Purification and Characterization of the DeoR Repressor of Bacillus subtilis

Abstract

Transcription of the Bacillus subtilis dra-nupC-pdp operon is repressed by the DeoR repressor protein. The DeoR repressor with an N-terminal His tag was overproduced with a plasmid under control of a phage T5 promoter in Escherichia coli and was purified to near homogeneity by one affinity chromatography step. Gel filtration experimental results showed that native DeoR has a mass of 280 kDa and appears to exist as an octamer. Binding of DeoR to the operator DNA of the dra-nupC-pdp operon was characterized by using an electrophoretic gel mobility shift assay. An apparent dissociation constant of 22 nM was determined for binding of DeoR to operator DNA, and the binding curve indicated that the binding of DeoR to the operator DNA was cooperative. In the presence of low-molecular-weight effector deoxyribose-5-phosphate, the dissociation constant was higher than 1,280 nM. The dissociation constant remained unchanged in the presence of deoxyribose-1-phosphate. DNase I footprinting exhibited a protected region that extends over more than 43 bp, covering a palindrome together with a direct repeat to one half of the palindrome and the nucleotides between them.

In Bacillus subtilis, the dra-nupC-pdp operon encodes three enzymes required for deoxyribonucleoside and deoxyribose utilization (12). Expression of the operon is induced by deoxyribonucleosides and deoxyribose. Transcription of this operon is negatively regulated by the DeoR repressor protein, which is encoded by the deoR gene located immediately upstream of the operon (12, 14). DeoR regulates the expression of the dra-nupC-pdp operon by binding to an operator sequence located in a region corresponding to −60 to −22 bp relative to the transcription start point (14). This site contains a palindromic sequence in the region from −60 to −43 bp and a direct repeat to the 3′ half of the palindrome located between the −35 and −10 regions. Previous studies with crude DeoR show that both the palindrome and the direct repeat are necessary for DeoR regulation of dra-nupC-pdp operon expression (14). Both deoxyribose-5-phosphate (dRib-5-P) and deoxyribose-1-phosphate (dRib-1-P) are suggested to be internal inducers for the expression of the operon, but dRib-5-P seems to be the preferred inducer (14).

In Escherichia coli, the expression of the deo operon is negatively regulated by the DeoR repressor protein and dRib-5-P is the effector molecule (1, 2, 9). B. subtilis DeoR shows no amino acid sequence similarity to E. coli DeoR, which belongs to the LacI-GalR family. Furthermore, there is no similarity in the DNA operator sites for these two repressors (14). In the present work, we describe the purification of DeoR of B. subtilis and show that the native DeoR repressor protein most likely exists as an octamer in solution. We also report the specific binding of DeoR to the operator DNA of the B. subtilis dra-nupC-pdp operon.

MATERIALS AND METHODS

Bacterial strains, plasmids, and media.

The bacterial strains and plasmids used in this work are listed in Table 1. B. subtilis was grown in Spizizen's salt-containing minimal medium (13) supplemented with 50 μg of l-tryptophan per ml and with 0.4% succinate as a carbon source. L broth (Difco Laboratories, Detroit, Mich.) was used as a rich medium for both E. coli and B. subtilis. Culturing of cells was performed at 37°C. For selection of antibiotic resistance, the following antibiotics and concentrations were used: ampicillin, 100 μg/ml; neomycin, 5 μg/ml; erythromycin, 1 μg/ml; lincomycin, 25 μg/ml; and phleomycin, 1 μg/ml. dRib-5-P and dRib-1-P are from Sigma.

TABLE 1.

Bacterial strains and plasmids used in this study^a

Bacterial strain or plasmid	Relevant genotype or description	Source or reference
Bacterial strains
B. subtilis
168	trpC2	C. Anagnostopoulos
XM15	trpC2 amyE::dra-lacZ	14
XM25	trpC2 amyE::dra-lacZ deoR::erm	14
XM251	trpC2 amyE::dra-lacZ deoR::erm pXM1000	Transformation of XM25 by pXM1000, Plr
E. coli
JOY999	TG1(pQE-30)	This work
JOY1000	TG1(pJOY1000)	This work
TG1	Wild type; lacIq	Laboratory stock
Plasmids
pEB112	Apr (E. coli) Plr (B. subtilis); multiple copy shuttle vector containing the pBR322 rep. origin for replication in E. coli and pC194 rep. origin for replication in B. subtilis	6
pJOY1000	BamHI-HindIII PCR fragment containing deoR generated by primers S3 and S4, ligated to pQE-30 digested with BamHI and HindIII	This work
pXM1000	PstI-HindIII PCR fragment containing deoR generated by primers S4 and S5, ligated to pEB112 digested with PstI and HindIII	This work
pQE-30	Apr, has a promoter and operator element consisting of the E. coli phage T5 promoter and two lac operator sequences, used for overexpressing deoR	Qiagen
pHH1002		12

