Nucleic Acid Ligands Based on Carbohydrates

MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 248

C'HIMIA 50 (I 99() Nr. () (Jlllli)

edizinisc edicinal Chemistry

Chimia 50 (1996) 248-256 Watson-Crick pairing rules. Translation is © Neue Schweizerische Chemische Gesellschaft arrested by this complexation and the RNA ISSN 0009-4293 strand is then degraded by RNase H, an enzyme which recognizes DNA-RNA het- eroduplexes, leading to catalytic turnover Nucleic Acid Ligands Based on of the antisense agent. In a different approach, single-stranded oligonucleotides Carbohydrates are designed such as to form a local triple- helical structure with the third strand associated in the major groove of double- Jiirg Hunziker* stranded DNA. In this case, the sequence specificity is derived from selective base Abstract. Sequence-specific nucleic acid ligands are important tools in chemistry and triplet formation on the Hoogsteen face of molecular biology and are thought to possess a considerable pharmaceutical potential. purine bases (Fig. 2). Since all such oligo- An overview of the structurally and mechanistically diverse approaches in the field with nucleotide analogs are linear polymers an emphasis on carbohydrates will be presented. their affinity and specificity can be altered Carbohydrates are just beginning to emerge as a novel class of nucleic acid binding at will by changing chain length and base compounds. The detailed study of the factors affecting the site-selectivity of some sequence. recently discovered antitumor antibiotics, e.g. calicheamicin, has shed a new light on Several clinical trials are currently un- the role thatoligosaccharides may play in nucleic acid recognition. Understanding these der way and although there have been factors may aid in the rational design of novel nucleic acid ligands and their therapeutic some promising results, difficulties have application. been encountered in getting antisense oligonucleotides to work the way they were intended to [le][2]. Selective delivery to 1. Introduction dance of the protein sequence and individ- the target tissue and potentially toxic side ual nucleotides. In addition, their bio- effects such as decreased blood clotting Nucleic acids are attractive targets for availability is low due to the relative ease and inreased blood pressure are the main the development of novel therapeutics. with which they are degraded and the slow obstacles. An improved understanding of The sequence specific, high affinity com- transport across cell membranes. And they the fate of antisense oligonucleotides in plexation of single-stranded messenger- are difficult and rather expensive to pro- vivo in conjunction with the development RNA or the corresponding double-strand- duce. Oligonucleotides or rather syntheti- of novel, more specific analogs will be ed genomic DNA can, in principle, pre- cally derived analogs, on the other hand, necessary to successfully generate thera- vent the production of illness related gene are obvious candidates for this task [I]. In peutics based on this concept. products at the stage of translation or tran- the antisense strategy, single-stranded oli- scription, respectively (Fig. 1). The inher- gonucleotides are designed to pair to an 1.2. Small DNA-Binding Molecules ent advantage of this approach lies in the RNA target (sense strand) by forming a In view of these difficulties, numerous low abundance of a certain gene or the local duplex structure according to the academic and industrial research groups resulting messenger-RNA within an af- fected cell. This theoretically requires a much lower dosage of a potential thera- RNA~.rget: inhibition of translation peutic agent. However, it also necessitates very high affinities and selectivities. Nu- cleic acid ligands based on this concept would act as direct antagonists to the cellular proteins that control translation and transcription.

1.1. Antisense Agents

Proteins themselves are poor candi- =====~>s DNA-target: inhibition of transcription antisense- or dates for these purposes. There is no gen- antigene agent eral code for the recognition of nucleic ~ acids by proteins, a one to one correspon- *\

*Correspondence: Dr. J. Hunziker Institut fUr Organische Chemie Universitat Bern Fig. 1. Schematic representation of a generalized antigene-concept. By complex formation with Freieslrasse 3 messenger-RNA or the corresponding DNA anti gene agents can shut down production of illness- CH-30 12 Bern related proteins at the stage of translation or transcription, respectively. MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 249

