A-tract and (+)-CC-1065-induced bending of DNA. Comparison of structural features using non-denaturing gel analysis, hydroxyl-radical footprinting, and high-field NMR

Modeling anticancer drug–DNA interactions via mixed QM/MM molecular dynamics simulations

The development of anticancer drugs started over four decades ago, with the serendipitous discovery of the antitumor activity of cisplatin and its successful use in the treatment of various cancer types. Despite the efforts made in unraveling the mechanism of the action of cisplatin, as well as in the rational design of new anticancer compounds, in many cases detailed structural and mechanistic information is still lacking. Many of these drugs exert their anticancer activity by covalently binding to DNA inducing a distortion or simply impeding replication, thus triggering a cellular response, which eventually leads to cell death. A detailed understanding of the structural and electronic properties of drug-DNA complexes and their mechanism of binding is the key step in elucidating the principles of their anticancer activity. At the theoretical level, the description of covalent drug-DNA complexes requires the use of state-of-the-art computer simulation techniques such as hybrid quantum/classical molecular dynamics simulations. In this review we provide a general overview on: drugs which covalently bind to DNA duplexes, the basic concepts of quantum mechanics/molecular mechanics (QM/MM), molecular dynamics methods and a list of selected applications of these simulations to the study of drug-DNA adducts. Finally, the potential and the limitations of this approach to the study of such systems are critically evaluated.

Evidence for a common molecular basis for sequence recognition of N3-guanine and N3-adenine DNA adducts involving the covalent bonding reaction of (+)-CC-1065

The antitumor antibiotic (+)-CC-1065 can alkylate N3 of guanine in certain sequences. A previous high-field 1H NMR study on the (+)-CC-1065d[GCGCAATTG*CGC]2 adduct (* indicates the drug alkylation site) showed that drug modification on N3 of guanine results in protonation of the cross-strand cytosine [Park, H.-J.; Hurley, L. H. J. Am. Chem. Soc. 1997, 119, 629.]. In this contribution we describe a further analysis of the NMR data sets together with restrained molecular dynamics. This study provides not only a solution structure of the (+)-CC-1065(N3-guanine) DNA duplex adduct but also new insight into the molecular basis for the sequence-specific interaction between (+)-CC-1065 and N3-guanine in the DNA duplex. On the basis of NOESY data, we propose that the narrow minor groove at the 7T8T step and conformational kinks at the junctions of 16C17A and 18A19T are both related to DNA bending in the drugDNA adduct. Analysis of the one-dimensional 1H NMR (in H2O) data and rMD trajectories strongly suggests that hydrogen bonding linkages between the 8-OH group of the (+)-CC-1065 A-subunit and the 9G10C phosphate via a water molecule are present. All the phenomena observed here in the (+)-CC-1065(N3-guanine) adduct at 5'-AATTG* are reminiscent of those obtained from the studies on the (+)-CC-1065(N3-adenine) adduct at 5'-AGTTA*, suggesting that (+)-CC-1065 takes advantage of the conformational flexibility of the 5'-TPu step to entrap the bent structure required for the covalent bonding reaction. This study reveals a common molecular basis for (+)-CC-1065 alkylation at both 5'-TTG* and 5'-TTA*, which involves a trapping out of sequence-dependent DNA conformational flexibility as well as sequence-dependent general acid and general base catalysis by duplex DNA.

DNA structural perturbation induced by the CPI-derived DNA interstrand cross-linker: molecular mechanisms for the sequence specific recognition

