By September 2023, there were 1,206 genome sequences of Stenotrophomonas in GenBank. A total of 727 unique genomes were downloaded. The downloaded genome sequences were then evaluated using CheckM (Parks et al., 2015). Only 160 genomes of S. maltophilia (>99.5% completeness and < 0.5% contamination) and 53 genomes of other species of Stenotrophomonas (>90.0% completeness and contamination <6.0%) were selected for subsequent comparative genomic analyses. More details about the 213 Stenotrophomonas genomes are listed in Supplementary Table S1. The global distribution and habitat preference of the genus Stenotrophomonas was evaluated using the analytical pipeline Microbe Atlas Project (MAP)¹ (Matias Rodrigues et al., 2017).

All the downloaded genome sequences were annotated using Prodigal (Hyatt et al., 2010). The faa files that were produced were used for pan- and core-genome analyses using the Bacterial Pan Genome Analysis tool (BPGA) pipeline v. 1.3 (Chaudhari et al., 2016). In a pan-genome analysis, the number of accumulated genes that are related to the number of genomes can be predicted by Heaps’ law as n = a . x b (Tettelin et al., 2008). In the equation, x is the number of genomes, while a and b are fitting parameters. When 0 < b < 1 indicates that the pan-genome is open, and only b < 0 indicates that the pan-genome is closed.

The core genome of Stenotrophomonas was constructed by aligning 213 genomes using USEARCH with a cut-off value 50% sequence identity. After the resulting core genes were concatenated and aligned using MUSCLE, a phylogenetic tree was constructed based on the concatenated core genes using the neighbor-joining (NJ) method (Chaudhari et al., 2016).

The amino acid sequences of the polyamines, IAA and cytokinin synthases were aligned using ClustalW (Hall, 2013). A phylogenetic tree based on these single coding sequences (CDS) was constructed using MEGA software v. 6.0 (Tamura et al., 2013) with the NJ algorithm (Saitou and Nei, 1987). The tree topology was evaluated using the bootstrap analysis based on 1,000 resampling replicates. The iTOL online server² was used to display the phylogenetic tree (Letunic and Bork, 2021). The ANI values between these strains were calculated using fastANI v. 1.33 (Jain et al., 2018). The dDDH values were calculated using the Genome to Genome Distance Calculator (GGDC 2.5)³ (Auch et al., 2010).

A COG analysis of the sequences of proteins from the core, accessory and unique gene families was used to categorize them according to the COG database using BPGA (v. 1.3). To ensure the accuracy and completeness of the results, only 33 complete Stenotrophomonas genomes were selected for analysis of their primary metabolic profile and the prediction of secondary metabolite biosynthetic genes using KAAS (KEGG Automatic Annotation Server) and the antiSMASH version 6.0 online⁴ (Blin et al., 2021), respectively. The carbohydrate active enzymes (CAZymes) were identified using the HMMER method of the dbCAN online server⁵ (Zheng et al., 2023) with the e-value thresholds ≤1e⁻¹⁵ and coverage >0.35.

The genes related to PGP, oxidative and osmotic stresses, as well as resistance to heavy metals, were firstly searched in 33 complete genomes, and the sequences obtained were compared with the reference sequences in GenBank to determine the correctness of the sequences. Subsequently, the sequences that were obtained were used as the reference sequences to search their counterparts in the remaining genomes using BLAST.⁶ Only sequences that had >40% identity with the reference sequences indicated that the genome had the gene (Zhao et al., 2023).

The antibiotic resistance genes were identified using resistance gene identifier (RGI) of the CARD⁷ (Alcock et al., 2019), and the sequences with Perfect, Strict and Loose hits identities >70% were selected for further analysis.

Integrated prophages were identified using the PHASTER online server⁸ (Arndt et al., 2016). The predicted prophages with completeness score > 90 were thought to be intact; those with a completeness score of 60–90 were questionable prophages, and those with a completeness score < 60 were considered to be incomplete prophages. The genomic islands (GIs) of the genomes were predicted using the IslandViewer 4 online server⁹ with the IslandPick (Langille et al., 2008), SIGI-HMM (Waack et al., 2006), IslandPath-DIMOB (Hsiao et al., 2003), and Islander (Hudson et al., 2014) methods (Bertelli et al., 2017).

To learn the original source of the biosynthetic genes for polyamines, IAA, and cytokinin, the topologic differences of the phylogenetic tree between the single genes and the core genome were compared. The DNA genomic G + C content of the single genes and the whole genomes were compared (Syvanen, 1994; Garcia-Vallve, 2000).

Among the 213 strains, 78 were isolated from environmental habitats, such as soil, plants, sludge, water, and biofilm reactors, and 135 were from materials from clinical settings, such as urine, blood, sputum, wounds, lungs, and animals among others. The species S. maltophilia, S. rhizophila, S. indicatrix, S. pavanii and S. geniculata can be found in both clinical and environmental habitats, while S. acidaminiphila, S. lactitubi, S. nitritireducens, and S. bentonitica were only found in environmental habitats.

