Functional Characterization of TcaA: Minimal Requirement for Teicoplanin Susceptibility and Role in Caenorhabditis elegans Virulence

Bacterial strains and culture conditions.The strains and plasmids used in this study are listed in Table 1. The strains were routinely cultured at 37°C in Luria-Bertani (LB) broth (Difco Laboratories, Detroit, MI) and on LB or sheep blood agar and stored as frozen stocks in skim milk at −80°C. Kanamycin (50 μg/ml), ampicillin (100 μg/ml), or erythromycin (5 μg/ml) was added to the medium when it was appropriate. Phage 80α was used for phage transductions.

Antibiotic resistance testing.Relative teicoplanin resistance levels were compared by swabbing 0.5 McFarland standard suspensions of strains across teicoplanin gradient plates. The MICs for teicoplanin were determined by an Etest method recommended to give high specificity and sensitivity for determining glycopeptide resistance levels (42). After 24 h of incubation at 35°C, Etest MICs were determined from brain heart infusion (BHI) plates containing kanamycin at 50 μg/ml that had been swabbed with 2.0 McFarland suspensions.

RNA extraction and Northern hybridization.RNA isolation and Northern blotting were performed as described previously (28). Ten micrograms of total RNA from each sample was separated through a 1.5% agarose-20 mM guanidine thiocyanate gel in 1× TBE (Tris-borate-EDTA) running buffer (15). The primers used for amplification of the digoxigenin (DIG)-labeled tcaA, produced by using a PCR DIG probe synthesis kit (Roche, Basel, Switzerland), have been described previously (25); and the position of the probe is shown in Fig. 1A.

• Open in new tab
• Download powerpoint

FIG. 1.

Transcript induction and TSS mapping of tcaA. (A) Map of tcaA and the surrounding genomic region showing the position of the tcaA DIG-labeled probe used for Northern blots and the positions of primers PE1 and PE2 used for primer extension. (B) Northern blot showing VraSR-dependent induction of tcaA by inhibitory concentrations of glycopeptide and β-lactam antibitiotics. N315 and its ΔvraSR deletion mutant KVR were grown to an OD₆₀₀ of 0.5 and were either unstressed (lanes −) or stressed with teicoplanin (10 μg/ml) (lanes T), vancomycin (10 μg/ml) (lanes V), or oxacillin (2× MIC) (lanes O) for 30 min before RNA extraction. Relative tcaA transcript levels are shown, and the corresponding ethidium bromide-stained 16S rRNA bands are shown below as an indication of RNA loading. (C) Determination of the TSSs of tcaA by primer extension. Two TSSs were detected and were located 2 nucleotides apart. The stronger signal detected with primer PE2 is indicated with a black arrow, and the weaker signal is indicated with a gray arrow. The same two TSSs with the same relative signal strengths were also obtained with primer PE1 (data not shown). The nucleotide sequences surrounding the TSSs are shown; the nucleotide corresponding to the stronger predicted TSS is in boldface and is indicated with an asterisk, and the nucleotide corresponding to the weaker predicted TSS is highlighted in gray. The predicted −10 box region is enclosed within a box.

Primer extension.RNA was extracted from cultures of strain BB1539 that were grown to an optical density (OD) of 1.0 and then stressed with 10 μg/ml of teicoplanin for 30 min. Primer extension was performed with 20 μg of total RNA and 3 pmol of primer PE1 (5′-GTACATATCTGTACATCATC-3′) or PE2 (5′-ATACATATAACAATTGAAGAG-3′) labeled at the 5′ end with biotin (Fig. 1A), by using Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. Sequencing reactions were performed with a Thermo Sequenase cycle sequencing kit (U.S. Biochemicals, Cleveland, OH). A biotin chromogenic detection kit (Fermentas, Burlington, Ontario, Canada) was used for biotin detection.

