Solution Conformation of the N-(Deoxyguanosin-8-yl)aminofluorene

Apr 1, 1995 - Jean Gauvin, Suse Broyde, and Robert Shapiro. Chemical ... Robert Shapiro, Stephen Ellis, Brian E. Hingerty, and Suse Broyde. Chemical ...
0 downloads 0 Views 1MB Size
Chem. Res. Toxicol. 1995,8, 369-378

369

Solution Conformation of the N-(Deoxyguanosin-8-yl)aminofluorene Adduct opposite Deoxyinosine and Deoxyguanosine in DNA by NMR and Computational Characterization Perlette Abuaf,tst Brian E. Hingerty,§ Suse Broyde," and Dezider Grunberger*lt Department of Biochemistry and Molecular Biophysics, College of Physicians and Surgeons, Columbia University, New York, New York 10032, Health Sciences Research Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, and Biology Department, New York University, New York, New York 10003 Received August 24, 1994@

Two-dimensional proton NMR and energy minimization computations have been employed to characterize the conformations of the N-(deoxyguanosin-8-yl)aminofluorene adduct [(AF)Gl positioned opposite deoxyguanosine in one, and opposite deoxyinosine in another DNA undecamer duplex in aqueous solution. The two oligomer duplexes used in this study are d[Cl-C2-A3-T4-C5-(AF)G6-C7-TS-A9-C 1O-C11l.[G12-G13-T14-A15-G16-X17-G18-A19-T20-G2lG221, where X17 was deoxyinosine in one duplex and deoxyguanosine in another. The exchangeable and nonexchangeable protons of the DNA are well resolved and narrow in the NMR spectra of the duplexes, and the base and sugar nucleic acid protons were assigned by NOESY and COSY data sets. All nine of the nonexchangeable aminofluorene ring protons were also assigned for the duplex that has deoxyinosine across from the modification site, and the (AF)GI structure was employed to model the (AF1G.G one. The NOE distance restraints establish t h a t the glycosidic torsion angle a t (AF')G6 is syn. All other glycosidic torsion angles are anti, Watson-Crick type A*T and G C base pairing is intact throughout the duplex except at the site of modification, and the helix maintains a n overall B-DNA conformation. The syn orientation at the (AF)G6 places the aminofluorene ring in the B-DNA minor groove in a conformation similar to t h a t found previously when the (AF)G was positioned opposite deoxyadenosine [Norman e t al. (1989) Biochemistry 28, 7462-74761.

Introduction The carcinogen N-acetoxy-2-(acetylamino)fluorenebinds with high specificity to the carbon-8 position of deoxyguanosine in DNA (1). This N-(deoxyguanosin-8-y1)-2(acety1amino)fluorene adduct is deacetylated by cellular enzymes to the N-(deoxyguanosin-8-yl)aminofluorene adduct (AF)G,' which is the major and most persistent adduct in mammalian systems (2-6). The mutagenic specificity of the (AF)G adduct has been studied by a number of groups (7-15). In bacterial studies using a modified plasmid, mainly point substitutions, particularly GC TA transversions, were reported (8). Other workers in bacterial systems have encountered additional types of mutations, including other substitutions, single base deletions, and a variety of frameshifts (9-12). Carothers et al. (13,14)using mammalian cells have found that point mutations, particularly GC TA transversions, predominate, but other kinds of point mutations are also observed, including GC CG transversions. These studies suggest that the structure of the adduct and its surrounding sequence is an important factor in

-

-

* To whom correspondence should be addressed.

-

Columbia University. Present address: Immunobiology Research Institute, Route 22 East, P.O.B. 999, Annandale, NJ 08801. Oak Ridge National Laboratory. 'I New York University. @Abstractpublished in Advance ACS Abstracts, March 1, 1995. l Abbreviations: (AF)G,N-(deoxveuanosin-8-yl)aminofluorene; AAF, __ (acety1amino)fluorene. +

*

mutagenesis. A number of experiment-based high resolution views have recently been provided of the structure of AF-modified DNA. Pate1 and co-workers, together with our group, have used high resolution NMR with potential energy calculations, to define a double-helic a l undecamer containing a deoxyguanosine-AFto deoxyadenosine mismatch (15). The G-A mismatch was selected for analysis in that work because of the predominance of GC TA transversions (OA mismatch) in the biological data. In this structure the modified deoxyguanosine is syn and the AF resides in the B-DNA minor groove. NMR studies on a n AF-modified 15-mer ras protooncogene sequence with C opposite the modification site revealed a mixture of conformers, a major one with anti modified deoxyguanosine and AF in the major groove, and a second one which manifested AF base stacking (16). Two conformers of these types have also recently been modeled from NMR data in an AF-modified c-H-rasl protooncogene sequence, again with C opposite the AF(G) (17). Interestingly, these two kinds of conformers had been predicted for AF-modified G.c duplexes in our early computations (18). These computations and those of Lipkowitz and co-workers (19) had also shown that both syn and anti domains are accessible to AFmodified dG. In the present work, we investigated the solution structure of the (AF)G.G mismatch in an undecamer duplex since the biological data in mammalian cells (13, 14) had indicated that this mismatch occurs as well. In the (AF')G-G structure the 2-amino protons in the partner G gave wide peaks which interfered with assignment of

-

0893-228x/95/2708-0369$09.00/0 0 1995 American Chemical Society

3'70 Chem. Res. Toxicol., Vol. 8, No. 3, 1995

Abuaf et al.