^aAp^r, ampicillin resistance; Pl^r, phleomycin resistance; rep., replication. 

DNA manipulations and genetic techniques.

Plasmid DNA was isolated by the alkaline-sodium dodecyl sulfate method (13). Transformation of E. coli and B. subtilis was performed as previously described (13). Treatment of DNA with restriction enzymes and T4 DNA ligase was performed as recommended by the supplier. A standard PCR was performed as described previously (14).

Construction of plasmids and strains.

The deoR gene was amplified by PCR using plasmid pHH1002, which carries deoR (12). The forward and reverse oligonucleotide primers were synthesized with BamHI and HindIII 5′-linked restriction sites, respectively (Table 2). The PCR product was digested with BamHI and HindIII and then ligated to BamHI- and HindIII-digested plasmid pQE-30, generating pJOY1000. The E. coli TG1 strain harboring pJOY1000 or pQE-30 is designated strain JOY1000 or JOY999, respectively. For in vivo complementation, the deoR gene with six histidine codons at the 5′ end from JOY1000 was amplified by PCR using plasmid pJOY1000 as template DNA. The forward and reverse oligonucleotide primers were synthesized with PstI and HindIII 5′-linked restriction sites, respectively (Table 2). The PCR product was digested with PstI and HindIII, ligated to PstI- and HindIII-digested plasmid pEB112, and transformed into E. coli TG1, selecting for ampicillin resistance. Plasmid extracted from E. coli was transformed into B. subtilis XM25 by selecting for phleomycin resistance, yielding XM1000 (Table 1).

TABLE 2.

Oligonucleotides used for the PCR amplifications

Primer	5′- or 3′-linked restriction site sequence	Nucleotide sequencea	Coordinatesb or source
Amplification of deoR
1	5′ BamHI	5′-CGCGGATCCATGGATCGGGAAAAACAG-3′	4052107–4052088
2	5′ HindIII	5′-GCCGAAGCTTTCACAAATCATTAACAAG-3′	4051166–4051187
3	5′ PstI	5′-GAACTGCAGATTAAAGAGGAGAAATTAAC-3′	Qiagen
Mobility shift assay
4	5′ EcoRI	5′-GCCGGAATTCGTGACACGTTCAAACCTT-3′	−80
5	5′ KpnI	5′-GCCGGGTACCATCCTTCGCACACTTCC-3′	+30
6	5′ KpnI	5′-CGGCGGTACCCTTTTGAACATATGTAAATTGGTAATTG-3′	−19
7	5′ EcoRI	5′-GCCGGAATTCTTCAATTACCAATTTACATATG-3′	−48
8	5′ KpnI	5′-CGGCGGTACCCATATGTAAATTGGTAATTG-3′	−27
9	5′ EcoRI	5′-GCCGGAATTCCCTTTCATTGAACAAAATTTCAATTACC-3′	−66

^aItalic letters indicate nucleotides of the linker sequences. Underlined letters indicate nucleotides of the restriction site sequence. 

^bThe number indicates the position of the 5′-proximal nucleotide (14) of the primer except for primers 1 and 2, for which the nucleotide numbering of genome sequence (5) has been used. 

Expression and purification of the DeoR repressor protein.

E. coli strain TG1 bearing pJOY1000 was grown in 3 liters of Luria broth. After the optical density at 600 nm reached 0.5, the culture was induced with 2 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 4 h. All the cells from the 3-liter cultures were harvested by centrifugation and stored at −80°C.

