• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Cell. Author manuscript; available in PMC Jul 11, 2008.
Published in final edited form as:
PMCID: PMC2443641



The structure of the E. coli β-clamp polymerase processivity factor has been solved in complex with primed DNA. Interestingly, the clamp directly binds the DNA duplex, and also forms a crystal contact the ssDNA template strand which fits binds into the protein binding pocket of the clamp. We demonstrate that these clamp-DNA interactions function in clamp loading, perhaps by inducing the ring to close around DNA. Clamp binding to template ssDNA may also serve to hold the clamp at a primed site after loading, or during switching of multiple factors on the clamp. Remarkably, the DNA is highly tilted as it passes through the β ring. The pronounced 22° angle of DNA through β may enable DNA to switch between multiple factors bound to a single clamp simply by alternating from one protomer of the ring to the other.


Chromosomal replicases achieve highly processive DNA synthesis through attachment to a ring-shaped sliding clamp processivity factor (reviewed in (Johnson and O'Donnell, 2005; McHenry, 2003). Sliding clamps in all three domains of life require a multi-protein clamp loader that assembles the clamp onto DNA. The bacterial β clamp is a homodimer; each monomer consists of three globular domains to yield a six-domain ring (Argiriadi et al., 2006; Kong et al., 1992). The eukaryotic sliding clamp, proliferating cell nuclear antigen (PCNA), is also a six domain ring with similar dimensions and chain-fold to bacterial β, except PCNA is a homotrimer and therefore each monomer consists of only two domains (Gulbis et al., 1996; Krishna et al., 1994). The protomers of both β and PCNA are arranged head-to-tail, which results in structurally distinct surfaces on the two “faces” of the clamps. The C-termini protrude from one face, sometimes referred to as the C-terminal face, and the N-termini from the other. Both β and PCNA contain a hydrophobic pocket on the surface of the C-terminal face to which the polymerase and clamp loader attach.

Although β and PCNA were originally identified as processivity factors for their respective chromosomal DNA polymerases, it is now clear that many proteins bind to sliding clamps. For example, all five E. coli DNA polymerases bind the β clamp, as do DNA ligase, MutS, MutL and the Hda cell cycle regulatory factor [reviewed in (Johnson and O'Donnell, 2005)]. Eukaryotic PCNA also interacts with diverse DNA polymerases and many other proteins involved in DNA repair and cell cycle control (Warbrick, 2000). Clamp-binding proteins all appear to bind the hydrophobic protein binding pocket on the C-terminal face of the clamp, although they likely have unique touch points as well (Bunting et al., 2003; Chapados et al., 2004; Gulbis et al., 1996; Jeruzalmi et al., 2001; Matsumiya et al., 2001; Shamoo and Steitz, 1999; Sutton and Walker, 2001; Sutton and Duzen, 2006). Proteins that bind PCNA and β typically do so via a conserved sequence (Dalrymple et al., 2001; Warbrick, 2000), and the homooligomeric structure of β and PCNA may provide the ability to bind more than one protein at the same time. Mechanisms by which multiple proteins coordinate action on sliding clamps is a topic of active investigation (Fujii and Fuchs, 2004; Goodman, 2002; Indiani et al., 2005; Sutton, 2004; Sutton and Walker, 2001).

DNA is generally modeled straight through β and PCNA, perpendicular to the plane of the ring. Interestingly, recent molecular simulations of PCNA suggest that DNA may adopt a tilt of up to 20° through the ring (Ivanov et al., 2006). However, a sliding clamp bound to DNA has not been directly observed and therefore we set out to obtain a co-crystal of the E. coli β clamp bound to DNA. We were surprised to find in biochemical studies reported here that β binds DNA specifically, and interacts with both double-strand (ds) DNA and single-strand (ss) DNA. These findings encouraged us to use small linear DNAs in attempts to crystallize the β-DNA complex. Conditions that lead to co-crystals of β bound to DNA were developed by devising a rapid visual screen for co-crystals that utilize DNA molecules labeled with a chromophore.

The co-crystal structure presented herein consists of β in complex with a primed DNA site. The structure reveals several interesting β-DNA contacts, including residues on surface loops. The DNA passes through the ring at a steep 22° angle. Moreover the template ssDNA of the primed site forms a crystal contact with the adjacent β where it binds the protein-binding pocket of the clamp. Study of site-specific β mutants confirm that binding of β to both dsDNA and ssDNA play key roles in the clamp loading mechanism.

The fact that the clamp recognizes primed DNA has implications for clamp loading, and also suggests mechanisms by which different proteins that bind the clamp may switch with one another on β and the DNA that it encircles.


Direct interaction of β with DNA

The β clamp binds DNA by encircling it and slides along the duplex. This topological binding mode conceptually precludes formation of ordered co-crystals of DNA bound to β. However, the experiments of Fig. 1A indicate that β binds DNA directly, independent of its circular shape. In this experiment, β was labeled with a fluorophore (βOG) and then circular duplex pET11a plasmid, or HaeIII digested pET11a, was titrated into βOG. The results indicate that the linearized DNA binds βOG, while circular DNA does not, consistent with the absence of a clamp loader in this assay. One may expect that β would bind linear DNA, since linear DNA can diffuse into the closed ring without need for a clamp loader. However, under the dilute conditions of these assays, β that binds linear DNA only by encircling it will rapidly slide back off the DNA, and the amount of DNA-β complex at equilibrium would be expected to be negligible. Yet the observed binding curve indicates saturation and therefore complete β-DNA complex formation.

Fig. 1
β Binds DNA Independent of the Circular Shape of the Clamp

Complete formation of a β-linear DNA complex suggests that β has direct contacts with DNA, irregardless of its ring shape. To test whether binding of β to DNA is independent of the ring shape of the clamp, we repeated the assay using a β monomer mutant (Jeruzalmi et al., 2001). The result, in Fig. 1B, shows that the β monomer mutant still binds linear DNA, and in fact binds circular DNA too. Binding of both linear and circular DNA to the β monomer mutant supports the conclusion that β interacts with DNA directly, independent of its circular shape. Moreover, the fact that the wt β ring does not bind circular DNA suggests that DNA must access the central cavity of β to exhibit binding.

