Specificity of Induction of the vanA and vanB Operons in Vancomycin-Resistant Enterococci by Telavancin

Abstract

Telavancin is a bactericidal, semisynthetic lipoglycopeptide indicated in the United States for the treatment of complicated skin and skin structure infections caused by susceptible Gram-positive bacteria and is under investigation as a once-daily treatment for nosocomial pneumonia. The related vanA and vanB gene clusters mediate acquired resistance to glycopeptides in enterococci by remodeling the dipeptide termini of peptidoglycan precursors from d-alanyl-d-alanine (d-Ala-d-Ala) to d-alanyl-d-lactate (d-Ala-d-Lac). In this study, we assessed the ability of telavancin to induce the expression of van genes in VanA- and VanB-type strains of vancomycin-resistant enterococci. Vancomycin, teicoplanin, and telavancin efficiently induced VanX activity in VanA-type strains, while VanX activity in VanB-type isolates was inducible by vancomycin but not by teicoplanin or telavancin. In VanA-type strains treated with vancomycin or telavancin, high levels of d-Ala-d-Lac-containing pentadepsipeptide were measured, while d-Ala-d-Ala pentapeptide was present at very low levels or not detected at all. In VanB-type strains, vancomycin but not telavancin induced high levels of pentadepsipeptide, while pentapeptide was not detected. Although vancomycin, teicoplanin, and telavancin induced similar levels of VanX activity in VanA-type strains, these organisms were more sensitive to telavancin, which displayed MIC values that were 32- and 128-fold lower than those of vancomycin and teicoplanin, respectively.

The glycopeptide antibiotics vancomycin and teicoplanin act by binding to the terminal d-alanyl-d-alanine (d-Ala-d-Ala) dipeptide of peptidoglycan precursors, preventing their incorporation into the bacterial cell wall (6, 11). Acquired resistance to these glycopeptides in enterococci results from the inducible production of peptidoglycan precursors terminating in the depsipeptide d-alanyl-d-lactate (d-Ala-d-Lac) and from the elimination of precursors that end with the d-Ala-d-Ala dipeptide. The substitution of d-Ala-d-Lac for the d-Ala-d-Ala dipeptide in susceptible bacteria results in a 1,000-fold-lower affinity for vancomycin (7). The two related gene clusters vanA and vanB mediate resistance by this mechanism and are the most frequently encountered genotypes among clinical vancomycin-resistant enterococci (VRE). The vanA genes confer high-level resistance to vancomycin and teicoplanin (4), while enterococci harboring vanB-type genes have a moderate level of resistance to vancomycin but remain susceptible to teicoplanin (20).

The vanA operon encodes a dehydrogenase, VanH, which reduces pyruvate to d-Lac (7); a ligase, VanA, required for the synthesis of d-Ala-d-Lac (7); and a d,d-dipeptidase, VanX, that hydrolyzes d-Ala-d-Ala, preventing its incorporation into cell wall precursor molecules (21) (Fig. 1). The expression of the vanHAX operon is regulated at the transcriptional level by the two-component regulatory system VanR-VanS in response to the presence of vancomycin or teicoplanin (3). VanS is a membrane-bound histidine kinase, and VanR is a cytoplasmic response regulator that acts as a transcriptional activator. The vanB gene cluster encodes three enzymes (VanH_B, VanB, and VanX_B) that are functionally equivalent to VanH, VanA, and VanX and display 67 to 71% amino acid identity to the vanA gene products (15). The VanR_B-VanS_B two-component regulatory system mediates the expression of the vanH_BBX_B operon in the presence of vancomycin but not teicoplanin (15). VanR_B and VanR are 34% identical, whereas VanS_B and VanS possess 23% sequence identity, with unrelated amino-terminal sensing domains (15).

FIG. 1.

Structure of the vanRSHAX gene cluster and biochemical pathway leading to the replacement of d-Ala-d-Ala with d-Ala-d-Lac at the terminus of the stem pentapeptide of the peptidoglycan precursor.

The vanA operon is transcribed as a single polycistronic message with the d,d-dipeptidase activity of VanX encoded at the 3′ end of the message (3). An endpoint assay of VanX d,d-dipeptidase activity coupled to d-amino acid oxidase and peroxidase has been utilized as a sensitive measure of the induction of the vancomycin resistance operon in Enterococcus faecalis (2, 5). More recently, a continuous assay using the alternative peptide substrate d-alanyl-(R)-phenylthioglycine has been developed to assay VanX activity (1).