extracted with water saturated diethyl ether. The crude AAFmodified oligomer was purified by C4 reverse phase HPLC. Next the pure AAF-modified oligomer was treated with 1 M NaOH E 9 containing 0.3% (v/v) 2-mercaptoethanol a t a concentration of 2 mg/mL at room temperature for 45 min and then neutralized with 1M hydrochloric acid. The crude AF-modified oligonucleotide was purged with nitrogen, desalted on a Sephadex G-25 column, and then purified by HPLC using a linear gradient of acetonitrile against 10 mM triethylammonium acetate buffer (pH 7.1) (0-25% in 30 min). The pure AF-modified oligomer was desalted on a Sephadex G-25 column and converted to 5' 3' sodium form on Dowex 50 x 8 cation exchange resin. The I stoichiometric ratio of the modified strand to the partner strand Ck G22 in the duplex was attained by titrating the AF-modified strand C2 G21 with its complementary strand as followed by proton NMR at A3 T20 elevated temperatures. T4 A19 NMR Sample Preparation. The NMR spectra of the C5 G1S duplexes were recorded in 0.1 M NaC1-10 mM phosphate-1 mM EDTA solutions of 99.96% DzO for DzO studies and in 90% AF G6 X17 X = G or I Hz0/10% DzO for HzO studies. The concentration of the samples C7 G16 was about 500 A260 units per 0.4 mL of solution. TS A15 Proton NMR Experiments. One- and two-dimensional A9 T14 proton NMR spectra were recorded on Bruker AM400 and C10 G13 AM500 spectrometers at one temperature, 18 "C. Twodimensional phase-sensitive NOESY spectra in DzO were colG12 lected with a sweep width equivalent to 10 ppm, repetition time I I 3' 5' of 2 s, and a single mixing time of 250 ms. The carrier frequency was placed on the residual HOD resonance, and the decoupler Figure 1. Adduct structure and sequence context investigated; channel was used to suppress the residual HOD signal. The dR represents deoxyribose. data sets were collected with 512 tl experiments using 1024 complex data points in the tz dimension and 32 scans per tl the AF protons. Therefore, we also synthesized and increment. Two-dimensional data sets were processed using the elucidated the structure of the duplex having a deoxyprogram FTNMR from D. Hare. The data sets were syminosine opposite (AF)G,which has a proton instead of an metrized prior to plotting. amino group a t carbon-2. In this duplex, the deoxyPhase-sensitive two-dimensional proton NOESY spectra in inosine carbon-2 proton aided in identifying all the HzO were collected using a n optimized jump and return pulse nonexchangeable aminofluorene ring protons. An energy sequence, with a mixing time of 120 ms and the carrier minimized structure of the (AF)GI duplex was computed frequency placed directly on the HzO. The time domain data sets were accumulated over a sweep width of 8064 Hz, using which then served as a model for the (AF)G-G one. 1024 complex data points, 512 tl experiments with 1s repetition Figure 1 shows the adduct structure and sequence condelay and 256 scans per increment. The free induction decays text investigated. Our results demonstrate that the were apodized with a 90" shifted sine bell function in both tl (AF)G-G and (AF)G-Iadducts are similar to the (AF)G-A and tz dimensions. Each dimension was base line corrected with one, with the modified deoxyguanosine syn and the AF a fifth-order polynomial base line fitting routine. sandwiched between the walls of the B-DNA minor Two-dimensional proton correlated spectra (COSY) were groove. recorded in magnitude mode in DzO solution. The data sets were collected with 512 tz increments and a sweep width of 4000 Hz with 1024 complex data points in the t 2 dimenMaterials and Methods sion. A repetition delay of 1.0 s and 32 scans per tl increment Caution:N-Acetoxy-2-(acetylamino)fluorene is a carcinogen. were used. Energy Minimization Computations. Minimized potenOligonucleotide Synthesis. The oligonucleotides d(Ctial energy calculations were carried out with DUPLEX, a CATCGCTACC), d(GGTAGGGATGG1, and d(GGTAG1GATGG) molecular mechanics program for nucleic acids that performs were synthesized on a Beckman System 1+ automated DNA potential energy minimizations in the reduced-variable domain synthesizer using solid-phase cyanoethyl phosphoramidite chemof torsion angle space (20). The advantage of torsion space, istry. The crude, 5'-dimethoxytritylated oligonucleotides were compared to Cartesian space, minimizations is the vast diminufirst deprotected and cleaved from the support by treatment with tion in the number of variables that must be simultaneously concentrated ammonium hydroxide for 36 h at room temperaoptimized, thereby permitting larger movements from a given ture, and then purified by HPLC on a semipreparative reverse starting conformation during minimization, as well as assurance phase C4 column (Rainin Dynamax C4 reverse phase, 300 A of realistic internal geometry. pore size, 12 pm particle size, length 25 cm, internal diameter 21.4 mm). Next the oligonucleotides were treated with acetic DUPLEX uses a potential set similar to one developed by acid and then repurified by HPLC using a linear gradient of Olson and co-workers for nucleic acids (211, and details have CH3CN against 0.1 M triethylammonium acetate buffer (pH 7.1) been published previously (20).Force field parameters were the (0-30% in 30 min). The oligomers were finally converted to same as those employed in the (AF)G-A study (15) except for the sodium form by treating them on a Dowex cation exchange those involving deoxyinosine. Bond lengths, bond angles, and resin. dihedral angles for this base were taken to be the same as for deoxyguanosine except that the NH2 group at carbon-2 was Preparation of Aminofluorene Adduct. The aminofluoreplaced with a hydrogen. Partial charges for deoxyinosine were rene adduct d[C-C-A-T-C-(AF)G-C-T-A-C-C]was prepared in two taken from the work of Ornstein and Fresco (22) and are steps. First the oligonucleotide d[C-C-A-T-C-G-C-T-A-C-Cl, consistent with the rest of the partial charge set employed in dissolved in 2 mM sodium citrate buffer, pH 7.1, was treated DUPLEX. with N-acetoxy-2-(acetylamino)fluorenedissolved in absolute ethanol (10 mg/mL) in a 1:8 molar ratio. The reaction mixture The DUPLEX hydrogen bond penalty function (20) was was vortexed and kept in the dark at 37 "C for 3 h and then employed in all first-stage minimizations to aid the minimizer

G

AF

!

-

-

-

c11 -

Chem. Res. Toxicol., Vol. 8, No. 3, 1995 371

2-Aminofluorene-DNA Adduct Solution Conformation

1

I

14.5

I

14.0

I

I

13.5

13.0

12.5

I

1

I

1

I

14.5

14.0

13.5

13.0

12.5

12.0 PPM

11.5

,

11.0

I

I

10.5

10.0

9.5

I

1

I

I

I

I

12.0

11.5

11.0

10.5

10.0

9.5

PPM

Figure 2. (A) Proton NMR spectrum (9.5-15.0 ppm) of the (AF'IG 11-mer duplex with G opposite the modification site, in HzO buffer, at 18 "C. Assignments are given in Table 1. (B) Proton NMR spectrum (9.5-15.0 ppm) of the (AF')G 11-mer duplex with I opposite the modification site, in HzO buffer, a t 18 "C. Assignments are given in Table 2. in locating the Watson-Crick hydrogen-bonded structures indicated by the NMR data, or as a strategy for orienting the (AF')G-I or (AF')G.G mismatched pair. To locate minimum energy conformations with interproton distances available from the experimental N M R data, pseudopotentials (permitting upper and lower bound restraints) are added to the energy, as described previously (15, 23-26). Briefly, the following functions can be employed:

The Ws are adjustable weights (30 kcal/(mol&) in the present work), d is the current value of the interproton distance, d~ is a target upper bound, and d m is a target lower bound. The functions are summed over all n experimental bounds. Equation l can be implemented when d is greater than dN, and eq 2 can be implemented when d is less than d m , if upper and lower bounds are employed. The present work usually employed only upper bounds, as did the earlier (AF1G.A study (15). However, lower bounds were used for selected proton pairs when initial trials produced structures with distances that were closer than the data suggested. All penalty functions were released in the last minimization steps to yield unrestrained final structures that are energy minima. Computations were carried out a t the Department of Energy's National Energy Research Supercomputer Center and the National Science Foundation's San Diego Supercomputer Center.

