Current Medicinal Chemistry 1157 Backbone Modification of ...€¦ · This has resulted in the...

Current Medicinal Chemistry, 2001, 8, 1157-1179 1157

Backbone Modification of Nucleic Acids: Synthesis, Structure andTherapeutic Applications

Jason Micklefield*

Department of Chemistry, University of Manchester Institute of Science and Technology,Faraday Building, Sackville Street, PO Box 88, Manchester, M60 1QD, UK

Abstract: Nucleic acids have been extensively modified by replacing the phosphodiestergroup or the whole sugar–phosphodiester backbone with alternative anionic, neutral andcationic structures. Several of these modified oligonucleotides exhibit improvedproperties including enhanced recognition and binding to RNA, duplex DNA andproteins. This has resulted in the development of new and more potent antisense andantigene agents, as well as aptamers. Furthermore, backbone modified oligonucleotideshave also been used in the development of several alternative strategies, which rely onaltogether different mechanisms of action and show significant promise for therapeutic intervention. In thisreview the latest advances in the synthesis and evaluation of the most promising backbone modified oligos willbe discussed, with a view to their future as novel pharmaceuticals.

1. INTRODUCTION Whilst significant advances have been made in the use ofmodified oligos in the antisense and antigene area, equally asexciting are the more recent developments of modified oligoswith alternative and novel modes of action. For examplesynthetic hammerhead ribozymes, capable of sequencespecific cleavage of a target RNA, can stabilised in biologicalfluids by chemical modification [18]. Aptamers, shortstrands of DNA or RNA generated by in vitro selection, canbe modified to improve their stability and binding affinity fortherapeutically relevant protein targets [19]. Alternativelymodified oligo decoys can function as structural RNAmimetics and compete with viral RNA structural elements inbinding to viral regulatory proteins [20]. Furthermore,modified oligos that mimic the high energy distortedstructures of dsDNA when bound to restriction enzymes, canbe used as transition state analogues which effectivelyinhibit the binding of the native dsDNA substrates [21].Even more provocative is the development of short modifiedoligos that can function as suicide inhibitors of self–splicingRNA introns from fungal pathogens, resulting in dead–endribosomal RNAs [22]. All of these approaches couldpotentially lead to new therapeutic strategies and all havearisen through the systematic evaluation of the effects ofchemical modifications, principally in the backbone ofoligonucleotides, on the structure, conformational and bio–physical properties of the oligo.

The last decade has seen an upsurge of interest in thechemistry and biology of nucleic acid analogues. This hasmainly been due to the realisation that modifiedoligonucleotides, with improved properties, can be used inthe sequence specific control of gene expression and haveimmense potential as therapeutic agents [1–5]. The principlemode of action of these modified oligos is through thebinding, via Watson and Crick base pairing, to a specificmRNA sequence associated with a diseased state and thesubsequent inhibition the translational event leading to adetrimental protein (antisense) [1–4]. In addition,transcription can be inhibited by the binding of an oligo toduplex DNA, principally through Hoogsteen base pairingand the formation of a triple helix (antigene) [5–7]. Theappeal of these approaches, particularly antisense, is thetremendous scope of these agents for use in the treatment ofmany diseases which are characterised by the expression ofunwanted genetic material, where target selection is governedprimarily by gene sequence. Thus a rational approach to thetreatment of cancers, viral and bacterial infections, as well asother diseases can be envisaged. To these ends a plethora ofnucleic acid analogues have been designed, synthesised andevaluated, with a view to the improving the efficacy of theseagents by altering their bio–physical properties comparedwith native DNA and RNA [8–17]. Desirable propertiesinclude improved stability to nuclease enzymes and uptakeinto cells. Increased affinity, kinetics and base pairingspecificity upon binding to nucleic acid targets can alsoimprove the efficacy of these agents. For the antisenseapproach the ability to activate enzymes, such as RNases Hor L, which are capable of irreversibly destroying a specificmRNA, is also an important factor.

As more is learnt about the effects of specificmodifications on oligonucleotide structures, then the morelikely it is that further novel targets and mechanisms fortherapeutic intervention, will be discovered utilisingmodified oligos. In parallel the increased availability anddiversity of synthetic modified oligos will continue to see anincrease in their use as tools to investigate fundamentalquestions pertaining to the structure–function relationship ofnucleic acids [23–25]. Moreover, modified oligonucleotidesare increasingly being used to address questions concerningthe origins of life. Notably, novel chemistry is being used to

*Address correspondence to this author at the Department of Chemistry,University of Manchester Institute of Science and Technology, FaradayBuilding, Sackville Street, PO Box 88, Manchester, M60 1QD, UK

0929-8673/01 $28.00+.00 © 2001 Bentham Science Publishers Ltd.

Dr. Mansoor Alam

1158 Current Medicinal Chemistry, 2001, Vol. 8, No. 10 Jason Micklefield

mimic the native nucleic acids and their inherent propensityfor self–replication via template based mechanisms [26,27].Finally, modified oligonucleotides are also findingincreasing use as probes and other tools in molecularbiology, particularly for sequencing by hybridisation andmedical diagnostic [28,29].

exploit that fact that only as few as 5 base pairs of aDNA.RNA heteroduplex are required to bind and activateRNase H [30]. These chimeras, or so called “gapmers”,usually retain a central sequence of DNA, orphosphorothioate DNA (PS–oligo), flanked by alternativelymodified oligos which are capable of protecting the centralregion from exonuclease cleavage (Fig. 1).Methylphosphonate and 2´–Oalkyl modifications have beenemployed to good effect in terminal regions of gapmers[31,32].

Because of the accelerated efforts in this area of researchover the last decade, it is not possible to cover the manydifferent modifications that have been performed onoligonucleotides, nor all the potential application of thesemodified oligos. Therefore, this review will focus on thelatest advances in the development of backbone modifiednucleic acids which show greatest promise for therapeuticapplications. Whilst most of the research discussed has beenconducted with a view to improving the efficacy of antisenseand antigene agents, where possible important examples ofbackbone modified oligos having alternative modes of actionwill be included. In this respect the review will be selectiveand the reader is referred to several excellent earlier reviewswhich cover extensively the many other types of backbone[8–14], sugar ring [15, 16], or base [17] modifications tooligonucleotides that have been reported.

Contrary to early opinion, activation of RNase H is notessential for potent antisense and in some cases can bedetrimental, since transient complexes with partiallycomplementary, non–target RNA, sequences can be cleavedresulting in non–specific effects [33]. Indeed several types ofbackbone modified oligos, which are incapable of activatingRNase H, have been shown to be effective antisense agents[34,35]. Such oligos are thought to act by a steric blockingmechanism, inhibiting the ribosome, initiation factors orother proteins from binding to specific sites on a RNAtranscript. It seems likely that strong binding to the targetmRNA is essential for steric blocking oligos, and that theseagents are more effective when targeted to non–codingsequences, such as 5´–leader sequences or splice junctions,given that an advancing ribosome can be expected todisplace an antisense oligo bound to a coding region [33].

1.1. Therapeutic Targets and Mechanisms of Action

Translational arrest is by far the most common mode ofaction for therapeutic intervention by modified nucleic acids.Several distinct types of modified oligos have beendeveloped which operate by different mechanisms and arecapable of inhibiting translation. These include those oligosthat activate RNase H, those that do not (steric blockingagents), those that recruit other cellular enzymes such asRNase L and finally ribozymes.

A third class of oligos capable of translational arrest arethe 2–5A–antisense chimeras [36]. These chimeras consist ofa 5´–monothiophosphorylated 2´→5´ linked tetraadenylate(2–5A) conjugated via a flexible alkyl spacer to an antisenseDNA, or a modified oligo. 2–5A chimeras were introducedbecause of their ability to activate RNase L, a ubiquitousenzyme in mammalian cells, which cleaves RNA whenactivated by binding to 2–5A, as part of the antiviralresponse induced by interferons. The antisense oligo portionof these chimeras is able to direct RNase L to cleave aspecific RNA sequence (Fig. 2a). A 2–5A–antisense agentwas recently shown to inhibit respiratory syncytial virus, inmonkey kidney cells and human HEP–2 cells, moreeffectively than phosphorothioate modified oligos of the samesequence, due possibly to the higher concentration of RNaseL than RNase H in the cytoplasm [37].

RNase H is a ubiquitous enzyme that cleaves the RNAstrand in a DNA.RNA heteroduplex [30]. Antisense oligosthat can recruit RNase H can cleave a target mRNA in acatalytic fashion, such that one antisense molecule caneffectively “knock out” several mRNA copies (Fig. 1). It istherefore not surprising that oligos that can recruit RNase Hare among the most potent antisense agents that have beendeveloped. However, only those backbone modified oligosthat most closely resemble native DNA, phosphorothioate,phosphorodithioate and boranophosphate modified oligos,can act as substrates for RNase H. This has prompted thedevelopment of chimeric (mixed backbone) oligos, which

mRNA

mRNA

ActivatedRNase L

Alkyl spacer

5´

a)

5´

3´

5´

2´

3´

✃

✃

ModifiedAntisense oligo

3´

NNNNNUHNNNN

NNNNNA

G

GG

5´

3´

3´

5´

b) mRNA cleavage site

2–5AChimera

2–5A NNNNC

N

UA

AP

AA

Y

NNN

N

NNN

NN N

3´3´

Gapmer 5´

5´

mRNA mRNA

3´

mRNA

5´

3´

✃RNase H

5´

3´

Modified oligo

mRNA

DNA

3´

5´

Fig. (1). RNase H binding to a mRNA–Gapmer heteroduplexfollowed by the sequence specific cleavage of the mRNA strand.

Fig. (2). a) RNase L activated by a 2–5A–antisense chimeraresulting in the sequence specific cleavage of a mRNA target. b)An hammerhead ribozymes [18]. N = any nucleotide; Y =pyrimidine; P = purine; H = A, C, or U.

Backbone Modification of Nucleic Acids Current Medicinal Chemistry, 2001, Vol. 8, No. 10 1159

Hammerhead ribozymes, the smallest of the known RNAenzymes, can also catalyse the cleavage of a target mRNAsequence–specifically resulting in translational arrest [18].These ribozymes base pair with the target RNA and fold intoa well defined structure consisting of a conserved catalyticcore flanked by three helical stems (Fig. 2b) [18]. Currentlymost advanced studies, in cell cultures and in clinical trials,are been conducted with ribozymes that are deliveredendogenously, using plasmids or viral vectors to express theribozyme within the cell. In addition, there is considerableinterest in exogenous delivery, which involves administeringthe ribozymes directly to cells. Clearly the difficultiesassociated with this latter approach, effective cellularpermeation and stability to enzymatic digestion, are thesame as other oligo therapeutic strategies. However sinceribozymes are generally larger, and RNA is more rapidlydegraded in biological fluid as well as more difficult tosynthesise than DNA, the technical problems are somewhatmagnified. Despite this there has been some degree ofsuccess in utilising chemical modifications to improve theproperties of ribozymes, whilst retaining and even enhancingtheir catalytic activity for therapeutic use [38,39]. Thesemodifications have mainly involved 2´–OH replacements inthe ribose ring, which will not be covered here. Howeverphosphorothioate backbone modifications have also beenused to stabilise ribozymes. Although this technology is notyet as advanced as other antisense strategies, the prospects ofutilising backbone modifications, which are discussed in thisreview, to improve the properties of ribozymes remains anintriguing and challenging possibility, that has yet to beexplored.

invested in the search for modified, or alternative baseswhich can recognise all four bases pairs in dsDNA [17].

All of the modes of action so far discussed involve thebinding of modified oligos to nucleic acid targets. Oftenoverlooked by those involved in the design and synthesis ofmodified oligos are protein targets for therapeuticintervention. The prospect that oligonucleotides could beused as therapeutic agents, through the binding to proteintargets, came to the fore with the development of methods forthe in–vitro selection of DNA or RNA ligands (aptamers),which can bind with high affinity to proteins [42,43]. Inprinciple, using techniques such as SELEX [43], aptamerscan be selected with high affinity for any protein target, froma pool of random DNA or RNA molecules generated andamplified using polymerase enzymes and PCR. Clearly theseoligos would be unstable in vivo, and to this end aptamerscan be stabilised by using 2´–fluoro or 2´–amino pyrimidinenucleoside triphosphates as substrates for in vitro selection[44]. In addition, post selection modification of aptamers hasalso been realised, using backbone as well as sugarmodifications, without compromising the high bindingaffinity for the protein target [19,44]. This therefore promisesto be another area in which oligonucleotide modification canbe used to improve the properties of therapeutic oligos. Inaddition strategies have been envisaged where modifiedoligos can be used as decoys [20] or transition stateanalogues [21], which mimic key nucleic acid structuralmotifs and bind certain enzymes that are responsible forprocessing the native nucleic acids. Whilst efforts intargeting proteins with modified oligos have only justbegun, it is clear that there are an awesome number ofpotentially distinct targets and mechanisms of action bywhich modified oligonucleotides might be used to curedisease.

Inhibition of translation by the antigene approach requiresthe binding of an oligo to a dsDNA target. This can beachieved by the formation of a triple helix, principally byHoogsteen hydrogen bonding of a third strand through theformation of T.AT and C+.GC triplets, or by reverseHoogsteen H–bonding and the formation of A.AT and G.GCtriplets [5,6]. Whilst the antigene approach has theadvantage that there are fewer copies of dsDNA than mRNAand therefore in principle a lower concentration of oligomight be required to effectively block transcription, theapproach is limited mainly to homopurine tracts in themajor groove. Moreover triple helices are generally lessstable than double helices, whilst the rate of binding of triplehelix forming oligos (TFOs) to dsDNA, is several orders ofmagnitude slower than the binding of an oligo to a singlestranded nucleic acid. Furthermore, any oligo operating in anantigene fashion must be able to penetrate within thechromatin structure of the cell nucleus [40]. These problemsmean that antigene strategies have not met with the samedegree of success as antisense strategies, in advancedbiological studies. Despite this promising results have beenachieved by improving the affinity of a triple helix formingoligos for homopurine tracts in dsDNA, using backbonemodified oligos, which has led to successful transcriptionalarrest in cell culture experiments [40]. Alternatively strandinvasion of dsDNA, by peptide nucleic acids, has alsoshown significant promise in antigene approaches [41].However, before this approach can match antisense in scope,problems associated with the recognition of dsDNA must beovercome. In this respect considerable effort has been

1.2. Design Criteria, Structure, Conformation andSynthesis

All therapeutic oligos must by stable to enzymaticdegradation and pass efficiently through the cell membrane tothe target site unaltered. Native oligos are not suitable forthis purpose because they are rapidly degraded by extra andintracellular nucleases. In the case of oligodeoxynucleotidesdegradation mainly occurs through the action of exonucleaseenzymes, with 3´–exonuclease occurring most rapidly. It isfor this reason that backbone modification strategies werefirst introduced. Every modification of the backbone that hasbeen investigated, which involves replacing thephosphodiester linkage with an alternative moiety, results inimproved stability to nuclease digestion. Indeed even thosemodifications that most closely resemble the nativestructures (e.g. phosphorothioates and boranophosphates) arestabilised significantly, whereas more radical backbonemodifications result in complete resistance to degradation.Efficient passage through the cell membrane presents muchmore of an obstacle. Although it has been speculated thatincreasing the lipophilicity of oligonucleotides by replacingthe anionic backbone with alternative neutral moieties mayresult in improved cellular penetration, this has still to beproved. In fact, whilst it depends on the particular cell typethat is targeted, to a large extent all modified oligos


irrespective of backbone chemistry, have trouble entering thecell unaided in high concentrations. For this reason anumber of delivery strategies are often employed to aid thepassage of oligos into cells. These include the use of:cationic liposomes [45]; microinjection [46]; the bacterialtoxin streptolysin O [31,35]; scrape loading [47]; andconjugation to small peptides [48] that are known to beinternalised within certain cell types. Whilst some of thesetechniques are not appropriate for human clinical studies,they have been used extensively in the efficient delivery ofmodified oligos in cell culture and animal experiments.

