ORIGINAL RESEARCH published: 03 June 2020 doi: 10.3389/fmicb.2020.01160 Genetic Variants of the DSF Quorum Sensing System in Stenotrophomonas maltophilia Influence Virulence and Resistance Phenotypes Among Genotypically Diverse Clinical Isolates Daniel Yero 1,2 , Pol Huedo 1,2 , Oscar Conchillo-Solé 1,2 , Sònia Martínez-Servat 1,2 , Uwe Mamat 3 , Xavier Coves 1,2 , Ferran Llanas 1 , Ignasi Roca 4 , Jordi Vila 4 , Ulrich E. Schaible 3 , Xavier Daura 1,5* and Isidre Gibert 1,2* Edited by: Jose L. Martinez, 1 Institut de Biotecnologia i de Biomedicina (IBB), Universitat Autònoma de Barcelona (UAB), Barcelona, Spain, Consejo Superior de Investigaciones 2 Departament de Genètica i de Microbiologia, Universitat Autònoma de Barcelona (UAB), Barcelona, Spain, 3 Cellular Científicas (CSIC), Spain Microbiology, Priority Research Area Infections, Research Center Borstel – Leibniz Lung Center, Borstel, Germany, 4 Reviewed by: Department of Clinical Microbiology-ISGlobal, Hospital Clínic, Universitat de Barcelona, Barcelona, Spain, 5 Catalan Ya-Wen He, Institution for Research and Advanced Studies (ICREA), Barcelona, Spain Shanghai Jiao Tong University, China Pablo Vinuesa, National Autonomous University The pathogenicity of Stenotrophomonas maltophilia is regulated in part by its quorum of Mexico, Mexico sensing (QS) system. The main QS signaling molecule in S. maltophilia is known as *Correspondence: diffusible signal factor (DSF), and the rpf gene cluster is responsible for its synthesis Xavier Daura and perception. Two cluster variants have been previously described, rpf-1 and rpf-2, xavier.daura@uab.cat Isidre Gibert which differ basically in the conditions under which DSF is produced. Here, correlations isidre.gibert@uab.cat between the rpf variant and antibiotic susceptibility, LPS electrophoretic profiles and virulence-related phenotypes were evaluated for a collection of 78 geographically and Specialty section: This article was submitted to genetically diverse clinical strains of S. maltophilia. In general there were associations Antimicrobials, Resistance between previously established genogroups and the genetic variant of the rpf cluster. and Chemotherapy, a section of the journal However, only few genotype-phenotype correlations could be observed. Resistance to Frontiers in Microbiology the β-lactam antibiotics ceftazidime and ticarcillin was associated with strains carrying Received: 03 February 2020 the rpf-1 variant, whereas strains of variant rpf-2, particularly those of genogroup C, Accepted: 06 May 2020 showed higher resistance levels to colistin. Strains of variant rpf-2 were also significantly Published: 03 June 2020 more virulent to Galleria mellonella larvae than those of rpf-1, most likely due to an Citation: Yero D, Huedo P, increased ability of rpf-2 strains to form biofilms. A comparative genomic analysis Conchillo-Solé O, Martínez-Servat S, revealed the presence of proteins unique to individual genogroups. In particular, the Mamat U, Coves X, Llanas F, Roca I, Vila J, Schaible UE, Daura X and strains of genogroup C share an operon that encodes for a new virulence determinant Gibert I (2020) Genetic Variants of the in S. maltophilia related to the synthesis of an alternative Flp/Tad pilus. Overall, this DSF Quorum Sensing System study establishes a link between the DSF-based QS system and the virulence and in Stenotrophomonas maltophilia Influence Virulence and Resistance resistance phenotypes in this species, and identifies potential high-risk clones circulating Phenotypes Among Genotypically in European hospitals. Diverse Clinical Isolates. Front. Microbiol. 11:1160. Keywords: quorum sensing (QS), diffusible signal factor (DSF), MLST (multilocus sequence typing), rpf cluster, doi: 10.3389/fmicb.2020.01160 biofilm, antibiotic resistance Frontiers in Microbiology | www.frontiersin.org 1 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes INTRODUCTION (Pompilio et al., 2011). S. maltophilia produces a wide range of extracellular enzymes, including lipases, proteases, chitinases, Stenotrophomonas maltophilia are metabolically and genetically esterases, DNases, and RNases, contributing to its colonization diverse Gram-negative bacilli belonging to the γ-proteobacteria and virulence in different organisms (Ryan et al., 2009; class that inhabit a wide range of environmental niches, mainly Trifonova and Strateva, 2019). Furthermore, it is well known in association with plant rhizospheres (Ryan et al., 2009; Brooke, that S. maltophilia synthesizes diverse O-specific polysaccharide 2012). Although S. maltophilia are thought to be low-grade structures that constitute immunodominant antigens of the pathogens and a rather uncommon cause of bacterial disease lipopolysaccharide (LPS) molecule and may contribute to the in humans, they are increasingly recognized as significant pathogenicity of the bacterium (Waters et al., 2007; Knirel, opportunistic pathogens in healthcare settings worldwide, being 2011). A diffusible signal factor (DSF)-mediated quorum sensing one of the ten most common organisms found in respiratory (QS) system, encoded by the rpf gene cluster, coordinates the tissue samples taken from intensive care unit (ICU) patients regulation of a number of phenotypes in S. maltophilia, such in Europe (Zuschneid et al., 2007; Gherardi et al., 2015). as biofilm formation, secretion of proteases, virulence, and Several nosocomial outbreaks of S. maltophilia infection and/or tolerance to a range of antibiotics (Fouhy et al., 2007; Ryan colonization have been reported (Denton and Kerr, 1998), et al., 2009; Trifonova and Strateva, 2019). Two genetic variants and the global prevalence and incidence of these bacteria of the rpf gene cluster (rpf -1 and rpf -2) divide the strains into as human pathogens have increased significantly during the two phenotypically and genotypically distinct subpopulations last decade (Chang et al., 2015; Nayyar et al., 2017; Rutter (Huedo et al., 2014). Only rpf -1 strains produce detectable DSF, et al., 2017), particularly in cystic fibrosis patients (Hatziagorou which correlates with their ability to control biofilm formation, et al., 2019). The most common clinical manifestations of motility and virulence (Huedo et al., 2014, 2015). However, S. maltophilia infections include pneumonia, bacteremia, wound no clear link between the presence of a specific rpf variant or urinary-tract infections (Looney et al., 2009; Gales et al., and the isolation source of the strains has been found so far 2019), with crude mortality rates ranging from 14 to 69% in (Lira et al., 2017). patients with bacteremia, being highest among patients receiving An additional important trait of S. maltophilia strains is inappropriate initial antimicrobial therapy (Paez and Costa, their high genetic diversity (Gherardi et al., 2015) that facilitates 2008; Falagas et al., 2009; Kim et al., 2019). Most important, adaptation to changing environmental conditions. Numerous S. maltophilia show low susceptibility to many antibiotics, genotypic studies of both clinical and environmental strains including those commonly used to treat infections by Gram- have revealed considerable variability among S. maltophilia negative opportunistic pathogens (Sánchez, 2015). Not only isolates (Nicoletti et al., 2011; Pompilio et al., 2011, 2016; have they a variety of intrinsic resistance mechanisms affecting Ochoa-Sánchez and Vinuesa, 2017; Patil et al., 2018; Rizek almost all antibiotic classes but they are also able to acquire et al., 2018; Steinmann et al., 2018; Bostanghadiri et al., 2019; new resistances via horizontal gene transfer and mutations. The Mojica et al., 2019). Due to these large genetic divergences, phenomenon of heteroresistance has also been described in and because previously proposed species are recognized to be S. maltophilia as a mechanism to withstand antibiotic treatment closely related to S. maltophilia, it is increasingly accepted (Martínez-Servat et al., 2018). For these reasons, this species to present it as the S. maltophilia complex (Svensson-Stadler has been classified as one of the leading multidrug resistant et al., 2012; Patil et al., 2018; Vinuesa et al., 2018). PCR- (MDR) organisms in hospital settings (Brooke, 2014) and has based multilocus sequence typing (MLST) and whole-genome now been included in the global priority list of the top 10 resistant sequencing (WGS) have clearly shown the existence of several microorganisms (TOTEM) isolated in ICUs (Rello et al., 2019). genomic or phylogenetic groups in the S. maltophilia complex, The pathogenesis of S. maltophilia infections involves several and have helped to clarify the taxonomic status of the species virulence factors, including extracellular enzymes, bacterial (Hauben et al., 1999; Kaiser et al., 2009; Ochoa-Sánchez and motility and biofilm formation (Trifonova and Strateva, 2019). Vinuesa, 2017; Steinmann et al., 2018; Gröschel et al., 2020; The ability to grow in matrix-enclosed biofilms is an important Mercier-Darty et al., 2020). Even strains isolated from the same virulence-related trait of pathogenic bacteria. Among other patient (Pompilio et al., 2016; Chung et al., 2017) or patients functions, biofilms facilitate the development of bacterial within the same hospital (Valdezate et al., 2004) can belong to persistence within the host and increase resistance against the distant phylogenetic groups with different phenotypes, probably host immune response (González et al., 2018) and antimicrobials due to the emergence of adaptive mutations as a result of selective (Stewart, 2015), including the last-resort antibiotic colistin pressure in the clinical environment or inside the host. In (Mulcahy et al., 2008; Kim et al., 2015; Chua et al., 2016). addition to interstrain variability, some isolates show cell-to-cell Thus, biofilm formation is regarded as a form of phenotypic variation within a clonal population as a mechanism to survive resistance (Olivares et al., 2013). S. maltophilia forms biofilms on environmental fluctuations (Abda et al., 2015). The complex a wide range of biotic and abiotic surfaces, such as indwelling genotypic and phenotypic heterogeneity between and within medical devices (Di Bonaventura et al., 2007). Still, very little S. maltophilia isolates hampers not only the effectiveness of is known about the genetic mechanisms that control biofilm antimicrobial therapies but also investigations into the virulence formation or virulence in these bacteria, and it is not clear and resistance determinants of this species. whether there is an intrinsic difference in biofilm formation To deepen the knowledge and understanding of the resistance among genomically diverse environmental and clinical isolates and virulence phenotypes displayed by S. maltophilia, in a way Frontiers in Microbiology | www.frontiersin.org 2 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes that is consistent with this high-diversity context, we have studied jackhmmer (Johnson et al., 2010) were used to identify homologs a panel of genetically diverse clinical S. maltophilia isolates with of the protein sequences using several databases. Prediction focus on the two genetic variants of the DSF-mediated QS system. of operons was done using the