GpC Sites Research Articles

P14.02 Influence of individual CpG methylation status on outcome in adult patients with glioblastoma multiforme receiving alkylating agent treatment

Abstract Background: Methylation of the O-6-methylguanine-DNA methyltransferase (MGMT) promotor causes gene silencing and has been associated with a favourable prognosis in patients with glioblastoma multiforme (GBM) receiving alkylating chemotherapy. However, analysis of MGMT promotor methylation is usually reported as a cut-off depending on the results of the correspondent CpG site testing. This approach disregards a possible heterogeneity concerning the methylation status within the individual CpG sites and its possible association with prognosis in GBM patients. The current study aimed at elucidating the association between methylation of CpG sites 74–98 within the MGMT promotor region and outcome in GBM patients receiving alkylating agents. Material and Methods: Individual methylation status of 230 patients with histologically proven GBM following concomitant radio-chemotherapy with TMZ after stereotactic biopsy or open tumor resection (OTR) was assessed by the Sanger sequencing (Sseq) approach. Methylation of CpG sites 74–98 within the MGMT promotor region was defined according to a ratio of cytosine /thymine peak >50%. The total number of methylated CpG sites as well as clinical factors such as age, Karnofsky Performance Score (KPS) and mode of surgical procedure were correlated with outcome using proportional hazards models. In a subset of 34 patients, a correlation between individual CpG methylation and MGMT mRNA expression was performed. Results: Median progression-free (PFS) and overall survival (OS) were 7.8 and 14.6 months, respectively. Alongside younger age, KPS> 80 and OTR, the cumulative total number of methylated loci within the CpG sites 74–98 was strongly associated with both PFS and OS and retained its prognostic influence on outcome in multivariate models (p <0.001). Furthermore, a linear coherence between the total number of methylated CpG sites 74–98 and survival parameters could be observed. Moreover, low number of methylated CpG sites was observed in tumor specimen with a high mRNA expression and vice versa (Spearman correlation coefficient: -0.62). Conclusion: In contrast to the concept of dichotomizing the MGMT promotor status into ‘methylated’ and ‘non-methylated’, our approach shows a clear heterogeneity within the methylation status of the CpG sites 74–98 within the GBM tumor specimens. Our data suggest a strong correlation between outcome and the total number of methylated CpG sites, thus an up-front analysis of the individual GpC site methylation status prior to initiation of alkylating chemotherapy might help to improve treatment response in GBM patients.

Potential Therapeutic Advantages of Doxorubicin when Activated by Formaldehyde to Function as a DNA Adduct-Forming Agent.

Doxorubicin has been in use as a key anticancer drug for forty years, either as a single agent or in combination chemotherapy. It functions primarily by interfering with topoisomerase II activity but in the presence of formaldehyde, it forms adducts with DNA, mainly with the exocyclic amine of guanine at GpC sites and these adducts are more cytotoxic than topoisomerase II induced damage. High levels of adducts form spontaneously from the endogenous level of formaldehyde in tumour cells (1,300 adducts per cell after a 4 hr treatment with doxorubicin), but substantially higher levels form with the addition of exogenous sources of formaldehyde, such as formaldehyde releasing prodrugs. The enhanced cytotoxicity of adducts has been confirmed in mouse models, with adduct-forming conditions resulting in much improved inhibition of tumour growth, as well as cardioprotection. Doxorubicin cardiotoxicity has been attributed to topoisomerase II poisoning, and the cardioprotection is consistent with a mechanism switch from topoisomerase II poisoning to covalent adduct formation. Although the adducts have a half-life of less than one day, a population remains as essentially permanent lesions. The capacity of doxorubicin to form adducts offers a range of potential advantages over the conventional use of doxorubicin (as a topoisomerase II poison), including: enhanced cell kill; tumour-selective activation, hence tumour-selective cell kill; decreased cardiotoxicity; decreased resistance to prolonged doxorubicin treatment. There is therefore enormous potential to improve clinical responses to doxorubicin by using conditions which favour the formation of doxorubicin-DNA adducts.

