Accepted Manuscript Title: DNA polymerase 3 →5 exonuclease activity; different roles of the beta hairpin structure in family-B DNA polymerases Author: Hariyanto Darmawan Melissa Harrison Linda J. Reha-Krantz PII: DOI: Reference:

S1568-7864(15)00050-6 http://dx.doi.org/doi:10.1016/j.dnarep.2015.02.014 DNAREP 2065

To appear in:

DNA Repair

Received date: Revised date: Accepted date:

4-11-2014 12-2-2015 13-2-2015

Please cite this article as: H. Darmawan, M. Harrison, L.J. Reha-Krantz, DNA polymerase 3 rightarrow5 exonuclease activity; different roles of the beta hairpin structure in family-B DNA polymerases, DNA Repair (2015), http://dx.doi.org/10.1016/j.dnarep.2015.02.014 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1 2

DNA polymerase 3’→5’ exonuclease ac vity; different roles of the beta hairpin structure in family-B DNA polymerases

4

Hariyanto Darmawan, Melissa Harrison, Linda J. Reha-Krantz*

cr

5

ip t

3

Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2E9

7 8

*Corresponding author at: Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada T6G 2E9. Tel,: +1 780 492 5383; fax: +1 780 492 9234.

9

E-mail address: [email protected]

an

us

6

10

12

M

11

Highlights

 T4 DNA polymerase proofreading requires a beta hairpin structure

14

 Yeast DNA polymerase delta proofreading does not require the hairpin

15

 The yeast DNA polymerase delta hairpin is needed for fitness

16

 DNA polymerase evolution

17

 Mismatch repair regulates the checkpoint when DNA pol delta replication is hindered

19

te

Ac ce p

18

d

13

20

Keywords:

21

DNA polymerase proofreading

22

Phosphonoacetic acid sensitivity

23

Hydroxyurea sensitivity

24

DNA polymerase evolution

25

Mismatch repair and checkpoint response 1 Page 1 of 48

26

DNA polymerase mutations in cancer

27 28

ABSTRACT

ip t

29

Proofreading by the bacteriophage T4 and RB69 DNA polymerases requires a β hairpin structure

31

that resides in the exonuclease domain. Genetic, biochemical and structural studies demonstrate

32

that the phage β hairpin acts as a wedge to separate the primer-end from the template strand in

33

exonuclease complexes. Single amino acid substitutions in the tip of the hairpin or deletion of

34

the hairpin prevent proofreading and create “mutator” DNA polymerases. There is little known,

35

however, about the function of similar hairpin structures in other family B DNA polymerases.

36

We present mutational analysis of the yeast (Saccharomyces cerevisiae) DNA polymerase δ

37

hairpin. Deletion of the DNA polymerase δ hairpin (hp) did not significantly reduce DNA

38

replication fidelity; thus, the β hairpin structure in yeast DNA polymerase δ is not essential for

39

proofreading. However, replication efficiency was reduced as indicated by a slow growth

40

phenotype. In contrast, the G447D amino acid substitution in the tip of the hairpin increased

41

frameshift mutations and sensitivity to hydroxyurea (HU). A chimeric yeast DNA polymerase δ

42

was constructed in which the T4 DNA polymerase hairpin (T4hp) replaced the yeast DNA

43

polymerase δ hairpin; a strong increase in frameshift mutations was observed and the mutant

44

strain was sensitive to HU and to the pyrophosphate analog, phosphonoacetic acid (PAA). But

45

all phenotypes - slow growth, HU-sensitivity, PAA-sensitivity, and reduced fidelity, were

46

observed only in the absence of mismatch repair (MMR), which implicates a role for MMR in

47

mediating DNA polymerase δ replication problems. In comparison, another family B DNA

48

polymerase, DNA polymerase , has only an atrophied hairpin with no apparent function. Thus,

Ac ce p

te

d

M

an

us

cr

30

2 Page 2 of 48

49

while family B DNA polymerases share conserved motifs and general structural features, the β

50

hairpin has evolved to meet specific needs.

51

1. Introduction

The role of the 3’ → 5’ exonuclease activity associated with many DNA polymerases

cr

53

ip t

52

was revealed in two seminal papers published in 1972. Brutlag and Kornberg [1] reported that

55

the 3’→5’ exonuclease activity of Escherichia coli DNA pol I was specific for mismatches at the

56

3’-end of the primer strand, which suggested that the 3’ → 5’ exonuclease of E. coli DNA pol I

57

functioned in error avoidance. A similar “proofreading” role was proposed for the 3’ → 5’

58

exonuclease activity of bacteriophage T4, and evidence was provided by biochemical studies of

59

“mutator” (increased replication errors) and “antimutator” (decreased replication errors) T4 DNA

60

polymerases [2]. Speyer and Drake discovered that several alleles of the T4 DNA polymerase

61

strongly altered replication fidelity by increasing or decreasing the spontaneous mutation

62

frequency [3-5]. While differences in the accuracy of nucleotide incorporation could explain the

63

mutator and antimutator phenotypes, Muzyczka et al. [2] found that the ratio of nuclease to

64

polymerase (N/P) activity was perturbed in the mutant DNA polymerases; the N/P ratio was

65

reduced for mutator DNA polymerases and increased for antimutator DNA polymerases. Further

66

characterization of mutator and antimutator DNA polymerases confirmed that mutator DNA

67

polymerases proofread mismatched primer-ends less efficiently while antimutator DNA

68

polymerases proofread mismatched and even matched primer-ends more efficiently than

69

observed for the wild type T4 DNA polymerase [2,6].

Ac ce p

te

d

M

an

us

54

70

DNA polymerase proofreading, however, requires more than removal of an incorrect

71

nucleotide. The primer-end with the terminal incorrect nucleotide must be physically moved 3 Page 3 of 48

from the polymerase to the exonuclease active site where the terminal nucleotide is removed,

73

then the trimmed primer-end is returned to the polymerase active site [7,8]. This is polymerase-

74

to-exonuclease and exonuclease-to-polymerase switching. Many details of the proofreading

75

pathway catalyzed by the T4 DNA polymerase have been revealed by a large collection of

76

mutants, which have amino acid substitutions throughout the DNA polymerase [7-18].

77

Proofreading is clearly a dynamic process that involves participation by all five protein domains.

cr

The first step in the proofreading pathway is recognition of a newly misincorporated

us

78

ip t

72

nucleotide at the primer-end in the polymerase active site, and then the “decision” is made to

80

extend the mismatch by incorporation of the next nucleotide or to transfer the primer-end to the

81

exonuclease active site [7,19,20]. Formation of exonuclease complexes requires separation of

82

the primer-end from the template strand and transfer of the primer-end to the exonuclease active

83

site, which is ~40 Å from the polymerase active site. T4 and RB69 DNA polymerases require a

84

β hairpin structure in the exonuclease domain for these tasks; the hairpin acts as a wedge and

85

inserts between the template and primer strands (14,21; Fig. 1A). T4 DNA polymerase mutants

86

with single amino acid substitutions in the tip of the hairpin, e.g. G255S, can degrade single-

87

stranded DNA, but have much less activity with double-stranded DNA unless there is a

88

preformed strand separation of at least three terminal base pairs [14,22]; this is because the

89

G255S-DNA polymerase has reduced ability to form stable, strand-separated exonuclease

90

complexes [22-24]. Single amino acid substitutions for Y254 in the T4 DNA polymerase and

91

deletion of the hairpin (hp) in the RB69 DNA polymerase also produce proofreading-deficient

92

DNA polymerases [25-27]. The RB69 hp mutant appears incapable of forming exonuclease

93

complexes with double-stranded, primer-template DNA, but still retains exonuclease activity

94

with single-stranded DNA [27]. Aller et al. [28] captured a well-ordered exonuclease complex

Ac ce p

te

d

M

an

79

4 Page 4 of 48

structure with the RB69 DNA polymerase that confirms the role of the β hairpin structure in

96

strand separation and in the stabilization of the primer-end in the exonuclease active site (Fig.

97

1B). Residues in the tip of the RB69 DNA polymerase hairpin, Y257 and M256, which are

98

analogous to T4 DNA polymerase residues Y254 and M253, are wedged firmly between the

99

primer and template strands. Note that structural studies were performed with the phage RB69

ip t

95

DNA polymerase, which is closely related in sequence (61%) and function to the phage T4 DNA

101

polymerase [29]. Thus, the inability to form exonuclease complexes with primer-template DNA

102

severely reduces exonuclease activity even though exonuclease activity with single-stranded

103

DNA substrates is retained.

us

an

If the mismatched primer-end cannot be extended, but exonuclease complexes also

M

104

cr

100

cannot be formed readily, then continued replication is stalled at the crossroads of the nucleotide

106

incorporation and proofreading pathways and the reverse reaction, pyrophosphorolysis [20,24].

107

Crossroad complexes are pre-translocation (pre-T) complexes, which are formed after nucleotide

108

incorporation and dissociation of pyrophosphate, just before the DNA polymerase translocates to

109

be in position to bind the next incoming nucleotide and form post-translocation (post-T)

110

complexes.

