REVIEW ARTICLE The multi-replication protein A (RPA) system – a new perspective Kengo Sakaguchi, Toyotaka Ishibashi*, Yukinobu Uchiyama and Kazuki Iwabata Department of Applied Biological Science, Tokyo University of Science, Chiba, Japan Replication protein A (RPA) is a single-stranded DNA (ssDNA)-binding protein complex comprising a hetero- trimeric combination of a large (70 kDa), middle (32 kDa) and small (14 kDa) subunit [1,2]. Function- ally, RPA corresponds to an alternative form of a bacterial ssDNA-binding protein (SSB). Until 2005, only one copy of the RPA complex was thought to be present in eukaryotes [1–9]. Indeed, preliminary analysis of the genomes of mammals and yeast indicated that they encoded a single copy of each subunit of the RPA complex [1,2]. However, we recently found that higher plants have at least three different species of complex (types A, B and C), each displaying a different biological function [10–12]. Orig- inally, we intended to investigate the plant repair system [13–43], but during the course of this study we Keywords convergent evolution; DNA polymerases; eukaryotic DNA metabolism; meiotic pairing and recombination; multi-RPA system; O. sativa and A. thaliana; paralog ⁄ ortholog/ analog/heterolog; Rad51 ⁄ DMC1 ⁄ Lim15; replication protein A; RPA subunits (70, 32 and 14 kDa) Correspondence K. Sakaguchi, Department of Applied Biological Science, Faculty of Science and Technology, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278 8510, Japan Fax: +81 471 23 9767 Tel: +81 471 24 1501 (ext. 3409) E-mail: kengo@rs.noda.tus.ac.jp *Present address Department of Biochemistry and Microbiology, University of Victoria, Victoria, Canada (Received 11 September 2008, revised 26 November 2008, accepted 5 December 2008) doi:10.1111/j.1742-4658.2008.06841.x Replication protein A (RPA) complex has been shown, using both in vivo and in vitro approaches, to be required for most aspects of eukaryotic DNA metabolism: replication, repair, telomere maintenance and homolo- gous recombination. Here, we review recent data concerning the function and biological importance of the multi-RPA complex. There are distinct complexes of RPA found in the biological kingdoms, although for a long time only one type of RPA complex was believed to be present in eukary- otes. Each complex probably serves a different role. In higher plants, three distinct large and medium subunits are present, but only one species of the smallest subunit. Each of these protein subunits forms stable complexes with their respective partners. They are paralogs as complex. Humans pos- sess two paralogs and one analog of RPA. The multi-RPA system can be regarded as universal in eukaryotes. Among eukaryotic kingdoms, para- logs, orthologs, analogs and heterologs of many DNA synthesis-related factors, including RPA, are ubiquitous. Convergent evolution seems to be ubiquitous in these processes. Using recent findings, we review the compo- sition and biological functions of RPA complexes. Abbreviations ATR, ataxia telangiectasia mutated and Rad3-related; dsDNA, double-stranded DNA; MMS, methyl methanesulfonate; NER, nucleotide excision repair; PCNA, proliferating cell nuclear antigen; pol a, DNA polymerase a; RPA, replication protein A; SC, synaptinemal complex; SSB, single-stranded DNA-binding protein; ssDNA, single-stranded DNA. FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS 943 serendipitously discovered the involvement of RPA [10–12]. Interestingly, RPAs are not necessarily com- pletely independent complexes. Only one copy of the small subunit was found, whereas there were three sets of the large and middle subunits [10–12]. The mode of action of these RPA complexes seems to be universal, at least in Plantae. Each RPA complex must be inde- pendently related to various DNA synthetic events within the plant. Because DNA replication and repair are generally very similar between animals and plants [13,44–66], the role of the RPA complex should be reconsidered in the light of this new finding. Therefore, we retrospectively searched reports about screening for RPA homologs in animals and fungi. Humans carry two homologs of the middle subunit (HsRPA2 and HsRPA4) [67–69]. Moreover, Richard et al. recently reported that the two human SSB homologs (hSSB1 and hSSB2) possess a domain organization that is closer to archaeal SSB than to RPA [70]. Although the genetic and biochemical characteristics of hSSB1 are totally different from those of human RPA, both are critical for genomic stability [70]. Thus, like Plantae, the human DNA repair enzymes also function as a multiple system. Furthermore, the multi-RPA or SSB– RPA mixed system is presumably universal in eukary- otes. Here, in the light of these recent discoveries, we review the function and structure of the RPA com- plexes. There are many reports in the literature concerning the role of RPAs. RPA is ubiquitous and essential for a wide variety of DNA metabolic processes, including DNA replication, repair and recombination [1]. In par- ticular, RPA is required for cross-over during meiosis [71–74]. According to a recent report [75], the large and middle subunits of human RPA may act as an independent prognostic indicator of colon cancer, as well as therapeutic targets for regulation by tumor sup- pressors involved in the control of cell proliferation. Thus, despite the previous studies on RPA, there are many new areas of research involving this complex that still need to be addressed. History of RPA studies We begin this review by summarizing studies that first identified RPA as a factor necessary for SV40 replica- tion in vitro [76–79]. RPA is required for activation of the pre-replication complex to form the