Functional dissection of the LysR-type CysB transcriptional regulator. Regions important for DNA binding, inducer response, oligomerization, and positive control.

CysB is a tetrameric LysR-type transcriptional regulator that acts as an activator of cys regulon genes and as an autorepressor. Positive control of cys genes requires the presence of the inducer N-acetylserine. Following random and site-directed mutagenesis of the cysB gene, 20 CysB variants were isolated. Six single amino acid substitutions within the N terminus of CysB abolished the DNA-binding ability of the protein. Seven mutations in the central region of CysB affected its response to the inducer. Four of these CysB mutants retained repressing activity, but lost their activating function in vivo. Their DNA binding characteristics were consistent with an inability to respond to acetylserine by a qualitative change in the DNA-protein interaction. Three of the single residue substitutions resulted in constitutive activity of CysB. The electrophoretic mobility of the complex formed by one of the CysBc variants with the cysP promoter suggested a dimeric state of this protein. Characteristics of six truncated CysB variants lacking 5-30 C-terminal residues indicated the involvement of the C terminus in the DNA binding, oligomerization, and stability of CysB. The single substitution Y27G resulted in the CysBpc variant, able to bind DNA and to respond to the inducer by a qualitative change in the DNA-protein complex, but defective in the positive control of the cysP promoter.

cysJ, cysK, and cysP promoters. The stable interaction of the CysB tetramer with a unique activatory site is critical for transcriptional activation of the promoter, and it may be required for RNA polymerase recruitment. In the cysB promoter, CysB protects the DNA region overlapping the RNA polymerase-binding site that results in repression of transcription (13). In this case, acetylserine reduces binding of CysB to DNA, thereby releasing the repression of the cysB gene. Sulfur limitation required for high-level expression of cys genes in vivo can be explained by the fact that sulfide and thiosulfate act as anti-inducers, competing with N-acetylserine for the CysB activator (18,19). In addition, cysteine is an inhibitor of serine transacetylase, the enzyme required for acetylserine biosynthesis.
CysB is an attractive representative of LTTRs to study the structure/function relationship for several reasons. First, the molecular basis of the interaction of CysB with cys gene promoters, including DNA bending and inducer response, is relatively well understood. This offers the tool for in vitro comparisons of properties of CysB mutants versus the wild-type protein. Second, the CysB protein is so far the only family member whose cofactor-binding domain (amino acids 88 -324) has been crystallized and subjected to three-dimensional analysis (20). The structure of the dimer formed by CysB monomer fragments (amino acids 88 -324) has shed some light on protein surfaces composing an inducer-and anti-inducer-binding cavity. However, the structure lacking the 87 N-terminal amino acids provide little information about the overall organization of the native tetrameric protein. Third, in some promoters of E. coli, including the tau promoter (21) and the ssu promoter (22), CysB acts in conjunction with another LysR-type activator, Cbl, which itself is highly homologous to CysB (23). The molecular basis of cooperation of these two proteins is not yet understood. A systematic mutational analysis of CysB might give further insight into the structure of this protein and related LysR family members and provide a valuable complement to the information gained from physicochemical studies. Here we describe the properties of several CysB variants resulting from single amino acid substitutions in the N-terminal and central regions of the protein as well as those resulting from C-terminal truncations. On the basis of the effects of these mutations on the repressing and activating functions of CysB in vivo and the DNA binding characteristics of selected CysB mutants in vitro, we discuss regions of the protein involved in DNA binding, inducer response, oligomerization, and possible RNA polymerase recruitment.

MATERIALS AND METHODS
Bacterial Strains and Growth Conditions-The E. coli strains and basic plasmids used in this study are listed in Table I. All strains were grown at 37°C in either LB medium or sulfur-free minimal medium (24) supplemented with glucose (0.2%), tryptophan (4 g/ml), and a sulfur source. Sulfate (0.5 mM) was used as a sulfur source in solid media, and L-djenkolic acid (1 mM) served as non-repressing sulfur source in liquid media for ␤-galactosidase assays. When required, the growth media contained ampicillin (100 g/ml), chloramphenicol (20 g/ml), or kanamycin (50 g/ml). The solid media (LB agar with X-gal or MacConkey/lactose agar) were used for selection of transformants of the desired phenotype.
DNA Manipulations-Plasmid DNA preparation, restriction enzyme digestion, and transformation were performed according to published procedures (25). Plasmid DNA was purified with QIAGEN midi-columns. Restriction endonucleases, DNA-modifying enzymes, and Taq polymerase were purchased from Life Technologies, Inc. and Promega and were used according to the manufacturers' recommendations. The dideoxy chain termination method (26) was used to sequence the cysB region from plasmid templates. Most sequences were obtained using an ABI PRISM 377 DNA sequencer, and some plasmids were sequenced manually using a sequencing kit and ␣- 35 S-dATP (both from Amersham Pharmacia Biotech).
Construction of the ⌬cysB Mutant-Deletion within the cysB gene was created by excision of the 821-bp NruI/HpaI fragment encompassing the promoter region and 216 codons of the cysB open reading frame (see Fig. 1) from plasmid pMH148, containing the 2.8-kilobase pair EcoRI/EcoRV fragment of pJOH2 cloned in pUC18 EcoRI/HincII. The deletion was tagged by insertion of a Kan r cassette excised from pUC4KIXX as a 1.35-kilobase SmaI fragment. The resulting plasmid, pMH161, was linearized and transformed into E. coli JC7623. Kan r Ap s transformants were selected, and correct replacement of the cysB gene in the genomic DNA was confirmed by PCR analysis. The ⌬cysB::kan mutation was transferred to other E. coli strains (see Table I) by P1-mediated transduction.
