526
`
`Biochemistry 1991, 30, 526-537
`
`An Induced-Fit Kinetic Mechanism for DNA Replication Fidelity: Direct
`Measurement by Single-Turnover Kinetics1"
`Isaac Wong, Smita S. Patel, and Kenneth A. Johnson* *
`Department of Molecular and Cell Biology, 301 Althouse Laboratory, The Pennsylvania State University, University Park,
`Pennsylvania 16802
`Received February 5, 1990; Revised Manuscript Received August 16, 1990
`
`An exonuclease-deficient mutant of T7 DNA polymerase was constructed and utilized in a series
`abstract:
`of kinetic studies on misincorporation and next correct dNTP incorporation. By using a synthetic oligo-
`nucleotide template-primer system for which the kinetic pathway for correct incorporation has been solved
`[Patel, S. S., Wong, I., & Johnson, K. A. (1991) Biochemistry (first of three papers in this issue)], the kinetic
`parameters for the incorporation of the incorrect triphosphates dATP, dCTP, and dGTP were determined,
`giving, respectively, kCit/Km values of 91, 23, and 4.3 M"1 s"1 and a discrimination in the polymerization
`step of 105—106. The rates of misincorporation in all cases were linearly dependent on substrate concentration
`up to 4 mM, beyond which severe inhibition was observed. Competition of correct
`incorporation versus
`dCTP revealed an estimated K, of ~6-8 mM, suggesting a corresponding km of 0.14 s_1. Moderate elemental
`effects of 19-, 17-, and 34-fold reduction in rates were measured by substituting the a-thiotriphosphate
`analogues for dATP, dCTP, and dGTP, respectively, indicating that the chemistry step is partially rate-
`limiting. The absence of a burst of incorporation during the first turnover places the rate-limiting step at
`In contrast, the incorporation of
`a triphosphate binding induced conformational change before chemistry.
`triphosphate, dCTP, on a mismatched DNA substrate was saturable with a Km of 87 juM
`the next correct
`for dCTP, 4-fold higher than the Kd for the correct
`incorporation on duplex DNA, and a fccat of 0.025 s_1.
`A larger elemental effect of 60, however, suggests a rate-limiting chemistry step. The rate of pyro-
`phosphorolysis on a mismatched 3'-end is undetectable, indicating that pyrophosphorolysis does not play
`a proofreading role in replication. These results show convincingly that the T7 DNA polymerase discriminates
`against the incorrect triphosphate by an induced-fit conformational change and that, following misincor-
`poration, the enzyme then selects against the resultant mismatched DNA by a slow, rate-limiting chemistry
`step, thereby allowing sufficient time for the release of the mismatched DNA from the polymerase active
`site to be followed by exonucleolytic error correction.
`
`IVIodels to explain the high fidelity of DNA replication
`have long been proposed in the literature [reviewed in Loeb
`and Kunkel (1982)]. For example, several models have in-
`voked a role for pyrophosphorolysis in fidelity by mechanisms
`sometimes referred to as “kinetic proofreading” (Brutlag &
`Kornberg, 1972; Deutscher & Kornberg, 1969; Flopfield, 1976;
`It is, however, worth noting that the original
`Ninio, 1975).
`models have been proposed as purely mathematical constructs,
`and the only evidence in support of a role for pyro-
`phosphorolysis is the observation that polymerization is less
`accurate at high pyrophosphate concentrations (Kunkel et al.,
`1986). Although induced-fit mechanisms have been suggested,
`it has also been argued that induced-fit models cannot account
`for increased selectivity (Fersht, 1974, 1985; Page, 1986).
`Thus, because of a void of mechanistic information, almost
`all of these models take considerable liberties with invoking
`specific but unsubstantiated intermediates in the kinetic
`pathway of polymerization. Furthermore, most of the pub-
`lished mechanistic studies available on misincorporation rely
`solely on steady-state kinetic data, which, as we will show, are
`at best difficult to interpret and at worst extremely misleading.
`The only transient kinetic study on the mechanism of DNA
`polymerization relied upon the use of the Klenow fragment
`of the DNA repair enzyme, Pol I, and some surprising results
`were obtained, the most notable of which is that the polymerase
`
`fThis work was supported by the Paul Berg Professorship from Penn
`State University (K.A.J.).
`*To whom correspondence should be addressed.
