Mol. Cells, Vol. 7, No. 3, pp. 419-424
Improvement of the 3'-5' Exonuclease Activity of Taq DNA
Polymerase by Protein Engineering in the Active Site
Yonghyun Park, Hyeja Choi, Dae Sit Lee l and Youngsoo Kim*
School of Chemical Engineering & Technology, Yeungnam University, Kyungsan 712-749,
Korea; 'Korea Research Institute of Bioscience and Biotechnology, Taejon 305-333, Korea
(Received on March 26, 1997)
Taq DNA polymerase from Thermus aquaticus has been shown to be very useful in the polymerase chain reaction method. Taq DNA polymerase has a domain at its amino terminus
(residue 1 to 291) that has a 5'-3' exonuclease activity, a 3'-5' exonuclease domain in the
middle (residue 292 to 423), and a domain at its C-terminus that catalyzes polymerase
reactions. Taq DNA polymerase is classified into the poll family which is represented by E.
coli DNA polymerase I. The three dimensional structural alignment of 3'-5' exonuclease
domains from the poll family, DNA polymerases leads us to understand why Taq DNA polymerase does not carry out proof-reading in the polymerase chain reaction. Three
sequence motifs, called ExoI, II, and III must be present in order to carry out proof-reading by the 3'-5' exonuclease reaction in DNA polymerization, but Taq DNA polymerase
contains none of them. The key catalytic module in the 3'-5' exonuclease is two metal ions
chelated by active-site carboxylic amino acids. In order to render the 3'-5' exonuclease
activity in Taq DNA polymerase, a catalytic module was constructured in the active site
by protein engineering. The mutant Taq DNA polymerase shows twice as much the 3'-5'
exonuclease activity as that of wild-type DNA polymerase.
Taq DNA polymerase from Thermus aquaticus has
been shown to be very useful in polymerase chain
reaction (PCR). It shows an optimum reaction temperature of 75 °C and keeps the activity for about one
hour at 94 °c. Taq DNA polymerase's high optimum
polymerization temperature at 75°C provides unique
advantages when comparing over mesophilic DNA
polymerases such as E. coli DNA polymerase I
(Holland et at., 1991; Myers and Gelfand, 1991;
Saiki et aI., 1985; 1988). Not only is Taq DNA polymerase highly useful commerically for PCR applications, but it is also important in studying DNA
replication. Taq DNA polymerase is apparently homologous to E. coli DNA polymerase I, which has long
been used for DNA replication studies. Taq DNA polymerase has a domain at its amino terminus (residue 1
to 291) that has a 5'-3' exonuclease activity and a
domain at its C-terminus that catalyzes polymerase
reactions (Longley et al., 1990). Unlike E. coii DNA
polymerase I, the intervening domain of Taq DNA polymerase has lost the editing activity of the 3'-5' exonuclease (Kornberg and Baker, 1992; Lyamichev et
ai., 1993). All the DNA polymerases can be grouped
into six families such as poll, pola, pol~, DNA-dependent RNA polymerase, reverse transcriptase, and
RNA-dependent RNA polymerase on the basis of amino acid homology (Joyce and Steitz, 1994). Taq
* To whom correspondence should be addressed.
DNA polymerase is classified into the pol I family,
which is represented by E. coii DNA polymerase I
(Delrarue et ai., 1990). The comparison of the structure of Klenow fragment (KF) with the corresponding
parts of Taq DNA polymerase shows that the polymerase domains are very nearly identical, whereas the
3'-5' exonuclease domains differ extensively (Lawyer
et ai., 1989). High resolution structural data from
crystallographic studies has been published on the polymerase and 3'-5' exonuclease domains, and even the
reaction mechanisms have nearly been identified
(Beese et at., 1991; 1993; Derbyshire et at., 1988;
1991; Freemont et at., 1986; Ollis et at., 1985; Pelletier et at., 1994). However, little was known about
the structural basis of the 5'-3' exonuclease activity
until the structure of Taq DNA polymerase was published at the 2.6 A resolution (Kim et ai., 1995). Afterward, the crystal structure of T5 5'-3' exonuclease
illustrates the structural features of 5'-3' exonuclease
even more clearly (Ceska et ai., 1996).
