INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

An Escherichia coli global regulatory protein, cyclic AMP (cAMP) receptor protein (CRP), is involved in the regulation of transcription, either positively or negatively, in many genes involved in carbon metabolism (4, 18). CRP, a homodimer, undergoes a conformational change when complexed with its allosteric effector cAMP and binds to a specific sequence located near or within target promoters to regulate transcription (18). The cellular level of cAMP is controlled by carbon sources, in part by the inhibition of adenylate cyclase activity by glucose (36). CRP is present predominantly in the cAMP-complexed, active conformation under non-catabolite-repressed conditions (e.g., growth with glycerol as the carbon source). The level of CRP-cAMP complex is reduced under catabolite-repressed conditions (e.g., growth with glucose), decreasing the activities of CRP-dependent promoters (4, 28).

The CRP binding sites lie at different locations with respect to transcriptional initiation sites of various promoters (44). Simple CRP-dependent promoters in which CRP-cAMP alone is sufficient for activation have been grouped into two classes based on the location of the recognition site for CRP and the corresponding mechanisms for activation (2, 13). For class I CRP-dependent promoters, such as lacZ and malT promoters, the CRP binding site is located upstream of the 35 region (2, 12). For class II CRP-dependent promoters like galP1, the CRP binding site is centered near the 35 region (7). The interaction between CRP and RNA polymerase has been shown to play a critical role in the activation of transcription in CRP-dependent promoters (6, 18).

SecB, an export-specific molecular chaperone in E. coli, promotes protein export across cytoplasmic membranes for a subset of precursor proteins. SecB binds to these precursors, thus keeping them in loosely folded translocation-competent conformations (22, 35, 46). Subsequently, SecB-precursor complexes target the membrane translocase composed of SecYEG and SecDF-YajC complexes via interaction through a soluble and membrane-associated translocation ATPase, SecA protein (5, 10, 16, 35).

Our previous studies showed that the synthesis of SecB is controlled by carbon nutrients, in contrast to those of other Sec proteins, which are growth rate dependent (39). The cellular amount of SecB is reduced in the presence of glucose. CRP-cAMP has also been shown to be involved in SecB synthesis. Exogenous cAMP partially compensates for the repressed level of SecB in the presence of glucose, and the compensatory recovery by the addition of cAMP is lost in cells lacking functional production of CRP from a cya and crp mutant strain. When plasmids carrying the wild-type crp gene were reintroduced into this cya and crp mutant strain, cells were shown to restore the response to the exogenous cAMP. Moreover, deletion studies on the upstream portion of secB suggest that the expression of secB is controlled by more than one promoter (39). In this study, we report the identification of two secB-specific promoters and their corresponding 5' transcriptional initiation sites.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. E. coli strains and plasmids used in this study are listed in Table 1.

Cell growth, media, and DNA manipulations. The medium used in this study was the modified minimal medium A (MinA) described by Davis and Mingioli (11) plus CaCl₂ (5 µg/ml), FeSO₄ · 7H₂O (0.25 µg/ml), and thiamine (10 µg/ml). All carbon sources were at 0.5% (wt/vol), and Casamino Acids (Difco) was at 0.1% (wt/vol). Antibiotics were used when required at the following concentrations: 100 µg/ml (ampicillin); 10 µg/ml (tetracycline).

Constructions of secB-lacZ transcriptional and translational fusion plasmids. To construct transcriptional fusions, restriction enzyme-digested fragments from cloned secB gene plasmid in pHK205 (20) were ligated into appropriate multicloning sites of pQF50. For translational fusions, either restriction enzyme-digested fragments or PCR products, with pHK205 as a template with corresponding primers (Table 2), were first cloned into SmaI-digested pUC18. Then, fragments between the XbaI site of pUC18 and the secB internal EcoRV site were inserted into XbaI-SmaI-digested pQF52. All these fusion plasmids were transformed into ZK4, and transformants were selected on 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal) plates. All fusion plasmids were reisolated, and DNA sequences were confirmed by DNA sequencing.

-Galactosidase assay. -Galactosidase activities were determined in permeabilized cells (29). Cells harboring lacZ fusion plasmids were normally grown in glycerol medium to exponential phase (OD₆₀₀, 1.2) and washed once with MinA medium. Cells were reinoculated into minimal medium with glycerol, glucose, or glucose plus cAMP (10 mM in final) to an OD₆₀₀ of 0.1. Cells were allowed to grow for 3.5 generations before the -galactosidase assay. For RH77 cells, cells from overnight culture in the glucose medium with 0.1% Casamino Acids were grown to exponential phase in the presence or absence of cAMP for about 3.5 generations.

