Syndecan-1 shedding is enhanced by LasA, a secreted virulence factor of Pseudomonas aeruginosa.

Microbial pathogens frequently take advantage of host systems for their pathogenesis. Shedding of cell surface molecules as soluble extracellular domains (ectodomains) is one of the host responses activated during tissue injury. In this study, we examined whether pathogenic bacteria can modulate shedding of syndecan-1, the predominant syndecan of host epithelia. Our studies found that overnight culture supernatants of Pseudomonas aeruginosa and Staphylococcus aureus enhanced the shedding of syndecan-1 ectodomains, whereas culture supernatants of several other Gram-negative and Gram-positive bacteria had only low levels of activity. Because supernatants from all tested strains of P. aeruginosa (n = 9) enhanced syndecan-1 shedding by more than 4-fold above control levels, we focused our attention on this Gram-negative bacterium. Culture supernatants of P. aeruginosa increased shedding of syndecan-1 in both a concentration- and time-dependent manner, and augmented shedding by various host cells. A 20-kDa shedding enhancer was partially purified from the supernatant through ammonium sulfate precipitation and gel chromatography, and identified by N-terminal sequencing as LasA, a known P. aeruginosa virulence factor. LasA was subsequently determined to be a syndecan-1 shedding enhancer from the findings that (i) immunodepletion of LasA from the partially purified sample resulted in abrogation of its activity to enhance shedding and (ii) purified LasA increased shedding in a concentration-dependent manner. Our results also indicated that LasA enhances syndecan-1 shedding by activation of the host cell's shedding mechanism and not by direct interaction with syndecan-1 ectodomains. Enhanced syndecan-1 shedding may be a means by which pathogenic bacteria take advantage of a host mechanism to promote their pathogenesis.

as soluble ectodomains and include cytokines, growth factors, and their receptors, and cell adhesion molecules such as selectins (4), CD14 (5), epidermal growth factor (6), TNF-␣ 1 (7,8) and its receptors (9,10), IL-6 receptor (11), and transforming growth factor-␣ (12), to name a few. These shed ectodomains play pivotal roles in diverse pathophysiological events including septic shock, host defense, and wound healing. Furthermore, because shedding itself has been found to be controlled by various extracellular ligands (13)(14)(15) and intracellular signaling pathways (3,12,15,16), it provides an additional level of regulation. Because protein kinase C agonists (phorbol ester) and peptide hydroxamates have been found to enhance and inhibit the shedding of most affected molecules, respectively, existence of a common shedding system has been proposed (3). However, shedding of some effectors is insensitive or only partially sensitive to hydroxamates (13,17) and additional regulators of shedding have been identified (12,16), 2 suggesting that certain components may be unique to individual shedding systems.
The genetic variability that pathogenic microorganisms can generate have allowed variant pathogens to take advantage of the host environment for their growth and survival. For example, a diverse group of pathogens including Yersinia spp. (19), Bordetella pertussis (20,21), and adenovirus (22) express RGDcontaining cell surface ligands and use these "molecular mimics" to interact with host integrin receptors for their colonization (23). Bacteria also produce molecules that can derange host homeostasis to their benefit. Several bacteria secrete toxins that can modify the host cell cytoskeleton (24) and secrete enzymes that can degrade extracellular matrix components, immunoglobulins and complement, either directly (25,26) or indirectly by activating the matrix metalloproteinases in the host (27). Furthermore, lipopolysaccharide from Gram-negative bacteria, the causative agent of endotoxic shock, affects the expression of host defense effectors such as TNF-␣ and IL-1, -6, -8, and -10 (28).
Recent studies indicate that bacterial pathogens may also utilize the host cell's shedding mechanism to enhance their virulence. For instance, the pore-forming toxins, streptolysin O and Escherichia coli hemolysin, trigger shedding of lipopolysaccharide (CD14) and IL-6 receptors (29). Culture supernatants from Pseudomonas aeruginosa, Staphylococcus aureus, Serratia marcescens, and Listeria monocytogenes can also augment shedding of the IL-6 receptor (30), and culture supernatants from S. epidermidis can activate shedding of TNF-␣ (31), although the responsible shedding enhancers were not defined in these studies. Furthermore, increased serum levels of soluble ectodomains of several surface effectors, such as CD14, TNF-␣, and IL-4 receptors, have been documented during infection (32)(33)(34). These findings suggest that bacteria-enhanced shedding can modulate the activation and function of host effectors, and play a role in bacterial pathogenesis.
The syndecans are a family of cell surface heparan sulfate proteoglycans which, along with the glypicans, are the major source of cell surface heparan sulfate (35). There are currently four mammalian syndecans known, syndecan-1 through -4, each encoded by distinct genes. Syndecans can bind and modulate the activity of a diverse group of soluble and insoluble ligands, such as extracellular matrix components, growth factors, chemokines, cytokines, and proteases, through the action of their heparan sulfate chains. Syndecans have also been proposed to act as adhesion and internalization receptors for pathogenic microorganisms (36,37).
