We have been studying the functions of PCNA (proliferating-cell nuclear antigen) for the assembly and reassembly of the replisome during replication fork progression. We have identified the functional interactions between PCNA and several proteins involved in DNA replication and repair from Pyrococcus furiosus. We recently reported that the activity of UDG (uracil–DNA glycosylase) in P. furiosus (PfuUDG) is stimulated by PCNA (PfuPCNA) in vitro, and identified an atypical PCNA-binding site, AKTLF, in the PfuUDG protein. To understand further the function of the complex in the BER (base excision repair) process, we investigated the AP (apurinic/apyrimidinic) endonuclease, which can process the BER pathway after uracil removal by UDG. Interestingly, one candidate ORF (open reading frame) for the AP endonuclease was found in the operon containing the gene encoding UDG in the P. furiosus genome. However, this ORF did not exhibit any activity. Instead, we identified the AP endonuclease activity from the other candidate gene products, and designated the protein as PfuAP. We discovered a physical interaction between PfuAP and PfuPCNA, suggesting the formation of a BER complex in one of the repair systems in P. furiosus.
To understand the molecular machinery for DNA replication and repair, the assembly and reassembly mechanisms of the replisome and several repair complexes must be investigated. Among the many proteins involved in these processes, PCNA (proliferating-cell nuclear antigen) is a key molecule, which may act as a conductor on the DNA, where it controls other proteins for their appropriate functions with the DNA . It is widely known that many replicative and repair enzymes interact with PCNA via a consensus sequence called the PIP (PCNA-interacting protein) box . The PIP box consists of the sequence QXX(L/I/M)XX(F/Y)(F/Y). In many cases, the hydrophobicity of the aromatic side chain is important for the interaction with PCNA-binding proteins [3–5]. We have been studying PCNA and PIPs in archaea, especially the hyperthermophilic archaeon Pyrococcus furiosus. The PIP box is generally located at the very terminus of the peptide chain (either the N-terminus or C-terminus) of the PCNA-binding proteins. We have identified functional PIP boxes in DNA polymerases B and D , RFC (replication factor C) , Hjc (Holliday junction resolvase) (S. Matsumiya, S. Ishino, S. Kiyonari, T. Nishino, K. Motikawa and Y. Ishino, unpublished work) and Hjm (Holliday junction migration) helicase  from Pyrococcus furiosus. Interestingly, no PIP box-like sequence was found in either the N- or C-terminal region of the DNA ligase protein (PfuLig). Therefore we investigated whether DNA ligase and PCNA interact with each other, and identified the PCNA-binding site in the middle of the DBD (DNA-binding domain) of the DNA ligase . This position is unusual, but our crystal structure of PfuLig  revealed that the interacting site is located in a region that protrudes from the surface of the molecule, so it can easily contact PCNA. The PCNA-binding site of the archaeal UDG (uracil–DNA glycosylase), which initiates the BER (base excision repair) pathway by catalysing the hydrolysis of the N-glycosyl bonds linking the uracil base to the sugar–phosphate backbone in DNA , is also quite interesting. In the present review, we focus on the UDG–PCNA interactions in the BER complex in P. furiosus.
UDG–PCNA interaction in P. furiosus
The UDG enzymes are currently classified into five families, and the thermostable family 4 enzymes are found in some thermophilic bacteria and archaea. Physical interactions between PCNA and a family 4 UDG from Pyrobaculum aerophilum (PaeUDG) and Sulfolobus solfataricus (SsoUDG) at the C-terminal PIP box-like motif have been reported [12,13]. The PCNA-binding motifs found in these crenarchaeal UDGs are not conserved in the archaeal UDG sequences, and, especially, several euryarchaeal UDGs lack the C-terminal region containing the PCNA-binding motif of the crenarchaeal UDGs . In addition, although direct interactions between PCNA and UDG from the crenarchaeal organisms described above were well documented, their functional significance was not fully understood. We therefore investigated the interaction between P. furiosus UDG (PfuUDG) and PCNA (PfuPCNA) .
We detected the physical interaction between PfuUDG and PfuPCNA using an SPR (surface plasmon resonance) analysis with a calculated Kd of 220 nM, although PfuUDG lacks the PIP box-like sequence at the C-terminus. Furthermore, we added the PfuPCNA to the in vitro uracil N-glycosylase assay, using a uracil-containing oligodeoxynucleotide duplex, and found that the glycosylase activity was stimulated by PfuPCNA in a concentration-dependent manner. A mutant PfuPCNA, which cannot form a stable trimeric ring, did not stimulate the reaction. From these results, we were eager to identify the PCNA-interacting site in PfuUDG. In the case of PfuLig, a shorter version of the PIP box, QKSFF, is critical for the physical and functional interaction with PfuPCNA . Moreover, a cluster of basic amino acids adjacent to the binding motif seemed to be important for the archaeal DNA ligase–PCNA interactions [9,15]. On the basis of these findings, we searched visually for a similar PCNA-binding motif in PfuUDG, and discovered the amino acid sequence, A152KTLF156. We then demonstrated that the predicted site is actually involved in the PCNA binding by site-specific mutagenesis. To confirm the formation of the PfuUDG–PfuPCNA complex in the cells, immunoprecipitation was performed, using a P. furiosus cell extract and antibodies against these two proteins. As shown in Figure 1, the PfuUDG band was clearly detected in the fraction precipitated with the anti-PfuPCNA antibody.