Results Exchangeable Nucleic Acid Protons. The proton spectra of the imino region of the (AF)G opposite G, and the (AF)G opposite I 11-mer duplexes in HzO are plotted in Figure 2. The (AF)G.G 11-mer duplex shows 12 partially resolved exchangeable protons in the 10- 14 ppm region. In the case of the (AF1G.I 11-mer duplex one imino proton a t about 10.6 ppm, and in the case of the (AF)W11-mer duplex two imino protons at about 10.5 and 11.2 ppm are shifted upfield from the rest of the imino protons resonating between 12-14 ppm. The imino protons have been assigned from an analysis of the NOESY spectrum (125 ms mixing time) of the (AF)G.I 11-mer duplex in HzO buffer (pH 6.95) a t 18 "C. The expanded NOESY contour plot establishing distance connectivities in the symmetrical imino-imino region is plotted in Figure 3A, and the 10.4-14 and 4.8-8.6 ppm regions are plotted in Figure 3B. The four deoxythymidine imino protons (13.3-13.6 ppm) exhibit NOEs to the deoxyadenosine H2 protons within individual A.T pairs. The nonterminal deoxyguanosine imino protons (12.212.8 ppm) exhibit NOEs to hydrogen-bonded and exposed deoxycytidine amino protons within individual G-Cpairs, while the terminal deoxyguanosine imino protons exhibit NOE crosspeaks only to the HzO resonance. Therefore, unambiguous assignment of the terminal deoxyguanosine imino protons was not possible. The deoxycytidine H5 assignment is established from the analysis of the COSY and NOESY spectra of the duplex recorded in DzO solution. The deoxycytidine H5 proton, in turn, shows

372 Chem. Res. Toxicol., Vol. 8, No. 3, 1995

Abuaf et al. Table 1. Exchangeable Proton Chemical Shifts in the CGC(AF)/GGG11-mer in 0.1 M NaCl-10 mM Phosphate-H2O, pH 6.92, 18 "Ca

A

chemical shifts (ppm) basepair ClG22 c24321 A3*T20 T4sA19 C54318 (AF)G6G17

e

C7G16 T8sA15 A9*T14 ClW13 Cll4312 a

G-N1H

T-N3H C-N4Hb C-N4He A-H2

NA 12.81

7.77 8.44

6.83 6.78

13.49 13.42 12.28 10.52 (G6) 11.23 (G17) 12.64

7.73 7.64 8.03

6.60

8.01

6.58

13.58 13.53 12.71 NA

7.41 7.46 8.13 7.79

6.59 6.83

NA not assigned; Hb: hydrogen bonded; He: exposed.

Table 2. Exchangeable Proton Chemical Shifts in the CGC(AF)/GIG 11-mer in 0.1 M NaC1-10 mM Phosphate-H20, pH 6.96, 18 "Ca 13:6

l2:8

1210

11:2

chemical shifts (ppm)

10:4

base pair

PPM

ClG22 C2G21 A3eT20 T4sA19 C5G18 (AF)G6*117 (D

Ln

C74316 T8eA15 A9T14 C1W13 Cll4212

G-N1H I-N1H T-N3H C-N4Hb C-N4He A-H2 NA 12.79

7.92 8.46

6.81 6.82

13.54 13.36 12.21 10.46 (G6) NA (117) 12.50

7.72 7.65 8.21

6.71 7.82

7.96

6.43

13.57 13.54 12.72 NA

7.38 7.49 8.14 8.04

6.61 6.90

NA: not assigned; Hb: hydrogen bonded; He: exposed.

1316

12:8

1210

11:2

10:4

PPM Figure 3. Phase-sensitive NOESY contour plots of the (AF')GI 11-mer duplex in HzO buffer, pH 6.95 at 18 "C. (A) Crosspeaks establishing distance connectivities in the symmetrical 10.414.0 ppm imino proton range. Thymine protons are N3H and guanine protons are N1H. (B)Crosspeaks establishing distance connectivities between the imino protons (10.4-14.0 ppm) and the amino protons (4.8-8.4 ppm): peak 1,C5 N4 exposedG18 N1H; peak 2, C5 N4 bondedG18 N1H; peak 3, C7 N4 exposed G16 N1H; peak 4, C7 N4 bondedGl6 N1H; peak 5, C10 N4 exposedG13 N1H; peak 6, C10 N4 bondedG13 N1H; peak 7, C2 N4 exposedG21 N1H; peak 8, C2 N4 bondedG21 N l H , peak 9, A19 H2/T4 N3H; peak 10, A15 H2/"8 N3H; peak 11,A9 H2/ T14 N3H; peak 12, A3 H2PT20 N3H.

NOES to the hydrogen-bonded and exposed deoxycytidine amino protons. These, in turn, exhibit NOES to the deoxyguanosine imino proton across the GC pair. The deoxyguanosine imino protons were assigned in this way. The deoxyadenosine H2 protons were established from an inversion-recovery experiment. The assigned deoxyguanosine imino protons of G.c pairs were correlated on the basis of observed NOES to the deoxythymidine imino (Figure 3A) and deoxyadenosine H2 (Figure 3B) of

adjacent A-T pairs in the duplex. The deoxyguanosine and deoxythymidine imino protons, the deoxycytidine hydrogen-bonded and exposed amino protons, and the deoxyadenosine H2 proton assignments in the (AF)G-G and (AF)G-I 11-mer duplexes are listed in Tables 1 and 2, respectively. The remaining deoxyguanosine imino proton a t 10.64 ppm for the AF(G).I duplex is assigned t o (AFIG6. This assignment suggests that the (AF)G6 imino proton exhibits a strong NOE to the HzO resonance (Figure 3B),indicating that it is accessible to solvent. Exchangeable Aminofluorene Proton. There is a single exchangeable NH proton on the aminofluorene ring which is linked to the C8 position of G6. However, we have not been able to locate this aminofluorene exchangeable proton either in the (AF)GI 11-mer-duplex or in the (AF)G-G 11-mer duplex. NonexchangeableNucleic Acid Protons. The 5.08.5 ppm region of the nonexchangeable proton spectra of the (AF)GG and (AF)GI 11-mer duplexes shown in Figure 4, panels A and B, depicts partially resolved base and sugar H1' protons. This region was further investigated by two-dimensional NOESY and COSY experiments. The phase-sensitive NOESY spectra (250 ms mixing time) of the (AF)G.G and (AF)G-I 11-mer duplexes in DzO buffer exhibit resolved crosspeaks such that NOE distance connectivities can be systematically utilized to yield assignments for nucleic acid base protons and a majority of the sugar protons, as well as several aminofluorene protons. The expanded NOESY plots establishing distance connectivities between base protons (6.88.5 p?m) and the sugar H1' and deoxycytidine H5 protons (5.0-6.5 ppm) in the (AF)G-G and (AF)OI 11-mer du-

Chem. Res. Toxicol., Vol. 8, No. 3, 1995 373

2-Aminofluorene-DNA Adduct Solution Conformation A

I

A

A L I -

0

a

OQ 0

8.5

8.0

7.5

7.0

6.5

6.0

5.5

5.0

5.5

5.0

00

b.