`
`0006-2960/91 /0430-526S02.50/0
`
`kr
`E,DNA„.,= Ep DNA„
`
`*"•"
`
`E„ DNA„.,
`
`E, DNAn>1 + dNMP
`
`bound correct and incorrect dNTPs with nearly equal affinities
`(Kuchta et al., 1987, 1988).
`In this report, we propose a mechanism for DNA replication
`fidelity as a part of our kinetic study of the T7 DNA polym-
`erase (Patel et al., 1991; Donlin et al., 1991). The T7 system
`lends itself especially well to our approach to fidelity because
`its kinetic scheme has been completely solved (Patel et al.,
`1991; Donlin et al., 1991). Consequently, our
`results can be
`interpreted by a direct comparison of the differences between
`In general, T7 DNA
`correct and incorrect polymerization.
`polymerase constitutes a nearly ideal model system for any
`DNA replication studies because it functions in vivo as a true
`replication enzyme with a minimal number of components.
`The problem of replication fidelity is summarized in Scheme
`I. Conceptually, we divide the issue of fidelity into two parts.
`First, we are concerned with the mechanism of making an
`this primarily involves solving the kinetics of misin-
`error;
`corporation. Second, we are interested in the kinetic conse-
`quences of the error; here, we must determine the relative
`©1991 American Chemical Society
`
`Downloaded via ALBERT EINSTEIN COLG MEDICINE INC on March 29, 2021 at 17:17:19 (UTC).
`
`See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
`
`Columbia Ex. 2087
`Illumina, Inc. v. The Trustees
`of Columbia University in the
`City of New York
`IPR2020-00988, -01065,
`-01177, -01125, -01323
`
`

`

`526
`
`Biochemistry 1991, 30, 526-537
`
`An Induced-Fit Kinetic Mechanism for DNA Replication Fidelity: Direct
`Measurement by Single-Turnover Kinetics1"
`Isaac Wong, Smita S. Patel, and Kenneth A. Johnson* *
`Department of Molecular and Cell Biology, 301 Althouse Laboratory, The Pennsylvania State University, University Park,
`Pennsylvania 16802
`Received February 5, 1990; Revised Manuscript Received August 16, 1990
`
`An exonuclease-deficient mutant of T7 DNA polymerase was constructed and utilized in a series
`abstract:
`of kinetic studies on misincorporation and next correct dNTP incorporation. By using a synthetic oligo-
`nucleotide template-primer system for which the kinetic pathway for correct incorporation has been solved
`[Patel, S. S., Wong, I., & Johnson, K. A. (1991) Biochemistry (first of three papers in this issue)], the kinetic
`parameters for the incorporation of the incorrect triphosphates dATP, dCTP, and dGTP were determined,
`giving, respectively, kCit/Km values of 91, 23, and 4.3 M"1 s"1 and a discrimination in the polymerization
`step of 105—106. The rates of misincorporation in all cases were linearly dependent on substrate concentration
`up to 4 mM, beyond which severe inhibition was observed. Competition of correct
`incorporation versus
`dCTP revealed an estimated K, of ~6-8 mM, suggesting a corresponding km of 0.14 s_1. Moderate elemental
`effects of 19-, 17-, and 34-fold reduction in rates were measured by substituting the a-thiotriphosphate
`analogues for dATP, dCTP, and dGTP, respectively, indicating that the chemistry step is partially rate-
`limiting. The absence of a burst of incorporation during the first turnover places the rate-limiting step at
`In contrast, the incorporation of
`a triphosphate binding induced conformational change before chemistry.
`triphosphate, dCTP, on a mismatched DNA substrate was saturable with a Km of 87 juM
`the next correct
`for dCTP, 4-fold higher than the Kd for the correct
`incorporation on duplex DNA, and a fccat of 0.025 s_1.
`A larger elemental effect of 60, however, suggests a rate-limiting chemistry step. The rate of pyro-
`phosphorolysis on a mismatched 3'-end is undetectable, indicating that pyrophosphorolysis does not play
`a proofreading role in replication. These results show convincingly that the T7 DNA polymerase discriminates
`against the incorrect triphosphate by an induced-fit conformational change and that, following misincor-
`poration, the enzyme then selects against the resultant mismatched DNA by a slow, rate-limiting chemistry
`step, thereby allowing sufficient time for the release of the mismatched DNA from the polymerase active
`site to be followed by exonucleolytic error correction.