111

of post-T complexes is hindered if the primer-end is not correct. This situation triggers initiation

112

of the proofreading pathway. However, if exonuclease complexes also cannot be formed readily,

113

then formation of pre-T complexes is favored. Pre-T complexes can be detected by anti-viral

114

drugs that mimic pyrophosphosphate, e.g. phosphonoacetic acid (PAA). Pre-T complexes are

115

trapped by binding PAA, which hinders further DNA replication [20,31]. The T4 G255S-DNA

116

polymerase is PAA-sensitive [18] as expected if polymerase-to-exonuclease switching is

117

compromised.

Ac ce p

te

d

105

Normally, pre- and post-T complexes are in rapid equilibrium [30], but formation

5 Page 5 of 48

118

After exonuclease complexes are formed, the terminal residue is removed rapidly [7,32]. T4 DNA polymerase mutants that have amino acid substitutions for conserved active site

120

residues are exonuclease and proofreading deficient as observed for other proofreading DNA

121

polymerases [33]. The last step in the proofreading pathway is to return the trimmed primer-end

122

from the exonuclease to the polymerase active site; this step is rapid for the wild type T4 DNA

123

polymerase, but slow for the W213S-DNA polymerase [34].

cr

Although much is known about proofreading by the T4 and RB69 DNA polymerases,

us

124

ip t

119

less is known about proofreading by other family B DNA polymerases. However, because

126

family B DNA polymerases share conserved motifs and overall general structure, similar

127

function is expected. For example, the L412M substitution in Motif A in the polymerase active

128

site of the T4 DNA polymerase produces a mutator DNA polymerase with enhanced ability to

129

extend mismatched primer-templates and the mutant is sensitive to PAA [12,18,20]. We

130

expected that both phenotypes would be observed in yeast (Saccharomyces cerevisiae) for the

131

L612M-DNA polymerase δ, which has the analogous L-to-M substitution in Motif A. The yeast

132

L612M-DNA polymerase δ is an excellent phenocopy of the T4 L412M-DNA polymerase;

133

reduced replication fidelity and PAA-sensitivity are observed for both mutant polymerases

134

[35,36]. But functional similarities were not observed for T4 DNA polymerase and yeast DNA

135

polymerase δ hairpin mutants. The G447S substitution in the tip of the yeast DNA polymerase δ

136

hairpin, proposed to be analogous to the G255S substitution in the T4 DNA polymerase (Table

137

1), had little effect on replication fidelity [37]. This is surprising because the yeast DNA

138

polymerase  hairpin resembles the phage T4 and RB69 DNA polymerase hairpins [Fig. 2; 38-

139

40]. There is amino acid sequence variation within the hairpins of family B DNA polymerases,

140

which could account for different function [37; Table 1], and certainly differences are expected

Ac ce p

te

d

M

an

125

6 Page 6 of 48

for yeast DNA polymerase ε which has only a truncated hairpin [41; Fig. 2]. But the lack of a

142

substantial proofreading defect for the G447S-DNA polymerase δ and the absence of a hairpin in

143

DNA polymerase ε raise the possibility that not all DNA polymerase hairpin structures are

144

required for proofreading.

Genetic selection strategies, modeled on T4 methods, were used to identify proofreading

cr

145

ip t

141

deficient yeast DNA polymerase δ mutants. Several yeast mutator DNA polymerases were

147

selected with amino acid substitutions in the exonuclease domain that reduced exonuclease

148

activity and increased mutation rates, but none of substitutions were in the hairpin region [42].

149

Just as mutator DNA polymerases can be selected in phage T4 and yeast, recent large-scale DNA

150

sequencing studies reveal that mutator DNA polymerase δ and ε mutants are also selected in

151

cancer cells [43-47]; this is because “mutator” DNA polymerases are proposed to provide a pool

152

of mutations that are needed to drive development of cancer cells [48]. While several of the

153

mutant DNA polymerase  and  mutants identified in human colon and endometrial tumors

154

have amino acid substitutions within the exonuclease domain that reduce 3’ → 5’ exonuclease

155

activity [47], no substitutions have been identified in the hairpin region.

an

M

d

te

Ac ce p

156

us

146

Having failed to find hairpin mutants that affect function of the yeast DNA polymerase δ,

157

we constructed hairpin mutants with major structural changes. A hairpin deletion (hpmutant

158

was made in which 7 amino acids were deleted (Table 1), as was done for the RB69 DNA

159

polymerase hp mutant that lacks 3’→5’ exonuclease activity with duplex, but not single-

160

stranded DNA [27] and displays a strong mutator phenotype in vivo [26]. A yeast/T4 chimeric

161

DNA polymerase δ was also engineered in which the yeast hairpin was replaced with the T4

162

DNA polymerase hairpin (T4hp) (Table 1). An additional hairpin mutant with a single amino

163

acid substitution, the G447D-DNA polymerase δ, was constructed because recent studies suggest 7 Page 7 of 48

164

that the G447D amino acid substitution curbs the strong mutator phenotype produced by an

165

exonuclease-deficient DNA polymerase δ [49]. The major conclusion from these studies is that, unlike the phageT4 and RB69 DNA

ip t

166

polymerases, yeast DNA polymerase δ does not need the hairpin for proofreading, but appears to

168

require the hairpin for optimum DNA replication efficiency. A second conclusion is that MMR

169

plays an important role in mediating replication problems caused by DNA polymerase δ mutants

170

that is independent of mismatch correction. We discuss the possibility that the two conclusions

171

are linked because T4 and RB69 phage do not have MMR and, thus, may have compensated for

172

this deficiency by using the hairpin to increase proofreading efficiency. Yeast, on the other

173

hand, may depend more on MMR instead of strong proofreading activity and, thus, the hairpin is

174

used instead to maintain optimum replication efficiency.

M

an

us

cr

167

2. Materials and methods

177

2.1. Yeast strains

178 179 180 181 182

Ac ce p

176

te

d

175

The yeast strains used in this study are listed in Supplementary Table S1. All strains are isogenic derivatives of MS71, which is a derivative of CG378/CG379.

2.2. Yeast culture conditions Standard yeast media were used (YPD, YPD plus selective antibiotics, SD medium

183

supplemented with essential nutrients or lacking specific nutrients (e.g. uracil, tryptophan,

184

histidine, lysine, or SD –Arg, plus canavanine); all media have been described [50]. PAA 8 Page 8 of 48

gradients are described by Li et al. [35]. Gradient plates are made in two steps. First, a pH 5-

186

adjusted solution of SD minimal medium made with high grade Noble agar and PAA (Sigma-

187

Aldrich) is poured into a square Petri dish and one end is propped up with a pencil. Second, after

188

the agar hardens the dish is set flat and a second layer of agar without PAA is poured gently over

189

the first layer to preserve the slanted first layer. If the PAA solution contains 4 mg/ml PAA (28

190

mM), then the final gradient is from ~0 to 4 mg/ml (~0 to 28 mM) PAA. HU gradient plates

191

were prepared as done for PAA gradients, but YPD and Difco agar were used; 0.2M HU was in

192

the slanted first layer and the second layer was YPD without HU.

an

us

cr

ip t

185

193

M

195

2.3 Site-directed mutagenesis and mutant construction

Single amino acid substitutions were made using an efficient single mutagenic primer

d

194

method that has been described [50]; details are available by request. The starting point was the

197

yeast integrating plasmid, Yip5, that contained the 2.2 kb fragment from the 5’-end of the POL3

198

gene (SalI-BamHI fragment) [37]. The DNA polymerase  hp mutant was constructed using

199

the same Yip5 plasmid. The mutagenic primer for the hp mutant was the following: 5ʹ-P-

200

GCAAGAAATTAAAGAGTCTGTGTTCGGTGAAACCAAAAATGTCAATATTGACGG-3ʹ.

201

The Yip5-pol3-hpΔ plasmid was sequenced to confirm the deletion. The plasmid was cleaved at

202

the unique HpaI restriction endonuclease site and then used to transform competent yeast cells

203

that were prepared as described [51]. PCR and DNA sequencing were done to confirm correct

204

integration of the plasmid into the chromosomal copy of the POL3 gene and restoration of a

205

single gene copy after selection for plasmid pop-out on 5-FOA.

Ac ce p

te

196

9 Page 9 of 48

206

The chimeric DNA polymerase  T4hp mutant was constructed starting with the Yip5 plasmid that contained the longer 2.5 kb fragment from the 5’-end of the POL3 gene (SalI-KpnI

208

fragment). A 170 bp fragment containing DNA encoding the yeast hairpin was removed leaving

209

PsiI and EagI ends. PCR was used to amplify DNA encoding the T4 hairpin; PsiI and EagI sites

210

were engineered into the 5’- and 3’-ends. Plasmid DNA with complimentary PsiI and EagI ends

211

was then ligated to the T4 hairpin fragment. The plasmid was sequenced to confirm construction

212

of the chimeric yeast/T4 DNA polymerase mutant. The plasmid was cleaved at the unique

213

BamHI restriction endonuclease site and transformed into competent yeast cells. PCR and DNA

214

sequencing were done to confirm correct integration of the plasmid into the POL3 gene and

215

restoration of a single gene copy after selection for plasmid pop-out on 5-FOA.