The genome size of the 213 strains of Stenotrophomonas ranged from 3.50 to 5.12 Mb with a G + C content of 63.8–69.3%. The number of CDS ranged from 3,099 to 5,180. The strains with the largest genome and CDS numbers were S. geniculata NWUBe21, which was isolated from seeds, and S. geniculate E119, which was isolated from feces, while the smallest was S. pictorum JCM 9942^T, which isolated from soil. S. acidaminiphila Au-Ste40, which was isolated from soil, contained the highest content of genomic DNA G + C, while S. nitritireducens 2001 that was isolated from lake sediments had the lowest.

The phylogenetic tree based on core genes showed that the 213 strains of Stenotrophomonas can be subdivided into 10 clusters (Figure 1A). Among them, Clusters I, II, III, IV, V and IX are composed of two (S. indicatrix and S. lactitubi), two (S. chelatiphaga and S. tumulicola), four (S. terrae, S. nitritireducens, S. pictorum and S. humi), two (S. acidaminiphila and S. nitritireducens), two (S. bentonitica and S. rhizophila) and two (S. maltophilia and S. geniculate) species, respectively. Clusters VI, VIII, and X all consist of S. maltophilia, while Cluster VII only comprises strains of S. pavanii. The ANI values between all the strains tested were higher than 81%, which confirmed that the 213 strains of Stenotrophomonas belonged to a single genus (Supplementary Table S2). The topology of the tree based on ANI values was similar to that based on the core genome with only minor differences (Figure 1B). Cluster VII formed a clade with Cluster VI in the ANI phylogenetic tree.

There were 14 different Stenotrophomonas species found in 10 clusters. Among them, S. lactitubi, S. indicatrix, S. acidaminiphila, S. nitritireducens, S. bentonitica, S. rhizophila, S. maltophilia, S. pavanii, and S. geniculata contain multiple strains. The ANI values of strains within the same species of S. lactitubi (95.1–96.9%), S. indicatrix (96.0–98.8%), S. acidaminiphila (97.1–98.9%), and S. pavanii (98.6–98.7%) were > 95.0% (Supplementary Table S2), which indicated that the non-type strains of these species were accurately classified. In contrast, the values between some non-type strains within the other species of Stenotrophomonas compared to their type strain were lower than the threshold value for the delineation of a species (ANI value <95% and dDDH value <70%), which indicated that these strains had been misidentified. For example, strains of S. nitritireducens are present in both Clusters III and IV, and some S. maltophilia strains tightly clustered with S. geniculata in Cluster IX. Strains 2001 and SCN18_13_7_16_R1_B_68_91, which are currently both identified as S. nitritireducens, are scattered in Clusters III and IV, respectively. The ANI and dDDH values of strain 2001 to the type strain of S. nitritireducens were only 84.9 and 27.0%, respectively, indicating that it had been misidentified. Nine of 12 S. rhizophila strains had <90% ANI and < 35% dDDH values to the type strain of S. rhizophila (DSM 14405^T), which suggested that these nine strains should not be identified as S. rhizophila. Strains of S. maltophilia were widely scattered in Clusters VI, VIII, IX, and X. Only those in Cluster X shared high ANI (>95%) and dDDH (>70%) values with the type strain, while the two values of the other three cluster strains with S. maltophilia NBRC 14161^T ranged from 91.2–94.0% and 41.0–52.0%, respectively. The low values indicate that these strains were misidentified as S. maltophilia (Supplementary Table S2). Among them, the strains in Cluster VIII share high values of ANI and dDDH with each other and low values with the other strains, indicating that they belong to an unidentified novel species, while the ANI and the dDDH values between the strains in Cluster VI were 90.7–99.9% and 39.5–99.5%, respectively, indicating that the strains in this cluster can be further divided into two or more different species, and the classification of the Cluster IX strains was similar to that of Cluster VI.

The pan-genome of the 213 Stenotrophomonas strains contained 27,186 gene families. There were 710, 15,437, and 11,039 core, accessory, and unique genes, respectively. Each of the 213 Stenotrophomonas genomes were composed of 15.1–24.3% core genes and 69.0–83.5% accessory genes, while the unique genes only accounted for 0–11.1% of the genomes. According to Heaps’ law, the pan-genome of Stenotrophomonas was open and increasing (b = 0.48) (Figure 2A).

The most common categories of COGs in the core genome of Stenotrophomonas (40.0%) were related to metabolism, while those of the accessory genomes (33.8%) and unique genomes (37.3%) were both related to information storage and processing (Figure 2B). The most abundant components of the core genome were related to the maintenance of primary cellular processes, such as class E (9.0%; amino acid transport and metabolism), class C (7.9%; energy production and conversion) and class J (7.1%; translation, ribosomal structure and biogenesis) (Figure 2C). The percentages of classes E, C, and J in the accessory and unique genomes were 4.9 and 4.8%, 3.6 and 3.1%, and 2.1 and 1.6%, respectively. In contrast to the core genome, the accessory and unique genomes were primarily related to the functions of adapting to environmental niches. They were the most enriched in class K (transcription), class U (intracellular trafficking, secretion, and vesicular transport), class T (signal transduction mechanisms), class Q (secondary metabolites biosynthesis, transport, and catabolism) and class V (defense mechanisms), which accounted for 10.2 and 9.7% (4.5%), 6.1 and 6.7% (2.6%), 6.6 and 5.9% (2.7%), 3.2 and 3.4% (1.9%), as well as 2.8 and 4.2% (1.6%), respectively (Figure 2C).