The extracellular domains of syndecans can be shed as soluble, intact heparan sulfate proteoglycan ectodomains which, because they bind the same ligands as their precursor proteoglycans on the cell surface, can serve as soluble effectors. For example, shed syndecan-1 ectodomains have been found to regulate the proliferative response of cells to FGF-2 (38) and potentiate the activity of neutrophil enzymes, such as elastase and cathepsin G (39), by binding to the enzymes and protecting them from inhibition by their physiological inhibitors. All syndecans are shed constitutively as part of normal syndecan turnover, but available evidence also indicates that syndecan shedding is a regulated host response to tissue injury and that shed syndecan ectodomains are regulators of inflammation (15). Thus, regulation of syndecan shedding by pathogenic bacteria may play a role in pathogenesis through alteration of the host response to infection and/or the pathogen's ability to colonize host tissue components.
We have studied whether pathogenic bacteria can modulate syndecan shedding and have found that culture supernatants from S. aureus and P. aeruginosa, but not from several other Gram-positive and Gram-negative bacteria, enhance shedding of syndecan-1 by host cells. Here we report the characterization of syndecan-1 shedding enhanced by P. aeruginosa. Syndecan-1 shedding augmented by overnight culture supernatants of P. aeruginosa is rapid, is seen with various types of host cells, and produces intact, soluble syndecan-1 ectodomains. A P. aeruginosa shedding enhancer has been purified from a clinical isolate and identified as the mature 20-kDa LasA protein, a known virulence factor of P. aeruginosa (40 -42). LasA-enhanced shedding produces syndecan-1 ectodomain core proteins identical in size to ectodomains shed endogenously, and is inhibited by inhibitors of the host cell's shedding mechanism. These results indicate that LasA enhances syndecan-1 shedding by activating the host cell's shedding machinery. Enhancement of syndecan-1 shedding by LasA may be a mechanism by which P. aeruginosa parasitizes a host system to aid their pathogenesis.

Materials-Affi-Prep
Hz Hydrazide affinity chromatography resins, Bio-Gel P-30 gel chromatography resins, Coomassie Brilliant Blue R-250, and pre-stained SDS-PAGE size standards were purchased from Bio-Rad. Bisindolylmaleimide I, genistein, and Tyrphostin A25 were from Calbiochem (La Jolla, CA). Heparan sulfate lyase (heparin lyase III, heparitinase) and chondroitin sulfate ABC lyase were obtained from Seikagaku (Falmouth, MA). Tryptic soy broth and tryptic soy agar were purchased from Remel (Lenexa, KS). The cationic polyvinylidene difluoride membrane, Immobilon N, was from Millipore (Bedford, MA) and ProBlott polyvinylidene difluoride membrane for N-terminal sequenc-ing was from Applied Biosystems (Foster City, CA). Tissue culture media and supplements other than serum were from Mediatech (Herndon, VA), fetal calf and calf serum were from HyClone (Logan, UT) and tissue culture plastics were from Costar (Corning, NY). Enhanced chemiluminescense (ECL) Western blotting detection reagents DEAE Sephacel were from Amersham Pharmacia Biotech, and molecular weight cutoff (MWCO) spin tubes were from Pall Filtron (Northborough, MA). TPCK-treated trypsin, soybean trypsin inhibitor, and all other materials were purchased from Sigma.
P. aeruginosa laboratory strains 7700 and 10145 were from the ATCC. The clinical P. aeruginosa isolates, BL1, BL2, CF1, CF2, and SP1, were from the Division of Infectious Diseases at Washington University School of Medicine (St. Louis, MO), CT4 was kindly provided by Dr. David Roberts at the NCI (Rockville, MD) (44) and PAO1 was from our culture collection (40 The rat monoclonal anti-mouse syndecan-1 ectodomain antibody (281-2) was generated in the our laboratory (45) and is now commercially available from Pharmingen (San Diego, CA). The purified LasA protein used for antisera production was prepared previously (40). Briefly, LasA was purified by ammonium sulfate precipitation (80%), DEAE ion exchange chromatography and further fractionation with a sulfopropyl column. C3H/Hen mice (Charles River Laboratories, Wilmington, MA) were immunized subcutaneously with 10 g of purified LasA emulsified in complete Freund's adjuvant. Mice were given a second injection of 5 g of LasA emulsified in incomplete Freund's adjuvant 2 weeks later. Two weeks after the second immunization, mice were bled via the dorsolateral tail vein and sera prepared. Non-immune sera were prepared by immunizing mice with complete Freund's adjuvant followed 2 weeks later with incomplete Freund's adjuvant as described above. Horseradish peroxidase-conjugated goat anti-rat secondary antibodies were obtained from either Jackson Immunoresearch (West Grove, PA) or Cappel (Durham, NC).