Research on Taq DNA polymerase has been initiated since it is a very useful commercial enzyme
for PCR. There have been two directions in PCR
The abbreviations used are: EDTA, ethylenediaminetetraacetic acid; IPTG, iso-propyl-l-thio-beta-galactopyranoside;
KF, Klenow fragment; PCR, polymerase chain reaction;
SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel
electrophoresis.
© 1997 The Korean Society for Molecular Biology
420
Mutagenesis of the 3'-5' Exonuclease Domain of Taq DNA Polymerase
research (Barnes, 1992). One has been to improve the
processivity and fidelity of Taq DNA polymerase.
T his line of research has been followed mainly by
Perkin Elmer since Taq DNA polymerase has been
the most popular enzyme in PCR experiments. The
other one is to introduce a new thermophilic DNA polymerase to substitute fo r Taq DNA polymerase. This
has been carried out mostly by biotech companies.
Thei r efforts have launched several thermophilic
DNA polymerases onto the market: Pfu DNA polymerase (Stragene), Vent DNA polymerase (New
England Biolab), and Takara DNA polymerase
(Takara Shuzo Co.) and so on. Presently, Taq DNA
polymerase has been used prevalently in PCR.
However, Taq DNA polymerase has a big disadvantage when compared to the other thermophilic
DNA polymerases : lower proof-reading activity. Apparently, Taq DNA polymerase does not carry the 3'5' exonuclease activity. This may be due to a shorter
3'-5' exonuclease domain, even though the domain exists (Flam an et al. , 1994). High fidelity DNA synthesis is the most necessary factor fo r all thermophilic
DNA polymerases in order to be preferred enzymes
fo r PCR (Eckert and Kunkel, 1990). On the market,
several enzy mes are competing with Taq DNA polymerase on the basis of higher fidelity. DNA polymerization fidelity is influenced by multiple factors,
but the presence of the 3'-5' exonuclease activity fo r
proof-readi ng is one of the most critical elements.
The presence of the 3'-5' exonuclease activity can remove mismatched nucl eotides and thereby reduce the
error rate in PCR reactions. The Taq, Pju, and Vent
DNA polymerases are major thermophilic DNA polymerases curren tly on the market. Their fidelities fo r
DNA synthesis vary such that the error rate of Taq,
Pfu , and ~e nt DNA p_qlymerases are r~~orted t? be
1.8 x 10 ., 0.2 x 10 ,and 1.2 x 10 repectIVely
(Flaman et al. , 1994). The reason why Taq DNA
polyme rase s hows th e high est value is that it
possesses only a background level of the 3'-5' exonucl ease activity. Disadvantage can be an attacking
point for other DNA polymerases to compete on the
market with Taq DNA polymerase. It was analyzed
in this study why Taq DNA polymerase carrys just a
minimal background level of the 3'-5' exonuclease activity while other thermophilic DNA polymerases
show a higher the 3'-5' exonuclease activity. In order
to increase the proof-reading function of Taq DNA
polymerase, we tried to introduce a catalytic module
fo r the 3'-5' exonuclease reaction by protein
engineering.
Materials and Methods
Materials
The mutagenesis kit and Taq DNA polymerase
were obtained from Bioneer Co. (Taejon, Korea). All
other enzymes were purchased from Promega. Ra32
dioactive compounds of [a_ P]dCTP for DNA la-
Mol. Cells
beling were purchased from Arnersham International
(Arnersham, UK).
