**3. Multiplicity of homologs**

176 Selected Topics in DNA Repair

Shereda, et al., 2008). These SSB complexes occlude up to 65 and 35 nucleotides of ssDNA, respectively (Filipkowski & Kur, 2007; Meyer & Laine, 1990; Shereda, et al., 2008). Other, less common SSB configurations have also been identified, including an alternative homopentameric Deinococcal SSB (Norais, et al., 2009; Sugiman-Marangos & Junop, 2010). In general, the N-terminal region of the SSB contains the oligomerization domain while the C-terminal region is implicated in heterologous protein interactions (Shereda, et al., 2008).

Fig. 1. RPA and SSB structures from eukaryotes, bacteria, and archaea. (A) The canonical eukaryotic RPA comprises three protein subunits, each containing OB-folds designated DBD

monomer. (C) Archaeal RPAs are diverse and display characteristics of both eukaryotic RPA and bacterial SSB. Examples from *H. salinarum* are shown. RFA6 resembles bacterial SSB, whereas RFA1 is uniquely archaeal. The genes encoding RFA2 and RFA7, which are found

The eukaryotic RPA is a heterotrimer composed of a large RPA1 (~72 kDa), medium RPA2 (~32 kDa), and small RPA3 (~14 kDa) subunit. RPA1 and RPA2 tend to be conserved subunits with a variable RPA3, though considerable variation can occur throughout domains (Lin, et al., 2008; Wold, 1997). RPA1 contains four OB-folds referred to as DNA binding domains (DBD). DBD A and B are centrally located, DBD C is in the C-terminal region, and DBD F is in the N-terminal region. RPA2 contains a single centrally located OBfold, DBD D, and RPA3 a single centrally located OB-fold, DBD E (Binz, et al., 2004; Oakley, et al., 2009; Pretto, et al., 2010). ssDNA binding occurs primarily through interaction with DBD A-D, which together are capable of occluding approximately 30 nucleotides (Blackwell & Borowiec, 1994; Iftode, et al., 1999; C. Kim, et al., 1992; Theobald, et al., 2003). A binding polarity of the complex is achieved through decreased ssDNA affinity from DBD A to D (Oakley, et al., 2009; Pretto, et al., 2010). The N-terminal domain of RPA1 is implicated in heterologous protein interactions and regulation, particularly through DBD F, while the Cterminal domain is involved in heterotrimer structure interactions (Broderick, et al., 2010;

A-F. Yellow OB-folds are DNA interacting domains, blue domains are involved in heterologous protein interactions, and gray domains are involved in maintenance of RPA structure. The N-terminus of RPA2 is the primary phosphorylation domain, denoted by red Ps. (B) The canonical bacterial SSB functions as a homotetramer and contains a single OBfold (yellow) and a C-terminal heterologous protein interacting domain (blue) per

in an operon, likely function as a multimeric protein, as do RFA3, RFA8, and RAL.

The highly conserved nature of the OB-fold has allowed identification of potential RPA/SSB homologs in the ever-expanding genome sequence database. To our knowledge, there have been no reports of genomes lacking an RPA/SSB homolog, with the possible exception of the crenarchaea *Pyrobaculum aerophilum* and *Aeropyrum pernix* (Luo, et al., 2007).

