Distamycins inhibit the binding of OTF-1 and NFE-1 transfactors to their conserved DNA elements

Targeting Transcription Factors for Cancer Treatment

Transcription factors are involved in a large number of human diseases such as cancers for which they account for about 20% of all oncogenes identified so far. For long time, with the exception of ligand-inducible nuclear receptors, transcription factors were considered as “undruggable” targets. Advances knowledge of these transcription factors, in terms of structure, function (expression, degradation, interaction with co-factors and other proteins) and the dynamics of their mode of binding to DNA has changed this postulate and paved the way for new therapies targeted against transcription factors. Here, we discuss various ways to target transcription factors in cancer models: by modulating their expression or degradation, by blocking protein/protein interactions, by targeting the transcription factor itself to prevent its DNA binding either through a binding pocket or at the DNA-interacting site, some of these inhibitors being currently used or evaluated for cancer treatment. Such different targeting of transcription factors by small molecules is facilitated by modern chemistry developing a wide variety of original molecules designed to specifically abort transcription factor and by an increased knowledge of their pathological implication through the use of new technologies in order to make it possible to improve therapeutic control of transcription factor oncogenic functions.

An overview of recent advances in duplex DNA recognition by small molecules

As the carrier of genetic information, the DNA double helix interacts with many natural ligands during the cell cycle, and is amenable to such intervention in diseases such as cancer biogenesis. Proteins bind DNA in a site-specific manner, not only distinguishing between the geometry of the major and minor grooves, but also by making close contacts with individual bases within the local helix architecture. Over the last four decades, much research has been reported on the development of small non-natural ligands as therapeutics to either block, or in some cases, mimic a DNA–protein interaction of interest. This review presents the latest findings in the pursuit of novel synthetic DNA binders. This article provides recent coverage of major strategies (such as groove recognition, intercalation and cross-linking) adopted in the duplex DNA recognition by small molecules, with an emphasis on major works of the past few years.

Targeting transcription factors by small compounds--Current strategies and future implications

Transcription factors are central regulators of gene expression and critically steer development, differentiation and death. Except for ligand-activated nuclear receptors, direct modulation of transcription factor function by small molecules is still widely regarded as "impossible". This "un-druggability" of non-ligand transcription factors is due to the fact that the interacting surface between transcription factor and DNA is huge and subject to significant changes during DNA-binding. Besides some "success studies" with compounds that directly interfere with DNA binding, drug targeting approaches mostly address protein-protein interfaces with essential co-factors, transcription factor dimerization partners, chaperone proteins or proteins that regulate subcellular shuttling. An alternative strategy represent DNA-intercalating, alkylating or DNA-groove-binding compounds that either block transcription factor-binding or change the 3D-conformation of the consensus DNA-strand. Recently, much interest has been focused on chromatin reader proteins that steer the recruitment and activity of transcription factors to a gene transcription start site. Several small compounds demonstrate that these epigenetic reader proteins are exciting new drug targets for inhibiting lineage-specific transcription in cancer therapy. In this research update we will discuss recent advances in targeting transcription factors with small compounds, the challenges that are related to the complex function and regulation of these proteins and also the possible future directions and applications of transcription factor drug targeting.

Iron(II) supramolecular helicates condense plasmid DNA and inhibit vital DNA-related enzymatic activities

The dinuclear iron(II) supramolecular helicates [Fe2 L3 ]Cl4 (L=C25 H20 N4 ) bind to DNA through noncovalent (i.e., hydrogen-bonding, electrostatic) interactions and exhibit antimicrobial and anticancer effects. In this study, we show that the helicates condense plasmid DNA with a much higher potency than conventional DNA-condensing agents. Notably, molecules of DNA in the presence of the M enantiomer of [Fe2 L3 ]Cl4 do not form intermolecular aggregates typically formed by other condensing agents, such as spermidine or spermine. The helicates inhibit the activity of several DNA-processing enzymes, such as RNA polymerase, DNA topoisomerase I, deoxyribonuclease I, and site-specific restriction endonucleases. However, the results also indicate that the DNA condensation induced by the helicates does not play a crucial role in these inhibition reactions. The mechanisms for the inhibitory effects of [Fe2 L3 ]Cl4 helicates on DNA-related enzymatic activities have been proposed.

Preclinical and clinical results with the natural marine product ET-743

ET-743 (Yondelis, trabectedin) is a natural marine product with antitumour properties derived from the tunicate Ecteinascidia turbinata. ET-743 binds to the N2 position of guanine in the minor groove of DNA with some degree of sequence specificity, altering the transcription regulation of induced genes. Cells that are deficient in nucleotide excision repair, hypersensitive to UV rays, cisplatin and conventional alkylating agents, are resistant to ET-743. This is a unique property of ET-743 and is of potential importance for the drug activity when administered alone or in combination with other drugs. ET-743 showed striking antitumour activity against sensitive and resistant human xenografts. The dose-limiting toxicities in animal models, hepatobiliary events, were of concern, but the pattern of the reversibility noted in monkeys and the evidence of a positive therapeutic index in tumour-bearing nude mice prompted its clinical development. The Phase I programme investigated different schedules of administration, with the dose-limiting toxicities being neutropenia and fatigue. As anticipated in the preclinical models, reversible non-cumulative transaminitis was a prevalent finding from one-third of the maximum tolerated dose level; long-lasting objective responses in pretreated resistant patients were noted, including consistent efficacy data in mesenchymal tumours. The Phase II data for ET-743 administered as a single agent has established a clinical role for the compound in advanced pretreated soft tissue sarcoma and a promising potential in pretreated ovarian and breast cancer. ET-743 combined with other drugs (i.e., cisplatin, paclitaxel or doxorubicin) showed more than additive effects in several preclinical systems and initial clinical results (e.g., a combination of ET-743 with cisplatin) appear to confirm the preclinical findings. In summary, ET-743 is a new drug with a novel mode of action, which has demonstrated activity in human tumours resistant to the available anticancer drugs. Further comparative studies are needed to define the role of ET-743 alone or in combination in cancer chemotherapy.

