HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hildebrand, E. L. Articles by Grossman, L. Search for Related Content PubMed PubMed Citation Articles by Hildebrand, E. L. Articles by Grossman, L. Social Bookmarking                      What's this?

J Biol Chem, Vol. 274, Issue 39, 27885-27890, September 24, 1999

Oligomerization of the UvrB Nucleotide Excision Repair Protein of Escherichia coli^*

Eric L. Hildebrand and Lawrence Grossman

From the Department of Biochemistry, School of Hygiene and Public Health, The Johns Hopkins University, Baltimore, Maryland 21205

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

A combination of hydrodynamic and cross-linking studies were used to investigate self-assembly of the Escherichia coli DNA repair protein UvrB. Though the procession of steps leading to incision of DNA at sites flanking damage requires that UvrB engage in an ordered series of complexes, successively with UvrA, DNA, and UvrC, the potential for self-association had not yet been reported. Gel permeation chromatography, nondenaturing polyacrylamide gel electrophoresis, and chemical cross-linking results combine to show that UvrB stably assembles as a dimer in solution at concentrations in the low micromolar range. Smaller populations of higher order oligomeric species are also observed. Unlike the dimerization of UvrA, an initial step promoted by ATP binding, the monomer-dimer equilibrium for UvrB is unaffected by the presence of ATP. The insensitivity of cross-linking efficiency to a 10-fold variation in salt concentration further suggests that UvrB self-assembly is driven largely by hydrophobic interactions. Self-assembly is significantly weakened by proteolytic removal of the carboxyl terminus of the protein (generating UvrB*), a domain also known to be required for the interaction with UvrC leading to the initial incision of damaged DNA. This suggests that the C terminus may be a multifunctional binding domain, with specificity regulated by protein conformation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In the characterization of any protein, fundamental questions include whether the monomeric structure is complete, and the identity of other macromolecules with which it can functionally interact. The answers to these take on added importance for the components of the nucleotide excision repair pathway (NER)¹ in Escherichia coli. Introduction of dual incisions surrounding a damage site in DNA by the NER proteins UvrA, UvrB, and UvrC, requires an ordered succession of association/dissociation steps, mediated at multiple points along the pathway by nucleotide binding and/or catalysis. As detailed in recent reviews (1-3), the following sequence of steps have been suggested, focusing here on the shifting identity and stoichiometry of the macromolecular components of the repair complex.

Dimerization of the UvrA protein, promoted by ATP binding, enables formation of a protein-DNA complex, initially at an undamaged site (4, 5). Recruitment of UvrB, which does not on its own bind DNA, generates a functional UvrA₂-UvrB₁-DNA complex, with this stoichiometry assumed from that demonstrated for an ATP-dependent UvrA-UvrB interaction (DNA not present) in solution (6). At this stage, the repair proteins can disengage from the initial DNA binding site, allowing translocation or delivery to the DNA lesion (7). Damage recognition is marked by formation of a stable protein-DNA complex (8), from which UvrA dissociates (6, 9, 10), leaving a UvrB-DNA pre-incision complex. The DNA in this complex is acutely kinked (11), a structural change necessary for later incision and either facilitated by or stabilized by contacts with UvrB (12). It is likely that conformational change(s) also occur in the UvrB protein to augment its affinity for DNA; in solution, nucleotide binding by UvrB has been shown to be limited to short oligonucleotides bearing damage (12). Addition of UvrC is enabled by conformational changes in the UvrB-damaged DNA pre-incision complex (13) and prompts a further conformational change that is dependent in rate on the nature of damage (14). These steps complete the endonucleolytic-competent complex that finally cleaves the DNA backbone in the hallmark pattern of NER, first four to five residues 3', and then eight phosphodiester bonds 5' to the lesion (15-19). As noted with DNA, UvrB has no or an extremely weak affinity for UvrC when the two are simply introduced in solution. The addition of UvrC to the repair complex is dependent both on the presence of the C-terminal domain of UvrB, and a conformational change that apparently renders this interaction surface accessible (20).

As a nexus in the progression of pre-incision steps, the malleable UvrB is thus called upon to display, at the appropriate times, binding affinities that are masked in the isolated protein. The catalytic properties of UvrB, as well as its associative properties, are markedly altered along the repair pathway. In particular, expression of the DNA-dependent ATPase of UvrB, an activity that is cryptic in the isolated protein, has been shown to be essential for release of the A₂-B-DNA from the initial binding site (7), in formation of the pre-incision complex (21), in supporting the 5' incision by the UvrB-UvrC nuclease (22), and likely in the translocation of the repair proteins to the damage site (evidence reviewed in Ref. 1).

