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

J. Biol. Chem., Vol. 276, Issue 22, 19182-19189, June 1, 2001

This Article Abstract Full Text (PDF) All Versions of this Article: 276/22/19182    most recent M100592200v1 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 Stewart, J. Articles by O'Donnell, M. Search for Related Content PubMed PubMed Citation Articles by Stewart, J. Articles by O'Donnell, M.

Mechanism of  Clamp Opening by the  Subunit of Escherichia coli DNA Polymerase III Holoenzyme^*

Jelena Stewart, Manju M. Hingorani^§, Zvi Kelman^¶, and Mike O'Donnell^**

From  The Rockefeller University and ^** Howard Hughes Medical Institute, Laboratory of DNA Replication, New York, New York 10021 and ^¶ Department of Microbiology, Cornell University Medical College, New York, New York 10021

Received for publication, January 22, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The  sliding clamp encircles the primer-template and tethers DNA polymerase III holoenzyme to DNA for processive replication of the Escherichia coli genome. The clamp is formed via hydrophobic and ionic interactions between two semicircular  monomers. This report demonstrates that the  dimer is a stable closed ring and is not monomerized when the  complex clamp loader (₃₁₁₁₁) assembles the  ring around DNA.  is the subunit of the  complex that binds  and opens the ring; it also does not appear to monomerize . Point mutations were introduced at the  dimer interface to test its structural integrity and gain insight into its interaction with . Mutation of two residues at the dimer interface of , I272A/L273A, yields a stable  monomer. We find that  binds the  monomer mutant at least 50-fold tighter than the  dimer. These findings suggest that when  interacts with the  clamp, it binds one  subunit with high affinity and utilizes some of that binding energy to perform work on the dimeric clamp, probably cracking one dimer interface open.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The DNA polymerase III (Pol III)¹ holoenzyme is primarily responsible for replicating the 4.4-megabase Escherichia coli genome (1, 2). Pol III holoenzyme performs this task with high speed and accuracy with the help of ten component subunits. These are  (the DNA polymerase (3)),  (the proofreading 3'-5' exonuclease (4)), and  (unknown function) that form the DNA polymerase III core (5);  (the sliding clamp (6, 7)); and the multisubunit DnaX complex (') that functions as the clamp loader (8-10) and contains at least two subunits of the  "organizer" that binds two core polymerases (11-13) and connects to the DnaB helicase at the replication fork (14, 15).

Rapid and processive DNA synthesis by Pol III holoenzyme is dependent on the interaction between the  subunit of the core polymerase and the  clamp (6).  is a ring-shaped protein that encircles double-stranded DNA and can slide freely along its length (6, 7). By itself, core polymerase can extend a primer by only a few nucleotides before dissociating from DNA (16). When  is bound to the polymerase and topologicaly linked to the primer-template, it serves as a mobile tether to keep the enzyme associated with DNA, facilitating replication of several thousand nucleotides at a time. Similar mechanisms for processive DNA synthesis by replicative polymerases have been discovered in a variety of other organisms (reviewed in Refs. 2, 17, 18, and 19), including eukaryotic DNA polymerase  (tethered to DNA by the PCNA sliding clamp (20, 21)) and bacteriophage T4 DNA polymerase, gp43 (tethered by the gp45 sliding clamp (22)).

The crystal structure of  shows it to be a ring-shaped dimer formed by the head-to-tail interaction of two semicircle-shaped monomers (7). A continuous -sheet forms a scaffold around the outer surface of the ring that supports 12 -helices lining the inside of the ring. The central cavity is about 35 Å in diameter, which is large enough to encircle double-stranded DNA as well as one or two layers of water molecules. Moreover, although the inside of the  ring is positively charged, it lacks specific contact with DNA, allowing  to form a stable topological link with the DNA and yet slide freely along the duplex. At the two identical dimer interfaces, a continuous -sheet formed by hydrogen bonding between  strands from each monomer stabilizes the ring structure in addition to a small hydrophobic core formed by packing of Ile²⁷² and Leu²⁷³ of one monomer with Phe¹⁰⁶ and Leu¹⁰⁸ on the other monomer. Charged amino acids at the interface are also in position to form six ion pairs (these interactions are detailed in Fig. 4). These numerous and potentially strong interactions between the two  subunits presumably underlie the highly stable dimeric structure of  and its ability to remain bound to DNA with a half-life of over 100 min (23, 24). Yet the closed circular clamp must be opened frequently during DNA replication for assembly on DNA to initiate processive replication as well as for disassembly of the  ring from DNA when replication is complete.

