[CANCER RESEARCH 56. 2130-2136.
May 1. 1996]
Deletions and Insertions in the p53 Tumor Suppressor Gene in Human Cancers:
Confirmation of the DNA Polymerase Slippage/Misalignment Model
Marc S. Greenblatt, Arthur P. Grollman, and Curtis C. Harris1
Laboratory of Human Carcinogenesis, National Cancer Institute, National Institutes of Health, Bethesda,
Pharmacological Sciences. Slate University of New York at Stony Brook, Stony Brook, New York 11794 [A. P. G.J
ABSTRACT
We analyzed all published deletions and insertions in the p53 gene to
assess the relevance of mutagenesis
models. Almost all deletions and
insertions can be explained by one or more of the following DNA sequence
features: monotonie base runs, adjacent or nonadjacent repeats of short
tandem sequences, palindromes, and runs of purines or pyrimidines (homocopolymer runs). Increased length of monotonie runs correlates posi
tively with increased frequency of events. Complex frameshift mutations
can be explained by the formation of quasi-palindromes, with mismatch
excision and replication using one strand of the palindrome as a template.
Deletions and insertions in the p53 tumor suppressor gene may reflect
both spontaneous and carcinogen-induced mutagenesis.
INTRODUCTION
The somatic mutations that accumulate during multistep carcinogenesis include major chromosomal events (such as translocations,
allelic deletions, and gene rearrangements) and point mutations (such
as base substitutions and small deletions or insertions; Ref. l). The
relative frequencies of different classes of mutations vary among
cancer-related genes. For example, mutations found in the APC tumor
suppressor gene are most often short deletions or insertions, whereas
activating mutations in the ras proto-oncogene are base substitutions
(1). The largest data base of somatic mutations in human cancers
exists for the p53 tumor suppressor gene (2, 3). All classes of muta
tions are found. Point mutations, often accompanied by deletion of the
second alÃ-ele,are the most common event. The majority of p53 point
mutations are base substitutions, but 10—15%of both somatic and
germ line p53 mutations are small, intragenic deletions or insertions
(1,4). This data set is large enough to test hypotheses regarding the in
vivo mechanisms of intragenic deletions and insertions.
In vitro mutagenesis studies have established that DNA sequence
context is the most important factor in determining the specificity of
deletions and insertions (reviewed in Refs. 5 and 6). Almost all small
deletions and insertions in model systems occur at monotonie runs of
two or more identical bases, or at repeats of 2-8-bp DNA motifs,
either in tandem or separated by intervening nucleotides. Several
mechanisms have been proposed. The most well-studied mechanism,
first proposed by Streisinger et al. (7) in 1966, involves slippage or
misalignment of the template DNA strands in iterated bases during
replication. This leads to either base deletion (if the nucleotides
excluded from pairing are on the template strand) or insertion (if they
are on the primer strand); replication "fixes" these mutations in the
daughter cells (5, 6). Misalignment can occur "spontaneously" in
monotonie runs or can be initiated by a spontaneous or adductinduced base misincorporation that creates a context for slippage (5,
8, 9).
Misalignment may also occur in repeat sequences that are not
Received 9/11/95; accepted 3/4/96.
The cosls of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
1To whom requests for reprints should be addressed, at Laboratory of Human
Maryland
20892
[M. S. G., C. C. H.J, and Department
of
monotonie runs. When a repeat sequence mispairs with a tandem
(adjacent) or nontandem (nonadjacent) complementary motif, the loop
formed by the intervening oligonucleotide sequence may be deleted or
duplicated (10-12). Such repeats have been identified at the sites of
spontaneous and ionizing radiation-induced deletions (13, 14) and at
the sites of deletions in synthetic DNA sequences with adducted bases
(9). Loops may also form between nonadjacent inverted repeats (pal
indromes), leading to deletion of the loop and sometimes to complex
frameshift mutations involving both deletion and base substitution
(12, 15-17). A review of short (<20-bp) germ line deletions and
insertions in genes associated with human genetic disease confirmed
that these DNA sequence features are the most important factors
determining the specificity of these events (11).
Quantitative mutational spectrum analysis in vitro shows that: (a)
the frequency of frameshift mutations increases as the length of the
monotonie run increases; and (b) single-base deletions are much more
frequent in vitro than are single-base insertions (reviewed in Ref. 5).
The proposed explanation is superior stability of the hydrogen bond
ing of extensive repeats, and G:C-rich regions, and intermediate DNA
structures associated with deletions. DNA polymerase infidelity may
lead to spontaneous frameshift errors; different polymerases vary in
the frequency of errors and the ratio of deletions:insertions (5). The
surrounding sequence context is another major factor affecting the
likelihood of misalignment mutations (9, 18-20).
Limited analysis of smaller p53 mutation data bases has confirmed
the presence of some of these high-risk sequence features, including
monotonie and tandem repeats, at the sites of p53 somatic mutations
(10). Long runs of purines or pyrimidines, known as homocopolymers
(e.g., AGAGAG, CCTTCC, and AAAGGG) have also been noted at
sites ofp53 deletions (21). We updated and reviewed the data base of
all reported somatic p53 mutations (2) and analyzed the subset of all
deletions and insertions in more detail to assess: (a) the extent to
which the mechanisms determined in experimental systems could
account for somatic p53 deletions and insertions; (b) how patterns of
observed p53 deletions and insertions might refine existing models;
and (c) whether relative frequencies of deletions and insertions pro
vide information to distinguish between spontaneous or carcinogeninduced events.
MATERIALS
AND METHODS
The p53 mutation data base has been described and is available via E-mail
from the European Molecular Biology Library (2). We updated the data base
to include all p53 mutations published before September 1994, as identified by
searches of MEDLINE and Current Contents. All intron and exon sequences at
the sites of, adjacent to, and up to 40 bp away from insertions and deletions
were examined, in some cases with the aid of the Gene Jockey sequence
analysis program (Biosoft, Cambridge, United Kingdom). Regression and x2
analyses were performed using standard statistical software.