C'HIMIA 50 (lqq6l N,.. 6 (Jllnil are joined in an effort to evaluate alterna- tives to oligonucleotides, compounds of major groove major groove different structural types, with respect to their anti gene properties. Studies of the binding modes of such molecules might lead to new strategies for the sequence- A T G c specific binding of nucleic acids. The vast pool of natural products provides many small molecules that bind to nucleic acids, most of them of fungal or microbial origin. Unlike oligonucleotides, minor groove minor groove the majority of these small molecules interact with double-stranded nucleic acids Fig. 2. Substitution pattern of H-bonding donor and acceptor groups for adenine-thymine and via contacts within the minor groove. Two guanine-cytosine base pairs in the major and minor groove examples of such compounds are netropsin and distamycin (Fig. 3). These crescent- shaped polyamides composed of N-methylpyrrole amino acids bind to stretches of double-stranded DNA consisting of adenine/thymine base pairs. The sequence selectivity derives from H-bonding between the amide protons and the nitrogen and carbonyl oxygen of adenine and thymine on the floor of the minor groove. It was shown in NMR studies by Wemmer and coworkers that in the complex two distamycin molecules are aligned side by side in an anti parallel manner [3a]. Based on these findings, Dervan and coworkers Netropsin Distamycin Im-Py-Py"y-Py-Py-Py successfully altered the sequence specificity of these molecules to include guanine/ Fig. 3. Structures of the minor groove binding drugs netropsin and distamycin. By substituting N- cytosine base pairs by substitutingN-meth- methylimidazole for N-methylpyrrole carboxamides at selected positions of and covalently linking ylimidazole for N-methylpyrrole carbox- two individual molecules to a dimer Dervan and coworkers developed distamycin analogs, e.g .. Im- Py-Py-y-Py-Py-Py, with altered sequence specificity and higher affinity [3). amides at selected positions [3]. The affinity could be improved to subnanomolar levels by covalently linking the two indi- charged DNA and thus helps aligning the chromophore part then leads to double- vidual halfs of these dimers [3e] (Fig. 3). other parts of the molecu Ie. Alternati vel y, strand cleavage. In 1985, Nishida and co- certain DNA-ligands form dimeric (or workers reported the structure of the neo- higher order) complexes upon binding of carcinostatin chromophore (Fig. 4) which 2. DNA Ligands a metal and only these can then interact is responsible for the observed biological with their target. However, these mecha- activity [6b]. This chromophore consisted Quite a few natural products have the nisms alone cannot account for the se- of a highly unusual bicyclo[7.3.0]dode- capability not just to form complexes with quence specificity displayed by the re- cadienediyne scaffold. Just two years lat- DNA but to chemically modify it. Many spective DNA-binding drugs. These as er, the discovery of two compounds with compounds with antitumor activity be- well as compounds without an aromatic potent cytostatic activity and a similar long into this category. Figs. 4 and 5 show moiety or an apparent metal binding site core structure, calicheamicin YI' and espe- some examples of such naturally occur- often contain additional carbohydrate side ramicin A I (Fig. 4), has been reported ring compounds. Their mechanism of com- chains which have to be responsible for simultaneously by researchers at Lederle plexation with nucleic acids can be distin- site selectivity. Carbohydrates might there- Laboratories [7] and Bristol-Myers [8]. guished into four main categories, interca- fore be an interesting class of potential Their unusual structure containing a bicy- lation, binding via metal complexes, and DNA ligands [5]. In recent years, several clic enediyne and their impressive anti- binding mediated by a protein component research groups have tried to develop syn- cancer activity intrigued the chemical and or carbohydrate side chains. Intercalation. thetic DNA ligands based on the saccha- pharmaceutical community. Immediate- consists of aligning an extended aromatic ride units of known DNA binders. ly, a race aimed at the total synthesis of chromophore in between two base pairs of Certainly the most extensively studied calicheamicin set on, culminating in two the double helix (Fig. 6). Driving force of group of DNA binding small molecules in reported synthesis by the groups of Nico- this kind of interaction is n-stacking that recent years are the enediyne antibiotics, laou [9] and Danishefsky [10]. In 1989, a extends over both strands and leads to an among them calicheamicin, esperamicin, fourth structure ofthis type, dynemicin A, overall stabilization of the double helix. and neocarcinostatin. Neocarcinostatin, a has been reported, again by researchers at Many naturally occurring DNA ligands I:] complex of a protein and a chromo- Bristol-Myers [II], followed by the dis- contain metal binding subunits. By com- phore compound has first been isolated covery of the kedarcidin [12] and C-I 027 plexing a metal, a net positive charge 1965 [6a]. The protein is thought to medi- [13] chromophores (Fig. 4) adding to the results which provides an attractive elec- ate the sequence-specific recognition of arsenal of enediyne antibiotics. Of these trostatic force towards the negatively double-stranded DNA, activation of the latter compounds, the synthesis of dyne- MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 250

C1HMIA 50 (144h) I\r. h (luni)

micin A has been reported by Myers and coworkers [14]. The mode of action of these enediyne antibiotics follows a similar pathway as shown in Scheme J for calicheamycin. An external nucleophile attacks the trisulfide moiety remote of the enediyne core. The Neocarzinoslalin Chromophore thiolate generated in this way undergoes an internal Michael addition leading to a conformational change of the chromophore. This brings the two acetylene units in close proximity and, in a spontaneous pericyclic process termed Bergman cyclization [15], a I ,4-aryl diradical is formed. This diradical then induces oxidative DNA cleavage by H-abstraction from 2-deoxy- ribose residues [161. For most of the enediyne antibiotics Esperamicin A1 Dynemicin A mentioned above the recognition of double-stranded DNA relies on intercalation by a heterocyclic aromatic moiety. except for calicheamycin.lnterestingly, it is cali- H,C°I("'(°Y cheamicin which displays the highest se- l""NAO quence specificity in the recognition of ° double-stranded DNA.

¢~%::~ OH ° ° 2.1. Calicheamicil1 ° ~ U Calicheamicin y,l targets oligopyrimi- CI dine-oligopurine tracts of DNA with a preference for the sequence d(TCCT) Kedarcidin Chromophore C-1027 Chromophore [17a]; others include d(TCTC) and Fig. 4. Structures of the recently discovered enediyne class of antitumor compounds. These com- d(TTTT) [17b]. Additionally, it could be pounds are specific cutters of double-stranded DNA. shown in competition experiments that calicheamicin binds in the minor groove [17a]. The presence of known minor groove binders, such as netropsin which binds to similar target sequences, prevents or greatly reduces the cleavage of double- stranded DNA by calicheamicin. Later, it became clear that most of the DNA recognition was due to the aryl-tetrasaccharide domain. Kahne and coworkers showed in NMR studies that the carbohydrate portion adopts a highly organized, extended conformation in solution [18]. They could Bleomycin A2 Hedamycin further show that the oligosaccharide is organized into a shape that complements the shape of the minor groove by the presence of the unusual hydroxylamine glycosidic linkage connecting rings A and B (Fig. 4). In a series of experiments Danishef- SkY, Crothers and coworkers could demonstrate that the oligosaccharide domain was indeed crucial for selective and tight binding to DNA [19]. In these experiments the cleavage properties of calicheamicinone, the aglycone of calicheamicin, were investigated. Calicheamicinone which cannot be obtained by degradation of the natural product, had been obtained in racemic or either enantiomeric form by Chromomycin A3 Aclarubicin total synthesis in the groups of Danishef- Fig. 5. Further examples of DNA-binding antibiotics with carbohydrate side chains sky [l9b][20] and Nicolaou [9cl. Dall- MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 251