Plasmid constructions
Site-directed mutagenesis was carried out to construct mutant 3'-5' exonucleases using pDS1 plasmid
(Kwon et aI. , 1991; Park et al. , 1993). The PCR
DNA fragment of mutant Taq DNA polymerase was
sequenced to confirm its amount of 1305 bases in order fo r the mutation to be carried out correctly. DNA
sequencing was performed by the thermal DNA
sequencing method using Taq DNA polymerase
(Innis et al., 1988). The constructed pDSl plasmid
was transformed into E. coli DH5a for protein
overexpression.
Purification of mutant Taq DNA polymerase
E. coli DH5a carrying pDSl plasmid was incubated at 37°C and shaken at 250 rpm. E. coli DH
5a in 300 ml LB medium was cultured overnight
with 80 Ilglml ampicillin. The induction of gene expression and the purification procedure proceeded as
described previously (Engelke et al. , 1990; Kim et al.,
1995). The two differences are that a Q-Sepharose
open colum was substituted for the FPLC mono Q as
the final purification step, and the host cell was E.
coli DH5a. The purified protein from procedure was
pure enough to carry out the 3'-5' exonuclease assay.
Assay of the 3'-5 ' exonuclease activity
The substrate of pUC18 DNA was digested with
XmaI and labeled at the 3'-terminus with [a_ 32P]dCTP
by Taq DNA polymerase to make two mismatched
base pairs like C-c. The labeled DNA was purified
by the Gene Clean kit (Bio101, U.S.A.) and unbound
radioactive nucleotides were discarded. The 3'-5' exonuclease reaction was fo llowed at 72 °C for 1 h in 50
III reaction mixture containing 25 mM Tris-HCl, pH
8.0, 1 mM ~-mercaptoethanol , 10 mM MgCI 2, 0.5 III
DNA substrate, and 7 Ilg of Taq DNA polymerase enzyme. After the reaction was stopped, the mixture
was cooled down on ice with addition of 2 III of 60
mM EDTA. The 52 III aliquot was spotted on a 2.3
cm-diametered DE-81 Whatman filter paper and dried
in a heat block for 10 min. The dried filter was washed twice with 0.5 M Na 2HP0 4 , pH 7.0, for 15 min
and then washed with 70% ethanol fo r 5 min. T he
radioactivity was counted in a Beckmann liquid scintillation counter model LS 6500.
Results
Comparison of the 3'-5' exonuclease domain between
Taq DNA polymerase and E. coli DNA polymerase I
Taq DNA polymerase is grouped into the Poll family of six nucleic acid polymerase classes such as
Poll, Po/a, Po/~, DNA-dependent RNA polymerase,
reverse transcriptase, and RNA-dependent RNA polymerase. E. coli DNA polymerase I, which has long
Vol. 7 (1997)
Yonghyun Park et al.
been studied for DNA replication, is also classified
into the Poll group like Taq DNA polymerase. It is
known that Taq DNA polymerase is very homologous to E. coli DNA polymerase I in both primary
and tertiary structures. The two polymerase domains
are almost identical to each other, but there is a major difference in terms of structural basis. E. coli
DNA polymerase I carries 3'-5' exonuclease activity
fo r proof-reading, but Taq DNA polymerase does not
possess this activity at a significant level. Taq DNA
polymerase has a minimal background level of the 3'5' exonuclease activity that causes lower proof-reading in DNA synthesis. However, Taq DNA polymerase does carry a 3'-5' exonuclease domain from
residue 292 to 423. The existing domain does not
play a proper role in correcting mismatched base
pairs. In order to introduce the 3'-5' exonuclease activity into Taq DNA polymerase, those two enzymes
must fi rst be compared in both primary and tertiary
structures. If the differences are figured out, it might
be possible to improve the 3'-5' exonuclease activity
of Taq DNA polymerase. One major difference in the
overall structure of the 3'-5' exonuclease domain of
Taq DNA polymerase, from that of KJenow fragment
(KF), was the deletion of four loops of lengths
between 8 to 27 residues as shown in Figure 1. In
KF, these loops pack together on one side of the 3'-5'
exonuclease domain as shown in Figure 2. Moreover,
all fo ur of the carboxylates (D424, D501, D355, E
357) known to be essential for divalent metal binding
and catalysis in the 3'-5' exonuclease domain of KF
have been replaced by residues incapable of binding
metal ions (L356, R405, G308, V31O) in the vestigal
3'-5' exonuclease domain of Taq DNA polymerase as
shown in Figure 2. In addition, several amino acids
involved in substrate binding in KF are replaced with
T.i'q. 292
l<1\IiEl\EW
~
314
E.C.
