`
`
`
`THE JOURNAL
`
`
`
`
`
`OF BIOLOGICAL CHEMISTRY
`
`Vol. 260, No. 9, Issue of May 10, pp. 5787-5796,1985
`Printed in U. S. A.
`
`The Mutational Specificity of DNA Polymerase-@ during in Vitro DNA
`Synthesis
`PRODUCTION OF FRAMESHIFT, BASE SUBSTITUTION, AND DELETION MUTATIONS*
`
`(Received for publication, November 9, 1984)
`
`Thomas A. Kunkel
`From the Laboratory of Genetics, National Institute of Environmental Health Sciences, Research Triangle Park,
`North Carolina 27709 .
`
`The frequency and specificity of mutations produced
`specificity of errors produced by purified eucaryotic DNA
`polymerases during a single round of DNA synthesis in vitro.
`in uitro by eucaryotic DNA polymerase-@ have been
`determined in a forward mutation assay using a 250-
`The fidelity of in uitro DNA synthesis has previously been
`base target sequence in M13mp2 DNA. Homogeneous
`determined by misincorporation assays with defined polynu-
`DNA polymerase-@, isolated from
`four different
`cleotide templates or using 4X174am3 DNA as a natural
`sources, produces mutations at a frequency of 4-6%/
`template (2). In the 4X assay, accuracy is determined by
`single round of gap-filling DNA synthesis. DNA se-
`measuring the reversion frequency of an amber mutation,
`quence analyses of 460 independent mutants resulting
`observed upon transfection of the product of the in uitro DNA
`from this error-prone DNA synthesis demonstrate a
`synthesis reaction into competent cells (3). This assay focuses
`wide variety of mutational events. Frameshift and base
`on base substitution errors
`that restore an essential gene
`substitutions are made at approximately equal
`fre-
`function and measures these events at one, two, or three
`quency and together comprise about 90% of all muta-
`template nucleotides. Since spontaneous and induced muta-
`tions. Two mutational “hot spots” for frameshift and
`tion rates reflect several different types of errors in addition
`base substitution mutations were observed. The char-
`to base substitutions, and since errors in DNA synthesis occur
`acteristics of the mutations at these sites suggest that
`at many different sites within a gene, it is desirable to monitor
`
`certain base substitution errors result from dislocation
`a wide spectrum of errors. To this end, a new system is
`of template bases rather than from direct mispair for-
`described here using M13mp2 DNA (4). The fidelity of in
`mation by DNA polymerase-@. When considering the
`vitro DNA synthesis is determined for a 250-base target
`entire target sequence, single-base frameshift muta-
`sequence in the lacZcu gene, scoring for any error causing loss
`tions occur primarily in runs of identical bases, usually
`pyrimidines. The loss of a single base occurs 20-80
`of a non-essential gene function (a-complementation). This
`times more frequently than single-base additions and
`forward mutational assay is thus capable of detecting frame-
`much more frequently than the loss of two or more
`shift, deletion, duplication, and complex errors in addition to
`bases. Base substitutions occur at many sites through-
`a large number of different base substitution errors at many
`out the target, representing a wide spectrum of mispair
`sites.
`formations. Averaged over a large number of pheno-
`The system has first been applied to an analysis of the
`typically detectable sites, the base substitution
`error
`frequency and specificity of mutations produced in vitro by
`frequency is greater than one mistake for every 5000
`eucaryotic DNA polymerase-& This class of DNA polymerase
`bases polymerized. Large deletion mutations are also
`has been implicated in repair synthesis in higher organisms
`observed, at a frequency more than 10-fold over back-
`(5, 6) and may participate in other processes as well. DNA
`ground, indicating that purified
`DNA polymerases
`polymerase-P (Pol-PI) has been purified, essentially to ho-
`alone are capable of producing such deletions. These
`mogeneity, from several sources and is well-characterized
`data are discussed in relation to the physical and ki-
`physically and kinetically (5-14). The purified enzyme con-
`netic properties of the purified enzymes and with re-
`sists of a single low molecular weight polypeptide (MI =
`spect to the proposed role for this DNA polymerase in
`30,000-45,000) containing no associated endo- or exonucleo-
`vivo.
