Download Errors in DNA Replication as a Basis of Malignant Changes1

[CANCER RESEARCH 34, 2311-2321,
September 1974]
Errors in DNA Replication as a Basis of Malignant Changes1
Lawrence A. Loeb,2 Clark F. Springgate,3 and Narayana Battala4
The Institute for Cancer Research, Fox Chase Center for Cancer and Medical Sciences, Philadelphia, Pennsylvania 19111
SUMMARY
In this essay we have considered possible relationships
between errors in DNA replication and malignant progression.
Theoretical and experimental studies designed to determine
the energy of interaction between different nucleotides have
been reviewed. The difference of the energy of interaction
between complementary and noncomplementary base pairs (1
to 3 kcal) is not sufficient to account for the exactness of
DNA replication that is observed in vivo or in vitro. Thus,
DNA polymerases must play a role in base selection. This
point has been established by the detection of polymerases
that copy polynucleotide templates with few errors and the
detection of others that make many errors. Different
mechanisms for base selection by polymerases have been
considered. Our findings that both extracts of human
leukemic lymphocytes and a purified DNA polymerase from
avian myeloblastosis virus have exceptionally high error rates
have been considered in relationship to oncogenesis. A
hypothesis is presented that relates mistakes in DNA replica
tion, as promoted by error-prone DNA polymerases, to tumor
progression. We have also considered a more general concept
encompassing early cellular changes that lead to oncogenesis.
The justification for this essay is that both of these hypotheses
are open to immediate experimental analysis.
INTRODUCTION
The purpose of this article is to consider possible
relationships between mistakes in DNA synthesis and malig
nant changes. Until recently, many of the theoretical concepts
relating to these phenomena were essentially mental exercises
not open to experimental verification. Since methods are now
at hand for determining the fidelity of polynucleotide
replication in vitro and for quantitating mutation rates of
animal cells in culture, it is now possible to analyze the
exactness by which DNA is replicated as cells become
1This investigation was supported by grants from the NIH
(CA-11524, CA-12818) and the National Science Foundation (GB36812), by giants to this Institute from NIH (CA-06927, RR-05539),
and by an appropriation from the Commonwealth of Pennsylvania.
2Also a member of the Department of Pathology, School of
Medicine, University of Pennsylvania, Philadelphia, Pa. 19104.
3Postdoctoral fellow of The Leukemia Society of America.
4Predoctoral associate, Graduate Group on Molecular Biology,
University of Pennsylvania, Philadelphia, Pa. 19104.
Received March 13,1974; accepted May 30, 1974.
SEPTEMBER
1974
malignant. There is little experimental evidence directly
relating mutational events to neoplasia and thus this article is
highly speculative. Hopefully, many of these speculations are
experimentally testable.
We have asked whether malignant changes are brought
about by mistakes in base pairing during DNA replication and
whether tumor progression ensues by an accumulation of these
errors in genetic content (76, 77). Our hypothesis was
formulated by considering experimental results on the fidelity
of DNA synthesis in vitro as well as some of the known
cellular changes that occur during tumor progression, in
particular: (a) the continuous evolution of new cell variants;
(b) the increasing ability of tumor cells to escape the host
mechanism for regulating cellular proliferation; (c) the
progressive accumulation of chromosomal abberations;and (d)
the increasing ease by which some tumors become resistant to
chemotherapeutic agents.
In this article, we first consider how the genetic information
in DNA is transmitted with few, if any, errors in normal cells
from parent to daughter. From an analysis of the energy of
interaction between different deoxynucleotides,
a role for
DNA polymerase in augmenting correct base selection appears
mandatory. That is to say, the exactness of DNA synthesis
observed in vivo and in vitro cannot be accounted for solely by
the energy differences between complementary and noncom
plementary base pairs (47).
Secondly, we consider the possible cellular consequences
resulting from mistakes in copying DNA. We have wondered
whether malignant progression, i.e., that process by which
cancer cells evolve into divergent phenotypes, might be a
manifestation of errors made during DNA replication. As
tumors evolve there would be an accumulation of errors in
genetic information. One possible mechanism by which this
might come about would be if enzymes, such as DNA
polymerase, were altered so as to permit errors in base pairing
during DNA replication and repair.
We suggest a unified theory in which diverse carcinogenic
agents affect DNA so as to cause errors in base pairing during
DNA replication. Some of these errors would cause alterations
in structural genes that code for enzymes that function in base
selection, leading to an expanding number of errors in genetic
information. Concisely, we are proposing that infidelity of
DNA replication may be responsible for tumor oncogenesis
and progression.
EXACTNESS OF BASE PAIRING
Embodied in the structure of DNA as proposed by Watson
2311
L. A. Loeb et al.
and Crick (90) was the suggestion that faithful duplication of
base sequence was mediated primarily by hydrogen bonding
between parental DNA and daughter nucleotides. A more
detailed model implies that the deoxynucleoside triphosphates
are prealigned by hydrogen bonding onto the DNA template
and that DNA polymerase would function as a "zipper" to
AMP interacted about equally with UMP and CMP. From these
and related studies, one can conclude that, even if nucleotidenucleotide interactions occur in aqueous solution at physio
logical conditions, the strength of the interaction is totally
inadequate to account for the exactness of DNA synthesis.
Theoretical estimates of the electrostatic energy of interac
join the bases together in phosphodiester linkage (20). In such tion between nucleotides have been made by quantum
a model the formation of correct base pairs, that is, mechanical calculations (50). These calculations yield large
deoxyadenosine to deoxythymidine, and deoxyguanosine to differences in energy between complementary and noncomple
deoxycytidine,
results solely from the differences in the mentary base pairs. For example, the electrostatic potential
physical interactions between correct and incorrect base pairs. energy for the complementary base pair A-U is estimated to be
This model does not seem adequate as'a 1st step in DNA —¿5.8kcal, while that for the noncomplementary base pair A-C
synthesis if one considers the exactness by which DNA is is +6.23 kcal. However, calculations from quantum mechanics
copied (see below) and the simple fact that deoxynucleotides
are far removed from biological realities. One has to assume
in physiological concentrations
fail to bind to DNA at that the dielectric constant is unity to neglect the effects of
physiological temperatures (37°).
solvents and to keep the coordinates of each base fixed and
As pointed out by Mildvan (47), a large difference in free independent of the base to which it is paired. Some possible
energy between complementary and noncomplementary base base pairs are illustrated in Chart 1. Chart \A is the
Watson-Crick A-T base pair. In Chart IB the thymidine is
pairs (at least 7.0 kcal/mole) would be necessary to account
for the exactness by which DNA is synthesized. Methods have replaced by cytidine without changing the position of the
atoms. The large energy difference may be in part due to the
been developed to analyze the energy differences between
complementary and noncomplementary base pairs. With few repulsion of similarly charged groups. However, as shown in
Chart 1C, this repulsion can be avoided by assuming a
exceptions the differences in free energy between complemen
tary and noncomplementary
base pairs that have been different position for cytidine ( 13) with little distortion of the
calculated amount to only 1 to 3 kcal. One error in 10" base helix. Recently, Jordan and Sostman (31) calculated the
pairs requires a free energy difference (AG) of 1.4 («) energies of interactions for different base pairs assuming a
variety of coordinates for incorrect base pairs. Surprisingly,
kCal/mole between the complementary and noncomplemen
when one compared the energy of an A-T base pair with a
tary base pairs (47). The small difference in energy between
A-C base pair, the complementary base
base pairing (1 to 3 kcal) is only sufficient to account for a noncomplementary
pair was found to be less stable. In contrast, the G-C base pair
level of discrimination of 1 error in 10 to 1 in 100. In contrast,
DNA synthesis in vitro is an extremely accurate process; with was found to be far more stable than was the noncomplemen
most cellular polymerases, using polynucleotide templates, the tary G-T base pair.
error rate is only 1 in 10,000 to 1 in 1,000,000. The difference
A large energy contribution to the stability of the DNA
in AG necessary to account for this error rate would be 7 to 9 double helix is obtained from the energy of interaction
kcal. Furthermore, DNA replication in vivo as measured from
between adjacent base pairs on the same chains, i.e., the
mutation rate data appears to be even more accurate, the error
stacking energy. Calculation of stacking interaction energies
rate being as little as 1 in IO9. For this process, a AG of yields values of 9 to 12 kcal between adjacent bases (31).