YINYVTIL
. PIIF'l\f1JIEll:
361
327
T.i'q. 315
E.c.
362
~RIIHRAP
[N[SANL~
T.i'q. 345
L
E.c.
L.
405
. VPMIP . . . .
EYiKlIIRD
344
.. ..... ~
404
~
T .i'q. 377
l'IlIlI!P
E.c.
SYIINS .. ~ . . . . . . • . . . . .
446
484
different ones in the corresponding residues of Taq
DNA polymerase. Although the 3'-5' exonuclease catalytic site has been destroyed and the size of the
domain reduced, the contact with the polymerase
domain and distance between the polymerase and 3'-5'
exon uclease domains remain similar in the two homologous polymerases.
Mutagenesis of the four active site residues to
carboxylic amino acids
Taq DNA polymerase shows a background level of
3'-5' exonuclease activity that causes lower proof-reading than other commercial thermophilic DNA polyme rases on the market. As suggested in the previous
section, the most significant feature that 3'-5' exonuclease must carry the activity is to have two metal
ions chelated by four active-site carboxyl ic amino
acids in its active site. It turned out that any mutation
in those fo ur residues reduced the 3'-5' exonuclease
activity by a factor of 10 - 3 to 10 - 5 (Derbyshire et al.,
1988; 1991). It was conceivable that the site-direct
mu tagenesis of L356, R405, G308 , and V310 to functionally active carboxylates might bring back some 3'5' exonuclease activity to Taq DNA pol ymerase.
Therefore, those 4 presumable active site residues
such as 308Gly (GGC), 310Val (GTG), 356Leu
(CTG), and 405Arg (CGG) were mutated to 308Asp
(GAC), 310Glu (GAG), 356Asp (GAC), and 405Asp
(GAC), respectively. The mutagenic DNA fragment
400
483
~423
T.i'q. 401
E.c .
421
....... ......
~
519
Figure 1. Structure-based alignment of the sequences of
the 3'-5' exonuclease domain of KF on the corresponding
domain of Taq DNA polymerase. The unaligned residues
are shown as dots in E. coli DNA polymerase I (E.c.) and
the unpaired missing residues are shown as blank in Taq
DNA polymerase (T.aq).
Figure 2. Superposition of the 3'-5' exonuclease domains
of E. coli DNA polymerase I and Taq DNA polymerase. E.
coli DNA polymerase I is in purple and Taq DNA polymerase in yellow. Two divalent metal sites are ca lled metal A and B, and four carboxylic amino acids are assigned
to divalent metals respectively. The four amino acids in the
active site of Taq DNA polymerase matched corresponding
ones of E. coli DNA polymerase I.
Mol. Cells
Mutagenesis of the 3'-5' Exonuclease Domain of Taq DNA Polymerase
422
1
2
3
4
5
enzyme
3' exonuclease
activity
( L1 cprn/ fJ g protein)
Wild - type Taq
DNA polymerase
505.3
100
mutant Taq DNA
polymerase
1005.0
203
Relative acti vity
(%)
220
200
Figure 3. SDS-PAGE of purified mutant Taq DNA polymerase from E. coli DH5a cells. Taq DNA polymerase is
shown at 94 kDa as a single thick band in lane 4. Lane 1,
molecular weight marker; lane 2, cell extract; lane 3, supernatant after heat treatment; lane 4, Q-Sepharose column
fraction.
