Inactivating pentapeptide insertions in the fission yeast replication factor C subunit Rfc2 cluster near the ATP-binding site and arginine finger motif Fiona C. Gray 1,2 , Kathryn A. Whitehead 1, * and Stuart A. MacNeill 1,2,  1 Wellcome Trust Centre for Cell Biology, University of Edinburgh, UK 2 Department of Biology, University of Copenhagen, Denmark Introduction The heteropentameric clamp loader replication factor C (RFC) plays a key role in chromosome replication in eukaryotic cells. RFC binds to nascent primer– template junctions and catalyses the loading of the ring-shaped sliding clamp, proliferating cell nuclear antigen (PCNA), onto DNA [1,2]. The homotrimeric PCNA complex encircles the DNA completely, form- ing a sliding clamp that tethers DNA polymerase d to the DNA, conferring upon it the processivity necessary to efficiently replicate the genome. PCNA also inter- acts with a large number of additional proteins implicated in DNA replication, DNA repair and DNA modification such as DNA ligase I, the nucleases Fen1 and XP-G, uracil-N-glycosylase and cytosine-5-methyl- transferase [3]. The five subunits of the RFC complex are related to one another but are not interchangeable [1,2]. The complex comprises a large subunit, Rfc1, and four Keywords AAA + protein; clamp loader; DNA replication; fission yeast; replication factor C Correspondence S. MacNeill, Centre for Biomolecular Sciences, University of St Andrews, North Haugh, St Andrews KY16 9ST, UK Fax: +44 01334 462595 Tel: +44 01334 467268 E-mail: stuart.macneill@st-andrews.ac.uk Website: http://biology.st-andrews.ac.uk/ macneill/ Present addresses *Department of Chemistry and Materials, Manchester Metropolitan University, UK Centre for Biomolecular Sciences, University of St Andrews, UK (Received 20 March 2009, revised 24 June 2009, accepted 26 June 2009) doi:10.1111/j.1742-4658.2009.07181.x Replication factor C (RFC) plays a key role in eukaryotic chromosome replication by acting as a loading factor for the essential sliding clamp and polymerase processivity factor, proliferating cell nuclear antigen (PCNA). RFC is a pentamer comprising a large subunit, Rfc1, and four small subunits, Rfc2–Rfc5. Each RFC subunit is a member of the AAA + family of ATPase and ATPase-like proteins, and the loading of PCNA onto dou- ble-stranded DNA is an ATP-dependent process. Here, we describe the properties of a collection of 38 mutant forms of the Rfc2 protein generated by pentapeptide-scanning mutagenesis of the fission yeast rfc2 gene. Each insertion was tested for its ability to support growth in fission yeast rfc2D cells lacking endogenous Rfc2 protein and the location of each inser- tion was mapped onto the 3D structure of budding yeast Rfc2. This analy- sis revealed that the majority of the inactivating mutations mapped in or adjacent to ATP sites C and D in Rfc2 (arginine finger and P-loop, respec- tively) or to the five-stranded b sheet at the heart of the Rfc2 protein. By contrast, nonlethal mutations map predominantly to loop regions or to the outer surface of the RFC complex, often in highly conserved regions of the protein. Possible explanations for the effects of the various insertions are discussed. Abbreviations PCNA, proliferating cell nuclear antigen; PSM, pentapeptide-scanning mutagenesis; RFC, replication factor C. FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS 4803 smaller subunits, Rfc2–Rfc5. Genetic analysis in yeast has shown that each of the five subunits is individually essential for chromosome replication [4]. Three RFC- like complexes have also been identified in eukaryotic cells with diverse but poorly understood roles in vari- ous aspects of checkpoint control, cohesion establish- ment and genome stability [5]. In these complexes, Rfc1 is replaced by one of Rad17, Ctf18 or Elg1. Two additional subunits are also present in the Ctf18–RFC complex. Each of the large and small RFC subunits is a mem- ber of the AAA + family of ATPase and ATPase-like proteins [6–8], and PCNA loading requires multiple ATP hydrolysis events that are discussed further below. The crystal structure of a variant form of bud- ding yeast RFC bound to PCNA has been solved in the presence of the nonhydrolysable ATP analogue ATP-cS [9]. The crystallized form of the RFC complex lacked N- and C-terminal sequences from Rfc1 (nei- ther of the missing parts of the protein is required for efficient clamp loading in vitro) and carried mutations in the so-called arginine finger motifs in Rfc2–Rfc5 [9]. The five subunits are located in a spiral arrangement in the order Rfc1–Rfc4–Rfc3–Rfc2–Rfc5. At the centre of the spiral, a cavity of sufficient size to accommodate the primer–template duplex is found. The pitch of the spiral matches that of the DNA, leading to a model in which RFC threads onto the 3¢OH group of the pri- mer like a screw-cap, preventing further primer exten- sion until PCNA loading has occurred [9]. In the yeast RFC crystal structure, a gap exists between Rfc1 and Rfc5 through which single-stranded DNA has been suggested to exit. PCNA loading onto DNA by RFC is entirely ATP dependent [1,2]. Biochemical analysis has shown that ATP binding, but