Distinct Roles of Phenol-Soluble Modulins in Spreading of Staphylococcus aureus on Wet Surfaces

Abstract

The human pathogen Staphylococcus aureus is renowned for the rapid colonization of contaminated wounds, medical implants, and food products. Nevertheless, little is known about the mechanisms that allow S. aureus to colonize the respective wet surfaces. The present studies were therefore aimed at identifying factors used by S. aureus cells to spread over wet surfaces, starting either from planktonic or biofilm-associated states. Through proteomics analyses we pinpoint phenol-soluble modulins (PSMs) as prime facilitators of the spreading process. To dissect the roles of the eight PSMs produced by S. aureus, these peptides were chemically synthesized and tested in spreading assays with different psm mutant strains. The results show that PSMα3 and PSMγ are the strongest facilitators of spreading both for planktonic cells and cells in catheter-associated biofilms. Compared to the six other PSMs of S. aureus, PSMα3 and PSMγ combine strong surfactant activities with a relatively low overall hydropathicity. Importantly, we show that PSM-mediated motility of S. aureus facilitates the rapid colonization of wet surfaces next to catheters and the colonization of fresh meat.

INTRODUCTION

Staphylococcus aureus is an opportunistic human pathogen that can cause a wide range of acute and chronic diseases, which range from superficial skin infections to life-threatening endocarditis and sepsis (1, 2). The ability of this Gram-positive bacterium to cause these infections depends on the production of secreted and cell wall-associated virulence factors. Of increasing concern is the ability of S. aureus to acquire resistance against antibiotics, as underscored by the global spread of methicillin-resistant S. aureus (MRSA) lineages.

Intriguingly, recent proteomics studies have revealed an enormous diversity in the production of virulence factors by different isolates of S. aureus, and only a few of these seem to be invariantly produced (3–5). Among the most commonly identified staphylococcal virulence factors, especially in the community-associated (CA)-MRSA lineages, are the so-called phenol-soluble modulins (PSMs) (6). These PSMs are short, amphipathic, α-helical peptides that have leukocidal activity and biosurfactant properties (7–9). The growth media of S. aureus cultures contain both N-terminally formylated and deformylated PSMs, suggesting that these virulence factors are substrates for the bacterial N-formylmethionine deformylase (9, 10).

To date, eight PSMs have been identified in S. aureus. These include the four PSMα1 to PSMα4 peptides (22 residues each), the PSMβ1 and PSMβ2 peptides (44 residues each), PSMγ (25 residues) and the recently reported PSM-mec (22 residues). The PSMα peptides are encoded by the psmα operon, the PSMβ peptides by the psmβ operon, and PSMγ by the hld gene. Notably, the hld gene is embedded within the regulatory RNAIII molecule that is encoded by the agr locus. The gene for PSM-mec was identified in MRSA strains carrying the staphylococcal cassette chromosome mec (SCCmec) types II or III. The expression of all psm genes is controlled by the Agr system for quorum sensing (8, 11–13). This system modulates gene expression such that cell wall-associated virulence factors (e.g., the immunoglobulin G-binding protein A) are most highly expressed at low cell densities and that secreted virulence factors (e.g., the PSMs) are most highly expressed at high cell densities (14–21).

Notably, PSMs have been implicated in the high virulence of CA-MRSA lineages, which are readily transmitted by direct contact with a carrier (9, 22). The investigated CA-MRSA isolates produce higher amounts of the PSM peptides than the generally less virulent nosocomial MRSA isolates (9, 22). The PSMα peptides have the strongest leukolytic, proinflammatory, and chemotactic activities (9). Consistently, a strain lacking psmα had a decreased ability to cause skin lesions in mice and rabbits (9, 23). In addition to this, Wang et al. (24) have shown that the mortality rates and the levels of the inflammatory cytokine tumor necrosis factor alpha in the blood of mice infected with psmα or psmγ mutant strains were substantially reduced. The PSMβ peptides appear less important for cytolysis and inflammation but, in low concentrations, they seem to promote biofilm formation by Staphylococcus epidermidis. High amounts of the same PSMβ peptides do, however, promote the detachment of staphylococcal cells from biofilms both in vitro and in vivo (24).