initiation com- plex, and for the ordered loading of essential initiator functions, such as DNA polymerase a–primase (pol a) complex, to the origins of replication [76–79]. The gen- eral role of RPA has been studied in great detail in mammals and yeasts [1,2]. It was originally thought that the RPA complex was evolutionarily conserved throughout eukaryotes and that the function is funda- mental irrespective of DNA synthesis. Many data were obtained on the assumption that there is just one RPA copy. RPA accumulates along stretches of ssDNA gen- erated during DNA replication and repair (Fig. 1A) [1,5–8,79–87]. RPA also plays an essential role in DNA repair and is required for nucleotide excision A B Fig. 1. (A) RPA in the DNA replication. (B) The role of RPA in NER. The multi-replication protein A system K. Sakaguchi et al. 944 FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS repair (NER) [88–90]. During strand elongation in DNA replication ⁄ repair, RPA stimulates the action of DNA polymerases such as pol a, pol d, pol e, pol k and pol j [5–8,80,81,85–87]. Conversely, pol f is not under the influence of RPA, suggesting that RPA- dependent ssDNA stretching is not always essential for DNA polymerization [88]. RPA interacts with XPA at sites of DNA damage, stimulating XPA–DNA contact and recruiting the incision proteins ERCC1 ⁄ XPF and XPG to the damaged site (Fig. 1B) [89–91]. These pro- cesses include damage detection and signaling, tran- scriptional responses, DNA damage checkpoints and apoptosis [4,7]. RPA is known to interact specifically with numerous transcription, replication and repair proteins including T antigen, the tumor suppressor p53, the transcription factors Gal4 and VP16, DDB, uracil DNA glycosylase, recombinases and the DNA helicases, Bloom’s and Werner’s proteins. RPA is also a checkpoint protein that has been iden- tified by the generation of a mutant in the large sub- unit in yeast [92]. In addition, RPA was found to be necessary for the removal of oxidized base lesions from genomic DNA in long-patch base excision repair [93,94]. RPA also interacts with Rad51 and Rad52, thereby playing a role in initiating homologous recom- bination events [95–111]. In the repair of double-strand breaks by homologous recombination in Saccharomy- ces cerevisiae, RPA stimulates DNA strand exchange by Rad51 protein, provided that RPA is added to a pre-existing complex of Rad51 protein and ssDNA. RPA is also implicated in forming the meiotic recom- bination nodules [112–118]. Furthermore, RPA has a specific interaction with the tumor suppressor p53 [119–121] and promotes DNA binding and chromatin association of ataxia telangiectasia mutated and Rad3- related (ATR) in vitro via ATR interacting protein [122]. RPA is also required to recruit and activate Rad17 complexes for checkpoint signaling in vivo [123]. Thus, the functions of RPA are surprisingly ambiguous. Namely, RPA functions in a wide range of systems from DNA replication to DNA damage and stress responses (biochemical and cell biological) as well as cross-over in meiosis [1,2]. It is thought that the major interaction between RPA and DNA occurs through the RPA70kDa sub- unit, and the role of the RPA32kDa and RPA14kDa subunits is supplementary [124]. Indeed, RPA70kDa is the major subunit of the complex having four ssDNA- binding domains in the middle of the subunit. By contrast, RPA32kDa and RPA14kDa each possess a single DNA-binding domain, displaying only weak binding affinity [2,125]. The contact surfaces in RPA have been elucidated for several of its binding part- ners. The results of these studies suggest that proteins from distinct processing pathways may use a small number of common sites to bind RPA and remodel the mode of DNA binding [124]. The RPA32kDa subunit is phosphorylated during progression of the cell cycle and in response to a wide variety of DNA-damaging agents, such as ionizing radiation, UV and camptothecin [120,126–128]. RPA phosphorylation stimulated by DNA damage promotes DNA binding and chromatin association of ATR in vitro via ATR interacting protein [83,122,129]. RPA is also required for recruitment and activation of the Rad17 complexes during checkpoint signaling in vivo. RPA may function in the sensing of DNA damage [111]. In budding yeast, the middle subunit (32 kDa) becomes phosphorylated in reactions that require the Mec1 protein kinase, a central checkpoint regulator and homolog of human ATR [71–74]. However, the meiosis-specific protein kinase Ime2 is required for normal meiotic progression [130]. A natural target of Ime2 activity is also the middle subunit of RPA [130]. Ime2-dependent RPA phosphorylation first occurs early in meiosis. The middle subunit is not supplemen- tary, but is a signal acceptor for sensing various struc- turally specific DNA sites. Furthermore, RPA32kDa is reportedly related to viral DNA replication [124,131]. There is almost no information concerning the molecular role of the RPA14kDa subunit. It is known that RPA14kDa contains one weak DNA-binding domain, which may slightly modify the mode of DNA binding of RPA. Consequently, it was generally believed that the major roles of RPA had been elucidated. However, at this stage, it was not known that RPA represented more than one molecular species. Thus, most research- ers did not consider the possibility of orthologs, para- logs, analogs and heterologs of the RPA complex. Multi-RPA systems In contrast to the intensive studies of RPA in mam- mals and yeasts, until 2001 little was known about this protein in plants. Plants are affected by various envi- ronmental stress factors. For example, DNA in plants is continuously damaged by UV irradiation from sun- light. UV is known to