`lytic activities. Pol-@ prefers gapped DNA as a primer-tem-
`plate and fills gaps to completion. The enzyme has also been
`shown to be error-prone for base substitutions, producing
`such errors at position 587 in 4X174am3 DNA at a frequency
`of one mistake for every 5000 correct incorporation events
`(15). These properties have prompted an analysis of Pol-p as
`the first to be examined with the forward mutational assay.
`
`The low spontaneous mutation rates observed in eucaryotes
`(1) suggest that these organisms maintain the genetic infor-
`mation with high fidelity. Replication, repair, and recombi-
`nation all contribute to this fidelity, and each of these proc-
`esses depends on the synthesis of new DNA by DNA polym-
`erases in association with other proteins. It is the primary
`effort of this laboratory to determine the mechanisms used
`by these proteins to achieve accurate DNA synthesis. The
`first step in this approach is to determine the frequency and
`
`* The costs of publication of this article were defrayed in part by
`the payment of page charges. This article must therefore be hereby
`marked “aduertisement” in accordance with 18 U.S.C. Section 1734
`solely to indicate this fact.
`
`EXPERIMENTAL PROCEDURES
`Materials-Escherichia
`coli strains NR9099 (A(pro-lac), recA-,
`ara-, thi-/F’(proAB, lacIBZ-PM15)) and S9Oc (A@ro-luc), recA56,
`ara-, thi-, strA-) were provided by Roeland Schaaper of this institute.
`E. coli CSH60 (Abro-lac), ara-, thi-/F’(truD36,proAB, lacI~ZAM15))
`and wild type bacteriophage M13mp2 were obtained from J. Eugene
`
`The abbreviations used are: Pol-P, polymerase-& Hepes, 4-(2-
`hydroxyethy1)-1-piperazineethanesulfonic acid.
`
`5787
`
`This is an Open Access article under the CC BY license.
`
`Columbia Ex. 2086
`Illumina, Inc. v. The Trustees
`of Columbia University in the
`City of New York
`IPR2020-00988, -01065,
`-01177, -01125, -01323
`
`

`

`
`
`
`
`THE JOURNAL
`
`
`
`
`
`OF BIOLOGICAL CHEMISTRY
`
`Vol. 260, No. 9, Issue of May 10, pp. 5787-5796,1985
`Printed in U. S. A.
`
`The Mutational Specificity of DNA Polymerase-@ during in Vitro DNA
`Synthesis
`PRODUCTION OF FRAMESHIFT, BASE SUBSTITUTION, AND DELETION MUTATIONS*
`
`(Received for publication, November 9, 1984)
`
`Thomas A. Kunkel
`From the Laboratory of Genetics, National Institute of Environmental Health Sciences, Research Triangle Park,
`North Carolina 27709 .
`
`The frequency and specificity of mutations produced
`specificity of errors produced by purified eucaryotic DNA
`polymerases during a single round of DNA synthesis in vitro.