All 33 strains with complete genome sequences harbored the intact pentose phosphate pathway, fructose metabolism pathway, tricarboxylic acid cycle (TCA cycle), fatty acid, lipopolysaccharide, and peptidoglycan biosynthetic pathways. Only S. geniculata E119 lacked an intact glycolysis pathway owing to a deficiency in the glyceraldehyde 3-phosphate dehydrogenase (gapA) gene that converts glyceraldehyde-3P to glycerate-1,3P₂ (Figure 3; Supplementary Table S3). In terms of nitrogen metabolism, all 33 strains harbored the Amt family ammonia transporter gene, which is responsible for the uptake of ammonia. In particular, S. indicatrix MGMM10, S. nitritireducens 2001, S. maltophilia strains PSKL2, FDAARGOS_92, and FDAARGOS_507, and S. rhizophila strains DSM 14405^T, Bl2-2, JC1, PI-27, and QL-P4 lacked both nitrate reductase and nitrite reductase, while the other 23 strains only lacked nitrite reductase. This indicated that all these strains cannot utilize nitrate as a nitrogen source. The metabolism of sulfur also varies between strains, and only S. acidaminiphila strains T0-18, SPDF1, and T25-65, and S. nitritireducens 2001 had an intact sulfate reduction pathway, while the remaining 29 strains lacked the genes involved in the conversion of sulfate to sulfite. However, they harbored the genes to convert sulfite to sulfide. In terms of the biosynthesis of amino acids, all 33 strains of Stenotrophomonas can synthesize 20 types of amino acids. There was only one exception. S. geniculata E119 cannot synthesize histidine because it lacks histidinol-phosphatase. In addition, the phosphate-specific ABC transporter complex PstSABC and the two-component system PhoRB were identified in the core genome. This indicates that the strains of Stenotrophomonas can take up inorganic phosphate and regulate phosphate homeostasis.

Bacterial chemotaxis is an important prelude to metabolism, competition, symbiosis, infection and other ecological interactions in bacterial communities (Kato et al., 2008). The genomic analysis revealed that all 33 strains of Stenotrophomonas had the genes for flagellar proteins and the chemotaxis signaling system. In particular, all 33 strains of Stenotrophomonas harbored the flg operon (flagellar motif), the fli cluster, and flhAB that encodes the structural proteins for the flagella, as well as the motAB and filA genes that encode the quorum unit MotAB and the sigma factor FliA, respectively. The chemotaxis genes had the cheABYWR and mcp genes that encode the chemotaxis and methyl-accepting chemotaxis proteins, respectively. The two-component system DesK-DesR genes, which regulate the biosynthesis of fatty acid desaturase (Des), was found in all 33 strains (Beranová et al., 2010). In addition, the genes for DNA mismatch repair, glutathione and folate biosynthesis, peroxidase, and resistance to cationic antimicrobial peptides (CAMP) and β-lactams were also found in the core genome, which enables the strains of Stenotrophomonas to adapt to complex and variable environments more efficiently.

Secondary metabolites can enhance the environmental adaptability of the strains and provide them with an evolutionary advantage (O’Connor, 2015). A total of 149 smBGCs of 14 major types were predicted in the 33 Stenotrophomonas genomes, and each strain contained three to seven smBGCs (Supplementary Table S4). Among them, S. maltophilia strains JZL8 and ISMMS2 had the highest number of smBGCs with seven, while each strain of S. nitritireducens 2001, and S. rhizophila strains Bl2-2, JC1, and QL-P4 had only three smBGCs. In detail, 55 RiPP-like smBGCs were found in 32 strains of Stenotrophomonas (Supplementary Figure S1), and 22 of the clusters were similar to the entolysin BGCs, which encode a cyclic lipopeptide (CLP) that can be fatal to bacteria, fungi, and viruses by disrupting their membranes (Oni et al., 2019). The clusters responsible for the biosynthesis of aryl polyenes (APEs), which help bacteria to evade the host immune system (Lee et al., 2021a,b), were found in 31 strains except for S. maltophilia CYZ and S. rhizophila JC1. The NRP-metallophore/NRPS heterozygous clusters were found in 29 Stenotrophomonas strains, and all of them were similar to the 2,3-dihydroxybenzoylserine (DHBS) BGCs, which is a cluster that is responsible for the biosynthesis of catecholate siderophore enterobactin (Ahire et al., 2011). In addition, some smBGCs are strain-specific. For example, the cluster responsible for ranthipeptide, lanthipeptide-class-i and lanthipeptide-class-iii biosynthesis was only detected in S. maltophilia strains CSM2, ISMMS2, and JV3, respectively.

Carbohydrates are the primary source of carbon for most microbes. Carbohydrate active enzymes (CAZymes), a class of enzymes that is involved in the metabolism of carbohydrates, is not only involved in the biosynthesis and degradation of biopolymers but also in the formation of bacterial biofilms and the glycosylation of proteins and lipids (Benini, 2020). A total of 118 different CAZyme genes family were identified in the 213 Stenotrophomonas genomes, including 63 glycoside hydrolases (GHs), 21 glycosyltransferases (GTs), 11 carbohydrate esterases (CEs), 9 polysaccharide lyases (PLs), 8 auxiliary activities (AAs), and 6 carbohydrate-binding modules (CBMs) (Supplementary Figure S2). S. rhizophila CFBP 13503 was identified as harboring the most CAZyme genes at 124, while S. pictorum JCM 9942^T had the fewest (68). The CAZyme number of the strains from clinical sites (69 to118, average 106) was similar to that of the strains from the environmental strains (68 to 124, average 104).