CHIMIA 50 (19%) NI'. b (JlIn;) ishefsky and Crothers found that calicheamicinone cleaves DNA with no apparent sequence selectivity and it primari- ly cleaves only one of the two target strands. These results suggested that the aryltetra- saccharide portion was indeed responsible for sequence selectivity. The hypothesis of Danishef.<;ky is further supported by the fact that esperamicin, which contains a similar carbohydrate frag- ment consisting of rings A, B, and E, does not display the same degree of sequence selectivity. Additionally, Kahne and coworkers investigated the cleavage selectivity of a calicheamicin analog termed calicheamicin Tcontaining only the sugar residues A and E [17b]. This derivative exhibits only minimal sequence selectivity implying that the full terasaccharide is required for specific DNA recognition. Several models have been put forth to rationalize the binding of calicheamicin to DNA [21]. It has been shown that a con- Fig. 6. St/"llcrure (!( 1I formational change of DNA is induced syn thetic 1mthral'yelin upon binding of calicheamicin. Specifi- wlfihiotic cOlllplexed with douhle-stranded cally, the oligopyrimidine regions of DNA DNA [4]. The aromat- display a certain flexibility and thus allow ic chromophore inter- for a distortion of the sugar-phosphate calates between two backbone to accomodate the relatively rigid consecutive base pairs. drug in the minor groove. The iodine of the aryl moiety is of particular importance to the binding event [2] b]. It is thought to Scheme I. Mechanism of the Activation and DNA-Damage Displayed by Calicheamicin r/ interact with the exocyclic amino group of o a guanine residue. On substituting the iodine the binding affinity is reduced in the 1) nucleophilic attack order I > Br > Cl > F > CH3 > H. More ------)0- HO structural information on the calicheam- 2) Michael addition icin-DNA complex was obtained by several NMR investigations [22]. Recently, OR Patel and coworkers published a detailed solution structure of calicheamicin YI' Bergman bound to a hairpin duplex of the sequence cyclization d(CACTCCTGGIIITICCAGGAGTG) by a combined NMRlmolecular dynamics approach [22d]. The complementary fit of the carbohydrate portion and the DNA DNA minor groove binding site d(TCCT) . d- (AGGA) creates numerous van der Waals contacts as well as H-bonding interactions. The iodine and the sulfur atoms -r-~ are engaged in H-bonds with the exposed DNA fragments proton of the amino groups of the 5'- and 3'- guanines, respectively. The sequence- specific carbohydrate binding orients site are reminiscent of a B-form duplex ies of the natural product indicated that the the enediyne aglycon such that it is while a widening of the minor groove is carbohydrate could not be obtained intact tilted relative to the helix axis and spans observed at the aglycon site. in this manner [23]. Nicolaou et al. were the minor groove placing the proradical Based on the results of these structural the first to accomplish this synthesis [24al centers adjacent to the cleavage sites. Spe- studies the Nicolaou and Danishefsky [9b]; to date, two other syntheses have cific localized conformational perturba- groups began to evaluate the binding ca- been reported [24b, c]. Scheme 2 shows as tions in the DNA have been identified pabilities of the saccharide moiety of cali- an example the synthesis devised by Kohne from imino proton chemical shift differ- cheamicin. This was possible only after and coworkers. It was then demonstrated ences and changes in sugar pucker pat- the chemical synthesis of the correspond- that the methyl glycoside 9 binds to the terns on complex formation. The helical ing methyl glycoside 9 (Scheme 2) had same sites in a similar way as the parent parameters for the carbohydrate binding been accomplished sinee degradation stud- natural product [21 b, h][25] (Scheme 3). MEDIZINISCHE CHEMIE • MEDICINAL CHEMISTRY 252

CHIMIA 50 (19%) Nr. 6 (lulli)

Schreiber, Danishefsky, and Crabtree reporter gene under the control of the thesis starts with the known trichloro- subsequently went on to demonstrate the transcription factor NF AT. Thus, the pro- acetimidate intermediate 10 of the total utility of the calicheamycin aryl-tetrasac- liferation ofT cells was inhibited in a dose synthesis of calicheamycin YI'. A diethyl- charide 9 as DNA ligand [26]. Calichea- dependent manner. These observations eneglycol unit is introduced at the ano- micin methyl glycoside 9 inhibits the mark the beginning of a novel strategy for meric position of the A ring and then formation of DNA-protein complexes at the development of antigene agents based coupled with a second tetrasaccharide unit micromolar concentrations in a sequence- on carbohydrate-mediated recognition of 10 leading to the dimer 11 shown in Scheme specific manner and rapidly dissociates DNA. 4. This head-to-head dimer was shown to preformed complexes. Similar results were Nicolaou et al. later reported on the bind to its target sequence d(TCCT-XX- obtained by the same authors in an in vivo synthesis and the DNA-binding proper- AGGA) with a IO-fold higher affinity as study. Calicheamicin methyl glycoside 9 ties of a covalently linked dimer of the compared to the monomer. The authors effectively blocked the expression of a calichearnicin carbohydrate [27]. The syn- could further demonstrate that dimer 11 is

Scheme 2. Synthesis of the Calicheamicin Aryl-Tetrasaccharide 9 Devised by Kahne and Coworkers [24c]. Two earlier syntheses were described by Nicolaou et al. and by Danishefsky and coworkers [24a, b].