Syndecan-1 Shedding Assays-Quantification of syndecan-1 shedding was performed as described previously (15). Briefly, confluent or 1-day post-confluent cultures of NMuMG and C127 cells in 96-well plates and LA-4 and NIH3T3 cells in 24-well plates were washed once with their respective culture media, and various test samples diluted in culture media were added to the cells. Cells were incubated at 37°C with the samples for 6 h or for the indicated times as described in the figure legends. Cell viability was measured with the tetrazolium salt (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) conversion assay (46). For quantification of shedding, the culture supernatants were collected, spun down to remove cells, and the cell-free supernatants were applied to Immobilon N membranes using a dot immunoblotting apparatus. The samples were acidified by adding NaOAc (pH 4.5), NaCl, and Tween 20 to final concentrations of 50 mM, 150 mM, and 0.1% (v/v), respectively. By acidifying the samples, only highly anionic molecules such as proteoglycans are retained by the cationic polyvinylidene difluoride membrane (Immobilon N) while most proteins pass through the membrane during dot blotting. To obtain measurements within the linear range of the dot immunoblotting method, 70 l out of 100 l in each well of 96-well plates were applied for NMuMG and C127 cells, and 200 l out of 500 l in each well of 24-well plates were applied for LA-4 and NIH3T3 cells. For quantification of cell surface syndecan-1, following removal of media with or without test samples, cells were washed once with ice-cold TBS (50 mM Tris, pH 7.5, 150 mM NaCl) containing 0.5 mM EDTA and incubated for 15 min at 4°C with ice-cold 10 g/ml TPCK-treated trypsin in TBS with 0.5 mM EDTA. Trypsin was subsequently inactivated with 100 g/ml soybean trypsin inhibitor and the reaction mixture was spun down to remove detached cells. For NMuMG cells, 15 l out of 100 l of trypsinate were blotted onto Immobilon N membranes as described above to obtain measurements in the linear range of the dot immunoblotting method. The blotted membranes were developed by sequential incubations at 4°C with (i) 10% (w/v) non-fat dry milk in TBS for 2 h or longer (blocking), (ii) 0.2 g/ml of anti-syndecan-1 antibody (281-2) in BLOTTO (TBS containing 0.5% non-fat dry milk and 0.1% Tween 20) for 14 -24 h, (iii) BLOTTO for 30 min ϫ 2 (wash), (iv) 1:8,000 dilution of horseradish peroxidase-conjugated goat anti-rat antibodies in BLOTTO for 14 -24 h, (v) TBS for 30 min ϫ 2 (wash), and (vi) the ECL development reagent. The developed blots were scanned and quantified using the public domain NIH Image (V. 1.60) software.
Ammonium Sulfate Precipitation and Bio-Gel P-30 Gel Chromatography-Overnight culture supernatant (1 liter) of P. aeruginosa, strain BL2, was mixed overnight at 4°C with ammonium sulfate at 80% saturation. The resulting precipitate was centrifuged at 15,000 ϫ g for 30 min at 4°C, dissolved in 60 ml of de-ionized H 2 O, and dialyzed twice against 4 liters of de-ionized H 2 O. The dialysate was freeze-dried, resuspended in 30 ml of buffer A (50 mM HEPES, pH 7.5, 50 mM NaCl), and 5 ml of this sample were applied to a 1 ϫ 115-cm Bio-Gel P-30 column pre-equilibrated with buffer A. The applied material was fractionated at a flow rate of 4.5 ml/h with buffer A and 24 1-h fractions were collected. Aliquots (300 l) of each fraction were Speed-vac dried, resuspended in 600 l of NMuMG culture media, filter sterilized, and tested for their ability to modulate syndecan-1 shedding from NMuMG cells. For gel analysis, fractions were dialyzed against de-ionized water, Speed-vac dried, resuspended in SDS-PAGE sample buffer, and fractionated by 12% reducing SDS-PAGE.
Immunoaffinity Chromatography-Carbohydrate moieties within the Fc region of anti-LasA IgGs were oxidized and reacted with hydrazide groups in the Affi-Prep coupling resin to form covalent hydrazone bonds. This coupling method was employed to orient the antigen-binding sites outwards from the resin to achieve higher antigen binding capacities. Mouse polyclonal anti-LasA IgGs were purified from sera by protein G affinity chromatography and dialyzed into oxidation buffer (0.1 M NaOAc, pH 5.5, 1 M NaCl). Anti-LasA IgGs (2 mg) in 5 ml of oxidation buffer were oxidized by incubation for 1 h at room temperature in the dark with 500 l of 180 mg/ml NaIO 4 in de-ionized H 2 O. The oxidized antibody was first dialyzed into H 2 O, then into coupling buffer (0.1 M NaOAc, pH 4.5, 1 M NaCl) and incubated overnight with 2 ml of Affi-Prep Hydrazide gel slurry at 4°C. The coupled affinity resin was transferred to a polypropylene column and the active fractions from gel chromatography, resuspended in binding buffer (50 mM HEPES, pH 7.5, 150 mM NaCl), were applied. The samples were re-cycled overnight at a flow rate of 5 ml/h through the affinity resin at 4°C. The flowthrough fraction was collected and the column was washed with binding buffer. The specifically bound materials were eluted with 0.1 M glycine (pH 2.8), neutralized with 1 M HEPES (pH 7.5), dialyzed into autoclaved de-ionized H 2 O, and concentrated by lyophilization. The concentration of purified LasA was determined by UV spectrophotometry based on the number of tyrosine and tryptophan residues in LasA (1 A 280 ϭ 0.41 mg/ml). This preparation of purified LasA, generated through ammonium sulfate precipitation, gel chromatography, and immunoaffinity chromatography, was used in the shedding assays.
