Proc. Natl. Acad. Sci. USA
`Vol. 92, pp. 6339-6343, July 1995
`Biochemistry
`
`A single residue in DNA polymerases of the Escherichia coli DNA
`polymerase I family is critical for distinguishing between deoxy-
`and dideoxyribonucleotides
`(DNA sequencing/dideoxynucleotides/T7 DNA polymerase/Taq DNA polymerase/fidelity)
`STANLEY TABOR AND CHARLES C. RICHARDSON
`Department of Biological Chemistry and Molecular Pharmacology, 240 Longwood Avenue, Harvard Medical School, Boston, MA 02115
`
`Contributed by Charles C. Richardson, March 28, 1995
`
`Bacteriophage T7 DNA polymerase effi-
`ABSTRACT
`ciently incorporates a chain-terminating dideoxynucleotide
`into DNA, in contrast to the DNA polymerases from Esche-
`richia coli and Thermus aquaticus. The molecular basis for this
`difference has been determined by constructing active site
`hybrids of these polymerases. A single hydroxyl group on the
`polypeptide chain is critical for selectivity. Replacing tyrosine-
`526 of T7 DNA polymerase with phenylalanine increases dis-
`crimination against the four dideoxynucleotides by >2000-
`fold, while replacing the phenylalanine at the homologous
`position in E. coli DNA polymerase I (position 762) or T.
`aquaticus DNA polymerase (position 667) with tyrosine de-
`creases discrimination against the four dideoxynucleotides
`250- to 8000-fold. These mutations allow the engineering of
`new DNA polymerases with enhanced properties for use in
`DNA sequence analysis.
`
`All known DNA polymerases can be placed in one of four
`families based on sequence homologies (1, 2). Although Esch-
`erichia coli DNA polymerase I, Thermus aquaticus DNA
`polymerase (Taq DNA polymerase), and bacteriophage T7
`DNA polymerase are members of the DNA polymerase I (Pol
`I) family, a number of their properties differ significantly (3).
`First, T7 DNA polymerase is responsible for the replication of
`a genome and, as such, interacts with other replication pro-
`teins, whereas E. coli DNA polymerase I and Taq DNA
`polymerase are mainly responsible for repair and recombina-
`tion. Second, the three polymerases have different exonuclease
`activities: T7 has only a 3' --
`5' activity, Taq has only a 5' --
`3' activity, and E. coli DNA polymerase I has both 3' --
`5' and
`5' --
`3' activities. The exonuclease activities reside in separate
`domains at the amino termini and can be inactivated selectively
`by genetic modification. Third, the thermostability of Taq
`DNA polymerase distinguishes it from the other two.
`A fourth difference is their relative abilities to distinguish
`between a deoxy- and a dideoxyribose in the nucleoside
`triphosphate. E. coli DNA polymerase I (4-6) and Taq DNA
`polymerase (7) incorporate deoxyiucleotides at a rate that is
`several hundred to several thousand times that of di-
`deoxynucleotides, while T7 DNA polymerase incorporates
`dideoxynucleotides much more efficiently, preferring de-
`oxynucleotides by only a factor of three (5, 6). To determine
`the molecular basis for this difference we constructed active
`site hybrids of the DNA polymerases and examined their
`ability to use ddNTPs relative to dNTPs. Inasmuch as chain
`termination by dideoxynucleotides is the basis of all commonly
`used enzymatic methods of DNA sequencing (5, 8), an un-
`derstanding of this difference has utility for improving the
`DNA polymerases used for this purpose.
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement" in
`accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`6339
`
`METHODS
`Construction ofHybrid Genes. Hybrid genes were constructed
`by using synthetic oligonucleotides and the polymerase chain
`reaction. Hybrids of the T7 DNA polymerase gene were ex-
`pressed under the control of the lac promoter in the vector
`pUC18; the parent vector (pGP5-12) contains the gene for T7
`DNA polymerase with a deletion that encodes amino acid
`residues 118-145, inactivating the exonuclease (9). Hybrids of the
`E. coli DNA polymerase I gene were constructed in pKLEN-1,
`which encodes the large fragment of E. coli DNA polymerase I
`(beginning at residue 324) under the control of a T7 RNA
`polymerase promoter. Hybrids of Taq DNA polymerase were
`constructed in pTQA-1, which encodes a truncated fragment of
`Taq DNA polymerase (beginning at residue 289) under the
`control of a T7 RNA polymerase promoter. For characterization
`of the purified Taq hybrid polymerase C-Q5 (see Table 1), a
`fragment containing the mutation in pTQA-1 was transferred into
`the full-length Taq DNA polymerase gene.