180
ec
:~
u
';;-"
"
160
140
120
100
.~
0;
cG
80
60
40
20
wild-type
was digested by KpnI at base 479 and BamHI at 1778
of the Taq DNA polymerase gene. In order to assure
correct mutations and no random mutagenesis during
the PCR mutagenesis reaction, 1305 bases were all
sequenced by the thermal DNA sequencing method.
No random mutation was fo und and those four amino
acids were correctly replaced with carboxylic amino
acids. The mutagenic fragment was ligated in the vector pDS1 , which contains tac promotor as a protein
expression system.
Expression and purification of mutant Taq DNA
polymerase
Wild-type and 3'-5' exonuclease-mutant genes of
Taq DNA polymerase were expressed in E. coli DH
Sa cells. Protein synthesis was induced with 0.2 mM
IPTG at 0.2 O.D . unit at 550 nm wavelength. A large
amount of protein was overproduced for wild-type
and 3'-5' exonuclease-mutant Taq DNA polymerase.
The purification procedure was simplified from the
one reported previously (Kim, 1995). Taq DNA polymerase was identified by SDS-PAGE of molecular
weight, where mutant Taq DNA polymerase was positioned as a single thick band at 94 kDa size as
shown in Figure 3, and by carrying out biochemical
reactions using mutant Taq DNA polymerases.
Assay of the 3-5' exonuclease activity of mutant Taq
DNA polymerase
The 3'-5' exonucleolytic activity of mutant Taq
DNA polymerase was compared to that of wild-type
Taq DNA polymerase. While the wild-type showed
505 cpm/~g protein corresponding to background levels of the 3'-5' exonuclease activity, mutant Taq DNA
polymerase resulted in 1005 cpm/~g protein. It doubled exactly the value of wild-type 3'-5' exonuclease.
mutant
Figure 4. Activity measurement of the 3'-5' exonuclease
of wild-type Taq DNA polymerase and mutant Taq DNA
polymerase which contains four carboxylic amino acids in
its active site.
The result of 3'-5' exonuclease assays for the wildtype and mutant Taq DNA polymerases are summarized in Figure 4. The insertion of catalytic carboxylates into the active-site of wild-type Taq DNA
polymerase increased the 3'-5' exonuclease activity by
two times. Since there are many factors involved in
the 3'-5' exonuclease reaction, the construction of
those carboxylic amino acids alone could not bring
up full activity. However, as expected, the introduction of a key catalytic module to the 3'-5' exonuclease reaction contributed to bringing up 3'-5' exonuclease activity to some extent.
Discussion
The 3'-5' exonuclease activity of thermophilic
DNA polymerase for PCR serves a proof-reading
function to increase DNA synthesis fidelity. This activity is present in most DNA polymerases, but not in
Taq DNA polymerase. It gives Taq DNA polymerase
a higher error rate in DNA synthesis. In general, the
exonuclease activity resides in a certain domain of
the whole DNA polymerase. However, there are
some exceptions like the single polypeptide in T4 exonuclease (Ceska et al., 1996). The 3'-5' exonuclease
is probably best understood at a mechanism level by
biochemical and X-ray crystallographical studies, the
so-called phosphoryl transfer mechanism.
A mechanism of phosphoryl transfer where two divalent metal ions are involved is becoming apparent
in several enzymes. There are several enzymes which
Vol. 7 (1997)
Yonghyun Park et al.
follow the same path such as alkaline phosphatase,
pyrophosphatase, and RNase H. This two metal-ion
mechanism seems to be the case for the 3'-5' exonuclease and polymerase domains of KF, being supported by structural and mechanical studies. From accumulated structural evidence and biochemical studies, this might be the case for all 3'-5' exonucleases
(Davies et ai., 1991; 1994; Esteban et ai. , 1994; Kankare et ai. , 1994; Kim and Wyckoff, 1991 ; Yang et
ai., 1990). For the phosphoryl transfer mechnism, the
four active-site carboxylic amino acids involved in
chelating the two divalent metal ions are 355Asp,
357Glu, Asp424, and 501Asp in KF. In addition to
those carboxylic amino acids, several amino acids
play significant roles in binding substrates such as
Leu361 , Phe473, Tyr497, Gln419, and Arg455 in KF.
