Gene, 121 (1992) 17-24 0378-I 119/92,1$05.00

GENE

17

06718

A simple and efficient method for the oligodeoxyribonucleotide-directed mutagenesis of double-stranded plasmid DNA (Nucleotide

sequence;

uracil N-glycosylase;

alkali-denatured

DNA;

nitrocellulose;

T7 DNA

polymerase;

large plasmids)

Rudolf Jung, M. Paul Scott, Luiz 0. Oliveira and Niels C. Nielsen United States Department

of‘Agriculture.Agricultural Research Service, Agronomy Department, Purdue University. West Lafuyette. IN 47907, USA

Received

2 March

by J. Messing:

1992; Revised/Accepted:

22 June/23

June 1992; Received

at publishers:

3 July 1992

SUMMARY

A method for the oligodeoxyribonucleotide-directed mutagenesis of double-stranded DNA without the necessity for phenotypic selection is described. Plasmids denatured with alkali and purified by adsorption to and elution from nitrocellulose have single-stranded regions where primers can hybridize and serve as templates for a T7 DNA polymerase-catalyzed synthesis of complementary mutant DNA strands. When this procedure was carried out such that the original nonmutant strand contained uracil [method of Kunkel, Proc. Natl. Acad. Sci. USA 82(1985)488-4921, mutation frequencies of between 3O”/b and 40% were obtained. The technique has been used to generate mutant genes in plasmids of a wide variety of sizes. The largest plasmid manipulated and successfully mutagenized was 22 kb. The method is rapid and efficient and is not dependent upon either fl phage vectors or the presence of restriction sites in the vicinity of the sequence targeted for mutation.

INTRODUCTION

Oligo-directed mutagenesis is a powerful tool to investigate associations between molecular structure and function. In addition to probing the regulatory regions of genes, site-directed mutagenesis permits proteins to be changed so that relationships between protein structure and function can be investigated.

Correspondence to; Dr. N. C. Nielsen,

USDA/ARS,

ment, Purdue University, West Lafayette, Tel. (3 17) 494-8090; Fax (3 17) 494-6508.

Agronomy

Abbreviations: bp, base pair(s); cfu, colony-forming units; methylsulfoxide; dNTP, deoxyribonucleoside triphosphate; strand(ed);

IPTG,

isopropyl-P-D-thiogalactopyranoside;

or 1000 bp; nc, nitrocellulose;

nt, nucleotide(s);

cleotide; on’, origin of DNA replication;

Depart-

IN 47907, USA.

DMSO, dids, double

kb, kilobase

oligo, oligodeoxyribonu-

PCR, polymerase

chain reaction;

PolIk, Klenow (large) fragment of E. coli DNA polymerase I; rpm, rotations per minute; ss, single strand(ed); T-DNA, uracil-free DNA; U-DNA, uracil-containing

DNA;

UNG,

chloro-3-indolyl-P-D-galactopyranoside.

uracil N-glycosylase;

XGal,

5-bromo-4-

A number of techniques have been developed that permit the controlled modification of DNA sequences. The early strategies relied on the production of a ss DNA template using an fl phage vector such as Ml3 (Zoller and Smith, 1983). In recent years variations have been developed to increase the yield of mutants. By applying biological selection for the mutant strand of DNA, efficiencies greater than 50% are routinely achieved (Kunkel, 1985; Nakamaye and Eckstein, 1986; Kramer and Fritz, 1987). However, two disadvantages limit the utility of protocols that rely on the use of fl phages. First, these procedures require that the modified DNA be subcloned both prior to and following mutagenesis. The second disadvantage is the instability of fl phages with inserts of more than 2000 bp (Maniatis et al., 1982). Occasionally, smaller inserts also tend to be deleted during phage growth. To deal with the latter of these two problems, phagemids, plasmids that contain the fl- ori to allow ss DNA rescue, are used (Dente et al., 1983). Phagemids contain a ColEl ori and it is used unless there is co-infection with a helper phage. Because the ss phagemids are noninfectious, the phagemids are un-

18

//::,::c[::,.,

able to participate in multiple cycles of amplification, and this results in a much greater stability of long inserts. None-

‘U ,.

theless, the yield of ss DNA from some phagemids is low after superinfection with helper phage, probably because secondary structures in the inserts interfere with amplification.

uracil-containing Boxes ung

DNA,

Methods:

for mutation.

solid biack

with an arrowhead

mutation.

of the mutagenesis

symbolize

3s

described

in Maniatis

dues (T) are replaced Immediately

primers

that contain

from BioRad,

et al. (1982). Occasional

devices ~Bioallalytical

the protocol,

Systems,

nitrocellulose

prepare

Inc. West Lafayette, filters, diameter

CA) and

thymine

two centrifuge-filter

considerably

in their ability to release bound

To avoid unacceptable

it is recommended

3

NC cantrifugation (optional)

that the filter material

mutagenic

primer

synthws of T-DNA (T7 ONA polymerase)

enrvchment of ds U-ONE-ONA hsteroduplex with NC centrifugation filter (optional)

has been can vary step,

upon release. Rinse

at 5000 rpm in

3

ds mutant

Repeat

Beckman

the elution

solution DNA

contains

and combine denatured

vortex and then incubate

at 37’C

2M for

the eluents

rpm.