However, differences in the stacking energy between adjacent
discrimination of 126 kcal would be required. The increased
exactness of DNA synthesis in vivo suggests that editing
purines or pyridines do not vary greatly (31). The essentially
"random" distribution of bases on a single strand of DNA
mechanisms in cells operate for correcting mistakes in base
argues against the significance of stacking energy between
pairing during or after DNA replication.
Prior to considering the investigations that have been carried
adjacent bases on one strand in the choice as to which
out with purified enzymes and the postulated consequences of nucleotide is incorporated into the other strand.
Nucleotide-Poly nucleotide Interactions. Considering that
mistakes in DNA synthesis, it is important first to review in
detail the nature of the interactions between nucleotides and base pairing during DNA synthesis occurs on a polynucleotide
polynucleotides. The strength of nucleotide interactions in template, studies of nucleotide-polynucleotide
interactions are
relationship to the fidelity of DNA replication has not been
perhaps more relevant than studies of nucleotide-nucleotide
considered in the literature, yet it forms the framework for interactions. The binding of guanosine, cytidine, and adenounderstanding the probability of mistakes in base pairing
sine to poly(U) has been measured by equilibrium dialysis (30,
58). Adenosine did not bind until its concentration reached a
during DNA replication.
Nucleotide Interactions. Evidence of hydrogen bonding threshold of 6 mM (58). At this high concentration stacking of
between both complementary and noncomplementary nucleo
bases between nucleotides begins to occur. In contrast, the
tide pairs in aqueous solution has been reported using proton
incorrect bases guanosine and cytidine did not bind to poly(U)
magnetic resonance (62). These interactions were detected at at concentrations as great as 20 mM, demonstrating Watson0° and at very high concentrations;
the differences in Crick preferential binding. However, guanosine, but not
interaction between complementary and noncomplementary
cytidine, was found to bind to poly(U) in the presence of
base-paired nucleotides were not very great. For example,
adenosine, presumably by costacking of guanosine with
Chart 1. Complementary and noncomplementaiy base pairing. A, Watson-Crick A-T base pair;5, substitution of cytidine for thymidine without
changing molecular coordinates; C, rotation of cytidine as proposed by Donohue (13).
2312
CANCER RESEARCH VOL. 34
a
A=T Base Pair
OP
:M
/
H H'
PO
Correct (Watson-Crick)
A adjacent to C
OP
in
H-V
1
A =C Base Pair
OP
I
H_N
.
/
&
\
c
/ H H.
,'
•¿'
»C~H
V
CP
SEPTEMBER 1974
2313
L. A. Loeb et al.
adenosine. Thus, the observed interaction of mononucleotides
with polynucleotides may occur only when the mononucleo
tides are stacked in groups and functionally behave as
oligonucleotides (58). Furthermore, many of these interac
tions involve the formation of triple-stranded complexes, the
specificities of which are not determined by base pairing.
Polynucleotide-Polynucleotide
Interactions. Studies on the
structure and stability of double helices containing noncomplementary base pairs were pioneered by Fresco and Alberts
(21). They studied the interaction of poly(U)5 with mixed
polynucleotides containing both uridine and adenosine. By
measuring absorbance at various input ratios, they found that
the maximum extent of base pairing occurred when there was
a stoichiometric equivalence between the number of nucleotides in the poly(U) strand and the number of adenosines in
the mixed polynucleotide. Their results raised the possibility
that the mismatched bases were simply looped out of the helix
by rotation about 2 adjacent phosphodiester bonds. These
defects have been extensively characterized by studies on
complexes of different copolynucleotides (5, 18, 68). The
cellular counterpart could be DNA with an added or deleted
nucleotide on 1 strand resulting in frame-shift mutations. In a
strict sense, with DNA templates this defect would not
represent errors in base pairing. Nevertheless, from the analysis
of Fink and Crothers (18), the "defect" free energy can be
estimated at 4.0 kcal/mole (a bulge defect of 2.8 kcal and a
loss of 1 base pair of 1.2 kcal). This free energy of
discrimination would predict an error frequency of about 1 in
1000 (1.4X4.0 kcal/mole).
Wang and Kallenbach (88) have studied in detail the
interaction of poly(I) with copolymers containing cytidine
with different mole ratios of inosine, adenosine, and uridine.
They estimated the difference in free energy between correctly
matched nucleotides poly(I)«poly(C) and the incorrect, mis
matched poly(I)'poly(U)
to be 1.7 kcal/base pair. With
poly(I)-poly(C)
versus poly(I)-poly(A)
the difference was
only 1.3 kcal. In either case, the difference in free energy
between complementary and noncomplementary
base pairs
would predict an error frequency of 1 in 10.
So far, with only 1 theoretical exception (50), studies using
a variety of methods indicate that the difference in free energy
between correct and incorrect base pairs is only 1 or 3 kcal
and thus appears insufficient to account for the fidelity of
DNA synthesis. Thus, DNA polymerase in vitro or perhaps
other enzymes that participate in the synthesis of DNA in vivo
must in some manner augment base selection.
see Refs. 15 and 41). However, the mechanisms for catalysis
by all these enzymes appear to be very similar if not identical
(Chart 2). They have the same fundamental requirements: 4
complementary deoxynucleoside triphosphates, Mg2+or Mn2+,
and a polynucleotide that serves as a template initiator (34).
They polymerize by adding deoxynucleotides
onto the
3'-hydroxy terminus of a DNA initiating strand. These
enzymes exhibit little specificity as to templates; they are able
to copy most natural DNA's as well as many synthetic
polynucleotide templates. Most cellular DNA polymerases are
considered "DNA dependent," in that they preferentially use
DNA or deoxyribonucleotide
templates. DNA polymerases
present in RNA tumor viruses ("reverse transcriptases") are
able to copy both RNA and DNA templates (3, 82). This
dichotomy may not be valid, since E. coli DNA polymerase I,
the best studied "DNA-dependent" polymerase (34), has been
shown to copy RNA's (44). Furthermore, polymerases that are
able to copy polyribonucleotides
have been described in
apparently normal cells (9, 56, 69, 80, 89), and others have
been reported to be stimulated with natural RNA templates
(8,32,57).
Experimental evidence that DNA polymerase participates in
base selection was first provided by Speyer (74), who showed
that T4 bacteriophage mutants with an altered DNA polymer
ase have an increased mutation frequency throughout their
genome. Hall and Lehman (25) extended this important
observation to show that T4 mutant polymerase carries out
DNA synthesis in vitro with a higher error rate than does the
enzyme obtained from its parent strain. Homogeneous DNA
polymerases from bacteria and bacteriophage have an associ
ated 3'-exonuclease activity that can excise mismatched bases
immediately after they are inserted. It has been suggested that
this exonuclease serves in a "proofreading" capacity (10).
