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 structu
Trang 1yeast replication factor C subunit Rfc2 cluster near
the ATP-binding site and arginine finger motif
Fiona C Gray1,2, Kathryn A Whitehead1,* and Stuart A MacNeill1,2,
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 slidform-ing 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.
Trang 2smaller 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 pripri-mer
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
Trang 3conserved regions of the protein Possible
explana-tions for the effects of the various inserexplana-tions 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
Trang 4origins, 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
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 K26 to create rfc2-K26a and rfc2-K26b, sequences encoding 10 extra amino acids were inserted into 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.
ValProProGlyLeuValPro ThrProGlyValProProGly LeuValPheHisGlu-Arg
GlyThrProAsp-Asn
ValGlyProGlyValGlyThr ProAsp-Asn
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.
Trang 5each 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, resres-idues 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
Trang 6Fig 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].
Trang 7ture 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.
Trang 8presence 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
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.
Trang 9strand 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
enpasses 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 formadisrup-tion 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.coliDH5a (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 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 AAATGTCTTTCTTTGCTCCA-3¢) and SPRFC2–3BAM (oligo sequence with BamHI site in lower case and rfc2+stop codon underlined: 5¢-TTGGTTGGggatccTTTTCAATGTA TAGACTAGC-3¢), restricted with BamHI and cloned into plasmid pBR322 to generate pBR322–Rfc2 The sequence
Trang 10of 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
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 32C 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 36.5C 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