The highly potent cytotoxic DNA-DNA cross-linker consists of two cyclopropa[c]pyrrolo[3, 4-3]indol-4(5H)-ones indoles [(+)-CPI-I] joined by a bisamido pyrrole (abbreviated to "Pyrrole"). The Pyrrole is a synthetic analog of Bizelesin, which is currently in phase II clinical trials due to its excellent in vivo antitumor activity. The Pyrrole has 10 times more potent cytotoxicity than Bizelesin and mostly form DNA-DNA interstrand cross-links through the N3 of adenines spaced 7 bp apart. The Pyrrole requires a centrally positioned GC base pair for high cross-linking reactivity (i.e., 5'-T(A/T)2G(A/T)2A*-3'), while Bizelesin prefers purely AT-rich sequences (e.g., 5'-T(A/T)4 or 5A*-3', where T represents the cross-strand adenine alkylation and A* represents an adenine alkylation) (Park et al., 1996). In this study, the high-field 1H-NMR and rMD studies are conducted on the 11-mer DNA duplex adduct of the Pyrrole where the 5'-TTAGTTA*-3' sequence is cross-linked by the drug. A severe structural perturbation is observed in the intervening sequences of cross-linking site, while a normal B-DNA structure is maintained in the region next to the drug-modified adenines. Based upon these observations, we propose that the interplay between the bisamido pyrrole unit of the drug and central G/C base pair (hydrogen-bonding interactions) is involved in the process of cross-linking reaction, and sequence specificity is the outcome of those interactions. This study suggests a mechanism for the sequence specific cross-linking reaction of the Pyrrole, and provides a further insight to develop new DNA sequence selective and distortive cross-linking agents.

Selective metal cation activation of a DNA alkylating agent: synthesis and evaluation of methyl 1,2,9, 9a-Tetrahydrocyclopropa[c]pyrido[3,2-e]indol-4-one-7-carboxylate (CPyI)

The synthesis of methyl 1,2,9,9a-tetrahydrocyclopropa[c]pyrido[3, 2-e]indol-4-one-7-carboxylate (CPyI) containing a one carbon expansion of the C ring pyrrole found in the duocarmycin SA alkylation subunit and its incorporation into analogues of the natural product are detailed. The unique 8-ketoquinoline structure of CPyI was expected to provide a tunable means to effect activation via selective metal cation complexation. The synthesis of CPyI was based on a modified Skraup quinoline synthesis followed by a 5-exo-trig aryl radical cyclization onto an unactivated alkene with subsequent TEMPO trap or 5-exo-trig aryl radical cyclization onto a vinyl chloride for synthesis of the immediate precursor. Closure of the activated cyclopropane, accomplished by an Ar-3' spirocyclization, provided the CPyI nucleus in 10 steps and excellent overall conversion (29%). The evaluation of the CPyI-based agents revealed an intrinsic stability comparable to that of CC-1065 and duocarmycin A but that it is more reactive than duocarmycin SA and the CBI-based agents (3-4x). A pH-rate profile of the addition of nucleophiles to CPyI demonstrated that an acid-catalyzed reaction is observed below pH 4 and that an uncatalyzed reaction predominates above pH 4. The expected predictable activation of CPyI by metal cations toward nucleophilic addition was found to directly correspond to established stabilities of the metal complexes with the addition product (Cu(2+) > Ni(2+) > Zn(2+) > Mn(2+) > Mg(2+)) and provides the opportunity to selectively activate the agents upon addition of the appropriate Lewis acid. This tunable metal cation activation of CPyI constitutes the first example of a new approach to in situ activation of a DNA binding agent complementary to the well-recognized methods of reductive, oxidative, or photochemical activation. Resolution and synthesis of a full set of natural product analogues and subsequent evaluation of their DNA alkylation properties revealed that the CPyI analogues retain identical DNA alkylation sequence selectivity and near-identical DNA alkylation efficiencies compared to the natural products. Consistent with past studies and even with the deep-seated structural change in the alkylation subunit, the agents were found to exhibit potent cytotoxic activity that directly correlates with their inherent reactivity.

Bifunctional alkylating agents derived from duocarmycin SA: potent antitumor activity with altered sequence selectivity

The series of four dimers derived from head to tail coupling of the two enantiomers of the duocarmycin SA alkylation subunit are described.