In particular, this study explored the CAZymes involved in the degradation of complex polysaccharides. Lignocellulolytic enzymes can be subdivided into cellulases, hemicelluloses, and lignin-modifying enzymes. Most cellulases and hemicelluloses are members of the GHs family, while the lignin-modifying enzymes are generally in the AAs protein family (López-Mondéjar et al., 2016; Zhou et al., 2018). Among them, all 213 strains produced GH3-family enzymes that degraded cellulose, while only some of the 213 strains harbored GH5- (45 strains), GH8- (44 strains), GH1- (19 strains), GH9- (4 strains) and GH74-family (5 strains) enzymes involved in the degradation of cellulose. The hemicellulose hydrolases produced by these species primarily include GH2 (211 strains), GH31 (209 strains), GH10 (162 strains), GH43 (100 strains), GH16 (13 strains) and GH39 (2 strains) families. Lignin-modifying enzymes were primarily represented by the families for laccase-like multi-copper (AA1) and lignin-modifying peroxidases (AA2) (Ma et al., 2021). Genes for the AA1 family were found in all 213 genomes, while AA2 was not found in any of them. In addition, the GH18, GH19, and GH20 families, which are responsible for the degradation of chitin (Swiontek Brzezinska et al., 2013), were detected in 186, 61, and 210 strains, respectively. The analysis revealed that the strains of Stenotrophomonas produced a large number of CAZymes involved in the degradation of polysaccharides, which indicates that these strains can easily obtain carbon and energy from their environment.

To survive in environments, particularly those that are extreme, microbes have evolved the ability to surmount oxidative stress and osmotic pressure (Zhao et al., 2023). Among these, oxidative stress is the result of an imbalance between the production of reactive oxygen species (ROS) and the ability of biological systems to detoxify them (Chautrand et al., 2022). Bacteria use two primary mechanisms to detoxify the ROS. One is to scavenge the ROS using superoxide dismutase (SOD) and catalase (CAT) (Arts et al., 2015). The sodB and sodC genes, which are responsible for catalyzing the decomposition of superoxide anion (O₂⁻) to O₂ and hydrogen peroxide (H₂O₂), were found in all 213 strains of Stenotrophomonas (Figure 4). The katE, ahpC and bcp genes, which are responsible for the conversion of H₂O₂, were also found in all the strains. The gpx gene encodes glutathione peroxidase (GPX), which catalyzes the reduction of many oxidants, including H₂O₂ and peroxynitrite (ONOO⁻) (Panday et al., 2020), was found in all the strains tested except for S. maltophilia HZ34. Another is the timely repair of the cysteine and methionine residues that have been damaged by ROS oxidation. The disulfide isomerase (DsbC) and methionine sulfoxide reductase (Msr) systems, which can repair oxidized cysteine and methionine residues, respectively (Darby et al., 1998; Grimaud et al., 2001). The dsbC, msrA and msrB genes were found in all 213 strains.

Bacteria usually adopt two main strategies to maintain their osmotic balance (Gunde-Cimerman et al., 2018). One is to take up and accumulate a large amount of potassium (Vyrides and Stuckey, 2017). All 213 strains of Stenotrophomonas were found to have K⁺ uptake transporter proteins genes, including the Kdp-ATPase system (kdp), the potassium efflux system (kef), and the potassium transport system gene (kup) (Sleator and Hill, 2002). Another strategy relies on the biosynthesis and accumulation of compatible solutes (Figure 3), including sugars, such as trehalose; amino acids, such as proline and glutamate; and their derivatives, such as betaine (Zhao et al., 2023). Glycine betaine, a common compatible solute, can be synthesized by choline dehydrogenase (BetA) and betaine-aldehyde dehydrogenase (BetB) (Lamark et al., 1991). The betA gene was detected in 211 strains and was only absent in S. rhizophila GN_RF3 and S. maltophilia HZ34. In contrast, the betB gene was identified in 208 strains and only absent in five strains of S. acidaminiphila. Trehalose can be produced from UDP-glucose via the OstA-OstB pathway (Paul et al., 2008). The ostA and ostB genes were only absent in S. maltophilia HZ34. In addition, the glutamate (gdhA, gltBD and glnA) and proline (proABC) biosynthetic genes were detected in the genomes of all 213 strains of Stenotrophomonas. These results indicate that all the strains of Stenotrophomonas can eliminate ROS and tolerate osmotic stress, which endows them with the ability to survive more effectively in extreme environments.

