Volume 17 Number 16 1989 Nucleic Acids Research

Voue1;ubr1 99NcecAisRsac

A simple method for site-directed mutagenesis using the polymerase chain reaction

Anne Hemsley+, Norman Arnheim, Michael Dennis Toney', Gino Cortopassi and David J.Galas*

Department of Molecular Biology, University of Southern California, Los Angeles, CA 90089-1340 and
'Department of Biochemistry, University of California, Berkeley, CA 94720, USA

Received June 5, 1989; Revised and Accepted July 12, 1989

ABSTRACT
We have developed a general and simple method for directing specific sequence changes in a plasmid
using primed amplification by the polymerase chain reaction (PCR). The method is based on the
amplification of the entire plasmid using primers that include the desired changes. The method is
rapid, simple in its execution, and requires only minute amounts of plasmid template DNA. It is
significant that there are no special requirements for appropriately placed restriction sites in the sequence
to be manipulated. In our system the yield of transformants was high and the fraction of them harboring
plasmids with only the desired change was consistently about 80%. The generality of the method
should make it useful for the direct alteration of most cloned genes. The only limitation may be
the total length of the plasmid to be manipulated. During the study we found that the Taq DNA
polymerase used for PCR adds on a single extra base (usually an A) at the end of a large fraction
of the newly synthesized chains. These had to be removed by the Klenow frgment of DNA polymerase
to insure restoration of the gene sequence.

INTRODUCTION
Site-directed mutagenesis is a powerful method that is commonly used in several areas
of molecular biology and biochemistry. Early methods for site-directed mutagenesis using
single-stranded plasmids (or segments of plasmids ) gave low efficiencies of obtaining the
desired mutant sequences (1). Improvements which employ selection against the wild-type
sequence (2,3) are costly in time and effort taken to isolate mutants. The development
of the polymerase chain reaction (PCR) (4,5,6) however, has led to new approaches to
site-directed mutagenesis (7,8,9). These methods use primers which contain one or several
bases which differ from the wild-type sequence, which, after PCR, are then incorporated
into the PCR product. This altered sequence is then cleaved with restriction enzymes and
inserted in place of the wild-type sequence. This is a relatively time-consuming approach,
and relies on the presence of conveniently located restriction sites flanking the wild-type
sequence being mutated.
We present here a rapid approach to PCR based site-directed mutagenesis, using an
adaptation of inverse PCR (10,11), whereby an entire circular plasmid is amplified and
a mutation inserted anywhere in the plasmid with no requirement for convenient restriction
enzyme sites. In this technique, the initial rounds of amplification are directed by two primers
located 'back-to-back' on the opposing DNA strands (see Figure 1.). One primer contains
one or a few mismatches which will generate the site-directed mutation. The first cycle
of PCR generates linear plasmid molecules, from one or both circular template strands.
Subsequent rounds of PCR further amplify the linear plasmid sequence generated by the
first PCR cycle, which is then isolated, ligated and used for transformation. This method

© IRL Press 6545

Nucleic Acids Research

primer 1 lacZ
CTTAACACTCGCCTATTOTTA4AGTGTGTCCTTTGTCGATACTGOTACTAAT W COTAAG
GAATTGTGAGCOGATAACAATTTCACACAGGAAACAGCTATOACCATGATTA C GCATTC
Met Thr Met ili Thre
primer 2

4.
lacl lcZ

|pBSIISK+

I first rounds of PCR

produce linear template

lacc 4 lacZ
primer 1 primtr 2
Figure 1.Schematic of the inverse PCR procedure. The expanded sequence shows the 'back-to-back' positions
of the two primers with respect to each other and the site of the introduced mutation (deletion of the boxed base
pair) in the coding sequence of the lacZ alpha peptide.

has the advantage of introducing mutations at any desired site, independent of whether
restriction sites are present in the surrounding sequence. In addition, the concentration
of the initial wild-type template after generation of the mutant plasmid PCR product is
negligible so that very few wild-type clones would be expected. We used the plasmid
pBluescript II SK+ as a model system, since it offers a convenient visual detection system
to monitor introduced changes in the alpha peptide sequence of lacZ. (Fig. 1.) Figure 2
shows the method in schematic form. The mutation we chose to introduce into the alpha
peptide gene resulted in a frameshift (single base-pair deletion).
MATERIALS AND METHODS
Primer synthesis
Synthetic oligonucleotide primers complementary to segments of the lacZ operator region
and alpha peptide coding sequence were prepared on an Applied Biosystems 381A DNA
6546
 Nucleic Acids Research

