Download Increased transversions in a novel mutator colon cancer cell line

Oncogene (1998) 16, 1125 ± 1130
 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00
Increased transversions in a novel mutator colon cancer cell line
James R Eshleman1,6, P Scott Donover2, Susan J Littman5, Sandra E Swinler2, Guo-Min Li4,7,
James D Lutterbaugh2, James KV Willson2, Paul Modrich4, W David Sedwick2,
Sanford D Markowitz2 and Martina L Veigl3
1
Departments of Pathology; 2Medicine; 3General Medical Sciences and Ireland Cancer Center, University Hospitals of Cleveland
and Case Western Reserve University, Suite 200, UCRC II, 11001 Cedar Road, Cleveland, Ohio 44106; 4Department of
Biochemistry and Howard Hughes Medical Institute; 5Department of Medicine-Oncology, Duke University Medical Center,
Nanaline Duke Building, Durham, North Carolina, 27710 USA
We describe a novel mutator phenotype in the Vaco411
colon cancer cell line which increases the spontaneous
mutation rate 10 ± 100-fold over background. This
mutator results primarily in transversion base substitutions which are found infrequently in repair competent
cells. Of the four possible types of transversions, only
three were principally recovered. Spontaneous mutations
recovered also included transitions and large deletions,
but very few frameshifts were recovered. When compared
to known mismatch repair defective colon cancer
mutators, the distribution of mutations in Vaco411 is
signi®cantly di€erent. Consistent with this di€erence,
Vaco411 extracts are pro®cient in assays of mismatch
repair. The Vaco411 mutator appears to be novel, and is
not an obvious human homologue of any of the
previously characterized bacterial or yeast transversion
phenotypes. Several hypotheses by which this mutator
may produce transversions are presented.
Keywords: colon cancer; HPRT; mutator phenotype;
transversions
Introduction
Mutator phenotypes have been proposed as a
prerequisite for multi-step carcinogenesis (Nowell,
1976; Loeb, 1991). By increasing the mutation rate, a
mutator phenotype permits the multiple independent
mutations in oncogenes and tumor suppressor genes
required for multi-step carcinogenesis. Direct evidence
that mutators play an important role in carcinogenesis
is provided by the recent discovery of a complex
mutator phenotype in colorectal cancer (CRC)
resulting from defects in mismatch repair (MMR), see
reviews (Eshleman and Markowitz, 1996; Modrich and
Lahue, 1996; Kolodner, 1995; Umar and Kunkel,
1996). This mutator is characterized by a 100 ± 1000fold increased spontaneous mutation rate in the hprt
gene target (Eshleman et al., 1995; Bhattacharyya et
al., 1994; Branch et al., 1995).
Correspondence: ML Veigl or WD Sedwick, Ireland Cancer Center
Research Laboratories, Suite 200, UCRC-II, 11001 Cedar Avenue,
Cleveland, Ohio 44106, USA
Current addresses: 6 Department of Pathology, Ross 632; Johns
Hopkins Medical Institutions, 720 Rutland Avenue, Baltimore,
Maryland 21205; 7 Department of Pathology, University of
Kentucky Medical Center, Lexington, Kentucky, 40536, USA
Received 18 June 1997; revised 7 October 1997; accepted 7 October
1997
In this report we describe the unique speci®city of a
new mutator defect, which is independent of MMR.
We have previously shown that the Vaco411 colon
cancer cell line displays a moderately (10 ± 100-fold)
increased spontaneous mutation rate (Eshleman et al.,
1995), and does not exhibit the RER phenotype which
generally accompanies MMR defects (Eshleman and
Markowitz, 1995). We now describe the spontaneous
hprt mutations arising in this new colon cancer
mutator, and compare these mutations to the repair
capacity exhibited by cell-free extracts of this cell line.
The spontaneous mutant collection lacks frameshifts,
which are commonly seen in MMR defective cell lines
(Eshleman et al., 1996; Bhattacharyya et al., 1995; Kat
et al., 1993). Instead, transversion-type base substitutions predominate, characterizing this cell line as a
novel transversion mutator.