not hydrolysis, is required for PCNA binding and opening of the PCNA ring. The RFC–ATP–open PCNA complex then associates with the primer–template DNA. This association appears to trigger ordered ATP hydrolysis in the different ATP- binding sites of RFC, closure of the PCNA ring and ejection of RFC from DNA, leaving the closed ring on the DNA. ATP is bound at four sites in RFC (designated ATP sites A–D) located at the subunit interfaces [1,2]. Each site is bipartite in nature, comprising elements pro- vided by adjacent subunits. Thus ATP site A is com- posed of the ATP-binding P-loop of Rfc1 (also known as RFC-A) and an arginine residue located in Rfc4 (RFC-B). The side chain of the arginine is referred to as an arginine finger and the finger protrudes into the ATP-binding site of the neighbouring subunit. The exact biochemical roles of the arginine fingers have not been precisely defined, but may involve sensing ATP binding in the P-loop and ⁄ or catalysing subsequent ATP hydrolysis. The fingers are not required for ATP binding [10]. In this study, we focus on the Rfc2 protein (also known as RFC-D). Rfc2 binds ATP in site D at the Rfc2–Rfc5 (RFC-D–RFC-E) interface and contributes an arginine finger to site C at the Rfc3–Rfc2 (RFC-C– RFC-D) interface. Biochemical analysis of the proper- ties of mutant yeast RFC complexes has shown that the Rfc2 arginine finger at site C is required for the RFC–ATP–open PCNA complex to bind DNA, lead- ing to the proposal that the conformational changes required for RFC to bind the primer–template DNA require that the Rfc2 arginine finger responds to the presence of ATP in site C [10]. ADP cannot substitute for ATP in these reactions. The Escherichia coli clamp loader, the c-complex, loads the b-sliding clamp onto DNA and is broadly analogous to RFC [1,2]. On the basis of analysis of c-complex subunits [11], three positively charged resi- dues in Rfc2 have been proposed to play a direct role in DNA binding by RFC by interacting with the phos- phate backbone of duplex DNA in the central cavity [12]. Consistent with this proposal, mutation of these arginine residues (arginines 101, 107 and 175 in yeast Rfc2) abolishes loading of RFC–ATP–open PCNA complex onto DNA [12]. Once loaded onto DNA, closure of the PCNA ring and release of RFC requires ATP hydrolysis. ATP sites C and D are particularly important for this [10]. Site C comprises the arginine finger from Rfc2 and the P-loop of Rfc3 (RFC-C), whereas site D comprises the P-loop of Rfc2 (RFC-D) and the arginine finger from Rfc5. Blocking ATP hydrolysis at sites C and D results in a significant inhibition of hydrolysis at sites A and B also, whereas blocking sites A and B has less of an effect on hydrolysis at sites C and D [10]. Taken together, these results underline the key role of Rfc2 (RFC-D) in RFC function. In this report, we describe the results of an extensive mutagenesis study of the Rfc2 protein of the fission yeast Schizosaccharomyces pombe. Using pentapeptide- scanning mutagenesis [13,14], a total of 38 mutant rfc2 alleles were isolated and tested for their ability to support chromosome replication in fission yeast cells carrying a deletion of the endogenous rfc2 + gene [15]. The majority of the inactivating mutations map in or around ATP sites C and D (arginine finger and P-loop respectively), or in the five-strand parallel b sheet at the core of Rfc2. By contrast, nonlethal mutations map predominantly to loop regions or to the outer surface of the RFC complex, often in highly Mutagenesis of the RFC small subunit Rfc2 F. C. Gray et al. 4804 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS conserved regions of the protein. Possible explana- tions for the effects of the various insertions are discussed. Results and Discussion Pentapeptide-scanning mutagenesis Pentapeptide-scanning mutagenesis (PSM) is a rapid method for the random insertion of variable five amino acid sequences into a target protein [14]. Here the system based on transposon Tn4330 was used [13]. Tn4430 contains cleavage sites for the restriction enzyme KpnI only 5 bp from its termini and duplicates five nucleotides of target site DNA during transposi- tion. By allowing Tn4430 to insert into a target DNA, then digesting the target with KpnI and re-ligating, the bulk of the transposon is deleted, leaving behind only 15 bp of sequence derived from the ends of the trans- poson and the target site duplication. Should Tn4430 insertion occur within an ORF, the 15-bp insertion will generally result in the encoded protein acquiring a five amino acid (pentapeptide) insertion. Insertion can also result in the generation of an inframe stop codon, but owing to the sequence constraints imposed by the sequence of the transposon ends, this is a relatively rare event [13]. In this study, a total of 46 independent PSM inser- tion alleles were isolated by allowing Tn4430 to trans- pose into a plasmid carrying the S. pombe rfc2 + gene (see Experimental procedures). Table 1 lists all 46 alleles and describes the location and nature of the pentapeptide insertions. DNA sequencing revealed that the 46 PSM insertions, despite their independent Table 