A Novel Class of CC-1065 and Duocarmycin Analogues Subject to Mitomycin-Related Reductive Activation

A new class of DNA alkylating agents is described that incorporate the quinone of the mitomycins, which is thought to impart tumor cell selectivity as a result of preferential reduction and activation in hypoxic tumors, into the AT-selective binding framework of the duocarmycins capable of mitomycin-like reductive activation and duocarmycin-like spirocyclization and subsequent DNA alkylation. Consistent with this design, the quinone prodrugs fail to alkylate DNA unless reductively activated and then do so with an adenine N3 alkylation sequence selectivity identical to that of the duocarmycins. Additionally, the agents exhibit a selectivity toward DT-Diaphorase (NQO1)-containing versus DT-Diaphorase-deficient (resistant) tumor cell lines, and they were shown to be effective substrates for reduction by recombinant human DT-Diaphorase. As such, the agents constitute effective duocarmycin and CC-1065 analogues subject to reductive activation. In addition, the solvolysis pH rate dependence of a series of reactive spirocyclopropanes revealed a unique and inverted order of reactivity at pH 7 versus pH 3. This behavior and the structural features responsible for it are consistent with an acid-catalyzed reaction at pH 3, but a direct uncatalyzed S(N)2 reaction at pH 7 that is not subject to acid catalysis.

An improved method for the rapid assessment of DNA bending by small molecules

Assessing the effects of non-covalently bound chemicals on DNA structure is particularly challenging. Traditional methods require the use of cumbersome electrophoretic techniques or that the compounds bind DNA with an extremely high affinity. Here we demonstrate that, by extending the use of the exonuclease Bal 31, we can rely on a standard cyclization assay technique and one dimensional gel electrophoresis to identify and quantitate chemical induced DNA bending. An important application of this method is to the study of small molecules that bind to DNA non-covalently and we illustrate the method with the antitumor antibiotic calicheamicin. Our results suggest that the distribution of circles resulting from the polymerization of a 21 bp DNA construct reflects the kinetics of the competing cyclization and dimerization reactions and provides a method for rapidly screening compounds for DNA bending.

Catalysis of the CC-1065 and duocarmycin DNA alkylation reaction: DNA binding induced conformational change in the agent results in activation

A number of indirect observations are summarized that suggest the rate acceleration for the CC-1065 and duocarmycin. DNA alkylation reaction is derived in part from a DNA binding-induced conformational change in the agents which substantially increases their inherent reactivity. This ground-state destabilization of the agent, which we suggest results from a binding-induced twist in the linking N2 amide and requires a rigid extended N2 amide substituent, disrupts the vinylogous amide stabilization and activates the agents for DNA alkylation.

A simple ligation assay to detect effects of drugs on the curvature/flexibility of DNA

Circular DNA molecules can readily be formed from the 169 bp tyrT fragment in the presence of T4 DNA ligase. We have analyzed the formation of DNA circles in the presence of the clinically important antitumour drugs amsacrine, mitoxantrone and daunomycin. All three are intercalating agents but they affect the closure reaction differently: daunomycin and mitoxantrone progressively inhibit the formation of circles whereas at low concentrations amsacrine strongly enhances the yield of circles suggesting that this drug can increase the flexibility and/or curvature of DNA. The ligation assay described here may prove useful and widely applicable for investigating the effects of small molecules on the secondary structure of DNA.

Specific targeting of protein-DNA complexes by DNA-reactive drugs (+)-CC-1065 and pluramycins

To gain insight into the interactions between transcriptional factor proteins and DNA, the DNA-reactive drugs (+)-CC-1065 and pluramycin were used to target specific protein-DNA complexes. The structural features of the complex between the transcriptional activator Sp1 and the 21-base-pair repeat of the early promoter region of SV40 DNA were examined using hydroxyl-radical footprinting; (+)-CC-1065, a sequence-specific minor groove bending probe; and circularization experiments. The results show that the 21-base-pair repeat region has an intrinsically in-phase bent structure that is stabilized upon saturation Sp1 binding by protein-DNA and protein-protein interactions to produce a looping structure. The intercalating drug pluramycin was used to probe the structural details of the interaction between the TATA binding protein (TBP) and the TATA box DNA sequence. TBP, which directs initiation of RNA transcription, exhibits two-fold symmetry and apparently interacts with the TATA box in a symmetrical fashion. However, the interaction results in an asymmetric effect, in that transcription is initiated only in the downstream direction. Using pluramycin as a probe, it was determined that TBP binding to the human myoglobin TATA sequences enhances pluramycin reactivity at a site immediately downstream of the TATA box. The implications on transcriptional control of ternary complexes comprised of transcriptional factors, DNA, and DNA-reactive compounds will be presented.

