Duplex recognition by oligonucleotides containing 2'-deoxy-2'-fluoro-D-arabinose and 2'-deoxy-2'-fluoro-D-ribose. Intermolecular 2'-OH-phosphate contacts versus sugar puckering in the stabilization of triple-helical complexes

To gain insight into the origins of the large binding affinity of RNA toward target duplexes, 2'-deoxy-2'-fluororibonucleic acid (2'F-RNA) and 2'-deoxy-2'-fluoroarabinonucleic acid (2'F-ANA) were tested for their ability to recognize duplex DNA, duplex RNA, and RNA-DNA hybrids. 2'F-RNA, 2'F-ANA, and the corresponding control single-stranded (ss) DNA strands were shown to form triple-helical complexes only with duplex DNA and hybrid DNA (Pu)-RNA (Py), but not with duplex RNA and hybrid RNA (Pu)-DNA (Py). In contrast, an RNA third strand recognized all four possible duplexes (DD, DR, RD, and RR) as previously demonstrated by Roberts and Crothers [(1992) Science 258, 1463-1466]. The 2'F-RNA (C3'-endo) strand exhibited significantly reduced affinity for duplexes compared to an unmodified RNA (C3'-endo) strand. These findings are consistent with the intermolecular 2'-OH-phosphate contact mechanism proposed by Escudé et al. [(1993) Nucleic Acids Res. 24, 5547-5553], as a ribo 2'-F atom should not interact with a negatively charged phosphate. In addition, they emphasize the role of the 2'-OH ribose as a general recognition and binding determinant of RNA. The 2'-F arabino modification (2'F-ANA, C2'-endo) led to a considerable increase in the binding affinity for duplex DNA, as compared to those of DNA and 2'F-RNA third strands. This is likely to be the result of a greater population of C2'-endo pucker of the 2'F-ANA compared to DNA. The enhancement observed for 2'F-ANA strands toward duplex DNA is comparable to that observed with 2'-OMe RNA. Since 2'F-ANA has been shown to be more resistant to nuclease degradation than DNA, these results are likely to stimulate experimental work on arabinose derivatives in laboratories concerned with targeting DNA sequences in vivo ("antigene" strategy).     

Single-Stranded DNA and RNA Targeted Triplex-Formation: UV,CD and Molecular Modeling Studies of Foldback Triplexes Containing Different RNA, 2′-OMe-RNA and DNA Strand Combinations

Abstract

We studied the influence of different 2′-OMe-RNA and DNA strand combinations on single strand targeted foldback triplex formation in the Py.Pu:Py motif using ultraviolet (UV) and circular dichroism (CD) spectroscopy, and molecular modeling. The study of eight combinations of triplexes (D D:D, R* D:D, D D:R*, R* D:R*, D R:D, R* R:D, DR:R*, and R*-R:R*; where the first, middle, and last letters stand for the Hoogsteen Pyrimidine, Watson-Crick [WC] purine and WC pyrimidine strands, respectively, and D, R and R* stand for DNA, RNA and 2′-OMe-RNA strands, respectively) indicate more stable foldback triplex formation with a DNA purine strand than with an RNA purine strand. Of the four possible WC duplexes with RNA/DNA combinations, the duplex with a DNA purine strand and a 2′-O-Me-RNA pyrimidine strand forms the most thermally stable triplex, although its thermal stability is the lowest of all four duplexes. Irrespective of the duplex combination, a 2′-OMe-RNA Hoogsteen pyrimidine strand forms a stable foldback triplex over a DNA Hoogsteen pyrimidine strand confirming the earlier reports with conventional and circular triplexes. The CD studies suggest a B-type conformation for an all DNA homo-foldback triplex (D.D.D), while hetero-foldback triplex spectra suggest intermediate conformation to both Atype and B-type structures. A novel molecular modeling study has been carried out to understand the stereochemical feasibility of all the combinations of foldback triplexes using a geometric approach. The new approach allows use of different combinations of chain geometries depending on the nature of the chain (RNA vs. DNA).     

Recognition of nucleic acid double helices by homopyrimidine 2', 5'-linked RNA.