`
`IVIodels to explain the high fidelity of DNA replication
`have long been proposed in the literature [reviewed in Loeb
`and Kunkel (1982)]. For example, several models have in-
`voked a role for pyrophosphorolysis in fidelity by mechanisms
`sometimes referred to as “kinetic proofreading” (Brutlag &
`Kornberg, 1972; Deutscher & Kornberg, 1969; Flopfield, 1976;
`It is, however, worth noting that the original
`Ninio, 1975).
`models have been proposed as purely mathematical constructs,
`and the only evidence in support of a role for pyro-
`phosphorolysis is the observation that polymerization is less
`accurate at high pyrophosphate concentrations (Kunkel et al.,
`1986). Although induced-fit mechanisms have been suggested,
`it has also been argued that induced-fit models cannot account
`for increased selectivity (Fersht, 1974, 1985; Page, 1986).
`Thus, because of a void of mechanistic information, almost
`all of these models take considerable liberties with invoking
`specific but unsubstantiated intermediates in the kinetic
`pathway of polymerization. Furthermore, most of the pub-
`lished mechanistic studies available on misincorporation rely
`solely on steady-state kinetic data, which, as we will show, are
`at best difficult to interpret and at worst extremely misleading.
`The only transient kinetic study on the mechanism of DNA
`polymerization relied upon the use of the Klenow fragment
`of the DNA repair enzyme, Pol I, and some surprising results
`were obtained, the most notable of which is that the polymerase
`
`fThis work was supported by the Paul Berg Professorship from Penn
`State University (K.A.J.).
`*To whom correspondence should be addressed.
`
`0006-2960/91 /0430-526S02.50/0
`
`kr
`E,DNA„.,= Ep DNA„
`
`*"•"
`
`E„ DNA„.,
`
`E, DNAn>1 + dNMP
`
`bound correct and incorrect dNTPs with nearly equal affinities
`(Kuchta et al., 1987, 1988).
`In this report, we propose a mechanism for DNA replication
`fidelity as a part of our kinetic study of the T7 DNA polym-
`erase (Patel et al., 1991; Donlin et al., 1991). The T7 system
`lends itself especially well to our approach to fidelity because
`its kinetic scheme has been completely solved (Patel et al.,
`1991; Donlin et al., 1991). Consequently, our
`results can be
`interpreted by a direct comparison of the differences between
`In general, T7 DNA
`correct and incorrect polymerization.
`polymerase constitutes a nearly ideal model system for any
`DNA replication studies because it functions in vivo as a true
`replication enzyme with a minimal number of components.
`The problem of replication fidelity is summarized in Scheme
`I. Conceptually, we divide the issue of fidelity into two parts.
`First, we are concerned with the mechanism of making an
`this primarily involves solving the kinetics of misin-
`error;
`corporation. Second, we are interested in the kinetic conse-
`quences of the error; here, we must determine the relative
`©1991 American Chemical Society
`
`Downloaded via ALBERT EINSTEIN COLG MEDICINE INC on March 29, 2021 at 17:17:19 (UTC).
`
`See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
`
`

`

`DNA Replication Fidelity
`
`Table I: Oligonucleotides
`25/36-mer
`5’-GCCTCGCAGCCGTCCAACCAACTCA
`CGGAGCGTCGGCAGGTTGGTTGAGTAGGTCTTGTTT-6’
`
`25A/36-mer
`
`5’-GCCTCGCAGCCGTCCAACCAACTCAA
`CGGAGCGTCGGCAGGTTGGTTGAGTAGGTCTTGTTT-5’
`
`contribution—the kinetic partitioning—between the four
`possible exit pathways from the central “enzyme-error” com-
`(1) pyrophosphorolysis, which represents the microscopic
`plex:
`reversal of polymerization, (2) incorporation of the next correct
`dNTP, (3) dissociation of the mismatched DNA from the
`transfer of the
`enzyme into free solution, and (4) direct
`mismatched DNA into the exonuclease site for repair.
`Paradoxically, the chief disadvantage of the T7 system lies
`in its high replication fidelity. Wild-type enzyme does not form
`detectable levels of stable misincorporations in vitro. This,
`as we will show, is due to the highly efficient 3'-5' proofreading
`exonuclease, which is particularly good at excising mismatches.