The  complex clamp loader (') assembles  clamps on primer-template DNA (where they can be used by the polymerase) and can also remove clamps from DNA when necessary (23-26). The process of clamp assembly requires that the  complex open the  clamp, guide DNA into the central cavity, and facilitate closure of the clamp around DNA. Crystal structure analysis,² and a recent biochemical study (27) reveals that the  complex contains three copies of ; the other subunits (, ', , ) are each present in a single copy (10, 13). The  subunit of  complex binds to  and destabilizes or opens the dimer interface (28, 29). The  subunits are the only ones that hydrolyze ATP (30-32). The ' subunit is homologous to  and appears to play a role in modulating the access of  to  (10, 33, 34). In the absence of ATP, the affinity between the  complex and  is low compared with the affinity between the  subunit and  (28). Clamp assembly initiates when ATP binds the  subunits and induces a change in conformation of the  complex that results in ability of  to bind  (28, 29, 32). The ' subunit appears at least partially responsible for modulating the access of  to , since a previous study indicated that ' and  compete for interaction with  (29). The ATP-induced conformational change of  complex may entail removing a surface of ' from , allowing  to bind and open the  clamp. In the presence of a nonhydrolyzable ATP analogue, the clamp loader- complex binds primer-template DNA with high affinity (32, 36). Interaction of  complex with DNA, especially primed template, triggers ATP hydrolysis and is stimulated by the presence of  (29, 32, 36, 37). ATP hydrolysis is coupled to closure of the clamp around DNA and  complex turnover. The  subunit of  complex binds to SSB and helps coordinate the switch between the primase, clamp loader, and polymerase proteins at the primer template (38, 39), and  enhances the stability of the  complex; however, these two proteins are not absolutely essential for clamp assembly (40-43).

Although all three subunits, , , and ', are required for loading  onto DNA, the single  subunit appears to be the predominant contact between  and the  complex (28). It remains possible that weaker interactions between  and the other  complex subunits exist.³ However, our previous studies demonstrated that  alone can open and remove  clamps from circular DNA molecules with nearly the same efficiency as  complex (k_{unloading complex} = 0.015 s¹; k _unloading = 0.011 s¹) (24). We were therefore curious as to how the  subunit generates the leverage required to part the apparently tightly closed  dimer interfaces. Previous studies indicate that  opening at just one interface is sufficient to allow passage of DNA into (or out of) the central cavity (29). Experiments herein measure the exchange of labeled  subunits as they are utilized by the  complex, and the results support the conclusion that the dimeric clamp is not split apart into monomers but rather stays intact during clamp assembly, presumably opening at only one interface for entry of DNA. In the simplest possible mechanism, the clamp loader could prompt clamp opening merely by perturbing one of the dimer interfaces and transiently reducing its stability.

Study of the - interaction in this report provides insight into how the  and  complex might open the  ring. We demonstrate here that the  ring retains its dimeric structure when bound by one  subunit. Furthermore, we have mutated two hydrophobic residues in the  dimer interface to produce a stable monomeric version of . Only one  subunit binds the  monomer, which is surprising, given the one /two  stoichiometry of the wild type - complex. This suggests that the binding site of  on the  ring is located primarily on one of the two  subunits. The affinity of  for the  monomer mutant is about 50-fold greater than for the  dimer, implying that the binding energy of  to a single  subunit of the dimer is harnessed to perform work, namely to force open one of the dimer interfaces. The  subunit binds  at the carboxyl terminus, which lies in the vicinity of the dimer interface (44). Therefore, it is conceivable that  binding to one  protomer disrupts the contacts in a nearby dimer interface that hold the ring closed.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Nucleotides, DNAs, and Buffers-- Radioactive nucleotides were purchased from PerkinElmer Life Sciences. Unlabeled nucleotides were purchased from Amersham Pharmacia Biotech. M13mp18 ssDNA was prepared by phenol extraction of purified M13mp18 phage that had been banded twice in CsCl gradients (45) and primed with a 30-nucleotide primer (Life Technologies, Inc.) as described (46). Buffer A contained 20 mM Tris-HCl (pH 7.5), 0.5 mM EDTA (pH 8.0), 100 mM NaCl, and 10% glycerol. DNA replication buffer contained 20 mM Tris-HCl (pH 7.5), 0.1 mM EDTA, 40 µg/ml bovine serum albumin, 5 mM dithiothreitol, 8 mM MgCl₂, 4% glycerol, 0.5 mM ATP, 60 µM dGTP, and 60 µM dCTP. Surface plasmon resonance (SPR) buffer contained 10 mM Hepes-NaOH (pH 7.4), 150 mM NaCl, 3.4 mM EDTA, and 0.005% Tween 20.

Proteins-- Proteins were purified as described: , ,  (46), , ', ,  (33),  (47), and SSB (46).  complex and Pol III* (a subcomplex of Pol III holoenzyme lacking the  subunit) were reconstituted from individual subunits and purified as described in Refs. 9 and 13, respectively. Mutant  proteins were constructed using DNA oligonucleotide site-directed mutagenesis. Various N-terminal tagged versions of  (described below) were purified according to the previously described protocol for wild type  (7). Radiolabeling of tagged  with ³²P was performed using [-³²P]ATP and cAMP-dependent protein kinase to a specific activity of ~100 cpm/fmol as described (48). The catalytic subunit of cAMP-dependent protein kinase produced in E. coli was a gift from Dr. Susan Taylor (University of California, San Diego). ³H- was labeled by reductive methylation as described (48).

