The Mutagenesis Protein MucB Interacts with Single Strand DNA Binding Protein and Induces a Major Conformational Change in Its Complex with Single-stranded DNA*

The MucA and MucB proteins are plasmid-encoded homologues of the Escherichia coli UmuD and UmuC proteins, respectively. These proteins are required for SOS mutagenesis, although their mechanism of action is unknown. By using the yeast two-hybrid system we have discovered that MucB interacts with SSB, the single strand DNA binding protein (SSB) of E. coli. To examine the interaction at the protein level, the MucA, MucA′, and MucB proteins were overproduced, purified in denatured state, and refolded. Purified MucA and MucA′ each formed homodimers, whereas MucB was a monomer under native conditions. RecA promoted the cleavage of MucA to MucA′, and MucB was found to bind single-stranded DNA (ssDNA), similarly to the properties of the homologous UmuD and UmuC proteins. Purified MucB caused a shift in the migration of SSB in a sucrose density gradient, consistent with an interaction between these proteins. Addition of MucB to SSB-coated ssDNA caused increased electrophoretic mobility of the nucleoprotein complex and increased staining of the DNA by ethidium bromide. Analysis of radiolabeled SSB in the complexes revealed that only a marginal release of SSB occurred upon addition of MucB. These results suggest that MucB induces a major conformational change in the SSB·ssDNA complex but does not promote massive release of SSB from the DNA. The interaction with SSB might be related to the role of MucB in SOS-regulated mutagenesis.

UV mutagenesis in Escherichia coli is a regulated process, controlled by the SOS stress response through its two global regulators, RecA and LexA. The mechanism underlying this process is trans-lesion replication by a DNA polymerase, most likely DNA polymerase III (for reviews see Refs. [1][2][3]. This process requires specifically two SOS-induced proteins, UmuD and UmuC (4 -6), whose mechanism of action is unknown. A prevailing hypothesis is that UmuDЈ, the active form of UmuD (7)(8)(9), along with UmuC are required to assist the DNA polymerase in replicating the damaged site (10). However, the nature of this assistance is not clear because purified DNA polym-erases can bypass DNA lesions unassisted (11)(12)(13)(14)(15)(16)(17)(18)(19). Moreover, in an in vitro system for UV mutagenesis carried out with crude protein extracts (20,21) or with purified proteins (22), we have found that UV mutations were effectively produced in the absence of UmuDЈ and UmuC.
Homologues of UmuD and UmuC have been identified in other bacteria, and some of them are encoded by conjugational plasmids (23,24). The most well-studied of these are the mucA and mucB genes, encoding homologues of UmuD and UmuC, respectively (25,26). An approach that has proven useful in the study of many proteins is to examine their interactions with other proteins. Employing this approach we present here in vivo and in vitro data that show, for the first time, that MucB interacts with SSB 1 and greatly changes the structure of the SSB⅐ssDNA complex.
Media-The media used in this study were M9, LB, and minimal A medium (27). Medium M9ZB was used to grow E. coli BL21(DE3) for overproduction of the Muc proteins (28). Antibiotics used were ampicillin (100 mg/ml) and kanamycin (70 mg/ml). Isopropyl-1-thio-␤-D-galactopyranoside was added to a final concentration of 0.4 mM.