Oligonucleotides and Plasmid Constructions-The 3Ј-coordinates indicated for all cysB-specific oligonucleotides used in this study are consistent with those of the published cysB sequence (27), where ϩ1 indicates A in the ATG start codon. Nucleotides that had been changed to create restriction sites, stop codons, or specific mutations are shown in boldface. Primers ECCB3 (CCCAAGCTTTTTTTCTTCACACCTAT-ACA-3Ј, Ϫ123) and ECCB4 (CATGCATGCGGTAATTAGACACTACT-T-3Ј, ϩ1101, complementary strand) were used for PCR amplification of the wild-type E. coli cysB gene from the pJOH1 plasmid. The fragment obtained was cleaved with HindIII/SphI and inserted into pBR322 HindIII/SphI to give plasmid pMH147. Plasmids pMH189, pMH191, pMH193, pMH195, pMH222, pMH223, pMH224, pMH225, pMH226, pMH228, and pMH229 are derivatives of pMH147 obtained by random mutagenesis and contain point mutations within cysB (see Table II). Plasmids pMH190, pMH192, pMH194, pMH196, and pMH197 were obtained from pMH189, pMH191, pMH193, pMH195, and pMH147, respectively, by excision of the HindIII/SphI fragments (containing the cysB region) and their insertion into HindIII/SphI-cleaved pACYC184. Primers ECCB8 (GGAATTCTAAGTGGATGGTTTAACA-3Ј, ϩ1) and ECCB9 (AAACTGCAGGCGGTAATTAGACAC-3Ј, ϩ1106, complementary strand) served to amplify the promoterless cysB gene on the pJOH1 template, and the PCR product was cleaved with EcoRI/PstI and placed under the p trc promoter in pTrc99A cleaved with EcoRI/PstI to give plasmid pMH199. Plasmid pMH199 served for overproduction of WT CysB. To overproduce CysB mutant proteins, plasmids pMH200, pMH201, pMH202, pMH203, and pMH205 were obtained by replacement of the 653-bp SmaI/BamHI fragment (see Fig. 1) of pMH199 with the SmaI/BamHI fragments excised from plasmids pMH189, pMH191, pMH193, pMAH01, and pMAH2, respectively. To construct C-terminally truncated CysB variants, oligonucleotide ECCB13 (TTACGTCCG-GTTAGGGC-3Ј, ϩ712) was used as an upper primer, and oligonucleotides ECCB14 (AAACTGCAGTTATTAATCAACGACATCACGCGT-3Ј, ϩ886, complementary strand), ECCB15 (AAACTGCAGTTAAGA-GCGCAATGCGAC-3Ј, ϩ910, complementary strand), and ECCB16 (AAACTGCAGTTATATATCTTTAAACATG-3Ј, ϩ942, complementary strand) were used as lower primers for PCR amplification of 3Ј-terminal fragments of cysB from the pJOH1 template. The PCR fragments were cleaved with BamHI-PstI and ligated with the pMH199 fragment obtained by excision of the 3Ј-portion of the WT cysB region (as a 377-bp BamHI/PstI fragment) to give plasmids pMH214, pMH215, and pMH217 (serving for overproduction of the corresponding CysB mutant proteins). Plasmids pMH231, pMH232, and pMH233 were obtained by replacement of the BamHI/SphI fragment of pMH147 (containing the 3Ј-portion of WT cysB) with the BamHI/SphI fragments excised from pMH214, pMH217, and pMH215, respectively. Other plasmids carrying truncated cysB (pMH218, pMH219, and pMH230) were obtained as follows. The BamHI/PstI fragment of pMH199 (containing the 3Ј-portion of WT cysB) was inserted into pUC19, and the insert was amplified by PCR with pUC-specific forward and reverse primers under mutagenic conditions (28). The PCR product was cleaved with BamHI/SphI and ligated with pMH147 devoid of the corresponding BamHI/SphI region. The ligation mixture was transformed into the EC2549 strain, and the transformants unable to grow on minimal sulfate plates were selected (3 of ϳ2000) and used to recover the corresponding plasmids.
Mutations resulting in C-terminal truncations of CysB are listed in Table II. To introduce the Y27G substitution into CysB, primers ECCB8 and ECCB19 (GATCCCGGGTTGTGAGGTACCAAGTCCTTCCGC-3Ј, ϩ67, complementary strand) were used for PCR amplification of DNA fragment from the pJOH1 template. The fragment obtained was cleaved with EcoRI/SmaI and used to replace the corresponding 113-bp EcoRI/SmaI fragment (see Fig. 1) of pMH199 to give plasmid pMH235. Plasmid pMH303, containing the translational fusion cysBЈ::ЈlacZ, was constructed as follows. The 450-bp BamHI/SmaI fragment (see Fig. 1), containing the cysB promoter and 30 codons of the cysB open reading frame, was excised from pJOH1 and inserted into pUC18 BamHI/ HincII. The cysB region was recovered as a BamHI/PstI fragment and inserted into pNM482 cleaved with BamHI/PstI to give pMH302. The cysBЈ::ЈlacZ region was excised from pMH302 as a 5.5-kilobase EcoRI/ StuI fragment and cloned in the low-copy vector pHSG576 cleaved with EcoRI/SmaI to give pMH303. Primers ECCB3 and ECCB11 (GTGCT-TGCCGCTTCGGGA-3Ј, ϩ148, complementary strand) served to amplify the cysB promoter fragment from the pJOH1 template for DNA binding assays.
