Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (220K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Bansal, M.
Right arrow Articles by Kozarich, J. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bansal, M.
Right arrow Articles by Kozarich, J. W.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 1997 Oxford University Press 1836-1845

Footnote

Mechanistic analyses of site-specific degradation in DNA-RNA hybrids by prototypic DNA cleavers

Mechanistic analyses of site-specific degradation in DNA-RNA hybrids by prototypic DNA cleavers Meena Bansal+, Jae S. Lee[sect], JoAnne Stubbe1 and John W. Kozarich*

Department of Chemistry and Biochemistry, University of Maryland, College Park, MD 20742, USA and 1Department of Chemistry, Massachusetts Institute of Technology, Cambridge, MA 02139, USA

Received November 25, 1996; Revised and Accepted February 20, 1997

ABSTRACT

Bleomycin (BLM) binding and chemistry are apparently sensitive to differences in nucleic acid conformation and could conceivably be developed as a probe for sequence-dependent elements of conformation. We report on the development of a new methodology to synthesize heterogeneous DNA-RNA hybrids of defined sequence and present the results of our comparative studies on the cleavage of DNA and DNA-RNA hybrids by four drugs: BLM, neocarzinostatin and esperamicins A1 and C. In the case of BLM with duplex DNA, purine-pyrimidine steps such as GT and GC, are consistently hit, as previously observed. However, in heterogeneous sequence hybrids, not all GC sites are recognized by the drug, although all GT sites are. Suppressed GC sites are consistently flanked by pyrimidines on both the 3' and 5' sides, suggesting that the BLM binding site in hybrids spans at least four bases. Kinetic isotope studies with specifically deuterated substrates (kH/kD = 1.2-4.0) and the effect of oxygen on the product profile are presented in support of a mechanism consistent with 4'-hydrogen abstraction in hybrids. The powerful double-labeled probe technique was extended to study the mechanism of action of other DNA degrading drugs on DNA-RNA hybrids. For neocarzinostatin, the sequence specificity lies in the AT-rich region for hybrids and is similar to that of DNA, however, the overall cleavage pattern for the hybrid is significantly different from that for the same sequence of DNA. In the hybrid, a stretch of AT residues is essential and the A sites are damaged to a greater extent than they are in DNA. However, no kinetic isotope effects are observed and, based on the product profile, the mechanism of degradation of the DNA strand of hybrids seems to be limited to abstraction of the 5'-hydrogen. For esperamicin A1, damage on the DNA strand of hybrids occurs exclusively via 5'-hydrogen abstraction in a non-rate determining step and primarily at A and T sites. Esperamicin C behaves similarily, exhibiting no isotope effects at 1', 4' and 5' positions. Overall, the differences observed in site-specific cleavage between the two substrates is proposed to be a result of conformational differences between the DNA strand of duplex DNA and DNA-RNA hybrids.

INTRODUCTION

Elucidation of the mechanisms of action of DNA cleaving agents, such as bleomycin (BLM) and the enediyne antibiotics, has been the focus of study of many laboratories. A central mechanistic theme is homolysis of carbon-hydrogen bonds in the deoxyribose moiety of DNA. Since the initial discovery of the antineoplastic activity of BLM (Fig. 1 ) (1 ), much research has focused on dissecting its mechanism of action. In the presence of Fe(II) and O2 or Fe(III) and H2O2, a ternary complex is formed with BLM resulting in the generation of activated BLM (2 -4 ). Recently, activated BLM has been identified by electrospray mass spectrometry as a ferric hydroperoxide and is the last detectable intermediate prior to DNA strand scission (5 ). Activated BLM is proposed to propagate a cascade of reactions via an initial rate limiting homolytic 4'-hydrogen abstraction step. Extensive mechanistic studies have supported a mechanism in which the 4'-carbon radical partitions to yield two monomeric products, the free base and the base propenal, and ultimately DNA strand scission products (7 -1 3 ). A range of 4'-hydrogen kinetic isotope effects are observed along the length of the DNA and are attributed to sequence-dependent variations in conformation, suggesting that the rate of site-specific cleavage is related to the polymorphic nature of DNA (14 ).

Another class of compounds, which include neocarzinostatin and the esperamicins (Fig. 1 ), contain an enediyne structure and can concomitantly cleave both strands of a DNA duplex. The esperamicins are a class of potent antitumor antibiotics that are characterized by the presence of a 1,5-diyne-3-ene as part of an unusual bicyclo[7.3.1] ring system. The drugs are activated by reduction of an allylic trisulfide followed by Michael addition of the resulting thiolate to the [alpha],[beta]-unsaturated ketone at the bridgehead of the bicyclic ring system and subsequent aromatization of the enediyne to a phenyl biradical (1 3 ,1 5 ). This carbon-centered biradical has been proposed to directly abstract a hydrogen from the deoxyribose of DNA to induce strand scission. Esperamicin A1 (esp A1) abstracts the 5'-hydrogen (1 6 ) and recent work by Yu and co-workers (17 ) suggests that a low level of 1'-hydrogen chemistry may also occur. Esperamicin C (esp C) abstracts both the 4'-hydrogen and 5'-hydrogen, resulting in double-strand DNA cleavage (1 6 ).


Figure 1. Stuctures of bleomycin, neocarzinostatin and esperamicins A1 and C.

Neocarzinostatin, an enediyne antibiotic isolated as a protein-bound complex, also generates a biradical species (NCS*) (18 ,19 ). NCS* shows high affinity for the 5'-hydrogens of A and T sites to generate 5'-aldehyde and 3'-phosphate termini (2 0,2 1 ), whereas 4'-hydrogen abstraction results in partitioning between a modifed abasic carbohydrate terminus and a 3'-phosphoglycolate terminus. The ratio of these two termini appears to be modulated by the structure of the thiol (2 2 ). Kinetic isotope effects have been reported for 4'- and 5'-hydrogen abstraction by neocarzinostatin (2 2 ,2 3 ).

Although DNA degradation by various drugs has received considerable attention, the mechanism of action of these compounds on DNA-RNA hybrids has been relatively unexplored. Since RNA is the key intermediate between the genetic information in DNA and expression of that information as protein, it is possible that the therapeutic activity of these drugs may also reside in their ability to cleave DNA-RNA hybrids. Haidle and Bearden first reported that BLM degrades only the DNA strand of hybrids (2 4 ). Our subsequent studies using DNA-RNA hybrid homopolymers confirmed that DNA strand scission occurs via 4'-hydrogen abstraction, but yielded different ratios of monomeric products (2 5 ,2 6 ). Morgan and Hecht have recently reported that both strands are cleaved in a substrate prepared by reverse transcription of single-stranded RNA (27 ). In addition, tRNA degradation by BLM has been reported to be conformation and not sequence dependent (28 ,29 ). Recently, Goldberg and co-workers have reported double-stranded damage in short oligonucleotide hybrids by neocarzinostatin (3 0).

