Chem. Res. Toxicol. 2000, 13, 839-845
839
Differential Destabilization of the DNA Oligonucleotide Double Helix by a T‚G Mismatch, 3,N4-Ethenocytosine, 3,N4-Ethanocytosine, or an 8-(Hydroxymethyl)-3,N4-ethenocytosine Adduct Incorporated into the Same Sequence Contexts Ja´nos Sa´gi,† Alex Perry, Bo Hang, and B. Singer* Donner Laboratory, Life Sciences Division, Lawrence Berkeley National Laboratory, University of California, Berkeley, California 94720 Received February 18, 2000
The T‚G mismatch and the exocyclic adduct 3,N4-ethenocytosine (C) are repaired by the same enzyme, the human G/T(U) mismatch-DNA glycosylase (TDG). This enzyme removes the T, U, or C base from duplex DNA. The rate of cleavage was found to differ with the lesion and was also affected by neighbor sequences [Hang, B., Medina, M., Fraenkel-Conrat, H., and Singer, B. (1998) Proc. Natl. Acad. Sci. U.S.A. 95, 13561-13566]. Since sequence influences duplex stability, we determined the thermodynamic stability of T‚G and C-containing 15mer duplexes in which the bases flanking the lesion were systematically varied. The duplexes contained central 5′-TTXTT, 5′-AAXAA, 5′-CCXCC, or 5′-GGXGG sequences, where X is T, C, or two closely related structural derivatives of C: 3,N4-ethanocytosine (EC) and 8-(hydroxymethyl)-C (8-HM-C). Each of the four lesions, incorporated opposite G, decreased both the thermal (Tm) and thermodynamic stability (∆G°37) of the 15-mer control duplexes. On the basis of the Tm and ∆G°37 values, the order of destabilization of the TTXTT sequence in 15-mer duplexes was as follows: 8-HM-C > EC > C > T‚G. The ∆Tm values range from -15.8 to -9.5 °C when Ct ) 8 µM. Duplexes with flanking AA or TT neighbors were more destabilized, by an average of 2 °C, than those with flanking GG or CC neighbors. The base opposite the modified base also influenced duplex stability. Within the TT context, of the four changed bases opposite the adducts, C had the greatest destabilizing effect, up to -18.4 °C. In contrast, a G opposite an adduct was generally the least destabilizing, and the smallest value was -3.0 °C. Destabilizations were enthalpic in origin. Thus, this work shows that independently changing the modified base, the sequence, or the base opposite the lesion each affects the stability of the duplex, to significantly varying extents. The potential contribution of the thermodynamic stability to repair efficiency is discussed.
Introduction In genomic DNA, the T‚G mismatch is formed mainly by deamination of 5-methylcytosine which is hydrogenbonded to the opposite G. The T‚G mismatch can also result from errors in replication. Such mismatches are recognized by the G/T(U) mismatch-DNA glycosylase (TDG)1 which excises the T as the first step in the repair process (1). This enzyme which removes T from the T‚G mismatch was recently found to also efficiently remove 3,N4-ethenocytosine (C) (Figure 1) (2, 3). The T‚G and C‚G mismatches are structurally unrelated and apparently do not induce similar changes in the local conformation (4-9). The T‚G mismatch forms a wobble pair with two hydrogen bonds, and there is a small and highly localized perturbation of the double helix. The base pair remains planar with the thymine projecting into the * To whom correspondence should be addressed. Telephone: (510) 642-0637. Fax: (510) 486-6488. † Permanent address: Institute of Chemistry, Chemical Research Center, Hungarian Academy of Sciences, H-1525 Budapest, P.O. Box 17, Hungary. 1 Abbreviations: C, 3,N4-ethenocytosine; 8-HM-C, 8-(hydroxymethyl)-3,N4-ethenocytosine; EC, 3,N4-ethanocytosine; A, 1,N6-ethenoadenine; TDG, G/T(U) mismatch-DNA glycosylase.
Figure 1. Structures of the exocyclic adducts studied in this work.
major groove and the guanine into the minor groove, on the basis of X-ray crystallography, NMR spectroscopy, and molecular modeling studies (5-7). In the C‚G pair, as determined by high-resolution NMR spectroscopy and
10.1021/tx000040g CCC: $19.00 © 2000 American Chemical Society Published on Web 08/01/2000
840
Chem. Res. Toxicol., Vol. 13, No. 9, 2000
molecular dynamics simulation (4), the adduct is displaced and shifted toward the major groove while the opposite G remains stacked. The bases in the pair are both horizontally and vertically displaced. Both bases are in anti orientations and are connected with one hydrogen bond. The refined structure is bent at the lesion. These changes are also described as localized changes (4, 9). Circular dichroism and intrinsic fluorescence studies also revealed only modest lesion-induced changes in the duplex (8). It is generally believed that enzymatic recognition and repair of DNA lesions are influenced by the geometry of the lesion and the resulting local conformational changes, which can be sequence-dependent (5, 10, 11). Experimental results showed that the T‚G and C‚G mismatches apparently do not either have similar geometry or induce similar structural changes (4-8). Therefore, the observation that the same enzyme recognizes two structurally dissimilar substrates (2, 3) raises the following question (11): What is the common motif in substrate recognition and selection by the human G/T(U) mismatchDNA glycosylase? The T‚G and C‚G mismatches have been described to be cleaved at different rates (2, 3) which are influenced by the flanking sequences (2). Sequence is an important factor in the thermodynamic stability of the duplex and influences the adduct-induced destabilization (12, 13). Thus, it appeared logical to also take into consideration thermodynamic stability as one further factor contributing to recognition and repair, in addition to the geometry of the lesion and the local conformational changes (5, 10). Recent publications report that various lesions lead to changes in flexibility and curvature of the duplex, which are also related to energy (14, 15). The rates of excision of 1,N6-ethenoadenine (A) by mammalian alkylpurineDNA-N-glycosylases (13) and also of a synthetic abasic site by Escherichia coli endonuclease IV (12) have been correlated with the thermodynamic stability of the duplex imposed by neighbor bases. These observations also support the view that thermodynamics may also influence repair of DNA, especially for those repair enzymes, such as the two examples mentioned above, which require double-stranded DNA substrates. In the study presented here, a comparison was made of the thermal and thermodynamic stabilities of duplexes containing a T‚G mismatch or an C adduct, as well as two related structures, 3,N4-ethanocytosine (EC), formed from haloethylnitrosoureas (16) and 8-(hydroxymethyl)-3,N4-ethenocytosine (8-HM-C), formed from glycidaldehyde (17). The latter three adducts derive from reactions with carcinogens (18). In contrast to C, EC in 25-mer duplexes is apparently not repaired by glycosylases (11). The repair of 8-HM-C is not known. The major objective of this study was to incorporate each of the four mismatches into the same sequence context for a comparative study of their effect on duplex stability. This study is also directed toward the determination of the effect of different doublet neighbor base sequences flanking the lesions and that of the opposite bases on duplex stability. All three factors (the adduct, sequence, and opposite base) were found to influence the oligonucleotide double helices to varying extents.