All purification procedures were performed at 4°C. The cells were resuspended in sonication buffer (50 mM sodium phosphate [pH 7.8], 300 mM NaCl) and disrupted by sonication on ice, and cell debris was removed by centrifugation. Streptomycin sulfate (0.11 volume of a 10% solution freshly prepared in sonication buffer) was added, and the precipitate was removed by centrifugation. The solution was dialyzed against sonication buffer. The entire sample was loaded onto an Ni-nitrilotriacetic acid (Ni-NTA) agarose column that had been equilibrated in sonication buffer. Ten column volumes of sonication buffer was allowed to flow through the column, and then 10 column volumes of washing buffer (50 mM sodium phosphate [pH 6.0], 300 mM NaCl, 10% glycerol) were allowed to flow through the column. DeoR was eluted using a 250-ml linear imidazole gradient from 100 to 500 mM in wash buffer. Fractions from the trailing half of the DeoR peak, which eluted at a conductivity equivalent to around 200 mM imidazole, were pooled. The pooled DeoR sample was then dialyzed against wash buffer containing 0.2 M imidazole (0.2 M imidazole was included to prevent precipitation of DeoR) and frozen at −80°C in 50-μl aliquots. Approximately 40 mg of DeoR was purified from 3 liters of culture.

Gel filtration analysis of DeoR.

A column (1-cm diameter, 95-cm height, and 75-cm³ bed volume) of Sephadex G-150 (Pharmacia Biotech Inc.) was used to determine the native molecular weight of DeoR. The buffer used contained 50 mM sodium phosphate (pH 6.0), 300 mM NaCl, 0.2 M imidazole, and 10% glycerol. The column was loaded with 0.5-ml samples of DeoR (concentrations varied from 2.5 to 5.0 mg/ml) and eluted at 4°C. Proteins used to construct an M_r standard curve for the column were myoglobin, chicken serum albumin, yeast hexokinase, and bovine gamma globulin. The protein concentrations in the eluted fractions were determined from their absorbance at 280 nm (A₂₈₀).

β-Galactosidase assay.

β-Galactosidase activity was measured by the method of Miller (8). Specific enzyme activities were expressed in units per milligram of protein. One unit is defined as 1 nmol of substrate converted per minute. The values shown are the means of at least two different experiments. The variation was less than 10%. The concentration of total protein was determined by the method of Lowry et al. (7).

Mobility shift assay.

The standard PCR mixtures containing 25 μmol of [α-³³P]dATP (25 μCi) were used to produce the radiolabeled operator-containing DNA probes. The following labeled fragments were generated: primers 4 and 5 (111-bp product), 6 and 7 (42-bp product), 7 and 8 (34-bp product), and 6 and 9 (30-bp product). Each binding reaction mixture contained 10 mM Tris-HCl, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol (pH 7.5), double-stranded poly(dI-dC) (1 U/ml) [1 U of poly(dI-dC) is 1 A₂₆₀ unit in a 1-cm light path], 250 μg of bovine serum albumin per ml, and 5% glycerol in a final volume of 10 μl. Approximately 10 to 100 pM of labeled DNA probe and various concentrations of DeoR were used in each binding reaction mixture. For the binding stoichiometry experiment, in addition to the labeled DNA, 1 μM nonradioactive DNA of the same fragment was added to each binding reaction mixture. After incubation for 20 min on ice, samples were loaded onto a 5% polyacrylamide gel and electrophoresed at 7 V/cm for 2 h at 4°C. Dried gels were visualized and quantitated with a Packard Instant Imager. Apparent K_d values were calculated from isotherms of free DNA at various repressor concentrations according to the Hill equation. The repressor concentration was calculated on the basis of the 35-kDa subunit.

DNase I footprinting.

The 111-bp DNA probe used for DNase I footprinting was similar to the probe used for gel shift assay except that a single strand was ³²P labeled at its 5′ end by T4 kinase. The DNA fragment was incubated with DeoR as described above for the mobility shift assay. For DNase I digestion, 1 μl of 50 mM CaCl₂ was added to the 10-μl DNA-protein mixture, followed by the addition of 1 U of DNase I. Digestion was stopped after 5 min on ice by the addition of 10 μl of a stop solution (200 mM NaCl, 30 mM EDTA, 0.1 μg of yeast tRNA per μl). Samples were precipitated with ethanol on dry ice for 20 min and centrifuged. Precipitated DNA was washed with 70% ethanol, dried, taken up in formamide sequencing gel buffer, and electrophoresed on an 8% polyacrylamide sequencing gel alongside a Maxam-Gilbert A+G sequencing ladder (11) for the same fragment.

