Conformational Heterogeneity of Arylamine-Modified DNA: 19F NMR evidence

Article abstract of DOI:10.1021/ja9632771

One- and two-dimensional ¹⁹F NMR spectroscopy was used to investigate the conformational heterogeneity of two arylamine-modified DNA duplexes, d[CTTCTTG*ACCTC]·d[GAGGTCAAGAAG], in which G* is either N-(deoxyguanosin- 8-yl)-4'-fluoro-4-aminobiphenyl (dG-C8-FABP) (I) or N-(deoxyguanosin-8-yl)- 7-fluoro-2-aminofluorene (dG-C8-FAF) (II). The ¹⁹F NMR spectrum of I showed a single peak, while that of II revealed two prominent signals with a 55:45 ratio, in good agreement with previous ¹H NMR results (Cho et al. Biochemistry 1992, 31, 9587-9602; 1994, 33, 1373-1384). Slow interconversion between the two conformations of II was established by temperature-dependent two-dimensional ¹⁹F NMR chemical exchange spectra. On the basis of magnetic anisotropy effects and isotopic solvent-induced shifts, the ¹⁹F signals at -117.31 and -118.09 ppm in the ¹⁹F NMR spectrum of 11 were assigned to a relatively undisturbed 'B-type' conformer and a highly perturbed 'stacked' conformer, respectively. Analysis of the temperature dependent (5-40°C) line shapes by computer simulation yielded an interconversion barrier (ΔG((+))) of 14.0 kcal/mol with a chemical exchange time of 2 ms at 30°C. This new ¹⁹F approach should be very useful in investigating the sequence-dependent conformational heterogeneity of arylamine-modified DNA.

Full text of DOI:10.1021/ja9632771

5384
J. Am. Chem. Soc. 1997, 119, 5384-5389
Conformational Heterogeneity of Arylamine-Modified DNA:
1
9
F NMR Evidence
†
‡
‡
,†
Li Zhou, Masoumeh Rajabzadeh, Daniel D. Traficante, and Bongsup P. Cho*
Contribution from the Departments of Biomedical Sciences and Chemistry, UniVersity of Rhode
Island, Kingston, Rhode Island 02881
X
ReceiVed September 17, 1996. ReVised Manuscript ReceiVed April 17, 1997
Abstract: One- and two-dimensional 19F NMR spectroscopy was used to investigate the conformational heterogeneity
of two arylamine-modified DNA duplexes, d[CTTCTTG*ACCTC]‚d[GAGGTCAAGAAG], in which G* is either
N-(deoxyguanosin-8-yl)-4′-fluoro-4-aminobiphenyl (dG-C8-FABP) (I) or N-(deoxyguanosin-8-yl)-7-fluoro-2-ami-
nofluorene (dG-C8-FAF) (II). The 19F NMR spectrum of I showed a single peak, while that of II revealed two
1
prominent signals with a 55:45 ratio, in good agreement with previous H NMR results (Cho et al. Biochemistry
1
992, 31, 9587-9602; 1994, 33, 1373-1384). Slow interconversion between the two conformations of II was
19
established by temperature-dependent two-dimensional F NMR chemical exchange spectra. On the basis of magnetic
1
9
19
anisotropy effects and isotopic solvent-induced shifts, the F signals at -117.31 and -118.09 ppm in the F NMR
spectrum of II were assigned to a relatively undisturbed “B-type” conformer and a highly perturbed “stacked”
conformer, respectively. Analysis of the temperature dependent (5-40 °C) line shapes by computer simulation
q
yielded an interconversion barrier (∆G ) of 14.0 kcal/mol with a chemical exchange time of 2 ms at 30 °C. This
new 19F approach should be very useful in investigating the sequence-dependent conformational heterogeneity of
arylamine-modified DNA.
Introduction
form covalent adducts in ViVo to give rise to C8-substituted
deoxyguanosine adducts as the major and the most persistent
forms (dG-C8-ABP and dG-C8-AF, respectively, Figure 1).1
Elucidation of three-dimensional molecular structures of DNA
containing these adducts is crucial to understanding the differ-
Arylamines and amides are an important class of chemical
carcinogens, some of which induce tumors in experimental
animals and in humans.1 Among the most extensively studied
arylamines are 4-aminobiphenyl (ABP) and 2-aminofluorene
4-6,7m
ences in their mutagenic and carcinogenic outcomes.
(8) (a) Mao, B.; Vyas, R. R.; Hingerty, B. E.; Broyde, S.; Basu, A. K.;
1-3
(
AF) and their analogs. DNA is considered to be a principal
Patel, D. J. Biochemistry 1996, 35, 12659-12670. (b) Cosman, M.; de los
Santos, C.; Fiala, R.; Hingerty, B. E.; Singh, S. B.; Ibanez, V.; Margulis,
L. A., Live, D.; Geacintov, N. E.; Broyde, S.; Patel, D. J. Proc. Natl. Acad.
Sci. U.S.A. 1992, 89, 1914-1918. (c) de los Santos, C.; Cosman, M.;
Hingerty, B. E.; Ibanez, V. Margulis, L. A.; Geacintov, N. E.; Broyde, S.;
Patel, D. J. Biochemistry 1992, 31, 5245-5252. (d) Cosman, M.; de los
Santos, C.; Fiala, R.; Hingerty, B. E.; Ibanez, V.; Luna, E.; Harvey, R. G.;
Geacintov, N. E.; Broyde, S.; Patel, D. J. Biochemistry 1993, 32, 4145-
4155. (e) Cosman, M.; Hingerty, B. E.; Luneva, N.; Amin, S.; Geacintov,
N. E., Broyde, S.; Patel, D. J. Biochemistry 1996, 35, 9850-9863. (f)
Fountain, M. A.; Krugh, T. R. Biochemistry 1995, 34, 3152-3161. (g)
Cosman, M.; Xu, R.; Hingerty, B. E.; Amin, S.; Harvey, R. G., Geacintov,
N. E.; Broyde, S.; Patel, D. J. Biochemistry 1995, 34, 6247-6260. (h)
Cosman, M.; Fiala, R.; Hingerty, B. E.; Amin, S.; Geacintov, N. E.; Broyde,
S.; Patel, D. J. Biochemistry 1994, 33, 11507-11517. (i) Cosman, M.; Fiala,
R.; Hingerty, B. E.; Amin, S.; Geacintov, N. E., Broyde, S.; Patel, D. J.