3´–O and 4´–O, which overcomes the anomeric effect (Fig.3b). In this respect, replacements of the phosphodiestergroup which result in 3´–substituent atoms with reducedelectronegativity and/or incorporation of electronegative 2´–fluorine, or 2´–Oalkyl groups on the sugar ring have beenused to good effect to produce modified oligos with highaffinity for complementary RNA [52,53,54]. All of theseconsiderations are important in the design of backbonemodified oligos, however it is not possible to consider theeffect of modifications with respect to one of these factor or athermodynamic parameter in isolation, since they are allintrinsically related.

Having effectively stabilised an oligo against degradationand assuming efficient delivery to the target site it isimportant that the modified oligo can interact with highaffinity and specificity with the target. Here modificationstrategies have had considerable success, particularly whentargeted to single stranded RNAs. In considering the effectsthat chemical modifications can have on binding affinity it isimportant to consider a simplified picture of the factorseffecting the thermodynamic stability of duplexes (Fig. 3a).The major stabilising forces in duplex formation arehydrogen bonding and base stacking (π–π stacking)interactions. Therefore base modifications that increase thenumber of hydrogen bond donors and acceptors, or increasethe π surface area of the bases can be used effectively toimprove affinity [46]. A major destabilising factor in duplexformation is electrostatic repulsion between backbones. Inthis case affinity can sometimes be enhanced by replacing thebackbone with neutral or even cationic backbones, whichalleviate the electrostatic repulsion. Often overlooked is therelative degree of hydration of the duplex. It is important thatany backbone modification is favourably hydrated in theduplex states [49]. All of these factors are primarily enthalpicand contribute to a typically large overall negative enthalpychange on hybridisation, which is the driving force for nativeduplex formation [50]. However, hybridisation of nativeoligos also results in a large loss of entropy which isunfavourable. Apart from the obvious loss of rotational andtranslational freedom inherent in any bimolecular interactionthe loss of free energy, associated with the internal bondrotations and vibrations, must also be considered. In thesingle stranded state the bonds are free to rotate and vibrate(random coil), but frozen out on formation of the more rigidduplex structure. By rigidifying the backbone in such a waythat the single stranded state is preorganised into a helicalbase stacked conformation prior to hybridisation, it ispossible to reduce the loss of entropy on duplex formationand increase binding affinity [50,51].

O

–

O

B

OH

OB

O

Electrostaticrepulsion

HO

✓

Preorganised ✓

✓

Random coil

✗

O

✗

b)

H -BondingBase Stacking

Hydration

– –

– – –

2´

(N) C3´–endo

✓

a)

(S) C2´–endo

σ*

– ∆H /

5´

3´

– ∆SH2O

H2O

H2O

Fig. (3). a) Factors effecting the stability of nucleic acidduplexes. b) C3´–endo and C2´–endo sugar conformationstypical of RNA and DNA respectively.

The effects of modifications on duplex and triplexstability are usually determined by measuring duplex andtriplex melting temperatures (Tm) using UV thermalanalysis. Typically values for the change of Tm permodification (∆Tm/mod), compared with an iso–sequentialnative reference duplex or triplex are quoted for comparison.It is however important to point out that binding affinity candepend significantly on the particular sequence used and asmany sequences as possible should be examined, includingmixed sequences, before definitive conclusions are reachedabout the effects of a particular modification. Most studiestend to use a number of standard sequences that have beenchosen for comparison and which give fairly representativeresults [55].

Similar considerations to those above have also resultedin the design of modified oligos which form more stabletriplexes. Once again modifications that more closelyresemble the conformation of RNA and adopt C3´–endosugar conformations, tend to form more stable triplexes,whilst alleviating electrostatic repulsion can also have goodeffects. Design criteria for modifying nucleic acids in order toenhance interactions with protein are less well established.Although it is important that modifications do not disruptmarkedly, secondary structure that may be important forbinding. In another respect, the negative charge of thephosphodiester group is more important in proteininteractions, than interactions with other nucleic acids, sincespecific contacts between the protein residues and thephosphodiester groups are often critical for binding. In thiscase alternative anionic replacements for phosphodiester can

In addition to these factors it is well established thatbackbone and sugar modifications that result in a preferencefor a C3´–endo or northern (N) conformation, typical of theribose sugar in RNA duplexes, tend to result in oligoswhich form more stable A–type duplexes withcomplementary RNA [51]. In RNA it is the presence of 2´–hydroxyl which drives the pseudorotational equilibrium ofthe ribose ring towards the N–conformer, which is attributedto a gauche effect between 2´–OH and 4´–O (ring oxygen)combined with an anomeric effect which operates when thebase is in a pseudoaxial position. In deoxyribose the absenceof the 2´–OH results in a more stable C2´–endo or southern(S) conformer as a consequence of the gauche effect between


be important [20]. Also, by systematically replacingphosphodiester groups with neutral linkers it is also possibleto probe which phosphodiester groups are and are not criticalfor binding, which can allow for further site specificmodifications to be investigated [19].

non–bridging oxygen atoms, in the phosphodiester group ofoligodeoxynucleotides, with a sulfur atom [58]. Theresulting phosphorothioate oligos (PS–oligos) 1 (Fig. 4)have been reviewed extensively [2–4, 8] and will thereforenot be discussed in detail here. PS–oligos are easilysynthesised on commercial DNA synthesiser, albeit as amixture of diastereoisomers at the phosphorus atom, and asalluded to earlier are resistant to nuclease cleavage, althoughless so than most other modified oligos. In addition they canact as substrates to RNase H. This has led to the widespreaduse of fully modified PS–oligos as antisense agent. Thesefirst generation antisense agents have been extensively testedin various human clinical trials against numerous targets.This has culminated in a PS–oligo called Vitravene (orFomivirsen), which is used in the treatment ofcytomegalovirus (CMV) induced retinitis in AIDS patients,and is the first antisense drug that has so far been approvedfor marketing. Despite this success PS–oligos have atendency to induce non–specific effects, through binding toextracellular and cellular proteins [59], as well as cleavage ofnon–target mRNAs, that are only partially complementary,due to activation of RNase H [33]. For these and otherreasons, including low binding affinity of PS–oligos forcomplementary nucleic acid targets, alternative modifiedoligos have been sought for antisense strategies for sometime. However the ease of synthesis of PS–oligos invariablymeans that they are frequently used against new antisensetargets, or are the first to be tried in the development ofalternative strategies requiring backbone modifications, suchas the 2–5A–antisense agents, ribozymes and aptamers aswell as other oligos designed for bind to specific proteintargets. In this respect PS–oligos are the standard by whichmost other backbone modified oligos, with therapeuticpretensions, are measured.

Finally any modification strategy must be syntheticallyfeasible, economically viable and rapid allowing access tosufficient quantities of material for detailed biologicalinvestigations. In this respect, it is important that thechemistry developed can be adapted for use on the solidphase, and preferably is compatible with existing syntheticprocedures, most notably the phosphoramidite or H–phosphonate chemistries which are most widely used for thesynthesis of native nucleic acids [56,57]. Whilst most baseand sugar modifications can be accommodated within theseprotocols, backbone modification present more of problemparticularly if the backbone functionality is significantlyremoved from the native phosphodiester. As a consequenceof this modified dinucleotides with 3´–phosphoramiditegroups and dimethoxytrityl (DMTr) protection at the 5´–endare employed extensively in the so called “block synthesis”approach, utilising standard DNA synthesisers. However, thechimeras produced by this method may be of limited utilityin therapeutic strategies where often a consecutive stretch ofmodifications might be required (e.g. gapmers). For thisreason synthetic strategies must be developed which useappropriately activated monomer synthons, that can becoupled with efficiency approaching that with which nativeoligos can be assembled. This can be technically demandingand take considerable effort and time in development. Forthis reason several promising modifications that havedisplayed favourable characteristics, when incorporated into afew positions within a chimeric oligo using the blocksynthesis approach, have taken considerable time fromconception before incorporation into relevant sequences foruse in biological studies.

OB

OBY

X

P

Y

O Z

Y

OB*O

O

PO

NC(CH2)2 O

H

5

1 X = O ; Y = O; Z = S –

2 X = NH ; Y = O; Z = O –

3 X = O ; Y = NH; Z = O –

4 X = O; Y = O; Z = NH2 or NHalkyl

Modified oligos can also be prepared by enzymaticprocedures using polymerases and modified nucleosidetriphosphates with various sugar, base and backbonemodifications (e.g. boranophosphate and phosphorothioateoligos). This can be particularly important in thedevelopment of therapeutic ribozymes and aptamers,although ultimately these structure must succumb to non–enzymatic synthesis.

2. BACKBONE MODIFICATIONS

2.1. Anionic Internucleoside Linkage

A large number of different backbone modifications havebeen investigated, particularly modifications that involveretaining the sugar ring and replacing the phosphodiesterlinkage with alternative anionic, neutral and cationic groups.In addition to these a number of nucleic acids analogues,where the whole sugar–phosphodiester backbone is replacedby alternative functionality, have been investigated. Tobegin with anionic phosphodiester replacements will bediscussed.

Fig. (4). The structures of phosphorothioate oligos (PS–oligos)1, phosphoramidate oligos 2 – 4 and a 5´–H–phosphonate 5used in the solid phase synthesis of N3´→P5´ phosphoramidateoligos. B = nucleoside base and B* = standard protectednucleoside base.

2.2. The N3´→P5´ Phosphoramidate Modification

Oligonucleotides with phosphoramidate, as opposed tophosphodiester linkages, have been the focus of muchThe whole area of research concerned with backbone

modified oligos began with the replacement of one of the


attention. These include the negatively charged, achiralN3´→P5´ phosphoramidates 2 with a bridging 3´–aminogroup [60,61], and the oppositely orientated P3´→N5´phosphoramidates 3 [62], as well as the neutral, but chiralphosphoramidates with non–bridging amino groups 4[63,64]. Of these the N3´→P5´ phosphoramidates,developed by Gryaznov and co–workers, have superior bio–physical properties and show greatest promise for therapeuticand other applications.

N3´→P5´ phosphoramidate oligodeoxynucleotides,including fully modified oligomers and phosphodiesterchimeras, form stable duplexes with both complementaryDNA and RNA [60,61,65]. On average duplexes with DNAand RNA, independent of nucleotide sequence and basecomposition, are stabilised by ca. 2.3–2.6 ˚C (increase inTm) per modification compared with native oligos. Inaddition duplexes, composed of self–complimentarysequences, in which both strands are uniformly modifiedwith N3´→P5´ linkages are similarly much morethermodynamically stable than the corresponding DNA orRNA duplexes. Furthermore N3´→P5´ phosphoramidatehomopyrimidine strands form more stable triplexes withboth dsDNA and dsRNA than corresponding native DNA[60,61,68]. By contrast the oppositely orientated P3´→N5´phosphoramidate 3 [62], and 3´NMe→P5´phosphoramidates 10 [61], with a N–methyl group replacingthe bridging NH group (Fig. 6) do not base pair with eithercomplementary DNA or RNA. Finally, it has been observedthat N3´→P5´ phosphoramidates with 2´–fluorosubstituents 11 [52] form even more stable duplexes withDNA and RNA, with a 4–5 ˚C increase in Tm permodification. Surprisingly 2´–Omethyl N3´→P5´phosphoramidates 12 form significantly less stable duplexeswith RNA and DNA [49]. It is expected, in both cases, thatthe additional presence of 2´–electronegative groups shouldfavour the 3´–endo sugar conformation and result in morestable A–type duplexes.

OB*O

NH2

OB*O

NHTr

PO(CH2)2 CNX

N

Me

Me

OB*

OB*O

NH

P

O

NC(H2C)2 O

NHTr

O

8

n

+

6

9 X =

2

7 X = N(iPr)2

CPG

Structural studies with a uniformly modified N3´→P5´phosphoramidate duplex, of sequenced(CGCGAATTCGCG)2, have been carried out using X–raycrystallography [49], molecular dynamics simulations[62,69] as well as NMR and CD spectroscopy [70,71].Unlike the Dickerson–Drew DNA dodecamer of identicalsequence which forms a classical B–form duplex, themodified duplex was shown by all methods to exhibitbackbone torsional angles including a C3´–endo sugarconformation typical of the iso–sequential A–form RNA.Furthermore, many of the helical parameters whichcharacterise the global duplex conformational state, derivedfrom both X–ray and the NMR solution structures, are closeto those observed for typical A–type DNA structures. Thefact that N3´→P5´ phosphoramidate modified DNA isstructurally much more similar to RNA than DNA, andforms remarkably stable duplexes is attributed to severalfactors [49,69–71]. Firstly replacement of the 3´–oxygen ofDNA with the more electropositive amino substituent,reduces the guache effect between the 3´–substituent and thesugar ring oxygen (O4´). This allows the anomeric effectbetween the lone pair on O4´ and the C1´–N glycosidicbond to stabilise the C3´–endo sugar conformation asopposed to C2´–endo which is typical in B–DNA.Secondly, the backbone nitrogen atom, which can adopt twopossible configurations, prefers to adopt a configurationwhich with an idealised all staggered A–type backboneplaces the N–H bond in an axial position (Fig. 6a). Thisconformation is stabilised because the nitrogen lone pair ispositioned antiperiplanar with the P–O5´ bond allowing fora strong anomeric effect involving overlap of the nitrogenlone pair with the P–O5´ antibonding orbital (σ*). Theanomeric effect further rigidifies the backbone, compared withphosphodiester linkages, by reducing the torsional flexibility

Fi.g (5). Solid phase synthesis of N3´→P5´ phosphoramidateoligos in a 5´→3´ direction, employing an amine–exchangereaction. Tr = trityl (triphenylmethyl).

