Mol Gen Genet (1986) 205:9-13 9 Springer-Verlag1986

Structure and function of dnaQ and mutD mutators of Escherichia coil Kyoko Takano, Yusaku Nakabeppu, Hisaji Maki, Takashi Horiuchi, and Mutsuo Sekiguchi Department of Biology, Faculty of Science and Department of Biochemistry, Faculty of Medicine, Kyushu University, Fukuoka 812, Japan

Summary. The nucleotide sequences of the recessive dnaQ49 and the dominant mutD5 mutator were determined. The dnaQ49 mutator has a single base substitution in the dnaQ gene, thus causing one amino acid change, 96Val ( G T G ) ~ Gly (GGG), in the DnaQ protein (e subunit of D N A polymerase III holoenzyme). The mutD5 mutator possesses two base substitutions in the same gene, resulting in two amino acid changes, 73Leu ( T T G ) ~ T r p (TGG) and 164Ala ( G C A ) ~ V a l (GTA), which were designated the mutD52 and mutD51 mutations, respectively. Construction of chimaeric genes carrying one or two of these mutations revealed: (1) either mutD51 or mutD52 alone causes the dominant mutator phenotype when present in a multi-copy plasmid; (2) mutD51, but not mutD52, exerts the dominant mutator phenotype when present in a low-copy plasmid; (3) the dominant mutD51 mutator activity is suppressed by the dnaQ49 mutation when both mutations are present in the same gene. Based on these findings, we devised a model for the action of these mutators.

Key words: Mutator

D N A polymerase III - dnaQ gene

Suppression

Introduction There is a special class of mutations that increases the frequency of spontaneous alterations of many genes. These mutations are collectively termed mutators and are useful tools for elucidating cellular mechanisms involved in replication and preservation of genetic information. In Escherichia coli, several classes of mutators have been characterized by genetic and biochemical methods (Cox 1976; Sekiguchi et al. 1982). The mutD5 and dnaQ49 mutators are among the most notable as they cause extremely high frequencies of mutation. Although they are located at almost the same position on the E. coli chromosome, they have different phenotypes. The dnaQ49 mutant shows defective D N A synthesis under restrictive conditions and exhibits high mutator activity at teperatures of 35 ~ C or higher (Horiuchi et al. 1978) whereas mutD5 shows no conditional lethality or temperature dependency for mutator activity (Degnen and Cox 1974; Cox and H o m e r 1982). Furthermore, the mutD5 mutator is dominant over the wild-type allele whereas dnaQ49 is recessive (Maruyama et al. 1983).

Offprint requests to: M. Sekiguchi

The nucleotide sequence of the dnaQ gene has been determined (Maki et al. 1983). In our attempt to clarify structural and functional relationships between the dnaQ and mutD mutators, we determined the nucleotide sequences of these mutators and constructed chimaeric genes carrying one or both of the mutations.

Materials and methods Bacteria, plasmids andphages. All the bacterial strains used in these experiments are derivatives of E. coli K12. Strains KHl171 ( F - , fecal, dnaQ +) and DH1 (Hanahan 1983) were used for measurement of mutation frequencies. Strain WB373 and vector phage mWB2342 (Barnes et al. 1983) used for D N A sequence analysis were obtained from K. Kanda. The pBR322-derived plasmids pMM5, pMM8 and pMM9 carrying dnaQ +, mutD5 and dnaQ49, respectively, have been described previously (Maruyama et al. 1983). The vector plasmid p M F 3 (Manis and Kline 1977) used to construct low-copy plasmids carrying the mutant alleles was obtained from H. Shinagawa. The vector plasmid pKT30 used for the cis-trans test was constructed by inserting the kanamycin-resistant fragment derived from p U C 4 K (Vieira and Messing 1982) into the BamHI site of pACYC184 (Chang and Cohen 1978).

DNA sequencing. The 1.6 kb EcoRI fragments derived from pMM5, pMM8 and pMM9 were inserted into the EcoRI site of mWB2342, in both directions. These hybrid phage replicative form (RF) DNAs were digested with the restriction enzymes shown in Fig. 1 and also with HindIII, blunt ended and ligated to form deletion phages. These phage DNAs were used as templates for D N A sequencing. D N A sequence determination was carried out by the dideoxy method (Barnes et al. 1983). Determination of mutation frequency. Cells harbouring various plasmids were grown overnight at 30 ~ C or 37 ~ C in L-broth. Aliquots of the cultures were plated on L-broth agar plates containing rifampicin or streptomycin at 100 lag/ ml. The plates were incubated at 30 ~ C or 37 ~ C for 2 days, and colonies formed were scored as resistant mutants. The number of viable cells was determined by plating aliquots of the cultures on L-broth agar plates and incubating at 30 ~ C or 37 ~ C for 1 day. The mutation frequency was calculated by dividing the number of resistant cells by the total number of viable cells.