M

an

us

cr

ip t

207

216

218

d

2.3 Determining spontaneous mutation rates

te

217

Mutation rates were determined as described previously [35]. The trp1-289 allele reverts by base substitution and the his7-2 and lys2::InsE(A8) alleles revert by +1 and -1 frameshift

220

mutations, respectively. Forward mutations to produce resistance to canavanine (CanR) can be

221

base substitutions, frameshifts or any mutation that inactivates function (insertion, deletion,

222

complex).

223

Ac ce p

219

224

3. Results

225

3.1. Replication fidelity by yeast DNA polymerase  hairpin mutants with single amino acid

226

substitutions: G447S, G447Q, G447D 10 Page 10 of 48

Single amino acid changes in the tip region of the phage T4 DNA polymerase β hairpin,

227

e.g. G255S, G255A, Y254G, Y254S, Y254Q, Y254K (14,25; Table 1, Fig. 1B), produce low

229

fidelity DNA polymerases that increase base substitution mutations to a level similar to

230

exonuclease-deficient DNA polymerase mutants [37], but this was not observed for analogous

231

amino acid substitutions in the yeast DNA polymerase δ hairpin. Little change in fidelity was

232

observed for the G447S- [37] or G447Q-DNA polymerase δ mutants, but a 7-fold increase in

233

-1fs mutations was detected with the Lys+ reporter for the G447D- DNA polymerase δ in the

234

absence of MMR and a 3-fold increase in mutations that confer CanR (Table 2).

an

us

cr

ip t

228

235

3.2 The pol3-G447D strain is sensitive to hydroxyurea (HU) but not to phosphonoacetic acid

237

(PAA)

d

M

236

We tested the possibility that the β hairpin in yeast DNA polymerase δ was needed for

239

efficient DNA replication by looking for a slow growth phenotype or HU- or PAA-sensitivity.

240

PAA inhibits DNA polymerases that favor formation of pre-T complexes and HU-sensitivity

241

identifies “weak” DNA polymerases that have difficulty replicating DNA when deoxynucleoside

242

triphosphate (dNTP) pools are reduced due to HU inhibition of ribonucleotide reductase. A slow

243

growth phenotype was not observed with rich or minimal media, but cell killing was observed in

244

the presence of HU in the absence, but not in the presence of MMR (Fig. 3A). HU induces a

245

checkpoint response for wild type yeast that severely slows DNA replication [52], but the

246

observed killing by HU for pol3-G447D cells that lack MMR suggests that MMR is needed to

247

activate the checkpoint response or to assist in recovery. This observation raises the possibility

248

that MMR interacts with the mutant DNA polymerase δ at stalled replication forks. This

Ac ce p

te

238

11 Page 11 of 48

proposal is elaborated further in the Discussion. Although the phage T4 G255S-DNA

250

polymerase is PAA sensitive [18], the similar yeast pol3-G447S, Q and D strains were not,

251

which indicates functional differences between the yeast and T4 hairpin mutants.

ip t

249

252

3.3. The yeast DNA polymerase  hp mutant replicates DNA with high fidelity but displays a

254

slow growth phenotype

us

cr

253

Seven amino acids from the tip region of the hairpin were deleted as was done to create

256

the proofreading-deficient RB69 DNA polymerase hp mutant [27; Table 1]. A strong mutator

257

phenotype is observed for the phage RB69 hp strain as expected if proofreading is

258

compromised [26]; a similar reduction in replication fidelity was expected for the yeast DNA

259

polymerase δ hp (pol3-hp) strain if the yeast hairpin is needed for exonucleolytic proofreading

260

activity. The pol3-hp strain replicated DNA almost as accurately as the wild type POL3 strain;

261

only a 2-fold increase in the CanR mutation rate was detected in the absence of MMR (Table 2),

262

which was due primarily to an increase in -1 fs mutations (Table 3). The small reduction in

263

replication fidelity by the DNA polymerase δ hp mutant was less than the weak mutator

264

phenotype observed for the pol3-G447D MMR- strain. In comparison, a 20-fold increase in the

265

CanR mutation rate is observed for the exonuclease-deficient pol3-5DV strain, and this is in the

266

presence of MMR [53]. The pol3-5DV strain is inviable in the absence of MMR because the

267

exceptionally high mutation rate produces error castastrophe; see discussion by Herr et al. [54].

268

Thus, the yeast DNA polymerase δ hairpin does not appear to function in proofreading because

269

replication fidelity is largely retained for the hpmutant.

Ac ce p

te

d

M

an

255

12 Page 12 of 48

270

A pronounced slow growth phenotype was observed for cells expressing the DNA polymerase δ hp mutant in the absence of MMR, which is shown for the pol3-hp msh6 msh3

272

strain (Fig. 4). Yeast cells that lack both Msh6 and Msh3 cannot form MutS α or β complexes

273

and, thus, are MMR-deficient [55]. Colonies were significantly smaller for the MMR-deficient

274

pol3-hpstrain than for the MMR-proficient pol3-hp strain. The slow growth phenotype for

275

the pol3-hp MMR- strain was also observed in liquid culture where the maximal cell titer

276

reached only about 2 x 107 cells/ml for the mutant compared to 108 cells/ml for wild type yeast.

277

While slow growth could be due to low viability caused by a high mutational burden, only a

278

weak mutator phenotype was observed for the pol3-hpΔ MMR- strain (Table 2); instead, an S-

279

phase check point appears to be activated as most cells resembled dumbbells as expected if the

280

cell cycle has a prolonged S-phase.

cr

us

an

M

The MRC1 gene was deleted in MMR proficient and deficient pol3-hpstrains to

d

281

ip t

271

determine if an S-phase checkpoint was responsible for the slow growth phenotype (Fig. 4).

283

Deletion of the MRC1 gene had little effect on pol3-hpcells in the presence of MMR, but

284

deletion partially rescued the slow growth phenotype in the absence of MMR; compare colony

285

sizes for pol3-hpMMR- andpol3-hpMMR- mrc1strains. Thus, there are problems with

286

DNA replication by the DNA polymerase δ hp mutant, but the problems are masked by MMR.

287

DNA replication activity of the hpΔ-DNA polymerase δ was further probed with HU and

Ac ce p

te

282

288

PAA. The pol3- hpstrain, with or without MMR, was not HU-sensitive (Fig. 3A), but PAA-

289

resistance was observed for the MMR-proficient pol3- hpstrain (Fig. 3B). A slow growth

290

phenotype is observed for wild type yeast cells at high concentrations of PAA after 2 days

291

incubation, but colonies were observed across the entire gradient (~0 to 4 mg/ml (28 mM) PAA) 13 Page 13 of 48

by 7 days. In contrast, cells expressing the PAA-sensitive L612M-DNA polymerase δ produced

293

only small colonies after 2 days incubation at the lowest PAA concentrations, and more colonies

294

were detected after 7 days, but only up to the midway position in the gradient. High PAA

295

concentrations killed cells that expressed the L612M-DNA polymerase δ, as observed previously

296

[35]. The pol3-hp strain, however, was more PAA-resistant than wild type cells (POL3) as

297

colonies were observed at the highest PAA concentrations after just 2 days incubation (Fig. 3B.).

298

Even the MMR-deficient pol3-hp strain appeared to be more resistant to PAA despite the slow

299

growth phenotype. The PAA-resistance of the hp-DNA polymerase δ indicates reduced ability

300

to form pre-T complexes compared to the wild type DNA polymerase δ and, thus, altered DNA

301

replication.

M

an

us

cr

ip t

292

302

3.4. A strong mutator phenotype for -1fs mutations was observed for the chimeric DNA

304

polymerase  T4hp mutant in the absence of MMR

Ac ce p

te

d

303

305

A chimeric yeast-T4 DNA polymerase δ mutant was constructed in which the T4 β

306

hairpin replaced the DNA polymerase δ hairpin (pol3-T4hp). The substitution extended from the

307

ExoII motif through the hairpin (Table 1). The T4 and yeast DNA polymerase δ hairpins are

308

similar (Fig. 2) and, thus, may be expected to substitute for each other. The yeast pol3-T4hp

309

strain was viable, but the converse construction, in which the yeast DNA polymerase δ hairpin

310

replaced the T4 hairpin in the phage T4 DNA polymerase, was not. Only small increases in

311

spontaneous mutation rates were observed for the yeast MMR-proficient pol3-T4hp strain, but

312

larger 5- to 20-fold increases in mutation rates were observed for the MMR-deficient pol3-T4hp

313

strain compared to the POL3 MMR- strain (Table 2). The CanR mutation spectrum for the 14 Page 14 of 48

MMR-deficient pol3-T4hp strain (pol3-T4hp MMR-) was predominated by -1 fs mutations

315

(Table 3) at sites within tracts of 4 to 6 Ts (As on the complementary strand) as observed for the

316

POL3 MMR- strain. More than a third of the total CanR mutants sequenced (15/42) for the

317

MMR-deficient pol3-T4hp strain were -1fs mutations at a single site within a tract of 6 Ts

318

(Supplementary Figure 1S). This “hotspot” mutation site was observed for only 2/20 CanR

319

mutants sequenced for the wild type DNA polymerase  MMR-deficient strain (POL3 MMR-)

320

(Supplementary Figure 1S). The mutation rate for -1 frameshift mutations in the absence of

321

MMR for the DNA polymerase  T4hp mutant was 5478 x 10-8 ([fraction of -1 fs mutations] x

322

[total CanR mutation rate]; 0.90 x 6087) compared to just 278 x 10-8 (0.85 x 363) for the wild

323

type DNA polymerase, which is almost a 20-fold increase in the rate for -1fs mutations for the

324

T4hp-DNA polymerase δ mutant.

