INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The control of the Escherichia coli phosphate (Pho) regulon by environmental inorganic phosphate (P_i) levels is a paradigm of a bacterial signal transduction pathway in which occupancy of a cell surface receptor (the P_i-specific binding protein PstS) regulates gene expression in the cytoplasm (reference 29 and references therein). This signal transduction pathway requires seven proteins, all of which probably interact in a membrane-associated signaling complex. The P_i signaling proteins include (i) two members of the large family of two-component regulatory systems, namely, response regulator PhoB (a transcriptional activator) and its partner, histidine sensor kinase PhoR (itself an integral-membrane protein); (ii) four components of the ATP-binding cassette family P_i-specific transport (Pst) machinery (PstA, PstB, PstC, and PstS); and (iii) a negative regulator of unknown function called PhoU.

We proposed elsewhere that the P_i signaling response involves three processes: activation, deactivation, and inhibition (30). Accordingly, activation occurs under conditions of P_i limitation and requires both PhoB and PhoR; activation involves autophosphorylation of PhoR by ATP, phosphotransfer to PhoB, and transcriptional activation of Pho regulon promoters by phospho-PhoB (P-PhoB). Deactivation is a distinct intermediate step-down process that occurs upon a growth shift from P_i-limiting to P_i excess conditions. It is required to reestablish inhibition and leads to the dephosphorylation of P-PhoB in a process requiring PhoR and an excess of either PhoU, a Pst component(s), or both. Inhibition prevents phosphorylation of PhoB when P_i is in excess; it requires all seven P_i signaling proteins (PhoB, PhoR, PhoU, PstA, PstB, PstC, and PstS) in an "inhibition complex" that, by insulating PhoB, interferes with its phosphorylation.

In order to study how PhoU and the Pst system interact with PhoB and PhoR, we had previously constructed mutants with defined deletions of the pstSCAB-phoU operon. This led to our discovery that a phoU (unlike a phoU missense) mutation causes a severe growth defect, resulting from an apparent P_i sensitivity phenotype (26). Growth inhibition by P_i has also been seen in mutants of the Pst transporter in which the P_i transport channel is permanently "switched on" (open), although in that case the growth defect is much less severe (34). The deleterious growth phenotype resulting from a phoU mutation also leads to the accumulation of compensatory mutants with lesions in phoB, phoR, or a pst gene under normal growth conditions (26). Indeed, the poor growth of a phoU mutant was apparently responsible for another laboratory concluding incorrectly that a phoU mutation abolishes P_i uptake (19), as their "phoU mutant" carries also a linked pst mutation (29). On the contrary, we proved that a phoU mutation has no effect on P_i uptake (26).

We have now found ways to overcome the difficulties of studying phoU and (presumably) open-channel pst mutations as well. The growth defect due to a phoU mutation is apparent only when the Pst system is synthesized at a high level, which in turn requires increased amounts of PhoB and PhoR, whose synthesis is autogenously controlled (10). We uncoupled phoB expression and phoR expression from their normal controls by placing them behind the foreign, arabinose-regulated P_araB or rhamnose-regulated P_rhaB promoter. Upon induction of the corresponding phoB or phoR mutant with arabinose or rhamnose, such strains show nearly normal P_i control of the Pho regulon. However, the fold induction is lowered. Importantly, a phoU mutation has no deleterious effect on growth of these strains, even in the presence of the respective inducer. Our methods for constructing and characterizing P_araB and P_rhaB fusions may also be generally useful.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Media and culture conditions. Luria-Bertani broth, tryptone-yeast extract, and M63 were routinely used as complex and minimal media. These were prepared as described elsewhere (28). To maintain plasmids, antibiotics (Sigma, St. Louis, Mo.) were added as follows: ampicillin at 100 µg/ml, gentamicin at 15 µg/ml, kanamycin at 50 µg/ml, and tetracycline at 12.5 µg/ml. Recombinants with a single-copy plasmid were selected with gentamicin at 4 µg/ml or kanamycin at 10 µg/ml. 5-Bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal; Bachem, Torrance, Calif.) was used at 40 µg/ml to detect -galactosidase. Tetracycline-sensitive (Tet^s) cells were selected on tetracycline-sensitive-selective (TSS) agar as described elsewhere (17). MacConkey agar (Difco, Detroit, Mich.) containing 1% L-arabinose, lactose, or L-rhamnose (Sigma) was used to test for the use of these carbon sources. Ara and Rha strains were verified by their inability to grow on M63 media containing 0.2% L-arabinose or L-rhamnose, respectively. Ara and Rha recombinants are distinguishable from arabinose- or rhamnose-sensitive ones by the inability of the latter to grow on MacConkey or glycerol agar in the presence of arabinose or rhamnose, respectively. In contrast, Ara and Rha recombinants are insensitive to these carbohydrates.