Plasmids and Other DNAs-The plasmids used in this study are presented in Table I. Plasmid p1-66 is a pBluescript SK ϩ derivative in which a 1.1-kilobase pair HincII-HincII fragment containing mucA and a part of mucB ( Fig. 1) from pGW1700 was inserted to the ApaI site in pBluescript SK ϩ under of the T7 RNA polymerase promoter. Plasmid p2 contains the same fragment in the opposite orientation (under the lac promoter). Plasmid p1-66-3 contains the entire mucAB operon in plasmid pBluescript SK ϩ under the T7 promoter. It was constructed by ligating the 1.2-kilobase pair BglII-AvaI fragment containing the 3Јterminal part of mucB ( Fig. 1)  Plasmid p1-66MucA1/B is similar to p1-66MucA1, except that the missing 3Ј-terminal fragment of mucB was added, to reconstitute the entire mucAB operon, with an NdeI site in front of its first ATG. Plasmid p1-66MucB1 is plasmid p1-66 with an NdeI site in front of the first ATG of mucB. Plasmid p1-66MucB is similar to p1-66MucB1, except that the 1.2-kilobase pair BglII-AvaI fragment containing the 3Ј-terminal portion of mucB was added, to reconstitute the entire mucB gene, with the NdeI site in front of it. Plasmid pETMucA is the overproducer of mucA. It was constructed by ligating the NdeI-BamHI fragment from p1-66MucA1 to plasmid pET3a that was digested with NdeI and BamHI, putting mucA under the T7 10 promoter. pETMucB is the overproducer of mucB. It was prepared by inserting the NdeI-BamHI fragment from p1-66MucB into to NdeI-and BamHI-cleaved pET3a. Plasmid pETMucAЈ is the MucAЈ overproducer. It was constructed by PCR amplification of mucAЈ from plasmid p1-66MucA1/B using primers 70 and 71 (see Table IV), followed by cleavage of the PCR fragment with NdeI and BamHI, and ligation to NdeI-and BamHIcleaved pET3a. The plasmids used in the two-hybrid system are presented in Table II. They were constructed by cloning PCR-amplified open reading frame of the tested genes into plasmids pACTII and pAS-CYH2 as described in Table III, using the primers described in  Table IV, which contained the restriction sites needed for cloning. Single-stranded DNA of phage M13mp8 was isolated as described (29). The 40-nucleotide long oligodeoxyribonucleotide used for the gel mobility shift assay was 5Ј-GGAAAACCCTGGCGTTACCCAACTTAATCGC-CTTGCAGCA-3Ј. All the oligonucleotides were synthesized by the Biological Services Department of the Weizmann Institute.
Yeast Methods-Yeast transformation was done using the lithium acetate method (30). Detection of ␤-galactosidase activity was done on filters as described (31).
Purification of MucA and MucAЈ-E. coli BL21(DE3) cells harboring plasmid pETMucA (1 liter) were grown at 37°C in a 5-liter flask with constant stirring. The cells were treated with isopropyl-1-thio-␤-D-galactopyranoside for 2 h to induce the synthesis of MucA. The cells were then centrifuged, resuspended in an equal volume of ST buffer containing 50 mM Tris⅐HCl, pH 7, and 15% sucrose, and frozen in liquid nitrogen. The cell suspension was thawed at 15°C, and its volume was increased to 100 ml with ST buffer. Cells were then disrupted by a 5-min sonication period in a Soniprep MSE sonicator equipped with a medium tip at 3 ⁄4 power. The lysate was spun at 25,000 rpm for 1 h in a Ti45 rotor. The supernatant was discarded, and the pellet, containing MucA inclusion bodies, was homogenized in water using a 40-ml glass homogenizer. The homogenate was spun at 15,000 rpm for 30 min in a Ti45 rotor and washed again with water. The pellet was homogenized again and treated with chicken egg lysozyme (0.3 mg/ml) for 1 h at 15°C. It was then treated with 2% sodium deoxycholate for 1 h at 15°C. The inclusion bodies were denatured with a buffer containing 7 M urea, 50 mM Tris⅐HCl, pH 7.4, 10 mM DTT, 1 mM EDTA, and 10 mM NaCl (buffer A). The solution of denatured protein was incubated for 4 h at 4°C with constant stirring, after which it was cleared by centrifugation at 4°C in a Ti45 rotor at 40,000 rpm for 2 h. Denatured MucA (70 mg) was passed through a phosphocellulose column (Whatman P11) in buffer A, eluting in the flow-through. This was done to remove contaminating DNA binding proteins, such as nucleases that bind phosphocellulose. MucA was then loaded on a DEAE-Sephacel (Pharmacia) column equilibrated with 7 M urea, 50 mM Tris⅐HCl, pH 7.5, 10 mM DTT, 0.1 mM EDTA, and 10 mM NaCl (buffer B), washed with 2 column volumes of buffer B, and eluted with 10 volumes of a linear gradient of 0 -1 M NaCl in buffer B. MucA eluted at 200 mM NaCl. Next MucA (22 mg) was purified on a 200-ml Sephadex G-100 column in buffer B containing 200 mM NaCl. MucA-containing fractions were collected, diluted to a concentration of 0.1 mg/ml protein, and dialyzed against a refolding buffer containing 0.1 M NaCl, 50 mM Tris⅐HCl, 0.1 mM EDTA, 10 mM ␤-mercaptoethanol, and 10% glycerol (buffer C). After dialysis the solution was cleared by a 2-h spin at 200,000 ϫ g. The supernatant contained the soluble MucA (80%; 13 mg). The purity was greater than 95% as estimated by Coomassie Blue staining. MucAЈ was purified from BL21(DE3) cells carrying plasmid pETMucAЈ using the same procedure.