Gel Filtration Analysis of , , L273A-, and I272A/L273A--- The , L273A-, and I272A/L273A- proteins (3 µM as dimer) were sized by gel filtration (at 4 °C) on an FPLC HR 10/30 Superose 12 column (Amersham Pharmacia Biotech) equilibrated with Buffer A. The proteins were incubated in a final volume of 200 µl of Buffer A for 15 min at 15 °C and then applied to the column. After collecting 6-ml, 170-µl fractions were collected, and 25-µl aliquots of the indicated fractions were analyzed by SDS-polyacrylamide gel electrophoresis (15% gels); proteins were visualized by Coomassie Blue staining. For size standards,  (130 kDa), bovine serum albumin (66 kDa), and  (39 kDa) were analyzed similarly.

Interaction between  and  was analyzed by incubating 9 µM  with 12.5 µM wild type  (as dimer) or 25 µM I272A/L273A-, (as monomer) for 15 min at 15 °C in a final volume of 200 µl of Buffer A, followed by gel filtration chromatography and SDS-polyacrylamide gel electrophoresis analysis as described above.

DNA Replication Assays-- Singly primed M13mp18 ssDNA (20 fmol), 0.8 µg of SSB, 75 fmol of Pol III*, and 750 fmol of  (wild type and mutant concentrations are calculated as monomer) were incubated at 37 °C for 2 min in 25 µl (final volume) of DNA replication buffer (this buffer contains ATP, dCTP, and dGTP). DNA synthesis was initiated upon the addition of the remaining two deoxyribonucleoside triphosphates (60 µM dATP, 20 µM dTTP (final concentrations), and 1 µCi of [-³²P]dTTP). After 20 s, reactions were quenched with 25 µl of 40 mM EDTA and 1% SDS. Aliquots (20 µl) of the quenched reactions were analyzed by electrophoresis on a 1% TBE-agarose gel, and the radiolabeled DNA was visualized on a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA). Synthesis was quantitated by spotting 20-µl aliquots of the reaction on DE81 filters, followed by liquid scintillation counting as described (49).

SPR Analysis of - Interaction-- The  subunit (10 µl of 0.6 µM) was immobilized on a carboxymethylated dextran matrix-coated sensor chip (CM5; Biacore) by carbodiimide coupling in 10 mM sodium acetate (pH 5.5). SPR analysis was performed at 23 °C by injecting 15 µl of  or I272A/L273A (0.25 and 1.23 µM; concentrations for both are given as monomer) in SPR buffer, at a flow rate of 5 µl/s. After each analysis was complete, the chip surface was regenerated by injecting 10 µl of 0.1 M glycine (pH 9.5) over the chip, which releases bound  with no significant effect on the binding capacity of the immobilized  protein.

The kinetic constants for interaction between  and  were determined by nonlinear curve fitting, using the BIAevaluation 2.1 software. The rate of dissociation (k_off) was calculated by fitting the curves to a single exponential decay described by Equation 1,

(Eq. 1)

where R₀ represents the response and t₀ represents the time at the start of the dissociation phase. The association rate (k_on) was calculated using the binding model A + B= AB and Equation 2,

(Eq. 2)

where R_eq is the response at steady state, C is the concentration of , and t₀ is the time at the start of the association phase. The dissociation constant (K_d) for interaction between  and  was calculated as k_off/k_on.

Protomer Exchange Assay-- The two  mutants for this assay were constructed by placing the  gene into either the pHKEp vector or the pHKEp^mut vector (50). Both of these vectors place a 34-amino acid tag onto the N terminus of the protein. The tags contain a protein kinase site (to label the protein with ³²P) and either a functional (pHKEp) or a nonfunctional (pHKEp^mut) hemagglutinin (HA) epitope. The nonfunctional epitope was formed by replacing two amino acids; YPYDVPDYA was changed to YPYDVPAAA. After expression and purification, one  contains a functional HA epitope (^ha₂) and the other  contains a nonfunctional HA epitope, which we use in this report in the phosphorylated form and refer to as ³²P-₂. The  with the mutated HA-epitope was labeled with ³²P (³²P-) as described (48). Titrations of these  variants showed that they were as active as wild type  in replication assays with Pol III* on SSB-coated M13mp18 ssDNA primed with a single oligonucleotide. Monoclonal antibody to the HA epitope was purchased from BabCo, and Protein A-Sepharose 4B was from Zymed Laboratories Inc. The HA antibody was conjugated to Protein A beads by incubation for 15 min at 25 °C in 400 µl of 20 mM Hepes (pH 7.4), 150 mM NaCl, 0.1% Triton X-100, 10% glycerol.