PCR

Kienow Treatment end-flushing

Isolation of PCR product

5' end-phosphorylation

Electrophoretic isolation of mutant plasmid DNA

Ligation and transformation

Figure 2. Schematic summary of the steps involved in the construction of mutants.

synthesizer. Primers were constructed so as to lie 'back to back' on the duplex, with 5'
ends apposing and 3' ends oriented for extension in opposite orientations around the plasmid
circle (see Fig. 1).
A pair of primers with sequences exactly complementary to the pBluescript II SK+
sequence were synthesized as a control to confirm that the amplification process was able
to faithfully regenerate the original sequence (designated wild-type primers, sequences
5' TGAAATTGTTATCCGCTCAC 3' and 5'CACAGGAAACAGCTATGACCAT-
GATTACGC 3'). In the experimental primer set, (designated mutant primers, sequences
5' TGAAATTGTTATCCGCTCAC 3' and 5' CACAGGAAACAGCTATGACCA-
TGATTAGC 3'), one primer has a deletion corresponding to the 14th nucleotide in the
coding region of the lacZ alpha peptide, designed to eliminate B-galactosidase activity.
Temnplate preparation
The DNA used as template for the PCR reaction was extracted from E. coli strain JM109
transformed with pBluescript II SK+ (Stratagene stock preparation). Plasmid DNA was
isolated by a standard aLkaline lysis mini-prep method (12).
PCR conditions
Template DNA (10 fmol) and primer sets (1 izM each) were incubated in a Perkin Elmer
Cetus Thermal Cycler in 100 jl reaction volumes containing 100 mM Tris-HCl pH 8.3,
500 mM KCl, 2.5 mM MgCl2, 100 ytg/ml gelatin, 0.2 mM of each dNTP and 2 units
Taq polymerase (Perkin Elmer-Cetus)(6). The amplification proceeded through a cycle
of denaturation at 940 C (1 min), annealing at 450 C (1 min) and primer extension at
720 C (12 min) for a total of 25 cycles. Shorter extension times gave poor results for
full length amplification.
Klenow treatment end-flushing
Ten microliter portions of the samples generated by PCR were analysed on a vertical agarose
gel (1.2%) to determine the efficiency of the amplification. The remainder (90 1l) from
3-4 samples was pooled and the four dNTP'S were added to a final concentration of
250 ztM each. Klenow fragment of E.coli polymerase I was added (20units) and the reaction
mix was incubated at 370 C for 30 min. This step was necessary to avoid added basepairs
at the ligation point (see results).
DNA purification and 5' end-phosphorylation
Double stranded DNA was isolated from the PCR reaction mix using the standard protocol
for Geneclean DNA purification (Geneclean ., BIO 101 Inc.) and phosphorylated at the
6547
Nucleic Acids Research

) 4 5 6:

kb

05
I
....) D.8...

D:. .OR
2 03

0 ,5.+

Figure 3. Agarose gel electrophoresis of PCR amplification products (10,1 of 100ulO reaction mixes) using wild-
type primers to regenerate unchanged pBluescript (lanes 3-5) and mutant primers to amplify plasmid DNA
containing a frame-shift within the alpha-peptide of the lacZ gene (lanes 6-8). Lane 2 shows an equivalent amount
of a control product, where the same amount of template DNA was used to amplify a 1kb fragment. PCR product
sizes are assessed against lkb DNA Ladder(BRL) (lane 1).

5' end by incubation with 15-20 units of T4 polynucleotide kinase (Amersham) in a 25

jil reaction volume containing 66 mM Tris-HCl, pH 7.6, 1 mM ATP, 1 mM spermidine-
HCl, 10 mM MgCl2, 10 mM DTT, 0.2 mg/ml BSA (13).
Isolation and ligation of the plasmid DNA
The phosphorylated DNA samples generated from the PCR reactions were then subjected
to electrophoresis in a low melting temperature agarose gel (1% Seaplaque agarose). The
ethidium bromide-stained bands of DNA corresponding to a size of 3kb were excised,
heated to 65°C, and 6 yd samples were added to standard blunt-end ligation buffer and
T4 DNA ligase (approximately 12 units, BRL) in a final volume of 10 1l. The ligation
mixes were incubated at 25°C for time periods ranging from 8 hours to overnight.
Transformation
The ligation mix was reheated to 65°C and 250 A1 of freshly prepared competent cells
(strain NM522) were transformed by a standard procedure(14).
RESULTS
Plasmid amplification
To test the method, we used a small (2964 bp) cloning vector carrying the alpha peptide
gene segment of B-galactosidase, a colorimetric indicator gene. We modified the usual
PCR protocol as shown in Materials and Methods. Successful PCR amplification of the
6548
 Nucleic Acids Research

