Crystal structure of the chi:psi subassembly of the Escherichia coli DNA polymerase clamp-loader complex Jacqueline M. Gulbis 1, *, Steven L. Kazmirski 2,3 , Jeff Finkelstein 4 , Zvi Kelman 4, , Mike O’Donnell 4 and John Kuriyan 2,3 1 Laboratory of Molecular Biophysics and 4 Laboratory of DNA Replication, Howard Hughes Medical Institute, The Rockefeller University, New York, NY, USA; 2 Department of Molecular and Cell Biology and of Chemistry, University of California, Berkeley, CA, USA; 3 Physical Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA The chi (v)andpsi(w) subunits of Escherichia coli DNA polymerase III form a heterodimer that is associated with the ATP-dependent clamp-loader machinery. In E. coli,thev:w heterodimer serves as a bridge between the clamp-loader complex and the single-stranded DNA-binding protein. We determined the crystal structure of the v:w heterodimer at 2.1 A ˚ resolution. Although neither v (147 residues) nor w (137 residues) bind to nucleotides, the fold of each protein is similar to the folds of mononucleotide-(v) or dinucleotide- (w) binding proteins, without marked similarity to the structures of the clamp-loader subunits. Genes encoding v and w proteins are found to be readily identifiable in sev- eral bacterial genomes and sequence alignments showed that residues at the v:w interface are highly conserved in both proteins, suggesting that the heterodimeric interaction is of functional significance. The conservation of surface-exposed residues is restricted to the interfacial region and to just two other regions in the v:w complex. One of the conserved regions was found to be located on v,distaltothew inter- action region, and we identified this as the binding site for a C-terminal segment of the single-stranded DNA-binding protein. The other region of sequence conservation is localized to an N-terminal segment of w (26 residues) that is disordered in the crystal structure. We speculate that w is linked to the clamp-loader complex by this flexible, but conserved, N-terminal segment, and that the v:w unit is linked to the single-stranded DNA-binding protein via the distal surface of v. The base of the clamp-loader complex has an open C-shaped structure, and the shape of the v:w com- plex is suggestive of a loose docking within the crevice formed by the open faces of the d and d¢ subunits of the clamp-loader. Keywords: clamp loader; DNA replication; processivity factor; sliding clamp. The replication of genomic DNA appears to be carried out in a fundamentally similar manner in prokaryotes, eukary- otes and archaebacteria [1]. In each case, the primary replicase is distinguished from other DNA polymerases by its ability to rapidly polymerize tens of thousands of nucleotides without dissociating from the template. This high level of processivity is conferred on the DNA polymerase by ring-shaped sliding clamps [the b subunit in bacteria, proliferating cell nuclear antigen (PCNA) in eukaryotes and archaebacteria] that tether the DNA polymerase to DNA [2]. The interaction between the DNA polymerase and the sliding clamp enables the active site of the polymerase to bind and release DNA rapidly during its spiral progression along the template strand, without actually dissociating from the template. Each strand of DNA at the replication fork is copied by a core DNA polymerase assembly that is attached to a sliding clamp. Although a single sliding clamp may remain attached to the polymerase during replication of the leading strand, each Okazaki fragment that is generated on the lagging strand requires a new clamp, and these must be rapidly loaded onto newly primed sites. It appears that proteins at the replication fork act in concert to rapidly and repetitively cycle the lagging strand polymerase between sliding clamps loaded at sites of Okazaki fragment synthesis [3]. DNA polymerase III (Pol-III), an archetypal replicase, from Escherichia coli, comprises several distinct subassem- blies. Polymerase-exonuclease cores (the a and e subunits) catalyze DNA synthesis and carry out proofreading, and these are localized to the template by association of the a subunit with the sliding clamp, b, which encircles DNA. The b clamp is loaded onto DNA by a clamp-loading, ATP- dependent machinery (the c or s complex),anintegralpart of the Pol-III holoenzyme [4,5]. Cellular isolates of the E. coli clamp-loader complex contain a mixture of proteins with the following stoichiometry: (c/s) 3 d 1 d¢ 1 v 1 w 1 [4,6]. Correspondence to J. Kuriyan, 16 Barker Hall, University of California, Berkeley, CA 94720-3202, USA. Fax: + 510 643 0159, Tel.: + 510 643 0137, E-mail: kuriyan@uclink.berkeley.edu Abbreviations: PCNA, proliferating cell nuclear antigen; Pol-III, DNA polymerase III; RFC, replication factor C; SIRAS, single isomorph- ous replacement and anomalous scattering; SSB, single-stranded DNA-binding protein. Present addresses: *Structural Biology Division, The Walter and Eliza Hall Institute of Medical Research, 1G Royal Parade, Parkville, Victoria 3050, Australia; University of Maryland Biotechnology Institute, Center for Advanced Research in Biotechnology, 9600 Gudelsky Drive, Rockville, MD 20850, USA. (Received 21 October 2003, revised 21 November 2003, accepted 27 November 2003) Eur. J. Biochem. 