Bacteria. All strains assayed are described in Table 1. Others included BT333 (endA::tetAR; from W. Wackernagel [4]), BW5045 (srlC300::Tn10; [18]), BW8078 (recA⁺ recA1; [28]), BW21391 (leu-63::Tn10; [12]), BW22773 (lacI^q rrnB_T14 lacZ_WJ16 proC::Tn5-132; [14]), CA10 (galU95; from M. Berlyn), JW383 (metF159 zii-510::Tn10 thi-1; from M. Berlyn), W3110trpB114 (trpB114::Tn10; from C. Yanofsky), and ZK1001 (cysC95::Tn10 rpoS::kan; from R. Kolter). Only relevant markers are given in parentheses.

Plasmids and phage. pAH85, pSK49, and pSK58 were constructed in an earlier study; each has the same backbone as pSK50uidA2 (13). pAH85 and pSK58 are similar, except that the former has a silent SpeI site at codon 113 of phoB. Each expresses phoB from its native promoter. pSK49 expresses phoB from P_tac. pAH136 and pAH150 are derivatives of pAH85 and pSK49 carrying P_rhaB and P_araB in place of P_phoB and the lacI^q-P_tac region, respectively; pAH151 and pAH152 are similar plasmids carrying P_rhaB-phoR⁺ and P_rhaS-phoR⁺ except that they have the attachment site of HK022 and encode gentamicin resistance (14). pBAD32 and pBAD33 (11) were from L.-M. Guzman. They are similar except that pBAD32 has an undefined deletion of ca. 0.6 kbp between the polylinker and the cat gene of pBAD33 (data not shown). pBC35 is a derivative of pBGS19+ (25) carrying the phoBR operon within a 4.7-kbp PstI-to-EcoRI fragment (32) (Fig. 1). pSK47 (16) has the same phoBR fragment cloned into the backbone of pSK50uidA2 (13). pSLF13 and pSLF22 were constructed in an earlier study (8). pSLF13 is a derivative of pSPORT1 (Gibco-BRL, Bethesda, Md.) that encodes a segment of VanS within a 0.4-kbp EcoRI-to-BamHI insert that has a start codon overlapping an NdeI site immediately downstream of a synthetic ribosome-binding site. pSLF22 is a derivative of pBAD32 containing a similar insert. pWJ17 and pWJ18 are derivatives of the lacZ transcriptional fusion vector pWJ13 containing a kanamycin resistance cassette as a PstI fragment, whose loss facilitates recognizing promoter-lacZ fusion plasmids as kanamycin-sensitive (Kan^s) derivatives (15). Fusions made with these plasmids were recombined onto the chromosome by allele replacement as described below.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Structure of the phoBR chromosomal region. The 4.7-kbp PstI-to-EcoRI fragment in pBC35 and pSK47 is shown. Asterisks mark the NdeI and BamHI sites that were previously introduced by site-directed mutagenesis upstream and downstream, respectively, of the phoB coding region. The phoB578 mutation was made in pLD82 (Table 2) and recombined onto the chromosome by allele replacement as described in the text. The construction of phoR574 and phoBR580 is described elsewhere (12). Arrows show gene orientations.