Although S. aureus was originally believed to be nonmotile, recent studies have shown that this organism is capable of spreading over wet surfaces (25–27). We have previously shown that a mix of the four PSMα peptides can promote this so-called colony spreading phenotype (27), and we hypothesized that this relates to their strong surfactant properties (9). The Agr system is an important determinant for colony spreading due to its control over the synthesis of PSMs (27). However, Agr is also needed for biofilm formation, which gives it a decisive role in the choice between motile and sessile lifestyles of S. aureus. Furthermore, it was shown that covalently cell wall-anchored proteins, like FnbpA, FnbpB, ClfA, and ClfB, can set a limit to the colony-spreading ability of S. aureus cells (28). To date, very little was known about the roles of individual PSM peptides in colony spreading and whether these are the main factors promoting spreading. Therefore, the present studies were aimed at dissecting the roles of the different PSM peptides in colony spreading. Furthermore, we wanted to test whether N-terminally formylated and deformylated PSMs are equally potent in colony spreading. To achieve these objectives, we constructed multiple psm mutant strains, which were then incubated in the presence or absence of synthetic PSMs. Importantly, our results show that PSMα3 and PSMγ have key roles in colony spreading. Furthermore, our observations link PSMα3 and PSMγ to the spreading of staphylococcal cells from catheter-related biofilms, and they suggest that PSM-mediated spreading plays a major role in the movement of S. aureus over biotic surfaces. Importantly, our findings show that the PSMs that most effectively promote the spreading of S. aureus combine high surfactant activities with low overall hydropathicity.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains and plasmids that were used in this study are listed in Table 1. Escherichia coli strains were grown in Lysogeny broth (LB) at 37°C under vigorous shaking. S. aureus strains were grown in tryptic soy broth (TSB) at 37°C under vigorous shaking, or on tryptic soy agar (TSA) plates. Where necessary, antibiotics were added at the following concentrations: ampicillin, 100 μg/ml (for E. coli); erythromycin, 5 μg/ml (for S. aureus); and chloramphenicol, 10 μg/ml (for S. aureus).

Table 1.

Bacterial strains and plasmids

Strain or plasmid	Descriptiona	Source or reference
Strains
E. coli
DH5α	λ− ϕ80dlacΖΔΜ15 Δ(lacZYA-argF)U169 recA1 endA hsdR17(rK− mK−) supE44 thi-1 gyrA relA1	29
TOP10	Cloning host for TOPO vector; F− mcrA Δ(mrr-hsdRMS-mcrBC) λ− ϕ80dlacΖΔΜ15 ΔlacX74 recA1 araD139 Δ(ara-leu)7697 galU galK rpsL (Strr) endA1 nupG	Invitrogen Life Technologies
S. aureus
RN4220	Restriction-deficient derivative of NCTC8325, cured of all known prophages	30
NCTC8325	HA-MSSA strain, agr+ ΔrsbU	31
NCTC8325−	HA-MSSA strain, Δagr ΔrsbU	27
NCTC8325 Δpsmα	NCTC8325 lacking the psmα genes	This study
NCTC8325 Δpsmβ	NCTC8325 lacking the psmβ genes	This study
NCTC8325 Δpsmα Δpsmβ	NCTC8325 lacking the psmα and psmβ genes	This study
SH1000	NCTC8325-4 derivative, rsbU+ agr+	32
SH1000−	NCTC8325-4 derivative, rsbU+ Δagr	27
SH1000 Δpsmα	SH1000 lacking the psmα genes	This study
SH1000 Δpsmβ	SH1000 lacking the psmβ genes	This study
SH1000 Δpsmα Δpsmβ	SH1000 lacking the psmα and psmβ genes	This study
HG001	NCTC8325 derivative, rsbU+ agr+	33
HG001−	NCTC8325 derivative, rsbU+ Δagr	27
HG001 Δpsmα	HG001 lacking the psmα genes	This study
HG001 Δpsmβ	HG001 lacking the psmβ genes	This study
HG001 Δpsmα Δpsmβ	HG001 lacking the psmα and psmβ genes	This study
Newman	ATCC 25904; high-level clumping factor production; σB+	34
Newman Δagr	Newman derivative, Δagr::tetM	35
Newman Δpsmα	Newman lacking the psmα genes	This study
Newman Δpsmβ	Newman lacking the psmβ genes	This study
Newman Δpsmα Δpsmβ	Newman lacking the psmα and psmβ genes	This study
LAC USA300	CA-MRSA	9
LAC USA300 Δpsmα	LAC USA300 lacking the psmα genes	9
Plasmids
pUC18	Apr; ColE1, ϕ80dlacΖ; lac promoter	36
TOPO	pCR-Blunt II-TOPO vector; Kmr	Invitrogen Life Technologies
pMAD	E. coli/S. aureus shuttle vector that is temperature sensitive in S. aureus and contains the bgaB gene; Emr Apr	37
pRIT5H	E. coli/S. aureus shuttle vector; Cmr; spa promoter	38
pSW4-GFPopt	pSW4 plasmid containing 729-bp AseI/BamHI-cloned gfpopt	39
GFPopt-pRIT5H	pRIT5H with gfpopt gene; Apr Kmr	This study
pAH9	sarA promoter Pi-RFP	40

^aCA, community acquired; HA, hospital acquired; MRSA, methicillin-resistant S. aureus. Cm^r, chloramphenicol resistance; Ap^r, ampicillin resistance; Em^r, erythromycin resistance; Km^r, kanamycin resistance.

Colony spreading assay.

The colony-spreading assay was performed essentially as described by Kaito et al. (25), but with minor previously described modifications (27). To detect colony spreading of the S. aureus strains SH1000 and Newman along the growth curve, these strains were grown in TSB for 24 h. Samples were collected at hourly intervals for the first 7 h and after 24 h of growth. All samples were immediately tested for colony spreading. Equal amounts of cells from each time point were spotted onto the 0.24% TSA plates. All spreading assays were repeated at least five times.