Experimental Procedures Oligonucleotides. The 5′-DMT-3′-phosphoramidites of 3,N4etheno-2′-deoxycytidine (dC) and 3,N4-ethano-2′-deoxycytidine
Sa´ gi et al. (EdC) (19) were purchased from Chem-Master International, Inc. (East Setauket, NY). The synthesis of the phosphoramidite of 8-(hydroxymethyl)-3,N4-etheno-2′-deoxycytidine (8-HM-dC) is described in ref 20. The 15-nucleotide long unmodified oligodeoxyribonucleotides were purchased in HPLC-purified form from Operon Technologies, Inc. (Alameda, CA). The C, EC, and 8-HM-C adduct-containing oligodeoxyribonucleotides were synthesized in an ABI 392 DNA synthesizer. For the Ccontaining oligonucleotides, the standard phosphoramidite chemistry was used, as described previously (13). The EC oligonucleotides were synthesized according to the method of Bonala et al. (21). Synthesis of the 8-HM-C oligonucleotides was carried out as described recently (20). Oligonucleotides were purified by HPLC and desalted, and the compositions were verified by enzyme digestion as previously described (13). Determination of the Thermodynamic Stability. Oligonucleotide concentrations were calculated from the high-temperature (80 °C) absorbance of the solution and from the singlestrand extinction coefficient for the oligonucleotide obtained by using the DNA/Oligo (long) Quantitation software of a Beckman DU 7400 diode array spectrophotometer. In the case of the adduct-containing oligonucleotides, the extinction coefficient for pdC [260 ) 7500 M-1 cm-1 (22, 23)] was replaced by the following values as determined: C, 260 ) 9240 M-1 cm-1; EC, 260 ) 3140 M-1 cm-1; and 8-HM-C, 260 ) 4900 M-1 cm-1 (20). For the non-self-complementary double helices, the oligonucleotide strands were mixed at a 1:1 molar ratio. Formation of duplexes was assessed with mixing curves (not shown). Duplex sequences are shown in Scheme 1.
Scheme 1 5′ AGCGG NN X NN GAGCT 3′ 3′ TCGCC NN N NN CTCGA 5′ where X is C, T, C, EC, 8-HM-C, or A, N is G, C, A, or T, and NN is TT, AA, CC, or GG. Double helices with a total strand concentration (Ct) of 8 µM were dissolved in 0.1 M NaCl, 10 mM sodium phosphate buffer, and 0.1 mM Na2EDTA (pH 7.0). The absorption-temperature melting profiles were recorded in a Beckman DU 7400 diode array spectrophotometer from 10 to 100 °C at 280 nm, essentially as described previously (12, 13). The thermal stability (Tm) and the thermodynamic parameters for duplex strand dissociation (∆H°, ∆S°, and ∆G°37) were calculated from the melting curves of the duplexes, using the Meltwin version 3.0 program (24). Thermodynamic parameters for melting were also calculated from the concentration dependence of the Tm values with each adduct in one sequence, as previously published (12), to ascertain whether the melting is a two-state process. One criterion is that enthalpy values obtained from the individual melt curve fits and from the concentration dependence agree within 15% (25) to 20% (26).
Results and Discussion The effects on duplex stability of four different lesions are compared in this study. Two lesions, the T‚G mismatch and the C adduct, are “functionally” related, as both the T and C are removed by the same enzyme in the repair process, the human G/T(U) mismatch-DNA glycosylase. The other two adducts, EC and 8-HM-C, are structurally related to C (Figure 1). All four lesions were site-specifically incorporated into the same sequence context and the duplex stability determined by thermodynamics. A comparison was then made of the effect of the sequence context on the lesion-induced duplex destabilization. For this purpose, four different doublet neighbor sequences were used: TT, AA, CC, and GG. The third aspect of the lesion-induced destabilization that was studied was the effect of the base opposite the lesion.