Primers ECCP1 (ATTGATGGCGGCAGCACACT) and ECCP2 (GAC-CAGCGCGAGTGAGTTCT, complementary strand) were used to amplify the cysP promoter region (from positions Ϫ321 to ϩ72 relative to the transcription start site) from the pMH1822 template.
Random Mutagenesis of cysB-Mutagenesis with hydroxylamine was carried out essentially as described (29). Briefly, ϳ1 g of purified pMH147 was mixed with 50 l of solution containing freshly prepared 1 M hydroxylamine, 50 mM sodium pyrophosphate, 100 mM NaCl, and 2 mM EDTA (pH 7.0) and incubated at 75°C for 45 min. The samples were dialyzed extensively against Tris/EDTA buffer (pH 7.6), and DNA was precipitated with ethanol and used to transform the EC2266 strain. Plasmid pMH147 was also used to transform the E. coli mutator strain mutD5. The transformants obtained were rinsed off the plates, and the plasmid DNA was re-isolated and used to transform the EC2266 strain. Transformants were screened for the desired phenotype on LB agar plates containing X-gal or on MacConkey/lactose agar plates.
Enzyme Assays-␤-Galactosidase activities were assayed in cells taken from mid-log phase cultures grown on minimal medium with L-djenkolic acid as a derepressing sulfur source according to the method of Miller (30) with o-nitrophenyl galactoside as a substrate. Duplicate assays were performed for each culture and were repeated at least three times.
Preparation of CysB-overproducing Extracts-E. coli extracts were prepared from strain DH5␣ containing plasmid pMH199 (WT cysB) or plasmids (derivatives of pTrc99A) encoding mutant alleles of cysB. Transformed cells were grown on LB/ampicillin, and isopropyl-␤-Dthiogalactopyranoside was added (0.5 mM final concentration) in the early exponential phase (A 600 ϭ 0.15). Growth was continued for a further 2 h, and cells were harvested and resuspended in buffer A (50 mM Tris-Cl (pH 7.5), 1 mM disodium EDTA, and 1 mM phenylmethylsulfonyl fluoride). Cell suspensions were sonicated on ice, and cellular debris was removed by centrifugation at 16,000 ϫ g for 30 min at 4°C. The supernatants were fractionated by streptomycin sulfate precipitation and ammonium sulfate precipitation as described previously (15), and fractions precipitated with ammonium sulfate (229 mg/ml) were dissolved in buffer A. All preparations in which the CysB protein was successfully overproduced served as a source of CysB proteins in DNA binding experiments. Highly purified CysB protein from Salmonella typhimurium (obtained from N. M. Kredich) was used as a reference for the E. coli WT CysB preparation in DNA binding and transcription runoff experiments.
DNA Binding Assays-DNA fragments containing the cysP and cysB promoter regions were generated by PCR (see "Oligonucleotides and Plasmid Construction"), 5Ј-labeled with [␥-32 P]ATP using T4 polynucleotide kinase (Promega), and purified with QIAquick spin columns (QIA-GEN). Conditions for binding reactions and electrophoretic mobility shift assay (EMSA) were essentially as described (17). Briefly, the reaction mixtures (20 l) contained ϳ10 ng of labeled DNA fragment and 50 ng of sonicated calf thymus DNA (to reduce nonspecific binding) in CysB buffer consisting of 40 mM Tris-HCl (pH 8.0), 10 mM MgCl 2 , 100 mM KCl, 1 mM dithiothreitol, and 100 g/ml bovine serum albumin. Incubation (5 min at 37°C) was initiated by including a given amount of CysB preparation and O-acetyl-L-serine where indicated. Samples were separated on 5% acrylamide/bisacrylamide (82:1) vertical gels in 0.05 M Tris borate/EDTA buffer (pH 8.3) for 1.5 h at 10 V/cm, and the radiolabeled bands were visualized by autoradiography. Although Nacetyl-L-serine is a better inducer of CysB-dependent transcriptional activation than O-acetyl-L-serine, the latter was shown to function better in EMSA experiments (19). The differences between the two are mostly operational and have been discussed in detail previously (14); the inducer is referred to as acetylserine below.
In Vitro Transcription Runoff Assay-The DNA fragment containing the E. coli cysP promoter used as a template for in vitro transcription was the same as that used for DNA binding experiments. Transcription initiation complexes were formed in 20-l mixtures containing ϳ20 ng of template DNA, the indicated amounts of purified S. typhimurium or E. coli CysB preparations and O-acetyl-L-serine or N-acetyl-L-serine, 1 g of nuclease-free E. coli RNA polymerase (Amersham Pharmacia Biotech) in CysB buffer containing 0.1 mM ATP. After 5 min at 37°C, transcription elongation was started by adding 2 l of a solution containing 20 M [␣-32 P]CTP (200 Ci/mmol), 2 mM ATP, 2 mM GTP, 2 mM UTP, and 0.5 mg/ml sodium heparin. After another 5 min at 37°C, the reaction was terminated by the addition of 0.2 ml of 10 mM disodium EDTA containing 50 g/ml yeast tRNA. Ethanol-precipitated radiolabeled transcripts were analyzed on sequencing gels.