Structure of the murine tenascin-R gene and functional characterisation of the promoter

The tenascin-R (TN-R) gene encodes a multidomain extracellular matrix glycoprotein belonging to the tenascin family. It is detectable mainly in oligodendrocytes and neuronal subpopulations of the central nervous system. In this report, we describe the structure of the 5'-region of the mouse TN-R gene and characterise the activity of its promoter. By in silico cloning and genome walking, we have deduced the organisation of the gene and identified the promoter sequence by 5'-RACE technology. TN-R transcripts in adult mouse brain contain non-coding exons 1 and 2 as demonstrated by the reverse transcriptase-polymerase chain reaction. The promoter displays its activity in cultured cells of neural origin, but not in a fibroblast-like cell line or an undifferentiated teratocarcimoma cell line. As for the human and rat genes, the elements required for the full and cell type-specific activity of the promoter are contained in exon 1 and 167 bp upstream of this exon. The mouse TN-R promoter sequence is similar to that of rat and human in that it displays similarly unusual features: it lacks any classical TATA-box or CAAT-box, GC-rich regions or initiator elements. The promoter contains consensus sequences for binding of a variety of transcription factors, notably p53/p73 and glucocorticoid receptors.

Pentamidine-induced alteration in restriction endonuclease cleavage of plasmid DNA

We have used restriction enzymes and DNaseI as probes to determine the specificity of pentamidine binding to plasmid DNA. Cleavage of plasmid pAZ130 by EcoRI, EcoRV and ApaI is inhibited by pentamidine, cleavage by XbaI, NotI and AvaI is unaffected, while cleavage by XhoI, which recognizes the same sequence as AvaI, is stimulated. DNaseI footprinting of DNA containing these restriction sites revealed that pentamidine protection is not strictly limited to AT-rich regions. We suggest that perturbation of the DNA micro- environment by pentamidine binding is responsible for its effect on nucleases.

Fabrication of Unimolecular Double-stranded DNA Microarrays on Solid Surfaces for Probing DNA-Protein/Drug Interactions

We present a novel method for fabricating unimolecular double-stranded DNA microarrays on solid surfaces, which were used to probe sequence-specific DNA/protein interactions. For manufacturing the unimolecular double-stranded DNA microarrays, two kinds of special single-stranded oligonucleotides, constant oligonucleotide and target oligonucleotide, were chemically synthesized. The constant oligonucleotides with internal aminated dT were used to capture and immobilize the target oligonucleotides onto the solid surface, and also to provide a primer for later enzymatic extension reactions, while target oligonucleotides took the role of harbouring DNA-binding sites of DNA-binding proteins. The variant target oligonucleotides were annealed and ligated with the constant oligonucleotides to form the new unimolecular oligonucleotides for microspotting. The prepared unimolecular oligonucleotides were microspotted on aldehyde-derivatized glass slides to make partial-dsDNA microarrays. Finally, the partial-dsDNA microarrays were converted into a unimolecular complete-dsDNA microarray by a DNA polymerase extension reaction. The efficiency and accuracy of the polymerase synthesis were demonstrated by the fluorescent-labeled dUTP incorporation in the enzymatic extension reaction and the restriction endonuclease digestion of the fabricated unimolecular complete-dsDNA microarray. The accessibility and specificity of the sequence-specific DNA-binding proteins binding to the immobilized unimolecular dsDNA probes were demonstrated by the binding of Cy3 labeled NF-κB (p50·p50) to the unimolecular dsDNA microarray. This unimolecular dsDNA microarray provides a general technique for high-throughput DNA-protein or DNA-drugs interactions.

Synergistic induction of the senescence-associated genes by 5-bromodeoxyuridine and AT-binding ligands in HeLa cells

5-Bromodeoxyuridine induces a senescence-like phenomenon in mammalian cells. This effect was dramatically potentiated by AT-binding ligands such as distamycin A, netropsin, and Hoechst 33258. The genes most remarkably affected by these ligands include the widely used senescence-associated genes and were located on or nearby Giemsa-dark bands of human chromosomes. We hypothesize that AT-rich scaffold/nuclear matrix attachment region sequences are involved in this phenomenon. In fact, upon substitution of thymine with 5-bromouracil, a rat S/MAR sequence reduced its degree of bending and became insensitive to cancellation of the bending by distamycin A. The S/MAR sequence containing 5-bromouracil also bound more tightly to nuclear scaffold proteins in vitro and this binding was not inhibited by distamycin A. Under the same conditions, the S/MAR sequence containing thymine easily dissociated from the nuclear scaffold proteins. Taken together, the synergistic induction of the genes may be explained not only by opening of condensed chromatin by distamycin A but also by increase in the binding of 5-bromouracil-containing S/MAR sequences to the nuclear scaffolds.

Distamycin A affects the stability of NF-kappaB p50-DNA complexes in a sequence-dependent manner

The effect of two different DNA minor groove binding molecules, Hoechst 33258 and distamycin A, on the binding kinetics of NF-kappaB p50 to three different specific DNA sequences was studied at various salt concentrations. Distamycin A was shown to significantly increase the dissociation rate constant of p50 from the sequences PRDII (5'-GGGAAATTCC-3') and Ig-kappa B (5'-GGGACTTTCC-3') but had a negligible effect on the dissociation from the palindromic target-kappaB binding site (5'-GGGAATTCCC-3'). By comparison, the effect of Hoechst 33258 on binding of p50 to each sequence was found to be minimal. The dissociation rates for the protein--DNA complexes increased at higher potassium chloride concentrations for the PRDII and Ig-kappaB binding motifs and this effect was magnified by distamycin A. In contrast, p50 bound to the palindromic target-kappaB site with a much higher intrinsic affinity and exhibited a significantly reduced salt dependence of binding over the ionic strength range studied, retaining a K(D) of less than 10 pM at 150 mM KCl. Our results demonstrate that the DNA binding kinetics of p50 and their salt dependence is strongly sequence-dependent and, in addition, that the binding of p50 to DNA can be influenced by the addition of minor groove-binding drugs in a sequence-dependent manner.

