Genetic assay of misincorporation

Mutation Research, 251 (1991) 201-216

201

© 1991 Elsevier Science Publishers B.V. All rights reserved 0027-5107/91/$03.50

MUT 05024

G e n e t i c assay o f m i s i n c o r p o r a t i o n Rogelio Maldonado-Rodriguez *, Paul H. Driggers ** and Kenneth L. Beattie ** * Verna and Marts McLean Department of Biochemistry, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030 (U.S.A.) (Received 26 July 1990) (Revision received 23 April 1991) (Accepted 15 May 1991)

Keywords: DNA polymerase I, DNA synthesis; Lacl gene; Misincorporation; (Escherichia coli)

Summary A system to characterize mutations arising from in vitro nucleotide misincorporation, which avoids the effects of in vivo mismatch repair on recovery of mutants, was constructed and evaluated. The lacI gene of Escherichia coli was inserted into phage M13 and the M13-1acI recombinant was introduced into a strain of E. coli lacking a resident lacI gene. In this system the function of the M13-bearing lacI gene can be detected by plaque color. Mutants in the 5'-region of the lacI gene (encoding operator-binding domain) are seen as blue plaques when the host strain is grown in the presence of chromogenic substrate, X-gal, in the absence of inducer. The use of uracil-containing single stranded DNA from M13-lacI as template for DNA synthesis avoids the contribution of mismatch repair (in transfection recipients) on the recovery of mutants. To demonstrate the usefulness of the M13-1acI system we produced nucleotide misincorporations by in vitro DNA synthesis in the N-terminal region of the lacI template in the presence of only 3 deoxynucleoside triphosphates (dNTPs). Such mutagenic reactions were conducted in the absence of dATP with 4 different primers and in the absence of dGTP with 2 primers. The type of mutants produced by these reactions were identified through sequencing of DNA from progeny phage after screening for i- (blue plaque) phenotype. Mutations recovered in this system consisted of single and multiple base substitutions in the region of the template near the 3'-terminus of the primer. Nearly all of the mutants induced by '-A' conditions were T--, C base substitutions, and those induced by '-G' conditions were C -~ T transitions. In general, the results were consistent with the spectrum of spontaneous mutants produced in strains deficient in mismatch repair, although some differences were noted. Several new base substitutions within the lacI gene (producing i- phenotype and unobserved by others) were isolated by the procedures described in this paper.

Correspondence: Dr. K.L. Beattie, Center for Biotechnology, Baylor College of Medicine, 4000 Research Forest Drive, The Woodlands, TX 77381 (U.S.A.). * Present address: Depto. de Bioquimica, Escuela Nacional de Ciencias Biologicas, Instituto Politecnico Nacional, Prol. de Carpio y Plan de Ayala, Mexico, D.F. 11340 (Mexico), Tel. (905) 396-3716.

** Present address: Laboratory of Developmental Molecular Immunity, National Institutes of Health, NICHD, Bldg. 6, Rm 2A01, Bethesda, MD 20892 (U.S.A.). *** Recipient of Research Career Development Award CA00891 from the National Cancer Institute.

202 Misincorporation during DNA synthesis has been examined by a variety of biochemical and genetic techniques, including measurement of (i) nucleotide misincorporation during in vitro DNA synthesis on synthetic polynucleotide templates (Gillin and Nossal, 1976); (ii) reversion mutations resulting from in vitro polymerizerion on OX174am3 template DNA (Weymouth and Loeb, 1978); (iii) mismatch editing efficiency on the same QX174 DNA (Sinha, 1987); induction of M13 lacZ mutants during in vitro gap filling (Kunkel, 1985a,b; Kunkel and Alexander, 1986; Roberts and Kunkel, 1986); (v) analysis of raisincorporation resulting from in vitro polymerization in the presence of only 3 of the 4 dNTPs (Hillebrand et al., 1984; Lai and Beattie, 1988); and (vi) analysis of numerous spontaneous mutants in E. coli strains deficient in certain repair functions (Schaaper et al., 1986; Fix et al., 1987; Schaaper and Dunn, 1987). The results of these studies suggest that the frequency and kind of misincorporation during DNA synthesis is affected by both the type of DNA polymerase and the DNA sequence, and that mismatch repair decreases the net error frequency in vivo. We have previously utilized biochemical techniques to characterize the frequency (Hillebrand et al., 1984) and type (Lai and Beattie, 1988) of misincorporation that occurs during in vitro primer elongation in the presence of only 3 of the 4 dNTPs. In the work reported here we developed a new genetic assay that is useful for investigating in vitro misincorporation, and used it to characterize mutations arising from in vitro polymerization catalyzed by E. coli DNA polymerase I Klenow enzyme in the absence of dATP. This new assay permitted us to simultaneously determine the extent of misincorporation (deduced by a gel electrophoretic assay) and the nature of the mutants that arose from that misincorporation (deduced by a genetic assay and DNA sequencing). Specifically, we used this system to characterize the type and frequency of mutations arising from misincorporation at different thymine and cytosine residues within the lacl gene of E. coli. Also, with the help of a Xhol restriction site created at a region of very low frequency of spontaneous i- mutation (Miller, 1980), we were able to examine base-substitution mutation at a

site where even nonconservative amino acid replacements produced no change in phenotype. Materials and methods

Reagents X-Gal (5-bromo-4-chloro-3-indolyl-/3-Dgalactoside), IPTG (isopropyl-/3-D-thiogalactoside), T4 polynucleotide kinase, and the restriction enzymes XhoI and Sinai were purchased from Bethesda Research Laboratories. T4 DNA ligase was obtained from New England Biolabs. [T-32p]ATP was from ICN. The four dNTPs (2'deoxynucleoside 5'-triphosphates) and ddNTPs (2',3'-dideoxynucleoside 5'-triphosphates) were obtained from Pharmacia. dNTPs were purified by two rounds of HPLC by a procedure previously described (Revich et al., 1984). DNA polymerase I large fragment was purified in this laboratory from a recombinant DNA overproducer kindly provided by Dr. Catherine Joyce. Oligodeoxynucleotides were synthesized by the solid phase phosphoramidite method (McBride and Caruthers, 1983) using a Cruachem PS200 semiautomated DNA synthesizer. Bacterial strains E. coli strain CJ236 dut, ung, thi, relA1 and plasmid pCJ105, F'(cm r) were kindly supplied by Dr. C. Joyce. E. coli strain JC10240 thr, recA56, srl, relA, ilu, thi, strA, spoT, Hfr(P045 of KL16) was kindly supplied by Dr. B. Bachmann. E. coli JM83 ara, strA, thi, lac-pro, 080dlacZdM15 Fand JM103 dlac-pro, supE, thi, strA, sbcl5, endA, hspR4, F'(traD36, proAB, laclqZdM15) were kindly provided by Dr. J. Messing. Bacteriophage, plasmids and template DNA Wild-type phage M13 was kindly provided by Dr. J. Messing. Plasmid pMC1 was kindly provided by Dr. M. Calos. Single-stranded template DNA for use in site-directed mutagenesis and for the genetic assay of misincorporation was obtained as the uracil-containing form (U-SS-DNA) by growth of the appropriate M13 strain in E. coli CJ236 in the presence of uridine (Kunkel et al., 1987), as follows. One phage plaque was resuspended in 3 ml of YT medium and heated at 65 °C for 10 min. The tube was then inoculated