RESULTS
Since the publication of our original p53 data base (2), the preva
lence of deletions and insertions among all p53 mutations has de
creased from 13 to 11%. This reduction probably reflects a detection
2130
Carcinogenesis. National Cancer Institute. National Institutes of Health. Building 37.
Room 2C05, Bethesda, MD 20892-4255. Phone: (301) 496-2048; Fax: (301) 496-0497.
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.
p53 DELETIONS AND INSERTIONS
Table I DNA sequence features associated with deletions and insertions of pS3 in human cancers
(%)Class
eventAllAllof
Sequence features associated with events
high-risksequence
feature9797951009794961009796Monotonieruns72807775868356867471Tandemrepeats57585667484467436683Tandemrepeats
only1112801083308gHomocopolymerruns565531426665521004738Nontandemrepeats192351700404
deletionsAll
insertionsComplex
frameshifts1
deletions1-bp
-bp
insertions2-bp
deletions2-bp
insertions3
deletions3
+ -bp
+ -bp insertionsn34925978121184827711324Any
bias, because most recent reports have examined only exons 5-8, in
which missense mutations are more common, and have neglected
exons 2-4 and 9-11, in which the majority of mutations are deletions
and insertions (1). The data base contains 370 deletions and insertions
in thep5J gene. Of these, 21 detected by sequencing of cDNA involve
entire exons or large deletions at splice sites, indicating aberrant
mRNA splicing. Because splicing mutations often result from either
point mutations in splice sequences at the intron-exon junction or
dysregulation of the splicing mechanism and not from loss or gain of
DNA, these mutations were not classified as DNA deletions or inser
tions. The patterns and DNA sequence features of the remaining 349
deletions and insertions (or "events") were analyzed.
Sizes and Relative Frequencies ofp53 Deletions and Insertions.
Simple deletions constituted 74% (259 of 349) of all events (71% of
one 1 bp, 76% of two bp, and 82% of &3 bp). Simple insertions
constituted 22% (78 of 349) of the events; the remaining 3% (12 of
349) were complex frameshift mutations that combined deletion,
insertion, and base substitution. In 49% (166 of 349, 118 deletions and
48 insertions) only 1 bp was affected; 10% (34 of 349, 27 deletions
and 7 insertions) involved 2 bp; and 39% (137 of 349, 113 deletions
and 24 insertions) altered 3 or more bp. Only six intragenic deletions
in p53 larger than 30 bp and unassociated with splice sites were
reported; the largest was 137 bp.
DNA Sequence Features at Sites of Deletions and Insertions.
We calculated the proportions of all DNA deletions and insertions
attributable to each of the following mechanisms (Table 1 and Fig. 1):
monotonie base runs, repeats of short tandem sequences, palindromes
(inverted repeats of dyad symmetry), and runs of four or more purines
or pyrimidines (homocopolymer runs).
Almost all [341 (98%) of 349] deletions and insertions can be
explained by one or more of these DNA sequence features. The most
common sequence motifs seen at the site of deletions or insertions are
monotonie runs (two to five consecutive identical bases), present at
the sites of 77% (269 of 349) of all events; only 46% of all p53 bp are
within these runs (odds ratio for an event occurring at a monotonie
run, 3.95; 95% confidence interval, 3.00-5.20; P < 0.0001). Re
peated tandem sequences, either adjacent to or distant from each
other, encompass or flank 30% of deletions or insertions, and palin
dromes flank another 5%. Homocopolymer runs of 4 or more bp occur
at the sites of 56% of events, but in only 3% are they the only
associated feature. Homocopolymer runs of 4 or more bp encompass
42% of all bp in p53-coding exons (odds ratio for an event occurring
at a homocopolymer run, 1.75; 95% confidence interval, 1.38-2.23;
P < 0.0001).
Short deletions and insertions ( 1 bp) are associated with a different
pattern of sequence features than are longer events (2 or more bp;
Table 1). For 1-bp changes (n = 118 deletions; n = 48 insertions),
monotonie runs and direct tandem repeat sequences account for al
most all events. Monotonie runs of 2-5 bp occur at 86 and 83% of
1-bp deletions and insertions, respectively. Where no monotonie
repeat exists (n = 24; 14% of 1-bp events), tandem repeat sequences
are present at 15 (94%) of 16 1-bp deletions and 7 (88%) of 8 1-bp
insertions. Overall, tandem repeats accompany 48% of 1-bp deletions
and 44% of 1-bp insertions; they are the only high-risk DNA se
quences present at 10% of 1-bp deletions and 8% of 1-bp insertions.
Runs of 4 or more homocopolymer bp occur at 64% of 1-bp events,
but in all but one instance (1%), they accompany other high-risk
sequence features.
Deletions of 2 bp (n = 27) are less likely to be associated with
monotonie runs (56%) and more likely to be associated with tandem
repeats (67%). In 33% of these 2-bp deletions, tandem repeats are the
only deletion-promoting DNA sequence features present. However,
six of seven insertions of 2 bp are at monotonie runs, and only three
are at tandem repeats. Homocopolymer runs accompany 52% of
deletions and all seven insertions of 2 bp. Of deletions of 3 or more
bp (n = 134), 68% were at monotonie runs, 58% were at direct
Fig. 1. Sequence features associated with p53 deletions and insertions.
MONO RUN. monotonie base run; TAN RPT. tandem repeat sequence;
HCP RUN, homocopolymer run of four or more bases; NONTAN RPT.
nonlandem repeated sequence (nonadjacent repeat sequence flanks
event); PAL, palindrome.
ANY
HCP
RUN
NONTAN
RPT
Sequence Feature
2131
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.
PAL
pS3 DELETIONS
AND INSERTIONS
DELETION FREQUENCY
0.6
0.5-
0.4 -J
El
0.3
1 BP DELETIONS
•¿
ALL DELETIONS
0.2
0.1
0.0
Fig. 2. Relationship between length of monotonie
repeat and frequency of deletions and insertions in p53.