PPM B

1

6.5

6.0

7.5

7.0

6.5

6.0

PPM

Figure 4. Proton NMR spectrum of the base and sugar H1' proton region (4.8-8.5 ppm) of the (A) (AF)G.G 11-mer duplex in DzO buffer at 25 "C; (B) (AF1G.I 11-mer duplex in D2O buffer a t 23 "C.

plexes are plotted in duplicate in Figures 5 and 6, respectively. Each base proton exhibits a n NOE crosspeak to its own sugar H1' and to the sugar H1' of the 5' linked base (27). This allows the chains to be traced from C1 to C5 and from C7 to C11 in the modified strands (Figures 5A and 6A) and from G12 to G22 in the unmodified strands (Figures 5B and 6B) of the (AF1G.G and (AF)G.I 11-mer duplexes, respectively. The tracing is interrupted in the modified strands a t the (AF)G6 because that base lacks a proton a t position 8 since the AF is covalently attached to carbon-8. This directionality of the NOEs is observed between the base protons and their own and 5'-flanking sugar H3' protons and sugar H2', H2" protons, permitting partial assignments of the sugar protons. The sugar proton assignments were also verified independently from the COSY spectra. An expanded magnitude COSY contour plot establishing coupling connectivities between the 4.2-6.4 ppm H1' and H3' region and the 1.8-3.8 ppm H2', 2" region of the (AF1G.I 11-mer duplex is plotted in Figure 7. A large chemical shift difference is observed between the H2' and H 2 protons of the (AF)G6residue, similar to the (AF1G.A 11-mer duplex case (15).The coupling pattern of the H2', 2" is also inverted for the (AF)G6 residue. These two observations indicate that the H2' proton of the (AFIG6 residue has undergone a downfield shift in all of the AFmodified duplexes. The deoxyadenosine H2 protons have long spin-lattice relaxation times. As a result, they are easily identified by inversion-recovery experiments. The deoxyinosine H2 protons behave the same way as the deoxyadenosine H2 protons. As a consequence, the (AF)GI 11-mer duplex shows five protons in the inversion-recovery experiment. Four of those five protons belong to deoxyadenosine H2 involved in the A-T pairs; these have been assigned from the observed NOES between deoxythymidine imino protons and deoxyadenosine H2 protons (Figure 3B). The base and sugar Hl', H2', 2", H3' nonexchangeable proton chemical shifts in the (AF)GG 11-mer duplex and (AF)GI 11-mer duplex are listed in Tables 3 and 4. In the case of the (AF)GI 11-mer duplex all of the nonexchangeable protons of the aminofluorene moiety are assigned, and they are listed in Table 4. It should be noted that the H2 proton of I17

v

A3 r

I

.4

8.0

I

7.6

I

7.2

PPM

f

Q

I P (D

8.4

8.0

7.6

7.2

PPM

Figure 5. Phase-sensitive expanded NOESY contour plots (mixing time 250 ms) of the (AF)G.G 11-mer duplex in D2O buffer, establishing distance connectivities between the base protons (6.8-8.4 ppm) and the sugar H1' and deoxycytidine H5 protons (4.8-6.4 ppm). The chain is traced from C1 to C11 in (A), and from G12 to G22 in (B). The deoxycytidine H5-H6 connectivities are marked by asterisks.

(7.82 ppm) at the lesion site is the most downfield of the H2 protons in the (AF)G.I 11-mer duplex. It should also be noted that the H8 protons of G16 and G18 flanking the lesion site are shifted upfield compared to the H8 protons of the other deoxyguanosines,in both the (AFM and (AF)GI 11-mer duplexes. Nonexchangeable Aminofluorene Protons. The aminofluorene protons of the (AF1G.I 11-mer duplex were assigned from an analysis of COSY and NOESY crosspeaks. The 7.90 ppm H2 proton of deoxyinosine 17 shows weak NOE crosspeaks to the 3.49 and 3.59 ppm geminal aminofluorene H9 protons, strong NOE to the 7.20 ppm aminofluorene H1 proton, and moderate to weak NOEs to the 6.87 and 7.32 ppm aminofluorene H3 and H4 protons, respectively. Figure S-1(supplementary mate-

374 Chem. Res. Toxicol., Vol. 8, No. 3, 1995

Abuaf et al.

A

‘ 1

1

CD

-tiE a

0 0

LD

0

pr)

CD

5

6.0

5.5

5.0 PPM

4.5

4.0

3.5

Figure 7. Expanded COSY contour plot of the (AF)G.I 11-mer

1

I

8.0

8.4

I

duplex establishing coupling connectivities between H1’ and H3’ protons (3.5-6.5 ppm) and the H2’,2” protons (1.5-3.8 ppm). H2’ and H2” of designated bases are connected by a line. Where no line is drawn, the protons have the same or very close chemical shifts.

7.2

7.6 PPM

B

Table 3. lH Chemical Shifts (ppm) for d[CCATC (AI?-G)CTACCl*[GGTAGGGATGGl spins H8 H2

c1

10 0

I

A15

G22 I

4

A

i

0 .4

1

T

0

4

I

8.0

I

7.6

7.2

PPM Figure 6. Phase-sensitive expanded NOESY contour plots (mixing time 250 ms) of the (AF)GI 11-mer duplex in DzO buffer, establishing distance connectivities between the base protons (6.8-8.4 ppm) and the sugar H1’ and deoxycytidine H5 protons (4.8-6.4 ppm). The chain is traced from C1 to C11 in (A), and from G12 to G22 in (B). The deoxycytidine H5-H6 connectivities are marked by asterisks.

rial) shows the distance connectivities between the 6.81 8.6 ppm and 3.014.8 ppm regions. Figure S-2 (supplementary material) shows the distance connectivities between the 6.8518.6 ppm and 6.48.2 ppm regions. Figure S-3 (supplementary material) shows the distance connectivities between the 6.8518.6 ppm and 4.416.2 ppm regions. The 6.87 ppm aminofluorene H3 proton exhibits weak NOEs t o the sugar H1’ protons of C7 and (AF)G6. The sugar H1’ protons are in the minor groove, and these NOEs help align the aminofluorene in the minor groove. The 7.32 and 7.35 ppm aminofluorene H4 and H5 protons, respectively, exhibit weak NOEs to the sugar H1’ protons of I17 and G18 on the partner strand. The 7.08 ppm resonance assigned to the aminofluorene H6 proton exhibits weak NOEs to the sugar H1’ protons of

H6 H5lCH3 H1’ H2’/H2” H3’ H4‘ H5’15”

7.70 c2 7.57 A3 8.36 7.81 T4 7.18 c5 7.39 G6* c7 7.35 T8 7.28 A9 8.30 7.55 c10 7.32 c11 7.57 G12 7.70 G13 7.82 T14 7.28 A15 8.15 7.48 G16 7.31 G17 7.80 G18 7.37 A19 8.01 7.71 T20 7.02 G21 7.79 G22 7.76

5.88 5.67 1.46 5.55 5.39 1.58 5.35 5.69 1.43

1.25

5.93 5.41 6.31 5.80 5.88 5.83 5.57 5.61 6.19 5.91 6.17 5.68 5.98 5.53 6.01 5.44 5.49 5.23 6.12 5.68 5.64 6.12