1. Pentapeptide insertion mutants: location and inserted sequences. Insertion number Domain Location Inserted sequence Identical isolates Low-level expression High-level expression K10 I 6–7 Pro-ArgGlyThrProPro-Arg K34 + + K12 I 8–9 Asn-LysGlyTyrProAsn-Lys + + K17 I 27–28 Pro-ArgGlyThrProPro-Lys – ) K32 I 36–37 Gln-GlyTyrProSerGln-Glu + + K33 I 42–43 Val-GlyValProGlnVal-Leu + + K33DK27 I 42–46 Val-GlyValProGlnLys-Thr + + K33DK22 I 42–49 Val-GlyValProLeuLeu-Ser + + K27 I 45–46 Lys-GlyValProGlnLys-Thr K24 + + K22 I 48–49 Leu-GlyValProLeuLeu-Ser + + K23 I 49–50 Ser-LysGlyTyrProSer-Asn ) ++ K23DK14 I 49–57 Ser-LysGlyTyrPro -Phe )) K3 I 50–51 Asn-ArgGlyThrProAsn-Asn K4, K6, K7, K30, K31 + + K14 I 57 Phe-Ter K16, K29, K36 )) K18 I 64–65 Gly-ArgGlyThrProGly-Lys )) K19 I 67–68 Ser-ArgGlyThrProSer-Thr )) K9 I 82–83 Met-GlyTyrProLeuMet-Lys K25 + + K11 I 94–95 Glu-GlyValProHisGlu-Arg + + K20 I 99–100 Ile-ArgGlyThrProIle-Ile + + F46 I 124–125 Phe-GlyValProLeuPhe-Lys + + F49 I 124–125 Phe-ArgGlyThrProPhe-Lys + + F45 I 146–147 Thr-ArgGlyThrProThr-Met )) K13 I 149 Ser-Ter )) F37 I 157–158 Cys-LeuGlyTyrProCys-Leu F38 )) F39 I 170–171 Leu-ArgGlyThrProLeu-Ser )) K5 I 171–172 Ser-GlyValProLeuSer-Ser )) F42 I 174–175 Cys-ArgGlyThrProCys-Ser )) K26 II 183–184 Asp-ArgGlyThrProAsp-Asn K28 + + K1 II 195–196 Ala-GlyValProLeuAla-Ala ) ++ F41 III 242–243 Val-GlyValProArgVal-Glu ) + K35 III 251–252 Tyr-ArgGlyThrProTyr-Asn + + K15 III 254–255 Ile-ArgGlyThrProIle-Arg + + F44 III 258–259 Leu-GlyValProLeuLeu-Asp + + F47 III 305 Lys-GYPSKVQNIHETF-Ter F48, F50 + + K8 III 331–332 Asp-GlyValProLeuAsp-Leu ) + F. C. Gray et al. Mutagenesis of the RFC small subunit Rfc2 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS 4805 origins, corresponded to only 31 different alleles (inser- tions indicated in bold type in Table 1). In addition to these insertions, generated by straight- forward Tn4430 transposition and excision, seven more alleles were constructed by further manipulation of the original set of mutants. These fell into two clas- ses. Three alleles were constructed bearing short sequence deletions by ligating together the 5¢ and 3¢ portions of different insertion alleles. For example, by ligating sequences 5¢ to the KpnI site in rfc2-K33 to sequences 3¢ to the KpnI site of rfc2-K27, a new allele (rfc2-K33D27) was created encoding a protein in which those amino acid residues between the K33 and K27 insertion sites were deleted (Table 1). Similar alleles were constructed using rfc2-K33 and rfc2-K22 (rfc2- K33D22) and using rfc2-K23 and rfc2-K14 (rfc2- K23D14). Note that construction of these alleles was only possible because the parental Tn4430 insertions were in the same reading frame. Note also that in the case of rfc2-K14, the stop codon created by the origi- nal insertion is lost in fusion to rfc2-K23 (Table 1). Also, the unique KpnI site remaining following Tn4430 excision was utilized to insert extra sequences to expand the pentapeptide insertions by 5, 10 or 20 residues. Four such alleles were created in this way: sequences encoding five or ten extra amino acids were inserted into the KpnI site of rfc2-K26 to create rfc2- K26a and rfc2-K26b, sequences encoding 10 extra amino acids were inserted into rfc2-K1 to create rfc2- K1a and sequences encoding 20 extra amino acids into rfc2-K11 to create rfc2-K11a (see Table 2 for details). Combining these seven new alleles with the original 31 pentapeptide insertions gave a total of 38 mutant alleles spread throughout the rfc2 gene (Fig. 1). Expression and functional analysis in fission yeast In order to facilitate expression and analysis of the mutant proteins in fission yeast, each mutant allele was cloned 3¢ to the thiamine-repressible nmt1 pro- moter in the expression vector pREP3XH6 [16] and Table 2. Extended insertions: location and inserted sequences. Insertion number Domain Location Inserted sequence Low-level expression High-level expression K11a I 94–95 Glu-GlyValProProGlyLeu ValProProGlyLeuValPro ThrProGlyValProProGly LeuValPheHisGlu-Arg ) + K26a II 183–184 Asp-ArgGlyThrProGlyVal GlyThrProAsp-Asn ++ K26b II 183–184 Asp-ArgGlyThrProGlyGly ValGlyProGlyValGlyThr ProAsp-Asn ++ K1a II 195–196 Ala-GlyValProProValGly LeuGlyProGlyLeuValPro LeuAla-Ala ++ Fig. 1. Location of pentapeptide insertions in Rfc2 protein. Schematic representation of the fission yeast Rfc2 protein showing the location of the pentapeptide insertion mutants generated in this study. Light grey box: domain I (amino acids 1–181). White box: domain II (amino acids 182–246). Dark grey box: domain III (amino acids 247–340). Open circles: functional proteins. Grey filled circles: partly functional proteins. Black circles: non-functional proteins. See text, Table 1 and Fig. 2 for further details. Mutagenesis of the RFC small subunit Rfc2 F. C. Gray et al. 4806 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS each plasmid transformed individually into an S. pom- be rfc2 + ⁄ rfc2::ura4 + diploid strain [15]. Individual transformant colonies were then induced to sporulate and the properties of the meiotic products examined following growth on minimal medium in the presence or absence of thiamine, i.e. with the