In vivo characterization of the type A and B vancomycin-resistant enterococci (VRE) VanRS two-component systems in Escherichia coli: A nonpathogenic model for studying the VRE signal transduction pathways

Abstract

Escherichia coli reporter strains modeling the high-level type A and B vancomycin resistances of Enterococcus faecium BM4147 and Ent. faecalis have been developed to study the respective VanR-VanS two-component regulatory systems. P_vanH-, P_vanRa-, P_vanY-, and P_vanRb-lacZ fusions report on expression from the vancomycin-resistant enterococci promoters of the type A vanRSHAXYZ and type B vanRSYWHBX gene clusters. These strains also express from single-copy chromosomal genes vanR_a, vanR_b, or vanRS_b behind their respective promoter (P_vanRa or P_vanRb) or vanS_a or vanS_b behind the rhamnose-inducible P_rhaB. Results show that activation (phosphorylation) of the response regulator VanR_a by its sensor kinase VanS_a leads to transcriptional activation of both P_vanH and P_vanRa. Additionally, VanR_b activates its cognate promoters P_vanY and P_vanRb, although this occurs only in the absence of VanS_b and presumably is caused by VanR_b phosphorylation by an unidentified endogenous E. coli kinase. Thus, VanS_b interferes with activation of VanR_b, probably by acting as a phospho-VanR_b phosphatase. Although both VanR_a and VanR_b activate their cognate promoters, neither activates the heterologous P_vanR, P_vanH, or P_vanY, arguing against the interchangeability of type A and B two-component regulatory switches in vancomycin-resistant enterococci. VanR_a also is activated by the nonpartner kinase PhoR. Because this occurs in the absence of its inducing signal (P_i limitation), PhoR autophosphorylation apparently is regulated in vivo. Furthermore, the activation of VanR_a caused by cross talk from PhoR in the absence of a signal allows distinction of cross talk from crossregulation as the latter, but not the former, responds to environmental cues.

Vancomycin-resistant enterococci (VRE) have become clinically problematic human pathogens with high mortality and incidence increasing alarmingly over 35-fold from 0.4% to 14% in hospital intensive-care units from 1989 to 1993 (1). Clinical isolates of VRE have been divided into three types based on the level of resistance (type A and B, high-level resistance; VanC, low- or moderate-level resistance), inducible versus constitutive resistance (type A and B, inducible; VanC, constitutive), and inducibility by both vancomycin and teicoplanin (type A) or inducibility by vancomycin only (type B; refs. 2–4). At the molecular level, types A and B resistance result from the same substitution of an ester linkage (d-Ala-d-lactate) for an amide linkage (d-Ala-d-Ala) at the termini of peptidoglycan chains involved in cell wall synthesis; VanC resistance instead involves a switch to d-Ala-d-Ser termini (5, 6).

Type A and B strains have six genes in common, ones encoding a two-component sensor kinase-response regulator system (VanS_a, VanR_a and VanS_b, VanR_b), three enzymes (VanH, VanA/VanB, and VanX) that bring about the switch to d-Ala-d-lactate peptidoglycan chains with a 1,000-fold lower affinity for vancomycin, and the d,d-carboxypeptidase (VanY). In type B VRE, vanY is upstream of vanH, whereas in type A the corresponding vanY comes after vanX. The consequence of this reordering is that while the promoters controlling the two-component regulatory system genes are P_vanRa and P_vanRb, respectively, the promoters for the structural genes are instead P_vanH and P_vanY (Fig. 1). Most previous genetic and enzymatic studies trying to define the molecular logic of how these proteins function in VRE have focused on the type A system. Courvalin and colleagues (7) also have demonstrated that mutations of type B VanS_b can result in teicoplanin-inducible resistance.

Figure 1.

Organization of type A and type B VRE gene clusters.

To characterize the two-component VRE signal transducing proteins, we previously have studied VanS_a and VanR_a in vitro and in vivo by using Escherichia coli (8, 9). We now report studies on the in vivo function and selectivity of VanS_a and VanR_a along with similar ones of VanS_b and VanR_b on expression from the type A and B promoters (Fig. 1). We examined VanR_a and VanR_b for transcriptional activation of its cognate and noncognate promoters. We tested the requirements of VanS_a and VanS_b for activation (phosphorylation) of VanR_a and VanR_b. Additionally, we tested for cross talk between these heterologous VRE signaling proteins, comparing interactions between them to those with the response regulator PhoB and the sensor kinase PhoR of the E. coli Pho regulon (10) as a basis for distinguishing specific and nonspecific (cross talk) interactions. Our results indicate that, unlike the VanH, VanA/VanB and VanX enzymes required for vancomycin resistance, the type A and type B two-component regulatory systems evolved independently.