Little is known in detail of the manifold protein-protein and protein-DNA interactions at any given point along the pathway and less of the regulation of transitions between associations, structurally or kinetically. Whether step-specific changes in stoichiometry may occur, as has been demonstrated qualitatively for subunit composition (e.g. dissociation of UvrA before recruitment of UvrC), has not yet been addressed. A more complete understanding of the potential for interactions of any of the repair proteins could only aid in our understanding of this complex mechanism by which damage of a remarkably broad spectrum is recognized and removed. In our purification of UvrB, we observed that its elution volume from a gel exclusion column, compared with that of protein standards, suggested a hydrodynamic volume approximately twice that predicted from the molecular weight of UvrB. This led us to broaden the scope of an earlier hydrodynamic study so as to re-examine the generality of the reported finding (6) that UvrB exists solely as a monomer, and whether one may then infer that UvrB-UvrB subunit interactions could play no role in complex assembly.

We present evidence here, derived from both hydrodynamic and chemical cross-linking techniques, that self-association, the formation of dimers and possibly higher order oligomers, is within the panoply of UvrB interactions and is likely to occur under physiological conditions. We also show that truncation of the protein by the ompT protease, which yields a product termed UvrB*, significantly impairs self-association. UvrB* is of interest in (at least) two ways. The residues eliminated are at the C terminus, and have been shown to be required for recruitment of UvrC to the incision complex (20), and because the cryptic DNA-dependent ATPase of UvrB is active in UvrB*, offering the possibility that it may serve as a model for conformational changes induced normally by interaction with UvrA.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Proteins-- The source, purification, and quantitation of UvrB protein have recently been described in detail (23). Briefly, the uvrB gene, under the regulation of its own promoters, was originally cloned into a pTZ19R plasmid from the E. coli strain AB1157 and expressed in N364 (W3110 gal⁺, sup⁰, F, (attB-bio-uvrB)), a uvrB deletion strain of E. coli K-12 (obtained from M. Gottesman, Columbia University). Specific proteolysis of the UvrB proteins, using the ompT expression system UT5600/pML19 (24), and purification of the UvrB* product also followed published protocol (7).

Small Zone Gel Filtration Chromatography-- The hydrodynamic radius and by implication the assembly state of UvrB and its potential dependence on initial protein concentration was examined using a 1.0 × 30 cm (10/30) Superose 12 HR column, operated at room temperature (22-24 °C) by a fast protein liquid chromatography system (Amersham Pharmacia Biotech). The buffer used with this column was 50 mM HEPES, pH 7.6, 300 mM KCl, 10 mM MgCl₂, 5% glycerol, 2 mM DTT. Glycerol content of the buffer had been reduced to keep operating pressure lower; this ranged from approximately 0.75-1.0 megapascals during sample and standard runs at a flow rate of 0.3 ml/min. Sample was loaded in a volume of 0.1 ml, and elution from the column was monitored by flow-through absorbance measurements at 280 nm, with 0.3 ml fractions collected as a check on elution volume and protein integrity (as judged by SDS-PAGE). Calibration standards were purchased from Sigma (MW-GF-200 kit, supplemented with apoferritin and thyroglobulin to extend the molecular weight range to 669,000), and included blue dextran (M_r 2,000,000) for void volume determination. The partition coefficient, K_av = (V_e  V₀)/(V_t  V₀), where V_e is the elution volume, V_0, the void volume, and V_t, the total gel volume, was calculated for standards and samples from triplicate runs (25, 26). The Stokes radii values for standards were taken from the literature (27), except for -amylase, for which a value of 5.18 nm was calculated from the published sedimentation coefficient and ancillary data (28).

Native Gel Electrophoresis-- Oligomers of UvrB were resolved on a 4-30% polyacrylamide gradient gel (375 mM Tris·Cl, pH 8.8); a 3% stacking gel (125 mM Tris·Cl, pH 6.8) was used to further improve resolution. The denaturant SDS was omitted from the gels and from both electrophoresis (25 mM Tris·Cl, 250 mM glycine, pH 8.3) and sample buffers (which contained a final concentration of 10% glycerol and 0.05% bromphenol blue). Electrophoresis was performed at 100 V to 2,000 V-h at room temperature, or at 150 V to 3,000 V-h at 4 °C. Protein bands were visualized by Coomassie Brilliant Blue R250 staining. Molecular weight standards were obtained from Amersham Pharmacia Biotech (HMW kit).