Western Immunoblotting of Partially Purified Syndecan-1 Ectodomains-Conditioned media from unstimulated NMuMG cells (contains constitutively shed syndecan-1 ectodomains), and from NMuMG cells stimulated for 14 h with crude P. aeruginosa supernatant (20%, v/v) or purified LasA (5 g/ml) were collected, and NaOAc (pH 4.5) and NaCl were added to final concentrations of 100 and 300 mM, respectively. The acidified conditioned media were incubated with DEAE-Sephacel for 2 h at 4°C. The mixtures were applied to disposable polypropylene columns, washed with 100 mM NaOAc (pH 4.5), 300 mM NaCl buffer and bound materials were eluted with 2 M NaCl. The eluates were dialyzed extensively against de-ionized H 2 O, concentrated by lyophilization, and the amount of partially purified syndecan-1 ectodomain in the samples was estimated by dot immunoblotting. Samples containing 30 ng of syndecan-1 were resuspended in digestion buffer (50 mM Tris, pH 7.5, 50 mM NaOAc, 5 mM EDTA, 2 mM phenylmethylsulfonyl fluoride) and digested with 10 milliunits/ml heparitinase and 20 milliunits/ml chondroitin sulfate ABC lyase for 3 h at 37°C with fresh enzymes added after 1.5 h. These digested samples and undigested samples containing 30 ng of syndecan-1 were fractionated by SDS-PAGE using 3.5-10% gradient acrylamide gels, electrophoretically transferred to Immobilon N (undigested) or nitrocellulose (digested) membranes, probed with monoclonal rat anti-mouse syndecan-1 antibodies (281-2), and then horseradish peroxidase-conjugated goat anti-rat IgGs, and developed by the ECL detection method as described above.

P. aeruginosa and S. aureus Secrete Soluble Enhancers of
Syndecan-1 Shedding-Overnight culture supernatants from several Gram-negative and Gram-positive bacteria were screened for their ability to alter shedding of syndecan-1 ectodomains by NMuMG cells. NMuMG cells were chosen initially since they express syndecan-1 abundantly (43) and because the epithelium is the target cell type of many bacterial pathogens (47,48). Overnight culture supernatants of bacteria were filter sterilized, diluted to 20% (v/v) with NMuMG culture media, and incubated with NMuMG cells for 14 h at 37°C. As shown in Fig. 1, culture supernatants from all tested P. aeruginosa (7/7 clinical, 2/2 laboratory) and the majority of S. aureus (3/5 clinical, 3/4 laboratory) strains enhanced shedding of syndecan-1 by more than 4-fold over control levels, whereas strains from several other Gram-negative (S. enteritidis, S. typhimurium, and K. pneumoniae) and Gram-positive (S. saprophyticus, S. xylosus, and S. pneumoniae) bacteria did not. Cellular extracts of P. aeruginosa and S. aureus strains did not affect shedding (data not shown). These results indicate that P. aeruginosa and S. aureus secrete a soluble enhancer(s) of syndecan-1 ectodomain shedding, and suggest that this property may be specific for certain bacterial species.
Enhancement of Syndecan-1 Shedding by P. aeruginosa Is Rapid and Dose-dependent and Affects Various Host Cells-Because all tested strains of P. aeruginosa augmented syndecan-1 shedding, we focused our subsequent studies on this Gram-negative bacterium. The clinical blood isolate, BL2, showed the greatest activity (ϳ14-fold enhancement, Fig. 1) and was therefore chosen for further studies. As shown in Fig.  2, stimulation of syndecan-1 shedding by BL2 culture supernatant was concentration-dependent ( Fig. 2A), rapid (6-fold increase by 2 h, Fig. 2B), and extensive (11-fold increase by 20 h, Fig. 2B). In contrast to normal turnover (constitutive) shedding, during which constant levels of cell surface syndecan-1 are maintained, P. aeruginosa-enhanced shedding reduced the amount of cell surface syndecan-1 (90% reduction at 20% su- Bacteria were grown overnight in tryptic soy broth at 37°C to stationary growth phase and culture supernatants were collected. Fresh NMuMG culture media (media control) or filter-sterilized bacterial supernatants diluted to 20% (v/v) with NMuMG media were incubated with confluent NMuMG cells in 96-well plates for 14 h at 37°C. At the end of incubation, conditioned media were collected, centrifuged to remove cells, acidified, and dot blotted onto cationic Immobilon N polyvinylidene difluoride membranes. Extent of syndecan-1 shedding was determined by the dot immunoblotting method using the anti-syndecan-1 ectodomain monoclonal antibody (281-2) as described under "Experimental Procedures." Each data point represents the mean of duplicate or triplicate measurements, and results are presented as fold over media control. The number and horizontal bar in P. aeruginosa (PA) and S. aureus (StaphA) samples indicate mean values for these bacteria.