Molecular Modeling-The homology model of E. coli CysB has been built on the basis of the crystal structure of CysB from Klebsiella aerogenes (fragment 88 -324; Protein Data Bank code 1AL3) by simple replacement of different residues and energy minimization using consistent valence force field (CVFF) force field as implemented in DIS-COVER Version 2.3.5 (MSI/Biosym, San Diego, CA). N-Acetyl-L-serine has been inserted in the CysB structure according to Tyrrell et al. (20).

RESULTS
Isolation and Initial Characterization of Random cysB Mutants-Assuming that the lack of positive control function of CysB can result from mutations impairing DNA-binding ability, inducer response, or RNA polymerase contacts, we used the cysPTЈ::lacZ cysB strain (EC2266) ( Table I) for selection of cysB mutants contained on plasmids. Since the expression of ␤-galactosidase in EC2266 is under the control of the CysB-dependent cysP promoter, the strain yielded white colonies on LB agar plates containing X-gal and dark-blue colonies in the presence of plasmid pMH147 carrying WT cysB. After transformation of EC2266 with several samples of mutagenized pMH147, 32 white or pale-blue transformant colonies were selected. To determine the location and nature of the mutations, the cysB genes contained on plasmids were sequenced entirely. This allowed us to exclude mutants resulting from generation of an early termination codon within the cysB open reading frame (with the exception of the Q5Ter mutant, which served as a control in some experiments) and several mutants resulting from identical nucleotide change. Ten plasmids containing cysB with mutations that caused single residue substitutions within the CysB protein (Table II) were analyzed further. None of these plasmids complemented the Cys Ϫ phenotype of two cysB strains tested (Table II). To distinguish between repressing and non-repressing cysB mutants, the corresponding plasmids were screened for their ability to affect the expression of the translational cysBЈ::ЈlacZ fusion, encoded by the low-copy number plasmid pMH303 in the ⌬cysB background (EC2549 strain). As shown in Table II, the presence of WT cysB reduced the expression of ␤-galactosidase in this strain to ϳ4%. Decreased levels of cysBЈ::ЈlacZ expression (2.6 -5.6%) were also observed with four plasmids encoding CysB variants M160I, T196I, A244V, and A247E. The same mutants cloned in pA-CYC184 reduced to ϳ5% the expression of the chromosomal fusion cysBЈ::Mud1(Ap,lacZ) in strain EC1171 (data not shown). We concluded that repressing CysB proteins M160I, T196I, A244V, and A247E (positions of mutations shown in Fig.  1) expressed from these plasmids retained a DNA-binding ability. The remaining six CysB variants with single residue substitutions in the N-terminal portion of the protein (region 11-48 containing the HTH motif and its close vicinity; see  (Table II). These variants were therefore assumed to be impaired in DNA binding. The repressing activities of plasmids carrying cysB mutants were also measured in strain EC1250 containing pMH303. The ␤-galactosidase activity in this strain represents partially repressed level of cysBЈ::ЈlacZ expression resulting from the activity of chromosomally encoded WT CysB; this value has been estimated as ϳ2.5 times lower than that in the EC2549 strain containing pMH303. The presence of plasmids encoding some of the nonrepressing CysB variants (E11K, S20L, and T22I) in the pMH303-containing EC1250 strain resulted in elevation of the cysBЈ::ЈlacZ expression to 230 -270% of the level characteristic for the strain (Table II). The effect observed could be consistent with the assumption of negative dominance of CysB proteins expressed from plasmids over chromosomally encoded WT CysB. In contrast, a plasmid expressing no CysB (Q5Ter mutant) and plasmids encoding CysB variants I48T, L44T, and E41K showed no such dominant-negative effect, as they did not change the level of cysBЈ::ЈlacZ expression in EC1250. We noticed that all four repressing and three non-repressing (E11K, S20L, and T22I) variants expressed from multicopy plasmids converted the wild-type strain EC1250 to the Cys Ϫ phenotype (Table II). We also checked three representative mutants of this group cloned in middle-copy plasmids (derivatives of pA-CYC184) for their effect on the expression of the chromosomal cysPTЈ::lacZ fusion in the EC2256 strain, WT for cysB. As shown in Fig. 2, plasmids carrying those cysB mutants markedly reduced ␤-galactosidase levels in EC2256, confirming the dominant-negative effects of mutant proteins expressed from plasmids over the WT CysB protein produced by the cell. Knowing that WT CysB is a tetramer, the trans-dominance of the multicopy CysB mutant forms may be explained by titration of WT CysB monomers and formation of inactive mixed heterooligomers. Therefore, it seems that none of the mutations altering residues 11,20,22,160,196,244, and 247 abolished the oligomerization ability of CysB. In contrast, CysB variants E41K, L44R, and I48T expressed from plasmids did not confer the Cys Ϫ phenotype to the EC1250 strain and had no effect on expression of the cysBЈ::ЈlacZ fusion in the pMH303-containing EC1250 strain (Table II). Therefore, mutations in region 41-48 of CysB could affect oligomerization of the protein (see also "Discussion").