However, a proofreading exonuclease does not appear to be an
integral part of eukaryotic (animal) DNA polymerases (11,12,
15, 66, 67). It can be argued that, in eukaryotic cells, DNA is
replicated by the concerted action of a number of enzymes
including a separate exonuclease that carries out a proofreadINITIATOR
~~ii i i n r 3'OH
dATP
dCTP
A G C A A C T
TCGTTGAC
Mg"
dGTP
dTTP
ATGTACTC
TEMPLATE
DNA polymerase
ROLE OF DNA POLYMERASE
Enzymes that copy polynucleotide templates are generically
referred to as polymerases (34). At least 3 DNA polymerases
have been identified in Escherichia coli, and the precise role of
any of these enzymes in either DNA replication or repair has
not been firmly delineated. In animal cells, multiple DNA
polymerases are also found and have been purified (for review,
5The abbreviations used aie: poly(U), polyuridylic acid; poly (I),
polyinosinic acid; poly(C), polycytidylic acid; poly(A), polyadenylic
acid; AMV, avian myeloblastosis viius.
2314
ELONGATED
PPI
,T"p„¿--
AGCAACTG
TCGTTGACATGTCCTC
l l t l l l l t l
l
l
l
l l
i
i
CHAIN
Chart 2. Model for synthesis of DNA based on known properties of
DNA polymerases. The concept of multiple nucleotide binding sites and
coordination of the or-phosphate with the added divalent cation Mg2+
or Mn2+ is taken from the data of Slater et al (71). E. coli DNA
polymerase I has been shown to be a zinc metalloprotein (70, 77) and
other polymerases have been shown to contain zinc (59, 78).
CANCER RESEARCH VOL. 34
On Mutagenic DNA Polymerases and Cancer
ing function. However, purified eukaryotic DNA polymerases
are able to copy polynucleotides in vitro with precision (Table
1). For example, the enzyme purified from sea urchin embryos
will incorporate less than 1 residue of dCMP for every 12,000
nucleotides polymerized using poly[d(A-T)] as a template
(75). An even greater precision has been reported for the small
polymerase from calf thymus (1 error in 100,000) (11). The
latter enzyme has been shown to be totally lacking in
exonuclease activity, even when presented with polynucleo
tides having a terminal incorrectly base-paired nucleotide.
Since polynucleotides can be copied with so few errors in
vitro, DNA polymerases must participate in base selection. The
finding that the purified DNA polymerase from AMV is
defective in that it makes many more mistakes than animal
polymerases (4) reinforces the concept that polymerases can
amplify base selection.
The mechanism by which animal or viral DNA polymerases
select bases without an associated exonuclease that functions
in augmenting base selection is unknown. At least 3
mechanisms are possible: the enzyme can selectively tighten
the binding of the correctly base-paired nucleotide to the
template (a Km effect), the enzyme can selectively weaken the
binding of the incorrect nucleotide (a KI effect), or the
enzyme can function by serving to orient preferentially the
correct nucleotide into a conformation suitable for polymeri
zation (a Vmax effect). Detailed studies on animal polymer
ases comparing binding constants between enzyme-substrate
complexes in the absence of template with Michaelis constants
determined in the presence of templates may help to
distinguish these alternatives.
INFIDELITY OF DNA SYNTHESIS
AND TUMOR PROGRESSION
In our initial studies on the exactness of DNA synthesis in
malignant cells, we chose to compare the ability of normal
stimulated lymphocytes and lymphocytes from patients with
acute lymphatic leukemia to copy polynucleotides of defined
structure (77). In order to study the fidelity of DNA synthesis
in such systems, it is necessary to use cell extracts and
synthetic polynucleotides. The errors observed in this artificial
situation may not be indicative of DNA replication. Lympho
cytes obtained from normal human peripheral blood are
essentially nondividing cells and have little DNA polymerase
activity (22,43,61,87). They can be transformed into rapidly
replicating cells by mitogenic agents, during which time (3
days in culture) DNA polymerase activity increases 20- to
200-fold (43). Lymphocytes from patients with acute
lymphatic leukemia have large amounts of DNA polymerase
activity (22, 56). Both cells have exceptionally small amounts
of DNase activity (42). Poly[d(A-T)] -poly [d(A-T)], a doublestranded polynucleotide consisting of only alternating residues
of deoxyadenylic acid and deoxythymidylic acid, was used as
a template. When copied without error only dAp and dTp
would be incorporated into the newly synthesized product
(see Chart 3A). However, if 2 residues of dCp were
incorporated for every 8 correct nucleotides polymerized, the
error rate would be 2 in 8. In these experiments we found that
nucleic acid-free extracts from leukemic lymphocytes incorpo
rated 10-fold greater amounts of dCTP than similar extracts
from normal lymphocytes stimulated with phytohemagglutinin (77). The error rate with 4 different leukemic extracts
varied from 1 in 250 to 1 in 850, while the range of 7 normal
extracts was 1 in 1,500 to 1 in 31,000.
In these experiments one measures the incorporation of an
exceedingly small fraction of the total radioactive incorrect
nucleotide present in the reaction mixture. Furthermore, there
is evidence for the presence of a terminal nucleotidyl-transferase in leukemic lymphocytes (46, 79) and such an activity
could catalyze the addition of dCMP onto polynucleotide
chains. Thus, to substantiate that the radioactivity in the
product represents errors in base pairing, a large number of
control experiments are mandatory. It was found that: (a) the
radioactivity in the product was indeed dCMP; (b) the dCMP
was evenly distributed throughout the poly[d(A-T)] product
and not simply located at the termini; and (c) the product
containing dCMP banded on CsCl gradients at a density
corresponding to poly[d(A-T)]. In these experiments, the
incorporation of dCTP required the presence of the comple
mentary nucleotides dATP and dTTP. Moreover, hydrolysis of
the product with a processive nuclease indicated that both the
correct and incorrect mononucleotides appeared simultane
ously, indicating that the incorrect nucleotides were distrib
uted evenly throughout the chain. The requirements for the
Table 1
Fidelity in copying polynucleotide templates
ofDNA
Source
polymeraseSea
urchinCalf
(3S)E.thymus
rate"<1/12,000<1/180,000<
-d(pT)600
TTpoly[d(A-T)]poly[d(A-T)]poly[d(A-T)]r(pA)
IE. coli polymerase
,000<1/100,000<1/12,0001/6001/6001/6,000Ref.75117725,77444
1/1 84
IBacteriophage
coli polymeiase
T4AMVAMVAMVTemplatepoly[d(A-T)Jd(pA)
-d(pT)2500
15poly(dA)-oligo(dT)poly[d(A-T)]NoncomplementarynucleotideincorporateddCTPdCTPdCTPdGTPdCTPdCTPdCTPd
0 Error rate, amount of noncomplementary
polynucleotide synthesized.
SEPTEMBER
1974
nucleotide incorporated over the amount of
2315
L. A. Loeb et al.
A.'ONA-dependtnl"
Polymerose
--dAp-dTp-dAp-dTp-dAp-dTp-dAp-dTp-dAp-dT-OH
,',' I!
i!
!¡ !'