We report here on a methodology to prepare heterogeneous DNA-RNA hybrids and on our results comparing site-specific cleavage in hybrids and in duplex DNA of identical sequence with BLM, neocarzinostatin and the esperamicins. Our results demonstrate that these agents cleave the DNA strand of DNA-RNA hybrids in a unique structure- and sequence-dependent manner as compared with duplex DNA. Possible explanations for the differences observed in site specificity and overall extent of cleavage are discussed.

MATERIALS AND METHODS

Materials

Blenoxane, a mixture of BLM A2 and B2, was a gift from Bristol-Meyers Squibb and was used without further purification. NTPs and dNTPs were purchased from Pharmacia and from US Biochemical Corp. (USB). Deuterated NaBH4 (99% atom D) and deuterated diborane (BD3[middot]THF, ~1 M) were purchased from Norell Inc. and Alfa respectively. [4'-2H]dTTP and [5'-di2H]dTTP were prepared as previously described (1 0,1 1 ,2 2 ). [4'-2H]dCTP was prepared as described by Zhang (3 1 ). [[gamma]-32P]ATP (>5000 Ci/mmol) was obtained from Amersham Inc.

Calf intestinal phosphatase, RQ1 RNase-free DNase, EcoRV, BamHI and recombinant ribonuclease inhibitor (rRNasin) were purchased from Promega. The cloning vector pT7/T3[alpha]-18, T7 RNA polymerase and RNase H- M-MLV reverse transcriptase were purchased from Bethesda Research Laboratories. T4 polynucleotide kinase and phenol:chloroform:isoamyl alcohol (25:24:1) mixture were obtained from USB. All primers were prepared by Oligos Inc. and acrylamide and N,N'-methylene bisacrylamide were from Research Organics. The Nuctrap columns used to remove excess nucleotides were from Stratagene.

Construction of pUM 346

DNA-RNA hybrid synthesis was accomplished by first constructing the plasmid pUM 346 via published cloning techniques (3 2 ) using the HindIII-BamHI fragment of pBR322 and the cloning vector pT7/T3[alpha]-18. This 346 bp region was chosen as the sequence of interest because it contains a large number of strong BLM cleavage sites. Plasmid pUM 346 contains the strong promoters for T3 and T7 RNA polymerase (Scheme 1).

Scheme 1. Strategy for the synthesis of DNA-RNA hybrids.

Preparation of RNA

To prepare RNA, pUM 346 was linearized with EcoRV, the site for which lies between the HindIII-BamHI restriction sites. The RNA strand was prepared in a 50 [mu]l reaction with 1.2 [mu]g linearized pUM 346. A typical reaction mixture contained 40 mM Tris-HCl, pH 8.0, 25 mM NaCl, 8 mM MgCl2, 2 mM spermidine-(HCl)3, 50 mM dithiothreitol (DTT) and 2 mM each NTP. Recombinant ribonuclease inhibitor (40 U) was added and, after incubating the reaction mixture for 5 min at 37oC, 250 U T7 RNA polymerase (BRL) was added. Transcription was carried out for 1 h and the resulting RNA and DNA mixture treated with 5 U RNase-free DNase. The reaction was terminated by the addition of 50 [mu]l 0.1 M EDTA and the proteins removed with two phenol:chloroform:isoamyl alcohol (25:24:1) extractions. The RNA was then precipitated via standard procedures and immediately used to prepare hybrids.

Preparation of DNA-RNA hybrids

The RNA was resuspended in 9 [mu]l water, mixed with 1 [mu]g primer complementary to the 3'-end of RNA, heated at 70oC for 10 min and then cooled on ice for primer annealing. In a final volume of 50 [mu]l, 1.7 mM dNTPs, 40 U rRNasin, 1000 U RNase H- M-MLV reverse transcriptase, 10 mM DTT, 50 mM Tris-HCl, pH 8.3, 75 mM KCl and 3 mM MgCl2 were added to the RNA-primer complex solution. The reaction was heated at 42oC for 90 min and the hybrid purified from excess primer on a 5% non-denaturing polyacrylamide gel. The gel was soaked in 1% (w/v) ethidium bromide solution and the fluorescent band corresponding to the hybrid cut out and the hybrid eluted into 10 mM Tris-HCl via standard procedures.

For 5'-end-labeling, 6 U T4 polynucleotide kinase was added to a 40 [mu]l solution containing hybrid, 40 U rRNasin, 50 mM Tris-HCl, pH 7.6, 10 mM MgCl2, 10 mM 2-mercaptoethanol and 180 [mu]Ci (>5000 Ci/mmol) [[gamma]-32P]ATP. The reaction was incubated at 37oC for 1 h. Hybrid was purified from free [[gamma]-32P]ATP via two procedures. First, the sample was loaded onto a Nuctrap nucleotide purification column (Stratagene) to remove the bulk of unreacted [[gamma]-32P]ATP. Second, since an accurate number of counts of hybrid probe per BLM reaction was required, all traces of [[gamma]-32P]ATP were removed by purification on a 5% non-denaturing polyacrylamide gel.

Preparation of deuterated DNA-RNA hybrids

In a procedure similar to that described above, a final concentration of 1.8 mM [4'-2H]dCTP or 2.9 mM [4'-2H]dTTP or [5'-di2H]dTTP was used in place of the protonated analog. The amount of RNA used in an individual hybrid preparation varied, ranging from 1 to 2.5 [mu]g/reaction.

Preparation of DNA

The plasmid pUM 346 linearized with EcoRV was dephosphorylated with calf intestinal phosphatase (CIP; Promega). Since the DNA was blunt-ended, the standard procedure for dephosphorylation was modified: 10 [mu]g DNA was incubated with 0.2 U CIP at 37oC for 15 min and at 56oC for another 15 min before a second aliquot of 0.2 U CIP was added. The mixture was again incubated at 37oC for 15 min and at 56oC for 15 min. The enzyme was heat inactivated at 65oC for 20 min and removed by phenol:chloroform:isoamyl alcohol (25:24:1) extraction twice. DNA was precipitated using standard techniques.

The DNA was radiolabeled under the same conditions as the hybrid described above. After the kinase reaction, however, and before precipitation, the DNA was reacted with BamHI to generate the labeled187 bp fragment used in BLM reactions and the longer 2500 bp labeled fragment was discarded. Three microliters of 3 M NaCl was added to the labeling mixture to optimize the reaction conditions for BamHI. These fragments were precipitated together and the 187 bp piece purified on a polyacrylamide gel as for the hybrids above.