Biochemistry 1994, 33, 11518-11527. (j) Cosman, M.; Fiala, R.; Hingerty,
B. E.; Laryea, A.; Lee, H. Harvey, R. G.; Amin, S.; Geacintov, N. E.;
Broyde, S.; Patel, D. J. Biochemistry 1993, 32, 12488-12497. (k) Cosman,
M.; Laryea, A.; Fiala, R.; Hingerty, B. E.; Amin, S.; Geacintov, N. E.;
Broyde, S.; Patel, D. J. Biochemistry 1995, 34, 1295-1307. (l) Cosman,
M., Hingerty, B. E.; Geacintov, N. E.; Broyde, S.; Patel, D. J. Biochemistry
1995, 34, 15334-15350. (m) Schurter, E. J.; Sayer, J. M.; Oh-hara, T.;
Yeh, H. J. C.; Yagi, H.; Luxon, B. A.; Jerina, D. M.; Gorenstein, D. G.
Biochemistry 1995, 34, 9009-9020. (n) Schurter, E. J.; Yeh, H. J. C.; Sayer,
J. M.; Lakshman, M. K.; Yagi, H.; Jerina, D. M.; Gorenstein, D. G.
Biochemistry 1995, 34, 1364-1375. (o) Yeh, H. J. C.; Sayer, J. M.; Liu,
X.; Altieri, A. S.; Ryrd, R. A.; Lakshman, M. K.; Yagi, H.; Schurter, E. J.;
Gorenstein, D. G.; Jerina, D. M. Biochemistry 1995, 34, 13570-13581.
(p) Feng, B.; Voehler, M.; Zhou, L.; Passarelli, M.; Harris, C. M.; Harris,
T. M.; Stone, M. P. Biochemistry 1996, 35, 7316-7329. (q) Belguise-
Valladier, P.; Fuchs, R. P. P. Biochemistry 1991, 30, 10091-10100. (r)
Koehl, P.; Valladier, P.; Lef e` vre, J.-F.; Fuchs, R. P. P. Nucleic Acids Res.
1989, 17, 9531-9541. (s) Veaute, X.; Fuchs, R. P. P. Nucleic Acids Res.
1991, 19, 5603-5606. (t) Seeberg, E.; Fuchs, R. P. P. Proc. Natl. Acad.
Sci. U.S.A. 1990, 87, 191-194. (u) Rodriguez, H.; Loechler, E. L.
Biochemistry 1993, 32, 1759-1769. (v) Drouin, E. E.; Lech, J.; Loechler,
E. L. Biochemistry 1995, 34, 2251-2259. (w) Jelinsky, S.; Liu, T.;
Geacintov, N. E.; Loechler, E. L. Biochemistry 1995, 34, 13545-13553.
target for the activated derivatives of these carcinogens, which
*
To whom correspondence should be addressed. Phone: (401) 874-
5
024. Fax: (401) 874-2181. E-mail: bcho@uriacc.uri.edu.
†
Department of Biomedical Sciences.
Department of Chemistry.
Abstract published in AdVance ACS Abstracts, June 1, 1997.
‡
X
(
1) Beland, F. A.; Kadlubar, F. F. In Handbook of Experimental
Pharmacology; Cooper, C. S., Grover, P. L., Eds.; Vol 94/I, Springer-
Verlag: Heidelberg, 1990; pp 267-325.
(
(
(
2) Kriek, E. J. Cancer Res. Clin. Oncol. 1992, 118, 481-489.
3) Heflich, R. H.; Neft, R. E. Mutat. Res. 1994, 318, 73-174.
4) Harris, T. M.; Stone, M. P.; Harris, C. M. Chem. Res. Toxicol. 1988,
1
, 79-96.
5) (a) Patel, D. J. Curr. Opin. Struct. Biol. 1992, 2, 345-353. (b) Krugh,
T. R. Curr. Opin. Struct. Biol. 1994, 4, 351-364.
6) Geacintov, N. E.; Cosman, M.; Hingerty, B. E.; Amin, S.; Broyde,
S.; Patel, D. J. Chem. Res. Toxicol. 1997, 10, 111-146.
7) (a) Norman, D.; Abuaf, P.; Hingerty, B. E.; Live, D.; Grunberger,
(
(
(
D.; Broyde, S.; Patel, D. J. Biochemistry 1989, 28, 7462-7476. (b) Cho,
B. P.; Beland, F. A.; Marques, M. M. Biochemistry 1992, 31, 9587-9602.
(
c) O’Handley, S. F.; Sanford, D. G.; Xu, R.; Lester, C. C.; Hingerty, B.
E.; Broyde, S.; Krugh, T. R. Biochemistry 1993, 32, 2481-2497. (d) Cho,
B. P.; Beland, F. A.; Marques, M. M. Biochemistry 1994, 33, 1373-1384.
(
e) Eckel, L. M.; Krugh, T. R. Nat. Struct. Biol. 1994, 1, 89-94;
Biochemistry 1994, 33, 13611-13624. (f) Shapiro, R.; Sidawi, D.; Miao,
Y.-S.; Hingerty, B. E., Schmidt, K. E.; Moskowitz, J.; Broyde, S. Chem.
Res. Toxicol. 1994, 7, 239-253. (g) Abuaf, P.; Hingerty, B. E.; Broyde,
S.; Grunberger, D. Chem. Res. Toxicol. 1995, 8, 369-378. (h) Shapiro, R.;
Ellis, S.; Hingerty, B. E.; Broyde, S. Chem. Res. Toxicol. 1995, 8, 117-
1
27. (i) Milhe, C.; Dhalluin, C.; Fuchs, R. P. P.; Lef e` vre, J.-F. Nucleic
Acids Res. 1994, 22, 4646-4652. (j) Milhe, C.; Fuchs, R. P. P.; Lef e` vre,
J.-F. Eur. J. Biochem. 1996, 235, 120-127. (k) Mao, B.; Cosman, M.;
Hingerty, B. E.; Broyde, S.; Patel, D. J. Biochemistry 1995, 34, 6226-
6238. (l) Mao, B.; Hingerty, B. E., Broyde, S.; Patel, D. J. Biochemistry
1995, 34, 16641-16653. (m) Broyde, S.; Hingerty, B. E. Biopolymers 1983,
22, 2423-2441.
