Contacts between the LexA repressor--or its DNA-binding domain--and the backbone of the recA operator DNA.

Using hydroxyl radical footprinting and ethylation interference experiments, we have determined the backbone contacts made by the entire LexA repressor and its amino-terminal fragment with the recA operator DNA. These techniques reveal essentially the same contacts between both proteins and one side of the DNA helix if one assumes that the DNA stays in the normal B-conformation. This result is somewhat unexpected because protection of guanine bases against methylation suggested a somewhat twisted recognition surface. The backbone contacts revealed by both methods are symmetrically disposed with respect to the center of the operator, providing further evidence that the operator binds two LexA monomers. Each half-operator contains seven interfering phosphates. These phosphates are found on both sides of the 5'-CTGT sequence that is believed to be the principal recognition target. On the side close to the center of the operator are found two phosphates, whereas the other five are clustered on the side apart from the dyad axis. We are not aware of such an extended cluster of interfering phosphates for any other DNA-binding protein. A quantification of the hydroxyl radical footprints allowed us to compare further the affinity of the LexA repressor for the recA operator with that of its isolated DNA binding domain. We find an only 13-fold higher binding constant for LexA than for its amino-terminal domain, which is in good agreement with our earlier results for the uvrA operator using a completely different binding assay.     

Repression of the E coli recA gene requires at least two LexA protein monomers

To analyze the DNA binding domain of E coli LexA repressor and to test whether the repressor binds as a dimer to DNA, negative dominant lexA mutations affecting the binding domain have been isolated. A large number of amino acid substitutions between amino acid positions 39 and 46 were introduced using cassette mutagenesis. Mutants defective in DNA binding were identified and then examined for dominance to lexA+. A number of substitutions weakened repressor function partially, whereas other substitutions led to a repressor with no demonstrable activity and a defective dominant phenotype. Since the LexA binding site has dyad symmetry, we infer that this dominance results from interaction of monomers of wild-type LexA protein with mutant monomers and that an oligomeric form of repressor binds to operator. The binding of LexA protein to operator DNA was investigated further using a mutant protein, LexA408, which recognizes a symmetrically altered operator mutant but not wild-type operator. A mixture of mutant LexA408 and LexA+ proteins, but neither individual protein, bound to a hybrid recA operator consisting of mutant and wild-type operator half sites. These results suggest that at least 1 LexA protein monomer interacts with each operator half site. We discuss the role of LexA oligomer formation in binding of LexA to operator DNA.     

A model for the LexA repressor DNA complex

A structural model for the interaction of the LexA repressor DNA binding domain (DBD) with operator DNA is derived by means of Monte Carlo docking. Protein–DNA complexes were generated by docking the LexA repressor DBD NMR solution structure onto both rigid and bent B-DNA structures while giving energy bonuses for contacts in agreement with experimental data. In the resulting complexes, helix III of the LexA repressor DBD is located in the major groove of the DNA and residues Asn-41, Glu-44, and Glu-45 form specific hydrogen bonds with bases of the CTGT DNA sequence. Ser-39, Ala-42, and Asn-41 are involved in a hydrophobic interaction with the methyl group of the first thymine base. Residues in the loop region connecting the two β-sheet strands are involved in nonspecific contacts near the dyad axis of the operator. The contacts observed in the docked complexes cover the entire consensus CTGT half-site DNA operator, thus explaining the specificity of the LexA repressor for such sequences. In addition, a large number of nonspecific interactions between protein and DNA is observed. The agreement between the derived model for the LexA repressor DBD/DNA complex and experimental biochemical results is discussed. © 1995 Wiley-Liss, Inc.     

Isolation and characterization of LexA mutant repressors with enhanced DNA binding affinity.