H3\,P~H3 1) TsOH. MeOH. rt Tf20, Et20 H3C'1 0 -78°C---Ht O. OCH3 2) BzCI, py, -50"C + o 65% 3) Tf20, py. H C C0J;:JFNI 0 CH CI .O"C 3 2 2 H3CO 80% 2 (7 steps) 3

H3C KHMDS BZS~O, DMF, -1Q°C + ~a_~_caac2H5

a'SDMS 81%

4 5 (12 steps)

1)TBAF. THF. rt 1) n-Bu3P. THF. rt + 2) MeOH. EtOH 2)DMAP, THF, rl NaOH, rt 79% 52% 2

7 8 (10 sleps)

CH 0 H3C3 ,«9" SOH, H3C ~ I ~a-~~OCH3 a aCH3 aH a H3C7-0-1 OCH3 r-~) HO~ H3C ~~

H.CO OH H3CO 9

Scheme 3. Binding Site and Orientation of the Calicheamicin r/ Aryl-Tetrasaccharide 9 in the Minor Groove of the Target DNA

Hy~OCH, H•••. y .....o, o HJy~O ••.NHO'" o~

'C s/"(.. K ,«CH. 0 H.C d(T-C-C-T) OH H"cH2CH~ OCH~ 9" SOH, H.C (A-G-G-A)d I f "lOCH, ./";r. . ~O_~~O\ OCH. ~ I ~ - / p(' o OCH, OH IO~ ° OC/H, C...... /'" ~---> .... /'" H,C-r--O-"! aCH, r--~j ° C ·G ... '''''''' HO~ H,CH.cHN~ HoC~Y.L<;:OH .••.•.•.•.•..... /"" HO /""'" •••••••••••••G/'" H,CO OH H,ca >r...... / 5' .•••..•.•.•.•.••.•.•...•.••..~

9 3'" MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 253

CHIMIA 50 (1996) Nr. 6 (lun;) a more effective inhibitor of polymerase II Scheme 4. Synthesis and Binding Site of the Ca/ichearnicin Saccharide Virner 11 Described by in an in vitro transcription system. The Nicolaou and Coworkers [27] effective concentrations of the oligosaccharide for inhibition of transcription factor binding and transcriptional inhibition are in the low micromolar range. It is noteworthy to realize, though, that at higher concentrations additional sequences such as d(TCTC), d(TTTT), and (TGGT) are recognized as well. The oligosaccha- 15 steps rides derived from calicheamicin target certain preferred DNA sequences but they lack the kind of specificity that can be observed with proteins or oligonucleotide analogs.

2.2. Chromomycin A binding mode completely different from the one displayed by calicheamicin was unveiled for the antitumor compound n d(T-C-C-T-X-X-A-G-G-A) chromomycin A3' Chromomycin A3(Fig. .u. (A-G-G-A- Y-Y-T -C-C- T)d 5) is a member of the aureolic acid family of antibiotics and was first isolated in 1960 [28]. However, the structure was fully established only in 1982 [29].It was shown ?~ that chromomycin binds to DNA prefera- ~~co~ 0 I bly at GC-rich sites and that, upon bind- o _ HO>-../S CH, ing, replication and transcription are in- o ~~~o 'ft ~'0. ~'I..o 0 Clio ,/'A ""''''''"."" hibited [30]. In 1989, Patel and coworkers ~~'H G"" )-T o 0 ~ /"" •••••••••••••• ./ showed that chromomycin binds in the ) G".. ''''''''C rO ./' "'"'''' ./ minor groove of DNA as an octahedral 2: 1 \ ./' """""·"c./' u,cy:?,o //li." " """ / drug-Mg2+ complex (Fig. 7) [31]. In the H: 'Y '0, X".... .>T complex, the drug acts as a bidentate li- Hoy..:>'O '1.0' J,/' """''''''''./' o ."'(. EX""""" )-Y gand. The metal is located in the middle of ~ OH ~Q1zCHN~ •••••••••••••• /' I r ~ OCH, /T"".". /Y ~ the minor groove at the central d(GC) step o - OOH, COo, """'''''''';;'A and is coordinated by the carbonyl group _01 / """"'''''''''''''''G/ and the adjacent phenolate anion of the u,c~HC.""""."". /' HO '-!:'" '''''''''''G/ chromophore. The planes of the chromo- 5, .•••T "'"'' ./' "''''''''''''''K phores span an angle of75° and are slight- 3'- ly tilted relative to each other. The extensively alkylated and acylated trisaccharide side chains adopt an extended confor- show that the aglycone, chromomycinone, connected via a-I,4-linkages. The sac- mation and are located along the minor itself preferably forms a I: I complex in charide side chains are in most cases re- groove. The sequence specificity of the the presence of magnesium and not a 2: I quired for optimal biological activity. chromomycin dimer for the d(GGCC) . complex [32], With a series of degradation Single crystal X-ray and NMR studies d(GGCC) binding site is associated with products of chromomycine A3they showed have shown that the saccharide side chains the formation of H-bonds between the that the trisaccharide side chain helps sta- are located in an extended conformation exocyclic amino groups of the guanine bilizing the 2: 1complex with magnesium. within the minor groove [34]. Upon drug bases and oxygens within the aromatic binding, preferably at GC-rich sequences, and carbohydrate moieties of the drug. 2.3. Anthracycline Carbohydrate Side the minor groove is widened and becomes Additional intermolecular interactions Chains shallow. In some cases an additional kink- consist of van der Waals contacts between The anthracyclines are another impor- ing of the double helix at the intercalation aliphatic protons of the trisaccharide resi- tant class of antitumor antibiotics. Some site is observed which leads the saccharide dues and the surface of the minor groove. of these compounds, e.g., daunorubicin tails protruding partly into the solvent The minor groove at the binding site is and doxorubicin are currently used in regIOn. widened and shallow and the individual chemotherapy. They exert their biological Based on the assumption that the an- nucleosides within this central tetranucle- activity by intercalation of a tetracyclic thracyclines have evolved with carbohy- otide segment share A-type conforma- aromatic core between base pairs of dou- drate side chains as a minor groove bind- tional parameters. ble-helical DNA (Fig. 6), thereby inhibit- ing motif, Harding and coworkers synthe- The sugar residues in chromomycin A3 ing access of transcriptional factors and sized the methyl glycoside 18 related to do not just play a role in the sequence topoisomerases [4][33]. Many of these the aclarubicin (aclacinomycin A) carbo- specificity towards DNA they also influ- compounds contain carbohydrate side hydrate moiety (Scheme 4) [35]. Prelimi- ence the stoichiometry of the complex chains consisting of between one and three nary binding studies were carried out by with Mg2+. Kahne and coworkers could 2,6-dideoxY-L-pyranoseresidues which are temperature-dependent UV melting curves MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 254