from the two steps. This

immediately

for mutagenesis

or stored

frozen for use at a later time. Ahrrenlin,r cmd ~~~r~zerj~ffri~~~reclcfiorrs. Defilter-pursued

DNA (8 ~1) is added to an Eppendorf

I ~1 of phosphorylated

mu~genic

tube. To this

primer (I - 10 pmoi, 10-40 times

excess) and 1 $ of 10 x annealing buffer (BioRad. Muta-Gene Kit). The mixture is incubated lo-15 min at 37°C and then chilled on ice. To initiate the polymerization Rad,

Muta-Gene

units/$,

reaction,

Kit; contains

I pl synthesis buffer (Bio1 pl T4 ligase (3

rapidly

add

dNTPs

and ATP)i

BioRad) /l ~1 T7 DNA polymerasc

Gene kit). Vortex the mixture

(0.5 units/PI. BioRad,

and then microfuge

for

Muta-

I s to collect the

solution in the bottom of the tube. Incubate 5 min at O’C. 5 min at 2O’C. and then 30 mm at 37’C. After in~ubati~~n. add 7 !il of 1 M NaCI. vorPipette the reaction

tube. Add 2 pi of freshly prepared

device and

DNA purified from traces of nondenaturcd

and may either be used

tex, and collect the solution

2 mM EDTA,

ONA

first for 2 min at 500 rpm and then for 2 min at 5000

volume of 18 pi in an Eppendorf containing

lnmmplete polymeruation products

25 ~1 of water to the surface of the filter in the centrifuge-filter centrifuge,

Model J21, and store at 3°C until use. To begin the denaturation process. suspend l-5 ug of U-DNA plasmids (0.2-1.0 pmol) in water to a final NaOH

.._,..., heteroduplex

selection agalnst parental DNA (transformation 1nt-3ung+ strain)

add

used be tested to ensure that at

the filters with 500 pi of 5 M NaCl, spin-dry

U-DNA

annealing

natured

DNA under low salt con-

DNA is recovered

filter

4

Maid-

losses of DNA during the filtration

least 50”,, of the bound denatured

4

denatured

double

5 mm (Whatman.

stone, UK or Schleicher & Schuell, Keene, NH). Our experience that different lots of nitrocellulose, even from the same vendor, ditions.

resi-

Denutururion.

IN) with

..~.j.)”

-.”

DNA.

the desired

Richmond,

,. ..

.,. ..

boxes

lines to thynline-containing

by uracil (U) in these DNA molecules.

before beginning

layer of 0.45qn

Dark

denaturation

lines refer to

Plusmidpreparution. Plasmids are grown in the E. coli

dur- host CJ236 (purchased

purified

protocol.

Light stippled

,”

denat”red enrichment ONA of wi,h

are una~railable. In this report an oligo-directed mutagenesis technique is described that involves the direct modification of ds DNA plasmids. The method avoids the introduction of the target sequence into fl phage vectors. Plasmid DNA dena-

targeted

.._

ds U-DNA alkaline

require that appropriate restriction sites be located in the proximity of the nt sequence targeted for mutation. Although PCR-derived protocols are suitable for many applications, their general applicability is limited. Protocols for the routine mutagenesis of plasmids with large inserts

description

“L, .., ....,,J

‘. ..

derived templates. Typically, these protocols involve multiple enzymatic reactions in vitro (Efimov et al., 1985; Foss and McClain, 1987; Sugimoto et al.. 1989; Olsen and Eckstein. 1990; Slilaty et al., 1990; Drutsa and Kaberdin, 1991) or are PCR-derived (Bowman et al., 1990; Higuchi et al., 1988; Kuipers et al., 1991; Shen et al., 1991) and frequently

Fig. 1. Schematic

.

: ;

u

Other methods for oligo-directed mutagenesis have been proposed to circumvent the requirement for fl phage-

refer to the sequence

‘: j: :

:’

new prerinsed

at the bottom

of the tube vviitb a microfuge.

mixture onto a filter-centrifuge

nitrocellulose

filters (see above).

device cquippcd Unprocessed

with

denatured

lo-15 min. After incubation, chill on ice. The following three steps should be completed in sequence as rapidly as possible: (I) Add 10 ul of ice cold

DNA will bind to the nitroccllulose. Centrifuge the solution first for 2 min S&Iat 500 rpm and then 2 min at 5000 rpm. Collect the flow-through

0.9 M Na.acctate

tion that contains

pH 4.8 to the chilled denatured

Remove 2 ul for electrophoretic tralized, denatured Denatured DNA

analysis

(optional).

plasmids

and vortex.