It is now accepted that all the DNA polymerases,
which have a proof-reading function due to the 3'-5'
exonuclease activity, possess three sequence motifs
called ExoI, II, and III. The ExoI motif contains the
core sequence DXE, where two carboxylic amino
acids like Asp355 and Glu357 are positioned in KF.
Exoll has the sequence NX2. 3(F/Y)D where Asn420
and Asp424 fit the case in KF. The ExoIII motif has
the sequence YX 3D, corresponding to active site residues Tyr497 and 501Asp in KF. The motifs are
necessary to keep the 3'-5' exonuclease in a fully active form , and the amino acids in the three motifs participate in chemical catalysis and substrate binding
(Derbyshire et ai., 1995). However, all three motifs
are absent in Taq DNA polymerase and thereby an
editing function is absent (Reha-Krantz, 1992).
The mutagenesis inserted four catalytic carboxylic
residues into the active site of Taq DNA polymerase
so that it can chelate two divalent metal ions that
make up a critical catalytic module. It is conceivable
that the 3'-5' exonuclease activity could be restorable
to a some extent once the phosphoryl transfer module
was placed. As expected, the 3'-5' exonuclease activity increased twice of that of wild-type Taq DNA
polymerase. If we presume that the improvement of
the 3'-5' exonuclease activity could contribute to increasing proof-reading as much, it would really half
the error-rate of DNA synthesis in peR reactions using Taq DNA polymerase. In order to directly correlate the 3'-5' exonuclease activity and values for
peR fidelity for DNA polymerases like Taq DNA
polymerase, more studies are yet to be done. Regardless of further study, if we assume that the 3'-5'
exonuclease activity could be extrapolated to peR
fidelity, and if we remember that the error rates of
Vent and Tag DNA polymerases are 1.2 X 10 - 5 and
5
1.8 X 10- , respectively, mutant Taq DNA polymerase may show proof-reading as reliable as Vent
DNA polymerase (Flaman et ai. , 1994). However, we
have to carry out a peR fidelity assay in order to examine whether the improved 3'-5' exonuclease activity
can really reflect into peR fidelity . In addition, spectroscopic analysis will be required to identify the two
423
newly inserted divalent metal ions in the active-site of
mutant Taq DNA polymerase. This would be direct evidence for active site reconstitution.
Acknowledgment
This work was supported by a grant from the
Genetic Engineering Research Fund of the Korean
Ministry of Education in 1996 and in part by the
Yeungnam University Research Grant in 1996 to
Youngsoo Kim .
References
Barnes, W. M. (1992) Gene 112, 29-35.
Beese, L. S., and Steitz, T. A. (1991) EMBO 1. 10, 25 -33.
Beese, L. S., Derbyshire, V., and Steitz, T. A. (1993) Science 260, 352-355.
Ceska, T. A., Sayers, 1. R., Stier, G. , and Suck, D. (1996)
Nature 382, 90-94.
Davies, J . F., Hostomska, Z., Homstomsky, Z., Jordan, S.
R., and Mattews, D. A. (1991) Science 252, 88-95.
Davies, 1. F., Almassey, R. 1., Hostomska, Z., Ferre, R. A.,
and Hostomsky , Z. (1994) Cell 76, 1123-1133.
Delarue, M. , Poch, 0 ., Tordo, N. , Moras, D. , and Argos, P.
(1990) Prot. Eng. 3, 461-467 .
Derbyshire, V., Freemont, P. S., Sanderson, M. R., Beese,
L. , Friedman, J. M. , Joyce, C. M. , and Steitz, T. A.
(1988) Science 240, 199-201 .
Derbyshire, V., Grindley, N. , and Joyce, C. M. (1991)
EMBO 1. 10, 17-24.