Construction of markerless tcaA deletion and plasmid deletions.The in-frame markerless deletion of tcaA from the chromosome of S. aureus COL was performed by using pKOR1, as described by Bae and Schneewind (3). Deletion of the region spanning nucleotides 2411159 to 2412541 in the S. aureus COL genome was confirmed by PCR and sequencing across the deleted region. Portions of the tcaA promoter and coding regions were deleted from plasmid pAW17-tcaA by PCR. Long-range PCR amplification, performed with an Expand Long Template PCR system (Roche) and primers flanking but facing outward from the region to be deleted, was used to amplify the entire pAW17-tcaA plasmid, excluding the portion of the gene or the promoter to be removed. Recircularization of the remaining portion of the plasmid was then performed by ligation, which was facilitated by the inclusion of HindIII restriction sites at the 5′ ends of both primers. All deletions were checked by sequencing across the region deleted. Deletion plasmids were first transformed into RN4220 and were then transduced into BB1372.

Membrane topology analyses.Membrane topology predictions were made by using several different web-based topology prediction algorithms, including TMHMM (version 2.0; http://www.cbs.dtu.dk/services/TMHMM-2.0/ ), DAS (10), TMpred (http://www.ch.embnet.org/software/TMPRED_form.html ), Split 4.0 (17), Phobius (18), and HMMTOP 2.0 (40).

PhoA fusions were constructed by cloning various regions of the tcaA gene downstream of the araB arabinose-inducible promoter and upstream of the signal sequence-less phoA gene in pHA-1. Regions of the S. aureus COL genome cloned to create the fusions spanned the following nucleotides: fusion 6, 2412539 to 2412396; fusion 5, 2412539 to 2412309; fusion 4, 2412539 to 2412012; and fusion 2, 2412539 to 2411160. The resulting plasmids were then transformed into Escherichia coli strain CC118. To assay PhoA activity, overnight cultures of fusion-containing strains were diluted 1:100 into fresh LB broth and grown to an OD of 600 nm (OD₆₀₀) of 0.5. The cultures were then split in two, 0.2% arabinose was added to one half, and both halves were incubated for a further 1 h. The cultures were then harvested, and PhoA activity was measured by a p-nitrophenylphosphate (Sigma-Aldrich, Steinheim, Germany) cleavage assay (30).

C. elegans survival assay.Nematode survival assays were performed as described previously (39). C. elegans strain Bristol N2 was maintained at 15°C on E. coli OP50 grown on nematode growth medium (8). To prepare the assay plates, an overnight saturated culture of S. aureus in tryptic soy (TS) broth was diluted 1:10 in fresh TS broth, and 10 μl of the diluted culture was spread on a 35-mm plate containing TS agar supplemented with 5 μg/ml nalidixic acid. The plates were incubated at 37°C for 4 h and then allowed to equilibrate at room temperature before being seeded with worms. Thirty hermaphrodite nematodes at the L4 developmental stage were transferred to each assay plate (three plates for each strain tested), and survival was monitored at 25°C by light microscopy. Nematodes that died because they crawled off the plate were censored from the analysis. Survival was calculated by the Kaplan-Meier product-limit method, and survival differences were tested for significance by using the log-rank test (GraphPad Prism, version 4.0). P values <0.05 were considered statistically significant. Similar results were obtained from four independently replicated experiments.

tcaA induction.The tcaA gene, a member of the S. aureus cell wall stress stimulon (26, 41), requires inhibitory antibiotic concentrations for induction (29). Uninduced tcaA transcription levels are very low, indicating weak tcaA expression under normal growth conditions (25). S. aureus strain N315 and S. aureus strain KVR, a ΔvraSR mutant of N315 (21), were used to determine if tcaA induction by cell wall antibiotics was dependent on the VraSR two-component sensor-transducer that is required for the induction of several members of the cell wall stress stimulon. Northern blots of RNA extracted from N315 and KVR unstressed and stressed with inhibitory concentrations of either teicoplanin (10 μg/ml), vancomycin (10 μg/ml), or oxacillin (concentrations corresponding to twice the MIC, i.e., 16 μg/ml for N315 and 2 μg/ml for KVR) showed strong tcaA induction by all three antibiotics in N315 but no induction in KVR (Fig. 1B).