`For this reason, we have constructed an exonuclease-deficient
`mutant enzyme (D5A,E7A) (Patel et al., 1991) on the basis
`of sequence homology studies (Reha-Krantz, 1988a,b; Leavitt
`& Ito, 1989; Bernad et al., 1989). Using this mutant polym-
`erase, we are able to study the kinetic mechanism and con-
`sequences of misincorporation. On the basis of these studies,
`we propose here that, in a normal cycle of polymerization
`incorporation, a rate-limiting conformational
`involving correct
`change step in an induced-fit mechanism bears the primary
`burden of substrate selectivity. The issues of (1) Kmax versus
`Km discrimination in substrate selectivity, (2) the role of py-
`rophosphorolysis in error correction, and (3) the kinetic par-
`repair will all be
`titioning mechanism for exonucleolytic error
`discussed in the context of this induced-fit model.
`
`Experimental Procedures
`Materials
`Bacterial Strains, Plasmids, and Phage. Escherichia coli
`A179 (Hfr-C)(X)trx/l::^an) and plasmids of pGP5-3 and
`pGPl-3 were obtained from S. Tabor and C. C. Richardson
`(Harvard Medical School; Tabor & Richardson, 1987). The
`plasmid containing the exo" T7 gene 5, pGAl-14, was con-
`structed in this laboratory as described in the preceding paper
`(Patel et al., 1991).
`Proteins and Enzymes. E. coli thioredoxin was purified as
`described in the preceding paper. Klenow fragment (KF) was
`kindly provided by R. Kuchta and C. Catalano (The Penn-
`sylvania State University). T4 polynucleotide kinase was
`purchased from New England Biolabs.
`Nucleoside Triphosphates. dNTPs were purchased from
`Pharmacia Molecular Biologicals at >98% purity. ATP was
`purchased from Sigma.
`[a-
`[«-32P]dTTP,
`[a-32P]dCTP,
`32P]dATP, and [y-32P]ATP were purchased from New Eng-
`land Nuclear.
`Synthetic Oligonucleotides. Synthetic oligonucleotides (see
`Table I) used were synthesized on either an Applied Biosys-
`tems 380A DNA synthesizer or a Milligen/Biosearch 7500
`DNA synthesizer and purified by electrophoresis through a
`denaturing gel (20% acrylamide, 1.5% bisacrylamide, and 8M
`urea in Tris-borate buffer). The major DNA band was vis-
`ualized by UV shadowing and excised. DNA was electroeluted
`from the gel slice in an Elutrap apparatus purchased from
`Schleicher & Schuell.
`Triethylammonium bicarbonate
`(TEAB; 2 M, pH 7.5) was added to a final concentration of
`0.5 M, and the eluate was applied to an Alltech Maxi-Clean
`C|g cartridge. After being washed with 10 mM TEAB, pH
`
`Biochemistry, Vol. 30, No. 2, 1991
`527
`7.5, purified DNA was eluted in 50% ethanol. Concentrations
`of purified oligonucleotides were determined by UV absor-
`bance at 260 nm in 8 M urea the following extinction calcu-
`lated coefficients: 20-mer, t = 202 450; 25-mer, e = 249 040;
`25A-mer, e = 261 040; 36-mer, e = 377 000 cm2//imol.
`Duplex Oligonucleotides. Duplex oligonucleotides were
`annealed at room temperature in TE buffer containing 100
`mM NaCl. They were
`then purified by electrophoresis
`through a nondenaturing gel (20% acrylamide and 1.5% bis-
`acrylamide in Tris-borate buffer). The major DNA band was
`visualized, excised, and electroeluted as above.
`Methods
`25A/36-mer. Template-primer containing a 3'-terminal
`A-A mismatch was enzymatically synthesized by using exo"
`T7 DNA polymerase. The 25/36-mer (1 ^M) was incubated
`with enzyme (500 nM) and dATP (2 mM) for 5 min at room
`temperature. Reaction was quenched by the addition of
`EDTA to 50 mM, followed by two extractions with buffer-
`saturated phenol-chloroform (1:1). Unincorporated dATP
`and EDTA as well as residual phenol-chloroform were re-
`moved by centrifugation through Bio-Spin 30 centrifuge
`desalting columns.
`3'-nP-Labeled 25A j36-mer. The 25/36-mer (1 /aM) was
`incubated with enzyme (500 nM) and [a-32P]dATP (3000
`Ci/mmol at a final concentration of 2-3 /aM) for 45 min at
`room temperature. Cold dATP (2 mM) and an additional
`aliquot of enzyme (250 nM) were added, and incubation was
`continued for an additional 5 min. Workup of labeled DNA
`was as described above.
`Reaction Buffer, 5' 32P Labeling, Reconstitution of T7 DNA
`Polymerase, Rapid-Quench Experiments, Product Analysis
`by Denaturing Acrylamide Gels, and PEI-Cellulose TLC.
