Intramolecular Free-Radical Cyclization Reactions on Pentose Sugars

From October 2004 to September 2010 Chuanzheng Zhou worked under the supervision of Prof. Jyoti Chattopadhyaya at the Chemical Biology Program, Biomed...
0 downloads 6 Views 9MB Size
Review pubs.acs.org/CR

Intramolecular Free-Radical Cyclization Reactions on Pentose Sugars for the Synthesis of Carba-LNA and Carba-ENA and the Application of Their Modified Oligonucleotides as Potential RNA Targeted Therapeutics Chuanzheng Zhou and Jyoti Chattopadhyaya* Chemical Biology Program, Department of Cell and Molecular Biology, Box 581, Biomedical Center, Uppsala University, SE-751 23 Uppsala, Sweden 8.2. α-L-Carba-LNA Modified AONs 9. Biological Evaluation of Antisense Oligonucleotides Containing carba-LNA Derivatives 10. RNAi Potency of siRNAs Containing Carba-LNA and Carba-ENA Modifications 11. Conclusions and Implications Author Information Corresponding Author Notes Biographies Acknowledgments References

CONTENTS 1. Introduction 2. Synthesis of Carba-ENA and Carba-LNA Nucleosides 2.1. Synthesis of Carba-ENA through Ring Closure Metathesis 2.2. Application of Intramolecular Free-Radical Cyclization on Constructing Bicyclic Nucleosides 2.3. Synthesis of Carba-LNA and Carba-ENA through Intramolecular Free-Radical Addition to CC 2.4. Synthesis of Carba-LNA and Carba-ENA through Intramolecular Free-Radical Addition to CN 2.5. Synthesis of Carba-LNA through Intramolecular Free-Radical Addition to CC 2.6. Synthesis of α- L -Carba-LNA Derivatives through Intramolecular Free-Radical Addition 2.7. Regio-Selectivity and Stereoselectivity of the Free-Radical Cyclization 3. Conformation of Carba-LNA and Carba-ENA Nucleosides 4. Orientation of Carbocyclic Moieties of Carba-LNA and α-L-Carba-LNA in Duplex Form 5. Thermal Stability of Modified AON/RNA Duplex 5.1. Carba-LNA Modification 5.2. α-L-Carba-LNA Modification 5.3. Carba-ENA Modification 6. RNA Selectivity of Modified AONs 7. Nuclease Resistance of Modified Oligonucleotides 7.1. Carba-LNA Modified Oligonucleotides 7.2. Carba-ENA Modified Oligonucleotides 7.3. α-L-Carba-LNA Modified Oligonucleotides 8. RNase H Elicitation of Carba-LNA and Carba-ENA Modified AONs 8.1. Carba-LNA and Carba-ENA Modified AONs © 2012 American Chemical Society

3808 3810

3825 3825 3826 3829 3830 3830 3830 3830 3830 3830

3810

3823 3823 3823 3824

1. INTRODUCTION Since 1950s, many different types of conformationally constrained nucleos(t)ides have been synthesized for structural studies or as potential antivirus agents.1−3 In the early 1990s, conformationally constrained nucleosides such as Bc-DNA (Figure 1) were incorporated into antisense oligonucleotides (AONs) with the aim to develop therapeutic AONs that effectively and specifically recognize target RNA with high affinity and selectivity.4 Altmann and Marquez subsequently showed that 4′,6′-methanocarba nucleotides5,6 having a Northtype (N-type) sugar pucker and 1′,6′-methanocarba nucleotides characterized by South-type (S-type) sugar pucker7 (Figure 1) have very different antiviral and antisense activities.8 4′,6′Methanocarba-T (N-type) showed excellent antiherpetic activity, but 1′,6′-methanocarba-T (S-type) was devoid of the activity. Incorporation of 4′,6′-methanocarba-T in DNA/RNA duplexes resulted in Tm increase (∼1.3 °C/modification), whereas 1′,6′-methanocarba-T modification induced a small decrease in Tm. This observation highlighted the importance of sugar conformation for RNA targeting, which in turn inspires an upsurge in the synthesis of sugar conformation constrained nucleotides for antisense therapeutics.9,10 An important compound that emerged during this upsurge is locked nucleic acid (LNA). The synthesis of LNA was reported first by Imanishi11 in 1997, whereas Wengel’s group12 reported an independent synthesis in 1998. In LNA, the 2′-O,4′-C-methylene- across the pentosesugar forms a fused five-membered ring (Figure 1), constraining the furanose sugar in a typical North conformation

3824 3824

Received: September 12, 2010 Published: April 24, 2012

3810

3812

3814 3814

3814 3815 3818 3818 3820 3820 3822 3822 3822

3808

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 1. The structures of early developed conformationally constrained nucleosides.

Figure 2. Structures of 2′,4′-locked nucleosides.

(phase angle P = 17.4°).11 Introduction of LNA to AONs results in unprecedented increases in the thermal stability of duplexes toward both DNA and RNA.13 Thermodynamic and structural studies suggest that LNA stabilizes the duplex by facilitating preorganization of AONs in RNA-like molecular structures and/or improving stacking.14,15 Short 2′,4′-linkage (up to eight atoms) always constrains the nucleos(t)ides into a typical North-type sugar conformation. Hence, many types of 2′,4′-locked nucleos(t)ides have been synthesized in the past decade and their structures are listed in Figure 2. LNA analogues such as 2′-thio-LNA16 (Figure 2B), 2′amino-LNA17−19 (Figure 2C−E), C3′-branched LNA20 (Figure 2F), C6′-branched LNA21,22 (Figure 2G,H), α-L-LNA,23 and 2′amino-α-L-LNA24,25 (Figure 2I−K) are 2′,4′-locked nucleosides containing a two-atom linkage. These nucleotides are constrained in a very similar conformation as that of LNA and also bestow similar high affinity toward complementary RNA. The pseudorotational phase angles (P) of 2′,4′-locked nucleos(t)ides containing a three-atom linkage such as 2′-C,4′-C-methyleneoxymethylene nucleoside26 (Figure 2N), 2′-O,4′-C-ethylenebridged nucleic acids (ENA)27 (Figure 2O), aza-ENA28−31 (Figure 2P), BNANC and derivatives32 (Figure 2Q), and N-Meaminooxy-BNA19 (Figure 2R) are also conformationally close to that of LNA. However, AONs modified with nucleotide containing 2′,4′ three-atom linkage generally exhibit slightly lower RNA affinity compared to those of AONs con-

taining 2′,4′ two-atom linkage nucleotide modification but higher selectivity toward target RNA compared to targeting DNA. Incorporation of the 2′,4′-locked nucleotide(s) containing a fouratom linkage such as BNAcoc33 (Figure 2S) and PrNA27 (Figure 2T) into AONs does not lead to an increase in the RNA affinity rendering these modifications unsuitable for modulation of AON or siRNA properties as a potential therapeutic or diagnostic. It is noteworthy that locked nucleos(t)ides described above have a heteroatom connected to C2′ of the sugar moiety; hence they are all classified as a group containing 2′,4′-heterocyclic linkages. Their carbocyclic analogues, locked nucleos(t)ides containing a 2′,4′ carbocyclic ring such as carba-LNA (Figure 2U), carba-ENA (Figure 2V), α-L-carba-LNA (Figure 2W) as well as their derivatives, have been synthesized in the past five years by our group34−37 and others.38−40 The striking feature of carba-LNA and carba-ENA derivatized AONs is that they render similar RNA affinity as AONs modified by incorporation of LNA and ENA but show very much improved nuclease resistance (survival in the blood serum for >48 h). The carbocyclic ring of carba-LNA and carba-ENA also allows for a range of substitutions thus providing an effective handle for engineering of new types of substitutions in the minor groove of AON/RNA or RNA/RNA duplexes to modulate important antisense or siRNA properties such as target RNA affinity, nuclease resistance, and delivery without impairing their RNase H or RISC recruitment capabilities. The present review focuses 3809

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 1. Synthesis of carba-ENA Derivatives through Ring-Closing Metathesis

Scheme 1) were first synthesized through the ring-closing metathesis method by Nielsen et al.38 The key intermediate 2 was obtained from uridine in 11 steps, and then it was treated with Grubbs' catalyst to furnish the unsaturated three-carbon 2′,4′-locked uridine 3 (unsaturated carba-ENA) in 96% yield. Carba-ENA uridine (6) was obtained by hydrogenation of the double bond of 3 following a known procedure.38 Treatment of unsaturated carba-ENA-U (3) with OsO4 gave the 6′,7′-di-OHcarba-ENA-U (4).39 Instead of this approach, a ring-closing enyne metathesis for compound 8 had been employed leading to the 6′-vinyl-unsaturated carba-ENA 9, which has further been transformed to C6′-hydroxymethyl-unsaturated carba-ENA-U (10) and C6′R/S-ethyl-carba-ENA-U (11).39

on the synthesis and structure of carba-LNA and carba-ENA as well as their derivatives that are decorated with various hydrophilic and hydrophobic groups at the C6′/C7′ position of the fused carbocylic ring in carba-LNA and C6′/C7′/C8′ position in carba-ENA. All carba-LNA derivatives and most of the carbaENAs, except for one case involving ring closure metathesis by Grubb’s catalyst, are synthesized through intramolecular freeradical cyclization reaction. The key synthetic and mechanistic features of the intramolecular free-radical cyclization reaction on the pentose sugar, developed in the Uppsala laboratory, will be briefly addressed. The recent advances in the therapeutic properties of AONs and siRNAs with these modifications will also be discussed. Readers can refer to older reviews2,3,41 about the synthesis and properties of other types of conformationally constrained nucleos(t)ides.

2.2. Application of Intramolecular Free-Radical Cyclization on Constructing Bicyclic Nucleosides

Ring closure of hex-5-enyl radical was first reported five decades ago by Lamb et al.42 It is an efficient approach for stereo- and regiocontrolled C−C bond formation, and its utility has been well recognized in natural product synthesis.43 In the early 1990s, this method was introduced by Chattopadhyaya and his co-workers to synthesize bicyclic nucleosides.44−48 The first reported bicyclic nucleoside 17 was obtained by treatment of the radical precursor 14 with Bu3SnH and AIBN in boiling benzene (Scheme 2).44 This treatment putatively generated

2. SYNTHESIS OF CARBA-ENA AND CARBA-LNA NUCLEOSIDES 2.1. Synthesis of Carba-ENA through Ring Closure Metathesis

The 2′,4′ carbocyclic rings in carba-LNA and carba-ENA are generally formed by coupling C2′ and C4′ tethered functions. Ring closure metathesis has been found to be very efficient for this C−C coupling. Carba-ENA series (compounds 3−7, 9−13, 3810

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 2. Synthesis of 2′,3′-Fused Bicyclic Nucleosides through Free-Radical Addition to CC

leading to a 3′,4′-lactone fused bicyclic nucleoside 33,48 which could be conveniently hydrolyzed under alkaline conditions to give 3′-Cbranched alkyl carboxylic acid 34. However, the stereochemistry of these two cyclizations is different given that C6′ of 31 is in S-configuration but 33 is in R-configuration. This suggests the stereoelectronic modulation on the linkage itself can significantly influence the stereoselectivity during free-radical cyclization. A 2′,3′lactone fused bicyclic nucleoside 36 has been subsequently synthesized from 35 using this method by Camarasa et al.50 Chattopadhayaya and co-workers reported that addition of C3′ radical to an alkynyl tethered to C2′ through a siloxane linkage (38, Scheme 5) is also efficient, generating a 2′,3′-fused five-membered ring containing siloxane with exocyclic methylene (CH2) function (39) through 5-exo cyclization mode.47 However, if trapping the C3′ radical using silicon-bearing allyl group (42), a 2′,3′-cis-fused seven-membered ring was formed (43). This indicates the radical cyclization took place in an unusual 7-endo mode in this case.45 Interestingly, by inverting the configuration of the C2′ tethered silicon-bearing allyl (as in 45, Scheme 5), a mixture (1/1) of cis and trans-fused sevenmembered rings (46) was obtained. The Si−O and Si−C bond in these siloxane-containing bicyclic nucleosides were cleaved by fluoride ion and by oxidation, respectively, affording Cbranched nucleosides 40, 44, 47, and 48. Thus, intramolecular cyclization of siloxane-bearing olefin or alkyne to a prochiral radical center, followed by opening the temporary silicon connection, represents an effective and wise strategy for synthesis of C-branched nucleosides. We envision further use of this intramolecular free-radical cyclization of siloxane-bearing olefin or alkyne for the synthesis of other C-branched carbohydrates. Following Chattopadhyaya’s work, many bicyclic carohydrates and branched nucleosides have been synthesized through freeradical cyclization strategy.51−53 Impressed by the stereo- and regiospecificity of the above free-radical cyclization reactions on the C2′ or C3′ chiral centers of the pentose ring in constructing these bicyclic nucleosides with predictable stereo outcome, in a preparative scale, vis-à-vis the biochemical success of Imanishi−Wengel’s LNA incorporated oligonucleotides, Chattopadhyaya realized that if the tethered olefin and the radical center are separated by a one-carbon bridge in the pentose ring of the nucleoside, free-radical cyclization will lead to a strained carbabicyclic nucleoside. This idea prompted the synthesis of carbaLNA and carba-ENA by free-radical cyclization, which we argued might open a new direction for new types of potential therapeutic

C2′ radical (16) and the addition of the radical to the CC of C3′-O-allyl occurred in the 5-exo mode, leading to a new 2′,3′α-fused heterocyclic five-membered ring. The pentose sugar of bicyclic nucleoside 17 was found in S-conformation. Instead, if starting from radical precursor 19, the obtained bicyclic nucleoside 21 exhibited N-type sugar puckering. Hence, it is not the 2′,3′-fused ring but the property of the 2′ or 3′ electronegative substituent that dictates the conformation of the pentose moiety. This is easy to understand given that (1) the fused ring is five-membered and so is flexible, and (2) electronegative substituent on C2′ or C3′ tends to occupy a pseudoaxial position because of the stabilizing gauche effect.49 If the 2′ (or 3′)-O-allyl lies in the β face in the radical precursor (23 and 25, Scheme 2), radical cyclization resulted in a bicyclic nucleoside with a 2′,3′-β-fused heterocyclic fivemembered ring (24 and 26).44 The 2′,3′-fused five-membered ring in the bicyclic nucleoside could also be efficiently formed by addition of the C2′ radical to the C3′-O-propynyl (Scheme 3).44 Scheme 3. Synthesis of 2′,3′-Fused Bicyclic Nucleosides through Free-Radical Addition to CC

This cyclization took place in the 5-exo mode, resulting in an exocyclic methlyene function. The success of this reaction was recently exploited for the synthesis of methylene-carba-LNA by Seth et al. (see Section 2.5).40 Another type of bicyclic nucleoside synthesized through freeradical cyclization is 3′,4′-β-fused six-membered bicyclic nucleosides 31 (Scheme 4).48 In the radical precursors 29, an allyl is attached to the 5′-O through an ether bond. After treatment with Bu3SnH and AIBN, intramolecular cyclization of the obtained hept-6-enyl radical (30) occurred exclusively in the 6-exo mode, formatting the 3′,4′-β-fused six-membered heterocyclic ring (31). The alkenyl function can also be introduced to the radical precursor through a bond other than an ether. Chattopadhyaya reported that after introduction of the allyl function to the 5′-O through an ester linkage, the obtained precursor (32, Scheme 4) was subjected to radical cyclization, 3811

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 4. Sythesis of Lactone Fused Bicyclic Nucleosides through Free-Radical Cyclization Reaction

Scheme 5. Synthesis of Siloxane-Containing Bicyclic Nucleosides through Free-Radical Cyclization Reaction

oligonucleotides, in which we should be able to create flexibility to steer required biophysical and biological properties.

Bu3SnH in toluene putatively gave radical intermediate 54. Freeradical cyclization, exclusively through a 5-exo pathway, produced 7′R/S-Me-carba-LNA-T (55) as a diastereoisomeric mixture. Through a similar strategy, 7′-Me-carba-LNA-5MeC/A/G have been successfully synthesized recently.54 On the other hand, subjecting 52 to another round of hydroboration−oxidation, Swern oxidation and Wittig reaction furnished 56 (Scheme 6), starting from which the radical precursor 57 was obtained in four steps. Treatment of 57 with AIBN and Bu3SnH in toluene led exclusively to 8′R-Me-carba-ENA nucleoside 59,34 putatively through 6-exo cyclization of radical intermediate 58. Starting from the aldehyde 50, C6′-modified carba-LNAs have also been synthesized.36 50 reacted with vinylmagnesium bromide to introduce 1S-hydroxy-allyl at C4′ as in compound

2.3. Synthesis of Carba-LNA and Carba-ENA through Intramolecular Free-Radical Addition to CC

The first carba-LNA and carba-ENA nucleosides synthesized in the Uppsala laboratory by the free radical approach are 7′R/SMe-carba-LNA-T (55, Scheme 6) and 8′R-Me-carba-ENA-T (59).34 The synthesis started from compound 49 whose primary alcohol was oxidized to aldehyde by Swern oxidation. The aldehyde was subjected to successive Wittig reaction, hydroboration−oxidation, Swern oxidation, and Wittig reaction, giving C4-allylated sugar 52, which was converted to radical precursor 53 in four steps. Treatment of compound 53 with AIBN and 3812

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 6. Synthesis of 7′-Me-carba-LNA and 8′-Me-carba-ENA through Free-Radical Cyclization to CC

Scheme 7. Synthesis of C6′ and C7′ Modified Carba-LNAs through Free-Radical Cyclization to CC

60 (Scheme 7A). Generation of C2′ radical (62) through radical precursor 61 followed by the 5-exo cyclization reaction led to a mixture of two main isomers, 6′S-OH-7′S-Me-carbaLNA-T (63) and 6′S-OH-7′R-Me-carba-LNA-T (64, 63:64 = 7:3) as well as about 10% of 6′-OH-carba-ENA-T (65) which was formed following the 6-endo cyclization pathway.35 The 6′-hydroxy group in 6′-OH-carba-ENA-T (65) was efficiently removed by radical deoxygenation resulting in parent carbaENA-T (66).35 Complete inversion of the configuration of

6′-OH in compound 64 was achieved by successive Dess− Martin oxidation and NaBH4 reduction, giving 6′R-OH-7′R-Mecarba-LNA-T (67, Scheme 7B). Similarly, oxidation the 6′-OH of 63 gave ketone 68. Reduction of the ketone by NaBH4 resulted in 6′R-OH-7′S-Me-carba-LNA-T (69). In 69, the 6′R-OH and 3′-OH are cis-orientated which allows coupling with thymidine-5′-phosphorodiamidite resulting in Sp and Rp-D2-CNA-T55−57 (70, Scheme 7C), which in fact are the first reported sugar and phosphate double locked nucleotides.58 3813

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 8. Synthesis of C6′ and C8′ Modified Carba-ENAs through Free-Radical Cyclization to CC

free-radical cyclization. The cyclization proceeded in the 6-exo cyclization mode to afford 8′R-benzyloxyamino-carba-ENA (102) as the sole product. Oxidative deamination of the 8′R-benzyloxyamino function of 102 was achieved by treatment with 3,5-di-tert-butyl-1,2benzoquinone, furnishing ketone 103. Starting from 103, 8′OH-carba-ENA 104 and 105, 8′-Me-carba-ENA 107, and parent carba-ENA 66 have been synthesized (Scheme 11B,C).61

Other reported carbocyclic nucleosides 6′S-OH-6′-Me-7′SMe-carba-LNA-T (71) and 6′R-OH-6′-Me-7′S-Me-carba-LNA-T (72)36 have been synthesized using the reaction of ketone 68 with methyl magnesium iodide (Scheme 7C). The 6′-OH group in nucleoside 71 was removed by radical deoxygenation reaction, giving 6′R/S-Me-7′S-Me-carba-LNA-T (73) as a diasteroisomeric mixture together with an unexpected bicyclo[2.2.1]2′,6′-methylene-bridged hexopyranosyl nucleotide (BHNA, 74), which was found to be formed by a radical rearrangement.59,60 C6′-modified carba-ENAs also have been synthesized by addition of the C2′ radical to C4′ tethered 1S-hydroxy-butenyl function as in 77 (Scheme 8).36 The radical cyclization reaction took place exclusively by the 6-exo pathway, giving 6′S-OH-8′RMe-carba-ENA-T (78). After oxidation the 6′-OH of 78, the resulting ketone 79 was transformed to 6′R-OH-8′R-Me-carbaENA-T (80), 6′R-OH-6′-Me-8′R-Me-carba-ENA-T (81), and 6′R/S-Me-8′R-Me-carba-ENA-T (82a/b), respectively (Scheme 8).

2.5. Synthesis of Carba-LNA through Intramolecular Free-Radical Addition to CC

Another member of the carba-LNA family, methylene-carbaLNA, was synthesized through intramolecular addition of the C2′ radical to a C4′-tethered CC by Seth et al. recently.40 Such an intramolecular addition of the C2′ radical to a C3′-Otethered alkyne, generating a 2′,3′-fused five-membered heteroring with exocyclic olefin function (28, Scheme 3), has been established in 1991 in the Uppsala lab.44 The synthesis of methylene-carba-LNA started from 5′-TBDPS-3′-(2-methylnapthalene)-allo-furanose derivative 108 (Scheme 12). Swern oxidation of the primary alcohol followed by a Wittig reaction provided an olefin, which was subjected to hydroboration− oxidation using 9-BBN/sodium perborate to give alcohol 109 in good yield. A Swern oxidation of 109 followed by a Coery− Fuchs reaction generated dibromo olefin 110. After deprotection of the naphthyl group, treatment with with n-BuLi furnished alkyne 111, which was further converted to thiobicarbonate 112 in four steps. Treatment of 112 with AIBN and Bu3SnH in toluene putatively gave radical intermediates 113 followed by intramolecular free-radical cyclization to give methylene-carbaLNA 114 in satisfactory yield.

2.4. Synthesis of Carba-LNA and Carba-ENA through Intramolecular Free-Radical Addition to CN

Carba-LNA derivatives containing C7′ hydrophilic substitutions have been synthesized by Chattopadhyaya et al. through intramolecular free-radical addition to CN.37 The synthesis started from 49 (Scheme 9), which was converted to 83 in six steps. Then the C4′-tethered aldehyde reacted with O-benzylhydroxylamine, giving oxime 84, which was further transformed to radical precursor 85. After treatment with Bu3SnH and AIBN in toluene, the generated C2′ radical was trapped by the CN (86, Scheme 9). This radical cyclization also took place exclusively in the 5-exo pathway, giving 7′R-benzyloxyaminocarba-LNA (88) as the major products plus a trace amount of 7′S-benzyloxyamino-carba-LNA (87). The 7′-benzyloxyamino function of 87 was converted to ketone by mCPBA oxidation to 7′-oxime followed by Dess−Martin oxidation. Reduction of ketone 89 furnished 7′S-OH-carba-LNA (90), which was subjected to radical deoxygenation to give parent carba-LNA (91). Subjecting the radical precursor 94 (Scheme 10), which was obtained by introduction of a hydroxyl group to the C4′ tethered oxime as in 84, to radical cyclization afforded 96 and 97.37 They are carba-LNA derivatives containing hydrophilic modifications at both C6′ and C7′. Through intramolecular radical addition to CN, carbaENA derivatives containing C8′ hydrophilic modifications have also been synthesized.61 The synthesis started from 51 (Scheme 11A). It was first transformed to 98 in five steps. The C4′ tethered nitrile group of 98 was reduced with DIBALH to aldehyde followed by oximation with O-benzylhydroxylamine to provide O-benzyl oxime 99. After esterification of the 2′-OH of 99 with phenyl chlorothioformate, the obtained 100 was subjected to

2.6. Synthesis of α-L-Carba-LNA Derivatives through Intramolecular Free-Radical Addition

All of the carba-LNA and carba-ENA nucleos(t)ides discussed above are in β-D form. It has been reported that α-L-LNA, just like β-D-LNA, showed striking biochemical features,62−64 which prompted us to synthesize α-L-carba-LNA nucleos(t)ides65 (Scheme 13). Swern oxidation of the primary alcohol of 115 followed by Grignard reaction with vinylmagnesium bromide afforded C4-hydroxylallyl ribofuranose 116 and 117 as diastereomers. After transformation to thiocarbonate 118 and 119, respectively, they were subjected to free-radical cyclization by treating with Bu3SnH and AIBN. Cyclization of 118 occurred with high stereoselectivity to give 6′R-OH-7′S-Me-α-L-carbaENA-T (121) as the only product, but cyclization of 119 led to 6′S-OH-7′S-Me-α-L-carba-ENA-T (122) as the major product plus tetracyclic minor product 123 (122/123 = 3/1). The 6′OH in 122 was efficiently removed by radical deoxygenation reaction, giving 7′R-Me-α-L-carba-ENA-T (124) plus a minor 3814

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 9. Synthesis of C7′ Modified Carba-LNAs through Free-Radical Cyclization to CN

Scheme 10. Synthesis of C6′ and C7′ Modified Carba-LNAs through Free-Radical Cyclization to CN

such as O, N, ester to linkage generally has no effect on the cyclization mode, but Si has been found to change the radical cyclization mode in hept-6-CC radical cyclization as shown in Scheme 5. For five- and six-member ring formation, the ring is always cisfused. The face (α or β on the pentose sugar) of the newly formed ring is always the same as that of the configuration of the 4′tethered AB (any double or triple bond). The configuration of the radical center in the radical precursor seems to play no role in determining the stereoselectivity of the carbon−carbon bond formation. However, a larger (e.g., seven-member ring) ring formation on the pentose sugar may lead to both cis- and trans-fused products, as stated above (41 to 43 vs 45 to 46 in Scheme 5). It has been proposed that hex-5-enyl radical cyclization generally proceeds through a chairlike or boatlike transition state.68 Different C5 stereochemistry will be reached through different transition states (Scheme 16). The chairlike transition state is more stable than the boatlike transition state, but the calculated difference between the two states is very small (less than 1 kcal/mol). Hence, many other effectors may compensate for the formation of the normally unfavorable boatlike transition state. For example, a big substituent X at C4 (Scheme 16) may favor the boatlike transition because it eclipses one of the vinylic hydrogen atoms in the chairlike transition state.

bicyclo[2.2.1]-2′,6′-methylene-bridged hexopyranosyl nucleoside 125 which was supposed to be formed by a radical rearrangement mechanism.65 Similarly, radical deoxygenation of compound 123 furnished 6,7′-methylene bridged-α-L-carba-LNA-T (126). Another α-L-carba-LNA nucleoside, α-L-methylene-carbaLNA, has been synthesized by Seth et al. recently.66 The synthesis strategy is the same as synthesis of methylene-carbaLNA 114 except using the sugar 127 as the starting material (Scheme 14). The primary alcohol was first transformed to alkyne (128) in six steps, and then to radical precursor 129 in four steps. Treatment of 129 with Bu3SnH/AIBN in refluxing toluene provided α-L-methylene-carba-LNA (130). 2.7. Regio-Selectivity and Stereoselectivity of the Free-Radical Cyclization

It is known that intramolecular cyclization in lower alkenyl and alkynyl radicals and related species occurs preferentially in the exomode, giving the thermodynamically less stable exoproduct as the major product (Scheme 15).67 The more rapid exocyclization is rationalized by better orbital overlap in the exotransition state. Radical cyclization on nucleoside to synthesize the bicyclic nucleoside follows the general rule: hex-5-CC(CN or CC) and hept-6-CC(CN) radical cyclization takes place predominantly in exocyclization mode. Introducing heteroatom 3815

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 11. Synthesis of C8′ Modified Carba-ENA through Free-Radical Cyclization to CN

Scheme 12. Synthesis of Methylene-carba-LNA through Free-Radical Cyclization to CC

In all the radical precursors (the C2′-O-thiocarbonate) for carba-LNA and carba-ENA, their JH1′, H2′ and JH2′, H3′ vary from 5.5 to 6.5 Hz, suggesting their sugar rings are predominantly in South-type conformation. After radical generation, the sugar ring must go over a pseudorotational cycle to take up Northtype conformation with the 3′-OBn in the axial position, since the sugar is constrained in this conformation in the final product. To reduce the total energy of the system, overlapping of the SOMO of C2′ radical with the HOMO of C4′-tethered AB (any double or triple bond) leads to C−C formation. For radical cyclization to carba-LNA derivatives, either a chairlike (Scheme 17A) or boatlike (Scheme 17B) transition state could be formed. The former leads to the 2′,7′-trans product while the latter leads to the 2′,7′-cis product. Actually, the 2′,7′-cis product was identified as the major product (around 70%), so the major pathway for the cyclization here is through a naturally

unfavorable boatlike transition state. The observation that a bulky B (NOBn) led to the significantly increased 2′,7′-cis product (to 94%, Scheme 17) suggests it is the steric clash between B and the axial 3′-OBn that makes the chairlike state unfavorable. However, the clash between B and the C6′ substitution in the boatlike state is also unfavorable. It was indeed found in the presence of both bulky B (NOBn) and bulky C6′ substitution (-OAc) that the ratio of cis/trans product decreased to 1/1. Hence, the interaction of B with surrounding substitution plays a very important role in influencing the stereochemical outcome for synthesis of carba-LNA. In turn, we can modulate the stereochemical outcome of the cis or/and trans product by introducing the proper substituent to C3′, C6′, and C7′. For radical cyclization to carba-ENA derivatives, a chairlike transition state is favorable. In this transition state (Scheme 18), 3816

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 13. Synthesis of α-L-carba-LNA Derivatives through Free-Radical Cyclization to CC

Scheme 14. Synthesis of α-L-Methylene-carba-LNA through Free-Radical Cyclization to CC

Scheme 15. Intramolecular Radical Cyclization Occurs Preferentially in the Exo-Modea

a

Scheme 16. Transition States for the Hex-5-enyl FreeRadical Cyclization

Here, n ≤ 5, AB is any double or triple bond.

the 4′ and 6′ substitutents adopt pseudoequatorial positions and 3′-OBn adopts a pseudoaxial position. Equatorial orientation of the bulky exocyclic B in this transition state (Scheme 18A) is favorable because pseudoaxial orientation of B may increase the system energy by introducing 1,3-pseudodiaxial interaction between B and 3′-OBn (Scheme 18B). This explains why the 2′,8′-cis product (Scheme 18A) is the sole product obtained from hept-6enyl radical (58, 77) and hept-6-oxime radical (101) cyclization.

As shown in Scheme 4, similar radical precursors 29 and 32 lead to similar bicyclic nucleosides 31 and 33 but their C6′ configuration is opposite. This cannot be explained using steric clash. Instead, this stereoselectivity could be imposed by the constraining of the ester bond (Scheme 19). Radical intermediate 30 from 29 should adopt a chairlike transition state in which the CC adopts an equatorial position. Cyclization of 3817

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Scheme 17. Transition States of Hex-5-enyl (ynyl) Radical Cyclization To Afford 2′,7′-trans (A) and 2′,7′-cis (B) Products

Scheme 18. Transition States of Hept-6-enyl (ynyl, oxime) Radical Cyclization To Afford 2′,8′-cis (A) and 2′,8′-trans (B) Products

(Table 1). The phase angle of parent carba-ENA is 15°, which is very similar to that of ENA and aza-ENA.70 The presence of the CC bond in the carbocyclic ring decreases the constrained force on the sugar since the nucleosides 5, 9, and 10 have much higher phase angle values, 27°, 30°, and 30° respectively, than the parent carba-ENA. Nucleoside 4 exhibits an extremely small phase angle of 7°, suggesting 6′,7′-di-OH substitution in carba-ENA remarkably triggers the sugar toward North conformation. Other 6′ and/or 7′ modifications on carbaENA have a medium effect on sugar puckering and their phase angle P varies from 14° to 22°. Comparatively, substitution on the carbocyclic ring of carbaLNA nucleosides led to less effect on sugar puckering than substitution on the carba-ENA. The phase angles of parent and all modified carba-LNA nucleotides vary in a narrow range from 15° to 26°. The puckering amplitudes Φm describes the maximum outof-plane pucker.69 The puckering amplitudes of carba-ENAtype nucleosides, just like ENA and aza-ENA, are limited to a narrow range from 45° to 49°,34,36,61,70 whereas the puckering amplitude of various carba-LNA nucleosides varies from 54° to 58°.34,36,37,70 Hence, the pucker degrees of carba-LNA nucleosides are slightly higher than that of carba-ENA nucleosides. It seems that different substitutions on the carbocyclic ring of carba-LNA and carba-ENA only lead to a marginal effect on the puckering amplitude.

Scheme 19. Stereoelectronic Effects Dictate the Stereochemical Selectivity of Free-Radical Cyclization

30 leads to 31 with C6′ S-configuration. On the contrary, the rigid ester linkage in 32 enforces only a boatlike transition state. Again, to minimize the system energy, the CC should adopt an equatorial position. Cyclization leads to 33 with C6′ R-configuration. In summary, the stereochemical selectivity during synthesis of bicyclic nucleoside through radical cyclization was mainly controlled by the steric clash, but some other factors such as stereoelectronic effects may also play an important role when steric clash is insignificant.

3. CONFORMATION OF CARBA-LNA AND CARBA-ENA NUCLEOSIDES Sugar conformation of nucleosides can be characterized by the pseudorotation phase angles (P) and puckering amplitude (Φm).69 Phase angle P characterizes the sugar puckering mode. P = 0° corresponds to the 2′-exo, 3′-endo puckering mode (North) and P = 180° corresponds to the 2′-endo, 3′-exo puckering mode (South).69 All of the carba-LNA and carbaENA nucleosides are constrained in the North conformation

4. ORIENTATION OF CARBOCYCLIC MOIETIES OF CARBA-LNA AND α-L-CARBA-LNA IN DUPLEX FORM The models of carba-LNA modified DNA/RNA and α-carbaLNA modified DNA/RNA duplex have been built based on the published structures of LNA modified DNA/RNA71 and 3818

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Table 1. The Pseudorotational Phase Angles (P) and Puckering Amplitude (Φm)a of Carba-LNA and Carba-ENA Nucleosides

a

All the phase angle P and puckering amplitude Φm are calculated from structures obtained by ab initio (HF/6-31G**) optimization.

α-carba-LNA modified DNA/RNA duplex,72 respectively. As shown in Figure 3A, the carba-LNA modified DNA/RNA duplex adopts an A-like duplex structure which has a wide, shallow minor groove and a narrow, deep major groove. The carbocylic moiety of carba-LNA locates on the top of the minor groove. The α-L-carba-LNA modified DNA/RNA duplex

resembles a native DNA/RNA hybrid, adopting a geometry intermediate between that of A- and B-type duplex forms (Figure 3B). The carbocyclic moiety of α-L-carba-LNA lies at the bottom of the deep major groove. According to the models, it is likely that bulky substitution on carba-LNA will be better tolerated than on α-L-carba-LNA because the bulky 3819

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 3. Models of carba-LNA/RNA (A) and α-carba-LNA/RNA (B) duplex. Sequence of the DNA strand: 5′-CTGATATGC. All three T are modified by carba-LNA-T or α-carba-LNA-T but only one is highlighted on the model.

negative effect on Tm for the AON/RNA duplex, suggesting the 2′-oxygen of LNA plays an important role in the enhanced thermodynamic stability of the AON/RNA duplex. The crystal structure of DNA duplex containing one LNA modification showed that the 2′-oxa of LNA is engaged in the hydrogen bonding to several water molecules.74 It is conceivable that the hydration network is likely to reduce the electrostatic repulsion of the internucleosidic phosphates, and thereby contribute to an increase in the thermodynamic stability for AON/RNA duplex (which in turn can have negative implications on the nuclease stability; see discussion below). Just like the parent carba-LNA itself, all C6′ and/or C7′ substituted carba-LNA modified AON/RNAs exhibit remarkably higher Tm than native duplex. However, it was found that substitution at the C6′ and/or C7′ of carba-LNA renders different effects on the thermodynamic stability of the AON/ RNA duplex depending both on the nature (hydrophobic versus hydrophilic) and stereochemical orientation of the substituent.36,37 Compared with parent carba-LNA, both the hydrophobic 7′-Me and hydrophilic 7′-NH2 substitution on carba-LNA result in Tm decrease and the extent of the decrease depends on the configuration. When the 7′-Me and 7′-NH2 substituents point to the vicinal 3′-phosphate (compounds 64, 67 and 87, 97), they cause a decrease of the Tm by about 2−3 °C compared to the parent carba-LNA modified AON/RNA, but when the 7′ substituent points away from the vicinal 3′-phosphate (compounds 63, 69 and 88, 96), a much less pronounced destabilization effect (around −1 °C) is observed.36,37 The Tm decrease caused by C7′ substitution could be explained by the influence of (A) steric hindrance, and/or (B) hydration effects.

group renders less steric clash in the wider, shallower minor groove. Hydrophobic substitution will lead to a more negative effect on carba-LNA than α-carba-LNA since it is known that the hydration network in the minor groove is important for the stability of the DNA/RNA duplex.73 Comparatively, C7′ is closer to the bottom of the groove than C6′ for both carba-LNA and α-L-carba-ENA. The models provide the basis for understanding how different substitutions on the carbocyclic moiety modulate the properties of modified oligonucleotides.

5. THERMAL STABILITY OF MODIFIED AON/RNA DUPLEX 5.1. Carba-LNA Modification

Carba-LNA-type thymidines that have been synthesized in Chattoapdhyaya’s laboratory have been incorporated as mono substitution but at four different sites in a 15mer AON sequence, 5′-d (CTT CAT TTT TTC TTC) (T denotes the modification sites). Relative thermal stability of the modified AON/RNA duplexes are compared with the native counterpart to give ΔTm which was calculated by subtraction of the duplex melting temperatures (Tm) of native AON/DNA from the Tm of the modified AON/DNA. ΔTm listed in Figure 4A for each modification is an average ΔTm value of four AONs with the modification at four different sites. In the 15mer AON sequences 5′-d (CTT CAT TTT TTC TTC), one LNA-T modification leads to Tm increase of about 4.5 °C for AON/RNA duplex, whereas one parent carba-LNAT (91) modification only leads to 3.6 °C increase in Tm.37 Thus, replacement of 2′-O- with 2′-CH2- function results in a 3820

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 4. Comparison of (A) thermal stability: [ΔTm (ΔTm = Tm of modified AON/RNA hybrid − Tm of native AON/RNA hybrid)], (B) RNA selectivity: [ΔΔTm (ΔΔTm = ΔTm of AON/RNA hybrid − ΔTm of AON/DNA duplex)], and (C) nuclease stability [t1/2 of carba-LNA and carbaENA modified AONs upon SVPDE digestion] of AONs containing different modifications. AON sequence is 5′-d (CTT CAT TTT TTC TTC); T denotes the carba-LNA and carba-ENA modification sites within the modified AONs.

groove, thereby contributing to the energetic stabilization or destabilization by promoting or perturbing the electrostatic interactions. Since hydrophilic C7′-NH2 substitution results in a more significant destabilization effect on Tm than hydrophobic C7′-Me substitution when they have the same steric orientation,37 the relative contribution of steric clash versus the perturbed hydration is apparently difficult to dissect.

On the one hand, the C7′ substituent could render a steric clash in the minor groove of the AON/RNA duplex, leading to slight induced-expansion of the minor groove to accommodate bulky substituents, which results in perturbation of local or global conformation with an energy penalty at the cost of reduction of the thermal stability. On the other hand, the C7′ substituent could potentially influence the hydration network in the minor 3821

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

sequence leads to Tm increase by 4.6 °C/modification. Hence, the destabilization effect caused by substituted α-L-carba-LNA should be attributed to C7′ substitution. Comparatively, C7′ hydrophobic substitution on α-L-carba-LNA seems much more harmful to the thermodynamic stability of the AON/RNA duplex than C7′ substitution on carba-LNA. It is conceivable that the C7′ substitution on α-L-carba-LNA, which locates in the bottom of a narrow, deep major groove, creates a more significant steric clash and hydration perturbation than C7′ substitution on carba-LNA that lies in the wide, shallow minor groove of the AON/RNA duplex.

The RNA affinity of methylene-carba-LNA modified AON has been studied by Seth et al. using a 12mer DNA sequence d(GCGTTUTTTGCT) (U denotes the methylene-carba-LNA-U modification).40 Methylene-carba-LNA modified AON/RNA showed similar thermal stability as the LNA modified counterpart. Hence, methylene modification at C7′ on carbaLNA seems better than both hydrophobic Me group and hydrophilic amino group. One obvious reason is methylene substitution renders less steric clash. Another explanation comes from the crystal structure of the duplex:40 it was found that the methylene group participates in the formation of a CH···O type hydrogen bond with the O4′ of the 5-terminal nucleotides of a neighboring duplex. Hence, the methylene-carba-LNA moiety retains a negative polarization in the minor groove, and the disruption of hydration network by methylene substitution is minimal compared to other modifications. Using the same sequence, they also found the RNA affinity of both 7′R-Me-carba-LNA and 7′S-carba-LNA is less than LNA. Comparatively, 7′R-Mecarba-LNA is higher than 7′S-carba-LNA.40 This result is consistent with the above conclusion obtained by Chattopadhyaya from the 15mer AON sequence.36 On the contrary, both the hydrophilic hydroxyl group and hydrophobic methyl group substitution at C6′ of carba-LNA can stabilize the duplex regardless of their orientation.36 For example, both 6′S-Me-7′S-Me-carba-LNA (63) and 6′R-Me-7′SMe-carba-LNA (69) modified AONs showed slightly higher RNA affinity than the parent carba-LNA modified counterpart. Though 6′S-OH, Me-7′S-Me-carba-LNA (71) and 6′R-OH, Me7′S-Me-carba-LNA (72) have opposite C6′ configuration, they have the same affinity toward complementary RNA, which is even as good as LNA. Methylation of C6′ on LNA (cEt) has also been found to result in no negative effect on RNA affinity no matter what the configuration of the C6′-Me is.22 Obviously, a C6′ substituent of carba-LNA is located at the edge of the minor groove, thereby resulting in much less steric clash compared with C7′ substituents, but how the C6′ substituents render stabilizing effect on AON/RNA duplex needs further elucidation. A practical conclusion that can be drawn from these observations36 is that introduction of modification at C6′ can be used as an efficient strategy to fine-tune the electrostatics of the backbone in order to further develop carba-LNA nucleotides with high affinity toward target RNA.

5.3. Carba-ENA Modification

Carba-ENA nucleosides that have been synthesized in Chattopadhyaya’s laboratory34,36,61 have also been incorporated as mono substitutions at four different sites of the 15mer AON [5′-d (CTT CAT TTT TTC TTC) where T denotes the sites of modifications]. Generally speaking, all AONs incorporated with the carba-ENA derivatives show slightly higher RNA affinity (Figure 4A) than the native AON but much lower DNA affinity compared to carba-LNA modified counterparts. One parent carba-ENA-T modification (66) increases the thermodynamic stability of the AON/RNA duplex on average by 1.4 °C/modification.61 Further substitution on the C6′ and/ or C8′ of the carbocyclic linkage of carba-ENA only slightly destabilizes (down to −1 °C/modification) or stabilizes (up to +1 °C/modification) the AON/RNA duplex depending mostly on the substitution site (C6′ or C8′) and orientation of the substituents but not so much on its nature (hydrophobic Me versus hydrophilic OH or NH2). For example, when C8′-OH or C8′-Me points at the 3′-phosphate (105 and 107, respectively), they lead to a more significant destabilization effect than when they point away from the 3′-phosphate (104 and 59).61 On the other hand, C6′-OH or C6′-Me substation leads to a slight stabilization effect when they point toward the 3′-phosphate, whereas when they point away from the 3′-phosphate, they exert an obvious destabilization effect (78 vs 80, 82a vs 82b).36 Carba-ENA nucleosides 3, 4, 6, 9, 10, 1138,39 have been incorporated as mono or triple substitution in 9mer AON sequences: 5′-d (GTG ATA TGC) (T denotes the modification sites). One parent carba-ENA-U (6) modification in this 9mer AON resulted in Tm increase by 4.0 °C/modification for the AON/RNA duplex.38 In view of the fact that one parent carbaENA-T (66) modification in the 15mer AON sequence only led to a 1.4 °C increase in Tm, we can clearly see how significant the Tm is sequence-dependent. C6′,C7′-di-OH substitution on carba-ENA-U (4) can further slightly increase the Tm (+0.5 to +1 °C/modification).39 On the contrary, introduction of the CC double bond on the carbocyclic ring of carba-ENA-U (3) leads to a slight decrease of the Tm (−0.5 to −1 °C/ modification) compared to parent carba-ENA.38

5.2. α-L-Carba-LNA Modification

Unlike C7′ and/or C6′ substituted carba-LNA modified AON/ RNAs which exhibit higher Tm than native duplex, C7′ and/or C6′ substituted α-L-carba-LNA modification in the 15mer AONs leads to Tm decrease for the AON/RNA duplex (Figure 4A).65 Though 6′R-OH-7′S-Me-α-L-carba-LNA (121) and 6′S-OH7′S-Me-α-L-carba-LNA (122) have an additional 6′-OH compared with 7′-Me-α-L-carba-LNA (124), AONs containing one 121 or 122 modification exhibit the same RNA affinity as the 124 modified counterpart, around −3 °C/modification. This observation is consistent with the results obtained by Seth et al. recently that 6′-Me-α-L-LNA exhibited the same RNA affinity as α-L-LNA and LNA.75 These results reveal that the C6′ substituent on α-L-carba-LNA and α-L-LNA which lies along the edge of the major groove in the AON/RNA hybrid does not impair RNA affinity. One 7′-methylene-α-L-carba-LNA modification in the 12mer DNA sequence d(GCGTTUTTTGCT) leads to a 10 °C decrease in Tm for the AON/RNA duplex compared to the native duplex.66 However, α-L-LNA modification in the same

6. RNA SELECTIVITY OF MODIFIED AONS For all of the carba-LNA, α-L-carba-LNA, and carba-ENA modified AON/DNA duplexes, the Tm is lower than that of the modified AON/RNA hybrid,34,36,37,61,65 so all of them are RNA selective. The magnitude of RNA selectivity (denoted by ΔΔTm, ΔΔTm = ΔTm of AON/RNA − ΔTm of AON/DNA duplex) are compared in Figure 4B. The 7′-NH2 substitution in carba-LNA (87 and 88) and 8′NH2 substitution in carba-ENA (102) significantly reduce the magnitude of RNA selectivity (Figure 4B).37,61 Presumably, the 7′-NH2 and 8′-NH2 group can reduce the repulsion of two 3822

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

nuclease resistance, which highlights that the substituent on the carbocyclic ring of carba-LNA can significantly modulate its nuclease resistance. Replacing the 2′-O in LNA with 2′-CH2, the obtained parent carba-LNA was found to be 60 times more stable toward SVPDE treatment than LNA.37 All C6′ and/or C7′ substituted carba-LNA except C6′R-OH-carba-LNA (compounds 67 and 69) modified AONs are more stable than the parent carba-LNA modified counterpart, suggesting the steric clash imposed by the substitution in the carbocyclic ring contributes greatly to the improved nuclease resistance. Substituent C6′R-OH decreases the stability,36,81 but C6′S-OH substitution, which points to the 5′-phosphate, can significantly improve the nuclease resistance. 36 It is conceivable that C6′R-OH substitution can assist the departure of the 3′-O− anion during SVPDE digestion through intramolecular hydrogen bond formation, whereas C6′S-OH substitution retards the binding of nuclease to the phosphate. Hydrophobic C6′-methyl substitution renders the best effect on nuclease resistance for carba-LNA and carba-ENA modified AON.36 Both hydrophobic (methyl) and hydrophilic (amino) substitution at C7′ of carba-LNA imparts a positive effect on nuclease stability with the C7′-Me-carba-LNA (55) being in this respect slightly better than C7′-NH2-carba-LNA (87 and 88).37 Generally for carba-LNA, C7′ substitution leads to a less significant effect on nuclease stability than substitution at C6′ when the substituents are the same. Unlike the substitution at C6′, the orientation of substituent at C7′ also seems to be not so important in view that C7′S-NH2 (87) and C7′S-Me (69) exert similar effect as C7′R-NH2 (88) and C7′R-Me (67) substitutions, respectively.37 To address why carba-LNA modified AONs exhibit considerably improved 3′-exonuclease stability than the LNA modified counterpart, a Michaelis−Menten kinetic analysis of SVPDE digestion of model dinucleotides 131−135 has been carried out by Chattopadhyaya and Zhou.81 The results of these experiments had shown that 7′S-Me-carba-LNA modified dimers 133 and 135 were 620- and 330-fold more stable respectively than native dimer 131, whereas LNA modified dimers 132 and 134 were only 6- and 50-fold more stable, respectively, than the native dimer 131 (Figure 5). Hence, 7′SMe-carba-LNA modified dimers are much more nuclease resistant than the LNA modified counterpart by ∼100 fold. A pKa measurement showed that the pKa of 3′-OH (13.53) of 7′SMe-carba-LNA is 1.4 unit higher than that of 3′-OH of LNA (12.10).81 Since the scission of the 3′O−P bond has been proven to be the rate limiting step upon SVPDE digestion,81 it seems it is the relatively lower acidity of the of 3′-OH of 7′SMe-carba-LNA that makes the internucleotide phosphate of 133 more nuclease resistant (decreased Kcat) than LNA modified counterpart 132. On the contrary, the improved stability of internucleotide phosphate for 3′-end modified dimer 135 compared with that of dimer 134 comes not from the decreased Kcat but increased KM, suggesting the 7′S-Me-carbaLNA modification retards binding of 3′-exonuclease to the 5′-OH linked phosphate more remarkably than LNA.81 In summary, both an electrostatic effect and steric clash contribute to the improved nuclease resistance for carba-LNA modified antisense oligonucleotides.

strands by neutralizing the negative charge on the phosphate backbone to a higher extent in the AON/DNA duplex than in the AON/RNA duplex because the minor groove of the former is much narrower than the latter. For Me and OH modified carba-LNA and carba-ENA, the RNA selectivity was found not to be significantly dependent on the chemical nature and orientation of substituent on the carbocyclic moiety. Generally, carba-ENA derivatives significantly decrease the Tm of AON/DNA (∼ −4 to −5 °C/modification). As a result, the RNA selectivity (ΔΔTm) of carba-ENA derivatives modified AONs is around 5 °C/modification.36,61 On the contrary, carba-LNA derivatives only slightly increase or decrease the Tm of AON/DNA, leading to medium RNA selectivity, around 3−4 °C/modification.36,37 Hence, carba-ENA-type modified AONs show lower RNA binding affinity but better RNA selectivity than carba-LNA-type modified AONs.34,36 It is well-known that a typical B-type DNA/DNA duplex has a narrower minor groove (ca. 5−6 Å) than its DNA/RNA hybrid (ca. 9−10 Å).76 Thus, the larger six-membered carbocyclic ring in carba-ENA may produce larger steric perturbations in the narrow minor groove of DNA/DNA than the relatively smaller five-membered carbocyclic ring of carba-LNA. This may and in fact has been shown to result in an overall larger destabilization effect for carba-ENA nucleosides modified AON/DNA duplexes compared to that of carba-LNA nucleosides modified AON/DNA duplexes. On the other hand, the AON/RNA hybrid has a relatively wider minor groove, and the steric clash bestowed by carba-ENA and carba-LNA lead to much small differences in the thermal stability of AON/RNA than AON/ DNA. The overall steric effect results in better RNA selectivity of carba-ENA over carba-LNA. α-L-Carba-LNA nucleosides modified AONs are also RNA selective,65 but the magnitude of their RNA selectivity (2−3 °C/ modification) is much lower than carba-LNA-type and carbaENA-type modified AONs (Figure 4B).

7. NUCLEASE RESISTANCE OF MODIFIED OLIGONUCLEOTIDES Therapeutic oligonucleotides should be stable enough during delivery and after internalization of cells to fulfill their role. Unmodified AONs (i.e., single-stranded native DNA) are very labile toward intracellular and extracellular nucleases.77 It has been demonstrated that the nuclease resistance of AON correlates well with the magnitude and duration of the gene silencing effect.77 Hence, satisfactory nuclease resistance is a requirement for good pharmacokinetics of modified AON in antisense technology.78 It was found that the predominate nuclease activity in human blood serum is 3′ exonuclease.79 Chattopadhyaya et al. have compared the nuclease stabilities of different carba-LNA nucleosides and carba-ENA nucleosides modified AON sequences, 5′-d (CTT CAT TTT TTC TTC) (T denotes the modification), by treating them with 3′ exonuclease, SVPDE under identical conditions.36,37,61 Their stabilities (t1/2) toward SVPDE treatment are compared in Figure 4C. 7.1. Carba-LNA Modified Oligonucleotides

It has been reported that LNA modification only slightly increases80 the nuclease resistance for AON. All of the carbaLNA modified AONs are significantly more stable than the LNA-modified counterpart. This is a striking merit of carbaLNA compared to standard LNA. Moreover, AONs containing different carba-LNA derivatives show completely different

7.2. Carba-ENA Modified Oligonucleotides

As shown in Figure 4C, parent carba-ENA (66) modified AON was found to be around 15 times more stable than parent 3823

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 5. Comparison of Michaelis−Menten parameters of digestion of native and 7′S-Me-carba-LNA and LNA modified dinucleotides monophosphates 131−135 by 3′-exonuclease, SVPDE.

carba-LNA (91) modified AON toward SVPDE treatment.61 This can be easily understood in terms of their relative hydrophobic character: Propylene linkage between C2′ and C4′ positions in carba-ENA is one carbon longer compared to the ethylene linkage in carba-LNA. Thus, the propylene linkage in carba-ENA is more hydrophobic than carba-LNA and also may render a more pronounced steric hindrance to prevent nuclease binding and attacking on the vicinal phosphodiester linkage. But unlike carba-LNA, neither hydrophobic nor hydrophilic substitution at the C6′ and/or C8′ of carba-ENA renders a significant effect on the nuclease resistance: C6′-Me substitution on carba-ENA (82) only results in a 2-fold increase in the stability;36 C8′-Me substitution (59 and 107) renders no obvious effect; hydrophilic substitution (OH, NH2) at C6′ and C8′ even leads to slightly decreased nuclease resistance.36,61

H1, and (2) the cleavage patterns are all very similar, independent of the nature of the modification type. Generally, as a result of modification in the AON strand, the cleavage of AON/RNA hybrid duplex by RNase H is suppressed within a 5−6 base pairs long region toward the 3′-end of the RNA strand starting from the base opposite to the modification site in the AON strand of the duplex. This is presumably because of the fact that the 5−6 basepairs from the site of modification with North-sugar constrained nucleotides take up a local RNA/ RNA type conformation.71 This local alteration of the conformation in the modified AON/RNA hybrid cannot be recognized and cleaved by RNase H. For carba-LNA and carba-ENA derivatives modified AON/ RNA, the RNase H digestion rates are however dependent on the site of modification within the AON strand.36 Generally, single modification at the 3′-end or 5′-end of AON strands show relatively higher RNase H recruitment ability compared to those containing modification in the middle.36 The RNase H recruitment of AON (5′-d (CTT CAT TTT TTC TTC, T denotes modification site) containing different carba-LNA and carba-ENA nucleosides modifications are compared in Figure 6. It seems that different types of carbaLNA and carba-ENA modifications in the AON strand can slightly influence the digestion rate. LNA modified AON/RNA was digested by RNase H with a slower digestion rate than digestion of native AON/RNA, but both parent carba-LNA (91) and parent carba-ENA (66) modified AON/RNA duplexes have been found to undergo digestion faster than the native counterpart.37,61 7′S-NH2-carba-LNA (87) and 8′RNH2-carba-ENA (102) modification in AON strands have also been found to accelerate the digestion of the complementary RNA strand in the AON/RNA hybrid. However, when the C7′NH2 substituent points away from the 3′ phosphate like in 7′RNH2-carba-LNA (88) and 6′S-OH-7′S-NH2-carba-LNA (96), AONs with these modifications show the lowest RNase H recruitment ability.61 Other types of substituted carba-LNA and carba-ENA generally render an insignificant effect on RNase H recruitment efficiency.36,37,61

7.3. α-L-Carba-LNA Modified Oligonucleotides

Upon SVPDE treatment, α-L-LNA modified AON is more stable than the LNA modified counterpart and the former is less stable than α-L-carba-LNA modified AONs.65 However, the nuclease resistance of modified α-L-carba-LNA is not as good as carba-LNA and carba-ENA derivatives, indicating the C6′ and/ or C7′ substitutions on α-L-carba-LNA lead to a less positive effect on the nuclease resistance than C6′ and/or C7′ substitutions on carba-LNA nucleotides.

8. RNASE H ELICITATION OF CARBA-LNA AND CARBA-ENA MODIFIED AONS 8.1. Carba-LNA and Carba-ENA Modified AONs

In antisense strategy, RNase H recruitment is a very important approach leading to target RNA degradation. Digestion of carba-LNA and carba-ENA modified AON/RNA duplex by Escherichia coli RNase H1 has been evaluated.36,37,61,65 It should be noted that the activity obtained by E. coli RNase H1 may not reveal the real activity of RNase H elicitation in human though E. coli RNase H1 have similar nuclease properties as human RNase H.82 It was found the carba-LNA-type and carba-ENAtype modified AONs are (1) good substrates for E. coli RNase 3824

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 6. Comparison of RNase H digestion rates of AON/RNA heteroduplexes containing different types of modifications. AON sequence is 5′-d (CTT CAT TTT TTC TTC) and T denotes the modification site. Dotted line indicates the rate of RNase H digestion of native AON/RNA hybrids.

8.2. α-L-Carba-LNA Modified AONs

hydrolysis of α-L-LNA and 7′R-Me-α-L-carba-LNA (124) modified AON/RNA65 are much higher than digestion of the native AON/RNA hybrid and ∼2 times higher than that of LNA and 7′-Me-carba-LNA (55) respectively, suggesting that the locked ring of LNA and carba-LNA that lies in the minor groove of AON/RNA hybrid may impart RNase H elicitation, but it is not the case for α-L-LNA and α-L-carba-LNA-type whose locked ring is located in the major groove. The reaction rates of RNase H-mediated hydrolysis of 6′S-OH-7′S-Me-α-Lcarba-LNA (122) and 6′R-OH-7′S-Me-α-L-carba-LNA (121) modified AON/RNA have been found to be slightly lower than that of the native counterpart.

Wengel et al. have reported that an AON sequence d(CACACTCAATA) fully modified by α-L-LNA can induce RNase H-mediated cleavage of complementary RNA, albeit with an efficiency that is substantially reduced compared with native DNA/RNA hybrid.23 A structural study based on the NMR showed that α-L-LNA modification does not alter the minor groove of AON/RNA too much, which provides a reasonable explanation for the observation given the fact that RNase H binds the AON/RNA duplex in the minor groove.83 However, Chattopadhyaya et al. found a α-L-LNA or α-L-carbaLNA modification in the 15mer AON sequence d(CTT CAT TTT TTC TTC) leads to RNase H activity suppressed within a 5−6 base pairs long region,65 just in the same way as for LNA and carba-LNA derivatives. It is conceivable that full modification of this AON with α-L-LNA will completely abolish RNase H recruitment. Given that the concentration of RNase H used in Wengel’s study is more than 100-fold higher than that Chattopadhyaya et al. used, the above contradiction could be attributed to the different enzyme concentrations used in the two experiments. If this is true, α-L-LNA modification in AON/ RNA does not abolish the RNase H elicitation but needs more enzyme for efficient RNA degradation. However, since the RNase H concentration in the actual biological system is supposed to be extremely low, heavy or full α-L-LNA modification should be avoided while designing therapeutic AON. The RNase H recruitment of AON sequences (5′-d (CTT CAT TTT TTC TTC, T denotes modification site) containing different α-L-carba-LNA derivatives modifications have been studied by Chattopadhyaya et al.65 Here, their RNase H recruitment ability is compared with that of carba-LNA and carba-ENA (Figure 6). The reaction rates of RNase H-mediated

9. BIOLOGICAL EVALUATION OF ANTISENSE OLIGONUCLEOTIDES CONTAINING CARBA-LNA DERIVATIVES Though more than 30 carba-LNA and carba-ENA derivatives have been synthesized, only a few including methylenecarba-LNA, 7′-Me-carba-LNA, and 8′-Me-carba-ENA have been subjected to biological evaluation for antisense and RNAi potency. The biological property of AONs containing methylenecarba-LNA and 7′R-Me-carba-LNA has been studied by Seth et al. recently.40 They designed two AONs (14mer and 18mer; see Figure 7) to target mouse PTEN mRNA, and two modified residues were introduced to both the 5′ end and 3′ end of these AONs (Figure 7). Down regulation of PTEN mRNA was first studied in brain endothelial cells. For the 14mer AONs, it was found that the activity follows this rank: LNA (IC50 = 2.8 μM) > methylene-carba-LNA (3.9 μM) > 7′R-Me-carba-LNA (7.9 μM). These activities just parallel the Tm's of corresponding AON/ RNA hybrids. The 18mer AONs were less active than the 3825

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 7. Targeting mouse PTEN mRNA by methylene-carba-LNA, 7′R-Me-carba-LNA, and LNA modified 14mer and 18mer AONs. Underlined letters indicate modified nucleosides. All internucleosidic linkages are phosphorothioate; IC50 values were determined in brain endothelial cells using electroporation.

linkages have been also prepared. It was found that PSsubstituted AON VII, VIII, IX still achieved potent and selective inhibition of HTT, but the difference between 7′-Me-carba-LNA modification and LNA, cEt became smaller. Just as expected, all of the PS-AONs displayed some toxicity. The expansion of a CAG repeat is also involved in other diseases such as Spinocerebellar ataxi-3. The CAG repeat in ATX3 gene is typically less than 31. Individuals with more than 52 repeats show full penetrance. Inhibition of WT and mutant ATX3 by 7′-Me-carba-LNA modified AON IV and cEt modified III (Figure 8) has also been studied by Corey et al.86 In contrast to allele-selective inhibition of HTT, both AON IV and III showed potent but nonselective inhibition of WT and mutant ATX3.

14 mer AONs. Comparatively, the methylene-carba-LNA modified 18mer (D1) was more potent than LNA (D3) and 7′R-Mecarba-LNA (D2) modified ones (Figure 7). The potency of the same AONs has also been tested in animal experiments. Dose-dependent down regulation of PTEN mRNA in liver tissue was observed for all the modified AONs.40 The methylene-carba-LNA modified 14mer and 18mer showed ED50 of 4.9 mg/kg and 4 mg/kg, respectively, which is very similar to that of LNA modified counterparts (Figure 7). 7′R-Me-carba-LNA was less potent. A very interesting observation is that at a high dose (e.g., 15 mg/kg), LNA modified AONs were found to be highly toxic, but 7′R-Mecarba-LNA and methylene-carba-LNA modified ones only showed modest toxicity. It is noteworthy that all of the AONs used in this work contain phosphorothioate (PS) internucleosidic linkages. Hence, the results cannot reveal the effect of better nuclease resistance of carba-LNA analogues compared to LNA. The expansion of a CAG repeat in exon 1 of the Huntingtin (HTT) gene leads to Huntington’s disease. However, expression of wild-type (WT) HTT is necessary to support normal development of function. Thus, potential therapeutic strategies include reducing expression of mutant HTT allele. LNA modified antisense oligonucleotides have been shown to successfully inhibit expression of the mutant HTT allele.84 Recently, AONs containing 7′-Me-carba-LNA have been subjected to evaluation for the same purpose.85 7′-Me-carbaLNA, LNA, and several other types of modification have been introduced into a 19mer AON consisting of repeating GCT sequence. The cells employed for this evaluation are patientderived fibroblast cells which contain 69 CAG repeats in the mutant allele and 17 in the WT allele. After transfection of these AONs into the cell followed by 4 days of incubation, IC50 values were calculated from Western blot quantification. As shown in Figure 8, native AON I, ENA modified AON V, and 2′F-RNA modified AON VI does not show any allele selectivity. However, the IC50 for 7′-Me-carba-LNA modified AON IV was 15 nM and revealed >6.6-fold selectivity, which is even better than standard LNA and cET modified counterparts. The AONs having the same sequence and modification but PS internucleosidic

10. RNAI POTENCY OF SIRNAS CONTAINING CARBA-LNA AND CARBA-ENA MODIFICATIONS siRNA has been proven to have the potency to knockdown virtually any specific genes in mammalian cells and hence represents an important new therapeutic strategy.87 However, native siRNA have many problems as potential therapeutics tools including nuclease degradation, off-targeting, delivery, and so on. An efficient solution to the challenges is chemical modification.88 Several of carba-LNA and carba-ENA nucleosides have been incorporated to siRNAs to target different mRNAs in cell culture. It was found that carba-LNA and carba-ENA nucleoside modification can render striking features for siRNA such as improved potency and nuclease reisitance and decreased off-target effect.89−92 Using the 2′-O-TEM based RNA synthesis strategy that has been developed in our lab,93,94 7′-Me-carba-LNA and 8′R-Mecarba-ENA modified siRNAs have been synthesized to target the eGFP gene in HeLa cells.89 These HeLa cells stably expressing eGFP were transfected with 10 nM siRNAs complexed with INTERFERin. After 72 h post-transfection, eGFP levels were evaluated and compared with that obtained in experiments employing siRNAs incorporated at different modification sites with 19 other types of modified nucleotides such as 2′-F, 2′-OMe,95 4′-modified RNA (HM),96 LNA; 3826

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 8. AONs containing LNA, 7′-Me-carba-LNA modifications support allele-selective HTT inhibition. Underlined letters indicate modified nucleosides. For AONs VII, VIII, and IX, all internucleosidic linkages are phosphorothioate.

α-L-LNA, aza-ENA,28 UNA,97 HNA,98 ANA,99 etc. This exhaustive study provides valuable guidance to chemically modify siRNAs as potential therapeutics.89 It was found that regardless of the modification type, the modifications in the passenger strand are well tolerated, and even siRNAs with heavily modified passenger strands support target gene knockout to below 20%. Modification at the 3′-end of passenger strand has been found to be a very effective strategy to improve siRNA performance especially when the modification can thermally destabilize siRNA duplex. It is likely that the 3′-end modification increases siRNA performance by favoring guide strand incorporation to RISC.89 However, not only modification type but also modification position as well as number of modifications in the guide strand have been found to have a strong impact on siRNA silencing activity:89 (1) Extensive modification of the guide strand is generally accompanied by reduced activity. (2) Multimodification in the seed region (positions 2−8 counting from 5′-end) result in reduced activity, but single modifications in the seed region are generally tolerated for HNA, ANA, 2′-OMe, 7′-Me-carba-LNA, and 8′R-Me-carbaENA, whereas single LNA, aza-ENA modification in the seed region have a negative effect on silencing activity. (3) Modification in the central position of guide strand is very sensitive to modification type. Modifications that are capable of perfect base-pairing to the target gene are generally well tolerated; otherwise, the modification in the central position of guide strand impairs silencing activity. (4) Most modification types are well tolerated in the 3′-region of the guide strand. Hence extensive modification in the 3′-region of the guide strand could be preferred to improve the 3′-exonuclease resistance. The potency of downregulation of eGFP by siRNAs containing different chemical modifications in the guide strand have been compared, and it was found that siRNAs containing 7′-Me-carba-LNA or 8′R-Me-carba-ENA modification are obviously higher than other types of modifications (Figure 9).89 siRNAs containing one 7′-Me-carba-LNA (JC-F1, Figure 9) or 8′R-Me-carba-ENA (JC-S1) modification at the position 3 (from 5′ end) of guide strand showed the best gene silencing

efficiency, which is much higher than that of siRNAs containing UNA, aza-ENA, or LNA modification in the same region. They are compatible with the siRNAs containing a HNA modification in the seed region as well as with unmodified siRNA (Figure 9). Interestingly, some siRNAs such as JC-F1/W131 and JC-S1/DO003 that formed by annealing JC-F1 and JC-S1 respectively with modified passenger strands showed even better silencing potency than the native counterpart.89 In this study, LNA modified siRNA was found to have significantly improved serum stability than native siRNA. The stability of 7′-Me-carba-LNA modified siRNA should be at least equal or better than the LNA modified one. However both LNA and 7′-Me-carba-LNA modified siRNA did not show better eGFP knockdown activity than native siRNA, suggesting the activity of siRNA in cell culture is not related to the serum stability of siRNA. Hence, we cannot attribute the improved activity of 7′-Me-carba-LNA modified siRNA compared to LNA modified counterpart to the better nuclease resistance of the former. It is known100 that interaction between the seed region of the guide strand and the complementary sites in the target mRNA is very important for the specificity of gene deregulation. Chemical modification in the seed region has been proven to be an efficient approach to reduce the off-target effects.101 Recently, 7′-Me-carba-LNA and 8′R-Me-carba-ENA together with other eight types of modifications have been incorporated into the seed region of the guide strand of siRNA to evaluate their off-target effects.90 Generally, siRNA potency was found to be positively correlated to off-target effects for most investigated siRNA. However, reduced off-targeting can be achieved by properly incorporating some types of modifications to specific positions in the seed region. For example, a single 2′OMe modification at position 2, a single 7′-Me-carba-LNA, 8′R-carba-ENA, aza-ENA, or HNA modification at position 3 can obviously reduce off-target effects. A single UNA modification, however, at position 7 resulted in the most potent activity. Since a single nucleotide modification of carba-LNA or carba-ENA in the seed region is well tolerated,89 it seems 3827

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

Figure 9. (A) Comparison of silencing eGFP activity of modified siRNAs in HeLa cells. The eGFP levels shown here are the values after 72 h post-transfection of 10 nM siRNAs complexed with INTERFERin. The sequences of siRNA and structures of modifications are shown in (B).