RNA preparation and primer extension. Total RNAs were prepared from ZK4 cells harboring pHK205, pTQ17, or pTQ46. Cells grown in glycerol minimal medium were harvested and transferred to minimal medium with either glycerol or glucose as the carbon source and were grown for 3.5 generations. Total RNAs were extracted by the acidic hot phenol method (33); briefly, cell pellets were suspended in lysis buffer containing 0.02 M sodium acetate (pH 5.2), 1 mM EDTA, 0.5% sodium dodecyl sulfate (SDS), and 0.2% diethylpyrocarbonate (DEPC). An equal volume of phenol equilibrated with 0.02 M sodium acetate (pH 5.2) was added, and the mixture was incubated at 65°C for 5 min with gentle shaking. The aqueous phase was reextracted with an equal volume of phenol-chloroform (1:1 [vol/vol]) and precipitated with 3 volumes of 100% EtOH (ethanol). The RNA pellet was dissolved in DEPC-treated deionized H₂O and treated with 10 U of RNase-free DNase I in the presence of 40 U of RNase inhibitor (Boehringer Mannheim, Indianapolis, Ind.). The RNA was precipitated by adding 1/10 of the volume of 3 M sodium acetate and then 2.5 times the volume of 100% EtOH.

Construction of pT7-Crp and CRP protein purification. To clone the E. coli crp gene behind bacteriophage T7 RNA polymerase promoter, the crp gene from pHA7E (1) was excised with BamHI and EcoRI digestions. The 1-kb BamHI-EcoRI fragment was inserted in pT7-6 (43) digested with BamHI and EcoRI. The resulting plasmid, pT7-Crp, was transformed into BL21(DE3), which carries an isopropyl--D-thiogalactopyranoside (IPTG)-inducible T7 RNA polymerase gene.

Site-directed mutagenesis and secB DNA fragment preparation. DNA fragments containing various secB upstream regions were amplified by PCR with appropriate oligonucleotides shown in Table 2, with pHK205 as the template. Amplified fragments were subcloned into the SmaI site of pUC18 cloning vector. Site-directed mutagenesis on the secB CRP binding site and putative binding sites was performed by PCR amplification with base pair-substituted primers (shown in Fig. 4A and Table 2). After subcloning into pUC18, the orientation of inserts and sequences of all cloned fragments were verified by DNA sequencing for both coding and template strands (AmpliTaq; Applied Biosystems Inc., Foster City, Calif.). DNA fragments were purified from agarose gel after 5'-HindIII and 3'-EcoRI digestions and labeled with [-³²P]dATP and dGTP (New England Nuclear, Boston, Mass.) by using Klenow enzyme (Promega, Madison, Wis.).

Gel-mobility shift assay and DNase I footprinting. DNA fragments employed in these experiments are shown in Fig. 4B. ³²P-labeled DNA fragments and purified CRP proteins were incubated in binding buffer (10 mM Tris-HCl [pH 7.8], 50 mM KCl, 1 mM EDTA, 50 µg of bovine serum albumin/ml, 1 mM DTT, 0.05% Nonidet P-40, 50 µM cAMP, 20 µg of salmon sperm DNA [Sigma Chemical Co., St. Louis, Mo.]/ml) for 30 min in a total volume of 20 µl. Then, 3 µl of loading buffer (binding buffer containing 50% glycerol and 0.1 mg of bromophenol blue/ml) was added, and the samples were immediately loaded on a 5% polyacrylamide gel (Protogel; National Diagnostics), with current applied. Then electrophoresis was initiated at a low voltage and progressively increased to 150 V over 60 to 75 min. The electrophoresis buffer contained 10 mM Tris-HCl (pH 7.8), 1 mM EDTA, and 50 µM cAMP and was replaced with fresh buffer a couple of times during the run.

9

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The expression of the secB gene is controlled by two distantly separated promoters, one of which is subjected to catabolic regulation. Previous studies on the SecB synthesis suggested that there might be separate promoters for secB expression with different regulatory mechanisms (39). To identify these putative secB promoters, secB-lacZ transcriptional fusions were constructed (Fig. 1). The presence of promoter activity was examined by measuring -galactosidase activities from cells carrying these fusion plasmids. The responses to either glycerol or glucose as a sole carbon source and the effect of added cAMP were determined (Fig. 1). Two significant promoter activities were observed, one upstream of the HindIII site (at the 315 bp, pTS101), and the other downstream of the HindIII site (pTS103). Cells carrying plasmid pTS101 essentially did not respond to different carbon sources or to the addition of cAMP in the presence of glucose. In contrast, -galactosidase activities in cells carrying plasmid pTS103 were significantly reduced in the presence of glucose, but not as severely if cAMP was also present. These responses were comparable to those of actual SecB synthesis observed with plasmid pKS101 containing a deletion of the HindIII upstream region (39). Plasmid pTS107, carrying a 5' end extended by about 200 bp, did not show significant differences from pTS103.