Spontaneous protomer exchange was measured (i.e. no other proteins besides ) in 50-µl reactions containing 2 pmol of ³²P-₂ and 2 pmol of ^ha₂ in 20 mM Hepes (pH 7.5), 150 mM NaCl, and 10% glycerol. Reactions were incubated at 37 °C for 0, 1, 2, 4, 6, or 8 h before the addition of 50 µl of HA antibody-conjugated beads and placed at 4 °C for a further 30 min. Beads were pelleted, washed three times with 1 ml of 50 mM Tris-HCl (pH 8.0), 500 mM NaCl, 5 mM EDTA, 0.1% SDS, and 0.1% Triton X-100; resuspended in Eco-Lume (ICN); and counted in a scintillation counter. Control experiments were performed similarly except that either no antibody was conjugated to the beads or the ^ha₂ was not added to the reaction.

To measure the effect of  complex on  protomer exchange during clamp assembly onto DNA, 250 fmol each of ^ha₂ and ³²P-₂ were incubated for 5 min at 37 °C with 500 fmol of  complex and 1.8 pmol of nicked pBS DNA in 70 µl of 20 nM Tris-HCl (pH. 7.5), 0.1 mM EDTA, 4% glycerol, and 8 mM MgCl₂. The reaction was then applied to a 5-ml A15 M gel filtration column equilibrated with the same buffer plus 0.15 M NaCl. Fractions of six drops each were collected, and those containing  on DNA were identified by scintillation counting and pooled (420 µl), and then the DNA was linearized upon treatment with 700 units of BamHI for 3 min at 37 °C to release . To confirm that linearization was complete within this time, an aliquot (20 µl) was removed, quenched with 20 µl of 1% SDS, 40 mM EDTA, and then analyzed in a native agarose gel. Then 50 µl of HA antibody beads were added to the reaction, and incubation was continued for a further 30 min at 4 °C. The beads were pelleted; washed three times with 1 ml of 50 mM Tris-HCl (pH 8.0), 500 mM NaCl, 5 mM EDTA, 0.1% SDS, and 0.1% Triton X-100; resuspended in Eco-Lume (ICN); and counted in a scintillation counter. In a control experiment, the above procedure was repeated except that heterodimeric  was used in the assembly reaction by first preincubating 250 fmol of each  in one tube for 5 h at 37 °C before adding to the reaction containing  complex and DNA.

Nickel Column Affinity Assay for -₂ Complex-- Reactions contained 67.5 pmol of ³H-₂ (wild type  labeled by reductive methylation), 1.7 nmol of ^his₂, which contained a six-residue histidine tag on a 23-residue N-terminal leader ( was cloned into the pHK vector in Ref. 50), and 6.6 nmol of  (where present) in 200 µl of 20 mM Tris-HCl (pH 7.5), 10% glycerol, 8 mM MgCl₂, and 100 mM NaCl. A control reaction utilized 1.7 nmol of unlabeled wild type ₂ in placed of the ^his₂ derivative. Reactions were assembled on ice and then shifted to 37 °C, and aliquots of 20 µl were removed at 2 and 24 h of incubation. Upon removal of an aliquot, NaCl was added to a final concentration of 0.5 M, and the reaction was applied to a 1-ml nickel chelate column (HiTrap; Amersham Pharmacia Biotech) equilibrated in 20 mM Tris-HCl (pH 7.9), 5 mM imidazole, 8 mM MgCl₂, and 10% glycerol. The column was washed with 5 ml of the same buffer and then eluted with 3 ml of 20 mM Tris-HCl (pH 7.9), 1 M imidazole, 8 mM MgCl₂, and 10% glycerol. Fractions of 1 ml were collected. The flow-through (wash) and bound (elution) fractions were analyzed by liquid scintillation counting and analyzed in a 10% SDS-polyacrylamide gel to confirm the presence of  with  in the bound fractions. The typical yield of ³H-₂ off the column was greater than 85%.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Is Not Monomerized during Assembly onto DNA-- We have shown previously that the  clamp is a tight dimer and remains a dimer even when diluted to a concentration of 50 nM (23). Nonetheless, it is possible that  complex dissociates the  dimer into monomers using the energy of ATP hydrolysis and then reassembles the  dimer onto DNA in a second step. To test this possibility, we constructed two chemically distinct  mutants; one was phosphorylated and contained a protein kinase tag (³²P-^pk), and the other had a hemagglutinin epitope tag (^ha). If the  complex monomerizes  dimers and reassembles them onto DNA, then it should act upon a mixture of ³²P-₂ and ^ha₂ to form ³²P--^ha heterodimers on DNA.