nmt1 promoter either repressed or derepressed (see Experimental pro- cedures for further details). Figure 1 summarizes the properties of the Rfc2 mutants determined from this analysis. When overexpressed (in cells grown in the absence of thiamine to fully induce the nmt1 promoter), 27 of the 38 mutant proteins were able to substitute for loss of wild-type Rfc2 function (20 insertions shown as unfilled or shaded circles in Fig. 1 plus three dele- tion alleles shown in Table 1 and four extended inser- tion alleles shown in Table 2). This decreased to 24 when thiamine was present in the growth medium, because three of the mutant alleles could only rescue when the Rfc2 protein was overproduced (shaded cir- cles in Fig. 1, plus the K11a extended insertion, see Table 2). Note that viable mutant strains were also examined for increased sensitivity to the DNA repli- cation inhibitor hydroxyurea, which might indicate a specific defect in replication checkpoint function. However, none of the strains displayed this property (data not shown). In the following discussion, the mutations are considered in three groups dependent on their location in the 3D structure of the Rfc2 protein. Mapping the insertions onto the structure of Rfc2: insertions into the N-terminal AAA + domain The 3D structure of the Rfc2 protein comprises three separate domains. Domains I and II together form the AAA + ATPase module, whereas domain III forms part of the RFC circular collar and is unique to clamps loaders [9]. Based on alignment with the bud- ding yeast Rfc2 protein, the N-terminal AAA + domain of fission yeast Rfc2 encompasses amino acid residues 1–181 (Fig. 1). In the Rfc2 crystal structure (PDB entry 1SXJ) this domain comprises 11 a-helical segments and 5 b strands [9]. Here the helices are des- ignated a1–a3, a3¢ and a4–a10 and the strands b1–b5 (Fig. 2A). Note that in PDB entry 1SXJ, the a3–a3¢ segment is designated as a single a helix (helix 69) despite there being clear structural discontinuity at res- idues 84–85. For this reason, residues 70–89 are denoted here as two separate a helices, a3 and a3 ¢ (Fig. 2A). Twenty-three insertions were located in this domain of the protein, the most N-terminal (K10) being located between residues 6 and 7, and the most C-ter- minal (F42) between residues 174 and 175. The aver- age distance between the insertions is eight residues, but some clustering was observed (Figs 1 and 2A) and the longest stretch without an insertion is 25 residues (amino acids 100–124 inclusive). All 10 inactivating mutations identified in this study were located in the N-terminal AAA + domain (Fig. 1). Three of these (insertions K17, K18, K19) were located in or around the P-loop that forms part of RFC ATP-binding site D (Fig. 3, left-hand panel). Insertions K18 and K19 are located in helix a3 close to the a-phosphate of ATP, whereas insertion K17 maps to the unstructured loop region that lies between a1 and a2, close to the adenosine of bound ATP (Figs 2A and 3). It is not unreasonable to postulate that these three mutations exert their effects by dis- rupting ATP binding in site D, thereby rendering the RFC complex inactive. Four of the remaining inactivating mutations (K5, F39, F42, F45) cluster close to the Rfc2 arginine finger (Fig. 3, right-hand panel) that forms part of ATP- binding site C at the interface of Rfc2 and Rfc3 (also known as RFC-C). As noted above, the arginine finger at this site is crucial for the association of the RFC– ATP–open PCNA complex with the primed template, and subsequent ATP hydrolysis, but not for ATP binding [10]. Insertions K5 and F42 flank the highly conserved SRC motif containing the arginine (RFC box VII), whereas insertion F39 is located one amino acid N-terminal to SRC. Both F39 and K5 insertions disrupt the a10 helix, whereas insertion F45 disrupts a8. These insertions are likely to affect the positioning of the arginine finger in site D, thereby blocking DNA binding by RFC–PCNA. Of the remaining three inactivating mutations, one (insertion F37) maps to b-strand b4 in the five-strand parallel b sheet that comprises the core of domain I (Figs 2A and 3A). The b4 strand is the central strand in the sheet [9]; it is likely that disruption of this strand by pentapeptide insertion will affect the entire sheet (Fig. 3B). The final two inactivating mutations in domain I (K13, K14) cause premature termination of the Rfc2 polypeptide chain. The remaining 13 insertions in domain I failed to dis- rupt Rfc2 function. Given that all the pentapeptide insertion sequences include a proline residue that might be expected to have a significant effect on the secondary structure in the vicinity of the insertion, it is perhaps sur- prising that only 8 of the 29 pentapeptide insertions investigated in this study (< 30% of the total) abolished Rfc2 function altogether (black circles in Fig. 1; these figures exclude the two insertions that resulted in prema- F. C. Gray et al. Mutagenesis of the RFC small subunit Rfc2 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS 4807 A Fig. 2. Location of insertion sites in conserved regions. Location of insertions in N-terminal AAA + domain (domain I, shown in A), central domain (domain II, B) and C-terminal collar domain (domain III, C) of fission yeast Rfc2. The aligned sequences of 10 Rfc2 proteins from diverse eukaryotic species are shown, corresponding to S. pombe Rfc2 residues 1–181 (domain I), 182–241 (domain II) and 242–340 (domain III). Insertions resulting in premature termination of translation are underlined. Abbreviations and RefSeq accession numbers: Sp (S. pombe, NP_594540), Sc (Saccharomyces cerevisiae, NP_012602), Hs (Homo sapiens, NP_852136), Dm (Drosophila melanogaster , NP_573245), Gg (Gallus gallus, NP_001006550), Xl (Xenopus laevis, NP_001082757), Dr (Danio rerio, NP_999902), At (Arabadopsis thaliana, NP_564148), Dd (Dictyostelium discoideum, XP_637900) and Ez (Encephalitozoon cuniculi, XP_955685). Positions conserved in at least eight of the ten pro- teins are shown boxed. Secondary structure elements in S. cerevisiae Rfc2 are shown above (based on PDB entry 1SXJ, see Fig. 3). The alignments were generated using CLUSTAL X [22,23]. Mutagenesis of the RFC small subunit Rfc2 F. C. Gray et al. 4808 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS ture polypeptide chain termination). Related studies of other proteins [13,17–19] have revealed similar ratios of functional to nonfunctional mutants, however, so this phenomenon is not specific to RFC. Four of the nonlethal insertions in domain I mapped within a helices (Fig. 2A): two in the a2 helix (inser- tions K33 and K27), one at the extreme C-terminal end of a3¢ (K9) and one in the middle of a4, close to the border of the a4 and a5 helices on the inner face of the RFC complex (K20). Two of the three con- served arginine residues implicated in DNA binding by Rfc2 (R101 and R107 in the budding yeast Rfc2 pro- tein, R95 and R101 in the fission yeast protein) are located in a4 [12]. If the K20 insertion disrupts either of these interactions, it does so without markedly dis- rupting RFC function. Simultaneous mutation of the three conserved arginine residues to alanines in the budding yeast Rfc2, Rfc3 and Rfc4 proteins (such that the resulting RFC complex carried nine arginine-to- alanine substitutions) abolished DNA binding, but the relative contributions of the individual subunits or the individual arginines in each subunit were not tested [12]. Our results may imply that the a4 arginines R95 and R101 are not absolutely required for in vivo DNA binding by Rfc2 or indeed that DNA binding by Rfc2 is nonessential for RFC complex function. Further biochemical and genetic analysis is required to resolve this issue. However, as a starting point, we used site- directed mutagenesis to construct seven new single, double and triple arginine-to-alanine mutations in fis- sion yeast Rfc2 (designated Rfc2–S1 to Rfc2–S7), at R95 and R101 in the a4 helix and at R165, the third residue previously implicated in DNA binding [12]. Each mutant allele (Table 3) was expressed from the nmt1 promoter in haploid rfc2D cells exactly as described above for the pentapeptide insertion mutants. The results of this are summarized in Table 3. All seven mutant proteins, including Rfc2–S7 with triple K95A, K101A and K165A substitutions, were able to rescue for loss of Rfc2 function when expressed at low level (nmt1 promoter repressed by the B C Fig. 2. Continued. F. C. Gray et al. Mutagenesis of the RFC small subunit Rfc2 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS 4809 presence of thiamine in the growth medium, see above). Clearly, if DNA binding by Rfc2 does play an essential role in vivo, then the three arginines are not required for these interactions to occur. Interestingly, when plated on plates containing 10 mm hydroxyurea, cells expressing Rfc2–S7 were somewhat less elongated and more heterogeneous in appearance than cells expressing wild-type Rfc2 or any of the single or dou- ble arginine-to-alanine mutants Rfc2–S1 to Rfc2–S6, suggesting that Rfc2 function may be mildly impaired in these cells. The remaining nine insertions are located in loop regions where it is reasonable to expect that the inserted sequences would be tolerated without disrupt- ing RFC function (Fig. 2A): insertions are found at the N-terminus of the protein (K10, K12), between a1 and a2 (K32), between a2 and b1 (K22, K23 and K3), between b2 and a4 (K11) and at the N-terminal end of b3 (F46, F49). Two domain I deletion constructs (K33DK27 and K33D22) were also viable, despite the latter appearing to delete almost all of the a2 helix (Table 1). This helix contains a pair of very well-conserved basic amino acids (lysines 44 and 45 in fission yeast Rfc2) and the K33D22 deletion removes these altogether, replacing the sequence LKKT with GVP (see Fig. 2A and Table 1). The function of a2 helix and the conserved basic residues is unclear: in the budding yeast RFC crystal structure, the side chains of the arginines pro- trude from the surface of the RFC complex (Fig.3A), perhaps suggesting an involvement in mediating protein–protein or protein–DNA interactions on the surface of the complex [9]. The a2 helix is not close to the proposed exit path for single-stranded DNA, how- ever, suggesting that the latter possibility