induce DNA damage [13], although plants generally have a higher tolerance for UV than animals. Field-grown crops such as wheat are also known to suffer continuous UV-induced DNA damage. Furthermore, the formation of reactive oxygen species in cells due to UV irradiation, biotic stresses and secondary metabolism, causes cellular components, including DNA, to be oxidized and there- K. Sakaguchi et al. The multi-replication protein A system FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS 945 fore susceptible to oxidative modification. In addition, the fidelity and integrity of DNA are constantly chal- lenged by chemical substances in the environment, ion- izing radiation and errors that occur during DNA replication or proofreading. This accumulated damage blocks a number of critical processes, such as tran- scription and replication, and can eventually cause cell death. Thus, UV damage can reduce the growth and yield of plant crops. Indeed, there is no difference between the abilities of animals and plants to remove damaged DNA [13]. Plants have evolved several DNA- repair pathways [13]. Whereas previous studies on DNA repair have focused mostly on animals and yeast cells, recent analyses of UV tolerance and DNA repair have addressed the responses of plants to environmen- tal factors and the mechanisms of stress resistance in plants [13]. An additional basis for molecular analyses has been provided by the completion of genome- sequencing projects in model plants such as rice and Arabidopsis. Completed genome sequences allow the identification of entire gene groups related to DNA repair in higher plants. In order to better understand the mechanisms of DNA protection and plant DNA repair systems, we attempted to isolate the gene encod- ing plant RPA. Surprisingly, analysis of rice revealed a new type of RPA complex gene [10–12]. In 1997, an ortholog of the RPA70kDa subunit (Os- RPA1) was isolated from deepwater rice (Oryza sativa L. cv. Pin Gaew 56), and its expression was induced by gibberellin [132]. To use the OsRPA1 protein for plant DNA replication studies, we aimed to clone the cDNA and obtain the recombinant protein from rice (O. sativa L. cv. Nipponbare). Although we failed to clone the OsRPA1 cDNA, we unexpectedly obtained cDNA of the RPA70kDa subunit alternative. The new alternative gene differed greatly from OsRPA1, having closer homology with its counterpart in Arabdop- sis thaliana reported in the database [10]. We found that A. thaliana also has a homolog of OsRPA1, sug- gesting that two different RPA types are universally present in seed plants [10]. Rice has two different types of RPA70kDa subunit, renamed OsRPA70a (newly found) and OsRPA70b (OsRPA1), respectively [10]. We discovered their homologs in A. thaliana, and described the substantial properties of the T-DNA insertion lines [11]. Transcripts of OsRPA70a are expressed in proliferating tissues, such as root tips and young leaves that contain meristem, but also more weakly in the mature leaves, whereas OsRPA70b is expressed mostly in proliferating tissues [10]. The existence of these genes gives rise to an intrigu- ing evolutionary question. Why do mammals and yeast have only one copy of the gene for the RPA70kDa subunit in their genome? Furthermore, is only the larg- est subunit of the RPA complex duplicated in plant, and what are the roles of the two RPA types? Interest- ingly, when the RPA70a subunit lacked the T-DNA insertion or RNA interference (RNAi), the line could be viable [10–12]. The surviving mutant was morpho- logically normal except for hypersensitivity towards some mutagens, such as UV and methyl methanesulfo- nate (MMS) [10–12]. Plants are naturally exposed to UV for much longer than animals or yeast [133–135] and depend on sunlight for their development. Because seed plants synthesize DNA under relatively high levels of UV irradiation, the RPA system might be more complicated in plants than in animals. Therefore, we attempted to screen for rice RPA genes in the genome (O. sativa L. cv. Nipponbare). We found three different genes encoding the largest (RPA70kDa) and middle subunits (RPA32kDa), but only one gene encoding the smallest (RPA14kDa) [12]. Each OsRPA70s and OsRPA32s gene was not a pseudogene or redundant gene. We designated the subunits from rice as OsRPA70a, OsRPA70b, OsRPA70c, OsRPA32-1, OsRPA32-2, OsRPA32-3 and OsRPA14 [12]. The RPA70bsubunit is the ubiquitous RPA70 subunit found in all eukaryotes [10]. The various subunits do not ran- domly associate with other subunits, but form a distinct complex. Three different RPA complexes (A, B or C type) were composed of these subunits in vivo. Types A, B and C were OsRPA70a–OsRPA32-2–OsRPA14, OsRPA70b–OsRPA32-1–OsRPA14 and OsRPA70c– OsRPA32-3–OsRPA14, respectively [11,12]. Only the smallest subunit is common to all the complexes. Because the system was also present in A. thaliana [11,12], these properties may be universal in higher plants. In conclusion, higher plants have a multi-RPA system [11,12]. The RPA complexes are spatially segregated in plants. Type A is localized to the chloroplast, whereas types B and C are found in the nuclear region [11]. In human and yeast cells, the middle subunit exists in the nucleus and cytoplasm, whereas the large subunit is present only in the nucleus [11]. The RPA32kDa sub- units probably exist as each protein alone (OsRPA32- 1, OsRPA32-2, OsRPA32-3 or OsRPA14) or as free heterodimer complexes such as OsRPA32-1–OsRPA14, OsRPA32-2–OsRPA14 and OsRPA32-3–OsRPA14 [11,12]. In rice, co-regulation of OsRPA70b and OsRPA32-1 during the cell cycle, and regulation of OsRPA32-1 in response to UV has been reported [43]. RPA70kDa has been reported to be unstable when not