Chemical Cross-linking-- The bifunctional imidoester dimethyl suberimidate (DMS) (29) is reactive primarily with lysine residues, and has long been used as a protein cross-linking reagent (30). UvrB and UvrB* proteins, at 5 µM, were allowed to equilibrate for 10 min at room temperature in a buffer of 20 mM HEPES, pH 7.6, 10% glycerol, 1 mM DTT and KCl at 25, 100, or 300 mM, and ± ATP to 8 mM, before addition of cross-linker, bringing the reaction volume to 20 µl. DMS (Fluka BioChemika) was then added to a final concentration of 5 mg/ml from a freshly prepared 25 mg/ml stock solution in 0.5 M triethanolamine buffer, pH 8.5 (31). An additional 1-µl aliquot of the DMS stock (an increment of 1.2 mg/ml) was added after 15 min of incubation, owing to the short lifetime of this reagent in aqueous solution. After 30 min, the reaction was quenched with Tris (pH 6.8), to a final concentration of 50 mM. Products were analyzed by SDS-PAGE and visualized by Coomassie Brilliant Blue R250 staining. Trials were also performed with glutaraldehyde (Sigma) and disuccinimidyl glutarate (DSG), obtained from Pierce. The latter is a homobifunctional reagent of the N-hydroxysuccinimide ester family that also targets primary amino groups (32). With these reagents, the protocol above was used with the following exceptions: glutaraldehyde was added as an aqueous solution to a final concentration of 0.1%; and DSG to 500 µM, from a freshly prepared 10 mM stock in Me₂SO (Me₂SO contributed no more than 5% of the total reaction volume of 20 µl).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Gel Filtration Chromatography-- Gel filtration chromatography is used in our laboratory in the purification of UvrB. This is performed at 4 °C using a 4.9 × 110-cm column of Sephacryl S-300, equilibrated with 50 mM potassium MOPS, pH 7.5, 500 mM KCl, 2 mM -mercaptoethanol, 15% glycerol. Molecular weight standards were run through this column, in duplicate, as a routine performance evaluation. Comparison of the single elution peak of UvrB from two purification runs with a calibration plot of the partition coefficients determined for the molecular weight standards yielded a predicted Stokes radius of 4.55 nm, and an apparent molecular weight estimate of 150,000. This suggested that, under the conditions described, UvrB occurred predominantly as a dimer (if roughly globular, as would be expected for the standards).

More rigorous gel filtration studies were performed with a 1.0 × 30-cm Superose 12 fast protein liquid chromatography column at room temperature in a buffer of 50 mM HEPES, pH 7.6, 300 mM KCl, 10 mM MgCl₂, 5% glycerol, 2 mM DTT. The performance of this column can be judged from the calibration curve shown in Fig. 1a, in which the log of the Stokes radius for standards ranging in molecular weight from 12,400 to 669,000 is plotted versus the observed partition coefficient K_av (error estimates for the mean partition coefficients for standards and protein samples expressed as the coefficient of variation ranged from 0.21 to 1.94%; the regression correlation coefficient, R², for the calibration curve was 0.979). Loaded onto this column at initial concentrations ranging from 1 to 35 µM (1 µM being the lower limit for a significant signal with our apparatus), UvrB eluted consistently as a single peak with minor tailing but with the peak position showing a strong concentration dependence. The effective radii of hydration (Stokes radius) for UvrB at the differing initial concentrations were calculated from the calibration curve shown and are plotted in Fig. 1b. At the lowest concentration, 1 µM, the partition coefficient coincided with a Stokes radius of 3.7 nm or an apparent molecular weight of 97,000 (estimated from an alternate calibration curve plotting K_av versus M_r). At an initial concentration of approximately 24 µM, the partition coefficient approached an asymptotic value corresponding to a Stokes radius of 4.67 nm or apparent molecular weight of 170,000. These results are indicative of associative behavior and suggest a monomer-dimer equilibrium (molecular weights predicted from the amino acid composition are 76,091 and 152,182, respectively). The single peak observed at any given concentration further suggests that equilibration is rapid with respect to the limits imposed by the elution experiment. The higher values for molecular weights obtained by gel filtration, compared with those predicted by sequence, could derive from a number of factors. The lower molecular weight predicted represents, of course, a limit of detection imposed by our experimental conditions; a lower plateau was not demonstrated but would presumably be intermediate between that observed and the value estimated for UvrB* (below). Partitioning in gel permeation, furthermore, is a function of molecular shape as well as size (27), hence deviation from the globular shape of the commonly used standards could lead to error. Interaction of protein with the gel matrix cannot be excluded, though the salt concentration used (300 mM) should minimize this possibility. Finally, the observed partition coefficient represents a mass averaged function of the species in equilibrium. A minor population of higher order oligomers could skew the partition function toward values suggesting a higher molecular weight.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Gel filtration chromatography of UvrB and UvrB* proteins. Data were obtained at room temperature using a Superose 12 column operated under small zone conditions, as detailed under "Experimental Procedures." The calibration plot shown (panel a) relates the log of the Stokes radius to the observed (mean from triplicate runs) partition coefficients (Kav) for molecular weight standards cytochrome c (12,400), carbonic anhydrase (29,000), bovine serum albumin (67,000), alcohol dehydrogenase (150,000), -amylase (200,000), apoferritin (443,000), and thyroglobulin (669,000). The line represents a least squares regression used to estimate the Stokes radius of UvrB () and UvrB* () proteins, applied to the column over the initial concentration range shown in panel b.