We have studied the effect of a 2',5'-RNA third strand backbone on the stability of triple helices with a 'pyrimidine motif' targeting the polypurine strand of duplex DNA, duplex RNA and DNA/RNA hybrids. Comparative experiments were run in parallel with DNA and the regioisomeric RNA as third strands adopting the experimental design of Roberts and Crothers. The results reveal that 2',5'-RNA is indeed able to recognize double helical DNA (DD) and DNA (purine):RNA (pyrimidine) hybrids (DR). However, when the duplex purine strand is RNA and the duplex pyrimidine strand is DNA or RNA (i.e. RD or RR), triplex formation is not observed. These results exactly parallel what is observed for DNA third strands. Based on T m data, the affinities of 2',5'-RNA and DNA third strands towards DD and DR duplexes were similar. The RNA third strand formed triplexes with all four hairpins, as previously demonstrated. In analogy to the arabinose and 2'-deoxyribose third strands, the possible C2'- endo pucker of 2',5'-linked riboses together with the lack of an alpha-2'-OH group are believed to be responsible for the selective binding of 2',5'-RNA to DD and DR duplexes, over RR and RD duplexes. These studies indicate that the use of other oligonucleotide analogues will prove extremely useful in dissecting the contributions of backbone and/or sugar puckering to the recognition of nucleic acid duplexes.     

Pyrimidine motif triplexes containing polypurine RNA or DNA with oligo 2'-O-methyl or DNA triplex forming oligonucleotides

Triplex forming oligonucleotides (TFOs) are potentially useful in targeting RNA for antisense therapeutic applications. To determine the feasibility of targeting polypurine RNA with nuclease-resistant oligonucleotides, TFOs containing 2'-deoxy or 2'-O-methyl (2'-OMe) backbones, designed to form pyrimidine motif triplexes with RNA, were synthesized. TFOs were made which can form trimolecular triplexes, or bimolecular, 'clamp' triplexes with polypurine RNA and DNA. It was found that the relative stabilities of the triplexes formed followed the order: M.DM(clamp)>D.DD approximately M.DD>M. RM>D.DM>M.RD approximately M.DM, where M is a 2'-OMe, D is a DNA and R is an RNA backbone. The third strand is listed first, separated by a dot from the purine strand of the Watson-Crick duplex, followed by the pyrimidine strand of the duplex. The results described here provide insight into the feasibility of using TFOs containing a 2'-OMe backbone as antisense agents.     

Structural basis of cleavage by RNase H of hybrids of arabinonucleic acids and RNA

The origins of the substrate specificity of Escherichia coli RNase H1 (termed RNase H here), an enzyme that hydrolyzes the RNA strand of DNA-RNA hybrids, are not understood at present. Although the enzyme binds double-stranded RNA, no cleavage occurs with such duplexes [Lima, W. F., and Crooke, S. T. (1997) Biochemistry 36, 390]. Therefore, the hybrid substrates may not adopt a canonical A-form geometry. Furthermore, RNase H is exquisitely sensitive to chemical modification of the DNA strands in hybrid duplexes. This is particularly relevant to the RNase H-dependent pathway of antisense action. Thus, only very few of the modifications currently being evaluated as antisense therapeutics are tolerated by the enzyme, among them phosphorothioate DNA (PS-DNA). Recently, hybrids of RNA and arabinonucleic acid (ANA) as well as the 2'F-ANA analogue were shown to be substrates of RNase H [Damha, M. J., et al. (1998) J. Am. Chem. Soc. 120, 12976]. Using X-ray crystallography, we demonstrate here that ANA analogues, such as 2'F-ANA [Berger, I., et al. (1998) Nucleic Acids Res. 26, 2473] and [3.3.0]bicyclo-ANA (bc-ANA), may not be able to adopt sugar puckers that are compatible with pure A- or a B-form duplex geometries, but rather prefer the intermediate O4'-endo conformation. On the basis of the observed conformations of these ANA analogues in a DNA dodecamer duplex, we have modeled a duplex of an all-C3'-endo RNA strand and an all-O4'-endo 2'F-ANA strand. This duplex exhibits a minor groove width that is intermediate between that of A-form RNA and B-form DNA, a feature that may be exploited by the enzyme in differentiating between RNA duplexes and DNA-RNA hybrids. Therefore, the combination of the established structural and functional properties of ANA analogues helps settle existing controversies concerning the discrimination of substrates by RNase H. Knowlegde of the structure of an analogue that exhibits enhanced RNA affinity while not interfering with RNase H activity may prove helpful in the design of future antisense modifications.     