CHIMIA 50(19%) Nr. 6 (lUll;) with calf thymus DNA. Addition of the cin B and related aminoglycosides bind to tures. RRE and TAR are of major impor- aclacinomycin trisaccharide 18 to the DNA the A site of the 16S ribosomal RNA tance in the viral replication cycle [4 11 and solution resulted in only minor variations which comprises the decoding function of therefore attractive targets for therapeutic of the melting temperature of the DNA « the ribosomes [37]. This affects the bind- intervention [42]. The aminoglycoside ] 0). This indicates that the binding affinity ing of aminoacylated transfer-RNA and binding sites are identical to the protein of the carbohydrate is weak. leads to inhibition of translocation of the binding sites as has been shown by chem- messenger-RNA and consequently to mis- ical footprinting and competition experi- coding. Aminoglycosides also inhibit the ments and share a set of unusual structural 3. RNA Ligands splicing of several group I introns [38]. features. RRE-RNA contains at this site Group I introns are a class of functional consecutive G-A and G-G base pairs, While quite a few classes of low-mo- RNA sequences that catalyze their own TAR-RNA a bulge consisting of three lecular weight compounds that bind to excision from a longer RNA strand as well pyrimidine nucleotides. double-stranded DNA have been charac- as the religation of the remaining parent From the studies on the binding prop- terized, rather little is known about corre- strand (exon) in a well characterized two- erties of aminoglycosides it seems evident sponding RNA ligands. One notable ex- step process [39]. They share a common that there is not a particular base sequence ception, though, are the aminoglycoside structural core with a binding site for the which is recognized by the structurally antibiotics. cofactor guanosine. Von Ahsen et al. have diverse class of aminoglycosides but rath- Aminoglycoside antibiotics are basic shown that aminoglycosides are competi- er a common structural motif. Experi- pseudo-oligosaccharides (Fig. 8) which tive inhibitors of cofactor binding and ments with combinatorial RNA-libraries are produced by microorganisms and have further disrupt structural contacts in the that were selected for the aminoglyco- found wide-spread use in therapy and as catalytically active core. sides neornycin, tobramycin, lividomy- tools in molecular biology [36]. It has The steady spread of illnesses caused cin, or kanamycin additionally support been known for quite some time that their by retroviruses, most prominently HIV, this notion [43]. However, no high-resolu- potency stems from inhibition of protein has lead to an intensified search for antivi- tion structural data is available yet for such synthesis. Streptomycin, for instance, in- ral compounds that selectively bind to a binding site or a complex between an terferes with translation by inducing mis- RNA. Recently, two groups independent- aminoglycoside and its RNA-target. reading of the genetic code and by inhibit- ly reported on the sequence-selective rec- It should further be noted that not only ing translational initiation. The miscoding ognition of certain substructures within the structurally complex aminoglycoside is believed to be the result of interference the HIV RNA-genome by aminoglyco- antibiotics mentioned above display high with a proof-reading step. Moazed and side antibiotics [40]. These substructures, affinity binding to RNA. Even chitosans- Noler showed in footprinting experiments Rev-response element (RRE) and trans- f3-1 ,4-linked glucosamine oligomers - act that streptomycin binds to a single site activation responsive region (TAR), are as ligands for double-stranded RNA and within] 6S ribosomal RNA in the vicinity binding sites for the viral regulatory pro- effect a dramatic thermal stabilization of of a pseudoknot structure [37a]. Neomy- teins Rev and Tat, rich in stem-loop struc- these complexes. Double-stranded DNA is complexed as well but undergoes a conformational change from B-form to A- form DNA [44].

4. Conclusions

There exist a large number of nucleic acid-binding natural products which contain carbohydrate components. Although the role of these sugars, for the most part, is still unclear, it was shown in a few cases that the saccharide portion itself mediates the sequence-specific recognition. A com- parison of the examples mentioned above, reveals similarities in the structure of the respective DNA- and RNA-binding gly- coconjugates that may prove useful in the design of novel antisense agents. DNA-binders, like for instance calicheamicin and chromomycin, have oligosaccharide side chains that are rather hydrophobic. There is rarely more than a single positively charged ammonium group present in the carbohydrate and Fig. 7. Structure of many of the sugars are of the 2,6-deoxy- the 2: I: 1 chromo· hexopyranose type. Often, the remaining mycin A,rma[?nesi- um-DNA complex hydroxy groups are further alkylated or described by Patel acylated. Thus, the recognition process and coworkers [31] relies onjust a few H-bonding interactions MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 255