There are numerous excellent and extensive reviews covering the history, genetics and biochemical characterization of *E. coli* SSB and its encoding gene (Lohman & Ferrari, 1994; Meyer & Laine, 1990; Shereda, et al., 2008). The single gene, *ssb*, is located adjacent to, but divergently transcribed from, the *uvrA* gene (Brandsma, et al., 1985). In *Bacillus subtilis*, there are two SSB genes, *ssb* and *ywpH* (Lindner, et al., 2004). Although the amino acid sequences of the two proteins are similar, YwpH lacks the C-terminal region found in both the *E. coli* and *B. subtilis* SSBs. *B. subtilis ssb* is essential, but unlike in *E. coli*, is found between ribosomal protein genes. Of 87 bacterial genomes analyzed, all contained at least one SSB gene (Lindner, et al., 2004). Based on gene organization, four groupings were proposed: Group I organisms display *B. subtilis-*type gene organization and more than one SSB. Group II organisms contain only one gene for SSB. *E. coli* represents the type organism for Group III bacteria, where the single *ssb* gene is divergently transcribed from the adjacent *uvrA* gene. Group IV consists of organisms whose SSB organization does not resemble either *B. subtilis* or *E. coli*. *Thermotoga maritima* and *T. neopolitana* likely fall into Group IV, whereas *Thermoanaerobacter tengcongensis* is in Group I (Olszewski, et al., 2010; Olszewski, et al., 2008).

Recently, a novel homopentameric SSB, DdrB, was identified in *Deinococcus radiodurans*, which may challenge previous notions regarding canonical SSB structure (Norais, et al., 2009; Sugiman-Marangos & Junop, 2010). Restricted to the Deinococcal lineage, it may represent an adaptation related to the unusually high DNA repair capacity of this organism. However, it should be kept in mind that other SSB proteins that deviate from the traditional consensus may yet be identified (Norais, et al., 2009; G. Xu, et al., 2010).

Three separate genes encode the canonical eukaryotic RPA, which is the subject of several recent review articles (Binz, et al., 2004; Broderick, et al., 2010; Richard, et al., 2009; Wold, 1997). However, in higher plants each subunit of RPA may be represented by multiple genes whose proteins function in distinct processes (Sakaguchi, et al., 2009).

Both *Saccharomyces cerevisiae* and *Schizosaccharomyces pombe* contain a single gene for each of the three RPA subunits. All three are essential in *S. cerevisiae*, but RPA3 appears to be nonessential in *S. pombe* (Brill & Stillman, 1991; Cavero, et al., 2010; Dickson, et al., 2009; Maniar, et al., 1997). Given that RPA3 is the most variable, and the essential role that the intact protein plays in replication, an additional RPA3 gene may yet be identified that carries out this role in *S. pombe*. In addition to heterotrimeric RPA, a protein complex, Stn1/Ten1, resembling an RPA2/RPA3 dimer has been found to be necessary for telomere maintenance

Role of RPA Proteins in Radiation Repair and Recovery 179

one additional homolog that shows different domain structure, as seen in Figure 1 (Komori & Ishino, 2001; Pugh, et al., 2008; Robbins, et al., 2005). Other euryarchaea, such as *Methanococcus jannaschii, Methanobacter thermoautotrophicus*, and the Methanosarcinae, have proteins with four OB-folds, which may act as monomers, or in a complex with an adjacently encoded RPA2 homolog (Chedin, et al., 1998; Kelly, et al., 1998; Komori & Ishino, 2001; Lin, et al., 2008; Robbins, et al., 2004). In addition, *M. jannaschii* contains a two-OB-fold homolog which appears to function as a homotrimer (Robbins, et al., 2005). The *Pyrococcus furiosis* RPA more closely resembles eukaryotic RPA, with three distinct subunits that function as a heterotrimer and whose genes compose an operon (Komori & Ishino, 2001). In *H. salinarum*, Rfa3, Rfa8, and Ral, and Rfa2 and Rfa7 resemble the *P. furiosis* RPA with respect to operon structure and sequence homology to eukaryotic RPA (Figure 1). Rfa6 resembles the crenarchaeal SSB/RPA in gene and OB-fold structure, while Rfa1 appears more uniquely archaeal, with three OB-folds (DeVeaux, et al., 2007; Robbins, et al., 2005).

The roles of the RPA-like homologs in *H. salinarum* will be discussed further below.