Characterization of tcaA promoter region.Primer extension with primers PE1 and PE2 (Fig. 1A) identified two predicted transcriptional start sites (TSSs). The stronger TSS signal was 166 bp upstream of the predicted tcaA ATG start codon, and the second, weaker signal was a further 2 nucleotides upstream (Fig. 1C). The best potential RNA polymerase sigma factor consensus promoter sequence preceding the TSS was TTGAAC-N₁₄-TATAAT, which closely resembles the housekeeping σ^A promoter consensus sequence (TTGACA-N_16-18-TATAAT), although the spacer region of 14 bp is unusually short.

To identify the regions of the tcaA promoter required for VraSR-mediated induction of tcaA in response to cell wall stress, a series of tcaA promoter deletions was constructed in pAW17-tcaA (Fig. 2A). Plasmids containing the promoter deletions were introduced into the tcaRAB deletion mutant BB1372 (7, 25). Northern blots were then used to determine if the promoter mutants had altered tcaA induction phenotypes when they were induced by inhibitory concentrations of teicoplanin (Fig. 2B). As expected, there was no induction in deletion 7, in which the entire promoter region had been removed. Deletion 6 was the only other deletion which abolished induction. As the induction of tcaA in deletion 1 was comparable to the induction in the wild-type pAW17-tcaA plasmid, the 27-bp region between deletions 1 and 6 must be required for induction. A notable feature of this region is that it contains half of a 13-bp inverted repeat (ΔG = −9.8), the other half of which covers the −35 region of the predicted promoter consensus (Fig. 2C). The untranslated tcaA leader region, between the TSS and the ATG translational start codon, contained several short direct and inverted repeats (Fig. 2C), indicating that this region may be prone to genetic alterations. A BLAST search analysis (http://www.ncbi.nlm.nih.gov/BLAST/ ), however, revealed that the entire 239-bp tcaR-tcaA intergenic region was highly conserved (>99% identity) in all currently available S. aureus genome sequences. Deletions in this region appeared to lead to increased tcaA transcription upon teicoplanin induction (Fig. 2B).

• Open in new tab
• Download powerpoint

FIG. 2.

tcaA promoter region required for induction. (A) Map of the intergenic region between tcaR and tcaA showing the positions of promoter deletions. The deletions encompass the regions between the following nucleotide coordinates from the genome of S. aureus COL: 1, 2412808 to 2412769; 2, 2412667 to 2412637; 3, 2412667 to 2412577; 4, 2412667 to 2412547; 6, 2412808 to 2412742; 7, 2412808 to 2412547; 9, 241269 to 2412577; and 10, 2412691 to 2412547. (B) Northern blot showing tcaA transcript induction from the intact pAW17-tcaA plasmid (BB1539) and each of the promoter deletion mutants. Strains were grown to an OD₆₀₀ of 0.5 and were either unstressed (lanes −) or stressed with teicoplanin (10 μg/ml) (lanes +) before RNA extraction. Ethidium bromide-stained 16S rRNA bands are shown beneath the transcripts as an indication of RNA loading. The gel at the bottom shows the lanes corresponding to promoter deletion 6 and is a longer exposure of a Northern blot showing that tcaA is still transcribed at a low level in deletion 6 and that transcription is not influenced by teicoplanin induction. (C) Expanded map of the region upstream of the TSS required for tcaA induction. The nucleotide positions marking the ends of deletion 1 and deletion 6 are indicated. The major predicted TSS is labeled, the predicted −10 and −35 promoter consensus sequences are underlined, and the arrows beneath the nucleotide sequence show the position of the inverted repeat. The positions of inverted and direct repeats present in the untranslated leader region between the TSS and the translational start site are also indicated by arrows, with each repeat differentially shaded. (D) Effect of promoter deletions on complementation of teicoplanin susceptibility. The resistance levels of BB1372 strains containing pAW17-tcaA promoter deletions were compared to those of BB1372 containing the intact plasmid (BB1539) and the empty pAW17 plasmid (BB1541) by swabbing 0.5 McFarland standard suspensions of the strains across a teicoplanin gradient of 0 to 5 μg/ml. MIC*, MIC values were determined by Etest by using 2.0 McFarland suspensions on BHI supplemented with kanamycin at 50 μg/ml.