View larger version (31K): [in this window] [in a new window]   FIG. 1.   Overview of secB and neighboring genes and diagrams of secB-lacZ fusion constructions. Boxes indicate ORFs for yibN, grxC, secB, and gpsA. The yibN product has not yet been identified; grxC has recently been suggested to code for glutaredoxin 3; gpsA encodes sn-glycerol-3-phosphate dehydrogenase. Solid lines show intergenic regions of these genes. For secB-lacZ fusion constructions, solid lines represent secB upstream regions fused to the corresponding lacZ gene, solid bars represent a promoterless lacZ gene or a promoterless lacZ gene without a ribosome-binding site, and MCS denotes multicloning sites. Promoter activities were determined by measuring -galactosidase activities for cells carrying corresponding fusion plasmids with different carbon sources and with the addition of exogenous cAMP. During exponential growth in the glycerol medium, cells were transferred into indicated media. The activities of -galactosidase (expressed in Miller units) were assayed. Data represent the average of three independent experiments, and standard error was typically within ±10%.

Determination of 5' ends of secB transcripts from the P1 and P2 promoters. The transcriptional initiation sites were determined by primer extension assays (Fig. 2). Total mRNAs were prepared from cells harboring the plasmid-encoded secB gene, pHK205 (21) or secB-lacZ fusion plasmids, pTQ46 or pTQ17, grown in glycerol or glucose minimal media. The 5' end of the secB transcript from the proximal promoter P2 was located 75 bp upstream of the initiation codon of SecB and subsequently designated +1 (Fig. 2B). Upon quantitation of primer extension products with mRNAs from glycerol- or glucose-grown cells, P2 transcripts were consistent with -galactosidase activities of pTQ46 measured at the time of total RNA preparations (data not shown). The 5' end of distal promoter P1 transcript was also determined by a separate primer extension experiment with mRNA extracted from plasmid pTQ17 and located at position 988 with respect to the newly identified +1 of P2 transcript (Fig. 2A). Both promoter sequences are shown in Fig. 2C. These 5' transcriptional initiation sites were in agreement with previous fusion studies (Fig. 1).

View larger version (56K): [in this window] [in a new window]   FIG. 2.   Primer extension for the identification of 5' ends of transcripts from the secB P1 and P2 promoters. Total RNAs were prepared from ZK4 harboring plasmid pTQ17 grown in glycerol or plasmid pTQ46 or pHK205 grown in glycerol or glucose media as indicated. Primers complementary to coding strand were extended at 42°C for 45 min with reverse transcriptase. Reference DNA sequencing was done with the same primers as those in primer extensions. (A) P1 promoter. (B) P2 promoter. Total RNA (2.5 and 5 µg, respectively) was used for panels A and B, respectively. Asterisks indicate 5' positions from each respective promoter. (C) Sequences of the secB P1 and P2 promoters.

Characterization of the roles of CRP-cAMP in the secB expression: direct interaction between the secB and the CRP-cAMP complex. The addition of cAMP in the glucose medium resulted in increased transcriptional activation at the secB P2 promoter (Fig. 1). This recovery from catabolic repression requires the presence of the crp gene (39). These results suggest the involvement of the CRP-cAMP complex on the expression of secB. The possible interaction between the CRP-cAMP complex and the secB P2 promoter regulatory region was examined directly, by measuring the binding activity of purified CRP protein on the DNA fragment containing the secB P2 promoter. Preliminary results showed that CRP effectively bound to the secB P2 promoter region in the gel electrophoretic mobility shift assay (data not shown; see Fig. 5A).

View larger version (73K): [in this window] [in a new window]   FIG. 3.   CRP binding site for the secB P2 promoter. DNase I footprinting experiments with DNA fragment containing the secB promoter regions were carried out with purified Crp proteins. Final Crp protein concentrations are 0, 0.2 × 109, 0.4 × 109, 0.8 × 109, 1.6 × 109, and 6.4 × 109 M in lanes 1 to 6, respectively. Vertical dark bars indicate protected binding regions by Crp proteins, and numbers are in respect to the +1 transcriptional initiation site from the P2 promoter.

