A genetic enrichment for mutations constructed by oligodeoxynucleotide-directed mutagenesis

73

Gene, 37 (1985) 73-8 1 Elsevier GENE

1344

A genetic enrichment for mutations constructed by oligodeoxynucleotide-directed mutagenesis (Recombinant

DNA, amber mutants;

Ml3 vectors;

gapped heteroduplex;

mismatch

repair;

phage 1 att site)

Carl E. Bauera, Steven D. Hessea, Daryle A. Waechter-Brulla”, Steven P. Lynna*, Richard I. Gumportb and Jeffrey F. Gardner=** “Department of Microbiology, ‘Department of Biochemistry and College of Medicine, Universityof Illinois. Urbana, IL 61801 (U.S.A.) Tel. “(217) 333-7287 and b(21 7) 333-2852 (Received

February

(Revision

received

(Accepted

March

8th, 1985) March

22nd, 1985)

3Oth, 1985)

SUMMARY

A genetic enrichment procedure for mutations constructed by oligodeoxynucleotide(oligo)-directed mutagenesis of DNA cloned in M13mp vectors is described. The procedure uses an Ml3 vector that contains the cloned target DNA and amber (am) mutations within the phage genes I and II. This vector cannot replicate in a suppressor-free (sup”) bacterial strain. A gapped heteroduplex is formed by annealing portions of a complementary ( - )strand containing wild-type copies of genes I and II to the am-containing template ( + )strand. The oligo is annealed to the single-stranded (ss) region and the remaining gaps and nicks are repaired enzymatically to form a closed circular heteroduplex structure. By transfecting the DNA into a sup ’ host we promote the propagation of heteroduplexes with the oligo-containing (- )strand since only this construction contains the wild-type copies of genes I and II. This procedure eliminates the need for any physical separation of the covalently closed circular DNA that contains the oligo from the ss template. Using this technique we have constructed 17 point mutations with mutation frequencies ranging from 2-20% for single base changes and from 0.3-9x for multiple base changes. In addition, we found that the mutation frequencies were affected by the state of DNA methylation in the ( + ) and ( - )strands.

INTRODUCTION Abbreviations:

The use of synthetic oligos as a tool for creating mutations was first demonstrated by constructing site-specific mutations within 4x174 (Hutchison

attachment

am,

double-stranded; N-terminal peptide

IPTG,

section

of Microbiology

and Molecular

School, 25 Shattuck

Street, Boston,

stranded;

address:

Department Medical

MA 02115 (U.S.A.) ** To whom

correspondence

and reprint

requests

should

be

virion;

( - )strand,

( + )strand.

addressed.

0378-l 119/85/$03.30

sup ‘,

0

1983 Elsevier

units;

Science

Publishers

1 ds,

suppressor-free;

for a

me + , methylated; Klenow

fragment

RF, replicative

( + )strand, DNA

IucZ’,

gene coding

me -,

acid; oligo, oligodeoxynucleotide; PolIk,

I; R, resistant;

indolyl-p-D-galactoside;

Tel. (617) 732-1917.

bacteriophage

bp, base pair(s);

isopropyl+D-thiogalactoside;

Nal, nalidixic

DNA polymerase

Harvard

atrP,

of the /I-galactosidase

pfu, plaque-forming Genetics,

mutation;

active in cc-complementation;

unmethylated; * Present

amber

site; PME, /I-mercaptoethanol;

strand

Xgal, DNA

of E. coli

form; ss, single

5-bromo-4-chloro-3strand

complementary

in the Ml3 to

the

74

et al., 1978; Razin 1979). Initially construction DNA.

et al., 1978; Gillam

this procedure

of selectable

However,

recent

nonselectable

mutations

et al., 1979;

M 13 cloning

by differential 198la),

constructing 1980;

mutagenesis

mutations 1981b;

the

hybridization of

1983) and their use (Zoller

and

Smith,

of techniques

within ds templates Morinaga

of

the construction

198 1; 1983), and the improvements et al.,

including

for the identification

vectors (Messing,

for site-directed

to the

within ss phage

advances

of techniques

(Wallace

was limited

mutations

development

and Smith,

et al.,

for

(Wallace

1984) have

made this an attractive procedure for constructing specific mutations within any cloned target DNA. In principle, oligo-directed mutagenesis is often the best method for constructing specific bp changes. However, the difftculty in successfully performing this technique has hampered its routine use. Two of the more time-consu~ning steps of the procedure

on the me ’ strand

favor of the base present man

et al.. 1980; Pukkila

(Kad-

et al., 1983). Since

the

in vitro synthesized

strand that contains the mutagenizing oligo is me it is likely that the oligo would be flanked

by hemimethylated

presence

of hemimethylated

mutation

frequency

the mismatch Kramer

sites. The

sites could decrease

by directing preferential

toward

the wild-type

et al. (1982) demonstrated

of methylation frequency

(GATC)

the

repair of

sequence. that the state

in ds DNA affects the marker rescue

of a mutation

born on a restriction

ment. They also suggested

that site-directed

fragmuta-

tion frequencies might be increased by incorporating an oligo into hemimethyIated gapped ds DNA which contains an me ( + )strand and an me + { - )strand. They speculated that this should increase the mutation frequency by directing mismatch repair in favor of the sequence present on the me ’ strand which the oligo is incorporated.