`in uitro by eucaryotic DNA polymerase-@ have been
`determined in a forward mutation assay using a 250-
`The fidelity of in uitro DNA synthesis has previously been
`base target sequence in M13mp2 DNA. Homogeneous
`determined by misincorporation assays with defined polynu-
`DNA polymerase-@, isolated from
`four different
`cleotide templates or using 4X174am3 DNA as a natural
`sources, produces mutations at a frequency of 4-6%/
`template (2). In the 4X assay, accuracy is determined by
`single round of gap-filling DNA synthesis. DNA se-
`measuring the reversion frequency of an amber mutation,
`quence analyses of 460 independent mutants resulting
`observed upon transfection of the product of the in uitro DNA
`from this error-prone DNA synthesis demonstrate a
`synthesis reaction into competent cells (3). This assay focuses
`wide variety of mutational events. Frameshift and base
`on base substitution errors
`that restore an essential gene
`substitutions are made at approximately equal
`fre-
`function and measures these events at one, two, or three
`quency and together comprise about 90% of all muta-
`template nucleotides. Since spontaneous and induced muta-
`tions. Two mutational “hot spots” for frameshift and
`tion rates reflect several different types of errors in addition
`base substitution mutations were observed. The char-
`to base substitutions, and since errors in DNA synthesis occur
`acteristics of the mutations at these sites suggest that
`at many different sites within a gene, it is desirable to monitor
`
`certain base substitution errors result from dislocation
`a wide spectrum of errors. To this end, a new system is
`of template bases rather than from direct mispair for-
`described here using M13mp2 DNA (4). The fidelity of in
`mation by DNA polymerase-@. When considering the
`vitro DNA synthesis is determined for a 250-base target
`entire target sequence, single-base frameshift muta-
`sequence in the lacZcu gene, scoring for any error causing loss
`tions occur primarily in runs of identical bases, usually
`pyrimidines. The loss of a single base occurs 20-80
`of a non-essential gene function (a-complementation). This
`times more frequently than single-base additions and
`forward mutational assay is thus capable of detecting frame-
`much more frequently than the loss of two or more
`shift, deletion, duplication, and complex errors in addition to
`bases. Base substitutions occur at many sites through-
`a large number of different base substitution errors at many
`out the target, representing a wide spectrum of mispair
`sites.
`formations. Averaged over a large number of pheno-
`The system has first been applied to an analysis of the
`typically detectable sites, the base substitution
`error
`frequency and specificity of mutations produced in vitro by
`frequency is greater than one mistake for every 5000
`eucaryotic DNA polymerase-& This class of DNA polymerase
`bases polymerized. Large deletion mutations are also
`has been implicated in repair synthesis in higher organisms
`observed, at a frequency more than 10-fold over back-
`(5, 6) and may participate in other processes as well. DNA
`ground, indicating that purified
`DNA polymerases
`polymerase-P (Pol-PI) has been purified, essentially to ho-
`alone are capable of producing such deletions. These
`mogeneity, from several sources and is well-characterized
`data are discussed in relation to the physical and ki-
`physically and kinetically (5-14). The purified enzyme con-
`netic properties of the purified enzymes and with re-
`sists of a single low molecular weight polypeptide (MI =
`spect to the proposed role for this DNA polymerase in
`30,000-45,000) containing no associated endo- or exonucleo-
`vivo.
`lytic activities. Pol-@ prefers gapped DNA as a primer-tem-
`plate and fills gaps to completion. The enzyme has also been
`shown to be error-prone for base substitutions, producing
`such errors at position 587 in 4X174am3 DNA at a frequency
`of one mistake for every 5000 correct incorporation events
`(15). These properties have prompted an analysis of Pol-p as
`the first to be examined with the forward mutational assay.
`
`The low spontaneous mutation rates observed in eucaryotes
`(1) suggest that these organisms maintain the genetic infor-
`mation with high fidelity. Replication, repair, and recombi-
`nation all contribute to this fidelity, and each of these proc-
`esses depends on the synthesis of new DNA by DNA polym-
`erases in association with other proteins. It is the primary
`effort of this laboratory to determine the mechanisms used
`by these proteins to achieve accurate DNA synthesis. The
`first step in this approach is to determine the frequency and
`
`* The costs of publication of this article were defrayed in part by
`the payment of page charges. This article must therefore be hereby
`marked “aduertisement” in accordance with 18 U.S.C. Section 1734
`solely to indicate this fact.
`
`EXPERIMENTAL PROCEDURES
`Materials-Escherichia
`coli strains NR9099 (A(pro-lac), recA-,
`ara-, thi-/F’(proAB, lacIBZ-PM15)) and S9Oc (A@ro-luc), recA56,
`ara-, thi-, strA-) were provided by Roeland Schaaper of this institute.
`E. coli CSH60 (Abro-lac), ara-, thi-/F’(truD36,proAB, lacI~ZAM15))
`and wild type bacteriophage M13mp2 were obtained from J. Eugene
`
`The abbreviations used are: Pol-P, polymerase-& Hepes, 4-(2-
`hydroxyethy1)-1-piperazineethanesulfonic acid.