Initially uniformly modified N3´→P5´ phosphoramidateoligodeoxynucleotides were synthesised on the solid phasefrom a CPG–supported nucleosides via a repetitive cycleinvolving 5´–detritylation and phosphitylation to give 5´–H–phosphonate diesters (e.g. 5 in Fig. 4) which can becoupled with standard protected 3´–amino nucleosides [65].In an effort to scale up the synthesis of N3´→P5´phosphoramidate modified oligos as well as chimeric oligosincorporating phosphoramidate and phosphodiester orphosphorothioate internucleoside linkages, for biologicalstudies, Fearon et al. [66] developed a more efficient andversatile synthetic procedure. The improved synthesis (Fig.5), which is fully compatible with standard DNAsynthesisers, takes place in the 5´→3´ direction andemploys an amine–exchange reaction between a 3´–aminonucleoside 6, and a 5´–phosphoramidite monomer 7.Oxidation of the intermediate phosphoramidite productfollowed by capping, allows for further detritylation andelongation steps resulting in oligo 8, which can be cleavedfrom the support and deprotected in the standard way.Further improvements of this chemistry have resulted fromthe use of more highly hindered–amine phosphoramidites(e.g. 9). By virtue of increased steric bulk, thesephosphoramidites 9 exchange essentially irreversibly withthe support bound 3´–amino group significantly reducing thenumber of equivalents of phosphoramidite monomer requiredfor efficient coupling, compared with the less hindereddiisopropylamine phosphoramidites 7 [67].


of the P–N3´ bond. Both these stereoelectronic factors serveto preorganise the phosphoramidate DNA into the requiredconformation for the formation of stable A–type duplexes. Athird significant factor is the H–bonding qualities of the 3´–NH group, which like the 2´–OH in RNA, can serve as bothdonor and acceptor increasing the hydration of the backboneor, as was observed in the X–ray crystal structure [49], thecoordination to anions. This can further contribute to theconformational rigidity of the backbone and the stability ofphosphoramidate modified duplexes. This model [49] alsoexplains the poor base pairing qualities of 3ŃMe→P5´phosphoramidate modified oligos, which are no longerfavourably solvated due to the lack of an acidic backbonehydrogen atom. Furthermore in the 2´–modified N3´→P5´phosphoramidates, the presence of a small 2´–Fluorosubstituents 11 allows for hydrogen bonding of water oranions, to the 3´–NH group and stable duplex formation. Incontrast the destabilisation of duplexes with 2´–OMe groupsis probably due to poor solvation or coordination of the 3´–NH due to steric constraints imposed by the more bulky 2´–OMe group (Fig. 6b). This is despite the fact that in bothcases the additional 2´–electronegative group increases thesugar conformational bias towards the 3´–endo sugar pucker.

phosphorothioate oligos. However, unlike PS–oligos,N3´→P5´ modified oligos are unable to activate RNase Hwhen bound to complementary RNA [73]. Presumably thisis a consequence of the failure of the enzyme to recognise themore RNA like conformation of the phosphoramidatebackbone in heteroduplexes.

Despite an inability to activate RNase H N3´→P5´phosphoramidate oligos have shown potent antisenseactivity in cell culture [72,73] and in vivo [34], whichpresumably must occur by a steric blocking mechanism.Studies in various leukaemia cell lines including primarybone marrow cells from patients with chronic myelogenousleukaemia (CML), were carried out with uniformly modifiedN3´→P5´ phosphoramidate oligos against bcr–abl, c–mycand c–myb oncogene mRNAs [72]. This revealed asequence–specific inhibitory effect on expression of thecorresponding proteins, which was significantly more potentthan the effect of iso–sequential PS–oligos. Thisantileukaemia effect was also examined in vivo, by treatingsevere combined immunodeficiency mice injected withhuman leukaemia cells (HL–60) [34]. Whilst both fullymodified iso–sequential phosphorothioate andphosphoramidate anti c–myc oligos were both effectivelytaken up by the leukaemia cells and had similar tissuedistributions in vivo, the survival rate of the mice treatedwith phosphoramidate anti c–myc oligo was significantlyhigher than those treated with the PS–oligos or controlmismatch oligos. Other antisense studies against the humanT cell leukaemia virus type–1 Tax protein [73], showed thata fully modified phosphoramidate oligo was effective ininhibiting tax gene expression in tax–transformed fibroplasts,whereas iso–sequential RNase H activating chimericgapmers, with terminal phosphoramidate orphosphorothioate modifications were both ineffective [73].Finally N3´→P5´ phosphoramidate oligos were shown toamongst the best inhibitors of HIV–1 reverse transcription invitro, compared with a series of other promising antisenseoligos when target against the TAR element of HIV–1 [74].

PN

O

O

H/F

HO

O

O

H

OHH

PN

O

O

O

H

O

O

H

OHH

CH3

B

B

O

OBO

N

P

O

O O

R2R1

σ*

3´

3´

A–type(C3´–endo)

–

–

(b)

(a)

5´

10 R1 = CH3 ; R2 = H

11 R1 = H ; R2 = F

12 R1 = H ; R2 = OCH3

–

σ*

Triple helix forming N3´→P5´ phosphoramidate oligoshave also shown significant promise in antigene basedstrategies due to their increased affinity for dsDNA, comparedwith other oligos. N3´→P5´ phosphoramidate oligoscontaining thymine and cytosine or guanine bases targetedagainst the polypurine tract (PPT) motifs of HIV proviralDNA were shown, through the formation of particularlystable triple helices, to effectively inhibit transcriptionalelongation in various in vitro assay systems [75]. Furtherstudies have shown that PPT binding phosphoramidateoligos, conjugated to a psoralen moiety are able to access,bind and cross link, after UV irradiation, to the PPT targetof the HIV proviral genome within the nuclear chromatinstructure in HIV–1 infected cell cultures [40].

Fig. (6). The structures of 3ŃMe→P5´ phosphoramidates 10 , 2´–fluoro 11 and 2´–Omethyl 12 N3´→P5´ phosphoramidates.Structures (a) and (b) show the preferred conformation of3Ń→P5´ phosphoramidates, with idealised all staggered A–type backbones [49].

In addition to superior qualities for binding to singleand double stranded nucleic acid targets, uniformly modifiedN3´→P5´ phosphoramidate are stable to nuclease enzymes(e.g. snake venom phosphodiesterase (SVPD) and alkalinephosphatase), and are resistant to hydrolysis in humanplasma and HeLa cell extracts for up to 8 hours [72]. Clearlystability of these oligos is due to electron donation from the3´–amino group which renders the phosphoramidate lesssusceptible to nucleophilic attack. Furthermore, these singlestranded phosphoramidate oligos were shown not to bind tonon–specific cellular proteins as was observed to occur with

In addition to antisense and antigene strategies N3´→P5´phosphoramidate DNA has been shown to be useful in thedevelopment of nuclease stable therapeutic oligos with alltogether different targets and modes of action [20,22,76].Particularly interesting is the idea that phosphoramidatemodified DNA because of the conformational similarity toRNA, can function as structural RNA mimetics or “decoys”and compete with viral RNA structural elements in binding


CGUGUG CGCAGCGCACAC GCGUCG3´

5´

U G

G G47 48

73 71

AGCCAGA GAGCAGCUCGGUCU CUCGUCG3´

5´ 23 25

40 39

a)

b)

A

U UU

Fig. (7). Duplex model of a minimal HIV–1 Rev binding element(RRE) (a) and Tat binding element (TAR) (b).

recent discovery that the hexamer d(ATGAC)rU, uniformlymodified with phosphoramidate linkages, can act as a suicideinhibitor of a self–splicing group–I intron from the fungalpathogen Pneumocystis carinii a common cause of death inimmunocompromised patients [22,76]. In this approach thenuclease stable hexamer was shown in vitro to compete withan iso–sequential 5´ exon in binding to the catalytic core ofthe group–I ribozyme contained within the pre–rRNA of P.carinii (Fig. 8). This binding event involves antisense base–pairing to an internal guide sequence which is significantlystabilised by tertiary interactions. Once bound the modifiedhexamer can undergo a trans–splicing reaction, throughattack by the terminal uridine 3´–OH group on thephosphodiester group between the 3´–exon and intron,leading to a dead–end product. Potentially, in vivo thiswould result in non–functional ribosomal RNA and woulddisrupt protein synthesis. Several other fungal pathogenscontaining similar self–splicing group–I introns could alsobe targeted, which makes this a particularly attractiveantifungal approach given that short oligos are more easilyprepared and more readily penetrate through the cellmembranes.

to viral gene regulatory proteins [20]. This was tested byGryaznov and Wilson et al. [20] who showed thatN3´→P5´ phosphoramidate DNA duplexes, which are iso–sequential to the minimal binding duplexes of the RRE andTAR elements of the HIV–1 RNA genome (Fig. 7), bindspecifically to peptides derived from Rev and Tat proteinsrespectively, with affinity comparable to the parent minimalRNA ligands. Since, Rev and Tat are regulatory proteinsessential for HIV–1 replication, these decoys are potentialand novel anti–HIV agents. Even more provocative is the

2.3. Boranophosphate Oligos

OH

OH

CUC

GUGCUC

UACUGG

AA TT GG AA CC UU

UCAGUA

GUGCUC

UACUGGAUGACUp

G–OH (3´)

5´

3´ 3´

5´

GUGpCUC

UACUGGAUGACUOH

3´

5´

pG

GUGOH

UACUGG

3´

pG

GUGpCUC

UACUGG

3´

5´

GUGOH

UACUGG

3´

AA TT GG AA CC UU

AUGACUCUC 3´5´AA TT GG AA CC UU 3´

+ +

UCAGUA5´

UCAGUA5´

(a)

(c)

(d)

3´

(b)

(f)

(e)

Boranophosphate DNA 13 (Fig. 9), introduced by Shawand co-workers, is derived by replacing one of the non–bridging oxygen atoms in the phosphodiester group of DNAwith borane (BH3) [77–79]. The boranophosphate diester isisoelectronic with phosphodiesters, isosteric with themethylphosphonate group, as well as chiral. Thesenegatively charged oligos are highly water soluble, but morelipophilic than DNA. NMR and CD studies show thatreplacing the phosphodiester group in dinucleotides with theboranophosphate diester results in only a slight change inconformational characteristics, such as sugar pucker, acyclictorsional angles and base stacking [80]. Furthermore, X–raycrystal structure data for dimethyl boranophosphate,(CH3O)2P(O)BH3

–, reveals close similarity in geometrywith dimethyl phosphate, except for a longer P–B bond(1.91 versus 1.51 Å) [81]. Initially it was supposed thatthese modified oligos might be used in the treatment ofcancers, using boron neutron capture therapy (BNCT), inaddition to potential antisense and antigene applications[78].

Boranophosphate modified DNA 13 can be prepared bysolid phase synthesis using an adaptation of H–phosphonatechemistry [79,82,83] and enzymatically using DNApolymerases and 2´–deoxyribonucleoside 5´–(α–P–borano)triphosphates 14. The α–borano dNTPs 14 can beprepared in good yields from the cyclotriphosphate 15 (Fig.9) [84]. Boronation using a N, N–diisopropylamine–boranecomplex gives 16, which can be ring opened by hydrolysisand deprotected upon ammonolysis. The diastereoisomericproducts 14 can easily be separated by HPLC, but theconfiguration of the α–phosphorus chiral centres have notbeen definitively assigned [84]. A single diastereoisomer ofα–borano dGTP, tentatively assigned Rp configuration atphosphorus, was incorporated into a DNA mixed sequence14mer by a primer extension reaction on a complementary14mer template using a DNA polymerase which resulted in a

Fig. (8). A schematic representation of the mechanism of suicideinhibition of the self–splicing group–I introns by N3´→P5´phosphoramidate hexamer d(ATGAC)rU [22]. The normal cis–splicing event follows the Pathway a → b → c → d. Thetrans–splicing event involves binding of the phosphoramidateoligo (outlined letters) and folding (a → e), followed bycleavage to give the dead–end product (f ). Nucleotides in boldtype face represent exons, whilst the intron sequence isrepresented by italicised letters and the tertiary interactionsresulting in the correctly folded ribozyme are signified by thedashed lines.


OB*

OR2

OP

OP

O

P O

O

O

O

O

BH3

OB

OBO

O

P

O

O BH3

O

OB

OH

OPOPOPO

O

O

O

O

O

BH3

R1

OB*

OR2

OP

OP

O

P O

O

O

O

O

R1

R1

–

14 R1 = H17 R1 = OH

15 R2 = Ac or Bz

–

13 R1 = H18 R1 = OH

–

16

–

–

––

–

–

Fig. (9). Structures of boranophosphate DNA 13 and RNA 18 . Synthesis of α–boranonucleoside triphosphates (14 & 17) forenzymatic synthesis of boranophosphate DNA and RNA.

single Sp boranophosphate linkage towards the 3´–end [85].The duplex formed with complementary DNA exhibited aTm that was 0.7 ˚C lower than the native DNA duplex,whilst the thermodynamic parameters of hybridisation (∆G˚,∆H˚ and ∆S˚) and duplex conformation, as determined byCD spectroscopy, were very similar to the native duplex[85]. A similar template–directed primer extension reactionwas also used to synthesise a 44–nucleotide stretch of mixedsequence, fully boranophosphate modified DNA [86].Furthermore, α–borano dNTPs were shown to becompatible with the polymerase chain reaction (PCR),allowing effective amplification of DNA incorporatingboranophosphates linkages, which has been used to developa novel PCR sequencing method that is complementary toother sequencing methods [29,87]. Finally, the synthesis ofribonucleoside α–boranotriphosphates 17 was recentlyreported [88], by an analogous route to that described abovefor the 2´–deoxy series (Fig. 9). It will be interesting to seeif these α–boranotriphosphates 17 are substrates for RNApolymerases allowing for the enzymatic synthesis ofboranophosphate RNA 18.