Mass spectrometric analyses of culture supernatants.

Strains Newman (agr⁺, Δagr, or Δpsmα), LAC USA300 (agr⁺ or Δpsmα), NCTC8325 (agr⁺ or Δagr) and HG001 (agr⁺ or Δagr) were grown in TSB. At an optical density at 600 nm (OD₆₀₀) of 2 and 6, 3-ml culture samples were collected, and cells were separated from the growth medium by centrifugation (8,000 × g, 4°C, 10 min). To precipitate the secreted proteins, the medium fractions were incubated at 4°C with 10% trichloroacetic acid overnight. The proteins were pelleted (20,000 × g, 4°C, 20 min), washed with acetone, and subsequently dissolved in 8 M urea. Protein concentrations were determined using the Bio-Rad DC protein assay according to the protocol of the supplier. The samples (2 μg) were reduced with 10 mM dithiothreitol (DTT; Duchefa Biochemie) for 30 min and alkylated with 10 mM iodoacetamide (Sigma-Aldrich) for another 30 min in the dark. Finally, the protein samples were incubated overnight at 37°C with 40 ng of trypsin (Promega). Peptides were purified using ZipTips (4), separated and analyzed by liquid chromatography-tandem mass spectrometry using an Easy-nLCII high-pressure liquid chromatography system (Thermo Fisher Scientific, Waltham, MA) coupled directly to an LTQ Orbi-Trap mass spectrometer (Thermo Fisher Scientific, Waltham, MA). The Easy-nLCII was equipped with a self-packed analytical column (C18-material [Luna 3u C18(2)100A; Phenomenex], 100-μm inner diameter by 200-mm column). Peptide elution was performed by application of a binary gradient of buffer A (0.1% [vol/vol] acetic acid) and B (99.9% [vol/vol] acetonitrile, 0.1% [vol/vol] acetic acid) over a period of 80 min with a flow rate of 300 nl/min. The mass spectrometry (MS) analyses were performed as described by Miller et al. (41). A nonredundant database was constructed that contained all available S. aureus protein sequences (Uni-Prot) that differ in at least one amino acid residue (see Table S2 and FASTA Files S1 and S2 in the supplemental material). This database was subsequently used for the database search (including a concatenated reversed database, 40,098 entries). The database search was performed with Sequest using a fragment ion mass tolerance of 1.00 Da and a parent ion tolerance of 10 ppm. Oxidation of methionine and carbamidomethylation of cysteine were specified in Sequest as variable modifications. Validation of MS/MS-based peptide and protein identifications was performed with Scaffold (version Scaffold 3.3.2; Proteome Software, Inc., Portland, OR). Peptide identifications were accepted if they exceeded specific database search engine thresholds. Sequest identifications required at least deltaCn scores of >0.10 and XCorr scores of greater than 1.9, 2.2, 3.8, and 3.8 for singly, doubly, triply, and quadruply charged peptides. All experiments have been conducted in independent duplicates. Proteins were only accepted as being identified if they were detected in both biological replicates per sample set. With these filter parameters, the false-positive rate was below 1%.

Construction of a GFP expressing vector.

For constitutive expression of GFP in S. aureus the gfpopt gene was amplified from plasmid pSW4-GFPopt using primers GFPoptFR and GFPoptRV (see Table S3 in the supplemental material). The amplified PCR product was then cloned in plasmid pRIT5H using the EcoRI and SalI restriction sites. Expression of GFP was detected using the IVIS Spectrum from Caliper Life Sciences using the specific filter for GFP (excitation, 465 nm; emission, 520 nm).

Construction of PSM mutant strains of S. aureus.

Mutants of S. aureus were constructed using the temperature-sensitive plasmid pMAD (37) and previously described procedures (42). Primers were designed using the genome sequence of S. aureus NCTC8325 (http://www.ncbi.nlm.nih.gov/nuccore/NC007795). To delete the psmα and/or psmβ operons, the primer pairs with the designations F1/R1 and F2/R2 were used for PCR amplification of the respective upstream and downstream regions (each ∼500 bp) (see Table S3 in the supplemental material). The R1 and F2 primers contain a 24-bp linker sequence to fuse the flanking regions by PCR prior to cloning in pMAD. The resulting plasmids were used to transform S. aureus strain RN4220 via electroporation. Next, the plasmids were isolated from the RN4220 strain and used to transform the S. aureus SH1000, HG001, NCTC8325, or Newman strains via electroporation in order to delete their psmα and/or psmβ operons through subsequent plasmid integration and excision steps (43). At the end of the procedure, white colonies were screened for the absence of the psmα and/or psmβ genes by colony PCR using primers F1 and R2.

Complementation of psm mutations by synthetic PSM peptides and determination of surfactant properties.