New insights into calicheamicin-DNA interactions derived from a model nucleosome system

Using the Xenopus borealis 5S RNA gene, we have identified several new features of the interaction of calicheamicin (CAL), an enediyne antitumor agent, with nucleosomal and naked DNA targets. CAL-mediated DNA damage was generally reduced by incorporation of the DNA into a nucleosome. However, in one instance, the frequency of DNA damage was enhanced in the nucleosome compared to naked DNA. This increase in CAL damage may result from bending-induced changes in the target site, while the association of histone proteins with DNA in the nucleosome may generally reduce the affinity of CAL for its targets by imposing dynamic constraints on the DNA, by altering target structure, or by steric hindrance. One implication of these observations is that new structural features created by incorporation of DNA into chromatin may produce 'hot spots' for CAL-mediated DNA damage not apparent in naked DNA studies. In a second set of experiments, the orientation of CAL at damage sites in naked 5S rDNA was determined. The results suggest that minor groove width per se is not a major determinant of CAL target selection. Our studies support the generality of an oligopurine recognition element, with the additional requirement that the purine tract is interrupted at the 3'-end by a pyrimidine(s). To account for these observations, we propose a model in which CAL recognizes the unique structural and dynamic features associated with the 3'-end of an oligopurine tract. Finally, we conclude that the dyad axis of pseudosymmetry of the 5S rRNA gene nucleosome cannot be determined with any degree of certainty. This places significant limitations on the interpretation of results from the study of drug-DNA interactions with reconstituted nucleosomes.

CC-1065 and the duocarmycins: unraveling the keys to a new class of naturally derived DNA alkylating agents

Key studies defining the DNA alkylation properties and selectivity of a new class of exceptionally potent, naturally occurring antitumor antibiotics including CC-1065, duocarmycin A, and duocarmycin SA are reviewed. Recent studies conducted with synthetic agents containing deep-seated structural changes and the unnatural enantiomers of the natural products and related analogs have defined the structural basis for the sequence-selective alkylation of duplex DNA and fundamental relationships between chemical structure, functional reactivity, and biological properties. The agents undergo a reversible, stereoelectronically controlled adenine-N3 addition to the least substituted carbon of the activated cyclopropane within selected AT-rich sites. The preferential AT-rich non-covalent binding selectivity of the agents within the narrower, deeper AT-rich minor groove and the steric accessibility to the alkylation site that accompanies deep AT-rich minor groove penetration control the sequence-selective DNA alkylation reaction and stabilize the resulting adduct. For the agents that possess sufficient reactivity to alkylate DNA, a direct relationship between chemical or functional stability and biological potency has been defined.

Binding of Sp1 to the 21-bp repeat region of SV40 DNA: effect of intrinsic and drug-induced DNA bending between GC boxes

The effect of the antitumor antibiotic (+)-CC-1065 on the binding of Sp1 to the 21-bp repeats of SV40 DNA has been investigated. (+)-CC-1065 alkylates N3 of adenine in DNA and resides in the minor groove. As a consequence of alkylation of the two 5'-AGTTA* sequences (* indicates covalent modification site), which reside between GC boxes III and IV, and boxes V and VI, protein binding to the 3' sites is completely abolished and there is a significant decrease in Sp1 binding to the other regions. The effect of substituting A5 tracts for the (+)-CC-1065-bonding sequence was intermediate between the unmodified 5'-AGTTA* and the drug-modified sequences. It is proposed that a structural distortion of DNA associated with stiffening of the helix induced by the drug-adduct formation is primarily responsible for the inhibition of binding of Sp1 molecules to 21-bp repeats, rather than steric hindrance due to the occupancy by drug molecules of the minor groove within that region.