RESULTS

Overexpression and purification of DeoR.

The B. subtilis deoR gene was cloned into pQE-30 to generate plasmid pJOY1000, in which the expression of deoR was driven from the E. coli phage T5 promoter containing two lac operator sequences, so that production of DeoR was induced by IPTG. His-tagged DeoR was overproduced in E. coli strain TG1 and purified in a single step by Ni-chelate affinity chromatography as described in Materials and Methods. The His-tagged DeoR protein was purified because the native DeoR protein was refractory to purification. Although DeoR was an abundant protein in cells following overexpression, a substantial fraction was insoluble. Nevertheless, DeoR comprised a significant fraction of the proteins in the soluble cell extract (Fig. 1, lane 2). Extract proteins were absorbed to Ni-NTA-agarose, and the repressor was eluted by approximately 0.2 M imidazole. Fractions from the trailing half of the DeoR peak were pooled to yield a preparation exceeding 95% homogeneity (Fig. 1, lanes 5 and 6). The yield was approximately 40 mg of purified protein from 3 liters of E. coli culture.

FIG. 1.

Purification of DeoR. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis results are shown. Lane 1, molecular mass standards (from top to bottom, 94, 67, 43, 30, 20, and 14 kDa); lane 2, soluble cell extract; lane 3, flowthrough from an Ni-NTA agarose column; lane 4, wash from an Ni-NTA agarose column with wash buffer; lanes 5 and 6, pooled fractions eluted by the imidazole gradient and dialyzed pooled fractions, respectively.

In vivo complementation by the six-histidine-tagged DeoR repressor.

Although the His tag does not usually interfere with the structure or function of purified proteins (4), we tested the His-tagged DeoR for in vivo complementation of a B. subtilis deoR mutant. The deoR gene with six histidine codons from pJOY1000 was subcloned into pEB112 under the control of inducible promoter P_tac as described in Materials and Methods. When transformed with this plasmid, the deoR strain XM251 was phenotypically DeoR⁺ in the presence of IPTG. β-Galactosidase activity showed that dra-lacZ expression in strain XM251 had a normal, approximately 15-fold DeoR regulation (14) similar to the wild-type XM15 when grown in minimal medium succinate containing (Table 3).

TABLE 3.

β-Galactosidase level of B. subtilis wild-type and deoR strains carrying a dra-lacZ fusion^a

Strain	Relevant genotype	Inducer added	Enzyme activity (nmol/min/mg)
XM15	Wild type	None	15
		Deoxyribose	262
XM25	deoR::erm	None	107
		Deoxyribose	110
XM251	XM25(pXM1000)	None	14b
		Deoxyribose	189b

^aCells were grown in minimal medium containing succinate. Inducer was added to a final concentration of 1 mg/ml. 

^bIPTG was added to a final concentration of 1 mM to induce deoR. 

Molecular mass of the DeoR repressor protein.

By comparing the mobility of purified repressor on sodium dodecyl sulfate-polyacrylamide gels with those of several other proteins of known molecular weight, the mass of the His-tagged DeoR subunit was found to be 35 kDa (Fig. 1). This agrees with the molecular mass of 34 kDa calculated from the derived amino acid sequence of the deoR gene (12).

The native molecular mass of DeoR repressor in wash buffer containing 200 mM imidazole, as determined from its elution profile from a Sephadex G-150 gel filtration column, is 280 ± 10 kDa. This estimate was based on a comparison with the elution pattern of several other proteins of known molecular weight. Assuming that the ratio of Stokes radius to mass of the DeoR protein and the size standard proteins is the same, this means that the native protein is most likely an octamer.

DNA binding to DeoR.