Preparation of 4'-deuterated DNA

The DNA used to determine isotope effects was 187 bp and was made by PCR. The final reaction mixture contained 10 mM Tris-HCl, pH 8.8 at 25oC, 50 mM KCl, 0.1% Triton X-100, 2.5 mM MgCl2, 1 mM BamHI primer, 2 mM EcoRV primer, 10 ng HindIII-BanII template DNA, 5 U Taq polymerase, 0.2 mM non-deuterated nucleotides (except for the analogous deuterated nucleotide) and 0.36 mM [4'-2H]dCTP or 0.7 mM [4'-2H]dTTP in a final reaction volume of 100 [mu]l. PCR cycling was initiated by heating at 92oC for 1 min. Then the following temperature/time conditions were cycled 35 times: 92oC for 1 min, 37oC for 2 min and 74oC for 1.5 min. The DNA was end-labeled as above and then digested with BamHI, precipitated and purified as above. The sequence was confirmed by Maxam-Gilbert sequencing (3 3 ).

Drug-polynucleotide reaction conditions

BLM reaction mixtures contained 200 000 c.p.m. hybrid or 100 000 c.p.m. 32P-labeled DNA probe, 2.5 [mu]M BLM, 8.75 [mu]M FeSO4[middot]7H2O, 0.125 [mu]g/[mu]l salmon sperm DNA in 50 mM Tris-HCl, pH 7.5, and were performed as described previously (1 2 ). Neocarzinostatin reaction conditions were the same as those used earlier (2 2 ). For esp A1 and C (1 6 )the concentration of esp A1 was increased 25-fold and 15-fold for esp C in hybrid reactions as compared with those with DNA. The hybrid was sequenced using Maxam-Gilbert sequencing methods to determine sequence specificities of BLM (3 3 ).

Product analysis

Denaturing polyacrylamide gels (12 or 20% 30:1 acrylamide:N,N'-methylene bisacrylamide, 1* TBE, 7 M urea) were pre-run at 2000 V for 30 min at 50oC. Equal amounts of radioactivity for each reaction sample were loaded on the gel after heat denaturation at 90oC for 2 min and subsequently electrophoresed for an additional 2 h. The gels were transferred to Whatman 3MM chromatography paper for drying and autoradiographed by two methods: first, the conventional technique using Fuji AIF-RX film and, second, PhosphorImager 400A (Molecular Dynamics) analysis in order to integrate the volume of the bands at each cleavage site using ImageQuant software (Molecular Dynamics). This procedure has been previously described (2 2 ).

Data analysis

Cleavage ratios at corresponding sequence sites for hybrid versus DNA were calculated by taking ratios for volumes of the bands for cleavage of hybrid (rNdN) to the volumes of the same bands for DNA at a particular sequence site. These values were then normalized to an appropriate internal standard. For example, in Table 1 the rNdN/DNA ratios are normalized to T245. At this site there is about equal cleavage in both substrates. However, C205 was chosen as an internal control for studying [4'-2H]dCTP substitution effects because this site is within the primer region of the hybrid and is not deuterated. C223 was chosen as the internal control for [4'-2H]dTTP studies. (Note that for isotope studies with [4'-2H]dCTP with DNA, T214 was the internal control and not C205. Either site could have been used; the observation that both values were ~1 indicates that the method of analysis is accurate.)

Table 1 . Normalized ratios of cleavage of hybrid (rNdN) to duplex DNA
Sequence Site rNdN/DNA Percent cleavage
 

  

  

 DNA

 Hybrid
Suppressed sites
5'-CGCC-3'

 C210

 0.03 +- 0.04

 186

 4
5'-CGCT-3'

 C236

 0.15 +- 0.1

 162

 18
5'-TGCT-3'

 C230

 0.6 +- 0.05

 142

 90
5'-TGCT-3'

 C227

 1.0 +- 0.2

 98

 96
Unsuppressed sites
5'-AGCA-3'

 C205

 3.0 +- 0.1

 106

 328
5'-AGCG-3'

 C234

 5.5 +- 0.6

 30

 159
5'-GGCG-3'

 C223

 1.7 +- 0.6

 89

 193
5'-TGCA-3'

 C251

 2.4 +- 1.0

 128

 277
5'-TGCG-3'

 C243

 2.4 +- 0.5

 86

 218
Controls
5'-AGTC-3'

 T214

 1.1 +- 0.2

 131

 158
5'-CGTG-3'

 T225

 0.7 +- 0.2

 161

 118
5'-CGTT-3'

 T245

 1.0 +- 0.0

 100

 100
5'-TATA-3'

 T239

 5.8 +- 0.4

 80

 148
5'-TATG-3'

 T241

 1.7 +- 0.2

 87

 774
5'-TATG-3'

 T220

 2.9 +- 0.3

 56

 168
The site attacked by BLM is underlined in the first column. In the second colulmn, the site position in pBR322 is listed. The experiment was repeated five times and the ratios obtained, along with the standard deviation, are reported in the third column. A value of zero indicates 100% suppression in the hybrid at that site relative to DNA, whereas a value of 1 indicates equal hitting in the hybrid and DNA. This table is arranged in three groups: first, GC sites that are suppressed in hybrids; second, GC sites that have about equal or greater hitting in hybrids than in DNA; third, sites which are used as internal controls. The percent cleavages, with respect to the internal control T245, are listed for each site analyzed in both substrates in the last two columns.

RESULTS

Synthesis of DNA-RNA hybrids

We have developed a new enzymatic method designed to prepare long hybrids of defined sequence with fidelity (Scheme 1 ). Starting from the site of initiation of transcription, the length of the RNA transcribed from linearized plasmid pUM 346 with T7 RNA polymerase was 214 bases. After reverse transcription reactions, the prepared hybrid was 5'-end-labeled and sequenced by Maxam-Gilbert reaction: a ladder of bands corresponding to the sequence of the pBR322 insert from which pUM 346 was made was the affirmative test of hybrid formation. Simpler methods to confirm hybrid synthesis were not dependable, since this particular hybrid did not migrate significantly differently from single-stranded RNA on agarose or non-denaturing acrylamide gels. Acridine orange was also not useful in differentiating single- from double-stranded species due to the highly folded nature of this RNA.