R6K

Molecular biology methods. PCR amplifications of ara, rha, and phoR sequences were carried out with Vent DNA polymerase (New England Biolabs, Beverly, Mass.) and oligonucleotide primers (Table 3; IDT Inc., Coralville, Iowa). Other enzymes were from New England Biolabs or Promega (Madison, Wis.). QIAGEN (Hilden, Germany) products were used for isolation of plasmid DNA, extraction of DNA fragments from agarose gels, and purification of PCR fragments. The phoR, P_araB, and P_rhaSB fragments were sequenced on both strands at the Dana Farber Cancer Institute Molecular Biology Core Facility (Harvard Medical School, Boston, Mass.). The `araD' and `rhaD fragments (the prime indicates that a gene portion is missing on the side of the prime) were verified only by restriction enzyme analysis.

Molecular genetics. Many new strains were constructed by generalized P1 transduction as described elsewhere (28). BW23321 (lac-169 hsdR514 uidA(MluI)::pir⁺ endA_BT333) was made by transduction of BW21116 (17) to tetracycline resistance (Tet^r) with P1 grown on BT333 followed by transformation of resultant transductant BW23298 with FLP expression plasmid pCP20 and excision of the tetAR genes as described elsewhere (4). The notation endA_BT333 signifies the resultant allele. In order to construct strains carrying a minimum number of antibiotic resistance markers, new mutations were usually introduced in two steps. In the first step, a linked auxotrophic marker was introduced by selecting Tet^r transductants. These were then made prototrophic with P1 grown on an appropriate donor; the resultant transductants were scored for loss of antibiotic resistance and other relevant phenotypes. Several new mutations and fusions were constructed on plasmids as described above and then recombined onto the chromosome by allele replacement as described elsewhere (17). The resultant chromosomal allele is often given a designation corresponding to the plasmid used in its construction. As necessary, transductants carrying a phoU mutation were verified by complementation with RZ5lacP-phoU⁺ upon introduction of phoB⁺ or phoR⁺ or by backcross of the phoU mutation into an appropriate recipient.

R6K

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Structures of the ParaB and PrhaB allele replacement plasmids. Arrows show gene and promoter orientations. Open boxes in pAH33, pAH54, and pLD78 are promoter regions; hatched boxes are coding regions. All sites are shown for the enzymes indicated. NdeI sites are subscripted to correspond to those indicated in Table 2. pAH33 and pAH54 contain unique PstI and SalI sites downstream of ParaB for construction of a ParaB fusion. Genes cloned into these sites require also a ribosome binding site. pLD78 has SalI, XbaI, BamHI, and SphI sites downstream of PrhaB for construction of a PrhaB fusion. pLD78 has the native rhaB ribosome-binding site upstream of the NdeI2 site, which corresponds to the normal met start codon of rhaB. Therefore, genes cloned into this site do not require a ribosome-binding site, although partial digestions are required to use this site. An asterisk marks SphI and BglII sites that were lost during cloning of the `araD' fragment. Arrows show gene and promoter orientations. bla, -lactamase gene; tetA and tetR, tetracycline resistance and repressor genes, respectively; oriT, origin of transfer from RP4; T14 and T1T22, transcription terminators; kan, kanamycin resistance gene; lacZ (op), promoterless lacZ gene for construction of transcriptional, i.e., operon (op), fusions (17). See text.

525 cya-161

Enzyme assays. -Galactosidase and bacterial alkaline phosphatase (BAP) assays were carried out as described elsewhere (28).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Construction of allele replacement vectors for tightly regulated protein synthesis from arabinose- and rhamnose-regulated promoters. Since expression of the phoBR operon is subject to positive autogenous control (10), we considered that it would be advantageous to uncouple PhoB synthesis and PhoR synthesis from their normal controls for new studies on the Pho regulon. We therefore constructed a derivative of P_araB plasmid pBAD32 (11) carrying a P_araB-phoR⁺ fusion (pAH21; Table 2). However, in preliminary studies, we found that substantial amounts of PhoR were apparently made in the absence of inducer. A phoR mutant carrying pAH21 synthesized BAP upon P_i limitation in the absence of arabinose even under conditions of catabolite repression due to glucose (data not shown). No doubt the "leakiness" of P_araB under these conditions reflects the small amount of PhoR required for activation (phosphorylation) of PhoB. To further reduce PhoR synthesis, we assembled an allele replacement, "suicide" vector system to recombine the P_araB-phoR⁺ fusion onto the chromosome in single copy at the ara locus. We also constructed an analogous allele replacement vector carrying the rhamnose-regulated P_rhaB promoter (6) for construction and recombination of a P_rhaB-phoB⁺ fusion onto the chromosome in single copy at the rha locus.