Telavancin is a rapidly bactericidal, semisynthetic lipoglycopeptide with a broad Gram-positive spectrum of activity that includes methicillin-resistant Staphylococcus aureus (13, 14). Although telavancin is active against VRE, MICs for VanA-type isolates are higher than those for VanB-type strains (17). While there is no proposed breakpoint for telavancin for VRE, the FDA-approved breakpoint for vancomycin-susceptible E. faecalis is 1 μg/ml, and most VanB-type VRE are inhibited by telavancin at or below this concentration (Table 1) (19-21). In this report, the ability of telavancin to induce the expression of vancomycin resistance operons in VanA- and VanB-type strains of VRE was investigated and compared to that of vancomycin and teicoplanin by use of a VanX activity assay, determining the relative levels of the VanX protein by Western blotting and quantifying the levels of soluble cytoplasmic peptidoglycan precursors.

TABLE 1.

Characteristics and origin of Enterococcus strains used in this study^c

Species	Strain	Phenotype	MIC (μg/ml)			Reference or source
			VAN	TEI	TLV
E. faecalisa	BM4110	Vans	2	0.5	1	10
E. faecalisb	BM4110/pIP816-1	VanA	1,024	256	8	2
Enterococcus faecium	ATCC 51559	VanA	1,024	256	8	ATCC
E. faecalis	ATCC 51299	VanB	16	1	1	ATCC
E. faecalis	ATCC 51575	VanB	256	0.5	0.5	ATCC

^aStreptomycin-resistant derivative of JH2.

^bPlasmid pIP816-1 carrying Tn1546.

^cVAN, vancomycin; TEI, teicoplanin; TLV, telavancin; Van^s, vancomycin susceptible; ATCC, American Type Culture Collection.

MATERIALS AND METHODS

Reagents.

Telavancin was synthesized by Theravance, Inc. (South San Francisco, CA). Teicoplanin was provided by P. Tulkens from the Université Catholique de Louvain, Brussels, Belgium. Vancomycin, mutanolysin, bovine serum albumin (BSA), sodium phosphate, and Tris-5,5′-dithiobis(2′-nitrobenzoic acid) (DTNB) were purchased from Sigma (St. Louis, MO). Tryptic soy agar (TSA) containing 5% sheep's blood and brain heart infusion (BHI) broth were obtained from Hardy Diagnostics (Santa Maria, CA) and from Difco, respectively. Bradford protein assay reagent and 4 to 12% Criterion polyacrylamide gels were obtained from Bio-Rad (Hercules, CA). The VanX substrate d-alanyl-(R)-phenylthioglycine was synthesized by Anaspec (San Jose, CA), while the VanX inhibitor N-[(1-aminoethyl)hydroxyphosphinyl]-d-alanine was obtained from Acme Bioscience (Belmont, CA). A polyclonal antibody to the VanX protein was produced and affinity purified by Covance Research Products (Denver, PA). The IRDye 800CW goat anti-rabbit secondary antibody was obtained from Li-Cor Biosciences (Lincoln, NE).

Strains and growth conditions.

The bacterial strains used in this work are described in Table 1. Strains were grown on TSA containing 5% sheep's blood at 37°C. MICs were determined by broth microdilution according to CLSI-approved methods (9).

VanX activity assay.

Strains were grown overnight on TSA containing 5% sheep's blood at 37°C and suspended in BHI broth to an A₆₂₅ of 0.5. BHI broth (5 ml) containing vancomycin, teicoplanin, or telavancin at various concentrations was inoculated with 100 μl of the cell suspension, and the cultures were incubated at 37°C on a revolving drum until the A₆₂₅ reached 0.7. Duplicate cultures for each concentration of antibiotic were prepared. Bacteria were collected by centrifugation (5,000 × g at 4°C for 5 min), washed in 1.0 ml of ice-cold 100 mM sodium phosphate (pH 7.0), and then stored overnight at −80°C. The frozen cells were thawed, resuspended in 0.1 ml of 20 mM MES (morpholineethanesulfonic acid) (pH 6.0) containing 1,000 U of mutanolysin, and then incubated at 37°C for 60 min, when 0.9 ml of 100 mM Tris (pH 7.5) was added. Cells were then lysed by sonication, and the insoluble material was removed by centrifugation (200,000 × g at 4°C for 10 min). The protein concentration of the supernatant was measured by using the Bio-Rad protein assay with BSA as a standard.