CHIMIA 50(19%) Nr. 6 (Juni) and is mostly driven by the hydrophobic- ular biology and as therapeutics in medi- C. Helene, Angew. Chem. 1993,105,697; ity of the DNA-binding saccharides [45]. cine. Carbohydrates are a recent addition d) R.S. Varma, Synlett 1993, 621; e) J.F. This suggests that rather the complemen- to the arsenal of such compounds. Their Milligan, M.D. Matteucci, J.e. Martin, J. Med. Chern. 1993,36, 1923; £)J. Hunziker, tary shapes, and hence specific van der capabilities in the recognition of nucleic C. Leumann, in Modern Synthetic Meth- Waals contacts between the carbohydrates acids and the underlying mechanisms are ods, 'Nucleic Acid Analogues: Synthesis and the minor groove, may determine the just now beginning to be explored. Under- and Properties', Eds. B. Ernst and C. Leu- site-selectivity displayed by many carbo- standing these factors is a prerequisite for mann, Verlag Helvetica Chimica Acta, hydrate-containing minor groove binders. the development of new strategies in the Basel, 1995, Vo\. 7, p. 331. Additionally, the relatively rigid carbohy- design of nucleic acid binding agents and (2) T. Gura, Science 1995, 270, 575. drates may take advantage of the sequence- their pharamaceutical application. [3] a) J.G. Pelton, D.E. Wemmer, Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 5723; b) M. dependent flexibility of double-stranded Received: March 26, 1996 Mrksich, P.B. Dervan, J. Am. Chem. Soc. DNA to bind their target sites [5]. Further- 1994,1 16,3663;c) Y.-H.Chen,J.W. Lawn, more, the hydrophobicity of DNA-bind- [I) a) E. Uhlmann, A. Peyman, Chern. Rev. ibid. 1994, 116, 6995; d) B.H. Geierstan- ing carbohydrates may differentiate these 1990,90,543; b) J. Engels, Nachr. Chern. ger, M. Mrksich, P.B. Dervan, D.E. Wem- from the hydrophilic and often negatively Tech. Lab. 1991,39, 1250; c)N.T. Thuong, mer, Science 1994, 266, 646; e) J. Cho, charged cell-surface saccharides which are involved in protein recognition and may enhance the cellular uptake of the Scheme 5. Synthesis of the Aclacinomycin A Trisaccharide 18 by Harding and Coworkers [35] respective drugs. The known RNA-binding carbohy- OCH3 drates, on the other hand, display a com- CI CH H3C-f-O-i pletely different chemical and physical H3C-r--O--t + H3C-r--OJ 3 ~CF3 behavior. They are extremely hydrophilic o ~ ~CF3 62% due to extensive substitution of hydroxy AcO HO H3C-r--O--t function by ammonium groups. Deoxy- ~ AcO genation or O-alkylation is not common 12 13 (4 sleps) 14 for these aminoglycosides. They share, though, a similar rigidity as the DNA- binding saccharides since they mainly OCH3 consist of pyranose subunits and a cy- H3C-r-O-i 1) NIS, MaCN, O'C clohexitol part. To date, there is no de- 2) Pd-C, H2' Et3N 1) NaOMa, MeOH, rt ~F. EtOH,rt tailed structural information available on o the 2) n-Bu2SnO, C6H6 + 3) NaOMa, MaOH, rt complexes of aminoglycosides with n·SU4N1, SnSr H3C-1'-O-i RNA. One can assume that enthalpy due reflux ~ 55% to charge neutralization is a major factor in 63% HO the interaction with RNA, in contrast to 15 16 the recognition process ofthe DNA minor groove binders where entropic factors play CH CH a key role. These attractive forces can H3 -r--OJ 3 H C-r--OJ 3 induce dramatic conformational changes C 3 ~CF3 as is seen for the transistion of B- to A- o~ form DNA upon binding of aminoglyco- H3C-r--OJ 3 sides. However, the mechanisms by which r H C-r--O-i ~ 2) Pd·C, H2' EtOAc ~ aminoglycosides achieve sequence spe- o rt o cificity in the recognition of RNA remains 3) aq. NaOH, rt C unclear and awaites further investigation. H3 pJ H3cpJ 60% o Ligands that sequence-selectively bind HO to nucleic acids have a number of impor- 17 18 tant uses as tools in chemistry and moIec-

NH2

HN=\ HO~NH2 0 NH o-r---i-r-OH NH HOf.t-1-.£-NHyNH2 2 H2N 6-r--f.-t - H3C~ OH II OHC NH HO~NH. o OOH l~~ o H HO N_ HO~OH ~ CH3 HO H.N

Streptomycin Neomycin B Tobramycin

Fig. 8. Structures of some representative examples of arninoglycoside antibiotics MEDIZINISCHE CHEMIE . MEDICINAL CHEMISTRY 256

CHIMIA 50 (1996) Nr. 6 (Juni)