The LexA repressor from Escherichia coli is a sequence-specific DNA binding protein that shows no pronounced sequence homology with any of the known structural motifs involved in DNA binding. Since little is known about how this protein interacts with DNA, we have selected and characterized a great number of intragenic, second-site mutations which restored at least partially the activity of LexA mutant repressors deficient in DNA binding. In 47 cases, the suppressor effect of these mutations was due to an Ind- phenotype leading presumably to a stabilization of the mutant protein. With one exception, these second-site mutations are all found in a small cluster (amino acid residues 80 to 85) including the LexA cleavage site between amino acid residues 84 and 85 and include both already known Ind- mutations as well as new variants like GN80, GS80, VL82 and AV84. The remaining 26 independently isolated second-site suppressor mutations all mapped within the amino-terminal DNA binding domain of LexA, at positions 22 (situated in the turn between helix 1 and helix 2) and positions 57, 59, 62, 71 and 73. These latter amino acid residues are all found beyond helix 3, in a region where we have previously identified a cluster of LexA (Def) mutant repressors. In several cases the parental LexA (Def) mutation has been removed by subcloning or site-directed mutagenesis. With one exception, these LexA variants show tighter in vivo repression than the LexA wild-type repressor. The most strongly improved variant (LexA EK71, i.e. Glu71----Lys) that shows an about threefold increased repression rate in vivo, was purified and its binding to a short consensus operator DNA fragment studied using a modified nitrocellulose filter binding assay. As expected from the in vivo data, LexA EK71 interacts more tightly with both operator and (more dramatically) with non-operator DNA. A determination of the equilibrium association constants of LexA EK71 and LexA wild-type as a function of monovalent salt concentration suggests that LexA EK71 might form an additional ionic interaction with operator DNA as compared to the LexA wild-type repressor. A comparison of the binding of LexA to a non-operator DNA fragment further shows that LexA interacts with the consensus operator very selectively with a specificity factor of Ks/Kns of 1.4 x 10(6) under near-physiological salt conditions.     

The LexA repressor and its isolated amino-terminal domain interact cooperatively with poly[d(A-T)], a contiguous pseudo-operator, but not with random DNA: a circular dichroism study

The interaction of the entire LexA repressor and its amino-terminal DNA binding domain with poly[d(A-T)] and random DNA has been studied by circular dichroism. Binding of both protein species induces an about 2-fold increase of the positive circular dichroism band at about 270 nm of both polynucleotides, allowing a precise determination of the principal parameters as a function of mono- and divalent salt concentration and pH. Both proteins interact much more strongly (about 2000-fold) with poly[d(A-T)] than with random DNA as expected from the homology with the specific consensus binding site of LexA (CTGTATATATATACAG). For both LexA and its DNA binding domain we find that the interaction with poly[d(A-T)] is cooperative with a cooperativity factor omega of about 50-70 for both proteins over a wide range of solvent conditions, suggesting that the carboxy-terminal domain of LexA is not involved in this type of cooperativity. On the contrary, no cooperativity could be detected for the interaction of the LexA DNA binding domain with a random DNA fragment. The overall binding constant K omega (or simply K in the case of random DNA) depends strongly on the salt concentration as observed for most protein-DNA interactions, but the behavior of LexA is unusual in that the steepness of this salt dependence (delta log K omega/delta log [NaCl]) is much more pronounced at slightly acidic pH values as compared to that at neutral or slightly alkaline pH. The behavior is not easily understood in terms of the current theories on the electrostatic contribution to protein-DNA interactions on the basis of polyelectrolyte theory. A comparison of the overall binding constant K omega of the entire LexA repressor and its DNA binding domain reveals that LexA binds only 20-50-fold stronger under a wide variety of salt and pH conditions. This result tends to demonstrate further that the additional energy due to the dimerization of LexA via the carboxy-terminal domain should be rather weak as expected from the small dimerization constant of LexA (2 X 10(-4) M-1).     

1H-NMR investigation of the interaction of the amino terminal domain of the LexA repressor with a synthetic half-operator.

A synthetic half-operator DNA-duplex, d(GCTACTGTATGT), containing a portion of the proposed recognition sequence (CTGT) of several "SOS" genes, has been synthesized. The dodecamer has been characterized through 1H-NMR spectroscopy. Complete assignment of exchangeable hydrogen bonded imino protons has been achieved by applying 1D NOE techniques and an analysis of the temperature dependence of the chemical shifts. In order to determine the specific role of the CTGT consensus sequence in the overall recognition process, the oligonucleotide duplex has been titrated with the amino terminal DNA binding domain of the LexA repressor. The observation of substantial changes of 1H-NMR chemical shifts in the imino proton region upon interaction with the protein strongly suggests that the protein binds specifically to the operator DNA. The largest deviations of 1H-NMR chemical shifts upon protein binding have been observed for protons assigned to the CTGT segment, thus strongly suggesting a direct involvement of this sequence in the binding process. At high potassium chloride concentrations the 1H-NMR chemical shift deviations are reverted which is consistent with the known drop in the affinity constant of LexA for operator DNA at high salt concentrations.     

DNA-binding site of lac repressor probed by dimethylsulfate methylation of lac operator.