Antibiotics are used extensively to treat bacterial infections in agriculture and animal husbandry. Bacteria usually harbor many antibiotic resistance genes (ARGs) to survive (Zhuang et al., 2021). The number of ARGs per strain of Stenotrophomonas ranged from 3 to 23. Among them, Stenotrophomonas sp. strain SH_67_7 harbored the most ARGs, and S. nitrodeiducens CN18_13_7_16_R1_B_68_91 harbored the least (Supplementary Table S5). There were somewhat more ARGs in the clinical strains (average 14 per strain) than in the environmental strains (average 11 per strain), and S. maltophilia (8–23), S. geniculata (9–20), and S. pavanii (15–16) species had dramatically higher number of ARGs than the other species (4–13). The antibiotic efflux pumps genes, including the resistance-nodulation-cell division (RND) type and the major facilitator superfamily (MFS) type, were identified in all the strains of Stenotrophomonas. rsmA, an RND efflux pump gene involved in the resistance to fluoroquinolones, diaminopyrimidines and phenol antibiotics, was found in all 213 strains except for S. geniculata NWUBe21 and S. rhizophila DE0483. smeDEF, an RND efflux pump gene cluster involved in the resistance to macrolide, fluoroquinolone, tetracycline, and phenol antibiotics (Alonso and Martínez, 2000), was identified in 201 strains of Stenotrophomonas. Another RND efflux pump gene cluster smeABC, which is involved in resistance to aminoglycoside, β-lactam, and fluoroquinolone antibiotics (Li et al., 2002), was found in most of the clinical strains (124) and a minority of the environmental strains (28).

The resistance of bacteria to aminoglycosides (AGs) is primarily caused by AG-modifying enzymes (AMEs), which include O-phosphotransferases (APHs), N-acetyltransferases (AACs) and O-adenyltransferases (ANTs) (Wright, 1999). The aph gene was identified in 134 clinical strains and 53 environmental strains, while the aac gene was found in 59 clinical and 17 environmental strains. In particular, the aph(9)-Ic, aph(3′)-IIc, aph(3″)-Ib, aph(6)-Id, aph(3′)-IIa, aph(3′)-VIa, and aph(6)-Ic genes that encode variants of APH were identified in 185, 183, 5, 4, 1, 1 and 1 strains, respectively, while the aac(6′)-Iz, aac(6′)-Ib8, aac(6′)-Iak, aac(6′)-Iap and aac(6′)-31 genes that encode variants of AAC were identified in 62, 12, 8, 3 and 1 strains, respectively. The ant(2″)-Ia gene was only found in S. maltophilia CV_2005, S. maltophilia 476BLSM_MAD, S. maltophilia 212 and S. maltophilia Smal28. In the Stenotrophomonas resistome, the highest numbers of types of genes for resistance were associated with aminocoumarin, phenicol, and tetracycline antibiotics, with 25, 14, and 12, respectively, while sulfonamide, monobactam, and mupirocin-like had the fewest types of resistance genes. The results indicated that the strains of Stenotrophomonas have multidrug resistance that can be adapted to respond to antibiotic pressure in the external environment.

Most heavy metals are lethal in the environment (Gadd, 2010). The CopAB proteins are responsible for resistance to copper. In particular, CopA can bind copper ions in the periplasm to sequester the copper, while CopB is a transmembrane protein that can serve as an exporter of copper (Ge et al., 2021). The copA gene was identified in all 213 strains of Stenotrophomonas, and the copB gene was found in 212 strains (Figure 4). The CzcABC-type efflux pump proteins are involved in exporting cobalt, zinc and cadmium from the cells (Hussain et al., 2022). All 213 strains of Stenotrophomonas also harbored czcA, czcB and czcC genes. The ars operon ensures that various bacteria are resistant to arsenic. Some bacteria can reduce As(V) (arsenate) to As(III) (arsenite), which is catalyzed by arsenate reductase (ArsC). The As(III) that is produced can be directly pumped out of the cell via ArsB. In addition, MAs(III) (methylarsenite) can be oxidized to the less toxic MAs(V) (methylarsenate) via ArsH (Hao et al., 2021). The arsC, arsB and arsH genes were identified in 213, 209 and 194 genomes of Stenotrophomonas, respectively. The chrR gene encodes chromate reductase, which catalyzes the conversion of soluble and toxic Cr(VI) to insoluble and less toxic Cr(III), respectively (Baldiris et al., 2018), was found in all strains. In contrast, the chrA gene that encodes the chromate efflux transport protein, which pumps chromate out of the cytoplasm (Díaz-Pérez et al., 2007), was found in 55 strains. These genes in strains of Stenotrophomonas endow the bacteria with tolerance to heavy metals in the environment.

MGEs are the direct evidence of the evolution and adaptation of microbial populations through the horizontal transfer of genes (Sobecky and Hazen, 2009). The MGEs in microbes usually include prophages, GIs, and insertion sequence (IS). A total of 637 prophage regions, including 79 questionable, 354 incomplete, and 204 intact prophages, were detected in these Stenotrophomonas genomes, thus, indicating that inducible or transferable functional prophages are widely distributed in strains of Stenotrophomonas. The number of prophages in each strain ranged from one to nine. Among them, 204 intact prophages were found in 60.6% (129/213 strains) of the strains of Stenotrophomonas, and they ranged from one to five intact prophages per strain (Supplementary Table S6). In particular, the largest proportion was composed of Verrucomicrobia phage P862 (NC_029047) followed by Burkholderia phage phi1026b (NC_005284), Vibrio phages VHML (NC_004456) and vB_VpaM_MAR (NC_019722). The major proteins of the prophages that were detected were the structural proteins of tail, capsid, portal, and head, and the functional enzymes terminase, integrase and lysin.