MATERIALS AND METHODS

Media, Chemicals, and Other Reagents.

Luria–Bertani, 3-(N-morpholino)propanesulfonic acid (Mops), and tryptone-yeast extract media are described elsewhere (11). Ampicillin was added at 100 μg per ml; kanamycin, at 12.5 or 50 μg per ml, gentamicin, at 5 or 15 μg per ml, and chloramphenicol, at 5 and 20 μg per ml to select antibiotic-resistant integrants or maintain plasmids, respectively. l-rhamnose was used at 1.1 mM for induction. After integration, the integrants were grown without an antibiotic. Pfu DNA polymerase (Stratagene) was used to generate DNAs for cloning. All PCR-amplified fragments were cloned and sequenced as double-stranded plasmid DNAs in a core facility.

Bacteria.

All bacteria are derivatives of E. coli K-12 BW13711 (11). Conditional replication (oriR_R6Kγ) plasmids were propagated at low or high copy number in the pir⁺ host BW23473, its isogenic pir-116 host BW23474, or similar hosts (12, 13). BW24381 (lacI^qrrnB_T14 ΔlacZ_WJ16 ΔphoBR580 ΔcreABCD154 rpoS (Am) ΔaraBAD_AH33 ΔrhaBAD_LD78 Δ(ackA pta)160) is a derivative of BW23660 (9) into which the ΔrhaBAD_LD78 (14) and Δ(ackA pta)160 (9) mutations were introduced by P1kc transduction essentially as described elsewhere (15). Its derivative BW25124 was made by recombining the P_vanH-lacZ fusion in pJS110 onto the chromosome via allele replacement (13) and then introducing the ΔphoR574 mutation (9) by P1kc transduction. pJS110 was made by cloning the PstI–BsiWI, P_vanH-lacZ DNA fragment from pJS79 into the allele-replacement lacZ transcriptional fusion vector pWJ18 (W. Jiang, A.H., and B.L.W., unpublished results).

Construction of lacZ Transcriptional Fusions.

The P_vanH and P_vanRa promoters were PCR-amplified by using pAT87 (16) as template and the primer JS-P27 (GCAGTCGAC-CGGAAAGCAATGATAACTATACGACG) with JS-P29 (GCAGGATCC-TCTGAAGAACGAAAACGGCTCGTTC) and JS-P31 (GCAGTCGAC-ATGTATCTAGGGCTTCATTATACAGG) with JS-P32 (GCAGGATCC-CTTAATAATTTATCAGATTATAGGGCCG), respectively (extensions containing a restriction site precede hyphens). The 240-bp P_vanH and 255-bp P_vanRa PCR products were cloned as BamHI–SalI fragments into pAH125 to make pJS79 and pJS81, respectively. The P_vanY and P_vanRb regions were amplified by using genomic DNA of Ent. faecalis V583 (from Daniel Sahm, Jewish Hospital, St Louis, MO; ref. 17) as template and the primer JS-P21 (GCAGTCGAC-TCACAGATATTCCAGCCGGACAAATTGTCC) with JS-P26 (GCAGGATCC-TTTGCAATAAAACTACGATTTGTGGC) and JS-P15 (GCAGTCGAC-TTAAACGGTATATTTCGGAAGAAC) with JS-P19 (GCAGGATCC-ATTTAAGAAGATAACATAACAGTCTG), respectively. The 554-bp P_vanY and 182-bp P_vanRb PCR products were similarly cloned into pAH125 to make pJS52 and pJS51, respectively. To integrate these plasmids at attP21, a fragment encoding chloramphenicol resistance of pCAH54 (18) and attP_P21 of pAH95 (A.H. and B.L.W., unpublished results) was introduced by exchanging the NheI–SphI fragments with the corresponding one in pIADL46 (from Ivan Lessard, Harvard Medical School, Boston, MA) to make pJS94, pJS96, pJS97, and pJS98, respectively. pAH125 is a derivative of pSK49 (12) for making lacZ transcriptional fusions (S.-K. Kim, A.H., and B.L.W., unpublished results).