into

involve the synthesis of the oligos and the construction of mutants. Although recent advances in automated DNA synthesis and the introduction of a segmented-support synthesis (Frank et al., 1983;

We have investigated the use of a genetic enrichment procedure, initially described by Messing (1983) for marker rescue of mutations from restriction fragments, for its usefulness in oligo-directed

Matthes et al., 1984) have reduced the problem of constructing oligos, there remains considerable difft-

mutagenesis. This technique uses an M 13 template which contains am mutations within two genes

culty in obtaining mutations at a high enough frequency to facilitate their routine isolation. One major problem encountered during mutagenesis is a high

essential for phage growth. A portion of a complementary wild-type M 13 strand is hybridized to the template to form a gapped heteroduplex which will

background

selectively replicate in a SUP’ host. The mutagenic oligo is then hybridized to the ss region and incorpo-

from template

of unmutagenized

phage which arises

that had not annealed

to the oligo or

was incompletely converted into a ds form. Previous techniques for increasing the mutation frequency, either by isolating the covalently closed ds DNA from the template or by selectively destroying the rem~ning ss DNA, have employed alkaline sucrosegradient centrifugation (Zoller and Smith, 198 1; 1983), CsCl density-gradient centrifugation (Baas (Simons et al., et al., 1981), gel electrophoresis 1982), hydroxylapatite chromatography (Kramer et al., 1982), or Sl nuclease digestion (Hutchison et al., 1978). While these procedures significantly increase the mutation frequency, we have found them to be both difficult and time-consuming. Another factor which may reduce the mutation frequency is the in vivo correction of the mismatch by the host mismatch repair systems to the wild-type sequence present on the template strand. One repair pathway is methylation-directed and in the presence of hemimethylated DNA will correct a mismatch in

rated into the wild-type (gene I and II) M 13 strand. When this DNA is transfected into a sup 1 strain we observe a 20-fold increase in the mutation frequency in comparison with a suppressor-containing strain. Furthermore, the mutation frequency was affected by the state of methylation of the gapped heteroduplex DNA. During the completion of this work two reports of related mutagenic techniques have appeared. Kramer et al. (1984b) have reported constructing a G -+ A mutation and its reversion using a simiIar am enrichment procedure. In addition, Marmenout et al. (1984) used gapped ds DNA with an antibiotic selection to create an A t C mutation. In general, our findings support and extend these observations. We have constructed 17 different mutants including double and triple mismatches and have determined the individual contributions of the am selection and the mismatch repair to the increased mutation fre-

quencies. We find that the am selection is the primary augmenting factor whereas the mismatch repair system acts as a minor effector. In addition, we find considerable variation from mutant to mutant in the frequencies we obtain.

MATERIALS

AND METHODS

(a) Strains

M13mplOam (gene Iam and gene IIam strain) and M13mplO + (gene I + and gene II + derivative) were obtained from Pharmacia P-L Biochemicals and Amersham Corporation, respectively. M13mplO + : AlacZ’ was constructed by BamHI + PvuI cleavage of M13mp 10 + RF followed by filling in the recessed 3’ end at the BamHI site and removal of the protruding 3’ end at the PvuI site with T4 DNA polymerase. The blunt-end fragment was subsequently circularized with T4 DNA ligase. This construction contains a BarnHI-PVuI deletion within 1acZ’ and, in addition, regenerates the BamHI recognition sequence. M13mplOam: lacZ’am317 contains a T -+ A transversion which destroys the Hind111 site and also creates an am stop codon in the IacZ’ reading frame. This mutation was mutagenesis of constructed by site-directed M13mplOam (C.E.B., unpublished data). Strain JM103 (supE) (Messing, 1983) and JM105 a sup ’ strain were obtained from P-L Biochemicals. GM 119 NalR[ F’KmR], a dam supE strain was constructed by mating F’KmR into a NalR derivative of the dam parent (GM1 19) (Pukkila et al., 1983) which was obtained from the E. coli Genetic Stock Center, Yale University. (b) Oligodeoxynucleotide