An electrophoretic gel mobility shift assay as described in Materials and Methods was used to measure the binding of DeoR to labeled operator DNA. In most cases, purified DeoR was used. The radioactive oligonucleotide used for the characterization of binding was a 111-bp fragment corresponding to nucleotides −80 to +30 relative to the dra-nupC-pdp operon transcription start point (14). This 111-bp fragment contains the operator for DeoR and was shown in preliminary studies to bind well to crude DeoR (14). The specificity of the interaction between DeoR and operator DNA was tested in two ways. First, to demonstrate that the DNA was bound specifically by DeoR, gel shift assays were performed using the 111-bp DNA fragment and crude extracts from either E. coli JOY1000 which overexpresses DeoR or JOY999 which carries the vector plasmid only. The crude extract containing overexpressed DeoR clearly contained a protein that binds to DNA: increasing amounts of this extract increased the amount of DNA bound (Fig. 2, lanes 2 to 5). In contrast, crude extracts from cells that carried the vector only contained no protein that bound to DNA (Fig. 2, lane 6). These results indicate that DeoR protein binds to DNA and rule out the possibility that an impurity in the DeoR preparation binds to the DNA instead.

FIG. 2.

Binding of the 111-bp dra-nupC-pdp operator DNA by crude extracts (15 mg/ml) of E. coli cells in which DeoR was overexpressed (JOY1000 [lanes 1 to 5]) and a control strain bearing the plasmid vector only (JOY999 [lane 6]). Lanes: 1, free DNA fragment (no extract); 2 to 5, JOY1000 extract dilution of 1:50, 1:20, 1:10, and 1:5, respectively; 6, JOY999 extract dilution of 1:5.

To demonstrate that DeoR binds specifically to the operator DNA, quantitation of binding affinity was studied with purified DeoR, which was not possible in earlier experiments with crude extracts. We have determined the apparent dissociation constant for binding of the purified repressor to the operator region of the dra-nupC-pdp operon. Binding isotherms were calculated from the increase in the levels of bound DNA, and apparent K_d values represented the DeoR concentration required for 50% saturation of the control site DNA. The apparent dissociation constant determined from this data was 22 nM (Fig. 3A). The binding of DNA to DeoR was described by a sigmoid curve (Fig. 3B), which suggested that the binding of DeoR to the operator DNA is cooperative. In other words, the binding of operator DNA to DeoR enhances the binding of additional operator DNA to the same DeoR molecule.

FIG. 3.

Binding of the DeoR repressor protein to the 111-bp operator DNA of the dra-nupC-pdp operon. A profile of a gel shift assay (A) and the calculated binding isotherm for DeoR with the operator DNA (B) are shown. DeoR concentrations (in nanomolar) are given.

Binding stoichiometry.

In order to determine the DeoR binding stoichiometry for operator DNA of the dra-nupC-pdp operon, gel shift assays using high concentrations of operator DNA (DNA concentration much greater than K_d) were performed as described in Materials and Methods. The binding stoichiometry was approximately four DeoR molecules per 111-bp DNA fragment, assuming all the DeoR protein was active (Fig. 4). This suggested that four DeoR subunits were needed for total binding to the operator DNA.

FIG. 4.

DeoR binding stoichiometry for the 111-bp operator DNA of the dra-nupC-pdp operon. DeoR concentrations (in micromolar) are given at the top of the gel. Nonradioactive 111-bp DNA fragment (1 μM) was added to each binding assay in addition to the radiolabeled DNA.

Three palindromic halves are required for DeoR binding.

It has been shown that both the palindrome and the direct repeat are necessary for the binding of DeoR to the operator DNA of the dra-nupC-pdp operon (14). To investigate the roles of these three palindromic halves, the binding affinity was quantitated with three DNA fragments containing different parts of the operator site and the apparent dissociation constant for DeoR binding was determined (Fig. 5). The apparent K_d value determined from these data was 20 nM for a DNA fragment containing the palindrome and the direct repeat (Fig. 5). No or very weak binding was found for DNA fragments containing either only the palindrome (Fig. 5) or containing the 3′ half of the palindrome and the direct repeat (data not shown). This result indicated that binding of the DeoR repressor to the operator DNA operon required both the palindrome and the direct repeat. In other words, three palindromic halves are needed for tight binding.

FIG. 5.

Binding of the DeoR repressor to DNA fragments containing different operator sites. Results with a 43-bp fragment containing the palindrome and the direct repeat and a 34-bp fragment containing only the palindrome are shown.

Effect of dRib-5-P on DeoR binding to the operator.