(2) Pipette the neu-

plasmid onto the surface of the cent~fuge-filter device. binds with high affinity to the nitrocellulose, while

nondenatured DNA does not. Centrifuge then at 5000 rpm for 2 min. Discard save for clectropboretic

analysis).

first for 2 min at 500 i-pm, and the Row-through solution (or

Wash the filter-bound

DNA by pass-

ing through 500 ~1 of 0.1 M NaCl at 5000 rpm for 2 min. Discard the wash solution. (3) To elute the DNA from the nitrocellulosc. apply

plasmids

be used to monitor

with a mutant

strand.

the DNA electrophoretically,

Part of the solution can while the remainder

can

either be stored or used for transformation. T~u~.~#~Y~Qz~~~. The solution that contains plasmids (l-10 ~1) with a mutant strand was used to transform 50-200 ~1 of high efficiency competent ung ’ cells (E. cwii TB 1, > 5 x 10"cfuiug pBR322). Sefertion. Colony hybridization with the radiolabelled

mutagenic

primer, nt sequence

analysis

of random

picked WI-

onies. or. if appropriate, restriction analysis may then be applied to screen for mutants (Maniatis et al.. 1982).

19 tured with alkali that contains ss regions in which mutagenic oligos can hybridize serves as the template in our

E. coli TBl

protocol. The mutant strand is synthesized with T7 DNA polymerase. By using alkali-denatured DNA templates that contain uracil (Kunkel, 1985) a strong selection against nonmutagenized DNA can be achieved. This ensures that plasmid DNA with the desired mutations is recovered with

ss U-DNA (uracil-containing DNA). The data also show that the transformation rates were substantially higher when the cells were transformed with ds T- and U-DNA instead

high efficiency. Restriction sites in the proximity of the targeted mutation site are not required, and plasmids as large as 22 kb can be successfully manipulated.

RESULTS

103-times more effec-

(thymine-containing

DNA) than with

of ss U-DNA. The higher rate of transformation with ds U-DNA compared with ss U-DNA was probably due to the repair of apyrimidinic sites that were generated upon removal of uracil by the glycosylase enzyme. When the same ds DNAs were denatured with alkali prior to transformation, the difference in transformation efficiency of ds U-DNA versus ds T-DNA decreased to that observed when ss DNA was used for transformation. This result

AND DISCUSSION

(a) Several steps are required for site-directed

cells were transformed

tively with ss T-DNA

mutation of

ds plasmids Fig. 1 presents a schematic overview of the protocol that was developed. The plasmid to be manipulated is propagated in E. coli CJ236. These cells have a urfg-dutgenotype, and plasmids produced by these cells will contain an occasional uracil in place of a thymine (Kunkel, 1985). After isolation, the uracil-containing plasmids are alkali denatured with 0.2 M NaOH, and the mixture is introduced into a centrifuge-filter device equipped with a double layer of nitrocellulose filters. Those plasmids that are ‘collapsed’ and contain ss DNA regions are retained on the filter, but ds DNA molecules that persist after alkali treatment pass through the filter and are discarded. After eluting the ‘collapsed’ plasmids from the nitrocellulose filter, appropriate oligo primers are added, and a mutant strand is synthesized with T7 DNA polymerase. When the synthesis is complete, the mixture is again introduced into a centrifuge-filter device equipped with two layers of nitrocellulose filters. This time, the ds DNA plasmids with the mutant strand pass through the filter and are collected, while the ss DNA retained on the filter is discarded. The nitrocellulose filtration step improves the recovery rate of mutants. The filtrate enriched for mutant strands is used to transform an ung+ strain of E. coli, a step that selects against uracil-containing parental DNA strands (Kunkel, 1985). Appropriate techniques can then be used to identify plasmids that carry the desired mutations. (b) Denatured uracil-containing ds DNA is inactivated in ung+ cells with the same efficiency as uracil-containing ss DNA The feasibility of using the Kunkel selection method with alkali-denatured DNA was tested by comparing the transformation efficiency of uracil-containing plasmids purified from E. co/i strain CJ236 with the transformation efficiency of DNA purified from ung+ cells. When transformed into cells with proficient UNG enzymatic activity, a certain fraction of DNA strands with uracil is inactivated. Consistent with this mechanism, the data in Table I show that

suggests that many plasmid molecules contained mainly ss DNA after denaturation or were in a ‘collapsed’ configuration. Consistent with this conclusion, alkali-denatured closed circular plasmids migrated similarly to ss DNA plasmids in electrophoretic gels (not shown). The use of denatured plasmids as templates for mutation demanded that the ss DNA configuration remain stable TABLE

I

Transformation

efficiency of U-DNA

and T-DNA

plasmids”

in E. coli

(ung + ) strain TB 1

Transformation pTZ

rate (cfu!ng)

1S-A

pBSKP

pLeMS

(2860 bp)

(2964 bp)

(4758 bp)

T-DNA

8.5 x 10“

n.d.

n.d.

U-DNA

9.5 x 10’

n.d.

nd.

