INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The chromosomal tetA(L) locus of Bacillus subtilis encodes a protein that confers resistance to low concentrations of tetracycline (Tc) by catalyzing efflux of a Tc-divalent metal ion complex in exchange for protons (Tc-Me⁺²/H⁺ antiport) (2, 8). Tc efflux does not occur in the absence of a divalent cation such as Co²⁺, Mg²⁺, or Mn²⁺, with Co²⁺ being the best. Conversely, the divalent cation does not efflux via TetA(L) without Tc (9, 12, 32). Modest amplification of the gene or changes in the promoter region lead to increased expression and Tc resistance as does expression of the gene from a multicopy plasmid (4, 16). During the past few years, studies in this laboratory have established that TetA(L) is a multifunctional antiporter. In addition to electrogenic Tc-Me⁺²/H⁺ antiport, TetA(L) catalyzes the exchange of cytoplasmic Na⁺ or K⁺ for a greater number of external H⁺ or K⁺ ions (7-9, 12, 13). The monovalent cation/H⁺ antiporter mode has roles in Na⁺ resistance and pH homeostasis (8). While not yet demonstrated, the net K⁺ uptake monovalent cation/K⁺ antiporter mode could also be of physiological importance.

The enumeration and evaluation of the roles of TetA(L) have been complicated by the variable phenotype of isolates carrying the same tetA(L) deletion. Initial observations on these tetA(L) deletion strains, which focused on the most severe phenotype (e.g., mutant JC112), led to the suggestion that tetA(L) cannot be deleted without compensatory changes that might involve second-site mutation(s) (8). These studies also indicated that the TetA(L) efflux protein had an additional function that accounts for a growth deficit observed in tetA(L) mutants at pH 7 even in the absence of Tc or elevated Na⁺ (8, 13). Initially, it was hypothesized that there might be an endogenous substrate for TetA(L) whose reduced efflux compromises growth (8). However, the more recent finding of the net K⁺ uptake mode of TetA(L) (13) raises the possibility that inadequate K⁺ acquisition is part or all of the basis for the growth defect of tetA(L) mutants at pH 7. The present study was directed toward categorization of the different phenotypes of tetA(L) deletion strains and clarification of the basis for the phenotype at pH 7. We further sought to begin the elucidation of the compensatory changes in tetA(L) deletion strains.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmids, bacterial strains, and growth conditions. The plasmids and B. subtilis strains used in this study are described in Table 1. The tetA(L) deletion mutants from the earlier study, including JC112 and JC112C, carried a chloramphenicol resistance (Cm^r) cassette which replaced the entire tetA(L) chromosomal coding sequence. The tetA(L) gene had been removed as a 1.8-kb ClaI-NdeI fragment in a previously described plasmid, which was then used to produce the mutants (8). One new tetA(L)-containing plasmid, pTL2, was used in this study. The coding sequence of tetA(L) was excised from pTL1 (8) by BamHI-HindIII digestion and was cloned into the BamHI and HindIII sites downstream of the ermC promoter of a pBK15 derivative named pVEB3 (Table 1). In this plasmid, an internal 750-bp MunI fragment of the Cm^r gene of pBK15 was replaced with a 1.2-kb EcoRI fragment of pSP2 containing a spectinomycin resistance (Sp^r) gene. All new constructs were verified by sequence analysis. The sequencing was performed by the Utah State Biotechnology Center (Logan), using an ABI-100 model 377 sequencer. Construction of bacterial strains carrying deletions and/or gene fusions is described below.

Determination of MICs of Tc, Co²⁺, and Na⁺. To determine the MIC of Tc, cells were grown in TKM (pH 7.0) with various Tc concentrations. The A₆₀₀ was measured after 17 h of growth. To avoid precipitation, the MIC for CoCl₂ was determined from the A₆₀₀ of wild-type, JC112, and JC112C cells after 8 h of growth in SpizKM (pH 7.0) containing various CoCl₂ concentrations. For determination of the MIC for NaCl, TKM at pH 7.0 or 8.3 was supplemented with various concentrations of NaCl, and the A₆₀₀ was measured after 15 h of growth. For each of these compounds or ions, the MIC was taken as the lowest concentration at which the A₆₀₀ after the indicated period of growth was below 0.1.