As a prerequisite for an experiment of this type, it is important that the ³²P-₂ and ^ha₂ mixture does not undergo spontaneous protomer exchange to form heterodimers during the time of the experiment. The time course for spontaneous heterodimer formation was measured in the experiment of Fig. 1 by mixing equal amounts of ³²P-₂ and ^ha₂, followed by removal of aliquots at time intervals and immunoprecipitation of the mixture using Protein A beads to which an antibody to hemagglutinin is attached. Initially, ³²P- will not be precipitated, since it lacks the epitope. But as protomer exchange occurs, the ³²P--^ha heterodimer will be formed, which should result in the appearance of radioactivity in the pellet. The result, shown in Fig. 1, demonstrates that the time scale of spontaneous subunit exchange is on the order of hours (t_1/2 ~ 2 h). As the clamp-loading reaction only requires 5 min, spontaneous protomer exchange during the reaction should be nearly negligible. Control reactions not shown here have been performed that demonstrate requirements for both the antibody and the presence of the ^ha to detect radioactivity attached to the beads in the pellet. In both controls, the pellets lacked radioactivity above background levels (0.5 fmol of ³²P-₂).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Time course of  protomer exchange. Two chemically distinct species of , one containing a hemagglutinin epitope and one labeled with 32P, were mixed together to initiate formation of heterodimers as indicated. After various times of incubation, aliquots were withdrawn, and beads to which antibody to the hemagglutinin tag were attached were added. Heterodimeric  consists of a 32P- protomer attached to a ha protomer that should be trapped by the hemagglutinin beads. Radioactivity in the pellet, representing heterodimeric , is plotted with respect to time. I.P., immunoprecipitation.

Next, we examined how the  complex loads a mixture of these two  variants onto DNA to determine whether it catalyzes protomer exchange during the clamp assembly process (i.e. whether  complex breaks  dimers apart and reassembles them onto DNA as illustrated in the scheme of Fig. 2). To test this possible action, ³²P-₂ and ^ha₂ were mixed together, and  complex was added immediately along with ATP and circular plasmid DNA containing a single nick to initiate clamp assembly. After 5 min, the reaction was applied to a gel filtration column to separate clamps that had been assembled on DNA from those remaining in solution. Following this, the isolated -DNA complex was treated with BamHI to rapidly linearize the DNA, allowing the clamps to slide off DNA into solution. Then the reaction was analyzed for heterodimer formation by immunoprecipitation using the hemagglutinin antibody beads (see the scheme in Fig. 2).

View larger version (29K): [in this window] [in a new window]   Fig. 2.   complex does not monomerize  during clamp assembly onto DNA. The  complex and ATP were used to assemble a mixture of 32P-2 and ha2 onto nicked circular plasmid DNA. As indicted, the -DNA complex was isolated from free  and then linearized to release  into solution. Hemagglutinin antibody bound to beads was used to quantitate heterodimers that were assembled onto DNA by the  complex. Lane 1, 32P-2 and ha2 were mixed immediately prior to their transfer to DNA by  complex; lane 2, 32P-2 and ha2 were premixed and incubated 5 h to form heterodimers before being transferred to DNA by  complex.

The results of this experiment (Fig. 2, lane 1) demonstrate that very little heterodimeric  is formed in the reaction, indicating that  complex does not catalyze  monomerization during clamp loading. In control reactions not shown here, we confirmed that  complex loads approximately equal amounts of ³²P-₂ and ³²P-^ha₂ on DNA, and both variants of  were as active as wild type  in replication assays with PolIII*. In another control experiment ³²P-₂ and ^ha₂ were premixed for 6 h to form the ³²P--^ha heterodimer prior to use by  complex in assembly onto DNA. The result, shown in Fig. 2, lane 2, demonstrates that the experimental strategy is functional in detecting heterodimers that are assembled on DNA. Thus, it would appear that  complex does not monomerize  but probably only opens one interface of the ring during the clamp opening process. This conclusion is consistent with a previous finding that showed that  complex was capable of assembling a  dimer onto DNA that was cross-linked at one interface by a disulfide bond (i.e. indicating that  complex does not need to open both interfaces to assemble  onto DNA (29)).

In the study of Fig. 3 we designed another experiment to examine the oligomeric state of  during clamp assembly, this time while it is in complex with , the clamp-opening subunit of the  complex. Previous studies indicated that one  monomer binds to the  dimer, consistent with the single copy of  in  complex (28). The  subunit is capable of removing  rings from circular DNA (24, 29) and thus must either destabilize one interface or perhaps transiently dissociate  into monomers. In either case, one may expect  to accelerate the rate of protomer exchange. We examined these possibilities in a variation of the protomer exchange assay. The assay utilized a hexahistidine-tagged ₂ (^his₂) and tritiated wild type ₂ (³H-₂). The ³H-₂ was mixed with a 25-fold molar excess of ^his₂ in the presence or absence of a 4-fold molar excess of  (over total ), and then the mixture was analyzed at either 2 or 24 h for heterodimer formation by nickel chelate chromatography. Homodimeric ³H-₂ should not bind to the column (flow-through fraction), and heterodimeric ³H--^his should be retained (bound fraction) and detected by elution from the nickel chelate column, followed by scintillation counting.