Next the DNA binding assay was used to quantitate the affinity between β and a short synthetic blunt DNA duplex. The result, in Fig. 1C, yields an apparent Kd value of 453 nM. The experiment was then repeated using a short primed template, with the result that β binds a primed site about 4-fold tighter than the blunt duplex (Kd value, 120 nM; Fig. 1C). This finding implies that β has binding sites for both dsDNA and ssDNA regions of a primed site, and this is confirmed in the crystal structure presented below.

Visual screen for co-crystals

β crystallizes easily by itself, and therefore we wished to develop a rapid screen for β-DNA co-crystals, ensuring the presence of DNA in the β crystal before taking it to a synchrotron. A simple and rapid screen for β-DNA co-crystals was provided by using DNA substrates tagged with a chromophore that allow rapid identification of co-crystals by simply inspecting for color. This made it possible to test a variety of conditions and different DNA substrates for co-crystal formation. In the current study, DNA was 5′ end-labeled with Cy5 (blue). DNA structures that vary in their lengths of duplex and ssDNA were examined for co-crystal formation with β. This method led to conditions that produced brightly colored blue crystals using a 10 bp duplex and 5′ ssDNA extensions of various lengths (Fig. 2A). To rule out fortuitous binding of the Cy5 moiety to β, a series of 5′-TAMN labeled DNAs were also tested. This time the resulting crystals were bright red (Fig. 2A). Since the Cy5 and TAMN molecules have very different structures, β likely binds DNA and not the label.

Fig. 2
Co-crystal color screen and electron density of DNA in the β-DNA complex

The strategy of using chromophoric ligands to visually screen co-crystals should generalize to other DNA binding proteins. It should also apply to other types of ligands that can be labeled or are intrinsically chromophoric. Examples are shown in Fig. 1S, in which β is co-crystallized with a TAMN-labeled peptide corresponding to the C-terminal sequence in α (DNA polymerase) which is known to bind β (Lopez de Saro et al., 2003). Fig. 1S also shows colored co-crystals of the yeast PCNA clamp with DNA, and of DNA co-crystals using β from a gram positive bacterium, Streptococcus pyogenes.

Structure of the β-DNA complex

The structure of β bound to a 10/14-mer primed site was solved to 1.92Å resolution by molecular replacement using native β (Kong et al., 1992) as the starting model (Table S1). Overall, the primer template is in reverse orientation and possible reasons for this are described later. However, the duplex DNA is perfectly 2-fold symmetric and thus has no forward or reverse. We therefore initiate description of the structure with how the duplex portion of DNA goes through the β clamp, then discuss how template ssDNA binds to β.

The helical density of all 10 bp was immediately visible from the difference map (Fig. 2B). The DNA is standard B-form even though no restraints were placed during refinement on sugar pucker, torsion angles, or Watson-Crick hydrogen bonding. Analysis with the program 3DNA (Lu and Olson, 2003) shows an average helical twist of 38.81° and rise of ~3.26 Å per bp. Several residues contact DNA, illustrated in Fig. 2C.

Interestingly, DNA is sharply tilted within the central channel, defining an angle of ~22° from the C2 rotation axis of β (Fig. 2B and and3A).3A). The tilt allows dsDNA to make contacts with exposed R24 and Q149 (see Fig. 3A, B), which lie on protruding loops on the C-terminal face of β, the face to which Pol III core and the clamp loader bind (Naktinis et al., 1995). The sharp tilt of DNA through β is consistent with recent PCNA-DNA modeling studies (Ivanov et al., 2006). Interaction with R24, Q149, and other β-DNA contacts (Fig. 2C) likely define the tilt, although a crystal contact to ssDNA (described later) may also contribute to the observed angle of DNA.

Fig. 3
Structure of the β-DNA Complex

The inner surface of β, like processivity clamps of other organisms, is lined by α helices that carry a net positive charge. Some of the basic side chains lining the inside channel of the β ring appear flexible (i.e. disordered, have high B factors, or exhibit different conformations in the two protomers), but become ordered in the co-crystal through contact with the phosphate backbone of DNA.

R24 and Q149 are functional in replication

Among the DNA binding residues of β, highly conserved R24, and less conserved Q149, lie on exposed surface loops of the C-terminal face of β to which proteins bind, suggesting they may be functional with other proteins. We therefore constructed alanine replacement mutants of R24 and Q149, and examined the β mutants in replication activity assays. In Fig. 3C, β is titrated into reactions containing Pol III* (a Pol III core-clamp loader complex) and primed M13 ssDNA coated with SSB. Activity in this assay requires both clamp loading and polymerase extension. The results show that the R24A/Q149A double mutant of β is greatly reduced in replication activity; the single mutants of β show intermediate defects. In the fluorometric assay for DNA binding (i.e. as in Fig. 1) the β mutants show only ~2–3 fold defects, consistent with the presence of other residues (on the inside of β) that interact with DNA (Fig. 2C).

Mutation of R24 and Q149 may alter diffusion of β along DNA, creating a drag on the polymerase during chain extension. To test β mutants in chain extension speed, Pol III core was preincubated with large amounts of β (or β mutants) and singly primed M13 ssDNA along with only two dNTPs, giving time for the Pol III-β complex to assemble on primed ssDNA (i.e. in case clamp loading is defective). Then synchronous chain extension was initiated by adding remaining dNTPs, and timed aliquots were analyzed in a native agarose gel. The product analysis demonstrates that the β mutants are as rapid as wt β in function with Pol III core, and thus R24 and Q149 are not defective in elongation speed with Pol III (Fig. 3D).

