Article

• The EMBO Journal (1999) 18, 771 - 783
• doi:10.1093/emboj/18.3.771

The internal workings of a DNA polymerase clamp-loading machine

Jennifer Turner¹, Manju M. Hingorani², Zvi Kelman^1,3 and Mike O'Donnell^2,4

1. Cornell University Medical College, 1300 York Avenue, New York, NY 10021, USA
2. The Rockefeller University, New York, NY 10021, USA
3. Sloan-Kettering Institute, 1275 York Avenue, New York, NY 10021, USA
4. The Howard Hughes Medical Institute, 1230 York Avenue, New York, NY 10021, USA

Correspondence to:

Mike O'Donnell, E-mail: odonnel@rockvax.rockefeller.edu

Received 11 September 1998; Accepted 27 November 1998; Revised 27 November 1998

• Keywords:

  • ATPase,
  • clamp loader,
  • DNA polymerase,
  • processivity,
  • sliding clamp

Introduction

Replicases are efficient DNA-synthesizing enzymes that duplicate long chromosomes with high speed and processivity. These biological machines are comprised of three functional components in both prokaryotic and eukaryotic organisms: (i) a DNA polymerase, (ii) a processivity factor or sliding clamp protein and (iii) a multiprotein clamp loader. Replicases from Escherichia coli, Saccharomyces cerevisiae, humans and bacteriophage T4 have these three components. The DNA polymerases lack high processivity alone but upon association with their respective sliding clamps, they can replicate several thousand bases continuously (reviewed in Kelman and O'Donnell, 1995). The sliding clamps are ring-shaped homo-oligomers that encircle duplex DNA (Stukenberg et al., 1991; Kong et al., 1992; Burgers and Yoder, 1993; Krishna et al., 1994; Gulbis et al., 1996; Yao et al., 1996). These protein rings are assembled onto DNA by their respective clamp loaders in an ATP-dependent reaction (Kelman and O'Donnell, 1995). Unlike sequence-specific DNA-binding proteins, the circular clamps form a topological link with DNA and slide along the duplex without localizing to a specific region. Therefore, when the DNA polymerase binds its clamp, it is endowed with high processivity, allowing it to move continuously along the template during chain extension.

The E.coli replicase, DNA polymerase III holoenzyme, consists of 10 different polypeptide chains (Kornberg and Baker, 1992; Kelman and O'Donnell, 1995). Within the holoenzyme are two core polymerases () and a clamp loader ( complex), that are held together by , a connector protein (Onrust et al., 1995a). Both core polymerases become highly processive when tethered to DNA by sliding clamps (Stukenberg et al., 1991). As the holoenzyme moves along the replication fork, the processive polymerase extends DNA continuously on the leading strand. On the lagging strand, the other polymerase releases its sliding clamp upon completion of each Okazaki fragment, and is cycled back to the replication fork by targeting to a new clamp assembled at an upstream RNA primer by the clamp loader (O'Donnell, 1987; Stukenberg et al., 1994).

The E.coli clamp loader, complex, consists of five different subunits: , , ', and (Maki and Kornberg, 1988). Within the complex are two to four subunits and one each of , ', and (Maki and Kornberg, 1988; Onrust et al., 1995b). The proteins are assembled such that both ' and bind directly to , while and bind the ' and subunits, respectively (Onrust et al., 1995b). Previous studies have shown that a ' complex is sufficient to load onto DNA (Onrust et al., 1991). The and subunits are not essential for loading onto DNA; binds single-stranded DNA binding protein (SSB) and facilitates displacement of the primase from RNA primers, an event that must occur prior to clamp loading on the lagging strand (Kelman et al., 1998; Yuzhakov et al., 1999). The subunit is the major contact point between complex and , and binds tightly to in the complete absence of the other clamp loader proteins (Naktinis et al., 1995). When is part of the complex, however, it exhibits only a low affinity for in the absence of ATP. In the presence of ATP, the complex undergoes a conformational change that now allows to bind with an affinity comparable with that of the free – interaction (Naktinis et al., 1995). is the only clamp loader subunit that binds and hydrolyzes ATP (Maki and Kornberg, 1988; Tsuchihashi and Kornberg, 1989; Onrust et al., 1991). Recent studies show that ATP binding powers a change in subunit conformation (Hingorani and O'Donnell, 1998), which may underlie the ATP-induced change in complex conformation that leads to tight interaction with (Naktinis et al., 1995; Hingorani and O'Donnell, 1998). Thus, the subunit transduces the energy from ATP to expose the subunit for interaction with the ring.

This report describes the mechanics of the complex-catalyzed process of assembly onto DNA. We show that the energy of ATP binding powers the complex machinery to bind and open the circular clamp at one interface. Once complex binds , its ATPase activity is supressed until it binds DNA. The correct DNA substrate stimulates ATP hydrolysis, which is coupled to the release of complex from on DNA. On DNA, reassumes its lowest free energy state and forms a closed ring, now with DNA passing through the center.

The molecular details that underlie the clamp loader activity have been explored and as a result, specific functions have been assigned to the three integral components of the complex: is the 'motor', is the 'ring opener' and ' is the 'modulator'. The subunits are the only components of complex to interact with ATP, and their function is to drive ATP-induced conformational changes of other subunits in the clamp loader. The subunit contains the intrinsic clamp-opening activity, as it can open the ring in the complete absence of ATP and the other subunits. The ' subunit binds and probably modulates the ability of to gain access to . The ATP-driven motor either moves or alters the ' modulator such that the clamp opener gains access to the clamp and opens the ring. This action requires only ATP binding to the complex. The resulting clamp loader–open clamp composite then binds primed DNA with high affinity. Upon interaction with primed DNA, two or three molecules of ATP are hydrolyzed, resulting in closure of the ring around DNA and ejection of complex off the DNA and back into solution.

The complex uses energy from ATP binding to open the ring

The sliding clamp is composed of two crescent-shaped protomers arranged in a tight head-to-tail dimer (Kong et al., 1992). This circular protein clamp is transferred onto circular DNA by the ATP-dependent clamp-loading activity of the complex. How does complex assemble the ring around DNA? How is ATP used in this reaction? Is ATP binding sufficient, or is hydrolysis necessary? What is the role of DNA in this process? To begin addressing these questions, we developed a novel 'Cys-labeling' assay to detect and characterize the ring-opening step in clamp assembly.

Leu273, a buried hydrophobic residue in the dimer interface, was substituted with cysteine ( crystal structure; Kong et al., 1992). This buried cysteine residue should not be accessible to a thiol-reactive reagent, such as a maleimide, when the L273C- ring is closed. When the interface is opened, however, the buried cysteine should become accessible to the surface and reactive to maleimide. To follow this reaction, we synthesized a radioactive reagent by linking the N-terminus of a ³²P-labeled peptide with an N-hydroxysuccinimide (NHS) ester coupled to maleimide (Figure 1). If the dimer interface opens, the interface cysteine residue will react with the [³²P]maleimide reagent, resulting in a radiolabeled clamp. One other modification was made in L273C- to ensure that only the interface cysteine residue would be labeled; the reactive Cys333 at the protein surface was changed to serine. The modified is a dimer and is fully active in DNA replication assays with complex and core polymerase (data not shown).

Figure 1.

complex opens the clamp on binding ATP. The scheme indicates the position of a cysteine residue (S) within the dimer interface that is accessible for reaction only when the ring is opened. The structure of the [³²P]maleimide reagent used to label cysteine is also shown. (A) (3 M) incubated with [³²P]maleimide in the absence or presence of complex (3 M), primed oligoDNA (3 M) or ATP (2 mM), and analyzed by SDS–PAGE. ³²P-labeled is indicated by an arrow, showing a 2- and 10-fold increase in labeling in the presence of complex and complex + ATP, respectively. (B) complex-catalyzed ring opening assayed for varying times in the presence of ATP (left) or ATP--S (right). (C) complex-catalyzed ATP hydrolysis compared with ATP--S hydrolysis. complex (0.1 M), incubated with (0.2 M), hydrolyzes ATP with a k_cat = 1.7/min () and hydrolyzes ATP--S with a k_cat = 410^-3/min ().

View full figure (68 KB)

The results of the ring-opening assay are shown in Figure 1A. The [³²P]maleimide reagent was incubated with in the absence or presence of complex, ATP and DNA, and then the reactions were analyzed on an SDS–polyacrylamide gel. The autoradiogram of the gel shows very low reactivity of alone (Figure 1A, lane 1). When is mixed with complex, there is a 2-fold increase in labeling (Figure 1A, lane 2). However, when both complex and ATP are present in the reaction, the is 10-fold more reactive with the [³²P]maleimide (Figure 1A, lane 4). The results indicate that in the presence of ATP, complex can open the ring or at least change the ring conformation enough to expose the cysteine residue buried in the dimer interface. When DNA is present in the reaction, is labeled to the same extent as in the absence of DNA (compare lanes 4 and 5), indicating that complex can open the ring even in the absence of DNA.