Results
Transversions are the most frequent spontaneous hprt
mutation recovered from Vaco411
Individually isolated, spontaneous, 6-thioguanine
(6TG) resistant mutants were initially characterized
by hprt RT ± PCR and cDNA sequencing (Table 1).
Base substitutions are the predominant type of
mutation recovered (70% of total mutations). Transversions (50% of total mutations and 71% of base
substitutions) occur more than twice as frequently as
transitions (20% of total mutants and 29% of base
substitutions). Of note, small (1 ± 4 bp) deletions, which
are prevalent in MMR defective human cell lines, are
almost absent (only 5%) among the Vaco411 mutations. Finally, 20% of the Vaco411 spontaneous hprt
mutants harbor large deletions that remove one or
more coding exons.
Sequence context of transversions
As discussed above, transversions predominate among
the spontaneous mutations seen in Vaco411. A
sequence level analysis of the transversions is depicted
in Table 2. All four types of transversions were
recovered, but the A : T?T : A transversion (only
mutant #1) is underrepresented (P=0.024, binomial
distribution). The three other types of transversions are
about equally represented, although the C : G?G : C
transversion occurs most frequently (40% of all
transversions).
A colon cancer with a novel transversion mutator
JR Eshleman et al
1126
Inspection of the sequences surrounding the mutated
base suggests that runs of adenine and thymidine
nucleotides (Table 2, characters in italics) are found
adjacent to most transversions. When de®ned as three
or more consecutive A or T bases, this motif is present
in 15 out of 20 transversions. The A/T run includes the
mutated base in ®ve of the six transversions involving
Table 1
Overall distribution of Vaco411 mutations
Mutation type
Number of
mutants
Percent of total
mutants (%)
28
20
8
2
2
8
70
50 (71)a
20 (29)a
5
5
20
Total base substitutions
Transversions
Transitions
Small deletions (1 ± 4 bp)
Complex small mutants
Large deletions
a
Values in parentheses are the percent of transversions or transitions
out of all base substitutions
an A:T base pair (e.g. mutant #1), and is adjacent to the
mutated base in ten of the 14 mutants involving a G : C
base pair (e.g. mutant #8). To test for signi®cance, all
transversion sites previously reported as selectable in
6TG (Cariello, 1994) were analysed for this A/T motif.
This analysis demonstrates that 50% (92 out of 184) of
the phenotypically selectable transversion sites are
¯anked by three or more consecutive A or T bases. In
contrast, 75% (15 out of 20) of the Vaco411
transversions displayed this motif. The di€erence
between sites in the database at which transversions
have been selected and those found in Vaco411 is
statistically signi®cant (P=0.021, binomial distribution). A search for other motifs surrounding transversion sites in Vaco411 using the MacVector DNA
software program (Oxford molecular group, version
5.0) was uninformative. The largest numbers of
transversions were found at position 403 where two
C : G?G : C transversions and one C : G?A : T trans-
Table 2 Distribution of transversions in the hprt gene of Vaco411
Mutant #
Mutation
Location of mutation
Sequence contexta
1
A:T?T:A
2
3
4
5
6
Siteb
E€ect
CGTTATGGCGA
2
Met?Lys
A:T?C:G
A:T?C:G
A:T?C:G
A:T?C:G
A:T?C:G
TCCTCATGGAC
TTGATTGTGGA
GTGAAAAGGAC
CAAAATACAAA
E7-GACT|GTAAGTG-I7
116
396
498
646
(39,864)
His?Pro