Plasmids.

pAH85, pAH66, and pAH67 (12) are derivatives of pSK49 that synthesize PhoB^wt, PhoB_{M17V, E87D}, and PhoB_T97A, respectively, under P_phoB control. pAH151 synthesizes PhoR under P_rhaB control (14) in a pSK49 derivative with attP_HK022 and encoding gentamicin resistance of pAH70 and pAH143 (A.H. and B.L.W., unpublished results). pMP1 is a derivative of pSK49 carrying the same P_rhaB-phoR⁺ fusion. pJS87 synthesizes full-length VanS_a with a C-terminal His₆ tag (vanS_H6) under P_rhaB control. pMP3 is a derivative of pSK49 carrying the same P_rhaB-vanS_H6 fusion. The vanS_H6 fragment was constructed by PCR using pSLF40 (9) and the primers JS-P3 (GAATTC-TGTACGGAGTCAAGCCATATGTTG) and JS-P7 (A ATAGTCGAC-TTATTAgtggtggtggtggtggtgGGACCTCCTTTTATCAACCAAGTC) according to conditions above (the His₆ coding region is in lowercase). The 1,216-bp vanS_H6 product was cloned into pUC19 as an EcoRI–SalI fragment to make pJS60 for sequencing and then into pAH151 as an NdeI–SalI fragment to make pJS87. pJS126 is a derivative of pSLF55 with attP_P21 and encoding chloramphenicol resistance pSLF55 synthesizes VanR_a from its native promoter (9). pMP2 is a derivative of pSK49 with the P_rhaB-vanS_b fusion. pJS107 and pJS109 are similar plasmids that express vanR_b and vanRS_b, respectively, behind P_vanRb. The vanR_b and vanS_b fragments were generated by using Ent. faecalis V583 genomic DNA and the primer JS-P15 (GCAGTCGAC-TTAAACGGTATATTTCGGAAGAAC) with JS-P16 (CGCTCTAGA-ATAAGACACAAATTGCTGTGC) and JS-P17 (GCACTGCAG-CATATGACCATGGCGGGTGTAGGTTACCGATTGG) with JS-P18 (TTCTCTAGA-TTGTTTCATATGCCGTTTGTG), respectively. The 964-bp vanR_b and 1,641-bp vanS_b PCR products were cloned with SalI and XbaI or PstI and XbaI into pAH143, to make pJS45 and pJS46, respectively. The P_vanRb-vanR_b fragment was subcloned as an SphI–BamHI fragment from pJS45 into pAH85 to make pJS107. The SacII–XbaI vanS_b fragment was subcloned from pJS46 into pJS45 to make the P_vanRb-vanRS_b plasmid pJS47. The P_vanRb-vanRS_b fragment then was subcloned as an SphI–BamHI fragment from pJS47 into pAH85 to make pJS109.

attP Plasmid Integration and PCR Testing of Copy Number.

Plasmids were integrated into the chromosome of cells expressing the respective integrase and verified by PCR to have a single integrated plasmid as described elsewhere (9, 12, 18).

Cell Growth and Enzyme Assay.

Cells were grown and assayed for bacterial alkaline phosphatase and β-galactosidase (BG) as described (11). Units are nanomoles of product made per min per cell culture at OD₄₂₀ at 37°C and 28°C, respectively.

Purification of VanS_b and Protein Phosphorylation Experiments.

A maltose binding protein (MBP) fusion to the cytosolic domain of VanS_b (′VanS_b, residues V146M to L447) was constructed by using pJS46 as template and the primer MKP5 (GCAGGTGTGGGATTGCTTCATATG-GGGCTGACAATTCGG) with MKP6 (GTACCCGGGGATCC-TCTAGATTGTTTCATATGC). The 1,224-bp ′vanS_b fragment was cloned with NdeI into a derivative of pMAL-c (New England Biolabs) called pIADL16 (from I. Lessard) to make pJS119. Overproduction, purification, and autophosphorylation of MBP-′VanS_b was accomplished as described for MBP-′VanS_a (19).

RESULTS

An Improved E. coli Reporter System for Studying VRE Two-Component Systems.

lacZ transcriptional fusions to P_vanH and P_vanRa were made to improve on the signal-to-noise ratio of previously engineered strains. The basal levels of earlier P_vanH- and P_vanRa-lacZ fusions ranged from 28.5 ± 9.8 to 69.3 ± 6.4 BG units (BGU; ref. 9). These elevated levels interfered with their use in experiments directed toward probing structural determinants for VanRS interactions. Their high background probably resulted from an RNase III processing site upstream of lacZ in pWJ19 that stabilizes the mRNA and is derived from pTL61 (20). The new fusions lack this site as the lacZ junction in pAH125 is from pRS415 (21). Basal levels of the new type A fusions in the absence of VanR_a were 0.7 ± 0.1 BGU for P_vanH and 12.8 ± 0.9 BGU for P_vanRa (Table 1). The basal levels of the type B P_vanY-lacZ and P_vanRb-lacZ fusions were 7.6 ± 1.2 and 8.5 ± 0.9 BGU, respectively (Table 2).

Table 1.