online software and database Genetic characterization was performed using MLST and rpf OperonDB (Pertea et al., 2009). typing, together with the analysis of shared and unique proteins. Phenotypic analyses included the determination of antibiotic Genotypic Characterization and resistance profiles, biofilm formation, protease secretion, cell Phylogenetic Analysis motility and LPS banding patterns. Moreover, the virulence Genotypic characterization was basically performed on the basis of the isolates was assessed using two different non-vertebrate of the MLST scheme for S. maltophilia (Kaiser et al., 2009). model hosts. To our knowledge, a systematic analysis of the In addition, determination of the rpf variant was included population structure based on the two genetic variants of the in the genotyping as previously described (Huedo et al., DSF-mediated QS system and its relationship with the resistance 2014). The detailed MLST procedure was obtained from the and virulence phenoptypes of clinical S. maltophilia strains has S. maltophilia MLST database3 . The allele numbers for each not yet been carried out. locus and the sequence type (ST) were determined online at the S. maltophilia MLST database. A combination of the allelic sequences of the seven genes yielded the allelic profile for MATERIALS AND METHODS each isolate. Concatenated sequences of the seven housekeeping genes for each strain were aligned, and a phylogenetic tree Bacterial Isolates and Growth Conditions was constructed using MEGA version 7.0.21 (Kumar et al., The strain collection used in this study consisted of 78 clinical 2016) based on the neighbor-joining algorithm (Tamura–Nei S. maltophilia isolates (Huedo et al., 2014) collected from point model). Phylogenetic trees were visualized and annotated using prevalence studies in ICUs of geographically distant European the interactive web platform iTOL v3 (Letunic and Bork, hospitals. The strain panel included isolates from sputum, blood 2016). To divide strains into clonal groups according to their and swabs from surgical wound, oropharynx, perineum, vascular MLST allelic profiles and to infer the population structure, ulcer, decubitus ulcer, or a bronchoaspirate (Supplementary BURST analysis (Feil et al., 2004) was used on the basis of Table S1). All clinical strains were isolated from different patients, five shared loci. except for five patients with two isolates each (Supplementary Table S1). Three strains of S. maltophilia, namely K279a, D457, Analysis of Antimicrobial Susceptibility and ATCC13637, were included as references. ATCC13637 was The minimum inhibitory concentrations (MICs) of antibiotics purchased from the American Type Culture Collection. Unless were determined using Etest strips following the manufacturer’s otherwise stated, cultures were routinely grown at 37◦ C in LB instructions (bioMérieux, Madrid, Spain). Briefly, a 0.5- broth (10 g/L tryptone, 5 g/L yeast extract and 10 g/L NaCl) McFarland suspension was used to grow a confluent bacterial on a rotary shaker at 200 rpm. The species of all isolates was lawn on Mueller–Hinton (MH) agar plates (bioMérieux). Seven confirmed by matrix-assisted laser desorption/ionization time- antibiotics with proven in vitro activity against S. maltophilia of-flight (MALDI-TOF) mass spectrometry (data not shown). (Falagas et al., 2008; Wang et al., 2014; Chang et al., 2015) were used: ceftazidime, minocycline, levofloxacin, trimethoprim- Genome Sequences and Comparative sulfamethoxazole, amikacin, ticarcillin-clavulanate, and colistin. Genomics The MIC values were determined according to the Etest reading Annotated genome sequences of the unique isolates studied guide after 18 h of incubation at 37◦ C. The susceptibility test here were generated in the study by Gröschel et al. (2020) results were interpreted according to clinical susceptibility and are available from NCBI repository under the accession and resistance breakpoints (S) suggested by the Clinical and numbers RASX00000000-RAVB00000000. Genome sequences Laboratory Standards Institute (CLSI) (CLSI, 2019). For from S. maltophilia strains K279a, D457, and ATCC13637 were S. maltophilia, clinical breakpoints are now available for obtained from the NCBI genome database1 . The encoded protein trimethoprim-sulfamethoxazole (S ≤ 2 µg/mL), ticarcillin- sequences were extracted from the genomes and compared clavulanate (S ≤ 16 µg/mL), ceftazidime (S ≤ 8 µg/mL), to identify core and exclusive proteins. Orthologous protein minocycline (S ≤ 4 µg/mL), and levofloxacin (S ≤ 2 µg/mL). sequences of the selected S. maltophilia strains were assessed by For amikacin and colistin, CLSI breakpoints for Pseudomonas complete reciprocal best hit using UCSC blat (Kent, 2002) as aeruginosa were used with susceptibility cutoffs of 16 and comparison tool and accepting hits with a complete graph as 2 µg/mL, respectively. Isolates with intermediate antibiogram previously reported (Callister et al., 2008) or the largest clique. results were considered resistant (non-susceptible). MDR was This algorithm clusters the gene-derived proteins into core (hard defined as non-susceptibility to at least one agent in three or and soft) and accessory proteins. Hard core proteins are those more antimicrobial categories (Magiorakos et al., 2012). The found in all genomes, while soft core proteins are found in MIC50 and MIC90 values represent the MIC value at which 95–99% of genomes. Blastp2 and the HMM search program ≥50 and ≥90% of the strains within a test population are inhibited, respectively. 1 https://www.ncbi.nlm.nih.gov/genome/ 2 3 https://blast.ncbi.nlm.nih.gov/Blast.cgi http://pubmlst.org/smaltophilia/ Frontiers in Microbiology | www.frontiersin.org 3 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes In addition to the Etest method, colistin MICs were with 1% crystal violet (w/v) to measure their diameters. All determined by the broth microdilution (BMD) method in cation- experiments were conducted in triplicate. adjusted MH broth (CAMHB) in accordance with CLSI and the European Committee on Antibiotic Susceptibility Testing Biofilm Formation Assays (EUCAST) joint recommendations (CLSI, 2019; EUCAST, 2019). To determine biofilm formation, overnight cultures of each Briefly, MICs were determined in sterile 96-well plates by two- isolate were grown aerobically (200 rpm) in LB at 30◦ C, followed fold serial dilutions of colistin sulfate (Apollo Scientific Ltd., by dilution of the cultures into fresh LB to an OD550 of 0.1. Batch AS405305) in 100 µL of CAMHB. Bacteria were first grown Sterile untreated 96-well microtiter plates (BrandTech 781662) overnight in CAMHB using CLSI-recommended incubation were inoculated with 200 µL of the bacterial suspensions and conditions. After that, 100 µL of the bacterial suspensions with incubated at 30◦ C for 24 h. Prior to biofilm quantification, cell a final optical density at 550 nm (OD550 ) of 0.05 were added growth was estimated in each well by measuring the optical to the wells containing the antibiotic dilutions, and the MIC density at 620 nm (OD620 ) using a microplate reader (Multilabel plates were read after 18 h of incubation at 37◦ C. The MIC was Plater Reader VICTOR3). Quantification of the amount of defined as the lowest antibiotic concentration that inhibited 80% biofilm was performed by crystal violet (CV) staining as described of growth (based on OD measurements) in comparison to the previously (Huedo et al., 2014). Briefly, wells containing adhered control (CLSI, 2015). cells were washed three times with water, fixed at 60◦ C for 1 h and stained for 15 min with 200 µL of a 0.1% CV solution. SDS–PAGE and Immunoblotting of The stained biofilms were rinsed with distilled water, allowed to dry at 37◦ C for 30 min and then extracted with 200 µL of Lipopolysaccharides 95% ethanol. The amount of biofilm was quantified by measuring For identification of identical ladder-like LPS patterns, LPS the OD550 of dissolved CV using a microplate reader. Biofilm preparations from proteinase K-digested whole-cell lysates of formation (OD550 of CV) was normalized to cell growth (OD620 ) S. maltophilia isolates were separated on 12% sodium dodecyl and reported as relative biofilm formation. For this quantitative sulfate polyacrylamide gel electrophoresis (SDS–PAGE) gels assay, we have used six to eight replicate wells for each isolate. and stained with silver nitrate according to the method Based on the ratio (OD550 of CV)/(OD620 of cell growth), the of Hitchcock and Brown (1983). The lysates were prepared isolates were grouped into three categories to classify their ability from the biomass of overnight cultures grown on agar to form biofilm: weak (0 to 0.85), moderate (0.85 to 1.5), or strong plates. For immunoblots, a rabbit polyclonal antibody against (>1.5) biofilm formers (Zhuo et al., 2014). heat-inactivated S. maltophilia K279a cells was used. In order to detect serological cross-reactivity between O-specific polysaccharides, the antiserum was adsorbed with heat-killed Virulence in Caenorhabditis elegans Determination of the virulence of S. maltophilia strains in the cells of a K279a 1rmlBACD mutant as previously described C. elegans CF512 infection model was based on the “slow killing” (Steinmann et al., 2018). The LPS samples were electrotransferred method (Tan et al., 1999). C. elegans CF512 [fer-15(b26)II; fem- from SDS–PAGE gels onto polyvinylidene difluoride (PVDF) 1(hc17)IV], a strain showing temperature-dependent sterility, membranes (Millipore, Merck, Darmstadt, Germany), followed was provided by the Caenorhabditis Genetics Center (CGC) and by incubation of the blots with adsorbed anti-K279a antibody. routinely maintained on NGM plates (1.7% agar, 50 mM NaCl, The immunoblots were then treated with alkaline phosphatase- 0.25% peptone, 1 mM CaCl2 , 5 µg/mL cholesterol, 25 mM conjugated AffiniPure goat anti-rabbit immunoglobulin (Ig)G KH2 PO4 , 1 mM MgSO4 ) seeded with Escherichia coli OP50 (H + L) (Dianova, Hamburg, Germany) and developed in (provided by CGC) at 16◦ C. Bacterial strains were grown in Brain the presence of nitroblue tetrazolium and 5-bromo-4-chloro-3- Heart Infusion (BHI) broth (BD Difco Bacto) overnight at 37◦ C, indolylphosphate substrate (Promega). and 100 µL of each culture was spread onto a 5.5-cm-diameter NGM agar plate and incubated at 37◦ C for 24 h. Before usage, Production of Extracellular Protease plates were cooled down to room temperature, and 15–20 adult Extracellular proteolytic activity was assessed by means of casein hermaphrodite CF512 worms were placed onto plates seeded hydrolysis on LB agar supplemented with 1% (w/v) skimmed with the bacterium of interest. Plates were incubated at 25◦ C milk. The bacterial suspensions (5 µL each with an OD550 of (at which worms are sterile) and scored for live worms every 0.5) were spotted onto the agar, and the plates were incubated 24 h. E. coli OP50 was used as a negative control. A worm was at 30◦ C for 24 h. Proteolytic capacity was evaluated by measuring considered dead when it no longer responded to touch. Three the diameter of the translucent halos. replicates per strain were prepared. Twitching Motility Virulence in Galleria mellonella Twitching was assessed using subagar stab inoculations (stab Larvae of Galleria mellonella were obtained from our own assay) from fresh overnight LB agar plates as previously described hatchery, which was established in collaboration with Professor (Rashid and Kornberg, 2000). To prepare twitching plates, LB Fernando García del Pino from the Zoology Department at medium containing 1% (w/v) Noble agar (DifcoTM ) was used. the Universitat Autònoma de Barcelona. To prepare bacterial After incubation at 30◦ C for 24 h, the agar was removed from inocula, S. maltophilia isolates were grown overnight at 37◦ C in the plates, and the twitching zones were visualized by staining 10 mL of BHI in a rotary shaker at 200 rpm. Then, cells were Frontiers in Microbiology | www.frontiersin.org 4 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes sedimented by centrifugation, washed in PBS and adjusted to contain approximately 5 × 105 CFU in a single dose of 5 µl. Bacterial burden of the doses was confirmed by plating on BHI medium. This inoculum size was selected based on previous studies (Betts et al., 2014; García et al., 2015) that have determined the optimal dose of S. maltophilia required to kill G. mellonella over a 24–96 h period. Ten larvae per group were infected through the left proleg using a 50-µl Hamilton MicroliterTM R syringe with the aforementioned inoculum and incubated in the dark at 30◦ C in empty petri dishes. Mortality was determined every 24 h. Larvae were considered dead when they no longer responded to touch, which correlates with total melanization (black colored larvae). Statistical Analysis Statistical comparisons between groups were performed using Welch’s t-test, i.e., assuming non-equal variance. The associations between two categorical phenotype and genotype variables were investigated using the z-test or Fisher’s exact test as appropriate (Fisher’s exact test works better with small sample sizes). The odds ratio (OR) with a corresponding 95% confidence interval (95% CI) for each variable was calculated. In all cases a p-value < 0.05 was considered statistically significant, and all tests were two-tailed. Data were analyzed using SPSS 15.0 for Windows (SPSS Inc., Chicago, IL, United States) or GraphPad Prism 5 (GraphPad Software, San Diego, CA, United States). For infection experiments, the time at which 50% of animals were scored as dead (LT50), was calculated using a non-linear regression based on the Hill equation (GraphPad Prism 5). RESULTS Genetic Diversity of the S. maltophilia Strains To evaluate the genetic diversity and possible relations between the S. maltophilia isolates involved in this study, their MLST profiles were first analyzed. The 78 strains, isolated from patients of different European countries, mapped to 37 STs (Figure 1 and Supplementary Table S1). At the date of submission to the S. maltophilia MLST database, 30 of these STs were new and designated ST-76 to ST-81 and ST-120 to ST-143. Fifteen isolates (19.2%) belonged to seven previously described STs, whereas 63 (80.8%) belonged to the 30 novel STs, corroborating previous reports on the genotypic heterogeneity of S. maltophilia. Although no predominant clonal group was identified by BURST FIGURE 1 | Correlation between genotypes and antibiotic resistance profiles. analysis, the three most frequent MLST genotypes were ST-77 Neighbor-joining tree based on the concatenated data for all seven MLST loci (8 isolates), ST-132 (8 isolates), and ST-133 (7 isolates). The of the 61 “unique” S. maltophilia strains from this study. The ST of each isolate remaining STs occurred each with a frequency lower than 5.0%. is indicated (NA means not assigned). Previously defined genomic groups and rpf variants are also indicated. Data for strain D457 was deduced from its When the entire S. maltophilia MLST database (522 isolates, whole genome sequence (NCBI Reference Sequence: NC_017671.1). The 367 STs) was analyzed with the BURST algorithm, some of our antibiotic resistance profiles are displayed as spheres of different sizes in isolates were part of discrete clonal groups, sharing six (clonal consecutive columns outside the tree. Data are presented as log2 fold complexes, the most exclusive group definition) to five identical increase in resistance with respect to the clinical breakpoint for each antibiotic alleles (Supplementary Figure S1). For instance, four strains (abbreviations as in Table 1). Highest values in the dataset have the largest size, and all others are scaled down proportionally. Groups of LPS profiles are isolated from three different countries could be grouped into also indicated (ND means undetermined). Presence of a K279a-like O-antigen one of the few clonal groups (ST-8) detected so far in the is indicated with black triangles. global population of S. maltophilia (Supplementary Table S1 and Frontiers in Microbiology | www.frontiersin.org 5 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes Supplementary Figure S1). Likewise, our predominant ST (ST- the so far largest number of strains in the MLST database 77) comprises strains from different European countries and is (Supplementary Figure S2) and includes the reference strain part of a genetically related group of the clonal complex ST-103 ATCC13637 and the reference genome strain K279a. The other (Supplementary Figure S1). In addition to MLST, rpf genotyping genogroups with a significant number of isolates in our collection (Huedo et al., 2014) revealed that the rpf -1 variant was present in were #2 (19%), #3 (13%) and C (15%), which is in agreement a total of 47 strains (60%), while rpf -2 was identified in 31 strains with the frequency of isolates belonging to these groups in (40%) (Supplementary Table S1). the worldwide population. In agreement with previous studies Among the 78 strains studied here, five pairs of isolates (Corlouer et al., 2017; Mercier-Darty et al., 2020), genogroup #2 were confirmed to be epidemiologically linked. They were either represents the most genetically diverse group in our collection. As cultured from two sources of the same patient on the same day proposed by Gröschel et al. (2020) and shown in Supplementary (four pairs) or originated from samples taken from the same Table S1, the strains could be further classified into the patient at different dates (pair OS91 and OS87) (Supplementary monophyletic lineages named Sm1–Sm18 (termed S. maltophilia Table S1). In most cases, the paired isolates share the same sensu lato) or the more distantly related lineages Sgn1–Sgn4 ST, but isolates PU101 and PU102 are genetically distinct, with according to Vinuesa et al. (2018). All our strains, except the only three alleles in common. Also, based on the MLST typing phylogenetically distant strain OU111 (genogroup A), belong to and data from the source of isolation, among the 35 strains the S. maltophilia sensu lato cluster. None of them clustered into collected from Germany, and coming from a single hospital, lineages in which there are isolates mislabeled as S. maltophilia there seems to be genetically related groups (Supplementary or with an unclear taxonomic status, for example the genomic Table S1). In order to avoid interference of clonal strains with groups #8 and #10 (Ochoa-Sánchez and Vinuesa, 2017). In correlation tests, those isolates displaying identical MLST type this respect, our results therefore could provide support for and isolation data (i.e., same source, patient, and date) were the multispecies concept to explain diversity among isolates of considered as the same clone, so that only one representative S. maltophilia. isolate was selected for further investigations (Supplementary Interestingly, there were statistically significant associations Table S1, highlighted in bold). However, those clonal strains between the major MLST-based genomic groups and the rpf that showed different antibiotic susceptibility profiles (see below) variants. In the genomic group #6, most of the strains contained were not excluded from further analyses. Thus, the analysis of the variant rpf -1 (P < 0.001), whereas the variant rpf -2 was correlations between the phenotypic and genotypic parameters predominant in genogroups C (P < 0.001) and #3 (P < 0.05). studied here was performed on 58 isolates. In addition, we have included the strains K279a, D457, and ATCC13637 as references in all experiments, yielding a total of 61 “unique strains” (36 rpf -1 Identification of Core and Unique and 25 rpf -2). Proteins Within S. maltophilia Strains and Genogroups In addition to the genotypic characterization, the genome- Classification of S. maltophilia Isolates derived protein sequences of 60 unique strains of S. maltophilia Into Genomic Groups were compared to identify core and exclusive proteins (assembled For cluster analysis, taxonomic status verification and assignment genome sequence of PR59 could not be determined). According of the S. maltophilia isolates to genomic groups, phylogenetic to our ortholog identification method, the S. maltophilia strains analyses were carried out based primarily on concatenation of included in this comparison share a core genome of 2,585 the seven MLST genes (Figure 1 and Supplementary Figure S2). proteins. If we consider the soft core genome (proteins present A neighbor-joining tree of the 522 isolates in the MLST database in 95–99% of genomes), 349 additional proteins could be added (Supplementary Figure S2) allowed us to cluster all these strains to this set of shared proteins. We also identified the number according to the 20 genogroups (#1 to #10 and A to J) as of exclusive proteins shared by all members of each genogroup previously established by Hauben et al. (1999), Kaiser et al. (Figure 2A and Supplementary Table S2). The analysis showed (2009), and Mercier-Darty et al. (2020). This phylogenetic that the number of unique proteins to each genogroup varies analysis showed a lack of clustering by isolation source or considerably from genogroup to genogroup. Genogroups #5, #7, geographical origin (Supplementary Figure S2), which stresses D and A clearly stand out having the highest number of unique again the heterogeneity of S. maltophilia isolates, in this case proteins; this is likely due to these groups being represented at a worldwide level. Even though most of the strains in the by a single strain. It is worth mentioning that all members of MLST database were of clinical origin, it is noteworthy that genogroup C share 17 unique proteins, 12 of which are part almost all environmental strains were clustered in groups #8, of an operon that encodes a protein translocation system for #9, and #10. The 78 strains in our panel could be clustered into the assembly of Flp or Tad (tight adherence) pilus (Figure 2B). 