Nucleic Acids Research, 1994, Vol. 22, No. 23 4869-4871
Human and E.coli excinucleases are affected differently by
the sequence context of acetylaminofluorene-guanine
David Mu, Elisabeth Bertrand-Burggraf1, Juch-Chin Huang, Bobert P.P.Fuchs1 and Aziz Sancar*
Department of Biochemistry and Biophysics, University of North Carolina, School of Medicine,
Chapel Hill, NC 2759-79260, USA and 1UPR 9003 Cancerogenese et Mutagenese, Moleculaire et
Structurale, CNRS, Ecole Superieure de Biotechnologie de Strasbourg, Boulevard Sebastien Brandt,
67400 Strasbourg/lllkrich-Graffenstaden, France
Received September 26, 1994; Revised and Accepted October 18, 1994
Synthetic DNA substrates containing an acetylamino-
fluorene (AAF) adduct at each of the three guanine in
the G1G2CG3CC sequence were constructed and
tested as substrates for reconstituted E.coli (A)BC
excinuclease and human excinuclease in HeLa cell-free
extract (CFE). The (A)BC excinulcease repaired the
three substrates with relative efficiencies of Gl :G2:G3
[Seeberg, E., and Fuchs, R.P.P. (1990) Proc. NatlAcad.
Sci. USA 87, 191 - 194]. The same lesions were
repaired by the human excinuclease with the strikingly
different efficiencies of Gl :G2:G3 as 38:100:68. These
results reveal that the human excinuclease is affected
by the sequence context of the lesion in a different
manner than its prokaryotic counterpart.
in agreement with an ealier report
Mutagenic potential ofa DNA adduct depends on several factors
ranging from the types oflesions, sequence context oflesion and
relative location ofthe lesion with regard to the replication fork,
to the repairability of the lesion (1, 2). Analyses of mutation
spectra ofAAF in E.coli revealed that NarI recognition sequence
(GGCGCC) was a particularly hot spot, with -2 frameshift
mutations predominating (3). A study with uniquely modified
DNA revealed that all of the mutations at this site arose from
the G3-AAF adduct at the NarI sequence (4). Recently these
uniquely modified substrates were tested in an in vitro human
replication/mutagenesis system (2) and again it was found that
only the G3 adduct was mutagenic. Thus the human and E.coli
systems behaved nearly identically in terms of mutagenicity for
AAF-NarI sequence adducts. Given the similar mutagenic
behavior, we wished to know whether in human as well there
was a discordance between the repairability and the mutagenicity
of AAF lesions at the NarI sequence. The results of this report
suggest that repairability is not an important determinant in the
mutagenicity ofAAF adducts in either humans or E.coli and that,
at least in this particular case, the sequence context of the lesion
has entirely different effects on human and E.coli excinucleases.
MATERIALS AND METHODS
The substrates were 138-bp duplexes which were constructed by
ligating an AAF-adducted 14-nt oligomer with four other
oligomers as described previously (5). In the particular batches
Table 1. Extents of excision by human and E. coli excinucleases on AAF adducts at 3 different
locations in the same sequence
'The reaction mixtures contained ca. lnM substrate (identical concentrations for all threesubstrates)
and saturating amount ofenzymeas determined experimentally.The actualpercentexcisions(< 10%)
for the human enzyme are comparable to those of excision assays with other substrates in our
laboratory, indicating the superiority of a defined system as in the case of Ecoli enzyme.
2Thew numbers were obtained by averaging the three experiments, raisingtheoptimal substrate
excision level to 100%, and normalizing the other two substrate excision level to that of theoptimal.
*To whom correspondence should be addressed
.-. 1994Oxford UniversityPress
Nucleic Acids Research, 1994, Vol. 22, No. 23
TGTdG&AATTGCAGACTCAqTA TcAcAccAcA CGAGCT
LTM Gi G2 G3UM GI G2 G3
Figure 1. (Top panel) The G-AAF substrates used in this study. The 138-bp
duplexes were constructed by ligating the AAF-adducted 14 mer (sequence in
lower case) with 4 other oligomers. The 32p label is indicated by an asterisk and
the position ofthe AAF adduct is marked. The brackets indicate the incision sites
ofthe E.coli excinuclease and the extrapolated incision sites ofhuman excinuclease
for the longest excision product. (Bottom panel) Excision of G-AAF adducts
by E.coli and human excinucleases. The lanes 1-4 contained 1/20th the amount
ofDNA loaded into lanes 5-8. The G-AAF is excised in the form of 12 mer,
consistent with the previous report of Seeberg and Fuchs (1). The band at 63
nt in lane 4 results from excision of the G3 substrate. The numbers 26, 27, 28
are the major excision products of human excinuclease although the relative
frequency varies depending on the substrate.
of 14-mers used in this study, the GI and G2 were free of
contaminants but G3 substrate was contaminated with GI by
about 20% as judged by T4-polymerase assay (6, 7). This
contamination was attributed to the incomplete separation ofG3
oligomer from GI oligomer on a C18 reverse phase column by
high performance liquid chromatography.
HeLa whole cell-free extract prepared by the method ofManley
et al. (8) was used as a crude source of human excinuclease.