in a com- plex. Because expression of OsRPA70a was observed at both the mRNA and protein levels, we suggest that The multi-replication protein A system K. Sakaguchi et al. 946 FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS the rice genome contains another protein, distinct from OsRPA32-2 that might form a stable complex with OsRPA70a. As described earlier, the RPA32kDa sub- unit is phosphorylated in response to cell-cycle phase transitions and a wide variety of DNA-damaging agents, suggesting that RPA activities are regulated by the extent of phosphorylation [120,126–128]. Rice had three different RPA32kDa subunits. This infers the existence of independent phosphorylation systems that control each type of RPA complex. Does the phos- phorylation occur on the same RPA complex? Are such phenomena limited in the RPA system? Drosophila has two paralogs of proliferating cell nuclear antigen (PCNA) and a ‘heterolog’ (Rad9– Rad1–Hus1) [65,136,137]. Moreover, the fungus Coprinus cinereus generates two different PCNAs by alternative splicing, although there is only a single copy of the gene in the genome [138]. Even the plural- izing recipe of PCNA is also phylogenetically diversi- fied. The roles of PCNA are probably diversified, and a division of labor occurs [65]. Like RadA and hSSB, we also found another FEN-1-like analog, SEND-1 and GEN [25,63,66]. All are transcribed and translated and therefore do not represent pseudogenes. Knock- down of one of their genes in the same category seems to lead to lethality, although there is little published data on this subject. The diversification must be closely related to the point at which biochemical control sys- tems divide [65]. Similar considerations probably apply to the multi-RPA system. Phylogenetic aspects of multi-RPA systems Sophisticated studies are required to verify whether a specific subunit (OsRPA32-1, OsRPA32-2 or OsRPA32-3) is responsible for phosphorylational control. Furthermore, which RPA complex corresponds to the RPA found in mammals and yeast? Are no other RPA types present in animals and yeasts? Whether mammals and yeasts evolved a multi-RPA system, which was subsequently lost over evolutionary time is so far unclear. We have investigated the plant multi-RPA system in terms of phylogenetics. Two large RPA subunits, RPA70 and RPA32, and a small subunit, RPA14, are relatively well conserved among eukaryotes (Fig. 2A). The deduced amino acid sequence among OsRPA70a, OsRPA70b and OsRPA70c showed low identity levels ( 50%) between them [12]. Similarly, the deduced amino acid sequence among OsRPA32-1, OsRPA32-2 and OsRPA32-3 was compared; each type also displayed low identity levels [12]. In the system, the sequence homologies among the OsRPA70kDa subunits and among the OsRPA32kDa subunits were low [12]. The B type complex was thought to be ubiquitous in eukaryotes [12]. RPA70kDa has two RPA ssDNA-binding domains, DBD-A and DBD-B for binding ssDNA, and a third, DBD-C, which displays only weak ssDNA-binding activity (Fig. 2B). RPA70kDa also contains the DBD- F domain, which has been shown to interact with multiple proteins and to interact weakly with DNA (Fig. 2B). The primary amino acid sequences of DBD-A, DBD-B, DBD-C and DBD-F domains are very similar [12]. RPA32kDa has only a single ssDNA- binding domain (DBD-D) [12]. Furthermore, all the domains have high levels of sequence homology with their counterparts in human and yeast RPAs [12]. The DBD-E domain is in the RPA14kDa subunit, and is also highly conserved [12]. In yeast, RPA1 (largest subunit) can only bind to the RPA2 ⁄ 3 dimer (middle and smallest subunit dimer). The DBD-C and DBD-D regions of rice are quite similar to the DBD-C and DBD-D regions of S. cerevisiae [139], but OsRPA14 has only low simi- larity to RPA3. This sequence divergence may account for the differences in binding observed between the yeast and rice proteins. Rice DBD-A and DBD-B domains are more conserved than DBD-C and DBD-F, implying that the primary function of OsRPA70a and OsRPA70b is to bind DNA, and that this function has been conserved during evolution, even though the secondary functions of these proteins may have diverged. Based on this analysis the B type complex corresponds to the mammalian and yeast RPA. In plant, human and yeast, the domains of DBD-A and DBD-B are more homologous than those of DBD-C and DBD-F, and the biochemical characteris- tics are common among OsRPA70a, OsRPA70b and OsRPA70c. It is well established that the RPA70kDa subunit accumulates along stretches of ssDNA gener- ated by stalled replication forks and ⁄ or DNA damage [1,82–84]. In the RPA70kDa subunit, DBD-A and DBD-B possess the strongest ssDNA-binding activity. Indeed, DBD-A and DBD-B were the first to be iden- tified as DNA-binding domains [12]. DBD-C and DBD-D have a weak ssDNA–binding activity [12], whereas DBD-F interacts physically with the tumor suppressor p53 and nucleosome remodeling complex FACT. The interaction with DBD-F can also contrib- ute to an additional binding of structurally distorted DNA (i.e. damaged DNA). By analogy, the primary function of all the OsRPA70kDa subunits must be to find special regions of DNA with which to bind. Is there a divergence in biochemical function among the K. Sakaguchi et al. The multi-replication protein A system FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS 947 various domains? What is the specialization of hSSBs (analogs of RPA), which appeared by convergent evolution [70]? Furthermore, why are the middle subunits diversified phylogenetically? As discussed earlier, the major role of the middle subunits is not to bind to DNA, although they may be involved in the controlling signal via phosphorylation. Indeed, in