Construction and Phenotypic Characterization of C-terminally Truncated CysB Mutants-The collection of CysB mutants isolated by random mutagenesis lacked representatives with changes in the C-terminal part of the protein. According to the structural model of the CysB dimer (20), 33 C-terminal amino acids (positions 292-324) form the loop exposed to the surface containing two helices. Therefore, this domain might be important for the overall structure of the native CysB molecule and its function. To pursue the possible effects of C-terminal mutations, we employed several approaches to construct mutated and/or truncated forms of CysB (see "Oligonucleotides and Plasmid Construction"). Six plasmids were obtained that encoded CysB variants lacking 5, 16, 19, 23, 26, or 30 Cterminal residues of the WT protein. Among these, only the CysB K320Ter variant conferred the Cys ϩ phenotype to both cysB-negative strains tested, EC2275 and EC2549. CysB K320Ter showed also significant repressing activity in relation to the cysBЈ::ЈlacZ fusion in the ⌬cysB strain, and it was able to activate the cysPTЈ::lacZ fusion to the level observed with WT CysB (Table II). Therefore, it appears that CysB K320Ter, lacking 5 C-terminal residues, retained sufficient activity for all CysB functions in vivo. All further truncated CysB variants (lacking 16 -30 C-terminal residues) were apparently not fully functional since plasmids carrying the corresponding cysB mutant genes did not complement the Cys Ϫ phenotype of the ⌬cysB strain EC2549. Among these, the CysB protein lacking 16 C-terminal amino acids (CysB N309Ter) showed a reduced repressing activity but almost wild-type positive control activity in vivo as measured with the cysBЈ::ЈlacZ fusion in the pMH303-containing EC2549 strain and with the cysPTЈ::lacZ fusion in the EC2559 strain, respectively. All further truncated CysB variants displayed no repressing activity and significantly reduced positive control function in the same tests (Table II). However, plasmids encoding CysB variants lacking 19 -30 C-terminal amino acids showed a curious complementation activity in strain EC2275 (cysB), observed as the appearance of a few transformant colonies over the background of abortive growth on minimal plates with sulfate. The presence of these plasmids in strain EC2266 (cysPTЈ::lacZ cysB) resulted in the appearance of a mixed population of transformants on MacConkey/lactose solid medium, seen as a few red colonies on the overall white background of bacterial growth. Because of this type of growth, ␤-galactosidase activities of EC2266 with these plasmids could not be reliably established. The cysB mutation, present in both the EC2275 and EC2266 strains, derives from the NK1 strain, in which it was identified as the  (37). It seems that some arrangement of subunits between CysB I33N and C-terminally truncated CysB variants allows formation of a protein able to positively control the cysP promoter. Our attempts to overproduce CysB mutants lacking Ͼ16 C-terminal residues for in vitro experiments failed, which may suggest an instability of these CysB variants. It is therefore possible that the C terminus contributes to the overall stability of the protein, and the instability of the C-terminally truncated CysB variants might explain the observed abortive complementation of the CysB I33N mutation. DNA Binding Characteristics of Repressing and Constitutive CysB Variants-E. coli WT CysB and those CysB mutants that displayed repressing activity in vivo were cloned in the pTrc99A vector and overexpressed, and obtained extracts were fractionated as described under "Materials and Methods." The extent of CysB overproduction in particular preparations was estimated by SDS-polyacrylamide gel electrophoresis (Fig. 3). The DNA binding characteristics of the E. coli WT CysB preparation used in this work (referred to as E. coli WT CysB) were tested by EMSAs with a 393-bp DNA fragment containing the E. coli cysP promoter region (p cysP ). The complexes formed by E. coli WT CysB with p cysP , shown in Fig. 4A, could be identified by strict correspondence to those formed by highly purified S. typhimurium CysB, analyzed in detail previously (16,18). When a single CysB tetramer binds to p cysP , it induces DNA bending of ϳ100°in the absence of acetylserine and of ϳ50°in its presence, giving rise to "slow" (C1s) and "fast" (C1f) primary complexes, respectively. The C1f (fast) primary complex represents a preferential interaction of the CysB tetramer with the  single activatory site located just upstream of the Ϫ35 region of the promoter. The inducer responsiveness of CysB can therefore be monitored by EMSA as a qualitative (mobility) and quantitative (amount) alteration of the primary complex (Fig.  4A, compare lanes 2 and 4 and lanes 5 and 6). We used EMSA to test three non-repressing CysB variants (A244V, A247E, and T196I) for DNA binding characteristics with p cysP . The results of this analysis showed that all three proteins formed complexes with p cysP (Fig. 4, B and C), but the mobilities of C1 complexes were unaffected by acetylserine. For CysB variants A244V and T196I, these mobilities were exactly the same as that of WT CysB in the absence of inducer (Fig. 4, B, compare lanes 3 and 5-7; and C, compare lanes 3 and 8 -10). In the case of CysB A247E, the C1 complex showed an intermediate mobility between slow and fast, irrespective of the presence of acetylserine (Fig. 4C, lanes 5-7), although the amount