The PSMα1 to PSMα4, PSMβ1 and PSMβ2, PSMγ, and PSM-mec peptides were synthesized as described previously with a C-terminal four-residue glycyl spacer and an ε-amino biotinyl lysine (27). All peptides were dissolved in dimethyl sulfoxide plus 10 mM DTT to a concentration of 12 mM. The peptides were then diluted 10-fold in phosphate-buffered saline, and 2 μl of each peptide solution was spotted in the center of a soft TSA plate prior to the inoculation of the S. aureus mutant strain to be tested for spreading. The surfactant properties of the PSMs were determined by spotting 2 μl of each peptide solution in the center of a soft TSA plate and by subsequently measuring the diameter of the resulting transparent halo.

Colony spreading of S. aureus from catheter-associated biofilms.

Catheters were positioned on soft agar plates and 2 μl of bacteria grown overnight were spotted next onto these catheters before the plates were incubated overnight at 37°C. Subsequently, the catheters were transferred to fresh soft TSA plates and incubated again overnight at 37°C, which resulted in the formation of catheter-associated biofilms. Catheters with associated biofilms were transferred to fresh soft TSA plates, where 2 μl of PSM peptide solution was spotted prior to the transfer of the catheters. The plates were then incubated overnight at 37°C. Images were recorded with a G:box (Syngene, Leusden, Netherlands).

Colony spreading of S. aureus on pork.

Pieces of pork meat were placed in sterile petri dishes and, approximately in the center, 2 μl of bacteria grown overnight in TSB were spotted. The meat was incubated at 37°C for 48 h, after which images were recorded with a Sony cyber-shot camera.

RESULTS

Secreted factors regulated by agr are responsible for colony spreading.

We have previously shown that the agr locus, which regulates the synthesis of secreted virulence factors, is required for colony spreading of S. aureus. This suggested that secreted virulence factors are of prime importance for colony spreading. Another indication for the role of secreted factors in spreading came from the observation that colonies of S. aureus cells that are capable of spreading were surrounded by a transparent halo (Fig. 1A). This halo was absent from nonspreading agr mutant colonies. Furthermore, when overnight grown agr-proficient bacteria were spotted on soft agar plates, spreading of the cultured cells was visible after a few seconds, suggesting that any secreted factors needed for spreading were already present in the culture. In fact, this area of rapid spreading remained clearly visible on the plates (marked by a box in Fig. 1B), which was due to the emergence of subsequent “waves” of spreading cells from the initial spreading zone. This suggested that the factors needed for spreading were not synthesized continuously. To investigate whether the spreading factors were synthesized growth phase dependently, S. aureus cells were grown to different growth stages and tested for the rapid spreading phenotype. Indeed, the rapid spreading was not observed for cells in the early exponential growth phase (Fig. 2A, time point t3), but it started when the cells reached the late exponential phase (t5) and continued in the stationary phase (t7). Consistent with our previous findings, these time points correspond with the activation of the Agr system (17). To verify that secreted factors promote colony spreading, Δagr cells from different strains were resuspended in filtered supernatants of overnight grown agr⁺ strains and the cell suspension was spotted on top of soft agar plates. As shown in Fig. 2B, the filtered supernatants were able to promote colony spreading of the Δagr cells. Notably, when the filtered supernatant of an agr⁺ strain was included within the soft agar, it did not promote the colony spreading of Δagr cells (Fig. 2C, plate 4). Together, these results demonstrate that secreted factors regulated via the Agr system are both needed and sufficient for colony spreading of S. aureus. Importantly, to promote colony spreading, these secreted factors need to be present on the surface of the soft agar on which the bacteria spread.

Fig 1.

Characteristic features of S. aureus spreading motility. (A) A transparent “halo” is produced by cells of spreading colonies. A similar halo is also generated by surfactants that are spotted on soft agar plates (not shown). (B) Spreading motility involves a rapid phase of spreading by the cells spotted on a plate. Subsequently, waves of cells emerge from the cells in the initial spreading zone. The boxes mark the rapid spreading zones that emerged from the sites of inoculation of strains SH1000 and Newman on soft agar plates.

Fig 2.

Colony spreading depends on the growth phase of inoculated S. aureus cells and on secreted factors produced by agr⁺ strains. (A) Colony spreading by planktonic cells of S. aureus strains SH1000 or Newman collected from cultures in different growth stages; t3 corresponds to the early exponential growth phase (OD₆₀₀ = 1.5), t5 to the late exponential phase (OD₆₀₀ = 6.8), and t7 to early stationary phase (OD₆₀₀ = 9.0). (B) Filter-sterilized culture medium of agr⁺ cells of S. aureus strain Newman promotes rapid spreading of Δagr cells of strains SH1000, NCTC8325, and Newman. As shown for Δagr cells of strain SH1000, no spreading is observed when fresh medium is used (left plate labeled “agr−”). (C) Culture supernatants of agr⁺ cells need to be applied on top of soft agar plates to promote spreading. Fresh soft agar plates were prepared and inoculated as follows: plate 1, regular soft agar with the wild-type strain Newman; plate 2, regular soft agar with the wild-type strain Newman resuspended in 2 μl of filter-sterilized supernatant of the wild-type strain Newman; plate 3, 200 μl of filter-sterilized supernatant of the wild-type strain Newman included within the soft agar prior to inoculation with the wild-type strain Newman; and plate 4, 200 μl of filter-sterilized supernatant of the wild-type strain Newman included in the soft agar prior to inoculation with Δagr cells of strain Newman. (D) Δagr cells carrying plasmid GFPopt-pRIT5H for expression of GFP were coinoculated with agr⁺ cells carrying plasmid pAH9 for expression of mCherry. GFP fluorescence of the Δagr cells is detectable at the edges of the spreading zone (green color), whereas mCherry fluorescence of the agr⁺ cells (red color) is detectable in the entire area covered by spreading. All spreading assays were repeated at least five times.