Interaction between UDG and PCNA in P. furiosus
The PCNA-binding motif found in PfuUDG was not located at the extreme C-terminus, unlike that in the crenarchaeal UDGs, but within the peptide chain. This PCNA-binding site should be located on the surface of the three-dimensional structure of PfuUDG, as in the case of PfuLig. To examine the position of the PCNA-binding site in PfuUDG, the crystal structure of the Thermus thermophilus UDG, belonging to family 5 but sharing sequence similarity with that of PfuUDG, complexed with duplex DNA  was utilized for structural model building. As shown in Figure 2(A), we obtained a PfuUDG–PfuPCNA–dsDNA (double-stranded DNA) complex structure model, in which the PCNA-binding site is precisely located near the IDCL (interdomain-connecting loop), a common place of interactions for PCNA-binding proteins.
The archaeal BER complex
The BER complex in the P. furiosus
The presence of the UNG2 (one of the two types of UDG in human cells, which is localized in the nucleus) -associated repair complex at replication foci has been shown in human cells [17–19]. The complex contains APE1 (apurinic/apyrimidinic endonuclease 1), DNA polymerases α, β, δ and ε, DNA ligase I and XRCC1 (X-ray repair complementing defective repair in Chinese hamster cells 1), in addition to UDG and PCNA . The UDG enzyme generates an apyrimidinic site, which is potentially mutagenic and cytotoxic because of DNA replication inhibition, and it therefore has to be repaired quickly. AP (apurinic/apyrimidinic) endonuclease is the major contributor to the subsequent repair system, after UDG. We investigated the AP endonuclease in P. furiosus. AP endonucleases are mainly classified into two families, on the basis of the amino acid sequence similarity. These are the Exo III family and the Endo IV family, represented by Escherichia coli exonuclease III and endonuclease IV respectively . Three Exo III family enzymes (Archaeoglobus fulgidus , Methanothermobacter thermoautotrophicus  and Thermoplasma volcanium ) and two Endo IV family enzymes (M. thermoautotrophicus  and Pyrobaculum aerophilum ) have been characterized to date in the archaeal domain. We searched for genes encoding amino acid sequences homologous with these enzymes in the total genome sequence database of P. furiosus, and found two candidates for the Endo IV family enzymes, but none for the Exo III family enzymes. The amino acid sequence similarities of the two candidate ORFs (open reading frames), PF0258 and PF1383, with E. coli Endo IV are 13.5 and 13.1% respectively. It was especially interesting that the gene encoding PF1383 is located at a position very close to that for UDG (PF1385), and these two genes seem to form an operon with two other genes. We cloned these genes independently, and expressed them in E. coli. Unexpectedly, only PF0258, which is not encoded by the gene near the udg gene, exhibited endonuclease activity to cleave the phosphodiester backbone immediately 5′ of the abasic site. We designated the gene product as PfuAP, and are currently characterizing it in vitro. Sequence analyses revealed that some archaea have both the Exo III and Endo IV family enzymes, and some have only the Endo IV family enzymes. It is interesting to analyse in more detail the distribution of AP endonucleases in archaeal genomes from an evolutionary viewpoint. Our most interesting issue is the interaction of PfuAP with PfuPCNA. Preliminary experiments have shown physical and functional interactions between the two proteins (S. Kiyonari, S. Tahara, S. Ishino and Y. Ishino, unpublished work).
Is there a ‘uracilosome’ in P. furiosus?
The rate of cytosine deamination to uracil is accelerated at high temperature , and it is therefore possible that a uniquely efficient mechanism for uracil excision repair may exist in hyperthermophilic archaea. We are currently interested in determining whether PfuUDG and PfuAP can bind simultaneously to the same PfuPCNA ring, so that the sequential cleavage of the uracil–glycosyl bond and the diester bond progress efficiently. Our structural model of PfuUDG–PfuPCNA–dsDNA (Figure 2A) shows that the dsDNA, piercing the hole of the PCNA ring, is tilted by 15° from the 3-fold axis of PfuPCNA. A recent report describing the crystal structure of the E. coli β-clamp–DNA complex showed that the DNA strand passes through the β-clamp ring with an orientation tilted by 22°, and the authors proposed that the DNA tilt actively contributes to the switching mechanism of Pol III (normal DNA synthesis) and Pol IV (translesion synthesis), which are bound to the same clamp ring . It would be interesting if the PfuUDG switches to PfuAP on the same PfuPCNA ring, in the same manner (Figure 2B).
It is now well known that the family B DNA polymerases from the hyperthermophilic archaea, including P. furiosus PolBI, specifically recognize uracil bases in the template strand and stall DNA polymerization. This property of the archaeal DNA polymerases has been implicated as an intrinsic activity for the removal of uracil bases [28–30]. We investigated whether PolBI and PfuUDG interact directly with each other by SPR analysis. The sensorgram showed no direct interaction between the two proteins, suggesting, at least, that a stable complex of PolB and UDG is not formed in vitro . However, it is possible that PolBI and UDG bind to the same PCNA ring to switch at the uracil site.
In addition to UDG and PolBI, P. furiosus has dUTPase, which probably contributes to precise DNA replication by preventing dUTP incorporation in the cells . Some functional association of the three uracil sensors, PolB, UDG and dUTPase, was proposed as a complex named ‘uracilosome’ for the efficient escape from uracil under hyperthermophilic conditions . It would be interesting to isolate the complex that protects the genome from uracil from the cell extract, to understand this efficient repair system that is probably unique to the hyperthermophilic archaea.
This work was supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (to Y.I.)
Molecular Biology of Archaea: Biochemical Society Focused Meeting held at University of St Andrews, U.K., 19–21 August 2008. Organized and Edited by Stephen Bell (Oxford, U.K.) and Malcolm White (St Andrews, U.K.).
Owing to exceptional unforeseen circumstances, this speaker was unable to give this presentation at the meeting. This paper is included in the interests of completeness of the session.