OpuA, an osmotically regulated binding protein-dependent transport system for the osmoprotectant glycine betaine in Bacillus subtilis.

Exogenously provided glycine betaine can efficiently protect Bacillus subtilis from the detrimental effects of high osmolarity environments. Through functional complementation of an Escherichia coli mutant deficient in glycine betaine uptake with a gene library from B. subtilis, we have identified a multicomponent glycine betaine transport system, OpuA. Uptake of radiolabeled glycine betaine in B. subtilis was found to be osmotically stimulated and was strongly decreased in a mutant strain lacking the OpuA transport system. DNA sequence analysis revealed that the components of the OpuA system are encoded by anoperon (opuA) comprising three structural genes: opuAA, opuAB, and opuAC. The products of these genes exhibit features characteristic for binding protein-dependent transport systems and in particular show homology to the glycine betaine uptake system ProU from E. coli. Expression of the opuA operon is under osmotic control. The transcriptional initiation sites of opuA were mapped by high resolution primer extension analysis, and two opuA mRNAs were detected that differed by 38 base pairs at their 5' ends. Synthesis of the shorter transcript was strongly increased in cells grown at high osmolarity, whereas the amount of the longer transcript did not vary in response to medium osmolarity. Physical and genetic mapping experiments allowed the positioning the opuA operon at 25 degrees on the genetic map of B. subtilis.


OpuA, an Osmotically Regulated Binding Protein-dependent Transport System for the Osmoprotectant Glycine Betaine in Bacillus subtilis*
(Received for publication, November 29, 1994, and in revised form, May 12, 1995) Bettina Kempf and Erhard Bremerz

From the Max-Planck-Institute for Terrestrial Microbiology, D-35043 Marburg, Germany
Exogenously provided glycine betaine can efficiently protect Bacillus subtilis from the detrimental effects of high osmolarity environments. Through functional complementation of an Escherichia coli mutant deficient in glycine betaine uptake with a gene library from B. subtilis, we have identified a multicomponent glycine betaine transport system, OpuA. Uptake of radiolabeled glycine betaine in B. subtilis was found to be osmotically stimulated and was strongly decreased in a mutant strain lacking the OpuA transport system. DNA sequence analysis revealed that the components of the OpuA system are encoded by an operon (opuA) comprising three structural genes: opuAA, opuAB, and opuAC. The products of these genes exhibit features characteristic for binding protein-dependent transport systems and in particular show homology to the glycine betaine uptake system ProU from E. coli. Expression of the opuA operon is under osmotic control. The transcriptional initiation sites of opuA were mapped by high resolution primer extension analysis, and two opuA mRNAs were detected that differed by 38 base pairs at their 5' ends. Synthesis of the shorter transcript was strongly increased in cells grown at high osmolarity, whereas the amount of the longer transcript did not vary in response to medium osmolarity. Physical and genetic mapping experiments allowed the positioning the opuA operon at 25 0 on the genetic map of B. subtilis.
Monitoring and adapting to changes in environmental conditions are critical processes that determine the survival of microorganisms and their successful long term competition for a given habitat. In its soil environment, Bacillus subtilis encounters often osmotic challenges due to frequent variations in the availability of water. Since the cell envelope is permeable to water, drying and wetting of the soil alters the osmotic conditions and hence triggers the flux of water across the cell membrane. Active and timely adaptation reactions are thus required to avoid cell lysis under low osmolarity or dehydration of the cytoplasm under high osmolarity growth conditions (L, 2). Despite the importance of changes in the environmental osmolarity for growth and survival of B. subtilis, the specific physiological and genetic adaptation mechanisms to such environmental challenges are in general rather poorly understood.
* This work was supported in part by the Max-Planck Society and grants from the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The exposure of B. subtilis to a hypersaline environment triggers the induction of a set of general and salt-specific stress proteins, indicating that increased salt concentration alerts the cell to adverse growth conditions (3). The expression of a number of genes in this general stress regulon is determined by the alternative transcription factor if3, which serves to control a regulatory network responsive to stationary phase signals and growth-limiting conditions (4)(5)(6). However, B. subtilis mutant strains lacking the if3 protein are not at a survival disadvantage compared with the wild type when exposed to osmotic shock or extreme desiccation under laboratory conditions (6). Therefore, it is uncertain whether members of the if3-controlled general stress regulon playa direct role in the adaptation of B. subtilis to high osmolarity environments.
A central part of the physiological response of B. subtilis to high osmolarity stress is the intracellular accumulation of inorganic and organic osmolytes that serve to counterbalance intracellular versus extracellular osmolarity and consequently help to maintain a turgor optimal for cell growth. An increase in medium osmolarity stimulates turgor-sensitive transport systems that mediate rapid accumulation of K+ in the cell, which, in turn, restores turgor and permits cell growth to resume (7,8). This initial reaction is followed by a cellular response that replaces ionic osmolytes, which are deleterious at high concentrations, with organic osmolytes, which are more compatible with the normal physiological and structural requirements of the bacterial cell (1,2). In B. subtilis proline is the predominant organic osmolyte synthesized in defined medium by cells exposed to a hypersaline environment (7). However, several hours are required to reach a proline level that is sufficient for osmoprotection, leaving the cell at a growth disadvantage in harsh high osmolarity environments (9). B. subtilis can more efficiently respond to high osmolarity by accumulating glycine betaine (9). This potent osmoprotectant is widely found in nature and has been adopted across the microbial, plant, and animal kingdoms as an effective compatible solute (L, 2). Glycine betaine can be synthesized by B. subtilis from its precursor choline or taken up directly from the environment (7,9,10). A strong increase in the growth rate and the proliferation under environmental conditions that are otherwise strongly inhibitory for B. subtilis can be attained when glycine betaine can be directly accumulated from the growth medium (7,9,10). The presence of uptake systems for glycine betaine has been reported for a variety of Gram-negative and several Gram-positive bacteria (11)(12)(13)(14)(15), but the details of such transport systems have been studied at the molecular level only in Escherichia coli and Salmonella typhimurium. Here, two glycine betaine transport systems, ProP and ProU, have been characterized (1,2,(16)(17)(18).
In B. subtilis the importance of exogenously provided glycine betaine for the efficient adaptation to a high osmolarity environment is firmly established, but the route of glycine betaine 16701 This is an Open Access article under the CC BY license. uptake in this model system for Gram-positive bacteria is unknown. We have begun to characterize the mechanisms of glycine betaine uptake in B. subtilis, and we report in this paper the identification and analysis of a binding proteindependent transport system (OpuA) for this osmoprotectant.