Purification of MucB-E. coli BL21(DE3) cells harboring plasmid pETMucB were grown and induced like the MucA overproducing cells. The overproduced protein formed inclusion bodies, and those were purified like the MucA inclusion bodies, except that no lysozyme was used. The inclusion bodies were denatured in a buffer containing 8 M urea, 50 mM Tris⅐HCl, pH 7.4, 10 mM DTT, and 1 mM EDTA. The solution of denatured protein was incubated for 2 h at 4°C with constant stirring. The solution was then cleared at 40,000 rpm for 2 h in a Ti45 rotor at 4°C. The denatured MucB was dialyzed against a buffer containing 50 mM Tris⅐HCl, 0.1 mM EDTA, 10 mM ␤-mercaptoethanol, 10% glycerol, and 7 M urea (buffer D), and 1 mg of protein was passed through a 30-ml DEAE-Sephacel (Pharmacia) column equilibrated with buffer D, and collected in the flow-through. MucB was then loaded on a 10-ml phosphocellulose column (Whatman P11) in buffer D, washed with 2 column volumes of buffer D, and eluted with 10 volumes of a linear gradient of 0 -400 mM NaCl in buffer D. MucB eluted at 200 mM NaCl. The fractions containing MucB were diluted to a concentration of 0.05 mg/ml protein and dialyzed against a refolding buffer containing 50 mM Tris⅐HCl, 0.1 mM EDTA, 10 mM ␤-mercaptoethanol, and 10% glycerol (buffer E). After dialysis the solution was cleared by a 2-h spin at 200,000 ϫ g. The supernatant contained 0.15 mg of soluble MucB (15% yield). The purity was 85-90% as estimated by Coomassie Blue staining. MucB was concentrated using a Centricon concentrator (Amicon) to 0.12 mg/ml.
Other Proteins-SSB was purified as described previously (32). DNA polymerase I, T4 DNA ligase, and alkaline phosphatase were purchased from Boehringer Mannheim. Restriction endonucleases were purchased   Radiolabeling of SSB-SSB was radiolabeled with NaB[ 3 H] 4 as described (33). A mixture (100 l) of SSB (100 M), H 3 BO 3 -NaOH (0.2 M, pH 9.0), formaldehyde (5 mM), and NaB[ 3 H] 4 (170 M, 24 Ci/mmol; Amersham Corp.) was incubated at room temperature for 25 min, after which the reaction was terminated by the addition of 100 l of Tris⅐HCl (50 mM, pH 7.9). The radiolabeled SSB was separated from the unreacted NaB[ 3 H] 4 by a brief spin at 480 g in a 0.8-ml Sephadex G-50 mini-column, followed by dialysis against 25 mM Tris⅐HCl, pH 8.0, 1 mM EDTA, and 0.3 M NaCl.
Sucrose Gradients-Fifteen g of 3 H-labeled SSB and 5 g of bovine serum albumin or 5 g of MucB were loaded on a 5-ml 5-20% sucrose gradient containing 2 mM MgCl 2 , 2 mM DTT, and 25 mM Tris⅐HCl, pH 7.5. The gradient was run for 8 h at 49,000 rpm in a SW 50.1 rotor at 4°C. Fractions of 200 l were collected from the top and analyzed by scintillation counting.