12 genogroups (Supplementary Table S1), including the new Moreover, members of genogroup #4 share a unique operon for group G (4 strains) by Mercier-Darty et al. (2020). Therefore, exopolysaccharide synthesis, and isolates in genogroup G share our strain collection provides a good representation of the 41 proteins including peptidases and a transcriptional regulator, global diversity of S. maltophilia. Nineteen (31%) of the 61 in addition to other unique enzymes and several hypothetical unique isolates were clustered in genogroup #6, including the proteins (Supplementary Table S2). On the other hand, the clonal complex ST-8 (Figure 1). Genomic group #6 comprises comparative analysis did not identify proteins shared exclusively Frontiers in Microbiology | www.frontiersin.org 6 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes FIGURE 2 | Comparative genomics of S. maltophilia genomic groups. (A) Flower plot of unique proteins in each genogroup and size of the hard-core proteome. The petals marked with diagonal lines represent the genomic groups where the strains are mainly of the rpf-2 variant. The identified unique proteins to each genogroup are listed in Supplementary Table S2. (B) Linear map showing the Flp/Tad pilus-encoding gene cluster exclusive to all strains of genogroup C (rpf-2), based on strain D457 (NCBI Reference Sequence: NC_017671.1) as the reference genome. The coding sequences (CDS) are numbered from 1 to 14 and the functions of the encoded proteins are shown below the map. Two CDS (empty arrows) were not detected as unique by our algorithm because they have paralogs within each genome. by strains of the rpf -1 or rpf -2 variants. This result highlights isolates with apparently unique LPS patterns (Supplementary the genetic heterogeneity among clinical S. maltophilia isolates Table S1). There were at least two different LPS profiles in and confirms, in addition to their distribution along genotypes, each of the genogroups with more than one member, where that the origin of the two rpf variants in S. maltophilia can be genogroups F and C with dominating LPS groups 21 and 15, explained by horizontal gene transfer. respectively, were the most homogeneous. In summary, apart from some exceptions, we could uncover a good correlation between the LPS profile and the ST of the isolates, which Diversity of S. maltophilia Isolates Based was largely in accord with the phylogenetic tree shown in on Their LPS Profiles Figure 2. Furthermore, the LPS samples were screened with The enormous structural diversity of surface-exposed O-specific a polyclonal antiserum raised against the O-antigen of strain polysaccharides of smooth-type LPS is assumed to result from K279a, which has been recently proposed to share similar adaptive mechanisms that enable the bacteria to withstand characteristics with the S. maltophilia O8-antigen of a branched specific conditions in various environmental niches, which in tetra-saccharide with three rhamnopyranosyl residues in the turn provides the possibility of using O-polysaccharides as an main chain and 3-O-methylxylose as substituent (Steinmann epidemiological marker to discriminate between related and et al., 2018) (Supplementary Figure S3). Our immunoblot unrelated strains within a given species (Rimler, 1990; Aucken analyses indicated serological cross-reactivity of the antiserum and Pitt, 1993; Santamaría et al., 1998). We therefore compared with the O-antigens of all strains of LPS profile group 2 the LPS profiles of our S. maltophilia isolates as an additional (including S. maltophilia K279a), which accounts for 14.5% of marker to cluster the strains on the basis of identical LPS all examined isolates. Interestingly, the LPS of all strains of patterns in SDS-PAGE, the ladder-like appearance of which is genogroup #4 (3.9%) cross-reacted as well with the antibody, caused by varying numbers of oligosaccharide repeating units suggesting that these isolates, despite showing different LPS in the O-specific polysaccharide (Supplementary Figure S3). As banding patterns, expressed an O-specific polysaccharide either expected, we could observe significant heterogeneity of the LPS structurally similar or even identical to that of the reference banding patterns, whereby none of the isolates investigated here strain K279a. Of note, each of the ten separate groups synthesized rough-type LPS lacking the O-specific polysaccharide comprising two to six clonal strains did not only display (Supplementary Figure S3). This was also valid for the strains an identical MLST type, but could also be assigned to one B5565, H5726 and PU126, but unequivocal evidence for the and the same LPS group, which supported the conclusion of presence of smooth-type LPS could only be obtained using epidemiologically linked strains that were isolated at the same increasing amounts of LPS in SDS-PAGE (Supplementary hospital in Germany (Supplementary Table S1). All strains Figure S3, inset). The three strains were therefore excluded from that reacted with the O-specific K279a antiserum carried the further analysis of LPS profiles. Based on the electrophoretic rpf -1 cluster, except for strain OC323 (genetically close to mobility of higher molecular mass LPS species, we assigned K279a) which encoded the rpf -2 variant (Figure 1). This could 76 strains to a total of 31 LPS groups, including 20 be in fact another indication that the two rpf variants in Frontiers in Microbiology | www.frontiersin.org 7 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes S. maltophilia were acquired by horizontal gene transfer as OU343, and OL16, all belonging to different and distant previously suggested (Huedo et al., 2015), at least for strain genogroups (Figure 1), were susceptible to all antibiotics tested OC323 that may have received the rpf -2 cluster more recently (Supplementary Table S1). Overall, no significant correlations by recombination events. were observed between the resistance profiles (antibiotic resistance combinations) and the MLST genotypes. However, statistically significant associations could be observed between Correlation Between Antibiotic specific genomic groups and the susceptibility of the isolates Resistance Phenotypes and Genotypes to certain antibiotics (Supplementary Table S3). For instance, The antibiotic resistance pattern of our isolates was evaluated almost all strains of the genogroup #6 were susceptible to with the β-lactam antibiotics ceftazidime and ticarcillin- ticarcillin-clavulanate, whereas strains of genogroups #2 and F clavulanate, and the non-β-lactam antimicrobials amikacin, were mostly susceptible to colistin but ceftazidime resistant. trimethoprim-sulfamethoxazole, levofloxacin, minocycline and The genetically related strains of genogroups #2 and F showed colistin (MIC values in Supplementary Table S1). The MIC similar resistance profiles and were therefore considered as a values of all antimicrobials were determined by the Etest, with whole in association studies. In addition, we found high-level the exception of colistin, which was additionally tested using resistance to levofloxacin in all strains of the new genogroup the BMD method (see section “Materials and Methods”). All G. Almost all strains of the genogroup C were susceptible strains were susceptible to minocycline, and only 6.5 and 19.7% to the β-lactams ceftazidime and ticarcillin-clavulanate, but of the unique strains were non-susceptible to trimethoprim- were quantitatively more resistant to colistin (Supplementary sulfamethoxazole and levofloxacin, respectively. Conversely, high Table S3 and Figure 1). Interestingly, the variant of the rpf resistance rates were detected against the antibiotics amikacin cluster showed some correlation with the resistance profile of the (54.1%), ticarcillin-clavulanate (62.3%), and ceftazidime (62.3%), isolates. Most of the rpf -1 strains turned out to be resistant to all of them with MIC values reaching >256 µg/mL (Table 1). ceftazidime (P = 0.014), and, from a quantitative point of view, For colistin susceptibility testing, as expected (Moskowitz the rpf -2 resistant strains inclined toward higher MIC values et al., 2010), the results of the BMD and Etest methods for colistin compared to the rpf -1 variant, although without were significantly different. For our strain collection, the Etest statistical significance (Figure 3A). This analysis also showed resulted in a colistin MIC50 value of 2 µg/mL corresponding higher resistance levels to ceftazidime and ticarcillin-clavulanate to susceptible strains (susceptibility breakpoint ≤ 2 µg/mL), among resistant strains of the rpf -1 variant. whereas the BMD method yielded a colistin MIC50 value above the susceptibility breakpoint (>2 µg/mL) (Table 1). Moreover, while the colistin MIC90 value determined by both methods Correlation Between Genotypes and was consistent with a resistant phenotype, it was eightfold lower Virulence-Related Phenotypes by the Etest than by the BMD method. This is due to the Virulence-associated phenotypes, including biofilm formation, phenomenon of heteroresistance to colistin displayed by most of production of extracellular enzymes and motility, were finally the S. maltophilia isolates (Martínez-Servat et al., 2018). Finally, assayed for our strain collection. All isolates were protease and taking the recommended BMD method as a reference, 67.3% of biofilm producers (Supplementary Figure S4). However, the the strains could be considered as non-susceptible to colistin. amount of biofilm formed in vitro varied considerably among To further characterize the resistance patterns, the criteria S. maltophilia isolates. The rpf -2 strain OU110 exhibited the suggested by Magiorakos et al. (2012) for related non- highest biofilm-forming ability. Most of the isolates (44.3%, fermenting Gram-negative bacteria were applied. Among the 27/61) produced moderate biofilms; with a mean of 1.14 and 61 unique isolates, 37 (60.7%) were found to be MDR strains, standard deviation of ±0.17 (see “Materials and Methods” whereas 24 (39.3%) were not. On the other hand, 22 different section for measurement and classification of biofilms). Strong antibiotic resistance profiles were observed among the 61 strains biofilm formation capacity (mean: 1.95, SD: ±0.43) was observed (Supplementary Table S1 and Figure 1), the most common in 24.6% (15/61) of the isolates, whereas the remaining being those that combine resistance to ceftazidime, amikacin, 19 strains (31.1%) showed weak biofilm formation capacity ticarcillin-clavulanate, and colistin. Notably, in three out of five (mean: 0.38, SD: ±0.24). With respect to genotypes, we found strain pairs isolated from the same patient, the members of the statistically significant associations (P < 0.005) between the same pair showed different resistance profiles (Supplementary isolates of genomic group C and their ability to form biofilm Table S1). The two genetically different strains isolated from (Supplementary Table S4). On the contrary, 90.9% of strains of the oropharinx and the perineum of the same patient, PU101 genomic groups #6, F and #2 were moderate to weak biofilm and PU102, displayed the same resistance profile. These strains producers, though without significant associations. Once again, were also assigned to different LPS groups. Strain OL11 was the despite variations in phenotypic traits, the biofilm-forming ability most resistant isolate in our panel (six out of seven antibiotics). of the strains could be related to the rpf variant. Quantitatively, It had been isolated with PL12 from different anatomical sites the biofilm formation capacity of rpf -2 strains (relative biofilm of the same patient. The two strains are genetically identical, mean 1.38) was significantly higher (P < 0.05) than that of possess the same LPS profile (Supplementary