of complex was higher with the inducer. It should be stressed that none of the tested mutants, T196I, A244V, and A247E, was able to respond to the inducer by a qualitative change in the DNA-protein complex (reflected by complex mobility), which is critical for transcriptional activation of the promoter by WT CysB. Defects in inducer responsiveness observed in DNA binding studies correlated with the inability of corresponding CysB mutants to activate expression of the cysP promoter (as the cysPTЈ::lacZ fusion) in vivo. Earlier studies on the autoregulatory function of S. typhimurium CysB (19) revealed that acetylserine reduces the affinity of the protein for a particular "repressory" binding site in the cysB promoter region. We have shown by EMSA (Fig. 5) that interaction of E. coli WT CysB with p cysB displayed the same characteristics, i.e. there was less complex formed by a given amount of CysB in the presence of acetylserine than in its absence. In contrast, CysB mutant derivatives T196I, A244V, and A247E formed similar amounts of complex with the cysB promoter fragment, irrespective of the presence of inducer. This result further confirmed the involvement of CysB residues 196, 244, and 247 in response to the inducer and the non-inducible character of the T196I, A244V, and A247E variants. We also analyzed three cysB mutants obtained previously in our laboratory and characterized as causing constitutive expression of sulfite reductase and O-acetylserine sulfhydrylase in E. coli (32). Sequencing of these mutants (cysB c 301, cysB c 302, and cysB c 303) identified the single amino acid substitutions A227D, Y164N, and Y197S, respectively, in encoded proteins ( Fig. 1 and Table II). Plasmids carrying cysB c mutants displayed wild-type repressing and activating functions in vivo in tests summarized in Table II, but they conferred constitutive (not inhibited by cysteine) expression of the cysPTЈ::lacZ fusion in strain EC2266 (data not shown). Two of the CysB c variants, Y164N and A227D, were overproduced, and they were characterized in EMSA with the cysP promoter fragment. The CysB c Y164N protein formed a fast primary complex (C1f) with p cysP in the absence of acetylserine, and the mobility of this complex corresponded exactly to that of the complex formed by WT CysB in the presence of inducer (Fig. 4D, compare lanes 4 and  5). It therefore seems that CysB c Y164N was able to adapt the conformation suitable to interact with an activatory site in p cysP without the aid of the inducer. However, an influence of acetylserine on CysB c Y164N could be noticed as a quantitative effect on C1 complex formation and a further subtle increase in complex mobility (Fig. 4D, compare lanes 5 and 6). This result suggests that CysB Y164N, albeit active in vivo without the inducer, can still respond to it. It is possible that acetylserine induces some conformational change in CysB c Y164N allowing formation of a complex in which DNA is bent less than in the fast complex formed by WT CysB. The complex formed by the other CysB c variant, A227D, showed an unexpectedly higher electrophoretic mobility than complexes formed by WT CysB and CysB c Y164N (Fig. 4D, lanes 7 and 8), and acetylserine affected neither the mobility of this "super fast" complex (C1f*) nor its amount. A secondary complex formed by CysB c A227D was detected (Fig. 4, D, lane 8; and E, lane 5), whose mobility was comparable to that of WT CysB without inducer. The secondary complex (C2) formed by WT CysB is shown in Fig.  4D (lane 3), and a similar C2 complex formed by the CysB T196I variant is shown in Fig. 4C (lane 9). As established by previous stoichiometry studies (16), the C2 complex of WT CysB results from binding of two CysB tetramers in two separate binding sites within p cysP . Considering the differences in mobility between complexes formed by the A227D variant and those of WT CysB, it is tempting to conclude that the A227D protein might bind DNA as a dimer, forming a super fast primary complex.
Among the C-terminally truncated CysB variants, CysB K320Ter was not tested further, as its phenotypic characteristics were similar to those of WT CysB. Of further truncated CysB mutants, only one, CysB N309Ter, could be overproduced to an amount sufficient for DNA binding experiments. The corresponding extract enriched with CysB N309Ter was tested in EMSA using the cysP promoter. As shown in Fig. 4E (lanes  7-9), this protein was able to bind p cysP , forming a complex of much higher mobility than that of WT CysB. Interestingly, the mobility of this super fast complex (C1f*) was comparable to that of the complex formed by CysB c A227D. The plasmid encoding CysB N309Ter (and other plasmids encoding trun- cated CysB variants) displayed no dominant-negative effect on the Cys phenotype of WT strain EC1250 (Table II) and did not reduce the ␤-galactosidase level in the EC2256 strain (data not shown). The lack of a dominant-negative effect on WT CysB in vivo can be most easily explained by an oligomerization defect in CysB N309Ter and the other C-terminally truncated vari-ants. Consequently, we can hypothesize that CysB N309Ter and CysB c A227D bound DNA as dimers because complexes formed by these proteins with the cysP promoter displayed a similar mobility (see also "Discussion"). However, in contrast to CysB c A227D, CysB N309Ter could not perform all functions of WT CysB in vivo, as it was unable to complement the cysteine requirement of the ⌬cysB mutant strain.