synthesis

The heptadecamer d(CCAGTGCCAAGCTTGGG) that was used for site-directed mutagenesis of the IacZ’ gene was synthesized with an Applied Biosystem Synthesizer Model 380A. Other oligos were synthesized either individually by solid-phase phosphate-phosphotriester chemistry (Ito et al., 1982) or simultaneously by a cellulose-segmented support method (Matthes et al., 1984). The oligos were deblocked and purified by Sephadex G50

chromatography followed by preparative polyacrylamide-gel electrophoresis. (c) DNA isolation The me+ RF and ss DNA were prepared from infected cells or phage particles, respectively, by infecting JM103 as described previously (Bauer et al., 1985). The me- RF DNA and phage were prepared by infecting 2 liters of LB medium mid-log GM 119 NalR[ F’KmR] containing = 0.6) with approx. 1012 pfu of Ml3 phage. (A600nm The template and RF DNA were isolated as described (Bauer et al., 1985). The isolated RF was shown to be me- at GATC sites by its sensitivity to cleavage by Mb01 and resistance to cleavage by DpnI (not shown). (d) Site-directed mutagenesis of lucZ ’

M13mplO’ :AlacZ’ RF (me* and me-) were digested with DpnI or MboI, respectively, extracted with phenol (saturated with 10 mM Tris . HCl pH 8.0, 1.0 mM EDTA) extracted with ether, precipitated with ethanol, and dried. The DNA pellet was resuspended to a final concentration of 0.5 pg/pl in 10 mM Tris. HCl pH 8.0, 1.0 mM EDTA. The gapped heteroduplex template used for mutagenesis was constructed by mixing 1.0 pg of the digested M13mplO+ : AlacZ’ RF with 0.5 pg of M13mp10am:lacZ’am317 template ss DNA in 30 ~1 of 100 mM NaCl, 40 mM Tris * HCl pH 7.5, 20 mM MgCl,, 2.0 mM jME. The template-RF mixture was denatured by heating to 100’ C for 3 min and allowed to cool over 20 min to 65 “C. After addition of 5’-P-oligo (50.0 pmol), the mixture was cooled slowly to 30’ C and then placed in ice water for 15 min followed by the addition of 70 ~1 of 22 mM Tris . HCl pH 7.5, 11 mM MgCl,, 1.0 mM BME, 0.83 mM dATP, 0.83 mM dCTP, 0.83 mM dGTP, 0.83 mM dTTP, 0.4 mM rATP, 0.5 units PolIk (Boehringer Mannheim), and 0.5 units T4 DNA ligase (Bethesda Research Laboratories). After an additional 30 min at 0 ’ C, the primer extension mixture was incubated at 23 “C for 2 h. An aliquot of the mixture (50 ~1) was transfected into competent JM105 cells (Maquat and Reznikoff, 1978) and plated immediately after heat shock onto fresh YT plates (Miller, 1971) containing 0.4 mg/ml

76

Xgal and 0.4 mM IPTG in the top agar. The mutation frequency was calculated as the number of Lac ’ plaques divided by the total number of plaques.

BumHI

(e) Mutagenesis of artP An Ml3 clone of attP (Bauer et al., 198.5) was mutagenized by hyb~dizing Hind111 + HaeIII-cut M13mpll+ RF to Ml3mp9~ : attP template. The resulting gapped ds DNA was annealed with the oligo, primer extended, and transfected as described above. The mutations were identified by selective hybridization and sequenced as described (Bauer et al., 1985).

ligase DNA polymerase anneal

oligonucleotide dNTPs PolK DNA ligose

I

A

RESULTS

{a) Rationale of the am enrichment procedure A major factor contributing to a low mutation frequency during oligo-directed mutagenesis is a high back~ound of unmutated phage caused by template which has not been converted to a covalently closed heteroduplex. To circumvent this problem we used two M 13 strains with different genotypes. The strain that provided a template for mutagenesis contains two am mutations (Messing et al., 1981; Messing, 1983). One mutation is within a gene essential for RF replication (gene II) and the other is located within a gene involved in head assembly (gene I). Thus, this phage cannot propagate in a sup” strain. To the template ( + )strand (Fig. 1) we anneal parts of the complementary ( - )strand that contain wild-type gene I and gene II to create a heteroduplex with a gap in the region to be mutagenized. The mutagenic oligo is incorporated into the (- )strand of the gapped heteroduplex by hybridizing it to the ss region followed by filling in the rem~ning gaps and sealing the nicks to form a covalently closed heteroduplex. Since the ( - )strand is transcribed into the message (Konings and Schoenmakers, 1978), the wildtype copies of gene I and gene II in the ( - )strand will complement the am mutations present in the template ( + )strand. Thus by transfecting the primer extension mixture into a sup ’ host, the heteroduplex DNA, some of which contains the oligo, will form