10 E

A

I

q

P

B

I

9

9 P

M

T

Other methods. Transformation and plasmid D N A preparation were performed according to Maniatis et al. (1982). Transfection and phage D N A preparation were performed according to Barnes et al. (1983).

E

t,

I

fl b

9

~

I

I

Results

200 bp

Fig. 1. Strategy of DNA sequencing. The line indicates a 1.6 kb EcoRI fragment carrying dnaQ +, dnaQ49 or mutD5. It was subcloned into mWB2342, and parts of the sequence were deleted with the use of appropriate restriction enzymes to prepare sequencing templates. The dnaQ coding region is indicated by a box. Arrows indicate the location, direction and length of each sequence determined. Symbols: E, EcoRI; A, AccI; P, PvuI B, BstEII; M, MluI

AccI 5 '- G - ~ G A C T T C C T G T A A T T G A A T C G A A C T G T A A A A C G A C A A G T C T G A C A T A A A T GACCGCT

52

ATG AGC ACT GCA ATT ACA CGC CAG ATC GTT CTC GAT ACC Met Ser Thr Ala Ile Thr Arg Gln Ile Val Leu Asp Thr

98

GAA ACC ACC GGT ATG AAC CAG ATT GGT GCG CAC TAT GAA GGC CAC GIu Thr Thr Gly Met Asn Gln Ile Gly Ala His Tyr Glu Gly His

143

AAG ATC ATT GAG ATT GGT GCC GTT GAA GTG GTG AAC CGT L y s I l e Ile G l u Ile G l y A l a V a l G l u Val V a l A s n A r g Pvul ACG GGC AAT AAC TTC CAT GTT TAT CTC AAA CCC GAT CGG T h r G l y A s n A s n P h e H i s V a l T y r L e u Lys P r o A s p A r g

188

CGC CTG Arg Leu

CTG GTG 233 Leu Val G(D52) GAT CCG GAA GCC TTT GGC GTA CAT GGT ATT GCC G~T GA~ TTT TTG 278 A s p P r o G l u A l a P h e G l y V a l H i s G l y Ile A l a A s p G I u P h e L ~ u CTC GAT AAG CCC ACG TTT GCC GAA GTA L e u A s p Lys Pro T h r P h e A l a G l u Val G(Q49) TAT ATT CGC GGC GCG GAG TTG GTG ATC Tyr Ile A r g G l y A l a G l u L e u V ~ I Ile

GCC G~T GAG TTC ATG GAC A l a A s p G l u Phe M e t A s p

323

CAT AAC GCA GCG TTC GAT His A s n A l a A l a P h e A s p

368

ATC GGC TTT ATG GAC TAC GAG TTT TCG TTG Ile G l y P h e M e t A s p T y r G l u P h e Ser L e u BstEII ~ C G A A G A C C A A T A C T T T C T G T A A G G--~ A C C P r o L y s T h r A s h T h r P h e C y s Lys V a l T h r

CTT AAG CGC GAT ATT L e u L y s A r g A s p Ile

413

GAT AGC CTT GCG GTG Asp Set Leu Ala Val

458

GCG AGG AAA ATG TTT CCC GGT AAG CGC AAC AGC CTC GAT GCG TTA A l a A r g Lys M e t P h e P r o G l y Lys A r g A s h S e r L e u Asp A l a L e u

503

TGT GCT Cys Ala T(D51) GCA TTA A[a Leu

CGC TAC GAA ATA GAT AAC AGT AAA CGA ACG CTG CAC GGG A r g T y r G l u Ile A s p A s n S e t L y s A r g T h r L e u H i s G l y

548

CTC GAT GCC CAG ATC CTT GCG GAA GTT TAT CTG GCG ATG L e u A s p A l a G l n Ile L e u A l a G l u Val T y r L e u A l a M e t

593

ACC GGT GGT CAA ACG TCG ATG GCT TTT GCG ATG GAA GGA GAG ACA Thr Gly Gly Gln Thr Ser Met Ala Phe Ala Met Glu Gly Glu Thr

638

CAA CAG CAA CAA GGT Gln Gln Gln Gln Gly MIuI GCA AGT AAG TTA C-~ A l a Ser L y s L e u A r g