Table S1) and rpf -1 strains (relative biofilm mean 0.92) (Figure 3B). With are resistant to almost all antibiotics, except that PL12 showed respect to the number of isolates, strains of variant rpf -1 susceptibility to colistin. On the other end, strains OU152, are mostly moderate to weak biofilm formers (P < 0.001) Frontiers in Microbiology | www.frontiersin.org 8 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes TABLE 1 | In vitro activity (µg/mL) of seven antimicrobial agents against the 61 “unique” strains of S. maltophilia studied here. Method Agent Class or sub-class MIC50a MIC90a Range % of susceptible isolatesb Etest Ceftazidime (CAZ) Cephalosporins 16 >256 1–>256 37.7 Minocycline (MIN) Tetracyclines 0.25 0.5 0.032–2 100 Levofloxacin (LEV) Fluoroquinolones 0.5 16 0.125–>32 80.3 Trimethoprim- Folate pathway inhibitors 0.25 0.5 0.125–>32 93.5 sulfamethoxazole (SXT) Amikacin (AMK) Aminoglycosides 32 256 2–>256 45.9 Ticarcillin-clavulanate (TTC) Penicillins 64 >256 0.5–>256 37.7 Colistin (COL) Polymyxins 2 8 0.064–32 75.4 BMDc Colistin Polymyxins 8 64 <0.25–>256 32.7 a MIC90 and MIC50 values were defined as the lowest concentration of the antibiotic on a two-fold scale at which 90 and 50% of the isolates were inhibited, respectively. b Susceptibility breakpoints for S. maltophilia established by the Clinical and Laboratory Standards Institute (CLSI, 2019). For amikacin and colistin, CLSI breakpoints for P. aeruginosa were used. Intermediate and resistant strains were grouped together into a non-susceptible category. c Broth microdilution. FIGURE 3 | Phenotypic resistance and virulence-related phenotypes within each rpf-variant subpopulation of S. maltophilia. (A) Box and whisker plots (range bars indicate maximum and minimum values) showing log2-fold change increase from susceptibility cut-off for each antibiotic: CAZ, ceftazidime; LEV, levofloxacin; AMK, amikacin; TTC, ticarcillin-clavulanate; COL, colistin. Only resistant isolates were included. (B) Profiles for virulence traits of isolates from different rpf-variant subpopulations (*p < 0.05; unpaired t-test with Welch’s correction). (Supplementary Table S5). On the other hand, there was no strains evaluated in this study were differentially virulent to statistically significant correlation of protease secretion and C. elegans in the slow killing assay, with no correlation twitching motility with the phylogenetic clustering, except for found with genomic groups (Supplementary Figure S4 and strains belonging to genogroup C or those that are of the Supplementary Table S4). The strains required 4–10 ten rpf -1 variant, which are less casein exoprotease producers days to kill 50% of the worms (mean LT50 = 6.9 days). (Supplementary Figure S4 and Supplementary Tables S4, S5). Complete killing by E. coli OP50, used as a negative control, The virulence capacity of the strains was investigated by was achieved after 15 days. No significant correlation was infecting both the nematode C. elegans and larvae of the insect found between the assayed virulence-related phenotypes and G. mellonella. The results were compared and correlated with the virulence of the strains in the C. elegans model. The two the assayed virulence-related phenotypes. The S. maltophilia least virulent strains in this model, PC274 (LT50 = 10.5 days) Frontiers in Microbiology | www.frontiersin.org 9 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes and the type strain ATCC13637 (LT50 = 10.0 days), showed practice (Hand et al., 2016). On the other hand, colistin is low motility and biofilm formation capacity and produced usually the last resort antibiotic for treatment of infections lower amounts of extracellular proteases. Nevertheless, the caused by MDR strains, despite its proven toxicity (Petrosillo most virulent strain EV5520 (LT50 = 3.8 days) displayed also et al., 2008; Wood et al., 2010). In the present study, in weak virulence-related phenotypes, indicating that virulence addition to a high proportion of colistin-resistant isolates, factors other than those examined here contribute to the we have shown that S. maltophilia was capable of developing infection process in the nematodes. In general the type colistin-resistance heterogeneity, deduced from the observed strain ATCC13637 showed obvious signs of attenuation in discrepancies between the two commonly used susceptibility all assays (Supplementary Figure S4), as previously reported testing methods (Etest vs. BMD) (Lo-Ten-Foe et al., 2007). using a zebrafish infection model (Ferrer-Navarro et al., 2013), This is in agreement with our previous, more detailed studies most likely due to numerous in vitro passages since it was on heteroresistance with four colistin resistant S. maltophilia isolated in 1960. strains (Martínez-Servat et al., 2018). Based on the strain All S. maltophilia isolates were virulent to G. mellonella larvae characterization presented here, this phenotypic heterogeneity in killing assays. The selected inoculum size of 5 × 105 CFU seems to be widely spread among clinical isolates regardless of resulted in 20–60% death at 24 h post-inoculation with most their genotypes and resistance profiles. Due to the emergence strains, while complete killing was observed after 96 h. Some of resistances to almost all antibiotic classes and the observed strains were considered to be highly virulent as they were heteroresistance phenomenon, combination therapies and novel able to kill approximately 60% of the larvae after 24 h of agents may soon be needed to treat infections caused by MDR inoculation and reached 100% after 48 h. These highly virulent S. maltophilia strains. strains belong to almost all genogroups and rpf variants. An important finding of this study is that resistance to some Nevertheless, on average, rpf -2 strains were more virulent than antibiotics appears to be influenced by the rpf variant. For rpf -1 strains (P = 0.05) on the basis of LT50 values (Figure 3B). instance, while the rpf -1 subpopulation displayed higher levels of The association results showed again no significant correlation resistance to the β-lactam antibiotics ceftazidime and ticarcillin, between virulence-related phenotypes and the capacity of the strains of the rpf -2 group showed higher resistance levels to strains to kill the larvae. colistin (Figure 3A). The differential resistance profiles found for the rpf -1 and rpf -2 subpopulations may be attributable to inherent DSF production and its lack in rpf -1 and rpf -2 DISCUSSION subpopulations, respectively (Huedo et al., 2014). While rpf -1 strains synthesize DSF under standard conditions activating In recent years S. maltophilia has received considerable attention. the QS, rpf -2 strains require detection of exogenous DSF to First, the frequency of MDR phenotypes found among isolates trigger DSF production (Huedo et al., 2015). In this line, it is particularly high, as demonstrated by the panel studied here. has been reported that DSF signals stimulate the production Second, S. maltophilia is capable of surviving in a wide variety of and release of outer membrane vesicles (OMVs), which may environmental niches, partly due to its capacity to form biofilms contain β-lactamases (Devos et al., 2015). Furthermore, some on diverse biotic and abiotic surfaces, including medical devices. antibiotic resistance mechanisms seem to be controlled by Finally, S. maltophilia is characterized by a high rate of genomic DSF in the plant-growth-promoting model strain S. maltophilia re-arrangements and hypermutator activity, which allows rapid R551-3 (rpf -1), including the biosynthesis of β-lactamases and adaptation to new niches (Berg and Martinez, 2015). With this multidrug efflux pumps (Alavi et al., 2013). On the other hand, presentation card, it does not come as a surprise that, despite not in time-kill curve experiments, S. maltophilia’s DSF seemed to being a particularly pathogenic or virulent species, S. maltophilia influence polymyxin tolerance in P. aeruginosa, with addition has found its way as an emerging nosocomial pathogen (Brooke, of DSF leading to an enhanced tolerance to polymyxin B 2012; Adegoke et al., 2017). and colistin (Ryan et al., 2008). We have recently shown Although 60.7% of the clinical strains investigated in this work that naturally occurring DSF signals and synthetic derivatives were MDR, most isolates were susceptible to trimethoprim- potentiate the activity of the polymyxin antibiotic colistin against sulphamethoxazole and all to minocycline. Monotherapy S. maltophilia and some members of the Burkholderia cepacia with trimethoprim-sulfamethoxazole, fluoroquinolones or complex (Huedo et al., 2019). Accumulation of DSF on the tetracyclines such as tigecycline, minocycline, and doxycycline surface may slightly increase the permeabilization of the bacterial remains the most effective antimicrobial treatment against this membrane due to its lipophilic nature, potentiating colistin organism (Chang et al., 2015; Hand et al., 2016). However, activity against rpf -1 strains. Worth mentioning, there appear resistance to levofloxacin was detected in almost one fifth of to be associations between MLST genotypes and antibiotic the strains evaluated in this study. Even though trimethoprim- resistance regardless of the rpf cluster variant. Strains of sulphamethoxazole is indeed the first choice antibiotic against genogroup C were mostly susceptible to ticarcillin-clavulanate S. maltophilia, resistance to this drug has been observed in in addition to ceftazidime. Moreover, all strains of genogroups 6.5% of the strains included in this work, as it has also been #4 (rpf -1) and G (rpf -2) were resistant to levofloxacin, and reported in other studies (Valdezate et al., 2001). Minocycline most strains of genogroups #2 and F (mostly rpf -1) were emerges as the best alternative to treat S. maltophilia, but more susceptible to amikacin. A significant proportion of β-lactam- studies are needed to introduce this antibiotic into clinical resistant isolates arise in S. maltophilia through altered expression Frontiers in Microbiology | www.frontiersin.org 10 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes of β-lactamases (Sánchez, 2015) or amino acid substitutions virulence factors (Berg and Martinez, 2015) significantly in efflux pumps (Blanco et al., 2019). Both intrinsic and impedes the establishment of relationships with epidemiological acquired resistance mechanisms are also involved in resistance relevance. Although we could not uncover significant to aminoglycosides and quinolones in S. maltophilia, such correlations between all phylogenetic groups and virulence- as expression of multidrug−resistant efflux pumps, presence related phenotypes, we have identified groups of genetically of antibiotic inactivating enzymes or mutations in the genes related strains that are scattered throughout European countries coding for antibiotic targets (García-León et al., 2014; Sánchez, and could represent high-risk clones. For instance, although 2015). The comparative genomic analysis found no unique the number of strains tested is small, it can be observed that proteins shared by members of a genomic group that would genogroup C comprises stronger biofilm-forming strains that be