Contribution of SecDF to Staphylococcus aureus resistance and expression of virulence factors
© Quiblier et al; licensee BioMed Central Ltd. 2011
Received: 20 January 2011
Accepted: 12 April 2011
Published: 12 April 2011
Skip to main content
© Quiblier et al; licensee BioMed Central Ltd. 2011
Received: 20 January 2011
Accepted: 12 April 2011
Published: 12 April 2011
SecDF is an accessory factor of the conserved Sec protein translocation machinery and belongs to the resistance-nodulation-cell division (RND) family of multidrug exporters. SecDF has been shown in Escherichia coli and Bacillus subtilis to be involved in the export of proteins. RND proteins can mediate resistance against various substances and might be of relevance in antimicrobial therapy. The role of RND proteins in Staphylococcus aureus has not yet been determined.
Markerless deletion mutants were constructed to analyze the impact of the so far uncharacterized RND proteins in S. aureus. While the lack of Sa2056 and Sa2339 caused no phenotype regarding growth and resistance, the secDF mutant resulted in a pleiotropic phenotype. The secDF mutant was cold sensitive, but grew normally in rich medium at 37°C. Resistance to beta-lactams, glycopeptides and the RND substrates acriflavine, ethidium bromide and sodium dodecyl sulfate was reduced. The secDF mutant showed an aberrant cell separation and increased spontaneous and Triton X-100 induced autolysis, although the amounts of penicillin-binding proteins in the membrane were unchanged. The impact of secDF deletion on transcription and expression of specific virulence determinants varied: While coagulase transcription and activity were reduced, the opposite was observed for the autolysin Atl. A reduction of the transcription of the cell wall anchored protein A (spa) was also found. The accumulation of SpA in the membrane and lowered amounts in the cell wall pointed to an impaired translocation.
The combination of different effects of secDF deletion on transcription, regulation and translocation lead to impaired cell division, reduced resistance and altered expression of virulence determinants suggesting SecDF to be of major relevance in S. aureus. Thus SecDF could be a potential target for the control and eradication of S. aureus in the future.
Staphylococcus aureus is a frequent colonizer of the human body as well as a serious human pathogen. It is known for its adaptability to diverse environments. It can cope with stress factors and acquire resistances to antibiotics thus rendering treatment difficult. S. aureus can cause a wide range of infections, mainly due to an impressive arsenal of virulence determinants comprising cell surface components and excreted factors interacting with the host system. Transport of proteins to the cell surface and secretion to the extracellular space is mediated through different transport systems  of which the general protein secretion system Sec plays a prominent role in protein export and membrane insertion.
Sec-mediated translocation has best been studied in Escherichia coli and is catalyzed by the essential SecYEG protein complex (reviewed in ). The motor ATPase SecA or a translating ribosome is believed to promote protein export by driving the substrate in an unfolded conformation through the SecYEG channel. The accessory SecDF-YajC complex facilitates protein export and membrane protein insertion efficiency in vivo , possibly via the control of SecA cycling . The large exoplasmic loops of the integral membrane proteins SecD and SecF have been shown to be required for increasing protein translocation by a yet unknown mode of action . While secDF disruption leads to a cold-sensitive phenotype and defects in protein translocation , the absence of YajC, which interacts with SecDF, causes only a weak phenotype . SecYEG has been shown to interact with the SecDF-YajC complex . YidC, a protein that is proposed to mediate membrane integration and the assembly of multimeric complexes, can also interact with SecDF-YajC to take over SecYEG-dependent membrane proteins .
Data on the S. aureus Sec system is scarce: SecA and SecY have been shown to be important, respectively essential, for growth by using antisense RNA . Deletion of secG resulted in an altered composition of the extracellular proteome, which was aggravated in a secG secY2 double mutant . Deletion of secY2 alone, which together with secA2 belongs to the accessory Sec system , did not show any effect on protein translocation. As in the Gram-positive bacterium Bacillus subtilis, in S. aureus the accessory SecD and SecF proteins are fused to form a single protein (Sa1463), which was identified in membrane vesicles . However, the chromosomal organization in S. aureus resembles the one of E. coli, with yajC lying immediately upstream of secDF. Furthermore, SecDF was identified in a surface-exposed peptide epitope screen by using a cell shaving technique  and expression was found to be slightly higher in a COL sigB deletion mutant . SecDF is postulated to be essential in S. aureus according to a mutagenic screen .