GAA GCA ACA ATT CAG CGC ATT GTA CGT CAG Glu Ala Thr s G l n A r g I l e Val A r g G l n

683

GTT GTT TTT GCG ACA GAT GAA GAG ATT GCA V a l V a l Phe. A l a T h r A s p G l u G l u Ile A l a

728

GCT CAT GAA GCC CGT CTC GAT CTG GTG CAG AAG AAA GGC GGA AGT A l a H i s G l u A l a A r g L e u A s p L e u Val G l n L y s L y s G l y G l y S e r

773

TGC CTC TGG CGA GCA TAA ATACCTGTGAAAGGCGCTAAAAA-3' C y s L e u T r p A r g A l a ***

814

Fig. 2. Nucleotide sequence of the dnaQ coding region, given 5' to 3'. The nucleotides are numbered from the AeeI cleavage site. The restriction sites for AeeI, PvuI, BstEII and M,!uI are shown. Open triangles show revised bases. Base substitutions observed in dnaQ49 and mutD5 and restriction sites are underlined

Construction of chimaeric genes. Plasmid pMM5, pMM8 or pMM9 was digested with BstEII and PstI to produce a 4.6 kb large fragment and a 1.6 kb small fragment. Each was purified to construct plasmids pKT14,' pKT15 and pKT16 (Fig. 4). The 1.4 kb EcoRI fragments of pMM8, pKT14 and pKT15 were then recloned into pMF3 to construct pKT22, pKT24 and pKT25. The 1.6 kb EcoRI fragments of pKT14 and p K T I 6 were recloned into pKT30 to construct pKT31 and pKT33.

Nucleotide sequences of dnaQ49 and mutD5 mutators The strategy for the sequence analysis is shown in Fig. 1. Over 80% of the dnaQ region was determined from both strands of the DNA. The nucleotide sequence of dnaQ + was slightly different from the sequence previously determined by the chemical degration method (Maki et al. 1983). There are five alterations in the dnaQ region, apparently due to ambiguity in some parts of the gels used in previous experiments. The revised nucleotide sequence of the wildtype dnaQ gene is shown in Fig. 2. The nucleotide sequences of dnaQ49 and mutD5 were determined simultaneously (Fig. 3). The dnaQ49 allele has one base change from the wild-type allele, resulting in alteration of one codon; 96Val ( G T G ) ~ G l y (GGG). With this change, a new BamHI site would be produced, and this was confirmed by restriction enzyme analysis (data not shown). The mutD5 allele has two base changes; 73Leu ( T T G ) ~ T r p (TGG) and 16~Ala ( G C A ) ~ V a l (GTA). We designated these two mutations, mutD52 and mutD51, respectively. These results are included in Fig. 2. Mutator activity of chimaeric genes To determine whether mutD51 or mutD52, or both, is responsible for mutator activity, we constructed chimaeric genes carrying one or other of the base substitutions (Fig. 4). Cells transformed with these plasmids were examined for their mutator activity (Table 1) and both mutD51 and mutD52 were found to cause a dominant mutator phenotype when present in a multi-copy plasmid. We also constructed a chimaeric plasmid, pKT16, carrying both the mutD51 and dnaQ49 mutations, the properties of which will be discussed below. Effect of copy number To determine whether or not the dominant characteristics of mutD51 and mutD52 are expressed when present in a low-copy plasmid, the D N A fragments carrying each of the mutations were recloned into the miniF plasmid pMF3 (Manis and Kline 1977). Plasmid pKT24 (mutD51) showed mutator activity in dnaQ + cells, indicating that mutD51 causes a dominant mutator phenotype, even in the merodiploid (Table 2). On the other hand, pKT25 (mutD52) induced only a low frequency of mutation. Unexpectedly, the control plasmid pKT22, a derivative of pMM8, carrying both mutD51 and mutD52, had no apparent mutator activity. Thus, mutD52 suppresses the effect of the mutD51 mutation when present in a low-copy vector. Suppression of dominant mutator mutD51 by recessive mutator dnaQ49 Cells harbouring the chimaeric plasmid p K T I 6 carrying both the dnaQ49 and mutD51 mutations showed only a low mutation frequency (see Table 1). This suggests that

11 Table 1. Mutation frequencies of bacteria harbouring multi-copy

wild

dnaQ49

mutD5

GCATGC

GCATGC

plasmids with mutators Plasmid (genotype)

Mutation freqeuncy ( x 10 -9)