To address whether the β mutants are deficient in clamp loading we designed a clamp loading assay using synthetic primed DNA immobilized to magnetic beads (see Fig. 3E). Primed DNA is attached to beads via a 3′ biotin-to-streptavidin linkage and SSB is used to block clamps from sliding off. To monitor clamp loading, β is labeled with 32P via a N-terminal kinase tag, then clamp loading is initiated and quenched at the indicated times. The beads are collected using a magnetic concentrator and 32P-β attached to the immobilized DNA is quantitated. The results demonstrate that the R24A/Q149 double mutant of β is highly deficient in clamp loading. The single amino acid mutants of β show intermediate levels of activity, and are approximately equally defective in the clamp loading assay which may be due to the high amount of β used in this assay, or to other differences between the clamp loading and replication assays. The Kd of each β mutant for binding to the clamp loader is nearly the same as wt β (Fig. S2), and thus the activity defect presumably lies in a central role for DNA-β interaction in the clamp loading mechanism itself (see Discussion).

It is interesting to note previous studies in which mutation of β residues 148–152 to Ala results in a noticeable defect with Pol III, and a more severe defect with Pol IV (Sutton, 2005). Hence, Q149, eliminated in the β148–152 mutant, is not essential and perhaps the deficiency with Pol III may occur through its defect in clamp loading rate, while the greater defect with Pol IV may be due to loss of a direct contact between Pol IV and a mutated residue(s) on β.

ssDNA binds to the protein-binding pocket of β

The ssDNA template of the primed DNA is visible in the electron density, and it forms a crystal contact with an adjacent molecule of β in the crystal lattice (see Fig. 4A and Fig. S4). The fact that the ssDNA-β interaction is a crystal contact (i.e. it is intermolecular) raises the question of whether it is simply a crystal artifact or has physiological significance. However, the contact is in a very interesting place; ssDNA enters the protein binding site of the adjacent β clamp. The reason the ssDNA binds an adjacent clamp (i.e. forms a crystal contact) is because the primed DNA goes through β in reverse of normal. We believe the reason that the primed site is in reverse orientation in the crystal relative to the physiological orientation provided by the clamp loader is that the intermolecular ssDNA-β crystal contact promotes growth of the crystal lattice, and that β is "blind" to the orientation of dsDNA, since the duplex is perfectly 2-fold symmetric.

Fig. 4
Interaction of β with the ssDNA region of the primed site

The clamp loader ensures that β is correctly oriented on DNA, such that the primed/ssDNA junction extrudes from the C-terminal face containing the protein binding pockets (as illustrated in Fig. 6 and Fig. 7). We present three lines of experimental evidence that β binds primed DNA in the correct orientation when it is in solution. One line of evidence has already been presented; β binds primed DNA tighter than dsDNA in dilute solution (Fig. 1), implying that both the ssDNA and dsDNA interact with β. The second line of evidence requires a closer look at the ssDNA-β interaction. ssDNA binds in a region on the surface of β that is lined with basic residues (Fig. 4B). This region also contains two adjacent tyrosines (Y153 and Y154; see Fig. 4C) which stack with nucleotide bases (T13 and A15) that guide the ssDNA bases T11 and T12 into the protein binding pocket, placing them next to conserved residues Val247 (3.6 Å) and Met362 (3.09 Å), and T11 may form a H-bond to Thr172 (2.95Å). The two tyrosines are nearly perpendicular, a configuration often observed when two tyrosines each stack with two independent ssDNA bases (Burley and Petsko, 1985). Stacking of Tyr153 with base A15 of DNA flips out the terminal base of the duplex opposite the 3′ base pair; the adjacent T14 forms a base pair with the 3′ terminal nucleotide of the primer strand.

Fig. 6
Pol III C-terminal 9mer peptide competes with primed template ssDNA
Fig. 7
Possible functions of β-DNA contacts

Although interaction of ssDNA with β is a crystal contact, superposition of the symmetry related β-ssDNA protomer onto the structure shows very little gap between the two DNA molecules (Fig. S3). Therefore, even though the ssDNA-β interaction in the crystal is intermolecular, the ssDNA would appear capable of binding intramolecularly to the protein binding pocket in solution. If β-ssDNA interaction is functional, and not a crystal contact artifact, mutation of the surface tyrosines that stack with nucleotide bases should be defective in replication assays. Indeed, substitution of these tyrosines for serine significantly lowers the replication activity of β, and single substitutions result in intermediate effects (Fig. 5A). Analysis of these mutants demonstrate that the defect lies in clamp loading, not polymerase speed (Fig. 5B and C). The Kd for interaction of these β tyrosine mutants to the clamp loader are minimal, 1.4–2.5 fold (Fig. S2), and therefore interaction of β with ssDNA, as with dsDNA, likely plays a role in clamp loading at some point after initial substrate binding (see Discussion).

Fig. 5
The ssDNA Binding Tyrosines of β a27e Important to Function

DNA competes with Pol III for the protein binding pocket of β

Next we developed an assay to test whether ssDNA uses the protein binding pocket of β to interact with primed DNA. In this experimental design we used a 9-mer peptide derived from the C-terminus of the Pol III α-subunit which is known to bind the hydrophobic pocket of the β clamp (Lopez de Saro et al., 2003). The co-crystal structure of β in complex with this Pol III 9-mer peptide shows that it localizes to the hydrophobic protein binding pocket of β and does not interfere with DNA binding residues R24, Q149 or those in the central cavity (unpublished). This is also the case in the co-crystal structure of E. coli β with a C-terminal peptide derived from Pol IV (Burnouf et al., 2004).

If β utilizes the protein binding pocket to bind primed DNA, then the Pol III 9-mer peptide should compete with DNA binding to β. To examine this possibility β was labeled with Alexa Fluor555 (βAF555) and then was complexed with a primed DNA containing a 3′ end-labeled quencher (see Fig. 6). The AF555 emission spectrum overlaps the quencher absorption spectrum, and is quenched by the labeled DNA. The concentrations of components were chosen to exploit the difference in affinity of β for primed DNA and dsDNA, such that βAF555-DNAQuencher complex formation largely depends on the ssDNA interaction with β. If the ssDNA of the primed site binds to the protein binding pocket of β, the Pol III 9-mer peptide should displace the ssDNA from the protein binding pocket, causing the DNAQuencher to dissociate and restoring the fluorescence of βAF555. The result, in Fig. 6, shows an increase in βAF555 fluorescence as the Pol III 9-mer peptide is titrated into the assay, consistent with ssDNA binding to the protein binding pocket of β.