Mismatch-Induced Duplex Destabilization
Chem. Res. Toxicol., Vol. 13, No. 9, 2000 841
Table 1. Thermodynamic Parameters for the Dissociation of the 15-mer Duplexes Containing a Central C Base, a T‚G Mismatch, or a EC, EC, or 8-HM-EC Adduct, Opposite Ga no. 1 2 3 4
5 6 7 8
9 10 11 12
13 14 15 16
17
central sequence
Tm (°C)
∆Tm (°C)
H280b (%)
MeltWin analysis ∆H° (kcal/mol) ∆S° (cal K-1 mol-1)
∆G°37 (kcal/mol)
TTCTT AAGAA AACAA TTGTT CCCCC GGGGG GGCGG CCGCC
60.7 ( 0.5
Control 15-mer Duplexes 27.7 98.4 ( 2.6 267.3 ( 10.0
60.6 ( 0.3
29.5
107.6 ( 6.2
296.4 ( 18.6
15.71 ( 0.49
71.9 ( 0.2
42.7
98.6 ( 2.0
259.6 ( 5.7
18.05 ( 0.21
70.4 ( 0.4
42.3
98.2 ( 1.7
259.6 ( 4.9
17.61 ( 0.21
TTTTT AAGAA AATAA TTGTT CCTCC GGGGG GGTGG CCGCC
51.2 ( 0.4
-9.5 ( 0.4
51.4 ( 0.1
-9.2 ( 0.1
29.6
91.4 ( 1.0
64.1 ( 0.2
-7.8 ( 0.2
39.7
63.2 ( 0.1
-7.2 ( 0.4
42.9
TTCTT AAGAA AACAA TTGTT CCCCC GGGGG GGCGG CCGCC
49.1 ( 0.6
-11.6 ( 0.6
C-Containing 15-mer Duplexes 27.0 89.9 ( 3.6 252.8 ( 10.8
49.6 ( 0.1
-11.0 ( 0.3
31.0
93.2 ( 2.3
62.1 ( 0.1
-9.8 ( 0.2
37.2
58.8 ( 0.2
-11.6 ( 0.3
37.7
TTECTT AAGAA AAECAA TTGTT CCECCC GGGGG GGECGG CCGCC
46.9 ( 0.6
-13.8 ( 0.7
EC-Containing 15-mer Duplexes 24.2 81.3 ( 3.2 228.0 ( 9.6
48.7 ( 0.1
-11.9 ( 0.3
32.3
90.8 ( 3.6
60.7 ( 0.1
-11.2 ( 0.4
37.0
60.6 ( 0.1
-9.8 ( 0.5
44.1
TTHMCTT AAGAA
44.9 ( 0.4
-15.8 ( 0.6
T‚G Mismatch-Containing 15-mer Duplexes 28.8 89.8 ( 5.8 249.6 ( 16.4
∆∆G°37 (kcal/mol)
15.15 ( 0.19
12.02 ( 0.33
-3.13 ( 0.4
255.5 ( 2.9
12.15 ( 0.07
-3.56 ( 0.27
92.8 ( 1.2
249.0 ( 3.9
15.54 ( 0.04
-2.51 ( 0.24
97.4 ( 0.8
263.4 ( 2.2
15.67 ( 0.06
-1.94 ( 0.16
11.47 ( 0.27
-3.68 ( 0.32
262.7 ( 7.1
11.72 ( 0.09
-3.99 ( 0.58
90.3 ( 5.2
243.2 ( 15.8
14.86 ( 0.40
-3.19 ( 0.37
90.8 ( 3.7
247.6 ( 11.0
14.04 ( 0.28
-3.57 ( 0.06
10.61 ( 0.24
-4.54 ( 0.38
256.0 ( 11.1
11.39 ( 0.12
-4.32 ( 0.39
90.6 ( 3.4
245.3 ( 10.4
14.53 ( 0.22
-3.52 ( 0.31
93.8 ( 2.9
254.9 ( 8.0
14.73 ( 0.22
-2.88 ( 0.22
8-HM-C-Containing 15-mer Duplex 26.6 73.1 ( 3.0 203.8 ( 9.1
9.92 ( 0.19
-5.23 ( 0.28
a Thermal transitions were determined in 0.1 M NaCl, 10 mM sodium phosphate, and 0.1 mM EDTA (pH 7). The total strand concentration of the duplexes was 8 µM. Data shown are averages ( the standard error of the data obtained from the measurements of three to six individually annealed samples. The average standard error for the H280 values was (1.5%. b H280 refers to the change in absorption at 280 nm ranging from 10 to 100 °C.
The absorption-temperature melting profiles for the 15-mer duplexes with or without a site-specifically incorporated lesion exhibited monophasic and cooperative transitions from duplex to single-stranded coils, and the processes were reversible. The data calculated from the melting curves are listed in Tables 1 and 2. These include thermal stability (Tm), the increase in absorption at 280 nm (H280), and the enthalpy (∆H°), entropy (∆S°), and free-energy change (∆G°37) values for the duplex strand dissociation. Differential values (∆Tm and ∆∆G°37) are also shown. The unmodified 15-mer duplexes melt according to a clear “two-state” (all-or-none) model (12, 13). This indicates that there are no thermodynamically significant intermediate structures formed during melting in addition to the double helices and single-stranded coils. Melting of lesion-containing duplexes fits within the range defined for two-state melting (25, 26). Comparison of the Effect of a Single T‚G Mismatch, C, EC, or 8-HM-C (X), on the Duplex Stability in the TTXTT Sequence Context. All four lesions (X) were first incorporated into the 5′-AGCGGTTXTTGAGCT sequence which was then annealed to the unmodified complementary strand (Scheme 1). This is
the sequence in which all four lesions were compared (Table 1). The T‚G wobble pair, which is known to be among the most stable mismatches (27-29), was found to destabilize the control duplex by 9.5 °C (Table 1, duplexes 1 and 5). This was the smallest among the effect of mismatches of T and also C (Table 2, duplexes 1823). The exocyclic adduct C was more destabilizing than the T‚G mismatch with a ∆Tm of -11.6 °C (Table 1, duplex 9). This indicates that the effect of C is in the range of the effect of all single-base mismatches. Furthermore, destabilization by C was similar to that of another etheno adduct, A, when incorporated into the same sequence context. The ∆Tm with A was -9.2 °C (13). On the basis of the free energy differences, the effects of two etheno adducts were similar to and greater than that of the T‚G. The ∆∆G°37 values were -3.13, -3.68, and 3.80 kcal/mol for the duplexes incorporating T‚G, C‚G, and A‚T (13) mismatches, respectively. The EC adduct contains a saturated exocyclic ring (Figure 1). Saturation adds two additional hydrogen atoms to the ring and decreases the aromatic character of C. The net result is a further destabilization, as compared to the effect of C, and the ∆Tm was -13.8 °C
842
Chem. Res. Toxicol., Vol. 13, No. 9, 2000
Sa´ gi et al.