`
`5787
`
`This is an Open Access article under the CC BY license.
`
`

`

`tions and plating and the scoring and confirmation of mutant (light
`blue or colorless) phenotypes were performed as described previously
`(17).
`DNA sequence analysis of mutants was by the chain terminator
`method (19), using the
`oligonucleotides described previously (17).
`The 32P label needed to observe the sequence ladder was incorporated
`onto the 5’ ends of the oligonucleotides prior to hybridization using
`[r-32P]ATP and T4 polynucleotide kinase.
`
`of DNA Polymerase-P in Vitro
`Mutational Specificity
`
`5788
`LeClerc, University of Rochester, Rochester, NY. DNA polymerase-
`B from rat Novikoff hepatoma (hereafter called rat Pol-@), homoge-
`neous fraction VI (7), and from HeLa cells, Fraction VI11 (9), were
`from Dale W. Mosbaugh, University of Texas, Austin, TX. DNA
`polymerase-@ from chick embryos (chick Pol-@), homogeneous frac-
`tion VI11 (12), was provided by Akio Matsukage, Aichi Cancer Center,
`Chikusa-Ku, Nagoya, Japan. Human liver Pol-@, homogeneous frac-
`tion VI1 (lo), was from T. S.-F. Wang and D. Korn, Stanford
`University, Palo Alto, CA. Restriction endonucleases AuaII, PuuI,
`and PuuII were from Boehringer Mannheim.
`RESULTS
`Preparation of Gapped M13mp2 DNA-Bacteriophage M13mp2
`The M13rnp2 Mutagenesis Assay-The
`hybrid bacterio-
`was plated on minimal plates as described below, using E. coli NR9099
`phage M13mp2, developed by Messing and co-workers (4),
`to 1 liter of 2 X YT
`as a host strain. A single plaque was added
`contains a small segment of the E. coli lac operon within the
`medium (containing, per liter, 16 g of Bacto-Tryptone, 10 g of Yeast
`Extract, 5 g of NaCl, pH 7.4) containing 10 ml of an overnight culture
`intergenic region of M13. This segment of DNA consists of
`of E. coli NR9099. Ml3mp2-infected cells were grown overnight at
`the C-terminal coding sequence of the lac1 gene, the lac
`37 “C with vigorous shaking. Cells were harvested by centrifugation
`promoter and operator regions and the DNA sequence coding
`a t 5000 X g for 30 min, and replicative form DNA was prepared by
`for the first 145 amino acids of the N-terminal end of the lacZ
`the method of Birnboim and Doly (16). Phage were precipitated from
`gene (the a region). The E. coli host strain used for the assay
`the clear culture supernatant by addition of polyethylene glycol 8000
`(E. coli CSH50) contains a chromosomal deletion of the lac
`
`to 3% and NaCl to 0.5 M. The phage pellet, obtained by centrifugation
`operon, but harbors an F’ episome to provide the remaining
`at 5000 X g for 30 min at 0 “C, was resuspended in phenol extraction
`buffer (100 mM Tris-HC1, pH 8.0 300 mM NaCl, 1 mM EDTA). The
`coding sequence of the lac2 gene. Functional EacZ gene prod-
`single-stranded phage DNA was extracted with phenol twice, followed
`uct, /3-galactosidase, is produced when the lacZ gene coding
`by two ch1oroform:isoamyl alcohol (24:l) extractions. The M13mp2
`information missing on the F factor (for the host used here,
`DNA was precipitated with ethanol and resuspended in TE buffer
`nucleotides +71 through +163 where +1 is the first tran-
`(10 mM Tris-HC1, pH 8.0, 1 mM EDTA).