Next, the assay was used to examine whether ATP binding or hydrolysis is required to stimulate ring opening. To separate the effects of ATP binding from ATP hydrolysis, we substituted ATP with its analog, ATP--S (Figure 1B). The turnover number for ATP hydrolysis catalyzed by complex (in the presence of ) is 1.5–2/min at 37°C (Figure 1C). Therefore, in a 10 min assay, each complex hydrolyzes 15–20 molecules of ATP. The complex binds ATP--S with similar high affinity as it does ATP (K_d^ATP =2 M and K_d^ATP--S = 5 M, respectively; Hingorani and O'Donnell, 1998), but the turnover number for ATP--S hydrolysis is very low (4–610^-3/min; see Figure 1C). Since complex catalyzes hydrolysis of only one ATP--S every 2–3 h, ATP--S can be considered a non-hydrolyzable analog of ATP within the time frame of the ring-opening assay.

The data shown in Figure 1B demonstrate that ATP and ATP--S facilitate complex-catalyzed ring opening to the same extent. Therefore, ATP binding appears sufficient to drive complex-catalyzed opening of the ring. Other non-hydrolyzable ATP analogs such as AMP-PNP and AMP-PCP were also tested in the assay, but ring opening was not detected (data not shown). However, this negative result may be explained by the fact that complex binds these analogs with much lower affinity than ATP (K_d 1 mM versus 2 M for ATP), and this binding energy may be insufficient to support ring opening (H.Xiao and M.O'Donnell, unpublished data).

Next, the ring-opening step was examined in greater detail to determine whether the clamp loader opens at both interfaces (e.g. monomerizes the dimer to assemble it around DNA), or if opening the dimer at only one interface is sufficient to allow entry of DNA into the ring.

The complex need open only one interface of the dimer for clamp loading

To determine whether complex needs to open one or both interfaces of the dimer for assembly of the clamp around DNA, we constructed a dimer in which the monomers could be cross-linked at the dimer interface. Arg103 and Ile305, two residues close to each other but on opposing sides of the dimer interface (C atoms 6 Å apart; cyrstal structure; Kong et al., 1992), were substituted with cysteine. Due to the head-to-tail arrangement of the dimer, the two interfaces are identical. It was expected that oxidation would result in formation of disulfide cross-links between the cysteine residues, and yield dimers covalently linked at one or both interfaces.

Initially, the modified (^S-S) was ³²P-labeled at an N-terminal kinase recognition site and analyzed on a non-reducing SDS–polyacrylamide gel to determine if disulfide cross-links had formed. Figure 2A shows that in the absence of dithiothreitol (DTT), more than half of the ^S-S migrates as an 80 kDa band of cross-linked dimers; the rest migrates as a 40 kDa band of monomers (lane 1). Native migrates as a single band at 40 kDa (lane 2). Incubation of ^S-S with DTT reduces the disulfide cross-links such that at 2 and 20 mM DTT essentially all ^S-S migrates as monomers on the SDS–polyacrylamide gel (lanes 3 and 4, respectively).

Figure 2.

complex opens the clamp at one interface. Cysteine residues were placed at the dimer interface to design a clamp with disulfide cross-links at the interface (^S-S). (A) Native and ^S-S analyzed on a non-reducing SDS–polyacrylamide gel with no DTT, 2 mM DTT and 20 mM DTT in the reaction. (B) Gel filtration analysis of native [³²P] () and [³²P]^S-S clamps assembled onto DNA by complex in the absence of DTT () and in the presence of DTT (). The next scheme shows a mix of ^S-S rings with two, one or zero cross-links at the interface, assembled on DNA and analyzed by gel filtration followed by non-reducing SDS–PAGE. (C) Analysis of ^S-S from a clamp-loading reaction without DTT [as in (B)]. (D) The same reaction analyzed with SDS present during gel filtration. Fraction numbers are indicated beneath the gels. The positions of non-cross-linked ^S-S (monomers) and cross-linked ^S-S (dimers) are indicated by arrows.

View full figure (90 KB)

Next we analyzed the cross-linked ^S-S for assembly around a circular DNA substrate. The ³²P-labeled ^S-S was incubated with complex and DNA in the presence of ATP, and then assayed by gel filtration over a Bio-Gel A-15m column (this large pore resin includes proteins but excludes the large DNA substrate as well as any protein bound to it). Figure 2B shows that about half of the ^S-S is assembled on DNA in the absence of DTT ( on DNA, fractions 9–13; free , fractions 18–27). Treatment of ^S-S with DTT prior to the clamp loading reaction results in assembly of nearly all the ^S-S on DNA (fractions 9–13), and is comparable with assembly of native on DNA (Figure 2B). Presumably in the absence of DTT the disulfide cross-links at the interface prevent some (or all) of the ^S-S dimers from being assembled on DNA.

To determine if the ^S-S clamps assembled onto DNA in the absence of DTT contain any cross-links, the column fractions were analyzed on a non-reducing SDS–polyacrylamide gel (Figure 2C). As expected, non-cross-linked ^S-S was assembled onto DNA (monomers in fractions 9–13). However, a large fraction of the cross-linked ^S-S was also assembled onto DNA (dimers in fractions 9–13), while the remainder eluted as free protein (dimers in fractions 18–27). Presumably the ^S-S dimers on DNA contain only one cross-link, and therefore can still be opened at one interface. The ^S-S dimers that were not assembled on DNA may contain two disulfide cross-links that preclude ring opening and assembly onto DNA.

To confirm that cross-linked ^S-S on DNA (dimers in fractions 9–13; Figure 2C) is cross-linked at only one interface, the experiment in the absence of DTT was repeated and split into two halves. One half was analyzed by gel filtration as described above (Figure 2C), to confirm that ^S-S was assembled on DNA. The other half was analyzed by gel filtration in the presence of SDS (Figure 2D). If the ^S-S on DNA contains one disulfide cross-link, then SDS denaturation will result in a dimer opened at the non-cross-linked interface that will dissociate from DNA during gel filtration. If the ^S-S on DNA were to contain two disulfide cross-links, the protein ring will remain covalently sealed even after SDS denaturation, and thus remain topologically linked to DNA in the presence of SDS. The results show that the cross-linked ^S-S dissociates from DNA during gel filtration in the presence of SDS, indicating that the rings were cross-linked at only one interface (dimers in fractions 18–27; Figure 2D). Thus, the complex can load clamps that are sealed at one interface, and does not need to monomerize the dimer to load it onto DNA.

is the ring opener

We performed the Cys-labeling assay described in Figure 1 with individual subunits and subassemblies of the complex, to determine the minimal subunit requirement for ring opening. The minimal subassembly that tested positive for ring opening was the three-subunit ' complex, which, like complex, needs ATP or ATP--S to crack open the dimer interface (data not shown). Individual subunits or combinations of two subunits did not give a detectable signal. This result implies that more than two proteins are needed to detect ring opening in the Cys-labeling assay, that ' is sufficient for ring opening and that the and subunits are not absolutely required for the mechanics of this process. This conclusion is consistent with earlier data showing that efficient stimulation of processive DNA polymerase activity requires the simultaneous presence of all three subunits, , and ' (Onrust et al., 1991). However, the Cys-labeling assay for ring opening does not rigorously exclude the possibility that a smaller subassembly than ', or even just one subunit, can crack open the dimer interface transiently. For example, the assay requires time for the [³²P]maleimide to react with an exposed cysteine residue, and is likely to proceed more efficiently the longer the dimer interface is held open. To investigate if any one subunit can open the ring at least transiently, we designed a more direct assay to detect opening of the dimer interface. This assay is based on the following rationale: once is assembled onto DNA, the –DNA complex is very stable (t_1/2 >1 h at 37°C; Yao et al., 1996). However, if the dimer interface were to open even for a short time, the ring could slip free of DNA. The experiment described below examines the ring-opening activity of the , and ' subunits by observing the unloading of circular clamps from a circular DNA substrate.

In this experiment, ³²P-labeled clamps were assembled onto circular DNA by complex, and the [³²P] on DNA was separated from free complex and free [³²P] by gel filtration. The [³²P]–DNA complex was incubated further with either , or ', in the absence or presence of ATP. Following this, the reactions were analyzed on an agarose gel to separate [³²P] trapped on DNA from the [³²P] released from DNA. The results in Figure 3A, show that the subunit by itself can release [³²P] from DNA, and this process does not require ATP. The lack of ATP dependence is consistent with earlier evidence that binds in the absence of ATP (Naktinis et al., 1995; note neither nor binds ATP). It also verifies that the [³²P] unloading (mediated by ) observed in this assay is not a result of a complex contaminant in the protein preparation, since complex requires ATP to release from DNA (Naktinis et al., 1996). The gel was also stained with ethidium bromide (Figure 3B), which showed that the circular DNA remained intact during incubation with each protein, including . The data confirm that release from DNA is a result of ring opening and not due to the ring sliding off linearized DNA, as the circular DNA substrate is not degraded during the assay.