Gel Mobility Shift Assay (31)-Purified MucB and MucAЈ proteins (2.7 and 27 pmol, respectively) were incubated with 32 P 5Ј-end-labeled 40-nucleotide long ssDNA (2 pmol) in a buffer containing 20 mM Tris⅐HCl, pH 7.5, 1 mM EDTA, and 6% glycerol in a total volume of 15 l at 4°C for 15 min. Then 15 l of 12% glycerol were added, and the mixture was fractionated by electrophoresis on a 6% native polyacrylamide gel in 4.5 mM Tris⅐HCl, 0.12 mM EDTA, 4.5 mM boric acid, pH 8.3, at 4°C for 1 h at 25 mA and 180 V. The binding of SSB to ssDNA was assayed by an agarose gel mobility shift assay as described (34). The binding mixture contained 25 mM Tris⅐HCl, pH 7.5, 10 mM DTT, 10% glycerol, 0.1 mM EDTA, M13mp8 ssDNA (230 fmol), and SSB. The mixtures were incubated for 15 min at 4°C, after which they were loaded on a 1% agarose gel containing ethidium bromide and run in a buffer containing 40 mM Tris⅐acetate, 1 mM EDTA, pH 8.1, at 40 V for 3 h. The gel was exposed to UV light and photographed.
Autoradiography of Radiolabeled Nucleic Acids and Proteins-Gels containing radiolabeled RNA or DNA were autoradiographed at room temperature using a Kodak XAR-5 x-ray film. Alternatively, the gels were dried and fluorographed with an intensifying screen at Ϫ80°C. Gels containing radiolabeled SSB were treated with amplifier (Amersham Corp.) at room temperature for 30 min, dried onto a DE81 filter, and fluorographed with an intensifying screen at Ϫ80°C using Kodak XAR-5 x-ray film.

Protein-Protein Interactions of the Muc Proteins in Vivo-
To gain a molecular insight into the mechanism of action of the mutagenesis proteins MucA and MucB, we sought to identify their involvement in novel protein-protein interactions using the yeast two-hybrid system. First, we examined self-interactions of the Muc proteins fused to the two-hybrid system re-porter proteins. Each of the muc genes was fused to the transcriptional activation domain or to the DNA binding domain of the yeast GAL4 gene, and pairs of the fused genes were introduced into yeast. The interactions among various combinations of the fused Muc proteins were tested using the filter assay for ␤-galactosidase (31). The results shown in Table V suggest that MucA and MucAЈ interact with themselves and with each other. MucB did not show self-interaction, but it did interact with both MucA and MucAЈ. These interactions parallel those of the homologous UmuD and UmuDЈ proteins (35,36). A truncated form of MucB (MucB⌬) lacking 30 amino acid residues from its C terminus was also fused to the GAL4 activation domain. This truncated protein did not interact with either MucA or MucAЈ in the two-hybrid system ( Table V), suggesting that the C-terminal 30 amino acids are required for the interaction of MucB with MucA or MucAЈ.
We looked for interactions between RecA and the Muc proteins in the two-hybrid system (Table VI). When RecA was fused to the transcription activation domain of GAL4, we found that it interacted with MucA and MucAЈ. No interactions were found between RecA and MucB. RecA did not interact with MucA when the former was fused to the DNA binding domain of GAL4. This result is most likely due to an altered structure of RecA within the fusion protein that prevents MucA recognition and/or binding. These results demonstrated that the known interactions of Muc proteins can be detected in the two-hybrid system (37)(38)(39), rendering it a useful tool for the detection of novel interactions of the Muc proteins.
In Vivo Evidence for the Interactions of Muc Proteins with SSB-Since Umu/Muc-dependent mutagenesis requires DNA synthesis, we tested additional proteins, associated with DNA synthesis, for interactions with the Muc proteins. We tested SSB and DnaB, the major replicative DNA helicase. MucA, MucAЈ, and MucB were found to interact with SSB (Table VI).
No interactions were found between the Muc proteins and DnaB. Similarly, no interactions were found between the p53 protein and any of the other proteins that were examined. Interestingly, MucB⌬, which did not interact with MucA and MucAЈ, showed a strong interaction with SSB. The finding that SSB interacted with the Muc proteins suggested it as a novel target for the action of the Muc proteins. In addition, the binding of MucB⌬ to SSB suggested that MucB has at least two domains. One domain, located at the C terminus, binds MucA or MucAЈ, and the other domain binds SSB. The interactions observed in the yeast two-hybrid system need to be verified by biochemical experiments with purified proteins. To that end we have overproduced the Muc proteins, purified them, and examined their interaction with SSB.