T-DNA

7.4 x 106

6.0 x 10”

1.2 x 10”

U-DNA

6.3 x IO4

8.5 x 10’

4.0 x 10J

ss plasmid

ds plasmid

Alkali-denaturated

plasmid’

T-DNA

1.5 x 10h

1.1 x 10h

1.3 x 10’

U-DNA

7.5 x 10’

1.5 x 10’

5.0 x 10’

Alkali-denatured,

renatured

plasmid‘

T-DNA

3.2 x IO6

1.4 x IO”

5.1 x lo5

U-DNA

9.0 x lo2

1.8 x lo2

1.9 x 10’

’ Plasmid

pTZ18-A

was purchased

from BioRad,

Inc. (Richmond,CA);

plasmid pBSKP was purchased from Stratagene (La Jolla, CA), and the preparation of plasmid pLeMS is described by Scott et al. (1991). The ss phagemid Mutagene

DNA was prepared kit.

as described

in the manual

of the BioRad

h DNA (1 ug) was denatured in 20 nl 0.2 M NaOH/0.2 mM EDTA, neutralized with 10 ~10.9 M Na,acetate pH 4.8, and precipitated with 75 nl ethanol. The precipitate in 10 mM Trisil ’ Denatured nealing

was collected by centrifugation

and suspended

mM EDTA.

DNA (0.5 ug) was incubated

buffer (BioRad,

thesis buffer (BioRad, another 30 min.

Mutagene Mutagene

15 min at 37°C in 10 nl an-

kit). After addition kit), the incubation

of 1 ~1 10 x synwas continued

for

20 long enough for the that mutant strand therefore important renatured rapidly or

oligo primers to interact with them so synthesis could commence. It was to determine whether the plasmids slowly. This was tested by incubating

are grown on plates containing IPTG and XGal. The oligo S-GGGTTTTCCCAGTCACGAC was used as a primer to restore function to the IacZr gene product. The oligo could potentially mediate a mutation from T-C, and mutants would give rise to blue colonies. When the 19-mer was annealed to alkali-denatured pTZ 18-A in the presence of T7 DNA polymerase, about 6% of the total colonies recovered was blue. Although this result already showed that the method could be used to generate site-directed mutations, the yield of transformants was less than expected, or about one-tenth that observed when ss U-DNA was used under exactly the same conditions. An electrophoretic analysis of the denatured plasmid and of the products after the mutagenesis reaction is shown in Fig. 2 and provides at least a partial explanation for the

the denatured plasmids for 15 min at 37 “C in annealing buffer, followed by incubation in polymerase buffer for 30 min at the same temperature. These conditions were chosen because they corresponded to those used in our mutagenesis protocol. The data in Table I show that only a slight increase in transformation efficiency occurred during this interval of incubation. We also compared the electrophoretic behavior of the plasmids before and after incubation and found no detectable difference (data not shown). These results demonstrated that alkali-denatured plasmids renatured slowly and that the ss regions should be available for hybridization

to oligo primers.

low recovery of mutants. Those data also demonstrate some features of the system that must be controlled to optimize the recovery of mutants. Lane 2 reveals that multiple molecular species of the plasmid are recovered following denaturation. Lane 4 demonstrates that only a low percentage of ss DNA was converted to ds DNA during the mutagenesis reaction. This suggests two explanations for the observed recovery rate of mutants: incomplete denaturation of the plasmids and/or poor synthesis of the mutant strand. As shown earlier, selection against U-DNA is

(c) Alkali-denatured U-DNA plasmids support site-directed mutagenesis Table II presents results obtained when plasmid pTZ 18-A was subjected to oligo-directed mutagenesis using alkali-denatured DNA as described above. Plasmid pTZ18-A contains a C+T transversion in IucZa, creating an amber mutation that gives rise to white colonies upon a-complementation when bacteria harboring the plasmids TABLE

II

Mutagenesis” Template

of the uracil-containing

plasmid

with a 19-bp primer h to reverse the amber mutation

pTZ18-A

DNA polymerase

(0.5 pmol)

Primer (pmol)

Transformants”

Mutants”

(cfu,;ug)

(“,,) 55’

ss U-DNA

Tl’

10

1.1 x 10s

B. ss U-DNA

Tl

0

7.5 x IO’

< 0.0 1

T7

IO

5.0 x 10’

0. I

T7

40

4.2 x lo2

T7

0

3.5 x lo2

clh < 0.0 I

T7

40

I.1 x

A.

C. Nondenatured D. Denatured

ds U-DNA

ds U-DNA.

E. Same as D F. Denatured ds U-DNA, with nitrocellulose

no pretreatment nondenaturcd

with nitrocellulose

filter

DNA removed

filter before polymerization

DNA G. Denatured ds U-DNA, nondenatured removed before polymerization/unprocessed denatured

IO2

I5

DNA Tl

40

1.1 x IO’

39

H. Same as G

T7

10

1.

T7

1.0 x lo? 1.5 x 10:

31 8

removed

after polymerization

with nitrocellulose

filter

Same as G

0.5

J. SamcasG K. Same as G

Scquenased T4”

40

2.5 x IO’

II

40

< 0. I

L. Same as G

PolIki

40

6.4 x 10’ 9.0 x lo2

,’ For experimental

conditions

see Fig. 1 legend.

h 5’-GGGTTTTCCCAGTCACGAC; ’ 0.5 units (BioRad Inc., Richmond, ’ 0.5 units of Sequenase, ’

version

point mutation CA).