'i
"
'
;
--dTp-dAp-dTp-dApdTp - dAp- dTp- dAp - dTp - dAp-
dATP
--¡JAp-dT-OH
dCTP
--dTp-dAp-dTp-OAp-dTp-dAp-dTp-dAp-dTp-dAp
Poly
d(A-T)»poly
dtA-TJ
Corrici Synth»»!»
dCTP-(*H)
M,"
Eniyme
^dAp-dTp-dAp-|dCp]-dAp-dTp-dAp-dTp-[dCp]-dA-OH
--dTp-dAp-dTp
- dAp - dTp-dAp-dTp
- dAp - dTp - dAp-
Error Rote. Z in 8
B. "RNA-dependent " Polymerote
<JT-32P-PP
-dTp-dT-OH
II
II
- Ã-Ap-rAp-rAp-rAp-rAp-iAp-rAp-rAp-rAp
2500 Nucleolides
P«'»'AP
^Q-
">[">°"T75~
dCTP-(*H)
M,"
-OTp-dTp-dTp-dTp-dTp-dTp-dTp-dTp-dTp-dT-OH
-rAp-rAp-rApCorrect
rAp- rAp - rAp - rAp - rAp - rAp - rAp Synlheii»
Enzyme
-- dTp-dTp- dTp- jdCfj-dTp-dTp-dTp-pCpJ-dTp-dT-OH
rAp-rAp-rApError
rAp-rAp-rApRote:
rAp - rAp - rAp-
rAp -
2 in 8
Chart 3. Assay foi fidelity using polynucleotide templates.
incorporation of both complementary and noncomplementary
nucleotides were the same and typical for a reaction catalyzed
by DNA polymerase. There are 2 direct but not mutually
exclusive interpretations for these observations: that the
infidelity of DNA synthesis in the leukemic extract is
catalyzed by altered cellular DNA polymerases in these cells;
or that a faulty polymerase, perhaps of viral origin, is present
in the leukemic cells, and not in normal cells.
We have been wondering whether DNA synthesis is less
exact in malignant cells than in normal cells. If so, this might
account for many of the aspects of tumor progression. The
concept of tumor progression as formulated by Foulds (19)
envisions that as tumors develop there is an ever-changing
array of cell phenotypes. This is manifested by changes in
antigenicity and increasing numbers of chromosomal aberra
tions as tumors evolve, etc. This concept could account for the
emergence of drug resistance to a variety of chemotherapeutic
agents with the progressive development of anaplasia. These
changes, which appear to be independent, multifocal, and in
some respects random, could represent a manifestation of a
cascading of errors in genetic information. One possible
mechanism for an expanding number of mutations would be if
DNA polymerases or other enzymes that guarantee the
accuracy of DNA replication were defective, so as to permit an
abnormally high number of errors in base pairing during DNA
replication. However, at present there is no direct evidence
that the frequency of mutations in tumor cells is greater than
it is in normal cells.
Altered DNA Polymerases. The 1st interpretation, that of
altered cellular DNA polymerases in leukemic cells, is open to
experimental attack. Methods have been developed for the
purification of DNA polymerases from human leukemic
lymphocytes as well as normal lymphocytes stimulated to
divide in culture with phytohemagglutinin (72). The exactness
by which these enzymes copy different polynucleotides can be
determined. However, if genes coding for polymerases are
subjected to random mutations, the polymerases might be
2316
altered differently in individual cells. During enzyme purifica
tion many of these minor species might go undetected.
Another approach that may be used in combination with
purification is to assay directly for the presence of altered
DNA polymerases in crude extracts. Several investigators (29,
36, 38) have reported evidence that random amino acid
substitutions frequently render enzymes more easily denatured
by heat. If one heats a mixture containing a single enzyme
species, the decay rate should be 1st order (29). The presence
of an altered species in the mixture of DNA polymerases
would be manifested by an increase in the rate of loss of
activity. What may be unique concerning the assays for errors
in a mixture of polymerases containing error-prone altered
enzymes is that the error rate should decrease at a greater rate
than the loss of total polymerase activity; i.e., heating would
be curative. This may supply an explanation for the unusual
sensitivity of some neoplastic cells both in culture and in vivo
to the lethal action of elevated temperatures (23).
The concept of expanding numbers of errors in genetic
information during malignant progression is in some ways
analogous to theories of error accumulation during aging (53,
81). Szilard (81) formulated a theory of somatic mutation in
which aging results from random chromosomal "hits" that
functionally destroy chromosomes of somatic cells, rendering
all genes carried by that chromosome inactive. Orgel (53, 54)
viewed aging as an accumulation of errors in protein synthesis,
leading to an increasing breakdown in the accuracy of the
protein-synthetic machinery and ultimately to lethal "error
catastrophe." Support for the accumulation of altered proteins
during aging in culture is provided by measurements of glucose
6-phosphate dehydrogenase and 6-phosphogluconate activity
in cultures of human fibroblasts (29). The amount of
thermolabile enzyme increased with the number of times the
cultures were propagated, suggesting that there was an
increasing heterogeneity of these enzymes presumably repre
senting altered amino acid sequences. If cellular aging is
mediated by increasing numbers of random errors in protein
CANCER RESEARCH VOL. 34
On Mutagenic DNA Polymerases and Cancer
synthesis, some of these errors may lead to altered enzymes
involved in base selection during DNA synthesis (29, 54). Thus
inaccuracies in protein synthesis during aging may be coupled
to somatic mutations during tumor progression. For both
concepts it is difficult to explain the immortality of
transformed
cell Unes without invoking mechanisms for
cellular selection. This may be a realistic explanation, for when
wild-type E. coli is cocultivated with mutants having increased
mutation rates it is the mutant progeny that emerge as the
dominating species (52).
From a practical viewpoint the presence of altered cellular
polymerases in malignant cells may provide a target for
chemotherapeutic attack (76). These altered enzymes, which
permit the incorporation of noncomplementary
nucleotides
into DNA, may also permit the incorporation of certain
nucleotide analogs. The nucleoside analogs could be phosphorylated by cellular kinases in both normal and malignant cells
but could be incorporated only into DNA of malignant cells
due to the altered polymerases. Wisely chosen, these analogs
might serve to terminate DNA synthesis. For example, they
might be blocked at position 3'' of the sugar and not be able to
form a phosphodiester link with the next incoming nucleoside
triphosphate. It could be argued that altered polymerases in
each tumor may have a different amino acid sequence. If so, it
may be necessary to screen the response of tumor cells from
each individual to a panel of different analogs.
Infidelity of Reverse Transcriptase. The 2nd explanation,
that a viral polymerase is present in leukemic extracts and is
responsible for the high error rate observed, immediately
directs one's attention to determining the fidelity of DNA
polymerases from known animal RNA tumor viruses (reverse
transcriptases).
The hypothesis that reverse transcriptases
allow more errors in base pairing is particularly attractive
because the statistical chance of introducing an error into a
large genome, such as that of a eukaryotic cell, is much greater
than that of introducing an error into a small one such as that
of a virus. Populations of viruses might be able to tolerate a
greater rate of errors than can populations of eukaryotic cells.
We have investigated the exactness by which the DNA
polymerase from AMV copies polynucleotide templates (4,
75). Since this polymerase can be extensively purified in
relatively large amounts, it has been possible to study in detail
the fidelity with which this enzyme copies different templates.
We found that AMV polymerase makes an exceptionally large
number of mistakes in copying a variety of homopolymer
templates. For studies with purified enzymes the fidelity assay
was designed to measure the concurrent incorporation of the
complementary and noncomplementary
nucleotides using a
template-initiator
complex of defined size. The assay is
illustrated in Chart 3B. The template was polyriboadenylic
acid, 2500 nucleotides long, the initiator was an oligomer of
thymidylic acid, and both dTTP-a-32P and dCTP-3H were
present in the reaction mixture. Since only a small amount of
the noncomplementary
dCTP would be incorporated, its
specific radioactivity is much greater than that of dTTP. Using
a variety .of homopolymers as templates the error rate was
found to be approximately 1 in 600, while with alternating
polynucleotides the error rate was 1 in 6000 (4). Even though
the viral polymerase copies polyribonucleotide template with
greater efficiency than polydeoxyribonucleotide
template, the
SEPTEMBER
1974
error rate appears to be the same. That the polymerase
catalyzes the incorporation of the noncomplementary nucleo
tides is evidenced by the coincidence of incorrect and correct
nucleotide incorporations
after the purified enzyme was
chromatographed in a variety of systems. With the exception
of nucleotide concentration, the error rate was independent of
variables in the reaction mixtures, such as the concentrations
of magnesium, template, and enzyme, as well as the number of
initiating sites per template. The errors appeared to be spaced
evenly in the DNA product, in that degradation of the product
of the reaction with a processive exonuclease simultaneously
renders correct and incorrect nucleotides acid soluble. The
ratio of correct to incorrect nucleotide incorporation was
constant as the product grew in length during the reaction.