CC-1065 bonding to intracellular and purified SV40 DNA: site specificity and functional effects

CC-1065 is a minor-groove bonding agent capable of forming covalent adducts with the N-3 position of adenines within A-T-rich regions of duplex DNA. By examining the formation and location of CC-1065 adducts within the simian virus 40 (SV40) DNA molecule, the present study marks the first time that the precise sites of CC-1065 lesions have been identified at the level of eukaryotic genomic DNA. In naked DNA preparations, r values (moles of drug/mole of nucleotide base pair) > or = 0.0015 effected, after thermal treatment, a measurable decrease in intact supercoiled form I, as well as increases in forms II and III, indicating that both single-strand and apparent double-strand damage had occurred. A similar pattern of damage was observed in SV40-infected cells, albeit at higher CC-1065 levels. The amount of CC-1065 required to produce a 50% loss in form I was > 2-fold higher in infected cells (r = 0.029) than with purified DNA samples (r = 0.013). The appearance of double-strand damage at low drug levels suggested a high specificity of CC-1065 bonding to localized regions of the genome. The precise location of these CC-1065 adduction sites was examined by three methods: sequence analysis of the entire genome (GenBank), DNA polymerase termination assay of specific fragments of SV40, and restriction enzyme digestion analysis of the entire SV40 molecule. When sequence analysis of the entire genome was performed by examining both strands for the presence of the consensus CC-1065 binding sequence 5'-A/T-A/T-A/T-A/T-A*-3'[Reynolds et al. (1985) Biochemistry 24, 6228-6247], 294 single-strand adduction sites were predicted, compared to 20 sites where CC-1065 should bond to both strands within a 30-base-pair window and at which, when heated, a double-strand break should occur. DNA polymerase termination assay of actual adduction sites was performed on restriction fragments of SV40 DNA pretreated with CC-1065 in infected cells or in purified supercoiled DNA preparations and selected on the basis of the sequence analysis (i.e., regions 2510-2730, 3701-3920, 4400-4659, 4020-4320, and 5163-65). In general, double-strand lesions were detected in similar regions of the genome by the DNA termination assay and by sequence analysis. When restriction enzyme digestion and the DNA polymerase termination assay were compared throughout the genome, nearly identical patterns of adduct formation were observed. Interestingly, similar alkylation patterns were observed with either naked or infected cell DNA.(ABSTRACT TRUNCATED AT 400 WORDS)

(+)-CC-1065 as a probe for intrinsic and protein-induced bending of DNA

(+)-CC-1065 is a biologically potent DNA-reactive antitumor antibiotic produced by Streptomyces zelensis. This antibiotic covalently modifies DNA by alkylation of N-3 of adenine in the minor groove. As a structural consequence of covalent modification of DNA, the helix axis is bent into the minor groove. The drug-induced bending of DNA has similarities to intrinsic A-tract bending and the 3' adenine of A-tracts shows a unique reactivity to alkylation by (+)-CC-1065. Upon covalent modification of A-tracts, the magnitude of bending is increased and the helix is stiffened. Using high-field NMR, hydroxyl-radical footprinting and gel electrophoresis, the molecular basis for the high reactivity of the bonding sequence 5'-AGTTA* (an asterisk indicates the covalent modification site) to (+)-CC-1065 has been shown to involve the inherent conformational flexibility of this sequence. Furthermore, these studies also demonstrate that after alkylation the drug-induced bending is focused over the TT region. By analogy with the junction bend model for A-tracts, a 'truncated junction bend model' is proposed for this structure. Last, the application (+)-CC-1065 entrapped/induced bending of DNA as a probe for the Sp1-induced bending of the 21-base-pair repeat and Mu transposase bending of the att L3 sequence is described.