A, deletions increase in a linear, positive relationship with
length of repeat [R = 0.930 for 1 bp (Q), 0.864 for all
(*)]. B, insertions increase in frequency only at monotonic runs of five bases.
LENGTH OF REPEAT
INSERTION FREQUENCY
B
0.5
0.4
0.3
s
ALL INSERTIONS
•¿
1 BP INSERTIONS
0.2
0.1
0.0
LENGTH OF REPEAT
repeats, and only 43% were accompanied by homocopolymer se
quences. Notably, 43% were flanked by nontandem repeat sequences,
and 10% were flanked by palindromes, a sequence feature that rarely
accompanies shorter events. The same general patterns are seen for
insertions of 3 or more bp (n = 27; Table 1).
Krawczak and Cooper (11) identified a consensus sequence of 6 bp
(TG A/G A/G G/T A/C), which occurs frequently at sites of germ line
deletions in a variety of genetic diseases. We identified this sequence
six times in the coding exons of the p53 gene; 14 somatic p53
deletions or insertions are associated with it, including 11 at one site
in exon 5 (codons 173-175, TG AGG C). With only one exception, at
least one other deletion-associated sequence feature (usually tandem
repeat or palindrome) also is associated with these events. The region
of p53 that most frequently contains deletions or insertions is codons
151-159: CGC CCG CGC CGC ACC CGC GTC CGC GCC. Thirty
mutations (9% of all deletions and insertions, and 1% of all p53
mutations) have been reported at this G:C-rich sequence with multiple
runs and direct repeats.
Increased Length of Monotonie Runs Correlates with Increased
Frequency of Events. To test the prediction of the DNA polymerase
slippage model that deletions and insertions should be more common
at longer repeats (5), we calculated a regression curve of the frequency
of events per site. We divided the number of deletions and insertions
occurring at monotonie runs of 2, 3, 4, and 5 bp by the number of runs
of each length in p53-coding exons times the number of bases per run.
The resulting number describes the frequency of events per bp at
monotonie runs. One run of 6 bp occurs in exon 11, but because this
region is infrequently sequenced (only one deletion here has been
reported), it was not included in this analysis. Fig. 2 shows that for
deletions of 1 bp, the relationship is positive and linear; deletions are
proportionally more frequent at longer runs (R = 0.930). This rela
tionship is present but weaker when deletions of all lengths are
2132
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.
p53 DELETIONS
Table 2 Features of deletions and insertions flanked by palindromes and tandems
nDeletionsInsertionsLength
(%)2 of paired sequence
bp3bp4bp5
bp>5
bpLength
(%)1-3
of deletion
bp4-9
bp10-20bp>20
bpLength
(%)1-3
of insertion
bp4-9
bp10-20
bp>20bpSegment
AND INSERTIONS
sequence context, with both stable palindromes and tandems absent
from the site.
Exogenous versus Endogenous Mutation Analysis by Tumor
Type. The mutational spectrum may be an index of the relative
frequencies of endogenous versus exogenous mutations in tumors and
may help detect patterns characteristic of exogenous mutagens. Pre
viously, we analyzed base substitution patterns among tumor types
and noted that substitutions that are thought to be due to exogenous
carcinogens are more frequent in tumors for which epidemiological
studies indicate a link with carcinogen exposure (1, 4). Endogenous
deamination of 5-methylcytosine at CpG sites results in G:C to A:T
transitions at CpG dinucleotides; CpG site transitions therefore have
been referred to as "spontaneous" mutations, which can occur in the
absence of exogenous mutagens. The prevalence of CpG site transi
tions varies widely among tumor types, generally correlating nega
tively with carcinogen association (1).
(%)Bothdeleted
Because DNA polymerase infidelity alone can produce frameshift
primersIntervening
mutations, deletions and insertions also have been suggested as a
onlyOne
loop
marker of spontaneous mutation (10). However, chemical carcinogens
loopPartial
primer +
loopOtherSegment
primer +
and their adducts can induce frameshift mutations in vitro (6, 18) and,
therefore, may contribute to frameshift mutations in human cancers.
(%)Bothinserted
We compared the prevalence of deletions and insertions among tu
primersIntervening
onlyOne
loop
mors to determine whether they correlated either positively or nega
loopPartial
primer +
tively
with carcinogen exposure and could thus be used as an indicator
loopOtherPalindromes1814401739222271443362500751421-290-757-710SO250250Tandems6560523129201533843130602020710-1257-6517-23000602020
primer +
of endogenous or exogenous mutations.
Deletions and insertions are most common in tumors of the thyroid
(17%), head and neck (15%), breast (15%), and sarcomas (15%); they
are least common in melanoma (5%) and other skin cancers (8%) and
considered, because these events are often associated with other
colon (8%), hepatocellular (8%), and cervical carcinomas (8%). Both
sequence features. Insertions of 1 or more bp are uncommon at runs
of these lists contain tumors that have been frequently (thyroid, head
of two, three, or four bases; this portion of the curve is flat. However,
and neck, skin, hepatocellular, and cervical) and infrequently (sarco
insertions increase sharply in frequency at 5-bp runs (Fig. 2). When
mas and colon) associated with environmental mutagens, and both
only mutations and sequences in the commonly studied exons 5-8
contain tumors that frequently (thyroid and colon) and infrequently
were considered, similar results were obtained (deletions at 4-bp runs
(head and neck and hepatocellular) contain CpG site transitions,
are slightly underrepresented; data not shown). Therefore, selection
markers of mutations unassociated with exogenous mutagens (1).
bias in detection of deletions is unlikely to affect these results.