In order to compare the structures of the DNA-binding sites on variants of the lac repressor, we have studied the influence of these variants on the dimethylsulfate methylation of the lac operator. Since a bound protein changes the availability of specific purines in the operator to this chemical attack, comparisons of the methylation patterns will show similarities or differences in the protein DNA contacts. We compared lac repressor, induced lac repressor (repressor bound to the gratuitous inducer isopropyl-β-d-thiogalactoside), mutant repressors having increased operator affinities (X86, I12 and the X86-I12 double mutant) and repressor peptides (long headpiece, residues 1 to 59 and short headpiece. residues 1 to 51). All of these repressors and repressor peptides exhibit the same general pattern of protection and enhancement in the operator; however, the short headpiece pattern differs most from that of the repressor while the induced repressor and the long headpiece show intermediate patterns that are strikingly similar to each other. The mutant repressors do not show an isopropyl-β-d-thiogalactoside effect but otherwise are almost indistinguishable from wild-type repressor. These results demonstrate that all molecules bind to the operator using basically the same protein-DNA contacts; they imply that (1) most and possibly all repressor contacts to operator lie within amino acids 1 to 51, (2) inducer weakens many contacts rather than totally disrupting one or even a few and (3) the tight-binding mutants do not make additional contacts to the DNA.These results are consistent with a model in which the amino-terminal portions of two repressor monomers make the DNA contacts. We show that one can understand the affinity of binding as related to the accuracy of the register of the two amino-terminal portions along the DNA. Furthermore, the action of inducer and the behaviour of the tight binding mutants can be accomodated within a two-state model in which the strongly or weakly binding states correspond to structures in which the amino-terminal regions are rigidly or loosely held with respect to each other.     

Effect of alkylating agents on the expression of inducible genes of Escherichia coli

Increasing doses of alkylating agents such as N-methyl-N'-nitro-N-nitrosoguanidine, diethyl sulphate and ethylmethane sulphonate cause an inhibition of the expression of the recA and sfiA genes of wild-type Escherichia coli. This behaviour was not observed in a lexA56 mutant which has a defective LexA repressor that is unable to bind to the SOS operator. Furthermore, an ada-1 mutant showed the same behaviour as the wild-type strain indicating that the adaptive proteins are not responsible for the inhibition of recA and sfiA at high doses of alkylating agents. These results suggest that the inhibitory effect of these alkylating agents may be found in the interaction between the LexA repressor and the control regions of sfiA and recA. On the other hand, high doses of either UV light or mitomycin C produced only a slight decrease in the induction of recA and sfiA, whereas bleomycin had no effect. The fact that a repressor structurally related to LexA repressor, such as LacI protein, showed the same behaviour as the LexA repressor when a Lac+ strain was treated with alkylating agents, suggests that these compounds can modify the binding abilities of repressors to DNA, producing a limited or even abolished release of repressors, and so decreasing the expression of inducible genes.     

Binding of the Bacillus subtilis LexA protein to the SOS operator

The Bacillus subtilis LexA protein represses the SOS response to DNA damage by binding as a dimer to the consensus operator sequence 5′-CGAACN₄GTTCG-3′. To characterize the requirements for LexA binding to SOS operators, we determined the operator bases needed for site-specific binding as well as the LexA amino acids required for operator recognition. Using mobility shift assays to determine equilibrium constants for B.subtilis LexA binding to recA operator mutants, we found that several single base substitutions within the 14 bp recA operator sequence destabilized binding enough to abolish site-specific binding. Our results show that the AT base pairs at the third and fourth positions from the 5′ end of a 7 bp half-site are essential and that the preferred binding site for a LexA dimer is 5′-CGAACATATGTTCG-3′. Binding studies with LexA mutants, in which the solvent accessible amino acid residues in the putative DNA binding domain were mutated, indicate that Arg-49 and His-46 are essential for binding and that Lys-53 and Ala-48 are also involved in operator recognition. Guided by our mutational analyses as well as hydroxyl radical footprinting studies of the dinC and recA operators we docked a computer model of B.subtilis LexA on the preferred operator sequence in silico. Our model suggests that binding by a LexA dimer involves bending of the DNA helix within the internal 4 bp of the operator.     

Bending of synthetic bacteriophage 434 operators by bacteriophage 434 proteins.

The extent of DNA bending induced by 434 repressor, its amino terminal DNA binding domain (R1-69), and 434 Cro was studied by gel shift assay. The results show that 434 repressor and R1-69 bend DNA to the same extent. 434 Cro-induced DNA bends are similar to those seen with the 434 repressor proteins. On approximately 265 base pair fragments, the cyclic AMP receptor protein of Escherichia coli (CRP) produces larger mobility shifts than does 434 repressor. This indicates that the 434 proteins bend DNA to a much smaller extent than does CRP. The effects of central operator sequence on intrinsic and 434 protein-induced DNA bending was also examined by gel shift assay. Two 434 operators having different central sequences and affinities for 434 proteins display no static bending. The amount of gel shift induced by 434 repressor on these operators is identical, showing that the 434 repressor bends operators with different central sequences to the same extent. Hence, mutations in the central region of the operator do not influence the bent structure of the unbound or bound operator.     