Molecular basis for sequence selective DNA alkylation by (+)- and ent-(-)-CC-1065 and related agents: alkylation site models that accommodate the offset AT-rich adenine N3 alkylation selectivity

A detailed evaluation of the DNA alkylation selectivity of (+)-CC-1065, ent-(-)-CC-1065 and a series of aborted and extended analogs possessing the CPI alkylation subunit is detailed and the refinement of a model that accommodates the offset AT-rich adenine N3 alkylation selectivity of the enantiomeric agents is presented. The natural enantiomers bind in the minor groove in the 3'-->5' direction starting from the adenine N3 alkylation site across a 2 base (N-BOC-CPI; i.e. 5'-AA), 3.5 base (CPI-CDPI1/CPI-PDE-I1; i.e. 5'-AAA), 5 base (CC-1065/CPI-CDPI2; i.e. 5'-AAAAA) or 6.5 base (CPI-CDPI3; i.e. 5'-AAAAAA) AT-rich site. In contrast, the unnatural enantiomers bind in the reverse 5'-->3' direction in the minor groove and the binding site necessarily starts at the first 5' base preceding the adenine N3 alkylation site and extends across the alkylation site to the adjacent 3' bases covering an AT-rich site of 2 bases (N-BOC-CPI; e.g., 5'-AA), 5 bases (CC-1065/CPI-CDPI2; eg. 5'-AAAAA), or 6.5 bases (CPI-CDPI3; e.g. 5'-AAAAAA). Notably, the model accommodates the unusual observation that both enantiomers of N-BOC-CPI alkylate the same sites within duplex DNA (5'-AA > 5'-TA) and the required reversed binding orientation of the enantiomeric agents. The reversed binding orientation is required to permit access to the electrophilic cyclopropane and the resulting offset AT-rich alkylation selectivity is the natural consequence of the diastereomeric relationship of the adducts. Three dimensional models of the natural and unnatural enantiomer alkylations are presented which clearly illustrate the offset binding sites. A fundamentally simple model for the CC-1065 DNA alkylation reaction, that accommodates the behavior of both enantiomers, is provided in which the sequence selectivity is derived from the noncovalent binding selectivity of the agents preferentially in the narrower, sterically more accessible AT-rich minor groove, the inherent steric accessibility to the adenine N3 alkylation site that accompanies deep penetration of the agent into the minor groove within an AT-rich site, and the 2 base-pair (N-BOC-CPI), 3.5 base-pair (CPI-PDE-I1/CPI-CDPI1), 5 base-pair (CC-1065/CPI-CDPI2), or 6.5 base-pair (CPI-CDPI3) site size required to permit agent binding in the minor groove at the alkylation site.(ABSTRACT TRUNCATED AT 400 WORDS)

(+)-CC-1065 as a structural probe of Mu transposase-induced bending of DNA: overcoming limitations of hydroxyl-radical footprinting

Phage Mu transposase (A-protein) is primarily responsible for transposition of the Mu genome. The protein binds to six att sites, three at each end of Mu DNA. At most att sites interaction of a protein monomer with DNA is seen to occur over three minor and two consecutive major grooves and to result in bending up to about 90 degrees. To probe the directionality and locus of these A-protein-induced bends, we have used the antitumor antibiotic (+)-CC-1065 as a structural probe. As a consequence of binding within the minor groove, (+)-CC-1065 is able to alkylate N3 of adenine in a sequence selective manner. This selectivity is partially determined by conformational flexibility of the DNA sequence, and the covalent adduct has a bent DNA structure in which narrowing of the minor groove has occurred. Using this drug in experiments in which either gel retardation or DNA strand breakage are used to monitor the stability of the A-protein--DNA complex or the (+)-CC-1065 alkylation sites on DNA (att site L3), we have demonstrated that of the three minor grooves implicated in the interaction with A-protein, the peripheral two are 'open' or accessible to drug bonding following protein binding. These drug-bonding sites very likely represent binding at at least two A-protein-induced bending sites. Significantly, the locus of bending at these sites is spaced approximately two helical turns apart, and the bending is proposed to occur by narrowing of the minor groove of DNA. The intervening minor groove between these two peripheral sites is protected from (+)-CC-1065 alkylation. The results are discussed in reference to a proposed model for overall DNA bending in the A-protein att L3 site complex. This study illustrates the utility of (+)-CC-1065 as a probe for protein-induced bending of DNA, as well as for interactions of minor groove DNA bending proteins with DNA which may be masked in hydroxyl radical footprinting experiments.