Enhancement of Syndecan-1 Shedding by P. aeruginosa pernatant, Fig. 2A). For at least 20 h of incubation, responding NMuMG epithelial cells remained morphologically normal by light microscopic examination and viable as measured by the tetrazolium salt conversion assay. However, viability and morphology of NMuMG cells were decreased and altered, respectively, when incubated for 35 h or longer with higher concentrations of the supernatant (Ͼ10%), suggesting that subtle morphological changes may not have been detected by light microscopy at earlier time points and at lower concentrations of the supernatant (data not shown).
To examine whether P. aeruginosa can enhance the shedding of syndecan-1 by other cell types, we tested the effects of BL2 culture supernatant (20%, v/v) on LA-4 lung and C127 mammary gland epithelia, and NIH3T3 fibroblasts. BL2 culture supernatants augmented syndecan-1 shedding by more than 5-fold during a 20-h incubation for all cell types tested. The extent of shedding stimulation was highest with NMuMG epithelia (ϳ13-fold), followed by C127 epithelia (ϳ10-fold), LA-4 epithelia (ϳ8-fold), and NIH3T3 fibroblasts (ϳ5-fold). These results demonstrate that although the epithelium, physiological target cell type of P. aeruginosa, respond most extensively to shedding enhancement by P. aeruginosa supernatant, other host cells such as fibroblasts also respond.
Identification of the Syndecan-1 Shedding Enhancer of P. aeruginosa as LasA-We next performed experiments to characterize the P. aeruginosa syndecan-1 shedding enhancer. We first examined whether the activity is susceptible to proteinase K treatment to determine whether the enhancer is a protein. BL2 supernatant was pretreated with 10 g/ml proteinase K for 30 min at 37°C, inactivated with 20 mM phenylmethylsulfonyl fluoride, and then tested for enhancement of syndecan-1 shedding. Proteinase K treatment abolished the activity of P. aeruginosa supernatant. We next fractionated the crude supernatant with molecular weight cutoff spin tubes to obtain a rough estimate of the enhancer's size. Using 3, 10, 30, and 100 kDa molecular mass cutoff tubes, we found that the size of the shedding enhancer is larger than 10 kDa but smaller than 30 kDa. These results suggest that the syndecan-1 shedding enhancer is a 10 -30-kDa protein.
Based on these properties of the shedding enhancer, proteins in the BL2 supernatant were collected by 80% ammonium sulfate precipitation and fractionated by Bio-Gel P-30 (fractionation range ϭ 2.5-40 kDa) gel chromatography in an effort to identify the enhancer. Fractions obtained from gel chromatography were assayed for their ability to enhance shedding of syndecan-1 ectodomains and analyzed by SDS-PAGE. As shown in Fig. 3, the shedding enhancing activity was isolated in one peak and two fractions, 12 and 13. Analysis of the active and inactive fractions by 12% SDS-PAGE and Coomassie staining revealed the presence of a single, major 20-kDa band in the active, but not in the inactive, fractions (Fig. 3, inset). To identify the putative 20-kDa shedding enhancer, N-terminal sequencing was performed. The first 10-amino acid sequence of the 20-kDa protein matched perfectly with mature LasA protein (Table I), a known virulence factor of P. aeruginosa (40 -42).
The hypothesis that LasA is a syndecan-1 shedding enhancer of P. aeruginosa was tested by fractionating the partially purified active peak obtained from Bio-Gel P-30 gel filtration by immunoaffinity chromatography using mouse polyclonal anti-LasA IgGs covalently coupled to a cross-linked agarose resin. The rationale behind this experiment was that if LasA is the shedding enhancer, then the active component in the partially purified material will be bound to the affinity column, and shedding activity will be seen only with the specifically bound fractions and not with the flow-through or wash fractions. As An overnight culture supernatant of strain BL2 was precipitated by 80% ammonium sulfate and fractionated by Bio-Gel P-30 gel chromatography at a flow rate of 4.5 ml/h. The collected fractions were Speed-vac dried, resuspended in NMuMG culture media, filter sterilized, and assayed for syndecan-1 shedding activity. Each data point in the activity chromatogram represents mean values from duplicate determinations. The active fractions (12 and 13) and inactive fractions in the vicinity (10, 11, 14, and 15) were subjected to 12% reducing SDS-PAGE and visualized by Coomassie Brilliant Blue R-250 staining (inset).

Enhancement of Syndecan-1 Shedding by P. aeruginosa
shown in Fig. 4, the specifically bound eluate (EL), but not the flow-through (FT) or wash (WSH) fractions, enhanced syndecan-1 shedding by NMuMG cells. The inactive flow-through fraction contained the contaminating smear seen in the active fractions partially purified by gel chromatography, and the eluate fraction contained the highly purified 20-kDa LasA protein (Fig. 4, inset). The purified LasA protein enhanced syndecan-1 shedding by various host cells (Fig. 5) and did not affect steady-state mRNA levels of syndecan-1 (data not shown). When culture supernatants from the P. aeruginosa strain lacking LasA (PAO-B1A1) (40) were subjected to the identical purification procedure, the resulting eluate fraction from anti-LasA immunoaffinity chromatography did not contain protein bands and did not enhance the shedding of syndecan-1 ectodomains from NMuMG cells (data not shown). Taken together, these results indicate that LasA is a syndecan-1 shedding enhancer of P. aeruginosa.