The prevalence of deletions and insertions varies in tobacco-asso
Nonadjacent Tandem Repeat Sequences and Palindromes. Six
ciated cancers (e.g., cervix, 8%; lung, 10%; bladder, 12%; and head
ty-five mutations occur at nonadjacent tandem repeats, and 18 occur
and neck, 15%). The hot spot region from codons 151-159 is affected
at inverted repeats (palindromes). In all but one of these events, the
in 16% of 36 lung cancer deletions and insertions, including 3 around
length of the repeated motif is 3 or more bp (Table 2). More than
codon 157. Codon 157 is a hot spot for base substitutions in lung
three-fourths of the sequences deleted or inserted between tandem
cancer (1); this sequence may be the target of an important promutarepeats are between 4 and 20 bp long; 40% are between 4 and 9 bp,
genic, tobacco-associated adduci.
and 43% are between 10 and 20 bp, but only 12% are more than 20
bp. Of sequences deleted or inserted between palindromes, however,
DISCUSSION
more than three-fourths are 10 bp or greater; 40% are greater than 20
bp, and few are between 4 and 9 bp. Deletions or insertions flanked by
This analysis confirms many of the prior observations of deletions
tandem sequences usually include one full copy of the repeat and the
and insertions in human disease (10, 11, 21), highlights the relevance
complete sequence between the repeats. Those between inverted re
of many of the mechanistic models for deletion and insertion derived
peats usually include a portion, but not all, of the palindrome; in about
from in vitro mutagenesis studies (5, 6, 9), and suggests refinement of
one-fourth of cases, only the sequence between the repeat is affected
some models.
(Table 2).
The regression curve of the frequency of deletions and insertions
Complex Frameshift Mutations. Complex frameshifts are muta
versus the length of a monotonie base run shows that as the length of
a run increases, the probability of deletions (especially 1-bp deletions)
tions which involve combinations of deletion, insertion, and base
substitution. In a model derived from in vitro studies (6, 17), these
increases in a linear relationship. However, a different relationship is
mutations are explained by formation of quasi-palindromic loops,
noted for insertions. A run of repeated bases must reach a threshold of
which are treated as double-stranded DNA. Deletions, insertions, and
5 bp before insertions are likely to occur in the p53 gene. This
substitutions are generated by excision of mismatches and templatethreshold pattern has been demonstrated for other systems in vitro,
directed replacement of bases complementary to one side of the loop.
such as deletions in DNA polymerase a replication of free DNA (5).
Twelve complex frameshift mutations appear in the data base. Nine
Sequence context surrounding runs of identical length can contribute
of these can be explained by the formation of palindromes with
to differences in deletion and insertion frequency (5, 19, 20); there
template-directed mutation as described by the above model (exam
fore, a simple length-frequency relationship is difficult to interpret.
However, the same sequences in p53 are at risk for both deletions and
ples in Fig. 3; Ref. 17). Two others occur at the sites of tandem repeats
insertions; the fact that their incidence-versHj-length curves differ
of 5-6 bp. Only one complex frameshift is not explained by DNA
2133
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.
pSJ DELETIONS
AND INSERTIONS
CODONS 235 - 241
CODONS 155-161
CC
T
AcxcG
G¡/Ac/
A
C*•
C
AC
fcC (r)
\G/
GT
CG
TC
del 4 C.GCC
ins 1 T
""
G—C
C—
GT©G
mutations fit a model
with DNA excision
of the palindrome as
circled; bold italics.
yAtoT^
AAC fc
^
A-T
C
A^v\cT\\TT
G.
A°T
Q
AG— GC
_AAdel
—¿ins
6 to
10
"~
CTG
/G\T—
—¿
fA\T—
[AG
/GG VA
AT^AA
*QT^lTTGGGGA
GT
C—
T
A— T
C
GA—
—¿
TTG
INTRON 6 - EXON 7, CODON 227
TA
CA
T
T
A
C
C—G
GCTCAC
TAG
jG ^^
CC
CA—
T
A—
TSs$£?GC Cde/3
C
Gr-A7"G—
CODONS 199-213
Fig. 3. Complex frameshift
of quasi-palindrome formation
and replication using one side
a template. Deleted bases are
new bases.
AA
CC
GT
T1
1A
A
C
C
C —¿
G
T
T
G
CTT
G
CT
C
A
C
TTGCGTGTGAGTATTTGG
1G^
T
:C
C
GG
TC
C/l
TC
TA
Ggc
GG
GA
AT
AC
CA
TTO
GAG
C
C
T
C
G
Cmult
fc
subs,
ñdelie
AC
T
AT
G
G
C
C —¿
VT/
C —¿
TT—¿
TC
GT
GT
TU'GGATTCQTCTATA
xG_L/'>»delie
suggests that the intermediate events involved in their generation
differ. One potential explanation is that repair of short heteroduplex
DNA structures predisposing to deletions is less efficient than that of
those leading to insertion (see below).
The proportion of p53 deletions and insertions that represent spon
taneous events due to DNA replication infidelity versus mutageninduced changes remains unclear. Frameshift mutations that seem to
form spontaneously in genomic DNA may be generated by DNA
polymerase-induced errors in replication or may arise as a result of
endogenous or exogenous damage to template DNA. Endogenous
DNA damage is generated by reactive oxygen species (22), products
of lipid peroxidation (23, 24), or other metabolic reactions that alter
DNA structure (25), thereby complicating the assignment of these
mutations to specific environmental agents.
We compared the prevalence of frameshift mutations in tumor
types with epidemiological and mutational spectrum evidence sup
porting carcinogen-induced and endogenous mutagenesis (1); associ
ation with cancers linked to carcinogens would strengthen the infer
ence that frameshifts were induced by exogenous damage. Deletions
and insertions did not correlate well with these indices. This suggests
that differences among tumor types in their prevalence may reflect
both endogenous and carcinogen-induced mutagenesis, such as organspecific carcinogens or tissue-specific properties that regulate repli
cation fidelity. The presence in lung cancers of a hot spot for frameshift mutations in the G:C-rich sequence at codons 151-159, near a
base substitution hot spot (1). suggests that carcinogen-mediated
frameshifts may occur in vivo.