Integration of a tetA(L) gene into the B. subtilis amyE locus. The tetA(L)-containing plasmid pAG4 (30) was digested with HindIII and then treated with mung bean nuclease. Subsequent digestion with ClaI released a 2.5-kb fragment containing tetA(L) and its promoter. This fragment was cloned into the SmaI and ClaI sites of pAC7 (25). The resulting plasmid, pTL6, was digested with NruI and used to integrate the tetA(L) gene into the amyE loci of mutant strains JC112 and JC112C. The strains were identified by initial screening for kanamycin resistance (Km^r) followed by identification of starch-negative strains (25). Strains BTK38 and BTK39 were confirmed to be strains of JC112 and JC112C, respectively, which had incorporated the tetA(L) gene and promoter into the amyE locus.

Northern analyses. Total RNA isolation and Northern analyses were conducted as described previously (6) to assess whether mRNA levels for a variety of genes were elevated. Primers used in PCR to amplify fragments for the probes employed in the analyses are listed in Table 2. A 400-bp PCR fragment (primers yqkI1 and yqkI2) was used as a probe for yqkI; a 400-bp PCR fragment (primers yusP1 and yusP2) was used as a probe for yusP; an internal 270-bp HindIII fragment of the PCR product (primers ycnB1 and ycnB2) was used as a probe for ycnB; a 310-bp NruI fragment of the PCR product (primers yhcA1 and yhcA2) was used as a probe for yhcA; a 700-bp SphI fragment from the mrpA locus was used as a probe for mrpA (14); a 180-bp fragment of pTCC1-25 (5) was used as a probe for yybF. A 780-bp HincII fragment from the yheL coding sequence (Fig. 1A) was used to probe the expression of yheL, an nhaC homologue. A 680-bp PCR fragment (primers Czcd3 and Czcd7) was cloned into pGEM3Zf(+) (Promega) via the HindIII and BamHI sites. The resulting plasmid, pCZCD2, was linearized by HindIII digestion and used as the DNA template for preparation of a riboprobe transcribed in vitro by T7 RNA polymerase. A 410-bp PCR fragment (primers yhaU1 and uhaU2) was used as a probe for yhaU; a 670-bp PCR fragment (primers ykqB1 and ykqB2) was used as a probe for yqkB; an internal 510-kb BclI-ClaI fragment of the PCR product (primers ykrM1 and ykrM2) was used as a probe for ykrM; a ClaI fragment (495 bp) from the 5' end of a PCR product (primers yuaA2 and yubG2) was used as a probe for yuaA and yubG.

View larger version (43K): [in this window] [in a new window]   FIG. 1.   (A) Schematic diagram of the B. subtilis yheKL region. This 2.5-kb fragment represents a reverse strand of nt 42801 to 45300 as reported in the B. subtilis genome project (GenBank accession no. Z99109 BSUB0006). The locations and relative sizes of the yheK and yheL coding sequences, and the oligonucleotides used in this study, are as indicated. The transcription start point mapped in this study is represented by an arrow at nt 44931. (B) Northern analysis of yheL in B. subtilis wild type (Wt) and JC112. A 780-bp HincII fragment from the yheL coding sequence (Fig. 1A) was 32P labeled and used as the probe. The positions of 23S and 16S rRNAs are indicated on the right. (C) Reverse transcription mapping of the yheL transcription start site. The RNA used in the primer extension reactions was isolated from wild type or JC112. Control lanes (labeled A, C, G, and T) are a DNA sequence ladder using the same primer K1 as used in primer extension. The complement of the sequence that contains the mapped transcription start site (+1) and 10 are presented on the right.

Deletion of czcD-trkA from the wild type, JC112 and JC112C. A fragment containing the putative czcD-trkA operon was amplified by PCR with primers Czd6 and Trka6 (Table 2). Primer Czcd6 contained additional nucleotides creating a HindIII site, and primer Trka6 contained additional nucleotides creating a KpnI site. The PCR product was digested with HindIII and KpnI and ligated into the HindIII and KpnI sites of pGEM3Zf(+) (Promega), resulting in plasmid pCZCD5. A 0.3-kb MunI fragment of pCZCD5, containing the 3' end of czcD and the ribosome binding site and 5' end of yheL, was replaced by a PCR-amplified Sp^r gene, producing plasmid pCZCD6. This plasmid was linearized by HindIII digestion and used to transform the wild type, JC112, and JC112C into Sp^r strains BTK33, BTK34, and BTK35, respectively. The disruption of the czcD and trkA genes was confirmed by PCR, restriction analyses, and sequencing.