In an earlier gel shift assay with crude DeoR, dRib-5-P was able to release DeoR from the DNA-protein complex (14), but no similar experiment has been performed with dRib-1-P in vitro. Here we have determined the apparent dissociation constant for binding of the purified repressor to the 111-bp fragment in the presence of dRib-5-P or dRib-1-P (Fig. 6). The results showed that K_d was increased 60-fold when 100 μM dRib-5-P (K_d > 1,280 nM) was present in the assay mixture, whereas almost no change was observed for K_d when 100 μM dRib-1-P (K_d = 25 nM) was present. These results confirmed the results of a previous report that dRib-5-P binds to DeoR in vitro and acts as an internal inducer for the expression of the dra-nupC-pdp operon (14). In contrast, dRib-1-P binds only very weakly, if it binds at all, to DeoR under the in vitro conditions tested.

FIG. 6.

Binding of DeoR to the 111-bp operator DNA of the dra-nupC-pdp operon in the presence of 100 μM dRib-5-P or dRib-1-P.

DNase I footprinting analysis of the interaction between DeoR and DNA.

DNase I footprinting was used to identify the precise locations of the DeoR binding sites. The same 111-bp DNA fragment that was used for the measurement of DeoR binding affinity (except that a single strand was end labeled) was used for the DNase I footprinting experiment. The labeled DNA fragment was incubated with or without 5 μM DeoR and partially digested with DNase I. The pattern of protection and hypersensitivity is shown in Fig. 7. In the absence of repressor, DNase I cleavage produced a distinct pattern of bands (Fig. 7, lane 3). Upon addition of the DeoR repressor, a protected region of 43 bp appeared covering most of the palindrome, the direct repeat, and all the nucleotides between them (Fig. 7, lane 4). This confirms the previous reports about the locations of DeoR binding sites from work with mutagenesis and gel shift assays (14).

FIG. 7.

DNase I footprinting of DeoR binding sites. Lane 1, G+A sequencing ladder; lane 2, T+C sequencing ladder; lanes 3 and 4, no DeoR (lane 3) and 5 μM DeoR (lane 4) was added. The nucleotide sequence of the sense strand of the operator DNA between nucleotides −60 and −22 relative to the transcription start site is given to the left of the gel. The palindrome and the direct repeat are marked by the vertical lines.

It is worth mentioning that the adenine residue at the 5′ end of the palindrome 5′-ATTGAACAAAATTTCAAT-3′ was found to be not protected or only weakly protected. Previous mutagenesis studies of this palindrome showed that this adenine residue had no effect in DeoR regulation in vivo (14). Moreover, the adenine residue at the 3′ end of the direct repeat 5′-TTCAA-3′ was only weakly protected, too.

DISCUSSION

We conclude from our studies that the purified B. subtilis DeoR protein is an octamer composed of 34-kDa subunits which binds cooperatively to dra-nupC-pdp operator DNA. The affinity of DeoR for operator DNA is greatly reduced by binding of dRib-5-P. These conclusions are based on studies of the DeoR protein bearing an N-terminal six-histidine tag, which was used because we were unable to purify the native DeoR protein. Thus, it is reasonable to ask whether the properties of the His-tagged DeoR are the same as those of the native DeoR. We believe that they are for the following reasons. Binding to dra-nupC-pdp operator DNA by overexpressed recombinant native DeoR and His-tagged DeoR in crude E. coli extracts was essentially the same with respect to affinity for DNA and the effect of dRib-5-P. Also, a plasmid-borne copy of the gene for the His-tagged DeoR protein could complement a B. subtilis deoR mutant just as well as the native deoR gene. Finally, the properties of the purified His-tagged DeoR account very well for previously described observations on the repression of the dra-nupC-pdp operon in vivo (12, 14).