View larger version (27K): [in this window] [in a new window]   FIG. 3.   Recombination of a ParaB-phoR+ fusion onto the chromosome at the araCBAD locus. The construction of pAH35 is described in Table 2. pAH35 can integrate into the chromosome via either of two homologous recombination events (event A or event B). Only an event A integrant is shown for simplicity. Subsequent recombination events result in loss of the plasmid, regenerating the parental strain (event A) or an allele replacement (event B segregant). These are selectable on TSS agar; ones with an allele replacement are recognizable as Ara colonies (see Materials and Methods). A phoR mutant was used to recombine the ParaB-phoR+ fusion onto the chromosome to avoid recombination at the phoBR locus; a wild-type strain can also be used. A black arrow shows the orientation of phoR. Grey arrows show the orientations of other genes. Grey boxes show araC or araD gene fragments and the pir-dependent vector origin (oriRR6K). A prime indicates that a gene portion is missing on the side of the prime. An arrowhead marks the nick site of oriT, the RP4 conjugative transfer origin.

View larger version (26K): [in this window] [in a new window]   FIG. 4.   Allele replacements at the chromosomal araCBAD and rhaRSBAD regions. (A) Structure of the araBAD mutant and a ParaB-phoR+ fusion at the araCBAD locus. (B) Structure of the rhaBAD mutant and a PrhaB-phoB+ fusion at the rhaRSBAD locus. Black arrows represent ParaB-phoR and PrhaB-phoB fusions. Light grey arrows signify genes belonging to the ara and rha gene clusters. Dark grey arrows show the yabI and polB genes flanking the ara locus and the sodA and yiiL genes flanking the rha locus. Dotted lines indicate regions where homologous recombination between the plasmid and the chromosome can occur. See text.

Measurements of arabinose and rhamnose regulation of single-copy promoter-lacZ fusions. To judge the regulatory capabilities of the arabinose- and rhamnose-regulated promoters in this system, we constructed P_araB-, P_rhaB-, and P_rhaS-lacZ fusions and recombined them onto the chromosome in single copy at the lac locus (Materials and Methods). We then examined lacZ expression by measuring -galactosidase levels when strains were grown in the presence and absence of the respective inducer on different carbon sources. Different carbon sources were used because these promoters are subject to catabolite repression. The results are shown in Table 4. In brief, arabinose or rhamnose led to induction of >100-fold to several thousandfold, depending upon the promoter, inducer, and carbon source. As expected, the lowest expression levels were obtained during growth on glucose, for which catabolite repression is the strongest; intermediate expression levels were obtained during growth on fructose, for which catabolite repression is less severe; the highest expression levels were obtained during growth on glycerol, for which catabolite repression is the weakest (7). Differences in both the basal (uninduced) and induced levels exist. In the absence of inducer, the basal level was about 10-fold higher for P_araB than for P_rhaB or P_rhaS. The induced levels were similar for P_araB and P_rhaB (compare BW22831 with BW22887), while the induced levels for P_rhaS were ca. 25% of those for P_rhaB (compare BW22886 with BW22888).