`SDS-DNA Activity Gel Analysis. SDS-DNA activity gel
`analysis was carried out as described (10, 11). Ten milliliters of
`induced cells was pelleted and resuspended in 0.3 ml of 50 mM
`Tris HCl (pH 6.8). Two to 20 ,ul of cells was added to 60 ,ul of
`50 mM Tris-HCl, pH 6.8/15% glycerol/100 mM mercapto-
`ethanol/0.02% bromphenol blue/0.5% SDS. Samples were
`heated at 37°C for 5 min prior to loading 20-,ul samples on
`duplicate gels. After electrophoresis (5 V/cm for 13 hr at
`13°C), SDS was removed from the gels by soaking four times
`in 800 ml of renaturation buffer (50 mM Tris-HCl, pH 7.5/15%
`glycerol/6 mM magnesium acetate/40 mM KCl/400 ,ug of
`bovine serum albumin per ml/1 mM dithiothreitol) at 4°C over
`16 hr. DNA polymerase activity in each gel was assayed in a
`6-ml mixture of renaturation buffer containing all four dNTPs
`(each at 1.5 ,uM), 30 ,uCi of [a-32P]dATP (1 Ci = 37 GBq), 30
`,uM ddTTP (where present), and 1 ,LM E. coli thioredoxin.
`Reactions were carried out for 2 hr at either room temperature
`(E. coli and T7) or 70°C (Taq). Unincorporated [a-32P]dATP
`was removed from the gels by soaking four times in 300 ml of
`5% trichloroacetic acid/1% PP; over 2 hr. The gels were
`transferred to filter paper, dried, and autoradiographed.
`Purification of Wild-Type (WT) and Hybrid Proteins.
`"WT" DNA polymerase refers to the enzyme that contains the
`WT residue (tyrosine in T7 and phenylalanine in E. coli and
`Taq) at the ribose selectivity site, while the mutant or hybrid
`enzyme has the opposite residue. "WT T7 DNA polymerase"
`has a deletion of 28 amino acid residues (118-145) in the
`exonuclease domain (9) and is a one-to-one complex of the T7
`gene 5 protein and E. coli thioredoxin. "WT E. coli DNA
`polymerase I" is the large fragment of E. coli DNA polymerase
`I, or Klenow fragment, missing the 5' -* 3' exonuclease
`activity. "WT Taq DNA polymerase" is full-length Taq DNA
`polymerase. WT Taq DNA polymerase was purchased from
`
`Abbreviations: Pol I, DNA polymerase I; WT, wild-type.
`
`Columbia Ex. 2059
`Illumina, Inc. v. The Trustees
`of Columbia University in the
`City of New York
`IPR2020-00988, -01065,
`-01177, -01125, -01323
`
`

`

`Proc. Natl. Acad. Sci. USA
`Vol. 92, pp. 6339-6343, July 1995
`Biochemistry
`
`A single residue in DNA polymerases of the Escherichia coli DNA
`polymerase I family is critical for distinguishing between deoxy-
`and dideoxyribonucleotides
`(DNA sequencing/dideoxynucleotides/T7 DNA polymerase/Taq DNA polymerase/fidelity)
`STANLEY TABOR AND CHARLES C. RICHARDSON
`Department of Biological Chemistry and Molecular Pharmacology, 240 Longwood Avenue, Harvard Medical School, Boston, MA 02115
`
`Contributed by Charles C. Richardson, March 28, 1995
`
`Bacteriophage T7 DNA polymerase effi-
`ABSTRACT
`ciently incorporates a chain-terminating dideoxynucleotide
`into DNA, in contrast to the DNA polymerases from Esche-
`richia coli and Thermus aquaticus. The molecular basis for this
`difference has been determined by constructing active site
`hybrids of these polymerases. A single hydroxyl group on the
`polypeptide chain is critical for selectivity. Replacing tyrosine-
`526 of T7 DNA polymerase with phenylalanine increases dis-
`crimination against the four dideoxynucleotides by >2000-
`fold, while replacing the phenylalanine at the homologous
`position in E. coli DNA polymerase I (position 762) or T.
`aquaticus DNA polymerase (position 667) with tyrosine de-
`creases discrimination against the four dideoxynucleotides
`250- to 8000-fold. These mutations allow the engineering of
`new DNA polymerases with enhanced properties for use in
`DNA sequence analysis.
`
`All known DNA polymerases can be placed in one of four
`families based on sequence homologies (1, 2). Although Esch-
`erichia coli DNA polymerase I, Thermus aquaticus DNA
`polymerase (Taq DNA polymerase), and bacteriophage T7
`DNA polymerase are members of the DNA polymerase I (Pol
`I) family, a number of their properties differ significantly (3).