SecDF belongs to the resistance-nodulation-cell division (RND) family of multidrug export pumps, that is conserved and widely distributed in all three major kingdoms of life . RND proteins have a wide substrate specificity and diverse functions ranging from the efflux of noxious host derived substances, such as bile salts by E. coli  to the involvement of eukaryotic efflux pumps in cholesterol homeostasis in humans . Multiple antibiotic resistance can be associated with these exporters, as they often recognize a broad range of substrates, thereby diminishing drug accumulation in the cell [20, 21]. S. aureus possesses two additional uncharacterized RND proteins, namely Sa2056, located downstream of the essential femX , and Sa2339 (MmpL homologue).
To evaluate the role and impact of the RND proteins in S. aureus, markerless deletion mutants were constructed in the sequenced and well-characterized clinical strain Newman. SecDF, Sa2056 and Sa2339 were found to be dispensable, as we obtained null mutants by allelic replacement of the corresponding genes using the pKOR1 system of Bae et al. . The mutants were confirmed to have generally retained genome stability and to carry the desired modification in the corresponding locus as described in methods.
Deletion of sa2056 and sa2339 had no apparent effect on S. aureus when evaluating growth and resistance properties (data not shown), suggesting that they may be important under other conditions than applied in this study. The following report is therefore focused on the secDF mutant and its phenotype.
As secDF knock out mutants in B. subtilis and E. coli show a cold sensitive phenotype [6, 24], growth of the S. aureus secDF mutant was tested at 15°C. The temperature drop affected the secDF mutant approximately after two generations, causing a notably reduced growth rate with a subsequent halt in growth after 24 h. The plasmid pCQ27, but not the empty vector pCN34, significantly restored growth at 15°C (Figure 1B).
To determine whether wild type and mutant bacteria produced different levels of hydrolases, their activity was analyzed in concentrated supernatant and cell wall extracts (Figure 5B). In the supernatant of the mutant, high molecular mass bands matching different forms of the major S. aureus autolysin Atl , were expressed similarly (>130 kDa, pro-Atl) or even stronger (~84 kDa, PP-AM) compared to the wild type and the complemented mutant (Figure 5B). Interestingly, the >130 kDa band migrated at a slightly higher position in the mutant, corresponding to the height of the pro-Atl band in the cell wall fractions, where the mutant showed overall stronger hydrolytic bands than wild type or complemented mutant.
We qualitatively assessed the amount of various Sec-dependent S. aureus virulence factors, such as coagulase, hemolysin and protease activities, as well as of the immunomodulatory protein SpA to determine whether they were affected in the secDF mutant as well.
SpA is one of the proteins predicted to be attached to the cell wall by sortase following export . SpA levels were determined in subcellular fractions during growth by Western blot analyses.
Surprisingly, secreted SpA amounts were fairly constant despite this translocation defect. Also in the wild type, SpA levels in the supernatant were constant, whereas the amount of cell wall-bound SpA increased during growth, suggesting constant shedding of this protein.
Efflux pumps play an important role in S. aureus resistance, virulence and pathogenicity. Yet the impact of the RND family of efflux pumps in staphylococcal resistance and fitness is still open (reviewed in ). To our knowledge, this is the first study to evaluate their role in S. aureus.
We found SecDF to contribute probably in part indirectly to resistance against several substances, including β-lactams and glycopeptides, making it an interesting target for increasing the efficacy of these standard antibiotics. In contrast Sa2056 and Sa2339 seemed not to be required for growth and resistance under the conditions tested. Banerjee et al. recently had found a conservative amino acid mutation in Sa2056 in a high-level β-lactam resistant mecA-negative strain . However in that strain PBP4 and Sa0013 were also mutated and the exact reason for the observed resistance phenotype was not identified.