β directly interacts with DNA

The current study reveals that β binds DNA directly, independent of the ring shape of β. This is apparent from solution studies showing that β forms a saturable β-DNA complex with linear DNA at low concentration, but not with circular DNA (Fig. 1). If β were to bind DNA only due to the ring shape of the clamp, it would slide off linear DNA rather than building up a substantial amount of β-DNA complex at equilibrium, especially at low concentrations. The conclusion that β binds DNA independent of its ring shape is supported by the fact that a β monomer mutant, which is no longer a ring, binds DNA (both linear and circular DNA) with the same affinity as wt β. This finding encouraged us to screen β and DNA for a β-DNA co-crystal.

A Visual Screen for Co-crystals

We present here a simple method to screen oligonucleotide binding proteins for co-crystals containing the DNA (or RNA). The method employs a chromophoric tag on the DNA. Crystals containing bound DNA are deeply colored (e.g. Fig. 2A and Fig. S1B and S1C) and easily distinguished from crystals that lack DNA. In the present study the DNA was 5′ end-labeled with either Cy5 or TAMN. This method is especially useful for proteins that bind weakly to an oligonucleotide substrate, as most conditions may not yield co-crystals, but would require time consuming analysis to verify whether the ligand is present. This method also generalizes to other ligands besides DNA, as long as the ligands can be tagged without disrupting its ability to form a complex with the protein. For example, peptides tagged with a chromophore can be used to screen for conditions that yield protein-peptide co-crystals (e.g. Fig. S1A).

Arrangement of primed DNA in β

The co-crystal structure reveals that β interacts with DNA such that the duplex is highly angled (22°) as it passes through the ring. Although the crystal contact made by the ssDNA could underlie, or contribute to the angle of DNA, we note that recent molecular simulations predict that DNA can achieve a similar steep angle though the PCNA clamp (Ivanov et al., 2006). The β clamp also binds template ssDNA, which resides in a hydrophobic protein binding pocket on the C-terminal face of the clamp. The ssDNA interaction is a crystal contact, which probably helps crystal growth, but it orients primed DNA through β the reverse of normal. We demonstrate here that ssDNA binds the hydrophobic pocket of β in solution, and contributes to clamp loading activity, indicating that this contact is intramolecular and the primed DNA is correctly oriented through β in solution. The evidence that the ssDNA-β contact is physiologically relevant is as follows. First, primed DNA binds to β about 4-fold tighter than dsDNA, implying that ssDNA unique to the primed site interacts with β. Second, mutation of the tyrosines that bind ssDNA significantly reduce replication activity and clamp loading rate, indicating that β-ssDNA interaction is functional. And third, a Pol III peptide that binds the hydrophobic pocket of β competes with primed DNA for β, indicating that DNA binds the hydrophobic protein binding site of β. Since the biochemical studies are performed in dilute solution, we presume primed DNA transits through the β clamp in the physiologically relevant orientation, enabling ssDNA to bind the protein binding pocket (see illustrations in Fig. 6 and Fig. 7).

One may presume that during function with DNA polymerase, the polymerase will bind the protein binding site of β and displace ssDNA from β (see Fig. 7B). Since the polymerase-β clamp complex must move on DNA, we expect that the several direct connections of β to DNA we observe here (Fig. 2C) are transient and easily broken, enabling the clamp to slide on DNA during function with polymerase. In the dynamic situation of β-polymerase function, DNA likely adopts many different conformations as it passes through β, besides the view shown in the co-crystal structure. This hypothesis is supported by molecular simulations of β-DNA which indicate many different conformations of DNA within β (Daniel Barsky, Lawrence Livermore National Labs, personal communication). Interestingly, the simulations show R24 in continuous contact with DNA, and Q149 is within the top ten residues that interact with DNA.

β-DNA Interaction Plays a Role In Clamp Loading

Mutation of residues in β that bind ssDNA and/or dsDNA show defects in the clamp loading step; polymerase elongation speed is not affected. These β mutants bind the clamp loader with nearly the same affinity of wt β (Fig. S2) and therefore the role of β-DNA binding appears to lie in the clamp loading mechanism at a step downstream of initial clamp-clamp loader complex formation.

Where in the clamp loading mechanism may clamp residues that bind DNA be important to function? One possibility is illustrated in Fig. 7A. The initial clamp loading intermediate that builds up prior to ATP hydrolysis consists of an open clamp in the form of spiral that matches the spiral clamp loading AAA+ domains of the five clamp loader subunits (Fig. 7A, first diagram) (reviewed in (Bloom, 2006)). This open clamp-clamp loader-ATP complex binds primed DNA and positions it through the open clamp. We propose that once DNA is positioned inside the clamp, the DNA interactive residues of the clamp are attracted to DNA and induce the clamp to close around DNA (Fig. 7A, middle diagram). The clamp, in transiting from an open spiral to a closed planar structure (and tilted), no longer matches the spiral surface of the clamp loader, thereby breaking its connection to some of the clamp loading subunits. The strongest β-clamp loading subunit interaction is mediated by the δ subunit of the clamp loader, which binds the protein binding pocket of β (Jeruzalmi et al., 2001). Hence, template ssDNA binding to the protein binding pocket of β may help release δ from β and thereby complete clamp loader ejection, a necessary prerequisite for polymerase to bind β (Fig. 7A, right diagram) (Naktinis et. al., 1996).

The β-ssDNA Contact May Act as a "Placeholder"

After the clamp loader dissociates from β, the clamp may remain at the primed site through its connection to template ssDNA (Fig. 7B). In the absence of this connection, β could slide away from the 3′ terminus. Thus the ssDNA-β connection may serve a “placeholder” role, keeping β at the primed site where it is needed for function with DNA polymerase. Pol III contains two regions that bind to β (Dohrmann and McHenry, 2005; Lopez de Saro et al., 2003). Since polymerase competes with ssDNA for the protein binding pocket of β (i.e. Fig. 6), the ssDNA-β interaction will be broken upon association of polymerase with β and this may facilitate β diffusion along DNA. A placeholder function also has implications for how multiple factors may switch on the clamp, as described below.