In contrast to UvrB, the elution behavior of UvrB* provides no indication of a potential for oligomeric self-association. Over a range in initial concentration from 1 to 30 µM, the partition coefficient corresponds to a Stokes radius of 3.4 nm, or apparent molecular weight of 77,000. Depending on which cleavage site is recognized by the ompT protease (20, 33), the molecular weight predicted from sequence for a monomer of UvrB* would be in the range of 68,588 to 71,080. A single elution peak was also observed for UvrB*, except at the highest concentration, where a minor peak coincided with the column void volume. It is likely that this material resulted from a large-scale aggregation of UvrB*. Such aggregates would have a Stokes radius in excess of that of the largest protein standard used, M_r 670,000, and would likely include ten subunits as a minimum. The propensity of UvrB* to form large aggregates in concentrated solution had been surmised from its excessive but variable ability to scatter light, as observed (data not shown) both in absorbance and fluorescence emission spectra (Rayleight scatter peaks). A tendency to form such aggregates has been reported both for UvrA and UvrC (6, 34).

Native PAGE of UvrB and UvrB*-- Electrophoresis through a nondenaturing polyacrylamide gel offers a relatively rapid yet sensitive technique for examination of the potential for self-assembly by a protein. Multiple bands can be resolved when UvrB is electrophoresed on such gels, the most useful to date being a 4-30% gradient gel, as shown in Fig. 2. Run at room temperature, five bands can be discerned (and corroborated by densitometric profiles, not shown) when 15 µg of UvrB is applied (lane 2). Three bands remain detectable by Coomassie staining with the protein load reduced to 5 µg (gel not shown). An equal loading (i.e. 15 µg) of this protein on SDS gels gives a single band when visualized by Coomassie staining, with an estimate of 98-99% purity from densitometer scans (cross-linking gels to be shown below). That the state of assembly is temperature-dependent is evident by comparison with the migration profiles seen in Fig. 2, lanes 5 and 6. With electrophoresis performed at 4 °C rather than at room temperature, the number of visible UvrB bands (lane 5) is apparently reduced (three are detectable; that of highest molecular weight marginal by eye but detectable by densitometry), the distribution is altered to more strongly favor the band with greatest mobility (68% of lane density, as compared with 36% at 22 °C), and the mobility of this major band is shifted relative to that of the molecular weight standards (toward an apparently higher molecular weight).

View larger version (74K): [in this window] [in a new window]   Fig. 2.   Native polyacrylamide gel electrophoresis of UvrB and UvrB*. The sample proteins, 15 µg of either UvrB (lanes 2 and 5) or UvrB* (lanes 3 and 6), were applied to 4-30% gradient polyacrylamide gels, run either to 2,000 V-h at room temperature (lanes 1-3), or to 3,000 V-h at 4 °C (lanes 4-6). Proteins were visualized by Coomassie Blue staining. Molecular weight standards (lanes 1 and 4) included: thyroglobulin (669,000), ferritin (440,000), catalase (232,000), lactate dehydrogenase (140,000), and albumin (67,000). Solid arrowheads mark the position of UvrB bands in lanes to the left; open arrowheads, the bands of UvrB*.

It is our suggestion that this most mobile band of UvrB is not populated by monomeric UvrB, but by monomer and dimer in rapid equilibrium and with the equilibrium position altered by temperature. If so, the additional bands may represent oligomers with from three to six subunits. Assignment of the degree of oligomerization to bands is not straightforward. The mobility of a given band in a nondenaturing gel depends on molecular charge as well as size and shape. By use of a gradient gel with up to 30% polyacrylamide, protein standards in a size range appropriate to the study of UvrB oligomerization will, if driven long enough by an applied voltage, reach a pore size small enough to block further migration. If electrophoresis is continued until all proteins, even those with low charge density, reach their terminal postions, then comparison of sample mobility with that of standards has merit (35). The use of standards, however, to estimate the molecular weight of a sample is subject to uncertainty unless it is known that the relation of both size and shape to molecular weight is invariant between sample and standards (36).