The structural basis of actinomycin D–binding induces nucleotide flipping out, a sharp bend and a left-handed twist in CGG triplet repeats

The potent anticancer drug actinomycin D (ActD) functions by intercalating into DNA at GpC sites, thereby interrupting essential biological processes including replication and transcription. Certain neurological diseases are correlated with the expansion of (CGG)n trinucleotide sequences, which contain many contiguous GpC sites separated by a single G:G mispair. To characterize the binding of ActD to CGG triplet repeat sequences, the structural basis for the strong binding of ActD to neighbouring GpC sites flanking a G:G mismatch has been determined based on the crystal structure of ActD bound to ATGCGGCAT, which contains a CGG triplet sequence. The binding of ActD molecules to GCGGC causes many unexpected conformational changes including nucleotide flipping out, a sharp bend and a left-handed twist in the DNA helix via a two site-binding model. Heat denaturation, circular dichroism and surface plasmon resonance analyses showed that adjacent GpC sequences flanking a G:G mismatch are preferred ActD-binding sites. In addition, ActD was shown to bind the hairpin conformation of (CGG)16 in a pairwise combination and with greater stability than that of other DNA intercalators. Our results provide evidence of a possible biological consequence of ActD binding to CGG triplet repeat sequences.

The binding of the Co(II) complex of dimeric chromomycin A3 to GC sites with flanking G:G mismatches

Some neurological diseases are correlated with expansion of (CXG)n trinucleotide repeats, which contain many contiguous GpC flanked by mismatched X/X base pair. This study focused on the binding of the Co(II) complex of dimeric chromomycin A3(Chro), CoII(Chro)2, to DNA with CXG trinucleotide repeats. The present study showed that GC sites with flanking G:G mismatches provide an excellent binding site for CoII(Chro)2 as shown by surface plasmon resonance and fluorescence analysis, compared to GC sites with flanking A:A, T:T, or C:C mismatches. In addition, we measured the ability of CoII(Chro)2 to act on the hairpin DNA of (CGG)16. We observed that CoII(Chro)2 could stabilize and trap the cruciform conformation of (CGG)16. Furthermore, two CoII(Chro)2 molecules may bind at the two GpC sites separated by at least one GC site in the hairpin structure of (CGG)16. In a synthetic self-priming DNA model, 5′-(CGG)16(CCG)6-3′, CoII(Chro)2 can interfere with the expansion process of CGG triplet repeats, as shown by a gel electrophoretic expansion assay. Here, we first report the acting of CoII(Chro)2, the groove-binding drug, to trinucleotide repeats. Our results provide the possible biological consequence of CoII(Chro)2 bound to CGG triplet repeat sequences.

7-Aminoactinomycin Binding to DNA Sequences Lacking GpC Sites: A Thermodynamic and Kinetic Study

The interaction of 7-aminoactinomycin (7AAMD) with selected DNA sequences (TAGTTA, R5, HP5, and HP1) of different lengths and secondary structures, all containing a 5'-TAGT-3' block, was studied at an ionic strength of 0.02 M and pH 7.7 by means of fluorescence equilibrium and kinetic (stopped-flow) measurements. Both approaches indicated that the antibiotic binds strongly to both the single-stranded and hairpin (HP1) structures, although the sequences lacked the canonical GpC sites favored by actinomycin. Binding isotherms and initial rate analyses revealed that the binding stoichiometry was 1:1 in all cases. While the single-stranded sequences displayed a simple monoexponential kinetic behavior, the binding of 7AAMD to HP1 at <30 degrees C was biphasic and could be rationalized in terms of a sequential formation of two isomeric bound forms or alternatively in terms of an ssHP1-hpHP1 equilibrium, with both HP1 forms reacting with 7AAMD. The rates of complex dissociation induced by the detergent SDS were also measured. After correction of the kinetic traces for spurious effects that can be attributed to the SDS, monoexponential traces were obtained, with relaxation times in agreement with the kinetics of complex formation.