203 with 20 /zl of an overnight culture of CJ236 grown in YT medium containing 3 8 / z g / m l chloramphenicol. Simultaneously, 50 ml of YT medium was inoculated with 200 /zl from the same overnight culture. After 3 h at 37 o C, both cultures were added to 1 litre of fresh YT-chloramphenicol medium containing 320 ng/ml uridine and 10 mM MgSO4 and incubated at 37 °C for an additional 7 h, then centrifuged at 6000 rpm for 15 min at 5 °C in a Sorvall GS-3 rotor. The phage in the supernatant were titered both in CJ236 and in JM103 host cells to confirm the incorporation of uracil. Single-stranded DNA was purified from this supernatant as described by Zinder (1982). The DNA was considered as uracil-containing only when the transfecting activity and phage titer were at least 1000-fold lower in JM103 than in CJ236. Template DNAs for sequencing of mutant phages were obtained by growing minicultures in E. coli JM103, as follows. Individual phage plaques were placed into tubes with 3 ml of YT medium, heated 10 min at 65 ° C, inoculated with 20 /zl of overnight JM103 cells grown in the same medium, and incubated 3 h at 37 o C. Then 9/zl of this culture was added to 150 /xl of log phase JM103 cells in 3 ml of fresh YT medium and incubated an additional 7 h. The supernatant from each miniculture was used to isolate single-stranded M13 DNA as described by Crouse et al. (1983). Oligonucleotide-directed mutagenesis and DNA sequencing Oligonucleotide-directed mutagenesis was carried out by the method of Zoller and Smith (1982), along with the uracil-template selection method of Kunkel et al. (1987). DNA sequencing was carried out using [5'-32p]primers, by the 'dideoxy' chain termination method of Sanger et al. (1977). Genetic and electrophoretic assay of misincorporation induced by 'minus A ' reactions Oligonucleotide primers (20 pmoles) were phosphorylated as recommended by the T4 polynucleotide kinase supplier with 0.3 mM ATP or with [y-32p]ATP (1 /zl, 2.5 mCi/0.04 ml). A mixture of non-radioactive (20 pmoles) and labeled (20 pmoles) phosphorylated primer was annealed with 2.5 pmoles of uracil-containing sin-

gle-stranded M13 DNA for 2 h at 55 o C in 0.5 M NaCI followed by slowly cooling down at room temperature. In the case of primer D the procedure was carried out with both commercially supplied ATP and twice-HPLC-purified ATP. Primer-template D was then purified by gel filtration through Sepharose 2B. Misincorporation in '-A' reactions was carried out for 30 min at room temperature in 30-~zl reaction mixtures containing 0.2 pmoles primer-template, 40 mM Tris-HCl (pH 7.4), 5 mM MgCI/, 0.3 mM dGTP, 0.3 mM dTTP, 0.3 mM twice-HPLC-purified dCTP and 6 units of E. coli DNA polymerase I (Klenow fragment). Extension and ligation of the DNA strand containing misincorporated base(s) was carried out by incubation overnight at 15 °C with addition of 0.3 mM ATP, 0.3 mM dATP and 200 units T4 DNA ligase. A control reaction was carried out for 12 h at 15°C with all the ingredients required for complete DNA synthesis and ligation (4 dNTPs, ATP, Mg 2+, buffer, ligase, Klenow enzyme). The extent of misincorporation in place of dATP was assessed by electrophoresis in 8% polyacrylamide denaturing gel, followed by autoradiography. Reaction aliquots (1 #1)were also used to transfect JM103 competent cells. In the misincorporation reaction conducted with X primer, part of the extended DNA was digested with Xhol enzyme before the transfection, as follows. The mixture was extracted twice with phenol, three times with ether, precipitated and washed with ethanol, and redissolved in 10/zl 10 mM Tris-HC1 (pH 7.5), 0.1 mM EDTA. A 5-~1 aliquot of this mixture was digested with Xhol restriction enzyme as suggested by the supplier. To discriminate between i- and i + mutants, phage plaques isolated in JM103 were picked with tooth picks and inoculated onto separate areas of 3 ml soft agar containing 24 /zl 2% X-Gal and 400/zl log phase PD8 cells solidified over 20 ml YT agar medium. In the case of the G primer, transfection of DNA and screening of mutants were simultaneously performed, as follows. Transfected JM103 cells were mixed with 3 ml YT soft agar, 26/zl of 2% X-Gal and 1 ml log phase PD8 cells. Representative clear and blue phage plaques were grown in 3-ml minicultures of JM103 to prepare DNA for sequencing, as described previously.

204

Reverse sequencing assay of misincorporation In order to directly determine the type of misincorporation occurring in the '-A' reaction, the in vitro synthesized nascent DNA strand was sequenced, as follows. A mixture of 'cold-phosphorylated' (18 pmoles) and 32p-phosphorylated (2 pmoles) primer C or primer X (two different Expts.) was annealed with 20 pmoles of Ml3laclSaXc U-SS-DNA. A 2-pmole quantity of primer was elongated under the '-A' conditions for 1 h at 37 °C, and then subjected to extens i o n / hydrolysis by incubation at 37 ° C for 60 min after addition of dATP and SmaI restriction enzyme. The 'complete' reaction was also incubated 1 h at 37 ° C in presence of all four dNTPs and SmaI. Nascent strands (142 nucleotides in the case of primer C and 162 nucleotides with primer X), arising from SmaI cleavage of 'complete' and 'minus A' reaction products, were purified from a denaturing 5% polyacrylamide gel. These fragments were sequenced using 32p-labeled reverse sequencing primer by the dideoxy method (Sanger et al., 1977). The detailed conditions for this protocol were described previously (Lai and Beattie, 1988).