involved, in a straightforward, direct way, in antibiotic are almost all resistant to the last-resort antibiotic colistin. resistance (Supplementary Table S2). Taken together, these However, all these strains are susceptible to other antibiotics findings confirm that mutations, together with horizontal gene tested herein and, thus; there remain good treatment options transfer, affect the resistance patterns in S. maltophilia isolates against infections caused by them. Nevertheless, strains of beyond genomic diversity. this group associated with human infections (Supplementary In addition to antibiotic resistance, the genetic variant of the Table S1) have also proven to be highly virulent in different DSF QS system has an impact on pathogenicity of S. maltophilia infection models (Ferrer-Navarro et al., 2013; Huedo et al., as well, and many virulence determinants are involved in 2014), and can quickly acquire a resistant phenotype (Alonso this process. Biofilm formation is associated with an increase and Martínez, 1997; Martínez-Servat et al., 2018). Interestingly, in virulence of S. maltophilia. Although the biofilm forming all strains of group C seem to express an additional type capacity varied considerably among the strains studied here, of the Flp/Tad pilus encoded by a unique cluster of genes consistent with previous observations from our group (Huedo (operon SMD_0948-SMD_0960 in D457). This operon seems et al., 2014), we have shown that strains carrying the rpf -2 to be different from the TadE-like pili gene cluster located at variant are stronger biofilm producers than strains carrying the Smlt2867-Smlt2875 in the K279a genome (Crossman et al., rpf -1 version (Figure 3B). In a previous study, we showed 2008), although some components are similar. In strain D457, that, although the protein variant RpfF-2 is functional in DSF the closest ortholog to the Flp/Tad pilus in K279a is the system synthesis, its permanent repression by RpfC-2 results in the encoded by operon SMD_2505-SMD_2513. The Flp/Tad pilus absence of DSF production under the conditions tested. In has been categorized as a subtype of the type II secretion contrast, the rpf -1 subpopulation produces larger amounts of system (Tomich et al., 2007), and the encoding cluster is present DSF and, in that case, DSF seems to contribute to biofilm in a wide variety of bacterial species, for which it functions dispersion (Huedo et al., 2014). Similar results were observed for as an important virulence factor. It is considered a target of the closely related bacterium Xanthomonas campestris (encoding horizontal gene transfer and is essential for biofilm formation a very close RpfF-1 variant), in which DSF also appears to (Tomich et al., 2007; Nykyri et al., 2013). To our knowledge, disperse biofilms (Dow et al., 2003). Likewise, in P. aeruginosa, it is the first time that this alternative structure has been found a related DSF signal, formerly termed cis-DA, displayed a strong encoded in the genome of strains of S. maltophilia, which dispersion capacity of mature biofilms of a large number of could contribute to the virulence of these strains. Besides, bacterial pathogens (Davies and Marques, 2009). Overall, it the genogroup #6 includes the largest subgroup of our strain seems likely that the absence of DSF production confers on collection. Of note, the predominance of clinical isolates in rpf -2 members a stronger ability to form biofilms, and this genogroup #6 has also been shown by others, mostly in cystic may partially explain why these strains are more virulent in fibrosis patients (Kaiser et al., 2009; Vasileuskaya-Schulz et al., larvae of G. mellonella. Pompilio et al. (2016) observed that 2011; Corlouer et al., 2017; Ochoa-Sánchez and Vinuesa, 2017; S. maltophilia biofilm formation is directly associated with Steinmann et al., 2018). Three of a total of four trimethoprim- mortality rate in G. mellonella. In general, DSF signaling in sulfamethoxazole resistant isolates of our collection belong to S. maltophilia regulates factors contributing to virulence, such genogroup #6, this antibiotic being a first-line therapeutic choice as motility, biofilm formation and colonization (Alavi et al., in S. maltophilia infections. 2013; An and Tang, 2018). However, no correlation could be Phylogenomic analyses linked with comparative detected when comparing virulence and biofilm formation with LPS profiling and O-serotyping have recently revealed casein proteolitic activity or twitching motility, suggesting that correlations between genomic classification of a number the latter attributes are not major virulence traits of S. maltophilia, of S. maltophilia isolates and electrophoretic banding at least in the two infection models tested. On the other hand, patterns of their LPS in silver-stained polyacrylamide gels no significant association between antibiotic resistance and the and immunoblots with an antibody raised against the amount of biofilm formed could be observed. Thus, our results O-antigen of the reference strain S. maltophilia K279a are similar to those reported by Pompilio et al. (2016) but (Steinmann et al., 2018). The results obtained here confirm diverge from those reported by Liaw et al. (2010), who concluded the usefulness of LPS profile typing in combination with that MDR isolates show higher levels of biofilm formation in immunoblotting as a supportive tool for identification of comparison to non-MDR strains. epidemiologically related S. maltophilia strains. The LPS- In S. maltophilia, the high intra-specific genetic variability, based approach was not only suited to supplement the together with a low diversity of the known resistance and genotypic characterization of our diverse strain panel, but Frontiers in Microbiology | www.frontiersin.org 11 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes could also corroborate previous data on the high variability of AUTHOR CONTRIBUTIONS O-polysaccharide structures among S. maltophilia strains. DY, PH, SM-S, OC-S, UM, XC, FL, and IR performed the experiments presented in the work. OC-S and DY did the CONCLUSION comparative genomics analyses. DY, PH, SM-S, OC-S, and UM participated in the writing of the manuscript. All authors Overall, this study shows that the QS system plays a pivotal role contributed to the design and interpretation of the results, revised in pathogenicity and persistence in S. maltophilia. Furthermore, the manuscript, and approved it for publication. since there are significant correlations between the rpf variant and the phylogenetic groups of the isolates based on their MLST profiles (Supplementary Table S4), this genotyping method could be valid as an epidemiological tool. Genome sequencing FUNDING studies also revealed genes that are exclusive to certain groups of isolates and could contribute to the observed phenotypes, This work was funded by the Spanish MICINN showing that other factors besides DSF synthesis contribute (BIO2015-66674-R). The authors also thank Catalan AGAUR for to the virulence and resistance of S. maltophilia. More in- support (2017 SGR 1062). IR was supported by the Department depth genomic analysis (e.g., single nucleotide polymorphism of Health, Generalitat de Catalunya, grant SLT002/16/00349. detection), together with phenotypic characterization of a larger collection of representative clinical isolates of S. maltophilia, will give a much deeper understanding of genetic diversity, ACKNOWLEDGMENTS phylogeny, population structure and epidemiology of this pathogen. For instance, this will allow the discovery of point We gratefully acknowledge Dörte Grella and Manuel Hein mutations involved in antibiotic resistance, for example, in clonal (Research Center Borstel) for their excellent technical assistance. strains isolated from the same patient. Altogether, this could The strain D457 was kindly provided by Dr. JM (Centro Nacional have significant implications in understanding the biology of de Biotecnología, CSIC, Madrid). this species and, even more important, in the management of S. maltophilia infections. SUPPLEMENTARY MATERIAL DATA AVAILABILITY STATEMENT The Supplementary Material for this article can be found The datasets generated for this study can be found in the https: online at: https://www.frontiersin.org/articles/10.3389/fmicb. //pubmlst.org/smaltophilia/. 2020.01160/full#supplementary-material REFERENCES against multidrug-resistant Stenotrophomonas maltophilia. Eur. J. Clin. Microbiol. Infect. Dis. 33, 1565–1572. doi: 10.1007/s10096-01 Abda, E. M., Krysciak, D., Krohn-Molt, I., Mamat, U., Schmeisser, C., Förstner, 4-2101-3 K. U., et al. (2015). Phenotypic heterogeneity affects Stenotrophomonas Blanco, P., Corona, F., and Martínez, J. L. (2019). Involvement of the RND maltophilia K279a colony morphotypes and β-lactamase expression. Front. efflux pump transporter SmeH in the acquisition of resistance to ceftazidime Microbiol. 6:1373. doi: 10.3389/fmicb.2015.01373 in Stenotrophomonas maltophilia. Sci. Rep. 9, 1–14. doi: 10.1038/s41598-019- Adegoke, A. A., Stenström, T. A., and Okoh, A. I. (2017). Stenotrophomonas 41308-9 maltophilia as an emerging ubiquitous pathogen: looking beyond contemporary Bostanghadiri, N., Ghalavand, Z., Fallah, F., Yadegar, A., Ardebili, A., Tarashi, antibiotic therapy. Front. Microbiol. 8:2276. doi: 10.3389/fmicb.2017.02276 S., et al. (2019). Characterization of phenotypic and genotypic diversity of Alavi, P., Müller, H., Cardinale, M., Zachow, C., Sánchez, M. B., Martínez, Stenotrophomonas maltophilia strains isolated from selected hospitals in Iran. J. L., et al. (2013). The DSF quorum sensing system controls the positive Front. Microbiol. 10:1191. doi: 10.3389/fmicb.2019.01191 influence of Stenotrophomonas maltophilia on plants. PLoS One 8:e67103. doi: Brooke, J. S. (2012). Stenotrophomonas maltophilia: an emerging global 10.1371/journal.pone.0067103 opportunistic pathogen. Clin. Microbiol. Rev. 25, 2–41. doi: 10.1128/CMR. Alonso, A., and Martínez, J. L. (1997). Multiple antibiotic resistance in 00019-11 Stenotrophomonas maltophilia. Antimicrob. Agents Chemother. 41, 1140–1142. Brooke, J. S. (2014). New strategies against Stenotrophomonas maltophilia: a serious doi: 10.1128/AAC.41.5.1140 worldwide intrinsically drug-resistant opportunistic pathogen. Expert Rev. Anti An, S., and Tang, J. (2018). Diffusible signal factor signaling regulates multiple Infect. Ther. 12, 1–4. doi: 10.1586/14787210.2014.864553 functions in the opportunistic pathogen Stenotrophomonas maltophilia. BMC Callister, S. J., McCue, L. A., Turse, J. E., Monroe, M. E., Auberry, K. J., Res. Notes 11:569. doi: 10.1186/s13104-018-3690-1 Smith, R. D., et al. (2008). Comparative bacterial proteomics: analysis of Aucken, H. M., and Pitt, T. L. (1993). Lipopolysaccharide profile typing as a the core genome concept. PLoS One 3:e1542. doi: 10.1371/journal.pone.00 technique for comparative typing of gram-negative bacteria. J. Clin. Microbiol. 01542 31, 1286–1289. Chang, Y.-T., Lin, C.-Y., Chen, Y.-H., and Hsueh, P.-R. (2015). Update on Berg, G., and Martinez, J. L. (2015). Friends or foes: can we make a distinction infections caused by Stenotrophomonas maltophilia with particular attention between beneficial and harmful strains of the Stenotrophomonas maltophilia to resistance mechanisms and therapeutic options. Front. Microbiol. 