humans, HsRPA2 interacts with uracil–DNA glycosylase and XPA, but HsRPA4 does not [67–69]. Moreover, the small sub- unit is presumably responsible for linking the other subunits (large and middle). The driving force behind the diversification of the small subunit is an interesting question that needs to be addressed. The phylogenetic data suggest that the multi-RPA (or the SSB–RPA mixed) systems are universal in eukaryotes. However, it is important to establish whether plants have paralogs or orthologs of hSSB. In particular, we need to investigate the in vivo functions of each of the A, B and C types of plant multi-RPA systems. In vivo roles of the multi-RPA system If the multi-RPA system is unique in plants, some of the in vivo roles may also be specific for plants. OsRPA70a (type A complex) is localized in the chloro- plast, but OsRPA70b (type B) and OsRPA70c (type C) are found in the nuclear compartment [12]. The type A system is thought to be plant specific, whereas types B and C could be universal. Fortunately, the homologs of OsRPA70a, OsRPA70b and OsRPA70c were found A B Fig. 2. (A) Pairwise comparison of each OsRPA subunit with human (HsRPA), Schizosaccharomyces pombe (SpRF-A) and Drosophila melanogaster (DmRPA). (B) Domain structures of OsRPAs. Each color box indicates each DBD domain shown as the lower half of the figure. DBD domain are classified into A, B, C, D, E and F. The multi-replication protein A system K. Sakaguchi et al. 948 FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS to be present in A. thaliana (AtRPA70a, AtRPA70b and AtRPA70c) [11,12]. Interestingly, the AtRPA70a deletion mutant (SALK017580) was lethal, but the AtRPA70b deletion mutant (SALK088429) was viable and hypersensitive to UV and MMS [12]. Therefore, type A may be essential for DNA replication and transcription (and also DNA repair) in the chloroplast. Type B may have at least some role in nuclear DNA repair [12]. Intrigu- ingly, the AtRPA70c deletion mutant does not appear to be viable. Type C shows nuclear localization, and the AtRPA70c deletion mutant may be lethal, suggest- ing that type C is essential for DNA replication and transcription (and possibly DNA repair) in the nucleus [12]. To investigate the function of the various proteins, RNAi of AtRPA70a and AtRPA70b were performed [140–143]. The RNAi-mediated knockdown of AtRPA70a also displayed lethality. However, RNAi of AtRPA70b was viable and did not differ in phenotype from wild-type. RT-PCR analysis was also carried out using total RNA extract from seedlings of atrpa70b mutant and the AtRPA70b RNAi line. No atRPA70b transcript could be detected. Furthermore, western blot analysis of total proteins from seedlings of wild-type and atrpa70b mutant indicated very little AtRPA70b [12]. These results indicated that AtRPA70a (probably, the AtRPA70a–AtRPA32-2–AtRPA14 complex) has an essential role, probably in DNA replication in the chloroplast, whereas AtRPA70b (the AtRPA70b–At- RPA32-1–AtRPA14 complex) is not essential under normal growth conditions. However, it is known that yeast rpa70 mutants are very sensitive to mutagens such as UV and MMS [11,12]. To determine whether AtRPA70b is similarly involved in mutagen tolerance, the mutagen sensitivity of atrpa70b mutant and the AtRPA70b RNAi line was tested. When 1-week-old seedlings were exposed to various UV-B doses and then grown for an additional week in the absence of UV-B, there were no remarkable morphological differ- ences between wild-type, atrpa70b mutant and AtRPA70b RNAi line seedlings, although leaf yellow- ing was somewhat increased in the mutant and RNAi seedlings [11,12]. Compared with wild-type, the amounts of chlorophyll (a + b) were decreased in atrpa70b and the AtRPA70b RNAi lines [11,12]. One- week-old seedlings were also grown on MS medium containing various concentrations of MMS or H 2 O 2 . After 1 week, growth of the wild-type plants was inhibited by UV-B, MMS or H 2 O 2 . Compared with wild-type plants, the growth of atrpa70b mutant and AtRPA70b RNAi line seedlings was more inhibited by UV-B, and was completely stopped by MMS [11,12]. Mutants showed little increase in sensitivity to H 2 O 2 . Like the yeast rpa70 mutants, the atrpa70b mutant and AtRPA70b RNAi line are more sensitive than wild-type to UV and MMS, suggesting that At- RPA70b is involved in the repair system for DNA damaged by these mutagens [11,12]. The lethality of both the T-DNA insertion mutant and the RNAi line of AtRPA70a indicate that the AtRPA70a–AtRPA32-2–AtRPA14 complex plays an essential role, such as DNA replication, in the chlorop- lasts of living cells (Fig. 3). By contrast, the mutant and RNAi line of AtRPA70b were viable but showed high sensitivity to UV and MMS, suggesting involve- ment of the AtRPA70b–AtRPA32-1–AtRPA14 com- plex in the repair of damaged DNA (Fig. 3). However, AtRPA70c deletion was thought to be lethal, suggest- ing that the AtRPA70c–AtRPA32-3–AtRPA14 com- plex may function mainly in nuclear DNA replication and transcription (Fig. 3). Subcellular localization analysis suggested that the type A RPA complex is required for chloroplast DNA metabolism, whereas types B and C function in nuclear DNA metabolism [12]. Recently, RPA70 and RPA32 subunits from plants have been reported to play a role in viral and transpo- son DNA syntheses [131,144]. It will be intriguing to investigate how the RPA complex functions in these mechanisms. Higher plants may have evolved the type A for the chloroplast to offer protection against high levels of UV irradiation. Indeed, as mentioned earlier, plants are exposed to UV radiation for much longer than animals or yeast. Higher plants depend on exposure to sunlight, including UV, for their