2.0 (US Biochemical,

is underlined. Cleveland,

OH).

I unit (BioRad).

’ 5 units of PolIk, (New England Biolabs Inc., Beverly, MA). 8 E. coli TBI. h Blue colonies

on XGal plates.

’ After retransformation ’ After retransformation

into E. co/i TBl. four clones out of ten segregated into E. di TB 1, nine clones out of 20 segregated

into blue and white colonies into blue and white colonies

on XGal plates. on XGal plates.

10.1

21

1

M

2

3

4

5

(d) Nitrocellulose

(NC) filtration substantially

improves the

recovery of mutants Two methods were tried to remove nondenatured plasmids after denaturation to reduce the nonmutant background. One involved the use of the restriction enzyme SauIIIA, an enzyme chosen to have a strict requirement

Fig. 2. Electrophoretic

profile of DNA

steps in the mutagenesis

protocol

standards; lane 1, nondenatured closed circular plasmid DNA, denaturation: denatured

obtained

Biolabs Catalogue, 1991). However, treatment of alkalidenatured plasmids with SauIIIA was unsuccessful because both ds DNA and denatured ds DNA were cleaved identically by the enzyme. As a control, ss DNA was treated

at various

in Fig. 1. Lane M, DNA size

pTZ 18-A DNA; oc and ccc, nicked and respectively; lane 2, DNA after alkaline

den1 and denI1, nicked plasmid,

in samples

described

respectively:

and closed

lane 3, denatured

circular

under the same conditions, but only a very small portion was digested (data not shown). This result led us to con-

forms of the

DNA after it has been

clude that the alkali-denatured DNA was not truly ss DNA but was a collapsed plasmid that contained both ss and ds

bound and eluted from nitrocellulose filters; lane 4, DNA after polymerization using template-directed duplex formation with T7 DNA polymerase;

lane 5. same DNA as in lane 4, but after passage

cellulose

to remove unprocessed

mutant-strand-containing,

denatured

DNA.

forms den1 and dcnI1 were identified

(not shown): Nicked and closed circular plasmid purifying

by cutting the appropriate

with Gcneclcan

as described

den11 plasmids circular

plasmid

in Table I and separated comigrated

and closed,

DNA was isolated after

bands from an agarose

DNA, were denatured

ss DNA and denatured ds DNA, but not for nondenatured ds DNA. Table IV shows that this approach can be used to remove nondenatured plasmids from the alkali-denatured mixture. After filtration, the ss DNA is easily eluted using either 10 mM Tris buffer pH 8 or water. Several different polymerases were evaluated for their ability to produce ds plasmids with mutant strands. The results of these experiments are given in Table II. Neither T4 DNA polymerase nor the Klenow fragment of the DNA polymerase I yielded any mutants among the colonies (10”) tested. This probably reflects the lower turnover rate for both of these enzymes in comparison with the very high turnover of T7 DNA polymerase (Tabor and Richardson, 1989). The T7 DNA polymerase was the most effective of those that we tried and has been previously shown to be superior for the mutagenesis of ss DNA (Bebenek and Kunkel, 1989). It polymerizes DNA on templates containing secondary structures where the polymerization with T4 DNA polymerase is inhibited. Native T7 DNA polymerase

gel and plasmid

with NaOH

on a 13” agarose gel. The den1 and

exactly with the isolated,

circular

plas-

as follows

(Bio 101, La Jolla, Ca). Both isolated

forms, as well as linearized

regions (see also legend to Fig. 2). The second method to remove ds DNA from the alkali-denatured mixture of DNA involved the use of nitrocellulose filters. Under conditions of high ionic strength, nitrocellulose has a high affinity for

nitroindicate

nicked, and closed, circular, heteroduplex

mid DNA. The plasmid electrophoresis

through

Arrowheads

plasmid

DNA.

plasmid moved slightly faster than the denatured

denatured,

The linearized,

nicked, denatured

ccc plasmid form. Under

the denaturation conditions used, no detectable circular or linear ssDNA was relcascd from the denatured nicked plasmid DNA.

less efficient for ds DNA (Table I), and plasmids that do not denature well should therefore yield a higher nonmutant background. The nonmutant background will cause a lower apparent mutation rate. We therefore attempted to improve the denaturation step. The data in Table III summarize results when pBSKP plasmids are exposed to several different denaturation conditions. These data show that treatment with 0.2 M NaOH was the most effective means to denature the plasmids we evaluated; however, this may not be true for all plasmids.