Development of distamycin-related DNA binding anticancer drugs

The relatively low therapeutic index of the clinically used alkylating agents is probably related to the fact that these compounds cause DNA damage in a relatively unspecific manner, mainly involving guanine-cytosine rich stretches of DNA present in virtually all genes, therefore inducing unselective growth inhibition and death, both in neoplastic and in highly proliferative normal tissues. These considerations explain why in the last twenty years there has been an increasing interest in the identification of compounds which can target DNA with a much higher degree of sequence specificity than that of conventional alkylators. Minor groove binders (MGBs) are one of the most widely studied class of alkylating agents characterised by a high level of sequence specificity. The prototype of this class of drugs is distamycin A which is an antiviral compound able to interact, non-covalently, in theminor groove of DNA in A-T rich regions. It is not cytotoxic against tumour cells and thus has been used as a carrier for targeting cytotoxic alkylating moieties in theminor groove of DNA. The benzoyl mustard derivative of distamycin A, tallimustine, was found to be able to alkylate the N(3) of adenine in theminor groove of DNA only in the target hexamer 5'-TTTTGA or 5'-TTTTAA. Tallimustine was investigated in the clinic and was not successful because it causes severe bone marrow toxicity. The screening of other distamycin derivatives, which maintain antitumour activity and exhibit much lower toxicity against human bone marrow cells than tallimustine led to the identification of brostallicin (PNU-166196) which is currently under early clinical investigation. Although MGBs which bind DNA in A-T rich regions have not fulfilled the expectations, it is too early to draw definitive conclusions on this class of compounds. The peculiar bone-marrow toxicity observed in the clinic both with tallimustine or with CC-1065 derivatives is not necessarily a feature of all MGBs, as indicated by recent evidence obtained with brostallicin and other structurally unrelated MGBs (e.g., ET-743).

DNA-binding activity and cytotoxicity of the extended diphenylfuran bisamidines in breast cancer MCF-7 cells

The DNA binding properties of three novel extended diphenylfuran bisamidines (1-3) possessing different dicationic terminal side chains were studied. The ultrafiltration assay showed that bisamidines 1-3 have significant affinity for DNA. The DNA-binding data for bisamidines 1-3 using homopolymers poly(dA-dT)- poly(dA-dT) and poly(dG-dC)- poly(dG-dC), indicated that these compounds show moderate specificity for AT base pairs. We studied the cytotoxicity effects of bisamidines 1-3, Hoechst 33258 and DAPI (4',6-diamidino-2-phenylindole) in cultured breast cancer MCF-7 cells. The bisamidines 1-3 showed comparable antitumour activity to Hoechst 33258, but were substantially more cytotoxic compared to DAPI. These data show that in broad terms the cytotoxic potency of bisamidines 1-3 in cultured breast cancer MCF-7 cells decreases with the size of the alkyl group substituent (cyclopropyl>isopropyl>cyclopentyl), in accord with their increases in DNA affinity, as shown by the binding constant values.

5-Bromodeoxyuridine suppresses position effect variegation of transgenes in HeLa cells

An ectopic gene integrated in the host genome is occasionally silenced due to a position effect of its adjacent chromatin structure. We found that 5-bromodeoxyuridine clearly activated such a transgene in HeLa cells. The transgene was also activated to various degrees by inhibitors of histone deacetylase, DNA topoisomerases, or DNA methyltransferase. The peptide antibiotic distamycin A potentiated markedly the effect of 5-bromodeoxyuridine. Transient expression of an artificial AT-hook protein termed MATH20 also potentiated its effect although significantly activated the transgene alone. Since distamycin A and MATH20 are able to displace histone H1 and other DNA-binding proteins bound to specific AT-rich sequences by a dominant, mutually exclusive fashion, these results suggest that 5-bromodeoxyuridine targets such an AT-rich sequence located adjacent to the silenced transgene, resulting in chromatin accessibility.

Intra- and intermolecular triplex DNA formation in the murine c-myb proto-oncogene promoter are inhibited by mithramycin

Mithramycin inhibits transcription by binding to G/C-rich sequences, thereby preventing regulatory protein binding. However, it is also possible that mithramycin inhibits gene expression by preventing intramolecular triplex DNA assembly. We tested this hypothesis using the DNA triplex adopted by the murine c-myb proto-oncogene. The 5'-regulatory region of c-myb contains two polypurine:polypyrimidine tracts with imperfect mirror symmetry, which are highly conserved in the murine and human c-myb sequences. The DNA binding drugs mithramycin and distamycin bind to one of these regions as determined by DNase I protection assay. Gel mobility shift assays, nuclease and chemical hypersensitivity and 2D-gel topological analyses as well as triplex-specific antibody binding studies confirmed the formation of purine*purine:pyrimidine inter- and pyrimidine*purine:pyrimidine intra-molecular triplex structures in this sequence. Mithramycin binding within the triplex target site displaces the major groove-bound oligonucleotide, and also abrogates the supercoil-dependent H-DNA formation, whereas distamycin binding had no such effects. Molecular modeling studies further support these observations. Triplex-specific antibody staining of cells pretreated with mithramycin demonstrates a reversal of chromosomal triplex structures compared to the non-treated and distamycin-treated cells. These observations suggest that DNA minor groove-binding drugs interfere with gene expression by precluding intramolecular triplex formation, as well as by physically preventing regulatory protein binding.

Interference of transcriptional activation by the antineoplastic drug ecteinascidin-743

Ecteinascidin-743 (ET-743) is a tetrahydroisoquinoline alkaloid isolated from the tunicate Ecteinascidia turbinata currently under phase II clinical trials for its potent anticancer activity. ET-743 binds DNA in the minor groove and forms covalent adducts with some sequence specificity. It selectively inhibits in vitro binding of the CCAAT box factor NF-Y. In this study, we assayed ET-743 function in vivo on the HSP70 promoter. On heat induction, the drug blocks transcription rapidly at pharmacological concentrations and in a CCAAT-dependent manner, whereas the activity of the CCAAT-less simian virus 40 promoter is not affected. The effect is exerted at the mRNA level. The distamycin-like alkylating tallimustine is inactive in these assays. Binding of NF-Y and of the heat-shock factor is normal in ET-743-treated cells. Run-on analysis of several endogenous genes further proves that the drug has rapid, profound, and selective negative effects on transcription. Thus, this marine-derived compound is a promoter-specific, transcription-interfering agent.

Bisanthracycline WP631 inhibits basal and Sp1-activated transcription initiation in vitro

An in vitro transcription assay was used to compare the capacity of the bisintercalating anthracycline WP631 (which displays a remarkably high DNA-binding affinity) and the monointercalating anthracycline daunomycin to inhibit transcription initiation of the adenovirus major late promoter linked to a G-less transcribed DNA template. Both drugs inhibit basal RNA synthesis in a concentration-dependent way, and the drug concentrations required to inhibit transcription initiation are similar. However, in this study WP631 was around 15 times more efficient at inhibiting transcription initiation when used with an adenovirus promoter containing an upstream Sp1-protein binding site under experimental conditions in which the Sp1 protein acted as a transactivator in vitro. The differences in the ability of each drug to inhibit transcription initiation were related to the competition between Sp1 and the drugs for the same binding site. Concentrations of WP631 as low as 60 nM could inhibit the Sp1-activated transcription initiation in vitro. In contrast, the concentration of daunomycin required to inhibit Sp1-activated transcription by 50% was almost the same as the concentration required to inhibit basal transcription. The efficiency of WP631 at displacing Sp1 from its putative binding site was confirmed using gel retardation and footprinting assays. These results are the first unequivocal example of a direct effect of an intercalator on activated transcription initiation.