9

11

9

View larger version (23K): [in this window] [in a new window]   FIG. 4.   CRP binding consensus sequence comparison and DNA fragments for CRP binding assay. (A) Bold face letters indicate base pairs replaced by those in parentheses by site-directed mutagenesis. (B) DNA fragments used in the CRP binding assay and their relative positions with respect to the secB P2 CRP binding site. The probe numbers correspond to the 5' primers for PCR amplifications, except Pr-102, which has the same 5' end as that in pTQ102 in Fig. 2. Pr, probe; BCRP, secB P2 CRP binding site.

View larger version (49K): [in this window] [in a new window]   FIG. 5.   CRP titration and mutational analysis on secB CRP binding sites. CRP binding affinities were determined with various secB DNA fragment probes carrying mutations on the CRP binding site as described in the legend for Fig. 4. (A) Binding of CRP protein on the secB gene. Gel-mobility shift assays were carried out with secB P2 containing DNA fragments and purified CRP. 32P end-labeled secB DNA probes were incubated with increasing amounts of purified Crp protein in the presence of 50 µM cAMP. DNA-protein complexes were analyzed in 5% native PAGE. (B) Site-directed mutagenesis on the BCRP site (secB P2 CRP). CRP proteins are 0, 0.2, 0.8, and 3.2 × 109 M in lanes 1 and 5, 2 and 6, 3 and 7, and 4 and 8, respectively. (C) Deletion and truncation on the BCRP site (secB P2 CRP). Abbreviations: B, bound-DNA probe; F, free-DNA probe; and C, unbound negative control DNA probe.

The effect of the binding site truncation and mutations on CRP binding and secB P2 promoter activity. The identified CRP binding site of the secB P2 promoter (hereafter designated BCRP) was somewhat surprising in that it is not very similar to previously known CRP binding sites (Fig. 4A) (12, 34). In addition, initial searches with Nucleotide Subsequence Search, MacVector, version 6.0 (Oxford Molecular, Campbell, Calif.), revealed two other potential candidates for CRP binding sites, one located at the 119-bp region and the other at the 44-bp region of the secB gene (respective to transcriptional start site +1). To further substantiate the actual BCRP site, deletions and site-directed mutations (Fig. 4A and B) were introduced and compared for CRP binding activities. A site-directed point mutation with a 4-bp substitution in the BCRP site (Fig. 4A) almost completely abolished the CRP binding affinity (Fig. 5B, lanes 5 to 8; Pr-BCRP-M) compared to that without mutations (Fig. 5B, lanes 1 to 4; Pr-BCRP-W). In addition, deletion upstream of BCRP, as in probe Pr-47, showed the same affinity for CRP (Fig. 5C, lanes 4 to 6; Pr-47) as in Pr-46 (Fig. 5C, lanes 1 to 3). However, when the left half of BCRP binding sequences was truncated (probe Pr-60), the ability of this fragment to bind CRP was almost completely abolished (Fig. 5C, lanes 7 to 9). Moreover, mutations at the 44-bp consensus sequence, TGTGA, did not have any effect on CRP binding (data not shown), and this site is not involved in CRP binding, even when the real BCRP site is altered or deleted (Fig. 5B and C). These data provide further evidence that the BCRP site is, indeed, the real CRP binding site for the secB gene, as shown by DNase I footprinting.

View larger version (29K): [in this window] [in a new window]   FIG. 6.   The effect of CRP binding site mutations on secB promoter activity. Promoter activities and their responses to cAMP were determined from wild-type ZK4 cells as well as in cya and crp mutant background RH77 cells carrying secB-lacZ fusion plasmids with mutations on the CRP binding site. Fusion plasmids pTQ60, pTQ78, and pTQ80 were constructed by subcloning of DNA probes used in the experiments shown in Fig. 5B and C into translational fusion vector pQF52 as described in Materials and Methods. All other fusion plasmids were described in the legend for Fig. 1. Cells carrying indicated fusion plasmids were grown overnight in minimal medium containing 0.5% glucose and 0.1% Casamino Acids and were transferred to the same fresh media by dilutions with or without cAMP. Activities were expressed in Miller units, and averages of two independent cultures and standard error were within ±10%.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, two spatially separated promoters for secB were identified by primer extension as well as by transcriptional and translational fusions. These results were in agreement with suggestions from previous physiological studies (39) as well as with E. coli genomic DNA sequence analysis (39, 41). Although these two promoters are separated by almost 1 kb, both lead to specific secB expression as determined by translational fusions (Fig. 1). The removal of the distal P1 promoter from fusion constructs results in a reduction of about twofold in overall secB--galactosidase expression, in agreement with the previous deletion analysis (39). Only the proximal P2 promoter is under the control of different carbon sources, while the distal P1 promoter appears to mediate expression in a constitutive manner under the conditions tested. Furthermore, glucose repression at the P2 promoter is shown to be regulated at the transcriptional level as evidenced by both secB-lacZ fusions and primer extension analysis (Fig. 1 and 2). Regulation at the transcriptional level was further substantiated by the finding that the stability of the SecB protein was unchanged in media containing different carbon sources (data not shown).