*E. coli* SSB mutants were first isolated in a screen for DNA replication mutants (Meyer & Laine, 1990). Although the two best-characterized of these temperature-sensitive (ts) mutations, *ssb-1* and *ssb-113*, fail to grow at 42o C due to inability of the labile SSB to participate in DNA replication, *ssb-1* mutants have essentially normal phenotypes at the permissive temperature. The mutation (H55Y) resides in the OB-fold and affects the ability of the protein to form homotetramers (Shereda, et al., 2008). In contrast, even at the permissive growth temperature, *ssb-113* mutants are severely compromised in their ability to survive numerous DNA damaging agents, and display recombination deficiencies (Chase, et al., 1984). The extremely pleiotropic phenotype conferred by the amino acid change (P176S of 177) in *ssb-113* suggests that the protein interaction capabilities of the C-terminal region are necessary for more than just replication. A protein lacking only 10 amino acids from the Cterminus is non-functional (Curth, et al., 1996). Indeed, mutation of the terminal phenylalanine is lethal, and disrupts protein-protein interactions (Genschel, et al., 2000). These phenotypes provided support not only for the commonality of protein function in recombination, repair, and replication, but also early evidence that these roles were separable within an individual protein. In contrast, a *B. subtilis* mutant in which SSB lacks the Cterminal 35 amino acids, involved in interaction with the helicase PriA, is viable, demonstrating that this conserved region is not essential in all bacteria (Lecointe, et al., 2007). In the budding yeast *S. cerevisiae*, all three subunits of RPA are essential (Brill & Stillman, 1991; Maniar, et al., 1997). Like *E. coli ssb-113* mutants, ts-mutants of RPA1 are profoundly UV- and ionizing radiation (IR)-sensitive at the permissive temperature (Parker, et al., 1997). Several ts-mutants display a mutator phenotype, which is likely related to defects in replication rather than repair (Chen, et al., 1998). Small insertions in either the RPA1 DNAbinding domain or the N-terminal protein-interaction domain result in a marked decrease in survival after UV exposure; however, only the latter is defective in the cell-cycle response to such damage. In addition, only the DNA-binding mutant is defective in repair of UVinduced lesions (Longhese, et al., 1996; Teng, et al., 1998). The N-terminal domain is required for interactions with the clamp loader. This binding is abolished in the *rfa1-t11* mutation, which contains a single change (K45E) (H. S. Kim & Brill, 2001; Majka, et al., 2006; Umezu, et al., 1998). This same allele is deficient in many pathways requiring interactions

**4. Characterized mutations of RPA and SSB** 

and protection in *S. cerevisiae*, *S. pombe* and *Candida albicans* (Sun, et al., 2009). Together with Cdc13, the Stn1/Ten1 complex may represent a highly conserved telomere binding RPA.

Mammals are thought to have only one nuclear RPA protein, which functions in both replication and repair (Fanning, et al., 2006; Iftode, et al., 1999; Wold, 1997). However, an alternative RPA2 homolog (RPA4) found only in mammals has been identified (Keshav, et al., 1995). RPA4 binds with the single RPA1 and RPA3 subunits to form an alternative RPA that does not interact with DNA polymerase α, and consequently, does not support replication (Haring, et al., 2010; Mason, et al., 2009; Mason, et al., 2010). Rather, its role is restricted to repair, particularly in quiescent cells. A mitochondrial SSB has also been identified in many eukaryotes, which has sequence and structural similarities to *E. coli* SSB (Curth, et al., 1994). In addition, two human proteins, hSSB1 and hSSB2, which structurally resemble bacterial SSB, have been reported. hSSB1 has been characterized, and appears not to be involved in replication, but is required for genome stability and DNA repair processes (Richard, Cubeddu, et al., 2011; Richard, Savage, et al., 2011).