develop- ment because their energy is derived from photosyn- thesis. Thus, the repair system in subcellular organelles is presumably much more efficient in plants than in animals and yeast. The human homologs of RPA32, HsRPA2 and HsRPA4 [67] may correspond to OsRPA32-1 (type B) and OsRPA32-3 (type C) of plants, respectively, although only the middle subunit is diversified. Inter- estingly, hSSB1 did not localize to replication foci in S-phase cells and hSSB1 deficiency did not influence S-phase progression [70]. Depletion of hSSB1 abro- gated the cellular response to DSBs, including activa- tion of ATM and phosphorylation of ATM targets, after ionizing radiation [70]. Ionizing radiation and anti-cancer drugs can induce DNA DSBs, which are highly cytotoxic lesions. Cells deficient in hSSB1 exhib- ited increased radiosensitivity, defective checkpoint activation and enhanced genomic instability coupled with a diminished capacity for DNA repair. Thus, K. Sakaguchi et al. The multi-replication protein A system FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS 949 hSSB1 must influence diverse endpoints in the cellular DNA damage response. In this way, hSSB1 resembles the type B system. Why are they not always found? The multi-RPA types may resemble each other biochemically because most of the subunits (large and ⁄ or middle) display a significant degree of similarity. In many eukaryotes, the multi-RPA system may diversify by exchanging some subunits. For example, some of the non-homolog(s) of hSSB1 are derived from convergent evolution. Further- more, ubiquitous RPA (type B) is dispensable and can easily be analyzed using the knockdown mutant, whereas the type C or HsRPA complex (or hSSB2) is lethal. However, very few researchers have studied these mutants. Interestingly, the same phenomena was found in Drosophila PCNAs, where the major PCNA is a homolog of the ubiquitous PCNA in eukaryotes but is dispensable [65]. Subsequently we analyzed the proper- ties of these proteins in more detail. The role of the miner subunit is not well understood because the knockdown mutant is, as yet, unavailable [65]. A new perspective for RPA complexes If multi-system RPAs are found to be universal each of the corresponding functions should be reconsidered. Nuclear RPAs may be divided into two categories: (a) replication ⁄ transcription (plant C type), and (b) repair ⁄ recombination (plant B type). The large subunit may function as an agent for ssDNA stretching [1,2], whereas the middle subunit may act as a signal trans- duction acceptor. The small subunit may be a connect- ing factor for forming the heterotrimeric complex. Indeed, the small subunit mostly exists as a hetero- dimer with the middle subunit, whereas the largest sub- unit can be stabilized by binding to the dimer [10–12]. Genetic knockdown of the type 1 RPA increases the lethality (i.e. the type C), but type 2 RPA can survive unless the DNA is damaged (i.e. type B). Therefore, subunit variety and function of the various subunits of RPA must be reconsidered in view of these new find- ings. For example, human RPA interacted with XPA at sites of DNA damage, stimulated XPA–DNA inter- action, and recruited the incision proteins ERCC1 ⁄ XPF and XPG to the damaged site [89]. The RPA must be a complex with HsRPA2, which corre- sponds to type B. In NER and long-patch base exci- sion repair, type B may be responsible for these functions in eukaryote kingdoms. The reported biological functions of mammalian and yeast RPA are mostly involved in meiosis. The middle subunit has an important role in regulating synaptine- mal complex (SC) formation and meiotic recombina- tion at meiotic prophase, mainly at zygotene and pachytene [71–74,114,115,130]. The protein factors, such as DNA polymerases and recombinases, are major proteins involved in meiotic prophase events. Nevertheless, RPA is known biochemically to interact in vitro with DNA polymerases and recombinases [6–8,13,31,40–42,44,72,85–88,138,145–169]. In fulfilling its biosynthetic roles in nuclear replica- tion and in several types of repair, DNA polymerase is assisted by RPA. In eukaryotes, recent investigations have revealed at least 14 types of DNA polymerase Fig. 3. Hypothetic model of the cellular function of A-, B- and C-type RPA com- plexes. The multi-replication protein A system K. Sakaguchi et al. 950 FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS (pol a, b, c, d, e, f, g, h, i, j, k, l, m and p) [45,170]. In a sense, all are analogs of each other. RPA is reported to interact with at least pol a, d, e, k and j [3,5–8,76,80,81,85–88]. RPA contributes to the high fidelity of the polymerases during DNA synthesis. Of the polymerase species, pol a, d and e replicate DNA during S phase, but pol a is replication specific [80]. All the other polymerases are involved in DNA repair and recombination [81]. We reported that in meiosis two categories of DNA polymerases (a) pol a complex and (b) pol k and l were expressed [165,168]. The former is for replication at zygotene (or SC formation) and the latter is for repair and recombination at late zygotene to pachytene (Fig. 4) [155,165,168,171–173]. Using a D-loop recombination intermediate substrate, we observed that either pol k or pol l can promote the primer extension of an invading strand present in a D-loop structure [168]. Both could fully extend the primer in the D-loop substrate, suggesting that the D-loop extension is an activity that is intrinsic to the polymerases [168]. Two orthologs of the recombinases, Rad51 and Lim15 ⁄ Dmc1, are present in meiosis [44,114,115,152– 154,161,162,167]. These recombinases occur at late leptotene to early zygotene (Fig. 4). The interaction of RPA and Rad51 