Table 2. Effect of the Opposite Base on Thermodynamic Stability of the 15-mer Duplexes Containing the Central TTXTT/ AANAA Sequence, Where X Is a C or T Base or a EC, EC, 8-HM-EC, or EA Adduct and N Is G, C, A, or Ta no. 1 18 19 20 21 5 22 23 9 24 25 26 13 27 28 29 17 30 31 32 33 34 35 36 37
central sequence
Tm (°C)
TTCTT AAGAA TTCTT AACAA TTCTT AAAAA TTCTT AATAA
60.7 ( 0.5
TTTTT AAAAA TTTTT AAGAA TTTTT AACAA TTTTT AATAA
59.7 ( 0.3
∆Tm (°C)
H280b (%)
MeltWin analysis ∆H° (kcal/mol) ∆S° (cal K-1 mol-1)
∆G°37 (kcal/mol)
27.7
All Four Opposite C 98.4 ( 2.6
267.3 ( 10.0
15.15 ( 0.19
∆∆G°37 (kcal/mol)
44.4 ( 0.2
-16.3 ( 0.6
26.9
80.3 ( 1.8
226.9 ( 9.2
9.95 ( 0.21
-5.20 ( 0.34
47.2 ( 0.1
-13.5 ( 0.3
28.1
86.0 ( 4.2
242.2 ( 11.2
10.84 ( 0.23
-4.31 ( 0.11
48.2 ( 0.2
-12.5 ( 0.3
28.5
93.6 ( 3.6
265.2 ( 5.2
11.34 ( 0.08
-3.81 ( 0.22
33.3
All Four Opposite T 101.4 ( 4.6
278.6 ( 11.3
14.99 ( 0.23
51.2 ( 0.2
28.8
89.8 ( 3.8
249.6 ( 6.3
12.02 ( 0.20
47.9 ( 0.1
-8.5 ( 0.2 -9.5c -11.8 ( 0.3
30.5
85.3 ( 2.6
239.7 ( 10.4
10.99 ( 0.19
-2.97 ( 0.2 -3.13c -4.0 ( 0.30
48.6 ( 0.2
-11.1 ( 0.4
32.3
80.9 ( 3.3
225.4 ( 2.6
11.0 ( 0.14
-3.99 ( 0.21
TTCTT AAGAA TTCTT AACAA TTCTT AAAAA TTCTT AATAA
49.1 ( 0.2
-11.6 ( 0.3
27.0
11.47 ( 0.15
-3.68 ( 0.33
43.0 ( 0.3
-17.7 ( 0.5
29.1
72.4 ( 2.6
203.0 ( 6.4
9.45 ( 0.03
-5.70 ( 0.24
46.2 ( 0.2
-14.5 ( 0.2
27.7
77.0 ( 1.4
215.1 ( 4.9
10.31 ( 0.11
-4.84 ( 0.21
47.1 ( 0.2
-13.6 ( 0.3
32.8
86.0 ( 3.2
242.5 ( 11.2
10.81 ( 0.10
-4.34 ( 0.22
TTECTT AAGAA TTECTT AACAA TTECTT AAAAA TTECTT AATAA
46.9 ( 0.6
-13.8 ( 0.7
24.2
10.61 ( 0.24
-4.54 ( 0.38
42.3 ( 0.2
-18.4 ( 0.4
27.1
68.2 ( 0.9
190.2 ( 3.0
9.24 ( 0.02
-5.91 ( 0.11
44.4 ( 0.1
-16.3 ( 0.3
23.8
72.8 ( 0.5
203.0 ( 1.4
9.78 ( 0.03
-5.37 ( 0.14
43.6 ( 0.2
-17.1 ( 0.4
29.0
71.4 ( 2.2
199.4 ( 7.0
9.58 ( 0.10
-5.57 ( 0.19
TTHMCTT AAGAA TTHMCTT AACAA TTHMCTT AAAAA TThmCTT AATAA
44.9 ( 0.4
-15.8 ( 0.6
All Four Opposite 8-HM-C 26.6 73.1 ( 3.0 203.8 ( 9.1
9.92 ( 0.19
-5.23 ( 0.18
42.3 ( 0.1
-18.4 ( 0.3
28.8
65.9 ( 4.3
182.7 ( 8.5
9.19 ( 0.12
-5.96 ( 0.22
44.0 ( 0.3
-16.7 ( 0.4
27.3
62.4 ( 3.2
170.6 ( 3.3
9.47 ( 0.23
-5.68 ( 0.21
45.9 ( 0.1
-14.8 ( 0.3
31.6
78.6 ( 1.2
220.4 ( 4.9
10.27 ( 0.02
-4.88 ( 0.12
TTATT AATAA TTATT AATAA TTATT AAGAA TTATT AACAA TTATT AAAAA
56.8 ( 0.3
32.1
All Four Opposite C 89.9 ( 5.6 252.8 ( 12.4
All Four Opposite EC 81.3 ( 3.2 228.0 ( 9.6
All Four Opposite A 100.1 ( 2.3 276.4 ( 11.6
14.30 ( 0.21
47.8 ( 0.2
-9.0 ( 0.8
28.5
79.7 ( 2.1
222.1 ( 6.4
10.74 ( 0.2
-3.56 ( 0.32
53.8 ( 0.2
-3.0 ( 0.4
30.6
92.0 ( 1.8
255.4 ( 6.8
12.83 ( 0.12
-1.47 ( 0.21
46.3 ( 0.1
-10.5 ( 0.2
26.8
77.3 ( 1.9
215.9 ( 16.2
10.33 ( 0.23
-3.97 ( 0.33
48.9 ( 0.2
-7.9 ( 0.3
28.3
86.8 ( 3.2
243.5 ( 11.2
11.29 ( 0.26
-3.01 ( 0.11
a Thermal transitions were determined in 0.1 M NaCl, 10 mM sodium phosphate, and 0.1 mM EDTA (pH 7). The total strand concentration of the duplexes was 8 µM. Data shown are averages ( the standard error of the data obtained from the measurements of three to six individually annealed samples. The average standard error for the H280 values was (1.5%. b H280 refers to the change in absorption at 280 nm ranging from 10 to 100 °C. c Compared to the C‚G pair-containing duplex (Table 2, duplex 1).