Mutations of Different Molecular Origins Exhibit Contrasting Patterns of Regional Substitution Rate Variation

Transitions at CpG dinucleotides, referred to as “CpG substitutions”, are a major mutational input into vertebrate genomes and a leading cause of human genetic disease. The prevalence of CpG substitutions is due to their mutational origin, which is dependent on DNA methylation. In comparison, other single nucleotide substitutions (for example those occurring at GpC dinucleotides) mainly arise from errors during DNA replication. Here we analyzed high quality BAC-based data from human, chimpanzee, and baboon to investigate regional variation of CpG substitution rates.We show that CpG substitutions occur approximately 15 times more frequently than other single nucleotide substitutions in primate genomes, and that they exhibit substantial regional variation. Patterns of CpG rate variation are consistent with differences in methylation level and susceptibility to subsequent deamination. In particular, we propose a “distance-decaying” hypothesis, positing that due to the molecular mechanism of a CpG substitution, rates are correlated with the stability of double-stranded DNA surrounding each CpG dinucleotide, and the effect of local DNA stability may decrease with distance from the CpG dinucleotide.Consistent with our “distance-decaying” hypothesis, rates of CpG substitution are strongly (negatively) correlated with regional G+C content. The influence of G+C content decays as the distance from the target CpG site increases. We estimate that the influence of local G+C content extends up to 1,500∼2,000 bps centered on each CpG site. We also show that the distance-decaying relationship persisted when we controlled for the effect of long-range homogeneity of nucleotide composition. GpC sites, in contrast, do not exhibit such “distance-decaying” relationship. Our results highlight an example of the distinctive properties of methylation-dependent substitutions versus substitutions mostly arising from errors during DNA replication. Furthermore, the negative relationship between G+C content and CpG rates may provide an explanation for the observation that GC-rich SINEs show lower CpG rates than other repetitive elements.

Crystal structure of actinomycin D bound to the CTG triplet repeat sequences linked to neurological diseases.

The potent anticancer drug actinomycin D (ActD) acts by binding to DNA GpC sequences, thereby interfering with essential biological processes including replication, transcription and topoisomerase. Certain neurological diseases are correlated with expansion of (CTG)n trinucleotide sequences, which contain many contiguous GpC sites separated by a single base pair. In order to characterize the binding of ActD to CTG triplet repeat sequences, we carried out heat denaturation and CD analyses, which showed that adjacent GpC sequences flanking a T:T mismatch are preferred ActD-binding sites, and that ActD binding results in a conformational transition to A-type structure. The structural basis of the strong binding of ActD to neighboring GpC sites flanking a T:T mismatch was provided by the crystal structure of ActD bound to ATGCTGCAT, which contains a CTG triplet sequence. Binding of two ActD molecules to GCTGC causes a kink in the DNA helix. In addition, using a synthetic self-priming DNA model, 5'-(CAG)4(CTG)(16)-3', we observed that ActD can trap the cruciform or duplexes of (CTG)n and interfere with the expansion process of CTG triplet repeats as shown by gel electrophoretic expansion assay. Our results may provide the possible biological consequence of ActD bound to CTG triplet repeat sequences.

Actinomycin D binds to d(TGTCATG) with 2:1 drug to duplex stoichiometry.

Despite the apparent single-stranded conformation and the absence of a GpC site, d(TGTCATTG) has been found to bind strongly to actinomycin D (ACTD) with 1:1 drug to strand binding stoichiometry. A hairpin binding model was speculated in which the planar phenoxazone chromophore inserts at the GTC site by pushing out the T-base while the terminal G folds back to form a G small middle dotC base pair so that the 3'-sides of both G-bases stack on the opposite faces of the phenoxazone plane. However, it was also suggested that a slipped duplex binding with similar binding principle could also be operative at higher DNA concentrations. To support such a contention, ACTD binding studies were made with d(TGTCATG) and related oligomers. This heptamer differs from the parent octamer d(TGTCATTG) by a mere removal of a T-base which should result in an enhancement of dimeric duplex formation and a concomitant reduction in monomeric hairpin contribution. It was found that ACTD binds well to d(TGTCATG) with 1 drug to 1 strand (or 2 drugs to 1 duplex) binding stoichiometry. These results are consistent with a slipped duplex binding model in which a dimeric duplex is formed at the self-complementary CATG tetranucleotide sequence with extruding TGT ends. Two drug molecules are bound at both ends of the duplex by pushing out the T-bases of GTC's so that the opposite faces of each phenoxazone are stacked by the 3'-sides of the two G-bases on opposite strands. Such a model provides a ready explanation for the observed enhancement in ACTD binding to d(TGTCATGTC) and d(TGTCATGTCA), where additional base pairs at the ends will stabilize GTC/GTC binding sites, and to d(TGTCAATTG) in which two additional base pairs facilitate the slipped-duplex formation. The observed ACTD affinity reductions for oligomers containing GTTC instead of GTC are also consistent with the T-base displacement model. These findings greatly expand the repertoire of ACTD binding to DNA and may have important implications on understanding the transcription inhibitory activities of this drug.