`First, T7 DNA polymerase is responsible for the replication of
`a genome and, as such, interacts with other replication pro-
`teins, whereas E. coli DNA polymerase I and Taq DNA
`polymerase are mainly responsible for repair and recombina-
`tion. Second, the three polymerases have different exonuclease
`activities: T7 has only a 3' --
`5' activity, Taq has only a 5' --
`3' activity, and E. coli DNA polymerase I has both 3' --
`5' and
`5' --
`3' activities. The exonuclease activities reside in separate
`domains at the amino termini and can be inactivated selectively
`by genetic modification. Third, the thermostability of Taq
`DNA polymerase distinguishes it from the other two.
`A fourth difference is their relative abilities to distinguish
`between a deoxy- and a dideoxyribose in the nucleoside
`triphosphate. E. coli DNA polymerase I (4-6) and Taq DNA
`polymerase (7) incorporate deoxyiucleotides at a rate that is
`several hundred to several thousand times that of di-
`deoxynucleotides, while T7 DNA polymerase incorporates
`dideoxynucleotides much more efficiently, preferring de-
`oxynucleotides by only a factor of three (5, 6). To determine
`the molecular basis for this difference we constructed active
`site hybrids of the DNA polymerases and examined their
`ability to use ddNTPs relative to dNTPs. Inasmuch as chain
`termination by dideoxynucleotides is the basis of all commonly
`used enzymatic methods of DNA sequencing (5, 8), an un-
`derstanding of this difference has utility for improving the
`DNA polymerases used for this purpose.
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement" in
`accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`6339
`
`METHODS
`Construction ofHybrid Genes. Hybrid genes were constructed
`by using synthetic oligonucleotides and the polymerase chain
`reaction. Hybrids of the T7 DNA polymerase gene were ex-
`pressed under the control of the lac promoter in the vector
`pUC18; the parent vector (pGP5-12) contains the gene for T7
`DNA polymerase with a deletion that encodes amino acid
`residues 118-145, inactivating the exonuclease (9). Hybrids of the
`E. coli DNA polymerase I gene were constructed in pKLEN-1,
`which encodes the large fragment of E. coli DNA polymerase I
`(beginning at residue 324) under the control of a T7 RNA
`polymerase promoter. Hybrids of Taq DNA polymerase were
`constructed in pTQA-1, which encodes a truncated fragment of
`Taq DNA polymerase (beginning at residue 289) under the
`control of a T7 RNA polymerase promoter. For characterization
`of the purified Taq hybrid polymerase C-Q5 (see Table 1), a
`fragment containing the mutation in pTQA-1 was transferred into
`the full-length Taq DNA polymerase gene.
`SDS-DNA Activity Gel Analysis. SDS-DNA activity gel
`analysis was carried out as described (10, 11). Ten milliliters of
`induced cells was pelleted and resuspended in 0.3 ml of 50 mM
`Tris HCl (pH 6.8). Two to 20 ,ul of cells was added to 60 ,ul of
`50 mM Tris-HCl, pH 6.8/15% glycerol/100 mM mercapto-
`ethanol/0.02% bromphenol blue/0.5% SDS. Samples were
`heated at 37°C for 5 min prior to loading 20-,ul samples on
`duplicate gels. After electrophoresis (5 V/cm for 13 hr at
`13°C), SDS was removed from the gels by soaking four times
`in 800 ml of renaturation buffer (50 mM Tris-HCl, pH 7.5/15%
`glycerol/6 mM magnesium acetate/40 mM KCl/400 ,ug of
`bovine serum albumin per ml/1 mM dithiothreitol) at 4°C over
`16 hr. DNA polymerase activity in each gel was assayed in a
`6-ml mixture of renaturation buffer containing all four dNTPs
`(each at 1.5 ,uM), 30 ,uCi of [a-32P]dATP (1 Ci = 37 GBq), 30
`,uM ddTTP (where present), and 1 ,LM E. coli thioredoxin.
`Reactions were carried out for 2 hr at either room temperature
`(E. coli and T7) or 70°C (Taq). Unincorporated [a-32P]dATP
`was removed from the gels by soaking four times in 300 ml of
`5% trichloroacetic acid/1% PP; over 2 hr. The gels were
`transferred to filter paper, dried, and autoradiographed.
`Purification of Wild-Type (WT) and Hybrid Proteins.