Ile?Met
Lys?Asn
Tyr?Asp
D Exon 7
7
8
9
10
11
12
13
14
C:G?G:C
C:G?G:C
C:G?G:C
C:G?G:C
C:G?G:C
C:G?G:C
C:G?G:C
C:G?G:C
AACCAGGTTAT
CATACCTAATC
TCATGGACTAA
TGCTCGAGATG
I5-GAAAG|GATATAATTG-E6
I5-GAAAG|GATATAATTG-E6
GAAAAGGACCC
GAAAAGCAAAA
46
74
119
152
403
403
500
640
Gly?Arg
Pro?Arg
Gly?Ala
Arg?Pro
Asp?His
Asp?His
Arg?Thr
Ala?Pro
15
16
17
18
19
20
C:G?A:T
C:G?A:T
C:G?A:T
C:G?A:T
C:G?A:T
C:G?A:T
CTGTAGATTTTATC
GTCTTGATTGT
I5-GAAAG|GATATAATTG-E6
TCAGGGATTTG
I2-CTGTAG|GACTG-E3
E7-T|GTAAGTGAAT-I7
292
393
403
601
(16,602)
(39,867)
Asp?Tyr
Leu?Phe
Asp?Tyr
Asp?Tyr
D E3, D E2-3
D Exon 7
a
Sequence context is shown on the sense strand only. The mutated base is depicted in bold and is underlined. Ten surrounding
bases are generally depicted unless the adjacent sequence involves an A/T run (italic characters). Introns (I), exons (E), and
junctions (|) are designated where necessary. b Site of mutation is generally identi®ed as the cDNA base position. Splice site
mutations occurring in introns are designated as genomic DNA base positions depicted in parentheses. Mutants 11, 12, and 17
occurring at position 403 are presented in bold
Table 3
Mutant
type
Transitions, 1 ± 4 bp deletions and complex mutations in Vaco411
Mutant #
Mutation
Transitions
21
22
23
24
25
26
27
28
A:T?G:C
A:T?G:C
A:T?G:C
C:G?T:A
C:G?T:A
C:G?T:A
C:G?T:A
C:G?T:A
1 ± 4 bp
Deletions
29
30
Complex
31
32
Location of mutation
Sequence context
Site
E€ect
E2-TATGGACAG|GT-I2
TGACCTGCTGG
E7-GCCAGACT|GTA-I7
E2-GGACAG|GTAAG-I2
TTGGTGGAGAT
GTATAATCCAAAG
ATAAGCCAGAC
I8-TATAG|CATGTT-E9
131
233
530
134
355
463
526
610
Asp?Gly
Leu?Pro
Asp?Gly
Arg?Lys
Gly?Arg
Pro?Ser
Pro?Ser
His?Tyr
D1 bp
D3 bp
ATATGCCCTTGACc
I3-TAACTAG|AATGACC-E4c
575-577
(27890-93)
Frameshift
DExon 4
D1 and D4 bp
T?A, G?A
CCGCAGCCCTGGCGTCGT
TTCCTTGGTCAGG
14, 18-21
446, 448
Frameshift
Leu?Stop
See legend to Table 2. c Denotes the DNA sequence where the small deletions reside in mutants #29 and #30. For these two mutants, an exact
identi®cation of the deleted base(s), within the designated area, cannot be made
A colon cancer with a novel transversion mutator
JR Eshleman et al
version were recovered. Interestingly, position 403
possesses the longest adjacent A/T run (eight bases) in
the series. Comparatively, as shown in Table 3, such A/
T motifs are generally not found near transition
mutation sites (only one out of eight transitions).
Vaco411 is MMR pro®cient
To further examine the bias in base substitutions,
extracts prepared from the Vaco411 cell line were
challenged with a series of mispaired substrates
(Drummond et al., 1995). Extracts from Vaco411 are
fully capable of repairing base ± base mispaired
substrates (Table 4 ± compare Vaco411 with the
Table 4 Base ± base mispair and loop repair by Vaco411 extracts
Mismatch
A⋅A
A⋅C
A⋅G
C⋅C
C⋅T
G⋅G
G⋅T
T⋅T
1 bp loop (5'/T\)
2 bp loop (5'/CA\)
2 bp loop (3'/CA\)
Vaco411
SO
6.5
4.2
6.5
5.6
11.3
6.1
8.7
3.1
4.9
11.7
12.8
8.5
7.1
6.6
5.0
13.9
6.6
9.7
5.2
3.7
9.8
12.1
Repair activity for Vaco411 and SO in fmoles/15 min at 378C. SO is
a control colon cancer cell line with stable microsatellites and
competent MMR
MMR competent cell line, SO). Likewise, as
expected, when Vaco411 extracts are challenged with
insertion-deletion heteroduplex substrates consisting of
either a 5' T loop, a 5' CA loop, or a 3' CA loop, they
repair these insertion ± deletion heteroduplex substrates
as well as the control SO cell line.