6:893. complex? Front. Microbiol. 6:241. doi: 10.3389/fmicb.2015.00241 doi: 10.3389/fmicb.2015.00893 Betts, J. W., Phee, L. M., Woodford, N., and Wareham, D. W. (2014). Chua, S. L., Yam, J. K. H., Hao, P., Adav, S. S., Salido, M. M., Liu, Y., et al. (2016). Activity of colistin in combination with tigecycline or rifampicin Selective labelling and eradication of antibiotic-tolerant bacterial populations Frontiers in Microbiology | www.frontiersin.org 12 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes in Pseudomonas aeruginosa biofilms. Nat. Commun. 7:ncomms10750. doi: 10. García, C. A., Alcaraz, E. S., Franco, M. A., and Passerini de Rossi, B. N. (2015). 1038/ncomms10750 Iron is a signal for Stenotrophomonas maltophilia biofilm formation, oxidative Chung, H., Lieberman, T. D., Vargas, S. O., Flett, K. B., McAdam, A. J., Priebe, stress response, OMPs expression, and virulence. Front. Microbiol. 6:926. doi: G. P., et al. (2017). Global and local selection acting on the pathogen 10.3389/fmicb.2015.00926 Stenotrophomonas maltophilia in the human lung. Nat. Commun. 8:14078. García-León, G., Salgado, F., Oliveros, J. C., Sánchez, M. B., and Martínez, J. L. doi: 10.1038/ncomms14078 (2014). Interplay between intrinsic and acquired resistance to quinolones in CLSI (2015). Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria Stenotrophomonas maltophilia. Environ. Microbiol. 16, 1282–1296. doi: 10. that Grow Aerobically. Approved Standard: M07 A10, 10th Edn. Wayne, PA: 1111/1462-2920.12408 Clinical and Laboratory Standards Institute. Gherardi, G., Creti, R., Pompilio, A., and Di Bonaventura, G. (2015). An overview CLSI (2019). Performance Standards for Antimicrobial Susceptibility Testing: of various typing methods for clinical epidemiology of the emerging pathogen M100-ED29, 29th Edn. Wayne, PA: Clinical and Laboratory Standards Stenotrophomonas maltophilia. Diagn. Microbiol. Infect. Dis. 81, 219–226. doi: Institute. 10.1016/j.diagmicrobio.2014.11.005 Corlouer, C., Lamy, B., Desroches, M., Ramos-Vivas, J., Mehiri-Zghal, E., González, J. F., Hahn, M. M., and Gunn, J. S. (2018). Chronic biofilm-based Lemenand, O., et al. (2017). Stenotrophomonas maltophilia healthcare- infections: skewing of the immune response. Pathog. Dis. 76:fty023. doi: 10. associated infections: identification of two main pathogenic genetic 1093/femspd/fty023 backgrounds. J. Hosp. Infect. 96, 183–188. doi: 10.1016/j.jhin.2017.02.003 Gröschel, M., Meehan, C., Barilar, I., Diricks, M., Gonzaga, A., Steglich, M., et al. Crossman, L. C., Gould, V. C., Dow, J. M., Vernikos, G. S., Okazaki, A., Sebaihia, (2020). The phylogenetic landscape and nosocomial spread of the multidrug- M., et al. (2008). The complete genome, comparative and functional analysis resistant opportunist Stenotrophomonas maltophilia. Nat. Commun. 11:2044. of Stenotrophomonas maltophilia reveals an organism heavily shielded by drug doi: 10.1038/s41467-020-15123-0 resistance determinants. Genome Biol. 9:R74. doi: 10.1186/gb-2008-9-4-r74 Hand, E., Davis, H., Kim, T., and Duhon, B. (2016). Monotherapy Davies, D. G., and Marques, C. N. H. (2009). A fatty acid messenger is responsible with minocycline or trimethoprim/sulfamethoxazole for treatment of for inducing dispersion in microbial biofilms. J. Bacteriol. 191, 1393–1403. Stenotrophomonas maltophilia infections. J. Antimicrob. Chemother. 71, doi: 10.1128/JB.01214-08 1071–1075. doi: 10.1093/jac/dkv456 Denton, M., and Kerr, K. G. (1998). Microbiological and clinical aspects of infection Hatziagorou, E., Orenti, A., Drevinek, P., Kashirskaya, N., Mei-Zahav, M., De associated with Stenotrophomonas maltophilia. Clin. Microbiol. Rev. 11, 57–80. Boeck, K., et al. (2019). Changing epidemiology of the respiratory bacteriology Devos, S., Van Oudenhove, L., Stremersch, S., Van Putte, W., De Rycke, R., Van of patients with cystic fibrosis-data from the European cystic fibrosis society Driessche, G., et al. (2015). The effect of imipenem and diffusible signaling patient registry. J. Cyst. Fibros. S1569–1993(19)30838-0. doi: 10.1016/j.jcf.2019. factors on the secretion of outer membrane vesicles and associated Ax21 08.006 proteins in Stenotrophomonas maltophilia. Front. Microbiol. 6:298. doi: 10.3389/ Hauben, L., Vauterin, L., Moore, E. R., Hoste, B., and Swings, J. (1999). Genomic fmicb.2015.00298 diversity of the genus Stenotrophomonas. Int. J. Syst. Bacteriol. 49(Pt 4), 1749– Di Bonaventura, G., Stepanović, S., Picciani, C., Pompilio, A., and Piccolomini, 1760. doi: 10.1099/00207713-49-4-1749 R. (2007). Effect of environmental factors on biofilm formation by clinical Hitchcock, P. J., and Brown, T. M. (1983). Morphological heterogeneity among Stenotrophomonas maltophilia isolates. Folia Microbiol. 52, 86–90. doi: 10.1007/ Salmonella lipopolysaccharide chemotypes in silver-stained polyacrylamide BF02932144 gels. J. Bacteriol. 154, 269–277. Dow, J. M., Crossman, L., Findlay, K., He, Y.-Q., Feng, J.-X., and Tang, J.-L. (2003). Huedo, P., Kumar, V. P., Horgan, C., Yero, D., Daura, X., Gibert, I., et al. (2019). Biofilm dispersal in Xanthomonas campestris is controlled by cell–cell signaling Sulfonamide-based diffusible signal factor analogs interfere with quorum and is required for full virulence to plants. Proc. Natl. Acad. Sci. U.S.A. 100, sensing in Stenotrophomonas maltophilia and Burkholderia cepacia. Future 10995–11000. doi: 10.1073/pnas.1833360100 Med. Chem. 11, 1565–1582. doi: 10.4155/fmc-2019-0015 EUCAST (2019). Breakpoint Tables for Interpretation of MICs and Zone Diameters Huedo, P., Yero, D., Martínez-Servat, S., Estibariz, I., Planell, R., Martínez, P., Version 9.0. Available online at: http://www.eucast.org/clinical_breakpoints/ et al. (2014). Two different rpf clusters distributed among a population of (accessed November, 2019). Stenotrophomonas maltophilia clinical strains display differential diffusible Falagas, M. E., Kastoris, A. C., Vouloumanou, E. K., Rafailidis, P. I., Kapaskelis, signal factor production and virulence regulation. J. Bacteriol. 196, 2431–2442. A. M., and Dimopoulos, G. (2009). Attributable mortality of Stenotrophomonas doi: 10.1128/JB.01540-14 maltophilia infections: a systematic review of the literature. Future Microbiol. 4, Huedo, P., Yero, D., Martinez-Servat, S., Ruyra, À, Roher, N., Daura, X., et al. 1103–1109. doi: 10.2217/fmb.09.84 (2015). Decoding the genetic and functional diversity of the DSF quorum- Falagas, M. E., Valkimadi, P.-E., Huang, Y.-T., Matthaiou, D. K., and Hsueh, P.- sensing system in Stenotrophomonas maltophilia. Front. Microbiol. 6:761. doi: R. (2008). Therapeutic options for Stenotrophomonas maltophilia infections 10.3389/fmicb.2015.00761 beyond co-trimoxazole: a systematic review. J. Antimicrob. Chemother. 62, Johnson, L. S., Eddy, S. R., and Portugaly, E. (2010). Hidden Markov model speed 889–894. doi: 10.1093/jac/dkn301 heuristic and iterative HMM search procedure. BMC Bioinformatics 11:431. Feil, E. J., Li, B. C., Aanensen, D. M., Hanage, W. P., and Spratt, B. G. (2004). doi: 10.1186/1471-2105-11-431 eBURST: inferring patterns of evolutionary descent among clusters of related Kaiser, S., Biehler, K., and Jonas, D. (2009). A Stenotrophomonas maltophilia bacterial genotypes from multilocus sequence typing data. J. Bacteriol. 186, multilocus sequence typing scheme for inferring population structure. 1518–1530. doi: 10.1128/jb.186.5.1518-1530.2004 J. Bacteriol. 191, 2934–2943. doi: 10.1128/JB.00892-08 Ferrer-Navarro, M., Planell, R., Yero, D., Mongiardini, E., Torrent, G., Huedo, Kent, W. J. (2002). BLAT–the BLAST-like alignment tool. Genome Res. 12, 656– P., et al. (2013). Abundance of the quorum-sensing factor Ax21 in four 664. doi: 10.1101/gr.229202 strains of Stenotrophomonas maltophilia correlates with mortality rate in a New Kim, E. J., Kim, Y. C., Ahn, J. Y., Jeong, S. J., Ku, N. S., Choi, J. Y., et al. (2019). Risk Zebrafish model of infection. PLoS One 8:e67207. doi: 10.1371/journal.pone. factors for mortality in patients with Stenotrophomonas maltophilia bacteremia 0067207 and clinical impact of quinolone-resistant strains. BMC Infect. Dis. 19:754. Fouhy, Y., Scanlon, K., Schouest, K., Spillane, C., Crossman, L., Avison, M. B., doi: 10.1186/s12879-019-4394-4 et al. (2007). Diffusible signal factor-dependent cell-cell signaling and virulence Kim, H. A., Ryu, S. Y., Seo, I., Suh, S.-I., Suh, M.-H., and Baek, W.-K. (2015). in the nosocomial pathogen Stenotrophomonas maltophilia. J. Bacteriol. 189, Biofilm formation and colistin susceptibility of Acinetobacter baumannii 4964–4968. doi: 10.1128/JB.00310-07 isolated from Korean Nosocomial samples. Microb. Drug Resist. Larchmt. N 21, Gales, A. C., Seifert, H., Gur, D., Castanheira, M., Jones, R. N., and Sader, 452–457. doi: 10.1089/mdr.2014.0236 H. S. (2019). Antimicrobial susceptibility of Acinetobacter calcoaceticus- Knirel, Y. A. (2011). “Structure of O-Antigens,” in Bacterial Lipopolysaccharides: Acinetobacter baumannii complex and Stenotrophomonas maltophilia clinical Structure, Chemical Synthesis, Biogenesis and Interaction with Host Cells, eds isolates: results from the SENTRY antimicrobial surveillance program (1997- Y. Knirel and M. Valvano (Vienna: Springer), 41–115. doi: 10.1089/mdr.2014. 2016). Open Forum Infect. Dis. 6, S34–S46. doi: 10.1093/ofid/ofy293 0236 Frontiers in Microbiology | www.frontiersin.org 13 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes Kumar, S., Stecher, G., and Tamura, K. (2016). MEGA7: molecular evolutionary Paez, J. I. G., and Costa, S. F. (2008). Risk factors associated with mortality genetics analysis version 7.0 for bigger datasets. Mol. Biol. Evol. 33, 1870–1874. of infections caused by Stenotrophomonas maltophilia: a systematic review. doi: 10.1093/molbev/msw054 J. Hosp. Infect. 70, 101–108. doi: 10.1016/j.jhin.2008.05.020 Letunic, I., and Bork, P. (2016). Interactive tree of life (iTOL) v3: an online tool for Patil, P. P., Kumar, S., Midha, S., Gautam, V., and Patil, P. B. (2018). the display and annotation of phylogenetic and other trees. Nucleic Acids Res. Taxonogenomics reveal multiple novel genomospecies associated with clinical 44, W242–W245. doi: 10.1093/nar/gkw290 isolates of Stenotrophomonas maltophilia. Microb. Genomics 4:e000207. doi: Liaw, S.-J., Lee, Y.-L., and Hsueh, P.-R. (2010). Multidrug resistance in 10.1099/mgen.0.000207 clinical isolates of Stenotrophomonas maltophilia: roles of integrons, efflux Pertea, M., Ayanbule, K., Smedinghoff, M., and Salzberg, S. L. (2009). OperonDB: pumps, phosphoglucomutase (SpgM), and melanin and biofilm formation. a comprehensive database of predicted operons in microbial genomes. Nucleic Int. J. Antimicrob. Agents 35, 126–130. doi: 10.1016/j.ijantimicag.2009. Acids Res. 37, D479–D482. doi: 10.1093/nar/gkn784 09.015 Petrosillo, N., Ioannidou, E., and Falagas, M. E. (2008). Colistin monotherapy Lira, F., Berg, G., and Martínez, J. L. (2017). Double-face meets the bacterial world: vs. combination therapy: evidence from microbiological, animal and clinical the opportunistic pathogen Stenotrophomonas maltophilia. Front. Microbiol. studies. Clin. Microbiol. Infect. 14, 816–827. doi: 10.1111/j.1469-0691.2008. 