Activation of P_vanH and P_vanRa by VanR_a

Regulator, kinase	lacZ fusion*	β-Galactosidase Sp Act
		−Rha	+Rha
None	PvanH	0.7  ± 0.1	N.D.
VanRa	PvanH	0.9  ± 0.3	N.D.
VanRa, PrhaB-vanSa	PvanH	0.8  ± 0.1	31.2 ± 4.1
None	PvanRa	12.8  ± 0.9	N.D.
VanRa	PvanRa	14.7  ± 1.7	N.D.
VanRa, PrhaB-vanSa	PvanRa	10.4  ± 1.7	46.7 ± 4.2

Cells were assayed after 16–20 hr growth in 0.10% glycerol-Mops-2 mM P_i medium without (−Rha) or with (+Rha) rhamnose. Specific activity (Sp Act) units are nanomoles of product formed per cell optical density at 420 nm. Means of three or more determinations with SD are given. N.D., not determined. 

^*Strains are integrants of BW24381 with single copies of the plasmids in parenthesis: JCS182 (pJS94), JCS184 (pJS96), JCS196 (pJS94, pSLF55), JCS207 (pJS96, pSLF55), JCS230 (pJS94, pSLF55, pJS87), and JCS345 (pJS96, pSLF55, pJS87). 

Table 2.

Activation of P_vanRb and P_vanY by VanR_b

Regulator, kinase	lacZ fusion*	Carbon source	β-Galactosidase Sp Act
			−Rha	+Rha
None	PvanY	Fru	7.6  ±  1.2	N.D.
VanRb	PvanY	Fru	42.5  ±  3.8	N.D.
VanRSb	PvanY	Fru	2.6  ±  0.5	N.D.
None	PvanRb	Fru	8.5  ±  0.9	N.D.
VanRb	PvanRb	Fru	40.8  ±  0.5	N.D.
VanRSb	PvanRb	Fru	2.3  ±  0.4	N.D.
VanRb, PrhaB-vanSb	PvanY	Gly	59.2  ±  3.0	6.7  ±  0.8
VanRb, PrhaB-vanSb	PvanY	Fru	44.7  ±  2.6	12.5  ±  1.3
VanRb, PrhaB-vanSb	PvanY	Glu	47.9  ±  0.4	40.5  ±  2.2
VanRa	PvanY	Fru	8.7  ±  1.3	N.D.
VanRa, PrhaB-vanSa	PvanY	Fru	7.4  ±  0.8	7.6  ±  1.0
VanRa	PvanRb	Fru	11.0  ±  1.4	N.D.
VanRa, PrhaB-vanSa	PvanRb	Fru	8.4  ±  0.9	8.7  ±  0.7
VanRb	PvanH	Fru	0.9  ±  0.5	N.D.
VanRb	PvanRa	Fru	13.6  ±  3.5	N.D.

Cells were assayed after growth in 0.06% glucose (Glu)-, 0.10% glycerol (Gly)-, or 0.06% fructose (Fru)-Mops-2 mM P_i medium as in Table 1. N.D., not determined. 

^*Strains are integrants of BW24381: JCS186 (pJS97), JCS188 (pJS98), JCS197 (pJS97, pSLF55), JCS198 (pJS98, pSLF55), JCS199 (pJS94, pJS107), JCS200 (pJS96, pJS107), JCS201 (pJS97, pJS107), JCS202 (pJS98, pJS107), JCS237 (pJS97, pJS107, pJS102), JCS240 (pJS97, pJS109), JCS241 (pJS98, pJS109), JCS347 (pJS97, pSLF55, pJS87), and JCS349 (pJS98, pSLF55, pJS87). 

It is generally more meaningful to study regulatory effects when the relevant genes are in single copy. This is especially true in studies of the VRE and Pho two-component systems. Abnormal regulatory effects have been repeatedly observed when using multicopy phoB, phoR, or vanS plasmids, which presumably result from effects of plasmid copy number on synthesis of their gene products (15, 22, 23). It therefore was critical to construct reporter strains in which both the promoter fusions and regulatory genes are in single copy. The use of strains with the vanR and vanS genes stably recombined on the chromosome also afforded tighter control of these regulatory genes and circumvented problems that can arise from inconsistent gene dosages.

All reporter strains express vanR alone or together with vanS from its native promoter (P_vanRa or P_vanRb). Many express vanR from its promoter and vanS or phoR under tight control from the rhamnose-inducible promoter P_rhaB (14). These genes and fusions were recombined onto the chromosome by use of conditional replication, integration, and modular plasmids that have different phage attachment (attP) sites and antibiotic resistances (refs. 9, 12, 14, and 18; Materials and Methods; A.H. and B.L.W., unpublished results). The promoter-lacZ fusions were recombined onto the chromosome either by use of an analogous lacZ vector or one that provides homologous sequences for allele replacement at the lac locus.

Activation of P_vanH and P_vanRa by VanR_a and VanS_a.