Several models involving
DNA damage-induced
template mis
alignment or slippage during DNA replication have been proposed as
mechanisms of frameshift mutagenesis (5, 6, 8, 9, 17-19, 26-29). The
following sequence of reactions may lead to base deletions: DNA
polymerase encounters a damaged or adducted base at the replication
fork, and DNA synthesis is blocked, often transiently, by helix de
formation or adduci bulk. If chain extension is delayed, the nucleotide(s) at the 3' primer terminus may pair with a complementary
base(s) 5' to the lesion. This misaligned template generates a repli
cation product with deleted bases. The complementary base pairing at
the primer-template junction facilitates extension of the misaligned
chain. Base substitutions occur when insertion of an "incorrect" base
opposite the lesion is followed rapidly by chain extension. Adductinduced base misincorporation may also increase the stability of a
misaligned intermediate.
Frameshift mutations are induced in prokaryotic models by some
bulky chemical mutagens, including aflatoxin B, and arylamines. The
ability of a damaged nucleotide to generate frameshift mutations is
determined by the nature of the base inserted opposite the adduci, the
surrounding sequence context (adduct-induced frameshift mutations
occur more frequently in monotonie or tandem base runs), and the rate
of translesional DNA synthesis estimated from kinetic analysis (9, 19,
20, 30-32). Studies with chemically defined DNA adducts have
established the mutational specificity of some carcinogen-induced
lesions in vitro (31, 33). Although the frequency and sites of muta
tions are not always predictable, the relative propensity for some DNA
lesions to promote frameshift deletions in vitro can be estimated (9).
2134
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.
p5J DELETIONS
AND INSERTIONS
In addition to the effects of specific mutagens, DNA repair capacity
is an important factor affecting mutagenesis. The rates and types of
spontaneous and induced mutations in Salmonella typhimurium can
vary dramatically depending on the DNA repair capacity of the strain
(17, 34). In eukaryotes, mismatch repair involves a complex of several
repair enzymes. One of these, MSH2, binds to palindromes with high
affinity, and to large and small insertions and G/T mismatches with
moderate affinity (35). MSH2 mutants are defective in the repair of
mismatches and small insertion/deletion mispairs (36); thus, this sys
tem is likely responsible for repair of the misalignments leading to
frameshifts. Germ line mutations in hMSH2 (37, 38) and other mis
match repair enzymes have been linked to hereditary nonpolyposis
colon cancer (39-41). Individuals who inherit these mismatch repair
abnormalities demonstrate genome-wide insertions and deletions in
repetitive DNA sequences (replication errors). This "mutator phenotype" may contribute to intragenic deletions in cancer-related genes
(42). However, replication errors and the mismatch repair status in
cases of colon or other tumors containing p53 mutations have not yet
been reported.
These mismatch repair enzymes also are involved in homologous
recombination, as are other proteins, including RecA and its eukaryotic homologues, which bind to homologous DNA regions and pro
mote pairing and strand exchange (43-46). The heteroduplex DNA
sequences formed during this process are generally longer than the
intermediate structures proposed for the mispairs that lead to frameshifts (47, 48), and it is uncertain how these recognition and repair
systems are involved in mutagenesis.
Palindromes and nonadjacent repeat sequences often flank dele
tions and insertions larger than 3 bp in vitro (5, 6, 16) and in this
report. The patterns we observed suggest that the intermediate path
ways differ for events initiated by these two sequence motifs. Palin
dromes generally flank larger deletions, suggesting that short loops
between palindromic dyad elements are less likely to be deleted than
are longer loops. The presence of short loops in the dyads associated
with complex frameshifts (Fig. 3) supports this explanation. The data
also indicate that, although direct repeats of 2-bp motifs can predis
pose to short deletions, more than 2 bp homology are required for both
palindromes and nontandem repeat elements to form stable structures.
Longer dyad sequences are required for palindromes than for tandems.
The deletions associated with tandem sequences usually exactly in
corporate the intervening loop and one copy of the repeat, confirming
the existing model (11, 13, 14, 49). However, the p53 deletions
associated with palindromes usually include portions of both flanking
motifs; the exact deletion generally cannot be predicted. This suggests
that different deletion mechanisms occur, perhaps involving variable
actions of DNA repair proteins (see above). The specific system or
mutagen also may influence repair. For example, the spectrum of
spontaneous and induced mutations in 5. typhimurium usually changes
when a gene for an error-prone DNA repair enzyme is inserted, but no
change is seen in mutations induced by the heterocyclic amine GluP-l (17, 26).
Analysis of complex frameshift mutations confirms that DNA
replication and repair mechanisms may be involved in their genesis.
Prior studies of complex frameshift mutations induced in S. typhi
murium generated a model of adduct-mediated mispairing, quasipalindromes, and DNA replication in palindromes (17). In this system,
the presence of an error-prone DNA repair gene enhances the forma
tion of complex frameshifts. We show that complex mutations in p53
are consistent with DNA excision repair and synthesis in palindromes,
with one strand of the palindrome used as a template (Fig. 3). This
may be triggered by an adduct-induced substitution mutation that
related genes, including APC in colon cancers (50) and pl6'"k4
(CDKN2 and MTS-1) in a variety of tumors (51-54). Interestingly,
1-bp deletions in APC are less often associated with monotonie runs
but are often associated with tandem repeats, perhaps reflecting the
influence of impaired mismatch repair due to abnormalities in the
hMLHl and HMSH2 genes in colon tissue. Of 24 deletions and five
insertions described in pl6'nt4, only 38% (11 of 29) affect 1 bp, and
48% (14 of 29) affect 3 or more bp. DNA sequence features in this
preliminary data base are similar to those seen in p53. Monotonie runs
(86%) and tandem sequences (72%) are the most frequent motifs
present at the sites of deletions.