Growth Conditions, Media, and Chemicals-Bacteria were grown aerobically at 37°C in LB, MMA glucose (0.2%), or Spizizen's minimal medium (SMM) supplemented with 0.5% glucose, 20 mg/liter L-tryptophan, 18 mg/liter L-phenylalanine, and a solution of trace elements (29,30). The osmolarity of the various media was increased by addition of NaCI from a highly concentrated (5 M) stock solution. The osmolarity of media was determined with a vapor pressure osmometer (model 5500; Wescor Inc., Logan, UT). Expression of the opuA genes under T7</>10 control was carried out in cells grown in M9 minimal medium (29) supplemented with 0.2% casaminoacids. To select derivatives of strain MKH13 synthesizing glycine betaine transporters encoded by cloned B. subtilis DNA, we used MMA minimal agar plates containing 0.2% glucose as the carbon source, 0.8 M NaCI and 1 mM glycine betaine. Plates for the detection of extracellular a-amylase (AmyE) activity contained 1% starch in LB agar plates. Production of AmyE by B. subtilis strains was detected by flooding the colonies grown on LB starch plates with Gram's iodine stain (0.5% (w/v) iodine, 1% (w/v) potassium iodide) for 1 min and scoring for zones of clearing around the colony after decanting the stain (31). The AroI phenotype of B. subtilis strains was tested by scoring the growth of colonies on SMM minimal plates either lacking or containing 20 mg/liter L-tryptophan, 18 mg/liter L-phenylalanine, and 20 mg/liter L-tyrosine; Aror: strains can not grow on minimal plates lacking these aromatic amino acids. The antibiotics ampicillin, chloramphenicol, tetracycline, and kanamycin were used in E. coli at a final concentration of 100, 30, 5, and 15 JLg/ml, respectively. Kanamycin, tetracycline, and erythromycin were used in B. subtilis at a final concentration of 5, 10, and 0.5 ug/ml, respectively. The cyclic peptide antibiotic globomycin, a specific inhibitor of signal peptidase II (32) [ 3 5 S]dATP at a final concentration of 1 J.LCi/m!. The size of the reaction products was determined on a 4% DNA sequencing gel under denaturing conditions and visualized by autoradiography. A sequencing ladder produced by using the same primer was run on the same sequencing gel to determine the precise 5' ends of the opuA mRNAs. Transformation of competent B. subtilis cells with plasmids and linear DNA fragments was done according to routine procedures (31).
Construction ofPlasmids-A library of chromosomal DNA segments from the B. subtilis wild-type strain JH642 was prepared by cleaving chromosomal DNA with EcoRI and ligating the resulting restriction fragments into the EcoRI site in the polylinker of the lacZ a-complementing plasmid pHSG575. The DNA of the recombinant plasmids was transformed into strain DH5a, and colonies were selected on LB plates containing chloramphenicol, isopropyl-1-thio-I3-D-galactopyranoside (1 mx), and x-gal (40 JLg/ml). Approximately 90% of the obtained transformants (40,000 colonies) carried plasmids with cloned B. sub til is DNA as judged from their LacZ" phenotype. All colonies were pooled and grown for 2 h in LB medium with chloramphenicol; the plasmid DNA was then extracted and used to transform the E. coli strain MKH13. Plasmids pBKB13, pBKB14, pBKB17, pBKB18, pBKB38, and pBKB46 were constructed by deleting defined restriction fragments from the opuA+ plasmids pBKB1 and religating the plasmid backbones ( Fig. 1). Plasmids pBKB15, pBKB35, and pBKB39 were isolated by cloning appropriate restriction fragments isolated from plasmid pBKB1 into the vector pHSG575 ( Fig. 1). Plasmid pBKB33 carries the entire opuA operon on a 4-kb EcoRI-EcoRV restriction fragment ( Fig. 1) that has been cloned into the polylinker sequence of the T7</>10 expression vector pPDlOO, thus positioning opuA under T7 </>10 contro!. The same restriction fragment was inserted in the reverse orientation with respect to the T7</>10 promoter present in the vector pPD101, yielding plasmid pBKB34. Plasmid pBKB44, which expresses the opuAA+ opuAB+ genes under T7</>10 control was constructed by deleting a 636-bp NsiI fragment carrying most of the opuAC gene ( Fig. 1) from the opuA+ plasmid pBKB33. To achieve expression of the opuAA+ gene under T7 </>10 control, a 1.7-kbEcoRI-HpaI restriction fragment ( Fig. 1) was inserted into the vector pPD100, yielding plasmid pBKB43. A plasmid expressing the opuAC+ gene under T7 </>10 control was constructed by isolating a 1.4-kb ApaLI-NotI fragment from pBKB1; the overhanging ends of the restriction fragments were filled in with Klenow enzyme and then ligated into the SmaI site of plasmid pPD100. Plasmids positioning the opuAC gene under T7</>10 control (pBKB58) or aligning it in the reverse orientation with the T7</>10 promoter (pBKB57) were identified by restriction analysis. To construct an opuAA-lacZ gene fusion, a 1.3-kb EcoRI-SnaBI restriction fragment from pBKB1 ( Fig. 1) was cloned into the EcoRI and SmaI sites of the'lacZ fusion vector pJL29, yielding plasmid pBKB54. In this plasmid, the reading frames of opuAA and 'lacZ are properly aligned across the SnaBI and SmaI junction, thus generating a hybrid protein fusion, 4;>(opuAA-lacZ)hybl. The entire hybrid gene was transferred from plasmid pBKB54 on a 4.4-kb EcoRI-DraI restriction fragment into the E. coli-B. subtilis shuttle vector pRB373, which had been cut with EcoRI and SmaI; this construction resulted in plasmid pBKB56.
Transport Assays for Radiolabeled Glycine Betaine-Uptake of glycine betaine in B. subtilis and E. coli was measured using [l-14C]glycine betaine (55 mCi/mmol) as a substrate. The cells were grown to midexponential phase (A 5 7 8 = 0.15-0.5) in minimal medium with glucose as the carbon source and used immediately for the transport assay. E. coli strains were grown in MMA or MMA with 0.2 M NaCl, and B. subtilis strains were grown in SMM or SMM with 0.4 M NaC!. The uptake assay contained [l-14C]glycine betaine at a final substrate concentration of 10 JLM (5.5 mCi/mmol) in a total reaction volume of 2 m!. Samples (0.3 ml) were taken at various times and filtered through 0.45 JLm-pore-size filters (Schleicher and Schuell GmbH, Dassel, Germany). The cells were washed with 20 ml of isotonic minimal salts, and the radioactivity retained on the filters was determined in a scintillation counter. Protein concentrations were determined from total cell extracts using the Bio-Rad protein assay with acetylated bovine serum albumin as the standard. The cell extracts were prepared by passing the B. subtilis cells five times through a French press cell at 103,000 kilopascals.