is well established. Another meiotic role of RPA was also found. At meiotic prophase (late leptotene to early zygotene), with RPA, the homology- search recombinase complex is involved in homologous chromosome synapsis, preventing the formation of superfluous reciprocal recombinant events (Fig. 4) [114,115]. Both Rad51 and Lim15 ⁄ Dmc1 were identi- fied as being involved in this process, although the specific function of each protein is not yet known [44]. Are the DNA polymerase and recombinase functions mediated by one species of RPA complex? Interestingly, dephosphorylation of transformed nod- ule-associated histone H2AX chromatin occurs at this time. This suggests annealing of single strands or repair of DSBs. By a similar mechanism, if the middle subunit of RPA is also dephosphorylated, RPA would lose the function of maintaining the noncross-over condition. We must also consider the role of the multi- RPA system during the meiotic prophase events. It is known that a small amount of DNA replicates at zygotene (pairing DNA synthesis) and that the repair synthesis of DNA occurs at pachytene (cross- over DNA synthesis) [172,173]. The two sequential DNA synthesis reactions play a role in the progression of meiosis. It is possible that a complex of RPA and pol a differs from the recombination-dependent RPA. Because DNA polymerase searches for the RPA– Fig. 4. Hypothetic model of meiotic cell cycle and its relation to RPA. K. Sakaguchi et al. The multi-replication protein A system FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS 951 ssDNA complex structure on the DNA, RPA complexed with pol a are probably functionally inde- pendent from RPA complexed with other repair polymerases. Pol k and l were thought to be involved in the ‘crossover DNA synthesis’ for DNA recombina- tion. Because the pol k(or the pol l)-deficient mutant is viable, RPA may be like the type B or HsRPA2 type. However, ‘pairing DNA replication’ appears to be specific for SC formation. At that stage, the DNA polymerase a-catalytic subunit and primase are pre- sumably also present [165]. This replication could be the basis for SC extension and formation of the transi- tion nodules [44]. Indeed, this process probably requires RPA, such as the type C form (Fig. 4). During prophase, DNA polymerases as well as paralogs and orthologs of PCNA, recombinases and perhaps RPA are required (Fig. 4) [42,44,45, 151,152,155,157,159,160,165,168,171]. Electron micros- copy data [115,117] suggest that meiotic functions in vivo are shared by each of the paralogs and ortho- logs, and maybe also the analogs and heterologs. Indeed, control of the biological process could be more finely tuned by sharing function amongst paralogs, orthologs, analogs and heterologs. Background for the screening of multiple protein systems involved in DNA metabolism We have studied many protein factors in DNA replica- tion ⁄ repair and their relation to the meiotic system in higher plants (O. sativa and A. thaliana) [13– 43,45,156,171], a fungus (C. cinereus) [44,138,145–155, 157–169] and an arthropod (Drosophila melanogaster) [44–66]. Each of the materials represents the biological kingdom of plant, fungus and animal, respectively. Our research aimed to comprehensively understand these DNA synthesis-related events in phylogenetically diverse species. In addition to RPA, we elucidated many of the related factors, such as Rad51, Lim15 ⁄ Dmc1, RadA, PCNA, DDB, XRCC1, Rad2 family nucleases and special nucleases, DNA polyme- rases, ORC1, RFC, RecQ, DNA ligases, CAF-1, mtTFA, Rrp1, Mer3, Snm1, Rad6, SUMOylation fac- tors (Aos1, Uba2, Ubc9, SUMO), leucine aminopepti- dase and 26S proteasome-related factors (Jab1, Sgt1, DnaJ) (Table 1). During the course of our experi- ments, we frequently observed that protein factors involved in the same DNA metabolic processes are not always homologs in eukaryotic cells. Although the paralogs and orthologs are ubiquitous, evolutionally different factors were often found to be involved in the same biosystems, which are referred to as ‘analogs’ and ‘heterologs’. Indeed, convergent evolution might be ubiquitous in eukaryotic DNA metabolic processes. According to definition, ‘homolog’ is a gene related to a second gene by descent from a common ancestral DNA sequence. ‘Ortholog’ is a gene in different species that evolved from a common ancestral gene. ‘Paralog’ is a gene related by duplication within a genome. Orthologs retain the same function in the course of evolution, whereas paralogs evolve new functions. ‘Analog’ is a gene that has common activity but not a common origin. ‘Heterolog’ is a gene that differs in both origin and activ- ity. Heterolog does not classify homolog, ortholog, par- alog or analog. It may be also said that heterolog is used as a synonym of ‘just different protein (gene)’, basically. For example, PCNA is not one copy [65,138,159]; two PCNA paralogs and one PCNA-like heterotrimer (Rad9–Rad1–Hus1) (‘analog’ or ‘heterolog’) were found in Drosophila [65,136,137]. Rad9–Rad1–Hus1 is found universally in eukaryotes. Plant SYCP1 and yeast Zip1 mediate the same role in meiosis, despite displaying no significant homology (‘analog’ or ‘het- erologs’) [174,175]. Similarly, human mus81–Eme1 is functionally the same as Escherichia coli RuvC (‘ana- log’ or ‘heterologs’) [176–178]. In plants, two recA-like protein paralogs (Rad51 and Lim15 ⁄ Dmc1) as well as a prokaryotic recA homolog (RadA) were found (‘ana- logs’) [42]. Furthermore, this is not the plastid compo- nent [42]. As described earlier, in addition to the two subtypes of RPA (HsRPA2 and HsRPA4) two human SSB homologs are also present (‘analogs’) [70]. More- over, in human, five Rad51 paralogs (Rad51B, Rad51C, Rad51D, Xrcc2 