MGG

Molec. gen. Genet. 150, I 12 (1977)

9 by Springer-Verlag 1977

Excision-repair in Mutants of Escherichia coli Deficient in DNA Polymerase I and/or Its Associated 5I- 3 ' Exonuclease Priscilla Cooper Department of BiologicalSciences, Stanford University,Stanford, California94305

Summary. The ultraviolet (UV) sensitivity of Escherichia coli mutants deficient in the 5~--,31 exonuclease activity of DNA polymerase I is intermediate between that of pol + strains and mutants which are deficient in the polymerizing activity of pol I (po/A1). Like polA1 mutants, the 5'-exonuclease deficient mutants exhibit increased UV-induced D N A degradation and increased repair synthesis compared to a pol + strain, although the increase is not as great as in polA1 or in the conditionally lethal mutant BT4113ts deficient in both polymerase I activities. When dimer excision was measured at UV doses low enough to avoid interference from extensive DNA degradation, all three" classes of polymerase I deficient mutants were found to remove dimers efficiently from their DNA. We conclude that enzymes alternative to polymerase I can operate in both the excision and resynthesis steps of excision repair and that substitution for either of the polymerase I functions results in longer patches of repair. A model is proposed detailing the possible events in the alternative pathways.

Introduction The excision-repair of ultraviolet (UV) induced pyrimidine dimers in Escherichia coli DNA proceeds by a series of well-characterized enzymatic steps, recently reviewed by Grossman et al. (1975). An initial incision step, performed by the damage-specific uvrAB endonuclease, produces a nick on the 5' side of the pyrimidine dimer (Braun and Grossman, 1974). The dimer and adjacent nucleotides are excised and resynthesis takes place using the intact complementary strand as template. It now seems likely that more than one enzymatic pathway is available for excision and resynthesis and that any or all of several different enzymes may be involved (Hanawalt et al., 1975). Ultimately the repaired region is covalently joined to parental DNA by polynucleotide ligase.

Purified DNA polymerase I with its associated 5' ~ 3 ' exonuclease activity has been shown to perform both dimer excision and resynthesis in vitro (Kelly et al., 1969). PolA1 mutants, deficient in the polymerizing activity ofpolymerase I, are UV sensitive (De Lucia and Cairns, 1969), although the fact that they are intermediate in sensitivity between pol + and uvr strains suggests thay they are able to perform some excisionrepair (Monk et al., 1971). Boyle et al. (1970) found thatpoIA1 is able to excise dimers after low UV doses, but as shown by Kanner and Hanawalt (1970) the rejoining step is delayed. We have previously shown that polA1 mutants perform an increased amount of repair synthesis relative to that in a pol + strain after the same UV dose (Cooper and Hanawalt, 1972b). The same result is obtained when semiconservative replication is selectively inhibited in a dnaB mutant held at the restrictive temperature to rule out any contribution from normal DNA synthesis (P. Cooper, unpublished results). We have interpreted this result as due to an increase in a large patch component of repair synthesis during operation of some repair pathway alternative to that involving polymerase I (Cooper and Hanawalt, 1972a). DNA polymerases II and III can participate in excision-repair in the absence of polymerase I both in vivo (Youngs and Smith, 1973 ; Tait et al., 1974) and in toluene-permeabilized cells (Masker et al., 1973; Masker et al., 1975) and it is likely that either or both of these enzymes are involved in the extensive repair synthesis seen in the polA mutants, The dimer excision step as well as the resynthesis step could be performed by DNA polymerase I in vivo, and in fact it is now known that poIA1 mutants have nearly normal levels of the pol I 5I~3 ' exonuclease (Lehman and Chien, 1973). In addition, there are at least two other known enzymes which have the in vitro capability of excising dimers : exonuclease VII, isolated by Chase and Richardson (1974), and a 5 ' ~ 3 ' exonuclease activity associated with DNA polymerase III (Livingston and Richardson, 1975). It would be of interest to know whether the 5'-~3 /

