Chemistry of neocarzinostatin-mediated cleavage of oligonucleotides

Competitive Ribose C5' and C4'. 5573. Hydroxylation. Hiroshi Sugiyama, Tsuyoshi Fujiwara, Hiroshi Kawabata, Nobuyuki Yoda,+. Noriaki Hirayama,+ and Is...
0 downloads 0 Views 715KB Size
J . Am. Chem. SOC.1992,114, 5573-5578

5573

Chemistry of Neocarzinostatin-Mediated Cleavage of Oligonucleotides. Competitive Ribose C5’ and C4’ Hydroxylation Hiroshi Sugiyama, Tsuyoshi Fujiwara, Hiroshi Kawabata, Nobuyuki Yoda,+ Noriaki Hirayama,+and Isao Saito* Contribution from the Department of Synthetic Chemistry, Faculty of Engineering, Kyoto University, Kyoto 606, Japan. Received December 30, I991

Abstract: A series of hexanucleotides possessing A-T, G-C, inosine (1)-C, and 2-aminoadenine (ANH2)-Tbase pairs at the 5’4de of the target thymine were prepared, and their selectivity for C-5’ and C4’ oxidation in the neocarzinostatin (NCS)-mediated degradation was investigated. Quantitative product analysis indicated that preferential C5’ oxidation of the deoxyribose moiety of the target T occurs at -5’-AT- and 5’-IT- sites, whereas C5’ and C4‘ oxidation occurs competitively at T of -5’-GT- and -5’-ANH2T-sites. On the basis of the cleavage data, a binding model that permits competitive hydrogen abstraction from C5’ and C4’ of the deoxyribose moiety has been proposed. Computer modeling of the hexamer duplex and the post-activated form of the NCS chromophore (3) suggested that, in the complex between d(GCATGC)2 and 3, the C6 carbon of 3 is close to the target C5’ @ro-S) hydrogen. In contrast, the tricyclic core of 3 is slightly lifted up so as to become closer to the C4’ hydrogen in the complex between duplex d(GCGTGC)/d(CGCACG) and 3 due to the van der Waals contact between the protruding guanine 2-amino group in the minor groove and the core moiety of 3.

Neocarzinostatin (NCS) is an antitumor antibiotic consisting of nonprotein chromophore (NCS-C) and its carrier protein.’ NCS-C undergoes irreversible reaction with thiols to generate a biradical species which is capable of cleaving DNA via hydrogen abstraction from the DNA sugar backbone with a high degree of base specificity (T > A >> C G) upon aerobic incubation.lb,cJ NCS-C consists of three main structural subunits, a substituted naphthoate group, an amino sugar (N-methyl-a-D-fucosamine) and a highly strained bicyclo[7.3.0]dodecadiendiyne epoxide unit (Figure l).3 NCS-C binds to double stranded DNA via intercalation of its naphthoate moiety and interaction of the diendiyne bicyclic core with the minor g r o ~ v e . ~ A ~ ,diradical ~,~ species 2 derived from the diendiyne epoxide moiety via nucleophilic addition of a thiol at C12 has been proposed to abstract hydrogen from the deoxyribose moiety of DNA to induce DNA strand cleavage in the presence of oxygen (Scheme I).lbs5 The prominent DNA damage is direct strand breaks at -5’-AT- sites resulting in the formation of thymidine 5’-aldehyde fragments at the 5’-termini and phosphate at the 3/-termini6 It has also been shown that oxidation at C1’ of the deoxyribose-giving 2- deoxyribonolactone abasic site also occurs at the -5’-AGC- sequence as a less prevalent lesion.’ Recently, partial incorporation of deuterium from C5’-deuterium-labeled oligonucleotide into the C6 position of 3 has been demonstrated.* We have very recently found that there are two distinct cycloaromatization pathways (paths A and B) in the activation of NCS-C by thiol under physiological conditions as outlined in Scheme IG9 Incubation of NCS-C with 2-mercaptoethanol in the presence of apoprotein in aqueous buffer solution produced a previously unobserved cyclization product 4 as a major product via ionic pathway (path B).9 We have also demonstrated that the use of hexanucleotides as a sequence-specific substrate for NCS provides a very useful tool for understanding the detailed chemistry of NCS-mediated DNA degradation.1° By using a self-complementary hexanucleotide d(GCATGC),, we were able to characterize the structure of the previously unidentified oxidized deoxyribose moiety associated with spontaneous free base release.Ioa We also observed that previously unobserved C4’ hydroxylation of the deoxyribose moiety does occur significantly at T, of a self-complementary hexanucleotide d(CIG2T3A4CSG6)2 in competition with C5’ oxidation at A4 (Scheme II).lob This type of C4’ hydroxylation has also been observed in NCS-mediated degradation of calf thymus DNA.” Specific detection methods

-

‘Tokyo Research Laboratories, Kyowa Hakko Kogyo Co., Ltd., Machida, Tokyo 194, Japan.