2.0412.48 4.63 2.1412.40 4.84 2.74f2.81 5.03 2.1312.43 4.83 2.0312.37 4.76 3.5312.29 4.88 1.8012.23 4.65 2.0912.39 5.02 2.7212.85 4.71 2.0312.39 4.72 2.2212.22 4.50 2.5212.65 2.6212.76 4.95 2.0912.36 4.86 2.7012.83 5.04 2.1312.44 2.6312.44 4.90 2.3112.50 4.72 2.5612.84 4.95 1.7912.19 4.80 2.5712.66 4.86 2.34/2.48 4.63

4.02

4.12

4.15 4.00

4.24 4.16 4.13 4.12

G16 and 117, to the sugar H2” protons of T8 and G16, and to the sugar H2’ proton of the (AFIGG. The 7.03 ppm resonance is assigned to the aminofluorene H7 proton. It exhibits weak NOEs to the sugar H1’ proton of G16, to the sugar H2’ and H3’ protons of 117, and to the sugar H2” proton of G18. Glycosidic Torsion Angle. The intensities of the NOEs between bases and their own sugar H1’ protons for the syn-glycosidic torsion angle, with interproton distance of about 2.5 A, are comparable to the NOE between the H5 and H6 protons of deoxycytidine, which has a fixed distance of 2.45 A (28). In the case of antiglycosidic torsion angle with base-sugar H1’ interproton distance of about 3.7 A,these NOE intensities are weaker than that of the deoxycytidine H5 and H6 protons. This comparison cannot be used to establish the glycosidic torsion angle of the AF-modified G6 since that deoxyguanosine lacks a base proton. However, the NOE intensities between the base protons and their own sugar H1’ protons for all other residues in the (AF)G-Gand (AF’)G-I

2-Aminofluorene-DNA Adduct Solution Conformation ll-mer duplexes are weaker than the NOE between the H5 and H6 protons of deoxycytidine. This establishes that all bases including the G17 and 117, but excluding the (AF)G6, adopt a n anti-glycosidic torsion angle. Since I17 has an anti orientation, its H2 proton which resonates at 7.90 ppm is located in the minor groove. The observed NOEs between the 7.90 ppm H2 proton of I17 and 7.20 and 6.87 ppm aminofluorene H1 and H3 protons establish that the aminofluorene ring is located in the minor groove of the (AF)GI ll-mer duplex. This can only occur if the (AF)G6 adopts a syn torsion angle about its glycosidic bond. Energy Minimization Computations and Modeled Structures. The starting conformation for the energy minimizations was the structure of the duplex 11mer with A opposite the (AF)G (151, since the data showed that the (AF)GI structure was very similar to the (AF)GA one. The DUPLEX hydrogen bond penalty function (20)for Watson-Crick base pairing was utilized at all base pairs except the modified G, since the data had indicated Watson-Crick pairing a t all sites other than the (AF)G. At this locus a single hydrogen bond, between 0 6 of the modified G and H l N l of I, was maintained during the early restrained trials, but was released in later stages. This strategy aids in positioning the G I pair to orient the (AF) in the B-DNA minor groove. The NOE-based distance restraints listed in Table 5 were included in these calculations. All restraints were released in terminal minimizations. The final structure with I opposite modified G was employed as the starting conformation in a minimization for the (AFIGG duplex. In this two-stage minimization the hydrogen bond penalty function was again used in the first stage in the same manner as for the (AF)GI adduct, but no distance restraints were employed. The penalty function was again released in the terminal minimization. The resulting central trimers in the (AF)GI and (AF)Wstructures are shown in Figure 8 and 9. Torsion angles for the (AFIGI central trimer are given in Table 6. Very similar values were computed for the (AF1G.G structure.

Discussion

-

Because of the presence of G C C.G transversions in AF-induced mutations (13,14),we were interested in the conformation of the (AF)G adduct positioned opposite G in the center of the DNA oligomer duplex. Even though the resonances were sufficiently narrow to assign the exchangeable and nonexchangeable base protons, and the majority of the sugar protons, the aminofluorene ring protons could not be assigned. As a consequence, no NOEs between the aminofluorene protons and nucleic acid could be obtained to define the orientation of the AF with respect to the DNA. Since deoxyguanosine and deoxyinosine are identical except for the substituent at carbon-2 (deoxyguanosine has a n amino group while deoxyinosine has a proton), we adopted the strategy of first investigating the (AF)GI ll-mer duplex. In this duplex, we made use of the proton assignment at carbon-2 of deoxyinosine to assign all of the aminofluorene ring protons, and we were able to obtain NOEs between the aminofluorene ring protons and the sugar protons on the modified and complementary strand. These observed NOEs were converted into distance restraints which were used in the energy minimization computations. We then employed the (AF)G.I ll-mer duplex conformation t o model the conformation of the (AF)GGll-mer duplex.

Chem. Res. Toxicol., Vol. 8, No. 3, 1995 375 Chemical Shifts (ppm) for

Table 4.

dlCCATC(AF’-G)CTACCHGGTAGIGATGGl spins H8 H2 H6 H5lCH3 H1’ H 2 ” 2

c1 7.68 c2 7.58 A3 8.36 7.79 T4 7.18 c5 7.37 G6* c7 7.37 T8 7.31 A9 8.27 7.56 c10 7.30 c11 7.54 G12 7.82 G13 7.80 T14 7.25 A15 8.10 7.45 G16 7.38 I17 7.99 7.90 G18 7.22 A19 8.01 7.73 T20 7.02 G21 7.77 G22 7.74 H1 7.20 H5 7.35 H8 7.24

5.85 5.67 1.46 5.55 5.30 1.55 5.34 5.60 1.42

1.27

5.91 5.43 6.31 5.82 5.92 5.85 5.20 5.57 6.20 5.90 6.15 5.66 5.97 5.53 6.00 5.49 5.74 5.09 6.16 5.67 5.63 6.11

H3’ H4’ H5’/H5”

2.0512.47 4.62 2.1712.42 4.85 2.7512.96 5.04 2.1312.45 4.84 2.2212.25 4.91 3.6012.27 4.91 1.8012.19 4.64 2.0812.39 5.01 2.7112.85 2.0212.38 2.2212.22 4.50 2.5312.66 4.79 2.6112.75 4.95 2.0512.32 4.85 2.6312.79 5.02 2.301254 4.93 2.3012.25 4.86 2.2012.48 4.69 25412.86 4.96 1.7812.19 4.81 2.6112.67 4.95 2.4812.34 4.63

AF peaks H3 6.87 H6 7.09 H9/H9 3.4913.59

H4 7.32 H7 7.03

Table 5. Comparison of Experimental Interproton Distance Bounds with Distances in Final Unrestrained Energy Minimized Structure of hWG.1 11-mer Dudex atom pair

exptl bounds

achieved distance

1: 117(H2)-AF(H9) 2: 117(H2)-AF(H9’) 3: 117(H2)-AF(H3) 4: 117(H2)-AF(Hl) 5: 117(H2)-AF(H4) 6: C7(Hl’)-AF(Hl) 7: G18(Hl’)-AF(Hl) 8: C7(Hl’)-AF(H3) 9: GG(Hl’)-AF(H3) 10: 117(Hl’)-AF(H4) 11: G18(Hl’)-AF(H4) 12: 117(Hl’)-AF(H5) 13: G18(Hl’)-AF(H5) 14: GlG(Hl’)-AF(HG) 15: 117(H3’)-AF(H6)