(Table 1, duplex 13). The same stability order was observed earlier with 25-mer duplexes containing C and EC in another sequence (9). An even greater effect was observed with the 8-hydroxymethyl derivative of C, 8-HM-C (Figure 1 and Table 1, duplex 17). The Tm dropped by 15.8 °C, which was 4.2 °C lower than the effect of C (-11.6 °C). The structural cause for this effect could be the perturbation of the stacking interactions; the ∆∆H° value was -25.3 kcal/mol (Table 1, duplexes 1 and 17). The average
enthalpy change for one base pair in this specific 15-mer duplex is 6.56 kcal/mol. Thus, the -∆∆H° value observed for the T‚G- and C-containing duplexes, in which ∆∆H° was -8.6 and -8.5 kcal/mol, respectively, may indicate the loss of approximately 1.3 base pair interactions. This is known to originate from the horizontal displacement of bases in the case of the T‚G mismatch and from the horizontal and vertical displacements in the case of C‚ G. However, these effects are considered as very localized changes (4-8). In the case of EC, the observed ∆∆H°
Mismatch-Induced Duplex Destabilization
(-17.1 kcal/mol) corresponds to about 2.5 base pair interactions. In the duplex containing the 8-HM-C, the observed ∆∆H° (-25.3 kcal/mol) corresponds to the loss of almost 4 base pair interactions. The conformational changes at the EC or 8-HM-C adducts are apparently higher and/or much less localized than at T‚G or C. Effect of the Flanking Sequences on Destabilization. The T‚G, C, and EC adducts were also incorporated into the middle of four different homooligomeric sequences (Scheme 1). The A or T and G or C doublet neighbors had different effects on duplex stability (Tm, ∆G°37) whether unmodified or with a lesion. This is a function of the GC content (Table 1). Exchange of the doublet TT neighbors for AA or GG for CC caused only small changes. The average differences between the two A or T and two G or C pair-containing control duplexes (Table 1, duplexes 1 and 2, and 3 and 4) were 10.5 °C and 2.4 kcal/ mol. However, the average differences between the T‚G, C, and EC lesion-containing duplexes with A or T and G or C flanking doublets were 12.1 °C and 3.34 kcal/mol. The higher values for the lesion-containing duplexes represent the differential effect of sequence on destabilization. With the T‚G mismatch, the symmetric flanking GG or CC neighbors led to a smaller destabilization of the duplex, by 1.9 °C, than did the AA or TT neighbors (Table 1, duplexes 7 and 8, and 5 and 6). In the case of the ECcontaining duplexes, the symmetric GG or CC flanking bases allowed smaller destabilization for the duplex by 2.4 °C than did the AA or TT doublets. There was, however, some overlapping in the ∆Tm values but not in the ∆∆G°37 values (duplexes 15 and 16, and 13 and 14). In the case of the C-containing duplexes, only a small effect of the flanking sequences was observed (duplexes 9-12). The only difference observed was between the CC and the GG neighbor-containing duplexes. The context dependence observed here for the G‚T mismatch was similar to that described by Allawi and SantaLucia (6), who used single-neighbor base changes in various sequences and different duplex lengths. Sequence dependence with the C adduct has also been described, by Gelfand et al. (8), with symmetric single-C or -G neighbors only and various opposite bases in 13mer duplexes. On this basis, they concluded that sequence context modulated the C-induced destabilization only modestly (8), similar to our observation (Table 1, duplexes 9-12). Gelfand et al. also showed that the C duplex was more destabilized with the single-G neighbors than with the single-C neighbors, similar to our data (Table 1, duplexes 12 and 11). We observed an average 20% loss in the ∆G°37 value of the control 15-mer duplexes in the presence of the C adduct, whereas a larger loss in ∆G°25 was described for the 13-mers containing C (8). The different values can originate primarily from the different duplex lengths and sequences as well as the conditions that were used. The results reported in this work show that the sequence flanking a lesion not only affects the stability of the duplex but generally also affects the extent of destabilization induced by the lesion. The sequencedependent effect of the T‚G mismatch and the EC duplexes was similar to that observed previously with A (13). In the case of T‚G, EC, and A, the symmetric AA or TT neighbors allowed more destabilization of the
Chem. Res. Toxicol., Vol. 13, No. 9, 2000 843
duplex by 1.9, 2.4, and 2.3 °C, respectively, than the GG or CC neighbors. Effect of the Base Opposite the Lesion. The bases opposite a lesion may induce different conformations for the lesion or for the whole mismatch. The accommodation into the double helix can also be reflected in the thermodynamic stability of the duplex. Effects of the base opposite C, T and C, EC, 8-HM-C, and also A (X) were compared only in the TTXTT context (Table 2). The best accommodation, that is, the smallest destabilization of the duplex, was observed with the A‚G mismatch (Table 2, duplex 35). The ∆Tm was -3.0 °C, and the ∆∆G°37 was -1.47 kcal/mol. This effect is much smaller than the effect of any other natural mismatch studied in 15-mer duplexes including the wobble pairs, such as T‚G or C‚A (Table 1, duplex 5, and Table 2, duplex 19). No such small destabilization was observed with any mismatch of the exocyclic derivatives of the C studied here as found with the A‚G pair. The least destabilized duplexes contained the central C‚G (∆Tm ) -11.6 °C), EC‚G (-13.8 °C), and 8-HM-C‚T (-14.8 °C) mismatches. The explanation for the relatively minor effect of the A‚G mismatch comes from X-ray crystallography (30) and NMR studies (31). The A adduct in the A‚G mismatch was observed to be in the syn range of the glycosidic torsion angles, versus the anti range common in the B-DNA, in two sequences: the self-complementary dodecamer containing two A adducts in the TTAGC motif (30) and in the 