The DNA-binding drugs mithramycin and chromomycin are powerful inducers of erythroid differentiation of human K562 cells

The human leukaemic K562 cell line can be induced in vitro to undergo erythroid differentiation by a variety of chemical compounds, including haemin, butyric acid, 5-azacytidine and cytosine arabinoside. Differentiation of K562 cells is associated with an increased expression of embryo-fetal globin genes, such as the zeta, epsilon and gamma globin genes. Therefore the K562 cell line has been proposed as a useful in vitro model system to determine the therapeutic potential of new differentiating compounds as well as to study the molecular mechanism(s) regulating changes in the expression of embryonic and fetal human globin genes. Inducers of erythroid differentiation which stimulate gamma-globin synthesis could be considered for possible use in the experimental therapy of those haematological diseases associated with a failure in the expression of adult beta-globin genes. In this paper we demonstrated that the G + C selective DNA-binding drugs chromomycin and mithramycin were powerful inducers of erythroid differentiation of K562 cells. Erythroid differentiation was associated with an increase in the accumulation of (a) Hb Gower 1 and Hb Portland and (b) gamma-globin mRNA.

Nitric oxide inhibits Oct-1 DNA binding activity in cultured vascular smooth muscle cells

Since Oct-1 is a ubiquitous DNA binding protein shown to play an important role in regulating cell proliferation and possess structural characteristics consistent with a nitric oxide (NO) target, we studied NO regulation of the DNA binding activity of Oct-1 in the A7R5 vascular smooth muscle cell (VSMC) line. Two NO donors, sodium nitroprusside (SNP) and S-nitroso-N-acetyl-penicillamine (SNAP) were directly added to the nuclear extract-oligonucleotide reaction mixture, respectively and the effect on Oct-1 DNA binding activity was evaluated by gel shift assay. Both NO donors (0.01-1 mM) inhibited the DNA binding activity of Oct-1. This inhibitory effect was not attenuated by dithiothreitol (DTT) (1 mM) while in contrast, DTT did antagonize the effect of diamide on Oct-1 DNA binding activity. The NO effect on Oct-1 has some specificity; as the NO donors had no effect on myc DNA binding activity. The inhibitory effect of NO donors was reproduced in A7R5 cells, without affecting their viability. These findings provide the first evidence that NO inhibits the DNA binding activity of Oct-1, probably through a cGMP independent mechanism and suggests that NO may inhibit mitogenesis in part through an effect on Oct-1 DNA binding activity in VSMCs.

Targeting of the HIV-1 long terminal repeat with chromomycin potentiates the inhibitory effects of a triplex-forming oligonucleotide on Sp1-DNA interactions and in vitro transcription

We have studied the effects of chromomycin and of a triple-helix-forming oligonucleotide (TFO) that recognizes Sp1 binding sites on protein-DNA interactions and HIV-1 transcription. Molecular interactions between chromomycin, the Sp1 TFO and target DNA sequences were studied by gel retardation, triplex affinity capture using streptavidin-coated magnetic beads and biosensor technology. We also determined whether chromomycin and a TFO recognizing the Sp1 binding sites of the HIV-1 long terminal repeat (LTR) inhibit the activity of restriction enzyme HaeIII, which recognizes a sequence (5'-GGCC-3') located within these Sp1 binding sites. The effects of chromomycin and the TFO on the interaction between nuclear proteins or purified Sp1 and a double-stranded oligonucleotide containing the Sp1 binding sites of the HIV-1 LTR were studied by gel retardation. The effects of both chromomycin and TFO on transcription were studied by using an HIV-1 LTR-directed in vitro transcription system. Our results indicate that low concentrations of chromomycin potentiate the effects of the Sp1 TFO in inhibiting protein-DNA interactions and HIV-1-LTR-directed transcription. In addition, low concentrations of chromomycin do not affect binding of the TFO to target DNA molecules. The results presented here support the hypothesis that both DNA binding drugs and TFOs can be considered as sequence-selective modifiers of DNA-protein interactions, possibly leading to specific alterations of biological functions. In particular, the combined use of chromomycin and TFOs recognizing Sp1 binding sites could be employed in order to abolish the biological functions of promoters (such as the HIV-1 LTR) whose activity is potentiated by interactions with the promoter-specific transcription factor Sp1.

Structure-activity relationship of a series of C-terminus modified aminoalkyl, diaminoalkyl- and anilino-containing analogues of the benzoic acid mustard distamycin derivative tallimustine: synthesis, DNA binding and cytotoxicity studies

As part of our investigations into the design of more cytotoxic analogues of the experimental anticancer drug tallimustine, 1, C-terminus modified aminoalkyl-, 2a-c, diaminoalkyl-, 3, and anilino-containing, 4, derivatives have been synthesized. Compounds 2a-c differ by 2, 3, or 4 methylene units in the C-terminus, respectively. Results from an ethidium displacement study on poly(dA-dT), poly(dG-dC), calf thymus DNA and T4 coliphage DNA showed that compounds 2-4 interact in the minor groove of the polynucleotides with a preference for poly(dA-dT) over poly(dG-dC). Compound 4 bound more weakly to the DNAs than 2a-c and 3. Using a CD dilution assay compounds 2a-c and 3 were demonstrated to bind irreversibly to calf thymus DNA. The sequence selectivity by which compounds 2-4 alkylate DNA was demonstrated using a Taq polymerase stop assay. All the compounds alkylated preferentially at the 3'-purine residue in a 5'-TTTTGPu-3' sequence (Pu = A or G). This observed sequence specificity is similar to that of tallimustine and a related compound 5. At an equimolar concentration the aminoalkyl compounds 2a-c (2b > 2a > 2c), and diaminoalkyl compound 3 were more efficient at alkylating these sequences than the anilino compound 4. Following a one hour exposure of human chronic myeloid leukemia K562 cells, compounds 2b and 3 have lower IC50 values (1.64 microM and 3.03 microM, respectively) than tallimustine (5 microM) and similar values to a related compound 5 (2.2 microM). The order of cytotoxicity for all the compounds is 2b > 5 > 3 > 2a > 1 > 2c = 4. These results indicate that the cytotoxicities of these compounds are related to their relative ability to alkylate the consensus DNA binding sequence.