S0002-7863(96)03277-5 CCC: $14.00 © 1997 American Chemical Society

Conformational Heterogeneity of DNA
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5385
8
a
<
AP (1-aminopyrene, ∼100%). This is presumably due to
9
differences in the stacking abilities of the carcinogen fragments,
which may account for the lower frameshift mutagenicity of
dG-C8-ABP in Salmonella typhymurium TA 1538 and for the
relative resistance of the same adduct as compared to dG-C8-
1
0
AF to DNA repair enzymes in the livers of dogs. Recent
1
1
mutation studies by Melchior et al. using plasmid pBR322
transfected into Escherichia coli have shown that the highly
planar AP produces an unusually high incidence of frameshifts,
and also a much higher mutational efficiency per adduct as
compared to the less planar ABP, further substantiating the
stacking argument.
Conformational heterogeneity appears to be the basis for the
6
,7,8a,e,f,k,l,n,o-w
diverse nature of adduct-initiated mutagenesis.
Since mutagenesis is a relatively infrequent biological event and
DNA adducts adopt multiple conformations, it is conceivable
that observed mutations are due to the replication of an adduct
1
2
1
in a minor conformation. Until now, H NMR spectroscopy
has been used as a principal tool to investigate the solution
4
-8
structure of carcinogen adducts in DNA.
However, when
an adduct induces multiple conformations, as is often the case
1
for carcinogen-modified DNA, H NMR data are extremely
difficult to interpret owing to spectral complexities and chemical
exchanges. It is therefore not surprising that prior NMR
structural studies of carcinogen-modified DNA have focused
Figure 1. Chemical structures of the C8-substituted deoxyguanosine
adducts and the base sequence of the modified 12-mer in which the
carcinogen is covalently attached at the C8-position of G* (I, FABP-
modified 12-mer duplex, G* ) dG-C8-FABP; II, FAF-modified 12-
mer duplex, G* ) dG-C8-FAF). Key: dG-C8-ABP, N-(deoxyguanosin-
6,7a-f,i-l,8a,d-f,k-p
mostly on the major conformations.
This neces-
sitates the development of novel techniques, in which one can
correlate mutations to a specific adduct conformation.12
In this paper, we have explored 19F NMR spectroscopy as
8
-yl)-4-aminobiphenyl; dG-C8-FABP, N-(deoxyguanosin-8-yl)-4′-
fluoro-4-aminobiphenyl; dG-C8-AF, N-(deoxyguanosin-8-yl)-2-
aminofluorene; dG-C8-FAF, N-(deoxyguanosin-8-yl)-7-fluoro-2-
aminofluorene.
an alternate method that takes advantage of the sensitivity of
19
the F chemical shifts to the tertiary structural environment of
13-16
nucleic acids.
The approach has been to prepare two DNA
duplexes containing fluorine-tagged model carcinogens (I and
II, Figure 1), and to investigate their static and dynamic 19
F
NMR characteristics.
Results and Discussion
Design of the Fluorine Probes. The model fluorine probes
chosen for the present study were two 12-mer DNA duplexes
containing either N-(deoxyguanosin-8-yl)-4′-fluoro-4-aminobi-
phenyl (dG-C8-FABP) or N-(deoxyguanosin-8-yl)-7-fluoro-2-
aminofluorene (dG-C8-FAF) (I and II, respectively, Figure 1).
Incorporation of fluorine at the longest axis (i.e., 4′-position of
ABP and 7-position of AF) has been shown to maintain
Figure 2. Schematic representations of the B-type conformer and the
stacked conformer observed for an aminofluorene-modified DNA
duplex.
carcinogenicity in rat livers, while the activity toward other
tissues is comparable to that of the parent compound.1
,17,18
The
DNA adduct profiles of FABP, FAF, and their corresponding
nonfluoro compounds are also similar. According to circular
We7b,d and others have shown that arylamine-modified DNA
7e
19
duplexes adopt two major prototype conformations: namely, a
(
9) Shapiro, R.; Underwood, G. R.; Zawadzka, H.; Broyde, S.; Hingerty,
B. E. Biochemistry 1986, 25, 2198-2205.
10) Beland, F. A.; Beranek, D. T.; Dooley, K. L.; Heflich, R. H.;
Kadlubar, F. F. EnViron. Health Prospect. 1983, 49, 125-134.
11) Melchior, W. B. Jr.; Marques, M. M.; Beland, F. A. Carcinogenesis
994, 15, 889-899.
12) Loechler, E. L. Carcinogenesis 1996, 17, 895-902.
(13) Craik, D. J. NMR in Drug Design; CRC Press: Boca Raton, FL,
996.
14) Gerig, J. T. Methods Enzymol. 1989, 177, 3-23.
15) Lu, P.; Metzler, W. J.; Rastinejad, F.; Wasilewski, J. In Structure
“
B-type” conformer, in which the carcinogen resides in the
major groove of a relatively undistorted B-DNA, and a “stacked”
conformer, in which the carcinogen is inserted into the helix at
the adduct site (Figure 2). Due to the highly perturbed nature
of the DNA adduct at the modification site, the stacked
(
(
1
(
7d,e
conformer is considered as a promutagenic conformer.
These
1
experimental findings are in accord with previous theoretical
predictions, which have suggested several possible AF-induced
dynamic states in DNA.7m Such conformational heterogeneity
(
(
and Function of Nucleic Acids and Proteins: Wu, F. Y.-H., Wu, C.-W.,
Eds.; Raven Press: New York, 1990; pp 19-35.
appears to be commonly associated with a wide range of
carcinogen-modified DNA adducts6-8 and is strongly modulated
(16) Rastinejad, F.; Evilla, C.; Lu, P. Methods Enzymol. 1995, 261, 560-
5
74.
by the nature (coplanarity, stereochemistry, etc.) of the
(17) Arcos, J. C.; Argus, M. F. Chemical Induction of Cancer; Academic
8
a-p
carcinogens,
as well as the base sequence context surround-
Press: New York, 1974; Vol. IIB.
ing the adduct.7
a,d,f-l,8a,e,f,l,p-w
(18) Miller, J. A.; Sandin, R. B.; Miller, E. C.; Rusch, H. P. Cancer
Res. 1956, 15, 188-199.