271, 439–449 (2004) Ó FEBS 2004 doi:10.1046/j.1432-1033.2003.03944.x The major constituents, the c and s subunits, are ATPases encoded by the DnaX gene [7,8]. Subunits c and s are identical in sequence, except that the C-terminal region of s is longer, and serves to interface with the polymerase catalytic subunit (a) [9] and principal helicase (DnaB) [10–12], thereby coupling replicase and primosome at the fork[13,14].Thecorestructureofthec subunit, a protein with an N-terminal RecA-like domain (domain I), flexibly linked to helical domains II (middle) and III (C-terminal), resembles that of the d and d¢ subunits [15] and of the replication factor C (RFC) clamp-loader subunits of euk- aryotes and archaebacteria [16]. Crystal structures of a functional E. coli clamp-loader complex, cdd¢ (stoichiometry 3 : 1 : 1), and of a complex between the b clamp and an isolated d subunit, have greatly facilitated our understanding of the clamp loading mech- anism [17,18]. The clamp-loader complex is pentameric, and the subunits are arranged such that the major connections occur between the five C-terminal domains, which form a ring-shaped collar. It has been shown in vitro that loading of the clamp onto DNA can be orchestrated by just these three subunits (c, d and d¢) alone [19]. Structural considerations suggest that conformational changes in the clamp-loader, which occur in response to ATP binding at the c–c and c–d¢ interfaces, toggle the b-interacting element of d between dormant and active conformations, and facilitate the formation of multiple contacts between the b clamp and the clamp-loader complex. The b clamp is stabilized in an open conformation by this means, allowing the entry of DNA. The smaller v and w subunits are not obligatory participants in the clamp loading process. The v and w subunits were first isolated from E. coli extracts in association with the c subunit, and were shown to bind to each other tightly [20]. The w subunit interacts with c [21], specifically with domain III of the c subunit [22]. On this basis it was proposed that the sparingly soluble w protein bridges between the v subunit and the c-ATPase subunits in the c-complex [21]. Current evidence suggests that the v:w subassembly plays an important role in the processive synthesis of Okazaki fragments. Single-stranded DNA-binding protein (SSB) binds to v in an interaction that is strengthened nearly 1000-fold when SSB is also bound to DNA [23,24]. SSB coats single-stranded DNA as it unwinds, protecting it from nucleases and melting out secondary structure, thereby circumventing barriers to replication. The interaction between v and the clamp loader stabilizes reconstituted DNA Pol-III holoenzyme at high salt concentrations (up to 800 m M potassium) [25], and is crucial for the rapid replication of the lagging strand. The v subunit disrupts an otherwise stable contact between SSB and primase at the replication fork [26]. This facilitates the dissociation of primase from the newly synthesized RNA primer, and primase is then free to be recycled to another site. The b clamp is assembled onto the newly primed DNA template concomitantly, in readiness for synthesis of the next Okazaki fragment. This switching mechanism ensures that priming and initiation of new fragments is coordinated smoothly with the assembly of b clamps onto DNA. We have determined the 3D structure of a 1 : 1 complex of the E. coli v and w subunits in order to explore the function of these two proteins. Subunits v and w interact to form an elongated heterodimer, of comparable size to that of the c, d,andd¢ protomers. Sequence comparisons amongst 12 bacterial species, containing genes for both v and w, provide some clues as to how the v:w subassembly might interact with the clamp-loader complex and with SSB. Experimental procedures Crystallization The v:w complex was reconstituted by combining the individual component proteins, purified as previously described [27], and separated from an excess of v by anion exchange chromatography. The heterodimer was concen- trated to 10 mgÆmL )1 after extensive dialysis against a buffer comprising 20 m M Tris/HCl (pH 7.5), 4 m M dithio- threitol, 0.5 m M EDTA, 100 m M NaCl and 10% (v/v) glycerol. Monoclinic crystals (spacegroup P2 1 ; a ¼ 64.4 A ˚ , b ¼ 65.7 A ˚ , c ¼ 73.4 A ˚ , b ¼ 116.2°)weregrownin hanging drops at 4 °C over a period of 7 days. One microliter of the protein solution was mixed with 1 lLofa reservoir solution containing 100 m M Hepes (pH 6.8), 25% PEG 4000, 8% (v/v) glycerol and 8% (v/v) 2,5-methyl- pentanediol, prior to equilibration by vapour diffusion. Crystals were directly mounted in nylon loops and flash frozen at 100 K. Measurable diffraction extends beyond 2.1 A ˚ on a laboratory detector system comprising an R-Axis II image plate in conjunction with a Rigaku rotating anode generator. X-ray crystallography The structure was solved by single isomorphous replace- ment with the inclusion of some multiple wavelength anomalous diffraction data. An initial derivative dataset, complete to 2.7 A ˚ , was collected from a crystal soaked for several days in a stabilizing solution supplemented with 1 l M ethyl mercury phosphate. The coordinates for four mercury atoms were derived manually from a difference Patterson map and verified using SHELXS -90 [28]. Multiple wavelength data, to a resolution of 3.2 A ˚ , were collected from a similarly treated crystal on Beamline X25 at Brookhaven National Synchotron Light Source. Mono- chromator positions, defining three discrete wavelengths corresponding to the inflection point, peak, and a remote high-energy point, were determined according to the X-ray absorption fluorescence spectrum of the derivatized crystal. Patterson maps, calculated using anomalous differences as coefficients, confirmed the presence of the same four sites. All images were processed using DENZO [29] and the integrated intensities were scaled and merged using SCALE- PACK [29]. Phasing and refinement Initial attempts at structure determination by MAD phasing were unsuccessful, owing to mediocre diffraction and a lack of high quality dispersive difference data. Instead, a method utilizing only the anomalous Df¢¢ data, measured at the synchrotron and single isomorphous replacement and anomalous scattering (SIRAS) phases obtained from in-house native and derivative datasets, resulted in an experimental map with clearly defined protein and solvent 440 J. M. Gulbis et al. (Eur. J. Biochem. 