Early efforts in the non–enzymatic synthesis ofboranophosphate modified DNA were carried out in solutionusing phosphoramidite chemistry with borane–dimethylsulfide (DMS.BH3) oxidation replacing the standardoxidative step in DNA synthesis [77]. However solid phasesynthesis by this method is potentially problematic due tobase damage occurring during the boronation step andincompatibility of the resulting boranophosphate triesterswith the standard dimethoxytrityl deprotection step. The H–phosphonate approach was found to be superior for solidphase synthesis [79,82,83], since H–phosphonate diesterscan be boronated after chain elongation and removal of the5´–dimethoxytrityl group, with efficiency approaching thatfor the conventional oxidation step to the phosphodiestergroup in DNA synthesis [79]. Using this approacholigothymidine boranophosphates 22 (Fig. 10) are preparedby silylation of the H–phosphonate groups of 19 to give themore reactive trimethylsilyl phosphite triesters 20 which canbe boronated with o–chloropyridine–borane complex, amilder boronating agent than DMS.BH3, which preventsside reactions with the base [79]. The resultingboranophosphate triester 21 is readily hydrolysed with

OT

OTHO

O

P

O

O H

OBH3

O

P

O

TMSO

O

P

O

TMSO

OT

OTHO

O

P

O

O BH3

OH

nn

2122

20

19

CPG

–

Fig. (10). Solid phase synthesis of polythymidine boranophosphate DNA via H–phosphonates.


concentrated aqueous ammonia during cleavage from theCPG support. A fully modified boranophosphate T12oligomer synthesised by this approach showed somewhatlower affinity for complementary d(A)12 (Tm = 14 ˚C) thand(T)12 oligos with phosphorothioate (Tm = 28 ˚C),methylphosphonate (Tm = 32 ˚C) or native phosphodiesterlinkages (Tm = 47 ˚C) [79]. In separate studies, T14 fullymodified boranophosphates DNA was shown to formduplexes of similarly reduced stability with d(A)14 andr(A)14 [83], whilst T15 boranophosphates DNA wasobserved not to hybridise [82], albeit under slightly differentconditions. These results are in contrast to the mixedsequence oligo prepared enzymatically [85], with a singleenatiotopically pure boranophosphate linkage, which mightpossibly be due to the effects of phosphorus stereochemistryon duplex stability. In addition hybridisation studies withpolythymidine oligomers are not always representative of theeffects of modifications on mixed sequence duplex stability[55]. Therefore, further hybridisation studies are requiredwith mixed sequence boranophosphate, as well as stereopureboranophosphates oligos. These results may be forthcomingfollowing the development of efficient solid phase synthesisprotocols, and a recent report detailing a method for thestereoselective synthesis of an Sp boranophosphatedinucleoside, in 98% d.e., using a chiral auxiliary [89].

with a bioreversible protecting group, resulting in neutralphosphotriester or phosphorothioate triester pro–drugs (pro–oligos). It is expected that once inside the cell the pro–oligowill be “unmasked” by ubiquitous enzymes, releasing theparent oligonucleotide (drug). In addition to being morestable to nuclease degradation these neutral pro–oligos arelikely to exhibit lower non–specific binding to proteins,which is attributed largely to backbone negative charge innative and phosphorothioate modified oligos. Furthermore,it is speculated that the neutral pro–oligos would be morelipophilic, improving passage through the cell membrane.

To test these ideas Imbach et al. initially synthesiseddinucleotide model systems “masked” with variousbioreversible protecting groups, including those usedpreviously with other phosphorylated drugs, andinvestigated their stability in biological fluids [91]. Fromthese studies the S–acyl–2–thioethyl (SATE) group showedthe most promise. SATE groups are removed bycarboxyesterase mediated cleavage of the thioester groupfollowed by episulfide formation and elimination of parentphosphodiester or phosphorothioate groups (Fig. 11a) [96].Preliminary synthesis of SATE masked oligos was achievedby alkylation of the more nucleophilic sulfur atom inphosphorothioate linkages, centrally placed in chimeric oligogapmers with flanking regions of methylphosphonatemodifications, using appropriate alkylating agents (e.g. 23 &24 in Fig. 11b) [93]. Due to problems envisaged withincomplete alkylation of longer oligos and desulfurizationside reactions occurring during this post–alkylationprocedure, more recent synthesis of SATE masked oligoswas achieved using a phosphoramidite approach [96]. Thisinvolves the use of (SATE)–phosphoramidites 25, which canbe incorporated into oligomers in much the same way asstandard cyanoethyl phosphoramidites. However, thephosphotriester products are sensitive to bases andnucleophiles which excludes the use of standard CPG–succinate linkers and base protecting groups. To circumventthese problems thymidine (SATE)–phosphoramidite 25,which requires no base protection, and a photo–cleavablesolid support were used to synthesise pro–oligothymidylates(T12 ) with either phosphotriester or phosphorothioate triesterlinkages, the later requiring oxidation with Beaucagereagent. These uniformly masked DNA andphosphorothioate pro–oligos were shown to be stable tonucleases, SVPD and calf spleen phosphodiesterases, and arestable in human serum as well as gastric juice, whilstexhibiting lower affinity to serum proteins than nativeoligos. Futhermore, they are effectively hydrolysed to theparent oligonucleotides in total cell extracts rich incarboxyesterases consistent with the typical intracellularenvironment [96]. Whilst the in vivo properties of these pro–oligos have yet to be determined, provided that protectinggroups for the other bases (C, G & A) can be developed,which are compatible with the SATE pro–oligo synthesis,then this approach might be particularly useful for improvingthe efficacy of existing antisense PS–oligos, which wouldhave altogether different pharmacokinetic/pharmacodynamiccharacteristics.

Although it remains to be seen if boranophosphate oligoswill have the prerequisite binding affinity for nucleic acidtargets of therapeutic importance, other biological propertiesare favourable. Firstly, boranophosphate DNA isconsiderably more stable to various nuclease enzymes thannative DNA, and on the whole more stable thanphosphorothioate DNA [79]. Secondly and particularlyinteresting is the discovery that boranophosphate DNA canactivate E. coli RNase H and induce the cleavage of RNA[83,90]. This significantly increases the likelihood that thismodification will prove useful in antisense applications.Moreover, given that α–borano dNTPs are good substratesfor DNA polymerases and PCR [29,85–87], it would beinteresting to see if boranophosphate DNA aptamers can begenerated by in vitro selection, for binding to importantprotein targets. Even more intriguing is the possibility thatboranophosphate RNA aptamers or ribozymes might begenerated by a SELEX type procedure [43]. This willdepend on ribonucleoside α–boranotriphosphates 17functioning as substrates for RNA polymerases and theefficiency of boranophosphate RNA 18 to serve as a templatefor reverse transcription. Provided that these criteria can befulfilled then this would provide a potential route toboranophosphate RNA aptamers and ribozymes withimproved stability in biological fluids, for therapeuticapplications. Although such technology would necessitatethe re–synthesis of quantities of these boranophosphateRNAs, stereospecifically using solid phase chemistry, whichwould be particularly challenging.

2.4. The Pro–Oligonucleotide Concept

Following from the now well established pro–drugconcept, Imbach and co–workers introduced the idea of pro–oligonucleotides [91–96]. In this approach phosphodiester,or phosphorothioate oligonucleotide linkages are “masked”

Clearly alternative pro–oligo strategies should also beinvestigated [97]. Moreover, new pro–oligo strategies neednot necessarily involve bioreversible protection of


OP

O

X

X

S

RO

OP

O

X

X

SH

OP

O

X

X

ISR

OO

TDMTrO

OP

N(i Pr)2

OS R

O

X = O or S R = CH3 or tert–Bu

–Esterase

23 R = CH3 24 R = tert–Bu

a)

b)

25

Fig. (11). a) SATE bioreversible phosphodiester and phosphorothioate protecting groups. b) Post–synthesis SATE alkylatingagents 23 & 24 and 3´–(SATE)–phosphoramidites 25 .

phosphodiester linkage. For example, one might considerbioreversible protection of the 2´–hydroxyl group ofoligoribonucleotides. Like 2´–Oalkyl modified oligos, theseoligos would be expected to have increased lipophilicity andstability to nucleases, however upon in vivo deprotectionwould reveal oligoribonucleotides. This approach, and theSATE pro–oligo approach described above, might beparticularly attractive as a method for improving the in vivoefficacy and exogenous delivery of those therapeuticribozymes, aptamers or 2–5A–antisense agents, which mayprove difficult to stabilise by alternative modificationmethods.

can be improved with neutral linkages, but only providedthat the linkage is pre-organised into the correctconformation for binding and is favourably hydrated in theduplex state.

One of the earliest examples of a neutral oligonucleotideanalogue was the methylphosphonate oligos (MP–oligos), inwhich one of the non–bridging oxygen atoms of thephosphodiester group is replaced by a methyl group 26 (Fig.12). MP–oligos have been reviewed earlier [2–4, 8] and soonly a brief background will be provided along with somemore recent studies. Like boranophosphate andphosphorothioate modified oligos, the phosphorus atom isrendered chiral in these oligos, which are easily synthesised,as diastereoisomers at the phosphorus atom, by standardsolid phase synthetic protocols from readily availablemethylphosphonamidites 27 [57]. As a consequence of theease of synthesis, and the fact that they are highly resistant tonuclease degradation, the methylphosphonates were amongstthe first modified oligonucleotides to be tested as antisenseagents in vitro and in vivo against targets such as herpessimplex virus (HSV) and vasicular stromatitis virus (VSV)[98,99]. However the progress in the development of MP–oligos as antisense agents, was retarded due to several

2.5. Neutral Internucleoside Linkages:Methylphosophonates

Of all the backbone modifications, those involving thereplacement of the phosphodiester group with neutral linkers,out number the rest. Initially it was supposed that thesemore lipophilic oligos might increase the efficiency ofpassage into cells, whilst former the electrostatic repulsionwith target nucleic acids and therefore improve affinity.Whilst the former has not been shown to be the case, affinity

OB

OBO

X

P

O

O CH3

O

OBDMTrO

OP

N(iPr)2

CH3O

PO

H3CO

O

BO

P

O

O CH3

3´

(Rp)

26 X = O30 X = NH

27 5´

29

28

Fig. (12). Structures of methylphosphonate oligos 26 , methylphosphonamidites synthon 27 , A Rp–methylphosphonate linkage 28 ,α–anomeric methylphosphonates oligos 29 and methanephophonamidates 30 .


problems. In addition to poor aqueous solubility, fullymodified MP–oligos are unable to activate RNase H andhave significantly reduced affinity for complementary nativenucleic acids compared with native oligos. For these reasonsthis modification featured less often than phosphorothioateoligos in later biological studies.

amides, alkyl, alkenyl and alkynyl chains [10,14,108].These dinucleotide analogues were then incorporated intoseveral DNA sequences using a block synthesis approachwith modified dinucleoside phosphoramidites, resulting inchimeric backbones with alternating phosphodiester andmodified linkages. The thermal stability of the duplexesformed between these chimeras and complementary RNA andDNA was then determined. Of all the modificationsinvestigated, the two isomeric amides 31 and 32 (Fig. 13)were the only replacements that resulted in duplexes thatwere as stable as native duplexes with an average change inTm of between – 0.5 and +0.5 ˚C per modification withcomplementary RNA [109,110]. Interestingly the affinity forcomplementary DNA is somewhat lower for these amidechimeras. It is argued that restricted rotation about the amidebond, with a preference for the trans isomer, preorganisesthese two backbones into a conformation resembling A–typeduplexes. Additional NMR [111] and molecular modellingstudies [112] with modified duplexes support thishypothesis and additionally show that the amidemodification in isomer 31 also results in a bias of theadjacent 5´–terminal sugar towards the C3´–endoconformation which further explains the stability of duplexeswith RNA. In addition to this, favourable hydration of theamide linkage in duplexes, may also be an important factor,since it was found that substitution of the acidic amidehydrogen atom in 31 with alkyl and aryl groups reduces theaffinity of chimeric oligos for RNA targets [110,113].

Despite these problems more recent findings have re-initiated some interest in this modification. Firstly Tidd andco–workers showed that chimeric oligodeoxy gapmers,comprising of a central sequence of RNaseH activatingphosphodiester linkages with flanking regions ofmethylphosphonate modifications, which have no solubilityproblems, were efficiently able to cleave target mRNAs invitro and in human cell cultures [100]. Notably, the samegroup recently showed that the methylphosphonate–phosphodiester gapmers, were more effective antisense agentsagainst c–myc mRNA in human leukaemia cells followingtreatment with streptolysin O, than other promisingantisense agents including similar gapmers with flankingregions of 2´–Oalkyl modifications and phosphorothioateoligos with C5–propyne pyrimidine modifications [31].Similar methylphosphonate antisense gapmers against bcr–abl mRNA combined with streptolysin O delivery arecurrently being used in clinical trials with patients sufferingfrom chronic myeloid leukaemia [101]. A second majorfactor which has maintained interest in methylphosphonateswas the discovery that oligos with Rp configuredmethylphosphonate linkages 28 have higher affinity forcomplementary DNA and RNA than the Sp configureddiastereoisomers and in some sequences exhibit virtually thesame Tm as native DNA duplexes [102,103]. In light of thisapproaches towards the stereocontrolled synthesis of Rpconfigured MP–oligos have been investigated [104], but sofar effective methods are limited to dinucleosides [105].

Given that the amide backbone of type 31 aresynthetically more accessible, further modifications to thissystem were carried out. For example additional 2´–methoxygroups in dimer 33, which restricts both sugars to C3´–endoconformations, contribute to an increase in Tm withcomplementary RNA of between 2–3 ˚C per dimermodification [53]. Furthermore, an additional C5´–methylgroup (S configuration) 34 rigidifies the backbone further,which when combined with 2´–methoxy groups as in dimer35 results in a dramatic increase in Tm of up to 4.4 ˚C(∆Tm/mod) with RNA complements [114]. Clearly the latter

Finally, two interesting variants of MP–oligos haverecently been introduced, the α–anomericmethylphosphonates 29 [106] and themethanephophonamidates 30 [107] which are N3´→P5´Phosphoramidate–methylphosphonate hybrid structures. Inthe former, inversion of the anomeric configuration wasshown to improve the affinity of MP–oligos forcomplementary single stranded DNA and RNA as well asdouble stranded DNA [106]. Whilst in the latter case DNAincorporating between one and four modified linkages 30 ofeither configuration (Rp or Rs) showed reduced affinity forsingle stranded DNA and RNA targets [107]. However, oneunassigned stereoisomer, showed improved affinity fordsDNA. Both modifications 29 and 30 are under furtherinvestigation as potential triple helix forming oligos forantigene strategies. O

B

OB

O

O

HN

OR1

R2

R1

NH

O

OB*

NHMMTr

CO2C6 F532

31 R1 = R2 = H 33 R1 = OCH3; R2 =H 34 R1 = H; R2 = CH3 35 R1 = OCH3; R2 = CH3

3´

36

5´

Fig. (13). Structures of amide linked oligos 31–35 , nucleosidepentafluorophenyl esters 36 for solid phase synthesis of amidelinked DNA (MMTr = Monomethoxytrityl).