Strategy for using P_araB and P_rhaB fusions to study regulation of the Pho regulon. We constructed a number of phoB and phoR mutants in which PhoB or PhoR is synthesized under the control of a foreign inducible promoter(s). We did this for two reasons. First, we considered that such strains may be useful in our studies on protein-protein interactions between their gene products, as reported elsewhere (13). Second, we suspected that the deleterious effects of a phoU mutation were in part due to high-level expression of the Pho regulon. Because expression of the phoBR operon is under positive autogenous control, high-level expression of the Pho regulon probably requires also high-level synthesis of PhoB and PhoR. We therefore expected to overcome these deleterious effects by down regulating PhoB or PhoR synthesis by using these foreign-regulated promoters.

Arabinose-independent expression of P_araB fusions. The above results show the P_araB promoter to be tightly regulated for expression of the lacZ structural gene. To assess whether P_araB is useful for tight control of a regulatory gene under similar conditions, we constructed strains carrying chromosomal P_araB-phoB⁺ and P_araB-phoR⁺ fusions. Each strain has also a precise deletion of the respective regulatory gene. They have in addition a deletion of creC (creBCD or creABCD) as well as genes for acetyl phosphate synthesis [(ackA pta)], thus eliminating activation of PhoB by the kinase CreC or acetyl phosphate (33). We then examined PhoB- and PhoR-dependent control by measuring BAP levels under conditions of P_i limitation in the absence of arabinose during growth on different carbon sources. As shown in Table 5, P_araB-phoB⁺ phoB strain BW24803 exhibits a null phenotype in the absence of arabinose on all carbon sources. In contrast, P_araB-phoR⁺ phoR strain BW22861 exhibits a PhoR⁺ phenotype in the absence of arabinose, even during growth on glucose. These results are consistent with less PhoR than PhoB being required for transcriptional activation of phoA. Therefore, P_araB appears to be sufficiently tightly controlled for the expression of phoB, but not for the expression of phoR. Yet, phoR expression is clearly limiting under these conditions, as much higher BAP levels are seen in the presence of arabinose (data not shown).

Arabinose- and rhamnose-dependent expression of P_araB, P_rhaB, and P_rhaS fusions. We compared strains carrying chromosomal P_araB-phoB⁺, P_rhaB-phoB⁺, P_araB-phoR⁺, P_rhaB-phoR⁺, and P_rhaS-phoR⁺ fusions to determine which promoter fusion(s) and growth conditions were appropriate for studying gene regulation in the Pho regulon. This was done with strains that are otherwise similar to ones described above. We examined PhoB- and PhoR-dependent control of phoA expression by measuring BAP levels when strains were grown on one of four carbon sources (glucose, mannitol, fructose, or glycerol), which were expected to result in different levels of catabolite repression (7). As shown in Table 7, strains carrying a P_araB-phoB⁺ or P_rhaB-phoB⁺ fusion express a null phenotype in the absence of inducer. The same ones synthesized ca. 12-fold to 2,400-fold more BAP in the presence of inducer, depending upon the carbon source and promoter. The relative expression levels correlate well with expectation in regards to catabolite repression. Catabolite repression is more severe with glucose than mannitol, more severe with mannitol than fructose, and more severe with fructose than glycerol. The BAP levels in these P_araB-phoB⁺ and P_rhaB-phoB⁺ fusion strains (ca. 280 to 490 U; BW24774, BW24777, and BW24773) (Table 7) during growth on glycerol are similar to that in an otherwise isogenic wild-type strain (ca. 375 U; BW24739) (Table 6). Strains with a P_rhaB-phoR⁺ fusion at two chromosomal sites were examined; no difference between them was seen.

Use of foreign promoters to bypass growth defect due to a phoU mutation. We showed above that we can modulate expression of the Pho regulon in our P_araB-phoB⁺, P_rhaB-phoB⁺, and P_rhaB-phoR⁺ fusion strains by growing them on different carbon sources in the presence of the respective inducer. We therefore tested whether these conditions can be used to overcome the deleterious effect of a phoU mutation. Whereas a phoU mutant such as BW17142 grows extremely poorly under all conditions tested, an otherwise isogenic (pstSCAB-phoU) mutant such as BW17335 grows reasonably well (Table 8) (26). These strains have a kanamycin resistance cassette in place of phoU and pstSCAB-phoU sequences, respectively. A similar strain with an unmarked phoU mutation (BW18897; Table 8) also shows a severe growth defect. Yet all these strains synthesize similar amounts of BAP. To determine whether modulating expression of phoB or phoR can be used to overcome the growth inhibition due to a phoU mutation, we constructed otherwise isogenic phoU and (pstSCAB-phoU) strains with a phoB and phoR mutation carrying the respective fusions. These strains have an unmarked phoU mutation, as each has a kanamycin resistance marker elsewhere (Table 1).