`scribed base) is provided by the information carried in the
`Replicative form M13mp2 DNA was digested at 37 “C with restric-
`tion endonucleases PuuI and PuulI in a reaction containing 10 mM
`M13mp2 DNA. In this case, the two partial proteins produced
`Tris-HC1, pH 7.8, 10 mM MgCl,, 50 mM NaC1, 1 mM dithiothreitol,
`within an M13mp2-infected host cell reconstitute enzyme
`replicative form DNA, and enough of each enzyme to achieve com-
`activity by intracistronic a-complementation. This is detected
`plete digestion (as determined in small preliminary digestions before
`by plating infected cells under conditions to monitor the
`the large-scale preparative digestion). The digested DNA was sub-
`production of blue color resulting from hydrolysis of the
`jected to electrophoresis in a 0.8% agarose gel in TAE buffer (40 mM
`indicator dye 5-bromo-4-chloro-3-indolyl-/3-~-galactoside
`by
`Tris acetate, pH 8.3, 2 mM EDTA) containing 0.5 mg/ml ethidium
`bromide. Bands were visualized by a brief illumination with UV light,
`,&galactosidase. Decreased a-complementation resulting from
`and the 6.8-kilobase fragment was excised. The DNA electroeluted
`a mutagenic event in the lacZa gene in M13mp2 will give rise
`from this gel fragment was purified and concentrated on an Elutip-d
`to M13mp2 plaques having lighter blue or no color.
`affinity column (Schleicher and Schuell) according to the manufac-
`The general outline used to assay the accuracy of in vitro
`turer’s instructions. The fragment was then precipitated with ethanol
`DNA synthesis is
`shown in Fig. 1. A gapped molecule is
`and resuspended in T E buffer. The gapped double-stranded circular
`constructed (see “Experimental Procedures”) in which the
`molecule was then prepared by mixing 500 pg each of the restriction
`gap contains the target sequence. This is filled by a single
`endonuclease fragment and single-stranded circular viral DNA (7196
`cycle of in vitro DNA synthesis using the desired DNA polym-
`bases) in 30 mM NaCI, 30 mM sodium citrate. The mixture was heated
`to 95 “C for 10 min, cooled in an ice bath, and incubated at 65 “C for
`erase and reaction conditions. A portion of the product is
`30 min. The gapped molecule was then purified by preparative agarose
`then analyzed to assure complete synthesis, and the remainder
`
`gel electrophoresis and affinity chromatography as for the restriction
`is used to infect cells and assay for a-complementation. Cer-
`endonuclease fragment. The final preparation, when analyzed
`by
`tain errors during the in vitro DNA synthesis result in altered
`agarose gel electrophoresis, contained approximately
`80% gapped
`
`production of a-peptide, and upon scoring and confirming the
`double-stranded circular DNA and 20% linear restriction endonucle-
`ase fragment, but no detectable single-stranded DNA.
`mutant phenotype, the exact nature
`of the error can
`be
`synthesis
`DNA Synthesis Reactions and Product Analysis-DNA
`determined by DNA sequence analysis of the single-stranded
`was performed in 50.~1 reactions containing 20 mM Hepes, pH 7.8, 2
`viral DNA.
`mM dithiothreitol, 10 mM MgC12, 500 p M each M T P , dGTP, dCTP,
`This assay system has several major advantages for mea-
`and [a-”P]dTTP (500-1000 cpm/pmol), 300 ng of gapped circular
`surements of in vitro accuracy. Most importantly, the assay
`M13mp2 DNA, and either 0.4 unit of rat DNA polymerase-@, 0.8 unit
`scores for loss of a non-essential gene function, allowing a
`of chick DNA polymerase-@, 0.10 unit of human liver DNA polym-
`erase-@, or 0.5 unit of HeLa cell DNA polymerase-0. In each instance,
`wide variety of mutations to be tolerated. The use of a
`the unit definition was that of the enzyme supplier, as determined in
`derivative of M13 permits the simple preparation of large
`each individual laboratory before shipment, as previously described
`amounts of pure viral DNA, from a small culture, in single-
`(7, 9, 10, 12). Incubation was at 37 “C for 60 min, and reactions were
`stranded form. This not only facilitates routine DNA se-
`terminated by addition of EDTA to 15 mM. Twenty 11 (120 ng of
`quence analysis of mutants, it also allows one to produce a
`DNA) of each reaction were mixed with 5 pl of sodium dodecyl sulfate
`specifically gapped molecule and to easily engineer sequence
`dye mix (20 mM Tris-HC1, pH 8.0,5 mM EDTA, 5% sodium dodecyl
`sulfate, 0.5% bromphenol blue, 25% glycerol) and subjected to elec-
`changes in the mutational target
`(see below). M13mp2 has
`trophoresis in a 0.8% agarose gel in TAE buffer containing 0.5 mg/
`been chosen over the other M13mp derivatives, as it produces
`ml ethidium bromide. Electrophoresis was at a constant 50 V for 16
`a darker blue color on the plates, potentially allowing a wider
`h. Bands were visualized by illumination with UV light and photo-
`spectrum of “down” mutants to be scored. This early deriva-
`graphed. The gel was then dried and used to expose Kodak XAR film
`tive of the M13mp series of vectors also does
`not contain
`to produce autoradiograms.