Preparation of Total Cell Extracts, SDS-Polyacrylamide Gel Electrophoresis, and Immunological Detection of the OpuAA' -f3-Galactosidase Hybrid Protein-Cultures (20 ml in a 100-ml Erlenmeyer flask) of strain JH642 carrying the opuAA-lacZ fusion plasmid pBKB56 were grown overnight at 37 DC in LB medium or LB medium with 0.5 MNaCl. The optical density (A 5 7 8 ) of the cultures was determined and adjusted to A 5 7 8=5. From each culture, 2-ml portions were withdrawn, the cells were collected by centrifugation and resuspended in 150 JLl of TE (10 mM Tris-HCl, 1 mx EDTA, pH = 8.0), and 15 JLl oflysozyme (10 mg/ml in H 2 0 ) was added. The cell suspension was then incubated for 8 min at 37 DC in a water bath, 50 JLl of 4-fold concentrated sample buffer (final concentration, 0.06 M Tris, pH 6.8, 5% SDS, 10% glycerol, 3% dithiothreitol, 0.001 % bromphenol blue) was added, and the cells were lysed by incubation for 5 min at 95 DC. To reduce the viscosity of the cell extract,2 JLl of benzon nuclease (Merck) was added and incubated for 10 min at 37 DC, followed by another short (5-min) incubation at 95 DC. Aliquots of the cell extracts were then immediately loaded onto 7% SDS-polyacrylamide gels (34), and the proteins were visualized by staining with Coomassie Brilliant Blue (33). For the immunological detection of the OpuAA'-f3-galactosidase hybrid protein, total cellular extracts were electrophoretically separated on 7% SDS-polyacrylamide gels and transferred onto a sheet of Immobilon (Millipore, Eschwege, Germany). The bound proteins were probed with a rabbit f3-galactosidase antiserum, and the antigen-antibody complexes were visualized with a second goat anti-rabbit immunoglobulin G alkaline phosphataseconjugated antibody (Sigma) using 5-bromo-4-chloro-3-indolyl phosphate and nitroblue tetrazolium chloride (Boehringer Mannheim) as substrates.
Expression of the opuA Gene Products under T7 Control-Plasmids carrying various genes from the opuA operon under the transcriptional control of the T7<1>10 promoter were transformed into strain BL2l(ADE3) to selectively visualize the opuA-encoded proteins. These plasmids are pBKB33 (opuAA + opuAB+ opuAC+), pBKB44 (opuAA + opuAB+), pBKB43 (opuAA +), and pBKB58 (opuAC+). Plasmids pBKB34 and pBKB57 carrying the opuA+ operon and opuAC+ gene, respectively, improperly aligned with respect to the T7 promoter were used as controls. Cultures (20 ml in 100-ml Erlenmeyer flasks) of strain BL2l(ADE3) carrying the various plasmids were grown in M9 minimal medium with 0.2% casaminoacids to mid-log phase (A 5 7 8 = 0.5-0.7); the cells were washed with M9 minimal medium, resuspended in 20 ml of M9 minimal medium supplemented with 0.2% methionine assay medium (Difco), and grown for 1 h at 37 DC. T7q,10-mediated gene expression was initiated by adding isopropyl-1-thio-,B-n-galactopyranoside to a final substrate concentration of 1 mM, after 30 min rifampicin was added (200 JLg/ml) to inhibit the E. coli RNA polymerase. After a 1-h incubation at 37 DC, 1-ml portions of the cultures were withdrawn, the proteins were labeled for three min with [ 3 5S]methionine (final concentration, 5.2 nCi/ml), and the cells were collected by centrifugation and resuspended in 50 JLl of SDS-sample buffer. Portions (30 JLl) of the cell extracts were loaded onto 12% SDS-polyacrylamide gels, the proteins were electrophoretically separated, and radiolabeled peptides were visualized by autoradiography. To inhibit signal peptidase II, the antibiotic globomycin was added to the appropriate cultures 10 min prior to the radiolabeling of the proteins.
Computer Work-Analysis of nucleotide sequences from the opuA region and the alignment and analysis of protein sequences was performed with the Lasergene program from DNA Star (DNASTAR, Ltd., London) on an Apple Macintosh II computer. Multiple sequence alignments were done at the National Center for Biotechnology Information (NCB!) using the BLAST programs and the current versions of the available data bases (November, 1994) (35). The nucleotide sequence of the opuA region (see Fig. 3) reported in this paper has been submitted to the GenBank™IEMBL Data Bank with accession number U17292.

Cloning of the Structural Genes for a Glycine Betaine
Transport System-To clone genes from B. subtilis that code for glycine betaine transporters, we capitalized on the growth properties of the E. coli mutant strain MKH13. This strain is defective for glycine betaine synthesis and also lacks the glycine betaine transport systems, ProP and ProU (21). Therefore, it is severely impaired in its ability to cope with a high osmolarity environment and, in contrast to its parental strain MC4100, cannot grow in high osmolarity media containing the osmoprotectant glycine betaine. We reasoned that it should be possible to functionally complement the deficiency of strain MKH13 in glycine betaine uptake with the appropriate B. subtilis genes. A gene bank of EcoRI restriction fragments was prepared from chromosomal DNA of the B. subtilis strain JH642 in the low copy number cloning vector pHSG575 (Cm") and transformed into MKH13 by selecting for Cm" colonies on LB agar plates. These transformants were then replica-plated on high osmolarity minimal plates (MMA agar plates with 0.8 MNaCl) containing 1 mM glycine betaine to search for MKH13 derivatives that could grow under these selective conditions. Such strains were readily obtained, and each of the 46 MKH13 derivatives analyzed carried the same pHSG575-derived plasmid containing a 5.2-kb EcoRI restriction fragment. A restriction map of one of these plasmids, pBKB1, is shown in Fig. 1.
The ability of strain MKH13(pBKB1) to grow at high osmolarity in the presence of 1 mM glycine betaine was shown to be dependent on the presence of plasmid pBKB1 by retransformation into strain MKH13. There was no growth of MKH13-(pBKB1) in high osmolarity minimal medium (MMA agar plates with 0,8 MNaCl) lacking this osmoprotectant. Osmoprotection by glycine betaine requires the intracellular accumulation of this compound (1,2). We therefore measured the initial [1-14C]glycine betaine uptake in cultures of strain MKH13-(pBKB1) grown in low osmolarity or high osmolarity minimal media at a final substrate concentration of 10 J.LM. Glycine betaine transport activity was readily detectable in cultures of MKH13(pBKB1), and we found that this transport activity was under osmotic control ( Fig. 2A). Thus, plasmid pBKB1 encodes an osmotically controlled uptake system for glycine betaine from B. subtilis. We designate this glycine betaine transporter as OpuA (Qsmo protectant uptake) and refer to its structural genets) as opuA.