TABLE

III

Denaturation

of plasmid

pBSKP

Conditions

Temperature

0.2 M NaOH10.2

37

mM EDTA

30

90

10 10 5

80 80 Full degradation

37

5 10

100

5

0.2 M NaOH/0.2 IO mM Tris.HCI

mM EDTA/IO”; DMSO pH 7.5/l mM EDTA

37 100

10 mM Tris.HCl

pH 7.511 mM EDTA

100

.I Monitored

by electrophoresis

on a I O0 agarose

gel.

plasmid”

80

31

mM EDTA

Denatured

90

glycerol

by 0.2 M NaOHl0.2

Time (min) 10

mM EDTA/IO”,

50”, formamidc

(“C)

20 0.2 M NaOH/0.2

followed

because it has been reported for ds DNA (New England

Partial

degradation

Full degradation IO

(9,)

22 TABLE

IV

Transformation

of unl: + Escherichiu coli TB I cells with uracil-containing

pTZ 18-A. Effect of nitrocellulose Transformation Nondenaturcd

conditions

4 5

6 7 M

8 9 l(

Rate cfu;‘ug

U-DNA

5.5 x 10J

Denatured

U-DNA,

nonfiltered

Denatured

U-DNA,

filtered

9.4 x 10’

Fraction

1, Howthrough

Fraction

2, nitrocellulose-bound

I’ See Fig.

3

filtration .’

in 0.3 M Na’acetatc

I for experimental

and eluted with water

6.0 x 10’ 5.0 x 10’

details

Fig. 3. The effect of UNG

(Form II) does not measurably perform net strand displacement synthesis with ds DNA templates (Lechner et al., 1983). We found, however, that the T7 DNA polymerase was able to convert up to 5’4 of the denatured ds DNA into the polymerized nondenatured product (Fig. 2). Probably only denatured DNA molecules with a low content of ds regions were polymerized completely. For comparison, SO-90% of ss DNA is converted under the same reaction conditions (data not shown). Because the undesirably high nonmutant background was considered to be attributable to unconverted DNA present in the mixture after polymerization, a second nitrocellulose filtration step was introduced into the protocol. As illustrated in Fig. 1, nonconverted denatured DNA was selectively bound to the filters, and the converted ds DNA was recovered in the flowthrough fraction from the filtration step. Although this step led to an improvement in the percentage of mutants among the transformants that were obtained, Table IV shows that a background of nonmutants was not eliminated. We concluded that the background was due either to ds U-DNAs that were repaired following transformation of the ung+ cells or to plasmids that lacked uracil.

V. Lanes 8-14

show thymine-containing

denatured

3 and 10, denatured

DNA

lanes 4 and 11, nondenatured treated

For reaction pBSKP

pBSKP

with UNG;

UNG and incubated bated at 95’C:

treated

conditions

plasmid

plasmid

DNA; lanes 2 and 9, denatured

see Table

DNA, while lanes

DNA.

Lanca

1 and 8,

DNA treated with UNG; lanes

with UNG

and incubated

at 95’C;

DNA; lanes 5 and 12, nondenaturcd

lanes 6 and

13, nondenatured

DNA

DNA

treated

at 95°C; lanes 7 and 14, nondenatured

with

DNA incu-

lane M. size standards.

in the case of U-DNA and that UNG plus heat treatment has little effect on the electrophoretic behavior of T-DNA. Table V summarizes the results of an experiment in which the effect of UNG and heat treatment on the transformation rate of several plasmids was tested. The results show that despite the marked changes in the electrophoretic behavior of U-DNA after treatment with UNG and UNG plus heat, similar changes in the transformation efficiency and mutation rate were not evident. We conclude from these results that the E. co/z’ ung’ cells were very efficient in destroying uracil-containing DNA and that nonmutated plasmids that lack uracil probably make a significant contribution to the high background of nonmutant transformants. Consequently, refinements to the protocol that inTABLE

(e) In vitro treatment with UNG does not improve mutant recovery Although plasmids that lacked uracil were expected to make only a small contribution to the total DNA present, they could account for the nonmutant background. To evaluate this point further, we tested the effect of adding UNG to the reaction mixture either before or after the polymerization step in the protocol. Fig. 3 shows that the introduction of UNG results in major modifications in the electrophoretic behavior of denatured and nondenatured ds U-DNA as compared with equivalent T-DNA and that this effect occurred irrespective ofwhether UNG was added before or after the polymerization reaction. Heat treatment after digestion with UNG magnified the difference in the response of U-DNA and T-DNA to the enzyme. Fig. 3 shows that no high-molecular-weight DNA was detectable

treatment.

1-7 show uracil-containing

V

Influence

of UNG

treatment

on transformation

Transformation

DNA

and mutagenesis

rate (cfu:ug)

Uracrl-pTZ mutants”

Nondenatured Nonden.!UNG,’ Nonden.UNG/95’Cb

pBSKP

Uracil-pBSKP

2.0 x IO”

8.0 x IO”

1.8 x IO”

9.0 x IOJ 2.0 x IO?

Denatured‘ Dcn.UNG

1.9 x IO” 1.0 x loF I .o x 10

Den.;‘UNG,‘95”C

1.0 x IO”

,’ I unitlug

DNA,

30 min at 37°C:

I.2 x 10‘ 1.4 x 10:

18-A (‘I,,)

n.d. l1.d. 1l.d. 5 7.5

4.0 x IO’ UNG

rates

6.5 from

Perkin

Elmer Corp.