View larger version (45K): [in this window] [in a new window]   Fig. 3.   does not monomerize 2. A mixture of 3H-2 and a histidine-tagged  (his) were mixed in the presence or absence of a 4-fold excess of  subunit (over total ). At either 2 h (lanes 2 and 3) or 24 h (lanes 5 and 6), aliquots were removed and loaded onto nickel-chelate columns. After washing the columns, bound protein was eluted with buffer containing 1 M imidazole. Since 3H- lacks a His tag, 3H- in the bound fraction represents 3H--his heterodimers. Controls in which wild type 2 was substituted for his are shown in lanes 1 and 4.

The results of this experiment, shown in Fig. 3, illustrate that similar amounts of heterodimer are formed within 2 h in the presence or absence of , indicating that  does not appreciably speed up protomer interchange. Also, the fact that ³H- is retained on the column in the presence of  supports the ₁-₂ stoichiometry, since if  monomerized ₂, heterodimer would not be present for retention on the column. As a control, wild type ₂ was substituted for ^his₂, which should form a ³H--^wt heterodimer, but should not bind the nickel chelate column. The result of this control showed that ³H- was not retained on the column, as expected (not shown).

How Does  Open One Interface of the  Ring?-- The experiments described above demonstrate that the  complex does not catalyze the exchange of  protomers during the clamp loading operation. The results also demonstrate that  does not monomerize the  dimer. These results support and extend earlier studies that indicate that only one interface of the  dimer ring is cracked open during assembly onto DNA. The  subunit is the clamp-opening subunit of  complex. How does  open an interface of the  ring? To gain insight into how  performs its ring opening task, we mutated  to form a stable  monomer. Initially, we set out to determine whether  mainly binds only one protomer of the  dimer, in which case  should still bind a  monomer about as well as a  dimer. Alternatively,  may need to associate with elements on both protomers of ₂ in order to establish a firm grip on the  ring. The results of this line of investigation were unexpected and provided significant insight into the clamp opening function of .

To form a stable monomer of , we utilized the crystal structure to design site-specific mutations that would destabilize the dimer interface. The crystal structure of the dimeric  clamp revealed a small interface between the two  subunits that, despite its size, has an abundance of potentially strong interactions (see the diagrams in Fig. 4, A and B) (7). These interactions facilitate the formation of a highly stable circular clamp that maintains its dimeric structure even at low nanomolar concentrations. In particular, a small hydrophobic core of four amino acid residues (Phe¹⁰⁶, Leu¹⁰⁸, Ile²⁷², and Leu²⁷³) at the dimer interface appears to play an important role in the stability of the clamp structure. Initially, we constructed three single residue mutants in which Ala was substituted in place of either Phe¹⁰⁶, Leu¹⁰⁸, or Leu²⁷³ (we could not obtain the I272A mutant). Each of these point mutants migrated as a dimer in gel filtration analysis and retained 70-100% activity with PolIII* (not shown). However, a double mutant, I272A/I273A, behaved as a monomer and lacked replication activity (explained below). The experiments to follow focus on the double mutant and compare it with wild type .

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Mutation of the  interface to form a stable  monomer. A, the crystal structure of the  dimer shows that it is a ring-shaped structure, with a 35-Å central hole, wide enough to encircle double-stranded DNA (7). The  dimer interface, in particular the amino acid residues forming the hydrophobic core, is highlighted on the right. B, the dimer interface consists of two  sheets, 8 and '4 (one from each subunit), that form an antiparallel sheet across the dimer interface and neighboring amino acid contacts that stabilize the clamp structure. The lines connect amino acid residues predicted to form six ion pairs across the interface. The hydrophobic core residues, Phe106, Leu108, Ile272, and Leu273, are indicated, and gray circles highlight Ile272 and Leu273, which were mutated to Ala in this study. C, analysis of wild type  and mutants of  on a sizing column followed by SDS-polyacrylamide gel electrophoresis. The results show that wild type  (top panel) and L273A (middle panel) elute as dimers (81.2 kDa). In contrast, the I272A/L273A- (bottom panel) elutes as a smaller, monomeric protein (40.6 kDa), indicating that the double mutation severely disrupts the  dimer interface. D, quantitation of DNA synthesis by Pol III* on SSB-coated singly primed M13mp18 ssDNA in the presence of either no , I272A/L273A-, or wild type .