The same promoter deletion plasmids were also tested for their ability to restore teicoplanin susceptibility in tcaRAB deletion mutant BB1372. The gradient plate and the MICs in Fig. 2D show the differences in teicoplanin resistance when strain BB1372 was complemented with intact pAW17-tcaA (BB1539), the empty pAW17 plasmid (BB1541), or the pAW17-tcaA promoter deletion mutants. Deletions 4, 7, and 10, all of which have disrupted ribosome binding sites, were the only deletions that could no longer complement teicoplanin susceptibility, presumably due to the abortive effect of the partial ribosome binding site deletion on TcaA translation. Deletion 6, which was defective in tcaA induction, was still able to complement teicoplanin susceptibility, suggesting that only small amounts of the TcaA protein, such as those present in uninduced cells, are sufficient to restore teicoplanin susceptibility.

TcaA protein structure prediction and membrane topology.Web-based analyses of the secondary structure and topology of the TcaA protein sequence indicated that it is a transmembrane protein, with a short 50-amino-acid (aa) N terminus containing a C-4-type zinc finger motif, a single 20-aa membrane-spanning domain, and a large 390-aa C-terminal domain. Four of the six topology prediction programs used predicted that the N terminus was cytoplasmic and that the C terminus was extracellular.

PhoA fusions were used to confirm that tcaA was membrane spanning and to determine the topology. A set of plasmids containing various portions of the tcaA gene fused to phoA (Fig. 3A) were constructed in plasmid pHA-1, which contains a signal sequence-less phoA gene and the araB arabinose-inducible promoter, and introduced into E. coli phoA mutant strain CC118. The expression of PhoA activity requires disulfide bond formation, which occurs only in the periplasm of E. coli. Therefore, PhoA activity is detected only when the portion of the protein cloned in front of the signal sequence-less PhoA is directed to the periplasm during arabinose induction.

• Open in new tab
• Download powerpoint

FIG. 3.

Structure prediction and membrane topology of TcaA. (A) Map of tcaA showing the predicted N-terminal, transmembrane, and C-terminal domains. The position of the C-X₂-C-X₁₀-C-X₂-C sequence in the N terminus is indicated by vertical lines. The positions where phoA was fused to portions of the tcaA gene, creating fusions 2, 4, 5, and 6, are indicated by arrowheads. (B) Relative PhoA activities of fusions 2, 4, 5, and 6 with and without arabinose induction. The values shown represent the means ± standard deviations from three independent assays.

No PhoA activity was detected from fusion 6, which contained only the N-terminus region up until just before the transmembrane region, which indicated that this region was cytoplasmic. Fusion 5 (which contained the N terminus and the transmembrane region), fusion 4 (which contained approximately the first half of the gene), and fusion 2 (which contained the entire tcaA gene minus the last codon) all produced high PhoA activities when they were induced with arabinose (Fig. 3B), indicating that all of these fragments directed the PhoA reporter protein into the periplasm. Therefore, the PhoA fusion results were in agreement with the majority of the topology prediction results.