is unlikely. The third deletion mutation tested here (K23DK14) deletes much of the a2–b1 loop region and the b1 strand and was not viable. Presumably deleting b1 dis- rupts the structure of the b sheet at the core of domain I (Fig. 3). As noted above, insertion F37 in the b4 Table 3. Arginine-to-alanine DNA-binding mutants. Allele Domain Amino acid changes Low-level expression High-level expression S1 I R95A + + S2 I R101A + + S3 I R165A + + S4 I R95A-R101A + + S5 I R95A-R165A + + S6 I R101A-R165A + + S7 I R95A-R101A-R165A + + A B Fig. 3. Mapping of insertion sites onto budding yeast Rfc2 protein structure. (A) 3D structure of the budding yeast Rfc2 protein bound to ATP-cS [9]. The locations of the eight inactivating insertion mutations are indicated by the blue circles. The locations of the three arginines implicated in DNA binding are shown as R95, R101 and R165 (fission yeast numbering). The coordinates of Rfc2 were extracted from PDB file 1SXJ and the structure drawn using MACPYMOL 0.99 (DeLano Scientific, Palo Alto, CA, USA). ATP-cS is shown in blue and the locations of the three domains of the protein indicated. The images of the left and right are rotated by  180° relative to one another. (B) Topology diagram of Rfc2 b sheet. For simplicity, a helices are not shown. Mutagenesis of the RFC small subunit Rfc2 F. C. Gray et al. 4810 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS strand also eliminates Rfc2 function (Table 1). No other insertions in the b sheet were identified in this study. Mapping the insertions onto the structure of Rfc2: insertions into the central AAA + domain The a-helical central domain (domain II) of fission yeast Rfc2 encompasses amino acid residues 182–246 (Figs 1 and 2B). In the yeast RFC structure [9], this comprises five a helices (here designated a11–a15). Only three of the insertions were located in this domain (Fig. 2B): two of these insertions map within a11 (K26, K1) and the third in a15 (F41). None of the mutants abolishes RFC function but rescue of rfc2D cells by Rfc2–F41 was seen only when the cor- responding mutant gene was overexpressed from the nmt1 promoter, suggesting that the Rfc2–F41 protein is functionally impaired (Fig. 1). Helix a11 lies on the outer surface of the RFC complex. Expanding the K1 and K26 insertions by addition of sequences encoding a additional 5 or 10 amino acids (see Table 2 for details) did not disrupt Rfc2 function either. Mapping the insertions onto the structure of Rfc2: insertions into the C-terminal collar domain The C-terminal domain (domain III) of Rfc2 encom- passes amino acid residues 247–340 (Fig. 2C) and com- prises six a helices (a16–a20). These domains form a right-handed spiral collar structure from which the AAA + domains I and II appear to hang [9]. Five insertions map within domain III (Fig. 2C). Insertions K35, K15 and F44 map within a16, which is located on the outer surface of the RFC complex (Fig. 3A), whereas insertion K8 maps centrally within a20 located at the Rfc2–Rfc5 interface. None of these mutations abolishes Rfc2 function but, as with Rfc2– F41 described above, Rfc2–K8 was required to be overexpressed in order to rescue rfc2D cells, implying that the K8 insertion causes a degree of functional impairment (Fig. 1). The insertion in F47 is located in a19 (Fig. 2C), but this mutation is unusual in that Tn4430 transposition and subsequent restriction enzyme cleavage and re-ligation left behind a 16 bp, rather than a 15 bp, insertion in the DNA sequence. This produces a frame-shift mutation that causes termination of Rfc2 following the addition of 13 random amino acids after lysine 305 (shown in single-letter code in Table 1). Despite this, however, the Rfc2–F47 protein is func- tional, even when expressed at normal levels (Fig. 1 and Table 1). It can be concluded from this that a19 and a20 are not required for Rfc2 function, regardless of the negative effect on the Rfc2 protein observed with insertion K8 in a20 described above. One possible explanation for the apparent discrepancy is that the K8 pentapeptide insertion may cause structural disrup- tion that is incompatible with collar formation and RFC function (consistent with the location of the a20 helix at the Rfc2–Rfc5 interface), whereas the trun- cated Rfc2–F47 protein forms RFC complexes nor- mally. Further analysis of this point will require detailed biochemical analysis of the complex forming properties of the mutant proteins. Conclusions This study confirms the importance of ATP site C (which involves the Rfc2 arginine finger) and ATP site D (involving the Rfc2 P-loop) for RFC function in vivo, and demonstrates that three arginine residues (R95, R101 and R165) previously implicated in DNA binding by Rfc2 are nonessential in vivo. In addition, several highly conserved regions of the Rfc2 protein that are surprisingly tolerant of pentapeptide insertions have been identified. Future work will focus on investi- gating in greater depth the roles these conserved regions play in RFC function. Experimental procedures Bacterial and yeast strains and media E.coli DH5a (Stratagene, La Jolla, CA, USA) was used for routine cloning steps and FH1046 