`"WT" DNA polymerase refers to the enzyme that contains the
`WT residue (tyrosine in T7 and phenylalanine in E. coli and
`Taq) at the ribose selectivity site, while the mutant or hybrid
`enzyme has the opposite residue. "WT T7 DNA polymerase"
`has a deletion of 28 amino acid residues (118-145) in the
`exonuclease domain (9) and is a one-to-one complex of the T7
`gene 5 protein and E. coli thioredoxin. "WT E. coli DNA
`polymerase I" is the large fragment of E. coli DNA polymerase
`I, or Klenow fragment, missing the 5' -* 3' exonuclease
`activity. "WT Taq DNA polymerase" is full-length Taq DNA
`polymerase. WT Taq DNA polymerase was purchased from
`
`Abbreviations: Pol I, DNA polymerase I; WT, wild-type.
`
`

`

`6340
`
`Biochemistry: Tabor and Richardson
`
`Proc. Natl. Acad. Sci. USA 92 (1995)
`
`Perkin-Elmer (AmpliTaq) and the other five enzymes were
`purified using standard procedures. Polymerase activities in
`extracts and with purified proteins were determined by mea-
`suring the incorporation of [3H]dTMP into a primed M13
`template (9).
`Incorporation Rate Ratios of dNMPs to ddNMPs. Relative
`incorporation rates were determined by gel analysis of a
`32P-end-labeled primer annealed to single-stranded M13 DNA
`extended using a fixed ratio of dNTPs to each ddNTP. React-
`ion mixtures (20 ,ul) contained 0.5 pmol of [5'-32P]GTTTTC-
`CCAGTCACGACGTTGTAAAACGACGGCCAGTGCCA
`(500,000 cpm/pmol) annealed to single-stranded M13 mGP1-2
`DNA (5)/25 mM Tris-HCl, pH 8.0/5 mM MgCl2/5 mM di-
`thiothreitol/all four dNTPs (each at 120 ,uM)/one of the four
`ddNTPs at 20 ,uM/10 ng of yeast inorganic pyrophosphatase
`(12)/10 ng of the indicated DNA polymerase. Reactions were
`carried out for 15 min at 37°C (E. coli and T7) or 70°C (Taq).
`The reactions were terminated by the addition of 20 ,ul of 80%
`formamide/10 mM EDTA/0.02% bromphenol blue, and the
`samples were heated at 90°C for 2 min immediately prior to
`loading onto an 8% denaturing polyacrylamide gel.
`The relative rate of incorporation of dNMPs to each ddNMP
`for the three DNA polymerases containing tyrosine at the
`ribose selectivity site was determined by quantitative analysis
`of the gel shown in Fig. 3. The radioactive bands in each lane
`were analyzed with a PhosphorImager, and the data were fit to
`an exponential decay curve to obtain the apparent ratio of
`dNTPs to ddNTPs. The incorporation rate ratio is the ratio of
`the apparent to the actual ratio of each dNTP to ddNTP. For
`each of the three DNA polymerases with phenylalanine, the
`relative rate of incorporation was determined by comparing
`the average length of fragments generated using varying dNTP
`to ddNTP ratios to those generated using the homologous
`polymerase with tyrosine; the increase in discrimination
`against ddNTPs with phenylalanine is the ratio of the dNTP/
`ddNTP ratios required to produce fragments of the same
`average length for the two polymerases.
`
`RESULTS
`Analysis of T7 DNA Polymerase/E. coli DNA Polymerase I
`Hybrid Genes. The three-dimensional structure ofE. coli DNA
`polymerase I is known (13-15). The strongest homology with
`T7 DNA polymerase is in the crevice responsible for binding
`DNA and dNTPs (16). For the five most conserved regions,
`ranging in size from 14 to 29 amino acid residues, we con-
`structed hybrid genes in which the DNA encoding each seg-
`ment in E. coli DNA polymerase I was substituted for the
`homologous segment in the gene for T7 DNA polymerase (Fig.
`1). Activities of extracts ranged from <1% that of WT T7 DNA
`polymerase for hybrids containing regions A, D, and E, 2% for
`region C, and 80% for region B.
`To screen all the hybrids and avoid interference by the host
`DNA polymerases, we used an SDS-DNA activity gel assay for
`initial characterization (10, 11). SDS was added to cells
`containing the overproduced hybrid polymerases and the
`proteins were separated by electrophoresis in a polyacrylamide
`gel containing single-stranded DNA. After electrophoresis,
`the SDS was removed, allowing the proteins to renature, and
`polymerase activity was determined directly in the gel by
`incubation with [a-32P]dNTPs. By comparing reactions carried
`out on duplicate gels in the presence and absence of ddTTP,
`we determined the ability of ddTTP to inhibit DNA synthesis
`by each of the hybrid polymerases (Fig. 2).
`In the absence of ddTTP, DNA synthesis was relatively
`efficient with the hybrid polymerases containing regions A, B,
`and C of E. coli polymerase when compared to WT T7 DNA
`polymerase, while those containing regions D and E of E. coli
`polymerase exhibited <1% activity. As shown in the lower
`panel of Fig. 2A, whereas WT T7 DNA polymerase was in-
`
`Structure of the large fragment of E. coli DNA polymerase
`FIG. 1.