3.5-5.5 3.5-5.5 3.0-5.0 2.5 -3.5 3.5-5.5 3.0-5.0 4.0 -6.0 3.5-5.5 3.5-5.5 4.0 -6.0 4.0 -6.0 4.0 -6.0 4.0 -6.0 3.5-5.5 4.0 -6.0

3.10 4.43 5.33 3.21 5.56 4.84 5.63 5.54 4.65 5.81 2.92 4.82 3.54 6.25 6.02

Since only one long mixing time, 250 ms, was employed, spin diffusion undoubtedly produced apparent interproton distances that are too short. We employed wide upper target bounds with this in mind. In addition, our modeling protocol, which releases restraints in terminal minimizations, has the advantage of the capability to compensate for this problem: final structures are unrestrained energy minima on the potential energy surface. Consequently, achieved interproton distances in our model that are longer than their selected target upper bound (Table 5) are plausible, in light of uncertainty in bound assignment due to spin diffusion. Spin diffusion would also account for other long model distances, such as the 7.6 A distance between H2 of I17 and H1’ of C7, for which a small crosspeak was observed (Figure 6A) (29). On the other hand, the short model distance between H1’ of G18 and H4 of AF in relation to the target bounds (Table 5) as well as the 4.6 A model distance between H2 of I17 and H1’ of G18, for which no NOE is observed, is unexplained. However, most model distances not employed as restraints are in reasonable accord with the observed NOEs, such as the 4.9 A distance between H2 and H1’ of I17 (Figure 6). Conse-

376 Chem. Res. Toxicol., Vol. 8, No. 3, 1995

Abuaf et al. A

Figure 8. Stereoviews of the [C5-(AF)G6-C7].[G16-117-G181 trinucleotide segment of the 11-mer duplex, prepared for use with a stereo viewer. (A) emphasizes the location of the AF in the minor groove and its extension toward the sugar-phosphate backbone on the partner strand; (B)emphasizes the interaction of the AF with the walls of the minor groove. \

A

\

B

Figure 9. Same as Figure 8, for [C5-(AF)G6-C7l.[Gl6-G17-G18]. Table 6. Backbone Torsion Angles, in Degrees, for the [CS-(AF)Gg-C?]~[G1B.117~18] Segment in the Energy Minimized (AF)G-I 11-me@ a B v P Y E t‘ c5 295 186 47 154 237 197 231

177 68 255 270 63 182 123 322 182 45 158 231 180 251 273 176 17 216 210 288 51 292 169 191 59 233 276 147 128 48 131 222 182 254 294 Backbone torsion angles are given in the IUPAC convention: PQ-05’p-C5’y-C4’c-C3r-03’e-P. Glycosidic torsions x are: 04’Cl’-Nl-Ca(pyr), 04‘-Cl’-N9-C4(pur). The sugar pucker is defined with the pseudorotation parameter P (33). Torsion angles defining the AF base linkage are as follows: a’: N9(G)-C8(G)N(AF)-C2(AF); /Y: C8(G)-N(AF)-CP(AF)-Cl(AF). (AF)G6 c7 G16 I17 G18

quently, our models offer structures whose overall features portray the qualitative aspects of the data, while some quantitative details remain uncertain. A recent incisive assessment of the accuracy of NMR solution structure models, which considers especially data uncertainties, demonstrates that “with present methods the

accuracy of NMR structures is at best of the order of 1to 2 A” (30).Achieved model distances that are within a n A of target values are therefore likely to be a t the level of attainable accuracy. Global Structural Features. Based on the observed NOE parameters of the exchangeable and nonexchangeable protons, several conclusions on the global structure of the (AF)G.I and (AF1G.G 11-mer duplexes can be deduced. In both duplexes the deoxythymidine and deoxyguanosine imino protons exhibit NOEs to the deoxyadenosine H2 protons and the deoxycytidine amino protons, respectively, except a t the termini; the only NOEs involving the terminal imino protons that could be assigned were those to H20, as can be seen in Figure 3B. This finding indicates that there is stable WatsonCrick type A-T and G.c base pairing throughout the helices except at the sites of modification, (AF)G64217 and (AF)G6*117,and a t the terminal base pairs, which appear to be labile. In both the (AJ?)G-G and (AF)G.I 11-mer duplexes, all of the base protons (purine H8 and pyrimidine H6),

2-Aminofluorene-DNA Adduct Solution Conformation

except for the AF modified G6 residue, exhibit NOEs to their own and 5'-flanking sugar H1' protons. This directionality indicates that both helices are righthanded, and regular from C1 to C5 and C7 to C11 on the modified strands (Figures 5A and 6A1, and from G12 to G22 on the unmodified strands (Figures 5B and 6B). However, (AF)G6 lacks a base proton a t carbon-8 since the covalently bound AF is attached to G6 a t that position. As a result, the handedness of the helix a t the modification site cannot be defined. In both duplexes, the H2' proton of the (AF)G6 residue resonates much further downfield than the rest of the sugar H2' protons. This downfield shift was earlier observed in the (AF)G.A ll-mer duplex and attributed to the proton being near the plane of the G6 ring and therefore selectively experiencing strong in plane ring current contributions (15). In both the (AF1C.G and (AF1G.I ll-mer duplexes the NOEs between the base and its own sugar H1' proton for the unmodified bases are much weaker than the NOE crosspeak between the H5 and H6 protons of deoxycytidine. This establishes that all unmodified bases possess a n anti-glycosidic torsion angle. In the case of the modified G6 residue the H8 proton is lacking due to the covalent attachment to the carcinogen a t that site. As a result, the glycosidic torsion angle a t the modified deoxyguanosine residue cannot be defined by this method. (AF)G6Modification Site. The observed NOE between the H2 proton of 117 and the H1 proton of the aminofluorene ring defines the orientation of the (AF)G6 and I17 in the (AF)G.I ll-mer duplex. Since I17 exhibits anti-glycosidic torsion angle, its H2 proton is directed toward the minor groove of the (AF1G.I ll-mer duplex. As a result, the observed NOE between the H1 proton of the aminofluorene ring and the H2 proton of the I17 residue dictates that the aminofluorene be in the minor groove, which can occur only if the G6 residue adopts a syn glycosidic torsion angle. The observed NOES between the H2 proton of I17 and the H9 and H1 protons of the aminofluorene ring indicate that the edge of the aminofluorene ring with the H1 and the geminal H9 protons is toward the interior of the helix while the edge with the H5 and H6 protons is directed toward the solvent. The imino protons of the aminofluorene-modified G6 residues in both the (AF1G.G and (AF1G.I ll-mer duplexes resonate further upfield (10.52 and 10.64 ppm, respectively) than the regular hydrogen-bonded imino protons. This upfield shift of these protons along with the fact that they exhibit a strong NOE with solvent HzO indicates that they are not hydrogen bonded and are therefore exchanging with the solvent. While it was not possible to identify the deoxyinosine H1 proton in the NMR data, the DUPLEX hydrogen bond penalty function was employed in the first-stage minimizations, to search for a hydrogen bond between (AFlG.06 and H1 of I; this is in analogy with the acidic pH structure of (AF)G.A (15) for which a n 06.*.Hl hydrogen bond, with N1 protonated adenosine, had been proposed. In our final, unrestrained structure the G-06...Nl-I distance is 3.41 A and the 06-H1-N1 angle is 148",indicating a very weak hydrogen bonding interaction in the model, like the one suggested for the acidic pH (AF)G.A structure. Acidic pH is required for N1 protonation of A in (AF)G.A, but in the (AF)G.I and (AF)G.G structures the hydrogen a t N1 is present a t neutral pH. Consequently, there was no need for em-