Analysis of the transcriptional start and promoter region of B. subtilis yheL. The SUPERSCRIPT preamplification system (GIBCO BRL, Life Technologies) and its published standard procedure were applied for mapping the transcription start point. For each reverse transcription reaction, 5 µg of total RNA isolated from the wild type or JC112 and 50 pmol of 5'-end labeled primer K1 or L1 (Fig. 1A and Table 2) were used. Half of each primer extension reaction was resolved on a 5% polyacrylamide-urea denaturing gel, along with a ³²P-labeled size marker. A band of about 200 nucleotides (nt) resulted from extension using primer K1, and a band of about 800 nt resulted from primer L1 (data not shown). The other half of the primer extension reaction from primer K1 was resolved on an 8 M urea-6% polyacrylamide gel along side a sequence standard prepared using the same primer, K1, and a template (fragment K2L7) prepared by PCR amplification using primers K2 and L7 (Table 2). Bands of the same size were detected in both the wild type and JC112 (Fig. 1C). Sequencing of promoter region of yheL was conducted in the wild type, JC112, and JC112C. Chromosomal DNA from each strain was used as a template for PCR amplification (primers K2 and L1 [Table 2]) of a segment containing the putative promoter region (Fig. 1A). Each PCR product was gel purified and then sequenced using primers L1 and K1 (Table 2). The sequence of the PCR products was identical in the three strains and corresponded to that in the B. subtilis genome database (GenBank accession no. Z99109 BSUB0006).

Construction of yheL-lacZ fusions in the wild type, JC112, and JC112C. For production of strains of the wild type, JC112, and JC112C with a yheL-lacZ fusion integrated into the amyE locus, a fragment containing the putative promoter region of the yheL operon, from 160 bp upstream of the mapped transcription start site to the codon for amino acid 299 of the yheL, was amplified by PCR with primers K4 and L3 (Fig. 1A; Table 2). Since there were no sequence differences in the yheL promoter regions among the three strains, only the chromosomal DNA isolated from the wild type was used as a template for the amplification. Primer K4 contains additional nucleotides for an EcoRI site, and primer L3 contains additional nucleotides for a BamHI site. It was thus designed for insertion into plasmid pAC7 (25), which was used to integrate the fusion gene into the amyE locus. The PCR product, K4L3, was digested with EcoRI and BamHI and then ligated into pAC7 at EcoRI and BamHI sites, resulting in an in-frame fusion of K4L3 and the lacZ gene (starting with the first codon of the lacZ coding sequence) in pAC7. Escherichia coli DH5 containing this yheL-lacZ fusion plasmid, pAC7-K4L3, formed blue colonies on Luria-Bertani plates with additions of 0.2 mM isopropylthio--D-galactoside (IPTG) and 0.04% (wt/vol) 5-bromo-4-chloro-3-indolyl--D-galactoside (X-Gal). Plasmid pAC7-K4L3 was digested with restriction enzyme NruI. The linearized pAC7-K4L3 was transformed into the wild type, JC112, and JC112C, and Km^r, starch-negative strains were isolated and verified from each starting strain; BTK15, BTK16, and BTK17 were, respectively, derivatives of the wild type, JC112, and JC112C.

Assays of -galactosidase activities of strains expressing yheL-lacZ fusions. Expression of yheL-lacZ fusion protein in strains BTK15, BTK16, and BTK17 and derivatives thereof was studied by analyzing -galactosidase activity, as described previously (8). -Galactosidase activity was expressed as nanomoles of o-nitrophenyl--D-galactopyranoside hydrolyzed per minute per A₆₀₀ (Table 4). Overnight cultures grown in the media specified in Table 4 were used in the -galactosidase activity assay.