LasA Enhances Syndecan-1 Shedding by Stimulating the Shedding Mechanism of the Host Cell-To begin to elucidate the mechanism by which P. aeruginosa LasA enhances shedding of syndecan-1 ectodomains, we first examined the molecular size of shed syndecan-1 ectodomains and their core proteins. Conditioned media from NMuMG cells cultured to confluency (constitutively shed) and from NMuMG cells stimulated with purified LasA or crude P. aeruginosa supernatant were subjected to DEAE ion exchange chromatography to obtain partially purified samples of syndecan-1 ectodomain. These undigested samples were directly analyzed by Western immunoblotting (Fig. 6, lanes 1-3) or digested by heparitinase and chondroitin sulfate ABC lyase, and then analyzed by Western immunoblotting (Fig. 6, lanes 4 -6) to determine the size of shed syndecan-1 ectodomain core proteins. Similar to the constitutively shed syndecan-1 ectodomain (lane 1), syndecan-1 ectodomains obtained from both purified LasA (lane 2) and crude supernatant (lane 3) conditioned media were intact proteoglycans decorated with glycosaminoglycans as evident from the smear of immunologically detected syndecan-1 ectodomains. Interestingly, the size of syndecan-1 ectodomain core proteins stimulated to shed by both purified LasA (lane 5) and crude P. aeruginosa supernatant (lane 6) was identical to that of the constitutively shed core protein (lane 4) by SDS-PAGE analysis.
Because the similarity in molecular size of the shed syndecan-1 ectodomain suggested that LasA enhances syndecan-1 shedding by a mechanism similar to that of the host cell's shedding mechanism, the effects of a hydroxamate derivative (BB1101), protein kinase C antagonist (bisindolylmaleimide I), and protein tyrosine kinase (PTK) inhibitors (genistein, Tyrphostin A25) were tested. Genistein (49) and Tyrphostin A25 (50) inhibit PTKs by competing for binding with ATP and tyrosine residues to PTKs, respectively. These general PTK inhibitors inhibit syndecan-1 and -4 shedding stimulated by all known agonists such as epidermal growth factor, thrombin, sphingomyelinase, ceramide, and stress conditions (e.g. heat, hyperosmolarity), whereas the antagonistic effect of the protein kinase C inhibitor, bisindolylmaleimide I, is restricted to syndecan-1 and -4 shedding induced by hyperosmolarity, ceramide, and phorbol esters (15). 2 Hydroxamate derivatives inhibit the activity of the putative cleaving enzyme by chelating its active site zinc atom (51). Thus, general PTK inhibitors are inhibitors of regulated syndecan shedding whereas hydroxamate derivatives are inhibitors of both regulated and constitutive shedding. As shown in Table II, when co-incubated, BB1101 and Tyrphostin A25 inhibited both purified LasA-and P. aeruginosa supernatant-enhanced syndecan-1 shedding by more than 70 and 60%, respectively, at the highest concentra- shedding enhancer isolated from P. aeruginosa The 20-kDa shedding enhancer, partially purified by ammonium sulfate precipitation and gel chromatography, was subjected to 12% SDS-PAGE and electrophoretically transferred to Problott polyvinylidene difluoride membrane for 1 h at 200 mA using CAPS transfer buffer (10 mM CAPS, pH 11, 10% MeOH in de-ionized H 2 O). The 20-kDa band was visualized by Coomassie Brilliant Blue R-250 staining, destained, washed extensively with de-ionized water, excised from the membrane, and sequenced directly using an Applied Biosystems 477A protein sequencer at the Department of Physiology Core Facility at Tufts University Medical School. . Immunoaffinity chromatography with anti-LasA IgGs demonstrates that the P. aeruginosa shedding enhancer is LasA. The partially purified material obtained from ammonium sulfate precipitation and gel chromatography of P. aeruginosa supernatant was fractionated by affinity chromatography using mouse polyclonal anti-LasA IgGs coupled to a cross-linked agarose resin. The flow-through (FT), wash (WSH), and eluate (EL) fractions were collected, and along with the starting material (fractions 12 and 13), tested in triplicates for their ability to enhance shedding of syndecan-1 by NMuMG cells as described previously. Results of the activity assay are presented as mean fold increase over media control Ϯ S.D. Results from analysis of the fractions by 12% SDS-PAGE and Coomassie staining are shown in the inset.