Most importantly, nearest-neighbor base frequency studies on
the product showed that 70% of the noncomplementary
dCMP's are adjacent to dTMP. These studies suggest that most
of the mistakes are randomly distributed throughout the
product. It is now crucial to determine whether this high error
rate is a characteristic of DNA polymerase in animal tumor
viruses. If so, this faultiness may be of value in identifying the
presence of tumor virus in malignant cells and may provide
clues on the mechanism of viral transformation.
hi the search for a human tumor virus, considerable effort
has been directed at finding viral genetic information (6, 27)
and viral polymerases (22, 63) in human cancer cells. Even
though the ability to copy RNA is not a unique property of
reverse transcriptases (44), it is very useful operationally for
their identification. The papers of Gallo et al. (22) and
Sarngadharan et al. (65) present evidence suggesting the
presence of a viral polymerase in leukemic cells. The enzyme is
able to copy RNA templates, is associated with particles
having a density characteristic of RNA tumor viruses (65), and
can be inhibited by antisera prepared against a partially
purified polymerase from a primate RNA virus (84). If
infidelity is a property diagnostic of an RNA tumor virus
polymerase, it will be crucial to determine how exactly the
RNA-dependent DNA polymerase in human leukemic cells
copies polynucleotide templates.
In general, DNA polymerases are nonspecific. They show
little preference for homologous DNA templates (41). It is
conceivable that the introduction of a faulty DNA polymerase,
i.e., a polymerase from an oncogenic virus, into a normal cell
could cause cellular mutations by inaccurately copying host
DNA (75). Such mutations could alter cellular control
mechanisms for division and proliferation leading to malignant
transformation and tumor progression. Among the mutations
may be those in structural genes that code for cellular
polymerases causing the synthesis of altered enzymes.
Even though "malignant" transformation in culture has
provided a powerful tool for studying how cultured cells
acquire the phenotypic changes associated with cancer, this
transformation cannot be entirely equated with the initiation
of cancer in animals. In culture, the phenotypic changes
characterizing malignant transformation are reversible (2,33).
However, there is no definitive evidence that the revertants are
nonmalignant when injected into appropriate hosts. Some
recent experiments suggest that viral polymerases are necessary
for successful transformation
of normal cells in culture.
Noninfectious mutants, or Rous sarcoma virus and mouse
2317
L. A. Loeb et al.
sarcoma virus that lack DNA polymerase, are unable to
transform cells in culture (26, 55). Ting et al. (83) showed a
direct correlation between the inhibition of virus polymerase
by rifamycin derivatives and transformation in vitro by mouse
sarcoma virus and Rauscher leukemia virus. Other experiments
(93) suggest that inhibition of Rauscher leukemia virus
polymerase is correlated with diminished leukemogenic ability
in mice. The above experiments might only indicate that the
polymerase in the virion is required for viral integration or
replication and that these processes are prerequisites for
malignant transformation. The finding of polymerases that are
temperature sensitive with respect to fidelity may allow one
directly to relate errors in DNA synthesis to mau'gnant
transformation. A number of temperature-sensitive mutants of
Rous sarcoma virus have been isolated that have different
properties with respect to viral replication and transformation
(2, 7, 33, 40, 45). It will be important to determine the
fidelity of DNA synthesis by purified polymerases from such
temperature-sensitive mutants.
UNIFIED
CONCEPTS
We (76) and others (1,17,51,92)
have proposed that errors
in DNA synthesis are causally related to malignant alterations.
We have focused on the incorporation of noncomplementary
nucleotides into cellular DNA by mutagenic polymerases since
this limited concept can be experimentally verified. We have
neglected to consider other altered enzymes that might effect
base selection during DNA replication. For example, a similar
hypothesis could be formulated on deficiencies in DNA repair;
however, insufficient information is available about DNA
repair in animal cells to design crucial experiments (39). We
have emphasized that base-pairing errors in DNA could be
promoted by altered cellular polymerases or by polymerases of
viral origin that are defective in base selection. In the latter
case we are suggesting that a viral polymerase is able to copy
host-cell DNA.
The concept that tumor progression proceeds by the
alterations of enzymes functioning in base selection so as to
produce increasing numbers of mutations is diagrammed in
Chart 4. It is assumed that some of these mutational events
would either bypass or damage the machinery of the cell for
repair and would be expressed as phenotypic alterations.
Selection for more rapidly proliferating cells would have to be
on the basis of phenotypic changes. Increased numbers of
lethal and neutral mutations also would be expected. Neutral
mutations would survive and be capable of undergoing
subsequent genotypic change; the nature of the resulting
phenotype would determine its importance in the progression
of malignancy (49).
If malignant progression proceeds as a result of enzymes
altered in their base-selection function, with consequent
increased numbers of errors in DNA, one can ask what initially
brings about these altered enzymes. The experiments of Prehn
(60), which show that each hydrocarbon-induced sarcoma in
mice exhibits a unique set of tumor-specific transplantation
antigens, could result from different mutational events in each
tumor. Most chemical carcinogens are mutagens or can be
activated (49, 91) to serve as mutagens (1, 28). As such, they
form conjugates with protein, RNA, and DNA. The conse
quences of nucleotide-carcinogen interactions are being ac
tively investigated (24, 35, 37, 85). For example, the
interaction of the carcinogen ./V-2-acetylaminofluorene with
polynucleotides has been extensively studied by Kramer et al.
(35) and Levine et al. (37). Their results clearly indicate that
the carcinogen binds covalently to the C-(8) position of
guanine and suggest that similar modifications of messenger
RNA would alter its function in ribosomal binding and
translation. In the context of this proposal it is important to
know whether interactions of chemical carcinogens with DNA
bring about errors in base selection when the modified DNA is
VIRUS
NEOPLASTIC
EVENTSSirLALTERED
TUMOR
PROGRESSI1RAPID
DNA
POLYMERASE-V*
*INCREASED
* *
Chart 4. Accumulation of genetic
error during tumor progression.
PROLIFERAT'
'PHENOTYPES
lCARCINOGEN\VN1ION
*ITH SELECTIVE
VALUE FOR
PROLIFERATION
MUTATION
FREQUENCY]iNEW
^^-
PROGENY
GENOTYPESPROGENY
PHENOTYPESPARENTAL
GENOTYPESLETHALMUTANTSUNAFFECTED—
IN HOSTPRIMARY
2318
PHENOTYPES
CANCER
RESEARCH
VOL. 34
On Mutagenic DNA Polymerases and Cancer
copied by different DNA polymerases. Studies on this aspect
of infidelity are in progress in our laboratory. Alternatively, it
has been suggested that chemical carcinogens may interact
with polymerases to alter their base-selective function (51).
The principal target protein of the azocarcinogen and also the
protein-carcinogen conjugate have been extensively studied
and purified by Sorof et al. (Ref. 73; Sani, B. P., Mott, D. M.,
Jasty, V., and Sorof, S. Properties of the Target Protein of
Azocarcinogens,
personal communication).