To identify the approximate position of the opuA genets) within the cloned DNA segment from the B. subtilis chromosome, we subcloned defined restriction fragments from plasmid pBKB1 into the low copy number vector pHSG575 and also constructed a number of deletion derivatives of pBKB 1 (Fig. 1). Each of these plasmids was introduced by transformation into strain MKH13, and the ability of these transformants to grow in high osmolarity minimal media in the presence of 1 mM glycine betaine was tested. The results from these complementation experiments are summarized in Fig. 1. It is apparent that a large portion of the cloned 5.2-kb EcoRI fragment is required to mediate glycine betaine uptake activity. Thus, OpuA is most likely a multicomponent glycine betaine transport system.
Mutations in opuA Strongly Impair Glycine Betaine Transport Activity in B. subtilis-OpuA mediates glycine betaine uptake in E. coli. To investigate the role of the opuA-encoded transport system for glycine betaine transport in B. subtilis, we constructed two opuA mutations on plasmid pBKB1 and then FIG. 1. Physical and genetic organization of the cloned opuA region. A map of a plasmid, pBKBl, which carries the entire opuA operon, is shown. The exact positions and transcriptional orientations of the opuAA, opuAB, and opuAC genes were inferred from DNA sequence analysis. The extent of the DNA segment present in the various deletion derivatives and subclones of plasmid pBKBl are indicated by the lines. Plasmids pBKBll and pBKB52 are deletion derivatives of pBKBl and carry, in addition, gene cartridges encoding kanamycin (han) or tetracycline (tet) resistance, respectively. The han and tet gene cartridges are not drawn to scale. Osmoprotection by glycine betaine was assayed by monitoring the growth of the E. coli mutant strain MKH13 lacking both the ProP and ProU glycine betaine transport systems and harboring the various pBKBl-derived plasmids on glucose minimal plates containing 0.8 M NaCl and 1 mM glycine betaine. Growth of the strains was scored after 3 days of incubation at 37°C.   (Fig. 1). The il(opuA::tet)2 mutation was constructed in an analogous way by removing an NsiI DNA fragment from plasmid pBKBl and inserting a tetracycline resistance gene as the selective marker, resulting in plasmid pBKB52 (Fig. 1). Both opuA mutations destroyed the plasmid pBKBl-encoded glycine betaine uptake activity in the E. coli strain MKH13 (Fig. 1). We isolated the b..{opuA::neo)l and the il(opuA::tet)2 constructs from plasmids pBKBll and pBKB52, respectively, as EcoRI restriction fragments and then used this DNA to transform the B. subtilis strain JH642 to either kanamycin resistance (b..{opuA::neo)l) or tetracycline resistance (il(opuA::tet)2). One transformant from each experiment was purified, and the proper integration of the il(opuA::neo)l (strain BKB4) or b..{opuA::tet)2 (strain BKB7) mutation into the B. subtilis genome via a double recombinational cross-over event was proven by Southern hybridization using a DNA probe derived from plasmid pBKBl (data not shown). We measured the initial uptake activity for radiolabeled [l-14 C]glycine betaine of the opuA + B. subtilis strain JH642 and its ilopuA derivatives BKB4 and BKB7 in cells grown in low osmolarity and high osmolarity media with low substrate concentration (10 J-tM). An efficient and osmotically stimulated glycine betaine transport activity was present in the wild-type strain JH642 (Fig. 2B). In contrast, both opuA mutations strongly impaired glycine betaine uptake; this is documented in Fig. 2B for the il(opuA::neo)l deletion present in strain BKB4. Thus, opuA encodes an osmotically controlled glycine betaine transport system in B. subtilis. We note that neither of the chromosomal opuA deletions present in the B. subtilis strains BKB4 and BKB7 abolish glycine betaine uptake entirely (Fig. 2B). Thus, besides OpuA, at least one additional glycine betaine transporter must exist in B. subtilis, and the pattern of glycine betaine uptake in the opuA mutants indicates that this transport activity is also under osmotic control (Fig. 2B). In the above described glycine betaine uptake experiments, the high osmolarity media were prepared by adding NaCI to the growth media. In analogous transport assays in which NaCI was replaced by an isoosmolar concentration of KCI, glucose, or maltose, glycine betaine uptake activity in strains JH642 (opuA +) and BKB4 (b..{opuA::neo)l) was stimulated to an extent similar to that shown in Fig. 2B (data not shown). Thus, stimulation of glycine betaine transport in B. subtilis growing in a high osmolarity environment is a true osmotic effect since it can be triggered with either ionic or nonionic osmolytes.
opuA Encodes a Binding Protein-dependent Transport System-To characterize the nature of the opuA-encoded glycine betaine transporter more closely, we determined the DNA sequence of a 3.4-kb DNA segment from pBKBl that covers the region necessary for glycine betaine uptake activity (Figs. 1 and 3). Analysis of the sequenced DNA segment revealed the presence of three open reading frames that are oriented in the same direction and constitute the opuA locus (Fig. 3). Downstream of opuA, a region is present that harbors a DNA structure with dyad symmetry. This inverted repeat is bracketed by runs of AT base pairs, indicating that it possibly could function as a Rho-independent bidirectional transcriptional terminator (36). The opuA locus consists of three structural genes (opuAA, opuAB, and opuAC), and their tight physical organization strongly suggests that they are genetically organized in an operon. The intergenic region between the opuAA stop codon (TAA) and the ATG start codon of the opuAB gene is only one nucleotide in length, and the ribosome-binding site for opuAB is thus present in the preceding opuAA coding region (Fig. 3). The genetic organization of the opuAB and the opuAC junction is even more tightly spaced; here, the ATG start codon of opuAC is part of the TGA stop codon for opuAB (Fig. 3). Each of the three genes is preceded at an appropriate distance by a potential ribosome-binding site, which for the opuAB and opuAC genes is entirely contained in the coding region of the preceding gene (Fig. 3).