`These protocols were performed exactly as described in detail
`in the first paper in this series (Patel et al., 1991).
`Results
`Determination of Km and kCitfor Incorporation of the In-
`correct dNTPs. We began our studies by attempting to de-
`termine the steady-state kinetic parameters, Km and kml, for
`incorrect dNTP incorporation. The DNA substrate used was
`a 5'-labeled 25/36-mer, and the time courses of incorporation
`were monitored by analysis of the products on denaturing
`sequencing gels. The misincorporation of dATP and dGTP
`(versus A in the template) led to elongation of the 25-mer by
`one base, while the misincorporation of dCTP led to a series
`of four bands of sizes 26-29 bases. This resulted from the fact
`that, after the initial misincorporation to generate the 26-mer,
`dCTP was the correct base for the next two additions (27- and
`28- mers); at higher concentrations of dCTP and at longer time
`points, a second misincorporation on the 28-mer generated the
`29- mer. All four bands were excised, counted, and summed
`to yield the total products, thus defining the kinetics of the
`first misincorporation.
`The rates of misincorporation for the three incorrect dNTPs
`were determined over a range of dNTP concentrations from
`5 jiM up to 10 mM. The results in the millimolar range are
`shown in Figure 1A. Misincorporation rates were found to
`be linearly dependent on dNTP concentrations up to 4 mM.
`Beyond 4 mM, severe inhibition was observed, presumably due
`to inhibition of DNA binding. Regression analysis gave linear
`best fits to the data with the slopes defining the apparent
`second-order rate constant, kM/Km. The kal/Km values for
`dATP, dCTP, and dGTP were 91, 23, and 4.3 M"1 s"1,
`re-
`spectively. Although there was no indication of curvature in
`the data prior to the onset of the severe
`inhibition, we can
`
`

`

`528 Biochemistry, Vol. 30, No. 2, 1991
`A
`
`Wong et al.
`
`B
`
`I: Misincorporations. Panel A shows a plot of rates of
`figure
`misincorporation onto 25/36-mer as a function of dNTP concentration.
`The slopes of the solid lines yield the values of kM/Km. (A) dGTP,
`kcJKm = 4.2 M'1 s'1; ( ) dCTP, kal/Km = 23 M'1 s'1; (•) dATP,
`koat/Km = 91 M'1 s'1. The concentration dependence continues to
`be linear until 4 mM, beyond which the rate of misincorporation is
`severely inhibited. The dashed line shows an attempted hyperbolic
`fit to estimated lower limits for
`of 0.2 s'1 for
`of 8 mM and a
`the dCTP data with a resultant kaJKm of 25 M'1 s'1. Reactions were
`all carried out under steady-state conditions with 1 qM 25/36-mer
`and 5 nM enzyme. Reactions were quenched by the addition of EDTA
`to 125 mM, and the products were analyzed by denaturing sequencing
`gels. Quantitation of products was by excision and liquid scintillation
`counting of gel bands. Panel B shows the time course of a single-
`turnover dCTP misincorporation at 250 nM enzyme, 250 nM DNA,
`and 1 mM dCTP. The best fit of the data to a single exponential
`yields a rate of 0.021 s'1,
`indicating that the rate during the first
`is the same as the steady-state rate.
`turnover
`estimate the Km on the basis of data describing the inhibition
`of correct
`incorporation by the incorrect dCTP (see below).
`The dashed line in Figure 1A represents the calculated hy-
`perbola for a Km of 8 mM at a kml of 0.2 s'1 (kM/Km = 25
`M'1 s'1), setting a lower limit on the magnitudes of Km and
`kcaf
`While these experiments were carried out under steady-state
`conditions in which DNA was in 200-fold excess over enzyme,
`an experiment performed under pre-steady-state conditions,
`with a 1:1 ratio of DNA to enzyme, indicated that the rate
`during the first turnover was the same as that for subsequent
`turnovers (Figure IB). Even given the subsaturating con-
`centration of dNTP used in the single-turnover experiments,
`biphasic kinetics would have been expected if the steady-state
`rate measured some rate-limiting step after chemistry, given
`a detection limit of 5%. Consequently, within this limit, we
`conclude that the steady-state rates reported in Figure 1 reflect
`the rate of misincorporation and not
`the rate of product
`
`Inhibition of correct incorporation by incorrect dNTP. Time
`figure
`2:
`course of correct
`incorporation at 10 qM dTTP in the presence of
`0 (O), 0.5 ( ), 1.0 (A), and 2.0 mM (V) of the incorrect dCTP. Solid
`curves are best fits of the data to single exponentials with rates of
`Inset shows rate of correct
`83, 76, 69, and 62 s'1, respectively.