I

!i -

v

sup0 strain

Fig. 1. Amber enrichment by reversion

for mutations

of an am mutation

derivativeofMl3mpl~+

the BarnHI-PvuI

M13mplO’

(MATERIALS

gapped

heteroduplex gene

M13mplO’

was :dlacZ’

was

(--x--). The IacZ am mutation template remaining

segment)

ss region or

of

a). A of the

Mbol-digested

to Ml3mplOam:iacZ’ by a black segment

segment. The mutagenic

was hybridized

by

and the oligo

to the ss region of the

into the ( - )strand

by tilling in the

gaps with Pollk and sealing the nicks with DNA ligase AND METHODS,

section d). When transfected

only the heteroduplex

with wild-type

The same procedure plate

DpnI

gene

section

in genes I and II are designated

and incorporated

(MATERIALS

appropriate

150-bp

hybridized

’ gene is designated

into a sup” strain ( - )strand

a

when

within it by a crossed

(crosshatched

constructed

of the 1ucZ

METHODS,

contining

am317. The am mutations

:&ucZ’)was

fragment AND

formed

in MI3 as exemplified

in the la&?’ gene. A deletion

(Ml3mplO’

by deleting

IacZ’

progeny transfect

can be applied

to any cloned

M 13 am phage by constructing

with the homologous

DNA containing

the

copies of genes I and II will replicate.

parent

Ml3

insert in the

the gapped wild-type

ds tem-

strain

(see

Table 11).

plaques since this construction contains the functional copies of gene I and II. In contrast, cells infected with only ss template that has not formed heteroduplex DNA will not form plaques due to the absence of functional copies of gene I and II. This procedure should select for mutants by preferentially allowing the ds heteroduplex DNA to form plaques and thus would be functionally equivalent to in vitro

17

techniques for physically isolating ds heteroduplex DNA. Finally, we note that upon transfection and replication a heteroduplex could give rise to mixed progeny containing either wild-type or mutant copies of genes I and II. Presumably, complementation would allow the replication of both phages. However, we believe that complementation does not alter the frequency of mutants since the cells are plated immediately after transfection and thus each plaque should arise only from a heteroduplex. The procedure we use to identify oligo-directed mutations is sensitive enough to detect mutations within plaques containing mixed progeny. In addition, replating the phage which arise from the initial transfection onto a fresh lawn of sup o cells shows no significant difference in the mutant frequency (C.E.B., unpublished observations).

M13mplOam:ZacZ’am317 (Fig. 1) it generates a gapped heteroduplex structure containing a SO-bp long ss region of the ZacZ’ gene to which the oligo can hybridize. We used the frequent cutting enzymes DpnI or Mb01 to cleave the M13mplO + : AZacZ’ RF into six fragments to reduce the possibility that the digested M13mplO + : AZacZ’ RF could reanneal and ligate to form viable phage and consequently increase the Lac _ background. Control experiments demonstrated that the digested M 13mp 10 + : AZacZ’ RF produced no plaques after primer extension in the absence of template (not shown). By transfecting the primer extension mixture into a sup’ strain we observed 11 y0 mutants (Table I). This is approximately a 20-fold enrichment over the mutation frequency observed (0.5 %) when transfecting a supE strain that confers no selective enrichment.

(b) Quantification

(c) Effect of mismatch repair on the am enrichment

of the am enrichment

To quantify the extent of the am enrichment we performed site-directed mutagenesis on the IacZ’ gene of M13mplO + . The oligo we used converts an am mutation within the 1acZ’ gene (lacZ’am317) to the wild-type sequence (Fig. 2) thereby allowing us to measure the mutation frequency by scoring the Lac phenotype of the Ml3 plaques on Xgal indicator plates. The phenotypic change from Lac - to Lac + allows a stringent test of the oligo-directed base change since mutations at secondary locations would not readily result in a Lac+ phenotype. To generate a heteroduplex with a gap in the 1acZ’ gene we constructed a IacZ’ deletion (Fig. 1) of M 13mp 10 + by removing the BamHI-PvuI fragment (see MATERIALS AND METHODS, section a). When RF of this vector (M13mplO’ : AlacZ') is cleaved with DpnI (or MboI) and is hybridized to

. . .THR M13mplOam:

M13mplO+

CYS SER PRO SER

STOP

5’ . . . ACC ‘TGC AGC CCA AGC TAG GCA CTG GCC GTC .. . -

lacZ’am317

Oligodeoxynucleotide

E. coli possesses a mismatch repair system that detects the methylation state of DNA and preferentially corrects mismatches in hemimethylated DNA in favor of the sequence present in the me + strand (Radman et al., 1980; Pukkila et al., 1983). Kramer et al. (1982) reported a significant increase in the marker rescue frequency of a mutant ZacZ gene in Ml3 with a IacZ’ + restriction fragment when using hemimethylated DNA. Presumably this increase in marker rescue frequency was due to directed repair of the mismatch. We tested the effect of the methylation-directed mismatch repair system on the am enrichment technique in order to determine if a further enhancement in the mutation frequency could be obtained. We coupled these procedures by isolating template or RF from hosts containing a dam + or dam - genotype. The dam gene encodes an

(17-mer)

3 '-G GGT TCG AAC CGT GAC C 5' . . .ACC TGC AGC CCA AGC ‘ITG GCA CTG GCC GTC.. . . .THR CYS SER PRO SER LEir ALA LEU ALA VAL..

.

.

Fig. 2. Oligo-directed mutagenesis of IucZ. M13mp10am:lacZ’am317 contains a T + A transversion (underlined base) which creates a stop codon at amino acid position 15. The mutagenic oligo, which is partially complementary to the 1acZ ’ am3 17 template, contains a mismatch which will revert the am stop codon to the wild-type LEU codon present in 1ucZ’ and regenerate a Hind111 site (overlined).

7x

TABLE

I

duplex which potentially

Combined

amber

and mismatch

repair

homomethylated

enrichments

methylation Methylation

Frequency

( + )strnnd/‘(

)strand

me

GATC

;‘me +

mc

;mc

frequency

methylated

20”,, 14”,,

or

on state of

DNA used. One

site is located within the ss gap 3’ to the oligo

mutation

7°C)

me

depending

of the RF or template

and thus could not be completely

I I c’,)

’ ,:mc

hemimethylated

duplex DNA at five of the six darn

sites (GATC)

methylation

’ mutants”

Lac

111 e ’ /me+

of

contains

repair

DNA,

to the

observed which

sequence

manipulated.

The

(Table I) for hemi-

should

direct

mismatch

of the

oligo

containing

( - )strand (Table I, line 3), is approx. two-fold higher I’ Number plaques

of Lac’

present

plaques

divided

when transfected

by the total

into JMl05

number

independent

primer extension

mately

SO0 plaques

scored

than

(MATERIALS

section d).The values represent

.4ND METHODS, ofthrec

of

experiments

an average

observed

for heteroduplex

from fully me + or fully me

per assay.

sponds to the conditions used previously in techniques for mutagenesis (Zoller et al., 1982; 1983; Simons et al., 1982) reduces the mutation frequency from 11 to 7%.

II

Mutation

frequency

in M 13mp9: attP using the am enrichment

procedure

Mismatch

Oligodeoxynucleotide”

Methylation ( + )strand/(

Frequency mutants

- )strand

TATTAGTAACCTCTAh

A/C

met/me+

4’

ATGCAGTyACTATGAh

T/G

me+/me+

4’

TGATAGTAACCTGTTh

A/C

met/me+

5’

TTATTTTA
A/C

met/me+

1’

AAATAAfiATTTTATh

A/C

me+/me+

3’

TTAGTACAAAAAAGC’

C/T

me+/me+

8d

TACTTAGGTGGTATTh

G/T G/T

me -/me+

3’

me-/me+

3”

me-/me+

2’

CTATGAATTCACTACTT”

A/A TC/GT

me-/me+

9’

AAGTAGGGGATTCATAG”

CC/AA

me-/me+

5’

TCAACTACCCAGATGGTAT”

CC/AA

me-/me+

3’

CAGTCACTTCGAATCAACTh

TC/TA

me-/me+

ATTGATAAGGCATGCTTTTIT”

CC/CT

me- /me+

5” 2’

GCAACAAAGGGATAAGCAA’

CC/AA

me+/me+

3‘

GAAACGTAGCATGCTATAAAT’ ~ _

GCC/AAT

me-/me+

0.3’ ____

TGAATCAGCTACTTAh CTATGAAACAACI’AC”

.’ The underlined

base within the oligo represents

’ Mutations

were identified

by screening

ct al.. 198la)

and confirmed

by dideoxy

c C.E.B. (unpublished

the position

from 100-500 plaques sequencing

(Table I, footnote

using the mutagenic

a).

observations).

(unpublished

observations).

’ D.A.W.-B. (unpublished

observations).

f Oligos synthesized

by phosphate-phosphotriester

‘I Oligos synthesized

using a segmented-support

’ Oligo constructed

on a commercial

synthesizer

chemistry method

(Ito et al., 1982).

(Matthes

(MATERIALS

of (“;,)b

of the mismatch. by selective hybridization

” Bauer et al. (1985). D J.F.G.

DNA DNA. A

construction that would bias the repair against the desired mutation (Table I, line 2) and which corre-

with approxi-

enzyme involved in deoxyadenosine methylation (Marinus, 1984). After cutting M 13mp 10 + : AfucZ’ with DpnI or MboI, we constructed a gapped hetero-

TABLE

the values

constructed

et al., 1984). AND METHODS,

section

b).

oligo as a probe (Wallace

79

(d) Mutagenesis

of a cloned target DNA

The value of this method is best demonstrated by its ability to rapidly construct many mutations within a cloned target DNA. Using a M13mp9am clone of the phage 1 art site (M13mp9am:attP; Bauer et al., 1985) we constructed 16 mutations within attP with the am enrichment procedure (Table II). The mutation frequencies varied from 2% to 8% for single mismatches and from 0.3% to 9% for multiple mismatches. These mutations were constructed with oligos that were synthesized and purified by a variety of techniques.

DISCUSSION

The am enrichment procedure offers several advantages over currently used techniques for increasing mutation frequencies during oligo-directed mutagenesis. One improvement is the simplicity of an in vivo genetic enrichment that uses common Ml3 and E. coli strains. In addition, the mutation frequencies we observed are comparable to those obtained using more difficult in vitro techniques that rely on the physical separation of ss template from covalently closed ds DNA. Using a somewhat different approach for constructing gapped heteroduplex DNA and for selecting the wild-type gene I and II Ml3 derivative Kramer et al (1984b) reported a three-fold higher mutation frequency than we have observed. The higher value that they reported may be due to either the particular mutation that was corrected with an oligo containing a CA mismatch (see below) or to the different procedures used to construct the gapped heteroduplex and select the Ml3 progeny containing the mutation. To determine what effect mismatch repair might have on oligo-directed mutation frequency, we combined the am selection with the directed mismatch repair procedure described by Kramer et al. (1982). They observed a 20-fold enrichment during marker rescue experiments with restriction fragments, which resulted in a GT mismatch, when using gapped duplex DNA with three of the five GATC sites hemimethylated. Using a similar hemimethylated gapped ds structure and an oligo containing an AA

mismatch, we detected a two-fold effect that was dependent upon the state of methylation of the duplex DNA. The disparity may be explained by considering several recent observations concerning mismatch repair. Kramer et al. (1984a) and Pukkila et al. (1983) demonstrated that methylation-dependent mismatch repair does not correct all types of mismatches with equal frequency. It should be noted that the more recent study by Kramer et al. (1984a) detected only a three-fold methylation-directed repair for GT mismatches and less than two-fold effect for AA mismatches. Fishel and Kolodner (1983) demonstrated that E. coli also possesses a second mismatch repair pathway that is recF-dependent and corrects either strand regardless of the state of DNA methylation. Furthermore, they concluded that the recF pathway is the primary pathway of mismatch repair. Thus, the low level of methylation-directed repair that we observed may be due to several factors including the inability of the methylation-dependent repair system to efficiently recognize an AA mismatch, the location of the mismatch relative to the dam methylation sites (GATC), the state of the DNA (nicked vs. covalently closed) upon transfection, or a high background of methylation-independent recF-mediated repair. We note that, even though we did not observe a dramatic difference in the mutation frequencies when using hemimethylated DNA, the slight effect does have a significant impact on the number of plaques which need to be screened for the presence of the mutation. For example, the two-fold methylation-dependent effect observed for the ZucZ’ mutation (11 y0 vs. 20%) would require that one-half as many plaques be screened to have > 90% probability of obtaining the desired mutation, e.g., 10 vs. 21 plaques (Kramer et al., 1982). This difference would be quite significant if the screening for mutations were by direct sequencing. Finally we emphasize that this procedure works well for mutating DNA fragments cloned into the commonly used M13mp vectors (mp7, mp8, mp9, mp 10, mp 11) containing am mutations in genes I and II. Using this procedure several groups have constructed mutations within cloned target DNA (Table II; Bauer et al., 1984; Mantsala and Zalkin, 1984; Paluh et al., 1985). It is apparent from the mutation frequencies reported that there are variations in the mutation frequencies observed with dif-

ferent clones and oligos. Furthermore, the average frequency

we found that

IlI,C.A.,

Philips. S., Edgell, M.H.,Gillam,

P. and Smith,

for creating single mismatches

in attP is significantly

Hutchison

lower than the value observed

M.: Mutagenesis

DNA sequence.

Ito, H., Ike, Y.. Ikuta, S. and Itakura,

the IucZ’ gene (Table II). These dif-

of polynucleotides,

ferences

have

copolymers

methods

several

used for synthesizing

degrees of template the location

including

the

the oligos, the varying

and oligo purity, the difference in

and the type of mismatch,

tial repair

of some

methylation

(GATC)

inefficient

causes

mismatches,

the existence

sites within the insert,

identification

tial hybridization.

the preferen-

of the mutations

Although

of

or the

by differen-

the mutation

frequency

observed

for the attP insert is lower than the value

observed

for 1ucZ’ it is still significantly

the values obtained richment.

higher than

(< 0.5 %) without

selective en-

VI.

in a

K.: Solid phase synthesis

Further

studies

on

polystyrene

Nucl. Acids Res. 10 ( 1982)

for the solid support,

1755-1769. Konings,

R.N.H.

and Schoenmakers.

the filamentous

phage genome,

J.G.G.:

Transcription

in Denhardt,

Harbor

Laboratory,

of

D.T., Dressier,

D. and Ray, D.S. (Eds.), The Single Stranded Cold Spring

DNA Phages.

Cold Spring

Harbor.

NY.

1978, pp. 507-530. Kramcr,W.,

Schughart,

of DNA

cloned

K. and Fritz, H.J.: Directed

in filamentous

phage:

mutagenesis

influence

methylated

GATC sites on marker

fragments.

Nucl. Acids Res. IO (1982) 6475-64X5.

Kramer,

B., Kramer,

mismatches 38 (1984a) Kramer,

recovery

W. and Fritz,

are corrected

methyl-directed

of hcmi-

from restriction

H.J.: Different

with different

DNA mismatch-repair

basc:‘basc

etlicicncies

by the

system of E. &i. Cell

879-887.

W., Drutsa,

M. and Fritz,

V., Janscn,

H.W., Kramer.

H.J.: The gapped

oligonucleotide-directed

ACKNOWLEDGEMENT

position

J. Biol. Chcm. 253 (1978) 6551-6560.

when mutating may

S., Jahnke.

at a specific

duplex

mutation

B., Pflugfelder,

DNA approach

construction.

to

Nucl. Acids

Res. 12 (1984b) 9441-9456.

This work was supported in part by NIH grant GM 28717. C. Bauer was supported by NIH training grant GM 07283.

Mantsala,

P. and

function:

direct mutagenesis. L.E. and

and

mutant

Marinus, Baas, P.D.,Teertstra,

W.R.,Van

Mansfeld,

Van der Mare], G.A., Veeneman, Construction

A.D.M., Jansz, H.S.,

G.H. and Van Boom, J.H.:

of viable and lethal mutations

bacteriophage

$X174 using synthetic

in the origin of

Bauer, C.E.,Hesse,

S.D.,Gardner,

interactions

during

ombination.

Reznikoff.

lactose

Cold

J.F. and Gumport,

bacteriophage Spring

RI.: DNA

lambda

site-specific

Symp.

Quant.

Harbor

J.F. and Gumport,

homology

required

specific recombination.

rec-

Biol.. 49

co/i: the identification

The extent lambda

R.: Gene conversion

oftwo repair pathways

in Friedberg,

Cellular Responses

RI.:

for bacteriophage

of site-

J. Mol. Biol. 181 (1985) 187-197.

Fishel, R.A. and Kolodner,

EC.

and

to DNA Damage,

in Escherichia for mismatched

Bridges,

B.A.

(Eds.),

UCLA Symp. on Mol.

and Cell. Biol., Vol. 11. Liss, New York, 1983, pp. 309-324. R., Heikens,

W., Heisterberg-Moutsis,

H.: A new general

approach

of large numbers

solid supports.

G. and Blocker,

for the simultaneous of oligonucleotides;

chemical segmental

Nucl. Acids Res. 1 I (1983) 4365-4377.

S. and Smith,

M.: Site-specific

thetic oligodeoxyribonucleotide tions and minimum (1979) 81-97.

and

W.S.:

In vitro

by

analysis

interaction

promoters.

Riggs,

site-

J. Mol.

of the

with wild type Biol.

125 (1978)

mutagenesis

primers,

oligodeoxyribonucleotide

DNA, m Razin, A., DNA

and Biological

New York,

1984, pp. 81-l 10. A., Remaut,

Significance.

mutagenesis:

Methylation,

Springer-Verlag,

E., Van Boom,

directed

by hemimethylation

(Eds.),

J. and

Fiers.

selection

of GATC-sequences.

W.:

of mutants

Mol. Gen. Genct.

195 (1984) 126-133. Matthes,

H.W.D.,

cal synthesis microscale. Messing,

Zenke,

W.M.,

Grundstrom.

M. and Chambon,

T., Staub,

P: Simultaneous

of over one hundred

A..

rapid chemi-

oligonucleotides

on :I

EMBO J. 3 (1984) 801-805.

.I.: New Ml3 vectors

for cloning.

Methods

Enzymol.

I01 (1983) 20-78. Messing,

J., Crea, R. and Seeburg,

DNA sequencing.

Miller, J.H.: Experiments Harbor Morinaga,

Laboratory,

genesis

in Molecular Cold Spring

Y., Franceschini,

Improvement

P.H.: A system Genetics.

Harbor,

T., Inouye,

of oligonucleotide-directed

using

for shotgun

Nucl. Acids Res. 9 (1981) 309-321.

double-stranded

plasmid

Cold Spring

NY, 1972. S. and

Inouye

site-specific DNA.

M.: muta-

Bio/Tech-

nology 2 (1984) 636-639.

using syn-

I. Optimum

of prokaryotic A.D.

Biochemistry

Wintzerith,

C.E., Gardner,

nucleotides,

H.

Ohgonucleotide

(1984) 699-705.

Gillam,

in glutamine

amidotransferase

J. Biol. Chem. 259 (1984) 14230-14236.

M.G.: Methylation

Ceder,

Marmenout,

oligodeoxyribonucleo-

tides. J. Mol. Biol. 152 (1981) 615-639.

synthesis

amidotransferase cysteine

467-490.

REFERENCES

Frank,

Glutamine

coli RNA polymerase

E.wherichia

sequence

H.:

of the active-site

phosphoribosylpyrophosphate Maquat,

Bauer,

Zalkin,

replacement

condi-

length. Gene 8

Paluh,

J.L., Zalkin,

anthranilate

using site-directed 1889-1894.

H., Betsch,

synthase

D. and Weith,

H.L.: Study

of

by replacement

of cysteine

84

function mutagenesis.

J. Biol. Chem.

260 (1985)

81 Pukkila, P.J., Peterson, J., Herman, G., Modrich, P. and Meselson, M.: Effects of high levels of DNA adenine methylation on methyl-directed mismatch repair in Escherichiu coli. Genetics 104 (1983) 571-582. Radman, M., Wagner Jr., R.E., Glickman, B.W. and Meselson, M.: DNA methylation, mismatch correction and genetic stability, in Alacevic, M. (Ed.), Progress in Environmental Mutagenesis. Elsevier, Amsterdam, 1980, pp. 121-130. Razin, A., Hirose, T., Itakura, K. and Riggs, A.D.: Efficient correction of a mutation by use of chemically synthesized DNA. Proc. Natl. Acad. Sci. USA 75 (1978) 4268-4270. Simons, G.F.M., Veeneman, G.H., Konings, R.N.H., Van Boom, J.H. and Schoenmakers, J.G.G.: Oligonucleotide-directed mutagenesis of gene IX of bacteriophage M13. Nucl. Acids Res. 10 (1982) 821-832. Wallace, R.B., Shaffer, J., Murphy, R.F., Bonner, J., Hirose, T. and Itakura, K.: Hybridization of synthetic oligodeoxyribonucleotides to $X174 DNA: the effect of single base pair mismatch. Nucl. Acids Res. 6 (1979) 3543-3557. Wallace, R.B., Johnson, P.F., Tanaka, S., Schold, M., Itakura K. and Abelson, J.: Directed deletion of a yeast transfer RNA intervening sequence. Science 209 (1980) 1396-1400.

Wallace, R.B., Johnson, M.J., Hirose, T., Miyake, T., Kawashima, E.H. and Itakura, K.: The use of synthetic oligonucleotides as hybridization probes, II. Hybridization of oligonucleotides of mixed sequence to rabbit B-globin DNA. Nucl. Acids Res. 9 (1981a) 879-894. Wallace, R.B., Schold, M., Johnson, M.J., Dembek, P. and Itakura, K.: Oligonucleotide directed mutagenesis of the human b-globin gene: a general method for producing specific point mutations in cloned DNA. Nucl. Acids Res. 9 (1981b) 3647-3656. Zoller, M.J. and Smith, M.: Oligonucleotide-directed mutagenesis using Ml3 derived vectors: an efficient and general procedure for the production of point mutations in any fragment of DNA. Nucl. Acids Res. 10 (1982) 6487-6500. Zoller, M.J. and Smith, M.: Oligonucleotide-directed mutagenesis ofDNA fragments cloned into Ml3 vectors. Methods Enzymol. 100 (1983) 468-500. Communicated by S. R. Kushner.