A total of 6,663 GIs were predicted in the 213 Stenotrophomonas strains. The number of GIs per strain ranged from 17 to 60 (Supplementary Table S1). S. rhizophila JC1 contained the highest number of GIs, while S. chelatiphaga DSM 21508^T contained the fewest (17). There are many vital functional protein genes in these GIs, including kinases (201 strains), MFS transporters (192 strains), GNAT family N-acetyltransferase (171 strains), ATP-binding proteins (166 strains), response regulator transcription factor (156 strains), efflux RND transporter (118 strains), siderophore receptor (112 strains), RNA polymerase (111 strains), nucleotidyl transferase (101 strains), DNA polymerase III subunit beta (70 strains), DNA replication/repair protein RecF (64 strains), lysozyme (54 strains), type II toxin-antitoxin system (52 strains), CusA/CzcA family heavy metal efflux RND transporter (68 strains), copper resistance protein (62 strains), arsenical resistance protein (40 strains), mercury reductase (31 strains) and chromate efflux transporter (25 strains). These genes are involved in the physiology at the transcriptional and translational levels, as well as confer pathogenicity, drug and heavy metal resistance to the strain (Supplementary Table S7). In addition, some Gls include the genes that encode integrases (205 strains), transposases (185 strains) and recombinases (163 strains), which may mediate the movement of GIs among the host bacteria. Thus, the prophages and GIs enhanced the genetic diversity of the Stenotrophomonas strains and enabled them to rapidly adapt to diverse ecological niches.

Colonization of the rhizosphere by bacterial strains is the first and the foremost step. Genes involved in motility, chemotaxis, adhesion and biofilm formation are thought to contribute to this colonization (Kandel et al., 2017). The two site-specific tyrosine recombinases XerC and XerD are associated with competitive colonization on the root surface (Martínez-Granero et al., 2005). The two genes xerC and xerD were both found in the genomes of all 213 strains of Stenotrophomonas (Figure 4). The SmeDEF efflux pump, which is responsible for the microbial resistance to quinolones, has been found to be involved in the endophytic colonization of plant roots (García-León et al., 2014). The smeDEF gene was identified in all 213 strains. In addition, the thuA gene, which is involved in the utilization of trehalose, enhances the ability of the strain to colonize plants at the early stages (Pinski et al., 2020). The thuA gene was detected in 15 strains, including 10 S. rhizophila strains, S. bentonitica DSM 103927^T, S. bentonitica VV6, S. chelatiphaga DSM 21508^T, S. tumulicola JCM 30961^T and Stenotrophomonas sp. AP15.006.

As a physical barrier, biofilm can protect the embedded bacteria. Thus, the ability of bacteria to form biofilms determines their ability to colonize the root surface. Cellulose is an important component of biofilms in some species. The bcsA and wssD genes are two genes that are responsible for the biosynthesis of cellulose. Their inactivation significantly reduced the formation of biofilm and decreased the ability of strains to colonize the plant host (Barak et al., 2007; Monteiro et al., 2012). However, these two genes were only found in 43 strains of Stenotrophomonas (21 environmental and 22 clinical strains). In addition, lipopolysaccharides (LPS) and exopolysaccharides (EPS) facilitate the attachment of bacterial cells to the root surface (Kandel et al., 2017). The lapA gene, a key gene responsible for the assembly of LPS, was found in 210 strains. Rhamnose is an important subcomponent of LPS. The rfbABCD gene, a cluster that is responsible for the biosynthesis of rhamnose (Balsanelli et al., 2010), was found in all the strains of Stenotrophomonas. The exoD gene that encodes the protein for the biosynthesis of EPS (Reed and Walker, 1991) was found in 212 strains. The gumD gene, which is also responsible for the catalysis of the biosynthesis of EPS, (Meneses et al., 2011) was found in the genomes of 50 strains of Stenotrophomonas (33 environmental and 17 clinical strains). The results of genomic analyses showed that all the strains of Stenotrophomonas have a range of biofilm formation genes that can promote their ability to colonize plants. In addition, the ability to form biofilms has also been shown to be one of the main pathogenesis-related virulence factors (Wall et al., 2019). Inhibiting the expression of virulence-associated genes is reportedly correlated with the attenuation of the biofilm formation ability of S. maltophilia (Kim et al., 2018).

The formation of biofilm and cell adhesion are usually regulated by QS at the bacterial community level. The diffusible signal factor (DSF) is a QS molecule that can regulate the genes related to plant colonization, such those involved in chemotaxis, motility, the formation of biofilm and production of the multidrug efflux pump (Alavi et al., 2013a). The rpf gene cluster, which encodes all the components of the DSF system, was detected in 213 strains of Stenotrophomonas (Figure 4). Carbamoyl phosphate synthase, which is encoded by the carAB genes, can be used to degrade DSF (Newman et al., 2008). The carA gene was found in 212 strains (S. maltophilia HZ34 lacked), while the carB gene was found in 211. S. maltophilia HZ34 and Stenotrophomonas sp. SH_67_7 lack this gene. pcoI is a key gene related to the population sensing signals. Its deletion led to significant deficiencies in the formation of biofilm at the wheat rhizosphere (Wei and Zhang, 2006). The pcoI gene was only found in seven environmental strains, including five strains of S. acidaminiphila, and S. nitritireducens SCN18_13_7_16_R1_B_68_91 and Stenotrophomonas sp. MES1.