View this table: [in this window] [in a new window]   TABLE 8.   Pho regulon control in phoU mutants by using ParaB-phoB+, PrhaB-phoB+, and PrhaB-phoR+ fusions

phoU phoB

phoU phoR

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Our primary goal was to develop a method(s) to study PhoU function. The severe growth defect of a phoU mutant is especially problematic due to the rapid accumulation of compensatory mutants with lesions in phoB, phoR, or a pst gene. A Pst⁺ phoU mutant is also exquisitely sensitive to extracellular P_i, suggesting that PhoU has, in addition to its function in P_i signaling, a role as an enzyme in intracellular P_i metabolism (26). As shown here, we were able to overcome the deleterious effect of a phoU mutation by uncoupling PhoB or PhoR synthesis from its normal autogenous control and expressing phoB or phoR from a foreign promoter(s). To do this, we constructed strains carrying a P_araB-phoB⁺, P_rhaB-phoB⁺, P_araB-phoR⁺, or P_rhaB-phoR⁺ fusion on the chromosome. We used strains with single-copy chromosomal fusions to avoid problems resulting from the use of plasmids, such as an antibiotic requirement, variable plasmid copy number, and even plasmid loss. We also examined ways for modulating expression levels from the P_araB and P_rhaB promoters. We already reported work with similar strains and growth conditions to study the regulatory genes of the Enterococcus faecium vancomycin resistance gene cluster in an E. coli model system (12).

We constructed the corresponding P_araB and P_rhaB fusion strains with conditionally replicative (pir-dependent) allele replacement plasmids pAH33, pAH54, and pLD78 (Fig. 2). Strains carrying these fusions are made via a simple two-step procedure as illustrated in Fig. 3. The resulting P_araB and P_rhaB fusions reside on the chromosome at the araCBAD and rhaRSBAD loci, respectively, as shown in Fig. 4. We made similar strains with the respective araBAD_AH33 and rhaBAD_LD78 chromosomal mutations as controls. In these ways, we constructed a number of strains that express various normal or mutant pho, cre, or van genes under the control of P_araB or P_rhaB in single copy on the chromosome at the respective loci (12-14, 16).

The P_araB promoter (also called P_BAD) has been shown elsewhere to be tightly regulated (11, 21). In those studies, genes that are normally expressed at moderate levels were examined. In preliminary studies, we found that a pBAD plasmid (11) carrying phoR did not provide sufficiently tight control for our purposes, suggesting that these plasmids are somewhat leaky (14). This is consistent with phoR normally being expressed at a very low level. In order to reduce this basal level of expression, we recombined several P_araB-phoR fusions onto the chromosome by using derivatives of P_araB fusion allele replacement plasmid pAH33. When single-copy fusion strains were used, particular P_araB-phoR fusions, e.g., one expressing only the C-terminal kinase domain of PhoR (13), no longer appeared to be leaky. Yet a single-copy P_araB-phoR⁺ fusion strain synthesizes sufficient PhoR for (partial) activation in the absence of arabinose even in the presence of glucose (Table 5). These results suggest that full-length PhoR is a more active kinase than its C-terminal domain. However, in the absence of arabinose, the amount of PhoR synthesis from a P_araB-phoR⁺ fusion is also clearly limiting because upon P_i limitation only ca. 10% as much BAP is made in this strain as in a wild-type strain. In contrast, a single-copy P_araB-phoB⁺ fusion strain shows no leakiness as it exhibits a PhoB phenotype in the absence of inducer on all carbon sources tested (Table 7). Therefore, more PhoB than PhoR is apparently required for expression of the Pho regulon. These data are consistent with PhoR acting catalytically as an autokinase and phosphotransferase in the activation (phosphorylation) of PhoB, which in turn acts as a transcription factor.