The catabolic repression at the secB P2 promoter involves the CRP-cAMP complex, since the addition of cAMP partially relieves glucose repression, and this compensatory recovery requires the presence of the crp gene (39). Gel-mobility shift assays with a DNA fragment containing the secB P2 promoter region and purified CRP protein clearly demonstrate that the effect of CRP-cAMP on secB expression results from a direct interaction between CRP and secB (Fig. 3). It also appears that there is only a single binding site for CRP on secB, since a single shifted band was observed over the range of CRP concentrations tested. The finding that the CRP binding site to the secB P2 promoter is centered at about 61.5 bp upstream of the +1 transcriptional initiation site indicates that it may be a class I CRP-dependent promoter like the lacP1 promoter (12). The estimated CRP binding affinity of secB P2 (0.8 × 10⁹ M) is also comparable to that of lacP1 (0.328 × 10⁹ M [45]), even though the secB site deviated more than lacP1 from the well-defined consensus sequences of CRP binding sites (Fig. 4A). Mutational analyses substantiate the finding that the site newly identified by DNase I footprinting is the real CRP binding site. Moreover, the consensus sequence (TGTGA) near 44 bp does not contribute or function, even in the absence of the real CRP binding sequences.

Two open reading frames (ORFs) upstream of secB also appear to be under the control of the distal P1 promoter. The distal putative ORF to secB is designated yibN, and the proximal ORF has recently been identified as a homologue to grxC, a minor glutaredoxin 3 (3). No other significant promoter activities have been identified for these ORFs, and SecB was expressed from the P1 promoter as efficiently as from the P2 promoter despite the distance between secB and the P1 promoter (Fig. 1). Therefore, it is likely that these four genes, i.e., the two upstream ORFs' genes, secB, and gpsA, form an operon from the P1 promoter whose regulation differs from that of P2. In contrast, the expression from P2 covers only secB and downstream gpsA with carbon source-dependent expression. The physiological and functional relationships among these genes are unclear.

Since the only known physiological function of the SecB protein is as a molecular chaperone in protein export, it was previously assumed that the production of SecB may depend on growth rates of cells to support the increased need for protein translocation in faster-growing cells. Nonviability of secB null strains on Luria-Bertani plates and their viability on the glycerol minimal plate may also be partially explained by growth rate-dependent needs of the SecB protein (20). However, our previous studies showed that SecB synthesis is not related to the growth rates of cells (39). Moreover, Shimizu et al. (40) found that the nonviability of secB null strains on Luria-Bertani plates is due to the insufficient expression of the gpsA gene that is located immediately downstream from secB and encodes biosynthetic sn-glycerol-3-phosphate dehydrogenase. In addition, the gpsA gene is translationally coupled to secB, as the start codon of GpsA overlaps with the stop codon of SecB (41), a finding we have confirmed. The secB P2 promoter region is required for the efficient expression of the gpsA gene (40). We conclude that the expression of secB does not depend on the growth rate of cells; rather, it is related to carbon sources in the media by means of catabolic repression mediated by the CRP-dependent P2 promoter.

The unique regulation of secB expression may be related to the fact that SecB is involved in the export of only a subset of precursor proteins. Thus, the export defect of OmpA caused by SecB deficiency is not as severe as other SecB-dependent precursors such as MBP (maltose binding protein) and LamB (20, 27). The expression of MBP and LamB is also activated by the CRP-cAMP complex (31), whereas the expression of OmpA is constitutive and thus affected little by CRP (30). Since SecB is present in relatively low amounts compared to SecB-dependent precursors (typically 100- to 400-fold in excess of SecB [39]) and SecB binds transiently to these precursors (19), small variations in SecB quantities (twofold overall) may be sufficient to modulate relatively large variations in precursors. Thus, it is possible that CRP-cAMP-activated expression of SecB from the proximal promoter P2 is designed to serve for similarly activated expression of precursors such as pLamB and pMBP. Another scenario, however, is also possible. Precursors of exported proteins are synthesized mainly from membrane-bound polysomes in fast-growing cells. They are synthesized equally from free polysomes and membrane-bound polysomes in slower-growing cells (9). Thus, in faster-growing cells in the presence of glucose, there may be less need for a chaperone for export. Accordingly, the requirement for and the production of SecB are less. Alternatively, other chaperones, such as trigger factor and GroEL (24), can function under these conditions in place of SecB.