M.E. Parks, P.B. Dervan, Proc. Natl. Acad. dou, B. Meunier, ibid. 1995,107,819; c) C. Liu, B.M. Smith, K. Ajito, H. Komatsu, Sci. U.S.A. 1995, 92, 10389; f) P.S. Zurer, SJ.Danishefsky,M.D.Shair,J. Org. Chern. L. Gomez-Paloma, T. Li, E.A. Theodor- Chern. Eng. News 1995, 74 (3), 18. 1996,61, 16. akis, K.C. Nicolaou, P.K. Vogt, Proc. Natl. [4J Y.-G. Gao, Y.-c. Liaw, Y.-K. Li, G.A. van [17J a) N. Zein, A.M. Sinha, WJ. McGahren, A cad. Sci. U.S.A. 1996,93,940. der Marel, 1.H. van Boom, A.H.-1. Wang, G.A. Ellestad, Science 1988, 240, 1198; b) [28] M. Shibata, K. Tanabe, Y. Hamada, K. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, S. Walker, R. Landovitz, W.D. Ding, G.A. Nakazawa, A. Miyabe, H. Hitomi, M. Mi- 4845. Ellestad, D. Kahne, Proc. Natl. Acad. Sci. yamoto, K. Mizuno, J. Antibiot., Ser. B [5] D. Kahne, Chemistry & Biology 1995, 2, 7. U.s.A. 1992,89,4608. 1960,13, I. [6J a) N. Ishida, K. Miyazaki, K. Kumagai, M. [18] a) S. Walker, K.G. Valentine, D. Kahne, J. [29] R. Riccio, K. Nakanishi, J. Org. Chen!. Rikimaru, J. Antibiot. 1965,18,68; b) K. Am. Chern. Soc. 1990, 112, 6428; b) S. 1982,47,4589. Edo, M. Mizugaki, Y. Koide, H. Seto, K. Walker, D. Yang, D. Kahne, D. Gange, [30] G. Wakisaka, H. Uchino, T. Nakamura, H. Furihata, N. Otake, N. Ishida, Tetrahedron ibid. 1991, J] 3, 4716; c) S. Walker, D. Sotobayashi, S. Shirakawa, A. Adachi, M. Lett. 1985, 26, 331. Gange, V. Gupta, D. Kahne, ibid. 1994, Sakurai, Nature (London) 1963, 198, 385. [7] a) M.D. Lee, T.S. Dunne, M.M. Siegel, 116,3197. [31] a) X. Gao, DJ. Patel, Biochemistry 1989, c.c. Chang, G.O. Morton, D.B. Borders,J. [19] a) 1. Drak, N. Iwasawa, D.M. Crothers, SJ. 28,751; b) X. Gao, DJ. Patel, ibid. 1990, Am. Chern. Soc. 1987,109,3464; b) M.D. Danishefsky, Proc. Natl. Acad. Sci. U.S.A. 29, 10940; c) X. Gao, P. Mirau, DJ. Patel, Lee, T.S. Dunne, C.C. Chang, G.A. Elles- 1991, 88, 7464; b) 1. Aiyar, SA Hitch- J. Mol. BioI. 1992, 223, 259; d) M.A. Ke- tad, M.M. Siegel, G.O. Morton, W.J. Mc- cock, D. Denhart, K.K.C. Liu, S.J. Dan- niry, D.L. Banville, P.M. Simmonds, R. Gahren, D.B. Borders, ibid. 1987, 109, ishefsky, D.M. Crothers, Angew. Chern. Shafer, ibid. 1993,231,753. 3466; c) M.D. Lee, J.K. Manning, D.R. 1994,106,925. [32] a) D.J. Silva, R. Goodnow, D. Kahne, Bio- Williams, N. Kuck, R.T. Testa, D.B. Bor- [20] a) 1.N. Haseltine, M.P. Cabal, N.B. Mantlo, chemistryI993,32, 463; b)D.J. Silva,D. E. ders, J. Antibiot. 1989,42, 1070; d) W.M. N. Iwasawa, D.S. Yamashita, R.S. Cole- Kahne, J. Am. Chern. Soc. 1993,115,7962. Maiese, M.P. Lechevalier, H.A. Lecheva- man, SJ. Danishefsky, G.K. Shulte,1. Am. [33] 1.W. Lown, Chern. Soc. Revs. 1993, 165. lier, J. Korshalla, N.A. Kuck, A. Fantini, Chem. Soc. 1991,113,3850; b) V.P. Roc- [34] a) A.H.-1. Wang, Curro Opin. Struct. BioI. M.J. Wildey, J. Thomas, M. Greenstein, co, S.J. Danishefsky, G. Shulte, Tetrahe- 1992,2,361; b) G.A. Leonard, T.W. Ham- ibid. 1989,42,558. dron Lett. 1991, 32, 6671. bley, K. McAuley-Hecht, T. Brown, W.N. [8) a) 1. Golik,J. Clardy, G. Dubay, G. Groene- [21] a) R.c. Hawley, L.L. Kiessling, S.L. Sch- Hunter, Acta Crystallogr., SeCT. D 1993, wold, H. Kawaguchi, M. Konishi, B. Krish- reiber, Proc. Natl. Acad. Sci U.S.A. 1989, 49,458; c) H. Zhang, Y.-G. Gao, G.A. van nan, H. Ohkuma, K. Saitoh, T.W. Doyle,J. 86, 1105; b) K.c. Nicolaou, S.-C. Tsay, T. der Marel, 1.H. van Boom, A.H.-1. Wang, Am. Chern. Soc. 1987, 109, 3461; b) J. Suzuki, G.F. Joyce, 1. Am. Chern. Soc. J. BioI. Chern. 1993, 268, 10095; d) D. Golik, G. Dubay, G. Groenewold, H. Ka- 1992,114,7555; c) M. Uesugi, Y. Sugiura, Yang, A.H.-J. Wang, BiochemisTry 1994, waguchi, M. Konishi, B. Krishnan, H. Oh- Biochemistry 1993, 32, 4622; d) G. Krishna- 33,6595. kuma, K. Saitoh, T.W. Doyle, ibid. 1987, murthy, W. Ding, L. O'Brien, G.A. Elles- [35] CJ. Shelton, M.M. Harding, J. Chem. Res. 109,3462. tad, Tetrahedron 1994, 50, 1341; e) S.C. (S) 1995, 158. [9] a) K.C. Nicolaou, Angew. Chem. 1993, Mah, M.A. Price, C.A. Townsend, T.D. [36] K.L. Rinehart, T. Suami, Eds., 'Aminocy- 105, 1462; b) R.D. Groneberg, T. Miyaza- Tullius, ibid. 1994, 50, 1361; f) D.R. Lan- clitol Antibiotics' ,American Chemical So- ki, N.A. Stylianides, T.J. Schulze, W. Stahl, gley, T.W. Doyle, D.L. Beveridge, ibid. ciety, Washington D.C., 1980. E.P. Schreiner, T. Suzuki, Y. lwabuchi, 1994,50, 1379; g) D.R. Langley, 1. Golik, [37] a) D. Moazed, H.F. Noller, Nature (Lon- A.L. Smith, K.C. Nicolaou, J. Am. Chern. B. Krishnan, T.W. Doyle, D.L. Beveridge, don) 1987,327, 389; b) D. Moazed, H.F. Soc. 1993,115,7593; c) A.L. Smith, E.N. J. Am. Chem. Soc. 1994,116, 15; h) T. Li, Noller, Biochimie 1987,69,879. Pitsinos, C.