2.6. Amide Linked DNA

In light of the chirality problems associated withmethylphosphonates, many groups have investigatedalternative neutral, achiral internucleoside linkages. One ofthe most extensive contributions in this area has come fromDe Mesmaeker and co–workers at Novartis who havesynthesised neutral dinucleotide analogues with manydifferent linker groups including various ureas, carbamates,


OB

OB

O

O

N

O

R

R

CH3

OB*

O

O

R

OB*

RO

CHO

OB*

O*B

N

O

O

RO

OB

OB

N

O

OH

RO

CH3

n

41 R = Phth43 R = TPS

42 R = Phth44 R = TPS

39 R = Phth40 R = NH2

45 R = H46 R = PO or PS oligo

nDeprotectReduce

Methylate

37 R = H38 R = OCH3 or F

Fig. (14). Methylene methylimino (MMI) oligos 37 & 38 . Solid phase synthesis of MMI oligos 45 and phosphodiester (PO) orphosphorothioate (PS) chimeras 46 (Phth = Phthalimide & TPS = tert–butyldiphenylsilyl).

system 35 is optimally preorganised into the required A–type conformation for stable heteroduplexes with RNA.Despite these promising results no biological studies havebeen reported for these modified oligos. This is most likelydue to the fact that synthesis of biologically relevantsequences, for antisense experiments, is not best achievedusing the block synthesis approach. However, a syntheticmethod which is compatible with standard solid phase DNAsynthesis has recently be developed which utilisesnucleoside pentafluorophenyl esters 36 [115]. Thesemonomers can be used to synthesise mixed sequence fullymodified amide linked oligomers or chimeras withcontinuous regions of phosphodiester as well as amidelinkages and should facilitate long awaited biological studieswith this class of modified oligos.

efficient solid phase synthesis of fully modified MMI oligos[117]. Their procedure (Fig. 14) depends upon 5´deprotection of the CPG linked 5´–O–phthalimidonucleosides 39 to reveal the 5´–O–amino groupof 40 which is coupled, under mild acid catalysis, with the3´–aldehyde 41 resulting in an oxime dimer 42. A repeat ofthe cycle gives oxime oligomers, with a coupling efficiencyof 96–99 %, which can be cleaved from the support and basedeprotected under standard conditions. Reduction of theoxime and backbone N–methylation gives the fully modifiedMMI oligos 45. In addition MMI/phosphodiester orphosphorothioate chimeras can be prepared from oligo–oximes 44, derived from a tert–butyldiphenylsilyl (TPS)protected nucleoside monomer 43, by desilylation andcoupling with standard phosphoramidite monomers.MMI/phosphorothioates antisense gapmers, synthesised bythis method, were shown to inhibit the expression of proteinkinase C–alpha (PKC-α), an anticancer target, as effectivelyas iso–sequential phosphorothioate oligos [117]. Furtherbiological studies with MMI oligos are currently inprogress.

2.7. Methylene(methylimino) Linkages

In parallel studies at ISIS, Sanghvi and co–workers[116,117] have investigated several other neutral achiralphosphodiester replacements, culminating in oligosincorporating the methylene(methylimino) (MMI) linkage 37(Fig. 14). A block synthesis approach was also used toconstruct MMI / phosphodiester chimeric oligos, whichshowed similar affinity for complementary RNA, as theamide modified oligos, with a change in Tm permodification, of between –0.2 and + 1.5 ˚C compared withnative DNA [116]. The affinity for RNA can be furtherincreased by as much as + 3.8 ˚C (∆Tm/mod) by theadditional incorporation of electronegative groups at the 2´position 38 [54,118]. NMR and modelling studies withMMI dimers also revealed that the MMI linkages arerestricted in conformation to either of two low energystructures which differ in the orientation of the N–methylgroup, both allowing the bases to take up a stacked helicalconformation for base pairing [119]. Again the presence of aless electronegative C3´–methylene group also gives rise toa preference for the C3´–endo conformation in the 5´–terminal sugar, which is further enhanced by 2´–electronegative modifications 38 [120]. To facilitate access tobiologically relevant sequences ISIS have also developed an

2.8. Formacetal and Thioformacetal Linkages

In addition to amide and MMI other neutralmodifications that are well accommodated within nucleicacid duplexes include the formacetal 47 and 3´–thioformacetal 48 phosphodiester replacements (Fig. 15),which are restricted in rotation freedom, by an anomeric effect[121–126]. Whilst the formacetal modification 47 has beenshown to induce a preferred C2´–endo sugar conformation,resulting in more favourable binding to ssDNA, thethioformacetal modification 48, incorporating the lesselectronegative C3´–sulfur atom, prefers C3´–endo sugarpucker and forms more stable duplexes with ssRNA. Thesefindings have earmarked the thioformacetal modification aspotentially useful for antisense applications, but progress inthis direction as yet to be reported. However, one altogetherdifferent therapeutic application of formacetal modified oligoswas recently reported [19]. This arose from the earlierdiscovery of a DNA aptamer, generated by in vitro selection,


OB

OBO

X

O

O

CH H G5G11

G1

G10 G6

G2G14

G15

(5´)

4947 X = O48 X = S

G8

T7

T3

T4

T13

T9

T1 2

Fig. (15). Structures of formacetal 47 and 3´–thioformacetal 48modified oligos. Structure of a thrombin binding DNA aptamer49 .

isomeric 3´–N–sulfamates 54 and related sulfamides 55[131,132]. Accordingly dimers were synthesised andbetween one and five modifications were incorporated into astandard sequence of DNA [55]. This revealed that thesulfamides result in lower affinity for complementary DNAand RNA (ca. – 3 ˚C ∆Tm/mod), whilst the 3´–N–sulfamates have approximately the same affinity as nativeDNA [131]. The 3´–N–sulfamates are similar to theN3´→P5´ phosphoramidate in that the 3´–amino groupwhich is less electronegative than the native 3´–oxygenresults in a bias of the sugar pucker towards the C3´–endoconformation [132]. Binding studies of 5´–N–sulfamateswith complementary RNA have not been reported, but meritexploration. However, it can be expected that the 3´–N–sulfamates will have higher affinity for RNA because they aremore likely to be preorganised in the required conformationfor A–type duplexes.

which binds with high affinity to thrombin [42]. Theminimal binding domain of the aptamerd(GGTTGGTGTGGTTGG) forms a chair–like structure 49in which two stacked G–tetrads are connected by two T2loops and TGT loop. Despite a short half–life the aptamerhas potent in vivo anticoagulant properties. Bysystematically replacing each of the phosphodiester linkageswith formacetal linkages 47, using block synthesis, thosepositions where negative charge is essential for binding tothrombin were determined by measuring the thrombininhibitory activity of the modified aptamers in vitro. Fromthis a modified aptamer with four formacetal groups wasselected and shown to have an increased anticoagulant effectand an extended half–life in monkeys, compared with theparent aptamer. This report opens up the possibility thatother phosphodiester replacements which cause minimaldisruption to the conformation of native DNA and RNAmight also be used to stabilise, and even improve thebinding affinity of other aptamers which have been selectedfor binding to medically important protein targets.

OB

OBO

X

S

Y

O O

O

50 X = Y = CH2

51 X = NH2; Y = CH2

52 X = O; Y = CH2

53 X = NH2; Y = NH2

54 X = NH2; Y = O

55 X = Y = NH2

Fig. (16). Structures of sulfonyl containing oligos 50–55 .

Despite poor affinity for complementary nucleic acids thedimethlene sulfone modification 50 has been particularlyinsightful in other respects [21,25,133–135]. Studies onshort RNA analogues with sulfone linkages (rSNA) hasrevealed that their physical properties and chemical reactivityis highly dependent on sequence and as such rSNAs moreclosely resemble proteins than nucleic acids [25,133,134].The further consequences of replacing the phosphodiesterwith neutral sulfones has also prompted a re–examination ofthe role of the anionic backbone in the structure and functionof nucleic acids [25,134]. Furthermore, X–raycrystallographic studies with sulfone dinucleotide analoguesreveal distorted backbone conformation, considerablyremoved from native duplex conformations [24,13], whichprobably explains the dramatic drop in affinity of DNAincorporating the sulfone linkage for complementary nucleicacids. Remarkably this preorganisation of a dimethylenesulfone linkage into a distorted conformation results in abend in DNA duplexes, incorporating a sulfone linkage,which resembles the high energy conformation of duplexDNA transition states when bound to restriction enzymessuch as Eco RV [21]. This fact was used to design duplexDNA with a sulfone linkage in the Eco RV recognition site.The resulting modified bent duplexes are not substrates forthe restriction enzymes, but instead are transition stateanalogues which act as efficient inhibitors of the nativesubstrates [21].

2.9. Backbones Containing Sulfonyl Groups

A number of modifications have been introducedinvolving replacing the PO2– group of native oligos withneutral isosteric and isoelectronic SO2 groups. These includedimethylene sulfones 50 [25,127], sulfonamides 51 [128],sulfonates 52 [129], 5´–N–sulfamates 53 [130], 3´–N–sulfamates 54 [131] and sulfamides 55 [132] (Fig. 16).Oligos incorporating these linkages have been prepared fromdinucleoside phosphoramidites using the block synthesisapproach and the thermal stability of duplexes have beendetermined. In most of these cases duplex stability isdiminished significantly, with the exception of thesulfamates. The most destabilising is the dimethylenesulfone group 50 which results in a dramatic drop in Tm ofduplexes with both DNA and RNA (ca. – 15 ˚C ∆Tm/mod)[127], followed by sulfonamides 52 (ca. – 5 ˚C ∆Tm/mod)[128], which suggest that these modifications will not beuseful in antisense or antigene experiments. An early paperon 5´–N–sulfamates 53, reports the incorporation of onesulfamate linkage into each of two complementary mixedsequence DNA 20mers, resulting a drop in Tm of – 1.5 ˚Cper modification [130]. Despite this no further reports havebeen published. This prompted an investigation into the

Many enzymes and proteins bind to specific DNAsequences resulting in bends or other conformations which


are ordinarily high energy and therefore unpopulated in thenative unbound state. This opens up the intriguingpossibility that backbone modification such as dimethylenesulfone, which may not be useful in the antisense or antigeneapproaches, but mimic the distorted high energy transitionstates of nucleic acids when bound to various proteins, caninstead be used to develop a novel class of biologicallyactive agents that inhibit the function of key nucleic acidbinding proteins. Further developments in this area can beexpected due to the increase in detailed NMR and X–raystructures of native nucleic acids bound to various proteins,as well as various backbone modified nucleic acid analogues[24].

the pentamer does not show any hyperchromic shift withnon–complementary homooligomers. The extremely highstability of DNG duplexes and triplexes with DNA as wellas RNA, is mainly attributed to electrostatic attractionbetween the oppositely charged complementary backbones.As expected the stability decreases with an increase in saltconcentration, due to attenuation of the electrostaticattraction. Kinetic studies have also revealed that the rate oftriple helix formation between DNG (T5) with shortoligodeoxyadenylates, is an order of magnitude faster thanthe formation of DNA triplexes at similar salt concentrations[146]. Also, investigations with DNG (T5 and T8), showthat Tm’s are significantly reduced by base pair mismatches[147,148]. This suggest that sequence specific base pairingis the predominant mode of binding between DNG andnucleic acid targets, although it has still to be establishedwhether or not this is accompanied by any non–specificbinding, via salt bridges, between the oppositely chargedbackbones. Clearly such a mode of non–specific binding,involving direct contact between the backbones of DNG andnucleic acids, could be detrimental for antigene and antisenseapplications. Since these interactions may not necessarily beaccompanied by a change in base–stacking, methods otherthan UV melting experiments (e.g. surface plasmonresonance), may be required to prove that base pairing is theexclusive mode of binding. In this respect furtherinvestigation into the binding properties DNGs with RNAand DNA targets of mixed sequence, are required before thepotential of this system can be fully appreciated. Suchstudies appear close at hand with the development of anefficient solid phase synthetic method for the synthesis ofDNG (Fig. 18) [148]. The synthesis involves the couplingof a 3´–amino nucleoside 58, immobilised on a polystyreneresin, with a 5´–carbodiimide intermediate which is derivedfrom the thiourea 59, upon treatment with HgCl2. Removalof the 3´–Fmoc protecting group, allows for furtherelongation with monomer units 59. The resulting oligomer60 is then cleaved from the resin with mild acid and theguanidine acyl protecting groups are removed usingcadmium in acetic acid.

2.10. Positively Charged Deoxyribonucleic Guanidine(DNG) Oligos

Several groups have reported the synthesis of zwitterionicand cationic oligodeoxynucleotides with pendant alkylaminoside chains attached to the 2´–position of the sugar ring[137], C5–position of pyrimidine base [138], or to variousphosphorus internucleoside linkages [139,140]. Thesependant amino groups, which are protonated at physiologicalpH, can in some cases form salt bridges with phosphodiestergroups in complementary nucleic acids, alleviatingelectrostatic repulsion and allowing the formation of stableduplexes and triplexes with complementary nucleic acids,whilst retaining good aqueous solubility.

Remarkably, only a few de novo modified oligos havebeen reported with positively charged backbones replacingthe native backbone [140,141]. By far the most interestingare the deoxyribonucleic guanidine (DNG) oligomers,introduced by T. C. Bruce and co–workers [143–145], inwhich the phosophodiester group is replaced by a positivelycharged guanidinium group 56 (Fig. 17). Pentamericthymidyl DNG oligomers have been shown to bind withpoly(rA) and poly(dA), with higher affinity than any othermodified oligonucleotide so far investigated. The pentamerbinds in a 2:1 triple helical fashion, as evidenced by twohyperchromic shifts in the UV melting curves, correspondingto the melting of the triple and double helical structures[143–145]. Remarkably at low salt concentration DNG/DNAor RNA duplexes have been shown to be stable attemperatures approaching the boiling point of water, whilst

Finally, in addition to DNG, Bruce et al. [149,150] havealso reported the synthesis of pentameric thymidyloligomers, with positively charged methylated thiourealinkages 57 (Fig. 17) replacing the phosphodiester group.These oligomers which are more hydrophobic than the DNGoligomers, also form duplexes and triplexes with DNA andRNA which are much more stable than the correspondingnative structures, albeit the affinity is slightly less whencompared with the prototype DNG.

OT

OTO

NH

C

HN

CH2

HN

NH

C

HN

SCH3

56

57

+

Fig. (17). Deoxyribonucleic guanidine (DNG) 56 and themethylthiourea modification 57.