Resistance against cell wall active antibiotics and cell separation is dependent on a tightly balanced regulation of cell wall synthetic and hydrolytic enzymes, including their timely localization to the septum [43, 44]. The amount of PBPs 1-4 and PBP2a was apparently not influenced, suggesting that other factors important for cell division and β-lactam resistance were affected. The increased hydrolytic activity in the secDF mutant may explain the observed differences in cell wall production and separation. Overproduction of a hydrolase has been observed to affect formation of the FtsZ-ring in Mycobacterium tuberculosis . This cytoskeleton structure recruits the other cell division proteins to the site of future cell separation. A similar indirect effect in the secDF mutant might have lead to an incorrect localization of the cell division machinery, including PBPs (for a general review see ), thereby causing reduced resistance against the cell wall active antibiotics oxacillin and vancomycin. The difference in Atl processing might have impeded proper cell separation in addition.
Like E. coli and B. subtilis secDF mutants [6, 24], the S. aureus secDF mutant displayed a cold-sensitive phenotype. In E. coli and B. subtilis SecDF has furthermore been shown to participate in membrane integration and secretion of proteins [6, 24, 47]. In S. aureus many physiological functions were affected by the secDF deletion. Analysis of the secretion of classical S. aureus virulence factors containing a Sec-type signal peptide revealed a complex picture. Coagulase and proteases were reduced in the supernatant in the secDF mutant. However, hemolysin activity under planktonic growth conditions was increased in the mutant, as was the case for (unprocessed) hydrolases, indicating that the secDF deletion did not lead to an overall reduction, but to an altered secretion and processing of proteins. In contrast hemolysin activity was reduced during sessile growth indicating that the deletion of secDF may have effects on overall metabolism.
SpA seemed to be impaired in reaching its destined subcellular localization. In the secDF mutant SpA accumulated in the membrane, was reduced in the cell wall fraction but was found in increased amounts in the supernatant. Altered secretion and processing of SpA might be due to impaired cell wall anchoring by the membrane protein sortase. However, Mazmanian et al. have shown that the extracellular enterotoxin B fused to the sorting signal of SpA accumulates in the cytoplasm and to a lesser extent in the membrane in a sortase mutant . Thus, SpA might migrate by an alternate mechanism into the supernatant, circumventing linking to the peptidoglycan.
A similar divergent effect on protein secretion as we observed in the secDF mutant was found in a secG mutant. There SpA was found in increased amounts in the exoproteome, despite unaffected transcription . In contrast, we found deletion of secDF to change mRNA levels for many of the analyzed genes, such as atl, coa, hla, hld and spa. The lack of secDF therefore seems to have a different impact on virulence factor expression than secG, influencing, most likely indirectly, transcription in addition to translocation. The absence of SecDF could especially cause a defective or reduced membrane insertion of sensor proteins belonging to one of the numerous S. aureus two component systems contributing to virulence factor regulation and to adaptations to different growth conditions (reviewed in [49, 50]). The reduced hld levels in the mutant suggests that the secDF deletion affected at least one two component system by impairing signaling via the agr quorum sensor .
This study and the work of Sibbald et al.  once more demonstrate that protein and mRNA levels do not necessarily correlate. Specific regulation at the protein level has been shown for certain transcription factors in S. aureus [52, 53]. Such a control of protein stability via chaperones and proteases might exist as well for virulence factors. Interestingly, in E. coli, secY, yidC and secD mutants were shown to induce the Cpx system, which up-regulates the expression of factors involved in folding and proteolysis in response to abnormal proteins in the outer membrane, the periplasmic space or the plasma membrane . The induction of similar systems in the S. aureus secDF mutant due to clogging of the membrane, as suggested by the increased amounts of SpA in this compartment, could be an additional factor influencing protein stability and lead to the partially incoherent mRNA and protein levels, as seen for hla, hld and spa during planktonic growth.