Comparison of DNA duplex and DNA-RNA hybrid cleavage


Figure 2. Autoradiograph of a 20% denaturing acrylamide gel comparing the product profile of the HindIII-BamHI sequenced duplex DNA with DNA-RNA hybrids (rNdN) with BLM. The 5'-32P-end-labeled substrates were cleaved with Fe(II) and BLM as described in Materials and Methods and the products resolved by electrophoresis. Samples were analyzed prior to and after piperidine (pip) treatment as indicated. The doublets observed before piperidine treatment are oligomers containing 3'-phosphoglycolate (PG) termini and a 3'-termini that is alkali labile (ALM). The band representing ALM disappears into a band representing 3'-phosphate termini (P) after piperidine treatment. The three oligomeric products of BLM-mediated damage are resolved and found to be the same for both substrates, however, degradation is not observed to occur at C40 in the hybrid substrate. The numbers correspond to the plasmid map of pBR322.

Figure 2 shows an autoradiograph of a high resolution 20% polyacrylamide gel that resolves the oligomeric products formed by BLM-mediated cleavage at specific sites in DNA and in hybrids of the HindIII-BamHI sequence of pBR322. Comparison of the electrophoretic profile for a particular site of damage in the hybrid of the DNA strand was identical to that of the corresponding DNA duplex. For example, the two bands observed at thymidine 46 (T46; the number corresponds to the base position in pBR322) before base treatment were equally intense for both substrates. After treatment with 1 M piperidine, the migratory pattern changed and the slow moving alkali-labile moiety (ALM) disappeared, while a new, faster band appeared, which represented the 3'-phosphate (P) oligomer (12 ,14 ). The fastest moving band at each site represented the 3'-phosphoglycolate (PG) termini. (The piperidine-treated hybrid lane had fewer c.p.m. loaded onto the gel and was therefore faint compared with the other lanes, however, this does not affect the analysis.) However, although all GT sites were cleaved in both the duplex DNA and in the hybrid, no cleavage was observed at the GC40 site (the base underlined is the site of attack) in the hybrid. Another site, GC64 (not shown), also exhibited decreased damage in the hybrid.


Figure 3. Autoradiograph of a 12% denaturing acrylamide gel comparing the product profile of the EcoRV-BamHI sequenced duplex DNA with DNA-RNA hybrids (Hyb). The 5'-32P-end-labeled substrates were cleaved with Fe(II) and BLM as described in Materials and Methods, treated with piperidine and the products resolved by electrophoresis. At the bottom of the gel, the 3'-phoshate and 3'-phospholycolate termini are resolved, but migrate as one band further up the gel. Damage occurs with sequence specificity for 5'-purine-pyrimidine-3' sites, but variability in GC site damage is observed. The Maxam-Gilbert sequencing reactions are shown for sequence clarity and the numbers correspond to the plasmid map of pBR322.

The initial experiments with the HindIII-BamHI sequence of pBR322 above revealed that GT sites were recognized and cleaved to the same extent in both the hybrid and duplex DNA substrates and damage at GC sites was suppressed. This led to subsequent investigation of BLM mediated-cleavage in the hybrid of the EcoRV-BamHI sequence. This region contains more GC and GT dinucleotide steps within a shorter span, compared with the original sequence, which allows the investigation of a 50 bp region containing nine GC sites and six GT and AT sites. Figure 3 shows a representative gel (12% polyacrylamide) of BLM cleavage of hybrid (rNdN) versus duplex DNA. Table 1 lists the normalized ratios of cleavage of the hybrid (rNdN) to duplex DNA. A dramatic suppression of cleavage at some GC sites and an apparent reversal at others was observed. For example, at GC210 there was almost complete suppression, whereas at GC205 an apparent 3-fold greater cleavage was observed in the hybrid compared with DNA. At all suppressed GC sites, the bases immediately 3' and 5' of the dinucleotide step (5'- X GC X-3') were pyrimidines. At other GC sites the 3' base was always a purine and the 5' base a purine in 60% of the sites analyzed. 5'-AT-3' sites, which usually are not preferred hit sites in DNA, were cleaved significantly in the hybrid. For example, at AT239 damage was almost 6-fold more in the hybrid than in DNA.

Oxygen modulation experiments with bleomycin

The proposed mechanism of BLM-mediated DNA damage (3 4 ) correctly predicts that changes in the concentration of O2 affect the ratio of 3'-phosphate (P) and 3'-phosphoglycolate (PG) termini, with high O2 concentrations favoring PG termini. Acrylamide gels (20%) were used to compare the product profiles of reactions of the EcoRV-BamHI sequenced DNA fragment with the corresponding hybrid in which the reactions were performed at 4oC under 1 atm O2 (data not shown). At the three sites analyzed in DNA, the amount of PG production dramatically increased at elevated O2 concentrations and the formation of ALM and phosphate was minimal before and after base treatment, as previously observed (1 2 ). In the hybrid, the amount of PG produced increased with elevated oxygen at C205 and T214 and little phosphate production was detected. At C210, a suppressed site, very little damage was observed, consistent with the results reported above.

Under anerobic conditions, the trend in product formation in the hybrid of the HindIII-BamHI sequence was reversed: the amount of phosphate produced was greater than PG (Fig. 4 ). For example, at C40, although the amount of P and PG increased, the proportional increase in phosphate and ALM was much greater than PG. At T46, PG production completely disappeared, with a simultaneous increase in ALM; after base treatment the ALM was converted exclusively into 3'-phosphate.


Figure 4. Autoradiograph of a 20% denaturing acrylamide gel comparing the products of hybrid degradation under ambient and anerobic (Anr) conditions. The 5'-32P-end-labeled substrates were cleaved with Fe(II) and BLM as described in Materials and Methods and the products resolved by electrophoresis. Samples were analyzed prior to and after piperidine (Pip) treatment as indicated. The production of a 3'-ALM and 3'-phosphate (P; after piperidine treatment) increases under anerobic conditions and at some sites (GT46) 3'-phosphoglycolate termini formation is suppressed. This suggests that DNA and DNA-RNA hybrid degradation by BLM follow the same proposed mechanism (12).