Figure 3.

A ring-unloading assay reveals that opens . [³²P] clamps were assembled onto circular DNA and incubated with , or ' in the absence or presence of ATP and analyzed by agarose gel electrophoresis (scheme). (A) An autoradiogram of the agarose gel shows the [³²P] on DNA separated from the [³²P] unloaded off DNA. (B) The gel stained with ethidium bromide shows that DNA is not degraded during the reaction.

View full figure (65 KB)

These results show that clamp opening is actually performed by only one subunit of the complex, and is the result of a simple protein–protein interaction. It should be noted that even though opens enough to allow it to slip free from DNA, it cannot load onto DNA. Thus, the and ' subunits play a critical role in assembly of the open clamp around DNA. The following experiments examine how the ' and subunits function in clamp loading.

' and compete for access to

Previous studies identified as the only subunit of complex with a detectable interaction with (Naktinis et al., 1995). When is part of the complex, the interaction with is not favorable in the absence of ATP (Naktinis et al., 1995). In the presence of ATP, however, complex undergoes a conformational change and binds with similar affinity as alone (Naktinis et al., 1995; Hingorani and O'Donnell, 1998). These observations suggest that is partially buried in the complex and that the ATP-induced conformational change in complex results in exposure of for interaction with . Which subunit of the complex is responsible for modulating the access of to ? Our previous studies have shown that ' binds stably to the subunit (Onrust and O'Donnell, 1993). Therefore, we examined whether ' might interfere with the interaction between and .

In the experiment shown in Figure 4, we examined the interactions between , ' and by gel filtration using a Superose 12 sizing column. Figure 4A, B and C shows the elution profiles of (25 M dimer), (5 M) and ' (5 M), respectively. Figure 4D shows that (25 M) mixed with the subunit (5 M) forms a complex, and Figure 4E shows that a 1:1 mix of (5 M) and ' (5 M) results in a stable ' complex. Finally, Figure 4F shows that when (25 M) is added to purified ' complex (5 M), the subunit interacts preferentially with (to form ), resulting in the displacement of '. The complex is favored over ' even when the reaction contains a 3-fold excess of ' over (data not shown).

Figure 4.

and ' compete for binding to . Protein–protein interactions were analyzed on a Superose 12 sizing column, and the column fractions were analyzed by SDS–PAGE. Fraction numbers are indicated above the gels, and the positions of , and ' are indicated. (A) The elution profile of alone (25 M dimer), (B) (5 M) and (C) ' (5 M). Protein mixtures were incubated at 25°C for 15 min prior to analysis. (D) The elution profile of a mixture of (5 M) and (25 M), (E) the ' complex (5 M) and (F) the profile of (25 M) mixed with purified ' (5 M).

View full figure (83 KB)

These protein–protein interaction studies indicate that competes with ' for interaction with , and provide some insight into how the complex subunits may work during clamp loading. If, within the complex, ' binds and blocks access of to , then the ATP-dependent change in complex conformation may help remove the ' block and favor – interaction over the –' interaction. Later in the reaction, when the complex binds DNA, the – interaction may be severed to allow the ring to close around DNA. Thus, ' may serve as a modulator during clamp assembly, first by facilitating access of to , allowing it to open the ring, and then by facilitating the dissociation of from , thereby allowing the ring to close around DNA (see Discussion; see Figure 7). Detailed studies with modified ' protein constructs are underway to explore further its role as a modulator in the mechanics of clamp loader activity.

Figure 7.

A model of subunit dynamics during complex-catalyzed assembly of onto DNA. (A) is the 'C'-shaped motor subunit of the clamp loader that harnesses the energy of ATP for assembly of onto DNA. On binding ATP, a conformational change in affects ', another 'C'-shaped protein, which leads to exposure of the ring opener. (B) binds and opens the ring at one interface. (C) The ATPase activity of this protein complex is suppressed until it binds DNA. (D) An encounter with primed DNA stimulates ATP hydrolysis, which is coupled to release of the clamp and DNA. The open dimer snaps shut to form a ring around DNA.

View full figure (71 KB)

No interaction between the subunit and or has been detected. Therefore, if were to completely severe the –' contact in ', the complex would be expected to dissociate from '. Since this does not occur, the or ' subunits presumably maintain some contact with and/or in the '– complex.

ATP binding signals complex– to bind DNA and ATP hydrolysis ejects complex from on DNA

Clamp loading is a multistep process involving more than simple opening of the ring. For example, after binds to and opens the ring, must be positioned at a primed DNA site and then closed to form a continuous protein ring around DNA. The experiment in Figure 1 demonstrated that energy from ATP binding to powers the events leading to ring opening. The next experiment examines interaction of complex and with DNA, and the role of ATP binding and ATP hydrolysis in the clamp-loading reaction. Mixtures of [³H] and complex, as well as and [³H] complex, were used to follow their presence in complex with a primed DNA substrate. In the absence of ATP, no binding of [³H] or [³H] complex to DNA is detectable (Figure 5A and B, respectively). In the presence of ATP--S, both [³H] and [³H] complex co-migrate with the DNA (Figure 5C and D, respectively). appears essential for stable binding of complex to DNA, as no interaction of complex with DNA can be detected by gel filtration in the absence of , whether or not ATP or ATP--S is present in the reaction (Hingorani and O'Donnell, 1998). In the presence of ATP, only [³H] is observed on the DNA, whereas [³H] complex elutes as free protein (Figure 5E and F, respectively; see also Stukenberg et al., 1991). These data indicate that ATP binding to complex locks the clamp loader, clamp and DNA in a ternary complex that is stable to gel filtration, while ATP hydrolysis leads to release of the complex, leaving the ring on DNA.

Figure 5.

ATP binding to complex facilitates interaction with DNA. Interactions of complex and with DNA were assayed in the presence of ATP or ATP--S. [³H] complex (500 fmol) + (500 fmol) or [³²P] + complex were incubated with primed M13mp18 DNA (400 fmol) and 2 mM ATP or ATP--S at 37°C, and analyzed by gel filtration. The fraction numbers are indicated beneath the elution profiles and the fractions corresponding to proteins 'on DNA' and 'free' protein are indicated above the profiles. (A and B) The elution profiles of [³²P] () and [³H] complex (), respectively, in the absence of nucleotide. (C and D) In the presence of ATP--S, () and complex () are bound to DNA. (E and F) In the presence of hydrolyzable ATP, is retained on DNA () and complex is released ().

View full figure (103 KB)

The next experiment was designed to measure the number of ATP molecules hydrolyzed by complex for each clamp assembled on DNA. The clamp-loading reaction was performed with ³H-labeled in the presence of [³²P]ATP. The reactions were analyzed by thin-layer chromatography (TLC) to determine the amount of ATP hydrolyzed, and by gel filtration to quantitate the number of clamps loaded on DNA. The experiment was repeated at four different concentrations of complex. In each case, 2–3 ATP were hydrolyzed for each clamp assembled onto DNA (Table I).

Table 1: ATP hydrolyzed per clamp assembled on DNA

View full table

There is some ambiguity regarding the stoichiometry of subunits per complex. The free subunit can form tetramers (Tsuchihashi and Kornberg, 1989; Dallmann and McHenry, 1995; Onrust et al., 1995a); however, within complex, the estimates range from two to four subunits per complex (Maki and Kornberg, 1988; Dallmann and McHenry, 1995; Onrust et al., 1995a, b). A recent analysis of the molecular mass of complex by multi-angle laser light scattering indicates a molar ratio of 2.5 subunits per complex (unpublished data). Furthermore, quantitative ATP binding experiments yield a stoichiometry of two ATP molecules per complex (Hingorani and O'Donnell, 1998). These data suggest that even if there are more than two subunits per clamp loader, only two bind ATP, which is consistent with hydrolysis of 2–3 ATP per clamp assembled on DNA as shown above. The results with the complex clamp loader are also consistent with an earlier study indicating that two ATP are hydrolyzed during formation of a processive DNA polymerase III holoenzyme complex on primed DNA (Burgers and Kornberg, 1982).