Vaco411 mutations lack frameshifts
Consistent with pro®cient loop repair found in the
extract assay, only two 1 ± 4 bp deletions are found
among the spontaneous hprt mutations recovered from
Vaco411 (Table 3, mutants #29, 30). The ®rst is a
single-base deletion in a short 3-base pair mononucleotide repeat. The second is a 3-base pair deletion
of either AGA or GAA in an area of non-repetitive
DNA. Notably, not a single frameshift mutation was
found in the 6-base poly-G mononucleotide repeat
located at positions 207 to 212. The paucity of this
class of mutants in Vaco411 is in stark contrast to the
spontaneous mutations reported in the MMR de®cient
cell lines, RKO and HCT116, where frameshifts
comprise about 50% of the recovered mutants
(Figure 1), and where the poly-G run is a signi®cant
hotspot site. When the Vaco411 spectrum of
mutations in Figure 1 is compared to that of RKO,
or HCT116, it is signi®cantly di€erent (P50.005, chisquare). It is also signi®cantly di€erent (P50.05) from
the DLD-1 cell line, which is de®cient in the MMR
gene, GTBP.
Figure 1 Distribution of Vaco411 mutations relative to other colon cancer mutators. (A) Dash symbol signi®es no mutations of
this type. aThe underlying defect, if known, is shown in parentheses under the cell name. bRKO is an RER colon cancer cell line
with an unknown defect in MMR (Eshleman et al., 1996). cCombined data from Bhattacharyya et al., 1995 and Eshleman et al
(unpublished). dFrom Bhattacharyya et al., 1995. (B) Bar graph showing the relative distribution of mutation types for the given cell
lines
1127
A colon cancer with a novel transversion mutator
JR Eshleman et al
1128
Other mutations in the Vaco411 mutation spectrum
In addition to the mutations described above, two
complex mutants were found (Table 3, mutants #31
and 32). The ®rst complex mutation involves a double
deletion and the second a double-base substitution (one
transversion and one transition). These double mutations are not an artifact resulting from a mixture of two
di€erent mutants because DNA sequencing shows no
wild type sequence mixed with the mutant sequence. The
®nal eight mutations listed in Table 1 are large deletions
which are resolved through a combination of cDNA
sequencing and PstI Southern blotting. The large
deletions range from 190 bp to several kilobases in size.
Discussion
Vaco411 de®nes a novel human mutator phenotype in
which the hprt mutation rate is elevated 10 ± 100-fold and
in which transversions predominate. Several lines of
evidence indicate that the Vaco411 mutator defect is
distinct from the MMR de®ciency previously described
in familial and sporadic CRC. The mutation rate in
Vaco411 is only 10 ± 100-fold elevated above four nonmutator colon cancers (Eshleman et al., 1995) in contrast
to the hprt mutation rates generally seen in documented
MMR de®cient cultures where the hprt mutation rate is
100 ± 1000-fold elevated (Eshleman et al., 1995; Bhattacharyya et al., 1994; Glaab and Tindall 1997). The
Vaco411 cell line also possesses stable microsatellites
[(Eshleman et al., 1995) designated Vaco410 therein].
Further, nuclear extracts of Vaco411 cells are competent
when challenged with mispaired and looped substrates.