Previous studies had shown that VanR_a and VanS_a both were required for ca. 6- to 45-fold activation of a P_vanH-lacZ fusion in E. coli (9). Although no activation of a P_vanRa-lacZ fusion was apparent in that study, the basal level was high. As shown in Table 1, the new P_vanH-lacZ fusion was activated ca. 40-fold only in the presence of both VanR_a and VanS_a and the new P_vanRa-lacZ fusion was similarly activated ca. 5-fold. Apparently, the high basal activities of earlier fusion strains interfered with detecting activation of P_vanRa by VanR_a. Activation of P_vanRa is consistent with phospho-VanR_a (P-VanR_a) binding P_vanRa with higher affinity than free VanR_a (19).

Activation of P_vanY and P_vanRb by VanR_b.

Both the P_vanY- and P_vanRb-lacZ fusions show ca. 5-fold activation by VanR_b. Unexpectedly, this activation occurs in the absence of VanS_b (Table 2). Many response regulators are phosphorylated by nonpartner histidine kinases (24). It therefore is likely that VanR_b activation results from its being phosphorylated by an unidentified E. coli kinase. Although VanR_b also might have been activated by acetyl phosphate (25), this was ruled out here because our reporter strains are genetically blocked in acetyl phosphate synthesis.

Evidence of an in Vivo VanS_b Phosphatase.

Many two-component sensor proteins act both as a kinase and as phosphatase toward their partner response regulators. Using analogous reporter strains, we previously showed that VanS_a acts as a P-VanR_a phosphatase by demonstrating that VanS_a blocked activation of VanR_a by PhoR or acetyl phosphate (9). Accordingly, VanS_b was tested for interference with the activation of VanR_b. As shown in Table 2, no activation of the P_vanY- or P_vanRb-lacZ fusion was apparent when vanR_b and vanS_b were expressed together from P_vanRb. Similar ca. 17-fold decreases were seen for both type B VRE promoters. Curiously, the expression levels of both fusions in presence of P_vanRb-vanRS_b were reduced ca. 3-fold below the basal levels in the absence of VanR_b, suggesting that free VanR_b might act as a repressor. Further evidence that VanS_b is responsible for inhibiting activation comes from experiments in which vanR_b and vanS_b were expressed independently. The P_rhaB-vanS_b strains also showed reduced activation of P_vanY with the levels varying inversely with catabolite repression conditions. Glycerol results in the lowest, fructose results in an intermediate, and glucose results in the highest catabolite repression of P_rhaB (14). In agreement, greatest inhibition occurred upon induction in glycerol medium and the lowest, in glucose medium (Table 2). Together, these data suggest that VanS_a and VanS_b are poised at different kinase-phosphatase set points; VanS_a acts as a kinase under conditions where VanS_b is a phosphatase. Finally, VanS_a and PhoR were tested for phosphatase function toward P-VanR_b. As expected, no interference by them was seen.

Promoter Specificity of the VRE Response Regulators.

Both type A and type B VRE promoters were tested for activation by their cognate and noncognate response regulators. Conditions leading to activation of P_vanH and P_vanRa by VanR_a (Table 1) were without effect on expression of P_vanY or P_vanRb (Table 2). Likewise, conditions leading to activation of P_vanY and P_vanRb by VanR_b were without effect on P_vanH and P_vanRa (Table 2).

In Vivo Competition Between VanR_a and PhoB.

Previous kinetic studies of MBP-′VanS_a with VanR_a and PhoB provided a quantitative estimate of the overall efficiency and specificity of phosphotransfer from P-VanS_a (8). By acting as a slow heterologous substrate for P-MBP-′VanS_a with a k_xfer of 0.2 min⁻¹ and K_M of 95 ± 30 μM, PhoB was a competitive inhibitor of P-VanR formation. In those studies, VanR_a had a k_xfer of ca. 60 min⁻¹ and K_M of 3 μM for phosphotransfer from P-MBP-′VanS_a. Experiments therefore were done to see whether competition occurs between VanR_a and PhoB in vivo. As shown in Fig. 2, VanS_a activated phoA and P_vanH-lacZ expression ca. 1,120- and 18-fold, respectively, when only PhoB or VanR_a was present. When both were present, VanS_a activated phoA and P_vanH-lacZ expression ca. 760- and 36-fold, respectively. Although a modest reduction of PhoB activation by VanS_a occurred in the presence of VanR_a, it is unclear why VanR_a was stimulated, as these experiments were carried out in the absence of deliberate stimulation of VanS_a.

Figure 2.