`incorporation as a function of dCTP concentration fitted to a hyperbola
`to yield an extrapolated Kt for dCTP of 4 ± 2 mM. Enzyme and DNA
`concentrations were 250 nM each. The degree of inhibition observed
`is so small that the error on the K, is correspondingly large. Taken
`together with the estimated lower limit of 8 mM observed in the
`experiment described in Figure 1, we would estimate a reasonable
`range for the Km of dCTP to be around 6-8 mM. See text for further
`descriptions on the practical constraints in designing these experiments.
`(DNA) dissociation, as is the case for correct
`incorporation.
`In a separate attempt to extract discrete values of kat and
`Km for misincorporation, a competition experiment was per-
`formed to determine a Kt for inhibition of correct incorporation
`(dTTP) by an incorrect dNTP (dCTP). DNA (5'-labeled
`25/36-mer at 250 nM) was preincubated with 250 nM exo'
`enzyme and then was reacted with 10 qM Mg-dTTP (correct
`the rate of the burst of correct
`dNTP) to measure
`incorpo-
`ration in the presence of 0, 0.5, 1, and 2 mM dCTP (Figure
`2). From the concentration dependence of the effect of the
`incorrect dNTP on the rate of correct incorporation, a Kt for
`the incorrect dNTP can be very roughly estimated at 4 ± 2
`mM. However, the overall degree of inhibition observed was
`slight, and therefore, the value of K{ thus derived was heavily
`extrapolated. Taken together with the limit of 8 mM estimated
`in Figure 1A (see above description), we estimate an ap-
`proximate range for Km of 6-8 mM to be within reasonable
`limits of experimental errors. Unfortunately, the severe in-
`hibition observed at higher dNTP concentrations precluded
`a better measurement of the Aim for misincorporation.
`Elemental Effect for Misincorporation.
`In order to estimate
`to which the chemistry step is rate-limiting, we
`the extent
`compared the rates of misincorporations of dATP, dCTP, and
`dGTP with their a-thio analogues dATP(aS), dCTP(aS), and
`dGTP(aS). A full elemental effect, resulting in a 100-fold
`reduction of rate when the thio analogues are substituted for
`the oxy-dNTPs, would indicate a completely rate-limiting
`chemistry step. For the incorporation of the correct nucleoside
`triphosphate, dTTP, a small elemental effect of 3 has been
`observed (Patel et al., 1991), indicating that the chemistry step
`is not rate-limiting during a normal cycle of polymerization.
`In these experiments, 5'-labeled 25/36-mer was used at a
`5-fold molar excess over enzyme. Data were quantitated by
`excision and liquid scintillation counting of bands from a
`denaturing sequencing gel. The time courses of incorporation
`of the oxy-dNTPs and the a-thio-dNTPs are shown in Figure
`3. Substitution by the thio analogues resulted in reduction
`in incorporation rates by factors of 19, 17, and 37 for dATP,
`dCTP, and dGTP, respectively. These moderate values may
`
`

`

`DNA Replication Fidelity
`dATP
`
`dATPaS
`
`Biochemistry, Vol. 30, No. 2, 1991
`dCTPaS
`
`dCTP
`
`529
`
`0 10 20 30 40 50
`
`2
`
`5 10 20
`
`0
`
`10 20 30 40 50 2
`
`5 10 20
`
`0 10 20 30 40 50
`
`2
`
`5 10 20
`
`0
`
`10 20 30 40 50 2 5
`
`10 20
`
`Time
`
`(min)
`
`Time
`
`(min)
`
`dGTP
`
`dGTPaS
`
`0
`
`10
`
`20 30 40 50
`
`2
`
`5 10 20
`
`0
`
`10 20 30 40 50 2
`
`5 10 20
`
`3: Elemental effects on misincorporations. Time course of misincorporation of dATP, dCTP, and dGTP (•) and their a-thio analogues
`figure
`dATP(oS), dCTP(oS), and dGTP(«S) (O). Panel A shows dATP and dATP(«S) misincorporations at 2.4 X 10'2and 1.3 X I0"3 s'1, respectively.