Binding of actinomycin D to DNA oligomers of CXG trinucleotide repeats.

Actinomycin D (ACTD) binding propensities of DNA with CXG trinucleotide repeats were investigated using oligomers of the form d[AT(CXG)n = 2-4AT] and their corresponding heteroduplexes, where X = A, C, G, or T. These oligonucleotides contain -CXGCXG-, -CXGCXGCXG-, and -CXGCXGCXGCXG- units that can form homoduplexes containing one, two, and three GpC binding sites, respectively, with flanking X/X mismatches. The corresponding heteroduplexes contain these same sites with flanking Watson-Crick base pairs. It was found that oligomers with X = G exhibit weak ACTD affinities whereas those with X not equal to G and n = 3 exhibit unusually strong ACTD binding affinities with binding constants ranging from 2.3 x 10(7) to 3.3 x 10(7) M-1 and binding densities of approximately 1 drug molecule/strand (or 2/duplex). These binding affinities are considerably higher than those of their shorter and longer counterparts and are about 2- and 10-fold stronger than the corresponding CAG.CTG and CGG.CCG heteroduplexes, respectively. The CTG-containing oligomer d[AT(CTG)3AT] stands out as unique in having its ACTD dissociation kinetics being dominated by a strikingly slow process with a characteristic time of 205 min at 20 degrees C, which is 100-fold slower than d[AT(CAG)3AT], nearly 10-fold slower than the corresponding heteroduplex, and considerably slower than d[AT(CTG)2AT] (63 min) and d[AT(CTG)4AT] (16 min). The faster dissociation rate of the n = 4 oligomer compared to its n = 2 counterpart is in apparent contrast with the observed 10-fold stronger ACTD binding affinity of the former. It was also found that d[AT(CCG)3AT] exhibits the slowest dissociation rate of the CGG/CCG series, being more than an order of magnitude slower than that of its heteroduplex (tau slow of 43 vs 2 min). The finding that a homoduplex d[AT-CXG-CXG-CXG-AT]2 can bind two ACTD molecules tightly is significant since it was thought unlikely for two consecutive GpC sites separated by a single T/T mismatch to do so.

Dissociation Kinetics of Actinomycin D from Individual GpC Sites in DNA

We have examined the kinetics of dissociation of actinomycin from GpC sites in several DNA fragments containing synthetic DNA inserts, by a variation of the footprinting technique. Complexes of the ligand with radiolabelled DNA fragments were dissociated by adding a large excess of unlabelled calf thymus DNA. Samples were removed from this mixture at subsequent time intervals and subjected to DNase I footprinting. The rate of disappearance of the footprints varied considerably between the GpC sites located in different sequence environments. Actinomycin dissociates more slowly from GpC sites flanked by (AT)n than An.Tn. Within regions of alternating AT, TGCA represents a better binding site than AGCT, and CGCA is a better binding site than GGCA. GpC sites flanked by (AC)n.(GT)n present good binding sites; in this context, dissociation from CGCG is faster than from TGCA.

A novel DNA structure induced by the anticancer bisplatinum compound crosslinked to a GpC site in DNA.

The bifunctional platinum compound, [(trans-PtCI(NH)3)2)2(H2N(CH2)4NH2)]2+, forms a stable adduct with the self-complementary DNA oligomer CATGCATG, with the two platinum atoms coordinated at the N7 positions of the two symmetrical G4 nucleotides. The NMR-derived structure shows that the DNA octamer forms a novel hairpin structure with the platinated G4 residue adopting a syn conformation and the guanine base in the minor groove. Two such hairpins stack end-over-end and are linked together by the butanediamine tether to form a dumbbell structure. Such unusual structural distortion is different from that of the anticancer drug cisplatin-DNA adduct and may provide clues to explain the distinct biological activities of the two compounds.