M

an

us

cr

ip t

314

We were concerned, however, that the large replacement of T4 DNA polymerase

326

sequence in the yeast DNA polymerase δ could affect exonuclease activity because non-

327

conserved residues in the ExoII motif in the chimeric pol3-T4hp strain were replaced along with

328

T4 DNA polymerase hairpin residues (Table 1). To test this possibility, the “ttn” chimeric pol3-

329

T4hp strain was constructed in which the yeast ExoII motif was restored to produce the MMR-

330

deficient pol3-T4hp,ttn strain. The ttn substitution reduced the CanR mutation rate about 4-fold

331

(Table 2), which indicates that the ttn substitution contributes to the mutator phenotype observed

332

for the MMR- pol3-T4hp strain. Furthermore, while fs mutations predominated in the CanR

333

spectrum for MMR- pol3-T4hp strain (Table 3, Supplementary Figure 1S), base substitution

334

mutations predominated for the MMR-deficient pol3-T4hp,ttn strain as expected if the mutant

335

DNA polymerase was defective in proofreading misincorporated nucleotides. Note that the

336

phage T4 G255S-DNA polymerase is a strong mutator for base substitutions, but not fs

Ac ce p

te

d

325

15 Page 15 of 48

mutations [37]. However, introducing the T4 ExoII motif into yeast DNA polymerase δ did not

338

create a mutator phenotype; only slightly elevated mutation rates were detected for the MMR-

339

pol3-ieg strain above rates observed for the POL3 MMR- strain (Table 2). Thus, the strong

340

mutator phenotype for fs mutations observed for the pol3-T4hp strain requires the T4 hp and the

341

T4 ExoII motif. If frameshift mutations arise because of DNA polymerase dissociation as

342

proposed by Kunkel et al. [56], then the strong mutator phenotype for -1fs mutations observed

343

for the MMR-deficient pol3-T4hp strain may be caused by increased polymerase dissociation

344

during attempted proofreading, while the increase in base substitution mutations observed for the

345

MMR-deficient pol3-T4hp, ttn strain is due to decreased proofreading.

an

us

cr

ip t

337

d

348

3.5. The chimeric DNA polymerase  T4hp strain is HU- and PAA-sensitive Like the MMR-deficient pol3-G447D MMR- strain, the MMR-deficient chimeric pol3-

te

347

M

346

T4hp MMR- strain was sensitive to HU (Fig. 3A); thus, the T4hp hairpin-ExoII

350

insertion/replacement may produce a weak DNA polymerase δ. A slow growth phenotype was

351

observed for the chimeric pol3-T4hp MMR- strain, which could indicate slow DNA replication

352

[Fig. 4], but the slow growth may also be due to reduced viability that is caused by error

353

catastrophe. The HU-induced checkpoint may exacerbate the high mutation rate [57] and

354

increase cell death. The pol3-T4hp MMR- strain was also sensitive to PAA, but the pol3-T4hp

355

MMR+ strain appears to be more resistant than the wild type POL3 strain, but not as resistant as

356

the pol3-hpMMR+ strain (Fig. 3). The MMR-deficient pol3-T4hp ttn mutant was not sensitive

357

to HU or to PAA, data not shown.

Ac ce p

349

358 16 Page 16 of 48

359

361

3.6 T4hp-DNA polymerase δ replication in the absence of Rad27 orExo1 The pol3-T4hp rad27 strain was not viable. Only 3 of 4 spores germinated; cell growth

ip t

360

362

arrested with 50 to 100 cells. The pol3-hp rad27 mutant, however, was viable.

363

DNA polymerase  3’→5’ exonuclease activity or Rad27 flap endonuclease is required to form

364

ligatable ends at Okazaki fragment junctions and synthetic lethality is observed in the absence of

365

both activities [53,58-60], the T4hp insertion/replacement appears to prevent proofreading. One

366

possibility is that the T4hp-DNA polymerase δ is prone to dissociation in the context of Okazaki

367

fragment junctions.

cr

us

an

M

368

Since either

The 5’→3’ exonuclease activity of Exo1 may fulfill a similar function as Rad27 in forming ligatable junctions, but at repair gaps. Increased fs mutations are observed for DNA

370

polymerase δ and ε mutants with reduced exonuclease activity in the absence of Exo1 [61]; this

371

is the exo1-dependent mutator (edm) phenotype [62]. An edm phenotype was observed for the

372

pol3-T4hp exo1 strain (Table 2); synergistic increases in +1fs and -1fs mutations were observed

373

for the double mutant. The CanR mutation spectrum for the pol3-T4hp exo1 strain was primarily

374

frameshift mutations (17/20): 10 were +1fs mutations and 7 were –1fs mutations in

375

mononucleotide repeat tracts.

377 378 379

te

Ac ce p

376

d

369

3.7 The T4hp,L612M-DNA polymerase  has little proofreading activity If the T4hp-DNA polymerase δ mutant has reduced 3’→5’ exonuclease activity, then the addition of a second amino acid substitution, L612M, that reduces exonuclease activity by 17 Page 17 of 48

380

reducing the opportunity to proofread [12,20], has the potential to create a proofreading-deficient

381

DNA polymerase even though the exonuclease active site is unchanged. This was observed. The pol3-T4hp,L612M strain has a strong mutator phenotype in the presence of MMR

ip t

382

(Table 4), similar to the strong mutator phenotypes observed for the exonuclease-deficient pol3-

384

01 and pol3-5DV strains [53,58]. Mutation rates could not be determined for the pol3-T4hp,

385

L612M strain in the absence of MMR because the double mutant was not viable as observed for

386

the pol3-01 and pol3-5DV strains. No pol3-T4hp,L612M msh6 mutants were recovered from 31

387

dissected asci. MMR was partially inhibited for the pol3-T4hp,L612M strain by the addition of

388

cadmium (2.5 µM), which inhibits MMR activity [63]. The double mutant was not viable at

389

higher cadmium concentrations likely because of an excessively high mutation burden. The

390

mutation rate for CanR increased 10-fold from 840 x 10-8 to >9000 x 10-8 in the presence of 2.5

391

µM cadmium (Table 4), which is likely at the limit for haploid cell survival [54]. Base

392

substitution and -1 and +1 frameshift mutations were produced for cadmium-treated pol3-T4hp,

393

L612M cells. The pol3-T4hp, L612M strain was also not viable in the absence of Exo1, as

394

observed for the exonuclease-deficient pol3-01 strain [59]; no pol3-T4hp, L612M exo1 mutants

395

were recovered from dissection of 27 asci. The strong mutator phenotype and synthetic lethality

396

of the pol3-T4hp, L612M strain in the absence of Msh6 or Exo1 indicates a severe loss of

397

proofreading activity.

399

us

an

M

d

te

Ac ce p

398

cr

383

4. Discussion

400

18 Page 18 of 48

Experiments presented here with the yeast hpΔ-DNA polymerase δ mutant demonstrate

402

that the yeast β hairpin is not required for proofreading as observed for the phage T4 and RB69

403

DNA polymerase β hairpins. Instead of the strong mutator phenotype that is observed for the

404

phage G255S- and hpΔ-DNA polymerases [14,25,26,37], only a 2-fold increase in frameshift

405

mutations was observed for the pol3-hpΔ strain in the absence of MMR (Tables 2,3), but the

406

slow growth phenotype [Fig. 4] indicates that loss of the hairpin compromises DNA replication.