In order to control PhoB and PhoR synthesis independently, we also constructed strains with a P_rhaB-phoB⁺ or P_rhaB-phoR⁺ fusion. Although both L-arabinose and L-rhamnose act directly as inducers for expression of regulons for their catabolism, important differences exist in regard to the regulatory mechanisms (Fig. 5). L-Arabinose acts as inducer with the activator AraC in the positive control of the arabinose regulon (23). However, the L-rhamnose regulon is subject to a regulatory cascade; it is therefore subject to an even tighter control. L-Rhamnose acts as an inducer with the activator RhaR for synthesis of RhaS, which in turn acts as an activator in the positive control of the rhamnose regulon (6). As shown in Table 7, both the P_rhaB-phoB⁺ and P_rhaB-phoR⁺ fusion strains show a PhoB and PhoR phenotype, respectively, in the absence of rhamnose. We also examined a P_rhaS-phoR⁺ fusion. Like the P_araB-phoR⁺ fusion strain, the P_rhaS-phoR⁺ fusion strain synthesized a substantial, though limited, amount of BAP in the absence of induction.

View larger version (22K): [in this window] [in a new window]   FIG. 5.   Transcriptional activation of the arabinose and rhamnose regulons. (A) Positive control of the L-arabinose regulon. The araCBAD region near 1.5 min on the E. coli chromosome is shown. The unlinked araE and araFGH genes encode arabinose transport systems (23). (B) Positive control of the L-rhamnose regulon. The rhaRSBAD region near 88.2 min is shown. The nearby rhaT gene is transcribed towards rhaR (Fig. 4) and encodes a rhamnose permease (3). Thick arrows show gene orientations. Curved ones with circled pluses indicate sites for transcriptional activation by AraC (A) or RhaR and RhaS (B). The lower bars show the ara and rha sequences in the respective ParaB and PrhaB fusion allele replacement vectors illustrated in Fig. 2. The dashed lines show the chromosomal regions deleted in the resultant recombinants. See text.

The L-arabinose and L-rhamnose regulons are also regulated by catabolite repression. We therefore modulated expression of P_araB, P_rhaB, and P_rhaS fusions by using various carbon sources (glucose, mannitol, fructose, and glycerol) that lead to different levels of catabolite repression (7). When strains are grown on these carbon sources with the inducer in excess, the expression of the Pho regulon can be modulated 70- to 200-fold in our P_araB-phoB⁺, P_rhaB-phoB⁺, and P_rhaB-phoR⁺ fusion strains (Table 7). As expected, the differences are much smaller in the P_araB-phoR⁺ and P_rhaS-phoR⁺ fusion strains as the expression of these fusions is also leaky under these conditions. In preliminary experiments, we had also attempted to modulate expression levels by using different inducer concentrations. However, we were unable to maintain a constant, steady state level of induction (data not shown) (5). An inability to maintain steady-state induction at intermediate levels by the limiting inducer concentration is expected as the presence of arabinose and rhamnose results also in increased synthesis of their respective transport systems (Fig. 5). This has now been substantiated by monitoring P_araB expression levels in cells grown in the presence of subsaturating inducer concentrations (24). Importantly, as shown here, expression of these promoters can be modulated over a wide range by using different carbon sources in the presence of saturating inducer concentrations.

Both PhoU and the Pst transporter have been implicated in the negative control of the Pho regulon, as mutations of either result in high constitutive Pho regulon gene expression. Presumably, they somehow act together. To study how they interact with the PhoB-PhoR system, we made strains that synthesize PhoB or PhoR under control of the tightly regulated P_araB and P_rhaB promoters. By using these strains, we were able to express phoB or phoR at a reduced level and thereby overcome the harmful effect of a phoU mutation. In these ways, it should now be possible to study further the function(s) of PhoU.