The Angle of DNA Through β Has Implications for DNA Switching

Another possible role of β-DNA interaction observed here is implied by the highly tilted orientation of DNA as it passes through the ring. β is a homodimer and thus has two identical protein binding sites, through with it can bind two different proteins at once. This has been demonstrated directly for Pol III and Pol IV binding to β, and may generalize to other β interactive factors (Indiani et al., 2005; Sutton, 2004). Two DNA polymerases that bind one β clamp must vie for the single primed terminus.

The structure of the β-DNA complex shows that DNA is steeply angled through the ring, and this presents a possible route by which DNA may switch among two different proteins bound to the same clamp. The reasoning is as follows. DNA interactive residues R24 and Q149 are “off center” relative to the circular clamp, and are located on loops of the face of the clamp to which proteins bind. Thus these residues likely contribute to the tilted orientation of DNA through the center of the ring. Since β is a homodimer, the DNA should partition equally among two orientations in which it contacts R24/Q149 of one protomer or the other (e.g. see Fig. 7C). Tilted DNA that switches between two protomers within one clamp may facilitate switching of the DNA between two different DNA polymerases bound to the same β clamp, thus relocating from the active site of one DNA polymerase to the active site of the other. Interestingly, PCNA from the archaeon, Sulfolobus solfataricus, is a heterotrimer and each protomer binds a different protein (DNA polymerase, flap endonuclease and DNA ligase) (Dionne et al., 2003), and switching among the different factors may be facilitated by DNA movements inside the ring, as proposed above for E. coli β. DNA switching among different proteins attached to the same ring have also been proposed in molecular simulation studies of PCNA (Ivanov et al., 2006).

Many proteins bind to the β clamp. For example, β binds all five E. coli DNA polymerases, and also MutS, MutL, Hda and ligase. The damage inducible polymerases, Pols II, IV and V, are less accurate than Pol III and are thought to traverse certain lesions that would otherwise block replication fork progression (Goodman, 2002; Fujii and Fuchs, 2004). These polymerases are thought to utilize overlapping sites on β and switch among themselves when Pol III stalls, thereby allowing different polymerases to "sample" β in a process called “replication fork management” (Sutton et al., 2000; Sutton and Walker, 2001; Maul et al., 2007; Sutton and Duzen, 2006). The ability of β to bind template ssDNA may also facilitate switching of different proteins on β. For example, when the stalled Pol III dissociates from the primed site, β may bind ssDNA to remain at the ss/ds junction, and this may increase the association of an error-prone polymerase with β for advancing the fork past a template block.

Intracellular importance of DNA binding to β is provided by the dnaN159 mutant of β which shows pleotropic defects in DNA replication and repair (Sutton, 2004; Sutton and Duzen, 2006). The dnaN159 mutant of β carries two amino acid changes, G66E and G174A, and the crystal structure shows one of these residues, G174, binds DNA. Study of a G174A β mutant reveals altered relative affinity of Pol II, Pol IV and Pol V for β compared to Pol III (Maul et al., 2007; Sutton and Duzen, 2006). These studies are consistent with the view that different polymerases have only partly overlapping binding sites on the β clamp and suggest that DNA interaction near the polymerase binding site(s) on β facilitates exchange of different polymerases on β.

Implications for Eukaryotic PCNA

Does eukaryotic PCNA form contacts to DNA, like E. coli β? We have used chromophoric-tagged DNA structures in co-crystal trials with yeast PCNA and observe deeply colored crystals (Fig. S1B). The fact that PCNA-DNA co-crystals can be formed indicates that PCNA may bind DNA in a specific fashion, as observed here for the E. coli β-DNA complex. In support of this proposal are recent molecular simulations of PCNA that indicate DNA can achieve an angle as steep as 20° through PCNA (Ivanov et al., 2006). Furthermore, like β, PCNA contains hydrophobic protein binding pockets on the C-terminal face of the clamp. Perhaps PCNA binds ssDNA in these protein binding sites, as demonstrated here for E. coli β. Whether DNA forms direct contacts to PCNA, and if so, whether they are involved in clamp loading, must await future studies.

The fact that β binds both the dsDNA and ssDNA of a 3′ primed site junction suggests that β may help determine specificity of clamp loading at a 3′ ss/ds junction relative to loading at a 5′ ss/ds junction. In this regard it is interesting to note that the eukaryotic RFC clamp loader has alternative forms, one of which loads a different clamp. The different clamps, PCNA and 9-1-1, are loaded onto either the 3′ or 5′ terminus of a ss/ds junction, respectively (Ellison and Stillman, 2003; Yao et al., 2000). Considering that β binds both dsDNA and ssDNA of a primed site, it is tempting to speculate that some of the directionality in clamp loading may be inherent in the clamp. Future studies of these issues are sure to yield exciting insights into these important cellular processes.

Despite the many proposals of clamp-DNA interactions discussed here, further studies are needed to document the roles of the various β-DNA contacts that occur during its dynamic function. The current study demonstrates that direct clamp-DNA contacts are involved in the clamp loading step, possibly in helping the clamp to close around DNA. Whether direct clamp-to-DNA contacts also serve as a placeholder to keep the clamp the primed site to help mediate polymerase switching, or whether the tilt of DNA through β is maintained and used to aid switching among multiple clamp factors bound simultaneously to the same clamp are important topics for future studies.