A comparison of the primary structure of the B. subtilis DeoR with protein sequences in the database showed that B. subtilis DeoR has significant similarity to several regulatory proteins which belong to the SorC family of transcriptional regulators from different organisms. Interestingly, the proteins with the highest degree of similarity can be divided into two groups. SorC (Klebsiella pneumoniae) (GenBank accession no. X66059), DalR (K. pneumoniae) (accession no. AF045245), SmoC (Rhodobacter sphaeroides) (accession no. AF018073), and EriD (Brucella abortus) (accession no. U57100) show high degree of similarity to the amino-terminal part of DeoR, which contains the DNA-binding domain. Much less similarity is found in the rest of the primary sequence. SorC, DalR, SmoC, and EriD all regulate the transcription of genes involved in sugar alcohol catabolism. The second group consists of GapR (Staphylococcus aureus) (accession no. AJ133520), YgaP (Bacillus megaterium) (accession no. M87647), YvbQ (B. subtilis) (accession no. Z99121), and ClyR (Leuconostoc mesenteroides) (accession no. Y10621). This group of regulators shows similarity to the carboxy-terminal region of DeoR. GapR, YgaP, and YvbQ encode regulators of operons containing gap, which encodes the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase. The ClyR protein is involved in control of citric acid cycle gene expression. Hence, this group of proteins regulates genes involved in glucose metabolism. We have found that DeoR most likely binds dRib-5-P, and we speculate that the binding site may include parts of both the amino-terminal region (perhaps overlapping the DNA-binding region) and the carboxy-terminal region. The DeoR amino-terminal part is similar to proteins that bind sugar alcohol phosphates as effector molecules, and the carboxy-terminal part is similar to proteins that most likely bind glyceraldehyde-3-phosphate, which is the product of dRib-5-P cleavage catalyzed by deoxyriboaldolase. Domains capable of binding sugar alcohol phosphates and glyceraldehyde-3-phosphate may have been incorporated into the DeoR structure in order to create a dRib-5-P-specific binding domain.

Although they both appear to contain an α-helix–turn–α-helix domain of the type commonly found in DNA-binding proteins (3, 10) and appear to exist as octamers in the native (DNA-free) state, B. subtilis and E. coli DeoR repressor proteins share little sequence similarity and the DNA sequences to which they bind are dissimilar. The E. coli DeoR is thought to bind simultaneously to two or three operators of the 16-bp palindrome, which are separated by hundreds of base pairs. There is no evidence that B. subtilis DeoR binds to more than one operator site, although the operator site to which it binds has a complex structure, as noted in the next paragraph. Furthermore, DeoR repression of the deo operon in E. coli is characterized by long-range cooperative regulation (2), whereas no more than 141 bp of DNA is enough for complete DeoR repression of the dra-nupC-pdp operon of B. subtilis (14).

Previous molecular genetic studies with the dra-nupC-pdp operon indicated that a palindromic sequence located between nucleotides −60 and −43 relative to the start of transcription and a direct repeat of the 3′ half of the palindrome located between the −35 and −10 regions were both required for repression of the operon by DeoR (14). The results of the present studies directly demonstrate that these DNA elements are required for binding to DeoR in vitro. Furthermore, the corresponding segment of DNA was protected by DeoR in DNase I footprinting studies. This is a highly unusual structural requirement for a DNA-binding protein. Typical operator sequences consist of palindromes only, and typical repressors are dimeric proteins in which each subunit binds to one of the halves of the palindrome. Repressor proteins that are tetrameric or larger sometimes bind to multiple palindromic operators, as with E. coli DeoR. In the case of B. subtilis DeoR, our titration studies indicate that four subunits bind to a single segment of operator DNA (Fig. 4). Assuming that all the DeoR protein was active in the titration studies, this stoichiometry suggests to us that each subunit binds to one half of the palindromic sequence, so that three of the four subunits are bound to DNA in the DeoR-operator complex. DeoR is an octamer in solution but may dissociate to a tetrameric form upon dilution to the concentration used in the gel shift experiments. Cooperativity as observed in the DeoR binding curves (e.g., Fig. 3 and 4) could reflect differences in the affinity of the subunits for the slightly different half-palindromic sequences.

High-affinity B. subtilis DeoR binding to DNA takes place in the absence of effector molecule. dRib-5-P is most likely the effector that modulates B. subtilis DeoR binding to DNA, acting as an inducer to inhibit the binding of a repressor protein to a control site. Although dRib-1-P has also been reported as an alternative inducer (12, 14), no effect on DeoR binding to DNA is observed in the presence of dRib-1-P in vitro. In E. coli, dRib-5-P but not dRib-1-P induces the expression of the deo operon (9), but no information is available with respect to protein-effector molecule interaction. More detailed studies of B. subtilis DeoR are needed to locate the inducer-binding domain.