Given this caveat, the estimated apparent molecular weights for the UvrB bands marked by arrows in Fig. 2, lane 2 (22 °C) are: 106,800 ± 4,100, 201,200 ± 11,900, 263,700 ± 10,800, 318,600 ± 9,100, and 394,800 ± 6,700. On their own, and not unexpectedly, these estimates do not permit unambiguous assignment of the bands. The estimate for the smallest UvrB species is intermediate between that of a monomer (76,091) and of a dimer (152,182). When run at 4 °C, this band migrates with an apparent molecular weight closer to that expected for the dimer, 145,400 ± 6,600. In gradient gels, the leading edge of the band, where molecules would encounter smaller pore sizes, is usually sharpened. The broadness of the major UvrB bands, at either temperature, and the diffuse appearance of the leading edge, in addition to the temperature dependence support the suggestion that these bands may represent a weight-averaged migration position of monomer and dimer in rapid equilibrium. This interpretation gains support from the single elution peak noted in the gel filtration studies with UvrB. If so, the slower migrating bands could be assigned to tri-, tetra-, penta-, and hexamers; with these assignments the estimates of apparent molecular weight would underestimate the values calculated from the amino acid composition by 12-16%.

With proteolytic truncation of the C terminus, the potential for self-association of UvrB is significantly reduced, yet potential oligomer bands can be seen when 15 µg of UvrB* is electrophoresed in native gels (Fig. 2, lanes 3 and 6). The major band has an estimated molecular weight of 70,500 ± 2,200 (22 °C run, lane 3) to 73,700 ± 1,500 (4 °C run, lane 6). This is close to that predicted for a monomer from the amino acid composition, approximately 68,500-71,080, depending, as noted above, on the site of cleavage. Cleavage at more than one site may in fact account for the additional bands not marked by arrows flanking the major UvrB* band, although contaminants derived from the exposure to ompT-expressing cells cannot be ruled out. The major band itself, on close examination, appears to be a doublet. Conceivably, this may result from near equal cleavage rates at either lysine 605 or 607, generating products differing in molecular mass by 230 Da, as recently suggested (20). The mobility of the putative oligomers of UvrB*, denoted also by open arrowheads, were used to estimate molecular weights ranging from 113,000 ± 6,500 (4 °C gel) to 138,100 ± 10,200 (22 °C), a potential dimer, and from 172,400 ± 11,000 (22 °C) to 216,000 ± 9,200 (4 °C), a potential trimer.

View larger version (56K): [in this window] [in a new window]   Fig. 3.   Cross-linking of UvrB protein by DMS. UvrB protein at 5 µM was incubated with DMS (5 mg/ml) for 30 min at room temperature; products were analyzed by SDS-PAGE with Coomassie Blue staining. Where present (lanes 4, 6, and 8) ATP was included at 8 mM; KCl concentration was 25 mM (lanes 3 and 4), 100 mM (lanes 2, 5, and 6), or 300 mM (lanes 7 and 8). The solid arrowhead indicates the migration position of the putative cross-linked dimer. Open arrowheads align with the migration positions of putative cross-linked trimers and hexamers. The reference sample of unmodified UvrB (lane 2) is from a control reaction identical in composition to the 100 mM KCl, no ATP cross-linking reaction except for the omission of DMS. Molecular weight standards (lane 1) included: myosin (205,000), -galactosidase (116,000), phosphorylase b (97,400), and bovine albumin (66,000).

Chemical Cross-linking-- The potential for self-association of UvrB and UvrB* was further examined with a nonhydrodynamic approach, chemical cross-linking. The purity of the UvrB protein used in experiments was estimated to be 98-99%, as judged by densitometer scans (Fig. 3, lane 2 and from replicate experiments). Covalently linked oligomers of UvrB, resolved by SDS-PAGE, were readily obtained with the lysine-reactive bifunctional reagents glutaraldehyde, DSG, and DMS. Results with DMS are shown in Fig. 3. With exposure of 5 µM UvrB for 30 min to 5 mg/ml of DMS, cross-linked products included apparent dimers (migration position marked by the solid arrow) and higher order oligomers.