Some prevalent bacterial transcriptional regulators are also associated with plant-rhizobacterial interactions. The sigma-28 factor encoded by the filA gene is involved in the regulation of the expression of flagellin, chemotaxis, and motility-related genes (Yi et al., 2017). The filA gene was identified in the genomes of all the strains of Stenotrophomonas. The greA gene, another important transcription factor, determines the ability of host to adapt to hyperosmolarity and salt stress. The strains that lack this gene cannot establish an effective symbiotic relationship with plants (Nogales et al., 2002). The greA gene was found in 212 strains and was only absent from S. maltophilia HZ34. Cyclic diguanylate (c-di-GMP), a universal secondary messenger in bacteria, is involved in the regulation of a variety of physiological functions, including cell differentiation, biofilm formation, motility, adhesion, colonization of host tissues and the generation of pathogenic factors (Pinski et al., 2020). The biosynthesis of c-di-GMP is catalyzed by the gene that encodes diguanylate cyclase, which was found in 205 strains of Stenotrophomonas. The results indicate that strains of Stenotrophomonas have a range of transcriptional regulators that are related to the colonization of plants.

Polyamines, including putrescine, spermidine, spermine and cadaverine, may be involved in the promotion and protection of plant growth (Couée et al., 2004; Xie et al., 2014). Bacteria could synthesize putrescine from arginine step by step catalyzing by arginine decarboxylase (SpeA), agmatine deiminase (AguA), and N-carbamoyl-putrescine amidase (AguB) (Burrell et al., 2010; Michael, 2016) (Figure 3). The speA, aguA, and aguB genes were present in 213, 212, and 212 strains, respectively. Only S. maltophilia HZ34 lacked both the aguA and aguB genes (Figure 4). The speE gene, which is responsible for the conversion of putrescine to spermidine (Lee et al., 2012), was found in all the strains of Stenotrophomonas. The MetK and SpeD enzymes are responsible for the conversion of methionine to S-adenosyl-methioninamine (dcSAM), which is required for the biosynthesis of spermine (Xie et al., 2014). The metK and speD genes are found in all 213 strains. In addition, lysine decarboxylase (LdcC) catalyzes the conversion of L-lysine to cadaverine (Liu et al., 2022). The ldcC gene was found in the genomes of 203 strains of Stenotrophomonas. The presence of these genes suggests that strains of Stenotrophomonas can produce multiple types of polyamines.

IAA and cytokinin (CK) are the two most common phytohormones that influence various traits of plant growth and development, including cell division and elongation, seed germination, fruit development and the delay of senescence (Grossmann, 2009; McSteen, 2010; Akhtar et al., 2020). There are five different pathways for the biosynthesis of IAA from tryptophan. Among them, the indole-3-acetamide pathway (IAM), the indole-3-pyruvic acid pathway (IPA/IPyA) and the tryptamine pathway (TAM) are the most common pathways in bacteria (Zhang et al., 2019; Tang et al., 2023). The vital genes for the biosynthesis of tryptophan, including trpAB, trpC, trpD, trpEG, and trpF, were present in all 213 genomes of Stenotrophomonas. For the IPA pathway, the gene that is responsible for the conversion of tryptophan to IAAld was not detected in all 213 genomes, but the gene for indole-3-acetaldehyde dehydrogenase (AldA) that converts IAAld to IAA was found in all these genomes (Figure 3). In addition, the iroA gene that encodes indole-pyruvate ferredoxin oxidoreductase (IorA), which can directly convert IPA to IAA (Imada et al., 2017), was identified in 207 strains. As the IAM pathway, all 213 strains lacked the tryptophan monooxygenase that catalyzes the initial conversion of tryptophan to IAM, but the amidase gene responsible for the conversion of IAM to IAA was detected in all 213 genomes. For the TAM and TSO (tryptophan side-chain oxidase) pathway, only the aldA gene of the final step was detected, but the gene responsible for the conversion of tryptophan to IAAld was not detected.

The miaA gene encodes a tRNA dimethylallyltransferase that is responsible for the conversion of dimethylallyl diphosphate to N₆-(dimethylallyl) adenosine (iPR). The miaB gene, which encodes tRNA-2-methylthio-N(6)-dimethylallyladenosine synthase, converts iPR to 2-methylthio-N₆-(dimethylallyl)adenosine (2MeSiPR). Both genes are involved in the biosynthesis and conversion of CK (Nascimento et al., 2020) and were identified in all the Stenotrophomonas genomes except for S. rhizophila IS26 and Stenotrophomonas SH_67_7, which both lacked miaB.