Competition between VanR and PhoB. Expression of phoA (encoding bacterial alkaline phosphatase, Bap) and a P_vanH-lacZ fusion report on PhoB and VanR_a activation, respectively. VanS_a or PhoR were synthesized under P_rhaB control in cells expressing phoB, vanR_a, or both behind its native promoter. Bap and BG were assayed as in Table 1 in medium with rhamnose and 2 mM (Excess) or 0.1 mM (Limited) P_i. Arrows on the right are in proportion to Sp Act values (nmol product made per min per OD₄₂₀). With no rhamnose, Bap and BG Sp Act values of all strains were ca. 0.1 ± 0.1 and 1.0 ± 0.2, respectively. Strains are integrants of BW25053 (ΔphoBR) or BW25124 (ΔphoR): JCS289 (pJS126, pMP1), JCS297 (pMP3), JCS304 (pJS126, pMP3), JCS338 (pMP1), and JCS342 (pMP3).

Results from similar experiments with PhoR provided convincing evidence of in vivo competition. With excess P_i, PhoR activated phoA and P_vanH-lacZ expression ca. 7- and 90-fold, respectively, when only PhoB or VanR_a was present (Fig. 2). Activation of phoA under these conditions was an apparent consequence of PhoR overproduction from P_rhaB (data not shown). When both PhoB and VanR_a were present, activation of VanR_a by PhoR was <2-fold, whereas activation of PhoB was unaffected. The ca. 40-fold reduced activation of P_vanH caused by PhoB indicates that a PhoB-PhoR interaction interferes with PhoR autophosphorylation when P_i is in excess.

Under limited P_i conditions, PhoR activated phoA and P_vanH-lacZ expression ca. 2,900- and 90-fold, respectively, when only PhoB or VanR_a was present (Fig. 2). High activation of PhoB under these conditions was anticipated. These results also showed that activation of VanR_a by PhoR is unresponsive to P_i limitation. In agreement, competition between PhoB and VanR_a occurred with limited P_i as well. An ca. 2,600-fold activation of phoA expression was accompanied with only a ca. 7-fold activation of P_vanH-lacZ expression when both response regulators were present. That is, ca. 13-fold reduction of P_vanH-lacZ expression resulted because of the presence of PhoB.

Absence of Cross Talk from VanS_b in Vivo.

VanS_b was tested for activation of both VanR_a and PhoB. No activation of VanR_a by VanS_b occurred under conditions leading to activation of VanR_a by VanS_a or PhoR (Table 3). VanS_b also was tested for activation of PhoB^wt, PhoB_{M17V, E87D}, and PhoB_T97A. The latter are altered recognition (AR) mutants showing 200- to 400-fold enhanced activation by ′VanS_a (12). Yet, no activation by VanS_b occurred (data not shown).

Table 3.

Activation of VanR_a and VanR_b by VanS_a, VanS_b, and PhoR

Regulator, kinase	lacZ fusion*	Carbon source	β-Galactosidase Sp Act
			−Rha	+Rha
None	PvanH	Fru	0.7  ±  0.1	N.D.
VanRa	PvanH	Fru	0.9  ±  0.3	N.D.
VanRa, PrhaB-vanSa	PvanH	Fru	0.7  ±  0.0	15.2  ±  1.9
VanRa, PrhaB-vanSb	PvanH	Fru	0.8  ±  0.1	0.8  ±  0.1
VanRa, PrhaB-vanSb	PvanH	Gly	0.7  ±  0.1	0.7  ±  0.1
VanRa, PrhaB-phoR	PvanH	Fru	0.7  ±  0.2	21.2  ±  1.3
VanRa, PrhaB-phoR	PvanH	Gly	0.8  ±  0.2	47.6  ±  8.4
None	PvanY	Fru	7.6  ±  1.2	N.D.
VanRb	PvanY	Fru	42.5  ±  3.8	N.D.
VanRb, PrhaB-vanSa	PvanY	Fru	38.7  ±  2.4	43.1  ±  1.8
VanRb, PrhaB-vanSa	PvanY	Gly	22.7  ±  2.7	37.4  ±  3.3
VanRb, PrhaB-phoR	PvanY	Fru	32.2  ±  2.3	34.2  ±  1.1
VanRb, PrhaB-phoR	PvanY	Gly	26.0  ±  5.2	37.0  ±  2.6

Cells were assayed as in Table 2. N.D., not determined. 

^*Strains are integrants of BW24381; JCS182 (pJS94), JCS186 (pJS97), JCS196 (pJS94, pSLF55), JCS201 (pJS97, pJS107), JCS229 (pJS94, pSLF55, pAH151), JCS230 (pJS94, pSLF55, pJS87), JCS231 (pJS94, pSLF55, pJS102), JCS238 (pJS97, pJS107, pJS87), and JCS239 (pJS97, pJS107, pAH151). 