(Norwalk, CT). h 10 min at 95’C. ’ See footnote a in Table I. ’ Mutagenesis formed

as D. in Table Il. UNG

after mutagenensis.

and heat trcatmcnt

were per-

23 volve an increase of polymerization efficiency would probably be of greater benefit than steps to reduce the ds

introduction

U-DNA further. The replacement

sequence of native

T7 polymerase

with Seque-

nase (version 2.0, US Biochemical Co., Cleveland, OH) led to a decreased recovery of mutants. Sequenase is the result of in vitro genetic manipulation of the gene for T7 DNA polymerase to remove all exonuclease activity associated

or elimination

volved in the mutation. analysis

of a restriction

site was in-

We now preferentially

of six independent

use DNA

clones because

some

colonies contain both wild-type and mutant plasmids (mixed clones). These are also found as a result of mutagenesis with ss DNA (Table II). The percentage of mixed clones ranged between 0% and SO?< of colonies with mutant DNA, and their occurrence seemed to be primer de-

with the enzyme (Tabor and Richardson, 1989). It has catalytic properties of a Form I T7 DNA polymerase and executes efficient strand displacement synthesis. The alkali-denatured plasmid DNA used in the mutagenesis

pendent

reaction contained both ss and ds regions in varying amounts (see above, section d). As a result of polymerasedriven primer elongation, structures resembling preformed replication forks (Lechner and Richardson, 1983) should be formed. Strand displacement synthesis on these structures catalyzed by Form I T7 DNA polymerase is terminated by template switching (Lechner and Richardson, 1983). This would explain the lower performance of the Sequenase versus native T7 DNA polymerase. In this regard, it would be interesting to investigate the addition of the T7 DNA gene 4-encoded protein (Engler et al., 1983) into a mutagenesis reaction with native (Form II) T7 DNA polymerase. The gene I-encoded protein functions as a helicase and has been shown to stimulate multiply the strand displacement synthesis catalyzed by Form II T7 DNA polymerase at preformed replication forks (Lechner and Richardson, 1983). This strand displacement synthesis is not terminated by template switching.

of mutants were nonetheless obtained (Table II, C). We concluded that pTZ18-A must occasionally present ss regions in the plasmid into which the primer could anneal. Among the mutagenesis experiments to which this protocol has been applied, two involved the 16-kb plasmid pBI2.3F (the construction is described in Lelievre et al., 1991) and a 22-kb plant transformation plasmid derived from Agvobacterium (V. Christov, personal communication). The latter is the largest plasmid known to us that has been mutagenized using the protocol. Although only six colonies were isolated, the nt sequence analysis showed that all six contained the desired mutation. The T7 DNA polymerase (Form II) is characterized by a very high fidelity of DNA synthesis (Mattila et al., 1991). The mutational frequencies observed in a base substitution fidelity assay were between 3 and 15 x 10h. Since these rates are only slightly higher than the mutational background for E. coli in vivo (3 x 10”) under the same assay conditions, it would not in most cases be necessary to resequence the entire mutant construct. Although the probability of recovering unexpected mutants will increase as the size of the plasmid becomes larger, this simple oligodirected mutagenesis protocol could be of general usefulness with plasmids of widely disparate sizes.

(f) It is possible to generate mutations in plasmids of a wide variety of sizes Plasmids to be manipulated by the described method have no special sequence requirements such as restriction sites in the proximity of the mutation. This, in addition to efficiency, is a major advantage of the technique. Any cloned DNA under investigation can be mutagenized without subcloning. The protocol has been used in our laboratory to produce about 20 mutant genes that encode seed storage proteins and that contain either point mutations or deletions of up to 8 bp. Additionally, it has been applied successfully in other laboratories (G. Iyer, personal communication; V. Christov, personal communication; S. Filichkin, personal communication). Most plasmids used were about 5 kb in size. The recoveries of mutants from threequarters of our experiments have ranged between 3072 and 40%. The lowest yield obtained was 10% and the highest, 70”,. The selection of mutants was initially either done by colony hybridization (Grunstein and Hogness, 1975) with the radiolabeled mutagenic primer or, if appropriate, by restriction analysis of randomly picked colonies when an

(not shown).

Resolution

of mixed clones is easily

achieved by transformation. In one experiment, the alkali denaturation step was inadvertently left out of the procedure, and a small number

REFERENCES Bebenek,

K. and Kunkel, T.: The use of native T7 DNA polymerase

site-directed Bowman,

mutagenesis.

S., Tischfield, J.A. and Stambrook,

method for producing (1990) 254-260. Dente,

L., Cesareni,

plasmids. Drutsa,

site directed

G. and Cortcse,

P.J.: An efficient and simple

mutations

by PCR. Technique

R.: pEMBL:

V.R.: Simple approach

directed mutagenesis of any double-stranded Lett. 283 (1991) 227-229. Efimov, V.A., Mirskikh, Yu.: A convenient

O.V., Chakhmakhcheva, modification

directed in vitro mutagenesis 407-411. Engler.

M.J.,

Lechner,

2

a new family of ss

11(1983) 1645-1655.