Chem. Res. Toxicol., VoE. 8, No. 3, 1995 377

ploying acidic pH to examine the 06*..H1 hydrogen bonding possibility in the present work. Aminofluorene Ring. The aminofluorene ring is positioned in the minor groove and does not overlap with the aromatic ring protons of the flanking C5*G18 or C7G16 base pairs, indicating that it does not stack into the helix. The aminofluorene ring protons H4, H5, H6, and H7 exhibit NOES with the sugar protons of the G16, 117, and G18 residues on the complementary strand, indicating that the aminofluorene ring interacts with the partner strand and spans the minor groove of the helix. Thus, a t the central trinucleotide segment of the duplex the aminofluorene ring is sandwiched between the walls of the minor groove of the helix, and the W I G 6 residue assumes a syn-glycosidictorsion angle, forming an AFGG[synI.I17[anti] pair. This AFG6[synI.I17[antil alignment does not disrupt the helix nor the flanking base pairs since the C5G18 and C7G16 base pairs are intact. These features are similar to those of the (AF)GA duplex investigated earlier (151, as shown in Figure 8. All of the oligonucleotide-related NOES that define the overall structure of the helix are observed in both the (AF1C.G and (AF1G.I ll-mer duplexes. Assuming that the AF orientation is the same on changing the base across the modification site from deoxyinosine to deoxyguanosine, we compute the energy minimized structure shown in Figure 9 for the (AF1G.G duplex. Torsion Angles. All DNA torsion angles in the energy minimized conformation of the (AF1G.I ll-mer duplex were within the range of average values observed in the B-DNA dodecanucleotide crystal structure (31) except for those underlined in Table 6. Differences are the same as in the (AF)GA ll-mer (15). The deoxyguanosine is syn a t the (AF)G6 modification site. The sugar ring pucker of G16 is C3' endo on the partner strand. The backbone conformation of I17 has changed from the more common BI conformation ( E , 5 = t, g-) present in the rest of the structure to the BII conformation ( E , 5 = g-, t). These two alternate B-DNA conformers have been observed to coexist in a number of B-DNA crystal structures (32). The values of the a', p' pair that govern the alignment of the AF in relation to its covalently bound deoxyguanosine (Table 6) adopt the values of 206", 27" in (AF1G.I while a', p', are 208", -43" (which is the same as 317") in (AF)GA. These are both in the domain with C8(G) and Cl(AF) approximately cis, centered a t 0". There are now two groups of (AF)G duplex solution structures in the literature. In one group the (AF)G has the normal partner, C (16,171,while in the second group the (AF)G is mismatched with A (15), or with I o r G, in the present work. In the mismatched structures the modified deoxyguanosine is syn, with the AF in the minor groove. In the structures with a normal partner for the WF)G the modified deoxyguanosine is anti, a t least in one conformer, which places the AF in the B-DNA major groove. The glycosidic torsion of the second, carcinogenbase stacked conformer in these systems is undetermined. Thus, the early prediction that both syn and anti deoxyguanosines are energetically feasible for AF-modified G (18, 19) has been verified. The local sequence context, in particular, the nature of the partner base, determines the glycosidic torsion in the solution structures observed so far. It is of interest that the mismatch a t the lesion site stabilizes the abnormal syn conformation in the solution structure. One might speculate that during replication the (AF)G may be syn or anti since the two states are

378 Chem. Res. Toxicol., Vol. 8, No. 3, 1995 close in energy (18,19). If the (AF)G happens to adopt the normal B-like anti conformation observed in solution (16, 17), when the replication machinery reaches the (AF)G, a C could be placed in the complementary strand. However, if the (AF)G were syn a t that moment (andor carcinogen-base stacked as in the second conformer found in the (AF)GC solution studies (16, 17))a mismatch, by placement of an A or G in the opposite strand, or another mutagenic event such as a deletion or insertion could ensue.

Acknowledgment. This work was carried out in the laboratory of Dinshaw Patel a t the Columbia University College of Physicians and Surgeons. We thank him for overall guidance and support. The work was supported by NIH Grants CA-21111 (D. J. Patel and D.G.), CA28038 (S.B.),US DOE Grant DE-FG02-90ER60931(S.B.), US DOE Contract DE-AC05-840R21400 with MartinMarietta Energy Systems, Inc., and US DOE OHER Field Work Proposal ERKP931 (B.E.H.). We thank Nicholas Geacintov and Robert Shapiro, Chemistry Department, New York University, for helpful discussions. SupplementaryMaterial Available: Three figures showing phase sensitive NOESY contour plots (mixing time 250 ms) of the (AF)GI l l m e r duplex (3 pages). Ordering information is given on any current masthead page.

References (1) Singer, B., and Grunberger, D. (1983) Molecular Biology of

Mutagens and Carcinogens, Plenum, New York. (2) Irving, C. (1966) Enzymatic deacetylation of N-hydroxy-2-acetylaminofluorene by live microsomes. Cancer Res. 26, 1390-1396. (3) Irving, C., and Veazey, R. (1969) Persistent binding of 2-acetylaminofluorene to rat liver DNA in vivo and consideration of the mechanism of binding of N-hydroxy-2-acetylaminofluoreneto rat liver nucleic acids. Cancer Res. 29, 1799-1804. (4) Kriek, E. (1969) On the mechanism of action of carcinogenic aromatic amines. I. Binding of 2-acetylaminofluorene and N-hydroxy-2-acetylaminofluorene to rat liver nucleic acids in vivo. Chem.-Bid. Interact. 1, 3-17. (5) Beland, F., Dooley, K., and Jackson, C. (1982)Persistence of DNA adducts in rat liver and kidney after multiple doses of carcinogen N-hydroxy-2-acetyl aminofluorene. Cancer Res. 42, 1348-1354. (6) Poirier, M., True, B., and Laishes, B. (1982) Formation and removal of (guan-8-yl) DNA-2-acetylaminofluorene adducts in liver and kidney of male rats given dietary 2-acetylaminofluorene. Cancer Res. 42, 1317-1321. (7) Basu, A. K, and Essigmann, J. M. (1988)Site specificallymodified oligonucleotides as probes for the structural and biological effects of DNA damaging agents. Chem. Res. Toxicol. 1, 1-18. (8) Bichara, M., and Fuchs, R. P. P. (1985)DNA binding and mutation spectra of the carcinogen N-2-aminofluorene in E. coli: a correlation between the conformation of the premutagenic lesion and the mutation specificity. J . Mol. Biol. 183, 341-351. (9) Gupta, P. K., Pandrangi, R. G., Lee, M.-S., and King, C. M. (1991) Induction of mutation by N-acetoxy-N-acetyl-2-aminofluorene modified M13 viral DNA. Carcinogenesis 12,819-824. (10) Gupta, P. K., Lee, M. S., and King, C. M. (1988) Comparison of mutagenesis induced in single and double-stranded M13 viral DNA by treatment with N-hydroxy-2-aminofluorene. Carcinogenesis 9, 1337-1345. (11) Gupta, P. K., Johnson, D. L., Reid, T. M., Lee, M.-S., Romano, L. J., and King, C. M. (1989) Mutagenesis by single site-specific arylamine-DNA adducts. J . Biol. Chem. 264, 20120-20130. (12) Reid, T. M., Lee, M.-S., and King, C. M. (1990) Mutagenesis by single site-specific arylamine adducts in plasmid DNA Enhancing replication of the adducted strands alters mutation frequency. Biochemistry 29, 6153-6161. (13) Carothers, A. M., Steigenvalt, R. W., Urlaub, G., Chasin, L. A,, and Grunberger, D. (1989)DNA base changes and RNA levels in N-acetoxy-2-aminofluorene-induced dihydrofolate reductase mutants of Chinese hamster ovary cells. J . Mol. Biol. 208, 417428.