Construction and Characteristics of the Putative CysB pc Mutant-One of the objectives of this study was to identify single amino acid substitution(s) in CysB that result in the loss of transcriptional activation of the cys promoter, but not loss of both DNA binding and inducer response. By these criteria, none of the cysB mutants described above could be regarded as defective specifically in "positive control" function. In one of the attempts to isolate a CysB pc mutant (where pc is positive control), we took advantage of the recent study on LysR-type GcvA protein (36) suggesting that single residue substitutions F31L and F31A in the recognition helix of the HTH motif are responsible specifically for the positive control function of GcvA. Since Phe 31 of GcvA corresponds to Tyr 27 of CysB in alignment of the putative HTH regions of these proteins, we introduced the mutation changing Tyr 27 of CysB to Gly by site-directed mu- tagenesis (see "Oligonucleotides and Plasmid Construction"). The plasmid expressing CysB Y27G showed the wild-type level of repression of the cysBЈ::ЈlacZ fusion, but was not able to activate the expression of the cysPTЈ::lacZ fusion (Table II). The Y27G variant was overproduced and used for DNA binding assay with the cysP promoter fragment. As shown in Fig. 6A, CysB Y27G was able to bind the cysP promoter region, giving rise to slow (C1s) and fast (C1f) primary complexes seen in the absence (lanes 6 and 7) and presence (lanes 8 and 9) of acetylserine, respectively. Therefore, the Y27G mutation apparently did not abolish the ability of CysB to respond to the inducer by a qualitative change in the DNA-protein complex. The mobilities of the slow and fast primary complexes formed by CysB Y27G were, however, different than those of the respective complexes formed with WT CysB; the similar differences concerned also migration of secondary complexes (C2). The reason for this discrepancy is not clear at present, and some differences in overall tertiary structure between WT CysB and the Y27G variant may be only hypothesized. To further characterize the CysB Y27G protein, we tested its activity versus WT CysB in a transcription runoff assay with the cysP promoter. At first, we compared the activity of the E. coli WT CysB preparation (used in all DNA binding tests in this study) with that of purified S. typhimurium CysB. As shown in Fig. 6B, the synthesis of identical transcripts was stimulated with both CysB preparations in the presence of O-acetylserine. The sizes of two major runoff products, 73 and 66 nucleotides, corresponded to transcription start sites at positions Ϫ1 and ϩ6, respectively, relative to the major in vivo start site of the E. coli cysP promoter. The appearance of multiple transcripts in the runoff experiment using the E. coli cysP promoter region is not surprising, as it was also observed with the S. typhimurium cysP promoter, in which two major runoff products (corresponding to transcription start sites at positions Ϫ2 and ϩ5 relative to the in vivo transcription start site) and several minor products were synthesized (18). In contrast to WT CysB, the CysB Y27G mutant protein showed very low activity in an analogous runoff assay (Fig. 6B). It should be noted that CysB Y27G was even more active than WT CysB in the DNA binding assay (Fig. 6A) when both preparations were used at equal protein concentrations. We therefore assume that the lack of activity of Y27G in the transcription assay in vitro could not result from an inadequate protein amount or its low stability. On the basis of these data, we conclude that the Y27G mutation abolished the positive control function of CysB, leaving the DNA-binding and inducer response functions unaffected. DISCUSSION In this study, we analyzed the properties of 20 newly isolated CysB mutants to define roles of particular protein regions (see Fig. 1 for proposed functional map of CysB).
Regions Important for DNA Binding-In addition to the previously identified mutations S34R (14) and I33N (37), six new mutations were obtained that are clustered in region 11-48. All of these mutations impaired the ability of CysB to negatively autoregulate and also to activate the cysP promoter. Therefore, it seems clear that the predicted helix-turn-helix motif (residues 18 -38) of CysB (13) and its close vicinity are crucial for the DNA-binding function. Most of the N-terminal residues identified in CysB as important for DNA binding are highly conserved among LTTRs. It was proposed (9) that highly conserved amino acids contact conserved base pairs of a generic TN 11 A motif present in the binding sites for most LTTRs (38). Some CysB residues (i.e. Glu 11 and Thr 22 ) are poorly conserved in other LysR family members; these residues may therefore confer the specificity of DNA-CysB interactions. We found that a C-terminal region of CysB may be also critical for the DNAbinding function, as removal of 19 -30 C-terminal residues resulted in loss of repressing ability of the protein in vivo. The possible involvement of the C-terminal region in DNA binding places CysB in one subclass of LTTRs along with AmpR (8), NahR (7), and OxyR (9), in contrast to other subclass of LTTRs, i.e. MetR (39) and Nac (40), which tolerate substantial Cterminal truncations. The role of the C-terminal end of CysB in DNA binding may be indirect, as truncations of 19 and more residues apparently affect the stability of the protein and possibly its oligomerization state (see below).
Regions Involved in Inducer Recognition/Response-It was shown previously that substitutions T149M/P and W166R result in constitutive or partially constitutive activity of CysB, respectively (41). We have identified seven additional residues of CysB (Fig. 1) that may contribute to inducer binding and response. Non-inducible CysB variants were unable to respond to the inducer, acetylserine, by the conformational change that allows the transition of the slow DNA-protein complex to the fast complex, of which only the latter represents the structure required for transcriptional activation of the cysP promoter (16,18). The interaction of these CysB variants with the cysB promoter region was also unaffected by acetylserine, in contrast to wild-type CysB, whose binding to p cysB was inhibited by the inducer. One of the constitutive CysB variants examined in this study (CysB c Y164N) formed the fast complex with the cysP promoter region independently of the presence of inducer. The characteristics of this mutant correlated well with those described for CysB c T149M (41). More interestingly, the other CysB c variant, A227D, formed a super fast complex with the cysP promoter, suggesting a different oligomerization state of this mutant protein compared with that of WT CysB. Although attempting to evaluate the nature of the super fast complex is speculative without stoichiometry data, two arguments allow us to favor the possibility that the protein binds DNA as a dimer. First, a similar super fast complex was also observed with the CysB N309Ter variant, whose other characteristics suggested an oligomerization defect. Second, another LysR family protein, OxyR, may serve as a precedent, as its wild-type form is tetrameric, but the constitutive A233V variant binds DNA as a dimer (9). The fact that CysB c A227D was fully functional in vivo also suggests that the tetrameric structure of CysB may be nonessential for the activatory function of the protein.