and DS941 for pentapep- tide mutagenesis [13]. E. coli was cultured on LB medium. S. pombe rfc2 + ⁄ rfc2::ura4 + leu1-32 ⁄ leu1-32 ura4-D18 ⁄ ura4- D18 ade6-M210 ⁄ ade6-M216 h ) ⁄ h + [15] was used for func- tional testing of rfc2 mutations. S. pombe was cultured on YE, EMM or ME media [20] as required, and transformed by electroporation [21]. Pentapeptide mutagenesis To mutagenise rfc2 + using the pentapeptide insertion method, an rfc2 + cDNA was first amplified by PCR from an S. pombe cDNA library using oligonucleotides SPRFC2– 5BAM (oligo sequence with BamHI site in lower case and rfc2 + start codon underlined: 5¢-TTGGTTGGggatcc AA ATGTCTTTCTTTGCTCCA-3¢) and SPRFC2–3BAM (oligo sequence with BamHI site in lower case and rfc2 + stop codon underlined: 5¢-TTGGTTGGggatccTTTTCAATGTA TAGA CTAGC-3¢), restricted with BamHI and cloned into plasmid pBR322 to generate pBR322–Rfc2. The sequence F. C. Gray et al. Mutagenesis of the RFC small subunit Rfc2 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS 4811 of the rfc2 + insert was confirmed by sequencing. PSM was performed using the Tn4430 system [13]. pBR322–Rfc2 was transformed into E. coli FH1046 and individual transfor- mant colonies were mated with E. coli DS941 after which the mating mixes were plated on medium to select for clones containing pBR322-Rfc2 carrying the Tn4430 transposon. Isolates containing a Tn4430 insertion within the rfc2 region were then identified by colony PCR and plasmid DNA prepared. This was restricted by KpnI and self-ligated to remove the bulk of the transposon, before being trans- formed into E. coli DH5a. Thirty-four individual trans- poson-free clones were isolated in this manner and the position of the 15-bp insertion determined by KpnI restric- tion site mapping and DNA sequencing. Twelve dupli- cated mutants were discarded at this point, leaving a collection of 22 different pentapeptide insertion alleles (rfc2-K1 to rfc2-K36). A second mutagenesis was carried out using pBR322– Rfc2–SB as the target plasmid. This plasmid was created by cloning the SalI–BamHI fragment carrying the 600 bp 3¢ region of the rfc2 + ORF from pBR322-Rfc2 into pBR322. pBR322–Rfc2–SB was then subjected to pentapep- tide mutagenesis as above. Twelve independent transposon- containing isolated were identified with insertions in the SalI–BamHI region; sequence analysis showed that these represented only nine different insertions. These alleles were designated rfc2-F37 to rfc2-F49. Generating deletion and insertion alleles Deletions alleles were constructed by digesting pREP3XH6– Rfc2–K plasmids with KpnI (a second KpnI site is located in the LEU2 gene) and ligating together complementary pieces. Three alleles were constructed in this way: rfc2-K33 ⁄ K27, rfc2-K33 ⁄ K22 and rfc2-K23 ⁄ K14. Four insertion alleles were constructed by digesting the relevant pREP3XH6–Rfc2–K plasmids with KpnI and ligating in oligonucleotide duplexes constructed by annealing together the following complemen- tary pairs of oligonucleotides, either SPRFC2-F1 (5¢-CCC CGGGGTTGGTAC-3¢) and SPRFC2-F2 (5¢-CAACCC CGGGGGATG-3¢) to produce a five amino acid extension (insertion K26a) or SPRFC2-F3 (5¢-CCCCGGTGGGGT TGGGCCCGGGGTTGGTAC-3¢) and SPRFC2-F4 (5¢-CA ACCCCGGGCCCAACCCCACCGGGGGTAC-3¢) to gen- erate a 10 amino acid extension (insertions K1a and K26b). Insertion K11a resulted from the fortuitous ligation of four copies of the F1 ⁄ F2 duplex. Site-directed mutagenesis Site-directed mutagenesis was performed using the PCR overlap extension mutagenesis with Pfu DNA polymerase (Promega, Madison, WI, USA) and plasmid pREP3XH6– Rfc2 as template. Oligonucleotide sequences are available from the corresponding author on request. Following the second round of PCR, the product was digested with BamHI, re-cloned into pREX3XH6 and sequenced to con- firm the absence of PCR errors. The resulting plasmids were then transformed into yeast as described below. Expression in fission yeast To express the mutant rfc2-K alleles in S. pombe, each was transferred, as a BamHI fragment, into plasmid pREP3XH6 [16]. The resulting pREP3XH6–Rfc2 plasmids express Rfc2 with a 13 amino acid N-terminal extension that includes a hexahistidine tag (sequence of extension: MRGSHH HHHHGIQ). For the rfc2-F alleles, a SalI– EcoRV fragment from pBR322–Rfc2–SB (the EcoRV site is located in the vector sequence,  200 bp from the BamHI site) was transferred into plasmid pREP3XH6–Rfc2 that had been cut with SalI and SmaI. The resulting plasmids were then transformed into S. pombe rfc2 + ⁄ rfc2::ura4 + leu1-32 ⁄ leu1-32 ura4-D18 ⁄ ura4-D18 ade6-M210 ⁄ ade6-M216 h ) ⁄ h + [15] by electroporation [21] and transformants obtained on EMM medium. Individual colonies were then patched overnight at 32 °C on ME medium to induce spor- ulation, before being treated overnight with helicase to break down the asci walls and eliminate vegetative cells. Spores were then washed with water before being plated on EMM plates supplemented with adenine (EMM + A), uracil and adenine (EMM + AU), adenine and thiamine (EMM + AT) and adenine, uracil and 5 lm thiamine (EMM + AUT) at 23, 32 and 36.5 °C. Leucine was omitted from all plates to