One can ask
whether these proteins are involved in DNA synthesis and
whether the interaction of the target protein with the
carcinogen affects the fidelity of DNA replication.
Similarly, we desire to know not only how ionizing
radiation and UV radiation alter the structure of macromolecules but, more importantly, what the consequences of these
random alterations are. Saffhill and Weiss (64) found that,
after DNA was exposed to UV radiation and used as a
template for E. coli DNA polymerase I, there was a change in
the base composition of the newly synthesized DNA product.
More complete analysis will be required to determine whether
this alteration in the base composition of the newly
synthesized DNA represents errors in base pairing or preferen
tial synthesis of particular regions of the template.
In this essay we have considered many speculative concepts
that suggest that tumors evolve by an expanding number of
errors in genetic information. The essential idea is that
malignant changes result from random mutations and as such
are for all practical purposes genotypically irreversible. We
have presented theoretical arguments and experimental results
indicating that DNA polymerases play a role in base selection.
Perhaps the most direct evidence is the simple demonstration
that different DNA polymerases have different error rates
when copying the same templates. If mutations produce
defective polymerases, the polymerases in turn will produce
further genetic defects.
Even though most carcinogenic agents have been shown to
be mutagenic, it has been difficult to consider tumor viruses in
this framework. Our finding that a DNA polymerase from an
animal tumor virus is error prone suggests the possibility that
oncogenic viruses cause random mutations in this manner. We
are proposing that infidelity of DNA replication may be
responsible for tumor oncogenesis and progression. This
infidelity may be brought about by the introduction of
error-prone viral polymerases into cells or by mutations in
genes coding for cellular polymerases.
Acad. Sei. U. S., 70: 2281-2285, 1973.
2. Bader, J. P. Temperature-Dependent Transformation of Cells
Infected with a Mutant of Bryan Rous Sarcoma Virus. J. Virol.,
10: 267-276, 1972.
3. Baltimore, D. Viral RNA-Dependent DNA Polymerase. Nature,
226:1209-1211,1970.
4. Hallula. N., and Loeb, L. A. The Infidelity of Avian Myeloblastosis
Virus DNA Polymerase in Polynucleotide Replication. J. Biol.
Chem., 249: 4086-4093, 1974.
5. Bautz, E. K. F., and Bautz, F. A. The Influence of Non-comple
mentary Bases on the Stability of Ordered Polynucleotides. Proc.
Nati. Acad. Sci. U. S.,52: 1476-1481, 1964.
6. Bayt, W., Yates, J. W., Wallace, H. J., Jr., Holland, J. F., and
Spiegelman, S. Leukemia-Specific DNA Sequences in Leukocytes
of the Leukemic Member of Identical Twins. Proc. Nati. Acad. Sci.
U. S., 70: 2629-2632, 1973.
7. Biquard, J. M., and Vigier, P. Characteristics of a Conditional
Mutant of Rous Sarcoma Virus Defective in Ability to Transform
Cells at High Temperature. Virology, 47: 444-455,1972.
8. Bobrow, S. N., Smith, R. G., Reitz, M. S., and Gallo, R. C.
Stimulated Normal Human Lymphocytes Contain a RibonucleaseSensitive DNA Polymerase Distinct from Viral RNA-Directed DNA
Polymerase. Proc. Nati. Acad. Sei. U. S., 69: 3228-3238, 1972.
9. Bolden, A., Fry, M., Muller, R., Citeralla, R., and Weissbach, A.
The Presence of Polyriboadenylic Acid-Dependent DNA Polymer
ase in Eukaryotic Cells. Arch. Biochem. Biophys., 153: 26-33,
1972.
10. Brutlag, D., and Kornberg, A. Enzymatic Synthesis of Deoxyribonucleic Acid. XXXVI. A Proof Reading Function for the 3' * 5'
11.
12.
13.
14.
15.
16.
17.
18.
ACKNOWLEDGMENTS
19.
We thank A. Mildvan and S. Sorof of this Institute for generous
counsel and criticism throughout the formulation of the concepts in
this essay. We thank M. W. Springgate of The Johns Hopkins University
for deriving equations relating error rate to statistical probability. We
also thank Dr. J. Glusker for counsel and for defining the coordinates
used in Chart 1.
20.
21.
22.
REFERENCES
23.
1. Ames, B. N., Durston, W. E., Yamasaki, E., and Lee, F. D.
Carcinogens Are Mutagens: A Simple Test System Combining Liver
Homogenates for Activation and Bacteria for Detection. Proc. Nati.
SEPTEMBER
1974
24.
Exonuclease Activity in Deoxyribonucleic Acid Polymerases. J.
Biol. Chem., 247: 241-248, 1972.
Chang, L. M. S. Low Molecular Weight Deoxyribonucleic Acid
Polymerase from Calf Thymus Chromatin. J. Biol. Chem., 248:
6983-6992, 1973.
Chang, L. M. S., and Bollum, F. J. A Comparison of Associated
Enzyme Activities in Various DNA Polymerases. J. Biol. Chem.,
248: 3398-3404, 1973.
Donohue, J. Hydrogen-Bonded Helical Configurations of Polynu
cleotides. Proc. Nati. Acad. Sei. U. S., 42: 60-65, 1956.
Ebbesen, P. Aging Increases Susceptibility of Mouse Skin to DMBA
Carcinogenesis Independent of General Immune Status. Science,
183: 217-218, 1974.
Pansier, B. S. Eucaryotic DNA Polymerases. Their Association with
the Nucleus and Relationship to DNA Replication. Intern. Rev.
Cytol., in press.
Pansier, B. S., and Loeb, L. A. Sea Urchin Nuclear DNA
Polymerase. Methods Enzymol., in press.
Farber, E. Carcinogenesis-Cellular Evolution as a Unifying Thread:
Presidential Address. Cancer Res., 33: 2537-2550, 1973.
Fink, T. R., and Crothers, D. M. Free Energy of Imperfect Nucleic
Acid Helices. I. The Bulge Defect. J. Mol. Biol., 66: 1-12, 1972.
Foulds, L. The Experimental Study of Tumor Progression: A
Review. Cancer Res., 14: 327-339, 1954.
Freese, E. B., and Freese, E. On the Specificity of DNA
Polymerase. Proc. Nati. Acad. Sei. U. S., 57: 650-657, 1967.
Fresco, J. R., and Alberts, B. M. The Accommodation of
Noncomplementary
Bases in Helical Polyribonucleotides and
Deoxyribonucleic Acids. Proc. Nati. Acad. Sci. U.S., 46: 311321,1960.
Gallo, R. C., Yang, S. S., and Ting, R. C. RNA Dependent DNA
Polymerase of Human Acute Leukaemic Cells. Nature, 228:
927-929, 1972.
Giovanella, B. C., Lohman, W. A., and Heidelberger, C. Effects of
Elevated Temperatures and Drugs on the Viability of LI 210
Leukemia Cells. Cancer Res., JO: 1623-1631,1970.
Grunbergen, D., Blobstein, S. H., and Weinstein, B. I. Modification
2319
L. A. Loeb et al.
of Ribonucleic Acid by Chemical Carcinogens. VI. Effect of
Ar-2-Acetylaminofluorene Modification of Guanosine on the Codon
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
2320
Function of Adjacent Nucleosides in Oligonucleotides. J. MoL
Bio1.,82: 459-468, 1974.
Hall, Z. W., and Lehman, I. R. An in vitro Transversion by a
Mutationally Altered T4-Induced DNA Polymerase. J. Mol. Biol.,
36: 321-333,1968.