The E.coli (A)BC excinuclease was reconstituted from purified
UvrA, UvrB, and UvrC proteins (9). Assay reactions were
conducted according to Huang etal. (10) for the human enzyme
and Sancar and Rupp (9) for the E.coli enzyme. The level of
excision was quantitated by scanning the sequencing gel with an
Ambis scanner. The area covering the 25-30 nt was scanned
for the human excinuclease and the intensity of the 12 nt band
was taken to be the measure of the E.coli excinuclease activity.
The final precent excision for each experiment was averaged over
two independent readings from the Ambis scanner.
RESULTS AND DISCUSSION
Figure 1 (top panel) shows the three substrates used in this study.
Excision of these lesions by the E.coli excinuclease is expected
to release a radiolabeled 12 mer from the Gi and G2 substrates
and generate a 63 nt long labeled fragment from the G3 substrate.
Figure 1 (bottom panel) shows that this is indeed the case except
that with the G3 substrate a 12 mer is also seen due to the
contaminating GI substrate (see 'Materials and Methods'). Figure
1 also reveals an important contrast between the human and E.coli
excinucleases: the human excinuclease prefers the G2 substrate
which is excised rather poorly by the E.coli enzyme (lane 3 vs
7, see ref. 1). Both the E.coli and human enzymes were tested
under a variety ofconditions. In the case ofhuman excinuclease,
different preparations of HeLa CFE's were used in the excision
assay, whereas in the case of E.coli the Uvr A, B, C proteins
were stored for different lengths oftime. In view ofthese various
conditions it is therefore not surprising to see a spectrum of
excision levels (Table 1). In spite of the variations of the actual
excison levels between experiments, it is important to point out
that the relative extents of repair ofthe three substrates remained
nearly constant. Furthermore, it is evident from Table 1 that the
human and E.coli enzymes have opposing substrate preferences
within this sequence context.
In vitro the effect of sequence context on repair has been
measured only with the E.coli excinuclease. Jones and Yeung
(11) reported that the sequence context was an important
determinant in what side (furan vs pyrone) was incised in
psoralen-interstrand crosslinks. Thomas et al. (12) found that,
in general, the excision ofUV photoproducts and 4-nitroquinoline
oxide-guanine adducts was sequence-independent; however it was
found that a subclass of both lesions was repaired at a slower
rate although no obvious consensus sequence could be inferred
for the slow-repaired sites. Svoboda et al. (13) systematically
varied the sequence within 10 bp around a T< >T and found
that, essentially, the repair rate was sequence-independent
although the repair rates at the two extremes differed by nearly
a factor of 2. The most dramatic sequence effect on Uvr (A)BC
excinulcease reported to date is that of Seeberg and Fuchs (1)
on the repair of G1, G2, and G3 substrates used in this report.
By normalizing the excision signals of the other two substrates
to that ofthe optimal G1 substrate, the three lesions were excised
at 100%, 15%, and 58%, respectively.
In contrast to the studies on E.coli excinuclease, this is the first
study on the effect of sequence context on repair by the human
Nucleic Acids Research, 1994, Vol. 22, No. 23 4871
excinuclease. Based on this result, we predict that the human
excinuclease has different sequence preference compared with
the prokaryotic enzyme even when the mutagenesis spectra of
the two organisms are similar or identical.
This work was supported by the NIH grant GM32833 and by
a Human Frontier Science Program Organization. D.M. is
supported by a fellowship (DRG-1319) from the Cancer Research
Fund of the Damon Runyon-Walter Winchell Foundation.
1. Seeberg, E., and Fuchs, R. P. P. (1990) Proc. Natl Acad. Sci. USA 87,
2. Thomas, D. C., Veaute, X., Kunkel, T. A., and Fuchs, R. P. P. (1994)
Proc. Natl Acad. Sci. USA 91, 7752-7756.
3. Koffel-Schwartz, N., Verdier, J. M., Bichara, M., Freund, A. M., Daune,
M. P., and Fuchs, R. P. P. (1984) J. Mol. Bio. 177, 33-51.
4. Bumouf, D., Koehl, P., and Fuchs, R. P. P. (1989) Proc. NatlAcad. Sci.
USA 86, 4147-4151.
5. Bertrand-Burggraf, E., Kemper, B., and Fuchs, R. P. P. (1994) Mutation
Res. 314, 287-295.
6. Koehl, P. Burnouf, D., and Fuchs, R. P. P. (1989) J. Mol. Biol. 207,
7. Fuchs, R. (1984) J. Mol. Bio. 177, 173-180.
8. Manley, J. L., Fire, A., Cano, A., Sharp, P. A., and Gefter, M. L. (1980)
Proc. Natl Acad. Sci. USA 77, 3855-3859.
9. Sancar, A. and Rupp, W. D. (1983) Cell 33, 249-260.
10. Huang, J. C., Svoboda, D. L., Reardon, J. T., and Sancar, A. (1992) Proc.
Natl Acad. Sci. USA 89, 3664-3668.
11. Jones, B. K., and Yeung, A. T. (1990) J. Biol. Chem. 265, 3489-3496.
12. Thomas, D. C., Husain, I., Chaney, S. G., Panigrahi, G. B., and Walker,
I. G. (1991) Nucleic Acids Res. 19, 365-370.
13. Svoboda, D. L., Smith, C. A., Taylor, J.-S., and Sancar, A. (1993) J. BioL
Chem. 268, 10694-10700.