As the spent growth medium of agr⁺ cells can promote the spreading of Δagr cells, we wondered whether the Δagr cells would also spread if they were grown in the presence of agr⁺ cells. To test this, an overnight culture of a nonspreading Δagr derivative of S. aureus HG001 expressing the green fluorescent protein (GFP) was mixed with an overnight culture of the authentic S. aureus HG001 strain (agr⁺) expressing the fluorescent mCherry protein. Next aliquots of the mixed culture were transferred to soft agar plates and incubated overnight. Images were taken using the IVIS spectrum with the specific filters for GFP (excitation, 465 nm; emission, 520 nm) and mCherry (excitation, 570 nm; emission, 620 nm) to monitor the spreading of the two different strains. GFP was only detected at the edges of the colony spreading area, indicating that the Δagr strain was very well able to spread over the soft agar together with the agr⁺ strain (Fig. 2D). Since the Δagr HG001 strain cannot spread by itself, this observation implies that nonspreading S. aureus cells can passively spread on the soft agar plates with the help of factors that are actively secreted by spreading S. aureus cells.

MS identification of secreted proteins potentially involved in colony spreading.

To investigate which proteinaceous factors are involved in colony spreading, culture supernatants of agr⁺, Δagr, and Δpsmα variants of the S. aureus strains Newman, LAC USA300, NCTC8325, and HG001 were analyzed using MS. As expected, the MS analyses revealed many strain-dependent differences between the investigated S. aureus strains (see Table S1 in the supplemental material). Importantly, the only proteins that were common in the media of all tested agr⁺ strains and absent from the media of all tested Δagr strains were PSMα2, PSMα3, PSMα4, PSMβ1, PSMβ2, and the staphylococcal lipase 1. Only the identified PSM peptides have potential surfactant properties that might promote colony spreading, and therefore we focused all our following studies on these peptides.

Dissection of PSM function with synthetic peptides reveals major roles for PSMα3 and PSMγ in colony spreading.

Surfactants lower the surface tension of liquid-to-air interfaces, and due to this activity they facilitate the motility of many different bacterial species (44–46). To determine which of the previously identified PSM peptides might be involved in the spreading of S. aureus, we tested synthetic PSMα1-4, PSMβ1-2, PSMγ, and PSM-mec peptides for their ability to promote colony spreading. To this end, all PSMs were individually spotted at the same concentrations on soft agar plates. This analyses showed that the PSMα3 and PSMγ peptides strongly promoted the colony spreading by otherwise nonspreading Δagr strains (Fig. 3). Other PSMs—such as PSMα1, PSMα2, and PSMβ1—promoted colony spreading to lesser extents, whereas PSMα4, PSMβ2, and PSM-mec did not promote spreading at all (Fig. 3). Based on genetic studies, it was previously proposed that PSM-mec might inhibit colony spreading (47). We therefore tested the synthetic PSM-mec peptide for possible inhibitory effects, but unfortunately, no such effects were detectable in our experimental setting (Fig. 3, plates in bottom row). Furthermore, our studies show that N-terminally formylated PSMs promote spreading equally well as the nonformylated peptides (Fig. 3). It thus seems that both the formylated and nonformylated forms of PSMα3 and PSMγ contribute to the movement of S. aureus cells over wet surfaces. Importantly, both forms of PSMα3 and PSMγ exerted their effect on spreading in a dose-dependent manner (see Fig. S1 in the supplemental material). This underscores the view that these two PSMs actively contribute to colony spreading.

Fig 3.

Particular synthetic PSM peptides facilitate spreading of Δagr cells. To identify factors that facilitate spreading of Δagr cells (labeled “agr−” in the figure), chemically synthesized PSMs from S. aureus (plate rows 2 to 5) were spotted in the center of soft agar plates prior to the inoculation with S. aureus SH1000 Δagr cells. Both N-terminally formylated PSMs (marked with an “f” prefix) and nonformylated peptides were used in the assay. In addition, the PSM-mec peptides were tested for a potentially inhibitory role in spreading by cells of the S. aureus SH1000 agr⁺ strain (bottom row). All spreading assays were repeated at least five times.