Results

The M13-1acl genetic system The M13 recombinant bearing the lacl gene of E. coli, M131acla, was constructed as illustrated in Fig. 1. Wild-type M13 phage RF I DNA was cleaved at position 5825 with restriction enzyme AvaI. Recessed ends were filled in using DNA polymerase I (Klenow fragment) as described by Maniatis et al. (1982). Plasmid pMC1 originally described by Calos (1978) was partially digested with restriction endonuclease HinclI, then electrophoresed in a 0.7% agarose gel. The restriction fragment af 1724-bp length was extracted from the gel and ligated to the linearized M13 DNA as described by Crouse et al. (1983). DNA from the ligation reaction was used to transfect strain PD8. Recombinant phages were isolated on the basis of their ability to produce blue plaques in the presence of X-Gal + IPTG. A recombinant containing one copy of the antisense orientation of the lacl gene within the phage ( + ) strand was designated M131acla. As a consequence of the antisense orientation of the lacl template, in vitro primer elongation will occur in

Ava I

P

,

I

,,, I/v,,, Ligate v

H i n d II (partial) I

~

I

Fig. 1. Construction of M131acla. Of the two possible orientations of the lacl gene, only the 'antisense' orientation (M131acla)was isolated. See text for detailed description.

205

the 3 ' ~ 5 ' direction along the lacI template (corresponding to the C ~ N direction of amino acid sequence). Three M131acla derivatives containing new restriction sites were constructed by oligonucleotide-directed mutagenesis, following the protocol described earlier. The nucleotide sequences between residues 90-95 (site a), 129-134 (site b) and 232-237 (site c) within the lacI gene (Farabaugh, 1978) were changed at positions 91 (C --* T), 94 (C ~ A), 130 (G ~ T), 133 (G ~ A) and 235 ( C ~ G ) to create Xhol sites and at position 94 (C ~ G) to produce the Smal site. The creation of strains M13laclSaXb (with Smal and Xhol recognition sequences at sites a and b, respectively), M131aclXaXb (with Xhol recognition sequence at sites a and b) and M131aclSaXc

(with Smal and Xhol recognition sequences at sites a and c, respectively) was confirmed by sequencing and also by digestion with the respective enzymes. The derivative containing two Xhol sites was also confirmed by creation of the deletion mutant upon restriction and re-ligation. The nucleotide sequence of M13lacla is displayed in Fig. 2, along with locations of the primers and engineered restriction sites employed in this study. The E. coli host for M131acla, PD8 ara, strA, thi, nalA, recA56, d(lac-pro), 080, DlacZdM15, F'(/ysA fuc), was constructed as follows. Spontaneous nalA mutants were isolated from strain JC10240 by plating cells on L-agar plates containing 10/zg/ml nalidixic acid as described by Miller (1972). These mutants were replated and crossed

Xhol-a site

I ~ 1 CT CGA G

tt

GTG AAA CCA GTA ACG "IrA TAC GAT GTC GCA GAG TAT GCC GGT GTC TCT TAT CAG ACC G'Fr TCC CGC GTG GTG I I I I I I I | I 30 40 50 60 70 80 90 t 100

I PrimerG

CC CGG GI I

Smal-a site

Xhol-b site I

i

CT CGA G

tt

AAC CAG GCC AGC CAC f i f e TCT GCG AAA ACG CGG GAA AAA GTG GAA GCG GCG ATG GCG GAG CTG AAT TAC A'l-r I I I I I I I 110 120 130 140 150 160 170 I Primer D I Primer B

Xhol-c site I C TCG AGI

t

CCC AAC CGC GTG GCA CAA CAA CTG GCG GGC AAA CAG TCG TTG CTG ATT GGC G'l-r GCC ACC TCC AGT CTG GCC I I I I I I I 180 190 200 210 220 230 240 I Primer C I Primer X

Fig. 2. Positions of engineered restriction sites and primers within the lacl gene of M131acla. The nucleotide sequence of the viral strand is shown, numbered according to position within the code sequence. The locations of the 3'-end of primers are shown below the template sequence. The base substitutions that were created by oligonucleotide-directed mutagenesis to introduce Xhol and Smal sites into the gene are indicated by vertical arrows.

206 32pro

with F JM83. Cells bearing strA and nalA mutations were selected and screened for the lac deletion of JM83. NalA strA recombinants were spot tested for ultraviolet light sensitivity as described by Miller (1972). One colony was picked and designated PD2. Finally, the F' factor of strain KL730 was transferred to PD2 in a mating cross• Cells containing the F' factor were selected on the basis of sensitivity to infection by phage M13. One colony, designated PD8, was chosen for further characterization. Ultraviolet light survival curves determined for PD8 and the parental strains JM83 and JC10240 (Miller, 1972) showed PD8 to contain the recA56 allele derived from parental strain JC10240. Use of the M13-lacI system to characterize in vitro misincorporation Fig. 3 outlines our initial use of the M13-lacI system to characterize in vitro misincorporation. To illustrate the usefulness of the system we carried out primer elongation within the N-terminal region of the E. coli lacI gene under conditions known to cause misincorporation. A [5'32P]oligonucleotide primer was annealed to M131acla uracil-containing template DNA to form the primer-template, 1. The primer was then elongated by addition of E. coli DNA polymerase I (Klenow fragment) either in the presence of all 4 dNTPs to form the duplex molecule, 2, or in the absence of dATP to yield a mixture of extended primers containing one or a few misincorporated bases (indicated by *) opposite template T residues near the end of the original primer, as shown in 4. An aliquot of the primertemplate was also used to prepare a set of ddAterminated primers, 3, to provide markers in the electrophoretic assay of misincorporation. After a period of primer elongation in the presence of 3 dNTPs, the missing dNTP was added, along with ATP and T4 D N A ligase, to prepare closed circular duplex DNA. The DNA products were then used in a transfection assay, 5, to distinguish between wild-type progeny (forming colorless plaques) versus i - mutant progeny (forming blue plaques in the absence of inducer) arising from these D N A products. As depicted at the bottom of Fig. 3, D N A products can also be analyzed by a gel electrophoretic assay (Revich et al., 1984) to

(-- ....