The deduced amino acid sequences of the opuAA, opuAB, and opuAC genes exhibit features characteristic for binding protein-dependent transport systems (37,38) and, in particular, show striking homology to the components of the glycine betaine binding protein-dependent transport system ProU from E. coli (18). The opuAA gene encodes a hydrophilic protein of 418 amino acid residues tM; 46,473), and a comparison of the OpuAA protein with protein sequences present in the data libraries revealed strong sequence identities to many prokaryotic and eukaryotic proteins involved in ATP hydrolysis. Such a close relatedness in the amino acid sequence is a hallmark of the energizing components of binding protein-dependent trans- port systems (37,38). The alignment of the OpuAA protein with the corresponding polypeptide (Pro'V) from the ProU system from both E. coli and S. typhimurium is shown in Fig. 4A. Approximately 58% of the amino acid residues are identical among all three proteins, and only a single gap needs to be introduced to achieve a good alignment of the protein sequences over their entire length. Sequence conservation is particularly apparent in the N-terminal half of the OpuAA and ProV proteins and is pronounced around the Walker A and B ATP-binding motifs (Fig. 4A).
The opuAB reading frame codes for a quite hydrophobic polypeptide (M; 30,250) that is homologous to the integral inner membrane protein ProW of the E. coli ProU transport system. Analysis of the topology of the E. coli ProW protein withphoA and lacZ fusions has revealed that ProW has seven transmembrane segments with the carboxyl terminus in the cytoplasm and the amino terminus in the periplasm (39 (47% identity) over their entire length and can be aligned without introducing a single gap into the amino acid sequence. Thus, the topology of OpuAB and ProW appears to be similar. The OpuAB protein (282 amino acids) is considerably smaller than ProW (354 amino acids). Most of the reduced size of OpuAB can be ascribed to a deletion removing 55 amino acids present in the N-terminal region of ProW thought to be exposed in the periplasmic space (39).2 A small segment (amino acids 183-203) ofOpuAB displays limited homology to integral inner membrane components of other binding protein-dependent transport systems from both Gram-negative and Gram-positive bacteria (40). It has been speculated that these residues contribute to an interaction site for the ATPases of the binding protein-dependent transporters (37,38).
The last gene in the operon, opuAC, encodes a 293-amino acid residue hydrophilic protein with a predicted M; of 32,218. The OpuAC protein is likely the substrate-binding protein component for the OpuA glycine betaine transport system. Its homologue, ProX, in the E. coli ProU system is a periplasmic protein (17, 41) that is initially synthesized with an N-terminal     4. Alignment of the sequences of the components of the OpuA transport system with those ofProU.A, the amino acid sequence of the opuA-encodedATPase fromB. subtilis is compared with the homologous protein ProV from E. coli (18) and S. typhimurium (64). signa l seque nce exte ns ion (18 ). Th e first 20 amino aci ds of t he op uAC-encoded pr otein exhibit the features of a secre te d pr otein a nd show th e characte ristic signat ures of a lipoprotein signal seque nce. Th ere is a positively charged N-terminal end, followed by a highl y hydroph obic st re tch of a mino a cids and a st ring of a mino acid s (Leu-Ala-Ala-Cys) t hat conforms to the consens us seque nce recogn ized by signal peptidase II (42,43 ). As a rule, the cystei ne residue constit utes th e N terminus of t he pro teolytically processed protein a nd is modifi ed through t he coval ent attachmen t of lipids. Thi s lipid modification of t he N terminus se rves to a nchor th e ext racellula r pr otein in the cytoplasmi c membrane, and such lipoprotein s ha ve been described as su bs t ra te -binding protein s for a number of binding protein -dep endent t ran sp ort sys te ms in Gra m-posi tive bacteri a (42,43). Th e subs t rate bindin g protein s, ProX a nd OpuAC, show th e least seque nce conserva tion (33% identity in a 46a mino acid segment) a mong t he components of the ProD a nd OpuA transport sys te ms (Fig. 4C). Only a limited number of residues in the cen t ral part of the OpuA C a nd ProX pr otein s ca n be aligned, wh er eas the N-terminal a nd C-te r mina l ends of both protein s a re entire ly different (Fig. 4C ). Identification of th e op uA Gene Prod ucts-To iden tify th e pr otein s encoded by th e op uA operon , we used th e T7 RNA polym erase a nd T7 eJ>10 pr omoter expression sys te m (22,24). We constructed a set of T7 expression plasmids carrying eit he r th e entire op uA operon (op uAA, op uAB, op uAC; pBKB 33), th e fir st t wo st ruct ural genes (op uAA , op uAB; pBKB 44 ), or just the first gene of th e op uA locus (op uAA ; pBKB43). Th ese pla smids were transformed into th e E . coli strain BL2 l(ADE 3), which ca rries a chromosomal copy of t he st r uctura l gene for t he T7 RNA polym erase under the control of t he lacPO regulatory sequences (22). T7 eJ> 1O pr omoter-mediated expression of the va rious op uA-encoded genes was initiated by adding isopropyl-1-thio-(3-D-gal actopyranosid e to the cult ures, and th e translated protein s were t he n lab eled with [ 3 5 S1methionine . We were able to express the op uA -encoded pro tein s under T 7eJ>10 control in E. coli, but many degr ad a tion products of th e pr otein s were visible (Fig. 5A ), indica ti ng th at th e OpuA pr otein s from B. subtilis were relatively un stabl e whe n pr oduced in th e heterologou s E . coli host . Such pr otein in stability has a lso been obse rve d whe n compone nts for t he binding protein-depend en t iron-hydroxam ate t rans port system from B . subtilis were expr essed in E . coli under T7 eJ>10 control (44). A compa riso n of t he plasmid pBKB 33-, pBKB 44-, a nd pBKB 43-en coded protein s a llowed us to visualize a nd ide ntify th e compone nts of t he OpuA system . Th e opu A + plasmid pBKB 33 direct ed t he synt hesis of three pr oteins wit h a a ppa re nt molecul ar mass of 47 ,000 dal tons (OpuAA), 24,000 daltons (OpuAB), a nd 30,500 dal tons (OpuAC). Th e 30,500-da lto n protein wa s abs ent whe n t he op uAA + op uAB+ plasmid pBKB 44 was used to mediate gene expression, thus iden tifyin g t his polyp eptide as the product of t he op uAC gene . Th e sa me protein is synthesized in st rain BL2l(ADE3 ) carryin g just th e op uAC gene under T7eJ> 1O cont rol on plasmid pBKB 58 (Fig. 5B , lan e 5 ). A 47,000 -da lto n protein was pr odu ced in cells expressing only op uAA from plasmid pBKB43, (Fig. 5A, lane 3 ), t hus iden tifyin g t his polypep tide as t he OpuAA pr otein . Both th e 47 ,000-d al ton protein (OpuAA) a nd t he 24,000-da lton pr otein were sy nthesized wh en the op uAA+ op icAb " genes (plasmid pBKB 44) wer e expressed under T7 eJ>10 control (Fig. 5A, lane 2 ). Thus, t he 24 ,000-dal ton pr otein mu st be OpuAB . None of th ese op uAencoded pr otein s were pr odu ced whe n a n op uA + -containing restriction fragm ent was clone d into t he T7 expression vecto r in an orie ntation reversed with resp ect to that presen t in plasmid pBKB 33 ( Fig. 5A ; comp are lan es 1 a nd 4 ). Th e apparen t molecul ar masses of the OpuAA (47,000-da lton) a nd t he OpuAC (30,500-dalton) pr otein s comp are fa vorabl y wit h t he molecul ar masses deduced for OpuAA (46,4 73 daltons) a nd OpuAC (30,235 dal tons for th e proteolyti call y processed but unmodified polyp ep tid e) from th e DNA seque nce (Fig. 3) of th eir st ruct ura l genes. In contrast, the appa re nt molecular ma ss of th e OpuAB pr otein (24,000 dal tons) as calculated from its elect rophoretic mobili ty on a 12% SDS-polyacrylamide gel, deviates cons iderably from t he molecular mass predi cted for this pr otein from th e op uAB DNA seque nce (30,250 dalt ons). Th e OpuAB pr otein cons tit utes a quite hyd roph obic in tegr al membrane pr otein, and its a ppa re nt electrophore tic behavior is therefore not too surprising .