View this table: [in this window] [in a new window]   TABLE 4.   -Galactosidase levels in B. subtilis BTK15, BTK16, and BTK17

Deletion of the yheL gene. First a fragment containing the yheL operon was amplified by PCR with primers K4 and L2 (Table 2; Fig. 1A). Primer K4 contains additional nucleotides for an EcoRI site, and primer L2 contains additional nucleotides for a BamHI site. The PCR product K4L2 was digested with EcoRI and BamHI and ligated into pGEM7Zf(+) (Promega) in the EcoRI and BamHI sites, resulting in plasmid pNH1. The 1-kb HindIII-PstI fragment of pNH1 containing the ribosome binding site and 5' end of yheL was replaced by a PCR-amplified Sp^r gene, producing plasmid pNH2. Plasmid pNH2 was digested with XhoI, and the linearized plasmid was then used to transform the wild type, JC112, and JC112C into Sp^r strains BTK21, BTK22, and BTK23, respectively. The disruption of yheL in the these strains was confirmed by PCR and restriction digestion analyses.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Phenotypes of the tetA(L) deletion strains. At the start of the study, six tetA(L) deletion strains that had been isolated earlier (8) were compared to the wild-type strain with respect to both MICs for Na⁺ and Tc and pH profiles for growth. All of the strains are identical in phenotype either to JC112 or to JC112C, which are distinct, stable phenotypic types; both types are often represented in a particular deletion experiment, but the JC112C type predominates and is sometimes the only phenotype formed. JC112 and JC112C were chosen for the studies reported here. First, the MICs for Tc, Na⁺, and Co²⁺ were determined. A Tc- and Na⁺-sensitive phenotype was expected, given the known functions of TetA(L). Both JC112 and JC112C were more sensitive than the wild type to low concentrations of Tc, exhibiting MICs of 0.09 and 0.12 µg/ml, respectively, versus 1.4 µg/ml for the wild type (Table 3). The expected Na⁺ sensitivity, by contrast, was exhibited by only one of the tetA(L) mutants, JC112, which showed elevated relative sensitivity even at neutral pH, where Na⁺ is less cytotoxic than at pH 8.3 (Table 3). JC112C did exhibit a modestly compromised ability to use either Na⁺ or K⁺ to support growth at pH 8.3, relative to the wild type, but JC112 exhibited a much more pronounced phenotype under such conditions (data not shown). Earlier work showed that this characteristic correlates with Na⁺(K⁺)/H⁺ antiport status in support of pH homeostasis at elevated pH (8).

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Growth of wild-type (wt) and mutant strains of B. subtilis at various concentrations of added K+. The cells were grown in modified TTM (pH 7.0) and adjusted to the indicated concentrations of K+ added to the medium (see Materials and Methods). The A600 was measured after 15 h of growth at 30°C.

Preliminary Northern analyses. From BLAST analyses (1), selection was made of candidate genes that might compensate for different functions of TetA(L). These genes, for which preliminary Northern analyses were conducted, were chosen from three different groups: (i) ycnB, yhcA, yusP, and yybF, four genes whose predicted products have sequence similarity with tetA(L); (ii) mrp (14, 18), a gene locus with known Na⁺/H⁺ antiporter activity, and three genes that had sequence similarity to Na⁺/H⁺ antiporter-encoding genes from other bacteria, yheL and yqkI, which had similarity to the nhaC gene from alkaliphilic Bacillus firmus OF4 (14, 17), and yhaU, which had sequence similarity to the napA gene from Enterococcus hirae (31); and (iii) a gene locus for which there is a putative K⁺ uptake function, czcD-trkA (28), and three genes or gene loci that show homology to one or both of the genes of the K⁺ uptake-encoding ktrAB locus of Vibrio alginolyticus (20), ykqB, ykrM, and yuaA-yubG. Bands of the anticipated size were observed on Northern blots probed for each of these genes or loci (data not shown). The mRNA abundance was clearly elevated for two of them, and for those two loci the increase in mRNA was evident in both JC112 and JC112C. The two loci whose mRNA levels were elevated in JC112 and JC112C were the czcD-trkA locus that had been proposed to be involved in K⁺ uptake (28) and the nhaC homologue yheL.