Coupling of ATP hydrolysis to clamp assembly depends on the DNA structure

The previous experiments demonstrated that ATP binding drives complex to open the ring and to bind DNA, whereas ATP hydrolysis ejects complex from –DNA. Next, we examined complex ATPase activity to determine the effects of clamp-loading substrates ( and DNA) on ATP hydrolysis. The ATPase rate was quantitated in the absence and presence of at 37°C (Table II). The data show that the complex ATPase activity is markedly inhibited by , dropping from a turnover rate of 12/min to 1.7/min, respectively (see also Figure 1C). The complex still binds ATP with high affinity in the presence of (K_d = 2 M; Hingorani and O'Donnell, 1998). Therefore, this result implies that binding of to complex stabilizes it in the ATP-bound state, leading to inhibition of ATP hydrolysis. Presumably, inhibition of ATPase activity is important in the clamp-loading pathway to minimize futile cycles of ATP hydrolysis until complex interacts with DNA.

Table 2: ATPase activity of the clamp loader

View full table

DNA substrates stimulate the ATP hydrolysis activity of complex (Onrust et al., 1991). In the absence of , single-stranded DNA and primed DNA increase the complex ATPase rate from 12/min ( complex alone) to 35 and 55/min, respectively (Table II). We have observed that complex dissociates from DNA after hydrolyzing ATP (Figure 5). Thus, DNA may hasten the turnover of complex, such that on binding DNA, the clamp loader rapidly hydrolyzes ATP and releases DNA to return to its original conformation.

Finally, the ATPase activity of complex was measured in the presence of both the clamp and DNA. Addition of DNA to the reaction containing complex– resulted in a large increase in complex ATPase activity (Figure 6A; Table II). Primed single-stranded DNA (ssDNA) is a better effector of ATP hydrolysis than ssDNA (k_cat = 110 and 25/min, respectively, at 37°C). The above experiments were performed with 1 mM ATP in the reaction to ensure maximum ATPase activity ( complex K_m for ATP ranges from 10 to 30 M with unprimed ssDNA and primed ssDNA substrates). Furthermore, to ensure saturating concentrations of the DNA substrates, we repeated the measurements in the presence of an excess of a synthetic ssDNA 100mer or a synthetic primed template DNA (28mer primer/100mer template), which serves as a template for loading (Bloom et al., 1996). Again, the complex hydrolyzes ATP at a higher rate in the presence of primed oligoDNA (350/min) than in the presence of ss oligoDNA (120 min; Table II).

Figure 6.

ATP hydrolysis is coupled to clamp assembly at a primed site on DNA. (A) Primed M13mp18 ssDNA stimulates the ATPase activity of complex (in the presence of ) from 1.7/min (Table I) to 110/min (), while ssDNA stimulates the ATPase rate to 25/min (). (B) A gel filtration analysis of complex-catalyzed assembly of [³H] on primed DNA () compared with assembly on ssDNA (). The fraction numbers are indicated beneath the elution profile, and the fractions corresponding to ' on DNA' and 'free ' are indicated above the profile.

View full figure (68 KB)

In Figure 6B, we investigated whether ATP hydrolysis stimulated by ssDNA leads to assembly of clamps on ssDNA. The results show that no loading of [³H] clamps can be detected on unprimed, single-stranded M13DNA. Thus, complex ATPase activity is not coupled to clamp assembly on this substrate. The specificity of complex activity for clamp loading on primed DNA reflects the need for sliding clamps at primed sites on DNA that can be extended by the DNA polymerase. The uncoupled ATPase activity at unprimed ssDNA may in fact play an important role during lagging strand synthesis by helping complex locate primed sites on DNA by scanning long stretches of ssDNA via repeated DNA binding and release events.

Key elements of the clamp-loading mechanism

The complex catalyzes assembly of the clamp onto a circular DNA molecule to form a protein–DNA catenane. In a process fueled by ATP, the clamp loader opens the clamp at the dimer interface, binds DNA, and then releases the clamp so that it forms a closed ring around DNA. The subunit is the key component of complex that opens the ring. The interaction between and yields sufficient energy to destabilize the dimer interface. Therefore, ring-opening does not require ATP binding or hydrolysis, but is a result of a simple protein–protein interaction. Nevertheless, the 'ring opener' cannot facilitate assembly of the open ring onto DNA by itself. This process requires the combined activities of , and ', indicating that clamp loading is a more complicated process than simply opening the ring. Thus, complex must perform several functions to complete the process of clamp assembly on DNA. For example, the open ring and DNA must be positioned correctly for topological linkage, the ring must be closed and, finally, complex must dissociate from the ring on DNA.

Dissociation of complex from is necessary for its catalytic activity and, more importantly, it is a prerequisite for processive DNA synthesis, as the clamp loader and polymerase compete for the same site on (Naktinis et al., 1996). Once complex releases , the polymerase can bind the clamp and initiate DNA synthesis (Naktinis et al., 1996). In DNA polymerase III holoenzyme, the complex and both leading and lagging strand polymerases are held together by a dimer (Onrust et al., 1995a). The complex catalyzes rapid assembly of sliding clamps onto DNA for use by both polymerases (Yuzhakov et al., 1996). This is particularly significant during lagging strand synthesis. On the lagging strand, the polymerase releases its clamp upon finishing an Okazaki fragment; it then rapidly targets to a new that has been assembled at a fresh upstream primed site by the complex. The complex also recycles the used clamps that have been abandoned on completed Okazaki fragments (Naktinis et al., 1996). Thus, complex and polymerase work in concert for clamp loading and primer extension, resulting in rapid and efficient duplication of E.coli genomic DNA.

A model for complex-catalyzed clamp loading

The complex can be described as a 'switch' protein that alternates between the 'on' state (ATP bound) and the 'off' state (either ADP or no nucleotide bound). It is, however, more than a simple switch protein because in each state, and in the transition between these states, the clamp loader machinery performs multistep functions that together result in clamp loading. A model summarizing the current information is presented in Figure 7 and the structure–function details are described below. The model shows only two subunits, although estimates of the stoichiometry of in complex range from two to four (Maki and Kornberg, 1988; Dallmann and McHenry, 1995; Onrust et al., 1995a, b).

The crystal structure of the ' subunit has been solved recently (Guenther et al., 1997). The protein is composed of three domains organized in the form of the letter 'C'. The top and bottom domains are barely connected through a small domain that may function as a hinge. The sequence identity between ' and predicts that also has a 'C' shape (Guenther et al., 1997). The structure of , modeled after ', predicts that the ATP-binding region lies very close to the hinge region between the top and bottom domains of the 'C'. This central position suggests that ATP binding to may perturb the 'C' shape, perhaps resulting in movement of the top and bottom domains, as speculated in Figure 7A. Recent reports of an ATP-induced conformational change in , and in complex, are consistent with perturbation of the 'C' shape (Naktinis et al., 1995; Hingorani and O'Donnell, 1998).

The ATP-induced changes in suggest that it functions as a 'motor', using the energy of ATP binding to power the conformational change in complex that takes from a buried position into one that is exposed and allows interaction with the ring (Naktinis et al., 1996). The present study indicates that ' may be the subunit that prevents from binding in the absence of ATP. Hence, ATP binding to may re-position ' and/or affect the ring opener , such that it prefers binding to rather than '. In Figure 7A, this is depicted as a change in the ' subunit 'C' shape in response to the ATP-induced change in the subunit 'C' shape (i.e. the ' conformation is guided by ). The resulting arrangement leaves complex in the 'on' state in which can bind the clamp and open the ring at one interface (Figure 7B). The resulting complex–ATP–open ring composite is stabilized by the decrease in ATPase activity of the complex (Table II). If ATP were hydrolyzed, the clamp loader would release the clamp and 'turn over'. Thus the inhibition of ATPase activity probably maintains the integrity of the complex–ATP–open ring composite for an encounter with a DNA substrate.

The complex–ATP–open ring composite binds primed DNA with high affinity to form a complex–ATP–open –DNA composite (Figure 7C). The DNA stimulates complex ATPase activity which leads to deposition of the clamp onto DNA, and ejection of complex from on DNA (Figure 7D). Two to three molecules of ATP are hydrolyzed per clamp-loading event. Accordingly, the model in Figure 7 indicates that both subunits hydrolyze ATP and that after hydrolysis (or upon ADP + P_i dissociation), they return to their original 'C' shape. Presumably, at this point, the – interaction is disrupted, possibly facilitated by the rebinding of ' to , thereby facilitating release of complex from and closure of the ring around DNA (Figure 7D).

The energy from ATP hydrolysis may be used simply to orchestrate dissociation of the complex from the clamp and DNA. Alternatively, ATP hydrolysis may fuel a more active process such as closing of the clamp around DNA. At this time, we favor the former, 'clamp release' mechanism, keeping in mind that ATP binding powers assembly of all three components ( complex, and DNA) in close proximity, and that the lowest free energy state of the free clamp is a closed ring. Therefore, when is released, it should be capable of closing around DNA without further input of energy from ATP hydrolysis. Pre-steady-state studies of clamp loading and complex ATPase activity are underway to define more precisely the role of ATP hydrolysis in clamp loading.