Finally, the spectrum of spontaneous hprt mutations
described herein shows a lack of frameshifts which
stands in contrast to mutations found in human MMR
de®cient cells (Figure 1, Eshleman et al., 1996;
Bhattacharyya et al., 1995; Kat et al., 1993). Instead
the mutator phenotype residing in Vaco411 primarily
results in the recovery of transversions.
Transversions may spontaneously occur through
DNA replication errors or as a consequence of base
modi®cations. For example, previous experiments in the
mutD5 strain of E. coli have suggested that MMR can be
fully competent, but that a DNA polymerase defect in
exonucleolytic proofreading can produce so many
Table 5
Defect
Human
Vaco411
Bacterial
Mut A
Mut C
Mut D5
Mut M
Mut T
Mut Y
Yeast
Rad 6
Rad 18
Rad 52
APN1
a
mispairs that normal MMR capacity is overwhelmed
(Schaaper and Radman, 1989). In this mutD5 strain an
increase in transversions was observed, however the
frequency of recovery of these mutations was far below
the frequency of transitions seen in the same organism.
Endogenous errors in nucleotide metabolism can also
compromise DNA polymerase ®delity leading to an
increase in transversions and/or transitions which are
re¯ective of speci®c DNA pool imbalances (Kohalmi and
Kunz, 1993; Meuth, 1989; Phear and Meuth, 1989). On
the other hand, it is clear that MMR in bacteria requires
at least ten individual components (Modrich, 1991) and
that the human system is even more complex (Modrich
and Lahue, 1996; Kolodner, 1996). Therefore, it is
possible that Vaco411 harbors an atypical MMR
de®ciency, due to a defect in a relatively minor
component of MMR that is not revealed in extract
assays. Alternatively, observed mutations may arise in
the presence of a functional MMR system because the
mispaired intermediates are shielded from repair,
possibly due to an unde®ned component which interacts
with neighboring A/T runs. Theoretically, this could
occur because the less tightly hydrogen bonded A : T
base pair runs might permit local melting of the DNA
duplex adjacent to the mispair, thereby preventing
recognition by the MMR system. In this regard,
bacterial MMR does not correct all base ± base mispairs
with equal eciency. For example, base ± base mispairs
which could result in transversions are not repaired as
well as other mispairs. Furthermore, DNA repair can be
modulated by the surrounding DNA sequence (Fazakerley et al., 1986), suggesting that a similar mechanism
could lead to an increase in mutant recovery in certain
areas such as A/T runs or the A/T cluster at position 403.
In this regard, the sites where transitions occur in the
present study are not surrounded by A/T motifs,
strengthening the suggestion that these motifs may be
involved in a reaction important to correction of
mispairs leading to transversions. Finally, it is possible
that the mispaired intermediates in these local areas of
the gene do not occur through replication errors and are
not normally repaired by components of MMR, but
rather may be repaired by another system which is
defective in this cell line.
Because the Vaco411 mutator phenotype appears
distinct from that associated with MMR defects, an
analogous mutator among those previously described
Transversion phenotypes
Mutation
ratea
Dominant transversion
Reference
10 ± 1006
C:G?G:C, C:G?A:T, A:T?C:G
This Study
206
106
5006b
0 ± 106
20 ± 1206
6 ± 206
C:G?A:T, A:T?T:A
C:G?A:T, A:T?T:A
A:T?T:Ab
C:G?A:T
A:T?C:G
C:G?A:T
(Michaels et al., 1990)
(Michaels et al., 1990)
(Schaaper, 1988)
(Michaels et al., 1992)
(Yanofsky et al., 1966)
(Nghiem et al., 1988)
56
36
36
46
C:G?A:T
C:G?A:T
C:G?G:C
A:T?C:G
Fold increase above background. b Applies only to minimal media conditions
(Kang
(Kunz
(Kunz
(Kunz
et
et
et
et
al.,
al.,
al.,
al.,
1992)
1991)
1989)
1994)
A colon cancer with a novel transversion mutator
JR Eshleman et al
in either bacteria or yeast was sought (Table 5). Six
bacterial transversion phenotypes are shown, but when
the resulting mutation rates and the predominant type
of transversion formed are considered, Vaco411 does
not functionally appear to harbor a defect which is an
obvious human homolog of a known bacterial
transversion mutator. Similarly, Vaco411 appears not
to be the human homolog of one of the four
characterized yeast transversion mutators.