Overproduction of the Muc Proteins-The Muc proteins were overproduced using the phage T7 expression system as follows. 1) An ApaI-HincII fragment from plasmid pGW1700 (25) containing mucA and a part of mucB was cloned into the ApaI site of plasmid pBluescript SK. 2) NdeI sites were created in front of the initiation ATG codons of mucA and mucB by site-directed mutagenesis. 3) A BglII-AvaI fragment containing the 3Ј end of mucB ( Fig. 1) was ligated into both constructs, and the in vivo activity of the cloned and mutated muc genes was verified by demonstrating their ability to restore UV mutability in a ⌬umuDC strain. 4) Finally, each gene was ligated separately to the pET3a vector to generate pETMucA and pETMucB that overproduce the Muc proteins. The mucAЈ gene fragment was amplified by PCR and cloned into pET3a to yield plasmid pETMucAЈ, which overexpresses MucAЈ. The muc genes in the three final overexpression plasmids were sequenced to ensure overproduction of the wild-type proteins.
Purification and Analysis of MucA-The overproduced MucA protein had the expected size of 16.5 kDa in SDS-PAGE (Fig.  2). N terminus analysis revealed that the sequence of its first 10 amino acids was identical to the amino acid sequence of MucA, as predicted from the nucleotide sequence of its gene (26). The overproduced MucA protein precipitated into inclusion bodies. Therefore, the purification procedure included the isolation of inclusion bodies, solubilization of the protein in 7 M urea, followed by chromatography in a denatured form on phosphocellulose, DEAE-Sephacel, and Sephadex G-100 columns.
Finally the protein was refolded by dialyzing out the urea. Typically, 15 mg of MucA were obtained from 1 liter of culture (Fig. 2).
The proper folding of the purified MucA protein was assayed by monitoring its ability to undergo RecA-promoted cleavage to MucAЈ. As can be seen in Fig. 3, RecA promoted the cleavage of MucA to a smaller protein, the latter showing the same electrophoretic mobility as MucAЈ. The cleavage was dependent on the presence of ssDNA, and on either ATP or its non-hydrolyzable analog ATP␥S, and could occur at alkaline conditions in the absence of RecA (data not shown). In fact, some autocleavage of MucA to MucAЈ occurred during purification and storage (Fig. 2). Thus it appears that our purified MucA was folded correctly.
Purification of MucAЈ-The overproduced MucAЈ protein had an apparent mass of 14 kDa on SDS-PAGE (Fig. 2). N-terminal amino acid sequence analysis revealed that the first 10 amino acids of the overproduced protein were identical to the sequence of MucAЈ, as predicted from the cleavage site of MucA by RecA (40). In addition we noted that the overproduced MucAЈ did not contain an N-terminal methionine, as expected from the additional ATG codon that we have engineered into the gene. It was probably removed in the cell. The overproduced MucAЈ precipitated into inclusion bodies. We have purified and refolded it using a procedure similar to that used for the purification of MucA. At this point our only indication that MucAЈ was folded properly was that it formed homodimers under native conditions (see below).
Determination of the Molecular Weight of Native MucA and MucAЈ Proteins-The molecular weight of native MucA was analyzed by HPLC size exclusion chromatography, using a TSK250 column. The elution of MucA was monitored by an UV detector at 214 nm. Fig. 4 (top) shows the elution profile of MucA from the HPLC column (retention time is 17.6 min). Based on calibration with HPLC protein standards, native MucA has a molecular mass of 25.7 kDa, representing most likely a dimer (the molecular mass of the monomer is 16.5 kDa). When resolved under the same conditions, MucAЈ (monomer size of 14.5 kDa) appeared to have an apparent molecular mass of 27.1 kDa (retention time of 17.5 min, Fig. 4, bottom). Therefore, MucAЈ is also, most likely, a dimer.