Derbyshire, V. , Pinsonneault, J. K. , and Joyce, C. M.
(1995) Methods Enzymol. 262, 363-385.
Eckert, K. A. , and Kunkel, T. A. (1990) Nucleic Acids Res.
18, 3739-3744.
Engelke, D. R. , Krikos, A., Bruck, M. E., and Ginsberg, D.
(1990) Anal. Biochem. 191, 396-400.
Esteban, J. A. , So eng as, M. S., Salas, M., and Blanco, L.
(1994) 1. BioI. Chem . 269, 31946-31954.
Fiaman, J.-M., Frebourg, T. , Moreau, V., Charbonnier, F.,
Martin, C., Ishioka, C., Friend, S. H. , and Iggo, R. (1994)
Nucleic Acids Res. 22, 3259-3260.
Freemont, P. S., Ollis, D. L. , Steitz, T. A. , and Joyce, C.
M. (1986) Proteins 1, 66-73.
Holland, P. M., Abramson, R. D. , Watson, R., and Gelfand,
D. H. (1991) Proc. Natl. A cad. Sci. USA 88, 7276-7280.
Innis, M. A. , Myambo, K. B., Gelfand, D. H. , and Brow,
M. A. D. (1988) Proc. Natl. Acad. Sci. USA 85, 94369440.
Joyce, C. M., and Steitz, T. A. (1994) Annu. Rev. Biochem.
63, 777-822.
Kankare, J., Neal, G. S., Salminen, T. , Glumhoff, T. , Cooperman, B. S. , Rahti, R., and Goldman, A. (1994) Prot. Eng.
7, 823-830.
Kim, E., and Wyckoff, H. H. (1991) 1. Mol. Bioi. 218, 449464.
Kim, Y. (1995) Mol. Cells 5, 452-460.
Kim, Y. , Eom, S. H., Wang, J., Lee, D. S., Suh, S. W. ,
and Steitz, T. A. (1995) Nature 376,612-616.
Kornberg, A., and Baker, T. (1992) DNA Replication , 2nd
Ed, Freeman and Company, San Francisco, CA.
Kwon, S. T. , Kim, J. S., Park, J. H., Koh, S., and Lee, D.
424
Mutagenesis of the 3'-5' Exonuclease Domain of Taq DNA Polymerase
S. (1991 ) Mol. Cells 1, 369-375.
Lawyer, F. C , Stoffel, S., Saiki, R. K , Myambo, K , Drummond, R., and Gelfand, D. H. (1989) 1. BioI. Chem. 264,
6427-6437 .
Longley, M. J., Bennett, S. E. , and Mosbaugh, D. W.
(1990) Nucleic A cids Res. 18, 7317-7322.
Lyamichev, V., Brow, M. A D. , and Dahlberg, S. E.
(1993) Science 260, 778-7835.
Myers, T w., and Gelfand, D. H. (1991) Biochemistry 30,
7661-7665.
Ollis, D . L. , Brick, P., Hamlin, R. , Xuong, N. G. , and
Steitz, T. A (1 985) Nature 313, 762-766.
Park, J. H. , Kim, J . S., Kwon, S. T., and Lee, D. S. (1993)
Mol. Cells
Eur. 1. Biochem. 214, 135-140.
Pelletier, H., Sawaya, M. R., Kumar, A, Wilson, S. H.,
and Kraut, J. (1994) Science 264, 1891-1903 .
Reha-Krantz, L. J. (1992) Gene 112, 133-137.
Saiki, R. K , Scharf, S., Faloona, F., Mullis, K B. , Hom,
G. T , Erich, H. A , and Amheim, N. (1985) Science 230,
1350-1354.
Saiki, R. K , Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Hom, G. T , Mullis, K B., and Erlich, H. A
(1988) Science 239, 487-49l.
Yang, W. , Hendrickson, W. A , Crouch, R. J ., and Satow,
y. (1 990) Science 252, 88-95 .