Four mutations affecting the response of CysB to acetylserine fit in regions 95-173 and 196 -206, respectively (Fig. 1), suggested to be responsible for the inducer response in several LTTRs (2). Three remaining mutations were localized closer to the C-terminal part of the protein. The importance of this additional region (residues 227-255) for inducer recognition/ response is also illustrated by mutations described more recently for OxyR (constitutive variants A233V/T and G253K) (4) and NahR (variant R248C) (5). The amino acid residues whose substitution results in either non-inducible or constitutively active CysB variants are presented in the model of E. coli CysB (Fig. 7). As can be seen, all these residues (perhaps with the exception of Met 160 ) are localized in the neighborhood of the cavity, which is believed to be an acetylserine-binding pocket (20).
We performed molecular modeling to compare the structure of the WT CysB dimer with that of either non-inducible or constitutive CysB variants. However, the observed differences were not of essential meaning, so one cannot expect fundamental structural changes in the protein caused by the particular mutations studied in this work.
Regions Important for Oligomerization-The standard test for negative trans-dominance (42) is frequently used to identify residues involved in oligomerization of multimeric transcrip-tional regulators. Examples of LysR family members, NahR (7), OxyR (9), and GcvA (36), also suggest the correlation between the oligomerization defect and the inability of the mutant protein to exert a dominant-negative effect (also called a "poisoning effect") on the activity of the wild-type counterpart produced by the cell. For CysB, the negative dominance was clearly seen with some non-repressing (non-binding) variants (with mutations in region [11][12][13][14][15][16][17][18][19][20][21][22] as well as with non-inducible variants (Fig. 1). In contrast, three non-repressing mutants (substitutions in region 41-48) and all of the C-terminally truncated CysB variants did not display a dominant-negative effect. One truncated CysB variant (CysB N309Ter) bound to the cysP promoter, and the mobility of the resulting primary complex was consistent with a dimeric rather than a tetrameric form of the protein. The most likely interpretation of these findings indicates the involvement of the 16 C-terminal residues in oligomerization of CysB (as was also suggested for OxyR and NahR), but also the importance of a region encompassing Leu 41 -Ile 48 . It is tempting to speculate that region 41-48, composing part of the DNA-binding domain, is important for appropriate contacts between two CysB monomers whose HTH motifs interact directly with dyadic activatory binding sites. An alternative explanation for the lack of a dominant-negative effect of the E41K, L44L, and I48T variants  (20) and is shown in ball and stick representation in green. Residues identified in this study whose mutations led to non-inducible CysB proteins are indicated in blue. Residues whose alterations resulted in constitutive activity of CysB are shown in yellow; among these, residues 149 and 166 investigated previously (41) are also included.
assumes that the mutant proteins possess the ability to oligomerize, but that mixed CysB oligomers are still functional (i.e. the wild-type monomers are dominant to the mutant monomers with respect to DNA binding). Another suggestion emerging from our study is that interactions between subunits within the CysB tetramer may be affected by an interaction with the inducer. Such a possibility is inferred by the characteristics of the CysB c A227D variant, most probably dimeric as discussed earlier. A model proposed for WT CysB (37) assumed that C-terminal portions of the four subunits maintain the tetramer in an inactive state and that acetylserine disrupts these contacts to release the protein surface for interaction with the RNA polymerase. The characteristics of the CysB c A227D variant would be consistent with this model if one assumes that the A227D mutation disrupts C-terminal contacts between subunits and that the resulting protein is active without an inducer. It has to be pointed out that if both CysB c A227D and CysB N309Ter bind DNA as dimers, the arrangement of subunits in these dimers may not be equivalent. The electrophoretic mobilities of the primary complexes formed by these proteins with the cysP promoter were similar but not identical, and the N309Ter variant (in contrast to CysB c A227D) apparently was not active with all CysB-dependent cys gene promoters in vivo, as it was unable to complement the ⌬cysB mutant. Perhaps the characteristics of the above CysB variants closely resemble those of dimeric OxyR mutants, for which the possibility of dimerization via either the "dimerization domain" or the "tetramerization domain" was suggested (9).
Region Involved in the Positive Control Function-The single amino acid substitution Y27G in CysB (Fig. 1) resulted in a phenotype expected for a CysB pc mutant: the Y27G protein was able to bind DNA and to respond to the inducer by a qualitative change in the DNA-protein complex, but it was still defective in transcriptional activation. These results confirm that an additional function beyond DNA binding and bending (more precisely, inducer-dependent release of DNA bending caused by CysB alone) is required for CysB to act as a transcriptional activator. By its characteristics, the CysB Y27G variant corresponds to the GcvA pc F31L/A mutants (36), raising the possibility that in LysR-type proteins, an "activation patch" composes part of the N-terminal HTH region. A similar position of the putative activating region (within or nearby the HTH motif of the DNA-binding domain) was demonstrated for transcriptional regulators belonging to other families, i.e. cI (43) and FIS (44) proteins.
CysB protein is likely to contact the C-terminal domain of the RNA polymerase ␣-subunit, as the rpoA341 mutation, changing K271E in the ␣-peptide, specifically affects CysB-dependent expression of the cysPTWA operon (45). It is possible that recruitment of RNA polymerase by CysB involves a direct protein-protein contact between the C-terminal domain of the RNA polymerase ␣-subunit and the putative activating region of CysB, in which Thr 27 seems to be essential.
In conclusion, the results of this study highlight the regions of CysB that are critical for DNA binding, inducer response, oligomerization, and positive control and demonstrate that these functions can be genetically separated. Further work will continue to focus on the purification and detailed biochemical characterization of selected CysB variants for better recogni-tion of CysB subunit interactions. Isolation of additional mutants for further characterization of the CysB activating region is also in progress.