MucB⌬ is a truncated protein lacking 30 amino acid residues from its C terminus. BETA2 (␤-cell E-box transactivator 2) and E2A (helixloop-helix transcription factor E12 (60)) were used as controls. They interact with each other but not with any of the Muc proteins. The finding that MucAЈ migrates slightly faster than MucA on a size exclusion column can be attributed to a change in the shape of MucAЈ as compared with MucA. This is supported by information derived from the crystal structure of UmuDЈ, the MucAЈ homologue showing that UmuDЈ assumes an extended structure as compared with the intact UmuD (41). The finding that MucAЈ and MucA are dimers under native conditions suggests that these proteins were correctly refolded.
Purification and Analysis of MucB-The overproduced MucB protein had the expected size of 45 kDa on SDS-PAGE (Fig. 2). N-terminal amino acid sequence analysis revealed that the first 20 amino acids were identical to those predicted from the nucleotide sequence of the mucB gene (26). Like the other Muc proteins, the overproduced MucB precipitated into inclusion bodies. The protein was purified in the denatured state and was refolded. Total yield was 0.5 mg of MucB from 1 liter of culture. To our knowledge, this is the first report of the purification of MucB. Previously UmuC has been shown to bind ssDNA (42,43). Based on its homology to UmuC, MucB was expected to bind ssDNA as well. This was examined using the gel mobility shift assay. Incubation of purified MucB protein with a single-stranded 40-mer oligonucleotide resulted in the appearance of a slower migrating band (Fig. 5). This band represented most likely the MucB-bound oligonucleotide. Purified MucAЈ or MucA did not demonstrate any DNA binding activity and had no effect on the binding activity of MucB in this assay (Fig. 5). The ability of MucB to bind ssDNA suggested that it was folded correctly.
Analysis of MucB-SSB Interactions on Sucrose Gradients-To test whether MucB physically interacts with SSB, we sedimented 3 H-labeled SSB in a sucrose gradient in the presence or absence of MucB (Fig. 6). When MucB and SSB were run together in the sucrose gradient the migration rate of SSB in the gradient was slightly enhanced. No change in the migration rate of SSB was observed when bovine serum albumin or MucA was added (not shown). This suggests that the alteration in the sedimentation rate of SSB caused by MucB, although modest, resulted from specific binding of MucB to SSB, consistent with the results obtained with the two-hybrid system (Table IV). The small effect may be due to a low affinity of MucB to SSB. Alternatively, MucB may recognize preferentially the DNA-bound form of SSB, or it may form a ternary DNA⅐SSB⅐MucB complex.
MucB Induces a Major Conformational Change in SSB-ssDNA Complexes-To characterize further the interactions of MucB and SSB in the presence of ssDNA, we used a gel mobility shift assay. Binding of SSB to M13mp8 ssDNA retarded the migration of the DNA when run in a neutral agarose gel (Fig.  7). SSB has been shown to possess at least two modes of binding to ssDNA (34,44). In the fully cooperative mode, the entire DNA molecule is covered with SSB (each SSB tetrameter binding 35 nucleotides), leading to a strong retardation (e.g. the 3 M lane in Fig. 7). In the non-cooperative mode, only part of the DNA is covered with SSB leading to a milder retardation which increases with the amount bound to the DNA (the 0.8 -2 M lanes in Fig. 7 (34)). At lower SSB concentrations both modes of binding were observed. The binding of SSB seems to inhibit the binding of ethidium bromide to the ssDNA, since the fully shifted bands were stained to a lesser extent than the partially covered DNA molecules (Fig. 7 (34)).