Table 2 . Isotope effects on 4'-hydrogen abstraction at thymidylate and cytidylate sites by Fe(II)-BLM in DNA and hybrids as determined by volume integration of bands on a 12% denaturing polyacrylamide gel
Sequence Site Isotope effect (kH/kD)
 

  

 DNA ([4'-2H]dC)

 rNdC

 DNA ([4'-2H]dT)

 rNdT
C sites
5'-CGCC-3'

 C210

 1.8 +- 0.1

 1.2 +- 0.4

 0.9 +- 0.1

 1.4 +- 0.6
5'-CGCT-3'

 C236

 3.1 +- 1.2

 1.3

 0.8 +- 0.1

 1.2 +- 0.2
5'-TGCT-3'

 C230

 3.4 +- 0.7

 2.6 +- 0.5

 0.9 +- 0.1

 1.2 +- 0.3
5'-TGCT-3'

 C227

 1.9 +- 0.2

 1.3

 0.8 +- 0.1

 1.0 +- 0.2
5'-AGCA-3'

 C205

 0.8 +- 0.1

 1.0 +- 0.0

 0.7 +- 0.1

 1.2 +- 0.2
5'-AGCG-3'

 C234

 1.3 +- 0.6

 1.8

 0.7 +- 0.3

 1.0 +- 0.1
5'-GGCG-3'

 C223

 0.7 +- 0.2

 1.7 +- 0.2

 1.0 +- 0.0

 1.0 +- 0.0
5'-TGCG-3'

 C243

 1.6 +- 0.6

 1.5 +- 0.4

 1.1 +- 0.3

 0.8
T sites
5'-AGTC-3'

 T214

 1.0 +- 0.0

 0.8 +- 0.1

 1.6 +- 0.4

 2.3 +- 0.7
5'-CGTG-3'

 T225

 0.9 +- 0.1

 N.D.

 1.4 +- 0.2

 1.2 +- 0.2
5'-CGTT-3'

 T245

 0.9 +- 0.3

 0.8 +- 0.2

 2.2 +- 0.7

 1.4
5'-TATA-3'

 T239

 ND

 0.9

 ND

 4.0 +- 0.7
5'-TATG-3'

 T241

 ND

 0.8

 ND

 2.6 +- 0.2
5'-TATG-3'

 T220

 ND

 0.9 +- 0.2

 ND

 2.5 +- 0.2
ND, not determined.The data are the average with standard deviation of three to five experiments. Isotope effect values for DNA at AT sites were not determined due to poor resolution at these sites.

Kinetic isotope effects in hybrids and DNA with bleomycin

Isotope effects were measured at a number of cleavage sites utilizing substrates specifically deuterated at the 4'-position of deoxyribose residues. As shown in Table 2 , isotope effects at [4'-2H]thymidylate residues of GT dinucleotide steps in hybrids ranged from 1.2 to 2.3. In DNA, the values were from 1.4 to 2.2. Isotope effects at T sites of AT steps in hybrids were 2.5-4.0, but could not be accurately determined in DNA due to band compression. GC223 served as the internal control for both substrates.

Kinetic isotope effect values in DNA at 4'-deuterated cytidylate sites ranged from 1.3 to 3.4. In hybrids, the values ranged from 1.2 to 2.6. Interestingly, GC230, a suppressed site in hybrids, had the highest isotope effect in both substrates and GC236, another suppressed site, exhibited a larger isotope effect in DNA than in hybrids. However, for GC223, which is not a suppressed site, an isotope effect was observed only in hybrids. The internal controls for these reactions had values of 1.

Degradation of hybrids and DNA with neocarzinostatin


Figure 5. Autoradiograph of a 12% denaturing acrylamide gel comparing the product profiles of degradation of DNA and hybrids. The 5'-32P-end-labeled substrates were cleaved with NCS-chrom as described in Materials and Methods and the products resolved by electrophoresis. Samples were analyzed prior to (lanes A and C) and after piperidine (lanes B and D) treatment. The sequence specificity lies in the AT-rich region for hybrids and is similar to that of DNA, however, the overall cleavage pattern for the hybrid is significantly different from that for the same sequence of DNA.

DNA and the hybrid corresponding to the HindIII-BamHI fragment of pBR322, containing thymidine, [4'-2H]dTTP or [5'-di2H]dTTP and a 5'-32P-end-label, were prepared. The autoradiogram of the cleavage pattern of the duplex DNA and DNA-RNA hybrid is shown in Figure 5 . The cleavage pattern of DNA (Fig. 5 , lanes A and B) corroborates previous results which indicated that NCS* cleavage occurred preferentially at AT-rich regions (3 5 ). The damage at T sites exceeded the damage at A sites in DNA, in which typically A and T sites surrounded by purine residues were cleaved and bands corresponding to 3'-phosphate and 3'-phosphoglycolate were observed.

As in DNA, NCS* degrades hybrids with sequence specificity for AT-rich regions. However, the cleavage pattern for the hybrid was dramatically different to that for DNA (Fig. 5 , lanes C and D). Degradation of the hybrid by NCS* was restricted to a region with at least four A and/or T residues, whereas this was not necessary for DNA cleavage. In terms of chemistry, only 3'-phosphate fragments were observed prior to alkali treatment in the hybrid, with no 3'-phosphoglycolate fragment. No kinetic isotope effects for 4'- and 5'-hydrogens were observed (data not shown). In DNA, the deuterium partition isotope effect has been determined to be 2.4-5.5 and 1.0-2.6 respectively (2 2 ).

Degradation of hybrids and DNA with esperamicins A1 and C


Figure 6. Autoradiograph of a 20% denaturing acrylamide gel comparing the product profiles of DNA and hybrid degradation by esperamicins A1 and C, as indicated on the gel. The 5'-32P-end-labeled substrates were cleaved with drug as described in Materials and Methods and the products resolved by electrophoresis. Samples were analyzed prior to (Pip -, lanes A-D) and after piperidine (Pip +, lanes E-H) treatment as indicated. Lane B shows the presence of phosphoglycolate ends, which is associated with abstraction of the 4'-hydrogen for esperamicin C; esperamicin A1 chemistry is limited to 5'-hydrogen abstraction, since only 3'-phosphate termini are observed (lane A).

The HindIII-BamHI fragment of pBR322 was also used to investigate the mechanism of cleavage by the esperamicins in hybrids. As shown in Figure 6 , lane A, esp A1 favors cleavage at T sites in the DNA duplex, supporting earlier findings (3 6 ,37 ). Esp A1 also cleaved at A sites, but with lower frequency than at T sites. No phosphoglycolate ends were present and analysis of the [1'-2H]-, [4'-2H]- and [5'-2H]DNA did not yield significant isotope effects (1 6 ).

The DNA strand of the hybrid was degraded by esp A1 in a manner similar to that in duplex DNA. The similarity in cleavage profiles for both substrates suggests the mechanism of action of the drug to be the same. The only difference between DNA and the hybrid was that ~25 times more drug was necessary to cleave the hybrid compared with DNA.