271) Ó FEBS 2004 regions. MLPHARE , from the CCP4 suite [30], was used to refine heavy atom parameters and to generate phases. Density modification procedures, including solvent flatten- ing, histogram matching, and the derivation of phase relationships using Sayre’s equation, were implemented with SQUASH [31], and the improved phase information yielded an interpretable electron density map. A partial model of v was built into this density using O [32], and this was used as a basis for positioning a second complex in the asymmetric unit using AMORE [33]. Twofold noncrystallo- graphic symmetry averaging of the experimental maps, using the program RAVE [34], enhanced map quality sufficiently to enable tracing of most structural elements and assignment of the amino acid sequence. Iterative cycles of building and refinement were required to place the remainder of the structure. Refinement was carried out by least-squares optimization and simulated annealing procedures, using X - PLOR [35], and by maximum likelihood methods, using CNS [36]. Strict noncrystallographic symmetry constraints were released in the final stages, and individual temperature factors were refined for all nonhydrogen atoms. The final refined model had no outliers in the Ramachandran diagram, and contained 516 residues and 256 water molecules with refined B-factors of less than 60 A ˚ 2 . Twenty-six residues at the N-terminus of each of the two crystallographically inde- pendent w molecules were omitted from the model because of poor electron density in those regions. Side-chains were not modeled beyond Ca for the following residues: (v1: Arg92, Lys132, Arg135, Lys147); (v2: Arg92, Asp121, Ser122, Lys132, Arg135, Lys147); (w1: Asp93, Glu94, Arg135, Asn136); (w2: Gln26, Gln81, Gln123, His130, Arg135). The atomic coordinates have been released in the protein data bank with the access code 1EM8. Sequence alignments and conservation scores The E. coli amino acid sequence for w was used in BLAST [37] to identify w sequences in other organisms. w Sequences that were more than 80% identical to sequences already in the set were excluded from further analysis. This search resulted in the inclusion of 12 w sequences in an alignment using CLUSTALX [38]. A sequence conservation score was calculated for each amino acid position in the aligned sequences, by pairwise comparisons between sequences for each amino acid position. For each pairwise comparison, the BLOSUM 62 matrix [39] gives a substitution probability score for the amino acid substitution. These scores are summed for each amino acid position, divided by the number of pairwise comparisons made and then scaled, so that a score of 100% reflects absolute conservation. Using thissetof12w sequences, v sequences from the same organisms were used to calculate the level of sequence conservation for v and for other subunits of the clamp- loader complex (c, d, d¢). Results and discussion Structure determination Although v is well behaved as an isolated protein in solution, w forms insoluble aggregates. w was therefore purified under denaturing conditions and solubilized in the presence of v [27]. The resulting 1 : 1 complex of v and w is soluble and monodisperse in solution, as determined by dynamic light scattering (data not shown). Monoclinic crystals, containing two v:w complexes in the asymmetric unit, diffract X-rays to % 2A ˚ Bragg spacings on a rotating anode X-ray source. The crystal structure of the v:w complex was determined by SIRAS and refined using data to Bragg spacings of 2.1 A ˚ (Tables 1 and 2). Refine- ment against 27 852 reflections converged at a free R-value of 0.265 and a conventional R-value of 0.229. The final crystallographic model for each independent v:w hetero- dimer contained the complete sequence of v (147 residues) except for the N-terminal methionine, and 110 of 137 residues of w (the N-terminal 26 resides of w were disordered in both molecules in the asymmetric unit). Structure of the v:w heterodimer The v and w subunits formed an elongated heterodimer, in agreement with predictions made on the basis of sedimen- tation equilibria [21,25] (Fig. 1A). The conformations of the Table 1. Crystallographic structure determination. R merge ¼ R j jI j ÀhIi    = R j I j ;R iso ¼ RjF P À F PH    =RjF P j. Dataset Sites Resolution (A ˚ ) Reflections (measured/unique) Completeness (%) (overall/outer shell) R merge (%) (overall/outer shell) R iso (%) (overall) Native 1 Data 20.0–2.1 218 771/32 326 97.9/95.2 6.6/19.4 16.6 EMP (1.5418 A ˚ ) 4 20.0–2.7 90 321/15 653 95.6/71.9 7.3/25.9 EMP k1 ¼ 1.0093 A ˚ 4 20.0–2.8 114 443/13 238 97.1/98.7 7.3/27.5 EMP k2 ¼ 0.9919 A ˚ 4 20.0–2.9 87 057/11 998 93.1/95.6 8.2/32.0 EMP k3 ¼ 1.0062 A ˚ 4 20.0–2.9 87 285/11 993 93.9/96.8 7.6/31.8 Table 2. Crystallographic structure refinement. Refinement statistics RMS deviations Data Resolution (A ˚ )R w R free No. of reflections (all/working/free) Bond lengths (A ˚ ) Bond angles (°) Native2 20.0–2.1 0.229 0.265 30 953/27 852/3101 0.01 3.2 Ó FEBS 2004 E. coli DNA polymerase clamp-loader subassembly (Eur. J. Biochem. 