Carcinogenesis requires mutation and is frequently
accompanied by mutator phenotypes. Identi®cation of
the hotspots spontaneously generated in these mutator
phenotypes may allow one to predict the critical
changes seen in oncogenes and tumor suppressor
genes. In addition to elucidating carcinogenic pathways, comprehensive knowledge of mutator phenotypes may aid in the molecular diagnosis of cancer, and
provide strategies for cloning the genes responsible for
the underlying defects. Better understanding of
mutator phenotypes may also identify environmental
exposures to be avoided in certain predisposed
individuals, and may predict speci®c therapeutic
responses in certain cancers. The described mutator
in which transversions predominate is unique, in that
transversion base substitutions are recovered less
frequently than transitions in repair competent cells
(Giver et al., 1993; McGregor et al., 1991; Nigro et al.,
1989; Powell et al., 1992). Insight into the prevalence of
this new mutator may be gained by ®nding additional
microsatellite stable (non-RER) tumors in which
multiple transversions are found in independent gene
targets. Since non-RER tumors are more common than
RER tumors, this new mutator has the potential to be
fairly prevalent. In this regard, Vaco411 may be
representative of a class of mutators in non-RER
CRCs, which we have shown displays moderately
elevated mutation rates. Therefore, the Vaco411
transversion mutator may have additional counterparts among other colon cancers. The discovery of yet
another CRC mutator adds further evidence in support
of the multi-step mutator hypothesis (Nowell, 1976;
Loeb, 1991).
Materials and methods
Cell culture and selection
The Vaco411 cell line was originally established from a
female patient with a villous colonic polyp with carcinoma
in situ (Markowitz et al., 1994; Wang et al., 1995).
Vaco411 cell culture and selection of independent 6TG
resistant mutants were performed essentially as described
(McBain et al., 1984; Eshleman et al., 1995). Brie¯y, based
on the elevated spontaneous mutation frequency (Eshleman et al., 1995, designated Vaco410 therein), Vaco411
cells are purged of pre-existing mutants by regrowth from
100 cells in independent ¯asks. After reaching about 10 7
cells per ¯ask, they are subcultured and selected for 6TG
resistant mutants in 1.5 mg/ml 6TG at 10 000 cells per well
in 96 well plates. A single mutant was isolated from each
¯ask.
Molecular characterization of mutants
Mutants described in this study are 6TG resistant clones
which have an alteration in their hprt cDNA, ¯anking
introns in genomic DNA, or alteration on Southern blot.
Following expansion of resistant cells, RNA and genomic
DNA are isolated using the guanidine/cesium ultracentrifugation method. Nucleic acid isolation, RT ± PCR, cDNA
sequencing, splice-site genomic DNA sequencing, and PstI
Southern blotting are performed as described (Eshleman et
al., 1996).
Mismatch repair assays
Mismatch repair in cell free extracts was evaluated as
described previously (Parsons, et al., 1993; Li and
Modrich, 1995). Brie¯y, mismatch repair assays were
performed with 100 mg of nuclear extract and 24 fmol of
heteroduplex DNA which contained a single strand break
125 nucleotides 5' or 181 nucleotides 3' to the mispair
(Fang and Modrich, 1993; Kat et al., 1993). The
heteroduplexes (Su, et al., 1988; Parsons, et al., 1993;
Drummond, et al., 1995) and the nuclear extracts used in
this assay were prepared as described (Holmes, et al.,
1990). All reactions were performed in 10 ml at 378C for
15 min. The results are an average of 2 ± 3 determinations.