2.11. Whole Backbone Replacements: MorpholinoOligos

Although conceptually straightforward, retaining theribose or deoxyribose sugar and investigating alternativelinker groups has some limitations. In most cases theappropriate activated monomer units require several syntheticsteps from relatively expensive deoxyribonucleosideprecursors and even where cheaper ribonucleosides are used2´–protection or 2´–modifictions necessitate severalexpensive extra steps. Furthermore, it is technically


OT

NH

C

HN

NH2

NR

OT

C

HN

NHFmoc

SHN

R

O

O CCl3

OT

NH

C

HN

NH

NR

FMoc

58 59

+

R =

60

n

Fig. (18). Solid phase synthesis of deoxyribonucleic guanidine (DNG) oligos.

challenging to develop efficient coupling chemistry for solidphase synthesis of oligomers, particularly if the chemistry issignificantly different to that used to prepare nativeoligonucleotides or peptides on solid supports. This makesthe synthesis of some of the modified oligos discussed so farboth expensive and time consuming. As a result thesynthesis of sufficient quantities of some of these oligos forbiological evaluation has been the major “bottle–neck”which has prevented the further exploitation of initiallypromising bio–physical properties. In this respect oligos,with radically altered structures in which both the sugar andphosphodiester linkage are replaced by alternativefunctionality, most notably morpholino oligos and peptidenucleic acids (PNA), offer several advantages over some ofthe modifications mentioned thus far.

form heteroduplexes with complementary RNA which aremore stable than the corresponding native DNA.RNAduplexes.

Fully modified morpholino oligos have been shown tobe more effective antisense agents than iso–sequential PS–oligos in cell–free systems and in various cultured cellsusing the scrape–loading method to allow for efficientcellular uptake [47,153]. This is attributed to the fact thatmorpholino oligos, do not bind to non–specific proteins,like PS–oligos, and also are much more sequence specific,because they depend on a steric blocking mechanism and donot activate RNase H, which can result in cleavage of manypartially complementary sequences other than the targetsequence. In one notable study morpholino oligos wereshown to be more effective than PS–oligos as sequencespecific antisense inhibitors of the tumor necrosis factor–α(TNF–α) in mouse macrophages, despite poor uptake intothese cells [154,155]. Most recently, a 28–mer morpholinooligo targeted at c-myc pre–mRNA, overlapping with theinitiation codon, was shown to completely inhibit c–MYCprotein expression in several human leukaemia cell lines,following treatment with streptolysin O [35]. Particularlyinteresting was the fact that the morpholino oligo, inaddition to inhibiting splicing, also induces missplicing ofthe c-myc pre–mRNA giving rise to a translatable mRNAwhich results in a shorter, N–terminal deleted, MYCprotein. In contrast, an iso–sequential fully modified 2´–methoxyethoxy oligo, which has high affinity forcomplementary RNA and is also incapable of activatingRNase H, did not result in inhibition of translation, splicingor induce missplicing of c–myc pre–mRNA in the same cell

Oligos with morpholino nucleosides linked together byphosphorodiamidate groups 61, are cheaper and moreefficiently assembled than many of the other promisingmodified oligos [33,151]. Morpholino nucleosides 63 areprepared from ribonucleosides 62 (Fig. 19), with baseprotecting groups, in an high yielding one–pot procedureinvolving diol cleavage with sodium periodate, followed byammonolysis of the intermediate dialdehyde and reductionwith sodium cyanoborohydride. N–tritylation andphosphitylation provides a monomer unit 64 that can beused to synthesise oligomers on a solid support with acoupling efficiency of 99.7%. Fully modified morpholinooligos, unlike some fully modified neutral oligos, have goodaqueous solubility and as expected are extremely stable todegradation in biological fluids [152]. Despite possessing achiral phosphorodiamidate linker morpholino oligomers

OB*HO

OH OH

NH

O B*HO

62

63N

O B*

64Tr

O

PO NMe2

Cl

N

O B*

61

O

PO NMe2

N

O BO

Fig. (19). Synthesis of morpholino oligos.


lines. These favourable results, combined with the lowproduction costs, have prompted more in depth evaluation ofthe properties of morpholino oligos in vivo.

reveals an altogether unusual structure (P–type) which ismuch wider (28 Å) and has a larger pitch, of 18 base pairsper turn, than native A– or B–type duplexes [166]. Thissuggests that the PNA backbone is more flexible than nativenucleic acid backbones allowing it to adapt to A–, B– andP–type conformations. Therefore PNA can not be viewed asoptimally preorganised, for binding to either RNA or DNA,despite the high stability of resultant heteroduplexes. It cantherefore be expected that structural modifications to thePNA backbone, rigidifying the structure into an optimalconformation for binding to nucleic acids, should improvefurther the stability of heteroduplexes. So far most attemptsto do this have failed and the prototype PNA structure 65still remains one of the best systems for selectivehybridisation with a target RNA sequences [159,163]. Thiscombined with resistance to degradation, by both nucleaseand peptidase enzymes, has led to considerable efforts todevelop these molecules as antisense therapeutics[28,160,161].

2.12. Peptide Nucleic Acids

Peptide Nucleic Acids (PNA) 65, developed by Nielsenand co–workers [156,157], are a remarkable example of asimple, neutral and achiral whole backbone replacement,which in many ways have surpassed other attempts to mimicthe native nucleic acid structures in terms of molecularrecognition properties (Fig. 20). It is not possible to reviewall the work in the PNA area, for this the reader is referred tothe many original reviews on PNA synthesis, structure,properties and applications [158–163]. However no review,on backbone modified oligonucleotides, would be completewithout a discussion of these molecules and so following abrief introduction to the important properties of PNA thelatest advances in the therapeutic applications of PNA willbe discussed. Homopyrimidine PNAs also form (PNA)2.DNA triplexes

with extremely high thermal stability. Remarkably, as aconsequence of this, homopyrimidine PNAs targeted topurine tracts of dsDNA invade the duplex by displacing oneof the DNA strands and form a triple helix, with one PNAstand binding by Watson and Crick base pairing in anantiparallel orientation and the other binding in a parallelHoogsteen fashion. The resulting stable structures are calledP–loops 66 (Fig. 20). The efficiency of this strand invasioncan be further increased using bis–PNAs, consisting of twohomopyrimidine PNA oligos connected via a flexible linker68. bis–PNAs have increased affinity presumably because ofa reduction in the loss of entropy upon binding comparedwith formation of a termolecular P–loop 66. Furthermore,bis–PNA with pseudoisocytosine bases 67, instead ofcytosine bases on the Hoogsteen half, bind more tightly todsDNA independently of pH. Recent evidence suggests, thatstrand invasion by bis–PNA is likely to occur by a two stepprocess where the Hoogsteen H–bonds form first to give aPNA.(DNA)2 intermediate 69. This is then followed byWatson and Crick base pairing and strand invasion,resulting in the bimolecular P–loop 70 [167]. Clearly this

PNA consists of N–(2–aminoethyl)glycine units with thebases attached via a side chain amide linkage. The twoamide groups in each monomer, in addition to a gaucheeffect between the two backbone NH groups, infer somerigidity on the acyclic backbone. The amide groups alsofacilitate the relatively straight forward synthesis ofmonomers and then oligomers using solid phase peptidesynthesis methods. PNA oligomers are poorly soluble inwater and for this reason a lysine residue is usually attachedto the C–terminal end. PNA forms duplexes withcomplementary RNA and DNA that are considerably morestable than the native duplexes. Furthermore, PNA.DNA andPNA.RNA duplexes exhibit equivalent or in some caseslowered tolerance for base pair mismatches compared withthe native duplexes. This means that the binding specificityof PNA, for complementary nucleic acids, is at least as highas DNA. Structural studies, using NMR, on PNA.nucleicacid heteroduplexes indicate that PNA is able to adopt bothA– and B–type structures with RNA and DNA respectively[164,165]. In contrast, an X–ray structure of a PNA2 duplex

N

B

O

O

HN

O

N

O

B

HN

65

N NH

O

NH2

67

Watson and CrickHoogsteen

P-Loop

66

706968

+

bis-PNA dsDNA

Fig. (20). Peptide nucleic acids 65 , P–loop structure 66 , pseudoisocytosine base 67 and the mechanism of bis–PNA strand invasion(68→69→70 ).


unique property has earmarked PNA has a potential antigeneagent [28,160,161].

where much less promising, due to poor uptake, whichsuggests that the efficacy of these approaches will behampered by cell wall permeability problems similar tothose experienced with eukaryotic cell cultures.Research into the development of PNA antisense agents

has met with some success and several groups have reportedeffective inhibition of the translation of various target RNAsin cell cultures [45,168]. Notably antisense PNAs weredesigned to bind to the RNA component of humantelomerase, a ribonucleoprotein responsible for maintainingtelomeres, which is implicated in the proliferation of tumors[45]. In these studies the antisense PNAs inhibitedtelomerase activity, after they were first hybridised with apartially overlapping DNA strand which allowed for efficientdelivery into a human prostrate tumor cell line usingcationic liposomes. In addition, there have been recentreports of successful antisene experiments using PNA inanimals. In one study PNA 21mers conjugated to a specifictransporter protein and targeted to the galanin–receptormRNA of rats were effective in reducing the number ofgalanin receptors, in the dorsal horns of the rat’s spinal cord,following intrathecal administration [48]. This was evidentby a decrease in the number of galanin binding sites, asdemonstrated by an electrophysiological response in the rats.The effective in vivo antisense response in rats was attributedto efficient delivery of the PNA, which is facilitated by ashort transporter proteins called pAntp. This protein belongsto class of related proteins which are known to be effectivelyinternalised by living cells [169]. In other studies, shorterunmodified PNAs targeted at opiod (mu) and neurotensin(NTR–1) receptor mRNAs, where microinjected into thebrains of rats. The behavioural responses of the rats tomorphine and neurotensin, along with competitive bindingassays with neuronal homogenates from sacrificed rats, wereconsistent with a reduction of receptor sites due to anantisense mechanism [170]. Further studies, also showedthat same “naked” antisense PNAs were effectively able tocross the blood–brain barrier and inhibit translation of NTR–1 mRNA, following intraperitonial injection into rats [171].These results suggest that PNA antisense agents, with andwithout carrier proteins, can be effectively used in vivo, atleast when targeted to neuronal cells. Two recent studiesalso report effective antisense experiments in E. coli. In thefirst study, PNAs targeted at β–galactosidase and β–lactamase mRNA showed inhibition of enzyme productionin an antibiotic permeable E. coli strain, that wastransformed with the β–lactamase gene which confersresistance to the β–lactam antibiotics (e.g. ampicillin) [172].Furthermore the inhibition of β–lactamase activity wassufficient to re–sensitize the bacteria to ampicillin. In thesecond report, bis–PNAs designed to bind in a triplexfashion to ribosomal RNA targets of the same permeable E.coli strain, were shown to inhibit protein synthesis andbacterial growth. This antibacterial effect was as strong asthat of the antibiotic tetracycline, which also functions bybinding to rRNA [173]. These reports offer excitingprospects that PNAs and possibly other modified oligosmight be used in antibiotic treatments, either by blockingexpression of resistance genes combined with conventionalantibiotics, or by the specific disruption of bacterialribosome function. These prospects are particularly appealingin light of the ever increasing threat caused by the evolutionof “super bugs” with resistance to known antibiotics. Itshould be pointed out that the results with wild type E. coli

Considerable effort has been invested in the developmentof PNAs and bis–PNAs as antigene agents. Several in vitrostudies have shown that the binding of PNA, or bis–PNA todsDNA can block replication [174], transcriptionalelongation [175,176] and inhibit other proteins which bindto dsDNA including helicases [177] which are thought toprecede polymerases. Nevertheless, there is some concernthat antigene PNAs would be less effective than antisensePNAs as consequence of poor kinetics. Strand invasion ofdsDNA is particularly slow at physiological saltconcentration [176,177]. However, efforts in this directionhave not be deterred, particularly given that it has beenshown that the rate of PNA binding to dsDNA is increasedby a passing RNA polymerase [176], and regions of negativesupercoiled DNA [179]. It can therefore be expected thatPNA strand invasion may be faster and more efficient in vivo,due to distorted DNA topology occurring duringtranscription and replication in the nuclear environment[178]. To some extent this has been born out and a fewstudies have reported transcriptional arrest, in human cellcultures [41,180] and in animals [171], using PNAs targetedto dsDNA. One study reports a 10 fold increase in themutational frequency of the supFG1 gene targeted with bis–PNAs, following introduction into mouse fibroblasts [41].Remarkably in another report, a “naked” mixedpurine/pyrimidine sense PNA was shown to reduce NTR–1mRNA levels by 50% following direct injection into thebrains of rats [171]. It is difficult to see how this particularPNA sequence could form stable complexes with dsDNA, byconventional triple helix binding or strand invasion, unlessthe target gene is distorted or transiently opened in vivo.

3. CONCLUSION

This review has focused on the major chemical andbiological advances in the development of backbonemodified oligonucleotides as therapeutic agents.Phosphorothioate modified oligos have evolved furthest withthe notable advent of the first antisense drug. Several morerecent modifications most notably N3´→P5´phosphoramidate, PNA and morpholino oligos have beenshown to be equally, if not more, effective as antisenseagents in eukaryotic as well as prokaryotic cell culture, andin animal experiments. Despite an inability to activateRNase H, all of these oligos exhibit enhanced bindingaffinity, increased stability in biological fluids, and reducednon–specific effects. In addition PNA and morpholino oligosare likely to be more economically viable in the long term,due to ease of synthesis from cheaper starting materials.Furthermore N3´→P5´ phosphoramidate and PNA, becauseof their superior qualities for binding to dsDNA targets, havealso been employed effectively as antigene agents in cellculture experiments.

The use of backbone modifications in the development ofalternative strategies, with therapeutic potential has beenequally as encouraging. Again the N3´→P5´


phosphoramidates have been particularly useful, showingpromise as novel antifungal agents which function as suicideinhibitors of self–splicing introns in fungal pathogens.Phosphoramidate modified DNAs are also structural RNAmimetic and by acting as decoys, which bind HIV regulatoryproteins, show promise as novel anti–HIV agents. Neutralphosphodiester replacements, formacetal and dimethylenesulfone, have also been used effectively in the development ofDNA aptamers and dsDNA transition state analogues thatcan bind to protein targets. In addition it can be envisagedthat backbone modifications might also be used to improvethe properties of ribozymes for therapeutic application.

[11] Matteucci, M. D. Perspectives in Drug Discovery andDesign 1996, 4, 1.

[12] Varma, R. S. Synlett 1993, 621.

[13] Nielsen, P. E. Annu. Rev. Biophys. Biomol. Struct. 1995,24 , 167.