Sequence-selective binding to DNA of bis(amidinophenoxy)alkanes related to propamidine and pentamidine

The DNA sequences targeted by a complete homologous series of aromatic diamidines have been determined at single-nucleotide resolution via protection from cutting by the endonucleases DNase I, DNase II and micrococcal nuclease. Propamidine, pentamidine and to a lesser extent hexamidine bind selectively to nucleotide sequences composed of at least four consecutive A-T base pairs. In contrast, the binding to DNA of butamidine, heptamidine, octamidine and nonamidine is poorly sequence-selective. Sequences composed of only three consecutive A-T base pairs do not afford a potential binding site for propamidine or the longer homologues, and none of the drugs tolerate the presence of a G-C base pair within the binding site. Experiments with DNA molecules containing inosine in place of guanosine and 2,6-diaminopurine in place of adenine reveal that the lack of binding of propamidine to GC-containing sites is attributable to an obstructive effect of the exocyclic 2-amino group of guanosine. The present data support the view that the local conformation of the double helix (in particular the width of the minor groove) plays a dominant role in the binding reaction and that the capacity of diamidines to recognize AT-rich sequences selectively varies considerably depending on the length of the alkyl chain. The evidence indicates that binding to AT-tracts in DNA must play a role in the biological activity of these diamidines, but there is no simple correlation between binding and pharmacological efficacy.

The anti-cancer agent distamycin A displaces essential transcription factors and selectively inhibits myogenic differentiation

The anticancer drug, distamycin A, alters DNA conformation by binding to A/T-rich domains. We propose that binding of the drug to DNA alters transcription factor interactions and that this may alter genetic regulation. We have analyzed the effects of distamycin A upon expression of the muscle-specific cardiac and skeletal alpha-actin genes which have A/T-rich regulatory elements in their promoters. Distamycin A specifically inhibited endogenous muscle genes in the myogenic C2 cell line and effectively eliminated the myogenic program. Conversely, when 10T1/2C18 derived pleuripotential TA1 cells were induced to differentiate in the presence of distamycin A, adipocyte differentiation was enhanced whereas the numbers of cells committing to the myogenic program decreased dramatically. Using the mobility shift assay distamycin A selectively inhibited binding of two important transcription factors, SRF and MEF2, to their respective A/T-rich elements. The binding of factors Sp1 and MyoD were not affected. The inhibition of factor binding correlated with a repression of muscle-specific promoter activity as assayed by transient transfection assays. Co-expression of the myoD gene, driven by a distamycin A-insensitive promoter, failed to relieve the inhibition of these muscle-specific promoters by distamycin A. Additionally, SRF and MEF2 dependent promoters were selectively down regulated by distamycin A. These results suggest that distamycin A may inhibit muscle-specific gene expression by selectively interfering with transcription factor interactions and demonstrate the importance of these A/T-rich elements in regulating differentiation of this specific cell type.

In vitro and in vivo binding of a CC-1065 analogue to human gene sequences: a polymerase-chain reaction study

In this paper we analyse the in vitro sequence selectivity of the CC-1065 analogue 2-[[5-[(1H-indol-2-yl]carbonyl)-1H-indol-2-yl] carbonyl]-7-methyl-1,2,8,8a-tetrahydrocyclopropa [c]-pyrrolo-[3,2-e]-indol-4-one (U-71184) employing the polymerase-chain reaction (PCR). In addition, we determined whether alteration of PCR by U-71184 is detected when DNA is isolated from cells cultured in the presence of this drug. As molecular model systems we employed the human estrogen receptor gene, the Ha-ras oncogene and the chromosome X-linked, (CGG)-rich fragile X mental retardation-1 gene. The first conclusion that can be drawn from the experiments reported in our paper is that U-71184 inhibits PCR in a sequence-dependent manner. A second conclusion of our experiments is that PCR performed on DNA from U-71184-treated cells is inhibited when the primers amplifying the estrogen receptor gene region are used. This approach might bring important information on both in vivo uptake of the drug by target cells and binding to DNA.

Effects of minor and major groove-binding drugs and intercalators on the DNA association of minor groove-binding proteins RecA and deoxyribonuclease I detected by flow linear dichroism

Linear and circular dichroic spectroscopies have been employed to investigate the effects of small DNA ligands on the interactions of two proteins which bind to the minor groove of DNA, viz. RecA protein from Escherichia coli and deoxyribonuclease I (bovine pancreas). Ligands representing three specific non-covalent binding modes were investigated: 4',6-diamidino-2-phenylindole and distamycin A (minor groove binders), methyl green (major groove binder), and methylene blue, ethidium bromide and ethidium dimer (intercalators). Linear dichroism was demonstrated to be an excellent detector, in real time, of DNA double-strand cleavage by deoxyribonuclease I. Ligands bound in all three modes interfered with the deoxyribonuclease I digestion of dsDNA, although the level of interference varied in a manner which could be related to the ligand binding site, the ligand charge appearing to be less important. In particular, the retardation of deoxyribonuclease I cleavage by the major groove binder methyl green demonstrates that accessibility to the minor groove can be affected by occupancy of the opposite groove. Binding of all three types of ligand also had marked effects on the interaction of RecA with dsDNA in the presence of non-hydrolyzable cofactor adenosine 5'-O-3-thiotriphosphate, decreasing the association rate to varying extents but with the strongest effects from ligands having some minor groove occupancy. Finally, each ligand was displaced from its DNA binding site upon completion of RecA association, again demonstrating that modification of either groove can affect the properties and behaviour of the other. The conclusions are discussed against the background of previous work on the use of small DNA ligands to probe DNA-protein interactions.

Targeting of the Sp1 binding sites of HIV-1 long terminal repeat with chromomycin. Disruption of nuclear factor.DNA complexes and inhibition of in vitro transcription

Sequence selectivity of DNA-binding drugs has recently been reported in a number of studies employing footprinting and gel retardation approaches. In this paper, we studied the biochemical effects of the sequence-selective binding of chromomycin to the long terminal repeat of the human immunodeficiency type I virus. Deoxyribonuclease I (E.C.3.1.21.1) footprinting, arrested polymerase chain reaction, gel retardation and in vitro transcription experiments have demonstrated that chromomycin preferentially interacts with the binding sites of the promoter-specific transcription factor Sp1. Accordingly, interactions between nuclear proteins and Sp1 binding sites are inhibited by chromomycin, and this effect leads to a sharp inhibition of in vitro transcription.