Clamp assembly in vivo

DNA polymerase III holoenzyme contains two copies each of and (Onrust et al., 1995a). The subunit is approximately the N-terminal two-thirds of the subunit, and it is created by a frameshift during protein expression from the dnaX gene (Tsuchihashi and Kornberg, 1989; Blinkova and Walker, 1990; Flower and McHenry, 1990). The frameshift allows addition of one unique amino acid to , followed by a stop codon. In vitro, both and subunits can be assembled into functional clamp loaders with the , ', and subunits (Onrust et al., 1995b). In addition, has the unique ability to bind the core polymerase (McHenry, 1982; Studwell-Vaughan and O'Donnell, 1991) as well as the DnaB helicase with high affinity (Kim et al., 1996; Yuzhakov et al., 1996). These properties require the extra 213 amino acids in , which are absent in . Specifically a dimer, not , dimerizes the two core DNA polymerases in the holoenzyme and contacts the DnaB helicase during DNA replication.

Which protein, or , functions as the motor subunit for the clamp loader within the holoenzyme, or do both and form the clamp loader (i.e. a complex)? This question has been addressed in a previous study of the and proteins. ATP-binding site mutants of and were constructed which, when constituted into either complex or complex with , ', and , were unable to hydrolyze ATP and were inactive in DNA replication assays with core polymerase and (Xiao et al., 1995). The mutant proteins were also assembled into DNA polymerase III holoenzymes. The holoenzyme containing mutant + wild-type was inactive in DNA replication. The holoenzyme containing mutant + wild-type was as active as the wild-type DNA polymerase III holoenzyme. These results demonstrate that the ATP site of is needed for holoenzyme function, and thus is the clamp loading motor in the holoenzyme.

In normal cells, the holoenzyme contains both and (Maki et al., 1988). However, if the frameshift signal in dnaX is mutated such that is not produced, the cells remain viable (Blinkova et al., 1993). Presumably the holoenzyme in these -less cells contains four subunits, where two subunits dimerize the polymerases and the other two substitute for as the clamp loader motor. However, even though the subunit can substitute for , its ATPase activity and clamp-loading activity is not essential when the subunit is present in the holoenzyme, as demonstrated by the ATP site mutant study (Xiao et al., 1995).

The complex has the unique ability to utilize ATP--S to provide the core polymerase with -dependent processivity for DNA synthesis, although the level of synthesis is much lower than in the presence of ATP; the complex does not function in this capacity (Dallmann et al., 1995). This phenomenon is interesting since neither complex nor complex hydrolyze ATP--S significantly, and ATP hydrolysis is required by both clamp loaders for the completion of clamp assembly onto DNA (Hingorani and O'Donnell, 1998; unpublished data). The results of this report offer a new insight into the molecular mechanism by which the complex plus core, or the holoenzyme, may use ATP--S for processive DNA replication. In as much as and are closely related, and ATP--S induces complex to bind and open the ring, it may be presumed that ATP--S has the same effect on complex. Since , unlike , can bind the core polymerase, it may form a bridge between the polymerase and the clamp, resulting in an enzyme assembly that can function processively. The inability of to bind core would prevent formation of such an assembly, and, therefore, complex cannot facilitate -dependent DNA synthesis by the core in the presence of ATP--S. Thus, even though the complex appears to be the primary clamp loader in DNA polymerase III holoenzyme, under artificial conditions, such as in the presence of ATP--S (Dallmann et al., 1995) or when is absent from the holoenzyme (Blinkova et al., 1993), is eminently capable of supporting clamp assembly, in addition to dimerizing the polymerases and maintaining contact with the helicase during DNA replication

A common mechanism for clamp loading?

Clamp loaders of three other well studied systems include the gp44/62 complex from bacteriophage T4, yRF-C from S.cerevisiae and hRF-C from humans. These clamp loaders use ATP to assemble their respective circular clamps, T4 gp45, yeast proliferating cell nuclear antigen (PCNA) and human PCNA, onto DNA.

Like complex, both the yeast and human clamp loaders are five-protein machines (O'Donnell et al., 1993). Earlier studies have shown that ATP--S leads to formation of an RF-C–PCNA–DNA composite, but this is not competent for extension by DNA polymerase in the absence of hydrolyzable ATP (Lee and Hurwitz, 1990; Burgers, 1991). A possible interpretation of this result is that RF-C functions similarly to complex, in that ATP binding is sufficient for ring opening and interaction with DNA, but ATP hydrolysis is needed to eject RF-C, allowing the PCNA ring to close around DNA. As in the E.coli system, DNA polymerase requires contact with the same face of PCNA as RF-C. Hence RF-C must release PCNA for the polymerase to gain access to the ring. Studies of human RF-C have shown that a three-subunit subassembly (36, 37 and 40 kDa subunits) as well as the 40 kDa subunit alone is capable of removing PCNA clamps from DNA in the absence of ATP (Cai et al., 1997). These observations indicate that the RF-C clamp loaders may utilize a similar modular design and mechanism for assembly of rings around DNA as the E.coli clamp loader. The complex subunits also provide ancillary functions; for example the subunit binds SSB and is necessary for displacement of the primase from RNA primers (Kelman et al., 1998; Yuzhakov et al., 1999). It is tempting to speculate that eukaryotic clamp loader subunits may also provide additional functions at the site of DNA synthesis, other than clamp assembly on DNA.

The bacteriophage T4 clamp loader contains two different subunits (gp44 and gp62), and a circular sliding clamp (gp45), which is a loosely associated trimer that lacks the stability of the and PCNA rings (Yao et al., 1996). Recent studies on the role of ATP in gp45 assembly on DNA suggest that ATP hydrolysis is needed to open the gp45 ring (Latham et al., 1996; Sexton et al., 1996; Pietroni et al., 1997). In contrast, in E.coli, ATP binding is sufficient for ring opening, and ATP hydrolysis is used to eject complex from the ring on DNA. The unstable nature of the trimeric gp45 ring compared with the dimer may account for this significantly different role for ATP hydrolysis in the T4 system compared with E.coli. For example, gp44/62-catalyzed opening of the gp45 trimer may enhance dissociation of the protomers and preclude reclosure of the clamp around DNA. Therefore, it is possible that the T4 system has evolved to allow assembly of the clamp loader and closed clamp on DNA before ATP hydrolysis, which powers both ring opening and loading of the ring onto DNA at the same time.

Clamp loaders in cell cycle regulation

Recently, a number of clamp loader or clamp loader-like proteins have been implicated in cell cycle checkpoint pathways (reviewed in Elledge, 1996; Mossi and Hubscher, 1998). Classical checkpoint proteins sense particular aspects of a cell's environment or condition and 'decide' whether to restrict or promote cell cycle progression from one phase to the next. For example, some checkpoint proteins monitor the integrity of DNA prior to and after replication. If DNA is damaged, the cell cycle can be arrested in G₁ phase, slowed in S phase and arrested in G₂ phase, to allow repair of the DNA prior to, during and after DNA replication, respectively. Similarly, if DNA synthesis is incorrect or incomplete, DNA replication checkpoints can inhibit transition of the cell cycle (e.g. from G₂ to mitosis) until the specific problems are fixed.

Three of the five subunits of the S.cerevisiae clamp loader, RF-C1, RF-C2 and RF-C5, are now thought participate in cell cycle checkpoint pathways involving DNA metabolism (Cullmann et al., 1995; Sugimoto et al., 1996; Noskov et al., 1998). Although the underlying mechanism by which these proteins regulate the cell cycle is not yet understood, it is possible that the regulatory pathways involve loading of PCNA onto DNA. For example, a clamp loader is crucial at an early stage in DNA replication for initiation of processive DNA synthesis; therefore, a shutdown of its clamp-loading activity can prevent entry into S phase. The large subunit of human RF-C (p140) has several putative phosphorylation sites for protein kinase A, protein kinase C and tyrosine kinases, and this has led to speculation that the PCNA loading activity of RF-C may be regulated by phosphorylation events in checkpoint pathways (Mossi and Hubscher, 1998). Alternatively, clamp loaders may serve as signal sensors in checkpoint pathways through recognition of specific DNA structures produced during DNA replication and/or repair processes. For example, recent evidence suggests that RF-C5 may interact with Spk1 (Rad53/Mec2/Sad1), an essential protein kinase for the transition of S phase to mitosis (Sugimoto et al., 1996).