`Panel B shows dCTP and dCTP(uS) misincorporations at 1.6 X 10~2 and 9.3 X I0-4 s
`, respectively. As noted in the text, because dCTP
`triphosphate for the next two incorporations, a ladder of products from 26 to 29 bases was observed. The difference in banding
`is the correct
`patterns between the dCTP and dCTP(nrS) lanes is illustrative of the difference in elemental effects on dCTP misincorporation and next correct
`incorporation. With dCTP, the incorporation of the correct dCTPs after the misincorporation is faster than misincorporation and therefore
`no net accumulation of the 25C-mer is observed. However, because the elemental effect on misincorporation of dCTP is smaller than that
`incorporation, the rate of addition of the next dCTP(oS) in the correctly base-paired positions is now actually slower than
`for the next correct
`the rate of misincorporation. resulting in the accumulation of the 25C-mcr. Panel C shows dGTP and dGTP(aS) misincorporations at 7.1
`X 10 3 and 1.9 X I O'4 s-1, respectively. Overall elemental effects for dATP, dCTP, and dGTP are 19, 17, and 37. Reaction incubations contained
`I |iM 25/36-mcr, 200 nM enzyme, and 250 mM. 750 jiM, and 1.5 mM dATP, dCTP, and dGTP, respectively, in standard reaction buffer.
`
`

`

`530 Biochemistry, Vol. 30, No. 2, 1991
`
`Wong et al.
`
`Scheme II
`E*D„
`
`i
`
`+ dNTP
`
`E*D„*dNTP =•
`
`E.D^.PPj
`
`E*D„*PPi
`
`In this experiment,
`minal mismatch by direct measurement.
`25/36-mer was enzymatically elongated to 3'-32P-labeled
`25A*/36-mer as described under Experimental Procedures.
`Single-turnover experiments with equimolar enzyme and DNA
`were carried out
`in reactions containing inorganic pyro-
`phosphate at concentrations ranging from 0.1 to 20 mM. The
`time courses of the reactions were monitored by TLC on
`PEI-cellulose plates, also described under Experimental Pro-
`cedures. Pyrophosphorolysis was monitored by the appearance
`of radioactive dATP. For reaction times of up to 30 min,
`pyrophosphorolysis was not detected (data not shown). On
`this time scale, a small amount of radioactive dAMP was
`detected, which could be attributed to the low residual exo-
`nuclease activity of the enzyme on mismatched DNA. This
`rate of excision was known to be in the range of 10-5— 10-4 s'1.
`Thus, an upper limit of 10'4 can be set for the rate of pyro-
`phosphorolysis of a mismatch at 20 mM pyrophosphate. This,
`in turn, yields an upper limit for an apparent second-order rate
`constant, kax/Km, of 0.05 M'1 s'1.
`In contrast, the rate of
`pyrophosphorolysis on a correctly base-paired duplex DNA
`proceeds at 0.5-2.0 s'1 with a Km of 2 mM, giving a k<.iX/Km
`of (0.25-1) X 103 M'1 s'1 (Patel et al., 1991). Thus pyro-
`phosphorolysis selects against the correctly base-paired primer
`terminus by a factor of at least 5000.
`In a separate experiment, 5'-labeled 25A/36-mer was in-
`cubated with exo" enzyme in the presence and absence of 2
`mM pyrophosphate. The reaction was monitored on a dena-
`turing sequencing gel (Figure 5). Without added pyro-
`the residual exonuclease removed the terminal
`phosphate,
`It
`mismatch to generate the 25-mer at a rate of 5 X 10'5 s'1.
`then paused, unable to excise the next base, which was properly
`base paired. Discrimination by the residual exonuclease of
`the mutant enzyme for mismatched base pairs is seen also with
`the wild-type enzyme (Donlin et al., 1991) but not to the extent
`seen here with the exo' mutant.
`In the reaction containing
`pyrophosphate, however, a ladder of oligomers <25 bases long
`first glance this may seem to suggest pyro-
`resulted. At
`phosphorolysis; but in actual fact, when the gel slices were
`quantitated by liquid scintillation counting and the results
`plotted, it was found that the sum of all the smaller bands in
`the reaction with pyrophosphate was at all times equal to the
`total accumulation of the 25-mer in the reaction without py-
`In other words, the removal of the terminal
`rophosphate.