This work provides evidence that although secDF is dispensable in S. aureus, its deletion leads to a pleiotropic phenotype. Lack of SecDF affected cell separation, resistance and virulence factor expression showing that this conserved RND protein plays a major role in the important human pathogen S. aureus. Thus SecDF could be a potential therapeutic target rendering S. aureus more susceptible to the currently available antibiotics.
Strains and plasmids used in this study
Relevant genotype or phenotype
Ref. or source
Clinical isolate (ATCC 25904), rsbU +
NCTC8325-4 r- m+
Newman pME2, Tcr, Mcr
NewmanΔsecDF pME2, Tcr, Mcr
Newman pCN34, Kmr
Newman pCN34 pME2, Kmr, Tcr, Mcr
NewmanΔsecDF pCN34, Kmr
NewmanΔsecDF pCN34 pME2, Kmr, Tcr, Mcr
NewmanΔsecDF pCQ27, Kmr
NewmanΔsecDF pCQ27 pME2, Kmr, Tcr, Mcr
Cloning strain, [F-Φ80lacZΔM15 Δ(lacZYA-argF)U169 recA1 endA1 hsdR17 (rk-, mk+) phoA supE44 thi-1 gyrA96 relA1 λ-]
Relevant genotype or phenotype
Reference or source
S. aureus-E. coli shuttle vector, pT181-cop-wt repC aphA-3 ColE1 Kmr
pCN34 derivative carrying secDF and its promoter (Newman), Kmr
pKOR1 derivative carrying 1 kb fragments of the region up- and downstream of sa2056 amplified from Newman, ligated together with EcoRI and recombined at the attP sites, Apr, Cmr
pKOR1 derivative carrying 1 kb fragments of the region up- and downstream of sa2339 amplified from Newman, ligated together with HindIII and recombined at the attP sites, Apr, Cmr
pKOR1 derivative carrying 1 kb fragments of the region up- and downstream of secDF amplified from Newman, ligated together with HindIII and recombined at the attP sites, Apr, Cmr
E. coli-S. aureus shuttle vector used to create markerless deletions; repF(Ts) cat attP ccdB ori ColE1 bla P xyl /tetO secY570, Apr, Cmr
pBUS1 derivative carrying mecA and its promoter (COLn), Tcr, Mcr
In-frame markerless deletions of sa2056 (NWMN 2'384'867-2'388'051), sa2339 (NWMN 2'696'046-2'698'531) and secDF (NWMN 1'706'584-1'708'866) from the chromosome of S. aureus Newman (accession number NC_009641) was performed using pKOR1  yielding single mutants CQ33, CQ65 and CQ66, respectively. Correct deletion was confirmed by PCR and by sequencing. Furthermore, strain stability was confirmed by pulsed field gel electrophoresis of total genome SmaI digests .
To complement the secDF mutant, secDF with its endogenous promoter was amplified from S. aureus strain Newman with primers listed in additional file 2 table S1.
The amplified region was ligated into the SalI/BamHI restriction sites of pCN34, a low copy (20-25 copies/cell) E. coli-S. aureus shuttle vector . The junction region was sequenced as a control. The resulting plasmid pCQ27 was electroporated into RN4220 with subsequent transduction into the strains of interest.
To construct MRSA strains, the plasmid pME2, containing the mecA promoter and gene from strain COLn , was either electroporated or transduced into the strains selected.
Promoter predictions were performed by BPROM http://linux1.softberry.com/berry.phtml. Rho-independent transcriptional terminators were retrieved from the CMR terminator list http://cmr.jcvi.org/tigr-scripts/CMR/CmrHomePage.cgi.
Cells were grown to exponential phase, harvested at OD600 0.5 and fixed for one hour in 2.5% glutaraldehyde in phosphate buffered saline (PBS) pH 7.4. Electron microscopy was performed by the Center for Microscopy and Image Analysis, University of Zurich.
For qualitative susceptibility comparisons, bacterial suspensions of McFarland 0.5 were swapped across LB agar plates containing antibiotic gradients and incubated at 35°C for 20-24 h. Glycopeptides were tested on Brain Heart Infusion (BHI) (Difco) agar with a bacterial suspension of McFarland 2 .
Cells were grown to an OD600 of 0.7, pelleted by centrifugation and washed with 0.85% NaCl. The cells were then resuspended in 0.01 M Na-phosphate buffer pH 7 and the OD600 was adjusted to 0.7. After splitting the cultures, 0.01% Triton X-100 (Fluka) or an equal volume of PBS pH 7 was added. Cultures were incubated at 37°C and the decrease of OD600 was measured.
Cultures were grown to an OD600 = 0.7, centrifuged and the filtered supernatants (pore size 0.45 μm, TPP) stored at - 20°C until further use. The cell wall peptidoglycan was digested in SMM buffer (0.5 M sucrose, 0.02 M maleate, 0.02 MgCl2 pH 6.5) supplemented with 72 μg/ml lysostaphin and 2 mM phenylmethylsulfonyl fluoride (PMSF) . Cell wall containing supernatant was separated from the protoplasts and stored at - 20°C until further use. Protein concentrations were measured by Bradford assay (BioRad).
Twenty μg of protein from each fraction was separated by SDS-10% polyacrylamide gel electrophoresis (PAGE) containing cell wall extract of heat-inactivated (1 hour at 100°C in 4% SDS) S. aureus (end concentration OD600 = 6). The gel was washed twice for 15 min in dH2O and incubated for 18 h at 37°C in 0.1 M Na-phosphate buffer pH 6.8. Afterwards the gel was incubated for 3 min in staining solution (0.4% methylene blue, 0.01% KOH, 22% EtOH) and destained in cold water for several hours. Murein hydrolase activities produced clear bands.
Overnight cultures were pelleted at full speed, 0.5 ml supernatant was transferred into fresh tubes and 2 mM PMSF was added. The supernatants were normalized to an OD600 of 1 of the original culture with PBS. 0.1 ml supernatant was added to 0.25 ml reconstituted rabbit plasma (BBL Coagulase Plasmas, BD) and incubated at 37°C. Every 30 min tubes were examined for coagulation.
Cells were grown overnight in Todd-Hewitt (TH) medium , which was originally developed for the production of streptococcal hemolysins . To visualize hemolysis production of sessile bacteria, overnight cultures were normalized to an OD600 = 1 in PBS pH 7.4. Fifty μl was dispensed into 5 mm wide holes punched into 5% sheep blood agar. Plates were incubated overnight at 37°C and then stored at 4°C. To determine hemolysis in liquid media, the overnight cultures grown in TH medium were normalized to the same OD600 with PBS and pelleted for 10 min at 5'900 g. The supernatant was filtered (pore size 0.22 μm, TPP) and 140 μl added to the holes in sheep blood agar. Plates were incubated as above.
Cells were grown for 24 h in TH medium and normalized with PBS pH 7.4 to the same OD600. After pelleting the cells, the filtered supernatants (pore size 0.22 μm, TPP) were diluted up to 1:50'000 in TH medium. Sterile sheep blood was treated with 26 mM sodium citrate and 15 mM NaCl and diluted 1:100 in PBS pH 7.4. After washing the erythrocytes four times in PBS pH 7.4, they were resuspended to a dilution of 1:100 in PBS pH 7.4. Five hundred μl of washed erythrocytes were added to 500 μl of the diluted supernatants and incubated for 30 min at 37°C, followed by 30 min at 4°C. Finally the samples were centrifuged for 1 min at 7'000 g and the absorption of hemoglobin in the supernatant was measured at 415 nm .
Skim milk agar plates were prepared as follows: Skim milk (Difco) and Bacto agar (Difco) were dissolved separately in 250 ml dH2O, each with an end concentration of 75 g/l and 15 g/l, respectively. After autoclaving for 15 min at 110°C and cooling down to 50°C, the skim milk and Bacto agar solutions were mixed together. Overnight cultures grown in LB broth were normalized to an OD600 = 1 with 0.85% NaCl and 50 μl was added into punched holes in skim milk agar. Skim milk agar plates were incubated at 37°C for 24 h and another 96 h at room temperature.