Structural homologs of clamp loader subunits are also known to participate in cell cycle regulation. RAD24 in S.cerevisiae, which is homologous to E.coli and ' proteins as well as all the RF-C subunits, is required for DNA damage checkpoint pathways, and its homolog in Schizosaccharomyces pombe, RAD17, is required for both DNA replication and damage checkpoints (Elledge, 1996). Perhaps RAD24 and RAD17 interact with RF-C and regulate its clamp-loading activity, or they may be subunits of other protein loaders that play a role in checkpoint pathways. A recent study has shown that Cdc6p, an S.cerevisiae protein essential for entry of cells into S phase, also shares homology with prokaryotic and eukaryotic clamp loader subunits (Perkins and Diffley, 1998). It has also been suggested that Cdc6p may be involved in assembly of Mcm proteins onto chromatin, and thereby serves a function analogous to that of a clamp loader in assembling proteins onto DNA (Perkins and Diffley, 1998). Although quite speculative in nature, these findings imply that the use of clamp loaders and their underlying mechanisms may generalize to other processes beyond loading sliding clamps onto DNA.

Studies of E.coli DNA replication have been a guide to understanding similar mechanisms in complex eukaryotes, including humans. This report provides a detailed mechanism for the function of the complex clamp loader, a multiprotein enzyme that assembles sliding clamps onto DNA for processive and rapid DNA replication. The structural and functional homology between complex, RF-C and other important cell cycle regulatory proteins suggests that insights gleaned from the molecular mechanism of the complex will generalize to eukaryotic clamp loaders, and perhaps to yet other critical processes in DNA metabolism.

Materials

Radioactive nucleotides were purchased from Dupont-NEN. M13mp18 ssDNA was prepared by phenol extraction of purified M13mp18 phage that had been banded twice in CsCl gradients as described (Turner and O'Donnell, 1995). Nicked circular DNA was prepared by incubating 200 g of doubly CsCl-banded supercoiled Bluescript plasmid (pBS) with 100 U of mung bean nuclease (NEB) for 15 min, followed by phenol extraction and ethanol precipitation. Mung bean nuclease is a single strand-specific nuclease that nicks supercoiled DNA once, and by relaxing the plasmid precludes further nicking. The DNA was 90% nicked. Multiprimed DNA was prepared by incubating 0.6 M DnaG primase and 0.6 M DnaB helicase with 1 mM NTPs for 30 min at 37°C, followed by phenol extraction and ethanol precipitation. This treatment results in 30 primers per DNA circle (assuming an average primer length of 25 nucleotides). Synthetic ssDNA 100mer and synthetic primed template DNA were prepared as described (Hingorani and O'Donnell, 1998).

Buffer A is 20 mM Tris–HCl pH 7.5, 0.1 mM EDTA and 10% glycerol. Superose 12 gel filtration buffer is 20 mM Tris–HCl pH 7.5, 0.1 mM EDTA, 10% glycerol, 5 mM DTT and 100 mM NaCl. Reaction buffer is 20 mM Tris–HCl pH 7.5, 0.1 mM EDTA, 4% glycerol and 8 mM MgCl₂. Bio-Gel A-15m gel filtration buffer is reaction buffer plus 100 mM NaCl. Loading buffer is 50 mM Tris–HCl (pH 6.8), 100 mM DTT, 2% SDS, 0.1% bromophenol blue and 10% glycerol.

Proteins

The complex was constituted from pure , , ', and , followed by isolation from the free subunits on a MonoQ column (Pharmacia) as described (Onrust et al., 1995b). Purification of (Studwell and O'Donnell, 1990), and ' (Dong et al., 1993), and (Xiao et al., 1993) was performed as described earlier. The ' and proteins were ³H-labeled by reductive methylation as described (Kelman et al., 1995a). The catalytic subunit of cAMP-dependent protein kinase produced in E.coli was the gift of Dr S.Taylor (University of California, San Diego, CA).

The modified versions of were constructed by PerImmune, Inc. using a DNA oligonucleotide site-directed method. The mutant dnaN genes were placed into vectors encoding an N-terminal hexa-histidine tag and a kinase recognition sequence (pHK vector; Kelman et al., 1995b). The specific modifications are described in detail later. Below is a summary of the purification protocol. Modified was overexpressed in E.coli BL21(DE3) pLysS cells by inducing cultures at OD₆₀₀ = 0.6 with 0.4 mM isopropyl--D-thiogalactopyranoside (IPTG). The cells were harvested, and heat lysed by the method of Wickner (1976). The cell lysate was precipitated sequentially with 35 and 70% ammonium sulfate and the pellet was resuspended in 20 mM Tris–HCl (pH 7.9) 500 mM NaCl and 5 mM imidazole. The protein was loaded onto a 5 ml chelating Sepharose resin (Pharmacia) charged with NiSO₄, and bound proteins were eluted with a linear gradient of imidazole (60–500 mM) in the above buffer. Fractions were mixed immediately with EDTA (50 mM final) to chelate any nickel in the eluate that tends to precipitate protein. was identified by SDS–PAGE, and peak fractions were pooled and dialyzed against buffer A. The pool was loaded onto an 8 ml MonoQ column equilibrated in buffer A, and the bound protein was eluted with a linear NaCl gradient (0–500 mM) in buffer A. Peak fractions were pooled and frozen at -70°C. Each of the modified subunits had replication activity levels comparable with wild-type .

Ring-opening assay

The clamp was modified by changing Leu273 to cysteine and Cys333 to serine. The resulting Cys273 residue is buried in the dimer interface (L273C), and the highly reactive surface Cys333 is removed. The remaining three cysteine residues in were found unreactive to thiol-labeling reagents and were left unchanged. The reagent used to probe the interface cysteine was prepared by labeling 300 nmol of peptide (NH₃-LRRASVP-COOH; Chiron mimotopes) with 0.03 U of cAMP-dependent protein kinase catalytic subunit and [-³²P]ATP 200 Ci (3000 Ci/mmol) in 150 l of reaction buffer (30 min at 37°C). Then, ATP was added to a final concentration of 10 mM and the mixture was incubated for an additional 30 min. The mixture was incubated with 5 mg of sulfosuccinimidylc 4-[N-maleimidomethyl] cyclohexane-1-carboxylate (Sulfo-SMCC; Pierce) at 25°C for 1 h, and the ³²P-labeled maleimide was purified on a 24 ml Superdex-Peptide column (Pharmacia) equilibrated in 0.1 M sodium phosphate (pH 7.2), 0.15 M NaCl, using the Pharmacia SMART system. After a 7 ml void volume, 500 l fractions were collected. The peak (fractions 17–18) containing [³²P]maleimide was identified by the co-incidence of the radioactive peak and the peak of absorbance at 305nm (_max for N-ethylmaleimide). [³²P]maleimide concentration was determined at 305 nm, using a molar extension coefficient of 620/M/cm for maleimide, and was typically at 400 M.

The ring-opening assay was performed in 15 l of reaction buffer containing 3 M , 3 M complex, 2 mM ATP and 3 M primed oligoDNA. Proteins were mixed on ice and the reaction was initiated with 5 l of [³²P]maleimide and a shift to 37°C. The reactions were quenched after 12 min with 5 l of loading buffer + 5 l of 1 M DTT. Samples were analyzed on a 10% SDS–polyacrylamide gel that subsequently was dried and analyzed on a PhosphorImager. The complex subunits exhibit some background labeling by [³²P]maleimide; however, they are easily resolved from during SDS–PAGE and do not interfere with the detection of ring opening.

In experiments using ATP--S, 5 M each of complex and were incubated in 30 l of reaction buffer with either 200 M ATP or ATP--S. Reactions were initiated by adding 10 l of [³²P]maleimide, and 4 l aliquots were removed and quenched at 0, 0.5, 1, 2, 5 and 10 min, and analyzed as described above.

Cross-linked loading assay

For this experiment, Arg103 and Ile305 in were changed to cysteine to create a disulfide-linked (^S-S). Native and ^S-S were radiolabeled with the catalytic subunit of cAMP-dependent protein kinase and [-³²P]ATP as described (Kelman et al., 1995a), except that DTT was omitted in order to maintain the disulfide cross-links in ^S-S. The complex used in this assay was constituted in the absence of DTT.

The assembly reaction was performed in 50 l of reaction buffer containing 500 fmol of pBS DNA, 19 pmol of , 5.4 pmol of complex and 500 M ATP, in the absence or presence of 20 mM DTT. The reaction was allowed to proceed for 10 min at 37°C and then loaded onto a 5 ml Bio-Gel A-15m gel filtration column (Bio-Rad), equilibrated in A-15m buffer in the presence or absence of 1% SDS. Fractions (200 l) were collected and 40 l aliquots were quantitated by liquid scintillation counting or analyzed on a 12% SDS–polyacrylamide gel under non-reducing conditions (no DTT). The gels were dried and exposed to film or analyzed on a PhosphorImager (Molecular Dynamics)

Clamp-unloading assay

(1.6 M), ³²P-labeled at a C-terminal kinase site (Naktinis et al., 1996), was incubated with 3.3 M complex and 5 pmol of pBS DNA in 75 l of reaction buffer containing 500 M ATP for 10 min at 37°C. The reaction was applied to a 5 ml A-15m gel filtration column and eluted with A-15m buffer. Peak fractions, containing on DNA, were pooled, and 25 l aliquots (containing 0.2 M ) were incubated with 1 M , or ' with or without 500 M ATP. After 10 min, each sample was mixed with glycerol (15% final) and analyzed on a 1% agarose gel (89 mM Tris–borate buffer with 10 ng/ml ethidium bromide). After electrophoresis for 2 h at 100 V, the agarose gel was photographed under UV light and exposed to film.