Klen+gment II

4dNTP

"-A" /

*

1

.

/

*

I

+

J

+ ATP T4 DNA ligase

ddATP chain termination

(~

+dATP

J + ATP T4 D N A I I i r g g a s e

Fig. 3. Genetic and electrophoretic assay of misincorporation. A [5'-32p]primer is annealed to uracil-containing singlestranded M13 DNA, to form the primer-template (1). Primer elongation is carried out with purified D N A polymerase (for example, E. coli D N A polymerase I Klenow fragment) in the presence of all 4 dNTPs to form circular duplex D N A (2), or in absence of d A T P to form a population of nascent primer strands containing one or more misincorporated bases, *, as depicted in (4). The nascent primers are then extended by addition of the missing nucleotide in the presence of D N A ligase to form closed circular D N A (5, in the case of the ' m i n u s ' reaction). A portion of the original primer-template is also used for the corresponding +dideoxy' sequencing reaction (ddATP reaction), forming in this case a nested set of elongated primers terminated at each successive A in the sequence, (3). Samples of all D N A products are analyzed by denaturing polyacrylamide gel electrophoresis to give the pattern of autoradiographic bands shown at the bottom. Other D N A samples are used for transfection of E. coli(dut + ung + ) and progeny phage are screened for I - mutants in medium containing X-Gal but lacking IPTG. D N A from blue (mutant) and clear (wild-type) phage plaques is sequenced to identify specific genotypic changes.

further characterize the extent of misincorporation. More detailed conditions of the genetic assay are given in Materials and methods. This system permits us to conveniently and • rapidly characterize mutations that arise at many different sites within the lacI gene during in vitro misincorporation reactions. Of course, not all po-

207 TABLE 1 RECOVERY AND IDENTIFICATION OF MUTATIONS INDUCED BY MISINCORPORATION Primer

Reaction conditions

Transfecting activity (pfu/ng DNA)

Mutation frequency

Progeny sequenced blue clear

Mutants confirmed blue

clear

G G

Complete '-A'

6.5 7.1

0.005 0.017

10 97

0 20

4 66

0 0

D D

Complete '-A'

6.3 7.2

0.05 0.18

20 163

60 60

0 156

0 0

X

Complete (Xhol) '-A' (Xhol)

0.06

0.05

5

8

0

0

1.1

0.06

9

191

0

64

10.0 7.6

0.01 0.01

10 27

10 33

0 2

0 0

X X X

Complete '-A'

As indicated in column 1, three different primers were used for the in vitro DNA synthesis reactions (see Fig. 2 for their positions on the template). These primers were elongated in vitro under 'complete' and 'minus A' conditions (indicated as C and '-A', respectively in column 2). The '-A' reaction products were further elongated in the presence of dATP before being used in the transfection assay. In the case of primer-X a portion of each reaction mixture was treated with Xhol restriction enzyme prior to transfection. DNA products were transfected in JM103 competent cells to isolate the M13 progeny, and the respective transfection activities are indicated in column 3. Individual phage plaques were placed with tooth picks onto lawns of PD8 in solid medium containing X-Gal and lacking IPTG to identify blue i - mutants, and the respective mutation frequencies are indicated in column 4. Some blue and clear phage plaques (numbers indicated in column 5) were analyzed by dideoxy sequencing to identify the sequence changes induced by each reaction. The last column gives the number of sequence-confirmed mutants, i.e., those containing one or more sequence changes in the region of in vitro primer elongation. The M131aclSaXb template was used for primers D and G, while the M131aclSaXc template was used for primer-X. The data presented for primers D, G and X correspond to the Expts. of Fig. 6.

sitions within this gene are targets for phenotypicaUy detectable mutation. In the application of the system reported here we characterized base substitution mutations at thymine residues in the lacl template, arising from misincorporation during primer elongation in the absence of dATP. Base substitutions at cytosine residues in the template, arising from elongation in the absence of dGTP, were also studied. Use of restriction enzyme selection to enrich for non-phenotypic mutants Most of the data collected were for phenotypic (i-) mutants, forming blue plaques in the absence of inducer. However, we developed a restriction enzyme selection protocol to characterize non-phenotypic (silent) mutations arising at template T residues within the recognition sequence of Xhol (CTCGAG). At the Xho l-c site (with thymine at position 233 as shown in Fig. 2) we were unable to recover i- mutants (blue

plaques) directly from the products of '-A' reactions. However, we did recover base-substitution mutations at this site using Xhol selection. To demonstrate that misincorporation opposite the template T within the Xhol recognition sequence (CTCGAG) inhibits cleavage by this enzyme, the following experiment was performed. DNA synthesis was carried out in a 'complete' reaction and in the absence of dATP ('-A' reaction) using as template M131acla-XaXb in which Xhol sites a and b are separated by 39 base pairs. [5 '- 32p]Primer B was annealed to this template, such that elongation begins at template position G134 within the Xhol-b site and proceeds toward the Xhol-a site (See Fig. 2). Thus, the first misincorporation in the '-A' reaction would introduce a mismatched nucleotide opposite T133 within the Xhol-b site. On the other hand, the normal nucleotide (dAMP) would be incorporated in the 'complete' reaction. As seen in fig. 4, after being elongated in the presence of all 4 dNTPs, the

I !

~

ia~~

!

i~i~

O

0 0

N!:~~;:~~! I~I!~¸¸ ~!

i|!'~

~II

iii|i ~ ~ I,III,!¸

209

-A

C I

-A

I

A G C T

I

I

A G

C T

C II

'

i

AG C T A G C T

G C

i

T

w

•

iii I T

I ~

C

! ~' :~ ~

~

~ ~ _j ~

A A C G

G G C G G T C A

A C

-~,T ~T G C A C A A A C G G G C G G T C A A C

Fig. 5. Reverse sequence of 'complete' and 'extended minus A' reactions. DNA from 'complete' and 'extended -A' reactions were hydrolyzed with Sma I, then the nascent DNA strands were purified from a 5% polyacrylamide denaturing gel and sequenced by the Sanger dideoxy method using a reverse primer. '-A' and 'C' represent the sequences of nascent strand from the 'minus A' and 'complete' reactions, respectively. Sequencing lanes are identified by A, G, C and T at the top. To the right of each autoradiogram is displayed the unaltered sequence of elongated primer. The dideoxy bands indicating base substitution are marked by > (ddT band from the 'complete' reaction, corresponding to normal incorporation of A, and ddC band from the '-A' reaction, corresponding to misincorporation of C). Data for the experiment with primer C is shown on the left, and that for primer X is shown on the right.