407

Deletion of MRC1 partially rescued the slow growth phenotype [Fig. 4], which suggests that the

408

intraS-phase checkpoint is activated. While the RB69 DNA polymerase β hairpin acts as a

409

wedge between the template and primer strands in exonuclease complexes (Fig. 1B) the hairpin

410

has also been observed to contact the single-stranded template strand in complexes in which

411

replication is stalled at DNA damage, perhaps in a pre-exonuclease complex [64]. The yeast

412

DNA polymerase δ hairpin has also been shown to contact DNA in polymerase complexes [40];

413

K444 at the tip of the hairpin (Table 1) protrudes into the major groove and the complex is

414

stabilized by interactions with other hairpin residues, F441 and Y446 [40,65]. The position of

415

the yeast β hairpin is consistent with a role in polymerase-to-exonuclease switching [40,65];

416

however, these hairpin residues are missing in the hpΔ-DNA polymerase, but proofreading is

417

largely intact. Thus, in view of the slow growth phenotype observed for the pol3-hpΔ strain in

418

the absence of MMR, the yeast β hairpin appears to play a role in optimizing DNA replication

419

efficiency, perhaps by stabilizing polymerase complexes, rather than in proofreading. If this is

420

the case, then the presence of only a truncated hairpin in DNA polymerase ε that is too short to

421

contact DNA (Fig. 2) may be explained by the presence of the processivity (P) domain, which is

422

proposed to stabilize polymerase complexes [41]. Thus, the β hairpin structures in the T4 and

423

RB69 DNA polymerases may function to stabilize pre-exonuclease and exonuclease complexes,

Ac ce p

te

d

M

an

us

cr

ip t

401

19 Page 19 of 48

both sides of polymerase-to-exonuclease switching, but the yeast DNA polymerase δ hairpin

425

may only stabilize polymerase complexes. DNA polymerase ε, on the other hand, does not need

426

the β hairpin and relies instead on the P-domain.

427

ip t

424

Although previous studies indicated little change in replication fidelity for single amino acid substitutions in the tip of the yeast DNA polymerase δ hairpin [37], we constructed the

429

G447D-DNA polymerase δ because the G447D substitution appears to rescue the synthetic

430

lethality of the pol3-01,L612M strain, presumably by reducing the mutation rate and preventing

431

error catastrophe [49]. DNA replication fidelity was determined in the absence of MMR for the

432

double mutant, pol3-G447D,L612M; no change in replication fidelity was observed for the

433

double mutant compared to the po3-L612M strain (data not shown). However, if the G447D

434

substitution increases polymerase dissociation, for example at mismatched primer-ends, this

435

would increase the opportunity for removal of the incorrect nucleotide by another exonuclease.

436

Another possibility is that the mismatched primer-end would be extended by a translesion DNA

437

polymerase, which would be mutagenic if not repaired by MMR, but DNA replication would be

438

able to continue. While HU-sensitivity is observed for the pol3-G447D strain only in the

439

absence of MMR (Fig. 3A), the pol3-G447D,L612M strain is HU-sensitive in the presence of

440

MMR and is killed by very low HU concentrations in the absence of MMR (data not shown).

441

Thus, the G447D substitution in the tip of the hairpin may interfere with DNA interactions

442

needed to stabilize polymerase complexes. HU-sensitivity indicates that the G447D- and

443

G447D,L612M-DNA polymerase δ mutants cannot form stable post-T complexes unless dNTP

444

pools are high.

445 446

Ac ce p

te

d

M

an

us

cr

428

The above discussion points to a role for the β hairpin in stabilizing polymerase complexes. But the slow growth phenotype for the pol3-hpΔ strain was only observed in the 20 Page 20 of 48

absence of MMR; similarly, HU-sensitivity was only observed for the pol3-G447D strain in the

448

absence of MMR. Thus, MMR normally masks replication problems that are caused by the hpΔ

449

and the G447D substitution, but in different ways. While MMR complexes may stabilize

450

polymerase complexes formed with the hpΔ-DNA polymerase δ so that an S-phase checkpoint is

451

not activated, MMR may be needed for pol3-G447D cells to activate an S-phase checkpoint or to

452

assist in recovery since HU kills these cells. We observed previously that MMR is needed to

453

activate the Rad9-dependent checkpoint in PAA-treated pol3-L612M cells [66]. Although

454

further study is needed to reveal the role of MMR in mediating DNA replication problems

455

caused by the hpΔ- and G447D-DNA polymerase δ mutants, MMR is clearly performing a

456

function that is independent of mismatch correction because the mutants have relatively high

457

replication fidelity (Table 2).

cr

us

an

M

The MMR-deficient pol3-T4hp strain proved to be complex because the strong mutator

d

458

ip t

447

phenotype for fs mutations (Tables 2,3) and the HU- and PAA-sensitivity (Fig. 3) depended on a

460

synergistic interaction between T4 DNA polymerase residues in the ExoII motif and in the T4hp.

461

All of the phenotypes observed for the pol3-T4hp MMR- strain disappeared when the yeast ExoII

462

motif was restored in the pol3-T4hp,ttn MMR- strain. A weak mutator phenotype for base

463

substitution mutations, however, was observed for the pol3-T4hp,ttn MMR- strain, which

464

suggests that the T4hp interferes with proofreading, possibly by decreasing polymerase-to-

465

exonuclease switching. These studies demonstrate the evolutionary fine-tuning that has taken

466

place to optimize β hairpin function for the yeast and T4 DNA polymerases.

467

Ac ce p

te

459

The pol3-T4hp strain also provided informative comparisons with the pol3-L612M strain.

468

Both mutants display mutator phenotypes in the absence of MMR and the edm-phenotype

469

(Tables 2,3); in addition, both mutant DNA polymerases cannot support DNA replication in the 21 Page 21 of 48

absence of Rad27 [35], but there are important differences. In the presence of MMR, the pol3-

471

L612M strain is PAA-sensitive, but the pol3-T4hp strain is slightly PAA-resistant (Fig. 3B), but

472

in the absence of MMR, PAA kills both mutant strains (Fig. 3B; 35). Another difference is that

473

HU kills pol3-T4hp cells in the absence of MMR (Fig. 3B), but MMR- pol3-L612M cells are

474

only slightly HU-sensitive (data not shown). We reasoned that proofreading was compromised

475

for both mutant DNA polymerases, but in different ways; thus, the double mutant, pol3-

476

T4hp,L612M, has two defects, which together severely reduce the initiation of the proofreading

477

pathway. As observed for exonuclease-deficient DNA polymerases, pol3-T4hp,L612M was not

478

viable in the absence of MMR or Exo1 and a strong mutator phenotype was observed when

479

MMR was partially reduced with cadmium (Table 4). Thus, the inability to perform polymerase-

480

to-exonuclease switching creates exonuclease deficiency as much as inactivation of exonuclease

481

active site residues.

cr

us

an

M

d

While family B DNA polymerases share common functions, there is divergence in the

te

482

ip t

470

use of the β hairpin. The T4 DNA polymerase employs the hairpin to optimize proofreading

484

efficiency, and is about 1000-fold more active on double-stranded DNA ([23]; kcat = 15-19 s-1)

485

than the hairpin-less Klenow fragment of Escherichia coli DNA polymerase I ([67]; kcat = 0.001

486

s-1). The importance of the β hairpin for proofreading by the T4 and RB69 DNA polymerases is

487

demonstrated by the high number proofreading-defective mutants that have been identified in the

488

hairpin region (Table 5), and equally striking is the absence of amino acid substitutions in the

489

hairpin regions of yeast and human DNA polymerase δ and ε mutants that create mutator DNA

490

polymerases. In contrast to the absence of hairpin mutants in yeast and human mutator DNA

491

polymerase δ and ε, several amino acid substitutions in the adjacent region produce mutator

492

DNA polymerases, but this region is nearly void of amino acid substitutions for the phage T4

Ac ce p

483

22 Page 22 of 48

DNA polymerase despite extensive mutational analysis (Table 5). The yeast mutator DNA

494

polymerase δ mutants have reduced exonuclease activity as judged by sensitivity to cadmium

495

[42]. Amino acid substitutions V411L and L424V in human DNA polymerase ε reduce 3’-

496

exonuclease activity [47] and are observed frequently in mutator DNA polymerase ε mutants

497

expressed in cancer cells [44-47]. Interestingly, the prognosis is often good for DNA polymerase

498

ε exonuclease domain mutants [46], which may be due to retention of MMR because reduced

499

viability is observed for exonuclease-deficient mutants in the absence of MMR due to a heavy

500

mutation burden [47]. But another possibility is that if MMR mediates DNA replication

501

problems for DNA polymerase ε mutants as we observe for DNA polymerase δ mutants, then

502

MMR will be retained to maintain efficient DNA replication.

cr

us

an

M

503

ip t

493

Even though the β hairpin is not essential for proofreading by DNA polymerase δ, other features of the exonuclease active site are conserved and essential. Besides three essential

505

conserved aspartate residues that bind two Mg2+ ions in the catalytic center [33,58], a

506

phenylalanine residue (F123) is observed to intercalate between the terminal bases at the primer-

507

end in exonuclease complexes formed with the RB69 DNA polymerase [39], or is in close

508

proximity to the primer-end (Fig. 1B). Intercalation is likely the normal mode of binding as

509

determined from studies with the fluorescent base analog, 2-aminopurine (2AP) [68]. Highly

510

fluorescent exonuclease complexes are formed with the T4 DNA polymerase with primers

511

labeled at the 3’-end with 2AP because binding unstacks the terminal 2AP and relieves base

512

stacking interactions that quench fluorescence [32]; unstacking is proposed to be mediated by

513

intercalation of F120 in the T4 DNA polymerase or F123 in the RB69 DNA polymerase [68].