Role of czcD-trkA in the wild-type and tetA(L) deletion strains. The B. subtilis czcD gene, whose product is 42% identical to that of Ralstonia (previously Alcaligenes) eutrophus (21), is in an apparent operon with a gene that we designated as trkA (28). This locus was identified in our laboratory in an earlier screen of B. subtilis genes that increased K⁺ acquisition by a K⁺ uptake-deficient E. coli strain. Although the B. subtilis trkA gene product exhibits only partial sequence similarity to small regions of the TrkA component of an E. coli K⁺ uptake system, the B. subtilis trkA alone complemented an E. coli carrying a mutation in its own trkA (28). The elevation of czcD-trkA expression in the tetA(L) deletion mutants thus appeared consistent with a compensatory change for purposes of K⁺ acquisition. Northern analyses (Fig. 3) indicated that (i) the czcD-trkA mRNA abundance is low in the wild type grown on SpizKM at pH 7.0; (ii) this basal level of mRNA abundance is elevated in both JC112 and JC112C, and the increase is not abolished upon reintroduction of a functional tetA(L) gene into the two mutant strains (BTK38 and BTK39, respectively); (iii) all strains had a much higher level of czcD-trkA mRNA when grown in the presence of Co²⁺, in the rank order BTK38 [JC112 with tetA(L) restored] > BTK39 [JC112C with tetA(L) restored] > JC112 > JC112C wild type.

View larger version (34K): [in this window] [in a new window]   FIG. 3.   Northern analysis of czcD mRNA in the wild type, JC112, JC112C, and the two mutant strains upon restoration of an active tetA(L) gene, BTK38 and BTK39, respectively. The cells were grown in SpizKM with or without added CoCl2 as shown.

Expression and role of yheL in the wild-type and tetA(L) deletion strains. The putative antiporter-encoding yheL gene is downstream of, and would appear to form a likely operon with, yheK, a gene that exhibits sequence similarity to diverse regulatory genes (Fig. 1A). The size of the RNA detected with a yheL probe was consistent with this expectation, as shown in Fig. 1B, and a transcriptional start was mapped upstream of yheK (Fig. 1A and C). The elevated level of yheL RNA in JC112 relative to the wild type is evident in Fig. 1B. JC112C showed a more variable elevation of yheL mRNA in Northern experiments but had consistently more yheL RNA than the wild type; moreover, the elevated levels of yheL RNA of both mutant strains remained higher than in the wild type upon restoration of a functional tetA(L) either in single copy or on a multicopy plasmid (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A major finding of this study is that the net K⁺ uptake mode (13) of the chromosomally encoded TetA(L) protein plays a significant physiological role in K⁺ acquisition by B. subtilis. The growth deficit of tetA(L) deletion strains, in the absence of a challenge by alkali, Na⁺ and Tc, correlates qualitatively with their [K⁺]-related growth profile. A small growth deficit in JC112C, and larger component in JC112, is not abolished when a functional tetA(L) is restored to these mutants. This irreversible component of the mutants' phenotype is probably an indirect result of a pattern of changes in the expression of other genes. Complex patterns of response to tetA(L) deletion are the other major finding of these studies. The complexity may relate to the diverse functions of TetA(L).

This study showed that tetA(L) deletion strains exhibit two, partially distinct patterns of change in the expression of other genes. Two genes whose expression is significantly elevated have been identified in both phenotypic types of tetA(L) deletion mutants, but the catalogue is almost certainly incomplete. We hypothesize that there may be a regulon whose member genes encode ion-translocating transporters with roles in monovalent cation, pH, and perhaps divalent cation homeostasis. The two gene loci found here to be up-regulated in JC112 and JC112C, i.e., czcD-trkA and yheL, would be members of this putative regulon. Another member is probably tetA(L) itself. It was earlier noted that the basal level of expression of a tetA(L)-lacZ fusion was markedly increased in JC112 (8). A unifying hypothesis would be that the putative regulon has one or more master control genes and it is the function of such a gene that is altered upon tetA(L) deletion. The experiments here show that the increased expression of czcD-trkA and yheL is not abolished by restoration of active tetA(L). This suggests that there has been a mutation or some kind of irreversible change. If one or more second-site mutations are in fact involved, the mutation(s) may not obligatorily or immediately accompany tetA(L) deletion. Preliminary data (J. Jin and D. H. Bechhofer, unpublished data) indicate that the frequency of tetA(L) deletion is just as high as that of a clearly nonessential gene in the same protocol. Thus, TetA(L) functions may not be essential for viability. However, their loss may create selective pressure that leads to the patterns of change that we have begun to characterize here. Further investigation will be required to clarify the timing and nature of the emergence of the irreversible adaptations to tetA(L) deletion. It will also be of interest to determine the full panoply of genes that are up-regulated when tetA(L) is deleted, e.g., by a DNA array technology. It should be noted that some compensatory adaptations to tetA(L) deletion might involve genes that show little or no transcriptional regulation. Even tetA(L) itself is predominantly regulated by posttranscriptional mechanisms (27).