For example, the population of
the promutagenic stacked conformer increases as the planarity
(19) Fuchs, R. P. P.; Lef e` vre, J.-F.; Pouyet, J.; Daune, M. P. Biochemistry
7b
7d,e
of the carcinogen increases: ABP (∼10%) < AF (∼50%)
1976, 15, 3347-3351.

5386 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Zhou et al.
Figure 4. 1H NMR (400 MHz, 1D) spectra (12-15 ppm) of (a) the
FABP- (I) and (b) the FAF-modified (II) 12-mer duplexes in 90% H
0% D O at 18 °C. The spectra were recorded using a 1-1 jump and
return pulse sequence and referenced relative to DSS (2,2-dimethyl-
-silapentane-5-sulfonate, sodium salt).
2
O/
1
2
28
2
FABP and dG-C8-FAF adducts exhibited characteristic multiplets
at -119.56 and -120.03 ppm, respectively.
The FABP- and FAF-modified 12-mers (d[CTTCTTG*
ACCTC], G* ) dG-C8-FABP and dG-C8-FAF, respectively)
were similarly prepared as described for the nonfluoro ABP-
Figure 3. Ultraviolet spectra of (a) dG-C8-FABP and the unmodified
and FABP-modified 12-mer and (b) dG-C8-FAF and the unmodified
and FAF-modified 12-mer, in the 240-360 nm range, monitored using
a photodiode array detector.
2
4
and AF-modified 15-mer DNA oligomers. The UV spectra
Figure 3) of the modified 12-mers exhibited a broad shoulder
(
dichroism19,20 and molecular modeling21 studies, dG-C8-FAF
and dG-C8-AF exert similar conformational alterations in DNA.
The 12-mer sequence (d[CTTCTTGACCTC]‚d[GAGGTCAA-
GAAG], Figure 1) employed in this study is essentially that of
the original 15-mer (d[TACTCTTCTTGACCT]‚d[AGGTCAA-
above 300 nm, which is characteristic of the extended conjuga-
tion expected from DNA containing C8-substituted arylamine
2
4
adducts. HPLC analysis of an enzymatic hydrolysate of the
FABP-12-mer to nucleosides displayed several peaks, one of
which was distinctively nonpolar and identified as dG-C8-FABP
on the basis of comparison of its UV spectrum and retention
time with those of a standard (Vide supra). The FAF-modified
12-mer was characterized in a similar manner.
1
7b,d
GAAGAGTA]) used in our previous H NMR work,
but
shortened to facilitate synthetic and spectral work without
perturbing the unique conformational features at the adduct site.
Preparation of the Fluorine Probes. The synthesis, puri-
fication, and chemical characterization of the monomeric dG
adducts (dG-C8-FABP and dG-C8-FAF, Figure 1) were carried
Optical Melting Curves. Optical melting curves of the
unmodified, FABP- and FAF-modified (I and II, Figure 1)
duplexes at 5 µM concentration showed a typical cooperative
helix-coil transition upon heating. The midpoint transition
temperatures (Tm) were found to be 46 ( 1, 36 ( 1, and 35 (
1 °C for the unmodified, FABP-modified (I), and FAF-modified
(II) duplexes, respectively. As compared to the 15-mer
2
2
out by application of the general method of Lee and King.
The synthesis involves a treatment of dG with the reactive
N-acetoxy-N-(trifluoroacetyl) derivatives of 4′-fluoro-4-amino-
biphenyl and of 7-fluoro-2-aminofluorene, with concomitant
solvolysis of the trifluoroacetyl fragment to produce the
deacetylated arylamine adduct at the C8 position of guanine.
The UV characteristics of the monomeric dG-C8-FABP and dG-
C8-FAF adducts (Figure 3) are similar to those of the corre-
2
4
sequences containing nonfluoro analogs, the Tm values were
slightly higher for the unmodified 12-mer, and relatively
unchanged for the modified duplexes, while the extent of the
adduct-induced thermal destabilization (10-11 °C) was similar.
The addition of a G:C pair at both the 5′- and 3′-terminals of
the 12-mer sequence may account for their relative stability.
Thermodynamic parameters for the three duplexes were
calculated at 25°C using the procedure of Marky and Breslauer:
sponding nonfluoro analogs.2
3,24
The H NMR spectra of these
1
monomeric adducts have been analyzed by COSY experiments
1
and by comparison with the H NMR spectra of the correspond-
ing nonfluoro parent adducts.2
5,26
The H NMR spectrum of
1
2
7
dG-C8-FABP exhibited all the resonances of dG (except for
the H8) and four groups of aromatic signals integrating for eight
protons, which is in good agreement with that reported by van
de Poll et al. for the same adduct.25 Likewise, the H NMR
spectral pattern of dG-C8-FAF was very similar to that of dG-
unmodified 12-mer, ∆H° ) -5.2 kcal/mol, ∆S° ) 22.8 eu,
∆G° ) -12.0 kcal/mol; FABP-duplex (I), ∆H° ) -5.2 kcal/
mol, ∆S° ) 18.4 eu, ∆G° ) -10.7 kcal/mol; FAF-duplex (II),
∆H° ) -4.1 kcal/mol, ∆S° ) 20.9 eu, ∆G° ) -10.3 kcal/
mol. The values calculated for II may not be accurate since it
involves a conformational exchange in the duplex state (Vide
infra). Despite the significant thermal destabilization, the small
difference in free energy (∆∆G° < 1.7 kcal/mol) between the
unmodified and modified 12-mer duplexes suggests that the
adduct-induced helix destabilization is relatively minimal.