514

Fluorescent exonuclease complexes are not observed, however, for the T4 D219A-DNA

515

polymerase [32] and, as shown in Figure 1B, F123 does not intercalate between the terminal

Ac ce p

te

d

504

23 Page 23 of 48

bases for complexes formed with the RB69 D222A-DNA polymerase, which is analogous to the

517

T4 D219A-DNA polymerase. Thus, the D219A/D222A substitution in the ExoII motif prevents

518

phenylalanine intercalation. Intercalation by the conserved phenylalanine residue appears to be

519

important for proofreading because amino acid substitutions for proline residues near the

520

conserved phenylalanine residue, P123L in the T4 DNA polymerase, P286R and P266H in

521

human DNA polymerase ε, and P327R in human DNA polymerase δ were identified by genetic

522

selection for mutator DNA polymerases [11,44-47]; human P286R- and P286H-DNA

523

polymerase  mutants are severely exonuclease-deficient [47].

cr

us

an

524

ip t

516

Different family B DNA polymerases appear to have tuned the balance between primer elongation and proofreading to meet the specific needs for DNA replication accuracy and

526

efficiency in different organisms. T4 DNA polymerase proofreading is highly efficient, in large

527

part due to the β hairpin, and this may be needed because T4 phage do not have MMR to repair

528

DNA polymerase errors. Yeast DNA polymerase δ does not appear to use its hairpin for

529

proofreading, but this is acceptable because MMR can correct DNA polymerase replication

530

errors missed by proofreading; however, the yeast polymerase δ hairpin does appear to be needed

531

for optimal DNA replication activity.

533 534

d

te

Ac ce p

532

M

525

Conflict of interest statement

The authors declare that there are no conflicts of interest.

535 536

Acknowledgements 24 Page 24 of 48

537

This work was supported with funds from the Canadian Institutes of Health Research and the Alberta Heritage Foundation for Medical Research. The authors gratefully acknowledge

539

technical assistance from M. Funnel, M. O’Carroll, and A. Radziwon.

ip t

538

540

cr

542

Supplementary information

Supplementary Table 1S and Figure 1S can be found in the online version, at http://

us

541

an

543

References

545

[1] D. Brutlag, A. Kornberg, Enzymatic synthesis of deoxyribonucleic acid. XXXVI. A

546

proofreading function for the 3’→5’ exonuclease activity in deoxyribonucleic acid polymerases,

547

J. Biol. Chem. 247 (1972) 241-248.

548

[2] N. Muzyczka, R.L. Poland, M. J. Bessman, Studies on the biochemical basis of spontaneous

549

mutation. I. A comparison of the deoxyribonucleic acid polymerases of mutator, antimutator, and

550

wild type strains of bacteriophage T4, J. Biol. Chem. 247 (1972) 7116-7122.

551

[3] J.F. Speyer, Mutagenic DNA polymerase, Biochim. Biophys. Acta 21 (1965) 6-8.

552

[4] J.W. Drake, E.F. Allen, Antimutagenic DNA polymerases of bacteriophage T4, Cold Spring

553

Harbor Symp. Quant. Biol. 33 (1968) 339-344.

554

[5] J.W. Drake, E.F. Allen, S.A. Forsberg, R-M. Preparata, E.O. Greening, Spontaneous

555

mutation, Nature (1969) 1128-1132.

Ac ce p

te

d

M

544

25 Page 25 of 48

[6] P. Spacciapoli, N.G. Nossal, A single mutation in bacteriophage T4 DNA

557

polymerase (A727V, tsL141) decreases its processivity as a polymerase and increases its

558

processivity as a 3’5’ exonuclease, J. Biol. Chem. 269 (1994) 438-446.

559

[7] L.J. Reha-Krantz, DNA polymerase proofreading: Multiple roles maintain genome integrity,

560

Biochim. Biophys. Acta, 1804 (2010) 1049-1063.

561

[8] L.J. Reha-Krantz, L.A. Marquez, E. Elisseeva, R.P. Baker, L.B. Bloom, H.B. Dunford, M.F.

562

Goodman, The proofreading pathway of bacteriophage T4 DNA polymerase, J. Biol. Chem. 273

563

(1998) 22969-22976.

564

[9] L.J. Reha-Krantz, M.J. Bessman, Studies on the biochemical basis of mutation. VI. Selection

565

and characterization of a new bacteriophage T4 mutator DNA polymerase, J. Mol. Biol. 145

566

(1981) 677-695.

567

[10] L.J. Reha-Krantz, E.M. Liesner, S. Parmaksizoglu, S. Stocki, Isolation of bacteriophage T4

568

DNA polymerase mutator mutants, J. Mol. Biol. 189 (1986) 261-272.

569

[11] L.J. Reha-Krantz, Amino acid changes coded by bacteriophage T4 DNA polymerase

570

mutator mutants. Relating structure to function, J. Mol. Biol. 202 (1988) 711-724.

571

[12] L.J. Reha-Krantz, R.L. Nonay, Motif A of bacteriophage T4 DNA polymerase: Role in

572

primer extension and DNA replication fidelity. Isolation of new antimutator and mutator DNA

573

polymerases, J. Biol. Chem. 269 (1994) 5635-5643.

574

[13] L.J. Reha-Krantz, Use of genetic analyses to probe structure, function, and dynamics of

575

bacteriophage T4 DNA polymerase, Methods Enzymol. 262 (1995) 323-331.

Ac ce p

te

d

M

an

us

cr

ip t

556

26 Page 26 of 48

[14] S.A. Stocki, R.L. Nonay, L.J. Reha-Krantz, Dynamics of bacteriophage T4 DNA

577

polymerase function: Identification of amino acid residues that affect switching between

578

polymerase and 3’→5’ exonuclease activities, J. Mol. Biol. 254 (1995) 15-28.

579

[15] L.J. Reha-Krantz, C. Wong, Selection of bacteriophage T4 antimutator DNA polymerases: a

580

link between proofreading and sensitivity to phosphonoacetic acid, Mutat. Res. 350 (1996) 9-16.

581

[16] V. Li, M. Hogg, L.J. Reha-Krantz, Identification of a new motif in family B DNA

582

polymerase by mutational analyses of the bacteriophage T4 DNA polymerase, J. Mol. Biol. 400

583

(2010) 295-308.

584

[17] B.W. Langhorst, W.E. Jack, L. Reha-Krantz, N.M. Nichols, Polbase: a repository of

585

biochemical, genetic and structural information about DNA polymerases, Nuc. Acids. Res. 40

586

(2012) D381-D387, doi: 10.1093/nar/gkr847.

587

[18] L.J. Reha-Krantz, R.L. Nonay, S. Stocki, Bacteriophage T4 DNA polymerase mutations that

588

confer sensitivity to the PPi analog phosphonoacetic acid, J. Virol. 67 (1993) 60-66.

589

[19] M.F. Goodman, S. Creighton, L.B. Bloom, J. Petruska, Biochemical basis of DNA

590

replication fidelity, Crit. Rev. Biochem. Mol. Biol. 28 (1993) 83-126.

591

[20] L.J. Reha-Krantz, S. Woodgate, M.F. Goodman, Engineering processive DNA polymerases

592

with maximum benefit at minimum cost, Frontiers in Microbiology (2014) doi:

593

10.3389/fmicb.2014.00380.

594

[21] L.J. Reha-Krantz, Regulation of DNA polymerase proofreading activity: studies of

595

bacteriophage “antimutator” DNA polymerases, Genetics 148 (1998) 1551-1557.

Ac ce p

te

d

M

an

us

cr

ip t

576

27 Page 27 of 48

[22] L. Marquez, L.J. Reha-Krantz, Using 2-aminopurine fluorescence and mutational analysis to

597

demonstrate an active role of bacteriophage T4 DNA polymerase in strand separation required

598

for 3’→5’ exonuclease activity, J. Biol. Chem. 271 (1996) 28903-28911.

599

[23] R.P. Baker, L.J. Reha-Krantz, Identification of a transient excision intermediate at the

600

crossroads between DNA polymerase extension and proofreading pathways, Proc. Natl. Acad.

601

Sci. U.S.A. 95 (1998) 3507-3512.

602

[24] C. Hariharan, L.J. Reha-Krantz, Using 2-aminopurine fluorescence to detect bacteriophage

603

T4 DNA polymerase-DNA complexes that are important for primer extension and proofreading

604

reactions, Biochemistry 44 (2005) 15674-15684.

605

[25] U. Subuddhi, M. Hogg, L.J. Reha-Krantz, Use of 2-aminopurine fluorescence to study the

606

role of the β hairpin in the proofreading pathway catalyzed by the phage T4 and RB69 DNA

607

polymerases, Biochemistry 47 (2008) 6130-6137.

608

[26] A. Trzemecka, D. Plochocka, A. Bebenek, Different behaviors in vivo of mutations in the β

609

hairpin loop of the DNA polymerases of the closely related phages T4 and RB69, J. Mol. Biol.

610

389 (2009) 797-807.