Statistical methods
To analyse the distribution of transversion types, the null
hypothesis is established where each transversion type has
an equal probability of occurring. The probability of
obtaining either zero or one A : T?T : A transversion
versus the other three transversion types grouped together
is then calculated according to a binomial distribution
(P=0.024).
To analyse the probability of ®nding A/T runs adjacent to
sites of transversion, the transversion selectable sites were
identi®ed from the Cariello hprt database (Cariello, 1994),
supplemented with the transversion sites reported in RKO
(Eshleman et al., 1996), and Vaco411 (this report). Runs were
tabulated as containing three or more consecutive A/T bases.
The sites of transversions in Vaco411 as reported in Table 2
and the phenotypically selectable transversion sites in the hprt
database were analysed according to this criterion. A
binomial distribution calculation was performed to determine whether the di€erences in distribution are statistically
signi®cant. A similar distribution analysis was performed for
the transition sites in hprt.
A chi square analysis is performed to compare the
distribution of mutation types. A summary of the
spontaneous mutations recovered from Vaco411, RKO,
HCT116 and DLD-1 are presented in Figure 1a. The
distribution of mutations in each cell line is compared
individually to the distribution seen in Vaco411.
Acknowledgements
The authors acknowledge Drs Stephan Lapic and Wilma
Mackay for statistical assistance. We also acknowledge Drs
Russell Anderson and Karen Tsuchiya for helpful discussions, and the valuable assistance of Dr Robert Ho€man.
These studies were supported with grants from the
American Cancer Society (642-94733, JRE; DHP-104,
MLV), National Cancer Institute (CA66628-01A1, JRE;
CA67409, SDM), and National Institute of Environmental
Health Sciences (ES05540, WDS).
1129
A colon cancer with a novel transversion mutator
JR Eshleman et al
1130
References
Bhattacharyya NP, Ganesh A, Phear G, Richards B,
Skandalis A and Meuth M. (1995). Hum. Mol. Genet., 4,
2057 ± 2064.
Bhattacharyya NP, Skandalis A, Ganesh A, Groden J and
Meuth M. (1994). Proc. Natl. Acad. Sci. USA, 91, 6319 ±
6323.
Branch P, Hampson R and Karran P. (1995). Cancer Res.,
55, 2304 ± 2309.
Cariello NF. (1994). Mutat. Res., 312, 173 ± 185.
Drummond JT, Li GM, Longley MJ and Modrich P. (1995).
Science, 268, 1909 ± 1912.
Eshleman JR, Lang EZ, Bower®nd GK, Parsons R,
Vogelstein B, Willson JKV, Veigl ML, Sedwick WD and
Markowitz SD. (1995). Oncogene, 10, 33 ± 37.
Eshleman JR and Markowitz SD. (1995). Curr. Opin. Onc.,
7, 83 ± 89.
Eshleman JR and Markowitz SD. (1996). Hum. Mol. Genet.,
5, 1489 ± 1494.
Eshleman JR, Markowitz SD, Donover PS, Lang EZ,
Lutterbaugh JD, Li GM, Longley M, Modrich P, Veigl
ML and Sedwick WD. (1996). Oncogene, 12, 1425 ± 1432.
Fang W-h and Modrich P. (1993). J. Biol. Chem., 268,
11838 ± 11844.
Fazakerley GV, Quignard E, Woisard A, Guschlbauer W,
van der Marel GA, van Boom JH, Jones M and Radman
M. (1986). EMBO J., 5, 3697 ± 3703.
Giver CR, Nelson SL and Grosovsky AJ. (1993). Environ.
Mol. Mutagenesis, 22, 138 ± 146.
Glaab WE and Tindall KR. (1997). Carcin., 18, 1 ± 8.
Holmes J, Clark S and Modrich P. (1990). Proc. Natl. Acad.
Sci. USA, 87, 5837 ± 5841.
Kang X, Yadao F, Gietz RD and Kunz BA. (1992). Genetics,
130, 285 ± 294.