Deletions and insertions in p53 are likely underreported, because
many investigators concentrate on exons 5-8, the highly conserved
DNA-binding region in which most missense mutations occur. More
complete analysis of all p5J-coding exons would give a more thor
ough picture of mutational patterns. Missense mutations constitute
77% of mutations in exons 5-8 but less than half of the mutations
outside of these regions (1). This is probably due to less frequent
selection of missense mutants in exons 2-4 and 9-11. Single-base
substitutions in the DNA-binding domain often alter p5J-induced
transactivation (1). However, single-amino acid substitutions in the
transactivation domain (amino acids 20-42) do not affect transacti
vation; some double substitutions do (55). Frameshifts occurring
within or 5' to the DNA-binding domain eliminate the p53 DNAbinding function and thus would alter p53 function and cell cycle
kinetics every time they occur. The carboxyl terminus (exons 9-11) of
p53 also is important in the oligomerization and nuclear localization
of the p53 protein (reviewed in Ref. 56) and regulates its sequencespecific DNA binding (57-60).
increases the stability of a palindrome (17).
Germ line and somatic mutation have been reported in other cancer-
ACKNOWLEDGMENTS
We are grateful to Dorothea Dudek for excellent editorial assistance. Peter
Shields for assistance with statistical analysis, and Thomas Kunkel for critical
review of the manuscript.
REFERENCES
1. Greenblatt. M. S., Bennett, W. P., Hollstein. M., and Harris. C. C. Mutations in the
p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis.
Cancer Res., 54: 4855-4878. 1994.
2. Hollstein. M.. Rice, K.. Greenblatt, M. S.. Soussi, T.. Fuchs, R.. Sorlie. T.. Hovig. E.,
Smith-Sorensen, B., Montesano, R., and Harris. C. C. Database of p53 gene somatic
mutations in human tumors and cell lines. Nucleic Acids Res., 22; 3547-3551. 1994.
3. Hollstein. M., Shomer, B.. Greenblatt. M.. Soussi, T., Hovig, E., Montesano, R., and
Harris, C. C. Somatic point mutations in the p53 gene of human tumors and cell lines:
updated compilation. Nucleic Acids Res., 24: 141-146. 1996.
4. Hollstein, M., Sidransky, D., Vogelstein, B., and Harris. C. C. p53 mutations in
human cancers. Science (Washington DC), 253: 49-53. 1991.
5. Kunkel. T. A. Misalignment-mediated
DNA synthesis errors. Biochemistry. 29.
8003-8011. 1990.
6. Ripley, L. S. Frameshift mutation: determinants of specificity. Annu. Rev. Genet., 24:
189-213, 1990.
7. Streisinger, G., Okada. Y., Emrich. 1, Newton. J.. Tsugita. A.. Terzaghi, E., and
Inouye. M. Frameshift mutations and the genetic code. This paper is dedicated to
Professor Theodosius Dobzhansky on the occasion of his 66th birthday. Cold Spring
Harbor Symp. Quant. Biol.. 31: 77-84. 1966.
8. Benebek. K.. and Kunkel. T. A. Frameshift errors initialed by nucleotide misincorporation. Proc. Nati. Acad. Sci. USA. 87: 4946-4950. 1995.
9. Shibulani. S.. and Grollman, A. P. On the mechanism of frameshift (deletion)
mutagenesis in vitro. !. Biol. Chem., 268: 11703-11710, 1993.
10. Jego, N., Thomas, G., and Hamelin. R. Short direct repeats flanking deletions, and
duplicating insertions in p53 gene in human cancers. Oncogene, 8: 209-213. 1993.
11. Krawczak. M.. and Cooper. D. N. Gene deletions causing human genetic disease:
mechanisms of mutagenesis and the role of the local DNA sequence environment.
Hum. Genet., 86: 425-441, 1991.
12. Albertini, A. M.. Hofer, M., Calos, M. P., and Miller, J. H. On the formation of
spontaneous deletions: the importance of short sequence homologies in the generation
of large deletions. Cell, 29: 319-328, 1982.
13. Lutze, L. H., Cleaver, J. E., Morgan, W. F., and Winegar, R. A. Mechanisms involved
in rejoining DNA double-strand breaks induced by ionizing radiation and restriction
enzymes. Mutât.Res., 299: 225-232, 1993.
14. Lutze, L. H., Winegar. R. A., Jostes, R., Cross, F. T., and Cleaver, J. E. Radon-
2135
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.
pS3 DELETIONS
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
AND INSERTIONS
induced deletions in human cells: role of nonhomologous strand rejoining. Cancer
Res., 52: 5126-5129, 1992.
Nalbantoglu. J.. Hartley. D., Phear, G., Tear, G.. and Meuth. M. Spontaneous deletion
formation a( the aprt locus of hamster cells: the presence of short sequence homologies and dyad symmetries at deletion termini. EMBO J., 5: 1199-1204. 1986.
Glickman, B. W.. and Ripley, L. S. Structural intermediates of deletion mutagenesis:
a role for palindromic DNA. Proc. Nati. Acad. Sci. USA, 81: 512-516, 1984.
Levine, J. G.. Schaaper. R. M.. and DeMarini, D. M. Complex frameshift mutations
mediated by plasmid pKM 101 : mutational mechanisms deduced from 4-aminobiphenyl-induced mutation spectra in Salmonella. Genetics, 136: 731-746, 1994.
Lambert. I. B., Napolitano. R. L., and Fuchs. R. P. Carcinogen-induced frameshift
mutagenesis in repetitive sequences. Proc. Nati. Acad. Sci. USA, 89: 1310-1314.
1992.
Belguise-Valladier, P., and Fuchs. R. P. Strong sequence-dependent polymorphism in
adduct-induced DNA structure: analysis of single /V-2-acetylaminofluorene residues
bound within the Narl mutation hot spot. Biochemistry. 30: 10091-10100, 1991.
Parris, C. N., Levy, D. D., Jessee, J.. and Seidman, M. M. Proximal and distal effects
of sequence context on ultraviolet mutational hotspots in a shuttle vector replicated in
xeroderma cells. J. Mol. Biol.. 236: 491-502. 1994.