0002-786319211514-5573$03.00/0

for C4’ hydroxylated abasic sites recently developed in our laboratory have indicated that C4’ hydroxylation is estimated to be a minimum of 17% of the total event that occurred by the action of N C S on calf thymus DNA.” More recently Goldberg and co-workers have indicated that NCS-C abstracts the C4’ hydrogen of the T residue a t a d(GT) step in d(TCTTTGA), d(TTCTCATGTTTGA), and the HindIII-BamH1 restriction fragment (322 bp) of ~ B R 3 2 2 . l ~ In order to get insight into the binding geometry that permits competitive hydroxylation at C5’ and C4’ of the deoxyribose moiety, a series of hexanucleotides possessing A-T, G-C, inosine (1)-C, and 2-aminoadenine (ANH2)-Tbase pairs at the 5’-side of the target thymine were prepared and their selectivity for C5’ and C4’ oxidation in the NCS-mediated degradation was investigated (Figure 2). Careful analysis of NCS-mediated degradations of (1) (a) Ishida, N.; Miyazaki, K.; Kumagai, K.; Rikimaru, M. J . Antibiotics, 1965,18,68. (b) Goldberg, I. H. Free Radical Biol. Med. 1987,3, 41 and references therein. (c) Goldberg, I. H. Acc. Chem. Res. 1991,24, 191. (2) (a) Kappen, L. S.; Goldberg, I. H. Nucleic Acids Res. 1978,5, 2959 and references therein. (b) D’Andrea, A. D.; Haseltine, W. A. Proc. Natl. Acad. Sci. U.S.A.1978,75, 3608. (c) Lee, S. H.; Goldberg, I . H. Biochemistry 1989,28, 1019.

(3) (a) Edo, K.; Mizugaki, M.; Koide, Y.; Seto, H.; Fujihara, K.; Otake, N.; Ishida, N. Tetrahedron Lett. 1985,26, 331. (b) Edo, K.; Akiyama, Y.; Saito, K.; Mizugaki, M.; Koide, Y.; Ishida, N. J . Antibiof. 1986,39, 1615. (c) Myers, A. G.; Proteau, P. J.; Handel, T. M. J . Am. Chem. SOC.1988,110, 7212 and references therein. (4) Napier, M. A.; Goldberg, I. H. Mol. Pharm. 1983,23, 500. (5) (a) Myers, A. G. Tetrahedron Lett. 1987,28,4493. (b) Myers, A. G.; Proteau, P. J. J . Am. Chem. SOC.1989,1 1 1 , 1146. (6) (a) Kappen, L. S.; Goldberg, I. H.; Liesch, J. M. Proc. Narl. Acad. Sci. U.S.A. 1982,79, 744. (b) Kappen, L. S.; Goldberg, I. H. Biochemistry 1983, 22, 4872. (c) Kappen, L. S.; Goldberg, I . H. Nucleic Acids Res. 1985,13, 1637. (7) (a) Povirk, L. F.; Goldberg, I. H. Proc. Narl. Acad. Sci. U.S.A. 1985, 82, 3182. (b) Kappen, L. S.; Chen, C.-G.; Goldberg, I. H. Biochemistry 1988, 27, 4331. (c) Kappen, L. S . ; Goldberg, I. H. Biochemistry 1989,28, 1027. (d) Kappen, L. S.; Goldberg, I. H.; Wu, S . H.; Stubbe, J.; Worth, L.; Kozarich, J. W. J . Am. Chem. SOC.1990,112, 2797. (8) Meschwitz, S. M.; Goldberg, I. H. Proc. Natl. Acad. Sci. U.S.A. 1991, 88,3047. (9) Supivama. H.; Yamashita. K.; Nishi, M.; Saito, I . Tetrahedron Leu. 1992,33,-5i5. (10) (a) Kawabata, H.; Takeshita, H.; Fujiwara, T.; Sugiyama, H.; Matsuura. T.; Saito. I. Tetruhedron Lett. 1989, 30, 4263. (b) Saito. I.; Kawabata, H.; Fujiwara, T.; Sugiyama, H.; Matsuura, T. J . Am. Chem. SOC. 1989,1 1 1 , 8302. (1 I ) Sugiyama, H.; Kawabata, H.; Fujiwara, T.; Dannoue, Y.; Saito, I . J . Am. Chem.Soc. 1990,112, 5252. (12) (a) Kappen, L. S.; Goldberg, I. H.; Frank, B. L.; Worth, L., Jr.; Christnen, D. F.; Kozarich, J. W.; Stubbe, J. Biochemisfry 1991,30, 2034. (b) Frank, 8.L.; Worth, L., Jr.; Christner, D. F.; Kozarich, J. W.; Stubbe, J.; Kappen, L. S.; Goldberg, I. H. J . Am Chem. SOC.1991, 113, 2271.