Table 1. Phenotypic and sequence analysis of the results of the directed mutagenesis protocol using wild-type
or mutant primers. OK indicates that the sequence observed was the correct pBluescript sequence.
Primers White colonies Blue colonies
Wild type 79 ( 7%) 1070 (93%)
Mutant 1326 (97%) 42 ( 3%)
Sequenced: 29 7
Mutation site 29 mutant 5 wt.
2 +G
Ligation site 24 OK 6 OK
4 del., 1 del.,
1 +A

entire pBluescript II SK+ plasmid is shown in Figure 3. The other bands in each lane
were not characterized further, and undoubtedly resulted from the low temperature of primer
annealing during PCR. The second lane of the gel shows amplification of a lkb segment
(using a different primer pair) from the same plasmid as a control, giving some indication
of the relative efficiency of amplification.
The protocol used for the experiments described here is outlined in the diagram in the
introduction with the details provided in the Materials and Methods. In all cases it is
necessary to begin with only minute amounts of plasmid DNA prepared by a frequently
used rapid 'mini-prep' procedure.
Colony fonnation
The re-circularized products of the amplified plasmids were transformed by a standard
protocol with good yield of colonies. The control experiment, using wild-type primers
to regenerate the original pBluescript plasmid generated 1149 colonies, 1070 (93%) of
which were blue, suggesting an unaltered lacZ sequence. Mutant primer amplification
generated 1368 colonies, 1326 (97%) of which were white, the preliminary diagnostic
criterion for successful insertion of the frameshift mutation into the lacZ alpha peptide.
(See Table 1 ).
Sequence data
The nucleotide sequence of the region shown in the inset of Figure 1 was examined for
36 colonies generated by using the mutant primers. Both white (29 colonies) and blue (7
colonies) colonies were checked to assess the spectrum of alterations in nucleotide sequence
which led to each phenotype. It was expected that only alterations introduced in the primer
would lead to a phenotypic change in the colony, since the point of ligation of the plasmid
was designed to lie outside of the coding region of the alpha peptide of lacZ in order that
changes in sequence at the ligation point could easily be distinguished from changes at
the intended position of mutation. Results of the sequence analysis are shown in Table
1. Of the 36 plasmid sequences examined, 30 (83%) restored the wild-type sequence at
the ligation point. Five out of the other six plasmids had small (1-8 bp) deletions at the
site of the ligation and one had an additional adenine (A) inserted at the ligation point.
One possible explanation of the small deletions is that they are the result of incorporation
into the PCR product of a primer which was not synthesized to full length, as the primers
used were not purified after synthesis.
In experiments without the Klenow fragment end-flushing step, the insertion of an extra
nucleotide at the ligation position was seen to occur at a much higher frequency; 73%
of the 24 sequences examined (data not shown). It is interesting that the extra base pair
6549
Nucleic Acids Research