Labeled nucleotides are from Dupont-NEN and unlabeled nucleotides are from Pharmacia-LKB. Streptavidin-coated Dynabead M-280 magnetic beads are from Dynal Biotech. Primed templates labeled with Cy-5 or TAMN were from Integrated DNA Technologies (Coralville, IA): Primers, 5′-/Cy5/CCCATCgTAT-3′, 5′-/Tamn56/CCCATCgTAT-3′; templates, 5′-[Tn]ATACgATggg-3′. To yield 10/10, 10/12, 10/14, 10/16, 10/18 and 10/20-mer primed templates, oligonucleotides were annealed using equimolar ratios using 10 mM Tris-HCl, 1mM EDTA, pH 7.8. Magnetic bead clamp loading assays used a 30-mer (5′-gCAATAACTggCCgTTTgAAgATTTCg-3′), and 3′ biotinylated 102-mer (5′-CCATTCTgTAACgCCAgggTTTTCgCAgTCAACATTCgAAATCTTCAAACgACggCC AgTTATTgCTCTTCTTgAgTTTgATAgCCAAAACgACCATTATAg/TEGBiotin/-3′) annealed 2:1. M13mp18 ssDNA was primed with a DNA 60-mer. β and SSB were purified as described (Kong et al., 1992). Pol III*, Pol III core, and γ complex were reconstituted as described (McInerney and O'Donnell, 2004).

β mutants

The dnaN gene encoding β was cloned into a modified pET16c vector which incorporates a short N-terminal hexahistidine and protein kinase tag (Stukenberg et. al., 1994). Single and double β mutants (βQ149A, βR24A, βR24A/Q149A, βY153A, βY154A, βY153A/Y154A) were constructed by the QuickChange XL protocol (Stratagene Inc.). Plasmids were introduced into E. coli BL21[DE3] and grown in LB to OD 0.6 at 37 °C, then lowered to 20°C and induced upon adding 0.1 mM IPTG followed by incubation at 20°C for 10 h. The β monomer mutant (βI272AL273A) was as described (Jeruzalmi et al., 2001).

Data acquisition and structure refinement

Conditions for co-crystal formation were 0.2 mM β dimer and 0.22 mM DNA in 10 mM Tris-Cl (pH 7.4), 50 mM NaCl, 0.5 mM EDTA and 10% glycerol. Protein-DNA solutions were mixed with an equal volume of 22 – 26 % PEG 400, 75 mM MES (pH 5.9), 75 mM CaCl2, 5% glycerol, 0.5 % DMSO and allowed to equilibrate by vapor diffusion. Crystals grew to approximate dimensions (0.2 × 0.3 × 0.3 mm3) in two weeks at room temperature.

Diffraction data were collected at 100K at the X4a beamline at the National Synchrotron Light Source, Brookhaven National Laboratory. Data were indexed and scaled with the HKL2000 program suite (Otwinowski, 1997). A data set of 1.61 Å resolution was collected for the uncomplexed β and the structure was solved by molecular replacement using the known β structure (Kong et al., 1992), and refined to 1.7 Å using CNS (Brunger et al., 1998) (see Table SI).

Crystals of β-DNA complex diffracted to 1.92 Å resolution and the structure was solved by molecular replacement using the newly refined native structure as the search model. Refinement was initiated by placing a 10bp duplex DNA with ideal B-type geometry built using ICM-Pro (Molsoft L.L.C.). The β-DNA structure was refined without placing any restraints on the DNA duplex. Changes in the β-DNA complex structure were made using the program O (Jones et al., 1991). All refinements were carried out against |F0| > 0σ data, using the CNS program suite (Brunger et al., 1998). The final model spans all protein residues (2 × 366 residues) and the entire primed-DNA template.

M13mp18 ssDNA Extension Assays

(i) β-clamp dependence

β (0, 15, 31, 62, 124, 248, 494, 988 and 1975 fmol) was incubated with SSB (425 pmol), and primed M13mp18 ssDNA (30 fmol) for 5 min at 37°C in 25 µl of Replication Buffer (25mM TrisCl (pH 7.5), 5mM DTT, 40 µg/ml BSA, 4% glycerol, 8 mM MgCl2, and 0.5 mM ATP) containing 60 µM each dGTP, dCTP, dATP and 20 µM [α-32P]dTTP. DNA synthesis was initiated upon adding Pol III* (100 fmol). After 2 min, reactions were quenched with 25 µl of 1% SDS/40 mM EDTA. One-half of the reaction was spotted on DE81 filters to quantitate DNA synthesis, while the other half was analyzed in a 0.8% neutral agarose gel.

(ii) Pol III-β rate of synthesis measurements

β was assembled onto primed M13mp18 ssDNA coated with SSB in 25 µl of Replication Buffer containing β (450 fmol), SSB (425 pmol), γ complex (20 fmol), primed M13mp18 ssDNA (30 fmol), 0.5 mM ATP and 8 mM MgCl2. Pol III core (850 fmol), and 60 µM each dGTP and dCTP were added, followed by incubation for 10 min at 37°C. DNA synthesis was initiated upon adding 60 µM dATP and 20 µM [α32P] dTTP. After 2, 4, 6, 8, 15, 30 or 60s, reactions were quenched and analyzed as above.

Clamp Loading Assay

The 30/102-mer biotinylated DNA was conjugated to M-280 Streptavidin Dynabeads in 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 150 mM NaCl and 5 mg/ml BSA at 22°C for 30 min and washed 3 times with the same buffer. The yield was 185 ± 35 pmol of DNA/mg of Dynabeads. Clamp loading was performed at 20°C in clamp loading buffer (30 mM Tris-HCl pH 7.5, 7mM Mg(OAc)2, 100 mM NaCl, 1mM DTT and 1mM CHAPS) containing 1mM ATP, 280 nM DNA, 600 nM E.coli SSB (tetramer), 140 nM (dimer) 32P-βHK (or 32P-βHK mutants). Reactions were incubated 1 min, then initiated by adding 70 nM γ complex and quenched with 21 mM EDTA (pH 7.5) at the indicated times. Beads were harvested in a magnetic concentrator and washed twice in clamp loading buffer containing 200 mM NaCl. Protein was stripped from beads using 1% SDS at 95 °C for 5 min. 32P-βHK was quantified by liquid scintillation, and visually verified in a 12% SDS-PAGE.