271) 441 two crystallographically independent v:w complexes were essentially identical, except for a small difference in tilt angle between the v and w subunits. The root mean-square deviation in the positions of the Ca atoms between the two structures, calculated by superimposing 256 Ca atoms, was 0.83 A ˚ . Both v and w have central, parallel b-sheets that are connected to a-helices (Fig. 1D); v resembles a classic mononucleotide-binding fold, whilst w more closely typi- fies dinucleotide-binding proteins [40] (Fig. 2). The struc- tures of both subunits were compared with structures in the protein databank [41], using the DALI server [42]. Fig. 1. Structure of the v:w heterodimer. (A) Ribbon diagram of the v:w heterodimer crystal structure. The w subunit is colored cyan and sits on top of the v subunit. The v subunit is colored green except for the stretch of residues that reside in the w-binding site, which have been colored red. (B) An enlarged view of the contiguous loop region of v and how it interacts within the cleft of w. This loop region has high sequence similarity with a DNA-dependent DNA polymerase from the bacteriophage PRD1. (C) A rotated view of the surface of w is shown. The w subunit has been rotated to show the cleft between a1anda4thatmakesupthev-binding surface. Residues 61–66 of v are shown in green. The side-chain of Phe64 of v inserts itself into a conserved hydrophobic pocket consisting of Val57, Leu121, Trp122 and Ile125. (D) A schematic diagram of the v:w heterodimer. 442 J. M. Gulbis et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Structural similarity between v and w was revealed to proteins that are primarily nucleotide-binding proteins but, like the d and d¢ subunits of the clamp-loader, neither v nor w contain any of the functional elements required for nucleotide binding. The topology of the w subunit resembles that of the bacterial two-component signaling protein, CheY [43], and the uracil DNA-glycosylases, UDG [44] and MUG [45] (Fig. 2A). The v subunit has a central b sheet with seven parallel strands, which curve in a left-handed twist. There is structural similarity between v and the non-ATP-binding subdomain 2A of DEAD box helicases such as PcrA [46], and Rep47 (Fig. 2B). There are only two significant deviations in topology between v and subdomain 2A of these proteins. One was observed at the w interface, where a compact glycine-rich loop in v replaces a large insertion (subdomain 2B) in PcrA and Rep. The other occurs where an extended a helix, bordering the nucleotide-binding site between subdomains 1A and 2A in the helicases, is truncated in v to a short loop incorporating a single turn of 3 10 helix. The functional significance of this structural similarity between v and DEAD box helicases is unclear. Two parallel helices on one side of a four stranded b-sheet of w form the sides of an extended hydrophobic crevice in the molecular surface. A single contiguous loop region from v (residues 52–79) inserts snugly into this cleft, placing Phe64 of v into the hydrophobic pocket on w, and burying 1256 A ˚ 2 of surface area at the subunit interface (Fig. 1B,C). Interestingly, a DNA-dependent DNA polymerase from the E. coli bacteriophage, PRD1, has high sequence simi- larity to this loop region alone of v, extending over 28 residues with 13 identities, suggesting that this DNA polymerase might couple to the clamp-loader complex via the w subunit. The functional significance of this interaction is unclear, and no other proteins with significant sequence similarity to v (or to w) were detected using a BLAST search [37]. Sequence conservation in v and w A query of the nonredundant protein sequence database with the sequences of E. coli v and w,using BLAST and PSI - BLAST , resulted in statistically significant matches that were restricted to the genomes of certain bacteria (Table 3). This sequence search identified more sequences for v than for w (some bacterial genomes contained identifiable sequences for v, but not for w,andthe genomes that contained both had only one instance of each). We restricted our analysis to 12 v and w sequences from genomes that contained clearly identifiable genes for both proteins (Table 3). The presence of genes for v in genomes that do not contain w was unexpected, because the w-binding site appears to be conserved in these Fig. 2. Structural comparisons of w and v with other proteins. (A) A side-by-side comparison of w with the mismatch specific DNA uracyl glycosylase, MUG [45]. Similar structural features are colored yellow. (B) A side-by-side comparison of the v DEAD box helicase, PcrA [46]. Ó FEBS 2004 E. coli DNA polymerase clamp-loader subassembly (Eur. J. Biochem. 271) 443 proteins. As discussed below, the sequence of w is not as highly conserved as the v sequences, and it is possible that the BLAST searches simply failed to identify w proteins that have diverged greatly in sequence. Residues located at the interface between v and w were highly conserved in both proteins (Fig. 3). In w,the hydrophobic pocket that binds v is situated between two a helices (a1, residues 52–61 and a4, residues 118–126) (Fig. 1C). Four hydrophobic residues in w that are within this pocket have high conservation scores [Val57, 93.4%; Leu121, 76.2%; Trp122, 100%; and Ile125, 87.1%) (see the Experimental procedures for a definition of the conserva- tion scores, which are based on the BLOSUM 62 substitution matrix) [39]. In v, the aromatic residue (Phe64) that is bound within the pocket of w, is absolutely conserved (Fig. 1B,C). Other interfacial residues that are highly conserved included Trp57 of v (conservation score ¼ 100%) and Ala119 of w (score ¼ 73.7%). The surface of v also contains a highly conserved region that is located distal to the region of interaction with w (Fig. 4). This conserved surface region comprises an a helix (a4, residues 124–135) and a b strand (b7, residue 139–143), between which there is a cleft. Interestingly, helix a4has four absolutely conserved basic residues that are exposed (Lys124, Arg128, Lys132 and Arg135). There is an additional, absolutely conserved, arginine residue within the helix (Arg130) that is buried and forms hydrogen bonds with the main-chain oxygen atoms of Phe116 and the side- chainofaburiedasparticacidresidue(Asp115). Interaction between v and SSB The interaction between v and SSB is mediated by residues located within the very C-terminal region of SSB [23]. This region of SSB includes a conserved sequence motif (173- DDDIPF-178) in E. coli SSB, including three negatively Table 3. Sequences for v:w sequence comparison. a Species wv Accession no. % Identity to E. coli Accession no. % Identity to E. coli Escherichia coli 16132190 100 15804851 100 Shigella flexneri 30065610 99 30065454 99 Salmonella typhimurium 16767798 78 16767721 95 Yersinia pestis 16120761 59 16123590 72 Haemophilus influenzae Road 16271986 32 16273306 49 Vibrio parahaemolyticus 28899216 31 28899419 39 Pasteurella multocida 15602824 34 15602687 50 Haemophilus somnus 32029392 27 23467212 50 Vibrio vulnificus 27365077 29 27364859 38 Vibrio cholerae 15640676 31 15642498 41 Candidatus Blochmannia floridanus 33519585 24 33519517 43 Haemophilus ducreyi 33151459 36 33151849 54 Actinobacillus pleuropneumoniae serovar 32035477 25 32034863 54 Sequences for v alone b Shewanella oneidensis 24374933 39 Pseudomonas putida 26987715 33 Azotobacter vinelandii 23104211 30 Xylella fastidiosa Dixon 22994951 32 Xanthomonas campestris 21230123 34 Nitrosomonas europaea 30248459 24 Microbulbifer degradans 23028461 27 Bordetella pertussis 33593407 25 Magnetococcus sp. MC-1 22998318 31 Neisseria meningitidis 15677419 30 Burkholderia fungorum 22986355 26 Chromobacterium violaceum 34498368 26 Ralstonia metallidurans 22976686 19 Magnetospirillum magnetotacticum 23016368 21 Rhodopseudomonas palustris 22964979 19 Caulobacter crescentus 16125938 21 Rhodospirillum rubrum 22968178 21 Rhodobacter sphaeroides 22959552 24 Brucella melitensis 17987543 20 Novosphingobium aromaticivorans 23108152 25 a Sequences in italics were not included in the sequence conservation calculations as they had higher than 80% sequence identity to E. coli w. b These sequences were not included in the sequence conservation calculations shown in Figs 3 and 4. 444 J. M. Gulbis et al. (Eur. J. Biochem. 271) Ó FEBS 2004 charged residues [48]. A conditionally lethal E. coli mutant, SSB-113, differs from the wild type SSB by a single amino acid substitution in which the penultimate residue, Pro177, is replaced with serine. Although SSB-113 binds single- stranded DNA as tightly as wild type SSB, it is unable to interact with the v subunit [23]. Furthermore, removal of the last 26 residues of SSB leads to the loss of interaction between SSB and v [49]. This negatively charged motif could potentially interact with the conserved and positively charged surface region of v that is distal to the v:w interface (Fig. 4). The atomic coordinates of several SSB variants are listed in the Protein Data Bank [50–52]. Unfortunately, none of these structures include models for the acidic C-terminal region that is responsible for binding v and, possibly, primase. This region is not part of the DNA-binding Fig. 3. Sequence conservation in v and w. The conservation score using the BLOSUM 62 substitution matrix (see the Experimental procedues) for each residue in v and w was calculated for the 12 pairs of sequences shown in Table 3. The surfaces of v and w, shown in this figure, are colored according to this conservation score. To the right, the binding surfaces of both proteins are shown. Both binding surfaces have been conserved in each protein. On w, little surface conservation is observed outside the v-binding surface. In v, a large amount of surface area is conserved distal to the w-binding site.Thisareaisproposedtobindtosingle-strandedDNA-bindingprotein(SSB). Fig. 4. Potential v:single-stranded DNA-binding protein (SSB) interaction. Aregionofv, with high sequence conservation, is shown (B). This surface is suggested to bind to the negatively charged C-terminal tail of SSB. Absolutely conserved and positively charged residues, located within this region, are shown on the left in a ribbon diagram in the same orientation (A). A schematic drawing of the inferred interaction between v and the C-terminus consensus sequence of SSB is shown on the right (C). Ó FEBS 2004 E. coli DNA polymerase clamp-loader subassembly (Eur. J. Biochem. 271) 445 domain of SSB and it contains many Gly, Gln and Pro residues, suggesting that it may not have a regular or defined structure. The stabilization of this C-terminal region of SSB by interaction with v might be responsible for the stable interaction of single-stranded DNA, SSB and the v:w complex. The N-terminal segment of w is a possible linker between v:w and the clamp-loader In the crystals of v:w, the N-terminal 26 residues are disordered and are not present in the structural model. While disordered regions are not uncommon within crystal structures, it is interesting that this region of w is highly conserved in sequence (Fig. 5). Within the disordered N-terminal region there are three absolutely conserved hydrophobic residues (Ile14, Trp17 and Pro22). In contrast, residues on the surface of the 3D model of w are not highly conserved, except for those involved in the interaction with v. This suggests to us that the conserved, but disordered, N-terminal segment of w may have functional importance for binding to the clamp-loader complex. It is known that w interacts with domain III of c [22], thus bridging the clamp-loader and v. To identify where the N-terminal region of w might interact with the clamp-loader complex, a sequence alignment was performed for each of the clamp-loader subunits (d, d¢ and c), using sequences from the same bacterial species that were used in the alignment of the sequences of w and v.Ford and d¢,there was little evidence for conserved residues on the surface, except for residues that are involved in binding to the c subunit and the b clamp. As expected, the c subunit shows a high degree of conservation in the first two domains that make up the AAA + ATPase portion of the subunit. The third domain of the c subunit is its oligomerization domain, and there is a high degree of sequence conservation in regions that interacted to form