was an A:T in all cases but one. This added base-pair at the ligation point was most likely
due to the non-templated addition of a base to the ends of the synthesized DNA by the
Taq polymerase (15). In preliminary experiments, using a different primer-template system,
we found that a large fraction of the PCR products had a single extra base present that
was removed when the double-stranded product was treated with Klenow fragment as
described in the present protocol (data not shown).
All 29 white colonies produced using the mutant primers had the intended mutant sequence
correctly inserted into the early coding region of the lacZ alpha peptide. The blue colonies
had two sequence variations; five of the seven were wild-type sequence, of unknown origin,
while the other two had an additional nucleotide inserted in the primer region, restoring
the reading-frame to that of wild-type. This suggests a mutation made by slipped mispairing
by the E.coli DNA polymerase, since a run of two guanine nucleotides (G's) was expanded
to three.
DISCUSSION
In this paper, we demonstrate that it is possible to generate linear fragments of at least
3 kb in length, constituting an entire plasmid, using intact circular plasmid DNA as a
template. We have further demonstrated that it is a simple matter then to site-direct changes
in the sequence. The level of efficiency of amplification is less than that observed when
smaller fragments are amplified, however sufficient DNA is synthesized for subsequent
ligation and transformation procedures. Based on the data obtained from analysis of 29
candidate colonies, insertion of the desired mutation and exact reconstitution of the plasmid,
occurs at a frequency of 82%. The desired mutation is inserted at a frequency of about
95%, so that improvements in procedures to reconstitute the ligation site might enhance
the overall efficiency.
This procedure is consistently efficient, rapid (mutant colonies can be generated in two
to three days) and generates substantial numbers of transformants. An additional advantage
is that it does not require highly purified DNA as the template for the PCR reaction and
there are no special requirements for the plasmid preparation. It should also be possible
to use plasmids liberated from small numbers of bacterial cells taken directly from plates,
which would further shorten the procedure. We suspect that open circular plasmid provides
the best template for inverse PCR, since it was observed that template derived from
minipreps was actually a better substrate for PCR amplification than was highly purified,
supercoiled plasmid (data not shown).
The relatively long PCR extension times we used for efficient elongation of the 3 kb
DNA fragment (12 min.) could have affected the frequency of non-template-directed
nucleotide addition at the 3' ends of the molecules. Removal of these nucleotides can be
effected by a brief treatment with Klenow treatment of the PCR reaction under the conditions
described in the Materials and Methods section.
The mutations generated in this study were localized to the lacZ gene to permit a simple
visual assay of successful mutagenesis. The independence of this method on the proximity
of unique restriction sites makes it a convenient and rapid method whereby mutations may
be introduced at any site within any circular DNA molecule at an extremely high mutant
to wild-type ratio. At the present time, it is c6nceivable that plasmids up to lOkb in length
could be used in our system (16).

6550
 Nucleic Acids Research

ACKNOWLEDGEMENTS.
This work was supported by NIH grants A119036 to D.J.G. and GM36745 to N.A.
*To whom correspondence should be addressed
+Present address: Department of Biological Sciences, University of California, Santa Barbara, CA 93106,
USA

REFERENCES.
1. Derbyshire, K.M., Salvo, J.J., Grindley, N.D.F. (1986) Gene 46, 145-152.
2. Kramer, W., Drutsa, V., Jansen, H.W., Kramer, B., Pflugfelder, M., Fritz,H-J. (1984) Nuc. Acids Res.
12, 9441-9456.
3. Taylor, J.W., Ott, J., Eckstein, F. (1985) Nuc. Acids Res. 13, 8765-8785.
4. Saiki, R.K., Gelfland,D.H., Stoffel, S., Scharf,S.J., Higuchi, R., Horn,G.T., Erlich, (1988) Science 239,
487-489.
5. Saiki, R.K., Scharf, S., Faloona, F., Mullis, K.B., Horn, G., Erlich, H.A., Arnheim, (1985) Science 230,
1350-1354.
6. Mullis, K.B. and Fakoona, F.A. (1987) Methods in Enzymology 155, 335-350.
7. Higuchi,R., Krummel, B., Saiki, R.K. (1988) Nuc. Acids Res. 16, 7351-7367.
8. Kadowaki, H., Kadowaki, T., Wondisford, F.E., Taylor, S.I. (1989) Gene 76, 161-166.
9. Valette, F., Mege, E., Reiss, A., Adesink, M. (1989) Nuc. Acids Res. 17, 723-733.
10. Triglia, T., Gregory Peterson, M., Kemp, D.J. (1988) Nuc. Acids Res.16, 8186.
11. Ochman, H., Gerber, A.S., Hard, D.L.(1988) Genetics 120, p621-623.
12. Birnboim, H.C. and Doly, J. (1979) Nuc. Acids Res. 7, 1513-1518.
13. Wu, W.R., Wu, T., Anurhada, R. (1987) Methods in Enzymology 152, 343-350.
14. Maniatis, T., Fritsch, E.F., Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring
Harbor Laboratory, Cold Spring Harbor, New York.
15. Clark, J.M. (1988) Nuc. Acids Res. 16, 9677-9686.
16. Jeffreys, A.J. Wilson, V. Neumann,R. and Keyte, J. (1988) Nuc. Acids Res. 16, 10953-10971.

This article, submitted on disc, has been automatically

converted into this typeset format by the publisher.

6551