Esp C degraded DNA with similar sequence specificity and reactivity to esp A1. However, unlike esp A1, esp C effects chemistry at the 4' position of the deoxyribose ring (1 6 ). Esp C degraded DNA in sets of three bases (the best examples are TTT and TTA) in which the middle base was the favored cleavage site. The first of the three base sequence only yielded 3'-phosphate ends. The second base, which was the most damaged site of the three, generated both phosphoglycolate and 3'-phosphate ends, although the phosphoglycolate end was generated in greater amounts. However, at the third base of such sequences, both ends were typically produced equally. There was no observed isotope effect with esp C at the 1'-, 4'- or 5'-sugar hydrogen position (data not shown). However, in the hybrid esp C does not react at the 4' position of the deoxyribose ring, as indicated by the lack of phosphoglycolate ends (Figure 6 , lane D).

DISCUSSION

Our initial studies with DNA-RNA hybrids used homopolymers to eliminate complicating sequence-dependent variations (2 5 ). We pursued the use of hybrids to evaluate the chemistry of DNA strand cleavage by antibiotics in a nucleic acid that is not a traditional B-form helix. The methodology developed to synthesize hybrids allowed for not only sequence-specific, but also for strand-specific substitution of base analogs and isotopes. Our present work compared a heterogeneous DNA-RNA hybrid with duplex DNA of the same sequence under identical drug reaction conditions. The results demonstrate the importance of nucleic acid conformation on site-specific drug cleavage.

In B-form DNA site-specific cleavage by BLM occurs predominately at 5'-GC-3' and 5'-GT-3' sites or, in general, at 5'-purine-pyrimidine-3' sites (38 ,39 ). Using the EcoRV-BamHI fragment of pBR322, we were able to corroborate these results, with most GC and all GT sites being preferentially hit (Fig. 3 ).In the hybrid, BLM-induced degradation occurred at all T residues following G residues, but some discrimination occurred at GC sites.

Although BLM-induced degradation sites on the hybrid were more limited than in the DNA duplex, the mechanism of strand cleavage appears to be the same for the hybrid and DNA duplex, for the following reasons. First, the electrophoretic profile at a site of damage in the hybrid was identical to that of the DNA duplex. For example, in Figure 2 for both substrates at T46 before base treatment the two bands observed represent oligomer with a 3'-phosphoglycolate terminus (the faster band) and oligomer with a 3'-ALM (the slower band). After treatment with 1 M piperidine, the ALM moiety disappeared with generation of a new faster band, which represents the 3'-phosphate oligomer (1 2 ). This is observed to occur with both substrates. Second, the ratio of these bands could be changed by varying the O2 concentration, which is consistent with 4'-hydrogen abstraction (1 2 ). Finally, kinetic isotope effects at 4' positions but not at the 1' position were observed (2 6 ). These data suggested that the BLM complexes generated from B-form DNA and from the hybrid converge to a common structural motif, in which the 4'-hydrogen is accessible to BLM attack. On the other hand, since some of the favorable C residues after G residues were comparatively less damaged in the hybrid than in DNA [i.e. 5'- CGCC-3' (C210)], the minor groove floor could be significantly different from that of B-form DNA, perhaps due to the structural modulation that must occur to accommodate the 2'-hydroxyl on the RNA strand of hybrids.

We suggest that two factors are major contributors to the variability in GC suppression: (i) flanking sequences; (ii) steric or electronic interference in BLM binding. In DNA-RNA hybrids the immediately flanking bases 3' and 5' of the GC sites were pyrimidines at all four suppressed GC sites [e.g. CGCC (C210)]. However, in sites where suppression did not occur, the base 5' of the GC site was a purine 60% of the time and the base 3' of the GC site was always a purine [e.g. AGCA (C205)]. This is in contrast to the studies of Murray and Martin with duplex DNA, which indicated that on a statistical basis the 3' base was not predictive of the degree of cleavage at a specific dinucleotide step (4 0). They did, however, conclude that the type of base immediately 5' of the purine-pyrimidine cleavage site affected the extent of damage. This, in combination with our results, suggests that flanking sequences contribute more to BLM site recognition in the hybrid than in duplex DNA. In addition, DNase I footprinting studies indicated that metallobleomycin binding spans at least 2-3 bp (4 1 ). The present results imply that BLM sequence specificity in the hybrid is due to a binding area encompassing 4 bp on the DNA strand.

The basis for GC suppression is difficult to clearly define at this time, since the binding constants for BLM to individual sequences in the hybrid have not been determined. However, on the basis of the average percent cleavage at each site of both DNA and the corresponding hybrid (Table 1 ), it appears that the average affinities of BLM for DNA and for the hybrid are comparable. Recent NMR and modeling studies from our laboratories suggest that a partial intercalation of the bithiazole group is a major determinant in the binding of BLM to DNA (4 2 -4 4 ). Preliminary modeling of BLM binding to a hybrid has not revealed any obvious steric interference with a similar intercalation mode. However, the possibility that steric barriers are present which prevent proper sequence-specific hydrogen bonding (i.e. hydrogen bonding of the 5'-G with the pyrimidine moiety of BLM) cannot be excluded. The proposition that the 2-amino group of the G on the RNA strand opposite the C cleavage site is responsible for the observed suppression is strengthened by the fact that its removal enhances cleavage at GC sites in hybrids (4 5 ).

Cleavage at the minor lesion site AT was also affected in DNA-RNA hybrids. As in DNA, purine-pyrimidine dinucleotide steps were preferred, but the order of preference changed. The reaction order was found to be AT > GT > GC in hybrids, in contrast to the order GT ~ GC > AT in duplex DNA (38 ,39 ).

The differences in the structure of DNA and the hybrids are also manifested in the chemistry mediated by NCS*. The identification of each band was established as previously described (2 2 ). The formation of 3'-phosphate fragments prior to alkali treatment suggested that the mechanism of DNA strand degradation in hybrids by NCS* involved abstraction of the 5'-hydrogen (2 2 ). Bands representing 3'-phosphoglycolate fragments were absent, suggesting that 4'-hydrogen abstraction does not occur.

The NCS chromophore cleaves the DNA strand of the hybrid and the structural differences between DNA and the hybrid alter the sequence specificity and the mechanism of degradation. The lack of an observed isotope effect in the hybrid at the 4'-position is consistent with a lack of 4'-hydrogen abstraction. However, the lack of a kinetic isotope effect at the 5'-position suggests that in the hybrid this step is not rate limiting. Furthermore, unlike DNA, in the hybrid the only pathway occurring is 5'-hydrogen abstraction. In DNA, observed KIEs were shown to partition between the 4' and 5' positions: isotope effects can be observed if one pathway is partially inhibited relative to a second pathway, thereby enhancing the latter (2 2 ). In DNA there have to be two modes of binding to observe such a partitioning. However, our results suggest that the hybrid-NCS* complex is of a single orientation.