NMR Studies. Figure 4 shows the imino proton spectra of
the FABP- and FAF-modified 12-mer DNA duplexes (I and
II, respectively) taken in H2O buffer at 18 °C. While I showed
the 12 expected sharp imino protons in the downfield region
1
26
C8-AF reported by Beland et al., except for the lack of a high-
field H7 aromatic signal. The 19F NMR spectra of the dG-C8-
(
20) Lef e` vre, J.-F.; Fuchs, R. P. P.; Daune, M. P. Biochemistry 1978,
1
7, 2561-2567.
(
(
(
21) Broyde, S.; Hingerty, B. Chem.-Biol. Interact. 1983, 47, 69-78.
22) Lee, M.-S.; King, C. M. Chem.-Biol. Interact. 1981, 34, 239-244.
23) Shibutani, S.; Gentles, R.; Johnson, F.; Grollman, A. P. Carcino-
genesis 1991, 12, 813-818.
24) Marques, M. M.; Beland, F. A. Chem. Res. Toxicol. 1990, 3, 559-
65.
25) van de Poll, M. L. M.; Tijdens, R. B.; Vondr a´ cek, P.; Bruins, A.
(
5
(
P.; Meijer, D. K. F.; Meerman, J. H. N. Carcinogenesis 1989, 10, 2285-
(27) Marky, L. A.; Breslauer, K. J. Biopolymers 1987, 26, 1601-1620.
(28) Sklenor, V.; Bax, A. J. Magn. Reson. 1987, 74, 469-479.
(29) Bodenhausen, G.; Kogler, H.; Ernst, R. R. J. Magn. Reson. 1984,
58, 370-388.
2
291.
(26) Beland, F. A.; Allaben, W. T.; Evans, F. E. Cancer Res. 1980, 40,
8
34-840 (1980).

Conformational Heterogeneity of DNA
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5387
Figure 5. 19F NMR spectra (376.5 MHz) of (a) I and (b) II taken in
9
0% H
2 2 3
O/10% D O at 18 °C, referenced to external CFCl .
(Figure 4a), the same sequence containing dG-C8-FAF (II)
exhibited at least 15 resolvable imino proton resonances with
varying degrees of intensity (Figure 4b). These results indicate
that I exists primarily in a single conformation and II adopts
19
multiple conformations. In line with this interpretation, the
F
NMR spectrum of I revealed a single resonance at -116.94
ppm (Figure 5a), while that of II exhibited two prominent
signals at -117.31 and -118.09 ppm (Figure 5b), each
representing a unique fluorine environment. Their intensity ratio
1
7d
was 55:45, in good agreement with previous H NMR results
that showed the presence of an approximately equal distribution
of the B-type and stacked conformers in a similar sequence
context.
The conformeric relationship between the two 19F signals in
the spectrum of II was established by temperature-dependent
NOESY/EXSY experiments. Figure 6a shows contour plots of
1
9
two-dimensional F NMR exchange spectra of II recorded at
8 °C, where the off diagonal cross peaks indicate a transfer of
1
magnetization from one conformer to another. When an
identical measurement was conducted at 5 °C (Figure 6b), no
such cross peaks were detected. This strong temperature
3
0
dependence confirms that the two signals arise Via chemical
exchange from two structurally distinct, slowly-interconvertible
conformations, presumably a B-type conformer and a stacked
conformer.
Figure 6. 19F NMR (2D) exchange spectra of II recorded in 100%
2
D O at (a) 18 °C and (b) 5 °C. The spectra were obtained in the phase-
On the basis of magnetic anisotropy effects and isotopic
solvent-induced shifts, the upfield signal (-118.09 ppm) in the
spectrum of II was assigned to that arising from the stacked
conformer. The large shift (+0.78 ppm, Figure 5b) relative to
that assigned to the B-type conformer (-117.31 ppm) can be
attributed to ring-current effects produced by intercalation of
the carcinogen moiety into the helix as depicted in Figure 2. A
comparable magnitude of shielding through this mechanism has
been observed for 5-fluorouracil residues located in helical
29
sensitive mode using a NOESY sequence: sweep width 4529 Hz,
number of complex data points in t 512, number of complex free
2
1
induction decays in t 128, number of scans 480, number of dummy
scans 16, recycle delays 1.0 s, and mixing time 400 ms. The data were
subjected to exponential apodization using a line broadening of 8 Hz
in both dimensions and then zero-filled before Fourier transformation
of the 512 × 256 data matrix. The data were not symmetrized. A slight
chemical shift difference of the downfield signal was due to a
temperature effect (see Figure 7): B ) B-type conformer; S ) stacked
conformer.
31
regions of tRNA. Such an intercalation also limits the motion
of the carcinogenic moiety, thus resulting in signal broadening.
More direct and compelling evidence for this assignment was
The spin-lattice relaxation times (T1 values) in H2O buffer
were also measured to gain further insight into the relaxation
nature of the two conformers of II. We anticipated that the
exposed fluorine of the B-type conformer would have a shorter
T1 as compared to that of the buried fluorine of the stacked
conformer, due to a relatively large dipolar contribution to the
32
obtained by observing a H-D isotope effect. As a result of
1
9
the interaction of the exposed fluorine with the solvent, the
F
resonance from the accessible FAF residue of the B-type
conformer was shifted upfield by 0.24 ppm when the deuterium
content of the solvent was increased from 10% to 100%. In
contrast, the chemical shift of the buried (intercalated) fluorine
resonance in the stacked conformer was relatively unchanged
3
3
relaxation from the solvent protons.
-
0
Unexpectedly, the
117.31 ppm F signal of the B-type conformer had a T1 of
.48 s, while that (-118.09 ppm) belonging to the stacked
1
9
conformer had a T1 of 0.52 s. The lack of significant difference
in the T1 values between the two conformers suggests that
intercalation of the carcinogen into the helix occurs without
complete exclusion of solvent molecules, or that the fluorine
nucleus is relaxed by mechanisms other than dipolar relaxation.
(
-
+0.08 ppm) upon the same solvent change. As expected, the
116.94 ppm signal in the spectrum of I (Figure 5a), which
adopts predominantly a B-type conformation, exhibited a similar
shielding (+0.23 ppm) under the same conditions.
(
30) Wemmer, D. E. In Biological Magnetic Resonance; Berliner, L. J.,
Reuben, J. Eds.; Plenum Press: New York, 1992; Vol. 10, pp 195-264.