the C-terminal collar of the clamp-loader. Interestingly, the three c subunits also have a conserved surface region inside the C-terminal collar, which includes an exposed hydrophobic patch (Fig. 6). This hydrophobic patch consists of Phe359 (absolutely con- served) and Leu327 (conservation score ¼ 71.2%), and represents a potential interaction surface for w because it does not appear to be involved directly in other interactions. Examination of surface charge distributions and hydro- phobicity on both the v:w heterodimer and the clamp- loader complex failed to reveal any obvious docking mode for the v:w heterodimer onto the clamp-loader. The structure of the clamp-loader complex is such that there is a prominent gap in the C-shaped base of the structure, between the d and d¢ subunits. It had been proposed that this gap would close during one stage of the clamp loading cycle [17], but recent fluorescence energy transfer meas- urements, made by our group on the c complex, indicated that this gap stays open at all stages (E. Goedken, M. Levitus, A. Johnson, C. Bustamante, M. O’Donnell & J. Kuriyan, unpublished observation). Maintenance of the open C-shape of the base of the clamp-loader complex is also consistent with crystal structures determined recently in our group, of a nucleotide-loaded c complex (S. L. Kazmirski, M. Podobnik, T. F. Weitze, M. O’Donnell & J. Kuriyan, unpublished observation) and a eukaryotic RFC complex bound to PCNA (G. D. Bowman, M. O’Donnell & J. Kuriyan, unpublished observation). Strikingly, in the RFC–PCNA complex, an additional domain of the RFC-1 subunit is located in the gap corresponding to the d–d¢ opening in the c complex. These results suggest that this gap does not close during the clamp loading cycle, and it is possible that the v:w unit may be located in this region (Fig. 6). The insertion of v:w into this prominent crevice on the surface of the clamp-loader would explain why only one v:w heterodimer is bound to one clamp-loader complex, even though there are three c subunits (each with a potential binding site for w) in the complex. The lack of sequence Fig. 5. Conservation of sequences in the N-terminal segment of w. An alignment of the first 26 residues of w,fromthelistofsequencesgivenin Table 3, is shown. The alignment is colored according to the degree of sequence conservation. These 26 residues are disordered in the crystal structure of the v:w complex, yet a high amount of conservation is observed. It is proposed that the this linker binds to the clamp-loader complex, tethering the v:w heterodimer to the complex. 446 J. M. Gulbis et al. (Eur. J. Biochem. 271) Ó FEBS 2004 conservation on the surfaces of the d and d¢ subunits suggests that the docking of the v:w unit onto the clamp- loader may be loose, mediated primarily by the flexible N-terminal segment of w. This general location for v,which interacts with SSB, is consistent with the fact that the b clamp will be opened in its vicinity, leading to the insertion of DNA into the clamp at this site. Conclusions We have presented the crystal structure of the v:w hetero- dimer, which, together, form an accessory factor for the clamp loading process in E. coli and certain other bacteria. Despite a clear structural similarity to nucleotide-binding proteins, v and w are incapable of binding nucleotides. Rather, the v:w complex functions as an adapter unit that couples the clamp-loader complex to SSB. A conserved and positively charged surface pocket on v is probably the region that interacts with the C-terminal acidic region of SSB. Structure-based sequence alignments suggest that w may bind to the C-terminal collar domain of the c subunit via its N-terminal segment, a region of % 26 residues that is highly conserved in sequence, but disordered in the crystal. The loose docking of the v:w heterodimer onto the c complex might explain why many bacterial genomes do not contain readily identifiable v and w sequences, which are also lacking in eukaryotes and archaebacteria. The adapter function imposes minimal sequence constraints on these proteins, which could be replaced functionally by highly divergent, but related, proteins, or even by completely unrelated proteins. Acknowledgements We thank Lore Leighton for preparing the illustrations. We are grateful to the members of the Kuriyan laboratory, and to Irina Bruck, Jerard Hurwitz, Elena Conti, Marjetka Podobnik, David Jeruzalmi and Declan Doyle, for assistance and insightful discussions. Lonnie Berman and the staff of the National Light Source, Brookhaven National Laboratory, willingly donated their time and expertise, for which we are much indebted. J. M. G. holds a Wellcome Trust ISRF Fellowship. This work was partially supported by grants from the NIH (GM 45547 to J. K., GM 38839 to M. O’D.). References 1. Kornberg, A. & Baker, T. (1992) DNA Replication, 2nd edn. W.H. Freeman, New York. 2. Jeruzalmi, D., O’Donnell, M. & Kuriyan, J. (2002) Clamp loaders and sliding clamps. Curr. Opin. Struct. Biol. 12, 217–224. 3. Stukenberg, P.T., Turner, J. & O’Donnell, M. (1994) An explan- ation for lagging strand replication: polymerase hopping among DNA sliding clamps. Cell 78, 877–887. Fig. 6. Potential clamp-loader:w interaction. (A) Two views of the Escherichia coli clamp- loader complex are shown [17]. An exposed hydrophobic region of the c subunit, which is highly conserved but not involved in nucleo- tide binding or intersubunit interactions, is indicated as a potential binding site for the N-terminal disordered region of w.