In Vitro Analysis of VanS_b.

Unable to demonstrate VanS_b kinase in vivo, an MBP-′VanS_b fusion protein was purified by using the same strategies used to study both VanS_a (26) and PhoR (M.K.P., J.C.S., B.L.W., and C.T.W., unpublished results). Consistent with the in vivo results, MBP-′VanS_b showed no autophosphorylation in the absence or presence of detergents intended to facilitate proper folding (0.002 to 1.0% SDS, Brij 58, or Nonidet P-40). To investigate whether the truncation or overproduction of MBP-′VanS_b was deleterious to its folding, full-length VanS_b also was overproduced as an N-terminal, His-tagged, or MBP-VanS_b fusion protein. In no case would VanS_b undergo autophosphorylation, even in the membrane fraction. Although it is possible that VanS_b kinase is particularly labile, it is also reasonable to suppose that under these conditions VanS_b is heavily biased as a P-VanR_b phosphatase.

DISCUSSION

Although we previously have reported in vitro studies on purified native VanR_a and an MBP-′VanS_a fusion protein to the VanS_a cytoplasmic domain (8, 19, 26) and in vivo studies on VanR_a and VanS_a in E. coli (9, 15), little information has been available describing in vitro and in vivo features of the type B regulatory proteins (17). When P_vanRa- and P_vanH-lacZ fusions were tested for activation by VanR_a and VanS_a in previous E. coli in vivo studies (9), activation of P_vanH, but not P_vanRa, was observed. Using reporter strains that provide improved signal-to-noise ratios, a 5-fold increase in P_vanRa expression was seen when VanS_a synthesis was driven from the tightly regulated P_rhaB (14). This P_vanRa response is about one-eighth as strong as the 40-fold activation of P_vanH. These E. coli strains now provide a more faithful readout of the type A VRE promoters.

Initial results on VanR_b in E. coli showed ca. 5-fold activation of both type B VRE promoters, P_vanRb and P_vanY, in the absence of VanS_b, consistent with the phosphorylation of VanR_b by an unidentified endogenous E. coli kinase. This cross talk to VanR_b is in contrast to the behavior of VanR_a in E. coli where no significant activation of P_vanRa or P_vanH has been observed in the absence of VanS_a. That VanR_a and VanR_b are recognized by different E. coli kinases was unexpected; it is explicable by the fact that VanR_a and VanR_b share only 34% sequence similarity whereas the VanH, VanA/VanB and VanX resistance enzymes share more than 67% sequence similarity (17).

A second distinction between the type A and B VRE systems in E. coli was the behavior of the sensor kinases. In these studies, VanS_b exhibited only a phosphatase and no discernable kinase activity toward its cognate response regulator. In contrast, VanS_a exhibited both kinase and phosphatase activity under similar conditions (9). The issue of balance between phosphoaspartyl-response regulator-phosphatase and phosphohistidinyl-autophosphorylation-kinase has been noted elsewhere, but only for VanS_a rather than VanS_b (Fig. 3). While studying control of VanR_a by VanS_a in Ent. faecium BM4147, Arthur et al. (27) showed that VanR_a was constitutively active in the absence of VanS_a, presumably because of cross talk from an unidentified enterococcal kinase. In the presence of VanS_a and absence of vancomycin or teicoplanin, VanS_a acted predominantly as a phosphatase to reduce P-VanR_a-dependent transcription. Only in the presence of a glycopeptide inducer did VanS_a shift from a phosphatase to kinase.

Figure 3.

Signaling in the VanS-VanR phosphorelays and activation of the type A and B VRE promoters. Sensor kinase reaction scheme involves its autophosphorylation and phosphotransfer to the response regulator (thick solid arrows) leading to activation of the respective promoters (thick dashed arrow). Phosphatase reaction of the sensor kinase dephosphorylates the cognate response leading to down-regulation of the promoter in the absence of external stimulus (thin dashed arrows). While VanS_a has both kinase and phosphatase activities, whether VanS_b has a kinase activity has not been established (see text).

In this study, VanS_b apparently was locked as a phosphatase. To what degree this reflects its true phosphatase/kinase balance in VRE remains to be determined. Along with the unrelieved VanS_b phosphatase, there was no indication that VanS_b phosphorylates VanR_a or PhoB. In contrast, VanS_a phosphorylates both. The inability of VanS_b to activate PhoB was especially surprising. Under similar conditions, five other nonpartner kinases activate PhoB^wt and PhoB^AR, making VanS_b the sole exception (A.H., S.-K. Kim, and B.L.W., unpublished results). Because of amplification of PhoB synthesis resulting from its autogenous positive control, PhoB activation should provide a very sensitive test for cross talk (10), providing further evidence that VanS_b is in a kinase inactive state under these conditions. Although the absence of kinase activity for VanS_b is an anomaly, more rigorous experiments are required to prove that VanS_b is not a kinase. Nevertheless, these results support the hypothesis that the VanS_b-VanR_b system is regulated largely or exclusively through a phosphatase on/off balance rather than phosphatase/kinase balance. Accordingly, VanS_b might act solely as a negative regulator of the type B system, whereby its phosphatase activity would be inhibited in the presence of vancomycin, thus permitting phosphorylation of VanR_b by a different kinase in VRE and allowing P-VanR_b to act as a positive regulator.