[14] De Mesmaeker, A.; Altmann, K.–H.; Waldner, A.;Wendeborn, S. Curr. Opin. Struct. Biol. 1995, 5, 343.

[15] Hunziker, J. Leumann, C. In Modern Synthetic Methods;Ernst, B.; Leumann, C, Eds.; VCH: New York &Weinheim, 1995, pp. 331–419.

Detailed studies of the effects of backbone modificationson the conformational, physical and biological properties ofnucleic acids are of crucial importance in realising thetherapeutic goals. These studies have been enabled by newand improved methods for the solid phase synthesis ofbackbone modified oligos. Most recently MMI, DNG andamide linked oligos have all been prepared with efficienciesapproaching that achieved in the synthesis of native nucleicacids. Whilst boranophosphate DNA has succumbed toefficient synthesis by both enzymatic and solid phasemethods. It can be expected that the availability of these andother modifications will provide further ammunition forimproving the efficacy of antisense and antigene agents, inaddition to developing new modes of action through whichmodified oligonucleotides can have beneficial therapeuticeffects.

[16] Wengel, J. Acc. Chem. Res. 1999, 32 , 301.

[17] Luyten, I.; herdewijn, P. Eur. J. Med Chem. 1998, 33 ,515.

[18] Rossi, J. J. Chem. & Biol. 1999, 6, 33.

[19] He, G.–X., Williams, J. P.; Postich, M. J.; Swaminathan,S.; Shea, R. G.; Terhorst, T.; Law, V. S.; Mao, C. T.;Sueoka, C.; Coutré, S.; Bischofberger, N.; J. Med. Chem.1998, 41 , 4224.

[20] Rigl, C. T.; Lloyd, D. H.; Tsou, D. S.; Gryaznov, S. M.;Wilson, W. D. Biochemistry 1997, 36 , 650.

[21] Blättler, M. O.; Wenz, C.; Pingoud, A.; Benner, S. A. J.Am. Chem. Soc. 1998, 120, 2674.

[22] Testa, S. M.; Gryaznov, S. M.; Turner, D. H. Proc. Natl.Acad. Sci. USA 1999, 96 , 2734.

REFERENCES [23] Earnshaw, D. J.; Gait, M. J. Biopolymers 1998, 48 , 39.

[1] Oligonucleotides as Therapeutic Agents. Proceedingsof the Ciba Foundation Symposium No. 209; John Wiley& Sons: Chichester, 1997.

[24] Egli, M. Angew. Chem. Int. Ed. Engl. 1996, 35 , 1894.

[25] Richert, C.; Roughton, A. L.; Benner, S. A. J. Am. Chem.Soc. 1996, 118, 4518.

[2] Stein, C. A.; Kreig, A. M. (Eds.) Applied AntisenseOligonucleotide Technology; John Wiley & Sons: NewYork, 1998.

[26] Gat, Y.; Lynn, D. G. Biopolymers 1998, 48 , 19.

[27] Böhler, C.; Nielsen, P. E.; Orgel, L. E. Nature 1995, 376,578.[3] Crooke, S. T. (Ed.) Handbook of Experimental

Pharmacology: Antisense Research and Applications,Springer–Verlag: Berlin & Heidelberg, 1998, vol. 131. [28] Nielsen, P. E. Current Opinion in Biotechnology 1999,

10 , 71.[4] Crooke, S. T. Therapeutic Applications of

Oligonucleotides, Springer–Verlag: Heidelberg, 1995. [29] Porter, K. W.; Briley, J. D.; Shaw, B. R. Nucleic AcidsRes.1997, 25 , 1611.

[5] Soyfer, V. N.; Potaman, V. N. Triple–Helical NucleicAcids, Springer–Verlag: New York, 1996. [30] Crooke, S. T.; Lemonidis, K. M.; Neilson, L.; Griffey, R.;

Lesnik, E. A.; Monia, B. P. Biochem. J. 1995, 312, 599.[6] Antigene: Thuong, N. T.; Hélène, C. Angew. Chem. Int.

Ed. Engl. 1993, 32 , 666. [31] Giles, R. V.; Spiller, D. G.; Grzybowski, J.; Clark, R. E.;Nicklin, P.; Tidd, D. M. Nucleic Acids Res., 1998, 26 ,1567.[7] Giovannangeli, C.; Hélène, C. Antisense Nucleic Acid

Drug Dev. 1997, 7, 413.[32] Monia, B. P.; Johnston, J. F.; Sasmor, H.; Cummins, L. L.

J. Biol. Chem. 1996, 271, 14533.[8] Uhlmann, E; Peyman, A. Chem. Rev. 1990, 90 , 543.

[9] Milligan, J. F.; Matteucci, M. D.; Martin, J. C. J. Med.Chem. 1993, 36 , 1923.

[33] Summerton, J.; Weller, D. Antisense Nucleic Acid DrugDev. 1997, 7, 187.

[10] De Mesmaeker, A.; Häner, R.; Martin, P.; Moser, H. E.Acc. Chem. Res. 1995, 28 , 366.

[34] Skorski, T.; Perrotti, D.; Nieborowska–Skorska, M.;Gryaznov, S. M.; Calabretta, B. Proc. Natl. Acad. Sci.USA 1997, 94 , 3966.


[35] Giles, R. V.; Spiller, D. G.; Clark, R. E.; Tidd, D. M.Antisense Nucleic Acid Drug Dev. 1999, 9, 213.

[56] Gait, M. J. (Ed.) Oligonucleotide Synthesis: A PracticalApproach; Rickwood, D.; Hames, B. D., Eds.; IRL Press:Oxford, 1990.

[36] Maran, A.; Maitra, R. K.; Kumar, A.; Dong, B.; Xiao, W.;Li, G.; Williams, B. R. G.; Torrence, P. F.; Silverman, R.H. Science 1994, 265, 789.

[57] Eckstein, F. (Ed.) Oligonucleotides and Analogues: APractical Approach; Rickwood, D.; Hames, B. D., Eds.;IRL Press: Oxford, 1991.

[37] Player, M. R.; Barnard, D. L.; Torrence, P. F. Proc. Natl.Acad. Sci. USA 1998, 95 , 8874. [58] DeClerq, E.; Eckstein, F.; Merigan, T.C. Science 1969,

165, 1137.[38] Heidenreich, o.; Benseler, F.; Fahrenholz, A.; Eckstein,

F. J. Biol. Chem. 1994, 269, 2132. [59] Stein C. Nature Med. 1995, 1, 1119.

[39] Beigelman, L.; McSwiggen, J. A.; Draper, K. G.;Gonzalez, C.; Jensen, K.; Karpeisky, A. M.; Modak, A.S.; Matulic–Adamic, J.; Direnzo, A. B.; Haeberli, P.;Sweedler, D.; Tracz, D.; Grimm, S.; Wincott, F. E.;Thackray, V. G.; Usman, N. J. Biol. Chem. 1995, 270,25702.

[60] Gryaznov, S. M.; Chen, J–K. J. Am. Chem. Soc. 1994,116, 3143.

[61] Gryaznov, S. M.; Lloyd, D. H.; Chen, J–K; Schultz, R.G.; DeDionisio, L. A.; Ratmeyer, L.; Wilson, W. D.Proc. Natl. Acad. Sci. USA 1995, 92 , 5798.

[62] Barsky, D.; Colvin, M. E.; Zon, G.; Gryaznov, S. M.Nucleic Acids Res. 1997, 25 , 830.

[40] Giovannangeli, C.; Diviacco, S.; Labrousse, V.;Gryaznov, S.; Charneau, P.; Hélène, C. Proc. Natl. Acad.Sci. USA 1997, 94 , 79.

[63] Peyrottes, S.; Vasseur, J.–J.; Imbach, J. L.; Rayner, B.Nucleic Acids Res. 1996, 24 , 1841.[41] Faruqi, A. F.; Egholm, M.; Glazer, P. M. Proc. Natl.

Acad. Sci. USA 1998, 95 , 1398.[64] Endo, M.; Komiyama, M. J. Org. Chem. 1996, 61 , 1994.

[42] Bock, L. C.; Griffin, L. C.; latham, J. A.; Vermaas, E. H.;Toole, J. J. Nature 1992, 355, 564. [65] Chen, J–K; Schultz, R. G.; Lloyd, D. H.; Gryaznov, S. M.

Nucleic Acids Res. 1995, 23 , 2661.[43] Gold, L.; Tuerk, G. Science 1990, 249: 505.

[66] Nelson, J. S.; Fearon, K. L.; Nguyen, M. Q.; McCurdy, S.N.; Frediani, J. E.; Foy, M. F.; Hirschbein, B. L. J. Org.Chem. 1997, 62 , 7278.

[44] Ruckman, J.; Green, L. S.; Beeson, J.; Waugh, S.; Gillette,W. L.; henninger, D. D.; Claesson–Welch, L.; Janjic, N.J. Biol. Chem. 1998, 273, 20556.

[67] Fearon, K. L.; Hirschbein, B. L.; Nelson, J. S.; Foy, M. F.;Nguyen, M. Q.; Okruszek, A.; McCurdy, S. N.; Frediani,J. E.; DeDionisio, L. A.; Raible, A. M.; Cagle, N.; Boyd,V. Nucleic Acids Res. 1998, 26 , 3813.

[45] Hamilton, S. E.; Simmons, C. G.; Kathiriya, I. S.; Corey,D. R. Chem. & Biol. 1999, 6, 343.

[46] Flanngan, W. M.; Wagner, R. W.; Grant, D.; Lin, K.–Y.;Matteucci, M. D. Nat. Biotechnol. 1999, 17 , 48. [68] Zhou–Sun, B.; Sun, J.; Gryaznov, S. M.; Liquier, J.;

Garestier, T.; Hélène, C.; Taillandier, E. Nucleic AcidsRes. 1997, 25 , 1782.[47] Partridge, M.; Vincent, A.; Matthews, P.; Puma, J.; Stein,

D.; Summerton, J. Antisense Nucleic Acid Drug Dev.1996, 6, 169. [69] Cieplak, P.; Cheatham, III, T. E.; Kollman, P. A. J. Am.

Chem. Soc. 1997, 119, 6722.[48] Pooga, M.; Soomets, U.; Hällbrink, M.; Valkna, A.; Saar,

K.; Rezaei, K.; Kahl, U.; Hao, J.–X.; Xu, X.–J.;Wiesenfeld–Hallin, Z.; Hökfelt, T.; Bartfai, T.; Langel, Ü.Nat. Biotechnol. 1998, 16 , 857.

[70] Ding, D.; Gryaznov, S. M.; Lloyd, D. H.;Chandrasekaran, S.; Yao. S.; Ratmeyer, L.; Pan, Y.;Wilson, W. D. Nucleic Acids Res. 1996, 24 , 354.

[71] Ding, D.; Gryaznov, S. M.; Wilson, W. D. Biochemistry1998, 37 , 12082.

[49] Tereshko, V.; Gryaznov, S. M.; Egli, M. J. Am. Chem.Soc. 1998, 120, 269.

[72] Gryaznov, S. M.; Skorski, T.; Cucco, C, Nieborowska–Skorska, M.; Chiu, C. Y.; Lloyd, D.; Chen, J.–K.;Koziolkiewicz, M.; Calabretta, B. Nucleic Acids Res.1996, 24 , 1508.

[50] Kool, E. T. Chem. Rev. 1997, 97 , 1473.

[51] Herdewijn, P. Liebigs Ann. 1996, 1337.

[52] Schultz, R. G.; Gryaznov, S. M. Nucleic Acids Res. 1996,24 , 2966. [73] Heidenreich, O.; Gryaznov, S.; Nerenberg, M.; Nucleic

Acids Res. 1997, 25 , 776.[53] De Mesmaeker, A.; Lasueur, C.; Bévièrre, M.–O.;

Waldner, A.; Fritsch, V.; Wolf, R. M. Angew. Chem. Int.Ed. Engl. 1996, 35 , 2790.

[74] Boulme, F.; Freund, F.; Moreau, S.; Nielsen, P. E.;Gryaznov, S.; Toulme, J.–J.; Litvak, S. Nucleic Acids Res.1998, 26 , 5492.

[54] Peoc’h, D.; Swayze, E. W.; Bhat, B.; Dimock, S.; Griffey,R.; Sanghvi, Y. S. Nucleosides & Nucleotides 1997, 16 ,959.

[75] Escudé, C.; Giovannangeli, C.; Sun, J.–S.; Lloyd, D. H.;Chen, J .–K.; Gryaznov, S. M.; Garestier, T.; Hélène, C.Proc. Natl. Acad. Sci. USA 1996, 93 , 4365.

[55] Freier, S. M.; Altmann, K.–H. Nucleic Acids Res. 1997,25 , 4429.


[76] Testa, S. M.; Gryaznov, S. M.; Turner, D. H. Biochemistry1998, 37 , 9379.

[100] Tidd, D. M. Perspectives in Drug Discovery and Design1996, 4, 51.

[77] Sood, A.; Shaw, B. R.; Spielvogel, B. F.; J. Am. Chem.Soc. , 1990, 112, 9000.

[101] Clark, R. E.; Grzybowski, J.; Broughton, C. M.; Pender,N. T.; Spiller, D. G.; Brammer, C. G.; Giles, R. V.; Tidd, D.M. Bone Marrow Transplantation 1999, 23 , 1303.

[78] Spielvogel B. F.; Sood, A.; Shaw, B. R.; Hall, I. H. Pure& Appl. Chem. 1991, 63 , 415. [102] Reynolds, M. A.; Hogrefe, R. I.; Jaeger, J. A.; Schwartz, D.

A.; Riley, T. A.; Marvin, W. B.; Daily, W. J.; Vaghefi, M.M.; Beck, T. A.; Knowles, S. K.; Klem, R. E.; Arnold Jr,L. J. Nucleic Acids Res., 1996, 24 , 4584.

[79] Sergueev, D. S.; Shaw, B. R. J. Am. Chem. Soc. 1998, 120,9417.

[80] Li, H.; Huang, F.; Shaw, B. R. Bioorg. Med. Chem. 1997,5, 787.

[103] Schweitzer, M.; Engels, J. W. J. Biomol. Struct. Dyn.1999, 16 , 1177.

[81] Summers J. S.; Roe, D.; Boyle, P. D.; Colvin, M.; ShawB.R. Inorg. Chem. 1998, 37 , 4158.

[104] Wozniak, L. A. Reviews Heteroatom Chemistry 1999,19 , 173.

[82] Zhang, J.; Terhorst, T.; Matteucci, M.D. Tetrahedron Lett.1997, 38 , 4957.