To investigate why different PSMs promote the spreading of S. aureus cells to different extents, we compared their surfactant properties on soft agar plates by spotting 2 μl of each peptide solution in the center of such a plate (without adding bacteria) and by subsequently measuring the diameter of the resulting transparent halo. This revealed that PSMα3, PSMγ, and PSMβ2 have the strongest surfactant properties (Table 2). Furthermore, we assessed other potentially relevant parameters of the investigated PSMs, namely, the grand average of hydropathicity (GRAVY), the hydrophobic moment (μH), the net charge (z), the number of polar residues plus glycine, and the number of nonpolar residues. The results show that the PSMs that most effectively promote the spreading of S. aureus, namely, PSMα3 and PSMγ, combine high surfactant activities with low hydropathicity (Table 2). The other investigated PSM properties seem to be of minor relevance for spreading activity. Notably, while PSMβ2 has relatively strong surfactant properties, it has a high hydropathicity, suggesting that this high hydropathicity may counteract the promotion of spreading. Consistent with this view, PSM-mec combines the lowest surfactant properties with the highest overall hydropathicity (Table 2).

Table 2.

Properties of S. aureus PSM peptides^a

Parametera	PSMα1	PSMα2	PSMα3	PSMα4	PSMβ1	PSMβ2	PSMγ	PSM-mec
Surfactant properties (mean length [cm] ± SD)	1.0 ± 0.4	1.9 ± 0.1	7.4 ± 0.5	0.6 ± 0.1	1.4 ± 0.4	4.3 ± 2.4	6.3 ± 0.4	0.6 ± 0.1
Grand avg of hydropathicity (GRAVY)	0.957	0.890	0.305	1.700	0.570	0.607	0.150	1.100
Hydrophobic moment, <μH>	0.551	0.562	0.563	0.599	0.308	0.214	0.587	0.530
Net charge z	2	3	2	2	–1	–1	–1	–1
Polar residues + GLY (no./%)	10/47.62	10/47.62	10/45.45	6/30.00	23/52.27	22/50.00	14/53.85	10/45.45
Nonpolar residues (no./%)	11/52.38	11/52.38	12/54.55	14/70.00	21/47.73	22/50.00	12/46.15	12/54.55

^aSurfactant properties were determined by spotting 2 μl of a PSM solution (6 to 1.2 mM) on a soft agar plate and measuring the diameter of the resulting transparent halo. The grand average of hydropathicity (GRAVY) was determined using the ProtParam tool provided by the ExPASy server. The hydrophobic moment (μH), the net charge (z), the amount of polar residues + GLY, and the amount of nonpolar residues were obtained by using Heliquest (http://heliquest.ipmc.cnrs.fr/cgi-bin/ComputParams.py).

Genetic dissection of PSM function reveals additive effects in colony spreading.

As underscored by our proteomics analyses, agr-deficient strains are completely defective in the synthesis of PSMs. To identify the contributions of the psmα and psmβ operons to spreading, the respective single- and double-mutant strains were constructed. Next, the spreading ability of these mutants was compared to the parental strain and to the equivalent agr mutant, which is fully PSM deficient. To quantify spreading activity, all images of spreading assays were analyzed with ImageJ, and the area covered by the cells was determined. Figure 4 shows that deletion of the psmβ operon had a moderate but reproducible effect on spreading. Deletion of the psmα operon had a more severe effect on spreading and the spreading activity of the psmα psmβ double mutant was even lower (Fig. 4). Notably, the spreading activity of the double mutant was still higher than that of the agr mutant, which suggests that the remaining activity of the double mutant is due to the production of PSMγ. This view is consistent with the above finding that the synthetic PSMγ is sufficient to restore spreading of agr mutant strains (Fig. 3). A comparison of the results obtained for the single and double psmα psmβ mutant variants of strains Newman, SH1000, and HG001 shows that the deletion of the psmα operon has the strongest negative impact on spreading, which is fully consistent with the findings obtained with the synthetic peptides added at equimolar concentrations. Furthermore, our genetic dissection of the function of PSM-encoding loci shows that they contribute additively to colony spreading.

Fig 4.

Additive effects of psmα and psmβ gene deletions on the spreading of S. aureus cells. The psmα and/or psmβ loci of the S. aureus strains Newman, SH1000, or HG001 were deleted, and the effects on colony spreading were compared with the effects of an agr mutation. Subsequently, the spreading areas of the investigated mutant and parental strains were determined by ImageJ, and statistical analyses were performed based on triplicate measurements for each individual strain. The graphs show the areas covered in arbitrary units (AU).

Spreading activity of S. aureus cells from catheter-associated biofilms or on pork meat.