Redston. M. S.. Caldas, C., Seymour, A. B.. Hruban, R. H., da Costa, L., Yeo. C. J..
and Kem, S. E. p53 mutations in pancreatic carcinoma and evidence of common
involvement of homocopolymer tracts in DNA microdeletions. Cancer Res.. 54:
3025-3033, 1994.
Breimer, L. H. Molecular mechanisms of oxygen radical carcinogenesis and mu
tagenesis: the role of DNA base damage. Mol. Carcinog., 3: 188-197. 1990.
Nath, R. G.. and Chung, F. L. Detection of exocyclic 1JV2-propanodeoxyguanosine
adducts as common DNA lesions in rodents and humans. Proc. Nati. Acad. Sci. USA,
91: 7491-7495, 1994.
Chaudhary, A. K., Nokubo, M.. Reddy, G. R.. Yeola, S. N., Morrow, J. D.. Blair.
I. A., and Marne»,L. J. Detection of endogenous malondialdehyde-deoxyguanosine
adducts in human liver. Science (Washington DC). 265: 1580-1582, 1994.
Lindahl. T. Instability and decay of the primary structure of DNA. Nature (Lond.).
362: 709-715, 1993.
Levine, J. G.. Knasmuller, S., Shelton. M. L., and DeMarini, D. M. Mutation spectra
of Glu-P-1 in Salmonella: induction of hotspot frameshifts and site-specific base
substitutions. Environ. Mol. Mutagen., 24: 11-22, 1994.
Gupta, P. K.. Johnson, D. L., Reid, T. M., Lee, M. S., Romano. L. J., and King, C. M.
Mutagenesis by single site-specific arylamine-DNA adducts. Induction of mutations
at multiple sites. J. Biol. Chem.. 264: 20120-20130, 1989.
Moriya, M., Takeshita. M., Johnson, F.. Peden, K.. Will. S., and Grollman. A. P.
Targeted mutations induced by a single acetylaminofluorene DNA adduci in mam
malian cells and bacteria. Proc. Nati. Acad. Sci. USA, 85: 1586-1589. 1988.
Wang, C. I., and Taylor, J. S. In vitro evidence that UV-induced frameshift and
substitution mutations at T tracts are the result of misalignment-mediated replication
past a specific thymine dimer. Biochemistry, 31: 3671-3681, 1992.
Burnouf. D., Koehl, P.. and Fuchs. R. P. Single adduct mutagenesis: strong effect of
the position of a single acetylaminofluorene adduct within a mutation hot spot. Proc.
Nati. Acad. Sci. USA, 86: 4147-4151, 1989.
Grollman. A. P.. and Shibutani. S. DNA Adducts: identification and Biological
Significance. Mutagenic specificity of chemical carcinogens as determined by studies
of single DNA adducts. IARC Sci. Pubi.. /25: 385-397, 1994.
Mendelman. L. V.. Petruska. J.. and Goodman. M. F. Base mispair extension kinetics.
Comparison of DNA polymerase alpha and reverse transcriptase. J. Biol. Chem., 265:
2338-2346, 1990.
Singer, B., and Essigmann, J. M. Site-specific mutagenesis: retrospective and pro
spective. Carcinogenesis (Lond.), 12: 949-955. 1991.
Bronstein, S. M.. Cochrane, J. E., Craft, T. R.. Swenberg, J. A., and Skopek. T. R.
Toxicity, mutagenicity, and mutational spectra of /V-ethyl-W-nitrosourea in human cell
lines with different DNA repair phenotypes. Cancer Res., 51: 5188-5197, 1991.
Alani. E., Chi, N. W., and Kolodner, R. The Sacchartnn\ces cerevisiae Msh2 protein
specifically binds to duplex oligonucleotides containing mismatched DNA base pairs
and insertions. Genes & Dev.. 9: 234-247, 1995.
Alani. E., Reenan. R. A., and Kolodner. R. D. Interaction between mismatch repair
and genetic recombination in Saccharomyces cerevisiae. Genetics, 137: 19-39, 1994.
Fishel, R., Lescoe, M. K., Rao, M. R.. Copeland, N. G., Jenkins, N. A.. Garber, J.,
Kane. M.. and Kolodner, R. The human mutator gene homolog MSH2 and its
association with hereditary nonpolyposis colon cancer. Cell, 75: 1027-1038, 1993.
Leach, F. S., Nicolaides, N. C., Papadopoulos, N., Liu. B., Jen, J., Parsons, R.,
Peltomaki. P.. Sistonen. P.. Aaltonen, L. A.. Nystrom-Lahti. M., Guan. X. Y.. Zhang.
J.. Meltzer. P. S., Yu, J. W., Kao. F. T.. Chen, D. J. Cerosaletti, K. M.. Foumier,
R. E., Todd. S., Lewis, T., Leach. R. J., Naylor. S. L., Weissenbach, J., Mecklin. J. P.,
Jarvinen, H.. Petersen, G. M.. Hamilton, S. R., Green. J.. Jass, J.. Watson. P., Lynch,
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
H. T., Trent. J. M., de la Chapelle. A., Kinzler. K. W.. and Vogelstein, B. Mutations
of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell, 75: 12151235, 1993.
Peltomaki, P.. Aaltonen, L. A., Sistonen. P.. Pylkkanen, L., Mecklin, J. P.. Jarvinen,
H.. Green, J. S., Jass, J. R., Weber, J. L., Leach, F. S., Peterson, G. M.. Hamilton.
S. R.. de la Chapelle, A., and Vogelstein, B. Genetic mapping of a locus predisposing
to human colorectal cancer. Science (Washington DC), 260: 810-812, 1993.
Aaltonen, L. A.. Peltomaki, P., Leach, F. S.. Sistonen. P.. Pylkkanen, L., Mecklin,
J. P., Jarvinen. H., Powell. S. M.. Jen, J.. Hamilton, S. R.. Peterson, G. M., Kinzler,
K. W., Vogelstein, B., and de la Chapelle, A. Clues to the pathogenesis of familial
colorectal cancer. Science (Washington DC). 260: 812-815, 1993.