Abuaf et al. (14) Carothers, A. M., Urlaub, G., Mucha, J., Yuan, W., Chasin, L. A., and Grunberger, D. (1993) A mutational hot spot induced by N-hydroxy-aminofluorene in dihydrofolate reductase mutants of Chinese hamster ovary cells. Carcinogenesis 14, 2181-2184. (15) Norman, D., Abuaf, P., Hingerty, B. E., Live, D., Grunberger, D., Broyde, S., and Patel, D. J. (1989) NMR and computational characterization of the N-(deoxyguanosin-8-yl)aminofluoreneadduct [(AF)Glopposite adenosine in DNA (AF)G[synl.A[anti]pair formation and its pH dependence. Biochemistry 28, 7462-7476. (16) Cho, B. P., Beland, F. A,, and Marques, M. M. (1994) NMR structural studies of a 15-mer DNA duplex from a ras protooncogene modified with the carcinogen 2-aminofluorene: conformational heterogeneity. Biochemistry 33, 1373- 1384. (17) Eckel, L. M., and Krugh, T. R. (1994) Structural characterization of two interchangeable conformations of N-2-acetylaminofluorene modified DNA oligomer by NMR and energy minimization. Biochemistry 33, 13611-13624. (18)Broyde, S., and Hingerty, B. E. (1983) Conformation of 2-aminofluorene-modified DNA. Biopolymers 22, 2423-2441. (19) Lipkowitz, K. B., Chevalier, T., Widdifield, M., and Beland, F. A. (1982) Forcefield conformational analysis of aminofluorene and acetylaminofluorene substituted deoxyguanosine. Chem.-Biol. Interact. 40, 57-76. (20) Hingerty, B. E., Figueroa, S., Hayden, T., and Broyde, S. (1989) Prediction of DNA structure from sequence: a build-up technique. Biopolymers 28, 1195-1222. (21) Taylor, E. R., and Olson, W. K. (1983) Theoretical studies of nucleic acid interactions. I. Estimates of conformational mobility in intercalated chains. Biopolymers 22, 2667-2702. (22) Ornstein, R., and Fresco, J. (1983) Correlation of crystallographically determined and computationally predicted hydrogen-bonded pairing configurations of nucleic acid bases. Proc. Natl. Acad. Sci. U.S.A. 80, 5171-5175. (23) Schlick, T., Hingerty, B. E., Peskin, C. S., Overton, M. L., and Broyde, S. (1990)Search strategies, minimization algorithms, and molecular dynamics simulations for exploring conformational spaces of nucleic acids. In Theoretical Chemistry and Molecular Biophysics (Beveridge,D., & Laverty, R., Eds.) pp 39-58, Adenine Press, Schenectady, NY. (24) Cosman, M., de 10s Santos, C., Fiala, R., Hingerty, B. E., Singh, S., Ibanez, V., Margulis, L., Live, D., Geacintov, N. E., Broyde, S., and Patel, D. J. (1992) Solution conformation of the major adduct between carcinogen (+I-anti-benzo[alpyrene diol epoxide and DNA. Proc. Natl. Acad. Sci. U S A . 89, 1914-1918. (25) Cosman, M., de 10s Santos, C., Fiala, R., Hingerty, B. E., Ibanez, V., Luna, E., Harvey, R., Geacintov, N. E., Broyde, S., and Patel, D. J. (1993) Solution conformation of the (+)-cis-anti-[BPIdG adduct in a DNA duplex: intercalation of the covalently attached benzo[alpyrenyl ring into the helix and displacement of the modified deoxyguanosine. Biochemistry 32, 4145-4155. (26) O’Handley, S. F., Sanford, D. G., Xu, R., Lester, C., Hingerty, B. E., Broyde, S., and Krugh, T. R. (1993) Structural characterization of an N-acetyl-2-aminofluorene(AAF)modified DNA oligomer by NMR, energy minimization and molecular dynamics. Biochemistry 32, 2481-2497. (27) Hare, D. R., Wemmer, D. E., Chou, S. H., Drobny, G., and Reid, B. R. (1983) Assignment of the non-exchangeable proton resonances of d(C-G-C-G-A-A-T-T-C-G-C-G) using two-dimensional nuclear magnetic resonance methods. J. Mol. Biol. 171,319-366. (28) Patel, D. J. (1982) Antibiotic-DNA interactions: intermolecular nuclear Overhauser effects in the netropsin-d(C-G-C-G-A-A-T-TC-G-C-G) complex in solution. Proc. Natl. Acad. Sci. U.S.A. 79, 6425- 6428. (29) Suri, A. K., and Levy, R. M. (1995) Distance constraint ranges for NOES in proteins at long mixing times: a relaxation matrix analysis. J . Magn. Reson., Ser. B 106, 24-31. (30) Zhau, D., and Jardetzky, 0. (1994)An assessment of the precision and accuracy of protein structures determined by NMR. J . Mol. B i d . 239, 601-607. (31) Drew, H., Wing, R., Takano, T., Broka, C., Tanaka, S., Itakura, K., and Dickerson, R. (1981) Structure of a B-DNA dodecamer: conformation and dynamics. Proc. Natl. Acad. Sci. U.S.A. 78, 2179-2183. (32) Dickerson, R. E., Goodsell, D. S., Kopka, M. L., and Pjura, P. E. (1987) The effect of crystal packing on oligionucleotide double helix structure. J . Biomol. Struct. Dyn. 5, 557-580. (33) Altona, C., and Sundaralingam, G. (1972)Conformational analysis of the sugar ring in nucleosides and nucleotides. A new description using the concept of pseudorotation. J . Am. Chen. Soc. 94, 8205-8212.

TX9401167