Though UvrB is a cryptic DNA-dependent ATPase, we recently demonstrated that it binds ATP, with an apparent K_D ~ 1 mM (23). Cross-linking efficiency was not, however, significantly altered by the addition of Mg-ATP to 8 mM. Variation in the potassium chloride concentration, from 25 to 300 mM, also was without effect on the yield of dimers or total cross-linked species, suggesting that the self-association of UvrB is largely hydrophobic in character. We do note for future study, though, that the distribution of the higher order oligomers may be dependent on salt concentration. As the KCl concentration is increased, there is an apparent shift toward a higher order of oligomerization (an increase in putative hexamers at the expense of trimers; tentative assignments marked in Fig. 3 by the open arrowheads). We also tested the effect of DTT on cross-linking efficiency (UvrB has one cysteine residue). No differences, qualitative or quantitative, were observed when DTT was omitted or increased from 1 to 10 mM in the incubation/reaction buffer (data not shown).

Estimated by densitometry, the extent of reaction (proportion of all cross-linked species) over all salt concentrations, ± ATP, was approximately 57 ± 3%. The efficiency of cross-linking was similar, but on average less with DSG (extent of reaction 59% that with DMS in direct comparisons), at 500 µM reagent, 5 µM UvrB. The distribution of oligomeric products was the same, however, and the lack of dependence on salt or the presence of ATP was also observed. Quantitative differences in the efficacy of the two reagents would be difficult to interpret: a higher concentration of the water soluble DMS could be and was used, but this is offset to an extent by the shorter lifetime of the reactive species. A pH of 8.5 was used to favor the labeling reaction over hydrolysis of DMS (29). With DSG, a lower pH (7.6) could be used in the reaction mixture. DMS could be more effective in trapping complexes because of a longer spacer arm (1.1 nm, as compared with 0.77 nm for DSG). At the same concentration of protein, glutaraldehyde, which can form polydisperse reactive species in solution (30), apparently linked UvrB quantitatively into high molecular weight aggregates, which did not enter the 8% polyacrylamide gel. At lower protein concentration, 1 µM, requiring visualization by silver staining (data not shown), only the apparent dimer was formed, and to an extent comparable with that obtained with DMS at the higher protein concentration.

The specific proteolysis of UvrB by ompT, generating the active DNA-dependent ATPase UvrB*, significantly reduces but does not eliminate its ability to form complexes susceptible to chemical cross-linking. The purity of the UvrB* preparations used was estimated as 97-98% of total lane intensity by densitometer scans of reference lanes (overloaded, with 15 µg protein, as in lane 3 of Fig. 4). Initially, cross-linking was not observed either at 1 µM UvrB* exposed to glutaraldehyde or at 5 µM with DSG. But, as seen clearly in Fig. 4, putative dimers can be trapped by DMS with UvrB* at 5 µM. In comparison with the reference reaction with intact UvrB (Fig. 4, lane 4), it is apparent that the extent of dimer formation is less with UvrB*, and higher order oligomers are not observed. Over all conditions and from triplicate experiments, the yield of dimers as estimated by densitometry was 14 ± 6.5%. As with UvrB, neither the addition of Mg-ATP nor variation in the salt concentration over a 10-fold range had any apparent effect on the yield of cross-linking with UvrB*.

View larger version (70K): [in this window] [in a new window]   Fig. 4.   Cross-linking of ompT-proteolyzed protein (UvrB*) by DMS. The experiment is identical to that in Fig. 3 (intact UvrB) except that UvrB*, also at 5 µM, was used in place of the intact UvrB. The holo-UvrB reference protein is in lane 2, and a positive control reaction for cross-linking of intact UvrB in lane 4, performed also with protein at 5 µM, 100 mM KCl, no ATP. The UvrB* proteolytic product, unmodified with DMS, is in lane 3. The solid arrowhead indicates the migration position of the putative cross-linked UvrB* dimer.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Taken together, transport (electrophoresis and gel filtration) and cross-linking data show that UvrB is capable of self-assembly to form stable and reproducible oligomeric structures, with a dimer likely to be a major species at protein concentrations in the low micromolar range in aqueous solution. The interaction between subunits in dimer formation appears to be largely hydrophobic in character and, unlike dimerization of UvrA protein (4-6), is independent of ATP occupancy in the range of protein concentrations studied. Whether a dimer or other higher order oligomeric structures occur in vivo and if so how this would affect its interactions with UvrA, DNA, and UvrC will require additional study. Evidence to date indicates that UvrB contributes a single subunit to a UvrA₂·UvrB·DNA complex and to the UvrB-damaged DNA pre-incision complex (6, 10, 37). Should UvrB participate in all repair steps exclusively as a monomer then dissociation may need to be considered as an obligate initial step with potential regulatory significance, as has been indicated for the assembly of the UvrA₂ dimer. The C terminus of UvrB, having been identified previously as involved in the association with the UvrC protein (20), also is shown here to strongly affect, directly or indirectly, self-association by UvrB. It is suggested that this domain either provides the contact surface that stabilizes homodimerization, or supports a conformation in which self-association is more favorable. The selectivity of the same interaction surface could be altered in the pre-incision complex by conformational change(s) to favor binding to UvrC, a property not demonstrable for the proteins free in solution.