(B)Onthe left are space filling structures of the E. coli clamp-loader and v:w heterodimer, with dif- ferent colors for different subunits: d (magen- ta), d¢ (orange), c1(green),c2(red),c3 (blue), w (cyan), and v (dark green). A schematic diagram showing a possible mode of inter- action between the v:w heterodimer and the clamp-loader complex is shown on the right. The v:w heterodimer is believed to sit in the gap between d and d¢, while the N terminus of w interacts with the proposed binding region of c inside the C-terminal collar of the clamp- loader complex. Ó FEBS 2004 E. coli DNA polymerase clamp-loader subassembly (Eur. J. Biochem. 271) 447 4. Onrust, R., Finkelstein, J., Naktinis, V., Turner, J., Fang, L. & O’Donnell, M. (1995) Assembly of a chromosomal replication machine: two DNA polymerases, a clamp loader, and sliding clamps in one holoenzyme particle. I. Organization of the clamp loader. J. Biol. Chem. 270, 13348–13357. 5. Stewart, J., Hingorani, M.M., Kelman, Z. & O’Donnell, M. (2001) Mechanism of beta clamp opening by the d subunit of Escherichia coli DNA polymerase III holoenzyme. J. Biol. Chem. 276, 19182–19189. 6. Pritchard, A.E., Dallmann, H.G., Glover, B.P. & McHenry, C.S. (2000) A novel assembly mechanism for the DNA polymerase III holoenzyme DnaX complex: association of dd¢ with DnaX (4) forms DnaX (3) dd¢. EMBO J. 19, 6536–6545. 7. Tsuchihashi, Z. & Kornberg, A. (1990) Translational frameshift- ing generates the c subunit of DNA polymerase III holoenzyme. Proc.NatlAcad.Sci.USA87, 2516–2520. 8. Flower, A.M. & McHenry, C.S. (1990) The c subunit of DNA polymerase III holoenzyme of Escherichia coli is produced by ribosomal frameshifting. Proc. Natl Acad. Sci. USA 87, 3713– 3717. 9. Studwell-Vaughan, P.S. & O’Donnell, M. (1991) Constitution of the twin polymerase of DNA polymerase III holoenzyme. J. Biol. Chem. 266, 19833–19841. 10. Yuzhakov, A., Turner, J. & O’Donnell, M. (1996) Replisome assembly reveals the basis for asymmetric function in leading and lagging strand replication. Cell 86, 877–886. 11. Kim, S., Dallmann, H.G., McHenry, C.S. & Marians, K.J. (1996) s Couples the leading- and lagging-strand polymerases at the Escherichia coli DNA replication fork. J. Biol. Chem. 271, 21406– 21412. 12. Kim, S., Dallmann, H.G., McHenry, C.S. & Marians, K.J. (1996) Coupling of a replicative polymerase and helicase: a s–DnaB interaction mediates rapid replication fork movement. Cell 84, 643–650. 13. Gao, D. & McHenry, C.S. (2001) s Binds and organizes Escheri- chia coli replication through distinct domains. Partial proteolysis of terminally tagged s to determine candidate domains and to assign domain V as the alpha binding domain. J. Biol. Chem. 276, 4433–4440. 14. Gao, D. & McHenry, C.S. (2001) s Binds and organizes Escheri- chia coli replication proteins through distinct domains. Domain IV, located within the unique C terminus of s, binds the replication fork, helicase, DnaB. J. Biol. Chem. 276, 4441–4446. 15. Guenther, B., Onrust, R., Sali, A., O’Donnell, M. & Kuriyan, J. (1997) Crystal structure of the d-subunit of the clamp-loader complex of E. coli DNA polymerase III. Cell 91, 335–345. 16. Oyama, T., Ishino, Y., Cann, I.K., Ishino, S. & Morikawa, K. (2001) Atomic structure of the clamp loader small subunit from Pyrococcus furiosus. Mol. Cell. 8, 455–463. 17. Jeruzalmi, D., O’Donnell, M. & Kuriyan, J. (2001) Crystal structure of the processivity clamp loader c (c) complex of E. coli DNA polymerase III. Cell 106, 429–441. 18. Jeruzalmi, D., Yurieva, O., Zhao, Y., Young, M., Stewart, J., Hingorani, M., O’Donnell, M. & Kuriyan, J. (2001) Mechanism of processivity clamp opening by the d subunit wrench of the clamp loader complex of E. coli DNA polymerase III. Cell 106, 417–428. 19. Naktinis, V., Turner, J. & O’Donnell, M. (1996) A molecular switch in a replication machine defined by an internal competition for protein rings. Cell 84, 137–145. 20. O’Donnell, M. & Studwell, P.S. (1990) Total reconstitution of DNA polymerase III holoenzyme reveals dual accessory protein clamps. J. Biol. Chem. 265, 1179–1187. 21. Xiao, H., Dong, Z. & O’Donnell, M. (1993) DNA polymerase III accessory proteins. IV. Characterization of v and w. J. Biol. Chem. 268, 11779–11784. 22. Gao, D. & McHenry, C.S. (2001) s Binds and organizes Escheri- chia coli replication proteins through distinct domains. Domain III, shared by c and s,bindsdd¢ and vw. J. Biol. Chem. 276, 4447– 4453. 23. Kelman, Z., Yuzhakov, A., Andjelkovic, J. & O’Donnell, M. (1998) Devoted to the lagging strand – the v subunit of DNA polymerase III holoenzyme contacts SSB to promote pro- cessive elongation and sliding clamp assembly. EMBO J. 17, 2436–2449. 24. Glover, B.P. & McHenry, C.S. (1998) The vwsubunits of DNA polymerase III holoenzyme bind to single-stranded DNA-binding protein (SSB) and facilitate replication of an SSB-coated template. J. Biol. Chem. 273, 23476–23484. 25. Olson, M.W., Dallmann, H.G. & McHenry, C.S. (1995) DnaX complex of Escherichia coli DNA polymerase III holoenzyme. The vwcomplex functions by increasing the affinity of s and c for d.d¢ to a physiologically relevant range. J. Biol. Chem. 270, 29570–29577. 26. Yuzhakov, A., Kelman, Z. & O’Donnell, M. (1999) Trading places on DNA – a three-point switch underlies primer handoff from primase to the replicative DNA polymerase. Cell 96, 153– 163. 27. Xiao, H., Crombie, R., Dong, Z., Onrust, R. & O’Donnell, M. (1993) DNA polymerase III accessory proteins. III. holC and holD encoding v and w. J. Biol. Chem. 268, 11773–11778. 28. Sheldrick, G.M. (1990) Phase annealing in SHELX-90: direct methods for larger structures. Acta Crystallogr. A 46, 467–473. 29. Otwinowski, Z. & Minor, W. (1997) HKL DENZO and SCALEPACK . University of Texas, Southwestern Medical Center at Dallas, Dallas. 30. CCP4 (Collaborative Computational Project, 4) (1994) The CCP4 suite: programs for protein crystallography. Acta Crystallogr. D 50, 760–763. 31. Zhang, K.Y.J. & Main, P. (1990) The use of Sayre’s equation with solvent flattening and histogram matching for phase extension and refinement of protein structures. Acta Crystallogr. A 46, 377–381. 