When increasing amounts of MucB were added to a mixture of SSB and ssDNA, an increase in the intensity of the staining of the ssDNA by ethidium bromide was observed, along with an enhanced mobility of some of the DNA (Fig. 8). The addition of MucB in the absence of SSB, both with or without MucA and MucAЈ, did not have any affect on the migration of ssDNA (not shown). The effect of MucB on the migration of ssDNA in the presence of SSB might have resulted from the removal of SSB from the DNA by MucB. Such a displacement can explain the observed increased mobility of the DNA in the presence of MucB and the enhanced binding of ethidium bromide to DNA (Fig. 8). An alternative explanation is that MucB interacted with SSB, therefore changing its binding mode to the ssDNA without displacing it. In the latter scenario, the conformational change in the SSB⅐DNA complex induced by the binding of MucB would increase the mobility of the nucleoprotein complex in the gel and enhance the staining of the ssDNA by ethidium bromide. To distinguish between these two possibilities we repeated the gel mobility shift assay using this time radiolabeled SSB. The distribution of the radiolabeled SSB along the gel followed that of the ssDNA, as revealed when the fluorogram and the ethidium bromide-stained picture of the gel are compared (Fig. 9). Quantitation of the autoradiogram by densitometry showed a marginal 10 -15% decrease in the amount of radiolabeled SSB bound to ssDNA in the presence of MucB, which is within the range of the experimental error (Fig. 9). Similar results were obtained when the complex was formed at 30 or 0°C (Fig. 9). This result suggests that MucB alters the conformation of the SSB⅐ssDNA complex, without causing a massive release of SSB from the DNA.

DISCUSSION
The yeast two-hybrid system had proven useful in identifying interactions of the Muc proteins. It was previously shown that the UmuDЈ and UmuD proteins form both homo-and heterodimers when assayed biochemically (35,36,45) and in the yeast two-hybrid system (45). Here we show that MucAЈ and MucA behave similarly and exhibit both self-interactions and an interaction with each other. Each of these proteins was found to interact also with MucB. The C-terminal 30 amino acids of MucB were critical in mediating this interaction, since their deletion abolished the interaction of MucB with both MucA and MucAЈ. The C terminus of the UmuC protein had been previously reported to be essential for UV mutagenesis (46). Based on the homology between UmuC and MucB, our results suggest that the interaction between MucB and MucAЈ is crucial for the activity of these proteins in UV mutagenesis, consistent with the current view that the active species in mutagenesis is a complex of MucAЈ and MucB (or UmuDЈ and UmuC) (reviewed in Ref. 47). Interestingly, it has been reported that UmuC interacts with UmuDЈ, but not with UmuD, in the yeast two-hybrid system (45). In our case, MucB was found to interact both with MucA and MucAЈ. At this stage it is not clear whether these differences are real or whether they reflect differences in the assay systems.
The interaction between MucA and RecA as revealed in the two-hybrid system was expected based on the finding that RecA promotes the cleavage of MucA to MucAЈ (37,38), similarly to the cleavage of UmuD to UmuDЈ (7)(8)(9). Interactions between RecA and MucAЈ or UmuDЈ were proposed previously based on chemical cross-linking experiments (39). Taken together with our in vivo results, it seems plausible that the MucA/AЈ-RecA interactions fulfill a role additional to mediating the cleavage of MucA, possibly the recruitment of MucAЈ to DNA, as previously suggested (39). We found no interaction between MucB and RecA. A related binding interaction was the retention of UmuC on activated RecA immobilized on a column, when an extract of E. coli cells was passed through the column (48). However, in this reported case, since an extract (rather than purified proteins) was used, the possibility that a third protein had mediated the interaction between RecA and UmuC hasn't been ruled out.
SSB, reported here to interact with MucB, is an essential protein, which is involved in DNA replication, recombination, and repair (49 -51). It is a homotetramer that binds ssDNA specifically and in a cooperative manner. Each SSB monomer contains a DNA-binding site interacting with 16 nucleotides. The extent of DNA bound by SSB strongly depends on salt and Mg 2ϩ concentrations. With the increase in their concentration, the number of subunits that interact with DNA increases from 2 (SSB 35 mode) to 4 (SSB 65 mode). Electron microscopic analysis revealed that at low SSB to DNA ratio, a beaded structure is observed, with the DNA wrapped around beads of single or double tetramers of SSB, leading to a reduction in the contour length of DNA. This binding is of limited cooperativity and represents most likely the SSB 65 mode. At higher SSB to DNA ratios, the binding to DNA is cooperative leading to the formation of a smooth SSB-DNA filament, where SSB is presumably in the 35 mode (reviewed in Refs. 44, 50, and 52).