Kd measurements of β-γ complex interaction

Wt β and β mutants were labeled using pyrene maleimide (Molecular Probes, Inc., Eugene, OR) to form βPy as described (Snyder et. al., 2004). Reactions contained βPy (50 nM or 100 nM) in 60 µl 20mM Tris-Cl (pH 7.5), 8 mM MgCl2, 1 mM DTT, 0.2 mM EDTA, 50 mM NaCl and 0.5 mM ATP or ATP-γS. βPy anisotropy is enhanced upon γ complex binding. Titrations and Kd measurements were performed as described (Anderson et al., 2007).

Kd measurements of β-DNA binding

Wt β and β mutants were labeled using OregonGreen488 maleimide (Molecular Probes, Inc., Eugene, OR) to form βOG as described (Lopez de Saro et al., 2003). The following oligonucleotides were used to construct 18/18mer blunt duplex and 18/28mer 3′ primed template: 5′- CCC ATC gTA TAg CAA ggg -3′ (18mer-primer), 5′- CCC TTg CTA TAC gAT ggg -3′ (18mer-template), 5′- TTT TTT TTT TCC CTT gCT ATA CgA Tgg g -3′(28mer-template). DNA titrations contained βOG (200 nM or 500 nM) in 60 µl of 20mM Tris-Cl (pH 7.5), 1 mM DTT, 0.2 mM EDTA, and 40 mM NaCl. Reactions were equilibrated at 22°C for 10 min and then transferred into a 3×3-mm cuvette. Fluorescence emission spectra were recorded from 500 to 630 nm (excitation at 490 nm); emission at 517 nm was used for analysis.

Peptide competition experiments

β was modified with a AF555 fluorophore (Molecular Probes) to yield βAF555 as described above for βOG. βAF555 (150 nM) was mixed with an 18/34mer primed DNA labeled at the 3′ terminus of the 18mer with a Black Hole quencher (IDT) in 60 µl of 20 mM Tris-HCl (pH 7.5), 1 mM DTT, 0.2 mM EDTA, and 40 mM NaCl. A 9-mer peptide corresponding to the C-terminus of Pol III α-subunit was titrated into the reaction. A control experiment between the free AF555 fluorophore and the Black-Hole labeled primed template, showed no quenching.

Supplementary Material



We are grateful to the staff at Brookhaven National Laboratory, for their help,Jeff Finkelstein (Rockefeller U., NYC) for technical assistance, Daniel Barsky (LLNL, Lawrence, CA) and Lance Langston (Rockefeller U., NYC) for critical reading of the manuscript. This work was supported by grants from the NIH, GM38839 (M.O.D), GM70841 (X-P.K.) and GM45547 (J.K.).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

