsible for cell suicide mechanisms, Sr14405 DX03_18795 and DX03_18990, are also Figure 6 The impact of heat shock (35°C) on the gene expression of S. rhizophila DSM14405 T . Genes responsible for translation and post-translational modification, the metabolism and transport of amino acids, nucleotides, lipids, and the production and conversion of energy are up-regulated while genes involved in cell motility and intracellular trafficking are strongly down-regulated. The values above each column correspond to the percentage abundance of the corresponding functional gene group relative to the total count of the up and down-regulated genes. The transcription fold change for each CDS corresponds to the ratio calculated for S. rhizophila treated with 35°C compared to 30°C. Data are presented as the mean value of two independent replicates. The error bar shown on each functional group corresponds to the mean value of errors for all genes belonging to that functional group.
up-regulated as the result of the shift from 30 to 35°C. In contrast, as shown in Table 4 genes coding for pili-driven cell motility are negatively affected at 35°C.

Discussion
In the present work, we characterized the genomic features of the plant growth promoting and biocontrol agent S. rhizophila DSM14405 T . In addition, the comparison between the genome of S. rhizophila with other environmental and clinical Stenotrophomonas model strains, S. maltophilia R551-3 and K279a, respectively revealed a great deal of homology shared among all three bacteria. Surprisingly, the comparison of S. rhizophila's genes with either of the other two strains showed that the total number of genes shared between S. rhizophila and the human-pathogenic S. maltophilia K279a is only slightly less than the total gene number shared with the plant-associated beneficial S. maltophilia R551-3 strain. Given this significant degree of similarity, we concluded that the immense difference between S. rhizophila and S. maltophilia K279a with regard to the habitat and lifestyle must be based upon the role of those genes that are specific to either of the strains and absent from the other. On this account, a number of S. rhizophila specific genes coding for products of known function were detected and classified according to their biological roles. These include genes involved in plant growth promotion, plant cell wall degradation, biofilm formation, resistance against salinity and the type VI secretion system (T6SS). Clearly as an environmental plant-associated strain, S. rhizophila faces particular habitat-specific factors such as those that impose some sort of osmotic stress on the cell including root exudates and salinity. In a study published recently, we used a transcriptomic approach to study the manner of response of S. rhizophila DSM14405 T to osmotic stress [42]. Surprisingly, the combination of both the genomic and transcriptomic data reveals that a significant number of the genes with known biological role that are greatly involved in osmotic stress protection including cbg1, xynB, ggpS, ycaD, algJ and the T6SS are specific to S. rhizophila and absent from S. maltophilia K279a. This finding confirms the importance of the strain specific genes in the adaptation and performance of S. rhizophila in its natural habitat. As a clinical humanpathogenic strain, S. maltophilia K279a also possesses particular genetic potentials that make possible its lifestyle (Table 3, Figure 5). The role of type IV secretion system, hemolysin and other adherence components in virulence and pathogenicity was described earlier. A yet more crucial feature of S. maltophilia K279a with regard to performance in its habitat, however, is perhaps its ability to grow at 37°C. Also in this regard, this strain possesses specific heat shock resistance genes that are absent from S. rhizophila.
As mentioned earlier, S. rhizophila DSM14405 T is not able to grow at 37°C. To understand the manner of response of S. rhizophila to severe growth-inhibiting temperature, a genome-wide gene expression analysis was performed at 35°C using transcriptomics. The positive impact of the growth temperature at 35°C compared to 30°C on several functional gene groups including those responsible for the metabolism and transport of amino acids, lipids and the energy providing genes is due to the heat stress that the cell is coping with. Furthermore, studying the expression of single genes revealed that those genes coding for general and heat shock specific chaperones, various proteases that break down stress-denatured, abnormal proteins are upregulated. For example at 35°C, the primary general stress sigma factor 24 coding gene, which is strongly expressed at an earlier stage under cell stress, and initiates the expression of rpoH (the heat shock specific sigma factor 32), specific chaperones (e. g. fkpA) and protease coding genes (e. g. degP) is down-regulated while rpoH, and other response mechanisms show upregulation. This confirms an advanced state of shock for S. rhizophila at 35°C, at which heat shock specific response mechanisms are strongly activated. In this regard, the up-regulation of Sr14405 DX03_18795 and DX03_18990 (absent from S. maltophilia K279a) which code for toxin/antidote-based suicide systems further illustrate the mechanisms used by S. rhizophila DSM14405 T to respond to the severe heat stress. Moreover, reduced cell motility, which is also part of these response mechanisms, is presumably due to the fact that costly processes such as motility are to be minimized during harsh stress times.