Fig. 4. Effect of misincorporation within the Xhol recognition sequence on the sensitivity to digestion by Xhol. Lane 1 shows the pattern of 'ddA' sequencing bands in the region of the template containing the Xhol-b and -a sites in M131aclXaXb. Along the left is displayed the template sequence, with lines connecting 'ddA' bands with corresponding T residues in the sequence. The location of both Xhol sites and of the 3'-OH end of primer B are also indicated (also see Fig. 5). Lanes 2-4 display the products of ddG, ddC and ddT sequencing reactions, respectively. Lane 5 displays the products of the 'complete' reaction and lane 6 represents the '-A' reaction. Lane 7 represents the Xhol-digested 'complete' reaction and lane 8 represents the Xhol-treated '-A' reaction. Lane 9 displays a mixture of DNA from 'complete' and '-A' reactions, digested with Xhol. Lane 10 represents a '-A' reaction to which dATP was subsequently added to give extensive elongation (termed the extended '-A' reaction). Lane 11 displays the DNA from the extended '-A' reaction after digestion with Xhol. Lane 12 represents a mixture of DNA from 'complete' and 'extended '-A' reactions, digested with Xhol.

210 G cC CC

c c c

c c

c

CC

c c

~ c

C C C C C C C C C C C C [Primer G C C C -GTG--GTA--GTTATA--ATGTC--GTATG--GTGTCTCTTATC--GTTTC

I

I

I

I

I

I

I

30

3g

45

52

62 C C C C

70

76

C C C C

C C C C C x

C c c

cc CC CC CC CC CC CC CC CC C ~ cc

c cc ccc cccc ccccc cccccc ccccccc ccccccc ccccccc CCCCCCC CCCCCCC CCCCCC CCCCC CCCC CCCC CCC CCC CC CC cc

c c C c lrPrimerD C C C C C ~ATGTC--GTATG--GTGTCTCTTATC--GTTTC--GTGGTG--GTTTCTG--GTG-

I

I

I

I

I

I

I

I

52

62

70

76

87

g6 C C C

117 C C C C

141

C C C C C C C C C C C C C C

C C C C C C C C C C C C

C--C CIC CIC C--C C C C

C C C

Fig. 6. Mutations arising from primer elongation within the lacl gene, catalyzed by E. coli D N A polymerase 1 (Klenow enzyme) in the absence of dATP. Misincorporation reactions ('-A') were performed at 4 different regions in the amino terminal region of the lacl gene, using primers G, D, C and X. D N A from mutant plaques was analyzed by 'dideoxy' sequencing, employing the same primers used in the original '-A' reactions. For each primer the template sequence (5' --+ 3') and position of the 3'-OH terminus is indicated. Nucleotide positions within the lacI gene are numbered. Symbols identifying base substitutions are positioned above template residues at which mutations were determined by sequencing of mutants. Multiple base substitutions are indicated below each sequence by symbols connected by horizontal lines. Except for on T --* G transversion found at position 72 (denoted by 'G'), all mutations were T --+ C transitions. In the case of primer X, all single base substitutions at position T233, double transitions at T226 + T233 and at T220 + T233 were silent (clear plaques, selected by Xhol hydrolysis). In addition, 7 blue plaques with double T --+ C transition at positions 216 + 233, 1 blue plaque with double T --0 C transition at positive 225 + 233, and 2 single transitions at T225 (all recovered directly from the '-A' reaction conducted with primer X) produced the i - phenotype.

211

(B)

c

c C c c c c C

I Primer C -ATTA--ATTC--GTG--CTG--GTCGTTGCTGATTGI

I

I

I

166

171

183 C C C C

195

C C

I

I

209 c c c--c 'c

216 c c

c ~ c c c '"c c 'c c ~ c c c C II C C ~ C C C C C C C C ~ C C C C C C ~ C C 'C C ' "C C C C ~ C C C C C C C C C C C C C C

C C C c C C C C C C C C C C C C C C C C C C C C C C C C C C

CCC CCCC CCCCC CCCCC CCCCC CCCCC CCCCC CCCC CCC CC CC CC CC CC c rll-~T~IIffr P x

C c

-ATTA--ATTC--GTG--CTG--GTCGTTGCTGATTG--GTTG--CTCGAGI I I I I I 166

171

~83

195

209

216

I

I

225

233 C C C C C C C C C C C C C C

C C C

C C C C

Fig. 6 (continued)

extensively elongated primer (bands at top of lane 5) was susceptible to digestion by Xhol (producing bands at lower positions in lane 7, as expected from cleavage at Xhol sites a and b). However, the primer elongated in the '-A' reac-

tion was resistant to hydrolysis by Xhol, as shown by identical electrophoretic patterns in lane 8, compared with undigested product in lane 6. When the mixture of 'complete' and '-A' reaction products were incubated with Xhol, only the D N A

212

from the complete reaction was again hydrolyzed (lane 9). When primer elongated in the '-A' reaction was further extended by addition of dATP, it became sensitive to digestion at the second Xhol sequence (site a), as seen in lane 11, compared with undigested product in lane 10. The mixture of 'complete' and extended '-A' reactions appeared as two bands after Xhol treatment (lane 12). The upper band corresponds to DNA cleaved at the Xhol-a site after elongation of the primer in the '-A' reaction. The lower band corresponds to DNA cleaved at the Xhol-b site after elongation in the 'complete' reaction. Although this DNA would be sensitive to cleavage at both Xhol sites, only the cleavage at the Xhol-b site is visible by autoradiography through retention of the end label. In addition, we conducted direct sequencing of the nascent strand by the method of Lai and Beattie (1988) to confirm that misincorporation occurred at position T233 during the '-A' reaction with primer X. The protocol is detailed in Materials and methods. Representative data are shown in Fig. 5 (primer C on left and primer X on right). Only one difference was observed in the electrophoretic pattern of bands. The ddT band, corresponding to the first target T ('complete' reaction) was substituted by a ddC band, corresponding to misincorporated C ('-A' reaction), as indicated by > in the autoradiograms. Mutations recot,ered using the M13-lacI system Table 1 summarizes the transfecting activity, frequency of i - mutation (among progeny arising from in vitro DNA synthesis) and number of mutants scored (by plaque color and by sequencing) for each of the three primers. The transforming activities of DNA products of '-A' reactions were similar to those of the respective 'complete' reactions. The frequency of i - blue mutants arising from the '-A' reactions with primers G and D was 3 - 4 times greater than that arising from the 'complete' reactions, whereas with primer X the frequency of i - mutation was about the same for ' - A ' and 'complete' reactions. We attempted to detect base substitutions by 'dideoxy' sequencing of phage DNA in both blue and clear plaques, using the same primers as in the in vitro polymerization reactions. No sequence changes were