It is clear from the combined results of complementary techniques that UvrB associates to form dimers in the absence of any cofactors, at concentrations in the low micromolar range, and possibly at lower concentrations. The concentration dependence observed from gel filtration experiments offers clear, though qualitative evidence for associative behavior, and quantitative estimates of the molecular weights (more accurately, hydrated radii) of the end points, i.e. monomer and dimer. The data are less suitable, however, for estimation of equilibrium constants because of continuous dilution of solute with transport down the column (26). We may assume, however, that a crude approximation for the equilibrium constant taken from the concentration versus Stokes radius curve, i.e. in the low micromolar range, would only underestimate the avidity of interaction. Still, an equilibrium constant of this apparent value may reflect a weaker interaction than observed for the dimerization of UvrA, where the threshold for concentration dependence was reported to be 1 nM (5). It is of further interest that the dimerization of UvrA, a requisite initial step in NER, is promoted by the binding of ATP (4, 5), whereas the monomer-dimer equilibrium of UvrB is unaffected by the presence of ATP, at a concentration 8-fold in excess over K_D (23).

Are the present data in conflict with the earlier report that UvrB is monomeric? In their hydrodynamic characterization of UvrB, Orren and Sancar (6) used a 1.0 × 30 cm, AcA34 gel filtration column, equilibrated in 50 mM Tris·Cl, pH 7.5, 300 or 500 mM KCl, 10 mM MgCl₂, 10 mM -mercaptoethanol, 20% glycerol. Elution of UvrB, applied to the column in the concentration range 0.5 to 5 µM (unclear whether at 4 or 25 °C), led to an estimate of the Stokes radius of 3.99-4.24 nm, unaffected by the presence of ATP, and equivalent to a predicted molecular weight of 91,000. Equivalent results were obtained from velocity sedimentation through a glycerol gradient. The high estimate for M_r (91,000 versus a predicted value of 76,091-Orren and Sancar (6) used a value of 78,116) was not questioned, and it was reported that no effect of concentration was found over the stated range; we noted above that molecular weight estimates can err because of differences in molecular shape. At 5 µM UvrB, our gel filtration data and the chemical cross-linking studies suggest that an appreciable fraction of UvrB is in dimer form. The estimate of Orren and Sancar (6) of the average Stokes radius over the range 0.5 to 5 µM is in fact close to and slightly higher than ours for initial concentrations of 1 µM (3.71 nm) and 5 µM (4.19). The data, then, do not appear to be in conflict, only the lack of an observed trend that could be accounted for by the narrower concentration range used in the study cited.

In addition to dimerization, it is also clear that UvrB may form higher order oligomers. Given the low extent of formation of any one of these species, the lack of detection in gel filtration chromatography may easily be explained by the lesser sensitivity of this technique, as employed. Yet even at 5 µM UvrB, the cross-linker dimethylsuberimidate trapped multiple oligomeric species. That these are not simply collision complexes is supported by the visualization of a ladder of bands, weak but reproducible, in native PAGE. If, as we suggest, both monomer and dimer are represented by the single most mobile band in the gels, then it would be most reasonable to propose that the remaining bands represent tri-, tetra-, penta-, and hexameric assemblies.

As noted in the introductory review, helicase activity by the UvrA-UvrB complex is essential for the formation of a pre-incision complex. As a growing number of DNA helicases have been characterized, a common propensity to oligomerize has been noted, with the hypothesis that assembly provides the multiple DNA binding sites required for function (38). In most cases, the active form is either as dimer or hexamer. Heterotrimers are unusual; the only case other than UvrA₂UvrB known at present is the RecBCD assembly, for which the functional form may be hexameric (39). For the UvrA-UvrB helicase, which can unwind short DNA duplexes with 5' to 3' directionality (40, 41), a need for multiple DNA binding sites could presumably be met by two UvrA subunits. A molar ratio of two UvrA molecules to one UvrB in the absence of DNA was suggested by the solution hydrodynamic studies of Orren and Sancar (6). Data that would rule out higher order complexes, such as a determination of the molecular mass of the assembly, are lacking, especially for a complex on DNA. A complex greater in mass than UvrA₂·UvrB could more readily explain the ATP-dependent introduction of supercoils in DNA that appears to result from helix tracking by these proteins (42, 43). A potential association between two UvrA₂·UvrB complexes could also account for the introduction of supercoils by providing an anchoring point for the DNA (1).