Enhancement of Syndecan-1 Shedding by P. aeruginosa tion tested (both p Ͻ 0.05). Genistein also significantly inhibited enhanced shedding by approximately 45% (p Ͻ 0.05), but bisindolylmaleimide I did not significantly inhibit syndecan-1 shedding enhanced by LasA and crude P. aeruginosa supernatant (both p Ͼ 0.05). In contrast, when the kinase inhibitors and BB1101 were preincubated with purified LasA and removed from the test samples prior to incubation with NMuMG cells, none of them significantly inhibited enhanced syndecan-1 shedding (p Ͼ 0.05). Taken together, these results indicate that the PTK inhibitors and BB1101 are acting on the host cell when inhibiting LasA-enhanced shedding, and that LasA enhances syndecan-1 shedding via activation of the host cell's shedding mechanism. DISCUSSION We report here that the major opportunistic bacterial pathogens, P. aeruginosa and S. aureus, secrete potent enhancers of syndecan-1 shedding. Although we have not yet identified the S. aureus shedding enhancer, we have found that the syndecan-1 shedding enhancer of P. aeruginosa is the 20-kDa LasA protein, a virulence factor in animal models of corneal (40) and lung (41,42) infections. LasA is secreted as a precursor protein of approximately 40 kDa, which is then processed to the mature 20-kDa form by unknown mechanisms (52,53). Mature LasA is a zinc metalloendopeptidase with strong staphylolytic and weak elastolytic activities (54,55). The alternative name of LasA, staphylolysin, is derived from its ability to lyse staphylococcal cells, and because of its elastolytic activity, LasA was first thought of as P. aeruginosa elastase. It is now known that the role of LasA in elastolysis is to render the insoluble elastin substrate more susceptible to cleavage by the true P. aeruginosa elastase and other elastolytic enzymes (52,56).
The ability to enhance shedding of syndecan-1 appears to be specific for P. aeruginosa and S. aureus since several other Gram-negative and Gram-positive bacteria failed to do so. Obviously, a larger sampling of bacterial pathogens needs to be performed to verify this hypothesis since the number of strains examined for the inactive bacteria was minimal in this study.
Nevertheless, the finding that evolutionarily diverse bacteria, such as P. aeruginosa and S. aureus, can enhance syndecan-1 shedding suggests that this activity may augment their pathogenesis at target host sites common to both. In this regard, it is interesting to note that P. aeruginosa and S. aureus are the dominant pathogens in cystic fibrosis and burn patients, and that syndecan-1 is the major syndecan of target cell types at these tissue sites, the lung epithelia and epidermal keratinocytes, respectively.
Host Effector Shedding by Pathogenic Bacteria-The current emergence of antibiotic-resistant strains has been driven mainly by overuse of antibacterial agents aimed at inhibiting essential aspects of bacterial metabolism, such as cell wall and protein synthesis, thereby placing selective pressure on bacteria to become rapidly resistant to these agents for their survival. Thus, to prevent development of resistance, it may be ideal to develop prophylactic and therapeutic agents that target specific host-pathogen interactions involved in bacterial pathogenesis. Enhanced host effector shedding may be one such target of the pathogenesis cascade. Many bacterial pathogens as diverse as P. aeruginosa, S. aureus, S. epidermidis, E. coli, S. marcescens, and L. monocytogenes have the ability to enhance shedding of host surface effectors, such as CD14, TNF-␣, and IL-6 receptor (29 -31). These bacteria are not only distinguished by their cell wall characteristics and sites of colonization, but also by their arsenal of genetically distinct virulence factors. Thus, the shared ability to enhance shedding of host molecules indicates functional convergence and suggests that bacterial stimulation of shedding may play a role in pathogenesis.
Mechanism of Syndecan-1 Shedding Enhancement by LasA-The capacity of LasA to hydrolyze protein substrates such as elastin, albeit weak, suggests that LasA may enhance syndecan-1 shedding by direct cleavage of the proteoglycan. However, several lines of evidence indicate that this is not the mechanism of syndecan-1 shedding enhanced by LasA. First, the proteolytic specificity of LasA is rather restricted in that potential substrates are those with Gly in the P1 and P2 positions, Gly, Ala, or Phe at P1Ј and apolar residues in the flanking sequences (54). This stringent requirement is exemplified by the fact that elastin and the cell wall peptidoglycan of S. aureus with these motifs are susceptible, but casein without these motifs is not hydrolyzed by LasA (54). Syndecan-1 also does not contain these LasA-susceptible motifs. Second, our results show that the size of the core protein shed by LasA and endogenous host cell mechanisms is identical by SDS-PAGE analysis, and that PTK and hydroxamate inhibitors of LasAmediated syndecan-1 shedding inhibit shedding by acting on the responding host cells and not on LasA. The PTK inhibitors and the hydroxamate derivative (BB1101) inhibit shedding only when LasA, the reagents and host cells are co-incubated, and not when the inhibitors are preincubated with LasA and removed prior to incubation of the pretreated LasA with host cells. Furthermore, hydroxamate inhibitors are thought to be specific for the HEXXH zinc-binding catalytic domain of metalloendopeptidases, such as the matrix metalloproteinases (51), but the zinc-binding motif of LasA is HXH (55). Taken together, these findings indicate that LasA augments shedding of syndecan-1 ectodomains by activating the host cell's shedding mechanism.