P. Cooper : Excision-repair in D N A Polymerase I Deficient E. coli

exonuclease of polymerase I is the only dimer excising activity in vivo. Recently a number of mutants deficient in the 5I~ Y exonuclease activity of pol I have become available (Glickman et al., 1973; Konrad and Lehman, 1974), including one that is deficient in both pol I polymerizing and 5' ~ 3' exonuclease activities (Olivera and Bonhoeffer, 1974). The polAl07 mutant of Glickman is the least deficient of these, having 10-20% of the wild-type level of the polymerase I Y-exonuclease activity (Heijneker et al., 1973). The 5'--+3' exonuclease deficient mutants polAexl and polAex2 isolated by Konrad and Lehman are slightly temperature-sensitive for enzyme activity, withpolAexl having 5-6% the normal level of 5 ' ~ 3 ' exonuclease activity a t 30 ~ and 2% or less at 43 ~ (Uyemura etal., 1976) and poIAex2 being somewhat more deficient than po/Aex 1 in 5'-exonuclease activity and showing a slight decrease in pol I polymerizing activity as well (Lehman, personal communication). The doubly-deficient mutant BT4113 ts of Olivera and Bonhoeffer has 15% of the wild type pol I 5 ' ~ 3 ~ exonuclease activity but only 0.5% of the normal polymerizing activity at 45 ~ and it is inviable at 45 ~ At 30 ~ the 5'-exonuclease activity is nearly normal (80% of wild type) while the polymerizing activity is 21% of the wild type level (Olivera and Bonhoeffer, 1974). In order to evaluate the possible participation of the various available nucleases in the excision step of repair, we have examined the repair capabilities of strains deficient in the pol I 5'-exonuclease activity in comparison to thepoIA1 polymerase-deficient strain following exposure to ultraviolet light. We find that both UV-induced DNA degradation and repair synthesis are increased in the 5'-exonuclease deficient mutants relative to that in wild type strains but that the increase is not as great as in mutants deficient in either the pol ! polymerizing activity alone or in both the polymerizing and 5'-exonuclease activities of pol I. All the mutants tested are able to excise dimers efficiently at doses for which extensive degradation does not interfere with the measurement. We conclude that other enzymes are able to substitute for both polymerase I activities in the excision-repair of pyrimidine dimers.

Materials and Methods

Table 1. Strains of Escherichia coli K 12 Strain

Source

Genotype

Konrad and Lehman (1974) and personal communication

W3110 F trpA33 rha thy-

RS5064

Konrad and Lehman (1974) and personal communication

W3110 F - trpA33 rha- thy- polAexl

RS5069

Konrad and Lehman (1974) and personal communication

W3110 F trpA33 rha thy- polAex2

JGi38

Monk et al (1971)

F rha lacZ thy- polA1

JG139

Monk et al. (1971)

F- rha lacZ thy-

KMBL 1788

Glickman (1974)

F argAlO3pheA97 bio87 endAlO1 thyA

KMBL 1789

Glickman (1974)

F argAlO3pheA97 bio87 endAlO1 thyA polA 107

BT4113 tr

Olivera and Bonhoeffer (1974)

W3110 thy- met (pol + by transduction)

BT4113 ts

Olivera and Bonhoeffer (1974)

W3110thy met(polymerase I - )

KS463

glucose-salts medium (RT minimal medium) as previously described (Hanawalt and Cooper, 1971). Growth medium routinely contained 0.5% glucose, 0.1% casamino acids, 2 gg/ml thymine, plus required amino acids at 20 gg/ml. In the case of the BT4113 strains, guanosine was necessary for growth and was supplied at 20 gg/ml. In certain experiments approximately 0.5 M NaC1 (3% w/v) was added to the medium. Most plating experiments for viability assay utilized Difco nutrient agar. Glucose minimal plates were identical to the liquid growth medium with added 1.5% Difco Bacto-agar. TGY plates contained, per liter: 3 g Difco yeast extract, 5 g Bacto-tryptone, and 1 g glucose; TYE plates contained, per liter: 5 g yeast extract, 10 g Bacto-tryptone, and 8 g NaC1; both contained 1.5% Bactoagar. Plates were incubated at 33 ~ or 43 ~