Thibodeau, S. N.. Bren. G., and Schaid. D. Microsatellite instability in cancer of the
proximal colon. Science (Washington DC), 260: 816-819, 1993.
Markowitz. S., Wang. J., Myeroff. L.. Parsons. R.. Sun, L.. Lutterbaugh, J., Fan, R. S.,
Zborowska, E., Kinzler, K. W., Vogelstein, B., Brattain, M.. and Willson, J. K.
Inactivation of the type II TGF-ßreceptor in colon cancer cells with microsatellite
instability. Science (Washington DC). 26S: 1336-1338, 1995.
Radman, M.. and Wagner, R. Mismatch recognition in chromosomal interactions and
speciation. Chromosoma (Beri.), 102: 369-373, 1993.
Hollingsworth, N. M.. Ponte, L.. and Halsey, C. MSH5. a novel MutS homolog,
facilitates meiotic reciprocal recombination between homologs in Saccharom\ces
cerevisiae but not mismatch repair. Genes & Dev., 9: 1728-1739. 1995.
Baker, S. M., Bronner, C. E., Zhang, L., Plug, A. W., Robatzek, M., Warren, G.,
Elliott, E. A., Yu, J., Ashley, T., and Amheim. N. Male mice defective in the DNA
mismatch repair gene PMS2 exhibit abnormal chromosome synapsis in meiosis. Cell,
82: 309-319, 1995.
de Wind, N.. Dekker, M.. Berns, A., Radman, M.. and te Riele. H. Inactivation of the
mouse Afs/i2 gene results in mismatch repair deficiency, methylation tolerance,
hypcrrecombination, and predisposition to cancer. Cell. 82: 321-330, 1995.
Conley, E. C.. and West, S. C. Underwinding of DNA associated with duplex-duplex
pairing by RecA protein. J. Biol. Chem., 265: 10156-10163, 1990.
Hsieh, P., Camerini-Otero. C. S.. and Camerini-Otero, R. D. The synapsis event in the
homologous pairing of DNAs: RecA recognizes and pairs less than one helical repeat
of DNA. Proc. Nati. Acad. Sci. USA, 89: 6492-6496. 1992.
Studier, F. W., Rosenberg, A. H., Simon. M. N., and Dunn. J. J. Genetic and physical
mapping in the early region of bacteriophage T7 DNA. J. Mol. Biol.. 135: 917-937,
1979.
Nagase. H., and Nakamura, Y. Mutations of the APC gene. Hum. Mutât..2: 425-434,
1993.
Yeudall, W. A., Crawford, R. Y., Ensley, J. F., and Robbins. K. C. MTSI/CDK4I is
altered in cell lines derived from primary and metastatic oral squamous cell carci
noma. Carcinogenesis (Lond.). /5: 2683-2686. 1994.
Ohta. M., Nagai, H., Shimizu. M.. Rasio, D., Berd. D., Mastrangelo, M.. Singh, A. D.,
Shields, J. A., Shields, C. L., Croce. C. M., and Huebner, K. Rarity of somatic and
germline mutations of the cyclin-dependent kinase 4 inhibitor gene CDK41 in mela
noma. Cancer Res., 54: 5269-5272, 1994.
Spruck, C. H., Gonzalez-Zulueta, M., Shibata, A., Simoneau, A. R., Lin, M. F..
Gonzales, F., Tsai, Y. C., and Jones, P. A. pió gene in uncultured tumours. Nature
(Lond.), 370: 183-184. 1994.
Okamoto, A., Hussain, S. P., Hagiwara, K., Spillare. E. A.. Rusin. M. R., Demetrick,
D. J., Serrano, M.. Hannon. G. J.. Shiseki, M., Zariwala, M., Xiong, Y., Beach, D. H..
Yokota, J.. and Harris. C. C. Mutations in the pl6INK4/MTSl/CDKN2.
pl5INK4B/
MTS2 and pl8 genes in primary and metastatic lung cancer. Cancer Res.. 55:
1448-1451. 1995.
Lin. J., Chen, J., Elenbaas, B.. and Levine, A. J. Several hydrophobic amino acids in
the p53 amino-terminal domain are required for tran.scriptional activation, binding to
mdm-2 and the adenovirus 5 E1B 55-kD protein. Genes & Dev., 8: 1235-1246. 1994.
Hupp. T. R., and Lane. D. P. Allosteric activation of latent p53 tetramers. Curr. Biol.,
4: 865-875, 1995.
Bayle, J. H., Elenbaas, B.. and Levine, A. J. The carboxyl-terminal domain of the p53
protein regulates sequence-specific DNA binding through its nonspecific nucleic
acid-binding activity. Proc. Nati. Acad. Sci. USA, 92: 5729-5733. 1995.
Foord, O. S.. Bhattacharya. P.. Reich. Z.. and Rotter, V. A DNA binding domain is
contained in the C-terminus of wild type p53 protein. Nucleic Acids Res., 19:
5191-5198, 1991.
Wolkowicz, R., Peled, A., Elkind, N. B., and Rotter, V. Augmented DNA-binding
activity of p53 protein encoded by a carboxyl-terminal alternatively spliced mRNA is
blocked by p53 protein encoded by the regularly spliced form. Proc. Nati. Acad. Sci.
USA. 92: 6842-6846. 1995.
Jayaraman, J.. and Prives, C. Activation of p53 sequence-specific DNA binding by
short single strands of DNA requires the p53 C-terminus. Cell, 81: 1021-1029. 1995.
2136
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.
Deletions and Insertions in the p53 Tumor Suppressor Gene in
Human Cancers: Confirmation of the DNA Polymerase
Slippage/Misalignment Model
Marc S. Greenblatt, Arthur P. Grollman and Curtis C. Harris
Cancer Res 1996;56:2130-2136.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/56/9/2130
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 15, 2017. © 1996 American Association for Cancer Research.