Presence of divalent cation is not mandatory for the formation of intramolecular purine-motif triplex containing human c-jun protooncogene target

Modulation of endogenous gene function, through sequence-specific recognition of double helical DNA via oligonucleotide-directed triplex formation, is a promising approach. Compared to the formation of pyrimidine motif triplexes, which require relatively low pH, purine motif appears to be the most gifted for their stability under physiological conditions. Our previous work has demonstrated formation of magnesium-ion dependent highly stable intermolecular triplexes using a purine third strand of varied lengths, at the purine?pyrimidine (Pu?Py) targets of SIV/HIV-2 (vpx) genes (Svinarchuk, F., Monnot, M., Merle, A., Malvy, C., and Fermandjian, S. (1995) Nucleic Acids Res. 23, 3831-3836). Herein, we show that a designed intramolecular version of the 11-bp core sequence of the said targets, which also constitutes an integral, short, and symmetrical segment (G(2)AG(5)AG(2))?(C(2)TC(5)TC(2)) of human c-jun protooncogene forms a stable triplex, even in the absence of magnesium. The sequence d-C(2)TC(5)TC(2)T(5)G(2)AG(5)AG(2)T(5)G(2)AG(5)AG(2) (I-Pu) folds back twice onto itself to form an intramolecular triple helix via a double hairpin formation. The design ensures that the orientation of the intact third strand is antiparallel with respect to the oligopurine strand of the duplex. The triple helix formation has been revealed by non-denaturating gel assays, UV-thermal denaturation, and circular dichroism (CD) spectroscopy. The monophasic melting curve, recorded in the presence of sodium, represented the dissociation of intramolecular triplex to single strand in one step; however, the addition of magnesium bestowed thermal stability to the triplex. Formation of intramolecular triple helix at neutral pH in sodium, with or without magnesium cations, was also confirmed by gel electrophoresis. The triplex, mediated by sodium alone, destabilizes in the presence of 5'-C(2)TC(5)TC(2)-3', an oligonucleotide complementary to the 3'-oligopurine segments of I-Pu, whereas in the presence of magnesium the triplex remained impervious. CD spectra showed the signatures of triplex structure with A-like DNA conformation. We suggest that the possible formation of pH and magnesium-independent purine-motif triplexes at genomic Pu?Py sequences may be pertinent to gene regulation.     

Intramolecular triple-helix formation at (PunPyn).(PunPyn) tracts: recognition of alternate strands via Pu.PuPy and Py.PuPy base triplets.

Triple-helical DNA shows increasing potential for applications in the control of gene expression (including therapeutics) and the development of sequence-specific DNA-cleaving agents. The major limitation in this technology has been the requirement of homopurine sequences for triplex formation. We describe a simple approach that relaxes this requirement, by utilizing both Pu.PuPy and Py.PuPy base triplets to form a continuous DNA triple helix at tandem oligopurine and oligopyrimidine tracts. [Triplex formation at such a sequence has been previously demonstrated only with the use of a special 3'-3' linkage in the third strand [Horne, D. A., & Dervan, P. B. (1990) J. Am. Chem. Soc. 112, 2435-2437].] Supporting evidence is from chemical probing experiments performed on several oligonucleotides designed to form 3-stranded fold-back structures. The third strand, consisting of both purine and pyrimidine blocks, pairs with purines in the Watson-Crick duplex, switching strands at the junction between the oligopurine and oligopyrimidine blocks but maintaining the required strand polarity without any special linkage. Although Mg2+ ions are not required for the formation of Pu.PuPy base triplets, they show enhanced stability in the presence of Mg2+. In the sequences observed. A.AT triplets appear to be more stable than G.GC triplets. As expected, triplex formation is largely independent of pH unless C+.GC base triplets are required.     

Oligonucleotide-directed triple helix formation at adjacent oligopurine and oligopyrimidine DNA tracts by alternate strand recognition.

A significant limitation to the practical application of triplex DNA is its requirement for oligopurine tracts in target DNA sequences. The repertoire of triplex-forming sequences can potentially be expanded to adjacent blocks of purines and pyrimidines by allowing the third strand to pair with purines on alternate strands, while maintaining the required strand polarities by combining the two major classes of base triplets, Py.PuPy and Pu.PuPy. The formation of triplex DNA in this fashion requires no unusual bases or backbone linkages on the third strand. This approach has previously been demonstrated for target sequences of the type 5'-(Pu)n(Py)n-3' in intramolecular complexes. Using affinity cleaving and DNase I footprinting, we show here that intermolecular triplexes can also be formed at both 5'-(Pu)n(Py)n-3' and 5'-(Py)n(Pu)n-3' target sequences. However, triplex formation at a 5'-(Py)n(Pu)n-3' sequence occurs with lower yield. Triplex formation is disfavored, even at acid pH, when a number of contiguous C+.GC base triplets are required. These results suggest that triplex formation via alternate strand recognition at sequences made up of blocks of purines and pyrimidines may be generally feasible.     