[105] Wozniak, L. A.; Chworos, A.; Pyzowski, J.; Stec, W. J. J.Org. Chem. 1998, 63 , 9109.

[83] Higson, A. P.; Sierzchala, A.; Brummel, H; Zhao, Z.;Caruthers, M. H. Tetrahedron Lett. 1998, 39 , 3899.

[106] Debart, F.; Meyer, A.; Vasseur, J.–J.; Rayner, B. NucleicAcids Res., 1998, 26 , 4551.

[84] Krzyzanowska, B. K.; He, K.; Hasan, A; Shaw, B. R.,Tetrahedron 1998, 54 , 5119.

[107] Nawrot, B.; Boczkowska, M.; Wójcik, M.; Sochacki, M.;Kazmierski, S.; Stec, W. J. Nucleic Acids Res., 1998, 26 ,2650.

[85] Li, H.; Porter, K.; Huang, F.; Shaw, B. R., Nucleic AcidsRes. , 1995, 23 , 4495. [108] De Mesmaeker, A.; Waldner, A.; Wendeborn, S.; Wolf, R.

M. Pure & Appl. Chem. 1997, 69 , 437.[86] Sergueev, D.; Hasan, A.; Ramaswamy, M.; Shaw, B. R.,

Nucleosides & Nucleotides 1997, 16 , 1533. [109] Leberton, J.; Waldner, A.; Fritsch, V.; Wolf, R. M.; DeMesmaeker, A. Tetrahedron Lett. 1994, 35 , 5225.

[87] He, K.; Porter, K. W.; Hasan, A.; Briley, J. D.; Shaw, B.R. Nucleic Acids Res., 1999, 27 , 1788. [110] De Mesmaeker, A.; Waldner, A.; Leberton, J.; Hoffmann,

P.; Fritsch, V.; Wolf, R. M.; Freier, S. M. Angew. Chem.Int. Ed. Engl. 1994, 33 , 226.[88] He, K.; Hasan, A.; Krzyzanowska, B.; Shaw, B. R. J. Org.

Chem. 1998, 63 , 5769.[111] Blommers, M. J. J.; Pieles, U.; De Mesmaeker, A. Nucleic

Acids Res., 1994, 22 , 4187.[89] Jin, Y; Just, G. Tetrahedron Lett. 1998, 39 , 6433.

[90] Rait, V. K.; Shaw, B. R. Antisense Nucleic Acid DrugDev. 1999, 9, 53.

[112] Wolf, R. M.; De Mesmaeker, A.; Waldner, A.;Wendeborn, S.; Fritsch, V.; Leberton, J. New. J. Chem.1997, 21 , 61.

[91] Barber, I.; Rayner, B.; Imbach J.–L. Bioorg. Med. Chem.Lett. 1995, 5, 563. [113] Waldner, A.; De Mesmaeker, A.; Wendeborn, S. Bioorg.

Med. Chem. Lett. 1996, 6, 2363.[92] Barber, I.; Tosquellas, G.; Morvan, F.; Rayner, B.; Imbach

J.–L. Bioorg. Med. Chem. Lett. 1995, 5, 1441. [114] De Mesmaeker, A.; Leberton, J.; Jouanno, C.; Fritsch, V.;Wolf, R. M.; Wendeborn, S. Synlett 1997, 1287.

[93] Tosquellas, G.; Barber, I.; Morvan, F.; Rayner, B.; ImbachJ.–L. Bioorg. Med. Chem. Lett. 1996, 6, 457. [115] von Matt, P.; De Mesmaeker, A.; Pieles, U.; Zürcher, W.;

Altmann, K.–H. Tetrahedron Lett. 1999, 40 , 2899.[94] Tosquellas, G.; Morvan, F.; Rayner, B.; Imbach J.–L.

Bioorg. Med. Chem. Lett. 1997, 7, 263. [116] Sanghvi, Y. S.; Swayze, E. W.; Peoc’h, D.; Bhat, B.;Dimock, S. Nucleosides & Nucleotides 1997, 16 , 907.

[95] Mignet, N.; Morvan, F.; Rayner, B.; Imbach J.–L. Bioorg.Med. Chem. Lett. 1997, 7, 851. [117] Morvan, F.; Sanghvi, Y. S.; Perbost, M.; Vasseur, J.–J.;

Bellon, L. J. Am. Chem. Soc. 1996, 118, 255.[96] Tosquellas, G.; Alvarez, K.; Dell’Aquila, C.; Morvan, F.;

Vasseur, J.–J.; Imbach, J.–L.; Rayner, B. Nucleic AcidsRes.1998, 26 , 2069.

[118] Swayze, E. W.; Bhat, B.; Peoc’h, D.; Sanghvi, Y. S.Nucleosides & Nucleotides 1997, 16 , 971.

[97] Mauritz, R. P.; Meier, C. Nucleosides & Nucleotides1997, 16 , 675.

[119] Mohan, V.; Griffey, R.; Davis, D. R. Tetrahedron 1995,51 , 6855.

[98] Smith, C. C.; Aurelian, L.; Reddy, M. P.; Miller, P. S.;Ts’o, P. O. P. Proc. Natl. Acad. Sci. USA 1986, 83 , 2787.

[120] Gousset, H.; Liquier, J.; Taillandier, E.; Sanghvi, Y. S.;Peoc’h, D. J. Biomol. Struct. Dyn. 1998, 15 , 931.

[99] Agris, C. H.; Blake, K. R.; Miller, P. S.; Reddy, M. P.;Ts’o, P. O. P. Biochemistry 1986, 25 , 6268.

[121] Jones, R. J.; Lin, K.–Y. Milligan, J. F. Wandwani, S.;Matteucci, M. D. J. Org. Chem. 1993, 58 , 2983.


[122] Pudlo, J. S.; Cao, X.; Swaminathan, S.; Matteucci, M. D.Tetrahedron Lett. 1994, 35 , 9315.

[146] Blaskó, A.; Dempcy, R. O.; Minyat, E. E.; Bruice, T. C. J.Am. Chem. Soc. 1996, 118, 7892;

[123] Zhang, J.; Shaw, J. T.; Matteucci, M. D. Bioorg. Med.Chem. Lett. 1999, 9, 319.

[147] Blaskó, A.; Minyat, E. E.; Dempcy, R. O.; Bruice, T. C.Biochemistry 1997, 36 , 7821.

[124] Veal, J. M.; Brown, F. K. J. Am. Chem. Soc. 1995, 117,1873.

[148] Linkletter, B. A.; Sabo, I. E.; Bruice, T. C. J. Am. Chem.Soc. 1999, 121, 3888.

[125] Rice, J. S.; Gao, X. Biochemistry 1997, 36 , 399. [149] Arya, D. P.; Bruice, T. C. J. Am. Chem. Soc. 1998, 120,12419.

[126] Cross, C. W.; Rice, J. S.; Gao, X. Biochemistry 1997, 36 ,4096. [150] Arya, D. P.; Bruice, T. C. Proc. Natl. Acad. Sci. USA

1999, 96 , 4384.[127] Baeschlin, D. K.; Hyrup, B.; Benner, S. A.; Richert, C. J.

Org. Chem. 1996, 61 , 7620. [151] Summerton, J.; Weller, D. Nucleosides & Nucleotides1997, 16 , 889.

[128] McElroy, E. B.; Bandaru, R.; Huang, J.; Widlanski, T. S.Bioorg. Med. Chem. Lett. 1994, 8, 1071. [152] Hudziak, R. M.; Barofsky, E.; Barofsky, D. F.; Weller, D.

L.; Huang, S.–B.; Weller, D. D. Antisense Nucleic AcidDrug Dev. 1996, 6, 267.[129] Huang, J.; McElroy, E. B.; Widlanski, T. S. J. Org. Chem.

1994, 59 , 3250.[153] Summerton, J.; Stein, D.; Huang, S.–B.; Matthews, P.;

Weller, D.; Partridge, M. Antisense Nucleic Acid DrugDev. 1997, 7, 63.

[130] Huie, M. E.; Kirshenbaum, M. R.; Trainor, G. L. J. Org.Chem. 1992, 57 , 4569.

[131] Fettes, K. J.; Howard, N.; Adah, S.A.; Player, M. R.Torrence, P. F.; Micklefield J. Chem. Commun. 2000,765.

[154] Taylor, M. F.; Paulauskis, J. D.; Weller, D. D.; Kobzik, L.J. Biol. Chem. 1996, 271, 17445.

[155] Taylor, M. F.; Weller, D. D.; Kobzik, L. AntisenseNucleic Acid Drug Dev. 1998, 8, 199.[132] Micklefield J.; Fettes, K. J. Tetrahedron 1998, 54 , 2129.

[133] Steinbeck, C.; Richert, C. J. Am. Chem. Soc. 1998, 120,11576.

[156] Neilsen, P. E.; Egholm, M.; Berg, R. H.; Buchardt, O.Science 1991, 254, 1497.

[134] Benner, S. A. et al. Pure & Appl. Chem. 1998, 70 , 263. [157] Egholm, M.; Buchardt, O.; Christensen, L.; Behrens, C.;Freier, S. M.; Driver, D. A.; Berg, R. H.; Kim, S. K.;Nordén, B.; Neilsen, P. E. Nature 1993, 365, 556.[135] Eschgfaller, B.; könig, M.; Boess, F.; Boelstreli, U. A.;

Benner, S. A. J. Med. Chem. 1998, 41 , 276.[158] Eriksson, M.; Nielsen, P. E. Quart. Rev. Biophys. 1996,

29 , 369.[136] Hyrup, B. Richerts, C.; Schulte–Herbrüggen, T.; Benner,S. A.; Egli, M. Nucleic Acids Res., 1995, 23 , 2427.

[159] Nielsen, P. E.; Haaima, G. Chem. Soc. Rev. 1997, 26 , 73.[137] Cuenoud, B.; Casset, F.; Hüsken, D.; Natt, F.; Wolf, R.

W.; Altmann, K.–H.; Martin, P.; Moser, H. E. Angew.Chem. Int. Ed. Engl. 1998, 37 , 1288.

[160] Hyrup, B.; Nielsen, P. E. Bioorg. Biomed. Chem. 1996,4, 5.

[138] Heystek, L. E.; Zhou, H.–Q.;Dande, P.; Gold, B. J. Am.Chem. Soc. 1998, 120, 12165.

[161] Nielsen, P. E. Perspectives in Drug Discovery andDesign 1996, 4, 76.

[139] Horn, T.; Chaturvedi, S.; Balasubramaniam, T. N.;Letsinger, R. L. Tetrahedron Lett. 1996, 37 , 743.

[162] Nielsen, P. E. Pure & Appl. Chem. 1998, 70 , 105.

[163] Nielsen, P. E. Acc. Chem. Res. 1999, 32 , 624.[140] Bailey, C. P.; Dagle, J. M.; Weeks, D. L. Nucleic Acids

Res. , 1998, 26 , 4860. [164] Brown, S. C.; Thomson, S. A.; Veal, J. M. Davis, D. G.Science 1994, 265, 777.

[141] De Mesmaeker, A.; Waldner, A.; Sanghvi, Y. S.;Lebreton, J. Bioorg. Med. Chem. Lett. 1994, 4, 395. [165] Eriksson, M.; Nielsen, P. E. Nat. Struct. Biol. 1996, 3,

410.[142] Hyrup, B.; Egholm, M.; Buchardt, O. Nielsen, P. E.

Bioorg. Med. Chem. Lett. 1996, 6, 1083. [166] Rasmussen, H.; Kastrup, J. S.; Nielsen, J. N.; Nielsen, J.M.; Nielsen, P. E. Nat. Struct. Biol. 1997, 4, 98.

[143] Dempcy, R. O.; Browne, K. A.; Bruice, T. C. J. Am. Chem.Soc. 1995, 117, 6140. [167] Kuhn, H.; Demidov, V. V.; Nielsen, P. E.; Frank–

Kamenetskii, M. D. J. Mol. Biol. 1999, 286, 1337.[144] Dempcy, R. O.; Browne, K. A. Bruice, T. C. Proc. Natl.

Acad. Sci. USA 1995, 92 , 6097. [168] Scarfi, S.; Giovine, M.; Gasparini, A.; Damonte, G.;Millo, E.; Pozzolini, M.; Benatti, U. FEBS Lett. 1999,451, 264.[145] Browne, K. A.; Dempcy, R. O.; Bruice, T. C. Proc. Natl.

Acad. Sci. USA 1995, 92 , 7051.[169] Prochiantz, A. Nat. Biotechnol. 1998, 16 , 819.


[170] Tyler, B. M.; McCormick, D. J.; Hoshall, C. V.; Douglas,C. L.; Jansen, K.; Lacy, B. W.; Cusack, B.; Richelson, E.FEBS Lett. 1998, 421, 280.

Bonham, M. A.; Au, K. G.; Carter, S. G.; Bruckenstein, D.A.; Boyd, A. L.; Noble, S. A.; Babiss, L. E.; Science,1992, 258, 1481.

[171] Tyler, B. M.; Jansen, K.; McCormick, D. J.; Douglas, C.L.; Boules, M.; Stewart, J. A.; Zhao, L.; Lacy, B.; Cusack,B.; Fauq, A.; Richelson, E. Proc. Natl. Acad. Sci. USA1999, 96 , 7053.

[176] Larsen, H. J.; Nielsen, P. E. Nucleic Acids Res. 1996, 24 ,458.

[177] Bastide, L.; Boehmer, P. E.; Villani, G.; Lebleu, B.Nucleic Acids Res. 1999, 27 , 551

[172] Good, L.; Nielsen, P. E. Nat. Biotechnol. 1998, 16 , 355.[178] Kurakin, A.; Larsen, H. J.; Nielsen, P. E. Chem. & Biol.

1998, 5, 81.[173] Good, L.; Nielsen, P. E. Proc. Natl. Acad. Sci. USA1998, 95 , 2073.

[179] Bentin, T.; Nielsen, P. E. Biochemistry 1996, 35 , 8863.[174] Taylor, R. W.; Chinnery, P. F.; Turnbull, D. M.;

Lightowlers, R. N. Nat. Genetics 1997, 15 , 212. [180] Boffa, L. C.; Morris, P. L.; Carpaneto, E. M.; Louissaint,M.; Allfrey, V. G. J. Biol. Chem. 1996, 271, 13228.

[175] Hanvey, J. C.; Peffer, N. J.; Bisi, J. E.; Thomson, S. A.;Cadilla, R.; Josey, J. A.; Ricca, D. J.; Hassman, C. F.;

Current Medicinal Chemistry 1157 Backbone Modification of ...€¦ · This has resulted in the...

Documents

Transcript of Current Medicinal Chemistry 1157 Backbone Modification of ...€¦ · This has resulted in the...