Mutagenicity and mutation spectra of 2-acetylaminofluorene at frameshift and base-substitution alleles in four DNA repair backgrounds of Salmonella

We used colony probe hybridization procedures to determine the mutations in ∼ 600 revertants of the −1 frameshift allele hisD3052 and ∼ 200 revertants of the base-substitution allele hisG46 of Salmonella typhimurium induced by 2-acetylaminofluorene (2-AAF) in the presence of Aroclor-induced rat liver S9. 2-AAF was primarily a frameshift mutagen, exhibiting 5 times more frameshift than base-substitution activity. The only frameshift mutation 2-AAF induced at the hisD3052 allele was a hotspot (−2) deletion within the sequence CGCGCGCG. The addition of the pKM101 plasmid had a small effect on the mutagenic potency of 2-AAF at this allele in a uvr + background and no effect on the mutation spectra in either a uvr + or uvr − background. The small amount of base-substitution activity exhibited by 2-AAF at the hisG46 allele required the presence of both the pKM101 plasmid and the uvrB mutation. The base substitutions were G · C → T · A transversions (86%) and G · C → A · T transitions (14%), and 85% of the substitutions were at the second position of the CCC target of the hisG46 allele; the remainder were at the first position. We propose that the hotspot frameshift may be initiated by N-acetyl-2-aminofluorene adducts located at the C(8) position of any of the guanines except the first one in the CGCGCGCG hotspot sequence. The mutation might then result from correct incorporation of cytosine opposite the adducted guanine, followed by a 2-base slippage according to our recently proposed correct-incorporation/slippage model. The hotspot mutation may also result from a 2-AAF-induced B- to Z-DNA transition at the repeating GpC site as well as by the action of enzymes involved in DNA metabolism, such as DNA resolvases or topoisomerases, on DNA structures that have been distorted by 2-AAF adducts. The small amount of 2-AAF-induced base-substitution activity may be due to mispairing of adenine opposite the minor aminofluorene adduct at the C(8) position of guanine.

Does adriamycin induce interstrand cross-links in DNA?

Under nonenzymatic conditions in vitro, Adriamycin appears to form interstrand cross-links with DNA over 1-2 days. This is the first report of such Adriamycin-induced interstrand cross-links in vitro. Cross-links were measured by a fluorescence based renaturation assay and also by gel electrophoresis. Both procedures revealed an increase of cross-linking with reaction time and with increasing Adriamycin concentration and a 5-6-fold enhancement in the presence of Fe3+ ions. The cross-link contains the Adriamycin chromophore, with a lambda max of 508 nm, intercalated at the GpC site of cross-linking. Maximal stoichiometry of the cross-link was one per 11-20 bp. The cross-link appears to involve adducts of the Adriamycin chromophore linked to the N2 of guanine, with no indication that N7 of guanine is involved. Given that the mode of action of Adriamycin still remains obscure, even after 20 years of clinical use, the possibility that interstrand DNA cross-links may be associated with the clinical mechanism of action of this drug should now be fully addressed.

Daunomycin modifies the sequence-selective recognition of DNA by actinomycin.

The antitumour antibiotic actinomycin D normally binds to DNA by intercalation at sequences containing the CpG step, but in the presence of daunomycin it has been reported to interact with poly(dA-dT). This observation has neither been confirmed nor explained. Here we have used a photoreactive 7-azido derivative of actinomycin to study the effect of daunomycin on its binding to three DNA fragments. Daunomycin did indeed alter the binding of actinomycin to the DNA, such that the antibiotic was displaced from its primary GpC sites onto secondary sites in the DNA, though not to AT regions especially. These findings suggest a possible scientific explanation for the increased toxicity seen during combination chemotherapy with these two drugs.

Interaction of Bleomycin with Deoxyribonucleic Acid Oligomer: Proton Nuclear Magnetic Resonance Titration Study Using Novel Blemycin Complexes with Ni2+ and VO3+.

Two metal complexes of bleomycin (BLM), BLM-Ni2+ and BLM-VO3+ are used for studying interactions between BLM and deoxyribonucleic acid (DNA) by nuclear magnetic resonance. Although these BLMs do not mediate DNA strand scission under the usual conditions, they bind to DNA in the same manner as the active metal complexes of bleomycin (BLM-Fe2+ and BLM-Co3+). A self-complementary dodecanucleotide, d(CCCCAGCTGGGG), having a single site for cleavage was synthesized. d(CCCCAATTGGGG), which contains no -GpC- sequence, was also synthesized. The BLM-metal complexes were shown to bind specifically to the GpC site by circular dichroism and fluorescence titration studies. We assigned all the resonances for imino protons and phosphorus, and most of the nonexchangeable proton resonances of d(CCCCAGCTGGGG). No substantial change in the chemical shifts of these signals was observed upon titration with either BLM-Ni2+ or BLM-VO3+. This result is not consistent with a model of the strong intercalation of the BLMs between the base-pairs. The BLMs bind to DNA in a different manner, and DNA does not change its conformation upon binding with BLMs.