High-resolution NMR study of a synthetic DNA-RNA hybrid dodecamer containing the consensus pribnow promoter sequence: d(CGTTATAATGCG).r(CGCAUUAUAACG)

The nonsymmetrical double-helical hybrid dodecamer d(CGTTATAATGCG).r(CGCAUUAUAACG) was synthesized with solid-phase phosphoramidite methods and studied by high-resolution 2D NMR. The imino protons were assigned by one-dimensional nuclear Overhauser methods. All the base protons and H1', H2', H2", H3', and H4' sugar protons of the DNA strand and the base protons, H1', H2', and most of the H3'-H4' protons of the RNA strand were assigned by 2D NMR techniques. The well-resolved spectra allowed a qualitative analysis of relative proton-proton distances in both strands of the dodecamer. The chemical shifts of the hybrid duplex were compared to those of the pure DNA double helix with the same sequence (Wemmer et al., 1984). The intrastrand and cross-strand NOEs from adenine H2 to H1' resonances of neighboring base pairs exhibited characteristic patterns that were very useful for checking the spectral assignments, and their highly nonsymmetric nature reveals that the conformations of the two strands are quite different. Detailed analysis of the NOESY and COSY spectra, as well as the chemical shift data, indicate that the RNA strand assumes a normal A-type conformation (C3'-endo) whereas the DNA strand is in the general S domain but not exactly in the normal C2'-endo conformation. The overall structure of this RNA-DNA duplex is different from that reported for hybrid duplexes in solution by other groups (Reid et al., 1983a; Gupta et al., 1985) and is closer to the C3'-endo-C2'-endo hybrid found in poly(dA).poly(dT) and poly(rU).poly(dA) in the fiber state (Arnott et al., 1983, 1986).     

Oligonucleotide-minor groove binder conjugates and their complexes with complementary DNA: effect of conjugate structural factors on the thermal stability of duplexes

Synthetic polycarboxamide minor groove binders (MGB) consisting of N-methylpyrrole (Py), N-methylimidazole (Im), N-methyl-3-hydroxypyrrole (Hp) and beta-alanine (beta) show strong and sequence-specific interaction with the DNA minor groove in side-by-side antiparallel or parallel orientation. Two MGB moieties covalently linked to the same terminal phosphate of one DNA strand stabilize DNA duplexes formed by this strand with a complementary one in a sequence-specific manner, similarly to the corresponding mono-conjugated hairpin structures. The series of conjugates with the general formula Oligo-(L-MGB-R)m was synthesized, where m = 1 or 2, L = linker, R = terminal charged or neutral group, MGB = -(Py)n-, -(Im)n- or -[(Py/Im)n-(CH2)3CONH-(Py/Im)n-] and I < n < 5. Using thermal denaturation, we studied effects of structural factors such as m and n, linker L length, nature and orientation of the MGB monomers, the group R and the backbone (DNA or RNA), etc. on the stability of the duplexes. Structural factors are more important for linear and hairpin monophosphoroamidates than for parallel bis-phosphoroamidates. No more than two oligocarboxamide strands can be inserted into the duplex minor groove. Attachment of the second sequence-specific parallel ligand [-L(Py)4R] to monophosphoroamidate conjugate CGTTTATT-L(Py)4R leads to the increase of the duplex Tm, whereas attachment of [-L(Im)4R] leads to its decrease. The mode of interaction between oligonucleotide duplex and attached ligands could be different (stacking with the terminal A:T pair of the duplex or its insertion into the minor groove) depending on the length and structure of the MGB.     

Protection of DNA sequences by triplex-bridge formation.