Regions of TcaA required for complementation of teicoplanin susceptibility.A series of in-frame deletions was constructed within pAW17-tcaA to identify the regions of TcaA needed to complement teicoplanin susceptibility in BB1372 (Fig. 4). The plasmid containing deletion A fully complemented BB1372, indicating that the N terminus of the protein, which contained the C-X₂-C-X₁₀-C-X₂-C motif, was not involved in the teicoplanin resistance phenotype. Conversely, deletion B, which encompassed the predicted transmembrane domain, abolished complementation, indicating that the membrane location of TcaA is important for teicoplanin susceptibility. Deletions C to H, sequential C-terminal deletion mutants, showed that removal of the C terminus up until the region between deletions E and F had minimal effects on complementation, while removal of the regions between deletion E and the transmembrane domain abolished complementation. Therefore, only the approximately 130-aa region indicated in Fig. 4A is needed to restore teicoplanin susceptibility in the tcaA mutant. Correspondingly, this essential region contained both of the previously described tcaA mutations in S. aureus clinical GISA isolates SA137/9G and MI (25).

• Open in new tab
• Download powerpoint

FIG. 4.

Determination of the TcaA region involved in glycopeptide resistance. (A) Maps showing the positions of tcaA deletions A to H (numbers on the right indicate the exact nucleotide positions of the deletions, according to the genome annotation of strain COL). The region of approximately 130 aa required for complementation of teicoplanin susceptibility in BB1372 is shown (*), and the positions of previously described tcaA mutations in clinical isolates SA137/9G and MI are indicated (25). (B) Teicoplanin gradient (0 to 5 μg/ml) plate comparing the resistance levels of BB1372 containing the intact pAW17-tcaA plasmid (BB1539), the empty pAW17 plasmid (BB1541), and pAW17-tcaA deletion plasmids A to H. MIC*, MIC values were determined by Etest by using 2.0 McFarland suspensions on BHI supplemented with kanamycin at 50 μg/ml.

Effect of TcaA on virulence in C. elegans.The nematode C. elegans can serve as a simple surrogate model host for the study of S. aureus infection (39). Recently, Bae et al. (2) listed tcaA and tcaB among 71 bursa aurealis transposon mutants of S. aureus strain Newman that were found to attenuate C. elegans killing. However, due to the complex transcriptional organization of the tcaRAB operon in the absence of antibiotic induction (25), we wanted to determine the effect of a precise tcaA deletion on virulence in C. elegans. The vector pKOR1 was used to create an in-frame markerless deletion of the entire TcaA-coding sequence from the genome of S. aureus COL. The resulting tcaA deletion mutant, NM278, had a teicoplanin resistance level comparable to that of tcaRAB deletion mutant BB1372 (Fig. 5A). Teicoplanin susceptibility was restored by complementation with pAW17-tcaA (Fig. 5A). The markerless tcaA deletion in NM278 did not appear to influence the expression of the upstream tcaR gene or the downstream tcaB gene, as there were no detectable changes in the transcription of either gene in strain NM278 compared to that in strain COL under induced or uninduced conditions (data not shown). This indicated that the markerless tcaA deletion in NM278 was unlikely to exert polar effects on the expression of tcaR or tcaB. The relative levels of virulence of COL, NM278, and BB1372 were then examined in the C. elegans infection model (39). As shown in Fig. 5B, C. elegans killing was highly attenuated in both the in-frame tcaA deletion mutant NM278 and the tcaRAB locus mutant BB1372 compared to that in wild-type strain COL.

• Open in new tab
• Download powerpoint

FIG. 5.

Survival of C. elegans on tcaA deletion mutants. (A) Teicoplanin gradient plate comparing the resistance levels of wild-type strain COL containing the empty plasmid pAW17, BB1372 containing pAW17 (BB1541) and the complementing plasmid pAW17-tcaA (BB1539), the in-frame, markerless tcaA deletion mutant NM278 containing pAW17, and NM278 complemented with pAW17-tcaA. MIC*, MIC values from Etests performed on BHI with 2.0 McFarland inocula are shown on the right of the gradient plate. (B) Survival of nematodes fed S. aureus COL (wild type; squares; n = 90), NM278 (COL ΔtcaA; circles; n = 90), and BB1372 (COL tcaRAB::ermB; triangles; n = 90). P was <0.0001 by pairwise comparisons by the log-rank test of the following strain pairs: COL versus NM278 and COL versus BB1372. Data are representative of one of four independent experiments.