found in clear phage plaques, except with primer X, in which case base substitutions were found in about 1 / 3 of the plaques after selection by Xhol. In general, mutations were identified by sequencing of progeny phage arising from products of only the '-A' reactions, except with primer G, in which case a few mutants were recovered from products of the 'complete' reaction. Overall, the sequence changes were identified in 224 out of 287 (78%) of the i phage analyzed, although the proportion of sequence-identified mutations was much lower in the case of primer X, which produced little or no mutation above spontaneous background when employed in the '-A' reaction. Since very few spontaneous mutants (from 'complete' reactions) showed sequence changes in the region of the lacl gene examined (about 100 bases from each primer terminus), the vast majority of mutations identified by sequencing must have arisen from misincorporation in the 'minus' reactions. Fig. 6 displays the identity of mutations recovered after conducting '-A' reactions with each of four different primers, as determined by sequencing of phage DNA. The mutations displayed in Fig. 6 came from experiments represented in Table I, plus additional experiments conducted with primer C, not included in Table 1. Out of 435 base substitutions characterized, arising from '-A' reactions, over half (260 mutations) were single T ~ C transitions. About 1/4 of the mutations (116/435) occurred as non-tandem double T ---, C transitions. Only one transversion (T ~ G) was isolated. In the case of Primer X nearly all of the mutations with a base substitution in the first thymine (50 single- and 6 double-base substitutions) were identified from clear phage plaques obtained from Xhol-treated DNA from '-A' reactions. Additionally, two mutants with single T -~ C substitution at position 225, one mutant with double T ~ C at positions 225 + 233 and seven mutants with double T--, C transitions at positions 216 + 233 were isolated from blue phage plaques arising from the non-Xhol selected '-A" reaction. Fig. 7 displays the identity of mutations at cytosine residues in the template, resulting from elongation of primers G and C under 'minus G' conditions. The results are similar to the '-A' results in that only transitions were produced

213 TT TTT TTTT TTTT TTTT TTTT TTT TTT TT TT TT T T

Discussion

T

IPrimer G

-ACG--ACG--TCG--TCTTATCI I I I 42 49 T T T 'T T T T T T ~ T T T T T T T T T T ' T T T T ........T T T

55 T T

75

''

TT TT TTT TTT TTT TTT TT TT TT TT T T T T T

T TT TT TT TT T T T T T

Iprirner C

-ACCGCG--GCA--GCG--TCGTTGCTGI

I

I

I

179

186

198 T T T T T T T

210 T T T T

T

T T

T T T 'T

Fig. 7. Mutations arising from primer elongation within the lacl gene, catalyzed by E. coli D N A polymerase I (Klenow enzyme) in the absence of dGTP. Experiments and representation of data were the same as in Fig. 6, except that misincorporation reactions were conducted under 'minus G' conditions, using primers G and C.

(C ~ T). Of the 137 base substitutions identified by sequencing, over half (80) were single-base transitions and most of the remaining (51) occurred as triple C ~ T transitions.

We have developed a new genetic assay (Figs. 2 and 3) for analysis of misincorporation during in vitro DNA-synthesis reactions. The new system employs the well known assay of mutations within the E. coli lacI gene. We chose the lacI system because of the convenient phenotypic detection of forward mutations and because of the wealth of information that exists for spontaneous and induced mutagenesis in this gene. The lac! system was made directly accessible for in vitro DNA-synthesis reactions by construction of M13lacl recombinants. To demonstrate the utility of the M13-1acI system we employed the mutagenic 'minus' reaction conditions, first described by Shortle et al. (1982) and previously used in this laboratory to biochemically examine the relative frequency of misincorporation at different positions along a DNA template (Hillebrand et al., 1984; Revich et al., 1984) and to determine the influence of template sequence on the specificity of misincorporation (Lai and Beattie, 1988). Single-stranded M13-1acI templates were annealed with synthetic oligonucleotide primers at specific locations within the 5' (N-terminal) region of the lacI gene (encoding amino acids 1-70, the operator-binding domain). This portion of the lacI gene is ideal for mutational analysis because it contains the highest density of phenotypically detectable target sites within the gene (Miller and Schmeissner, 1979; Miller, 1984). The primers were then elongated in the presence of only 3 of the 4 dNTPs, to target misincorporation to specific regions of the template (near the 3'-end of each primer), a-Complementation of /3-galactosidase was used to identify i- mutants as blue plaques in E. coli PD8 (a recombinationless host lacking a resident lacI gene) grown in medium containing X-Gal but lacking IPTG. An increase in the frequency of blue plaques among progeny M13-lacl phage in this assay reflects the occurrence of mutations that inactivate the lac repressor (most commonly mutations within the amino terminal region of the protein). 'Dideoxy' sequencing of these mutants with the same primer used in the misincorporation reaction directly identifies the base changes that occur during in vitro DNA synthesis. Uracil-containing template