`I showing the regions substituted into T7 DNA polymerase. The
`regions A-E (shown in different colors) were removed from E. coli
`DNA polymerase I and placed into T7 DNA polymerase. The specific
`residues substituted were as follows: A, 666-682 (E. coli) -- 427-444
`(T7); B, 710-734 (E. coli) -* 480-506 (T7); C, 754-767 (E. coli) --
`518-531 (T7): D, 843-867 (E. coli) -- 609-633 (T7); E, 914-928 (E.
`coli) -- 690-704 (T7). The structure shown was derived from the
`coordinates determined by Beese et at (15). The letters within
`parentheses define specific helices (13, 15). The residue shown in helix
`0 (region C) is phenylalanine-762.
`
`hibited several 100-fold by a 20-fold excess of ddTTP over
`dTTP, DNA synthesis by the hybrid containing region C was
`resistant to this level of ddTTP. To define more precisely the
`residues responsible for this difference, we analyzed additional
`
`A T7DNAPol
`(-) W rA B C
`
`.....
`
`B Pol I
`
`D
`
`E
`
`K K6
`
`-ddTTP
`
`+ddTTP
`
`::. ...
`.....-. ......
`
`...
`
`.....
`
`:...w"
`
`FIG. 2. SDS-DNA activity gel analysis of hybrid DNA poly-
`merases. Cell aliquots were loaded on duplicate gels, and after
`electrophoresis, the SDS was removed and the polymerases were
`assayed for their ability to incorporate [a-32P]dAMP into single-
`stranded calf thymus DNA in the absence (-) or presence (+) of
`ddTTP. (A) Analysis of WT T7 DNA polymerase I (WT) and hybrids
`in which regions A-E of T7 DNA polymerase were replaced with the
`homologous regions from E. coli DNA polymerase I (Fig. 1). C-T8
`corresponds to cells containing the T7 DNA polymerase hybrid with
`the single Y526F mutation (Table 1). In the leftmost lane (-), the cells
`contain the control vector pUC18. The leftmost lane and those
`containing hybrids A, C, D, and E contain 10 times the amount of
`induced cells as in the other lanes to compensate for the low activity of
`these polymerases; the band visible at the top of each of these lanes is the
`result of DNA synthesis by the host E. coli DNA polymerase I (arrow-
`head). (B) Analysis of WT E. coli DNA polymerase I (WT) and E. coli
`DNA polymerase I hybrids C-K1 (substitution of 14 residues from T7
`DNA polymerase) and C-K6 (single F762Y mutation) (Table 1).
`
`

`

`..=_,...
`._ _wr_._ _
`........:..::
`
`Proc. Natl. Acad. Sci. USA 92 (1995)
`
`6341
`
`effect of substituting region C of T7 DNA polymerase for the
`homologous region in Taq DNA polymerase (Table 1). Substi-
`tutions that include tyrosine for phenylalanine at residue 667 of
`Taq DNA polymerase increased dramatically the ability of
`ddTTP to inhibit DNA synthesis.
`Analysis of Purified Proteins with Tyrosine Versus Phenyl-
`alanine at Ribose Selectivity Position. To analyze quantita-
`tively the effect of tyrosine versus phenylalanine at the ribose
`selectivity site, we purified the three hybrid polymerases, T7
`Y526F, E. coli F762Y, and Taq F667Y, and compared their
`properties to the WT DNA polymerases. The specific poly-
`merase activities of each of the three hybrid proteins were
`within 25% of the respective WT enzymes. The autoradio-
`graph in Fig. 3 shows the relative ability of the three WT and
`hybrid polymerases to incorporate each of the four ddNMPs.
`A 5'-32P-labeled primer annealed to a single-stranded M13
`template was extended by each of the DNA polymerases in the
`presence of a fixed (6:1) ratio of dNTP to each of the four
`ddNTPs. For all three DNA polymerases, the presence of
`phenylalanine resulted in strong discrimination against each of
`the four ddNTPs, while the presence of tyrosine resulted in
`efficient incorporation of all four ddNMPs.