Binding of actinomycin D to DNA: evidence for a nonclassical high-affinity binding mode that does not require GpC sites.

We have employed a combination of temperature-dependent UV absorption spectroscopy, circular dichroism, and batch calorimetry to characterize the binding of actinomycin D to a series of oligomeric DNA duplexes. We find the duplex [d(CGTCGACG)]2 to be unique in its ability to bind actinomycin D strongly despite the absence of a classic GpC site. We present evidence that this non-GpC-containing duplex binds two actinomycin D molecules in an apparently cooperative manner to form a complex that exhibits aberrant spectroscopic and calorimetric behavior. We propose that these observations are consistent with actinomycin D exhibiting a high-affinity, sequence-dependent DNA-binding mode distinct from its classic binding to isolated GpC sites.

DNA sequence specificity of mitomycin cross-linking.

Using a gel electrophoresis assay, we show that the target DNA sequence cross-linked by N-methylmitomycin A, its aziridinomitosene, and mitomycin C is CpG, in strong preference over GpC. The yield per CpG site increases as the number of successive CpG sequences increases. Molecular modeling reveals no systematic difference between the energies of mitomycin cross-links at CpG in comparison with GpC. However, the distance between guanine amino groups in CpG sequences is nearly the same as the distance in the cross-linked adduct, whereas the amino group separation at GpC sites is substantially larger in the starting DNA than in the adduct. We suggest that the favorable placement of the second reaction center in CpG greatly accelerates the second step in the cross-linking reaction. As shown by a competition assay, mitomycins bind A-T and G-C sequences noncovalently equally well, even though the only sequence that yields appreciable cross-linking is CpG. N-Methylmitomycin A and its aziridinomitosene are found to be better cross-linking agents than is mitomycin C.

Transcriptional analysis of multisite drug-DNA dissociation kinetics: delayed termination of transcription by actinomycin D

An in vitro transcription assay was used to measure the relative occupancy, sequence specificity, and dissociation kinetics of six actinomycin D binding sites on DNA during conditions of active transcription of the DNA from the lac UV5 promoter. Five of the sites contained a GpC sequence, with three of these having a common AGCT sequence that differed by up to an order of magnitude in affinity for the drug, as indicated by their relative occupany and dissociation kinetics. Positive cooperativity was observed by higher occupancy and slower dissociation kinetics for neighboring GpC sites on a different DNA fragment (UV5-lambda PL). Termination of transcription was observed at some drug binding sites, while complete drug-induced termination of transcription was seen 7-10 nucleotides downstream of two drug sites. This delayed termination was minimized when ITP was incorporated into the transcripts and suggests that a time delay is required to enable stable RNA hairpin helices to form. A model is presented of the role of RNA hairpin helices in delayed, drug-induced termination of transcription. The classical picture of DNA-binding drugs as inhibitors of transcription now appears too simplistic as it does not accommodate this phenomenon. It will be important to gain a greater understanding of the mechanism of this phenomenon of drug-induced termination of transcription, as there are many implications for the design of DNA-acting drugs.

2D NMR investigation of the binding of the anticancer drug actinomycin D to duplexed dATGCGCAT: conformational features of the unique 2:1 adduct.