Hanafusa, H., Baltimore, D., Smoler, D., Watson, K. F., Yaniv, A.,
and Spiegelman, S. Absence of Polymerase Protein in Virions of
Alpha-Type Rous Sarcoma Virus. Science, 177: 1188-1191, 1972.
Hehlman, R., Kufe, D., and Spiegelman, S. RNA in Human
Leukemic Cells Related to the RNA of a Mouse Leukemia Virus.
Proc. Nati. Acad. Sei. U. S., 69: 435-439, 1972.
Heidelberger, C. Current Trends in Chemical Carcinogenesis.
Federation Proc., 32: 2154-2161,1973.
Holliday, R., and Tarrant, M. Altered Enzymes in Ageing Human
Fibroblasts. Nature, 238: 26-30, 1972.
Huang, W. M., and Ts'o, P. O. P. Physico-Chemical Basis of the
Recognition Process in Nucleic Acid Interactions. 1. Interactions of
Polyuridylic Acid and Nucleosides. J. Mol. Biol., 16: 523-543,
1966.
Jordon, F., and Sostman, H. D. Molecular Orbital (CNDO/2 and
MINDO) Calculations on Protonated Deoxyribonucleic Acid Bases.
The Effects of Base Protonation on Intermolecular Interactions. J.
Am. Chem. Soc., 95: 6544-6554, 1973.
Kang, C.-Y., and Temin, H. M. Endogenous RNA-Directed DNA
Polymerase Activity in Uninfected Chicken Embryos. Proc. Nati.
Acad. Sei. U. S., 69: 1550-1554, 1972.
Kawai, S., and Hanafusa, H. The Effects of Reciprocal Changes in
Temperature on the Transformed State of Cells Infected with a
Rous Sarcoma Virus Mutant. Virology, 46: 470-479, 1971.
Kornberg, A. Active Center of DNA Polymerase. Science, 163:
1410-1418,1969.
Kramer, B. S., Fanselan, C. C., and Ludlum, D. B. Reaction of
BCNU (l,3-Bis(2-chloroethyl)-l-Nitrosourea
with Polycytidylic
Acid. Substitution of the Cytosine Ring. Biochem. Biophys. Res.
Commun., 56: 783-788, 1974.
Langridge, J. Genetic and Enzymatic Experiments Relating to the
Tertiary Structure of 0-Galactosidase. J. Bacteriol., 96: 17111717, 1968.
Levine, A. F., Fink, L. M., Weinstein, B., and Grunberger, D. Effect
of jV-2-Acetylaminofluorene Modification on the Conformation of
Nucleic Acids. Cancer Res., 34: 319-327,1974.
Lewis, C. M., and Holliday, R. Mistranslation and Ageing in
Neurospora. Nature, 228: 877-880, 1970.
Lieberman, M. W. Approaches to Analysis of Fidelity of DNA
Repair in Human Diploid Cells. Intern. Rev. Cytol., in press.
Limai, M., and Mason, W. S. Characterization of Two Conditional
Early Mutants of Rous Sarcoma Virus. Virology, 53: 258-273,
1973.
Loeb, L. A. Eucaryotic DNA Polymerases. In: P. D. Boyer (ed.),
The Enzymes, Vol. 10, pp. 173-209. New York: Academic Press,
Inc., in press.
Loeb, L. A., Agarwal, S. S., and Ewald, J. DNA Polymerase and
DNA Replication during Lymphocyte Transformation. Cancer
Res., 30: 2514-2520,1970.
Loeb, L. A., Agarwal, S. S., and Woodside, A. M. Induction of
DNA Polymerase in Human Lymphocytes by Phytohemagglutinin.
Proc. Nati. Acad. Sei. U. S., 61: 827-834, 1968.
Loeb, L. A., Tartof, K. D., and Travaglini, E. C. Copying Natural
RNAs with E. coli DNA Polymerase I. Nature, 242: 66-69, 1973.
Martin, G. S. Rous Sarcoma Virus: A Function Required for the
Maintenance of the Transformed State. Nature, 227: 1021-1023,
1970.
McCaffrey, R., Smoler, D. F., and Baltimore, D. Terminal
Deoxynucleotidyl Transferase in a Case of Childhood Acute
47.
48.
49.
50.
Lymphoblastic Leukemia. Proc. Nati. Acad. Sei. U. S., 70.
521-525,1973.
Mildvan, A. S. Mechanism of Enzyme Action. Ann. Rev. Biochem.,
in press.
Miller, E. C., and Miller, J. A. The Mutagenicity of Chemical
Carcinogens: Correlations, Problems, and Interpretations. In: A.
Hollander (ed.), Chemical Mutagens, Principles, and Methods for
Detection, Vol. 1, pp. 83-119. New York: Plenum Publishing
Corp,, 1971.
Mintz, B. Neoplasia and Gene Activity in Allophenic Mice. In:
Genetic Concepts and Neoplasia. M. D. Anderson Hospital and
Tumor Institute at Houston, Twenty-third Annual Symposium on
Fundamental Cancer Research, pp. 477-517. Baltimore: The
Williams & Wilkins Company, 1970.
Nash, H. A., and Bradley, D. F. Interactions between Trinucleotides: The Electrostatic Contribution and Its Possible Relation to
the Mechanism of Translation of the Genetic Code. Biopolymers,
.?.-261-273, 1965.
51. Nelson, R. L., and Mason, H. S. An Explicit Hypothesis for
Chemical Carcinogenesis. J. Theoret. Biol., 37: 197-200, 1972.
52. Nestman, E. R., and Hill, R. F. Population Changes in Continu
ously Growing Mutator Cultures of Escherichia coli. Genetics, 73
(Suppl.).-41-44,
1973.
53. Orgel, L. E. The Maintenance of the Accuracy of Protein Synthesis
and Its Relevance to Ageing. Proc. Nati. Acad. Sei. U. S., 49:
517-521,1963.
54. Orgel, L. E. Ageing of Clones of Mammalian Cells. Nature, 243:
441-445, 1973.
55. Peebles, P. T., Haapala, D. K., and Gazdar, A. F. Deficiency of
Viral Ribonucleic Acid-Dependent Deoxyribonucleic Acid Polym
erase in Non-Infectious Virus-Like Particles Released from Murine
Sarcoma Virus-Transformed Hamster Cells. J. Virol., 9: 488-493,
1972.
56. Penner, P. E., Cohen, L. H., and Loeb, L. A. RNA-Dependent DNA
Polymerase in Human Lymphocytes during Gene Activation by
Phytohemagglutinin. Nature, 232: 58-61,1971.
57. Penner, P. E., Cohen, L. H., and Loeb, L. A. RNA-Dependent DNA
Polymerase: Presence in Normal Human Cells. Biochem. Biophys.
Res. Commun., 42: 1228-1234, 1971.
58. Pitha, P. M., Huang, W. M., and Ts'o, P. O. P. Physicochemical
59.
60.
61.
62.
63.
64.
65.
66.
Basis of the Recognition Process in Nucleic Acid Interactions. IV.
Costacking as the Cause of Mispairing and Intercalation in Nucleic
Acid Interactions. Proc. Nati. Acad. Sei. U. S., 61: 332-339, 1968.
Poiesz, B. J., Battuta, N., and Loeb, L. A. Zinc in Reverse
Transcriptase. Biochem. Biophys. Res. Commun., in press.
Prehn, R. T. Tumor-Specific Antigens of Putatively Nonviral
Tumors. Cancer Res., 28: 1326-1330,1968.