-K. Hwang, Y. Mizuno, H. Sai m- Z. Zeng, V.A. Estevez, K.U. Baldenius, [38] a) U. von Ahsen, R. Schroeder, Nucleic oto, G.R. Scarlato, T. Suzuki, K.c. Nico- K.c. Nicolaou, G.F. Joyce, ibid. 1994, 116, Acids Res. 1991,19,2261; b) U. von Ah- laou, ibid. 1993,115,7612; d) K.C. Nico- 3709. sen, J. Davies, R. Schroeder, NaTure (Lon- laou, C.W. Hummel, M. Nakada, K. [22] a) S. Walker, J. Murnik, D. Kahne, J. Am. don) 1991, 353, 368; c) U. von Ahsen, 1. Shibayama, E.N. Pitsinos, H. Saimoto, Y. Chem. Soc. 1993, 115,7954; b) L. Gomez- Davies, R. Schroeder, 1. Mol. Bioi. 1992, Mizuno, K.-U. Baldenius, A.L. Smith, ibid. Paloma, J.A. Smith, W.J. Chazin, K.C. 226, 935; d) U. von Ah~en, H.F. Noller, 1993, 115, 7625. Nicolaou, ibid. 1994, 116, 3697; c) S.L. Science 1993, 260, 1500; e) J. Rogers, 1. [10] SA Hitchcock, S.H. Boyer, M.Y. Chu- Walker, A.H. Andreotti, D. Kahne, Tetra- Davies, Nucleic Acids Res. 1994,22,4983. Moyer, S.H. Olson, SJ. Danishevsky, An- hedron 1994, 50, 1351; d) N. Ikemoto, [39] T.R. Cech, Annu. Rev. Biochem. 1990,59, gew. Chem. 1994, 106, 928. R.A. Kumar, T.-T. Ling,GA Ellestad, S.J. 543. [11 J a) M. Konishi, H. Ohkuma, K. Matsumoto, Danishefsky, DJ. Patel, Proc. Natl. Acad. [40] a) M.L. Zapp, S. Stem, M.R. Green, Cell T. Tsuno, H. Kamei, T. Miyaki, T. Oki, H. Sci. U.S.A. 1995, 92, 10506. 1993,74,969; b) H.-Y. Mei, A.A. Galan, Kawaguchi, G.D. VanDuyne, J. Clardy, J. [23] a)J. Golik, H. Wong, B. Krishnan, D. Vyas, N.S. Halim, D.P. Mack, D.W. Moreland, Antibiot.1989,42, 1449; b) M. Konishi, H. T.W. Doyle, Tetrahedron Lett. 1991,32, K.B. Sanders, H.N. Truong, A.W. Czarnik, Ohkuma, T. Tsuno, T. Oki, G.D. Van- 1851; b) R.L. Halcomb, M.D. Wittman, Bioorg. Med. Chem. Lett. 1995,5, 2755. Duyne,1. Clardy, J. Am. Chem. Soc. 1990, S.H. Olson, S.J. Danishefsky, J. Galik, H. [41] a) M.J. Gait, J. Karn, Trends Biochem. Sci. 112,3715. Wong, D. Vyas, 1. Am. Chem. Soc. 1991, 1993,18,255; b) D.A. Mann, 1.Mikaelian, [12] J.E. Lee, D.R. Schroeder, D.R. Langley, 113,5080. R.W. Zemmel, S.M. Green, A.D. Lowe, T. K.L. Colson, S. Huang, S.E. Klahr, M.S. [24] a) K.C. Nicolaou, R.D. Groneberg, T. Mi- Kimura,M. Singh, P.J.G. Butler, MJ. Gait, Lee,J. Golik,J. Am. Chem. Soc. 1993, 115, yazaki, N.A. Stylianides. T.J. Schulze, W. 1. Karn, J. Mol. Bioi. 1994,241, 193. 8432. Stahl, J. Am. Chem. Soc. 1990,112,8193; [42] a) MJ. Gait, J. Karn, Trends Biotechnol. [ 13] Y. Minami, K. Yoshida, R. Azuma, M. b) R.L. Halcomb, S.H. Boyer, SJ. Dan- 1995, 13, 430; b) E. De Clercq, J. Med. Saeki, T. Otani, Tetrahedron Lett. 1993, ishefsky,Angew. Chem.1992, 104, 314; c) Chem. 1995,38, 2491. 34,2633. S.-H. Kim, D. Augeri, D. Yang, D. Kahne, [43] a) Y. Wang, R.R. Rando, Chernistry & [14] a) A.G. Myers, M.E. Fraley, N.J. Tom, J. J. Am. Chem. Soc. 1994, 116, 1766. Biology 1995, 2, 281; b) S.M. Lato, A.R. Am. Chem. Soc. 1994, 116,11556; b) A.G. [25] J. Aiyar, S.J. Danishefsky, D.M. Crothers, Boles, A.D. Ellington, ibid. 1995,2,291; c) Myers, M.E. Fraley, N.J. Tom, S.B. Cohen, J. Am. Chern. Soc. 1992,114,7552. M.G. Wallis, U. von Ahsen, R. Schroeder, D.J. Madar, ChemisTry & Biology 1995, 2, [26] S.N. Ho, S.H. Boyer, S.L. Schreiber, SJ. M. Farnulok, ibid. 1995,2,543. 33. Danishefsky, G.R. Crabtree, Proc. NaTI. [44] T. Motomura, Y. Aoyama, Bull. Chen!. [15] R.G. Bergman, Ace. Chern. Res. 1973, 6, A cad. Sci. U.S.A. 1994,91,9203. Soc. Jpn. 1992, 65, 1755. 25. [27] a) K.c. Nicolaou, K. Ajito, H. Komatsu, [45] W. Ding, G.A. Ellestad, J. Am. Chem. Soc. [16] a) K.C. Nicolaou, W.-M. Dai,Angew. Chem. B.M. Smith, T. Li, M.G. Egan, L. Gomez- 1991,113,6617. 1991, 103, 1453; b) G. Pratviel, 1. Berna- Paloma, Angew. Chem. 1995,107,614; b)