GCATGC Rifampicinresistant pBR322 pMM5 (dnaQ+) pMM8

Streptomycinresistant

5.1 + 3.6 0.55 + 0.55 13 _+ I 0 41,000 _+ 7,000 11,000 _+1,000

(mutD51, mutD52) pKT14 (mutD51) pKT15 (mutD52) pKTI6

70,000 +12,000 30,000 __. 7,000 21 _+ 5

13,000 _+5,000 3,700 + 1,000 0.63_+ 0.63

(mutD51, dnaQ49) The Escheriehia coli strain DHI (dnaQ+) was transformed with the plasmids shown. Triplicate cultures were grown overnight in L-broth containing 50 gg/ml ampicillin at 30 ~ C. The mutation frequency was determined as described in Materials and methods Table 2. Mutability of bacteria harbouring low-copy plasmids with mutators

Plasmid (genotype)

Mutation frequency ( x 10-9)

pKT22 (mutD51, mutD52) pKT24 (mutD51) pKT25 (mutD52) pMF3

14 1,200 49 2.9

The Eseherichia coli strain KH1171 (dnaQ+) was transformed with the plasmids, and grown overnight in L-broth containing 50 gg/ml ampicillin at 37~ C. Mutation to rifampicin resistance was measured Table 3.

Cis-trans test for suppressor activity of dnaQ49

Plasmid (genotype) A

B

Mutation frequency ( X 10 -9)

PBR322 pBR322 pMM9 (dnaQ49) pKT16

pKT30 pKT31 (mutD51) pKT30 pKT30

73 + 24 14,000_+2,000 600 + 600 900_+ 900

(dnaQ49, mutD51) pBR322

pKT33

pMM9 (dnaQ49)

(dnaQ49, mutD51) pKT31 (mutD51)

120 +_

20

46,000 _+10,000

Strain DHI was transformed with a set of series A and B plasmids, which are derived from pBR322 and pACYC184, respectively. Triplicate cultures were grown at 37~ in L-broth containing 50 ktg/ml each of ampicillin and kanamycin. Mutation to rifampicin resistance was measured

the dnaQ49 m u t a t i o n suppresses the d o m i n a n t m u t a t o r activity caused by rnutD51. To determine whether such suppression occurs only when the two m u t a t i o n s are present in the same gene, we performed a cis-trans test. F o r this purpose, mutD51 and dnaQ49 were placed on separate plasmids belonging to different compatibility groups. As shwon in Table 3, cells harbouring such plasmids ( p M M 9 and pKT31) exhibited high m u t a t o r activity. Thus, the effect o f dnaQ49 is cis specific.

Fig. 3. Parts of DNA sequencing gel patterns showing the differences in the dnaQ+, dnaQ49 and mutD5 alleles. Arrows indicate the altered sites of the mutants

Discussion

W e had previously cloned the dnaQ + gene, identified its p r o d u c t and suggested that it might be the s subunit o f the D N A polymerase I I I holoenzyme (Horiuchi et al. 1981). Subsequently, Echols and associates found that the dnaQ gene p r o d u c t is indeed the e subunit and carries tile 3'--*5' exonuclease activity (Echols et al. 1983; Scheuermann and Echols 1984). They further d e m o n s t r a t e d that enzyme preparations derived from dnaQ49 a n d mutD5 m u t a n t s are defective in the 3'--*5' exonuclease activity. This and other evidence strongly suggested that the mutD5 and dnaQ49 mutators are m u t a t i o n s that have arisen in a single gene, dnaQ (mutD) ( M a r u y a m a et al. 1983).

12 A

a-subunit

EooRI

Accl

T

"

Pvul

T

BstEII

Mlul

T

T

G-subunit plasmid

P

rnutator

EcoRI

j wild

Pstl

dnaQ49

mutD51 pMM8 pKT14

mutD52

+

mutD51

i

pKT15

I

pMM9

r////,,,~'/////I//l

pKT16

I///JIV/ 9 dnaQ49

Fig. 4A, B. Organization of chimaeric genes. A Structure of mutD5 plasmid pMM8. A filled box and an arrow indicate the dnaQ coding region and direction of transcription, respectively. B Chimaeric genes. Open boxes, filled boxes and hatched boxes indicate the sequences derived from pMM5 (wild type), pMM8 (mutD5) and pMM9 (dnaQ49), respectively