`mismatch occurred at the same rate in both reactions and can
`therefore be attributed solely to the residual exonuclease ac-
`tivity. Since the product of the initial mismatch excision is
`a correctly base-paired duplex DNA, a suitable substrate for
`pyrophosphorolysis, further degradation of the 25-mer in the
`reaction containing pyrophosphate yielded a ladder of small
`oligomers.
`Km and ka(( of a 3'-Mismatched DNA Substrate.
`In the
`initial misincorporation experiments with dCTP, the enzyme
`was observed to polymerize over a mismatch when the next
`correct dNTP was present. We took advantage of this fact
`in the following experiment in order to determine the Km for
`Increasing concentrations of DNA,
`the mismatched DNA.
`ranging from 5 nM to 1 jiM of 5Mabeled 25A/36-mer, were
`reacted with 1 nM enzyme and 200 /uM dCTP. The reaction
`was quenched after various times, and the products were an-
`
`Inhibition of misincorporation by pyrophosphate. The rates
`figure
`4:
`of misincorporation of dATP (500 mM) into 25/36-mer (500 nM)
`at 500 nM enzyme at 0 (•), 0.5 (A), 2.5 (O), and 5.0 mM (V)
`Inset shows the plot of the rates fitted to a hyperbola
`pyrophosphate.
`to yield a K-x of 4 mM. Note that the inhibition is competitive, as
`only the rate of incorporation is reduced while the full amplitude of
`misincorporation is obtained at all pyrophosphate concentrations.
`imply that the chemistry step may be partially rate-deter-
`mining. Compared with the incorporation of the correct
`dNTP, where the chemistry step is much faster than the
`conformational change, the larger elemental effects observed
`here for the incorrect nucleoside triphosphates would argue
`for a chemistry step only slightly faster than the conforma-
`tional change. Nonetheless, since a full elemental effect is not
`observed, it can be concluded that the conformational change
`is at least 3-5-fold slower than chemistry.
`Elementary effect experiments were also carried out for
`dCTP(aS) at 2 mM and 0.2 mM (data not shown) with the
`same results, indicating that (1) the a-thio-dNTPs were bound
`with roughly the same affinities (Kd values) as the oxy-dNTPs
`and (2) incorrect dNTP binding was not rate-limiting.
`Inhibition by Inorganic Pyrophosphate. To address the issue
`of a possible contribution of pyrophosphorolysis toward fidelity,
`we monitored the inhibitory effects of pyrophosphate on mis-
`In a single-turnover experiment with equimolar
`incorporation.
`enzyme and DNA, the time course of misincorporation of
`dATP into 5'-labeled 25/36-mer was monitored in the presence
`of increasing concentrations of inorganic pyrophosphate. The
`results are shown in Figure 4. The addition of 0.5, 2.5, and
`5.0 mM pyrophosphate inhibited the rates of misincorporation
`by factors of 1.1, 1.6, and 2.1, respectively, giving an estimated
`apparent Kx of 4 mM for pyrophosphate (see inset).
`However, regardless of the concentration of pyrophosphate
`the same reaction amplitude was observed. This
`present,
`contrasted with results of an analogous experiment performed
`with dTTP, the correct nucleoside triphosphate (Patel et al.,
`1991), where pyrophosphate reduced both the rate of dTTP
`incorporation (K-x of 2 mM) and the amplitude of incorpora-
`tion. While the reduction in the rate of incorporation can be
`thought of as competitive binding between pyrophosphate and
`substrate dNTP, as shown on the left-hand side of Scheme
`II, the reduction in the amplitude of incorporation reflects an
`internal equilibrium driven backward by the addition of py-
`rophosphate from the right side of the equation. Consequently,
`the full amplitude attained for misincorporation in the presence
`of high concentrations of pyrophosphate suggests that the
`reaction is not kinetically accessible. This implies that
`reverse
`pyrophosphorolysis did not remove
`the misincorporated nu-
`cleotide on the time scale of the single turnover.
`Pyrophosphorolysis of a Mismatched 3' dNMP. We then
`sought to determine the rate of pyrophosphorolysis on a ter-
`
`

`

`DNA Replication Fidelity
`
`+PFi
`
`-pPi
`
`0
`
`2.5 5.0
`
`10
`
`15
`
`20
`
`30
`
`0
`
`15
`
`30
`
`Biochemistry, Vol. 30. No. 2, 1991
`
`531
`
`7: Rate of dissociation of 25A/36-mer from enzyme. Exo"
`figure
`enzyme was preincubatcd with 25A/36-mer. The reaction was in-
`itiated with the addi