PSMs allow S. aureus to move across wet surfaces. In the assays described above this was demonstrated starting with planktonic cells growing in a broth. However, we wanted to know whether spreading would be detectable also in assays that mimic clinically relevant surfaces and conditions. In one approach, we therefore tested whether S. aureus cells present in a catheter-associated biofilm have the ability to spread. As shown in Fig. 5, both agr⁺ and Δagr strains were able to form biofilms on catheter material (plates 8 and 2, respectively). Interestingly, when the catheters were transferred to fresh TSA plates, the agr⁺ strains were well able to detach and spread away from the catheter material (plate 9). This phenomenon was not observed for the tested Δagr strains (plate 3). Importantly, however, when these catheters were transferred to fresh plates on which 2 μl of PSMα3 or PSMγ were spotted beforehand, the Δagr strains were also able to detach and spread away from the catheter (plates 4 and 5, respectively). In contrast, the PSMβ1 did not facilitate the spreading of Δagr cells from the catheter material (plate 6). These findings suggest that the ability of PSMs to promote spreading is important for S. aureus to move away from a biofilm and to colonize the surrounding wet surface. This idea was further tested by studying colony spreading on fresh pork meat. As predicted, agr⁺ cells were substantially more efficient than Δagr cells in colonizing pieces of pork meat upon 48 h of incubation at 37°C. In the experiment shown in Fig. 6A, the agr⁺ cells colonized an area that was on average ∼2.5-fold larger than the area colonized by Δagr cells. Importantly, the spreading phenotypes of psm mutant strains on fresh pork meat (Fig. 6B) resembled by-and-large the respective spreading phenotypes on soft agar plates (Fig. 4). While psmβ mutant cells were not significantly inhibited in their spreading on pork, the psmα mutant cells showed a significant spreading defect. Spreading on pork meat was most severely affected by deletion of both the psmα and psmβ genes, even to a slightly higher extent than agr mutant cells. It thus seems that PSM-mediated colony spreading has a general role in the colonization of wet surfaces by S. aureus.

Fig 5.

The spread of S. aureus cells from catheter-associated biofilms is facilitated by particular PSM peptides. (A) To investigate whether the spreading of cells from a catheter-associated biofilm is facilitated by PSMs, biofilms of Δagr cells of S. aureus Newman were grown on ∼1-cm-long strips of catheter material. These strips were then incubated on soft agar plates under differing conditions. Plate 1, 1egative control plate, showing that the used S. aureus Newman Δagr strain is unable to spread; plate 2, biofilms of Δagr cells were grown on catheter strips placed on a soft agar plate as shown with plate 2; plate 3, catheter strip with a biofilm (as on plate 2) transferred to a fresh soft TSA plate without further additions. Some “outgrowth” of the cells is observed but no spreading. Plate 4, catheter strip with a biofilm (as on plate 2) transferred to a soft TSA plate to which the PSMα3 peptide was added prior to the positioning of the catheter strip; plate 5, catheter strip with a biofilm transferred to a plate with the PSMγ peptide (as in plate 4); plate 6, catheter strip with a biofilm transferred to a plate with the PSMβ1 peptide (as in plate 4). (B) Control experiments with agr⁺ cells of S. aureus Newman. Plate 7, positive control plate showing colony spreading of agr⁺ cells; plate 8, colony spreading of Newman agr⁺ cells from a catheter strip; plate 9, catheter strip from plate 8 transferred to a fresh soft TSA plate without added PSMs. All spreading assays were repeated at least five times.

Fig 6.

Spreading of S. aureus on meat. (A) Overnight grown S. aureus SH1000 agr⁺ or Δagr (labeled “agr−” in the figure) cells were spotted on pork meat, which was subsequently incubated 48 h at 37°C. SH1000 agr⁺ cells covered a 2.5-fold larger area than SH1000 Δagr cells. The spreading areas are marked with dashed lines. (B) Spreading areas of agr⁺ and Δagr S. aureus SH1000 cells, and single or double psmα and/or psmβ mutant strains upon spotting onto fresh pork meat. All spreading assays were repeated at least three times.

DISCUSSION

Colony spreading and PSMs.

The present studies have focused attention on the role of secreted factors that are needed for the rapid colony spreading phenotype of S. aureus. The first indication for an important role of secreted factors in colony spreading was the presence of a transparent halo around the expanding colonies. Subsequent analyses showed that culture supernatants of spreaders were sufficient to make nonspreaders move on a wet soft agar surface and that the main common components in the media of spreaders were PSMs. Importantly, culture supernatants of spreaders facilitated the spreading of nonspreaders only when these supernatants were added on top of the soft agar. This is in accordance with the apparent need for high surfactant properties of the PSMs in order to be able to promote spreading. In contrast, when the culture supernatants of spreaders were included within the soft agar, they were unable to promote the spreading of non-spreaders and they even seemed to inhibit the spreading by spreaders (Fig. 2C). With synthetic PSM peptides we subsequently demonstrated conclusively that several PSMs of S. aureus are sufficient to promote colony spreading of otherwise nonspreading strains. This view was confirmed by mutagenesis experiments in which the psmα and/or psmβ operons were deleted. Taken together, our present findings show that PSMα3 and PSMγ are the key players in colony spreading, and that the other PSMα's and PSMβ's have minor roles in spreading. We observed no role for the PSM-mec neither in the promotion nor the inhibition of spreading. The latter observations are intriguing, because it has been reported that the PSM-mec and/or the PSM-mec mRNA can inhibit colony spreading (47). Since the synthetic PSM-mec peptides gave no phenotype, whereas other synthetic PSM peptides were active, it seems most likely that the previously reported effects do not relate to a translated product but rather to a regulatory effect of the psm-mec gene. We were unable to assess this possibility with the strains used in our studies, because they lack type II or type III SCCmec elements that encode PSM-mec. In any case, it is safe to conclude from our studies that addition of the PSM-mec peptide did not interfere with the function of the main spreading-promoting PSMα3 and PSMγ peptides. Indeed, an independent study confirms that deletion of the psm-mec gene does not affect the production of the other PSMs compared to the respective parental strains (11). Lastly, it was previously reported that S. aureus secretes both N-terminally formylated and deformylated PSM peptides. Our present data show that the removal of the N-terminal formyl group has no consequences with respect to the activity of PSMs in colony spreading. This indicates that the surfactant properties of PSM peptides are not substantially influenced by N-terminal formylation or deformylation.