`We determined quantitatively the relative rate of incorpo-
`ration of each ddNMP compared with the corresponding
`dNMP for each of the DNA polymerases shown in Fig. 3
`(Table 2). Whereas with WT T7 DNA polymerase the average
`rate of incorporation of deoxynucleotides is three times that of
`dideoxynucleotides, with the hybrid T7 DNA polymerase
`Y526F the average rate of incorporation of deoxynucleotides
`is 8000 times that of dideoxynucleotides. On the other hand,
`whereas with WT E. coli DNA polymerase I and Taq DNA
`polymerase the average rate of incorporation of deoxynucle-
`otides is 600 and 3000 times that of dideoxynucleotides, respec-
`
`T7 DNA
`Polymerase
`
`526Y Y526F
`(MUT)
`(WT)
`
`E. coli DNA
`Polymerase I
`
`Taq DNA
`Polymerase
`
`762F
`(WT)
`
`F762Y
`(MUT)
`
`667F F667Y
`(MUT)
`(WT)
`
`- m
`
`--U'
`
`--B 4.
`
`.0io.........
`if-F.7.740
`
`PHE
`
`PHE
`
`TYR
`
`L
`PHE
`
`W
`
`W
`TYR
`
`444r 4*|jb
`uSo w}w
`.::::,
`.: ::¢a '-u:. :u:*
`
`Stl}f
`
`'.., .;
`
`:.:::. ot .: ._: __
`*: t0f:>
`
`.:_rittsSj _
`....."n_
`
`*- :
`_ :::: :::_
`_-_
`:: .::.:
`
`_........
`
`*,: 4_
`.* .
`_.]_jk3w. __
`
`_.
`
`_
`
`J
`
`TYR
`
`Effect of phenylalanine versus tyrosine on discrimination
`FIG. 3.
`against dideoxynucleotides. Purified T7 DNA polymerase, E. coli
`DNA polymerase I, and Taq DNA polymerase with either phenylal-
`anine (F) or tyrosine (Y) at the ribose selectivity site were compared
`for their ability to extend a radioactive primer in the presence of a 6:1
`ratio of dNTPs to each of the four ddNTPs. Each set of reactions were
`carried out in the presence of ddGTP, ddATP, ddTTP, or ddCTP
`(from left to right). Unextended primer is shown on the left.
`
`Biochemistry: Tabor and Richardson
`
`hybrids containing smaller segments of region C (Table 1 and Fig.
`2A). All of the hybrids that contained the replacement of
`tyrosine-526 in T7 DNA polymerase with phenylalanine, the
`residue at the homologous position in E. coli DNA polymerase I
`(position 762), were resistant to ddTTP. Thus modification of this
`single site is critical for increasing the ability of T7 DNA poly-
`merase to distinguish between deoxy- and dideoxynucleotides.
`A reciprocal series of hybrids in E. coli DNA polymerase I
`were constructed to determine whether substitution of region
`C from T7 DNA polymerase would decrease their ability to
`distinguish between deoxy- and dideoxynucleotides (Table 1
`and Fig. 2B). SDS-DNA activity gel analysis of these hybrids
`defined the same single position as the sole determinant: when
`phenylalanine-762 in E. coli DNA polymerase I was replaced
`by tyrosine, DNA synthesis was strongly inhibited by ddTTP,
`reflecting a diminished ability of the hybrid polymerase to
`distinguish between deoxy- and dideoxynucleotides.
`Analysis of Taq DNA Polymerase/T7 DNA Polymerase
`Hybrid Genes. Of the 16 known members of the Pol I family
`(1, 2, 17), one other besides E. coli DNA polymerase I, Taq
`DNA polymerase, is known to strongly prefer deoxy- over di-
`deoxynucleotides (7). Taq DNA polymerase has a phenylalanine
`at residue 667, the site that corresponds to phenylalanine-762 in
`E. coli DNA polymerase I. We have therefore determined the
`
`Table 1.
`Localization of the domain responsible for distinguishing
`between dNTPs and ddNTPs
`
`*.