0 1992 American Chemical Society

5574 J. Am. Chem. SOC.,Vol. 114, No. 14, 1992

Sugiyama et al.

Scheme I

4

Ro

R =

OH

R'

-CH2CO2CH3 (3) -CH2CH20H

Scheme I1

I

OpACG

ii) s.v. PDE ii$AP

T I

these oligomers indicated that the ratio of the selectivity for C5' vs C4' oxidation is strongly dependent on the depth of the minor groove at the 5'-side of the target thymine. Described herein are detailed experimental results, in conjunction with computermodeling studies, which may provide important insight into the NCS-DNA association and explain the base specificity for the DNA cleavage. In a typical experiment, a mixture containing a self-comple-

mentary hexanucleotide d(G1CZA3T4GSC6)2, NCS, and 4hydroxythiophenol (HTP)Io as an activator in Tris-HC1 buffer (pH 7.2) was incubated at 0 OC for 12 h under aerobic conditions. Direct HPLC analysis of the reaction mixture indicated a clean formation of d(GCA)p and thymidine 5'-aldehyde fragment d(T*GC) together with minor amounts of spontaneously released thymine and adenine, showing that NCS recognizes the -5'-ATsite and the thiol-activated NCS-C diradical abstracts hydrogen from the deoxyribose moiety specifically at T4of this hexamer.'& 96% (4'

3

3%. 5'

97%)

94% (4'

3

41% 5'

91% (5'

59%)

I

100%)

GCGTGC CGCACG

GCATGC CGTACG I A

I

6%

9% 78% (4'

37%

I'

56%

Figure 1. Structure of the neocarzinostatin chromophore (NCS-C).

6%.

5'

38%)

I

1

"I

GCATCG CGTAGC

=

4'

5%

I

5'=95K

GC~TCG CGTAGC t

22% (5'

=

100%)

Figure 2. Cleavage sites of NCS-mediated degradation of hexanucleotides. The arrows represent the location and extent of cleavage. Selectivity for C4' vs C5' oxidation is shown in parentheses.

Neocarrinostatin-Mediated Cleavage of Oligonucleotides

J. Am. Chem. SOC.,Vol. 114. No. 14, 1992 5575

Scheme 111

C5’ d(GCA)p

1

d(G1C2A3T4G5C6)

NCS I RSH IO2

+

+

H&f$

G ;

: T

T4 am&

T

1

OH

+ T OpGC

5 :N=C 6 :N=G 7 :N=T 8 :N=A 9 : N = 2-amino-A 10 : N = l

ii) s. v. PDE AP

Table I. Quantitative Analysis of Products Formed in Neocarzinostatin-Mediated Degradation of Hexanucleotides free base,“ pM pyridazine,bp M ratio A total 5(C) 6(G) 7(T) 8(A) 9(ANH2) lO(1) total (C5’vs C4’) substrate T 0.8 2.3 (30.4) 1.2 (1.3) 3.5 (31.7) 0 0 0 1 d(GCATGC)* 0.8 97:3 2 d(GCGTGC) 8.0 (23.2) 0 (1.6) 8.0 (24.8) 0 9.4 0.3 0.2 9.9 59:41 d(CGCACG) 1.3 1oo:o 0 0 0.9 0 3 d(GC1TGC) 4.3 (17.9) 0 (0) 4.3 (17.9) 0.4 d(CGCACG) 0 0 1.3 1.3 95:5 4 d(GCATCG) 1.4 (26.2) 0 (2.0) 1.4 (28.2) 0 d(CGTAGC) 5 d(GCANH2TCG) 11.0 (25.3) 0 (0)c 11.0 (25.3) 0 0 0 12.2 12.2 38:62 d(CGTAGC) 0 (0) 0 0 1oo:o a Spontaneously released bases were determined by HPLC. The values in parentheses are the amount of free bases after alkali treatment (0.5 N NaOH, 90 “C). Quantitated as corresponding nucleosides after treatment with snake venom phosphodiesterase and AP by reverse phase HPLC. Formation of 2-aminoadenine.

run

no.