UV Irradiation Exponentially growing cultures at a density of about 2 x 108 cells/ml were filtered onto 9-cm Schleicher and Schuell membrane filters (pore size 0.45 It) and washed with RT buffer. Ceils were resuspended in RT buffer with no added glucose or other supplements and cell densities of parallel cultures were adjusted to give equal readings on a Klett-Summerson colorimeter with blue filter. Cell density for irradiation was usually 2-3 x 108 cfu/ml; 20 ml volumes were irradiated in 14-cm diameter petri dishes with a Philips 15-W germicidal lamp. The incident dose rate was 8.8 ergs/mm2/s as determined by an IL254 Germicidal Photometer (International Light). Cell suspensions were stirred during irradiation by a rotating platform shaker. Irradiated cultures were kept in the dark or under yellow light during all subsequent operations to prevent photoreactivation.

Bacterial Strains and Growth Conditions The E. coli K12 strains used are listed in Table 1; they are all derivatives of W3110 except KMBL 1788 and 1789 which were derived from W1485. Bacteria were diluted from overnight cultures and grown at 33 ~ (37 ~ for KMBL 1788 and 1789) with shaking in a Tris-buffered

DNA Degradation Cultures were labelled during at least three generations of exponential growth with 0.1 gCi/ml 14C-thymine; unlabelled thymine was supplied at 2 gg/ml. Cell suspensions were irradiated as described

P. Cooper: Excision-repair in D N A Polymerase I Deficient E. coli

3

above. Irradiated cells were then placed in prewarmed flasks (33 ~ or 43 ~) containing supplements required for growth and incubated with shaking. Immediately after irradiation and at intervals thereafter 0.5 ml aliquots were withdrawn into excess cold 5% trichloroacetic acid ; duplicate samples were taken for each time point. Precipitated sampIes were collected on MilIipore HA filters. Radioactivity was determined in a Packard "friCarb scintillation spectrometer with a toluene-based counting solution.

assayed for radioactivity in a mixture consisting of 1 part Triton X-100 and 2 parts toluene-based scintillation fluid.

Repair Synthesis Cultures were labelled for at least three generations with low levels of 14C-thymine, 0.025 gCi/ml in some experiments and 0.0125 gCi/ ml in others ; in either case, 2 gg/ml unlabelled thymine was supplied. Irradiated and unirradiated portions of each culture (10 ml volumes) were added to prewarmed flasks containing growth supplements. Thymine was replaced by 10 gg/ml 5-bromouracil; high specific activity methyl-3H-5-BrUra ( ~ 2 5 Ci/mM) was supplied as repair label at either 5 gCi/ml or 10 gCi/ml depending on the experiment. Lysis and isopycnic gradient analysis were performed essentially as described by Hanawalt and Cooper (1971) except that lysozyme was used at 100 ~tg/ml for 30 rain at 37 ~ incubation with pronase was continued for two hours, and 8.2g of technical grade CsC1 (Kawecki Berylco Industries, Inc., Penn Rare Metals Division) was used for each gradient.

Results

Viability and UV Sensitivity Although Konrad and Lehman (1974) reported that the polAexl and polAex2 mutants are inviable at 43 ~ we have found that under several plating conditions (i.e., minimal medium, nutrient agar, or tryptone-yeast extract) these strains are viable at 43 ~ Lethality occurs at the high temperature only when salt is added to the medium, as in the TYE plates used by Konrad and Lehman (Table 2). As shown in Figure 1, the polAex2 mutant is sensitive to increasing NaC1 concentrations on RT-minimal media even when incubated at the "permissive" temperature, and it is extensively killed (0.001% survival) by 3% added NaC1 (w/v) at 43 ~ Under the same conditions, the pol + strain KS463 is unaffected by added salt at 43 ~ and actually I00