Analysis of , and ' interactions

A ' complex was formed by mixing 2 mg of and 3 mg of ' in buffer A followed by purification on a 1 ml MonoQ column (with a linear gradient of 50–500 mM NaCl in buffer A). Then 5 M of the pure ' complex was mixed with 25 M (dimer) in 200 l of buffer A for 15 min at 25°C, and analyzed by gel filtration on a Superose 12 column equilibrated with buffer A containing 100 mM NaCl. After a void volume of 7 ml, 180 l fractions were collected, and 20 l aliquots were analyzed by SDS–PAGE and visualized by Coomassie staining of the gels. The control gel filtration profiles of alone, alone, ' alone, with , and with ' were also obtained under the conditions described above for the Superose 12 column chromatography.

Interaction of complex and with DNA

To monitor complex interaction with DNA, 500 fmol of [³H-'] was incubated with 500 fmol of and 400 fmol of multiply primed M13mp18 DNA for 10 min at 37°C in reaction buffer containing 2 mM ATP or ATP--S (75 l total volume). To monitor assembly onto DNA, [³H] was incubated with unlabeled complex. The reactions were applied to 5 ml A-15m gel filtration columns equilibrated in reaction buffer containing 0.2 mM ATP or ATP--S. Column fractions were quantitated by liquid scintillation counting, and the amount of complex and in the fractions was determined from the known specific activity.

complex ATPase activity

The number of ATP molecules hydrolyzed by complex for each assembled onto DNA was measured by incubating 25, 50, 100 or 200 nM complex with 1 M [³H] at 25°C, in reaction buffer containing 40 mM glucose. At time zero, 1 M [-³²P]ATP was added to the reaction and a 1 l aliquot was removed and quenched with 1 l of 0.5 M EDTA for quantitation of ATP hydrolysis. After 15 s, another 1 l aliquot was removed to quantitate of the amount of ATP hydrolyzed, and 1 pmol of nicked DNA was added as substrate for complex-catalyzed loading of [³H] clamps. After an additional 15 s, a final 1 l aliquot was removed for ATP analysis and the assembly reaction was quenched by addition of 0.5 U of hexokinase. The reaction was applied to an A-15m gel filtration column to quantitate the amount of [³H] assembled onto DNA. ATP hydrolysis was quantitated by TLC on PEI cellulose F sheets (EM Science), using 0.6 M potassium phosphate buffer, pH 3.4. The TLC sheet was dried and the molar amount of ADP + P_i formed was determined by analysis on a PhosphorImager.

ATPase assays (or ATP--Sase assays) measuring complex turnover rate were performed at 37°C with 0.1 M complex, with or without 0.2 M and with 1 mM ATP + [-³²P]ATP (or 0.5 mM ATP--S + [-³⁵S]ATP--S). At various times, 5 l aliquots of the reaction were quenched with equal volumes of 0.5 M EDTA and analyzed as described above. To compare the effect of primed and ssDNA on complex– -catalyzed ATP hydrolysis, similar reactions were performed, except with 0.1 M M13 ssDNA coated with 17 M SSB or with 0.1 M primed M13 DNA coated with SSB in the reaction (or 0.5 M ss-oligoDNA and primed oligoDNA). The molar amounts of products formed were plotted versus time of reaction, and the slope of each time course was divided by complex concentration to yield the turnover number.

We are grateful to Dr Susan Taylor for the catalytic subunit of the cAMP-dependent protein kinase. This work was supported by NIH grant GM38839. M.M.H. was supported by a grant from the Charles H.Revson Foundation.

Blinkova AL and Walker JL (1990) Programmed ribosomal frameshifting generates the Escherichia coli DNA polymerase III subunit from within the subunit reading frame. Nucleic Acids Res, 18, 1725–1729. | PubMed | ISI | ChemPort |

Blinkova AL, Hervas C, Stukenberg PT, Onrust R, O'Donnell M and Walker JR (1993) The Escherichia coli DNA polymerase III holoenzyme contains both products of the dnaX gene, and , but only is essential. J Bacteriol, 175, 6018–6027. | PubMed | ISI | ChemPort |

Bloom LB, Turner JT, Kelman Z, Beechem JM, O'Donnell M and Goodman M (1996) Dynamics of loading the sliding clamp of DNA polymerase III onto DNA. J Biol Chem, 271, 30699–30708. | Article | PubMed | ISI | ChemPort |

Burgers PMJ (1991) Saccharomyces cerevisiae replication factor C. II. Formation and activity of complexes with the proliferating cell nuclear antigen and with DNA polymerases and . J Biol Chem, 266, 22698–22706. | PubMed | ISI | ChemPort |

Burgers PMJ and Kornberg A (1982) ATP activation of DNA polymerase III holoenzyme from Escherichia coli. II. Initiation complex: stoichiometry and reactivity. J Biol Chem, 257, 11474–11478. | PubMed | ChemPort |

Burgers PMJ and Yoder BL (1993) ATP-independent loading of the proliferating cell nuclear antigen requires DNA ends. J Biol Chem, 268, 19923–19926. | PubMed | ISI | ChemPort |

Cai J, Gibbs E, Uhlmann F, Phillips B, Yao N, O'Donnell M and Hurwitz J (1997) A complex consisting of human replication factor C p40, p37 and p36 subunits is a DNA-dependent ATPase and an intermediate in the assembly of the holoenzyme. J Biol Chem, 272, 18974–18981. | Article | PubMed | ISI | ChemPort |

Cullmann G, Fien K, Kobayashi R and Stillman B (1995) Characterization of the five replication factor C genes of Saccharomyces cerevisiae. Mol Cell Biol, 15, 4661–4671. | PubMed | ISI | ChemPort |

Dallmann HG and McHenry CS (1995) DnaX complex of Escherichia coli DNA polymerase III holoenzyme. J Biol Chem, 270, 29563–29569. | Article | PubMed | ISI | ChemPort |

Dallmann HG, Thimmig RL and McHenry CS (1995) DnaX complex of Escherichia coli DNA polymerase III holoenzyme. J Biol Chem, 270, 29555–29562. | Article | PubMed | ISI | ChemPort |

Dong Z, Onrust R, Skangalis M and O'Donnell M (1993) DNA polymerase III accessory proteins. I. holA and holB encoding and '. J Biol Chem, 268, 11758–11765. | PubMed | ISI | ChemPort |

Elledge SJ (1996) Cell cycle checkpoints: preventing an identity crisis. Science, 274, 1664–1672. | Article | PubMed | ISI | ChemPort |

Flower AM and McHenry CS (1990) The subunit of DNA polymerase III holoenzyme Escherichia coli is produced by ribosomal frameshifting. Proc Natl Acad Sci USA, 87, 3713–3717. | Article | PubMed | ChemPort |

Guenther B, Onrust R, Sali A, O'Donnell M and Kuriyan J (1997) Crystal structure of the subunit of the clamp-loader complex of E.coli DNA polymerase III. Cell, 91, 335–345. | Article | PubMed | ISI | ChemPort |

Gulbis JM, Kelman Z, Hurwitz J, O'Donnell M and Kuriyan J (1996) Structure of the C-terminal region of p21 (WAF1/CIP1) complexed with human PCNA. Cell, 87, 297–306. | Article | PubMed | ISI | ChemPort |

Hingorani MM and O'Donnell M (1998) ATP binding to the E.coli clamp loader powers opening of the ring-shaped clamp of DNA polymerase III holoenzyme on DNA. J Biol Chem, 273, 24550–24563. | Article | PubMed | ISI | ChemPort |

Kelman Z and O'Donnell M (1995) DNA polymerase III holoenzyme: structure and function of a chromosomal replicating machine. Annu Rev Biochem, 64, 171–200. | Article | PubMed | ISI | ChemPort |

Kelman Z, Naktinis V and O'Donnell M (1995a) Radiolabeling of proteins for biochemical studies. Methods Enzymol, 262, 430–442. | PubMed | ISI | ChemPort |

Kelman Z, Yao N and O'Donnell M (1995b) Escherichia coli expression vectors containing a protein kinase-recognition motif, His₆-tag and hemagglutinin epitope. Gene, 166, 177–178. | Article | PubMed | ISI | ChemPort |

Kelman Z, Yuzhakov A, Andjelkovic J and O'Donnell M (1998) Dedicated to the lagging strand—the subunit of DNA polymerase III holoenzyme contacts SSB to promote processive elongation and sliding clamp assembly. EMBO J, 17, 2436–2449. | Article | PubMed | ChemPort |

Kim S, Dallmann HG, McHenry CS and Marians KJ (1996) Coupling of a replicative polymerase and helicase: a –DnaB interaction mediates rapid replication fork movement. Cell, 84, 643–650. | Article | PubMed | ISI | ChemPort |

Kong XP, Onrust R, O'Donnell M and Kuriyan J (1992) Three-dimensional structure of the subunit of E.coli DNA polymerase III holoenzyme: a sliding DNA clamp. Cell, 69, 425–437. | Article | PubMed | ISI | ChemPort |

Kornberg A and Baker T (1992) DNA Replication. 2nd edn. W.H.Freeman and Company, New York, NY, pp. 165–194.