`
`Inhibition of
`DNA synthesis
`by ddTTP
`-
`-
`+
`
`Polymerase
`Sequence
`Poll
`754 RRSAKAINFGLIYG
`Taq
`659 RRAAKTINFGVLYG
`518 RDNAKTFIYGFLYG
`T7
`Consensus
`R
`AK
`G
`YG
`T7 WT
`RDNAKTFIYGFLYG
`T7 C-T2
`RBRAKAINFGLIYG
`RfIaAKTFIYGFLYG
`T7 C-T3
`T7 C-T4
`RDNAKAINFGFLYG
`T7 C-T5
`RDNAKAI IYGFLYG
`RDNAKTFUEGFLYG
`T7 C-T6
`T7 C-T7
`RDNAKTFNYGFLYG
`T7 C-T8
`RDNAKTFIEGFLYG
`PolI WT
`RRSAKAINFGLIYG
`Pol I C-K1
`RDNAKTFIYGFYG
`Pol I C-K2
`RRSAKTFIYGLIYG
`RRSAKTFNFGLIYG
`Pol I C-K3
`Pol I C-K4
`RRSAKAIIYGLIYG
`Pol I C-K5
`RRSAKAIIFGLIYG
`RRSAKAINYGLIYG
`Pol I C-K6
`Taq WT
`RRAAKTINFGVLYG
`RINAKTINFGVLYG
`Taq C-Q1
`Taq C-Q2
`RRAAKTFIYGFLYG
`Taq C-Q3
`RRAAKTIIYGVLYG
`Taq C-Q4
`RRAAKTIIFGVLYG
`Taq C-Q5
`RRAAKTINYGVLYG
`Specificity residue
`1
`The three aligned sequences at the top correspond to region C in
`Fig. 1; the number of the first residue shown for each polymerase is
`indicated at the left. Three sets of hybrids are presented, correspond-
`ing to substitutions of E. coli DNA polymerase I sequences into T7
`DNA polymerase (T7), T7 DNA polymerase sequences in E. coli DNA
`polymerase I (Pol I), and T7 DNA polymerase sequences into Taq
`DNA polymerase (Taq). Substituted residues are underlined. Each
`hybrid protein was tested qualitatively for inhibition of DNA synthesis
`by ddTTP using the SDS-DNA activity gel assay described in the
`legend to Fig. 2; inhibition was categorized as >20-fold (+), reflecting
`efficient incorporation of a chain-terminating ddTMP, or <2-fold (-),
`reflecting strong discrimination against ddTTP. Hybrids T7 C-T8, Pol
`I C-K6, and Taq C-QS were purified for quantitative comparison to the
`WT proteins (Table 2).
`
`+
`-
`+
`-
`+
`-
`+
`-
`-
`+
`+
`-
`+
`-
`+
`-
`-
`+
`+
`-
`+
`
`

`

`6342
`
`Biochemistry: Tabor and Richardson
`
`Proc. Natl. Acad. Sci. USA 92 (1995)
`
`Effect of phenylalanine versus tyrosine at the ribose selectivity site on discrimination
`Table 2.
`against ddNTPs
`
`Incorporation rate ratio
`Average
`dC/ddC
`dG/ddG
`dA/ddA
`dT/ddT
`dNMP/ddNMP
`Polymerase
`WT T7 DNA polymerase (526Y)
`3.0
`3.7
`2.8
`3.3
`3.2
`8000
`11,000
`7300
`8400
`6400
`T7 DNA polymerase Y526F
`WT E. coli DNA polymerase I (762F)
`600
`250
`720
`1100
`140
`0.6
`0.75
`0.54
`E. coli DNA polymerase I F762Y
`0.72
`0.56
`WT Taq DNA polymerase (667F)
`3000
`2,600
`4700
`4500
`1400
`0.5
`Taq DNA polymerase F667Y
`0.32
`0.56
`0.59
`0.45
`T7 DNA polymerase, E. coli DNA polymerase I, and Taq DNA polymerase were compared with either
`phenylalanine (F) or tyrosine (Y) at the ribose selectivity site. In each case the unmodified or WT DNA
`polymerase (WT) is presented first. The relative rate of incorporation of each dNMP to ddNMP was
`determined by gel analysis of the average extension lengths at different ratios of dNTP to ddNTP (see text).
`
`tively, the hybrid DNA polymerases E. coli F762Y and Taq
`F667Y actually prefer ddNTPs over dNTPs -2-fold.
`
`DISCUSSION
`Mechanism of Discrimination. The critical phenylalanine/
`tyrosine residue defining the ability of Pol I-type DNA poly-
`merases to distinguish between deoxy- and dideoxyribose is
`located on the "O" helix facing into the crevice responsible for
`binding dNTPs and DNA (Figs. 1 and 4). In a crystal structure of
`the binary complex of E. coli DNA polymerase I with dCTP, the
`closest residue to the 3'-hydroxyl group of the dCIP is this
`Ehenylalanine (residue 762), separated by a distance of about 4.5
`A (14). While this binary complex must be viewed with caution
`since a primer-template is required for specific binding of the
`correct dNTP, genetic and structural data support the juxtapo-
`sition of these two moieties in the catalytically competent com-
`plex (17-20).
`What is the mechanism that accounts for efficient incorpora-
`tion of a nucleotide when there is a hydroxyl moiety on either the
`3' position of the ribose or on the aromatic residue of the poly-
`merase but inefficient catalysis when both hydroxyls are absent?
`ddTTP binds to E. coli DNA polymerase I in a binary complex
`with the same affinity as dTTP, suggesting that discrimination
`
`0 helix
`
`dTTP
`
`FIG. 4. Model showing possible relationship between ribose selec-
`tivity residue and a dNTP in Pol l-type DNA polymerases. On the left
`is helix 0 of E. coli DNA polymerase 1 (13); the four residues shown
`all face into the crevice responsible for binding DNA and dNTPs (Fig.