facilitate selection of pREP3X plasmids which carry the LEU2 selectable marker. Adenine is required to permit the growth of haploid cells either the ade6-M210 or ade6-M216 alleles. The addition of uracil permits growth of rfc2 + haploids; in the absence of uracil, only rfc2::ura4 + haploids expressing functional Rfc2 proteins can grow. The presence of 5 lm thiamine represses the nmt1 promoter in pREP3X, reducing rfc2 expression by a factor of 80–100 compared with cells grown on EMM without thiamine. Acknowledgements We would like to thank Dr Finbarr Hayes (University of Manchester, UK) for supplying the strains necessary for PSM. This research was funded by a Wellcome Trust Senior Fellowship in Basic Biomedical Research. KAW was the recipient of a BBSRC-funded postgrad- uate studentship. References 1 Indiani C & O’Donnell M (2006) The replication clamp-loading machine at work in the three domains of life. Nat Rev Mol Cell Biol 7, 751–761. Mutagenesis of the RFC small subunit Rfc2 F. C. Gray et al. 4812 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS [...]... for replication factor C in DNA replication checkpoint function in fission yeast Nucleic Acids Res 27, 462–469 Gray FC & MacNeill SA (2000) The Schizosaccharomyces pombe rfc3+ gene encodes a homologue of the human hRFC36 and Saccharomyces cerevisiae Rfc3 subunits of replication factor C Curr Genet 37, 159–167 Hayes F, Hallet B & Cao Y (1997) Insertion mutagenesis as a tool in the modification of protein...F C Gray et al 2 Johnson A & O’Donnell M (2005) Cellular DNA replicases: components and dynamics at the replication fork Annu Rev Biochem 74, 283–315 3 Moldovan GL, Pfander B & Jentsch S (2007) PCNA, the maestro of the replication fork Cell 129, 665–679 4 MacNeill SA & Burgers PMJ (2000) Chromosomal DNA replication in yeast: enzymes and mechanisms In The Yeast Nucleus (Fantes P & Beggs... Martin J (2002) AAA proteins Curr Opin Struct Biol 12, 746–753 9 Bowman GD, O’Donnell M & Kuriyan J (2004) Structural analysis of a eukaryotic sliding DNA clamp-clamp loader complex Nature 429, 724–730 10 Johnson A, Yao NY, Bowman GD, Kuriyan J & O’Donnell M (2006) The replication factor C clamp loader requires arginine finger sensors to drive DNA binding and proliferating cell nuclear antigen loading... Hayes F (1997) Pentapeptide scanning mutagenesis: random insertion of a variable Mutagenesis of the RFC small subunit Rfc2 14 15 16 17 18 19 20 21 22 23 five amino acid cassette in a target protein Nucleic Acids Res 25, 1866–1867 Hayes F & Hallet B (2000) Pentapeptide scanning mutagenesis: encouraging old proteins to execute unusual tricks Trends Microbiol 8, 571–577 Reynolds N, Fantes PA & MacNeill SA (1999)... Insertion mutagenesis as a tool in the modification of protein function Extended substrate specificity conferred by pentapeptide insertions in the omega-loop of TEM-1 beta-lactamase J Biol Chem 272, 28833–28836 Cao Y, Hallet B, Sherratt DJ & Hayes F (1997) Structure–function correlations in the XerD site- speci c recombinase revealed by pentapeptide scanning mutagenesis J Mol Biol 274, 39–53 Zhang Y, Altshuller... loading J Biol Chem 281, 35531–35543 11 Goedken ER, Kazmirski SL, Bowman GD, O’Donnell M & Kuriyan J (2005) Mapping the interaction of DNA with the Escherichia coli DNA polymerase clamp loader complex Nat Struct Mol Biol 12, 183–190 12 Yao NY, Johnson A, Bowman GD, Kuriyan J & O’Donnell M (2006) Mechanism of proliferating cell nuclear antigen clamp opening by replication factor C J Biol Chem 281, 17528–17539... receptor regulation by a phospholipase D1 mutant unresponsive to protein kinase C EMBO J 18, 6339–6348 Moreno S, Klar A & Nurse P (1991) Molecular genetic analysis of fission yeast Schizosaccharomyces pombe Methods Enzymol 194, 795–823 Prentice HL (1992) High-efficiency transformation of Schizosaccharomyces pombe by electroporation Nucleic Acids Res 20, 621 Thompson JD, Gibson TJ, Plewniak F, Jeanmougin... (2004) The PCNA-RFC families of DNA clamps and clamp loaders Prog Nucleic Acid Res Mol Biol 78, 227–260 6 Neuwald AF, Aravind L, Spouge JL & Koonin EV (1999) AAA(+): a class of chaperone-like ATPases associated with the assembly, operation, and disassembly of protein complexes Genome Res 9, 27–43 7 Ogura T & Wilkinson AJ (2001) AAA+ superfamily ATPases: common structure – diverse function Genes Cells... Higgins DG (1997) The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools Nucleic Acids Res 25, 4876–4882 Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R et al (2007) Clustal W and Clustal X version 2.0 Bioinformatics 23, 2947–2948 FEBS Journal 276 (2009) 4803–4813 ª 2009 The. .. McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R et al (2007) Clustal W and Clustal X version 2.0 Bioinformatics 23, 2947–2948 FEBS Journal 276 (2009) 4803–4813 ª 2009 The Authors Journal compilation ª 2009 FEBS 4813 . Inactivating pentapeptide insertions in the fission yeast replication factor C subunit Rfc2 cluster near the ATP-binding site and arginine finger motif Fiona. produce a five amino acid extension (insertion K26a) or SPRFC2-F3 (5¢-CCCCGGTGGGGT TGGGCCCGGGGTTGGTAC-3¢) and SPRFC2-F4 (5¢-CA ACCCCGGGCCCAACCCCACCGGGGGTAC-3¢)