`nonsense codons and can be grown on a suppressor minus
`remaining
`Expression of DNA Product in Competent Cells-The
`product of the polymerization reactions was used for transfection of
`host. This allows detection of nonsense mutations in the
`E. coli S9OC recA56 cells made competent as previously described
`target sequence.
`using 75 mM CaC1, (17) or more recently using the procedure of
`The gapped M13mp2 molecule depicted in Fig. 1 is a good
`Hanahan (18). With Ca2+-treated cells, the ratio of DNA molecules
`primer-template for most purified DNA polymerases. The
`to cells was held constant at 10:1, while with cells treated by the
`specific gap used for
`these studies, constructed
`using two
`procedure of Hanahan the ratio was either 1:l or 2:l. The transfec-
`
`

`

`Mutational Specificity of DNA Polymerase-@ in Vitro
`A B
`
`5789
`
`C
`
`M13 mp2
`Gapped D N A
`
`Plate
`
`n
`
`I
`
`Wet I ,
`
`RF II
`Gapped
`
`Primer
`
`RFI
`
`ss
`
`Score Mutation
`
`and Sequence Frequency and
`Confirm Mutant
`Phenotypes
`
`Mdtant (light blue or
`colorless)
`FIG. 1. Experimental outline of M13mp2 mutagenesis as-
`say. The five pairs of dots on the gapped molecule indicate the
`position of the five sites for adenine methylation used to instruct
`mismatch correction (20-23). The gap extends from positions +174
`to -216 (where +1 is the start of transcription) and is determined by
`cleavage with restriction endonucleases PuuI and PuuII. The 5' end
`(on the left) of the 390-base gap is therefore more than 100 bases
`away from the end of the target (which is position -84, the first
`nucleotide after the lac1 gene termination codon), but still within the
`lac DNA. The 3"OH primer terminus (on the right) is nucleotide
`+175, the middle nucleotide of the lacZa codon 45. The lacZa target
`is indicated by the darker line within the gap. The direction of DNA
`synthesis within the gap is right to left. The square (made with dashed
`lines) represents a competent E. coli cell.
`
`restriction endonucleases, has been chosen for several rea-
`sons. The complementary (minus) strand is almost full ge-
`nome length and perhaps more likely to survive transfection
`in the host cell than partial minus strands (2,3). The double-
`stranded portion of the molecule retains all adenine methyl-
`ation sites used to instruct mismatch correction (20,21). The
`end product of the in vitro reaction is therefore fully methyl-
`ated, and polymerase errors in the newly synthesized DNA
`are less likely to be repaired in vivo by the mismatch correc-
`tion repair system (22, 23). The 3'-OH primer terminus of
`the gap (at +174) is 11 bases beyond the last nucleotide
`missing (+163) on the F' (lacZ-AM15) of the host cell, thus
`including most of the coding sequence of the a-peptide needed
`for a-complementation. The 5'-phosphoryl terminus is at
`position -216, which
`is more than 100 bases beyond the
`regulatory sequences which control a-peptide production, but
`still within the non-essential DNA sequence in M13mp2. The
`gap thus contains the entire mutational target plus over 100
`bases of non-essential sequence in which mutations are ex-
`pected to be phenotypically silent. It is therefore not necessary
`to completely fill the entire gap in order to see mutations
`throughout the target. Only 250 of 390 bases need be incor-
`porated to observe a complete spectrum.