O V E A G P M W T A I A T G S A D A S L S A W L P N T H K A Y A
OpuAC Probably Is a Lipoprot ein-A s outli ne d above, t he OpuAC pr otein is likely to carry lipid modifications at its N terminus, a ncho ring it in the membran e. On e characteristi c fea ture of such lip opr otein s is t he inhibition of th e proteolyti cal pr ocessing of the ir signal seq ue nce by th e cyclic pep tide a ntibiotic globomyci n (32,45 ). To test whethe r OpuAC is ind eed a lip oprotein , we ex pressed t he entire op uA operon a nd th e opuAC ge ne a lone under T7 </>1O control in t he pr esenc e or abse nce of globomycin . Th e pr esenc e of globomycin inhibited complete ly t he processing of th e OpuA C pro tein a nd resulted in the accumulation of its precursor molecul e, pro-OpuAC (Fig.  5B ). In contrast, globomy cin had no influenc e of the elect rophoretic mobility of the OpuAA a nd OpuAB pr otein s (Fig. 5B,  lan es 1 a nd 2 ). Thus, th e selective block impose d by globomycin on pro-OpuAC pr ocessing strongly indicates th a t OpuAC is a lip opr otein with a n a mino -te r minal cysteine-lipid a nchor for th e ma ture pr otein .
Osm oregulat ion of op uA Exp ression-T he OpuA-medi ated glycine beta ine t ra ns port activity is osmotically modulated (Fig. 2). To test wh eth er this was du e (at least in part) to osmotic control of opuA t rans cription a nd to iden tify t he op uA promoterts ), we mapped t he t ra ns criptio n initiation sites by primer extensi on analysis. A 1.3-kb EcoRI-SnaBI restricti on fragment ca rryin g t he op uA up stream region a nd most of the opuAA coding seque nce was used to construct a (1)(opuAA-lacZ )hybl pr otein fus ion in t he E. coli-B. subti lis sh ut tle vector pRB 373. Th e resul ting plasmid, pBKB 56, was tran sform ed into t he B . subtilis st rai n JH642 to increase t he gene dosage for the op uA regul atory region . Total RNA was th en pr epared from log-ph ase cultures gro wn eithe r in LB medium or in LB medium with increased osmola ri ty (LB + 0.5 M Na C!). An op uAspecific primer was annealed to t he RNA isola te d from th e low a nd high osm olarity gr own cells a nd exte nded with avian myeloblastosis vir us rever se transcriptase in the pr esenc e of [ 3 5 SJdATP . Th e reaction pr odu cts were th en se pa rated on a DNA seque ncing gel a nd visualize d by a utoradiography . Two opuA-specific mRNA species were det ect ed that differed in size (38 bp ) at th eir 5 ' ends (Fig. 6C ). Syn th esis of the shorte r transcript (mRNA-1) is under osmotic cont rol, a nd the amount of t his mRNA increases st rongly in high osmola rity grown cells . In contrast , pr oduction of th e longer t ransc ript (mRNA-2) was not influenc ed by th e osmola rity of t he gro wth medium (Fig.  6C). Thus, expression of th e op uA operon is medi ated by two promoters that resp ond differently to changes in medium osmolarity. Inspection of t he DNA seque nce up stream of the ini ti ation sites of mRNA-1 a nd mRNA-2 reveal ed th e pr esen ce of pu tative -35 (consens us seque nce: TTGA CA) a nd -10 (conse ns us seque nce: TATAAT) seque nces that could possibl y constitute pr omoters recogni zed by a form of RNA polym erase complexe d with th e main vege tati ve (J" fact or «,A) of B. subtilis (Fig. 6A ) (46 ).
We mon it ored th e influenc e of medium osm olarity on op uA expression with the a id of th e < 1>(opu AA-lacZ ) hybl hybrid gene pr esent on plasmid pBKB 56. Th is protein fusi on encodes a hybrid pr otein th a t car ries at its a mi no terminus a large segment of the OpuAA pro tein a nd at its carboxyl terminus a n a lmos t compl ete J3-galactosidase. To test wh ether sy nt hesis of th e OpuAA '-J3Ga l hybrid protein was under osmo ti c control, we gre w th e B. subtilis strai n JH642(pBKB 56) overnight in LB medium a nd LB medium with 0.5 M Na Cl, pr ep ared tota l cell ext racts, a nd se pa rated t he pr otein s elect rophore tically on a 7% SDS -polyacrylam ide gel (Fig. 7A ). Cons iste nt with the influenc e of medium osmola ri ty on op uA-directe d transcription , we det ected a strong increase in the producti on of the hybrid OpuAA '-J3Gal pr otein in hi gh osmola rity grown cells. Thi s hybrid pr otein cross-reacted with an a ntise ru m direct ed agains t J3-gal actosid ase (Fig. 7B). We obse rve d t hat increased sy nt hesis of th e la rge OpuAA '-J3Gal hybrid protein resulted in the form a-A Eco R I TAC AATC ATATAG GAGGAT TAC AGAG CATT TAGAAGC ATAAATAAGAT CAT GT GG TCAC   ·35  · 10  -35 AT GGA TGTTTATAAAGAAATGGTACAGAATAAAAGAG AATA TGCTGTTTGTGTGGGAAG -10 cbs m~A-2 for th e opuA mRNA-l are in dicated by arrows. B , compa rison of th e pro mote r regions of the osmoregulated proU an d proP from E. coli with tho se of the osmotically controlled opuA P I pr omoter from B. subtilis. C, ma ppi ng of th e start sites for the opuA mRNAs. Total RNA was prepared from cells of strain JH642(pBKB56) grown in LB or LB with 0.5 M NaGl, hybridized to a prim er complementary to the op us: mRNA , a nd extended wit h re vers e tra nsc riptase in the presen ce of ra diolab eled [ 3 5 S1dATP. DNA seque ncing reactions primed wit h the same synthetic oligonucleotide used for the primer exte ns ion reaction wer e employed as a standard to size the opuA mRNAs. tio n of in solubl e aggregates, whi ch displayed no J3-galactosi dase activity . Th e protein from the E. coli Pro U sys te m, P roV, analogou s in function to OpuAA from B. subtilis , is a membr an e-a ssociated protein , a nd th e clumpin g of a ProV'-J3Ga l hybrid pr otein h as a lso been repor ted (47).