Conclusions
Overall, there is a great deal of similarity between the beneficial and human-pathogenic Stenotrophomonas model strains. Nevertheless, the adaptation to the habitat and lifestyle to guarantee the survival of the species is cared for by mechanisms specific to either of the groups, and many genes underlying these mechanisms in S. rhizophila DSM14405 T are strain-specific. Furthermore according to the genomic and transcriptomic analyses together with its physiological characteristics, S. rhizophila poses no threat to human health and could hence be safely applied in biotechnology.
Genome sequencing, assembly, annotation of S. rhizophila DSM14405 T The genome of S. rhizophila DSM14405 T was sequenced using a combination of next generation sequencing platforms. A first draft assembly based on 905,689 reads of an 8kbp paired-end library (Roche 454 GS, FLX Titanium, Helmholtz Center Munich, Germany) with a total of 167.1 Mbps (36-fold coverage) was generated with Newbler 2.6 (Roche Diagnostics, Penzberg, Germany). This assembly consisted of 175 contigs, 122 of which could be joined into a single circular scaffold. Gaps resulting from repetitive sequences were resolved by in silico gap filling, remaining gaps were closed by PCR followed by Sanger sequencing or by long reads from a Pacific BioSciences sequencing run (PacBio RS, 150,305 reads, 174.8 Mbps, 38-fold coverage, GATC, Konstanz, Germany), yielding a draft genome of 4,648,936 bps. To improve the quality of the sequence by eliminating 454 sequencing artefacts in homopolymer stretches, the genome was subsequently sequenced using the Illumina paired-end method (Illumina HiSeq 2000, 15,086,654 reads, 1508 Mbp; 324-fold coverage, Ambry Genetics, Aliso Viejo, CA, USA). The Illumina reads were aligned to the draft genome with CLC Genomics Workbench 4.7.2 (CLC bio, Aarhus, Denmark). The final consensus sequence was derived by counting instances of each nucleotide at a position and then letting the majority decide the nucleotide in the consensus sequence. Genes were identified with the Prodigal gene finder [43], ARA-GORN [44], and RNAmmer 1.2 [45]. Functional annotation of the predicted genes was performed using BASys [46], which provides annotations with respect to Clusters of Orthologous Groups (COG) [47], Pfam [48] and Gene Ontology (GO) [49]. The final genome includes 4,648,976 bases with a GC content of 67.26%.

Comparative genomics and bioinformatic analyses
Whole genome comparisons between S. rhizophila DSM14405 T , S. maltophilia R551-3 and K279a were performed using Mauve 2.3 [50] and Artemis Comparison Tool (ACT) [51]. In the approach using Mauve, the Progressive Mauve algorithm was used to score the genome alignment. DNAPlotter [52] was used for circular genome visualization. Orthologous coding DNA sequences (CDS) shared between S. rhizophila DSM14405 T and the other two Stenotrophomonas were assessed by performing reciprocal BLASTp best hits with an identity and e-value threshold of 30% and 10 −6 , respectively.
Cell culture growth conditions for transcriptomic analyses S. rhizophila DSM14405 T was grown in 100 ml Erlenmeyer flasks containing 50 ml CAA minimal medium at 30°C under agitation until OD 600 of 0.9 was reached. To introduce heat shock prior to RNA extraction, some of the flasks containing the culture were exposed to 35°C for 2 h under identical agitation conditions while the rest of the flasks remained at 30°C for the same period of time (control).

RNA extraction and transcriptomic analyses
RNA was extracted from 250 μl of each of the cultures exposed to 30°C and 35°C in duplicates using the RNeasy Mini Kit and the RNAprotect® Bacteria Reagent according to the manufacturer's protocol (Qiagen, Hilden, Germany). Total RNA was sent to a sequencing service offered by GATC Biotech (Konstanz, Germany) where the samples were processed according to company's proprietary protocols including depletion of rRNA (Ribo-Zero RNA Removal Kit, Epicentre, Madison, USA), fragmentation of mRNA, random-primed synthesis of cDNA, double strand synthesis and library preparation. Sequencing was performed using Illumina HiSeq 2000 and 50 bp single read mode resulting in 15,202,000 to 37,966,600 quality reads per sample. Reads were mapped to the reference genome of S. rhizophila DSM14405 T and only the corresponding normalized values for the reads that uniquely mapped to each CDS were used to assess the changes in gene transcription. The transcription fold change as the result of the temperature shift was assessed for each CDS by dividing its read number for the cell culture treated at 35°C by the value from the culture grown at 30°C. Of the total genes either up or down-regulated, only those showing fold changes greater than or equal to 1.5 and less than or equal to 0.6 were considered as significantly impacted.

Availability of supporting data
The genome of Stenotrophomonas rhizophila DSM14405 T was deposited in the NCBI database and can be accessed under the BioProject Nr. CP007597.