To measure the d,d-dipeptidase activity of VanX, a continuous assay was employed by using the synthetic substrate d-alanyl-(R)-phenylthioglycine (1). The assay was performed with 100 mM Tris (pH 7.5) containing 2.5 mM DTNB and 5 mM d-alanyl-(R)-phenylthioglycine. The reaction was initiated by the addition of soluble cell extract to the mixture and was monitored continuously at 412 nm by using an S-Max plate reader (Molecular Devices, Sunnyvale, CA) at 22°C for 30 min. The concentration of 5-thio-2-nitrobenzoic acid (TNB) liberated was calculated by using the following extinction coefficient: ɛ₄₁₂ = 13,600 M⁻¹ cm⁻¹. Specific activity was defined as the number of nmoles of TNB liberated per min per mg of protein in the extract.

VanX inhibition assay.

Assays to measure the inhibition of VanX activity by the d-Ala-d-Ala phosphonamidate analog N-[(1-aminoethyl)hydroxyphosphinyl]-d-alanine (22) were carried out as described above, except that d-alanyl-(R)-phenylthioglycine concentrations were varied between 0.125 mM and 0.75 mM and the inhibitor was added over a concentration range of 25 μM to 324 μM. The K_i value was determined by fitting initial rate data to a competitive model of inhibition by using GraphPad (San Diego, CA) Prism 5 software.

Western blotting.

A polyclonal rabbit antibody was raised against an 18-mer peptide from the VanX protein with the sequence CNEAQNRRRLRSIMENSG, designed through an inspection of the crystal structure (8). The antibody was affinity purified before use. Soluble cell extract containing 35 ng of protein was loaded onto a 4 to 12% polyacrylamide gel, and separation by electrophoresis followed by transfer onto a polyvinylidene difluoride (PVDF) membrane was performed by using standard methods. The membrane was probed by using a 1/5,000 dilution of the primary antibody. The secondary antibody was goat anti-rabbit IgG conjugated to the infrared fluorescent dye IRDye 800CW and was used at a 1/10,000 dilution. Fluorescence detection was carried out with a Li-Cor Odyssey system according to the manufacturer's instructions (Li-Cor Biosciences, Lincoln, NE).

Analysis of peptidoglycan precursors.

Extraction and analysis of peptidoglycan precursors were performed as previously described (19). Enterococci were grown in BHI broth at 37°C in the absence or presence of subinhibitory concentrations of vancomycin or telavancin to mid-exponential phase (A₆₀₀ = 1.0). Ramoplanin (3 μg/ml) was added to inhibit peptidoglycan synthesis, and incubation was continued for 15 min. Bacteria were harvested, and the cytoplasmic precursors were extracted with 8% trichloroacetic acid (15 min at 4°C), desalted, and analyzed by high-performance liquid chromatography (HPLC). Results were expressed as the percentages of total late peptidoglycan precursors represented by UDP-MurNAc-tetrapeptide, UDP-MurNAc-pentapeptide, and UDP-MurNAc-pentadepsipeptide that were determined from the integrated peak areas.

RESULTS AND DISCUSSION

VanX inhibition assay.

The hydrolysis of d-alanyl-(R)-phenylthioglycine by cell extracts was monitored continuously, and representative progress curves of the reaction are shown in Fig. 2. The d,d-dipeptidase activity present in the extracts could be inhibited by the d-Ala-d-Ala phosphonamidate analog N-[(1-aminoethyl)hydroxyphosphinyl]-d-alanine. The fitting of the initial rate data to a competitive model of inhibition yielded a K_i value of 62 ± 3 μM, which is in close agreement with the value of 36 μM previously determined with purified VanX enzyme and fitted to the same model (22). A double reciprocal plot of the inhibition data illustrates the competitive nature of the inhibition and is shown in Fig. 3. The turnover of d-alanyl-(R)-phenylthioglycine by the cell-free but otherwise unfractionated extract and the inhibition of this activity by a VanX-directed inhibitor strongly suggest that the observed d,d-dipeptidase activity can be attributed to VanX.

FIG. 2.

Progress curves for the background and VanX-catalyzed reactions using d-alanyl-(R)-phenylthioglycine as a substrate and ATCC 51559 cell extract as the VanX source. No product accumulation occurred in the absence of cell extract, while a basal level of VanX activity was observed in the extract of the VanA-type strain cultured without antibiotic, which increased when the strain was cultured in the presence of telavancin: background (no bacterial extract) (…), 0 μg/ml telavancin (- - - - ), or 3 μg/ml telavancin (—).

FIG. 3.