611

[27] M. Hogg, P. Aller, W. Konigsberg, S.S. Wallace, S. Doublié, Structural and biochemical

612

investigation of the role in proofreading of a β hairpin loop found in the exonuclease domain of a

613

replicative DNA polymerase of the B family, J. Biol. Chem. 282 (2007) 1432-1444.

614

[28] Aller, P., S. Duclos, S.S. Wallace, S. Doublié, A crystallographic study of the role of

615

sequence context in thymine glycol bypass by a replicative DNA polymerase serendipitously

616

sheds light on the exonuclease complex, J. Mol. Biol. 412 (2011) 22-34.

Ac ce p

te

d

M

an

us

cr

ip t

596

28 Page 28 of 48

[29] H. Hogg, W. Cooper, L. Reha-Krantz, S.S. Wallace, Kinetics of error generation in

618

homologous B-family DNA polymerases, Nuc. Acids Res. 34 (2006) 2528-2535.

619

[30] E. Fidalgo da /Silva, S.S. Mandal, L.J. Reha-Krantz, Using 2-aminopurine fluorescence to

620

measure incorporation of incorrect nucleotides by wild type and mutant bacteriophage T4 DNA

621

polymerases, J. Biol. Chem. 277 (2002) 40640-40649.

622

[31] K.E. Zahn, E.P. Tchesnokov, M. Götte, S. Doublié (2011) Phosphonoformic acid inhibits

623

viral replication by trapping the closed form of the DNA polymerase, J. Biol. Chem. 286 (2011)

624

25246-25255.

625

[32] L.B. Bloom, M.R. Otto, R. Eritja, L.J. Reha-Krantz, M.F. goodman, J.M. Beechem, Pre-

626

steady-state kinetic analysis of sequence-dependent nucleotide excision by the 3’-exonuclease

627

activity of bacteriophage T4 DNA polymerase, Biochemisty 33 (1994) 7576-7586.

628

[33] E. Elisseeva, S.S. Mandal, L.J. Reha-Krantz, Mutational and pH studies of the 3’→5’

629

exonuclease activity of bacteriophage T4 DNA polymerase, J. Biol. Chem. 274 (1999) 25151-

630

25158.

631

[34] E. Fidalgo da Silva, L.J. Reha-Krantz, DNA polymerase proofreading: active site switching

632

catalyzed by bacteriophage T4 DNA polymerase, Nuc. Acids Res. 35 (2007) 5452-5463.

633

[35] L. Li, K.M. Murphy, U. Kanevets, L.J. Reha-Krantz, Sensitivity to phosphonoacetic acid: a

634

new phenotype to probe DNA polymerase  in Saccharomyces cerevisiae, Genetics 170 (2005)

635

569-580.

Ac ce p

te

d

M

an

us

cr

ip t

617

29 Page 29 of 48

[36] S.A. Nick McElhinny, C.M. Stith, P.M.J. Burgers, T.A. Kunkel, Inefficient proofreading

637

and biased error rates during inaccurate DNA synthesis by a mutant derivative of Saccharomyces

638

cerevisiae DNA polymerase , J. Biol. Chem. 282 (2007) 2324-2332.

639

[37] M.I. Hadjimarcou, R.J. Kokoska, T.D. Petes, L.J. Reha-Krantz, Identification of a mutant

640

DNA polymerase  in Saccharomyces cerevisiae with an antimutator phenotype for frameshift

641

mutations, Genetics 158 (2001) 177-186.

642

[38] J. Wang, P. Yu, T.C. Lin, W.H. Konigsberg, T.A. Steitz, Crystal structures of an NH2-

643

terminal fragment of T4 DNA polymerase and its complexes with single-stranded DNA and with

644

divalent metal ions. Biochemistry 35 (1996) 8110-8119.

645

[39] J. Wang, A.K.M.A. Satter, C.C. Wong, J.D. Karam, W.H. Konigsberg, T.A. Steitz, Crystal

646

structure of a pol α family replication DNA polymerase from bacteriophage RB69. Cell 89

647

(1997) 1087-1099.

648

[40] M.K. Swan, R.E. Johnson, L Prakash, S. Prakash, A.K. Aggarwal, Structural basis of high-

649

fidelity DNA synthesis by yeast DNA polymerase delta, Nat. Struct. Mol. Biol. 16 (2009) 979-

650

986.

651

[41] M. Hogg, P. Osterman, G.O. Bylund, R.A. Ganai, E-B. Lundström, A.E. Sauer-Eriksson, E.

652

Johansson, Structural basis for processive DNA synthesis by yeast DNA polymerase , Nat.

653

Struct. Mol. Biol. 21 (2014) 49-55.

654

[42] K. Murphy, H. Darmawan, A. Schultz, E. Fidalgo da Silva, L.J. Reha-Krantz, A method to

655

select for mutator DNA polymerase ös in Saccharomyces cerevisiae, Genome 49 (2006) 403-

656

410.

Ac ce p

te

d

M

an

us

cr

ip t

636

30 Page 30 of 48

[43] R. Yoshida, K. Miyashita, M. Inoue, A. Shimamoto, Z. Yan, A. Egashira, E. Oki, Y. Kakeji,

658

S. Oda, Y. Maehara, Concurrent genetic alterations in DNA polymerase proofreading and

659

mismatch repair in human colorectal cancer, Eur. J. Human Genet. 19 (2010) 320-325.

660

[44] C. Palles, J-B. Cazier, K.M. Howarth, E. Domingo, A.M. Jones, P. Broderick, Z. Kemp,

661

S.L. Spain, E. Guarino et al., Germline mutations affecting the proofreading domains of POLE

662

and POLD1 predispose to colorectal adenomas and carcinomas, Nat. Genet. 45 (2013) 136-144.

663

[45] D.N. Church, S.E.W. Briggs, C. Palles, E. Domingo, S.J. Kearsey, J.M. Grimes, M.

664

Gorman, L. Martin, K.M. Howarth, S.V. Hodgson et al., DNA polymerase  and δ exonuclease

665

domain mutations in endometrial cancer, Hum. Mol. Genet. 22 (2013) 1127-1133.

666

[46] B. Meng, L.N. Hoang, J.B. McIntyre, M.A. Duggan, G.S. Nelson, C.H. Lee, M. Kobel,

667

POLE exonuclease domain mutation predicts long progression-free survival in grade 3

668

endometroid carcinoma of the endometrium, Gynecol. Oncol. 134 (2014) 15-19.

669

[47] E. Shinbrot, E.E. Henninger, N. Weinhold, K.R. Covington, A.Y. Göksenin, N. Schultz, H.

670

Chao, H.V. Doddapaneni, D.M. Muzny, R.A. Gibbs et al., Exonuclease mutations in DNA

671

polymerase epsilon reveal replication strand specific mutation patterns and human origins of

672

replication, Genome Res. (2014) doi: 10.1101/gr.174789.114.

673

[48] L. Loeb, Human cancers express mutator phenotypes: origin, consequences and targeting,

674

Nature Reviews Cancer 11 (2011) 450-457.

675

[49] A.J. Herr, S.R. Kennedy, G.M. Knowels, E.M. Schultz, B.D. Preston, DNA replication

676

error-induced extinction of diploid yeast, Genetics 196 (2014) 677-691.

Ac ce p

te

d

M

an

us

cr

ip t

657

31 Page 31 of 48

[50] L. Zhang, A. Radziwon, L.J. Reha-Krantz, Targeted mutagenesis of a specific gene in yeast,

678

Yeast Protocols, Methods in Molecular Biology 1163 (2014) 109-129.

679

[51] M. Knop, K. Siegers, G. Pereira, W. Zachariae, B. Winsor, K. Nasmyth, E. Schiebel,

680

Epitope tagging of yeast genes using a PCR-based strategy: more tags and improved practical

681

routines, Yeast 15 (1999) 963-972.

682

[52] G.M. Alvino, d. Collingwood, J.M. Murphy, J. Delrow, B.J. Brewer, M.K. Raghuraman,

683

Replication in hydroxyurea: it’s a matter of time, Mol. Cell. Biol. 27 (2007) 6396-6406.

684

[53] Y.H. Jin, R. Obert, P.M.J. Burgers, R.A. Kunkel, M.A. Resnick, D.A. Gordenin, The 3’→5’

685

exonuclease of DNA polymerase δ can substitute for the 5’-flap endonuclease Rad27/Fen1 in

686

processing Okazaki fragments and preventing genome instability, Proc. Natl. Acad. Sci. U.S.A.

687

98 (2001) 5122-5127.

688

[54] A.J. Herr, M. Ogawa, N.A. Lawrence, L.N. Williams, J.M. Eggington, M. Singh, B.D.

689

Preston, Mutator suppression and escape from replication error-induced extinction in yeast,

690

PLoS Genet. doi: 10.1371/journal.pgen.1002282.

691

[55] G.T. Marsischky, N. Ford, M.F. Kane, R. Kolodner, Redundancy of Saccharomyces

692

cerevisiae MSH3 and MSH6 in MSH2-dependent mismatch repair, Genes & Dev. 10 (1996) 407-

693

420.