However, because o f the d o m i n a n t phenotype of the

mutD5 mutant, these possible m u t a t i o n s are difficult to examine by genetic complementation tests. This p r o b l e m has now been overcome by the present nucleotide sequence analysis. Both dnaQ49 and mutD5 mutations were detected in the region o f the dnaQ gene which codes for the e subunit o f D N A polymerase III holoenzyme. The dnaQ49 mutation, a transversion T: A ~ G :C, seems to be caused by another mutator, rout-21, present in an original dnaQ m u t a n t and its parental strain (Horiuchi et al. 1978). The rout-21 m u t a t o r is considered to be an allele o f the m u t t gene, a strong m u t a t o r specifically producing T : A ~ G : C transversions, as determined from its m a p position and m u t a t i o n a l specificity. Unexpectedly we found two base substitutions, C : G ~ T : A (mutD51) and T : A ~ G : C (mutD52), in the mutD5 D N A sequence. By constructing chimaeric genes carrying either mutD51 or mutD52, we noted that each m u t a t i o n causes a d o m i n a n t m u t a t o r phenotype when present in a multi-copy plasmid. However, when each m u t a t i o n was cloned in a low-copy plasmid, only mutD51 showed the strong m u t a t o r phenotype. Moreover, the low-copy plasmid p K T 2 2 carrying both mutD51 and mutD52 did not increase the spontaneous m u t a t i o n frequency o f the host cell, thereby indicating that the mutD5 analysed in this w o r k was not a d o m i n a n t mutator. These findings plus previous observations suggest that the original mutD5 m u t a n t possessed only one mutation, mutD51, and later acquired a second mutation, mutD52, which abolished the d o m i n a n t character o f mutD51. To explain the action of d o m i n a n t and recessive m u t a t o r mutations in the dnaQ gene, we p r o p o s e d a model in a previous report ( M a r u y a m a et al. 1983); the mutD5 muta-

dnaQ49 rnutD51

Fig. 5a-d. A model explaining the dominant and recessive effects of mutators, a dnaQ+ plasmids in the wild-type strain. The e subunit (small open circle) which possesses editing nuclease activity binds to the ~ subunit (large circle), to produce DNA polymerase III which can perform high fidelity DNA replication, b dnaQ49 plasmids in the wild-type strain. The e subunit of dnaQ49 (open square) is not able to bind to tightly as that of dnaQ+, thus yielding the normal polymerase with the normal e (dnaQ+) subunit, e mutD51 plasmids in the wild-type strain. The e subunit of mutD5I (closed circle) can bind to the ~ subunit as tightly as the wild-type subunit, thus yielding the mutator polymerase with a low fidelity.

d dnaQ49-mutD51 chimaeric plasmids in the wild-type strain. The e subunit of the chimaera (closedsquare) which possesses abnormal nuclease activity is unable to bind to the ~ subunit when the normal counterpart is present. Replication can be performed by the polymerase with the normal e subunit derived from the wild-type chromosome

tion reduces the fidelity function o f D n a Q protein while the dnaQ49 m u t a t i o n reduces its ability to bind other components of the D N A replication enzyme. I f this model is valid, it would be expected that the dnaQ49 m u t a t i o n would diminish the d o m i n a n t character o f the mutD5 m u t a t i o n when both were present in the dnaQ gene. The present study revealed that such is indeed the case. By constructing the chimaeric plasmid p K T 1 6 which carries b o t h dnaQ49 and mutD51, it was shown that the dnaQ49 m u t a t i o n suppresses the d o m i n a n t m u t a t o r phenotype o f mutD51. N o such suppression was observed when the two mutations were present in different plasmids in a cell, thereby indicating that the effect o f dnaQ49 is cis specific. Thus, it is likely that the mutD5 m u t a t i o n alters the 3'--*5' exonuclease activity of the e subunit o f polymerase I I I holoenzyme and the dnaQ49 m u t a t i o n decreases interaction o f the e subunit with other subunits, p r o b a b l y the a subunit. This o f course does n o t exclude the possibility that the dnaQ49 m u t a t i o n also effects exonuclease activity. The model is illustrated in F i g . 5. A n o t h e r difference between mutD5 and dnaQ49 mutants is in their growth characteristics at high temperatures. The dnaQ49 m u t a n t was uanble to produce colonies at 44.5 ~ C, on NaCl-free L - b r o t h plates and showed a decreased rate o f D N A synthesis under restrictive conditions (Horiuchi