Other work from our laboratory with a putative A-form 782 bp HindIII-BamHI fragment of DNA (4 6 ) and NCS* showed no evidence of 4'-hydrogen abstraction, as well as no isotope effect at the 4' or 5' positions (data not shown). The products of degradation were most probably derived from 5'-hydrogen abstraction. These results were in contrast to our published results with putative B-form DNA (2 2 ). More interestingly, however, the results with the 782 bp HindIII-BamHI fragment are similar to the results with DNA-RNA hybrids, since the DNA strands of hybrids have some A-form character (4 7 ). Since the minor groove is wider in A-form DNA than in B-form DNA, the hydrogens are exposed to different extents. It is feasible that the 4'-hydrogens in polymers with A-form character are positioned such that the NCS chromophore cannot abstract them, whereas the 5'-hydrogens remain accessible. Since A-form polymers also do not show an isotope effect, it is possible that when the minor groove is widened, as in the hybrid, the kinetics of the reaction change and some step other than hydrogen abstraction becomes rate limiting. In B-form polymers, the 4'- and 5'-hydrogens are positioned towards the interior of the groove, allowing abstraction from either position. A change in the geometry of these hydrogens, combined with the change in the minor groove dimensions in A-form type polymers, could cause a shift in the damage profile by NCS* between the two structurally different DNA probes. This is in sharp contrast to BLM, which abstracts the 4'-hydrogen in a rate determining step regardless of conformation (this is also true for the DNA studies with BLM using the 782 bp HindIII-BamHI fragment).

In duplex DNA, esp A1 abstracts the 5'-hydrogen in a non-rate determining step at predominately T sites. In hybrids, this preference remains the same, with some increased damage at A sites. Esp C, on the other hand, changes the chemistry of degradation from predominantly that of 5'-hydrogen abstraction to 4'- and 5'-hydrogen abstraction. Kinetic isotope effects are not observed for either substrate. The removal of the fucosyl anthrinilate moiety from esp A1 (to make esp C) effects 4'-hydrogen abstraction in DNA-RNA hybrids, as was proposed by Christner and co-workers to occur in DNA (1 6 ). This proposal was based on the need to remove the hindrance around the C-7 radical of esp A1 and, additionally, to allow the drug flexibility to orient itself in either of two possible ways in the minor groove. This results in the binding step being slower than a chemical step and explains the lack of observed kinetic isotope effects in DNA. Additionally, this also explains the observation of 4'- and 5'-hydrogen chemistry on the same strand (two independent reactions) or opposite strands (double-strand cleavage by the same drug molecule). The loss of this 4'-hydrogen activity in hybrids suggests that the binding/orientation of this drug in the minor groove has changed such that the second (C7) radical reacts to either make a product that is not detectable by our method or leads to some other minor side reaction. Again, this suggests that the local structure of the DNA strand in hybrids is significantly different from that in duplex DNA.

Summary

In order to study sequence-specific, conformation-dependent recognition and cleavage by various drugs we have compared DNA-RNA hybrids to duplex DNA. We have demonstrated that BLM cleaves hybrids with some similarities in sequence preference to DNA, as is expected if the structure of the DNA strand of the hybrid is similar to the analogous strand in duplex DNA. The varying degrees of suppression are explained by at least three interdependent factors: (i) the flanking sequences dictate unfavorable local conformations in the hybrids with greater consequences than in duplex DNA; (ii) there is steric interference with BLM binding; and/or (iii) the angle between the 4'-hydrogen and the activated drug is unfavorable.

Overall, in hybrids (as compared with DNA): BLM shows GC suppression with change in sequence preference; esp A1 shows an increased preference for A sites; esp C loses its 4'-hydrogen chemistry; NCS* chemistry is limited to the 5'-hydrogen in a non-rate determining step. Each of these drugs binds and reacts differently with the new substrates and they can conceivably be developed as probes for nucleic acid conformation.

Solution and structure studies on DNA-RNA hybrids have varied depending on the sequence analyzed, the conditions used and/or the choice of a model, i.e. homopolymer versus short oligomer. However, on the basis of 2D-NOE studies and J-coupling analysis, recent work has shown the conformation of the DNA strand to be neither A- or B-form, but an intermediate O4'-endo conformation (4 8 ). As a result, the chemical groups in the face of the minor groove are positioned to interact uniquely with different DNA binding proteins and drugs. Understanding the interactions of nucleic acids with different conformations increases our understanding of the dynamic role that they play in controlling interactions with small molecules, leading to an increased understanding of how drugs can be better designed to target their substrates.

REFERENCES

1 Umezawa,H., Maeda,K., Takeuchi,T. and Okami,Y. (1966) J. Antibiot. Ser. A, 19, 200.

2 Sausville,E.A., Peisach,J. and Horwitz,S.B. (1976) Biochem. Biophys. Res. Commun., 73, 814. MEDLINE Abstract

3 Sausville,E.A., Stein,R.W., Peisach,J. and Horwitz,S.B. (1978) Biochemistry, 17, 2746. MEDLINE Abstract

4 Burger,R.M., Peisach,J. and Horwitz,S.B. (1981) J. Biol. Chem., 256, 11636. MEDLINE Abstract

5 Sam,J.W., Tang,X.-J. and Peisach,J. (1994) J. Am. Chem. Soc., 116, 5250-5256.

6 Wu,J.C., Kozarich,J.W. and Stubbe,J. (1983) J. Biol. Chem., 258, 4694. MEDLINE Abstract

7 Wu,J.C., Kozarich,J.W. and Stubbe,J. (1985) Biochemistry, 24, 7562.

8 Wu,J.C., Stubbe,J. and Kozarich,J.W. (1985) Biochemistry, 24, 7569. MEDLINE Abstract

9 Rabow,L.E., Stubbe,J., Kozarich,J.W. and Gerlt,J.A. (1986) J. Am. Chem. Soc., 108, 130.

10 Ajmera,S., Massof,S. and Kozarich,J.W. (1986) J. Labelled Compd Radiopharm., 23, 963.

11 Ajmera,S., Wu,J.C., Stubbe,J., Worth,L.,Jr, Stubbe,J. and Kozarich,J.W. (1986) Biochemistry, 25, 6586. MEDLINE Abstract

12 Worth,L.,Jr, Frank,B.L., Christner,D.F., Absalon,M.J., Stubbe,J. and Kozarich,J.W. (1993) Biochemistry, 32, 2601. MEDLINE Abstract

13 Golik,J., Clardy,J., Dubay,G., Groenewald,G., Kawaguchi,H., Konishi,M., Krishnan,B., Onkuma,H., Saitoh,K.-I. and Doyle,T.W. (1987) J. Am. Chem. Soc., 109, 3461.