214 D N A was used to enrich for progeny phage arising from the new strand synthesized in vitro. An important feature of the use of uracil-containing templates is that products of D N A synthesis are not subject to mismatch repair in the transfection hosts (Kunkel et al., 1987; Geisselsoder et al., 1987; Maldonado-Rodriguez and Beattie, 1991), so that the results reflect misincorporation events that occur during in vitro D N A synthesis. Another useful feature of this assay system is the possibility to simultaneously perform the electrophoretic assay of misincorporation (Hi[lebrand et al., 1984) along with the genetic assay, using the D N A products from the same in vitro polymerase reaction. Sequence identification of 435 base substitutions in the lacl gene induced by '-A' misincorporation reactions conducted with E. coli D N A polymerase I (Klenow enzyme) indicated a vast predominance of T - - * C transitions. Similarly, identification of 137 base substitutions arising under ' - G ' polymerization conditions revealed C --* T transitions, exclusively. All of the mutations identified by sequencing occurred as single or non-tandem multiple-base substitutions in the region of the template near the 3'-terminus of the primer. Several of the i - mutations recovered have not been previously observed in the lacl gene (Gordon et al., 1988). These include single T ~ C transitions at positions T96, T99 and T225 (producing Val ~ Ala in the lac repressor) and T ~ C transition at position T213 (producing Leu Ser in the repressor). Also recovered were several translationally silent base substitutions, linked to phenotypically visible mutations. The loss of sensitivity to digestion by Xhol after misincorporation within the recognition sequence for this enzyme (Fig. 4) suggests that the existence of a single mismatch in the restriction site is enough to prevent hydrolysis by the restriction enzyme. This p h e n o m e n o n was used to select for non-phenotypic mutants at a site within the lacl gene where spontaneous i mutants lacking binding activity to D N A or to I P T G are rare (Miller, 1980). Using inhibition of XhoI by misincorporation, we recovered 50 silent mutations (producing clear phage plaques) from products of the '-A' reaction, in which T ~ C transition occurred at the template T233, the first

target residue from the 3'-end of primer X. Interestingly, this base substitution changes Ser69 to Pro69 and yet this nonconservative amino acid substitution produces no phenotypic effect. Notably, polymerization under 'minus' conditions did not yield any multiple tandem-base substitutions, presumably because of the high efficiency of excision of 3'-terminal multiple base mismatches by the 3'-exonuclease activity of the Klenow enzyme ( M a l d o n a d o - R o d r i g u e z and Beattie, 1991). In work presented elsewhere (Maldonado-Rodriguez et al., 1991) we showed that if small quantities of the 'missing' dNTP are added to the polymerase reaction mixture, the frequency of mutant recovery in the M31-lacl system is considerably increased and the region of template over which mutations are recovered is broadened. Close inspection of the data of Fig. 6 reveals a puzzling phenomenon, i.e., several mutations were recovered at template positions (e.g., T54 and T62) which were not linked to base substitutions that should have occurred via previous misincorporation (at T64, T70 in the case cited above). As explained elsewhere (Maldonado-Rodriguez et al., 1991) we believe that this phenomenon is due either to the presence of small quantities of contaminating dNTP or to the combined action of 3'-exonuclease proofreading (generating PP~) and pyrophosphorolytic degradation of the primer to generate some d A T P in a '-A' reaction. The example reported here represents only one potential application of the M13-1acl system for studies of misincorporation by D N A polymerases. We believe that M13-1acl is a useful model system for analysis of mispairing events under several other controlled situations. For example, in the assessment of the base mispairing potential of chemically modified nucleotides during in vitro D N A synthesis the Ml3-lacl genetic assay can serve as confirmatory evidence (to complement the electrophoretic assay of misincorporation) for ambiguous base pairing of dNTP analogs during in vitro primer elongation. Thus, if an analog such as O%alkyl-dGTP were found to stimulate the extent of primer elongation in a "-A' reaction, it would likewise be expected to increase the frequency of i ~- mutations in the M13-1acl system when added to the '-A' reaction,

215

giving rise to a predictable kind of base substitution mutation. This application of the system is only possible if the mutation frequency is signifantly stimulated when the analog is added to a 'minus' reaction. The M13-1acl system should also be useful for studying the influence of controllable reaction parameters on in vitro misincorporation events - - parameters such as the presence of error-inducing agents (e.g., metal ions), identity of the DNA polymerase, and the composition of protein components of in vitro DNA replication systems. For example, we are currently using the M13-1acl system to biochemically assess the base-pair recognition abilities of engineered forms of DNA polymerases. Finally, the M13-1acI system is adaptable for assessment of the mispairing potential of certain nucleotide analogs in vivo. In cases where analogs are readily incorporated during in vitro DNA synthesis [such as 1,N6-Etheno-dATP in place of dATP (Revich and Beattie, 1986) or 6-thio-dGTP in place of dGTP (Ling et al., 1991)] then the use of analog-substituted M13-1acI DNA in the transfection assay should give rise to mutations if the template-bearing analog mispairs during in vivo replication. Sequence analysis of such mutations would reveal the specificity of in vivo mispairing in such experiments. Again, this application is useful only if the frequency of analog-induced mutation is significantly greater than the spontaneous frequency. The M13-1acI system may also serve to investigate the influence of DNA-repair functions (e.g., those involved in the SOS-repair pathway) on specific in vivo mispairing events following in vitro incorporation of modified bases into the lacl gene. It should be kept in mind, however, that the replication system of bacteriophage M13 differs from that of the E. coli host chromosome, as recently pointed out by Yatagai and Glickman (1990), who utilized the lacI gene cloned into phage M13 to characterize spontaneous mutation in the M13 environment versus that in the F' episome environment. These authors found that in M13 base substitutions accounted for 80% of spontaneous events, compared with only 11% in F'. Despite such differences, we believe that the M13-lacI system is a useful model system for characterization of spe-

cific in vitro and in vivo mispairing events. Lively philosophical arguments will persist a long time regarding the relative validity of results obtained with extrachromosomal versus chromosomal replication systems, in vitro versus in vivo replication, and purified polymerases versus intact, isolated replication complexes.

Acknowledgments This research was supported by N.I.H. Grant GM30590 and by Grant Q-1006 from the Robert A. Welch Foundation.