Functional domains of the penicillinase repressor of Bacillus licheniformis.

The penicillinase repressor (PENI) negatively regulates expression of the penicillinase gene (penP) in Bacillus licheniformis by binding to its operators located within the promoter region of penP.penI codes for a protein with 128 amino acids. Filter-binding analyses suggest that the active form of the repressor is a dimer. Genetic analyses of PENI derivatives showed that the repressor carrying either a 6-amino-acid deletion near the N terminus or a 14-amino-acid deletion at the C terminus was functionally inactive in vivo. A repressor derivative carrying a 6-amino-acid deletion within its N-terminal region was extensively purified and used in DNA footprinting and subunit cross-linking analyses. The results of these studies showed that the repressor derivative had lost its ability to bind operator specifically even though it could dimerize effectively. In similar studies, we demonstrated that an N-terminal portion of PENI with a molecular mass of 10 kDa derived by digestion with papain was able to bind operator specifically but with reduced affinity and had completely lost its ability to dimerize. These data suggest that the repressor has two functional and separable domains. The amino-terminal domain of the repressor is responsible for operator recognition, and the carboxyl-terminal domain is involved in subunit dimerization.     

A comparative three-dimensional model of the carboxy-terminal domain of the lambda repressor and its use to build intact repressor tetramer models bound to adjacent operator sites

A model for residues 93-236 of the lambda repressor (1gfx) was predicted, based on the UmuD(') crystal structure, as part of four intact repressor molecules bound to two adjacent operator sites. The structure of region 136-230 in 1gfx was found to be nearly identical to the independently determined crystal structure of the 132-236 fragment, 1f39, released later by the PDB. Later, two more tetrameric models of the lambda repressor tetramer bound to two adjacent operator sites were constructed by us; in one of these, 1j5g, the N-domain and C-domain coordinates and hence monomer-monomer and dimer-dimer interactions are almost the same as in 1gfx, but the structure of the linker region is partly based on the linker region of the LexA dimer in 1jhe; in the other, 1lwq, the crystalline tetramer for region 140-236 has been coopted from the crystal structure deposited in 1kca, the operator DNA and N-domain coordinates of which are same as those in 1gfx and 1j5g, but the linker region is partly based on the LexA dimer structures 1jhe and 1jhh. Monomer-monomer interactions at the same operator site are stabilized by exposed hydrophobic side chains in beta-strands while cooperative interactions are mostly confined to beta(6) and some adjacent residues in both 1gfx and 1j5g. Mutational data, existence of a twofold axis relating two C-domains within a dimer, and minimization of DNA distortion between adjacent operator sites allow us to roughly position the C-domain with respect to the N-domain for both 1gfx and 1j5g. The study correlates these models with functional, biochemical, biophysical, and immunological data on the repressor in the literature. The oligomerization mode observed in the crystal structure of 132-236 may not exist in the intact repressor bound to the operator since it is shown to contradict several published biochemical data on the intact repressor.     

Large-scale purification, oligomerization equilibria, and specific interaction of the LexA repressor of Escherichia coli

A rapid large-scale procedure for the purification of the LexA repressor of Escherichia coli is described. This procedure allows one to get more than 100 mg of purified protein from 100 g of bacterial paste with a purity of at least 97%. This method is comparable to earlier, far more complicated purification procedures giving clearly smaller yields. It is shown that the LexA protein may be identified spectroscopically by a large A235/A280 ratio and very pronounced ripples in the absorption spectrum arising from a high amount of phenylalanine residues with respect to that of the other aromatic amino acids. Polyacrylamide gel electrophoresis has been used to study the specific interaction of LexA with a recA operator fragment. The quaternary structure of LexA has been studied by equilibrium ultracentrifugation and sedimentation velocity measurements. The sedimentation coefficient increases with increasing LexA concentration, indicating that LexA is involved in self-association. This finding has been confirmed by equilibrium ultracentrifugation. The results are best described by a monomer-dimer and a subsequent dimer-tetramer equilibrium, with an association constant of 2.1 X 10(4) M-1 for the dimer and 7.7 X 10(4) M-1 for the tetramer formation. These relatively small association constants determined under near-physiological pH and salt conditions suggest that in vivo LexA should be essentially in the monomeric state. The degree to which LexA decreases the electrophoretic mobility of a 175 base pair fragment harboring the recA operator suggests that the recA operator interacts nevertheless with a LexA dimer. However, our results may be also explained by the binding of a LexA monomer with a simultaneous bending of the DNA fragment.