When the reaction mixture was treated with hot alkali (90 ‘C, 5 min), quantitative release of thymine from d(T*GC) was observed. Therefore, the total event that occurred at T, was easily determined by quantitation of the total amount of released thymine after hot alkali treatment. The amount of the deoxyribose C4’ oxidation at T4 was determined by quantitation of A-pyridazine 8 after hydrazine treatment and subsequent enzymatic digestion as described previous1y.lobJ1 Under this condition, the 3’phosphoglycolate termini such as GCA-glycolate was not detected, suggesting that the C4’-hydroperoxide species formed by trapping of the C4’-radical by O2is efficiently reduced by HTP to afford the C4’-hydroxy species before hydroperoxide can fragment to phosphoglycolate. These results also indicate that the extent of C4’ oxidation at T4is correctly estimated by the amount of 8. No phosphoglycolate termini were detected in the cleavage of all duplex hexamers tested in this study utilizing HTP as an activator.12bTherefore, in all cases C4’ oxidation was assayed by the amount of pyridazine produced. It was also reported that no significant amount of phosphoglycolate termini is produced at T9 in the cleavage of d(TTCTCATGTTTGA) when HTP was used as an aCtivator.lh The portion of C5’ oxidation at T4was obtained from the total amount of released thymine after hot alkali treatment (total event occurring at T4) minus the amount of 8 (C4’ oxidation), which corresponds to the amount of d(GCA)p produced (Scheme 111). The amount of 8 was always less than that of spontaneously released thymine. This may be due to the Criegee type decomposition of C5’ hydroperoxide intermediate which was reported previously.l@ The ratio of C5’ vs C4’ oxidation at T4 of d(GCATGC) thus obtained was 97:3 (Table I, run 1). For the analysis of the cleavage of modified base-containing oligomers such as d(CGITCG)/d(CGCACG) and d(GCANH2TCG)/d(CGTAGC), 2-amino-3’-(3-pyridazinylmethyl)-2’-deoxyadenylate ( 9 ) and 3-pyridazinylmethyl 2’deoxyinosine-3’-monophosphate(10) were prepared independently.

Under the standard enzymatic digestion conditions for the assay of C4‘ hydroxylated abasic site, both 9 and 10 were stable. The selectivity ratios for C5’ vs C4’ oxidation of various hexanucleotides were examined as described above. The results are summarized in Figure 2 and Table I. When a heteroduplex d(GIC2G3T4G5C6)/d(Cl2GI 1Cl&9CsG7) was incubated with NCS under the same conditions, the d(GCGTGC) strand was selectively oxidized with a remarkable increase of spontaneously released thymine (run 2). In fact, a comparable amount of G-pyridazine 610b911 to that of the released thymine was detected. The selectivity ratio for C5’ vs C4’ oxidation at T4 was 59:41 (run 2). In a marked contrast, when d(G IC213T4G5C6) /d( C I ZGI ,C d 9 C 8 G 7 )possessing the I-C base pair instead of the G-C pair at the 5’-side of the target T4 was used as a substrate, cleavage at T, occurred exclusively via the C5’ pathway (run 3). These results indicate that the presence of guanine 2-amino group of G3 of d(GCGTGC)/d(CGCACG) dramatically increases the ratio of C4’ to C5’ oxidation. Since the guanine 2-amino group of the G-C base pair was protruded into the minor groove thus shallowing the minor groove, the binding of activated NCS-C to the minor groove of d(GCGTGC)/d(CGCACG) would be sterically more hindered and the binding orientation would be slightly different from that for the binding to d(GCITCG)/d(CGCAGC) or d(GCATGC)2 which does not possess the 2-amino group in the minor groove of the 5’-side of the cleavage site (Figure 3). A more distinct example for the dramatic effect of the 2-amino group in the minor groove on the ratio for C5’ vs C4’ oxidation was the cleavage of a heteroduplex d(GlC2A3NH2T4C5G6)/d(C,2G,,TloA9G8C7) (run 5 ) in comparison with the cleavage of d(GCATCG)/d(CGTAGC) (run 4). The former duplex has the 2-aminoadenine (ANH2)-Tbase pair instead of the A-T base pair of the latter. Thermodynamic experiments of the oligomers indicated that d(GCANH2TCG)/d(CGTAGC)forms a more stable

5516 J. Am. Chem. SOC.,Vol. 114, No. 14, 1992

Sugiyama et al.

major grwve

I

n

H

I

n'

minor gmove

A""~-T baDe pair

A-T bEW pair

n

0 .,,,,,,H-N /N

~

~

/

N

0 -

J ~F N - H~

..,.,,,.. ~ N

~ 5

~

~

+

j

~

b N , 0

N3

N