VanR_a and VanR_b are much less similar than the type A and B VanH, VanA/VanB and VanX enzyme pairs. Likewise, VanS similarities are low, with only 23% similarity between them (17). The disparity between the high similarities for the triad of peptidoglycan-modifying enzymes VanHAX and the low similarities for the regulatory proteins suggests that there was a single origin for VanHAX in VRE, e.g., from one of the vancomycin-producer organisms (28), but that distinct signaling systems were recruited independently for inducible resistance. This teleological explanation for the differences between type A and type B VRE is now supported by several complementary findings. Fig. 3 illustrates different inducers. The failure of type B VRE to acquire teicoplanin resistance results from an inability of teicoplanin to cause induction (7), presumably reflecting different ligands binding to VanS_a and VanS_b. Additionally, no apparent cross talk occurs (Fig. 4), so that mismatches of VanS_a-VanR_b or VanS_b-VanR_a would not produce peptidoglycan alterations. Finally, even though both pairs of VRE promoters function in E. coli, there was no cross activation.

Figure 4.

Type A and B VRE two-component regulatory circuits. A signal transduction event from one sensor (VanS_a or VanS_b) to its cognate response regulator (VanR_a or VanR_b) leads to activation of the type A (P_vanH or P_vanRa) and type B (P_vanY or P_vanRb) promoters without apparent cross talk between the two systems.

It remains uncertain whether VanS_a or PhoR activates VanR_b by cross talk. Because VanR_b is activated by an unknown kinase in E. coli, only modest effects because of VanS_a (1.6-fold) or PhoR (1.4-fold) were seen (Table 3). Greater effects might not have been evident if P_vanY expression were already near maximally activated under these conditions.

Activation of VanR_a by PhoR and PhoB by VanS_a are clear examples of cross talk (Table 3, Fig. 2). It is noteworthy that high activation of VanR_a by PhoR occurred only in the absence of PhoB. In an earlier study, PhoB was activated by the kinase EnvZ only in the absence of its cognate response regulator OmpR (23). Although activation of PhoB by EnvZ failed to respond to an environmental signal(s), the significance of that finding was unclear as the ligand(s) that stimulates EnvZ is unknown (29). It therefore is intriguing that activation of VanR_a by PhoR was strongly inhibited by PhoB under both excess and limited P_i conditions. These results also show that PhoR autophosphorylation (or P-PhoR phosphotransfer) is somehow down-regulated under P_i excess conditions in a PhoB-dependent manner. Such an effect supports the model that PhoB and PhoR exist in an “inhibition complex” under these conditions (30). The absence of normal (ca. 2,000-fold) control by P_i over activation of VanR_a by PhoR suggests that interactions between nonpartner proteins because of cross talk are unregulated. Yet, interactions between some nonpartner proteins are clearly physiologically regulated. Activation of PhoB by the kinase CreC responds to unknown catabolites (31). Likewise, nonpartner interactions between the NarX-NarL and NarQ-NarP two-component proteins are regulated (32). Hence, nonpartner interactions indicative of cross talk would appear to be uncontrolled whereas those indicative of “cross-regulation” would be physiologically regulated (25, 33).

There is substantial interest in developing inhibitors to one or more of the five necessary and sufficient proteins in the type A and type B VRE pathways to obviate vancomycin resistance (34). The sensor kinase-response regulator transcription factors are reasonable sites to target for interdiction of the signal that leads to glycopeptide antibiotic resistance. This work to elucidate the apparent differences between the VanRS_a and VanRS_b systems ultimately may help to define a compound that will inhibit both type A and B VRE. Thus, inhibitors of the Bacillus subtilis KinA-SpoOF two-component regulatory system recently have been shown to be antibacterial against Ent. faecalis ATCC 29212 and Ent. faecalis OC3041. It will be interesting to see whether compounds such as RWJ-49815 inhibit the VanRS_a and VanRS_b pairs equivalently, studies that will be enabled by using E. coli reporter systems similar to ones described here in in vivo assays.

ABBREVIATIONS

BG

β-galactosidase

MBP

maltose binding protein

VRE

vancomycin-resistant enterococci