14 Kozarich,J.W., Worth,L.,Jr, Frank,B.L., Christner,D.F., Vanderwall,D.E. and Stubbe,J. (1989) Science, 245, 1396. MEDLINE Abstract

15 Long,B.H., Golik,J., Forenza,S., Ward,B., Rehfuss,R., Dabrowiak,J.C., Catino,J.J., Musial,S.T., Brookshire,K.W. and Doyle,T.W. (1989) Proc. Natl. Acad. Sci. USA, 86, 2. MEDLINE Abstract

16 Christner,D.F., Frank,L.F., Kozarich,J.W., Stubbe,J., Golik,J., Doyle,T.W., Rosenberg,I.E. and Krishnan,B. (1992) J. Am. Chem. Soc., 114, 8763.

17 Yu,L., Goldberg,I.H. and Dedon,P.C. (1994) J. Biol. Chem., 269, 4144-4151. MEDLINE Abstract

18 Goldberg,I.H. and Kappen,L.S. (1994) In Borders,D.B. and Doyle,T.W. (eds), Enediyne Antibiotics as Antitumor Agents. p. 327.

19 Myers,A.G., Proteau,P.J. and Jandel,T.M. (1988) J. Am. Chem. Soc., 110, 7212.

20 Kappen,L.S., Goldberg,I.H. and Liesch,J.M. (1982) Proc. Natl. Acad. Sci. USA, 79, 744.

21 Kappen,L.S. and Goldberg,I.H. (1983)Biochemistry, 22, 4872. MEDLINE Abstract

22 Frank,B.L., Worth,L.,Jr, Christner,D.F., Kozarich,J.W., Stubbe,J. and Goldberg,I.H. (1991) J. Am. Chem. Soc., 113, 2271.

23 Kappen,L.S., Goldberg,I.H., Frank,B.L., Worth,L.,Jr, Christner,D.F., Kozarich,J.W. and Stubbe,J. (1991) Biochemistry, 30, 2034. MEDLINE Abstract

24 Haidle,C.W. and Bearden,J.,Jr (1975) Biochem. Biophys, Res. Commun., 65, 815. MEDLINE Abstract

25 Krishnamoorthy,C.R., Vanderwall,D.E., Kozarich,J.W. and Stubbe,J. (1988) J. Am. Chem. Soc., 110, 2008.

26 Absalon,M.J., Krishnamoorthy,C.R., McGall,G., Kozarich,J.W. and Stubbe,J. (1992) Nucleic Acids Res., 20, 4179. MEDLINE Abstract

27 Morgan,M.A. and Hecht,S.M. (1994) Biochemistry, 33, 10286-10293. MEDLINE Abstract

28 Magliozzo,R.S., Peisach,J. and Ciriolo,M.R. (1989) Mol. Pharmacol., 35, 428-432. MEDLINE Abstract

29 Carter,B.J., de Vroom,E., Long,E.C., van der Marcel,G.A., van Boom,J.H. and Hecht,S.M. (1990) Proc. Natl. Acad. Sci. USA, 87, 9373. MEDLINE Abstract

30 Zeng,X., Xi,Z., Kappen,L.S., Tan,W. and Goldberg,I.H. (1995) Biochemistry, 34, 12435. MEDLINE Abstract

31 Zhang,G. (1993) Doctoral dissertation,University of Maryland.

32 Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

33 Maxam,A.M. and Gilbert,W. (1980) Methods Enzymol., 65, 499. MEDLINE Abstract

34 Stubbe,J. and Kozarich,J.W. (1987) Chem. Rev., 87, 1107.

35 Poon,R., Beerman,T.A. and Goldberg,I.H. (1977) Biochemistry, 16, 486. MEDLINE Abstract

36 Sugiura,Y., Uesawa,Y., Takahashi,Y., Kuwahara,J., Golik,J. and Doyle,T.W. (1989) Proc. Natl. Acad. Sci USA, 86, 7672. MEDLINE Abstract

37 Lu,M., Guo,Q., Krishnan,B., Golik,J., Rosenberg,I.E., Doyle,T.W. and Kallenbach,N.R. (1991)J. Biomol. Struct. Dyn., 9, 285. MEDLINE Abstract

38 Takeshita,M., Grollman,A.P., Ohtsubo,E. and Ohtsubo,H.(1978) Proc. Natl. Acad. Sci USA, 75, 5983. MEDLINE Abstract

39 D'Andrea,A. and Haseltine,W.A. (1978) Proc. Natl. Acad. Sci USA, 75, 3608.

40 Murray,V. and Martin,R.F. (1985) Nucleic Acids Res., 13, 1467. MEDLINE Abstract

41 Kuwahara,J. and Sugiura,Y. (1988) Proc. Natl. Acad. Sci. USA, 85, 2459. MEDLINE Abstract

42 Wu,W., Vanderwall,D.E., Lui,S.M., Tang,X.-J., Turner,C.J., Kozarich,J.W. and Stubbe,J. (1996) J. Am. Chem. Soc., 118,1268-1280.

43 Wu,W., Vanderwall,D.E., Turner,C.J., Kozarich,J.W. and Stubbe,J. (1996) J. Am. Chem. Soc., 118,1281-1294.

44 Stubbe,J., Kozarich,J.W., Wu,W. and Vanderwall,D.E. (1996) Acc. Chem. Res., 29, 322-330.

45 Bansal,M., Stubbe,J. and Kozarich,J.W. (1997) Nucleic Acids Res., 25, 1846-1853.

46 Mei,H.-Y. and Barton,J.K. (1988) Proc. Natl. Acad. Sci. USA, 85, 1339.

47 Arnott,S., Chandrasekaran,R., Millane,R.P. and Park,H.-S. (1986) J. Mol. Biol., 188, 631-640. MEDLINE Abstract

48 Salazar,M., Oleg,Y.F., Miller,J.M., Ribeiro,N.S. and Reid,B.R. (1993) Biochemistry, 32, 4207. MEDLINE Abstract

Return

*To whom correspondence should be addressed at present address: Merck Research Laboratories, Rahway, NJ 07065-0900, USA. Tel: +1 908 594 3249; Fax: +1 908 594 3695; Email: john_kozarich@merck.com

Present addresses: +ABL-Basic Research Program, Frederick Cancer Research and Development Center, Frederick, MD 21702-1201, USA and [sect]Pfizer, Central Research Division, Groton, CT 06340, USA
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?



This Article
Right arrow Abstract Freely available
Right arrow Print PDF (220K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Bansal, M.
Right arrow Articles by Kozarich, J. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Bansal, M.
Right arrow Articles by Kozarich, J. W.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?