Our results suggest that the interaction with MucB affects the cooperative mode of SSB binding (presumably the 35mode), which is the initial binding state of the SSB under our conditions. The exact nature of the MucB-SSB interaction is not clear yet, but it causes a major change in the SSB⅐ssDNA complex, without causing massive release of the SSB from DNA. This is evident from the increase in the mobility of the nucleoprotein complex and its increased staining with ethidium bromide. The quantitation of radiolabeled SSB is accurate within 10 -15%. Thus, while not causing a massive displacement from DNA, it is possible that MucB does replace some SSB molecules. This, however, must have a major effect on the conformation of the DNA, as evident from its increased mobility. Several attempts to examine whether the nucleoprotein complex contains the MucB protein, using anti-MucB antibodies, were unsuccessful. At this point it is not clear whether this is a technical problem or whether MucB is released after rearranging the SSB⅐ssDNA complex. The increased staining can be explained by at least the following two mechanisms: 1) an increase in the accessibility of the DNA bases to ethidium bromide and 2) a change in the stacking of the DNA bases, which enables better intercalation of the dye. Noteworthy, although MucB bound a ssDNA 40 nucleotides long, we found no effect of MucB alone, or in combination with MucA or MucAЈ, on the migration of naked M13mp8 ssDNA in agarose gel electrophoresis. The reasons for this result are not clear yet.
What are the consequences of the changes in the structure of the nucleoprotein complex caused by MucB? It is possible that such changes allow easier bypass of DNA lesions by a DNA polymerase or else they provide binding sites for other proteins required for the bypass reaction. A natural candidate for being involved in this reaction is the RecA protein. It has been shown to compete with SSB for binding to ssDNA and under certain conditions can form a ternary SSB⅐RecA⅐ssDNA complex (reviewed in Refs. 53 and 54). MucB interacted with SSB but not with RecA (Table IV). On the other hand, MucAЈ did interact with RecA. This may mean that a MucAЈB complex interacts with both RecA and SSB. The interactions of MucB with SSB and with MucAЈ must occur via different regions of the protein.
The C terminus of MucB is involved in the interaction with MucA/AЈ, whereas an additional domain is involved in the interaction with SSB. Thus, a putative 4-protein RecA-MucAЈ-MucB-SSB complex can be imagined, in which RecA is bound to MucAЈB via MucAЈ, and SSB is bound via MucB. So far there is no evidence for such a complex. An alternative possibility is that the various interactions appear transiently during the mutagenic bypass reaction.
One of the paradoxes in the field of SOS mutagenesis is that in vivo bypass of DNA lesions requires the UmuDЈ and UmuC proteins (or their homologues) (reviewed in Ref. 3), whereas in vitro bypass of blocking lesions (2,(11)(12)(13)(14)(15)(16)(17)(18)(19) and in vitro UV mutagenesis (20 -22) can occur without these proteins. The report on umuC-independent UV mutagenesis in phage S13 (55) and the finding that umuC-independent UV mutagenesis is observed when a screening rather than selection procedure is used (56) support the view that UmuDЈC stimulates the mutagenic reaction, rather than being absolutely required. An obvious difference between the in vivo and in vitro situations is that in the former there exist many more proteins that may affect bypass. One such family includes DNA damage binding proteins, which usually function in error-free DNA repair (3). We have recently shown that DNA damage binding proteins regulate induced mutagenesis via a mechanism that does not involve the removal of DNA damage. We found that DNA damage binding proteins directly inhibit trans-lesion replication by binding to lesions present on ssDNA (57). The current study points toward SSB as another candidate that might affect SOS mutagenesis. In vivo studies have shown that E. coli strains carrying the ssb1 mutation, encoding a temperature-sensitive SSB, had reduced UV mutagenesis at the non-permissive temperature (58, 59). However, this was attributed to the inability of the mutant to fully induce the SOS response. Of course, this analysis does not rule out the possibility of a direct involvement of SSB in the mutagenic reaction. We have previously reported that SSB facilitates unassisted bypass of UV lesions by DNA polymerase III holoenzyme (29). It is possible that the MucB⅐SSB complex increases bypass, especially when additional proteins are present. The exact mechanism of the involvement of SSB in SOS UV mutagenesis needs further biochemical and in vivo investigations.