Prewarmed LB broth was inoculated with an overnight culture to an OD600 0.05 and incubated at 37°C. Cells were harvested at OD600 0.2, 0.5, 1, 3 and 6, centrifuged for 5 min at 20'000 g and 4°C. Cells were immediately snap frozen in liquid nitrogen and stored at - 80°C. Total RNA was extracted as described in . Seven μg RNA was separated in a 1.5% agarose gel containing 20 mM guanidine thiocyanate in 1× TBE . RNA was transferred onto a positively charged nylon membrane (Roche) using the downward capillary transfer method. The blots were hybridized with specific digoxigenin (DIG)-labeled DNA probes (Roche). Primers used are listed in Additional file 2 Table S1.
Cells were sampled as described for transcription analyses and culture supernatant was collected as described for zymographic analysis. Cells were fractionated basically according to Schneewind et al. . Briefly, cells were digested in SMM buffer supplemented with each 72 μg/ml lysostaphin and lysozyme, 36 μg/ml DNase and 2 mM PMSF. Protoplasts were separated from the cell wall containing supernatant by centrifugation for 4 min at 16'000 g. Protoplasts were resuspended in membrane buffer (0.1 M NaCl, 0.1 M Tris-HCl, 0.01 MgCl2 pH 7.5) and lysed by three cycles of freezing in liquid nitrogen/thawing at 20°C. Cell membranes were separated from the cytoplasm by centrifugation for 30 min at 20'000 g and 4°C. Membrane pellets were solubilized in buffer B (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM MgCl2, 30% glycerol) supplemented with 1% Triton X-100 and 0.5% N-lauroylsarcosine, by gently mixing end-over-end at 4°C. Where necessary, protein fractions were concentrated with Amicon Ultra-15, -4 or -0.5 centrifugal filter units (MWCO 10 kDa, Millipore). Cell fractions were kept at - 20°C.
Five μg of protein was separated by SDS-10% PAGEs and either stained with Coomassie Imperial™ Protein Stain (Thermo Scientific) or blotted onto a PVDF-membrane (Immobilon-P, Millipore). For detection of SpA, membranes were blocked with 5% milk powder in PBS and then incubated with goat anti-human IgA conjugated with horseradish peroxidase (HRP, Sigma-Aldrich), 1:10'000 in 0.5% milk powder/PBS, 0.05% Tween 20 (AppliChem). After washing three times with PBS pH 7.4, HRP was detected with SuperSignal West Pico Chemiluminescent substrate (Thermo Scientific). PBP2a was detected as described in . For detection of PBP4, membranes were blocked with 5% milk powder in PBS. Membranes were pre-incubated with 40 μg/ml human IgG in 0.5% milk powder/PBS. Rabbit anti-PBP4 antibodies (1:2000, ) and 0.05% Tween 20 were then added. After incubation for 1 h, membranes were washed three times with PBS before addition of goat anti-rabbit IgG-HRP (Jackson ImmunoResearch), 1:10'000 in 0.5% milk powder/PBS/0.05% Tween 20. After washing three times with PBS, HRP was detected as described for SpA. Molecular weights of PBP2a, PBP4 and unprocessed SpA are 76 kDa, 48 kDa and 56.7 kDa, respectively.
Hundert μg of cell membrane fraction were incubated for 30 min at 35°C with Bocillin-FL (Invitrogen) as described by  before separation by SDS-7.5% PAGE. Fluorescence was visualized with the FluorChem™ SP imaging system (AlphaInnotech).
We thank S. Burger for her technical help. We are thankful to U. Luethy (Center for Microscopy and Image Analysis, University of Zurich) for TEM analysis. We are grateful to Hitoshi Komatsuzawa for kindly donating the rabbit anti PBP4 antibodies.
This study was supported by the Swiss National Science Foundation grant 31-117707 to B. Berger-Bächi, the Gottfried und Julia Bangerter-Rhyner Stiftung as well as the Olga Mayenfisch Stiftung to C. Quiblier, and the Stiftung für Forschung an der Medizinischen Fakultät der Universität Zürich to A. S. Zinkernagel.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.