In Fig. 4C, wild type  and the  mutants, L273A- and I272A/L273A-, were examined by gel filtration to determine their oligomeric state. Fig. 4C shows the SDS-polyacrylamide gel electrophoresis analysis of column fractions from the gel filtration analysis of wild type , L273A-, and I272A/L273A-, in the top, middle, and bottom panel, respectively. Wild type  elutes as a dimer in peak fraction 21, as does the L273A- mutant (calculated mass = 81.2 kDa). In contrast, the double mutant I272A/L273A- migrates more slowly through the column, indicative of a smaller size, and elutes as a monomer (calculated mass = 40.6 kDa). The gel filtration experiments were performed with 3 µM  (as dimer). Therefore, even at high protein concentration, I272A/L273A- is unable to form a stable dimer.

It has long been presumed that the circular structure of the  dimer is required for its action as a DNA polymerase processivity factor. There are, however, single subunit processivity factors that do not appear to encircle DNA, particularly the herpes simplex virus UL42 protein, which in fact is structurally similar to the eukaryotic PCNA clamp but does not oligomerize into a ring (51). To determine if a monomeric form of  can serve as a processivity factor, the monomeric I272A/L273A- mutant was tested for DNA replication activity with PolIII* using primed M13mp18 ssDNA as substrate. The result, in Fig. 4D, demonstrates that the monomeric  mutant is inactive with PolIII*. The dimeric single mutants (L273A, L108A, and F106A) retained 70-100% the activity of wild type  (not shown). Thus, a  monomer that does not form a circular clamp is not capable of tethering Pol III* to DNA for processive DNA replication.

Binds the  Monomer with Higher Affinity than the  Dimer-- Only one copy of the  subunit is present in the  complex, consistent with the stoichiometry of one  to two  in the - complex. The stoichiometry of only one  subunit per  dimer invokes the question of whether  interacts with both  protomers or can stably attach to one  protomer, perhaps somehow preventing a second  from binding the other  protomer (e.g. by steric occlusion). Interaction of  with the  monomer was tested in Fig. 5 by mixing  with an excess of either wild type ₂ or the momeric ₁ mutant, followed by gel filtration analysis on a sizing column. The elution profiles of the proteins were analyzed by SDS-polyacrylamide gel electrophoresis. As expected from previous studies, Fig. 5A shows that  and wild type ₂ form a stable complex with an apparent molecular mass of 111 kDa, consistent with the ₁₂ complex observed in our previous study (38.7-kDa  + 2 × 40.6-kDa  = 119.9 kDa) (28). Fig. 5B shows that  and the I272A/L273A- mutant also interact, forming a smaller ₁₁ complex that migrates at an intermediate position between the ₁₂ complex (Figs. 5A) and the free I272A/L273A- monomer (Fig. 5E). This result reveals that the binding site for  on the  clamp resides within one monomer and demonstrates that  need not bind both subunits of the dimer to form a stable contact with the clamp.

View larger version (64K): [in this window] [in a new window]   Fig. 5.   binds the  monomer mutant. - interaction was analyzed by gel filtration on a sizing column, and the proteins were visualized in column fractions by SDS-polyacrylamide gel electrophoresis and Coomassie staining. The complex of the wild type  dimer with  (A) elutes faster than free  (C) and free  (D), indicative of its large size (81.2 + 38.7-kDa 12 complex). B, the monomeric I272A/L273A- protein also interacts stably with , forming a 11 complex that elutes in an intermediate position between the 12 complex (A) and either  (C) or free  monomer (E). The elution positions of molecular weight standards are shown at the bottom of the gel.

The gel filtration analysis revealed that  can bind a single  protomer, but the possibility remained that the affinity of  for  may be affected by disruption of the dimeric structure. In particular, we noticed that during gel filtration  trails as free protein from a complex with wild type  (fractions 24-31 in Fig. 5A), whereas in Fig. 5B most of the  appears in complex with the  monomer, suggesting that  may bind ₁ tighter than ₂. Next, we used the SPR technique to examine more closely the relative affinity between  and the  dimer versus the I272A/L273A- monomer mutant (Fig. 6). The  subunit was immobilized on a sensor chip, and a solution of  in buffer (at different concentrations) was passed over it. The increase in mass (response units) resulting from interaction between  and  was measured over time; this is the association phase from which the association rate (k_on) can be calculated. Next, buffer lacking  was passed over the - complex on the chip, and the resulting decrease in mass over time provides information from which the dissociation rate (k_off) can be calculated. Fig. 6, A and B, shows sensorgrams of the interaction between  and two different concentrations of  and I272A/L273A-, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   binds the  monomer tighter than the  dimer. The sensorgrams of wild type  (A), and I272A/L273A- (B) were obtained by measuring the increase in response units when a 0.25 or 1.23 µM solution of either  (as dimer for wild type , as monomer for mutant ) was passed over  immobilized on a sensor chip. The sensorgrams were analyzed for kinetic and equilibrium parameters of the - interaction as described under "Experimental Procedures" (summarized in Table I). The Kd values indicate that  binds I272A/L273A- with ~80-fold higher affinity than the wild type  dimer.