Rabinowitz, Y., McCluskey, I. S., Wong, P., and Wilhite, B. A.
DNA Polymerase Activity of Cultured Normal and Leukemic
Lymphocytes. Response to Phytohemagglutinin. Exptl. Cell Res.,
57:257-262,1969.
Raszka, M., and Kaplan, N. O. Association by Hydrogen Bonding
of Mononucleotides in Aqueous Solution. Proc. Nati. Acad. Sci.
U. S., 69: 2025-2029, 1972.
Robert, M. S., Smith, R. G., Gallo, R. C, Sarin, P. S., and Abrell, J.
W. Viral and Cellular DNA Polymerase. Comparison of Activities
with Synthetic and Natural RNA Templates. Science, 176:
798-800, 1972.
Saffhill, R., and Weiss, J. J. Effects of Ionizing Radiation on in
vitro Replication of DNA by Polymerase I. Nature New Biol.,241:
69-71,1973.
Sarngadharan, M. G., Sarin, P. S., Reitz, M. S., and Gallo, R. C.
Reverse Transcriptase Activity of Human Acute Leukemic Cells:
Purification of the Enzyme, Response to AMV 70S RNA, and
Characterization of the DNA Product. Nature, 240: 67-72, 1972.
Sedwick, W. D., Wang, T. S. F., and Korn, D. Purification and
CANCER
RESEARCH
VOL. 34
On Mutagenic DNA Polymerases and Cancer
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
Properties of Nuclear and Cytoplasmic Deoxyribonucleic Acid
Polymerases from Human KB Cells. J. Biol. Chem., 247:
5026-5033, 1972.
Sedwick, W. D., Wang, T. S. F., and Korn, D. The DNA
Polymerases of KB Cells. Methods Enzymol., in press.
Seidel, H. Vergleichende Unterssuchengen überdie Stabilitatsanderung der doppelstrange Poly(A) Poly(U) and Poly(I) Poly(C)
nach partieller N-Oxydation des A-Beziehung-Sweise C-Stranges.
Biochim. Biophys. Acta, 729: 414-416, 1966.
Slater, L, and Slater, D. W. DNA Polymerase Potentials of Sea
Urchin Embryos. Nature New Biol., 237: 81 -85, 1972.
Slater, i. P., Mildvan, A. S., and Loeb, L. A. Zinc in DNA
Polymerase. Biochem. Biophys. Res. Commun., 44: 37-43, 1971.
Slater, J. P., Tamir, L, Loeb, L. A., and Mildvan, A. S. The
Mechanism of E. coli DNA Polymerase I: Magnetic Resonance and
Kinetic Studies of the Role of Metals. J. Biol. Chem., 247:
6784-6793, 1972.
Smith, G. R., and Gallo, R. C. DNA-Dependent DNA Polymerases I
and II from Normal Human Blood Lymphocytes. Proc. Nati. Acad.
Sei. U. S., 69: 2879-2884, 1972.
Sorof, S., Kish, V. M., and Sani, B. Purification and Properties of
the Principle Liver Protein Conjugate of a Hepatic Carcinogen.
Biochem. Biophys. Res. Commun., 48: 860-865, 1972.
Speyer, J. F. Mutagenic DNA Polymerase. Biochem. Biophys. Res.
Commun., 27:6-8, 1965.
Springgate, C., Battula, N., and Loeb, L. A. Infidelity of DNA
Synthesis by Reverse Transcriptase. Biochem. Biophys. Res.
Commun., 52: 400-406, 1973.
Springgate, C. F., and Loeb, L. A. On the Infidelity of DNA
Synthesis in Human Lymphocytes. Res. Commun. Chem. Pathol.
Pharmacol., 4: 651-655, 1972.
Springgate, C., and Loeb, L. A. On Mutagenic DNA Polymerases in
Human Leukemic Cells. Proc. Nati. Acad. Sei. U. S., 70: 245-249,
1973.
Springgate, C. F., Mildvan, A. S., Abramson, R., Engle, J. L., and
Loeb, L. A. E. coli DNA Polymerase I, a Zinc Metalloenzyme. J.
Biol. Chem., 248: 5987-5993, 1973.
Srivastava, B. I. S., and Minowada, J. Terminal Deoxynucleotidyl
Transferase Activity in a Cell Line (Molt-4) Derived from the
Peripheral Blood of a Patient with Acute Lymphoblastic Leukemia.
Biochem. Biophys. Res. Commun., 51: 529-535, 1973.
Stavrianopoulos, J. G., Karkas, J. D., and Chargaff, E. DNA
SEPTEMBER
1974
81.
82.
83.
84.
85.
Polymerase of Chicken Embryo: Purification and Properties. Proc.
Nati. Acad. Sei. U. S., 69: 1781-1785,1972.
Szilard, L. On théNature of the Ageing Process. Proc. Nati. Acad.
Sci. U. S., 45: 30-45,1959.
Temin, H. M., and Mizutani, S. RNA-Dependent DNA Polymerase
in Virions of Rous Sarcoma Virus. Nature, 226: 1211-1213, 1970.
Ting, R. C., Yang, S. S., and Gallo, R. C. Reverse Transcriptase,
RNA Tumor Virus Transformation, and Derivatives of Rifamycin
SV. Nature New Biol., 236: 163-166,1972.
Todaro, G. J., and Gallo, R. C. Immunologie Relatedness between
a DNA Polymerase from Human Acute Leukemic Cells and Primate
and Mouse Leukemia Virus Reverse Transcriptase. Nature, 244:
206-209, 1973.
Troll, W., Belman, L., Berkowitz, E., Chmielewicz, Z. F., Ambrus,
J. L., and Bardos, T. J. Differential Responses of DNA and RNA
Polymerase to Modifications of the Template Rat Liver DNA
Caused by Action of the Carcinogen Acetylaminofluorene in 'vivo
and in vitro. Biochim. Biophys. Acta, 157: 16-24, 1968.
86. Tuppy, H., and Kuchler, E. Evidence for Specific Interactions
between Complementary Nucleosides in a Chromatographie Sys
tem. Biochim. Biophys. Acta, SO: 669-671, 1964.
87. Tyrsted, G., Munch-Petersen, B., and Cióos,L. DNA Polymerase
Activity in Phytohemagglutinin-Stimulated
and Non-Stimulated
Human Lymphocytes. Exptl. Cell Res., 77: 415-427, 1973.
88. Wang, A. C., and Kallenbach, N. R. Helical Complexes of
Polyriboinosinic Acid with Copolymers of Polyribocytidylic Acid
Containing Inosine, Adenosine, and Uridine Residues. J. Mol. Biol.,
62:591-611,1971.
89. Ward, D. C., Humphreys, K. C., and Weinstein, I. B. Synthetic
RNA-Dependent DNA Polymerase Activity in Normal Rat Liver
and Hepatomas. Nature, 237: 499-503, 1972.
90. Watson, J. D., and Crick, F. H. C. The Structure of DNA. Cold
Spring Harbor Symp. Quant. Biol., 18: 123-131, 1953.
91. Weisburger, J. H., and Weisburger, E. K. Biochemical Formation
and Pharmacological Properties of Hydroxylamines and Hydroxamic Acids. Pharmacol. Rev., 25: 1-66, 1973.
92. Wheldon, T. E., and Kirk, J. An Error Cascade Mechanism for
Tumour Progression. J. Theoret. Biol., 42: 107-111, 1973.
93. Wu, A. M., Ting, R. C., and Gallo, R. C. RNA-Directed DNA
Polymerase and Virus-Induced Leukemia in Mice. Proc. Nati. Acad.
Sei. U. S., 70: 1298-1302, 1973.
2321