8:2190. doi: 10.3389/fmicb.2017.02190 02061.x Looney, W. J., Narita, M., and Mühlemann, K. (2009). Stenotrophomonas Pompilio, A., Crocetta, V., Ghosh, D., Chakrabarti, M., Gherardi, G., Vitali, L. A., maltophilia: an emerging opportunist human pathogen. Lancet Infect. Dis. 9, et al. (2016). Stenotrophomonas maltophilia phenotypic and genotypic diversity 312–323. doi: 10.1016/S1473-3099(09)70083-0 during a 10-year colonization in the lungs of a cystic fibrosis patient. Front. Lo-Ten-Foe, J. R., Smet, A. M. G. A., de Diederen, B. M. W., Kluytmans, J. A. J. W., Microbiol. 7:1551. doi: 10.3389/fmicb.2016.01551 and van Keulen, P. H. J. (2007). Comparative evaluation of the VITEK 2, disk Pompilio, A., Pomponio, S., Crocetta, V., Gherardi, G., Verginelli, F., Fiscarelli, E., diffusion, Etest, broth microdilution, and Agar dilution susceptibility testing et al. (2011). Phenotypic and genotypic characterization of Stenotrophomonas methods for colistin in clinical isolates, including heteroresistant Enterobacter maltophilia isolates from patients with cystic fibrosis: genome diversity, biofilm cloacae and Acinetobacter baumannii strains. Antimicrob. Agents Chemother. formation, and virulence. BMC Microbiol. 11:159. doi: 10.1186/1471-2180-11- 51, 3726–3730. doi: 10.1128/AAC.01406-06 159 Magiorakos, A.-P., Srinivasan, A., Carey, R. B., Carmeli, Y., Falagas, M. E., Rashid, M. H., and Kornberg, A. (2000). Inorganic polyphosphate is needed for Giske, C. G., et al. (2012). Multidrug-resistant, extensively drug-resistant swimming, swarming, and twitching motilities of Pseudomonas aeruginosa. and pandrug-resistant bacteria: an international expert proposal for interim Proc. Natl. Acad. Sci. U.S.A. 97, 4885–4890. doi: 10.1073/pnas.060030097 standard definitions for acquired resistance. Clin. Microbiol. Infect. Off. Publ. Rello, J., Kalwaje Eshwara, V., Lagunes, L., Alves, J., Wunderink, R. G., Conway- Eur. Soc. Clin. Microbiol. Infect. Dis. 18, 268–281. doi: 10.1111/j.1469-0691. Morris, A., et al. (2019). A global priority list of the TOp TEn resistant 2011.03570.x microorganisms (TOTEM) study at intensive care: a prioritization exercise Martínez-Servat, S., Yero, D., Huedo, P., Marquez, R., Molina, G., Daura, based on multi-criteria decision analysis. Eur. J. Clin. Microbiol. Infect. Dis. 38, X., et al. (2018). Heterogeneous colistin-resistance phenotypes coexisting in 319–323. doi: 10.1007/s10096-018-3428-y Stenotrophomonas maltophilia isolates influence colistin susceptibility testing. Rimler, R. B. (1990). Comparisons of Pasteurella multocida lipopolysaccharides Front. Microbiol. 9:2871. doi: 10.3389/fmicb.2018.02871 by sodium dodecyl sulfate-polyacrylamide gel electrophoresis to determine Mercier-Darty, M., Royer, G., Lamy, B., Charron, C., Lemenand, O., Gomart, C., relationship between group B and E hemorrhagic septicemia strains and et al. (2020). Comparative whole genome phylogeny of animal, environmental, serologically related group a strains. J. Clin. Microbiol. 28, 654–659. and human strains confirms the genogroup organization and diversity of the Rizek, C. F., Jonas, D., Garcia Paez, J. I., Rosa, J. F., Perdigão Neto, L. V., Stenotrophomonas maltophilia complex. Appl. Environ. Microbiol. 86:e02919- Martins, R. R., et al. (2018). Multidrug-resistant Stenotrophomonas maltophilia: 19. doi: 10.1128/AEM.02919-19 description of new MLST profiles and resistance and virulence genes using Mojica, M. F., Rutter, J. D., Taracila, M., Abriata, L. A., Fouts, D. E., Papp- whole-genome sequencing. J. Glob. Antimicrob. Resist. 15, 212–214. doi: 10. Wallace, K. M., et al. (2019). Population structure, molecular epidemiology, 1016/j.jgar.2018.07.009 and β-Lactamase diversity among Stenotrophomonas maltophilia isolates in the Rutter, W. C., Burgess, D. R., and Burgess, D. S. (2017). Increasing incidence of United States. mBio 10:e00405-19. doi: 10.1128/mBio.00405-19 multidrug resistance among cystic fibrosis respiratory bacterial isolates. Microb. Moskowitz, S. M., Garber, E., Chen, Y., Clock, S. A., Tabibi, S., Miller, A. K., Drug Resist. Larchmt. N 23, 51–55. doi: 10.1089/mdr.2016.0048 et al. (2010). Colistin susceptibility testing: evaluation of reliability for cystic Ryan, R. P., Fouhy, Y., Garcia, B. F., Watt, S. A., Niehaus, K., Yang, L., et al. (2008). fibrosis isolates of Pseudomonas aeruginosa and Stenotrophomonas maltophilia. Interspecies signalling via the Stenotrophomonas maltophilia diffusible signal J. Antimicrob. Chemother. 65, 1416–1423. doi: 10.1093/jac/dkq131 factor influences biofilm formation and polymyxin tolerance in Pseudomonas Mulcahy, H., Charron-Mazenod, L., and Lewenza, S. (2008). Extracellular DNA aeruginosa. Mol. Microbiol. 68, 75–86. doi: 10.1111/j.1365-2958.2008.06132.x chelates cations and induces antibiotic resistance in Pseudomonas aeruginosa Ryan, R. P., Monchy, S., Cardinale, M., Taghavi, S., Crossman, L., Avison, biofilms. PLoS Pathog. 4:e1000213. doi: 10.1371/journal.ppat.1000213 M. B., et al. (2009). The versatility and adaptation of bacteria from the genus Nayyar, C., Thakur, P., Tak, V., and Saigal, K. (2017). Stenotrophomonas Stenotrophomonas. Nat. Rev. Microbiol. 7:514. doi: 10.1038/nrmicro2163 maltophilia: an emerging pathogen in paediatric population. J. Clin. Diagn. Res. Sánchez, M. B. (2015). Antibiotic resistance in the opportunistic pathogen 11, DC08–DC11. doi: 10.7860/JCDR/2017/24304.9318 Stenotrophomonas maltophilia. Front. Microbiol. 6:658. doi: 10.3389/fmicb. Nicoletti, M., Iacobino, A., Prosseda, G., Fiscarelli, E., Zarrilli, R., De Carolis, 2015.00658 E., et al. (2011). Stenotrophomonas maltophilia strains from cystic fibrosis Santamaría, M., Gutiérrez-Navarro, A. M., and Corzo, J. (1998). patients: genomic variability and molecular characterization of some virulence Lipopolysaccharide profiles from nodules as markers of bradyrhizobium determinants. Int. J. Med. Microbiol. 301, 34–43. doi: 10.1016/j.ijmm.2010.07. strains nodulating wild legumes. Appl. Environ. Microbiol. 64, 902–906. 003 Steinmann, J., Mamat, U., Abda, E. M., Kirchhoff, L., Streit, W. R., Schaible, Nykyri, J., Mattinen, L., Niemi, O., Adhikari, S., Kõiv, V., Somervuo, P., U. E., et al. (2018). Analysis of phylogenetic variation of Stenotrophomonas et al. (2013). Role and regulation of the Flp/Tad pilus in the virulence maltophilia reveals human-specific branches. Front. Microbiol. 9:806. doi: 10. of pectobacterium atrosepticum SCRI1043 and pectobacterium wasabiae 3389/fmicb.2018.00806 SCC3193. PLoS One 8:e73718. doi: 10.1371/journal.pone.0073718 Stewart, P. S. (2015). Antimicrobial tolerance in biofilms. Microbiol. Spectr. 3. Ochoa-Sánchez, L. E., and Vinuesa, P. (2017). Evolutionary genetic analysis doi: 10.1128/microbiolspec.MB-0010-2014 uncovers multiple species with distinct habitat preferences and antibiotic Svensson-Stadler, L. A., Mihaylova, S. A., and Moore, E. R. B. (2012). resistance phenotypes in the Stenotrophomonas maltophilia complex. Front. Stenotrophomonas interspecies differentiation and identification by gyrB Microbiol. 8:1548. doi: 10.3389/fmicb.2017.01548 sequence analysis. FEMS Microbiol. Lett. 327, 15–24. doi: 10.1111/j.1574-6968. Olivares, J., Bernardini, A., Garcia-Leon, G., Corona, F. B., Sanchez, M., and 2011.02452.x Martinez, J. L. (2013). The intrinsic resistome of bacterial pathogens. Front. Tan, M. W., Mahajan-Miklos, S., and Ausubel, F. M. (1999). Killing of Microbiol. 4:103. doi: 10.3389/fmicb.2013.00103 caenorhabditis elegans by Pseudomonas aeruginosa used to model mammalian Frontiers in Microbiology | www.frontiersin.org 14 June 2020 | Volume 11 | Article 1160 Yero et al. S. maltophilia rpf Variants and Phenotypes bacterial pathogenesis. Proc. Natl. Acad. Sci. U.S.A. 96, 715–720. doi: 10.1073/ Waters, V. J., Gómez, M. I., Soong, G., Amin, S., Ernst, R. K., and pnas.96.2.715 Prince, A. (2007). Immunostimulatory properties of the emerging pathogen Tomich, M., Planet, P. J., and Figurski, D. H. (2007). The tad locus: postcards Stenotrophomonas maltophilia. Infect. Immun. 75, 1698–1703. doi: 10.1128/IAI. from the widespread colonization island. Nat. Rev. Microbiol. 5, 363–375. doi: 01469-06 10.1038/nrmicro1636 Wood, G. C., Underwood, E. L., Croce, M. A., Swanson, J. M., and Trifonova, A., and Strateva, T. (2019). Stenotrophomonas maltophilia – a low-grade Fabian, T. C. (2010). Treatment of recurrent Stenotrophomonas pathogen with numerous virulence factors. Infect. Dis. Lond. Engl. 51, 168–178. maltophilia ventilator-associated pneumonia with doxycycline and doi: 10.1080/23744235.2018.1531145 aerosolized colistin. Ann. Pharmacother. 44, 1665–1668. doi: 10.1345/apn. Valdezate, S., Vindel, A., Loza, E., Baquero, F., and Cantón, R. (2001). 1P217 Antimicrobial susceptibilities of unique Stenotrophomonas maltophilia clinical Zhuo, C., Zhao, Q., and Xiao, S. (2014). The impact of spgM, rpfF, strains. Antimicrob. Agents Chemother. 45, 1581–1584. doi: 10.1128/AAC.45.5. rmlA gene distribution on biofilm formation in Stenotrophomonas 1581-1584.2001 maltophilia. PLoS One 9:e108409. doi: 10.1371/journal.pone.010 Valdezate, S., Vindel, A., Martín-Dávila, P., Del Saz, B. S., Baquero, F., and Cantón, 8409 R. (2004). High genetic diversity among Stenotrophomonas maltophilia strains Zuschneid, I., Schwab, F., Geffers, C., Behnke, M., Rüden, H., and Gastmeier, despite their originating at a single hospital. J. Clin. Microbiol. 42, 693–699. P. (2007). Trends in ventilator-associated pneumonia rates within the doi: 10.1128/jcm.42.2.693-699.2003 German nosocomial infection surveillance system (KISS). Infect. Control Hosp. Vasileuskaya-Schulz, Z., Kaiser, S., Maier, T., Kostrzewa, M., and Jonas, D. (2011). Epidemiol. 28, 314–318. doi: 10.1086/507823 Delineation of Stenotrophomonas spp. by multi-locus sequence analysis and MALDI-TOF mass spectrometry. Syst. Appl. Microbiol. 34, 35–39. doi: 10.1016/ Conflict of Interest: The authors declare that the research was conducted in the j.syapm.2010.11.011 absence of any commercial or financial relationships that could be construed as a Vinuesa, P., Ochoa-Sánchez, L. E., and Contreras-Moreira, B. (2018). potential conflict of interest. GET_PHYLOMARKERS, a software package to select optimal orthologous clusters for phylogenomics and inferring Pan-Genome phylogenies, used for Copyright © 2020 Yero, Huedo, Conchillo-Solé, Martínez-Servat, Mamat, Coves, a critical Geno-Taxonomic revision of the genus Stenotrophomonas. Front. Llanas, Roca, Vila, Schaible, Daura and Gibert. This is an open-access article Microbiol. 9:771. doi: 10.3389/fmicb.2018.00771 distributed under the terms of the Creative Commons Attribution License (CC BY). Wang, Y. L., Scipione, M. R., Dubrovskaya, Y., and Papadopoulos, J. (2014). The use, distribution or reproduction in other forums is permitted, provided the Monotherapy with fluoroquinolone or trimethoprim-sulfamethoxazole for original author(s) and the copyright owner(s) are credited and that the original treatment of Stenotrophomonas maltophilia infections. Antimicrob. Agents publication in this journal is cited, in accordance with accepted academic practice. No Chemother. 58, 176–182. doi: 10.1128/AAC.01324-13 use, distribution or reproduction is permitted which does not comply with these terms. Frontiers in Microbiology | www.frontiersin.org 15 June 2020 | Volume 11 | Article 1160