32. Jones, T.A., Zou, J Y., Cowan, S.W. & Kjeldgaard, M. (1991) Improved methods for the building of protein models in electron density maps and the locations of errors in these models. Acta Crystallogr. A 47, 110–119. 33. Navaza, J. (1994) AMoRe: an automated package for molecular replacement. Acta Crystallogr. A 50, 157–163. 34. Kleywegt, G.J. & Jones, T.A. (1994) Halloween … masks and bones. In From First Map to Final Model (Bailey, S., Hubbard, R. & Waller, D., eds), pp. 59–66. SERC Daresbury Laboratory, Daresbury. 35. Bru ¨ nger, A.T. (1992) XPLOR : a System for X-Ray Crystallography and NMR. Yale Academic Press, New Haven. 36. Bru ¨ nger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., Read, R.J., Rice, L.M., Simonson, T. & Warren, G.L. (1998) Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921. 37. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. & Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol. 215, 403–410. 38. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. & Higgins, D.G. (1997) The CLUSTAL _ X windows interface: flexible strategies for multiple sequence alignment aided by quality ana- lysis tools. Nucleic Acids Res. 25, 4876–4882. 39. Henikoff, S. & Henikoff, J.G. (1992) Amino acid substitution matrices from protein blocks. Proc. Natl Acad. Sci. USA 89, 10915–10919. 40. Schulz, G.E. (1992) Binding of nucleotides by proteins. Curr. Opin. Struct. Biol. 2, 61–67. 448 J. M. Gulbis et al. (Eur. J. Biochem. 271) Ó FEBS 2004 [...]... Conservation of sequences in the N-terminal segment of w An alignment of the first 26 residues of w, from the list of sequences given in Table 3, is shown The alignment is colored according to the degree of sequence conservation These 26 residues are disordered in the crystal structure of the v:w complex, yet a high amount of conservation is observed It is proposed that the this linker binds to the clamp-loader. .. be involved directly in other interactions Examination of surface charge distributions and hydrophobicity on both the v:w heterodimer and the clamploader complex failed to reveal any obvious docking mode for the v:w heterodimer onto the clamp-loader The structure of the clamp-loader complex is such that there is a prominent gap in the C-shaped base of the structure, between the d and d¢ subunits It... the right The v:w heterodimer is believed to sit in the gap between d and d¢, while the N terminus of w interacts with the proposed binding region of c inside the C-terminal collar of the clamploader complex conservation on the surfaces of the d and d¢ subunits suggests that the docking of the v:w unit onto the clamploader may be loose, mediated primarily by the flexible N-terminal segment of w This... to the clamp-loader complex It is known that w interacts with domain III of c [22], thus bridging the clamp-loader and v To identify where the N-terminal region of w might interact with the clamp-loader complex, a sequence alignment was performed for each of the clamp-loader subunits (d, d¢ and c), using sequences from the same bacterial species that were used in the alignment of the sequences of w... is proposed that the this linker binds to the clamp-loader complex, tethering the v:w heterodimer to the complex Ó FEBS 2004 E coli DNA polymerase clamp-loader subassembly (Eur J Biochem 271) 447 Fig 6 Potential clamp-loader: w interaction (A) Two views of the Escherichia coli clamploader complex are shown [17] An exposed hydrophobic region of the c subunit, which is highly conserved but not involved... of this C-terminal region of SSB by interaction with v might be responsible for the stable interaction of single-stranded DNA, SSB and the v:w complex The N-terminal segment of w is a possible linker between v:w and the clamp-loader In the crystals of v:w, the N-terminal 26 residues are disordered and are not present in the structural model While disordered regions are not uncommon within crystal structures,... (2001) s Binds and organizes Escherichia coli replication proteins through distinct domains Domain IV, located within the unique C terminus of s, binds the replication fork, helicase, DnaB J Biol Chem 276, 4441–4446 15 Guenther, B., Onrust, R., Sali, A., O’Donnell, M & Kuriyan, J (1997) Crystal structure of the d-subunit of the clamp-loader complex of E coli DNA polymerase III Cell 91, 335–345 16 Oyama,... frameshifting generates the c subunit of DNA polymerase III holoenzyme Proc Natl Acad Sci USA 87, 2516–2520 8 Flower, A.M & McHenry, C.S (1990) The c subunit of DNA polymerase III holoenzyme of Escherichia coli is produced by ribosomal frameshifting Proc Natl Acad Sci USA 87, 3713– 3717 9 Studwell-Vaughan, P.S & O’Donnell, M (1991) Constitution of the twin polymerase of DNA polymerase III holoenzyme... Phe64 of v into the hydrophobic pocket on ˚ w, and burying 1256 A2 of surface area at the subunit interface (Fig 1B,C) Interestingly, a DNA- dependent DNA polymerase from the E coli bacteriophage, PRD1, has high sequence similarity to this loop region alone of v, extending over 28 residues with 13 identities, suggesting that this DNA polymerase might couple to the clamp-loader complex via the w subunit The. .. the d and d¢ subunits of the clamp-loader, neither v nor w contain any of the functional elements required for nucleotide binding The topology of the w subunit resembles that of the bacterial two-component signaling protein, CheY [43], and the uracil DNA- glycosylases, UDG [44] and MUG [45] (Fig 2A) The v subunit has a central b sheet with seven parallel strands, which curve in a left-handed twist There . Crystal structure of the chi:psi subassembly of the Escherichia coli DNA polymerase clamp-loader complex Jacqueline M. Gulbis 1, *,. archaebacteria] that tether the DNA polymerase to DNA [2]. The interaction between the DNA polymerase and the sliding clamp enables the active site of the polymerase