Dimer Excision For each experiment, 40 ml cultures of each bacterial strain were heavily labelled by at least three generations of growth in RT medium containing 2 gg/ml thymine and high specific activity methyl-3H-thymidine ( ~ 5 0 Ci/mM) at a concentration of 10 1-~Ci/ ml. Each culture was irradiated in two 20-mI portions which were then pooled and added to prewarmed flasks containing 40 ml RT medium with twice the concentration of growth supplements. The resulting cell density was about 108/ml. For each time point, duplicate 10 ml samples were withdrawn into ice-cold centrifuge tubes containing one of two different stop mixes, as follows. In the first method of preparing samples for hydrolysis, the stop mix consisted of an equal volume of NET buffer (0.1 M NaCI, 0.01 M EDTA, 0.01 M Tris-HC1, pH 8.0). Cells were collected by centrifugation, resuspended in 1 ml of a 1 : 10 dilution of NET and incubated at 50~ for 30 min with 50 gg/ml proteinase K (E. Merck). Samples were made i0% in tricholoroacetic acid, chilled and the acid-precipitable fraction was collected by centrifugation in the HB4 swinging bucket rotor of a SorvaI RC2-B centrifuge. In the second method, which gave lower backgrounds on the resulting chromatograms and was developed for use in the low dose experiments, the stop mix consisted of 0.2 ml 10 N KOH, giving a final alkali concentration of 2N. The acid-precipitable fraction was then collected by adding 1 ml 100% trichloroacetic acid (w/v), chilling, and centrifuging for I5 min at 10 K rpm. The pellet was carefully drained and resuspended in 2 ml buffer containing 50 m M Tris, 10 m M EDTA (pH 8) and 50 p_g/ml proteinase K. Incubation and subsequent acid-precipitation were as described above. The TCA pellets were dried thoroughly, then dissolved in 0.25 ml 98% formic acid and transferred to Pyrex ignition tubes which were then sealed. Hydrolysis was carried out for 40 min at 180~ Hydrolysates were evaporated to dryness, dissolved in 10 gl distilIed ~{20, and spotted on chromatography plates. The chromatographic separation of pyrimidine dimers and monomers was carried out by the one dimensional method of Cook and Friedberg (1976) using Brinkmann M N siIica gel plates (Polygram Sil G/UV2s4) and a solvent system consisting of ethyl acetate saturated with H20 and n-propanol. Cut up chromatograms were

10

1.0

>o

0.1 ~1

84

9

pol +, 43 ~

A

ex2, 33 ~

9

ex2, 43 ~

0.01

0.001

0

I

I

i

I

2

I

3 4 % ADDED NoCI (w/v)

i

5

Fig. 1. Effect of NaC1 on Cell Survival. Exponentially growing cultures of KS463 Col +) and RS5069 (po/Aex2) at 33 ~ in RT minima1 medium were diluted in RT buffer and plated on glucose minimal medium containing varying concentrations of added NaCI. Plates were incubated for 60 h at either 33 ~ o1 43 ~ and colonies were counted. 9 KS463 pol +, 43 ~ incubation; z~ RS5069 polAex2, 33~ incubation ; 9 RS5069 poIAex2, 43 ~ incubation

P. Cooper: Excision-repair in D N A Polymerase I Deficient E. coli

4 Table 2. Colony forming ability on various media. Cultures growing exponentially at 33 ~ in RT minimal medium were diluted in RT buffer and plated as describe in Methods Medium

Nutrient agar

cfu/ml x 10- ~

KS463pol +

RS5064po/Aex 1

RS5069polAex2

33 ~

43 ~

33 ~

43 ~

33 ~

43 ~

221

218

276

316

262

230

TGY

216

TYE

406

145 (46 %) 0.169 (0.05%)

140 (61%) 0.0436 (0.02%)

ibo

ments to be reported here have been performed in the absence of added salt. When plated on nutrient agar, a medium which gives no temperature effect on plating efficiency (Table 2), both the polAex 1 and poIAex2 mutants are intermediate in sensitivity to UV between a pol + strain and the polA1 mutant (Fig. 2) in agreement with an earlier report for the poIAl07 mutant (Glickman, 1974). This is true whether incubation is at 33 ~ or 43~ both mutants are slightly more sensitive at the high temperature. Although it is not possible to assess UV sensitivity of the conditionally lethal BT4113ts mutant at the restrictive temperature, at the permissive temperature we have found slightly increased UV sensitivity compared to the pol § transductant BT4113tr. Under these conditions the pol § strain shows 30% survival on nutrient agar plates compared to 10% for the mutant after exposure to a UV dose of 200 ergs/mm 2.