31) Hardin, C. C.; Gollnick, P.; Kallenbach, N. R.; Cohn, M.; Horowitz,
J. Biochemistry 1986, 25, 5699-5709.
(32) Hansen, P. E.; Dettman, H. D.; Sykes, B. D. J. Magn. Reson. 1985,
62, 487-496.
(33) O’Connor, T. P.; Coleman, J. E. Biochemistry 1982, 21, 848-854.
(

5388 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Zhou et al.
-
1
found to be 76 and 500 s , which are equivalent to chemical
exchange lifetimes of 13 and 2 ms, respectively. On further
raising of the temperature, the coalescent peak became sharper,
and eventually narrowed at 60 °C (-117.10 ppm), showing fine
H-F splitting patterns (inset in Figure 7), which indicates a
rapidly rotating FAF moiety in a single-strand 12-mer.
The dynamic energetic results obtained for the FAF-modified
1
2-mer duplex (II) suggest that the conformeric equilibrium
between the B-type and the stacked conformers is physiologi-
cally accessible; thereby, repair enzymes are likely to be
confronted with the two conformations in ViVo. This situation
has been previously labeled as a “two-way” mutagenic switch
7
e
by Eckel and Krugh. In a similar manner, the term “three-
way” mutagenic switch was used to describe mutations derived
from a duplex containing propanodeoxyguanosine (PdG) op-
posite a -2 deletion site in the hisD3052 sequence, in which
the possibilities include successful bypass of PdG, miscoding
leading to base substitution, or strand slippage leading to a
3
6
frameshift mutation.
Conclusions
In this report we present conclusive evidence that the FABP-
1
2 mer (I) exists exclusively as a B-type conformer, while the
FAF-12 mer (II) adopts a 55:45 mixture of the B-type and
stacked conformers at room temperature. While detailed
structural interpretation of 19F NMR is limited, the sensitivity
1
9
of F chemical shifts to the DNA helix environment clearly
offers unique advantages, especially in dealing with multiple
conformations that are commonly associated with carcinogen-
1
9
modified DNA. With this F approach, one should be able to
monitor the conformeric ratio as a function of the base sequence
surrounding the adduct site, the pH, and the salt concentrations,
all of which are important factors in predicting mutagenic
outcomes. Such data will hopefully provide valuable quantita-
tive insights about the role of conformational heterogeneity in
1
9
arylamine carcinogenesis. This newly developed F approach
may provide a tool for understanding the conformational basis
of site-specific mutagenesis of arylamines and other related
carcinogens.
Experimental Section
Figure 7. Temperature-dependent 19F NMR spectra of II in 90% H
0% D O. The temperature range was 5-60 °C. Before Fourier
transformation, all data were apodized with an exponential window
function using a line-broadening factor of 5 Hz, except for the inset
spectrum at 60 °C, for which a Lorentzian to Gaussian resolution
enhancement (LB ) -1 Hz and GB ) 0.2) was used. The inset shows
fine H-F splitting patterns of a rapidly rotating FAF moiety.
2
O/
1
2
Warning. FABP and FAF deriVatiVes are mutagens as well as
carcinogens in experimental animals, and should be handled with
caution.
12-mer oligodeoxynucleotides (d[CTTCTTGACCTC] and d[GAG-
GTCAAGAAG]) were obtained from Keystone Laboratory, Inc. (Menlo
Park, CA). HPLC grade solvents were obtained from Fisher Scientific
(
Pittsburgh, PA). DNase I, phosphodiesterase I, and alkaline phos-
phatase were purchased from Sigma Chemical Co. (St. Louis, MO).
-Fluorobiphenyl was purchased from Pfaltz & Bauer, Inc. (Waterburg,
The dynamic interconversion between the B-type conformer
4
and the stacked conformer of II was probed by temperature-
1
9
34,35
CT). All other reagent chemicals were obtained from Aldrich Chemical
Co. (Milwaukee, WI).
dependent F NMR spectroscopy.
The results of 376.5
MHz 19F NMR are shown in Figure 7. While the two fluorine
Melting points were measured using a B u¨ chi melting point apparatus
and are uncorrected. Low- and high-resolution electron impact mass
spectra (LRMS and HRMS) were obtained on a Finnigan-MAT CH5
and a 731 instrument, respectively, at the University of Illinois Mass
Spectrometry Laboratory, Urbana-Champaign, IL. Electrospray mass
spectra of monomer dG adducts were measured on a Micromass VG
Quattro triple quadrupole mass spectrometer at the Northeastern
University Barnett Mass Spectrometry Laboratory, Boston, MA.
HPLC Methods. Samples were concentrated using a Model AES
1000-120 SpeedVac concentrator (Forma Scientific, Inc., Marietta, OH).
HPLC data were obtained on a Waters Associates system equipped
with Model 501 pumps, a U6K injector, a 680 automated gradient
controller, and a Hitachi L-3000 photodiode array detector. Separations
were conducted using either an Ultrasphere C18 ODS analytical column
signals were in slow exchange at 5 °C, they became exchange
broadened and moved closer together as the temperature was
raised, giving rise to one time-averaged coalescent signal
(-117.43 ppm) at 40 °C. The two signals showed line-shape
changes characteristic of a two-site exchange process, in which
the population of the stacked conformer increased steadily with
increasing temperature. Analysis of the temperature-dependent
(
5-40 °C) line shapes by computer simulation yielded the
q
activation parameters for exchange: ∆G (30 °C) ) 14.0 kcal/
mol, ∆H ) 16.0 kcal/mol, and ∆S ) 6.7 eu. The rates of
interconversion of the two conformers at 18 and 30 °C were
q
q
(
34) Friebolin, H. Basic One- and Two Dimensional NMR Spectroscopy,
nd ed.; VCH Publishers: New York, 1993; pp 263-291.
35) Sandstr o¨ m, J. Dynamic NMR Spectroscopy; Academic Press:
London, 1982.