9-mer duplex containing the ACACA motif (31). One can assume that A also adopts the syn conformation in the TTATT context used in this work. However, the syn orientation observed in a given sequence may not be present in others or may not result in a similar good accommodation as found with the A(syn)‚G(anti) pair. The orientation of the C adduct was also found to be syn when the opposite base was T in the ACCCA motif of a non-self-complementary 11-mer duplex (32, 33). However, the results presented here showed that the C‚T mismatch in the TTCTT sequence caused a major destabilization, where the ∆Tm was -13.6 °C (Table 2, duplex 26). The glycosidic orientation of C in this sequence is not known. When the largest destabilization effects are compared, another interesting stability feature can be observed. Among the mismatches of the unmodified C base, the opposite C, forming a C‚C mismatch, caused the greatest reduction of duplex stability (Table 2, duplex 18). The ∆Tm was -16.3 °C, and ∆∆G°37 was -5.20 kcal/mol. Similarly, the C base opposite exocyclic adducts C, EC, and 8-HM-C as well as opposite A showed the greatest destabilization effect of the four normal opposite bases. Unexpectedly, exocyclic modifications of the C base contributed only a small extent to the large destabilization caused by the C‚C mismatch itself. The order was as follows: C‚C (∆Tm ) -16.3 °C) < C‚C (-17.7 °C) < EC‚C (-18.4 °C) ) 8-HM-C‚C (-18.4 °C) (Table 2, duplexes 18, 24, 27, and 30). Thus, the effects of the opposite base and the lesion on destabilization are apparently not additive. The C‚C mismatch has been characterized as an “open” mismatch, and the least stable among the wobble, “weak”, and “open” mismatches (27, 28). A recent NMR and molecular dynamics study on the solution structure for the C‚C (34) and for the C‚C (33) mismatches showed
844
Chem. Res. Toxicol., Vol. 13, No. 9, 2000
that the bases were only horizontally displaced; i.e., the mismatches remained fairly planar. There was one hydrogen bond observed between the O2 of C and N4 of the opposite C (33), and this was one of the possible sheared structures also proposed for the C‚C (34). In this structure, the exocyclic ring of C does not disrupt the hydrogen bond and the C remains stacked. On this basis, it seems reasonable that there was only a small further duplex destabilization observed with the C‚C mismatch, as compared to the effect of the C‚C mismatch (Table 2, duplexes 18 and 24). Saturation or substitution of the exocyclic ring of C probably does not change this sheared and hydrogen-bonded structure, which may explain the similar destabilization effects observed by the C‚C, EC‚ C, and 8-HM-C‚C mismatches. Biochemical Implications. This study was initiated by the recent observation that two structurally unrelated lesions, the T‚G mismatch and the C adduct, were repaired by the same enzyme: the human G/T(U) mismatch-DNA glycosylase (2, 3). It has been postulated that removal of a base by glycosylases generally requires the flipping of the base out into an extrahelical position in the major groove to fit into the active center cavity of the enzyme (35-38). Enzyme facilitated (active) flipping is believed to occur through bending and/or kinking of the DNA backbone (35-38). This requires energy, which is probably not the same for lesions that cause different changes in the duplex thermodynamic stability. In this study, the T‚G mismatch and the C adduct destabilized the 15-mer duplexes by averages of 2.8 and 3.6 kcal/mol, respectively, in four different sequences. C was earlier found to be removed with higher efficiency from the C‚ G mismatch than was T from the T‚G mismatch by the human G/T(U) mismatch-DNA glycosylase (2, 3). It can be speculated that the more destabilized C‚G mismatchcontaining duplex can be kinked and then the C flipped out with higher efficiency, that is, with less energy, than the T from the T‚G mismatch-containing duplex. However, the kinetic factor using the two different substrates can be important, and the contribution of the kinetics and the duplex stability in the efficiency of base excision is not known. There was no repair of two close structural derivatives of C (EC and 8-HM-C) by the human G/T(U) mismatchDNA glycosylase. Thus, repair cannot be predicted from only structural similarities (11). EC and 8-HM-C adducts induced similar and higher duplex destabilization than C, respectively. It thus appears that repair also cannot necessarily be predicted from comparative thermodynamic data. However, within a given group of substrates with different neighbor bases, such as an abasic site or A, thermodynamic data could be related to cleavage efficiency (12, 13). In summary, four different, but related, lesions were site-specifically incorporated into the same sequence context of 15-mer duplexes, and the effects on the thermodynamic stability of duplexes were compared. The T‚G mismatch and the C, EC, and 8-HM-C adducts all significantly decreased duplex stability, although the extent of decrease differed with the modified bases. The extent of destabilization increased in the following order: T‚G < C < EC < 8-HM-C. These effects fell in the range of those of single-base natural mismatches. Sequences flanking the lesion, which were systematically changed, also influenced the destabilization. Duplexes with AA or TT neighbors were more destabilized by the
Sa´ gi et al.
lesion than were duplexes with GG or CC neighbors. Opposite bases also made a difference in the accommodation of the mismatch into the duplex. The G opposite A caused minimal destabilization, while C opposite each adduct was strongly destabilizing. It is concluded that the thermodynamic stability of the lesion-containing duplex is a result of the effects of the lesion, its neighbor bases and opposite base, and is a factor in the efficiency of repair of double-stranded substrates. Conformational aspects of these mismatches in the same sequence context are under investigation.