and Xrcc3) have been found [179–181]. Two FEN-1 paralogs (FEN-1a and FEN- 1b) and one analog (SEND-1) were found in plants [25,26], and another FEN-1 analog occurs in Drosoph- ila (GEN) [63,66]. DNA polymerases, especially for DNA repair, are greatly diversified in eukaryotes [76,182,183]. DNA polymerase b (pol b) for short patch base excision repair are found only in verte- brates [45]; plant short patch base excision repair uses pol f instead [33,39,45]. However, as yet, a recBCD homolog has not been found in the eukaryotic recom- bination process. Prokaryotic homologs such as RadA and hSSB are often found in eukaryotes (‘analog’ or ‘heterolog’), although there are the eukaryotic func- tional alternatives [42,70]. All the protostomic animals lack any X family DNA polymerases essential for development of the nervous and immune system [45]. In Drosophila, AP endonuclease 1 homolog (Rrp1) binds to pol f [64]. Plant XRCC1 lacks the polymer- ase-binding domain [33,39]. Therefore, factor variation (orthologs, paralogs, ‘analogs’ and ‘heterologs’) seems to be ubiquitous in eukaryotic DNA metabolism. The multi-replication protein A system K. Sakaguchi et al. 952 FEBS Journal 276 (2009) 943–963 ª 2009 The Authors Journal compilation ª 2009 FEBS [...]... Escherichia coli DNA polymerase I from a higher plant, rice (Oryza sativa L.) Nucleic Acids Res 30, 158 5–1 592 Yanagawa Y, Kimura S, Takase T, Sakaguchi K, Umeda M, Komamine A, Tanaka K, Hashimoto J, Sato T & Nakagawa H (2002) Spatial distribution of the 26S proteasome in meristematic tissues and primordia of rice (Oryza sativa L.) Planta 214, 70 3– 707 Uchiyama Y, Hatanaka M, Kimura S, Ishibashi T, Ueda T, Sakakibara... 211 9–2 128 167 Namekawa SH, Iwabata K, Sugawara H, Hamada FN, Koshiyama A, Chiku H, Kamada T & Sakaguchi K (2005) Knockdown of LIM15 ⁄ DMC1 in the mushroom Coprinus cinereus by double-stranded RNA-mediated gene silencing Microbiology 151, 366 9–3 678 168 Sakamoto A, Iwabata K, Koshiyama A, Sugawara H, Yanai T, Kanai Y, Takeuchi R, Daikuhara Y, Takakusagi Y & Sakaguchi K (2007) Two X family DNA FEBS Journal... 1627, 4 7–5 5 165 Namekawa S, Hamada F, Sawado T, Ishii S, Nara T, Ishizaki T, Ohuchi T, Arai T & Sakaguchi K (2003) Dissociation of DNA polymerase alpha–primase complex during meiosis in Coprinus cinereus Eur J Biochem 270, 213 7–2 146 166 Namekawa S, Ichijima Y, Hamada F, Kasai N, Iwabata K, Nara T, Teraoka H, Sugawara F & Sakaguchi K (2003) DNA ligase IV from a basidiomycete, Coprinus cinereus, and its... Drosophila homologue of human abasic endonuclease 1 J Biol Chem 281, 1157 7– 11585 Ruike T, Takeuchi R, Takata K, Oshige M, Kasai N, Shimanouchi K, Kanai Y, Nakamura R, Sugawara F & Sakaguchi K (2006) Characterization of a second proliferating cell nuclear antigen (PCNA2) from Drosophila melanogaster FEBS J 273, 506 2– 5073 Kanai Y, Ishikawa G, Takeuchi R, Ruike T, Nakamura R, Ihara A, Ohashi T, Takata K,... Commun 232, 45 4–4 60 The multi-replication protein A system 158 Takami K, Matsuda S, Sono A & Sakaguchi K (1994) A meiotic DNA polymerase from a mushroom, Agaricus bisporus Biochem J 299(Pt 2), 33 5–3 40 159 Hamada FN, Koshiyama A, Namekawa SH, Ishii S, Iwabata K, Sugawara H, Nara TY, Sakaguchi K & Sawado T (2007) Proliferating cell nuclear antigen (PCNA) interacts with a meiosis-specific RecA homologues,...K Sakaguchi et al The multi-replication protein A system Table 1 The main role of DNA synthesis-related factor Protein Function Reference RPA Required for DNA recombination, repair and replication The activity of RPA is mediated by ssDNA binding and protein interactions May participate in a common DNA damage-response pathway associated with the activation of homologous recombination and double-strand... Yanagawa Y, Yamamoto T, Nakagawa H, Tanaka I, Hashimoto J & Sakaguchi K (2001) Characterization of plant proliferating cell nuclear antigen (PCNA) and flap endonuclease-1 (FEN-1), and their distribution in mitotic and meiotic cell cycles Plant J 28, 64 3–6 53 Kimura S, Uchiyama Y, Kasai N, Namekawa S, Saotome A, Ueda T, Ando T, Ishibashi T, Oshige M, Furukawa T et al (2002) A novel DNA polymerase homologous... stimulate its strand transfer activity Biochem Biophys Res Commun 352, 83 6–8 42 160 Ishii S, Koshiyama A, Hamada FN, Nara TY, Iwabata K, Sakaguchi K & Namekawa SH (2008) Interaction between Lim15 ⁄ Dmc1 and the homologue of the large subunit of CAF-1: a molecular link between recombination and chromatin assembly during meiosis FEBS J 275, 203 2–2 041 161 Iwabata K, Koshiyama A, Yamaguchi T, Sugawara H, Hamada... 323, 102 4–1 031 Ishikawa G, Kanai Y, Takata K, Takeuchi R, Shimanouchi K, Ruike T, Furukawa T, Kimura S & Sakaguchi K (2004) DmGEN, a novel RAD2 family endoexonuclease from Drosophila melanogaster Nucleic Acids Res 32, 625 1–6 259 Takeuchi R, Ruike T, Nakamura R, Shimanouchi K, Kanai Y, Abe Y, Ihara A & Sakaguchi K (2006) Drosophila DNA polymerase zeta interacts with recombination repair protein 1, the Drosophila... Sakakibara Y, Matsumoto T, Furukawa T, Hashimoto J & Sakaguchi K (2002) Characterization of DNA polymerase delta from a higher plant, rice (Oryza sativa L.) Gene 295, 1 9–2 6 Furukawa T, Kimura S, Ishibashi T, Mori Y, Hashimoto J & Sakaguchi K (2003) OsSEND-1: a new RAD2 nuclease family member in higher plants Plant Mol Biol 51, 5 9–7 0 Kimura S, Furukawa T, Kasai N, Mori Y, Kitamoto HK, Sugawara F, Hashimoto . 151 , 366 9–3 678. 168 Sakamoto A, Iwabata K, Koshiyama A, Sugawara H, Yanai T, Kanai Y, Takeuchi R, Daikuhara Y, Tak- akusagi Y & Sakaguchi K (2007) Two X family. 213 7–2 146. 166 Namekawa S, Ichijima Y, Hamada F, Kasai N, Iwaba- ta K, Nara T, Teraoka H, Sugawara F & Sakaguchi K (2003) DNA ligase IV from a basidiomycete,