Double reciprocal plot illustrating competitive inhibition of VanX by the phosphonamidate N-[(1-aminoethyl)hydroxyphosphinyl]-d-alanine (inset). The inhibitor was present at 0 μM (□), 25 μM (⧫), 42 μM (▾), 69 μM (▴), 127 μM (▪), and 325 μM (•). The data were fit to a competitive inhibition model, giving a best-fit K_i value of 62 ± 3 μM.

Expression of glycopeptide resistance mediated by vanA and vanB.

Vancomycin-susceptible strain BM4110 does not harbor vancomycin resistance genes (Table 1) and, as expected, did not display VanX activity above the assay baseline for each of the antibiotics tested (Fig. 4). Subinhibitory concentrations of vancomycin induced the synthesis of d,d-dipeptidase activity in both VanA- and VanB-type strains (Fig. 4A). In the two VanA-type strains, a basal level of VanX activity was observed in the absence of vancomycin, consistent with data from a previous report (2). The activity increased upon the addition of the glycopeptide in a concentration-dependent manner (Fig. 4A). d,d-Dipeptidase activity also increased with the vancomycin concentration in the VanB-type strains, but no basal level of activity was detected in the absence of vancomycin. The level of VanX activity in these strains was below that of the two VanA-type strains at every concentration tested when expressed as a multiple of the MIC and was reflected in lower MIC values than those for the VanA-type strains. Teicoplanin induced high levels of d,d-dipeptidase activity in the VanA-type strains (Fig. 4B) but did not induce activity in the VanB-type strains at concentrations up to 0.6-fold the MIC values. Cultures of the VanA-type strains in the presence of telavancin induced levels of VanX activity similar to those observed for vancomycin and teicoplanin, while the activity was not detected in VanB-type strains at concentrations up to 0.3-fold the MIC (Fig. 4C). Previous reports also demonstrated that teicoplanin does not induce vanB gene expression (2, 5). The structural features of vancomycin and teicoplanin that are required to prompt or circumvent the induction of resistance in an E. faecalis VanB-type strain were described previously (12). The addition of the vancosamine disaccharide to the teicoplanin aglycone caused a loss of activity toward the VanB-type strain compared to teicoplanin, indicating the induction of resistance. Conversely, replacing the vancosamine disaccharide of vancomycin with the lipidated monosaccharide of teicoplanin increased the activity of vancomycin toward the VanB-type strain, indicating a loss of vanB induction. It was observed that while the vancomycin and teicoplanin aglycones are the minimal structural features required to induce resistance, their ability to do so can be blocked by the addition of a lipid-substituted carbohydrate. The presence of the decylaminoethyl lipid tail linked to the nitrogen of the vancosamine sugar on telavancin and the lack of induction of vanB genes are consistent with this observation. The mechanism by which lipid-containing carbohydrate substituents circumvent vanB induction remains to be elaborated, but one possibility is that glycopeptides are less accessible to the sensor kinase when localized near the bacterial membrane (15). The concentration-response of the d,d-dipeptidase activity from VanA-type strains to each glycopeptide antibiotic followed the same pattern, where VanX activity increased before reaching a plateau at concentrations close to the MIC. At these high concentrations, the capacity of the bacterial cell to produce VanX appears to approach a limit. Despite remodeling of the terminus of peptidoglycan cell wall precursors from d-Ala-d-Ala to d-Ala-d-Lac, the growth of enterococci was inhibited by glycopeptide when it was supplied at a sufficiently high concentration. Teicoplanin and telavancin appeared to be at least as efficient as vancomycin in inducing VanX activity in VanA-type strains, since similar levels of VanX activity were observed at lower absolute concentrations of these antibiotics. Although telavancin efficiently induced vanHAX expression, it also had the greatest antibacterial activity. The MICs of telavancin were 32- and 128-fold lower than those of teicoplanin and vancomycin, respectively.

FIG. 4.

Specific activity of VanX or VanX_B in bacterial culture extracts grown in the presence of various concentrations of glycopeptide antibiotics. (A) Vancomycin induces dipeptidase activity in both VanA- and VanB-type strains of enterococci. (B) Teicoplanin induces dipeptidase activity in VanA- but not in VanB-type enterococci. (C) Telavancin induces dipeptidase activity in VanA- but not in VanB-type strains, which remain susceptible to telavancin and teicoplanin. ▴, ATCC 51559 (VanA); ▪, BM4110/pIP816-1 (VanA); ⧫, ATCC 51575 (VanB); ▾, ATCC 51299 (VanB); •, BM4110 (Van^s).

VanX protein.