That the equilibrium between monomer and dimer is rapid is suggested from our results by the single elution peak with tailing of UvrB from gel filtration columns, and the single broad peak with a diffuse leading edge seen in native PAGE. The indication that self-assembly is driven largely by hydrophobic interactions arises from the observation that a 10-fold variation in salt concentration had no effect on the monomer-dimer equilibrium position, as evidenced by cross-linking performed at a concentration of UvrB at which both monomer and dimer populations were significant. The importance of hydrophobic interactions has been noted for several of the interactions involving NER components, including the binding of UvrB to UvrC (20) and to damaged DNA bases (12, 44), as well as the interaction of UvrA with DNA, both nonspecifically and at damaged bases (45).

An additional and more specific clue as to the nature of the interaction between UvrB subunits comes from the greatly diminished capacity of UvrB* to undergo self-assembly. No oligomeric interactions could be detected by gel filtration, up to concentrations at which aggregation (variable complexes with a minimum of ten subunits) began to pose a problem (30 µM). Dimers were detected, however, by chemical cross-linking. The appearance of weak, putative multimers in overloaded native gels supports the contention that the cross-linked species are not simply trapped collision complexes. An interaction surface for the recruitment of UvrC to the UvrB·DNA pre-incision complex exists in the C-terminal domain of UvrB, likely in a stretch of residues predicted to form a coiled-coil (20). Though the UvrB* peptide is capable of interaction with UvrA and can enter into formation of a pre-incision complex (18, 46), the ability to stably bind UvrC is reduced to a level immeasurably low by gel retardation assay, which correlates with the >98% loss in ability to perform the 3' incision (20). We know now that proteolytic elimination of the C terminus of UvrB curtails, to a similar extent, the dimerization potential of UvrB. Two possibilities are envisioned: residues in the C terminus may provide contacts that stabilize self-assembly, and/or presence of the domain may impose or support a conformation that is more favorable to association. A putative interaction of the coiled-coil domains in the two UvrB subunits provides a rationale for the view that the C terminus includes the interaction surface, as it would for UvrB-UvrC binding. That the same surface could serve either for homodimerization, or for association with UvrC suggests that the affinity of this domain may vary in response to distant, though linked, conformational changes in UvrB induced by step-specific allosteric signals.

	FOOTNOTES

^* This work was supported by National Institutes of Health Merit Award GM-22846 (to L.G.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Dept. of Biochemistry, School of Hygiene and Public Health, The Johns Hopkins University, 615 N. Wolfe St., Baltimore, MD 21205. Tel.: 410-614-4226; Fax: 410-955-2926; E-mail: lg@welchlink.welch.jhu.edu.

	ABBREVIATIONS

The abbreviations used are: NER, nucleotide excision repair; PAGE, polyacrylamide gel electrophoresis; DTT, dithiothreitol; MOPS, 4-morpholinepropanesulfonic acid; Me₂SO, dimethyl sulfoxide; DMS, dimethylsuberimidate; DSG, disuccinimidyl glutarate.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				H. Wang, M. J. DellaVecchia, M. Skorvaga, D. L. Croteau, D. A. Erie, and B. Van Houten UvrB Domain 4, an Autoinhibitory Gate for Regulation of DNA Binding and ATPase Activity J. Biol. Chem., June 2, 2006; 281(22): 15227 - 15237. [Abstract] [Full Text] [PDF]

				N. A. Pavlov, D. I. Cherny, I. V. Nazimov, A. I. Slesarev, and V. Subramaniam Identification, cloning and characterization of a new DNA-binding protein from the hyperthermophilic methanogen Methanopyrus kandleri Nucleic Acids Res., February 1, 2002; 30(3): 685 - 694. [Abstract] [Full Text] [PDF]

				R. Rebeil and W. L. Nicholson The subunit structure and catalytic mechanism of the Bacillus subtilis DNA repair enzyme spore photoproduct lyase PNAS, July 19, 2001; (2001) 161278998. [Abstract] [Full Text] [PDF]

				R. Rebeil and W. L. Nicholson The subunit structure and catalytic mechanism of the Bacillus subtilis DNA repair enzyme spore photoproduct lyase PNAS, July 31, 2001; 98(16): 9038 - 9043. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hildebrand, E. L. Articles by Grossman, L. Search for Related Content PubMed PubMed Citation Articles by Hildebrand, E. L. Articles by Grossman, L. Social Bookmarking