T T AC A T A A A TG T TACGGT A AT A A A G A T T GC T T A A T AT GGAGGG A A A A~G T GT A..
Phy sical and Genetic Mapp ing of th e opu A Operon-A comparison of th e nu cleotide seque nce of the 3.4-kb op uA region (Fig. 3) to th e DNA data base revealed two short DNA sequ en ces of 81 a nd 31 bp t hat matched (with th e exceptio n of a si ngle mi sm a tch in ea ch DNA segment) t he DNA seq ue nce from 2910 -2991 bp a nd 333 9-33 70 bp , resp ectively (Fig. 3). Th ese matchin g seque nces a re located up stream of t he amyE gene at 25 0 on t he B. subtilis genetic map (48). Th ey represent j unction points of re peati ng un its a mplified in a mu tant st rai n showing hyperp roduction of an extrace llula r a-amylase (AmyE ) a nd increased resistance to t he a ntibiotic tun icamycin (tm rB ) (49,50). Th e iden ti fica tion of t hese junction point sequ en ces within th e op uA regio n suggested that t he op uA OpuAA'.pGal operon is located in the vicinity of the amyE gen e. We the re fore carried out both ph ys ica l and gene t ic mapping expe rime nt s to te st this ass u mption a n d to position the op uA operon on the B. subtilis gene t ic map. Using a DNA probe covering op uA (Fig. BA), we performed South ern hybridization expe rime nts with chromosomal DNA prepa r ed from strains JH642 (op uA + amyli"), BKB4 (!:l(opuA ::neo)] amyli" ), an d M01099 (op uA +~(amyE : : ery )) . As expe cte d , the opuA probe hybridized to a sing le 5.2·kb EcoR I fragment in chr omosoma l dig ests of the opuA + strains JH642 and M01099 but r ecognized a smaller EcoRI restriction fra gment (4 kb ) in the chromosomal digest of the~(opuA : :neo )l st r ain BKB4 (Fig. 8B ). Thus, the DNA probe used detects s pecifica lly the opuA region in the B. subtilis genome . Two closely s paced PstI restriction sit es are pre sent in the amyE gen e (F ig. BA), both of which wer e re moved during the construction of the MamyE::ery ) mutation (19 ) (Fig. BA). Th e op uA DNA probe det ect ed an approximately 12·kb PstI r estriction fragm en t in a chromosomal digest of strain JH642 (a nd a correspondingly sm alle r 10.8·kb r estriction fragm ent from strain BKB4), but an approximately 20-kb Pst I fragment was found in the~(amyE ::ery ) st r ain M01099 (Fig. 8B ). The larger size of the hybridizing chromosomal PstI restriction fragmen t from st rain M01099 results from the fusion of two adjacent Pst I fragments (Fig. BA). Taken together, thes e data sh ow that opuA and amyE are physically locat ed close to on e another . Th e amyE gen e h as be en positioned by DNA hybridiza tion next to t he end of a NotI restriction fragment on t he physical map of t he B. subtilis chromosome (51). Consistent with this previou s re port, we found a N ot I re striction site down st r eam of the op uA operon (Fig. 1).
The linkage between opuA and amyE wa s also apparent wh en we performed a genetic mapping experiment using the DNA transformation technique. Ch r omosom al DNA from   anchored via lipid modifications in the cytoplasmic membrane can serve the physiological function of periplasmic proteins from Gram-negative bacteria (42,43). The components of the OpuA transport system show homology to those from the binding protein-dependent glycine betaine uptake system ProU from E. coli (Fig. 4). The amino acid sequence of the ATPases (OpuAA and ProV) and the integral inner membrane components (OpuAB and ProW) from both systems show extensive identity, but the substrate-binding proteins from the OpuA and ProU systems are not well conserved (Fig. 4). Such low level conservation of the ligand-binding proteins has also been observed for several other pairs of ABC transporters of Gramnegative and Gram-positive microorganisms (54). The OpuAC protein is essential for the OpuA-mediated glycine betaine transport in B. subtilis (Fig. 1), and its processing is inhibited by globomycin (Fig. 5B), indicating that OpuAC is a lipidmodified and extracellular substrate-binding protein. The overall organization and subunit composition of the B. subtilis OpuA system is shown in Fig. 9 and is compared with its counterpart, ProU, from the Gram-negative bacterium E. coli.
The ll(opuA::neo)l mutation (Fig. 1) present in strain BKB4 removes both the genes for the substrate-binding protein (OpuAC) and the integral inner membrane component (OpuAB), thus inevitably destroying the functioning of the OpuA system entirely. A strongly reduced glycine betaine transport at low substrate concentration (10 /LM) reflects the loss of the OpuA-mediated transport activity (Fig. 2B). However, the presence of the ll(opuA::neo)l deletion does not completely abolish glycine betaine uptake and hence uncovers the existence of a second transport pathway for this osmoprotectant in B. subtilis. This second glycine betaine transport system is under osmotic control (Fig. 2B), implying that it is also involved in the defense against the deleterious effects of high osmolarity.