At present, we do not know at what concentrations exactly the different PSMs need to be present locally to promote spreading. In this respect, one has to bear in mind that both the synthetic and the natural PSM peptides investigated in the present studies will diffuse not only over the soft agar surface but also into the agar underneath the surface. In addition, in a regular spreading assay, the S. aureus cells applied to a soft agar plate will initially make use of the PSMs present in the growth medium, which explains the first observed wave of “rapid” spreading. Subsequently, these cells will produce additional PSMs, allowing them to spread further over the plate. Therefore, it is very difficult to make statements about local concentrations of the PSM peptides on plates and how these correlate with the concentrations present in the spent medium. The situation is somewhat different when synthetic PSM peptides are used for a spreading assay, since this essentially represents the first wave of ‘rapid’ spreading. To this end, relatively high amounts of PSM peptides were initially tested. However, as shown by dose response curves with the PSMs that promote effective spreading (see Fig. S1 in the supplemental material), we can mimic effective spreading with concentrations that are only ∼10-fold higher than the concentrations measured in the growth medium (9). We believe that this is a realistic situation, because bacteria that are spreading with the aid of PSMs secreted into the growth medium can employ a mix of several PSMs for spreading. Moreover, the Agr-deficient cells that were used for the spreading assays cannot produce any PSMs by themselves in the course of an assay, while Agr-proficient cells will produce their own PSMs. In this context it is important to note that when the concentrations of PSMs were stepwise increased, higher spreading levels were achieved. This mimics in some way the situation where Agr-proficient cells produce several “waves of PSMs” during the course of a colony spreading experiment.

Is colony spreading by S. aureus clinically relevant?

A key question in the analysis of colony spreading is whether this property is clinically relevant. As a first approach to answer this question, we recently tested the spreading ability of 500 different clinical isolates that are representative for invasive S. aureus infections in Europe. More than 85% of these strains were able to spread (our unpublished observations). While this does not tell us that spreading was important for the actual infections, this finding does show that spreading is a very common feature of strains that have caused invasive infections in humans. Accordingly, it is conceivable that invasive strains make use of their spreading ability to move away from catheters and other implanted devices so that they can efficiently colonize wet surfaces of the human body. Clearly, our experiments with catheter material show that S. aureus cells originating form a biofilm have a similar spreading ability as planktonic cells. In this light it is not surprising that we find the same PSMs (i.e., PSMα3 and PSMγ) to be most effective in both types of assays. These findings therefore suggest that spreading may be a clinically relevant staphylococcal trait.

It has been observed that most clinical S. aureus isolates are agr⁺, but Δagr isolates are also isolated from patients. This population heterogeneity is likely to be advantageous for S. aureus since agr⁺ strains are more potent in initiating infections (48), while the Δagr strains are more potent in the establishment of chronic infections through biofilm formation (49). Our finding that nonspreading cells can “hitch-hike” along with the spreaders on soft agar plates would suggest that a similar phenomenon might “help” Δagr cells also in the colonization of wet surfaces, either in the human host or in other habitats. One of these other habitats might be animal meat intended for human consumption. There are many reports on the contamination of meat products with S. aureus, including MRSA (50–52). This is highly unwanted, first, because S. aureus is renowned as a causative agent of food poisoning (53). In addition, the contamination of food products with S. aureus, MRSA in particular, is a high-risk factor for frail and immunocompromised individuals who are more susceptible for staphylococcal infections. As shown by our experiments with pork, agr⁺ cells can colonize larger surfaces of the investigated meat in shorter periods of time than Δagr cells, reflecting the differences in their spreading ability on soft agar plates. Also, spreading on pork meat was reduced when the psmα genes were deleted, and it was even more reduced when both the psmα and psmβ genes were deleted. Thus, we hypothesize that spreading is an important parameter at least in food spoilage and, consequently, in food poisoning by S. aureus. It will remain a challenge for future studies to verify this hypothesis and to pinpoint any other potentially clinically relevant roles of staphylococcal spreading.