How PTK activities are involved in syndecan-1 shedding enhanced by LasA is not understood. In fact, the role which PTK activities play in stimulation of syndecan shedding by all other known physiological agonists (e.g. epidermal growth factor, thrombin, stress-related agents) is not known. PTKs may activate shedding by affecting the syndecan substrate, the FIG. 6. Syndecan-1 ectodomains stimulated to shed by purified LasA and crude P. aeruginosa supernatant are intact proteoglycans and the size of their core proteins is identical to that of the constitutively shed ectodomain. Conditioned media from unstimulated NMuMG cells (lanes 1 and 4) and from NMuMG cells stimulated with 5 g/ml purified LasA (lanes 2 and 5) or 20% (v/v) crude P. aeruginosa supernatant (lanes 3 and 6) were incubated with DEAE-Sephacel for 2 h at 4°C, and bound materials were eluted with 2 M NaCl. Undigested samples were analyzed by 3.5-10% gradient SDS-PAGE and Western immunoblotting using the 281-2 anti-syndecan-1 ectodomain monoclonal antibody (undigested, lanes 1-3) or samples were digested with 10 milliunits/ml heparitinase and 20 milliunits/ml chondroitin sulfate ABC lyase and then analyzed by SDS-PAGE and Western immunoblotting (digested, lanes 4 -6). Molecular masses of the immunoreactive proteins were approximated from the migration pattern of pre-stained size standards.
Enhancement of Syndecan-1 Shedding by P. aeruginosa cleaving enzyme and/or other unidentified components of the shedding mechanism. Among these hypotheses, regulation of shedding by phosphorylation of the syndecan cytoplasmic domain is an attractive possibility since it has been shown that pervanadate inhibition of tyrosine phosphatases can augment shedding (57), highly conserved syndecan cytoplasmic domains can form a complex with Src family PTKs (58), unspecified tyrosine residues within cytoplasmic domains of syndecan-1 and -4 can be phosphorylated by Src-like PTKs (59), and importantly, this tyrosine phosphorylation is maintained in an off state in NMuMG cells (59). Based on these data, it is tempting to speculate that LasA enhances shedding by binding to a putative host determinant and activating Src family PTKs, which can then phosphorylate tyrosine residues in the syndecan-1 cytoplasmic domain to render the cell surface syndecan-1 molecules susceptible to cleavage for shedding. We are currently testing these hypotheses by studying whether (i) LasA binds specifically to host cells, (ii) specific Src family PTKs are activated, and (iii) specific tyrosine residues in the cytoplasmic domain of syndecan-1 are involved in transducing the signal from LasA.
Implications of LasA-enhanced Syndecan-1 Shedding in P. aeruginosa Pathogenesis-At present, we do not know whether enhanced syndecan-1 shedding by LasA is beneficial for the bacteria or the host. Based on the finding that LasA is a virulence factor in animal models of corneal (40) and lung infections (41,42), two tissue sites where syndecan-1 is the predominant syndecan on target epithelia, we hypothesize that LasA-enhanced shedding of syndecan-1 ectodomains promotes bacterial pathogenesis. There are several ways by which syndecan-1 shedding can contribute to P. aeruginosa pathogenesis. First, our results show that enhancement of syndecan-1 shedding by P. aeruginosa not only dramatically increases the amount of soluble ectodomains, but is also accompanied by a significant decrease in the level of cell surface syndecan-1. This property may be pathologically significant since in a previous study, we have found that antisense induced depletion of cell surface syndecan-1 altered cell morphology and organization, expression of other adhesion molecules like E-cadherin and ␤ 1 -integrins, and anchorage-dependent growth characteristics in NMuMG cells (18). Because highly polarized epithelia are thought of as an effective barrier against microbial colonization (47,48), the concomitant decrease in cell surface syndecan-1 levels observed during LasA-enhanced shedding could enhance P. aeruginosa colonization by altering the morphology of target epithelia, disrupting the epithelial barrier, and exposing intercellular, basolateral, and subepithelial adhesive components.
Syndecan-1 shedding enhanced by LasA may also contribute to P. aeruginosa pathogenesis by interfering with host defense mechanisms. Syndecan shedding is a host response activated during tissue injury, and is thought of as a defense mechanism against insults including wounding and cellular stress (e.g. hyperosmolarity, mechanical shear, heat). In this mechanism, excess shed syndecan-1 ectodomains, via their heparan sulfate chains, bind and neutralize the activity of potentially deleterious proinflammatory mediators such as proteases, chemokines, and cytokines (35). Thus, enhancement of syndecan-1 shedding by some bacterial pathogens, such as P. aeruginosa and S. aureus, may be a pathogenetic mechanism that takes advantage of a normal defense mechanism to promote their survival in the host environment. In support of this hypothesis, we have found in a separate study that shed syndecan-1 ectodomains can inhibit the antibacterial activity of Pro/Arg-rich antimicrobial peptides by binding to the peptides and preventing them from interacting with target bacterial cells. 3 These findings suggest that shed syndecan-1 ectodomains are a host-derived component of the virulence mechanism mediated by LasA. To decipher the role of enhanced shedding in bacterial pathogenesis further, we are currently evaluating the role of syndecan-1 shedding in murine models of bacterial infection using specific agonists and antagonists of shedding, and also using mice lacking syndecan-1 or overexpressing a constitutively shed form of syndecan-1.