Phosphorus (P) is an essential macronutrient for plant growth and development. Although there are adequate amounts of P in the soil, most of the organic and inorganic phosphate that is found in the soil is not directly available to plants. Many bacteria that are associated with plants can solubilize P into a form that can be utilized by plants (Sharma et al., 2013). Organic phosphates are primarily degraded by the enzyme acid phosphatases, alkaline phosphatases and phytase that are produced by bacteria (Sharma et al., 2013), while inorganic phosphates can be solubilized by the production of organic acids, such as gluconic acid, citric acid, malic acid, and succinic acid (Vassilev et al., 2006). There are two categories of phosphatases that are based on their optimal pH, including acid phosphatase and alkaline phosphatase, which both dephosphorylate phosphate ester or the phosphoric anhydride bonds in organic compounds (Jorquera et al., 2008). The acid phosphatase and alkaline phosphatase genes were found in 213 and 197 strains, respectively (Figure 4). The phytase encoded by phyA is responsible for the catalysis of the degradation of phytate to release P, and this gene was identified in 212 strains. S. nitritireducens SCN18_13_7_16_R1_B_68_91 does not produce phyA. In addition, the gdh gene encodes a phosphate starvation-inducible glucose dehydrogenase (GDH), which may be involved in the solubilization of mineral phosphate (Sharma et al., 2005). This gene was identified in 205 strains of Stenotrophomonas. Among them, five S. acidaminiphila strains, S. pictorum JCM 9942^T, Stenotrophomonas sp. MES1 and S. maltophilia HZ34 lacked gdh gene. An exopolyphosphatase and an inorganic pyrophosphatase are encoded by the ppx and ppa genes, respectively. They catalyze the hydrolysis of inorganic pyrophosphate. Both genes were found in all 213 genomes of Stenotrophomonas. These genes enable these strains of Stenotrophomonas to solubilize soil phosphate, which provides adequate amounts of P for their own metabolism and that of their host plant.

Similar to P, iron is an essential nutrient for plant and microbial growth, but its bioavailability is limited owing to the low solubility of ferric oxide (Fe³⁺) ions (Behnke and LaRoche, 2020). Siderophores, iron chelators that are produced by microorganisms, can convert the iron into soluble complexes that can be utilized by plants (Ahmed and Holmström, 2014). The comparative analysis revealed that most strains of Stenotrophomonas possessed the genes to biosynthesize catecholate siderophores. In particular, 26 and 189 strains possessed the catecholate siderophore biosynthetic genes fiu and cirA, respectively, and 171 and 187 strains had the biosynthetic genes entAFH and fepA, respectively, for the catecholate siderophore enterobactin (Patzer et al., 2003; Pakarian and Pawelek, 2016). In Gram-negative bacteria, the uptake and transfer of iron ions (Fe³⁺) by siderophores is powered by the Ton complex (TonB-ExbB-ExbD) (Faraldo-Gómez and Sansom, 2003). The tonB, exbB and exbD genes were present in all 213 genomes (Figure 4). In addition, many bacteria can take up ferrous iron (Fe²⁺) using the Feo system under anaerobic and acidic environments (Lau et al., 2016). The Feo system is primarily composed of FeoA and FeoB, (Kalidasan et al., 2018). The feoB gene was detected in all 213 strains of Stenotrophomonas except for S. maltophilia HZ34, while the feoA gene was not found in Stenotrophomonas SH_67_7 and S. maltophilia HZ34. The results suggest that the strains of Stenotrophomonas can enhance the availability of Fe³⁺ in an environment deficient in iron, which benefits both their survival and the growth of plant.

The biological control of microbes in the rhizosphere provides their hosts with natural protection from soilborne phytopathogens (Pandit et al., 2022). The production of cell wall-degrading enzymes by rhizobacteria is one of the primary mechanisms that they use to destroy pathogens. The chitinase encoded by the chiA gene can degrade the chitin in the cell walls of pathogenic fungi (Kharade and McBride, 2014). The chiA gene was detected in 204 genomes of Stenotrophomonas. In addition, the chitin-binding proteins (CBPs), encoded by the cbpD gene, can increase the efficiency of some specific chitinases (Manjeet et al., 2013). The cbpD gene is found in 183 strains of Stenotrophomonas (Figure 4). Proteases and lipases also play an important role in biocontrol (Friedrich et al., 2012). Among them, protease genes were identified in 205 strains of Stenotrophomonas, while lipase genes were identified in 200 strains. Six strains lacked the chitinase, protease, and lipase genes, including S. acidaminiphila T0-18, S. acidaminiphila T25-65 and S. nitritireducens SCN18_13_7_16_R1_B_68_91 isolated from bioreactors, S. pictorum JCM 9942^T isolated from the soil and S. acidaminiphila SPDF1 isolated from cell culture media.

Polyhydroxybutyrate (PHB), the most abundant and well-characterized polymer of bacterial polyhydroxyalkanoate (PHA) (Jendrossek, 2009), has proven to have both antibacterial and antifungal activities (Ma et al., 2019). The biosynthesis of PHB is a three-step process that begins with acetyl-CoA acetyltransferase (PhaA), which catalyzes the condensation of two molecules of acetyl-CoA to acetoacetyl-CoA; this compound is then reduced to 3-hydroxybutyryl-CoA by acetoacetyl-CoA reductase (PhaB) and finally catalyzed by PhaC (poly(3-hydroxyalkanoate) polymerase) to produce PHB (Jendrossek and Pfeiffer, 2014). The three genes are found in the genomes of all strains of Stenotrophomonas. The keratinases KerSMD and KerSMF have broad substrate specificity and can be used as pesticides (Gupta and Ramnani, 2006; Yue et al., 2011). Both genes kerSMD and kerSMF were detected in 205 strains of Stenotrophomonas. The results suggest that Stenotrophomonas may be a useful tool for the integrated management of various plant pathogens and pests.