O

UV-induced DNA Degradation \\\ X i.o

\\

03 k9

33 ~ 4 3 ~ 0 9 KS [] 9 R S 5 0 6 4 e• RS5069 ex2 Ax 9 JGI 5 8 polAl

QO

\ \

\ x\ o.ool

I

0

I00

I

I

I

I

200 300 400 500 E r g s / m m 2 , INCIDENT

I

600

Fig. 2. Comparative sensitivites of different strains to ultraviolet irradiation. Cells grown to late log phase at 33 ~ were irradiated in RT buffer as described in Materials and Methods and plated on prewarmed nutrient agar. Incubation was at either 33 ~ or 43 ~ o KS463 pol +, 33 ~ incubation, 9 43 ~ incubation; [] RS5064 polAexl, 33 ~ incubation, 9 43 ~ incubation; zxRS5069 polAex2, 33 ~ incubation, 9 43 ~ incubation; x JG138 polA1, 33 ~ incubation

exhibits about a 50% increase in ptating efficiency at 33 ~ when 1-2% NaC! is added (data not shown). The purified 5' ~ Y exonuclease has not yet been tested for adtivity in the presence of added NaC1 (I.R. Lehman, personal communication), so it is not known whether salt-sensitivity of the enzyme at high temperature is responsible for the lethality. Most of the experi-

PolA1 mutants degrade their D N A extensively after UV irradiation (Boyle et al., 1970; Cooper and Hanawait, 1972 b). That this degradation is probably not exclusively due to action of the recB, C gene product, exonuclease V, was shown by Strike and Emmerson (1974), who found that the double mutants polA1 r e c B 2 1 and polA12 recB21 irradiated with high UV doses degrade their D N A to approximately the same final extent as thepol- single mutants, though at lower initial rates. These authors suggested that the remaining pol I 5'~3' exonuclease activity is responsible for UV-induced breakdown when the pol I polymerizing activity is deficient. To test this, we determined D N A degradation after a dose of 200 ergs/mm 2 in the three exonuclease deficient mutants and in polA1 (Fig. 3). The pol § strains KS463 and BT4113tr (not shown) show no measurable loss of acid-precipitable ~4C-thymine at either 33 ~ or 43 ~ while poIA1 degrades approximately 50% of its D N A in 60 rain of post-UV incubation at either temperature. The poIAexl and poIAex2 mutants lose less than 10% of their acid-precipitable D N A during this time at 33 ~ and only slightly more at 43 ~ The presence of 3% added NaC1 in the post-UV incubation medium has no appreciable effect on the D N A degradation in potAex2 (data not shown). 9 BT4113ts also exhibits limited degradation at 33 ~ but solubilizes 30-35% of its D N A in an hour at 43 ~ after the 200 ergs/mm 2 dose. The degradation induced by higher UV doses was compared at 43 ~ for polAex2 and polA1 (Fig. 4). The maximum rate of degradation in polA 1 is reached after 200 ergs/mm 2 and no further increase is seen up to

P. Cooper: Excision-repair in D N A Polymerase I Deficient E. coli

I O 0 ~ A I ~ "~..~.~==~"~.!=~ 1-: 9 0 - [ - ~ ~ ~ nQ" + ~

5

,..~.KS465 pol+ 33eRS5064 polA exl ~:]-BT4II5ts RS5069 polA ex2

&8o

,~ I \

Z~60-

Z

9

0-

~" BT41[3 fs

JGI38poIAf

5o-

JGI38 pOIAI

w 40< 30Z

Fig. 3 a and b. D N A Degradation during incubation at (a) 33 ~ or (b) 43 ~ after exposure to a U V dose of 200 ergs/mm 2. The experiment was

performed as described in Methods.

E{ 3 2 0 -

100 % = 5600 - 9263 l,C_cpm, zx KS463

,~-~ IO-

pol+ ; 9 RS5064 polAexl ; [] RS5069 polAex2; 9 BT4113ts; 9 JG138 polAl

I0

I0

20 30 40 50 60

20 30 4 0 5 0 6 0

TIME (rain.)

(a) pol-t-

(b) polAex2

0

(c) polAI

n~ IOO (3s