Krishna TSR, Kong XP, Gary S, Burgers P and Kuriyan J (1994) Crystal structure of the eukaryotic DNA polymerase processivity factor PCNA. Cell, 79, 1233–1243. | Article | PubMed | ISI | ChemPort |

Latham GJ, Pietroni P, Dong F, Young M and von Hippel PH (1996) Fluorescence monitoring of T4 polymerase holoenzyme accessory protein interactions during loading of the sliding clamp onto the template–primer junction. J Mol Biol, 264, 426–439. | Article | PubMed | ISI | ChemPort |

Lee SH and Hurwitz J (1990) Mechanism of elongation of primed DNA by DNA polymerase , proliferating cell nuclear antigen and activator 1. Proc Natl Acad Sci USA, 87, 5672–5676. | Article | PubMed | ChemPort |

Maki S and Kornberg A (1988) DNA polymerase III holoenzyme of Escherichia coli. II. A novel complex including the subunit essential for processive synthesis. J Biol Chem, 263, 6555–6560. | PubMed | ISI | ChemPort |

Maki H, Maki S and Kornberg A (1988) DNA polymerase III holoenzyme of Escherichia coli. IV. The holoenzyme is an asymmetric dimer with twin active sites. J Biol Chem, 263, 6570–6578. | PubMed | ISI | ChemPort |

McHenry CS (1982) Purification and characterization of DNA polymerase III: identification of as a subunit of the DNA polymerase III holoenzyme. J Biol Chem, 257, 2657–2663. | PubMed | ISI | ChemPort |

Mossi R and Hubscher U (1998) Clamping down on clamps and clamp loaders, the eukaryotic replication factor C. Eur J Biochem, 254, 209–216. | Article | PubMed | ISI | ChemPort |

Naktinis V, Onrust R, Fang L and O'Donnell M (1995) Assembly of a chromosomal replication machine: two DNA polymerases, a clamp loader and sliding clamps in one holoenzyme particle. II. Intermediate complex between the clamp loader and its clamp. J Biol Chem, 270, 13358–13365. | Article | PubMed | ISI | ChemPort |

Naktinis V, Turner J and O'Donnell M (1996) A molecular switch in a replication machine defined by an internal competition for protein rings. Cell, 84, 137–145. | Article | PubMed | ISI | ChemPort |

Noskov VN, Araki H and Sugino A (1998) The RFC2 gene, encoding the third-largest subunit of the replication factor C complex, is required for an S-phase checkpoint in S.cerevisiae. Mol Cell Biol, 18, 4914–4923. | PubMed | ISI | ChemPort |

O'Donnell M (1987) Accessory proteins bind a primed template and mediate rapid cycling of DNA polymerase III holoenzyme from Escherichia coli. J Biol Chem, 265, 16558–16565.

O'Donnell M, Onrust R, Dean FB, Chen M and Hurwitz J (1993) Homology in accessory proteins of replicative polymerases, E.coli to humans. Nucleic Acids Res, 21, 1–3. | Article | PubMed | ChemPort |

Onrust R and O'Donnell M (1993) DNA polymerase III accessory proteins. II. Characterization of and '. J Biol Chem, 268, 11766–11772. | PubMed | ISI | ChemPort |

Onrust R, Stukenberg PT and O'Donnell M (1991) Analysis of the ATPase subassembly which initiates processive DNA synthesis by DNA polymerase III holoenzyme. J Biol Chem, 266, 21681–21686. | PubMed | ISI | ChemPort |

Onrust R, Finkelstein J, Turner J, Naktinis V and O'Donnell M (1995a) Assembly of a chromosomal replication machine: two DNA polymerases, a clamp loader and sliding clamps in one holoenzyme particle. III. Interface between two polymerases and the clamp loader. J Biol Chem, 270, 13366–13377. | Article | PubMed | ISI | ChemPort |

Onrust R, Finkelstein J, Naktinis V, Turner J, Fang L and O'Donnell M (1995b) 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. | Article | PubMed | ISI | ChemPort |

Perkins G and Diffley JFX (1998) Nucleotide-dependent prereplicative complex assembly by Cdc6p, a homolog of eukaryotic and prokaryotic clamp-loaders. Mol Cell, 2, 23–32. | Article | PubMed | ISI | ChemPort |

Pietroni P, Young MC, Latham GJ and von Hippel PH (1997) Structural analyses of gp45 sliding clamp interactions during assembly of the bacteriophage T4 DNA polymerase holoenzyme. I. Conformational changes within the gp44/62–gp45–ATP complex during clamp loading. J Biol Chem, 272, 31666–31676. | Article | PubMed | ISI | ChemPort |

Sexton DJ, Carver TE, Berdis AJ and Benkovic SJ (1996) Protein–protein and protein–DNA interactions at the bacteriophage T4 replication fork. Characterization of a fluorescently labeled DNA polymerase sliding clamp. J Biol Chem, 271, 28045–28051. | Article | PubMed | ISI | ChemPort |

Studwell PS and O'Donnell M (1990) Processive replication is contingent on the exonuclease subunit of DNA polymerase III holoenzyme. J Biol Chem, 265, 1171–1178. | PubMed | ISI | ChemPort |

Studwell-Vaughan PS and O'Donnell M (1991) Constitution of the twin polymerase of DNA polymerase III holoenzyme. J Biol Chem, 266, 19833–19841. | PubMed | ChemPort |

Stukenberg PT, Studwell-Vaughan PS and O'Donnell M (1991) Mechanism of the sliding -clamp of DNA polymerase III holoenzyme. J Biol Chem, 266, 11328–11334. | PubMed | ISI | ChemPort |

Stukenberg PT, Turner J and O'Donnell M (1994) An explanation for lagging strand replication: polymerase hopping among DNA sliding clamps. Cell, 78, 877–887. | Article | PubMed | ISI | ChemPort |

Sugimoto K, Shimomura T, Hashimoto K, Araki H, Sugino A and Matsumoto K (1996) Rfc5, a small subunit of the replication factor C complex, couples DNA replication and mitosis in budding yeast. Proc Natl Acad Sci USA, 93, 7048–7052. | Article | PubMed | ChemPort |

Tsuchihashi Z and Kornberg A (1989) ATP interactions of the and subunits of DNA polymerase III holoenzyme of Escherichia coli. J Biol Chem, 264, 17790–17795. | PubMed | ISI | ChemPort |

Turner J and O'Donnell M (1995) Cycling of E.coli DNA polymerase III from one sliding clamp to another: a model for the lagging strand. Methods Enzymol, 262, 442–449. | PubMed | ISI | ChemPort |

Wickner S (1976) Mechanism of DNA elongation catalyzed by Escherichia coli DNA polymerase III, DnaZ protein and DNA elongation factors I and III. Proc Natl Acad Sci USA, 73, 3511–3515. | PubMed | ChemPort |

Xiao H, Dong Z and O'Donnell M (1993) DNA polymerase III accessory proteins. IV. Characterization of and . J Biol Chem, 268, 11779–11784. | PubMed | ISI | ChemPort |

Xiao H, Naktinis N and O'Donnell M (1995) Assembly of a chromosomal replication machine: two DNA polymerases, a clamp loader and sliding clamps in one holoenzyme particle. IV. ATP-binding site mutants identify the clamp loader. J Biol Chem, 270, 13378–13383. | Article | PubMed | ISI | ChemPort |

Yao N, Turner J, Kelman Z, Stukenberg PT, Dean F, Shechter D, Pan ZQ, Hurwitz J and O'Donnell M (1996) Cycling of DNA sliding clamps of human, E.coli and T4 replicases. Genes Cells, 1, 101–113. | Article | PubMed | ISI | ChemPort |

Yuzhakov A, Turner J and O'Donnell M (1996) Replisome assembly reveals the basis for asymmetric function in leading and lagging strand replication. Cell, 86, 877–886. | Article | PubMed | ISI | ChemPort |

Yuzhakov A, Kelman Z and 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. | Article | PubMed | ISI | ChemPort |

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising