A rapid and efficient method for targeted random mutagenesis

313

Gene, 64 (1988) 313-319 Elsevier GEN 02351

A rapid and efficient method for targeted random mutagenesis (Recombinant

DNA;

RNA component

2’-deoxynucleoside

of RNase

5’-O-( 1-thiotriphosphate);

P; m@ gene; targeted

exonuclease

III; multiple

cloning-sites;

misincorporation)

Hideaki Shiraishi and Yoshiro Shimura Department of Biophysics, Faculty of Science, Kyoto University, Kyoto 606 (Japan) Received

28 October

Accepted

12 November

1987

Received

by publisher

1987 25 January

1988

SUMMARY

We describe a new rapid method for random introduction of single-nucleotide (nt) substitutions into a small segment of cloned DNA. A DNA fragment containing a sequence to be mutagenized is inserted into a multiple cloning site sequence of a vector plasmid. The plasmid is linearized with two adjacent cuts (generating 5’ and 3’ protruding ends) and then synchronously and unidirectionally digested with exonuclease III (Exo III) so that the 3’ termini generated are localized within the target region. A non-complementary cc-thiophosphate nucleotide is misincorporated into the 3’ terminus generated by Exo III. Since the nucleotide analogue is resistant to the 3’-5’ exonuclease activity of DNA polymerase I, its misincorporation into the 3’ termini is irreversible. Then, the single-stranded region is filled-in with four canonical nucleotides, and the plasmid is recircularized. This procedure was used to mutagenize a specific region of the m@ gene of E. coli. By sequencing 72 randomly selected clones, we found that 27 clones (37.5%) had nucleotide substitutions distributed within the desired region of a 55-nt-long segment of the gene. The procedure is simple and is applicable to any DNA molecule.

Targeted mutagenesis is a useful technique to study structure and function of nucleic acids and proteins (for review, see Botstein and Shortle, 1985).

Although oligodeoxynucleotide-directed mutagenesis is widely used to obtain specific mutants, targeted random mutagenesis is still preferred for the initial survey of functionally important sequences. In targeted random mutagenesis, several proce-

Correspondence

triphosphate;

INTRODUCTION

Faculty

to: Dr. Y. Shimura,

of Science,

Tel. (075)751-2111,

Kyoto

Department

University,

Kyoto

of Biophysics, 606

(Japan)

ext. 4230.

phosphate); phate); MCS,

Abbreviations:

bp, base pair(s);

dNTP,

2’-deoxynucleoside

5’-

dNTP[aS], dTTP[ctS],

Exo III, exonuclease multiple

(large) fragment

0378-l I19/88/$03.50 0 198X Elsevier Science Publishers B.V. (Biomedical Division)

2’-deoxynucleoside 2’-deoxythymidine

cloning

5’-O-(l-thiotri5’-O-(l-thiotriphos-

III; kb, kilobase

site; nt, nucleotide(s);

of E. coli DNA polymerase

or 1000 bp; PolIk,

I.

Klenow

314

dures have been used to limit the action of mutagens

they require numerous

laborious

to specific

ing biochemical gapped DNA.

for the preparation

DNA

segments

Hirose et al., 1982; Kadonaga

(Shortle

et al.,

1980;

et al., 1985). A major

complication of these procedures is that multiple nucleotide substitutions tend to occur under the conditions could

of heavy

be overcome

restricted respect,

mutagenesis.

This

if the target

gap misrepair

1982; Abarzua

of mutagenesis

mutagenesis

and Marians,

We report here the procedure of targeted

problem

to just 1 nt for each DNA molecule.

steps

misincorporation

and time-consumof the

for a novel method

mutagenesis

in which

linearized plasmid DNA is used as a substrate the misincorporation reaction.

is

for

In this

(Shortle

et al.,

1984; Shortle and Lin,

1985) which utilizes the misincorporation reaction of an excision-resistant nucleotide, dNTP[aS], is an

EXPERIMENTAL

excellent method since the target of this mutagenesis

(a) Outline of the method

resides only at the 3’ termini of gapped circular DNA. Several procedures have been devised to localize gaps within a predetermined region (Shortle et al., 1980; Abarzua and Marians, 1984). However,

The method is illustrated in Fig. 1. A DNA fragment containing a sequence to be mutagenized is inserted into an MCS sequence of a vector plasmid.

restriction enzymes

AND DISCUSSION

w

Pol Ik

-

dNTPbSl

Pollk

,/

/‘sdNTPs Si

/’ without

sequencing Fig. 1. Schematic boxes

indicate

mutagenized protruding

illustration

of targeted

misincorporation

inserts.

The tilled boxes represent

is inserted

into an MCS sequence

ends) and then unidirectionally

An excision-resistant

a-thionucleotide

filled-in with four canonical See EXPERIMENTAL

of a vector.

AND DISCUSSION,

selection Open boxes represent

to be mutagenized.

The plasmid

the vector sequences,

A DNA

is linearized

fragment

into the 3’ terminus

is recircularized. section

Mutant

a, for details.

generated

plasmids

containing

with two adjacent

with Exo III so that the 3’ termini generated

is misincorporated

nt, and the plasmid

/segregation

mutagenesis.

the segment

digested

I igase

-

dotted

cuts (generating

are localized

by sequencing

to be

5’ and 3’

within the target region.

by Exo III. The single-stranded

are isolated

and tilled

a sequence

randomly

region is then selected clones.

315

The plasmid restriction

carrying

the insert

sites in the MCS

is digested

sequence,

at two

which

are

located on the same side of the insert. The digestion generates

a 5’-protruding

or blunt end at the proxi-

mal side and a 3’-protruding end at the distal side. The linearized plasmid is synchronously digested

-

with Exo III (Guo and Wu, 1982) until the region to P-

be mutagenized becomes single-stranded. Since Exo III is specific to double-stranded 3’ termini, digestion occurs proceeds Perron

only at the recessive in the direction et

al.,

1985).

or blunt

3’ end and

of the insert

(Yanisch-

A

non-complementary

dNTP[crS] is misincorporated into the 3’ termini generated by Exo III by the action of PolIk in the absence of the other three nucleotides. Since the cr-thiophosphate nucleotide analogue is resistant to the 3’-5’ exonuclease activity of PolIk, its misincorporation into the 3’ termini is irreversible. Although the rate of misincorporation seems to be slow, misincorporation events may eventually accumulate so that almost all the recessive 3’ ends have mismatched nucleotides (Shortle et al., 1982). After the removal of free dNTP[aS], the single-stranded region is filled-in with 4 canonical nt by the action of PolIk. Then, the distal terminus (3’-protruding end) is converted to a blunt end using Sl nuclease. DNA is recircularized with T4 DNA ligase and introduced into E. coli. After segregation of the mutagenized DNA, nucleotide sequences of randomly selected clones are determined by the dideoxy chain-terminating method (Sanger et al., 1977) using universal primer that hybridizes near the MCS sequence. of plasmids

In the study of Ml RNA, the catalytic subunit of P (Guerrier-Takada et al., 1983) encoded by the rnpB gene (Sakamoto et al., 1983a), we found that a nucleotide substitution at nt 329 from the 5’ end of the RNA greatly reduces its catalytic activity (Shiraishi and Shimura, 1986). We applied our procedure to mutagenize the region around nt 329 of the rnpB gene. A plasmid, pPR4273 (Shiraishi and Shimura, 1986) which carries the gene for Ml RNA (rnpB) was digested with SnaBI + PvuII (Fig. 2A). A SnaBI-PvuII 1.7-kb fragment, which carries the coding sequence of M 1 RNA, and a SnaBI O.l-kb fragment, which contains the transcription terminator,

E. coli RNase

OrI ,i

PSW Fig. 2. Construction

of pSW plasmid.

Pv, PvuII; flanking

S, SnuBI.

regions.

(A) Physical

map of the rnpB gene and its

The promoter

pSW. HindIII-linkers contains

containing (Melton carries

and terminator

by p and t, respectively. were attached

the transcription

fragment

was inserted sequence

sequences

of the

(B) Structure

of

to a SnuBI O.l-kb fragment terminator,

and

the linker-

into the Hind111 site of pSP65

et al., 1984). A PvuII-SnaBI the coding

restriction

E, EcoRI; K, KpnI; Ps, PstI;

rnpB gene are indicated which

The following

B,BanII;

sites are shown: A,AccI;

1.7-kb fragment

of Ml RNA was inserted

which into the

XbaI site ofthis plasmid after filling-in the cohesive ends. A KpnI 0.62-kb fragment plasmid

was deleted from this plasmid

was designated

introduced

pSW. The transcription

to avoid possible

tion by the RNA transcribed rnpB gene (Sakamoto binding

interference

and the resulting terminator

from plasmid

from the strong

promoter

was

replicaof the

et al., 1983b). The wavy line represents

site of a sequencing

the

primer.

were separately inserted into the X&I site and the Hind111 site of pSP65 (Melton et al., 1984), respectively. The resulting plasmid was designated pSW (Fig. 2B). (c) Exo III digestion to localize the 3’ ends of DNA Plasmid

(b) Construction

-A

t

pSW was cleaved

at the AccI and PstI

sites located in the MCS sequence of the vector (Fig. 2B). The linearized plasmid (12.5 pg, about 5 pmol) was digested with 120 units of Exo III (Takara Shuzo Co.) at 23 “C in 250 ~1 of 66 mM Tris * HCl (pH 8.0) 90 mM NaCl, 5 mM MgCl,, 10 mM dithiothreitol. Aliquots (20 ~1) were removed at 1-min intervals. Under these conditions, digestion proceeded at a rate of approx. 10 nt/min, consistent with the results of Guo and Wu (1982) from the 3 ’ termini generated by AccI toward the coding region of Ml RNA. After digestion with Exo III, distribution of the 3’ termini generated by the nuclease was determined (Fig. 3). In the sample digested for 6 min, the 3’ termini are distributed almost symmetrically around nt 329, the position which is in the center of the target region in the rnpB gene.

316

(min) G A TC o I z

3

4

6

6

7 6 6

101112 -392 -384 i373,375,376 -367 :jg363 -348 iEtE:~.343.345

,.\..s

m

~325.326 -321

&ma?Ju

*\_

-315

u~us--

-308

.‘,xaWW~

*

*

-301 -298 -290 -283 -279

-265

-253 -247 -242 -238

Fig. 3. Time course ofExo III digestion was incubated

ofpSW

linearized

(5 PCi), and PolIk (0.2 unit). After the incubation, to generate

homogeneous

and 2 ~1 of the mixture

5’ termini.

Six microliters

were electrophoresed

end of the MI RNA sequence by the method

withAcc1

at 23°C for 8 min in 4 ~1 of a buffer containing

and Purl. A portion ofDNA

7 mM Tris

1 ~1 of Ban11 (1 unit) was added, of 95 “/, formamide,

on a 5% polyacrylamide

(0.2 pg) recovered

from Exo III digestion

HCl (pH 8.0), 20 mM NaCl, 7 mM MgCl,, and the mixture

0.1 “/, xylene cyanol,

gel containing

was incubated

and 0.1 y0 bromophenol

8.3 M urea. Positions

[a-32P]dCTP

for 30 min at 37°C blue were added,

of C-residues

are shown on the right side of the figure (Ml RNA is 377 nt long). The sequence

ladder

from the 5’

was generated

of Guo and Wu (1982).

(d) Misincorporation of a non-complementary nucleotide and elongation from mismatched termini It has been shown that each of the four dNTP[crS]‘s can be used effectively for the misincorporation reaction (Abarzua and Marians, 1984; Shortle and Lin, 1985). However, since the proportion of uridine residues in Ml RNA is low (15.1%) (Sakamoto et al., 1983a), dTTP[srS] was used for mutagenesis in the present study to make the nucleotide substitution most effective. A portion of

the DNA (0.2 pg) recovered from Exo III digestion was incubated for 12 h at 25°C in 30 ~1 of 130 mM Hepes-NaOH (pH 7.6) 0.2 mM MnCl,, 2 mM 2-mercaptoethanol, bovine serum albumin (0.1 mg/ml), 0.1 mM dTTP[ CtS] (Pharmacia), and PolIk (Takara Shuzo Co., 0.5 unit). After phenol extraction and ethanol precipitation, the precipitate was dissolved in 20 ~1 of a buffer containing 60 mM Tris . HCl (pH 8.0) 5 mM MgCl,, 0.2 mM MnCl, bovine serum albumin (0.1 mg/ml), 20 mM 2-mercaptoethanol, and four dNTPs (0.2 mM each). PolIk

317

(0.5 unit) was added, and the mixture was incubated

(f) Distribution

of the nucleotide

Distribution

of the nucleotide

substitutions

at 16’ C for 30 min. It is worth noting that compared to the conditions for gap misrepair mutagenesis performed by Shortle and Lin (1985), the incubation

shown

contained

single-nt

temperature of the filling reaction was raised from 0’ C to 16°C to accelerate the primer-extension

that all possible

reaction,

to T) were

and the concentration

of manganese

ion

was reduced by ‘15 to avoid the introduction of deletions and multiple nucleotide substitutions encountered

by the previous

authors.

remaining

of plasmid

and recovery

of

mutants The DNA recovered from the filling reaction was incubated at 30°C for 20 min in 40 ~1 of a buffer containing 30 mM sodium acetate (pH 4.6), 5 mM ZnSO,, 300 mM NaCl, S 1 nuclease (Takara Shuzo Co., 10 units). S 1 nuclease is ineffective in digesting single-nt mismatches under these conditions. After phenol extraction, DNA was precipitated with isopropanol. Then, the termini of the DNA were repaired with PolIk, and the DNA was recircularized with T4 DNA ligase. The circularized DNA was used to transform E. cofi DH5 (recA ‘) to allow segregation of mutant plasmid from wild type. Segregated plasmid was extracted from a pool of transformants (about 480 independent colonies) and used for the second transformation of strain DHl (recA -). Plasmids were prepared from 72 randomly selected transformants, and the nucleotide sequence of each plasmid was determined by the chain-termination method (Sanger et al., 1977). Among the 72 transformants, 27 (37.5%) contained rying nucleotide substitutions.

u

plasmids

u

U (2)

substitutions.

substitutions

encountered

mutant

obtained,

is 26

It is worth noting

(G to T, A to T and C in these

contained

mutants.

a double substitution

The at

nt 3 16 and 3 17 (Fig. 4). In this mutant, the nucleotide sequence 5’-GATT-3’ spanning the nt positions 316-319

(d) Recircularization

substitutions

in Fig. 4. Of the 27 mutants

in

5’-TTTT-3’.

the

m@

It is possible

gene

is

changed

that the second

to mis-

incorporation at nt position 317 was positively affected by the nature of the nucleotide at the adjacent downstream positions (318 and 319). Such a phenomenon, ‘pull-through’ misincorporation, has been observed in the misincorporation reaction of reverse transcriptase (Skinner and Eperon, 1986). In the mutants we isolated, substitutions are distributed almost randomly between nt 295 and 349. We noticed, however, that there seem to be two mutational hot spots at nt positions 316 and 344 (Fig. 4). Presumably, these hot spots would reflect the positions where the Exo III reaction pauses along the DNA strand. In Fig. 3, there are a few bands which remain unchanged for prolonged periods of time during the Exo III digestion (e.g., a band at nt position 315 and triple bands at nt 342, 343, and 345). There is good correspondence between the positions of these bands and the observed hot spots. The reason the Exo III reaction pauses is not known. No special sequences are apparent around these sites. One possibility is that some secondary or tertiary structure(s) of the single-stranded region produced by the enzymatic reaction may interfere with the interaction of Exo III and DNA.

car-

uuu i-7)

It

I

I

t t i tt1 tt t ....CCCGGGUAGGCU~~~UUGAGCCAGUGAGCGAUU~~~UGGCCUAGAUGAAUGACU~~~CCACGACA~~~ACCC.... 330

310

290

11 uu

Fig. 4. Distribution positions

of nucleotide

316 and 317 is shown

the 5’ end of the RNA. Numbers

substitutions.

Substitutions

by a bracket. Numbers in parentheses

are mapped

below the sequence

show the numbers

on the sequence

of Ml RNA. A double

are the nt positions

of clones isolated

(aligned

independently.

substitution

with the second

at nt

digit) from

318

(g) Conclusions

ACKNOWLEDGEMENTS

It has previously been shown that gap misrepair mutagenesis is effective for obtaining single-nt sub-

We are grateful to Karin Knisely cussions. This work was supported

stitutions

Aid for Scientific

molecules Marians,

within

target

(Shortle

regions

et al.,

of given

1982;

DNA

Abarzua

and

Education,

Research

Science,

for helpful disby a Grant-in-

from the Ministry

and Culture

of

of Japan.

1984; Shortle and Lin, 1985). The salient

feature of the procedures of gap misrepair mutagenesis reported thus far is that the misincorporation reactions

of

excision-resistant

nucleotides

were

directed at the target regions which were, in all cases, the 3’ termini of gapped circular DNAs. However, the construction

of gapped

circular

DNA

requires

some elaborate and cumbersome manipulations, such as heteroduplex formation or the formation of D-loops mediated by the recA protein and the purification of the gapped circular molecules by chromatography or density-gradient centrifugation (Shortle et al., 1980; Abarzua and Marians, 1984). In addition, the presence of appropriate restriction site(s) near the target region was required in the previous procedures. This has been a major drawback of these procedures and has considerably limited the applicability of the method. In contrast to these procedures, the experimental protocol presented in this paper is simple and rapid, and contains no laborious steps. In spite of its simplicity, it is as effective as the previous procedures for isolating localized single-nt substitutions. Above all, the most advantageous feature of our procedure is that it is applicable to any portion of the DNA molecule, regardless of the presence of appropriate restriction site(s) near the target regions. In E. coli, adenine residues in GATC sequences are methylated. Mispaired nucleotides are removed preferentially by the repair system from an unmethylated or newly synthesized strand (Radman and Wagner, 1986). In our procedure, a stretch of DNA is degraded by Exo III and re-synthesized after the misincorporation reaction. If there are any GATC sequences in the DNA portion digested by Exo III, methylation of the DNA with the dam methylasc before transformation would be required to avoid the elimination of the misincorporated nucleotides.

REFERENCES Abarzua,

P. and Marians,

isolation

of random

frequency. Botstein,

K.J.: Enzymatic

single-base

techniques

substitutions

Proc. Natl. Acad. Sci. USA 81 (1984) 2030-2034.

D. and Shortle,

vitro mutagenesis. Guerrier-Takada,

D.: Strategies

and applications

of in

Science 229 (1985) 1193-1201.

C., Gardiner,

K., Marsh,

Altman,

S.: The RNA moiety ofribonuclease

subunit

of the enzyme.

T., Pace,

N. and

P is the catalytic

Cell 35 (1983) 849-857.

Guo, L. and Wu, R.: New rapid methods based

for the

in vitro at high

on exonuclease

III digestion

for DNA sequencing

followed

by repair

syn-

thesis. Nucl. Acids Res. 10 (1982) 2065-2084. Hirose, S., Takeuchi,

Y. and Suzuki, Y.: In vitro characterization

of the Iibroin gene promoter tution

mutants.

Proc.

by the use of single-base

Nat].

Acad.

Sci. USA

substi-

79 (1982)

7258-1262. Kadonaga,

J.T. and Knowles,

for chemical

J.R.: A simple and efficient method

mutagenesis

ofDNA.

Nucl. Acids Res. 13 (1985)

1733-1745. Melton,

D.A., Krieg, P.A., Rebagliati,

M.R., Maniatis,

K. and Green, M.R.: Efficient in vitro synthesis active RNA and RNA hybridization containing

a bacteriophage

probes

SP6 promoter.

T., Zinn,

ofbiologically from plasmids

Nucl. Acids Res.

12 (1984) 7035-7056. Radman,

M.

Escherichia

Sakamoto,

and

Wagner

coli. Annu.

H., Kimura,

otide sequence P from

N., Nagawa,

repair

in

Y.: Nucleof RNase

a temperature-sensitive

transcription

Mismatch

20 (1986) 523-538.

F. and Shimura,

H., Kimura,

ponent

R.E.:

and stability ofthe RNA component

Acids Res. 11 (1983a) Sakamoto,

Jr.,

Rev. Genet.

of E. coii. Nucl.

8237-8251. N. and

products

of RNase

mutant Shimura,

Y.: Processing

of the gene encoding

P. Proc. Nat]. Acad.

of

the RNA com-

Sci. USA 80 (1983b)

6187-6191. Sanger, F., Nicklen,

S. and Coulson,

chain-terminating

inhibitors.

A.R.: DNA sequencing

with

Proc. Nat]. Acad. Sci. USA 74

(1977) 5463-5461. Shiraishi,

H. and Shimura,

functions

Y.: Mutations

of the RNA component

affecting

of RNase

two distinct

P. EMBO J. 5

(1986) 3673-3679. Shortlc.

D. and

nuclease: sors

Lin, B.: Genetic

identification

of nuclcase-minus

analysis

of three intragenic mutations.

of staphylococcal ‘global’ suppres-

Genetics

110 (1985)

539-555. Shortle,

D., Koshland,

Segment-directed

D., Weinstock, mutagenesis:

G.M. and Botstein,

Construction

D.:

in vitro ofpoint

319

mutations

limited to a small predetermined

lar DNA

molecule.

Proc. Natl. Acad.

region of a circu-

Sci. USA

77 (1980)

S.J. and Botstein,

D.: Gap

D., Grisafi,

P., Benkovic,

misrepair

mutagenesis:

transition,

transversion,

Proc. Natl. Acad. Skinner,

J.A.

of template

and

efficient

site-directed

and frameshit?

induction

mutations

of

and substrate

C., Vieira, J. and Messing,

phage cloning vectors of the

Ml3mp18

in vitro.

103-l 19.

by AMV

Communicated

and host strains:

and

pUC19

Sci. USA 79 (1982) 1588-1592. Eperon,

reverse transcriptase

I.C.:

Misincorporation

shows strong dependence

on the combi-

nucleotides.

Nucl. Acids

Res. 17 (1986) 6945-6964. Yanisch-Perron,

5375-5379. Shortle,

nation

by H. Yoshikawa.

J.: Improved

nucleotide

vectors.

Gene

Ml3

sequences 33 (1985)