`1). On the right is dTTP; the position of the ribose moiety is ap-
`proximately that observed in a binary complex of dCTP with E. coli
`DNA polymerase 1 (14). The two critical hydroxyl moieties on residue
`762 (tyrosine) and the 3' position of the dNTP are shaded. A hypothetical
`interaction of these hydroxyl ions with the Mg2+ responsible for stabi-
`lizing the pentavalent intermediate on the a-phosphate on the dNTP
`during catalysis is indicated.
`
`between a deoxy- and dideoxyribose occurs in a step subsequent
`to the initial binding (21). One model is that a hydroxyl moiety is
`required to restrict the space occupied by the nucleotide and thus
`stabilize a catalytically productive orientation. However, it is hard
`to reconcile this model with the fact that two hydroxyls in the
`same space, one on the enzyme and the other on the nucleotide,
`does not affect polymerization. Another model is that at least one
`of the hydroxyl moieties is necessary for binding an essential
`ligand. While this could be a water molecule, a more attractive
`possibility is a divalent cation. The number and location of metal
`ions required for polymerization in Pol I-type DNA polymerases
`are not known, although several conserved acidic residues have
`been implicated genetically as important in binding the divalent
`cation(s) (17-20). It is thought that one divalent cation acts as a
`Lewis acid to promote deprotonation of the 3' hydroxyl of the
`primer while a second promotes the formation of the pentaco-
`valent transition state at the a-phosphate of the dNTP (18).
`Perhaps the hydroxyl ion either on the nucleotide or on the
`enzyme is required to stabilize the binding of the metal ion that
`interacts with the a-phosphate of the dNTP (Fig. 4).
`What is the effect of residues other than phenylalanine and
`tyrosine at the ribose selectivity site? The substitution of the
`polar uncharged residues cysteine, serine, and asparagine at
`residue 526 of T7 DNA polymerase reduces its activity by 10-
`to 50-fold, as measured by SDS-DNA activity gel analysis; the
`residual activity of these mutant enzymes is unable to distin-
`guish between deoxynucleotides and dideoxynucleotides (data
`not shown). The substitution of the nonpolar residue leucine
`reduces T7 DNA polymerase activity 1000-fold; the residual
`activity incorporates deoxynucleotides preferentially over
`dideoxynucleotides. The only substitution that retains the high
`level of polymerase activity observed with phenylalanine and
`tyrosine is tryptophan, which results in a strong preference for
`deoxynucleotides over dideoxynucleotides. While these results
`show that a polar group at this site reduces the ability of a
`polymerase to distinguish between deoxy- and dideoxynucle-
`otides, they also suggest a critical role of an aromatic moiety
`for efficient incorporation of all nucleotides. Consistent with
`these results, Astatke et at (17) have shown that replacement
`of phenylalanine-762 ofE. coli DNA polymerase I with alanine
`increases the Km for dTTP by >100-fold, without affecting
`significantly the kcat.
`Relationship to Other DNA Polymerases. Mutations that
`increase the ability of a DNA polymerase to incorporate
`dideoxynucleotides into DNA have not, to our knowledge,
`been previously described. On the other hand, mutations that
`decrease the ability of DNA polymerases to incorporate
`dideoxynucleotides relative to deoxynucleotides have been
`described in T7 DNA polymerase (22), herpes DNA poly-
`merase (23), and human immunodeficiency virus (HIV) DNA
`polymerase (24, 25). The nature of the changes observed in
`these examples suggests that the decreased rate of incorpora-
`tion of dideoxynucleotides is mediated by an indirect mecha-
`nism. For example, in the T7 DNA polymerase mutants E480D
`
`

`

`Biochemistry: Tabor and Richardson
`and Y530F, the rate of incorporation of ddTMP relative to
`dTMP is reduced 45- and 100-fold, respectively, but these
`mutations also reduce the binding constant for dTTP by
`10-fold (22). Similarly, the hybrid polymerases described here
`containing regions A, D, and E that exhibit significantly lower
`overall activity reduce the rate of incorporation of ddTMP to
`a much greater extent than that of dTMP. In the case of HIV
`reverse transcriptase, the location of some of the mutations
`that decrease the rate of incorporation of dideoxynucleotides
`suggests that they affect the interaction of polymerase with a
`primer-template (25). In contrast, the presence or absence of
`a hydroxyl group at the site identified here appears to affect
`specifically analogs missing the 3'-hydroxyl moiety. Interest-
`ingly, this residue lies within a highly conserved motif shared
`by Pol II-type DNA polymerases and T7 RNA polymerase (2),
`suggesting that this motif