Acknowledgment. This work was supported by NIH Grants CA 47723 and CA 72079 (to B.S.) and was administered by the Lawrence Berkeley National Laboratory under Department of Energy Contract DE-AC0376SF00098.
References (1) Lindahl, T. (1982) DNA repair enzymes. Annu. Rev. Biochem. 51, 61-68. (2) Hang, B., Medina, M., Fraenkel-Conrat, H., and Singer, B. (1998) A 55-kDa protein isolated from human cells shows DNA glycosylase activity toward 3,N4-ethenocytosine and the G/T mismatch. Proc. Natl. Acad. Sci. U.S.A. 95, 13561-13566. (3) Saparbaev, M., and Laval, J. (1998) 3,N4-Ethenocytosine, a highly mutagenic adduct, is a primary substrate for Escherichia coli double-stranded uracil-DNA glycosylase and human mismatchspecific thymine-DNA glycosylase. Proc. Natl. Acad. Sci. U.S.A. 95, 8508-8513. (4) Cullinan, D., Johnson, F., Grollman, A. P., Eisenberg, M., and de los Santos, C. (1997) Solution structure of a DNA duplex containing the exocyclic lesion 3,N4-etheno-2′-deoxycytidine opposite 2′deoxyguanosine. Biochemistry 36, 11933-11943. (5) Allawi, H. T., and SantaLucia, J., Jr. (1998) NMR solution structure of a DNA dodecamer containing single G‚T mismatches. Nucleic Acids Res. 26, 4925-4934. (6) Allawi, H. T., and SantaLucia, J., Jr. (1997) Thermodynamics and NMR of internal G‚T mismatches in DNA. Biochemistry 36, 10581-10594. (7) Hunter, W. N., Brown, T., Kneale, G., Anand, N. N., Rabinovich, D., and Kennard, O. (1987) The structure of guanosine-thymidine mismatches in B-DNA at 2.5 Å resolution. J. Biol. Chem. 262, 9962-9970. (8) Gelfand, C. A., Plum, G. E., Grollman, A. P., Johnson, F., and Breslauer, K. J. (1998) The impact of an exocyclic cytosine adduct on DNA duplex properties: significant thermodynamic consequences despite modest lesion-induced structural alterations. Biochemistry 37, 12507-12512. (9) Sa´gi, J., and Singer, B. (1999) Thermal destabilization of DNA oligonucleotide duplexes by exocyclic adducts on adenine or cytosine depends on both the base and the size of adduct. In The Role of Cyclic Adducts in Carcinogenesis and Mutagenesis (Singer, B., and Bartsch, H., Eds.) Vol. II, pp 191-196, International Agency for Research on Cancer Scientific Publication 150, International Agency for Research on Cancer, Lyon, France. (10) Hunter, W. N. (1992) Crystallographic studies of DNA containing mismatches, modified and unpaired bases. Methods Enzymol. 211, 221-231. (11) Singer, B., and Hang, B. (1997) What structural features determine repair enzyme specificity and mechanism in chemically modified DNA? Chem. Res. Toxicol. 10, 713-732. (12) Sa´gi, J., Hang, B., and Singer, B. (1999) Sequence-dependent repair of synthetic AP sites in 15-mer and 35-mer oligonucleotides: role of thermodynamic stability imposed by neighbor bases. Chem. Res. Toxicol. 12, 917-923. (13) Hang, B., Sa´gi, J., and Singer, B. (1998) Correlation between sequence-dependent glycosylase repair and the thermal stability of oligonucleotide duplexes containing 1,N6-ethenoadenine. J. Biol. Chem. 273, 33406-33413. (14) Marathias, V. M., Jerkovic, B., Arthanari, H., and Bolton, P. H. (2000) Flexibility and curvature of duplex DNA containing mismatched sites as a function of temperature Biochemistry 39, 153-160. (15) Marathias, V. M., Jerkovic, B., and Bolton, P. H. (1999) Damage increases the flexibility of DNA. Nucleic Acids Res. 27, 18541858.