2
(
(36) Weisenseel, J. P.; Moe, J. G.; Reddy, G. R.; Marnett, L. J.; Stone,
M. P. Biochemistry 1995, 34, 50-64.

Conformational Heterogeneity of DNA
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5389
(
4.6 × 250 mm, 5 µm) or a semiprep column (10.0 × 250 mm, 5 µm)
Synthesis of dG-C8-FABP and dG-C8-FAF. The monomer dG
23
from Beckman Instruments, Inc., San Ramon, CA, with one of the
following systems: system 1, 90% methanol/10% H O, isocratic (1
mL/min); system 2, a 60-min linear gradient of 5-35% acetonitrile/
.1 M ammonium acetate (pH 7.0) (2 mL/min); system 3, a 20-min
adducts were prepared by following a general literature method, except
that the adduct was extracted with ethyl acetate. To a pH 7.0 citrate
buffer (10 mM) solution containing dG (10 mg) was added an ethanol
solution of the appropriate diester (∼25 mg) obtained above. The
2
2
0
linear gradient of 25-60% acetonitrile/0.1 M ammonium acetate (pH
7
mixture was stirred at 45 °C under a N atmosphere. After 18 h, the
.0) (2 mL/min).
2
reaction mixture was concentrated to dryness, redissolved in H O, and
NMR Methods. 1H NMR spectra were obtained on either a
extracted with ethyl ether (3 × 5 mL). The adduct in the aqueous
layer was then extracted with ethyl acetate (3 × 5 mL) and isolated by
reversed-phase HPLC. The yield was in the range of 10-30%. Data
for dG-C8-FABP: HPLC (system 3) t 16.80 min; UV λ 305 nm;
DPX400 or AM300 Bruker NMR spectrometer, operating at 400 and
00 MHz, respectively, with tetramethylsilane (TMS) as the internal
3
standard. Chemical shifts for oligomer samples are reported in parts
per million downfield from internal 2,2-dimethyl-2-silapentane-5-
sulfonate, sodium salt (DSS). F NMR spectra were measured on a
DPX400 or AM300 Bruker NMR spectrometer, operating at 376.5 and
R
max
1
H NMR (400 MHz, methanol-d
.7 Hz), 7.59 (dd, 2H, H2′,6′, J2′,3′ ) J5′,6′ ) 8.7 Hz, J2′,F ) J6′,F ) 5.6
Hz), 7.50 (d, 2H, H2,6), 7.14 (dd, 2H, H3′,5′, J3′F ) J5′,F ) 8.7 Hz,
3′,5′ ) 1.9 Hz), 6.47 (dd, 1H, G1′, J1′,2′ ) 9.3 Hz, J1′,2′′ ) 6.0 Hz), 4.59
4
) δ 7.72 (d, 2H, H3,5, J2,3 ) J5,6 )
1
9
8
2
3
82.4 MHz, respectively, and referenced relative to the external CFCl .
J
(
2
)
19
The available 5 mm probes did not allow proton decoupling. F spin-
lattice relaxation times (T ) were determined on a Bruker AM300
instrument by using the inversion-recovery method. Two-dimensional
exchange experiments were carried out in the phase-sensitive mode
using a NOESY pulse sequence. Detailed NMR parameters are
included in the figure captions.
dd, 1H, G3′, J2′,3′ ) 6.6 Hz, J3′,4′ ) 1.5 Hz), 4.03 (d, 1H, G4′, J4′,5′
)
1
.3 Hz), 3.92 (m, 2H, G5′,5′′), 2.71 (m, 1H, G2′, J1′,2′ ) 9.3 Hz, J2′,2′′
-13.4 Hz, J2′,3′ ) 6.6 Hz), 2.17 (m, 1H, G2′′, J2′′,3′ ) 1.7 Hz); 19F
NMR (methanol-d
FAF: HPLC (system 3) t
400 MHz, methanol-d ) δ 7.96 (s, 1H, H1), 7.67 (dd, 1H, H5, J5,6
.3 Hz, J5,F ) 5.0 Hz), 7.64 (d, 1H, H4, J3,4 ) 8.3 Hz), 7.55 (d, 1H,
H3), 7.25 (d, 1H, H8, J8,F ) 9.1 Hz), 7.05 (dd, 1H, H6, J5,6 ) 8.3 Hz,
6,F ) 9.3 Hz, J6,8 ) 2.3 Hz), 6.47 (dd, 1H, G1′, J1′,2′ ) 9.1 Hz, J1′,2′′
6.0 Hz), 4.59 (dd, 1H, G3′, J2′,3′ ) 6.4 Hz, J3′,4′ ) 1.6 Hz), 4.03 (d,
H, G4′), 3.93 (m, 2H, G5′,5′′), 3.87 (s, 2H, H9), 2.72 (m, 1H, G2′,
4
) δ -119.56 ppm (multiplet). Data for dG-C8-
2
9
1
R
16.93 min; UV λmax 300-335 nm; H NMR
(
8
4
)
For NMR experiments, the FABP-modified 12-mer (d[CTTCTTG-
(
1
FABP)ACCTC]) was dissolved in a pH 7.0 H
2
O buffer containing
00 mM NaCl, 10 mM sodium phosphate, and 100 µM tetrasodium
J
)
1
J
EDTA, and annealed with an equimolar amount of the complementary
strand (d[GAGGTCAAGAAG]) to form a 0.6 mM solution of I. A
2
.5 mM solution of II was prepared in a similar manner.
Line Shape Analysis. Theoretical line shapes for a two-site
1′,2′ ) 9.1 Hz, J2′,2′′ ) -13.4 Hz, J2′,3′ ) 6.4 Hz), 2.17 (m, 1H, G2′′,
19
J
2′′,3′ ) 1.7 Hz); F NMR (methanol-d
4
) δ -120.03 ppm (multiplet).
exchange system with different populations were generated by using a
Windows program (WinDNMR: Dynamic NMR Spectra for Windows,
Version 1.30, J. Chem. Educ. Software Series D, Volume 3, Number
The identities of dG-C8-FABP and dG-C8-FAF were also confirmed
by electrospray mass spectrometry. The spectrum of dG-C8-FABP
+
showed [M + H] at m/z 453.3 and a fragment ion at m/z 337.1, which
2
; Reich, H. J., Department of Chemistry, University of Wisconsin,
3
4,35
is consistent with loss of the deoxyribose moiety. Likewise, the
Madison, WI) and fitted with experimental spectra.
spectrum of dG-C8-FAF exhibited peaks at m/z 465.6 and 349.4, which
Optical Melting Experiments. Temperature-dependent absorption
changes for duplex samples (5 µM) at 260 nm were analyzed on a
Hitachi U-2000 spectrophotometer equipped with a Model 9000 Isotemp
refrigerated circulator (Fisher Scientific). The midpoint transition
temperature (T
from the plot of R versus T. The T
equals 0.5 where R is the fraction of strand in the double helix. At
each temperature point, R was calculated from the temperature-
dependent UV absorption curve. Thermodynamic parameters (∆H°,
∆
K
+
+
correspond to [M + H] and [MH - dR] , respectively.