Although much less is known about RPA genes in plants, the recent sequencing of higher plant genomes has revealed a large diversity of homologs (Sakaguchi, et al., 2009). Rice contains three genes each for RPA1 and RPA2, but only a single gene for RPA3 (Ishibashi, et al., 2006; Shultz, et al., 2007). Each RPA1 associates with a particular RPA2, and RPA3 is common among the different complexes. One complex is unique to chloroplasts, whereas the other two are nuclear. Like rice, *Arabidopsis* contains multiple homologs for the RPA1 and RPA2 subunits, but also has two RPA3 genes (Shultz, et al., 2007). In both plants, one of the large subunits has been shown to be non-essential for vegetative growth, but is required for meiosis, demonstrating specialization of function among the multiple species of protein (Chang, et al., 2009; Osman, et al., 2009). This homolog is involved in response to DNA damage as well as maintenance of telomeres (Takashi, et al., 2009). The degree of similarity among the multiple homologs suggests that they arose through duplication of ancestral genes after the establishment of the plant lineage.

By far the most diversity in RPA/SSB homologs is found in the archaea. Although many hypothetical RPA proteins have been identified in the numerous genomes analyzed, there is no quaternary structure that is common to all. In the crenarchaea, the RPA from *Sulfolobus solfataricus* is encoded by a single gene and exists as a monomer or a homotetramer (Haseltine & Kowalczykowski, 2002; Kerr, et al., 2001; Kerr, et al., 2003; Wadsworth & White, 2001). The different quaternary structures confer distinct binding capabilities (Rolfsmeier & Haseltine, 2010). Despite the similarity to bacterial SSB, and the ability of the gene to complement the lethality of an *E. coli ssb* mutation, the DNA-binding domain of this protein more closely resembles those of eukaryotic RPA1 (Haseltine & Kowalczykowski, 2002; Kerr, et al., 2003). Although the genomes of two related crenarchaea, *P. aerophilum* and *A. pernix*, have been reported to contain no obvious RPA or SSB homolog (Luo, et al., 2007), a previous study found a *Sulfolobus*-like SSB in *A. pernix* (Haseltine & Kowalczykowski, 2002). Given the recent identification of a novel SSB in *D. radiodurans*, and the fundamental role that ssDNA binding proteins play in replication and repair, it is likely that a protein carrying out these essential functions will be found in all organisms.

RPAs in euryarchaea present a multitude of OB-fold conformations that presumably provide unique functions (Lin, et al., 2005; Robbins, et al., 2005). Genomes of several euryarchaea, including *Thermoplasma acidophilum*, *Archaeoglobus fulgidus, Ferroplasma acidarmanus*, and *Halobacterium salinarum*, contain genes with similar organization to the crenarchaeal SSB/RPA gene, with one OB-fold; however, in each case, there is also at least

and protection in *S. cerevisiae*, *S. pombe* and *Candida albicans* (Sun, et al., 2009). Together with Cdc13, the Stn1/Ten1 complex may represent a highly conserved telomere binding RPA. Mammals are thought to have only one nuclear RPA protein, which functions in both replication and repair (Fanning, et al., 2006; Iftode, et al., 1999; Wold, 1997). However, an alternative RPA2 homolog (RPA4) found only in mammals has been identified (Keshav, et al., 1995). RPA4 binds with the single RPA1 and RPA3 subunits to form an alternative RPA that does not interact with DNA polymerase α, and consequently, does not support replication (Haring, et al., 2010; Mason, et al., 2009; Mason, et al., 2010). Rather, its role is restricted to repair, particularly in quiescent cells. A mitochondrial SSB has also been identified in many eukaryotes, which has sequence and structural similarities to *E. coli* SSB (Curth, et al., 1994). In addition, two human proteins, hSSB1 and hSSB2, which structurally resemble bacterial SSB, have been reported. hSSB1 has been characterized, and appears not to be involved in replication, but is required for genome stability and DNA repair processes