Characterization of a protein recognizing minor groove binders-damaged DNA

By using electromobility shift assay (EMSA), we have identified a protein able to recognize the DNA only if it was previously reacted with minor groove binders. This protein binds with very high affinity AT containing DNA treated with minor groove binders such as distamycin A, Hoechst 33258 and 33342, CC-1065 and ethidium bromide minor groove intercalator, but not with major groove binders such as quinacrine mustard, cisplatin or melphalan, or with topoisomerase I inhibitor camptothecin or topoisomerase II inhibitor doxorubicin. This protein was found to be present in different extracts of human, murine and hamster cells, with the human protein which appears to have a molecular weight slightly lower than that of the other species. This protein was found to be expressed both in cancer and normal tissues. By using molecular ultrafiltration techniques as well as southwestern analysis it was estimated that the apparent molecular weight is close to 100 kDa. We can exclude an identity between this protein and other proteins, with a similar molecular weight previously reported to be involved in DNA damage recognition/repair, such as topoisomerase I, mismatch repair activities such as the prokaryotic MutS protein and its human homologue hMSH2 or proteins of the nucleotide excision repair system such as ERCC1, -2, -3 and -4.

Systematic study on the chemical stability of the prodrug antitumor agent carzelesin (U-80,244)

The chemical stability of the novel anticancer agent carzelesin in aqueous buffer/acetonitrile (1:1, v/v) mixtures has been investigated utilizing a stability-indicating reversed-phase high-performance liquid chromatographic assay. The degradation kinetics of carzelesin has been studied as a function of pH, buffer composition, ionic strength, and temperature. Degradation of carzelesin follows (pseudo-) first-order kinetics. A pH-rate profile, using rate constants extrapolated to zero buffer concentration, was constructed demonstrating that carzelesin is most stable in the pH region 1-4. The degradation rate of carzelesin was not significantly affected by buffer components and by the ionic strength. In addition to the formation of the degradation products U-76,073, U-76,074, and aniline in alkaline medium and in acetate buffer solution, another degradation product was formed in acetate buffer solution. In perchloric acid buffer solution (pH* < 3), U-76,073 and U-76,074 could not be detected as degradation products.

Effect of DNA-binding drugs on early growth response factor-1 and TATA box-binding protein complex formation with the herpes simplex virus latency promoter

Adjacent binding sites for early growth response factor-1 (EGR1) and TATA box-binding protein (TBP) were identified on the herpes simplex virus latency promoter in previous work. The binding of EGR1 to the GC-rich region prevented TBP binding to the AT-rich region. With the simultaneous addition of both EGR1 and TBP, the intercalator nogalamycin prevented EGR1 complex formation, resulting in a dose-dependent increase of the TBP.DNA complex. The minor groove binder chromomycin A3 inhibited EGR1 complex formation but resulted in a smaller increase of the TBP complex. In contrast, an alkylating intercalator hedamycin strongly inhibited binding of both proteins. The ability of these GC-binding drugs to prevent EGR1.DNA complex formation was in the following order: hedamycin > nogalamycin > chromomycin A3, and the specificity was nogalamycin > chromomycin A3 > hedamycin. With transcription factor IIA (TFIIA) in the assay, TBP was able to bind the promoter whereas formation of the EGR1.DNA complex was reduced. An AT minor groove-binding drug, distamycin A, disrupted the TBP.TFIIA.DNA complex and restored the EGR1.DNA complex. We conclude that the binding motif and sequence preference of DNA-interactive drugs are manifested in their ability to inhibit the transcription factor-DNA complexes.

Evidence for distinct regulation processes in the aclacinomycin- and doxorubicin-mediated differentiation of human erythroleukemic cells

Human erythroleukemic K 562 cells were induced to were induced to differentiate along the erythroid lineage by anthracycline antitumor drugs, such as aclacinomycin (ACLA) and doxorubicin (DOX). Subsequent stimulation of heme and globin synthesis led to a differential quantitative expression of hemoglobins. Gower 1 (epsilon2, zeta2) was the major type for ACLA and X (epsilon2, gamma2) for DOX. Although ACLA and DOX increased both the expression of gamma-globin and porphobilinogen deaminase mRNAs, striking differences were observed in the expression of erythropoietin receptor mRNAs and in erythroid transcription factors GATA-1 and NF-E2, known to play a key role in erythroid gene regulation. Indeed, ACLA induces an increase either in the binding capacity of GATA-1 and NF-E2 or in the accumulation of erythropoietin receptor, GATA-1 and NF-E2 transcripts. In contrast, their expression with DOX was not significantly modified compared to uninduced cells, except for a slight decrease in NF-E2 expression on day 3. In conclusion, these data show that: 1. increased expression of erythroid transcription factors and erythroid genes are associated only with ACLA treatment, and 2. although cytotoxicity of both ACLA and DOX is certainly dependent on DNA intercalation, regulation of differentiation processes by these two drugs involves distinct mechanisms.

Distamycin-induced inhibition of formation of a nucleoprotein complex between the terminase small subunit G1P and the non-encapsidated end (pacL site) of Bacillus subtilis bacteriophage SPP1

The small subunit of the Bacillus subtilis bacteriophage SPP1 terminase (G1P) forms a sequence-specific nucleoprotein complex with the SPP1 non-encapsidated end (pacL site) during initiation of DNA encapsidation. Gel mobility shift assay was used to study the G1P-pacL interaction. Distamycin, a minor groove binder that induces local distortion of the DNA, inhibits G1P-pacL complex formation. The competition of G1P with distamycin for DNA binding at the pacL site is independent of the order of addition of the reactants. Other minor groove binders, such as spermine or Hoechst 33258, which do not distort DNA, failed to compete with G1P for pacL DNA binding. Cationic metals, which generate a repertoire of DNA structures different from that caused by the minor groove binders, can partially reverse the distamycin-induced inhibition of G1P binding to pacL DNA. The major groove binder methyl green, which does not distort sequence-directed bending of pacL DNA, competes with G1P for binding at the pacL site. Our data suggest that the natural sequence-directed bend that exists within the pacL site is the architectural element that facilitates assembly of a nucleoprotein complex and hence initiation of DNA encapsidation by bacteriophage SPP1.