694

[56] T.A. Kunkel, S.S. Patel, K.A. Johnson, Error-prone replication of repeated DNA sequences

695

by T7 DNA polymerase in the absence of its processivity subunit, Proc. Natl. Acad. Sci. U.S.A.

696

91 (1994) 6830-6834.

Ac ce p

te

d

M

an

us

cr

ip t

677

32 Page 32 of 48

[57] A. Datta, J.L. Schmeits, N.S. Amin, P.J. Lau, K. Myong, R.D. Kolodner, Checkpoint-

698

dependent activation of mutagenic repair in Saccharomyces cerevisiae pol3-01 mutants, Mol.

699

Cell 6 (2000) 593-603.

700

[58] A. Morrison, A.L. Johnson, L.H. Johnston, A. Sugino, Pathway correcting DNA replication

701

errors in Saccharomyces cerevisiae, EMBO J. 12 (1993) 1467-1473.

702

[59] R.J. Kokoska, L. Stefanovic, H.T. Tran, M.A. Resnick, S.A. Gordenin, T.D. Petes,

703

Destabilization of yeast micro- and minisatellite DNA sequences by mutations affecting a

704

nuclease involved in Okazaki fragment processing (rad27) and DNA polymerase δ (pol3-t), Mol.

705

Cell. Biol. 18 (1998) 2779-2788.

706

[60] Y.H. Jin, R. Ayyagari, M.A. Resnick, D.A. Gordenin, P.M.J. Burgers, Okazaki fragment

707

maturation in yeast. II. Cooperation between the polymerase and 3’→5’ exonuclease activities

708

of pol δ in the creation of a ligatable nick, J. Biol. Chem. 278 (2003) 1626-1633.

709

[61] H.T. Tran, D.A. Gordenin, M.A. Resnick, The 3’→5’ exonucleases of DNA polymerases

710

delta and epsilon and the 5’→3’ exonuclease ExoI have major roles in postreplication mutation

711

avoidance in Saccharomyces cerevisiae, Mol. Cell. Biol. 19 (1999) 2000-2007.

712

[62] N.S. Amin, M-N. Nguyen, S. Oh, R.D. Kolodner, exo-1 dependent mutator mutations:

713

model system for studying functional interactions in mismatch repair, Mol. Cell. Biol. 21 (2001)

714

5142-5155.

715

[63] Y.H. Jin, A.B. Clark, R.J.C. Slebos, H. Al-Refai, J.A. Taylor, T.A. Kunkel, M.A. Resnick,

716

D.A. Gordenin, Cadmium is a mutagen that acts by inhibiting mismatch repair, Nat. Genet. 34

717

(2003) 326-329.

Ac ce p

te

d

M

an

us

cr

ip t

697

33 Page 33 of 48

[64] M. Hogg, S.S. Wallace, S. Doublié, Crystallographic snapshots of a replicative DNA

719

polymerase encountering an abasic site, EMBO J. 23 (2004) 1483-1493.

720

[65] S. Doublié, K.E. Zahn, Structural insights into eukaryotic DNA replication, Front.

721

Microbiol. 5 (2014) 444-454. doi: 10.3389/fmicb.2014.00444.

722

[66] L.J. Reha-Krantz, Md.S.P. Siddique, K. Murphy, A. Tam, M. O’Carroll, S. Lou, A. Schultz,

723

C. Boone, Drug-sensitive DNA polymerase δ reveals a role for mismatch repair in checkpoint

724

activation in yeast, Genetics 189 (2011) 1211-1224.

725

[67] R.D. Kuchta, P. Benkovic, S.J. Benkovic, Kinetic mechanisms whereby DNA polymerase I

726

(Klenow) replicates DNA with high fidelity, Biochemistry 27 (1988) 6716-6725.

727

[68] J.M. Beechem, M.R, Otto, L.B. Bloom, R. Eritja, L.J. Reha-Krantz, M.F. Goodman,

728

Exonuclease-polymerase active site partitioning of primer-template DNA strands and

729

equilibrium Mg2+ binding properties of bacteriophage T4 DNA polymerase, Biochemistry 37

730

(1998) 10144-10155.

732 733 734

cr

us

an

M

d

te

Ac ce p

731

ip t

718

735 736

34 Page 34 of 48

737 738

Figure legends

ip t

739

cr

740

Fig.1. (A) Illustration of the T4 DNA polymerase hairpin acting as a wedge to separate the

742

primer and template strands in exonuclease complexes; adapted from [21]. Phage T4 DNA

743

polymerase residues in the tip of the β hairpin, M253, Y254 and G255, correspond to RB69

744

DNA polymerase residues M256, Y257, and G258. (B) A crystallographic snapshot of the

745

hairpin in exonuclease complexes formed with the RB69 DNA polymerase [27]; republished

746

with permission from the authors. The exonuclease active site is identified by residues D114,

747

E116, D222A and D327A. Note that F123 is adjacent to the 3’-terminal adenine nucleotide on

748

the primer strand.

749

Fig.2. Comparison of the hairpin structures in RB69 DNA polymerase (yellow), yeast DNA

750

polymerase δ (cyan), and the atrophied hairpin in yeast DNA polymerase ε (light green, red

751

arrow). The background structure of DNA polymerase ε is in white [41]; republished with

752

permission from the authors.

753

Fig.3. Yeast DNA polymerase δ mutants display different sensitivities to HU and PAA. (A)

754

About 50 cells of each strain were spotted across a gradient of HU from 0 to 0.2M HU. The

755

plates were scanned after 4 days incubation at 30 °C. (B) About 50 cells of each strain were

756

spotted across a gradient of PAA from 0 to 4 mg/ml (0 to 28 mM). The plates were scanned

757

after 2 and 7 days of incubation. The C (control) lane shows cells that were not exposed to PAA.

Ac ce p

te

d

M

an

us

741

35 Page 35 of 48

Fig.4. The yeast hp-DNA polymerase δ mutant displays a slow-growth phenotype. About 100

759

cells were streaked in different sectors; the plate was incubated for 3 days. The pol3-hp strain

760

in the absence of MMR produced the smallest colonies.

ip t

758

Ac ce p

te

d

M

an

us

cr

761

36 Page 36 of 48

799

The authors declare that there are no conflicts of interest.

Ac ce p

te

d

M

an

us

cr

ip t

800

44 Page 37 of 48

Ac

ce

pt

ed

M

an

us

cr

i

Graphical Abstract (for review)

Page 38 of 48

cr

β 255 β 212 GWNIEGFDVPYIMNRVKMILGERSMKRFSPI----GRVKSKLIQNMYGSKEIYSID--GVSILDYL

b

M an

T4

us

Table 1 Construction of yeast DNA polymerase  hairpin mutants. DNA pol Protein Sequencea

ip t

Table

β 447 β ypolδ G447S,Q,D 400 GYNTTNFDIPYLLNRAKALKVNDFPYFGRLKTVKQEIKESVFSSKAYGTRETKNVNIDGRLQLDLL S,Q,D

400 GYNiegFDIPYLLNRAKALKVNDFPYFGRLKTVKQEIKESVFSSKAYGTRETKNVNIDGRLQLDLL YYYt4tYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYYY

yhpΔ

400 GYNTTNFDIPYLLNRAKALKVNDFPYFGRLKTVKQEIKESVF-----G--ETKNVNIDGRLQLDLL

chimeric ypolδ T4hpc

400 GYNIEGFDVPYIMNRVKMILGERSMKRFSPI----GRVKSKLIQNMYGSKEIYSID--GRLQLDLL YYYt4t4t4t4t4t4t4t4t4t4t4t4t4t4----t4t4t4t4t4t4t4t4t4t4t--YYYYYYYY

chimeric ypolδ T4hp ttnc

400 GYNTTNFDVPYIMNRVKMILGERSMKRFSPI----GRVKSKLIQNMYGSKEIYSID--GRLQLDLL YYYYYY4t4t4t4t4t4t4t4t4t4t4t4t4----t4t4t4t4t4t4t4t4t4t4t--YYYYYYYY

ce pt

Ac

a

ed

ypolδ ieg

Highly conserved residues are indicated in bold font [37]. The ExoII motif is underlined. Arrows indicate β hairpin strands. c Yeast residues in the chimeric DNA polymerase are indicated by “Y”, T4 residues are indicated by “t4”. b

Page 39 of 48

ip t

Table

Table 2 Spontaneous mutation rates for yeast pol3-hp strains.

Mutation rates per 108 cells His+ Lys+ (+1 fs) (-1 fs)

Strain Properties

Trp (Base subst.)

POL3 MMR+ a, b

2.9 (2.7-3.6)c

1.1 (0.9-1.6)

6.1 (5-10) N=35e

53 (47-72) N=29

us

cr

+

CanR (Base subst., fs, complex) 31 (27-39)

148 (104-300) N=15

363 (260-538) N=35

20

nd

350

23 (16-38) N=15

nd

333(133-384) N=15

42 (34-71) N=15

1074 (586-1605) N=15

936 (633-1461) N=15

1.2 (0.8-1.7) N=19

0.4 (