Although much less is known about RPA genes in plants, the recent sequencing of higher plant genomes has revealed a large diversity of homologs (Sakaguchi, et al., 2009). Rice contains three genes each for RPA1 and RPA2, but only a single gene for RPA3 (Ishibashi, et al., 2006; Shultz, et al., 2007). Each RPA1 associates with a particular RPA2, and RPA3 is common among the different complexes. One complex is unique to chloroplasts, whereas the other two are nuclear. Like rice, *Arabidopsis* contains multiple homologs for the RPA1 and RPA2 subunits, but also has two RPA3 genes (Shultz, et al., 2007). In both plants, one of the large subunits has been shown to be non-essential for vegetative growth, but is required for meiosis, demonstrating specialization of function among the multiple species of protein (Chang, et al., 2009; Osman, et al., 2009). This homolog is involved in response to DNA damage as well as maintenance of telomeres (Takashi, et al., 2009). The degree of similarity among the multiple homologs suggests that they arose through duplication of ancestral

By far the most diversity in RPA/SSB homologs is found in the archaea. Although many hypothetical RPA proteins have been identified in the numerous genomes analyzed, there is no quaternary structure that is common to all. In the crenarchaea, the RPA from *Sulfolobus solfataricus* is encoded by a single gene and exists as a monomer or a homotetramer (Haseltine & Kowalczykowski, 2002; Kerr, et al., 2001; Kerr, et al., 2003; Wadsworth & White, 2001). The different quaternary structures confer distinct binding capabilities (Rolfsmeier & Haseltine, 2010). Despite the similarity to bacterial SSB, and the ability of the gene to complement the lethality of an *E. coli ssb* mutation, the DNA-binding domain of this protein more closely resembles those of eukaryotic RPA1 (Haseltine & Kowalczykowski, 2002; Kerr, et al., 2003). Although the genomes of two related crenarchaea, *P. aerophilum* and *A. pernix*, have been reported to contain no obvious RPA or SSB homolog (Luo, et al., 2007), a previous study found a *Sulfolobus*-like SSB in *A. pernix* (Haseltine & Kowalczykowski, 2002). Given the recent identification of a novel SSB in *D. radiodurans*, and the fundamental role that ssDNA binding proteins play in replication and repair, it is likely that a protein

RPAs in euryarchaea present a multitude of OB-fold conformations that presumably provide unique functions (Lin, et al., 2005; Robbins, et al., 2005). Genomes of several euryarchaea, including *Thermoplasma acidophilum*, *Archaeoglobus fulgidus, Ferroplasma acidarmanus*, and *Halobacterium salinarum*, contain genes with similar organization to the crenarchaeal SSB/RPA gene, with one OB-fold; however, in each case, there is also at least

(Richard, Cubeddu, et al., 2011; Richard, Savage, et al., 2011).

genes after the establishment of the plant lineage.

carrying out these essential functions will be found in all organisms.

one additional homolog that shows different domain structure, as seen in Figure 1 (Komori & Ishino, 2001; Pugh, et al., 2008; Robbins, et al., 2005). Other euryarchaea, such as *Methanococcus jannaschii, Methanobacter thermoautotrophicus*, and the Methanosarcinae, have proteins with four OB-folds, which may act as monomers, or in a complex with an adjacently encoded RPA2 homolog (Chedin, et al., 1998; Kelly, et al., 1998; Komori & Ishino, 2001; Lin, et al., 2008; Robbins, et al., 2004). In addition, *M. jannaschii* contains a two-OB-fold homolog which appears to function as a homotrimer (Robbins, et al., 2005). The *Pyrococcus furiosis* RPA more closely resembles eukaryotic RPA, with three distinct subunits that function as a heterotrimer and whose genes compose an operon (Komori & Ishino, 2001). In *H. salinarum*, Rfa3, Rfa8, and Ral, and Rfa2 and Rfa7 resemble the *P. furiosis* RPA with respect to operon structure and sequence homology to eukaryotic RPA (Figure 1). Rfa6 resembles the crenarchaeal SSB/RPA in gene and OB-fold structure, while Rfa1 appears more uniquely archaeal, with three OB-folds (DeVeaux, et al., 2007; Robbins, et al., 2005). The roles of the RPA-like homologs in *H. salinarum* will be discussed further below.