13 et al. 1978). These phenotypes were not observed with the mutD5 mutant. M a k i and K o r n b e r g (1985) isolated the subunit o f the D N A polymerase I I I holoenzyme which possesses D N A polymerase activity but not 3 ' ~ 5 ' exonuclease activity. Since the free ct subunit showed a higher thermolability than a catalytic core assembly, consisting o f the e, e and 0 subunits, the e subunit a n d / o r the 0 subunit seem to stabilize and stimulate the polymerase activity o f the ct subunit. Thus, the temperature-sensitive growth and D N A replication observed in the dnaQ49 m u t a n t is p r o b a b ly due to the p o o r association o f the altered e subunit with the e subunit. Since an a m b e r m u t a t i o n and Tn3 insertion mutations in the dnaQ gene also cause temperature-sensitive phenotypes (unpublished d a t a ; M a k i et al. 1983), the binding of the e subunit to the e subunit seems to be nonessential for D N A synthesis at low temperature but is required for the stabilization o f the ct subunit function at a high temperature. The temperature-dependent m u t a t o r activity o f the dnaQ49 m u t a n t m a y also be explained by dissociation o f the e subunit from the holoenzyme. The different phenotypes shown by the two alleles o f the dnaQ gene lead to the p r o p o s a l that the e subunit has two functional domains, the 3 ' ~ 5' exonuclease d o m a i n and the e subunit-binding domain. It is necessary to analyse the three-dimensional structure of the e subunit protein and correlate the amino acid substitutions with the functional domains.

Acknowledgements. We thank M. Ohara for comments on the manuscript. This work was supported in part by grants from the Ministry of Education, Science and Culture of Japan.

References Barnes WM, Bevan M, Son PH (1983) Kilo-sequencing: Creation of an ordered nest of asymmetric deletions across a large target sequence carried on phage M 13. Methods Enzym ol 101 : 98-122 Chang ACY, Cohen SN (1978) Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the P15A cryptic miniplasmid. J Bacteriol 134:1141-1156 Cox EC (1976) Bacterial mutator genes and the control of spontaneous mutation. Annu Rev Genet 10:135-156 Cox EC, Homer DL (1982) Dominant mutators in Escherichia coli. Genetics 100:7-18 Degnen GE, Cox EC (1974) Conditional mutator gene in

Escherichia coli: Isolation, mapping, and effector studies. J Bacteriol 117:477-487 Echols H, Lu C, Burgers PMJ (1983) Mutator strains of Escherichia coli, mutD and dnaQ, with defective exonucleolytic editing by DNA polymerase III holoenzyme. Proc Natl Acad Sci USA 80:2189 2192 Hanahan D (1983) Studies on transformation of Escherichia coli with plasmids. J Mol Biol 166:557-580 Horiuchi T, Maki H, Maruyama M, Sekiguchi M (1981) Identification of the dnaQ gene product and location of the structural gene for RNase H of Eseherichia eoli by cloning of the genes. Proc Natl Acad Sci USA 78 : 3770-3774 Horiuchi T, Maki H, Sekiguchi M (1978) A new conditional lethal mutator (dnaQ49) in Escherichia coli K12. Mol Gen Genet 163:277-283 Maki H, Horiuchi T, Sekiguchi M (1983) Structure and expression of the dnaQ mutator and the RNase H genes of Eseheriehia coli: Overlap of the promoter regions. Proc Natl Acad Sci USA 80:7137-7141 Maki H, Kornberg A (1985) The polymerase subunit of DNA polymerase III of Eseherichia coli. J Biol Chem 260: 12987-12992 Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York ManiS JJ, Kline BC (1977) Restriction endonuclease mapping and mutagenesis of the F sex factor replication region. Mol Gen Genet 152:175-182 Maruyama M, Horiuchi T, Maki H, Sekiguchi M (:[983) A dominant (mutD5) and a recessive (dnaQ49) mutator of Escherichia coli. J Mol Biol 167:757-771 Scheuermann RH, Echols H (1984) A separate editing exonuclease for DNA replication: The e subunit of Escherichia coli DNA polymerase III holoenzyme. Proc Natl Acad Sci USA 81: 7747-7751 Sekiguchi M, Horiuchi T, Maki H, Maruyama M, Oeda K (1982) Cloning of mutator genes and identification of their products. In: Miwa M, Nishimura S, Rich A, Soll DG, Sugimura T (eds) Primary and tertiary structure of nucleic acids and cancer research. Japan Scientific Societies Press, Tokyo, pp 181-188 Vieira J, Messing J (1982) The pUC plasmids, an M! 3mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19:259-268 C o m m u n i c a t e d by M. T a k a n a m i Received May 7, 1986