Mismatch-Induced Duplex Destabilization (16) Ludlum, D. B. (1990) DNA alkylation by the haloethylnitrosoureas: nature of modifications produced and their enzymic repair or removal. Mutat. Res. 233, 117-126. (17) Golding, B. T., Slaich, P. K., Kennedy, G., Bleasdale, C., and Watson, W. P. (1996) Mechanisms of formation of adducts from reactions of glycidaldehyde with 2′-deoxyguanosine and/or guanosine. Chem. Res. Toxicol. 9, 147-157. (18) International Agency for Research on Cancer (1987) IARC Monographs on the Evaluation of the Carcinogenic Risks to Humans, International Agency for Research on Cancer, Lyon, France. (19) Zhang, W., Rieger, R., Iden, C., and Johnson, F. (1995) Synthesis of 3,N4-etheno, 3,N4-ethano, and 3-(2-hydroxyethyl) derivatives of 2′-deoxycytidine and their incorporation into oligomeric DNA. Chem. Res. Toxicol. 8, 148-156. (20) Chenna, A., Perry, A., and Singer, B. (2000) The synthesis of 8-hydroxymethyl-3,N4-etheno-2′-deoxycytidine, a potential carcinogenic glycidaldehyde adduct and its site-specific incorporation into DNA oligonucleotides. Chem. Res. Toxicol. (in press). (21) Bonala, R. R., Rieger, R. A., Shibutani, S., Grollman, A. P., Iden, C. R., and Johnson, F. (1999) 3,N4-Ethano-2′-deoxycytidine: chemistry of incorporation into oligomeric DNA and reassessment of miscoding potential. Nucleic Acids Res. 27, 4725-4733. (22) Borer, P. N. (1975) In CRC Handbook of Biochemistry and Molecular Biology, 3rd ed., Vol. I, p 589, CRC Press, Boca Raton, FL. (23) Gray, D. M., Hung, S.-H., and Johnson, K. H. (1995) Absorption and circular dichroism spectroscopy of nucleic acid duplexes and triplexes. Methods Enzymol. 246, 19-34. (24) McDowell, J. A., and Turner, D. H. (1996) Investigation of the structural basis for thermodynamic stabilities of tandem GU mismatches: solution structure of (rGAGGUCUC)2 by twodimensional NMR and simulated annealing. Biochemistry 35, 14077-14089. (25) Xia, T., SantaLucia, J., Jr., Burkard, M. E., Kierzek, R., Schroeder, S. J., Jiao, X., Cox, C., and Turner, D. H. (1998) Thermodynamic parameters for an expanded nearest-neighbor model for formation of RNA duplexes with Watson-Crick base pairs. Biochemistry 37, 14719-14735. (26) SantaLucia, J., Jr., Allawi, H. T., and Seneviratne, P. A. (1996) Improved nearest-neighbor parameters for predicting DNA duplex stability. Biochemistry 35, 3555-3562. (27) Aboul-ela, F., Koh, D., Tinoco, I., Jr., and Martin, F. H. (1985) Base-base mismatches. Thermodynamics of double helix formation for dCA3XA3G + dCT3YT3G (X, Y ) A, C, G, T). Nucleic Acids Res. 13, 4811-4824. (28) Werntges, H., Steger, G., Riesner, D., and Fritz, H. J. (1986) Mismatches in DNA double strands: thermodynamic parameters and their correlation to repair efficiencies. Nucleic Acids Res. 14, 3773-3790.
Chem. Res. Toxicol., Vol. 13, No. 9, 2000 845 (29) Sa´gi, J., Chenna, A., Hang, B., and Singer, B. (1998) A single cyclic p-benzoquinone adduct can destabilize a DNA oligonucleotide duplex. Chem. Res. Toxicol. 11, 329-334. (30) Leonard, G. A., McAuley-Hecht, K. E., Gibson, N. J., Brown, T., Watson, W. P., and Hunter, W. N. (1994) Guanine-1,N6ethenoadenine base pairs in the crystal structure of d(CGCGAATT(dA)GCG). Biochemistry 33, 4755-4761. (31) de los Santos, C., Kouchakdjian, M., Yarema, K., Basu, A., Essigmann, J., and Patel, D. J. (1991) NMR studies of the exocyclic 1,N6-ethenodeoxyadenosine adduct (dA) opposite deoxyguanosine in a DNA duplex. dA(syn)‚dG(anti) pairing at the lesion site. Biochemistry 30, 1828-1835. (32) Cullinan, D., Korobka, A., Grollman, A. P., Patel, D. J., Eisenberg, M., and de los Santos, C. (1996) NMR solution structure of an oligodeoxynucleotide duplex containing the exocyclic lesion 3,N4etheno-2′-deoxycytidine opposite thymidine: comparison with the duplex containing deoxyadenosine opposite the adduct. Biochemistry 35, 13319-13327. (33) Cullinan, D., Eisenberg, M., and de los Santos, C. (1999) Solution structures of DNA duplexes containing the exocyclic lesion 3,N4etheno-2′-deoxycytidine. In The Role of Cyclic Adducts in Carcinogenesis and Mutagenesis (Singer, B., and Bartsch, H., Eds.) Vol. II, pp 179-189, International Agency for Research on Cancer Scientific Publication 150, International Agency for Research on Cancer, Lyon, France. (34) Boulard, Y., Cognet, J. A. H., and Fazakerley, G. V. (1997) Solution structure as a function of pH of two central mismatches, C‚T and C‚C, in the 29 to 39 K-ras gene sequence, by nuclear magnetic resonance and molecular dynamics. J. Mol. Biol. 268, 331-347. (35) Mol, C. D., Parikh, S. S., Putnam, C. D., Lo, T. P., and Tainer, J. A. (1999) DNA repair mechanisms for the recognition and removal of damaged DNA bases. Annu. Rev. Biophys. Biomol. Struct. 28, 101-128. (36) Stivers, J. T., Pankiewicz, K. W., and Watanabe, K. A. (1999) Kinetic mechanism of damage site recognition and uracil flipping by Escherichia coli uracil DNA glycosylase. Biochemistry 38, 952963. (37) Parikh, S. S., Mol, C. D., Slupphaug, G., Bharati, S., Krokan, H. E., and Tainer, J. A. (1998) Base excision repair initiation revealed by crystal structures and binding kinetics of human uracil-DNA glycosylase with DNA. EMBO J. 17, 5214-5226. (38) Barrett, T. E., Scha¨rer, O. D., Savva, R., Brown, T., Jiricny, J., Verdine, G. L., and Pearl, L. H. (1999) Crystal Structure of a Thwarted Mismatch Glycosylase DNA Repair Complex. EMBO J. 18, 6599-6609.
TX000040G