We have demonstrated that the DNA sequence between two triplex-forming polypurine.polypyrimidine (Pu.Py) tracts was protected from DNA modifying enzymes upon formation of triplex DNA structures with an oligodeoxyribonucleotide in which two triplex-forming Pu or Py tracts were placed at the termini (triplex-bridge formation). In model experiments, when two triplex structures were formed between double-stranded DNA with the sequence (AG)17-(N)18-(T)34, and an oligodeoxyribonucleotide, (T)34-(N)18-(GA)17, not only the Pu.Py tracts but also the 18 bp non-Pu.Py sequence in the duplex DNA between the tracts was protected from restriction enzymes, HpaII methylase and DNase I. This protection occurred only when both of the Pu.Py tracts were involved as triplexes. The length of the tracts could be as short as 21 bp, while the difference in length between the non-Pu.Py sequences on the duplex and the oligodeoxyribonucleotide should be within 10 nucleotides. The efficiency of protection was enhanced in the presence of a cationic detergent, cetyltrimethylammonium bromide, during triplex formation. Protection was also observed with another type of the triplex bridge formed between (G)34 and (T)34 tracts with an oligodeoxyribonucleotide, (T)34-(N)20-(G)34. These findings suggest that the protection of specific DNA sequences from enzymes by triplex-bridge formation can be applied to any DNA sequence by placing it between two triplex-forming sequences.     

Exclusion of RNA strands from a purine motif triple helix.

Research concerning oligonucleotide-directed triple helix formation has mainly focused on the binding of DNA oligonucleotides to duplex DNA. The participation of RNA strands in triple helices is also of interest. For the pyrimidine motif (pyrimidine.purine.pyrimidine triplets), systematic substitution of RNA for DNA in one, two, or all three triplex strands has previously been reported. For the purine motif (purine.purine.pyrimidine triplets), studies have shown only that RNA cannot bind to duplex DNA. To extend this result, we created a DNA triple helix in the purine motif and systematically replaced one, two, or all three strands with RNA. In dramatic contrast to the general accommodation of RNA strands in the pyrimidine triple helix motif, a stable triplex forms in the purine motif only when all three of the substituent strands are DNA. The lack of triplex formation among any of the other seven possible strand combinations involving RNA suggests that: (i) duplex structures containing RNA cannot be targeted by DNA oligonucleotides in the purine motif; (ii) RNA strands cannot be employed to recognize duplex DNA in the purine motif; and (iii) RNA tertiary structures are likely to contain only isolated base triplets in the purine motif.     

Probing the relationship between RNA-stimulated ATPase and helicase activities of HCV NS3 using 2'-O-methyl RNA substrates

The hepatitis C virus (HCV) NS3 protein contains an amino terminal protease (NS3 aa. 1-180) and a carboxyl terminal RNA helicase (NS3 aa. 181-631). NS3 functions as a heterodimer of NS3 and NS4A (NS3/4A). NS3 helicase, a nucleic acid stimulated ATPase, can unwind RNA, DNA, and RNA:DNA duplexes, provided that at least one strand of the duplex contains a single-stranded 3' overhang (this strand of the duplex is referred to as the 3' strand). We have used 2'-O-methyl RNA (MeRNA) substrates to study the mechanism of NS3 helicase activity and to probe the relationship between its helicase and RNA-stimulated ATPase activities. NS3/4A did not unwind double-stranded (ds) MeRNA. NS3/4A unwinds hybrid RNA:MeRNA duplex containing MeRNA as the 5' strand but not hybrid duplex containing MeRNA as the 3' strand. The helicase activity of NS3/4A was 50% inhibited by 40 nM single-stranded (ss) RNA but only 35% inhibited by 320 nM ss MeRNA. Double-stranded RNA was 17 times as effective as double-stranded MeRNA in inhibiting NS3/4A helicase activity, while the apparent affinity of NS3/4A for ds MeRNA differed from ds RNA by only 2.4-fold. However ss MeRNA stimulated NS3/4A ATPase activity similar to ss RNA. These results indicate that the helicase mechanism involves 3' to 5' procession of the NS3 helicase along the 3' strand and only weak association of the enzyme with the displaced 5' strand. Further, our findings show that maximum stimulation of NS3 ATPase activity by ss nucleic acid is not directly related to procession of the helicase along the 3' strand.     

Repairing the sickle cell mutation. I. Specific covalent binding of a photoreactive third strand to the mutated base pair.