Kinetic analysis of the SPR data yielded the association (k_on) and dissociation (k_off) rate constants, from which the equilibrium dissociation constant for the - interaction could be calculated (K_d = k_off/k_on). The parameters, summarized in Table I, reveal that  binds the  monomer mutant with substantially higher affinity than the wild type  protein. The average K_d value for interaction between  and  is about 0.46 µM (average of values determined at 0.25 and 1.23 µM  concentrations), which is ~57-fold higher than the K_d for the interaction between  and I272A/L273A- (average K_d = 0.0075 µM). The tighter interaction between  and the  monomer is particularly striking because the  monomer has only one potential  binding site, in contrast to the  dimer.

View this table: [in this window] [in a new window]   Table I Kinetic and equilibrium constants describing the interaction between  and

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The  Subunit Remains Dimeric during Clamp Loading-- This study has examined whether  complex monomerizes  during the clamp loading operation but could detect no evidence for splitting of  dimers during their assembly onto DNA. Consistent with retention of the  dimeric state, the  subunit does not appear to monomerize ₂ or to significantly increase the rate of protomer exchange among  dimers. Hence, it seems likely that the  subunit opens only one interface of the  dimer during clamp loading, consistent with the ability of  complex to load a  dimer onto DNA that has been cross-linked via a disulfide bond across one of the two interfaces (29).

The  subunit can open  and remove it from DNA but cannot load  onto DNA. The  and ' subunits of the  complex, along with , may orient DNA inside the open ring. The ' subunit, possibly also assisted by , must also sever the -to- contact, allowing the ring to close around DNA. Release of the  complex and closure of the  clamp around the DNA are tied to ATP hydrolysis and are probably coordinated with sensing the appropriate structure of DNA.

The Critical Role of Hydrophobic Interface Contacts in the  Clamp Structure-- Two  monomers contact each other in a head-to-tail fashion at two small, identical interfaces to form the ring-shaped, dimeric  clamp (Ref. 7; see also Fig. 4). A central feature of the  clamp structure is the continuous layer of sheet around the entire molecule, including the dimer interfaces. Further, particular hydrophobic amino acid side chains, contributed by each monomer, pack to form a small hydrophobic core within each interface. There are also six potential ion pairs formed at the interface, which may further strengthen the dimer. Earlier studies demonstrated that the  clamp retains its dimeric structure even when it is highly dilute (23). It is possible that the  clamp may "breathe" by alternate opening and closing of one or the other interface. However, the observed long lifetime of  on circular DNA when topologically linked to it (23, 24) indicates that if there is breathing at the interfaces, opening a wide distance (i.e. to slip off DNA) is a rare occurrence.

This study confirms an important role of the hydrophobic residues at the interface in maintaining a stable  clamp structure. Although  retained its dimeric status upon mutation of only one hydrophobic residue at the interface,⁴ mutation of two of the amino acids at the hydrophobic core, Ile²⁷² and Leu²⁷³, to Ala destabilizes the two interfaces to such an extent that the  dimer exists as a stable monomer in solution.

The -to- Binding Energy Opens the  Clamp-- In this study, we show that  binds the  monomer about 50-fold tighter than the  dimer. Moreover, the tight - monomer complex has a 1:1 stoichiometry, indicating that  has a binding site for only one  protomer. Thus, in the ₁-₂ complex,  probably binds only one of the  subunits. The apparent higher affinity of  for the  monomer mutant compared with wild type  also indicates that some of the binding energy of  to a  protomer is put into performing work on the  dimer, thus lowering the observed affinity. Given that upon binding of  to , the dimer opens, the work of the "lost" binding energy is probably utilized to part one of the  dimer interfaces.

The amount of work to open one interface can be calculated to be ~2.4 kcal, assuming a difference of 57-fold in the equilibrium binding constants for  binding to either ₁ or to ₂ (since  does not monomerize, there is no entropy component, and the free energy represents work). It is interesting to note that  remains a dimer well below 50 nM (23), and thus the free energy for dissociation to monomers is in excess of 10 kcal. These calculated free energies imply that the amount of work required to open one interface (i.e. ~2.4 kcal) is far less than the free energy to open the second interface (i.e. the full 10 kcal needed for  to monomerize). These results imply that the  dimer is constructed in such a way as to ensure preservation of a dimeric structure, even after one interface has been pried open.

The  Complex Mechanism-- The high stability of the wild type  dimeric clamp explains the need for a clamp loader/unloader protein during DNA replication. The  complex serves this function by binding and opening the  clamp when it must be loaded onto primer-template DNA or unloaded from a newly replicated duplex (25, 45). The <