References calos, M. (1978) DNA sequence for a low-level promoter of the lac repressor gene and an 'up' promoter mutation, Nature (London), 274, 762-765. Crouse, G.F., A. Frischauf and H. Lehrach (1983) An integrated and simplified approach to cloning into plasmids and single stranded phages, Meth. Enzymol., 101, 78-89. Farabaugh, P.J. (1978) Sequence of the lacI gene, Nature (London), 274, 765-769. Fix, D.F., P.A. Burns and B.W. Glickman (1987) Spontaneous mutations in PolA1 strain of Escherichia coli, Mol. Gen. Genet., 207, 267-272. Geisselsoder, J., F. Witney and P. Yuckenberg (1987) Efficient site-directed in vitro mutagenesis, BioTechniques, 5, 786-791. Gillin, F.D., and N.G. Nossal (1976) Control of mutation frequency by bacteriophage T4 DNA polymerase, II. Accuracy of nucleotide selection by the L88 mutator, CBI20 antimutator, and wild-type phage T4 DNA polymerases, J. Biol. Chem., 251, 5225-5232. Gordon, A.J.E., P.A. Burns, D.F. Fix, F. Yatagai, F.L. Allen, M.J. Horsfall, J.A. Halliday, J. Gray, C. Bernelot-Moens and B.W. Glickman (1988) Missense mutation in the lacl gene of Escherichia coli, Inferences on the structure of the repressor protein, J. Mol. Biol., 200, 239-251. Hillebrand, G.G., A.H. McCluskey, K.A. Abott, G.G. Revich and K.L. Beattie (1984) Misincorporation during DNA synthesis, analyzed by gel electrophoresis, Nucl. Acids Res., 12, 3155-3171. Kunkel, T.A. (1985a) The mutational specificity of DNA polymerase-fl during in vitro DNA synthesis, Production of frameshift, base substitution, and deletion mutants, J. Biol. Chem., 260, 5785-5796. Kunkel, T.A. (1985b) The mutational specificity of DNA polymerases-a and -y during in vitro DNA synthesis, J. Biol. Chem., 260, 12866-12874. Kunkel, T.A., and P.S. Alexander (1986) The base substitution fidelity of eucaryotic DNA polymerases, Mispairing frequencies, site preferences, insertion preferences and base substitutions by dislocation, J. Biol. Chem., 261, 160166.

216 Kunkel, T.A., J.D. Roberts and R.A. Zakour (1987) Rapid and efficient site-specific mutagenesis without pheuotypic selection, Meth. Enzymol., 154, 367-382. Lai, M.-D., and K.L. Beaffie (1988), Influence of DNA sequence on the nature of mispairing during DNA synthesis, Biochemistry, 27, 1722-1728. Ling, Y.-H., J.A. Nelson, D. Farquhar and K.L. Beattie (1991) Utilization of 2'-deoxy-6-thioguanosine 5'-triphosphate in DNA synthesis catalyzed by DNA polymerase I Klenow fragment of Escherichia coli, Nucleosides and Nucleotides, in press. Maldonado-Rodriguez, R., and K.L. Beattie (1991) In vitro mutagenesis in the lacl gene of Escherichia coli: Fate of 3'-terminal mispairs versus internal base mispairs in a transfection assay, Mutation Res., 247, 5-18. Maldonado-Rodriguez, R., M. Espinosa-Lara and K.L. Beattie (1991) Influence of neighboring base sequence on mutagenesis induced by in vitro misincorporation in the lacl gene of Escherichia coli, Mutation Res., 251,217-226. McBride, L.J., and M.H. Caruthers (1983) An investigation of several deoxynucleoside phosphoramidites useful for synthesizing deoxyoligonucleotides, Tetrahedron Lett., 24, 245-248. Maniatis, T., E.J. Fritsch and J.M. Sambrook (1982), Molecular Cloning, A Laboratory Manual, Cold Spring Harbor, New York, p. 113. Miller, J.H. (1972) Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Miller, J.H. (1980) The lac 1 gene: its role in the lac I operon control and its use as a genetic system, in: R. Miller and W.S. Reznikoff (Eds.), The Operon, Cold Spring Harbor, NY, pp. 30-87. Miller, J.H. (1984) Genetic studies of the lac repressor, XII. Amino acid replacement in the DNA binding domain of the Escherichia coli lac repressor, J. Mol. Biol., 180, 205212. Miller, J.H., and U. Schmeissner (1979) Genetic studies of the lac repressor, X. Analysis of missense mutations in the lacl gene, J. Mol. Biol., 131, 223-248. Revich, G.G., and K.L. Beattie (1986) Utilization of 1,N 6etheno-2'-deoxyadenosine 5'-triphosphate during DNA synthesis on natural templates, catalyzed by DNA polymerase I of Escherichia coli, Carcinogenesis, 7, 1569-1576.

Revich, G.G., G.G. Hillebrand and K.L. Beattie (1984) Highperformance liquid chromatographic purification of deoxynucleoside 5'-triphosphates and their use in a sensitive electrophoretic assay of misincorporation during DNA synthesis, J. Chromatogr., 317, 283-300. Roberts, J.D., and T.A. Kunkel (1986) Mutational specificity of animal cell DNA polymerases, Environ. Mutagen, 8, 769-789. Sanger, F., S. Nicklen and R.A. Coulson (1977) DNA sequencing with chain-terminating inhibitors. Proc. Nat[. Acad. Sci. (U.S,A.), 74, 5463-5467. Schaaper, R.M., and R.L. Dunn (1987) Spectra of spontaneous mutations in Escherichia coli strains defective in mismatch correction: The nature of in vivo DNA replication errors, Proc. Natl. Acad. Sci. (U.S.A.), 17, 622[)-6224. Sbaaper, R.M., B.N. Danforth and B.W. Glickman (1986) Mechanism of spontaneous mutagenesis: An analysis of the spectrum of spontaneous mutation in the Escherichia coli lac I gene, J. Mol. Biol., 189, 273-284. Shortle, D., P. Grisafi, S.J. Benkovic and D. Botstein (1982) Gap repair mutagenesis: Efficient site-directed induction of transition, transversion, and frameshift mutation in vitro, Proc. Natl. Acad. Sci. (U.S.A.), 79, 1588-1592. Sinha, N.K. (1987) Specificity and efficiency of editing of mismatches involved in the formation of base-substitution mutations by the 3' --* 5' exonuclease activity of phage T4 DNA poiymerase, Proc. Natl. Acad. Sci. (U.S.A.), 88, 915-919. Weymouth, L.A., and L.A. Loeb (1978) Mutagenesis during in vitro DNA synthesis, Proc. Natl. Acad. Sci. (U.S.A.), 75, 1924-1928. Yatagai, F., and B.W. Glickman (1990) Specificity of spontaneous mutation in the lacl gene cloned into bacteriophage M13, Mutation Res., 243, 21-28. Zinder, N.D., and J.D. Boeker (1982) The filamentous phage as vectors for recombinant DNA - a review, Gene, 19, 1-10. Zoller, M.J., and M. Smith (1982) Oligonucleotide-directed mutagenesis using M13-derived vectors: an efficient and general procedure for the production of point mutations in any fragment of DNA, DNA 3, 479-488.