3'-Azido-3'-deoxythymidine inhibits erythroid-specific transcription factors in human erythroid K562 leukemia cells

The present study examines genetic mechanism(s) possibly involved in the observed 3'-azido-3'-deoxythymidine (AZT)-induced inhibition of globin gene transcription by evaluating the direct phenotypic erythroid effects of AZT on erythroid-specific transcription factors which regulate globin gene promoters. In vitro binding of GATA-1 or NFE-2 to its consensus sequence was decreased in the presence of AZT reaching a maximum inhibition as early as 24 h after AZT treatment. Nuclear extracts from butyric acid-induced K562 cells treated with an IC50 concentration of AZT exhibited a decrease in GATA-1 and NFE-2 binding by approximately 30% and 35%. In contrast, 2',3'-dideoxycytidine which inhibits cell growth without affecting hemoglobin synthesis, had no effect on binding of GATA-1 and NFE-2 factors. Northern blot analysis revealed a 25% decrease by AZT in GATA-1 mRNA steady-state levels at 24 h and this inhibitory effect was maintained until 72 h after drug addition. A similar decrease in NFE-2 mRNA steady-state levels was observed at 72 h after AZT treatment. This study suggests that AZT inhibition of erythroid differentiation is subsequent to a decrease of nuclear factors gene expression which affect their DNA binding.

CCAAT-box binding protein NF-Y (CBF, CP1) recognizes the minor groove and distorts DNA

The CCAAT box is one of the most common promoter elements. The evolutionarily conserved heteromeric factor NF-Y binds this sequence with high affinity and specificity. By comparing the methylation interference patterns of different sites, performing electrophoretic mobility shift assays (EMSA) with IC-substituted oligonucleotides and competition experiments with the minor groove binding (MGB) drugs distamicin A, tallimustine and Hoechst 33258 we show that NF-Y makes key minor groove interactions. Circular permutation assays on four CCAAT boxes, MHC Class II Ea, HSP70, epsilon-globin and MSV, indicate that NF-Y is able to distort the double helix by angles of 62-82 degrees, depending on the site used, and suggest that nucleotides flanking the CCAAT pentanucleotide influence the degree of bending.

Competition between netropsin and restriction nuclease EcoRI for DNA binding

We find that netropsin and netropsin analogue protect DNA from EcorI restriction nuclease cleavage by inhibiting the binding of EcoRI to its recognition site. The drug -- EcoRI competitive binding constants measured by a electrophoretic gel mobility shift assay are in excellent agreement with the nuclease protection results for the netropsin analogue and in reasonable agreement for netropsin itself. Crystal structures of complexes show that netropsin and EcoRI recognize different regions of the DNA helix and would not be expected to compete for binding to the restriction nuclease site. The large distortions in DNA structure caused by EcoRI binding are most likely responsible for an indirect structural competition with netropsin binding. The structural change in the netropsin binding region induced by EcoRI binding to its region essentially prevents drug association. Given the reciprocal nature of competition, binding of netropsin to a minimally perturbed structure then also makes the association of EcoRI energetically more costly. Since many sequence specific DNA binding proteins significantly bend or distort the DNA helix, drugs that compete indirectly can be as effective as drugs that act through a direct steric inhibition.

Alteration of the expression of human estrogen receptor gene by distamycin

The effects of distamycin on the expression of the estrogen receptor gene were determined in the MCF7 human breast cancer cell line. Estrogen receptor (ER) RNA transcripts were analyzed by Northern blotting and RT-PCR using specific oligonucleotides for the 5' upstream region and for ER cDNA. After ex vivo distamycin treatment of the cells the expression of the canonical ER mRNA isoform of 6.3 kb is strongly inhibited, without appreciable alteration of the accumulation of 5' upstream ER mRNA isoforms. These results suggest that distamycin alters the transcriptional activity of the ER gene causing a change in the ratio between the canonical transcript and other isoforms containing 5' upstream regions.

Binding of distamycin and chromomycin to human immunodeficiency type 1 virus DNA: a non-radioactive automated footprinting study

Sequence-selectivity of DNA-binding drugs was recently reported in a number of studies employing footprinting and gel retardation approaches. In this paper we studied sequence-selectivity of the binding of chromomycin and distamycin to DNA by performing DNase I footprinting and analysis of the cleaved fragments by the Pharmacia ALF DNA Sequencing System. As a model system we employed the long terminal repeat of the human immunodeficiency type 1 virus. The main conclusion of our experiments is that automated analysis of DNase I footprinting is a fast and reliable technique to study drugs-DNA interactions. The results obtained suggest that distamycin and chromomycin differentially interact with the long terminal repeat of the human immunodeficiency type 1 virus; this differential binding depends upon the DNA sequences recognized. The data presented are consistent with a preferential binding of distamycin to DNA sequences of the binding sites of nuclear factor kappa B and transcription factor IID. By contrast, distamycin exhibits only weak binding to DNA sequences recognized by the promoter-specific transcription factor Sp1. Unlike distamycin, chromomycin preferentially interacts with the binding sites of the promoter-specific transcription factor Sp1.

Comparison of cell-cycle phase perturbations induced by the DNA-minor-groove alkylator tallimustine and by melphalan in the SW626 cell line

Tallimustine or N-deformyl-N-[4-N-N,N-bis(2-chloroethylamino)benzoyl], a distamycin-A derivative (FCE 24517), is a novel anti-cancer agent which alkylates N3 adenine in the minor groove of DNA. The cell-cycle phase perturbations induced by the drug were investigated and compared with those caused by melphalan (L-PAM) in SW626 human ovarian-cancer cells. By coupling bromodeoxyuridine (BUdR) immunoreaction with biparametric flow-cytometric (FCM) analysis, we investigated the cell-cycle phase perturbation induced by tallimustine or L-PAM, considering separately the cells which, during the 1-hr treatment, were in the S phase or in G1-G2/M phases of the cell cycle. L-PAM delayed the S-phase progression of cells exposed to the drug when they were in S phase, with a consequent accumulation of cells as soon as they reached the G2 phase. In contrast, the S-phase cells treated with tallimustine were not perturbed during the DNA-synthesis phase progression, and were blocked in G2 only after they had passed through the G1/S transition of a new cell cycle. In cells which were in G1 or G2/M phases during drug treatment, tallimustine and L-PAM caused similar accumulation in G2. The differences in the cell-cycle perturbation caused by tallimustine and L-PAM may well be related to the different DNA damage the 2 drugs produced. These findings emphasize the different properties of DNA-minor-groove alkylating agents and conventional ones.