`Mutation Frequency of Control and Pol-0-copied DNA-
`DNA polymerase-0 prefers gapped DNA as a primer template
`and is capable of filling gaps to completion (9,ll). An agarose
`gel analysis of the product of a DNA synthesis reaction by
`homogeneous rat Pol-/3 on M13mp2 DNA containing a 390-
`base gap is shown in Fig. 2. The major band after a 60-min
`reaction, visualized either by fluorescence (lane B ) or auto-
`radiography (lane C), is in the position of fully double-
`stranded DNA, representing a clear shift from the position of
`the unfilled gapped molecule (lane A). Results similar to those
`
`FIG. 2. Agarose gel analysis of the product of the rat hepa-
`toma Pol-@ copying reaction with gapped M13mp2 DNA. Lane
`A, standards, including an amount of uncopied gapped DNA equiva-
`lent to the copied DNA shown in lane B (Primer refers to the linear
`fragment used to make the gapped template); lane B, rat hepatoma
`Pol-@ copied, 60 min, 37 "C; lane C, Pol-@ copied, autoradiograph of
`32P incorporation. The analysis was performed as described under
`"Experimental Procedures." Lanes B and C were obtained from two
`separate agarose gel electrophoretic analyses of the same DNA prep-
`aration. RF, replicative form; SS, single strand.
`
`in Fig. 2 were obtained with each of the four Pol-0 prepara-
`tions used. The products of these reactions thus represent a
`relatively homogeneous population of copied molecules con-
`taining putative mutations resulting from one round of in
`vitro synthesis.
`The results of transfection of rat Pol-0-copied DNA into
`competent cells are shown in Table I, compared to transfec-
`tions of several uncopied control DNA molecules (see legend
`to Table I for description). The frequency of light. blue or
`for the
`colorless mutant plaques is 2.8 X
`to 6.6 X
`for DNA
`control DNA molecules, compared
`to 640 X
`copied by Pol-& This 100-fold increase is due to mutations in
`the newly synthesized strand, since denaturation to eliminate
`the biological activity of this (unligated) strand returns the
`mutation frequency to the background level.
`The unexpectedly high mutation frequency resulting from
`synthesis by this enzyme led
`to an examination of three
`additional preparations of DNA polymerase+?, isolated from
`different sources. The mutation frequencies of DNA copied
`by Pol-@ from chick embryo, HeLa cells, or human liver were
`all similar to results with the rat polymerase (Table 11). Thus,
`not just a single enzyme but rather the &polymerase class is
`highly inaccurate in vitro.
`DNA Sequence Analysis of Mutants-Since
`
`loss of a-com-
`
`

`

`
`
`DNA Polymerase-p in Vitro
`
`5790
`
`
`
`Mutational Specificity of
`
`DNA
`
`Plaques scored Mutation
`
`Total Mutant” frequency
`
`
`
`
`
`
`TABLE I
`Mutation frequency of control and rat Pol-@-copied DNA
`The rat hepatoma DNA polymerase-@ copying reactions, transfec-
`tions, and plating were as described under “Experimental Procedures”
`and in Ref. 17. Mutation frequencies of viral and replicative form
`DNA were determined from the same preparations of DNA used to
`construct the gapped molecule.
`Number of
`determi-
`nations
`
`Viral
`7
`10,597
`2
`Replicative form
`28,655
`2
`8
`Nicked constructb
`128
`199,655
`3
`Pol-@-copied
`669
`10,474
`5
`Pol-@-copied, denatured’
`
`3
`
`3,099 9.7
`1
`Mutants include colorless plaques as well as those having lighter
`blue color than wild type mp2. Several light blue phenotypes were
`observed varying in intensity from almost colorless to almost wild
`type. More than 95% of the mutants, when carefully removed from
`the plate, diluted, and replated, were of only a single phenotype.