A DNA third strand with a 3'-psoralen substituent was designed to form a triplex with the sequence downstream of the T.A mutant base pair of the human sickle cell beta-globin gene. Triplex-mediated psoralen modification of the mutant T residue was sought as an approach to gene repair. The 24-nucleotide purine-rich target sequence switches from one strand to the other and has four pyrimidine interruptions. Therefore, a third strand sequence favorable to two triplex motifs was used, one parallel and the other antiparallel to it. To cope with the pyrimidine interruptions, which weaken third strand binding, 5-methylcytosine and 5-propynyluracil were used in the third strand. Further, a six residue "hook" complementary to an overhang of a linear duplex target was added to the 5'-end of the third strand via a T(4) linker. In binding to the overhang by Watson-Crick pairing, the hook facilitates triplex formation. This third strand also binds specifically to the target within a supercoiled plasmid. The psoralen moiety at the 3'-end of the third strand forms photoadducts to the targeted T with high efficiency. Such monoadducts are known to preferentially trigger reversion of the mutation by DNA repair enzymes.     

DNA, RNA and hybrid RNA-DNA oligomers of identical sequence: structural and dynamic differences

A 27-mer sequence was synthesised as DNA duplex (DD), RNA duplex (RR), and RNA-DNA (RD) hybrid in order to characterise their structural and dynamic features. The hydrodynamic radius (Rh) and the rise (b) values of the three samples were consistent with the conformations predicted by CD analysis. The value of the torsional constant (alpha) of the samples containing RNA was approximately twice that of the DD sample and followed the order: DD < RD < RR. The same order was observed in the thermodynamic stability and in the reduction of the electrophoretic mobility. gamma-Ray footprinting analysis was carried out to resolve the individual strand conformation in the hybrid. The RNA strand preserved its conformation, while the DNA strand showed local deformations mainly at TA and TG steps.     

Solution structure of an arabinonucleic acid (ANA)/RNA duplex in a chimeric hairpin: comparison with 2′-fluoro-ANA/RNA and DNA/RNA hybrids

Hybrids of RNA and arabinonucleic acid (ANA) as well as the 2′-fluoro-ANA analog (2′F-ANA) were recently shown to be substrates of the enzyme RNase H. Although RNase H binds to double-stranded RNA, no cleavage occurs with such duplexes. Therefore, knowledge of the structure of ANA/RNA hybrids may prove helpful in the design of future antisense oligonucleotide analogs. In this study, we have determined the NMR solution structures of ANA/RNA and DNA/RNA hairpin duplexes and compared them to the recently published structure of a 2′F-ANA/RNA hairpin duplex. We demonstrate here that the sugars of RNA nucleotides of the ANA/RNA hairpin stem adopt the C3′-endo (north, A-form) conformation, whereas those of the ANA strand adopt a ‘rigid’ O4′-endo (east) sugar pucker. The DNA strand of the DNA/RNA hairpin stem is flexible, but the average DNA/RNA hairpin structural parameters are close to the ANA/RNA and 2′F-ANA/RNA hairpin parameters. The minor groove width of ANA/RNA, 2′F-ANA/RNA and DNA/RNA helices is 9.0 ± 0.5 Å, a value that is intermediate between that of A- and B-form duplexes. These results rationalize the ability of ANA/RNA and 2′F-ANA/RNA hybrids to elicit RNase H activity.     

Structure of bleomycin-induced DNA double-strand breaks: predominance of blunt ends and single-base 5' extensions

In order to examine the structure of bleomycin-induced DNA double-strand breaks, defined-sequence DNA was labeled in each strand at a single restriction site and treated with bleomycin. Various double-stranded fragments resulting from bleomycin-induced double-strand breaks were isolated, denatured, and run on sequencing gels to determine the sites of cleavage in each strand. For virtually every double-strand break, the cleavage site in one strand was a pyrimidine in a G-Py sequence, reflecting a specificity similar to that of bleomycin-induced single-strand cleavage. However, the cleavage site in the complementary strand was seldom a G-Py sequence, and was usually a site where single-strand cleavage was infrequent. When the sequence at the double-strand break was G-Py-Py', the break at Py was usually accompanied by a break at the base directly opposite Py, resulting in blunt ends. When the sequence was G-Py-Pu, the break at Py was usually accompanied by a break at the base opposite Pu, resulting in single-base 5' extensions. Double-strand breaks with 3' extensions, such as would result from cleavage of two C residues in a self-complementary G-C sequence, were conspicuously absent. These data provide further evidence that bleomycin-induced double-strand breaks do not result from coincidence of independent site-specific single-strand breaks.(ABSTRACT TRUNCATED AT 250 WORDS)