The VanX protein present in cell extracts of E. faecalis VanA-type strain BM4110/pIP816-1 was detected by Western blotting (Fig. 5 A). A protein with an M_r of between 22,000 and 36,000 was detected in the soluble cell extract when probed with the antibody. This protein migrated close to the predicted VanX M_r of 23,380 (Fig. 5A). The protein was absent from the extract of vancomycin-susceptible, isogenic strain BM4110, which does not carry the vancomycin resistance gene encoding VanX (Table 1 and Fig. 5). The VanX protein was present at a basal level in the absence of antibiotic in the VanA-type strain, which correlates with the basal level of dipeptidase activity observed for these strains (Fig. 4). The amount of VanX in the extracts increased with increasing concentrations of telavancin in the culture medium (Fig. 5B). A 4-fold increase in VanX over the basal level was observed at 0.4-fold the MIC. This value is comparable to the 6-fold increase observed for the d,d-dipeptidase activity for this strain at 0.4-fold the MIC of telavancin.

FIG. 5.

(A) Western blot of bacterial culture extracts grown with various concentrations of telavancin. BM4110/pIP816-1 is a VanA-type strain, while BM4110 is the vancomycin-susceptible isogenic strain that does not harbor the gene encoding VanX. (B) Quantification of the VanX protein detected by Western blotting. No VanX protein was detected in vancomycin-susceptible strain BM4110, while the amount of VanX detected in VanA-type strain BM4110/pIP816-1 was dependent on the concentration of telavancin present in the bacterial culture.

Cytoplasmic peptidoglycan precursors.

The profile of soluble peptidoglycan precursors in strain BM4110 was the same under uninduced, vancomycin, and telavancin conditions where the d-Ala-d-Lac-containing pentadepsipeptide was not detected (Table 2). Under uninduced conditions, the two VanA-type strains had similar distributions of precursors, where most of this pool was present as a tetrapeptide or d-Ala-d-Ala-containing pentapeptide. The pentadepsipeptide constituted 18% of precursors in uninduced strain BM4110/pIP816, consistent with the basal level of VanX activity and protein expression described above. However, the pentadepsipeptide was not detected in uninduced ATCC 51559 cultures. Culturing of VanA-type strains in the presence of subinhibitory concentrations of vancomycin or telavancin resulted in the production of high levels of pentadepsipeptide with a concomitant loss of pentapeptide to low or undetectable levels. The pentadepsipeptide was not detected in the two VanB-type strains in the absence of vancomycin and telavancin, while cultures of both strains with vancomycin induced high levels of pentadepsipeptide production and a reduction of pentapeptide levels. The d-Ala-d-Lac-containing precursor was not detected in the two telavancin-treated VanB-type cultures, indicating that telavancin did not induce the switch from pentapeptide to pentadepsipeptide production, in contrast to vancomycin.

TABLE 2.

Cytoplasmic peptidoglycan precursors in Enterococcus strains^a

Strain and growth condition	% peptidoglycan precursors
	UDP-MurNAc-tetrapeptide	UDP-MurNAc-pentapeptide (d-Ala)	UDP-MurNAc-pentadepsipeptide (d-Lac)
BM4110 (Vans)
Uninduced	ND	100	ND
VAN (1 μg/ml)	ND	100	ND
TLV (0.1 μg/ml)	ND	100	ND
BM4110/pIP816 (VanA)
Uninduced	24	58	18
VAN (64 μg/ml)	27	ND	73
TLV (4 μg/ml)	26	ND	74
ATCC 51559 (VanA)
Uninduced	29	71	ND
VAN (64 μg/ml)	41	ND	59
TLV (4 μg/ml)	40	8	52
ATCC 51299 (VanB)
Uninduced	ND	100	ND
VAN (3 μg/ml)	6	19	75
TLV (0.1 μg/ml)	ND	100	ND
ATCC 51575 (VanB)
Uninduced	ND	100	ND
VAN (16 μg/ml)	4	10	86
TLV (0.1 μg/ml)	ND	100	ND

^aND, not detectable; VAN, vancomycin; TLV, telavancin; Van^s, vancomycin susceptible.

In summary, telavancin induced the expression of Van resistance genes in VanA- but not in VanB-type strains of enterococci. Telavancin antibacterial activity was retained at concentrations at least 32-fold lower than those of vancomycin and teicoplanin in the VanA-type strains studied, while the VanB-type VRE were inhibited by telavancin at or below the vancomycin-susceptible E. faecalis breakpoint of 1 μg/ml. The greater antibacterial activity of telavancin is attributed to its multivalent interaction with the membrane-bound peptidoglycan precursor lipid II and its disruptive effects on the barrier function of the bacterial membrane (16, 18).