The phenotypic type of mutant represented by JC112C presumably exhibits up-regulation of one or more additional genes beyond those genes that are up-regulated in both JC112 and JC112C. In addition to the probable temperature-sensitive Co²⁺ efflux system, JC112C must have an elevated compensatory transport activity(ies) that accounts for its lack of Na⁺ sensitivity and only modest alkali sensitivity relative to JC112.

One of the two loci whose expression is markedly elevated in both JC112 and JC112C is the czcD-trkA locus. The best-studied homologue, the first czcD gene reported, was initially described as having roles in sensing the substrates of the czcCBA locus of R. eutrophus and in regulating expression of the locus (21, 29). CzcCBA is a toxic divalent cation efflux antiporter whose substrates are cadmium, zinc, and cobalt; the divalent cation is exchanged for external H⁺ (22). CzcD is a member of a family of membrane proteins called cation diffusion facilitators (22, 24). Some of these proteins, including CzcD, have most recently been shown to exclude their substrates when expressed in Ralstonia (3). The B. subtilis CzcD is not as closely related to the Ralstonia homologue as many other homologues, including some B. subtilis gene products. Paulsen and Saier (24) noted that functions of such less related proteins may not involve or may not solely involve transport of toxic divalent cations. In view of the findings here, it will be of interest to examine the actual transport properties of CzcD and CzcD-TrkA in vitro and, in particular, to determine whether the complex catalyzes divalent cation/K⁺ antiport. Such an activity would account for its apparent in vivo contributions to K⁺ acquisition and Co²⁺ resistance.

The YheL protein, like its homologue in alkaliphilic B. firmus OF4 (15), appears to have a modest role in Na⁺-dependent pH homeostasis but makes no detectable contribution to Na⁺ resistance. We propose that yheL be designated nhaC. It is striking that despite the significant up-regulation of yheL in both tetA(L) deletion mutants, deletion of yheL has little effect on these mutants. This is consistent with yheL being part of a regulon containing other genes that are more important compensatory genes for TetA(L) functions, e.g., czcD and trkA. Thus, the strong up-regulation of yheL would be a by-product of an effect on the regulon as a whole. In fact, a plausible but tentative hypothesis for the basis of the Co²⁺ sensitivity of JC112 and JC112C (albeit more modest) is an adverse by-product of the elevation of YheL and other proteins. The barrier function of the membrane may be compromised with respect to divalent cation exclusion when one or more of the membrane proteins is elevated as much as YheL is in JC112. The plausibility of this possibility is supported by the marked decrease in the MIC for Co²⁺ in wild-type B. subtilis cells transformed with a multicopy plasmid expressing yheL. Expression of the gram-negative TetA(C) gene has been correlated with an increase in cadmium sensitivity, and the mechanism is not yet understood (10, 11).

Finally, apart from the enhanced Co²⁺ sensitivity of tetA(L) mutants, the demonstrated effect of Co²⁺ on expression of both czcD-trkA and yheL is notable. It is reasonable to hypothesize that just as Na⁺ stress is exacerbated at elevated pH and K⁺ insufficiency (23), there may be an intersection among stresses related to TetA(L) function and the stress caused by inhibitory concentrations of toxic divalent cations. For example, Co²⁺ toxicity may be strongly dependent on pH, K⁺ status, or Na⁺ levels. Any such intersection may be clarified when all genes whose expression is significantly altered in tetA(L) deletion mutants have been catalogued.