Experiments with a opuAA-lacZ protein fusion showed that the amount of OpuA is responsive to changes in the osmotic strength of the environment. High osmolarity growth conditions induce opuA expression, and two differently regulated promoters direct the transcription of the opuA operon; one is OpuA B. subtilis: E. coli: ProU 1M OM strain BKB4 (ll(opuA::neo)l amyli" arol") was used to transform the B. subtilis strain TIBS57 (opuA + amyE3 aroIlO) to kanamycin resistance. This latter strain was used as the recipient because it carries both an amyE mutation and an alteration in the aroI, gene which is closely linked to the tmrB locus (Fig. SA). Transformants of strain TIBS57 were selected on LB agar plates containing kanamycin and were then tested for both their AmyE phenotype on starch-containing agar plates and their AroI phenotype on minimal plates lacking the aromatic amino acids Tyr, Phe, and Trp. From 197 kanamycinresistant transformants characterized, 176 (89%) were found to be Amyli;" and Arol ", attesting to the tight genetic linkage between the opuA operon and the amyE gene. Consistent with the expected greater genetic distance between opuA and aroI (Fig. SA), only a minor portion (21 of 197) of the kanamycinresistant transformants of strain TIBS57 had acquired simultaneously both the amyli" and arol" genes from strain BKB4. Taken together, these physical and genetic mapping experiments allow us to unambiguously position the opuA operon at 25°on the B. subtilis genetic map, and we conclude from the physical map of opuA that this operon is transcribed in a clockwise fashion on the B. subtilis chromosome.

DISCUSSION
The uptake of glycine betaine confers a high level of osmotic tolerance in B. subtilis and thus is an important facet in the stress response of this soil microorganism to high osmolarity (9). Glycine betaine is a preferred osmoprotectant in B. subtilis because the endogenous accumulation of proline is strongly reduced under high osmolarity growth conditions when glycine betaine is present in the culture medium (7). The data presented here show that glycine betaine transport in B. subtilis is under osmotic control and involves at least two transport systems. We have characterized one of these transporters in some detail and identified it as a multicomponent, binding proteindependent transport system, OpuA.
Bacterial binding protein-dependent transport systems are members of a superfamily of prokaryotic and eukaryotic transporters, known as ATP-binding cassette (ABC) transporters or traffic ATPases (37,38). These transporters couple hydrolysis of ATP to nutrient and ion uptake or to the translocation of drugs, polysaccharides, peptides and proteins across biological membranes. Because the substrate for the opuA-encoded ABC uptake system is metabolically inert in B. subtilis and serves an osmoprotective function (9), OpuA can be classified as part of the cellular defense machinery that permits this soil microorganism to cope with high osmolarity environments. Binding protein-dependent transport systems exhibit a very high affinity toward their substrate and can mediate unidirectional solute accumulation against a steep concentration gradient (37,38). Consequently, transporters such as the OpuA system are especially well suited to scavenge their substrate effectively from the environment even when it is present at a very low concentration and still attain a high intracellular level of the transported compound. Glycine betaine is synthesized by plants (52) and is brought in a varying supply into the habitat of B. subtilis through the degradation of plant tissues, thus necessitating effective mechanisms for the active acquisition of this important osmoprotectant.
Characteristic for the binding protein-dependent transport systems of Gram-negative bacteria is the presence of a soluble, ligand-binding, periplasmic protein that serves to capture the substrate and deliver it to the membrane-bound components. Binding of glycine betaine to the periplasmic ProX proteins (Fig. 9) from E. coli and S. typhimurium has been demonstrated directly (17,41,53). Since Gram-positive bacteria have no periplasm, it has been proposed that extracellular proteins osmotically controlled (opuA pol), whereas the second (opuA P-2) does not respond to the osmotic stimulus (Fig. 6). The putative -10 and -35 regions of the opuA P-1 and opuA P-2 promoters show homology (Fig. 6A) to the consensus sequence of aA-dependent promoters (46), and hence both promoters are likely transcribed by an RNA polymerase complex containing the main vegetative a factor (aA). The alternative transcription factor vB is an important regulatory element for a large network of stress proteins of B. subtilis whose synthesis increases after exposure of the bacterial cell to salt (4)(5)(6). We have tested glycine betaine uptake in several sigB mutants and found no difference from their sigB+ parents, indicating that vB does not play a central role in the regulation of the glycine betaine uptake systems of B. subtilis.' The distance of 17 bp between the -35 and -10 boxes in the opuA P-2 promoter matches the ideal distance between -35 and -10 regions of aA-dependent promoters, whereas the osmoregulated opuA P-1 promoter has a suboptimal spacing of 18 bp (Fig. 6A). Although both opuA promoters can direct the synthesis of substantial amounts of mRNA (Fig. 6C), they do not conform closely to the -35 and -10 consensus sequences of aA-dependent promoters (Fig. 6, A  and B). In particular, the osmoregulated opuA P-1 promoter is unusual since it contains in its -10 region a string of three consecutive GC base pairs. Interestingly, both the osmoregulated proU and proP promoters from E. coli also exhibit -10 regions rich in GC base pairs (Fig. 6B), and each of these promoters contains a TG motif characteristic for an extended -10 region that can partially compensate for inefficient -35 regions (55,56). A point mutation in the E. coli proU -10 region altering one of the GC base pairs to an AT base pair increases the basal level of proU expression at low osmolarity but does not alter its osmotic regulation (47). It is thus likely that the unusual -10 region of the B. subtilis opuA P-1 promoter makes an important contribution to the low basal level of the opuA P-1 transcript in the absence of osmotic stress (Fig. 6C).
In addition to their unusual -10 regions, both the osmotically regulated proU and opuA P-1 promoters deviate in the length of their spacer regions between the -35 and -10 sequences from the consensus 17-bp distance and contain suboptimal spacings of 16 and 18 bp, respectively (Fig. 6B). Expression of the osmotically regulated proU operon from E. coli is sensitive to changes in DNA topology (57,58). RNA polymerase appears to make specific contacts with both the -10 and -35 regions, and the relative orientation of these sequences is an important determinant for the efficiency of transcription initiation (59). Promoters with a 16-or 18-bp spacer sequence might therefore respond sensitively to environmentally controlled alterations in DNA topology, and both the E. coli proU and B. subtilis opuA P-1 promoters might thus be members of a class of DNA twist-sensitive promoters (60).
DNA sequences located upstream and downstream of the osmoregulated E. coli proU promoter and the nucleoid-associated DNA binding protein H-NS and HU contribute to the finely tuned genetic control of proU expression in response to changes in medium osmolarity (25,58,(61)(62)(63). Our identification of the osmoregulated opuA P-1 promoter from B. subtilis is an important first step in identifying the DNA sequences required in cis to mediate osmotically controlled transcription and in unravelling the signal transduction pathway that allows B. subtilis to sense changes in the environmental osmolarity and convert this information into a genetic response that finally leads to increased opuA expression.