One- and two-dimensional NMR studies on the oligomer dA1T2G3C4G5C6A7T8, with and without actinomycin D (ActD), were conducted. Analysis of the NMR data, particularly 2D NOE intensities, revealed that the free oligonucleotide is a duplex in a standard right-handed B form. At the ratio of 1 ActD/duplex (R = 1), 1D NMR studies indicate that two 1:1 unsymmetric complexes form in unequal proportions with the phenoxazone ring intercalated at a GpC site, in agreement with previous studies [Scott, E.V., Jones, R.L., Banville, D.L., Zon, G., Marzilli, L.G., & Wilson, W.D. (1988) Biochemistry 27, 915-923]. The 2D COSY data also confirm this interpretation since eight cytosine H6 to H5 and two ActD H8 to H7 cross-peaks are observed. At R = 2, both COSY and NOESY spectra confirm the formation of a unique 2:1 species with C2 symmetry. The oligomer remains in a right-handed duplex but undergoes extreme conformational changes both at and adjacent to the binding site. The deoxyribose conformation of T2, C4, and C6 shifts from primarily C2'-endo in the free duplex to an increased amount of C3'-endo in the 2:1 complex as revealed by the greater intensity of the base H6 to 3' NOE cross-peak relative to the intensity of the H6 to H2' NOE cross-peak. This conformational change widens the minor groove and should help alleviate the steric crowding of the ActD peptides. The orientation of the ActD molecules at R = 2 has the quinoid portion of the phenoxazone ring at the G3pC4 site and the benzenoid portion of the phenoxazone ring at the G5pC6 site on the basis of NOE cross-peaks from ActD H7 and H8 to G5H8 and C6H6. All base pairs retain Watson-Crick type H-bonding, unlike echinomycin complexes [e.g., Gao, X., & Patel, D.J. (1988) Biochemistry 27, 1744-1751] where Hoogsteen base pairs have been observed. In contrast to previous studies on ActD, we were able to distinguish the two peptide chains.(ABSTRACT TRUNCATED AT 400 WORDS)

Proton and phosphorus NMR studies of d-CpG(pCpG)n duplexes in solution. Helix-coil transition and complex formation with actinomycin-D.

AbstractThe Watson–Crick imino and amino exchangeable protons, the nonexchangeable base and sugar protons, and the backbone phosphates for d‐CpG(pCpG)n, n = 1 and 2, have been monitored by high‐resolution nmr spectroscopy in aqueous solution over the temperature range 0°–90°C. The temperature dependence of the chemical shifts of the tetramer and hexamer resonances is consistent with the formation of stable duplexes at low temperature in solution. Comparison of the spectral characteristics of the tetranucleotide with those of the hexanucleotide with temperature permits the differentiation and assignment of the cytosine proton resonances on base pairs located at the end of the helix from those in an interior position. There is fraying at the terminal base pairs in the tetranucleotide and hexanucleotide duplexes. The Watson–Crick ring imino protons exchange at a faster rate than the Watson–Crick side‐chain amino protons, with exchange occurring by transient opening of the double helix. The structure of the d‐CpG(pCpG)n double helices has been probed by proton relaxation time measurements, sugar proton coupling constants, and the proton chemical shift changes associated with the helix–coil transition. The experimental data support a structural model in solution, which incorporates an anti conformation about the glycosyl bonds, C(3) exo sugar ring pucker, and base overlap geometries similar to the B‐DNA helix. Rotational correlation times of 1.7 and 0.9 × 10−9 sec have been computed for the hexanucleotide and tetranucleotide duplexes in 0.1 M salt, D2O, pH 6.25 at 27°C. The well‐resolved 31P resonances for the internucleotide phosphates of the tetramer and hexamer sequences at superconducting fields shift upfield by 0.2–0.5 ppm on helix formation. These shifts reflect a conformational change about the ω,ω′ phosphodiester bonds from gauche‐gauche in the duplex structure to a distribution of gauche‐trans states in the coil structure. Significant differences are observed in the transition width and midpoint of the chemical shift versus temperature profiles plotted in differentiated form for the various base and sugar proton and internucleotide phosphorous resonances monitoring the d‐CpG(pCpG)n helix–coil transition. The twofold symmetry of the d‐CpGpCpG duplex is removed on complex formation with the antibiotic actinomycin‐D. Two phosphorous resonances are shifted downfield by ∼2.6 ppm and ∼1.6 ppm on formation of the 1:2 Act‐D:d‐CpGpCpG complex in solution. Model studies on binding of the antibiotic to dinucleotides of varying sequence indicate that intercalation of the actinomycin‐D occurs at the GpC site in the d‐CpGpCpG duplex and that the magnitude of the downfield shifts reflects strain at the O‐P‐O backbone angles and hydrogen bonding between the phenoxazone and the phosphate oxygens. Actinomycin‐D is known to bind to nucleic acids that exhibit a B‐DNA conformation; this suggests that the d‐CpG(pCpG)n duplexes exhibit a B‐DNA conformation in solution.