RCSB Protein Data Bank Accession number PDB code: 3BEP


  • Anderson SG, Williams CR, O'Donnell M, Bloom LB. A function for the chi subunit in loading the Escherichia coli DNA polymerase sliding clamp. J Biol Chem. 2007;282:7035–7045. [PubMed]
  • Argiriadi MA, Goedken ER, Bruck I, O'Donnell M, Kuriyan J. Crystal structure of a DNA polymerase sliding clamp from a Gram-positive bacterium. BMC Struct Biol. 2006;6:2. [PMC free article] [PubMed]
  • Bloom LB. Dynamics of loading the Escherichia coli DNA polymerase processivity clamp. Crit Rev Biochem Mol Biol. 2006;41:179–208. [PubMed]
  • Brunger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS, et al. Crystallography & NMR system: A new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr. 1998;54:905–921. [PubMed]
  • Bunting KA, Roe SM, Pearl LH. Structural basis for recruitment of translesion DNA polymerase Pol IV/DinB to the beta-clamp. Embo J. 2003;22:5883–5892. [PMC free article] [PubMed]
  • Burley SK, Petsko GA. Aromatic-aromatic interaction: a mechanism of protein structure stabilization. Science. 1985;229:23–28. [PubMed]
  • Burnouf DY, Olieric V, Wagner J, Fujii S, Reinbolt J, Fuchs RP, Dumas P. Structural and biochemical analysis of sliding clamp/ligand interactions suggest a competition between replicative and translesion DNA polymerases. J Mol Biol. 2004;335:1187–1197. [PubMed]
  • Chapados BR, Hosfield DJ, Han S, Qiu J, Yelent B, Shen B, Tainer JA. Structural basis for FEN-1 substrate specificity and PCNA-mediated activation in DNA replication and repair. Cell. 2004;116:39–50. [PubMed]
  • Dalrymple BP, Kongsuwan K, Wijffels G, Dixon NE, Jennings PA. A universal protein-protein interaction motif in the eubacterial DNA replication and repair systems. Proc Natl Acad Sci U S A. 2001;98:11627–11632. [PMC free article] [PubMed]
  • Dionne I, Nookala RK, Jackson SP, Doherty AJ, Bell SD. A heterotrimeric PCNA in the hyperthermophilic archaeon Sulfolobus solfataricus. Mol Cell. 2003;11:275–282. [PubMed]
  • Dohrmann PR, McHenry CS. A bipartite polymerase-processivity factor interaction: only the internal beta binding site of the alpha subunit is required for processive replication by the DNA polymerase III holoenzyme. J Mol Biol. 2005;350:228–239. [PubMed]
  • Ellison V, Stillman B. Biochemical characterization of DNA damage checkpoint complexes: clamp loader and clamp complexes with specificity for 5′ recessed DNA. PLoS Biol. 2003;1:E33. [PMC free article] [PubMed]
  • Fujii S, Fuchs RP. Defining the position of the switches between replicative and bypass DNA polymerases. Embo J. 2004;23:4342–4352. [PMC free article] [PubMed]
  • Goodman MF. Error-prone repair DNA polymerases in prokaryotes and eukaryotes. Annu. Rev. Biochem. 2002;71:17–50. [PubMed]
  • Gulbis JM, Kelman Z, Hurwitz J, O'Donnell M, Kuriyan J. Structure of the C-terminal region of p21(WAF1/CIP1) complexed with human PCNA. Cell. 1996;87:297–306. [PubMed]
  • Indiani C, McInerney P, Georgescu R, Goodman MF, O'Donnell M. A sliding-clamp toolbelt binds high- and low-fidelity DNA polymerases simultaneously. Mol Cell. 2005;19:805–815. [PubMed]
  • Ivanov I, Chapados BR, McCammon JA, Tainer JA. Proliferating cell nuclear antigen loaded onto double-stranded DNA: dynamics, minor groove interactions and functional implications. Nucleic Acids Res. 2006;34:6023–6033. [PMC free article] [PubMed]
  • Jeruzalmi D, Yurieva O, Zhao Y, Young M, Stewart J, Hingorani M, O'Donnell M, Kuriyan J. Mechanism of processivity clamp opening by the delta subunit wrench of the clamp loader complex of E. coli DNA polymerase III. Cell. 2001;106:417–428. [PubMed]
  • Johnson A, O'Donnell M. Cellular DNA replicases: components and dynamics at the replication fork. Annu Rev Biochem. 2005;74:283–315. [PubMed]
  • Jones TA, Zou JY, Cowan SW, Kjeldgaard M. Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr A. 1991;47(Pt 2):110–119. [PubMed]
  • Kong XP, Onrust R, O'Donnell M, Kuriyan J. Three-dimensional structure of the beta subunit of E. coli DNA polymerase III holoenzyme: a sliding DNA clamp. Cell. 1992;69:425–437. [PubMed]
  • Kornberg A, Baker T. DNA Replication. second edn. New York: W.H. Freeman and Company; 1992.
  • Krishna TS, Kong XP, Gary S, Burgers PM, Kuriyan J. Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell. 1994;79:1233–1243. [PubMed]
  • Lopez de Saro FJ, Georgescu RE, Goodman MF, O'Donnell M. Competitive processivity-clamp usage by DNA polymerases during DNA replication and repair. Embo J. 2003;22:6408–6418. [PMC free article] [PubMed]
  • Lu XJ, Olson WK. 3DNA: a software package for the analysis, rebuilding and visualization of three-dimensional nucleic acid structures. Nucleic Acids Res. 2003;31:5108–5121. [PMC free article] [PubMed]
  • Matsumiya S, Ishino Y, Morikawa K. Crystal structure of an archaeal DNA sliding clamp: proliferating cell nuclear antigen from Pyrococcus furiosus. Protein Sci. 2001;10:17–23. [PMC free article] [PubMed]
  • Maul RW, Scouten Ponticelli SK, Duzen JM, Sutton MD. Differential binding of Escherichia coli DNA polymerases to the beta-sliding clamp. Mol Microbiol. 2007;65:811–827. [PubMed]
  • McHenry CS. Chromosomal replicases as asymmetric dimers; studies of subunit arrangement and functional consequences. Mol Microbiol. 2003;49:1157–1165. [PubMed]
  • McInerney P, O'Donnell M. Functional uncoupling of twin polymerases: mechanism of polymerase dissociation from a lagging-strand block. J Biol Chem. 2004;279:21543–21551. [PubMed]
  • Naktinis V, Turner J, O'Donnell M. A molecular switch in a replication machine defined by an internal competition for protein rings. Cell. 1996;84:137–145. [PubMed]
  • Otwinowski Z, Minor W. Processing of the X-ray diffraction data collected in oscillation mode. Methods Enzymology. 1997;276:307–326.
  • Pritchard AE, Dallmann HG, Glover BP, McHenry CS. A novel assembly mechanism for the DNA polymerase III holoenzyme DnaX complex: association of deltadelta' with DnaX(4) forms DnaX(3)deltadelta'. Embo J. 2000;19:6536–6545. [PMC free article] [PubMed]
  • Shamoo Y, Steitz TA. Building a replisome from interacting pieces: sliding clamp complexed to a peptide from DNA polymerase and a polymerase editing complex. Cell. 1999;99:155–166. [PubMed]
  • Snyder AK, Williams CR, Johnson A, O'Donnell M, Bloom LB. Mechanism of loading the E. coli DNA polymerase III sliding clamp: II. Uncoupling the beta and DNA binding activities of the gamma complex. J. Biol. Chem. 2004;279:4386–4393. [PubMed]
  • Stukenberg PT, Turner J, O'Donnell M. An explanation for lagging strand replication: polymerase hopping among DNA sliding clamps. Cell. 1994;78:877–887. [PubMed]
  • Sutton MD. The Escherichia coli dnaN159 mutant displays altered DNA polymerase Usage and chronic SOS induction. J Bacteriol. 2004;186:6738–6748. [PMC free article] [PubMed]
  • Sutton MD, Duzen JM. Specific amino acid residues in the beta sliding clamp establish a DNA polymerase usage hierarchy in Escherichia coli. DNA Repair (Amst) 2006;5:312–323. [PubMed]
  • Sutton MD, Duzen JM, Maul RW. Mutant forms of the Escherichia coli beta sliding clamp that distinguish between its roles in replication and DNA polymerase V-dependent translesion DNA synthesis. Mol Microbiol. 2005;55:1751–1766. [PubMed]
  • Sutton MD, Smith BT, Godoy VG, Walker GC. The SOS response: recent insights into umuDC-dependent mutagenesis and DNA damage tolerance. Annu Rev Genet. 2000;34:479–497. [PubMed]
  • Sutton MD, Walker GC. Managing DNA polymerases: coordinating DNA replication, DNA repair, and DNA recombination. Proc Natl Acad Sci U S A. 2001;98:8342–8349. [PMC free article] [PubMed]
  • Warbrick E. The puzzle of PCNA's many partners. Bioessays. 2000;22:997–1006. [PubMed]
  • Yao N, Hurwitz J, O'Donnell M. Dynamics of beta and proliferating cell nuclear antigen sliding clamps in traversing DNA secondary structure. J Biol Chem. 2000;275:1421–1432. [PubMed]


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...