Differential effects of distamycin analogues on amplification of human gene sequences by polymerase-chain reaction

In this report we analyse the effects of distamycin and five distamycin analogues on amplification by polymerase-chain reaction (PCR) of two gene sequences displaying a different A+T/G+C content. The first was a 5' region of the human oestrogen receptor (ER) gene, containing a (TA)26 stretch; the second was a CG-rich sequence of the human Ha-ras oncogene. The results obtained unequivocally demonstrate that the addition of one pyrrole ring significantly improves the ability of distamycin derivatives to interfere with PCR-mediated amplification of the human ER genomic region carrying a (TA)26 stretch. The distamycin analogues analysed differ in the number of pyrrole rings and in the presence of an N-formyl, an N-formimidoyl or a retroamide group at position X1. Among compounds carrying the same number of pyrrole rings, those carrying an N-formyl or an N-formimidoyl group retain a similar inhibitory activity. The retroamide analogues, on the contrary, are much less efficient in inhibiting PCR-mediated amplification of the 5'ER region. With respect to sequence selectivity both distamycin and distamycin analogues exhibit a sequence preference, since they do not inhibit PCR amplification of Ha-ras CG-rich gene regions, with the exception of a distamycin analogue carrying four pyrrole rings.

Distamycin A and tallimustine inhibit TBP binding and basal in vitro transcription

The antibiotic distamycin A is a DNA minor groove binding drug (MGB) that recognizes a stretch of at least four ATs. The alkylating benzoyl mustard derivative tallimustine (FCE 24517) has powerful anti-tumor activity. Using the electrophoretic mobility shift assay (EMSA) we determined that both compounds can prevent binding of TBP and, with 10-fold higher concentration, TBP-TFIIA (DA) and TBP-TFIIA-TFIIB (DAB) to a TATA box. Once formed, the DA and DAB complexes are more resistant to MGB challenge. Both drugs can inhibit basal in vitro transcription of a minimal TATA-containing promoter and similar concentrations are necessary for binding and transcriptional inhibition. Tallimustine shows strong selectivity by decreasing only correctly initiated transcripts. Even at high doses (20 microM), however, they cannot disturb a competent pre-initiation complex or Pol II progression. This functional in vitro model will provide a way to investigate the activity of sequence-specific DNA binding drugs with potential anti-viral and anti-tumour activity and to develop novel more selective compounds.

The effect of AT and GC sequence specific minor groove-binding agents on restriction endonuclease activity

The ability of the naturally occurring A/T specific DNA minor groove binders netropsin and diastamycin A and two synthetic G/C selective oligopeptide analogues (1 and 2), to interfere with the catalytic activity of restriction endonucleases has been investigated. Enzymes were chosen to have A/T rich (EcoRI, EcoRV) or G/C rich (BalI, NruI) recognition sequences. An agarose gel assay was used to measure the cleavage of 32P-labelled DNA and ligand-DNA binding data was obtained using methidium-propyl EDTA footprinting. Netropsin and distamycin bind at the recognition sites, and dose-dependently inhibited cleavage by, EcoRI and EcoRV, (EcoRI > EcoRV). They were also more effective at inhibiting the catalytic activity of BalI than either 1 or 2. NruI was inhibited by distamycin and 2, but not by netropsin or 1. DNA footprinting revealed that neither 1 or 2 bound to the BalI or NruI recognition sequences under the conditions used whereas netropsin and distamycin footprint at adjacent sites. 1 binds to two of the three recognition sequences for the enzyme Fnu4HI (GCNGC) in the fragment studied and was shown to inhibit DNA cleavage only at these two sites. 2 binds strongly to two GGGCTC sequences which are recognition sites for the enzyme BanII. In this case a pronounced stimulation of cleavage was observed in the presence of 2 over a wide dose range. The results indicate that enzyme inhibition does not necessarily result from simultaneous occupancy of a common site, or at nearby flanking sequences, and in some circumstances, a pronounced stimulation of enzyme cleavage can occur.

Pharmaceutical development of a parenteral formulation of the novel anti-tumor agent carzelesin (U-80,244)

The aim of this study was to design a parenteral dosage form for the investigational cytotoxic drug carzelesin. A stable formulation in PET (Polyethylene glycol 400/absolute ethanol/Tween 80, 6:3:1, v/v/v) was developed. The prototype, containing 0.50 mg carzelesin in 2.0 ml PET formulation, was found to be the optimal formulation in terms of solubility, stability and dosage requirements in phase I clinical trials. Quality control of the formulation showed that the pharmaceutical preparation of carzelesin in PET is not negatively influenced by the manufacturing process. Shelf life studies demonstrated that the formulation is stable for at least 1 year, when stored at -30 degrees C in the dark. In addition, the stability of carzelesin in the PET formulation is discussed as a function of temperature, additives and after dilution in infusion fluids.

Polymerase-chain reaction as a tool for investigations on sequence-selectivity of DNA-drugs interactions

Sequence-selectivity of DNA-binding drugs was recently reported in a number of studies employing footprinting and gel retardation approaches. In this paper we performed polymerase-chain reaction (PCR) experiments to study the in vitro effects of distamycin, daunomycin, chromomycin and mithramycin. As model systems we employed the human estrogen receptor (ER) gene and the Harvey-ras (Ha-ras) oncogene, in order to obtain PCR products significantly differing for the A + T/G + C frequency ratio. Distamycin, daunomycin, chromomycin and mithramycin are indeed known to differentially bind to different DNA regions depending upon the DNA sequences recognized. The main conclusion of our experiments is that distamycin, daunomycin, chromomycin and mithramycin inhibit polymerase-chain reaction in a sequence-dependent manner. Distamycin inhibits indeed PCR mediated amplification of AT-rich regions of the human estrogen receptor gene, displaying no inhibitory effects on PCR-mediated amplification of GC-rich sequences of Ha-ras oncogene. By contrast daunomycin, chromomycin and mithramycin were found to inhibit PCR-mediated amplification of the Ha-ras GC-rich oncogene sequences. We propose that polymerase-chain reaction technique could be applied to study the in vivo interactions of DNA-binding drugs to specific genes in intact cells.