Kat A, Thilly WG, Fang W-h, Longley MJ, Li G-M and
Modrich P. (1993). Proc. Natl. Acad. Sci. USA, 90, 6424 ±
6428.
Kohalmi SE and Kunz BA. (1993). Mutat. Res., 289, 73 ± 81.
Kolodner R. (1996). Genes Dev., 10, 1433 ± 1442.
Kolodner R. (1995). TIBS, 20, 397 ± 401.
Kunz B, Henson E, Roche H, Ramotar D, Numoshiba T and
Demple B. (1994). Proc. Natl. Acad. Sci. USA, 91, 8165 ±
8169.
Kunz BA, Kang X and Kohalmi L. (1991). Mol. Cell. Biol.,
11, 218 ± 225.
Kunz BA, Peters MG, Kohalmi SE, Armstrong JD, Glattke
M and Badiani K. (1989). Genetics, 122, 535 ± 542.
Li G-M and Modrich P. (1995). Proc. Natl. Acad. Sci. USA,
92, 1950 ± 1954.
Loeb LA. (1991). Cancer Res., 51, 3075 ± 3079.
Markowitz SD, Myero€ L, Cooper MJ, Traico€ J, Kochera
M, Lutterbaugh J, Swiriduk M and Willson JKV. (1994).
J. Clin. Invest., 93, 1005 ± 1013.
McBain JA, Weese JL, Meisner LF, Wolberg WH and
Willson JKV. (1984). Cancer Res., 44, 5813 ± 5821.
McGregor WG, Maher VM and McCormick JJ. (1991). Som.
Cell Mol. Gen., 17, 463 ± 469.
Michaels ML, Cruz C, Grollman AP and Miller JH. (1992).
Proc. Natl. Acad. Sci. USA, 89, 7022 ± 7025.
Michaels ML, Cruz C and Miller JH. (1990). Proc. Natl.
Acad. Sci. USA, 87, 9211 ± 9215.
Modrich P. (1991). Ann. Rev. Genet., 25, 229 ± 253.
Modrich P and Lahue R. (1996). Annu. Rev. Biochem., 65,
101 ± 133.
Meuth M. (1989). Exp. Cell Res., 181, 305 ± 316.
Nghiem Y, Cabrera M, Cupples CG and Miller JH. (1988).
Proc. Natl. Acad. Sci. USA, 85, 2709 ± 2713.
Nigro JM, Baker SJ, Preisinger AC, Jessup JM, Hostetter R,
Cleary K, Bigner SH, Davidson N, Baylin S, Devilee P,
Glover T, Collins FS, Weston A, Modali R, Harris CC
and Vogelstein B. (1989). Nature, 342, 705 ± 708.
Nowell PC. (1976). Science, 194, 23 ± 28.
Parsons R, Li G-M, Longley MJ, Fang W-h, Papadopoules
N, Jen J, de la Chapelle A, Kinzler KW, Vogelstein B and
Modrich P. (1993). Cell, 75, 1227 ± 1236.
Phear G and Meuth M. (1989). Mol. Cell. Biol., 9, 1810 ±
1812.
Powell SM, Zilz N, Beazer-Barclay Y, Bryan TM, Hamilton
SR, Thibodeau SN, Vogelstein B and Kinzler KW. (1992).
Nature, 359, 235 ± 237.
Schaaper RM. (1988). Proc. Natl. Acad. Sci. USA, 85, 8126 ±
8130.
Schaaper RM and Radman M. (1989). EMBO J.,8, 3511 ±
3516.
Su S-S, Lahue RS, Au KG and Modrich P. (1988). J. Biol.
Chem., 263, 6829 ± 6835.
Umar A and Kunkel TA. (1996). Eur. J. Biochem., 238, 297 ±
307.
Wang CY, Eshleman JR, Willson JKV and Markowitz S.
(1995). Cancer Res., 55, 5101 ± 5105.
Yanofsky C, Cox EC and Horn V. (1966). Proc. Natl. Acad.
Sci. USA, 55, 274 ± 281.