`Occasionally, a plaque having sectors of both colorless and blue
`phenotypes was observed, which when replated yielded both wild type
`blue plaques and mutant plaques, in approximate proportion to the
`size of the sectors in the original infective center. The ratio of light
`blue to colorless mutants was 2:l for the nicked construct transfec-
`tions but 1:2 for the Pol-@-copied DNA. This is consistent with the
`sequence analysis (Table IV), since many of the Pol-@ frameshift
`mutants, produced at high frequency, are colorless.
`The nicked construct was made by cleaving replicative form DNA
`with restriction endonuclease AvaII and then hybridizing the full
`genome length complementary strand to the
`viral strand as
`for
`formation of the gapped molecule. The resulting completely double-
`stranded circular molecule contains a nick at position -264, only 48
`bases from the position of the nick in the Pol-@ gap-filled molecule.
`Having been subjected to manipulations similar to the gapped mole-
`cule and having a similar configuration (i.e. completely double-
`stranded circular with a single nick outside the mutational target but
`within the non-essential DNA), this construction was deemed most
`appropriate for subsequent analyses of spontaneous mutants (Table
`111).
`For the final transfection shown, the product of the Pol-@ copying
`reaction was denatured at 95 “C for 3 min and then diluted in TE
`buffer and used for transfection.
`
`coding sequence of the hcZa gene. These were observed at a
`run of 4 consecutive T residues at positions +70 through +73
`and at 2 T residues at positions +lo3 and +104. Both hot
`spots have similar dinucleotide neighbors, a CG on the 5‘ side
`and an AC on the 3’ side. Two types of mutations were
`produced at these sites: the loss of a single T in the run (a
`colorless phenotype) or a T --.* G transversion at +70 or +103,
`the 5’-most T residue (resulting in a valine + glycine change,
`a light blue phenotype). The relative proportion of frameshift
`and transversion errors is different for the two sites. These
`hot spots are indeed a result of errors by Novikoff hepatoma
`Pol$, since among 128 spontaneous mutants sequenced, not
`one error was observed at these two sites. An analysis of 20
`independent colorless mutants generated in the chick embryo
`Pol-@ copying reactions showed a similar high frequency of
`deletion frameshifts at +70 through +73. Thus, the hot spots
`seem to be a common feature of Pol-@ mutational spectra
`with this target. The frequency of -T frameshifts at +70 to
`+73 (51 of 159 mutants analyzed) was sufficiently high
`to
`interfere with the generation
`of an extensive spectrum of
`other Pol-@ errors. For this reason, site-specific mutagenesis
`techniques (24) were used to change the T residue at +72 to
`a C. This genetically silent change interrupts the run of 4 T
`residues and should effectively reduce the frequency of -T
`frameshifts at this hot spot. Using this target, differing by a
`single silent change from wild type, an additional 137 Novikoff
`hepatoma Pol-@ mutants were analyzed. As expected, the total
`forward mutation frequency was reduced (from 640 x
`to
`490 X IOw4), and the proportion of -T frameshifts at this site
`relative to the total mutants analyzed was decreased (18/137).
`This target was also used to generate the mutational spectrum
`of errors produced by chick Pol-@ (144 mutants sequenced).
`Spectrum of Errors by Pol-@-The complete spectra of
`single-base frameshifts (shown below each line of primary
`viral DNA sequence) and
`single-base substitutions (above
`each line) are shown in Fig. 3 (rat Pol-@) and Fig. 4 (chick
`Pol-@). Mutations are distributed over both the coding and
`regulatory sequences,
`consistent with
`DNA synthesis
`throughout the target. Only two of 440 Pol-@ mutants ana-
`lyzed did not contain an error in the target within the gap.
`While in many instances mutations were found consistent
`with previously observed
`phenotypes (17, 25), many new
`detectable sites for mutation were observed. These new mu-
`tations were the only changes detected within the 254 bases
`shown in Figs. 3 and 4. As most of these have been observed
`several times as independent
`events’ and have the same
`phenotypes (ie. intensity of blue color