Nucleic Acids Res.

V.L. and Kaberdin,

for

Nucleic Acids Res. 17 (1989) 5408.

for oligonucleotidecircular

DNA.

FEBS

O.G. and Ovchinnikov,

of the method

for oligonuclcotide-

of cloned DNA. FEBS Lett. 181 (1985)

R.L. and Richardson.

C.C.: Two forms of the

24 DNA polymerase 11165-11173.

of bacteriophage

T7. J. Biol. Chem.

258 (1983)

M. and Hogncss,

isolation Acad.

D.: Colony hybridization:

of cloned cDNAs

that contain

in plasmids.

Nakamaye,

R., Krummel,

a method

a specific gent.

for the

Proc. Natl. of in vitro

7367.

Biolabs

mutagenesis.

1990-1991

of restriction

endonuclcase

groups in vitro and its application Nucleic Acids Res. 14 (19X6)

Catalogue.

New England

Biolabs Inc.,

Beverly, MA, 1990. p. 135. Olsen, D.B. and Eckstein F.: High-efficiency oligonuclcotidc directed plasmid mutagenesis. Proc. Natl. Acad. Sci. USA 87 (1990) 14511455.

Kunkcl,

T.A.:

Rapid

phenotypic

and

selection.

efficient

Proc.

site-specific

Natl. Acad.

mutagenesis

Sci. USA

without

82 (1985) 488-

492. Kramer,

W. and

gapped Kuipers,

Fritz,

H.-J.:

duplex DNA.

Oligonucleotide-directed

Methods

Enzymol.

mutations

tagenesis replication

mu-

M.J. and Richardson,

displacement

synthesis

J. Bid. Gem.

polymerase. J.-M.,

Oliveira.

elements modulate

catalyzed

CC.:

stable

L.O.

and Nielsen,

the expression

Characterization

by bacteriophage

258 (1983) 11174-l

of

T7 DNA

Laboratory Harbor.

of glycinin genes. Plant Physiol.

98

synthesis

E.F.

Manual.

and

Sambrook.

Cold Spring

Harbor

J.: Molecular Laboratory,

Cloning.

A

Cold Spring

J., Tenkanen,

T. and Pitkanen,

by the Thermococcus lirovuli.~DNA

K.: Fidelity of DNA polymerasc

linkage

mutagenesis.

of a conserved

Asn-Gly

of soybean.

otide and double-stranded 194-200. Sugimoto,

synthesis DNA templates.

M., Esaki, N., Tanaka.

Biochem.

-

an ex-

Proc.

method

Natl.

for am-

Gene 103 (1991) 73-77.

esis by complementary-strand

using a closing Anal. Biochem.

H. and Soda.

mutagenoligonuclc185 (1990)

K.: A simple and elhmutagenesis

and phosphorothioate-modified

using plas-

nucleotide.

Anal.

179 (1989) 309-3 11.

S. and Richardson.

activity of bacteriophage

C.: Selective inactivation T7 DNA polymerase

of the exonuclcasc

by in rirro mutagcncais.

J. Biol. Chem. 264 (1989) 6447-6458. Zoller.

NY, 1982, p. 51.

P., Korpela,

directed

Tabor,

T‘., Fritsch,

respon-

cleavage

in glycinin, the major seed storage protein Acad. Sci. USA 89 (1991) 658-662.

mid DNA template

5’-CATGCAT-3’

(1991) 387-391. Maniatis.

K. and Nielsen, N.C.: A protease

cient method for the oligonuclcotide-directed

1184.

N.C.:

sible for post-translational

Slilaty, S.N., Fung, M., Shen, S.-H. and Lebel, S.: Site-directed

topologically

fork. J. Biol. Chem. 258 (1983) 11185- 11196.

R.L., Englcr,

strand

C.C.: A preformed,

Scott, M.P., Jung, R., Muntr,

Shen. T.J., Zhu, L.Q. and Sun, X.: A marker-coupled

site-directed

using PCR. Nucleic Acids Res. 19 (1991) 4558.

R.L. and Richardson,

Lcchner,

via

154 (1987) 350-403.

O.P., Boot, H.J. and Dcvos, W.M.: Improved

Lechner,

Mattila,

F.: Inhibition

by phosphorothioate

to oligonucleotide-directed 9679-9698. New England

B. and Saiki, P.K.: A general method

preparation and specific mutagenesis of DNA fragments: study of protein and DNA interactions. Nucleic Acids Res. 16 (1988) 7351-

Lelievre.

K.L. and Eckstein,

‘VcciIcleavage

Sci. USA 72 (1975) 3961-3965.

Higuchi,

activity. Nucleic Acids

Res. 19 (1991) 4967-4973.

Foss, K. and McClain, W.H.: Rapid site specific mutagenesis Gene 59 (1987) 285-290. Grunstein,

tremely heat stable enzyme with proofreading

M.J.

and

DNA fragments (1983) 468-500.

Smith,

M.: Oligonuclcotidc-directed

cloned into Ml3 vectors.

Methods

mutagenesis Enrymol.

of IOOB