2
Synthesis and Characterization of the FABP-Modified and FAF-
Modified 12-mers. The synthesis of the FABP- and FAF-modified
12-mer strand was carried out as described for the corresponding
nonfluoro ABP- and AF-modified 15-mers.24 This involved treating
d[CTTCTTGACCTC] with the N-trifluoroacetyl derivatives of FABP
and FAF prepared above. Each modified 12-mer was separated from
the unreacted 12-mer by reversed-phase HPLC using system 2.
Average yields with greater than 98% sample purity were approximately
10% and 30% for the FABP and FAF-modified 12-mers, respectively
(see Figure 3 for their UV spectra). Data for unmodified 12-mer:
m
) was obtained by the procedure of Marky and Breslauer
values are the temperature for R
m
S°, and ∆G°) were calculated using the following equations:27 log
eq ) -∆H°/2.3RT + ∆S°/2.3R and ∆G° ) ∆H° - T∆S°, where Keq
2
-6
)
2R/c(1 - R) ; c ) 5 × 10 M; R ) gas constant.
Synthesis of Reactive Carcinogens. N-Acetoxy-N-(trifluoroacetyl)-
HPLC t
mer: HPLC t
for FAF-modified 12-mer: HPLC t
R
R
17.85 min; UV λmax 268 nm. Data for FABP-modified 12-
24.68 min; UV λmax 268, 295-345 nm (shoulder). Data
24.36 min; UV λmax 268, 295-
4
′-fluoro-4-aminobiphenyl was prepared starting from 4′-fluoro-4-
R
2
2
nitrobiphenyl by the method of Lee and King: HPLC (system 1) t
3
3
CH
R
1
.49 min; H NMR (400 MHz, CDCl
3
) δ 8.4-7.2 (m, 8H), 2.27 (s,
350 nm (shoulder). The modified oligonucleotides were further
+
+
24
H, -OCH
3
); UV λmax 268 nm; LRMS m/z 341 (M , 16), 299 (M -
characterized by a standard enzymatic degradation analysis. The
+
+
2
CO, 27), 282 ([M - CH
3
+
CO
2
], 16), 213 ([M - CH
3
CO
2
- CF
3
],
presence of the nonpolar monomeric dG-C8-FABP or dG-C8-FAF
adduct in the enzymatic hydrolysate mixture was confirmed by
comparison of their HPLC retention times and UV spectra with those
of authentic standards.
+
+
2
4), 185 (ArN , 45), 171 (Ar , 27), 43 (CH
3
CO , 100); HRMS calcd
for C16
fluoro-4-nitrobiphenyl was synthesized by nitration of 4-fluorobiphe-
nyl: 1.72 g (10 mmol) of 4-fluorobiphenyl in 100 mL of CH Cl was
treated with excess N (20 mmol) in CH Cl at room temperature.
After 24 h, 10 mmol of N was added and the mixture was stirred
for an additional 24 h. The excess N was evaporated, and the residue
was dissolved in ethyl ether, washed with H O, and recrystallized in
5% ethanol to give long white needles: 1.94 g (89%); mp 124.5-
3 4
H11NO F 341.0675, found 341.0672. The nitro precursor 4′-
2
2
2
O
4
2
2
Acknowledgment. Dedicated to Professor Elie Abushanab
on the occasion of his 60th birthday. We thank Drs. F. A. Beland
and M. M. Marques for helpful comments and Dr. Paul Vouros
for providing electrospray mass spectra of the monomer adducts.
This work was supported by a grant from the American Cancer
Society (CN-130).
2
O
4
2 4
O
2
9
1
(
3
7
1
25.5 °C (lit. mp 120-121 °C); H NMR (300 MHz, CDCl ) δ 8.31
3
d, 2H, H3,5, J2,3 ) 8.9 Hz), 7.70 (d, 2H, H2,6), 7.61 (dd, 2H, H2′,6′,
2′,3′ ) 8.8 Hz, J2′,F ) 5.1 Hz), 7.20 (dd, 2H, H3′,5′, J3′,F ) 9.7 Hz);
J
UV λmax 310 nm. N-Acetoxy-N-(trifluoroacetyl)-7-fluoro-2-aminof-
luorene was synthesized in a similar manner from 2-fluoro-7-nitro-
Supporting Information Available: 1H NMR spectra of
the monomer adducts (dG-C8-FABP and dG-C8-FAF), optical
melting profiles of the three 12-mer duplexes (unmodified, I
1
fluorene (Aldrich Co.): HPLC (system 1) t
MHz, CDCl
OCH
R
3.69 min; H NMR (400
3
) δ 7.8-7.1 (m, 6H), 3.95 (s, 2H, H9′,9′′), 2.25 (s, 3H,
1
9
+
+
and II), F NMR spectra of II in 10% and 90% D O, and a
-
3
); UV λmax 275 nm; LRMS m/z 353 (M , 36), 311 (M - CH
2
-
2
+
+
19
CO, 32), 294 ([M - CH
3
CO
2
+
], 39), 225 ([M - CH
3
CO
2
- CF
3
],
plot of F NMR chemical shifts of II as a function of
+
+
1
00), 197 (ArN , 66), 183 (Ar , 56), 43 (CH
3
CO , 87); HRMS calcd
temperature (2 pages). See any current masthead page for
ordering and Internet access instructions.
for C17
3 4
H11NO F 353.0675, found 353.0675.
(37) Marler, E. E. J; Turner, E. E. J. Chem. Soc. 1931, 1359-1363.
JA9632771