Among the numerous molecular machines involved in the process of DNA replication are the ring-shaped sliding clamp and the clamp loader. Intriguing structural details of their interaction are now revealed.
The double-helical structure of DNA is an icon of our time, appearing almost daily as a backdrop to news stories about medical advances and in myriad sci-fi movies. The molecule itself consists of two inter-wound strands, the backbones of which are composed of alternating phosphate and sugar groups. Extending from the backbone into the heart of the double helix are the bases — guanine, adenine, thymine and cytosine — whose order encodes the information stored by DNA. The bases on one strand must pair precisely with those on the opposite strand, adenine with thymine and guanine with cyosine1, and this provides a simple and elegant mechanism for copying this genetic material. The double helix simply unwinds and unzips, with both strands then being used as templates for the enzyme DNA polymerase, together with a large assembly of accessory factors, to make two daughter DNA molecules. Papers published on page 724 of this issue2 and in Nature Structural and Molecular Biology3 now add to our understanding of a crucial molecular contributor to this process.
Because of a chemical asymmetry in the arrangement of DNA, one daughter strand, termed the leading strand, is synthesized continuously during DNA replication. The other, the lagging strand, is made in short segments (called Okazaki fragments) that are later joined together. Central to this process is the sliding clamp, a ring-shaped, multi-subunit molecule that encircles the DNA and binds to the DNA polymerase. As the name suggests, the sliding clamp can slide along DNA, and so provides a mechanism for tethering the DNA polymerase to the template. On the leading strand this is extremely important, because the polymerase might have to continue synthesizing DNA for more than a million bases. But the clamp is also important on the lagging strand, because several of the factors involved in processing the Okazaki fragments use it as a scaffold for their assembly4.
The ring-shaped nature of the sliding clamp presents a simple topological problem: how is DNA introduced into the hole in the clamp's centre? This is the job of a complex molecular engine termed the clamp loader, which must somehow open the sliding clamp, pass DNA into the ring and then reseal it. These requirements are a common theme in several DNA-based processes. Higher organisms, for instance, also possess a DNA-repair-specific sliding clamp and clamp loader, and it is likely that similarities will be found in earlier stages in the replication process, when the DNA-unwinding helicase proteins are loaded.
Bowman et al.2 and Miyata et al.3 now offer new insight into the clamp-loading reaction, through structural analyses of the complexes formed between a sliding clamp and its cognate clamp loader in yeast2 and in the archaeon Pyrococcus furiosus3. The clamp loader in question is named replication factor-C (RFC). In yeast, RFC has one large subunit and four smaller components with related sequences; in archaea (microbes that have a simpler replication machinery than yeast), it has one large and four identical small subunits. The sliding clamp, meanwhile, is named proliferating cell nuclear antigen (PCNA) and consists of three identical subunits.
It has been known for many years that RFC can, in the presence of the high-energy cofactor adenosine triphosphate (ATP), mediate the loading of PCNA onto DNA5,6. The RFC clamp loader then hydrolyses the ATP, resulting in the clamp being handed off to DNA polymerase. Taking advantage of the fact that a stable complex can be formed between DNA, PCNA and RFC in the presence of a poorly hydrolysable ATP analogue, Miyata et al.3 were able to purify such a complex from Pyrococcus and subject it to electron microscopy followed by three-dimensional reconstruction.
The resulting image is intriguing. The structure resembles a horseshoe stacked on a doughnut (Fig. 1), with the doughnut corresponding to the clamp, PCNA, and the horseshoe to the clamp loader, RFC. It seems as though the large subunit of RFC contacts one subunit of PCNA, and two or three of the RFC small subunits contact the remaining two PCNA subunits. DNA passes through the PCNA ring and emerges through the open channel in the horseshoe of RFC. Thus, this structure presumably corresponds to the loaded complex, just before PCNA binds DNA polymerase.
Meanwhile, Bowman et al.2 describe a stunning, high-resolution atomic structure of yeast RFC in complex with PCNA, again in the presence of a poorly hydrolysable ATP analogue. In this case, DNA was not present in the crystallized complex itself but was incorporated in the three-dimensional reconstruction. As in the archaeal image, RFC sits on top of PCNA, with the large subunit contacting one PCNA subunit. But there are some important differences in the interactions of the RFC small subunits. Most significantly, instead of lying flat on PCNA, contacting the remaining two PCNA subunits, the RFC small subunits spiral up and away from the plane of the PCNA ring and contact only one further PCNA subunit (Fig. 1). Remarkably, the pitch of the spiral produced by RFC matches the geometry of the DNA double helix, and the authors present a persuasive model for how the subunits might contact the DNA so as to direct loading.
So why is there this difference? Setting aside trivial explanations of species differences or methods of sample preparation, could it be that these two complexes represent distinct stages in the clamp-loading process? Might the archaeal structure represent the initial complex and the yeast one a later stage? If so, might the differences arise by virtue of an isomerization of RFC that occurs before ATP is hydrolysed and PCNA is handed over to DNA polymerase? And could the fact that one subunit of PCNA is exposed in the yeast structure therefore be indicative of the route that the DNA polymerase takes to access the ring? Finally, it remains to be seen how RFC actually opens the PCNA ring. Might the spiral of RFC subunits indicate that this clamp loader pries PCNA open by pulling a subunit or subunits out of the plane of the ring, popping it open like a lock washer? Or does RFC wrench PCNA open within the plane of the ring, creating a horseshoe structure that mirrors its own?
These structures provide exciting snapshots of the manner in which RFC and PCNA interact. Although small differences have been identified, conserved themes in this essential reaction are coming to the fore. With continuing efforts to study other stages of clamp loading, a full molecular view of the process seems to be on the horizon.
Watson, J. D. H. & Crick, F. H. C. Nature 171, 737–738 (1953).
Bowman, G. D., O'Donnell, M. & Kuriyan, J. Nature 429, 724–730 (2004).
Miyata, T. et al. Nature Struct. Mol. Biol. doi:10.1038/nsmb788 (2004).
Warbrick, E. BioEssays 22, 997–1006 (2000).
Tsurimoto, T. & Stillman, B. Proc. Natl Acad. Sci. USA 87, 1023–1027 (1990).
Gomes, X. V. & Burgers, P. M. J. Biol. Chem. 276, 34768–34775 (2001).
About this article
Proceedings of the National Academy of Sciences (2005)