Stenotrophomonas maltophilia Encodes a Type II Protein Secretion System That Promotes Detrimental Effects on Lung Epithelial Cells

Abstract

The Gram-negative bacterium Stenotrophomonas maltophilia is increasingly identified as a multidrug-resistant pathogen, being associated with pneumonia, among other infections. Despite this increasing clinical problem, the genetic and molecular basis of S. maltophilia virulence is quite minimally defined. We now report that strain K279a, the first clinical isolate of S. maltophilia to be sequenced, encodes a functional type II protein secretion (T2S) system. Indeed, mutants of K279a that contain a mutation in the xps locus exhibit a loss of at least seven secreted proteins and three proteolytic activities. Unlike culture supernatants from the parental K279a, supernatants from multiple xps mutants also failed to induce the rounding, detachment, and death of A549 cells, a human lung epithelial cell line. Supernatants of the xps mutants were also unable to trigger a massive rearrangement in the host cell's actin cytoskeleton that was associated with K279a secretion. In all assays, a complemented xpsF mutant behaved as the wild type did, demonstrating that Xps T2S is required for optimal protein secretion and the detrimental effects on host cells. The activities that were defined as being Xps dependent in K279a were evident among other respiratory isolates of S. maltophilia. Utilizing a similar type of genetic analysis, we found that a second T2S system (Gsp) encoded by the K279a genome is cryptic under all of the conditions tested. Overall, this study represents the first examination of T2S in S. maltophilia, and the data obtained indicate that Xps T2S likely plays an important role in S. maltophilia pathogenesis.

INTRODUCTION

Stenotrophomonas maltophilia is a Gram-negative bacterium found ubiquitously in soil, water, and plants and is increasingly being identified as an opportunistic and nosocomial pathogen (1–3). The most common type of infection is pneumonia followed by bloodstream infections although the bacterium has been associated with many other types of infection as well. S. maltophilia accounts for 4.5% of nosocomial pneumonia and 6% of ventilator-associated pneumonia and is reported to be among the 11 most isolated organisms in intensive care units (ICUs) in the United States (1, 3). Mortality rates for patients with S. maltophilia pneumonia are between 23 to 77%, while a separate study found that the overall attributable mortality rate for S. maltophilia infections is 37.5% (1, 4). Some of the risk factors for S. maltophilia infection are prolonged mechanical ventilation, presence of indwelling devices, compromised health status, malignancy, exposure to broad-spectrum antibiotics, and long-term hospitalization or ICU stays (1, 3). The incidence and prevalence of S. maltophilia are also increasing in cystic fibrosis (CF) patients in North America and Europe, with the prevalence of S. maltophilia being as high as 25% (1, 3, 5). Additionally, chronic S. maltophilia infection in CF patients is an independent risk factor for lung exacerbations (3, 6). Another reason for clinical concern is the intrinsic antibiotic resistance that S. maltophilia possesses, making infections difficult to treat (1, 3, 7, 8).

Despite the increasing clinical importance of S. maltophilia, our understanding of this bacterium's pathogenicity and virulence is very minimal. Phenotyping of S. maltophilia strains suggests that the organism has traits that are linked to the virulence of other bacteria (3, 5). Inoculation of S. maltophilia into the lungs of mice results in bacterial replication and a marked inflammatory response (9–11). However, documentation of the genetic basis of S. maltophilia pathogenicity is in its infancy. From the sequencing of the clinical isolate K279a, S. maltophilia is predicted to encode four types of protein secretions systems; i.e., types I, II, IV, and V (2, 12). Based upon myriad studies in other Gram-negative pathogens, one or more of these secretion systems is likely encoding virulence determinants.

Type II protein secretion (T2S) systems are common, although not universal, among Gram-negative bacteria (13). T2S is a multistep process (14–16). Proteins that are to be secreted are translocated across the inner membrane. In most cases, unfolded substrates cross that membrane via the Sec pathway; however, in some cases, folded substrates cross via the twin-arginine translocon. Once in the periplasm, unfolded substrates take on their tertiary conformation and may oligomerize. Finally, substrates are transported across the outer membrane by a complex of proteins that is dedicated to T2S. The T2S apparatus consists of 12 core proteins: a cytosolic ATPase (T2S E), inner membrane proteins that form a platform for T2S E (T2S F, L, and M), major and minor pseudopilins that form a pilus-like structure which spans the periplasm (T2S G, H, I, J, and K), an inner membrane peptidase that processes pseudopilins (T2S O), an outer membrane “secretin” that oligomerizes to form the secretion pore (T2S D), and a protein that appears to bridge inner and outer membrane factors (T2S C). The overall model is that substrates are recognized by the T2S apparatus, and then, using energy generated at the inner membrane, the pseudopilus acts like a piston to push the proteins through the secretin pore. T2S promotes the growth of environmental bacteria as well as the virulence of many human, animal, and plant pathogens (13–15, 17). Therefore, we initiated studies aimed at assessing the functionality of T2S in S. maltophilia and now report that the Xps T2S system of strain K279a mediates, among other things, detrimental effects on lung epithelial cells.

MATERIALS AND METHODS

Bacterial strains, media, and growth assays.

S. maltophilia strain K279a (American Type Culture Collection [ATCC] strain BAA-2423) served as our wild-type strain (Table 1). K279a is a multidrug-resistant strain that was isolated from the blood of a cancer patient (18). Mutants of K279a that were used in this study are listed in Table 1. Clinical isolates of S. maltophilia that had been previously obtained from patients were also included in this study (Table 1). Because of variations in the secreted activities produced by these isolates (see below), we used 16S rRNA sequencing to confirm the identity of the strains as S. maltophilia (data not shown). In performing this analysis, a fifth isolate (i.e., UPSm4) that had been previously reported to us as being S. maltophilia (9) proved to be a strain of Achromobacter xylosoxidans and therefore was not studied further here. S. maltophilia was routinely cultured at 37°C on Luria-Bertani (LB) agar (Becton, Dickinson, Franklin Lakes, NJ). When appropriate, medium was supplemented with chloramphenicol at 10 μg/ml, gentamicin (Corning, Tewksbury, MA) at 20 μg/ml, tetracycline at 20 μg/ml, norfloxacin at 5 μg/ml, 1 mM isopropyl β-1-thiogalactopyranoside (IPTG), or 10% sucrose. Growth of S. maltophilia was assessed by incubating strains in 25 ml (in a 125-ml flask) of buffered yeast extract (BYE) broth (19) at 37°C with agitation and monitoring optical densities of the cultures at 600 nm (OD₆₀₀) using a DU 720 spectrophotometer (Beckman Coulter, Indianapolis, IN). Escherichia coli DH5α (Life Technologies, Carlsbad, CA) was used as a host for recombinant plasmids. E. coli was grown in LB medium at 37°C. When appropriate, the medium was supplemented with ampicillin (Research Products International, Mt. Prospect, IL) at 100 μg/ml, chloramphenicol at 30 μg/ml, gentamicin at 5 μg/ml, tetracycline at 10 μg/ml, or 10% sucrose. Chemicals were from Sigma-Aldrich (St. Louis, MO), unless otherwise noted.

Table 1.

S. maltophilia strains used in this study

Strain	Description	Source or reference
K279a	Clinical isolate from blood	9, 18
NUS1	gspF mutant of K279a	This study
NUS2	xpsD mutant of K279a	This study
NUS3	gspF xpsD mutant of K279a	This study
NUS4	xpsF mutant of K279a	This study
NUS4(pBxpsF)	Complemented xpsF mutant	This study
UPSm1	Clinical isolate, tracheal aspirate	9
UPSm2	Clinical isolate, sputum	9
UPSm3	Clinical isolate, respiratory sinus	9
UPSm5	Clinical isolate, respiratory sinus	9

Detection of secreted enzymatic activities.

Initially, casein hydrolysis was determined by spotting a 5-μl aliquot of a bacterial suspension on Mueller-Hinton agar (Becton, Dickinson) containing 3% (wt/vol) skim milk (Nestle Carnation, Solon, OH) while gelatinase activity was tested by spotting bacteria on LB agar containing 4% (wt/vol) gelatin (20). Prior to being spotted on the indicator plates, bacteria were grown overnight on LB agar, resuspended in phosphate-buffered saline (PBS; Corning) to an OD₆₀₀ equal to 0.1, and diluted 1:10 in PBS. Plates were incubated at 37°C for 2 to 3 days. Whereas clearing due to caseinolytic activity was visible to the eye, the visualization of clearing due to gelatinase was aided by flooding the plate with ammonium sulfate for 10 min (21). In order to quantitate the levels of secreted enzymes, bacteria were grown in 25 ml of BYE broth for various periods of time, and then following centrifugation of the cultures, cell-free supernatants were obtained by filtering the culture supernatants through 0.22-μm-pore-size syringe filters (EMD Millipore, Billerica, MA). To measure caseinolytic activity, 100 μl of supernatant was incubated with 100 μl of 25 mg/ml azocasein powder in 0.1 M potassium phosphate buffer (pH 7.6). After 30 min, 800 μl of 5% (vol/vol) trichloroacetic acid was added, and then after centrifugation at 2,000 × g for 10 min, 100 μl of the supernatant was combined with 100 μl of 0.5N NaOH; absorbance was read at 440 nm using a Synergy H1 plate reader (BioTek, Winooski, VT) (22). Serine protease activity was determined as described previously (23). Briefly, 25 μl of supernatant was mixed with 100 μl of 0.5 mM N-succinyl-Ala-Ala-Pro-Phe-p-nitroaniline (pNA) in 20 mM sodium phosphate (pH 9.0) containing 400 mM NaCl. After 2 h of incubation at 37°C, the absorbance of the sample was read at 405 nm. The background absorbance from samples containing no substrate was subtracted from each sample determination.

Detection of secreted proteins.

S. maltophilia strains were grown in 300 ml of BYE broth (in 1-liter flasks) to early stationary phase, and culture supernatants were obtained as noted above. A total of 100 ml of supernatant was precipitated by the addition of 2 volumes of isopropanol at −20°C. After centrifugation of the sample at 10,000 × g, pellets were suspended in 10 ml of water and then concentrated another 20- to 40-fold by passage through a 30-kDa Amicon filter (EMD Millipore). Samples (100 μl) were then treated with a ReadyPrep 2D Cleanup Kit (Bio-Rad, Hercules, CA), according to the manufacturer's specifications. Finally, proteins were suspended in 2× Laemmli sample buffer containing 5% β-mercaptoethanol. The amount of protein in each sample was quantified using the RC DC (reducing agent and detergent compatible) Protein Assay (Bio-Rad). Samples containing equivalent amounts of protein were boiled for 5 min and subjected to electrophoresis through a 10% SDS-polyacrylamide gel. Protein bands were stained with a SilverQuest staining kit (Life Technologies) and compared with molecular weight standards. Images were converted to grayscale, and brightness and contrast were adjusted using Adobe Photoshop.

Examination of lung epithelial cells.

The human A549 cell line (ATCC CCL-185) was passaged in RPMI medium (Corning) containing 10% heat-inactivated fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA) at 37°C in 5% CO₂ (19). In order to obtain bacterial supernatants to be used on A549 cells, S. maltophilia strains were brought to an OD₆₀₀ of 0.215 in PBS and then diluted in RPMI medium to approximately 5 ×10⁵ CFU/ml. After 24 h of static incubation at 37°C in 5% CO₂, cell-free supernatants were obtained as above. Supernatants were used immediately or were stored at −20°C for periods of up to 1 month. To examine the effect of the supernatants on host cell rounding, monolayers containing 5 ×10⁵ A549 cells were established in the wells of a 24-well tissue culture plate (BD Falcon, Franklin Lakes, NJ), washed three times with fresh RPMI medium, and treated with 1 ml of supernatant for various periods of time; images were then captured using the 40× objective of an EVOS system (AMG, Life Technologies, Carlsbad, CA). Brightness and contrast of images were adjusted using ImageJ (NIH). To examine the effect of live bacteria on host cell rounding, monolayers containing 5 ×10⁵ A549 cells were established as above, washed three times with fresh RPMI medium, and treated with 1 ml of bacterial strains or medium alone. S. maltophilia strains were brought to an OD₆₀₀ of 0.215 in PBS and then diluted in RPMI medium to approximately 5 ×10⁵ CFU/ml. After 24 h of coculture, images were captured using a 40× objective.

To monitor effects on the actin cytoskeleton, 5 × 10⁵ A549 cells were first seeded onto a glass coverslip (Fisher Scientific, Pittsburgh, PA) placed within a 24-well tissue culture plate and incubated overnight. After the monolayers were washed three times with RPMI medium, 1 ml of supernatant was added for 1 h. The treated cells were fixed with 4% paraformaldehyde (vol/vol) (Electron Microscopy Sciences, Hatfield, PA) in PBS and then incubated for 20 min with a 1:300 dilution (in PBS) of Alexa Fluor 488-phalloidin (Life Technologies) and a 1:1,000 dilution (in PBS) of 5 mg/ml 4′,6-diamidino-2-phenylindole (DAPI) (Life Technologies). Coverslips were then mounted on slides with ProLong Gold (Life Technologies), and images were captured using the 60× objective of a Nikon C2+ confocal microscope (Nikon, Melville, NY). To monitor the effect of supernatants on the ability of A549 cells to remain attached, monolayers containing 2.5 × 10⁵ cells/well were maintained for 24 h, washed with RPMI medium, and then treated with supernatants. After 3 h, the monolayers were washed three times to remove the nonadherent cells. As described previously (24), the cells that had remained attached were treated with a 1:10 dilution of 1× trypsin (Corning) in PBS and then counted on a hemocytometer. To assess effects on host cell viability, A549 cells were seeded into wells within a 96-well tissue culture plate (BD Falcon) at a density of 5 ×10⁴ cells/well. After a series of washes, 90 μl of supernatant was added for an incubation period of 23 h. Ten microliters of PrestoBlue (Life Technologies) was added in a 1:10 dilution and incubated with cells at 37°C for 1 h. Fluorescence was read with an excitation of 560 nm and emission of 590 nm, and values obtained were normalized to those of untreated cells. A549 cell viability was also monitored using AlamarBlue (Life Technologies). Following the addition of 180 μl of supernatant and incubation for 20 h, 20 μl of AlamarBlue (1:10 dilution) was added, the plates were incubated for 4 h at 37°C, and then samples were read in the Synergy H1, as noted above.

DNA, RNA, and protein sequence analysis.

S. maltophilia DNA and RNA were isolated from early-stationary-phase BYE broth cultures using methods and reagents previously described (17). Primers used for sequencing, PCR, and reverse transcription-PCR (RT-PCR) were obtained from Integrated DNA Technologies (Coralville, IA). Primer names and sequences are listed in Table S1 in the supplemental material. For RT-PCR, the primer pair SK106 and SK107 was used to examine transcription of gspF, and the pair SK179 and SK180 and the pair SK181A and SK181B were used for xpsF. Control experiments in which the reverse transcriptase was omitted from the reaction mixture were done to rule out contributions from contaminating DNA. For PCR amplification of 16S rRNA gene sequences, the primers SK148 and SK149 were used (25, 26). The reaction mixtures for sequencing of the rRNA genes included the amplification primer pair, as well as internal sequencing primers SK150, SK151, SK152, SK153, SK154, SK155, SK156, SK157, SK158, and SK159. DNA sequences were analyzed using Lasergene (DNASTAR, Madison, WI). BLASTP homology searches were done using GenBank at the NCBI and the K279a database on the GenoList server (genodb.pasteur.fr/cgi-bin/WebObjects/GenoList).

Mutant construction and complementation.

Mutants of S. maltophilia K279a were constructed using the gene replacement vector pEX18Tc as described previously (27, 28). To obtain a mutant (NUS1) specifically lacking the gspF gene (i.e., Smlt2740 in the K279a genome database), the 5′ and 3′ ends of the gene and flanking DNA were separately PCR amplified from K279a DNA by using the primer pair SK87 and SK89 and the pair SK88 and SK90, respectively. Each of the generated fragments was ligated into pGEM-T Easy (Promega, Madison, WI), and then the two resulting plasmids were digested with SmaI and SacI. A trimolecular ligation was performed by placing a gentamicin resistance (Gm^r)-containing cassette, obtained from pX1918 (29) digested with PvuII and HincII, between the beginning and end of gspF. The plasmid thus obtained (i.e., pGΔgspF) carried a 920-bp deletion in the central gspF coding region. The ΔgspF fragment was ligated into pEX18Tc by digestion with EcoRI, yielding pEXΔgspF. pEXΔgspF was moved into E. coli S17-1 (30) and mobilized into S. maltophilia K279a via conjugation. Transconjugants were selected on LB agar supplemented with tetracycline, gentamicin, and norfloxacin. Emerging resistant colonies were streaked on LB agar supplemented with 10% sucrose and gentamicin. Mutation of gspF was confirmed by PCR using mutant strain DNA as a template with primers SK87 and SK88.

A similar strategy was employed for constructing an xpsD (Smlt0697) mutant (NUS2) and a gspF xpsD double mutant (NUS3). The 5′ and 3′ ends of the gene and flanking DNA were separately PCR amplified from K279a DNA using the primer pair SK91 and SK93 and the pair SK92 and SK94, respectively. Each of the generated fragments was ligated into pGEM-T Easy, and then the two resulting plasmids were digested with SacI and StuI. A trimolecular ligation was performed by placing a chloramphenicol resistance (Cm^r)-containing cassette, obtained from pRE112 (17) digested with ApaI, between the beginning and end of xpsD. This plasmid thus obtained (i.e., pGΔxpsD) carried a 1,972-bp deletion in the central xpsD coding region. The ΔxpsD fragment was PCR amplified out of pGEM-T Easy with primers SK112 and SK113. The resultant product was digested with KpnI and HindIII and ligated into pEX18Tc digested with the same enzymes, yielding pEXΔxpsD. pEXΔxpsD was moved into E. coli S17-1 and mobilized into K279a (for xpsD mutant) and gspF mutant NUS1 (for gspF xpsD double mutant) by conjugation. Transconjugants were selected on LB agar supplemented with tetracycline, chloramphenicol, and norfloxacin. Resistant colonies were streaked on LB agar supplemented with 10% sucrose and chloramphenicol. Mutation of xpsD was confirmed by PCR using primers SK91 and SK92.

A deletion mutant (NUS4) of xpsF (Smlt0688) was constructed using Flp-mediated excision as previously described (31). Briefly, xpsF with approximately 500 bp flanking on either side was PCR amplified with primers SK209 and SK210 and ligated into pGEM-T Easy, resulting in pGxpsF. The entire coding sequence for xpsF was replaced by a Flp recognition target (FRT)-flanked chloramphenicol cassette from pKD3 (31) amplified using primers SK205 and SK206 using recombineering and E. coli DY330 (31), resulting in the plasmid pGΔxpsF. The ΔxpsF construction was PCR amplified using primers SK211 and SK212. The resultant product was religated into pGEM-T Easy, and then ΔxpsF was digested out with BamHI and HindIII and ligated into pEX18Tc digested with the same enzymes, yielding pEXΔxpsF. pEXΔxpsF was moved into E. coli S17-1 and mobilized into K279a via conjugation. Transconjugants were selected on LB agar supplemented with tetracycline, chloramphenicol, and norfloxacin. Resistant colonies were streaked onto LB agar supplemented with 10% sucrose and chloramphenicol. Replacement of xpsF with chloramphenicol-flanked FRT sites (xpsF::frt-cat-frt strain SK3.2) was confirmed by PCR using primers SK213 and SK214. To perform Flp-mediated excision of the Cm^r cassette, pBSFlp (31) was electroporated into SK3.2, and transformants were selected on LB agar supplemented with gentamicin and IPTG. Individual colonies were patched onto LB agar containing either chloramphenicol or gentamicin or no selection. Colonies which were either chloramphenicol or gentamicin sensitive were streaked onto LB agar with 10% sucrose. Deletion of xpsF was confirmed by PCR using primers SK213 and SK214.

For trans-complementation of the xpsF mutant NUS4, a 1.3-kb PCR fragment containing the xpsF coding region plus 24 bp upstream was amplified from K279a DNA using primers SK213 and SK214. The resulting fragment was A-tailed using T4 polymerase (Life Technologies) and ligated into pGEM-T Easy, resulting in pGxpsFC′. pGxpsFC′ was digested with ApaI and SacI, and the resulting fragment was cloned into pBBR1MCS-5 (32) cut with the same enzymes, yielding pBxpsFC′. pBxpsFC′ was electroporated into the xpsF mutant NUS4 and transformants were selected on LB agar supplemented with gentamicin. Gm^r clones carrying pBxpsFC′ were confirmed by PCR utilizing SK214 and the vector-specific primer SK74.

RESULTS

S. maltophilia strains have two T2S loci.

Examination of the genome of the clinical isolate K279a (12) revealed the presence of two unlinked loci (i.e., gsp and xps) predicted to encode a T2S apparatus (Fig. 1A and B). Each locus had 11 T2S genes, corresponding to the core components T2S C through T2S M. In the gsp locus, gspCHIJ and gspFEDMLKG were separated by three unrelated genes (Fig. 1A). In the xps locus, the T2S genes occurred without interruption (Fig. 1B). Elsewhere in the K279a chromosome the Smlt3760 open reading frame encodes a predicted prepilin peptidase (T2S O) (Fig. 1C). All of the T2S genes were full-length, indicating that K279a has the potential to express two T2S apparatuses. RT-PCR analysis determined that both gsp and xps genes are expressed when K279a is grown on bacteriological medium (data not shown). We identified a similar set of T2S genes in the genomes of four other sequenced S. maltophilia strains, i.e., R551-3, D457, JV3, and SKA14 (see Table S2 in the supplemental material) (33, 34). Whereas SKA14 had its T2S genes arranged in a unique pattern, R551-3, D457, and JV3 exhibited gene synteny with K279a. Thus, two T2S loci were evident in both clinical (K279a and D457) and environmental (R551-3, JV3, and SKA14) isolates of S. maltophilia. When K279a protein sequences were used as the query in BLASTP analysis, the T2S proteins of S. maltophilia were most similar to those from species of Pseudomonas and Xanthomonas. This result is compatible with the genetic relationship that exists between Pseudomonas, Stenotrophomonas, and Xanthomonas (5). In summary, our bioinformatic analysis indicated that T2S is conserved in the S. maltophilia species. Furthermore, S. maltophilia strains appear to uniformly encode two T2S systems, as is the case for Pseudomonas aeruginosa and Xanthomonas campestris (35, 36).

Fig 1.

T2S loci in S. maltophilia strain K279a. Horizontal arrows denote the relative size and orientation of genes within the gsp locus (A), the xps locus (B), and a third locus encoding the prepilin peptidase (T2S O) gene (C). Genes shown in gray are those that encode T2S-related proteins, with the name of the gene being designated by its single-letter abbreviation; e.g., F for gspF or xpsF. Genes shown in white are those that encode a protein that does not contribute to the T2S apparatus and are identified by their locus numbers (e.g., Smlt2729). In panel A, the absence of open reading frames Smlt2736 and Smlt2739 reflects the designations given in the current genome database. In panel B, we indicated the presence of xpsC (rather than xpsN) in order to follow the precedent set in X. campestris whereby the gene originally annotated as xpsN was renamed xpsC (14).

Xps T2S mediates the secretion of multiple proteolytic activities.

When K279a was inoculated onto plates containing skim milk or gelatin, a zone of clearing developed around the areas of bacterial growth (Fig. 2), indicating that K279a secretes caseinolytic and gelatinase activities. To determine if one or both of the T2S loci contribute to these secreted activities, we made mutants of K279a that contained a mutation in the gsp and/or xps locus and then compared them to the parental wild type on the indicator plates (Fig. 2). A gspF mutant (i.e., NUS1) behaved as the wild-type K279a did, indicating that the Gsp T2S system is not required for the protease activities. In contrast, an xpsD mutant (NUS2) and an xpsF mutant (NUS4) displayed no or very small zones of clearing on the skim milk and gelatin plates. This reduction in secreted activities was also exhibited by NUS3, a mutant lacking both xpsD and gspF. These data indicated that, at least under these test conditions, Xps T2S mediates the secretion of caseinolytic and gelatinase activity. When an intact copy of xpsF was introduced into the NUS4 mutant, there was a restoration of secreted activity (Fig. 2), confirming that T2S is functional in S. maltophilia and that Xps T2S is required for secretion.

Fig 2.

Caseinolytic and gelatinase activities associated with wild-type and T2S mutant S. maltophilia strains. Strains as indicated were spotted onto an agar medium containing 3% skim milk (A) or 4% gelatin (B). Areas of growth and the associated presence or absence of zones of clearing were photographed after 2 (A) or 3 (B) days of incubation at 37°C. The data presented are representative of at least three independent experiments.

To quantitate the degree to which Xps T2S is responsible for secreted protease activities, we grew K279a and its various mutants in liquid medium, collected cell-free culture supernatants, and measured the levels of activity using two standard protease assays. In the course of this experiment, the xps and gsp mutants grew comparably to the parental K279a (see Fig. S1 in the supplemental material), indicating that mutations in the T2S loci do not result in a generalized growth defect. Wild-type supernatants effectively hydrolyzed azocasein, and gspF mutant supernatants displayed no loss of that activity (Fig. 3A). In contrast, the levels of activity in supernatants of the xpsD mutant, xpsF mutant, and gspF xpsD mutant were significantly lower than those of the wild type and equivalent to levels seen with medium alone. The complemented xpsF mutant had a level of activity that was comparable to that of the parental K279a (Fig. 3A). A similar result was obtained when we examined supernatants for serine protease activity; i.e., the ability to cleave Suc-Ala-Ala-Pro-Phe-pNA (Fig. 3B). Taken together, these data indicate that Xps T2S is responsible for all of the secreted caseinolytic activity and serine protease activity in K279a supernatants.

Fig 3.

Caseinolytic activity and serine protease activity in culture supernatants of wild-type and T2S mutant S. maltophilia. (A) Strains as indicated were grown in BYE broth at 37°C to late stationary phase, and then cell-free supernatants were examined for their ability to hydrolyze azocasein. (B) Strains as indicated were grown in BYE broth at 37°C to early stationary phase, and then cell-free supernatants were examined for activity against the substrate Suc-Ala-Ala-Pro-Phe-pNA. Data are the means and standard deviations from duplicate culture supernatants, and the results presented are representative of at least three independent experiments. In both panels, the levels of activity exhibited by the xps mutants were significantly less than those of the wild type and the other strains examined (P < 0.02; Student's t test).

Xps T2S mediates the secretion of multiple proteins.

As an alternate means of judging the contribution of Xps and Gsp to secretion by S. maltophilia, supernatants from broth cultures of K279a and its mutants were examined by SDS-PAGE (Fig. 4). Wild-type K279a exhibited the presence of seven or more protein bands, ranging in size from approximately 25 to 66 kDa. All of these protein species were absent from supernatants of the xpsD mutant and xpsF mutant (Fig. 4A). Complementation of the xpsF mutant resulted in a pattern of secreted proteins that was akin to that of the wild type (Fig. 4A). When the gspF mutant's supernatant was examined, there was no loss of protein bands (Fig. 4B). As expected, the gspF xpsD double mutant produced a supernatant that was devoid of protein species (data not shown). These data indicate that S. maltophilia secretes at least seven different proteins and that Xps T2S is responsible for the secretion of all of those protein species.

Fig 4.

Proteins present in culture supernatants of wild-type and T2S mutant S. maltophilia. Strains as indicated were grown in BYE broth at 37°C to early stationary phase, and then concentrated supernatants were electrophoresed through a 10% SDS-polyacrylamide gel and silver stained. The gels were loaded with 8 μg (A) or 5 μg (B) of total protein. The migration of molecular mass standards (in kDa) is indicated to the left of the gel images. Some of the protein bands present in the wild-type samples but not the xps mutant samples are denoted by asterisks. The very faint bands that appear in the xpsD mutant lane are likely the result of spillover from the adjacent wild-type lane since these bands were not observed in other gels that had the xpsD mutant lane placed farther apart from the wild-type lane. Overall, the data presented are representative of at least three independent experiments.

Xps T2S causes structural and viability changes in lung epithelial cells.

As a first step toward assessing the role of T2S in S. maltophilia pathogenesis, we compared the wild type and mutant strains for their effects on a human cell line. Given the rise of S. maltophilia as a pathogen involved in pneumonia, we used as the cell target A549 cells, a human cell line of type II lung epithelial cells that is widely used to investigate respiratory pathogens (19, 24, 37–39). Initially, we examined the effect of coculturing bacteria with the A549 cells. Whereas monolayers incubated with wild-type K279a or the gsp mutant were rounded after 24 h, monolayers incubated with an xps mutant appeared analogous to monolayers treated with medium alone (see Fig. S2 in the supplemental material). When supernatants obtained from K279a cultures were added to a monolayer of A549 cells, the epithelial cells began to round up after 1 h of incubation, and after 3 h of incubation all cells appeared rounded (Fig. 5A). Whereas supernatants from gspF mutant cultures produced a pattern of rounding similar to that of the wild type, supernatants obtained from the xpsD mutant, the xpsF mutant, and the gspF xpsD double mutant did not cause the A549 cells to round (Fig. 5A) even after 24 h of incubation (data not shown). The complemented xpsF mutant gave results similar to those with the wild type (Fig. 5A). Together, these data indicate that Xps T2S is required for the ability of S. maltophilia to trigger rounding of A549 cells.

Fig 5.

A549 cell morphology after treatment with supernatants from wild-type and T2S mutant S. maltophilia strains. Strains as indicated were grown in RPMI medium at 37°C in the presence of 5% CO₂ for 24 h, and then cell-free culture supernatants (as well as medium controls) were obtained and added to A549 cell monolayers (n = 3). (A) Following 1 and 3 h of incubation, the morphology of the treated A549 cells was determined by phase-contrast light microscopy. Scale bar, 100 μm. The data presented are representative of at least three independent experiments. (B) After 1 h of incubation with supernatants or medium control, A549 cells were fixed and treated with Alexa Fluor 488-phalloidin to label actin (left column) and DAPI to label nuclei (right column) and then visualized by confocal fluorescence microscopy. Scale bar, 20 μm. Data are representative of at least two independent experiments.

Cell rounding is often associated with changes in the actin cytoskeleton (24). To determine if K279a and its T2S-dependent proteins induce changes in host actin cytoskeleton, A549 cells were treated with supernatants and then stained for F-actin. Whereas untreated A549 cells exhibited typical actin filaments and stress fibers, cells treated with K279a supernatants displayed an absence of stress fibers, and the actin was marginalized to the periphery of the cell (Fig. 5B). These data indicate, for the first time, that S. maltophilia secretes a factor(s) that results in major changes in the host actin cytoskeleton. Supernatants obtained from the gspF mutant gave a pattern similar to that of the wild type (Fig. 5B), indicating that Gsp T2S is not required for cell rounding or cytoskeletal rearrangements. In contrast, supernatants from the xpsD mutant, the xpsF mutant, and the gspF xpsD double mutant did not lead to changes in the actin cytoskeleton (Fig. 5B). The inability of the xpsF mutant to trigger changes was reversed when an intact copy of xpsF was introduced into the mutant (Fig. 5B). Thus, Xps T2S mediates the secretion of the factor(s) responsible for rounding and actin rearrangement.

Given the effect of Xps T2S on cell morphology and actin, we posited that secreted factors would also cause A549 cells to detach from their substrata. After 3 h of incubation with supernatants from K279a, only about 10% of A549 cells remained attached (Fig. 6). Whereas supernatants obtained from the gspF mutant caused a level of detachment similar to that of the wild type, supernatants from the xpsD mutant, xpsF mutant, and gspF xpsD double mutant failed to cause any detachment (Fig. 6). The complemented xpsF mutant gave results similar to those of the wild type (Fig. 6). Together, these data indicate that the Xps T2S system of S. maltophilia can cause host cells to detach from surfaces.

Fig 6.

Attachment of A549 cells after treatment with supernatants from wild-type and T2S mutant S. maltophilia strains. Strains as indicated were grown in RPMI medium at 37°C in the presence of 5% CO₂ for 24 h, and then cell-free culture supernatants (as well as medium controls) were obtained and added to A549 cell monolayers. Following 3 h of incubation, the numbers of A549 cells that still remained attached to the wells of the microtiter plate were determined. Values were normalized to the monolayers that had been treated with medium alone. Data are the means and standard deviations from three treated monolayers, and the results presented are representative of at least three independent experiments. Levels of attachment after treatment with supernatants from the wild type, the gspF mutant, or the complemented xpsF mutant were significantly reduced compared to those of the untreated controls (P ≤ 0.001; Student's t test).

Next, to determine whether T2S-dependent activities promoted a cytotoxic effect, we assessed the viability of the A549 cells following their treatment with supernatants (Fig. 7). At 24 h, 30 to 40% of cells exposed to K279a products lost reactivity with vital stains, indicating that the wild-type strain secretes a factor(s) that, directly or indirectly, promotes loss of viability. Supernatants from the gspF mutant elicited a similar level of cell death, implying that Gsp T2S does not encode the cytotoxic factor(s). However, supernatants from the xpsD mutant, xpsF mutant, and gspF xpsD mutant failed to induce a loss of cell viability. The complemented xpsF mutant triggered cell death to a degree that was comparable to levels with the wild type and the gspF mutant. Thus, Xps T2S of S. maltophilia mediates the secretion of a factor(s) that leads to the death of lung epithelial cells.

Fig 7.

Viability of A549 cells after treatment with supernatants from wild-type and T2S mutant S. maltophilia. Strains as indicated were grown in RPMI medium for 24 h, and then cell-free culture supernatants were obtained and added to A549 cell monolayers. After 24 h of incubation, the numbers of viable A549 cells that still remained in the well were determined by staining with PrestoBlue. Values were normalized to cells treated with medium alone. The levels of cell viability after treatment with supernatants of the wild type, the gspF mutant, or the complemented xpsF mutant were significantly reduced compared to those of the untreated controls (P ≤ 0. 008; Student's t test). Data are the means and standard deviations from three treated monolayers and are representative of two experiments using PrestoBlue. A similar result was obtained in a third trial that utilized the vital stain AlamarBlue (data not shown).

Xps-dependent phenotypes are expressed to different degrees by other clinical isolates.

Four isolates of S. maltophilia obtained from the respiratory tract (Table 1) were examined for caseinolytic activity, serine protease activity, and the ability to cause A549 cells to round (Table 2). In terms of caseinolytic activity, strains UPSm1 and UPSm2 displayed zones of clearing on skim milk plates that were similar in size to those of K279a. In contrast, UPSm3 exhibited no clearing, and UPSm5 gave more clearing. For serine protease activity, UPSm1, UPSm2, and UPSm5 had levels of activity similar to those of strain K279a. However, UPSm3 displayed a reduced level of serine protease. Lastly, when the ability of bacterial supernatants to cause rounding was examined, UPSm1, UPSm3, and UPSm5 behaved similarly to K279a, whereas supernatants from UPSm2 produced rounding only after 3 h or more of incubation. Together, these data indicate that multiple activities ascribed to Xps T2S in K279a are expressed by the other clinical isolates and that the T2S system is likely to be functional in many strains of S. maltophilia. That a single activity was absent or poorly expressed in certain strains (e.g., the lack of casein hydrolysis in UPSm3) suggests that the gene encoding the Xps-dependent exoprotein is not always present or well expressed.

Table 2.

T2S-dependent activities expressed by strains of S. maltophilia

Strain	Relative level of the indicated Xps-dependent activitya
	Casein hydrolysis	Serine protease	A549 rounding
K279a	++	++	++
UPSm1	++	++	++
UPSm2	++	++	+
UPSm3	−	+	++
UPSm5	+++	++	++

^a+++, greatest activity; ++, intermediate activity; +, lower activity; −, no activity.

DISCUSSION

The data presented here represent the first experimental definition of a protein secretion system in S. maltophilia. That Xps T2S was shown to promote damage of lung epithelial cells also marks an important development in our understanding of the genetic basis of S. maltophilia pathogenicity. Although many advances have been made in our understanding of the genetic basis of antibiotic resistance in S. maltophilia (40), only a few previous studies have used mutants to define (potential) virulence factors of S. maltophilia (3). Key earlier studies focused on the role of flagella in bacterial adherence to bronchial epithelial cells in vitro, the impact of a diffusible, signal factor (cis-Δ2-11-metyl-dodecenoic acid) and a secreted cell-signaling protein (Ax21) in nematode and moth models of toxicity and virulence, the connection between phosphoglucomutase and lipopolysaccharide (LPS) and virulence in a rat lung model of infection, and the need for the RNA chaperone Hfq in adherence to bronchial cells (27, 41–44). Thus, embarking upon a systematic assessment of protein secretion systems in S. maltophilia is significant when one considers the current state of the S. maltophilia pathogenesis field.

Based upon our analysis of culture supernatants from strain K279a, Xps T2S mediates the secretion of at least seven proteins and three types of proteolytic activity. Given the conservation of xps genes among four other S. maltophilia genomes, we infer that Xps T2S is active in other strains of S. maltophilia. In support of this hypothesis, we along with other investigators have found protease activities in supernatants from a variety of other strains of S. maltophilia (3, 20, 23, 45, 46). The ∼47-kDa protein that we observed in the K279a supernatant but not in the xps mutants' supernatants is likely to be StmPr1, a 47-kDa serine protease whose gene sequence encodes a signal sequence (23). That Xps T2S of K279a mediates the secretion of this number of proteins and these types of degradative enzymes as well as having substrates in the size range of 25 to 66 kDa is entirely compatible with what is known about T2S in other Gram-negatives (13). From work done in other bacteria, where the number of T2S-dependent substrates can be ≥25 (47–50), we hypothesize that the output of S. maltophilia Xps T2S should prove to be greater than what we found in this initial report. Future studies using different growth conditions (e.g., other media or temperatures) and incorporating additional enzymatic assays (e.g., degradation of lipid, nucleic acids, and carbohydrates) are likely to uncover more Xps substrates.

We have documented that strain K279a secretes an Xps-dependent factor(s) that damages A549 cells. The detrimental effects were cell rounding, actin rearrangement, and cell detachment within a 3-h time period as well as cell death after 24 h. Compatible with these data are two past studies that reported that supernatants from other S. maltophilia strains elicit rounding and death of HEp-2 (human larynx), HeLa (human cervix), and Vero (African green monkey) cells as well as rounding and detachment of human fibroblasts (23, 45). Though it is tempting to speculate that the protease activities that we identified in K279a supernatants are responsible for the effects on A549 cells, one of the previous studies found that the toxic activity, though sensitive to heat (56°C) treatment, was resistant to protease inhibitors, including those that impede serine proteases (45). On the one hand, it is possible that the different effects on A549 cells that we observed are interconnected and due to a single secreted protein. For example, an Xps-dependent substrate might enter into the host cell and alter the actin cytoskeleton, and then this triggers rounding and detachment and ultimately death. Alternatively, a single exoprotein acts externally and affects matrix material and/or a surface receptor, and this leads to changes in internal signaling and ultimately morphological and viability changes. On the other hand, it is also possible that the effects seen derive from the independent action of multiple Xps substrates; e.g., one exoprotein triggers relatively rapid morphological changes, and another more slowly induces the cell death. Thus, future work will be aimed at identifying new Xps substrates that are responsible for damaging lung epithelial cells and determining how they achieve their effect on the host cell.

We have found that all S. maltophilia strains examined encode two T2S loci. However, all of the secreted proteins/activities of K279a that were identified in this paper were dependent upon Xps T2S; nothing was ascribed to Gsp T2S. This finding is akin to the current situation in X. campestris pv. vesicatoria (36). In contrast, in P. aeruginosa, both the Xcp and Hxc systems are known to be active, with Xcp being more broadly expressed and Hxc showing activity under phosphate-limiting growth conditions (35). Thus, it is quite possible that the Gsp T2S system of S. maltophilia is functional but under different growth conditions.

T2S contributes to the virulence of several human pathogens, including E. coli, Legionella pneumophila, P. aeruginosa, Yersinia enterocolitica, and Vibrio cholerae (14). Notable T2S-dependent substrates include cholera toxin of V. cholerae, exotoxin A of P. aeruginosa, and heat-labile toxin of enterotoxigenic E. coli (15). Other pathogenic processes linked to T2S include adherence to host cells, resistance to complement, and biofilm formation by various pathogenic E. coli, as well as lung infection, intracellular infection of macrophages, and dampening of the host innate immune response by L. pneumophila (13, 19, 51). Given this precedent as well as the connection between Xps T2S and damage to lung epithelial cells, the T2S system(s) of S. maltophilia likely promotes human infection and bacterial virulence. The Xps substrate(s) that damages A549 cells in vitro likely also damages epithelial cells in the infected respiratory tract, perhaps leading to increased bacterial spread and loss of lung function. Other plausible targets for Xps T2S are polymorphonuclear leukocytes (PMNs) and other phagocytes that attempt to contain S. maltophilia in the lung, bloodstream, or other body sites. Examining the behavior of the S. maltophilia T2S mutants in mammalian (murine) models of infection (9–11) will be an important next step in testing this hypothesis.