to increased silencing potency. Serum stability study revealed that 7′-Me-carba-LNA modification confers exceptional stability in a position dependent manner: 91 Double modification at position 20 and 1 is the best, followed by single modification at position 20, and single modification at position 1 showed the least stability though it is still 2 times more stable than the unmodified one. Comparatively, LNA modified siRNAs were found similar or only slightly less stable than 7′-Me-carba-LNA modified counterparts in serum, which is contradictory to the fact that 7′-Me-carbaLNA modified antisense oligonucleotides is significantly more stable LNA modified ones. This is because degradation of double-stranded siRNA in serum follows different mechanisms than degradation of single-stranded antisense olionucleotides.102−104 In this study,91 the toxicity of these modified siRNAs was studied by MTT assay. It was found all the 7′-Me-carba-LNA and LNA modified siRNAs were just the same as native siRNA, so none of them confer cellular toxicity.

that incorporating a single carba-LNA or carba-ENA modification at position 3 of the guide strand can produce siRNAs with high knockdown activity but reduced off-target effects. Silencing HIV-1 by carba-LNA modified siRNA in HEK293T cells culture has been studied by us recently.91 One 21mer siRNA, 5′-U1A2G3C4C5A6G7A8G9A10G11C12U13C14C15C16A17G18G19U20U21 (guide strand), was selected as the target against the HIV-1 TAR1 region. 7′-Me-carba-LNA and LNA were introduced into position U1, U13, U20 or U1 + U20 in the guide strand. After 48 h post-transfection using Lipofectamine, the IC50 values were evaluated by p24 ELISA (Figure 10). 7′-Mecarba-LNA modified siRNA showed very similar silencing potency as the LNA modified counterpart when their modification positions are the same. However their silencing potency varied a lot with the change of the modification positions. Modification at position 1 and position 13 resulted in a minimum effect and the silencing efficiency was found similar as native siRNA, whereas modification at position 20 and especially 20 plus 1 led 3828

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

This is one of the advantages of carba-LNA and carba-ENA over LNA. Another advantage of carba-LNA and carba-ENA over LNA is the carba-LNA and carba-ENA modified antisense oligonucleotides are significantly more stable in serum than the LNA modified counterpart. It is known the antisense activity is correlated directly with nuclease resistance for antisense oligonucleotides.77 So far, the limited biological evaluation data showed when containing phosphorothioate (PS) internucleosidic linkages, carba-LNA and LNA modified AONs have similar silencing efficiency,40,85 but when containing normal phosphate internucleosidic linkages, carba-LNA modified AONs showed slightly better silencing potency than LNA modified ones.85 However, it is not clear how much of the increased silencing potency should be attributed to its improved nuclease stability. Hence, the effect of nuclease stability of carbaLNA modified AON on the in vitro and in vivo kinetics of gene silencing needs to be addressed and compared with the LNA modified counterpart. Since PS-modified oligonucleotides often show some level of toxicity, development of nuclease stable oligonucleotides with high affinity to target RNA but without PS modification continues to be a perspective trend. From this point, carba-LNA modifications deserve more attention in the field of antisense oligonucleotides based therapeutics. Though single-strand RNA is very susceptible to intracellular nucleases, double-strand siRNA is fairly stable once entering into the cells.105 The major degradation of siRNA occurs during delivery.106 The mostly used delivery agents such DOTAP and Lipofectamine 2000 do not protect the siRNA from degradation,107 so improvement of nuclease stability for siRNA is also important for successful therapeutics. Not like in antisense oligonucleotides, one or two carba-LNA modifications in the 3′ end and/or 5′ end of siRNA does not significantly improve the nuclease stability compared with standard LNA. This is because siRNA is digested in serum by RNase A like endonuclease but not by exonuclease. Comparison of nuclease stability of carbaLNA heavily modified siRNA with the LNA modified counterpart and studying the effect of siRNA nuclease stability on in vitro and in vivo gene silencing remain to be done. On the other hand, the interesting observation that significant diastereomeric bias with C6′-O-Tol-carba-LNA modified siRNA 92 suggests modification on the carbocyclic moiety of carba-LNA is an efficient strategy to modulate the RNAi potency. A detailed structure−activity relationship based on the available carba-LNA derivatives may provide valuable principles on how to modify siRNA to get high silencing potency. In this aspect, practical conclusion which can be drawn from above observations is that introduction of various types of modifications at C6′ and C7′ in carba-LNA provides unique possibilities, compared to the LNA type molecules, to mine differently active molecules for “designer” thermodynamic and nuclease stability and delivery, which are essential for synthetic oligonucleotides to be successful as potential therapeutics. This is because there is a choice to fine-tune the electrostatics of the backbone in order to further develop carba-LNA nucleotides with high affinity toward target RNA and to prepare favorable conjugates built on the carba-bridge for one of the carba-LNA derivatives to be successful in the ADME-Tox studies.

Figure 10. Targeting HIV-1 TAR1 region by LNA and carba-LNA modified siRNA. IC50 values were determined from p24 ELISA of HEK 293T cells culture supernatant. The sequence of guide sequence is 5′-U 1 A 2 G 3 C 4 C 5 A 6 G 7 A 8 G 9 A 10 G 11 C 12 U 13 C 14 C 15 C 16 A17G18G19U20U21 and modifications locate in position 1, 13, 20 or both 1 and 20.

A pair of members of the carba-LNA family, 6′R-O-Tol-7′SMe-carba-LNA and 6′S-O-Tol-7′S-Me-carba-LNA, have also been incorporated to siRNA to targeting the HIV-1 TAR1 region.92 Compared to 7′-Me-carba-LNA, both 6′R-O-Tol-7′SMe-carba-LNA and 6′S-O-Tol-7′S-Me-carba-LNA have a additional bulky and hydrophobic C6′-O-Tol group. When modified at the end of the guide strand, for example, position 1 and position 20, the silencing potency of 6′R-O-Tol-7′S-Mecarba-LNA and 6′S-O-Tol-7′S-Me-carba-LNA is more or less than 7′S-Me-carba-LNA. However, modification in the middle, position 13, 6′R-O-Tol-7′S-Me-carba-LNA was found to be 2.4-fold more efficient than 7′-Me-carba-LNA, but 6′S-O-Tol-7′S-Me-carba-LNA was found to be 10-fold less efficient than 7′-Me-carba-LNA, so the different configuration of C6′-O-Tol substitution leads to a 24-fold difference in the silencing potency. A model study showed the 6′R-O-Tol is exposed toward the edge of the RNA duplex, while 6′S-O-Tol is located in the minor groove. 6′S-O-Tol-7′S-Me-carba-LNA modification at the position 13 is in the vicinity of the RISC cleavage site, and it may inhibit RISC mediated hydrolysis. This study highlights the possibility of using carba-LNA derivatives as molecular probes to address the mechanism of some important biology processes.

11. CONCLUSIONS AND IMPLICATIONS Synthesis of carba-LNA- and carba-ENA-type modified nucleosides has been achieved so far by free-radical cyclization strategy and ring-closing metathesis. Free-radical cyclization approach also allowed us to synthesize α-L-carbaLNA nucleos(t)ides. In the duplex form, the carbocyclic moiety of carba-LNA and carba-ENA modifications is found located in the minor groove, whereas the carbocyclic moiety of α-L-carba-LNA lies in the major groove. The C7′ of carbaLNA and α-L-carba-LNA, C8′ of carba-ENA are in the center of the grooves, which provide very good scaffolds for further functionalization to study how different chemical environments in the grooves modulate target RNA recognition as well as how they affect interaction with different enzymes. 3829

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

AUTHOR INFORMATION

(http://www.boc.uu.se). He has received many awards including Humboldt Research Award and Sorm Award.

Corresponding Author

*E-mail: [email protected].

ACKNOWLEDGMENTS Generous financial support from the Swedish Natural Science Research Council (Vetenskapsrådet), the Swedish Foundation for Strategic Research (Stiftelsen för Strategisk Forskning), and the EU-FP6 funded RIGHT project (Project No. LSHB-CT2004-005276) and Uppsala University is gratefully acknowledged. Authors thank very warmly their co-workers who have been involved in this carba-LNA/ENA project. Authors also thank Oleksandr Plashkevych for his proofreading of the manuscript.

Notes

The authors declare no competing financial interest. Biographies

REFERENCES (1) Herdewijn, P. Liebigs Ann. 1996, 9, 1337. (2) Zhou, C. Z.; Chattopadhyaya, J. Curr. Opin. Drug Discovery Dev. 2009, 12, 876. (3) Mathe, C.; Perigaud, C. Eur. J. Org. Chem. 2008, 9, 1489. (4) Tarkoy, M.; Leumann, C. Angew. Chem., Int. Ed. Engl. 1993, 32, 1432. (5) Rodriguez, J. B.; Marquez, V. E.; Nicklaus, M. C.; Barchi, J. J. Tetrahedron Lett. 1993, 34, 6233. (6) Altmann, K. H.; Kesselring, R.; Francotte, E.; Rihs, G. Tetrahedron Lett. 1994, 35, 2331. (7) Altmann, K. H.; Imwinkelried, R.; Kesselring, R.; Rihs, G. Tetrahedron Lett. 1994, 35, 7625. (8) Marquez, V. E.; Siddiqui, M. A.; Ezzitouni, A.; Russ, P.; Wang, J. Y.; Wagner, R. W.; Matteucci, M. D. J. Med. Chem. 1996, 39, 3739. (9) Herdewijn, P. Biochim. Biophys. Acta 1999, 1489, 167. (10) Orum, H.; Wengel, J. Curr. Opin. Mol. Ther. 2001, 3, 239. (11) Obika, S.; Nanbu, D.; Hari, Y.; Morio, K.; In, Y.; Ishida, T.; Imanishi, T. Tetrahedron Lett. 1997, 38, 8735. (12) Singh, S. K.; Nielsen, P.; Koshkin, A. A.; Wengel, J. Chem. Commun. 1998, 4, 455. (13) Koshkin, A. A.; Singh, S. K.; Nielsen, P.; Rajwanshi, V. K.; Kumar, R.; Meldgaard, M.; Olsen, C. E.; Wengel, J. Tetrahedron 1998, 54, 3607. (14) McTigue, P. M.; Peterson, R. J.; Kahn, J. D. Biochemistry 2004, 43, 5388. (15) Eichert, A.; Behling, K.; Betzel, C.; Erdmann, V. A.; Fuerste, J. P.; Foerster, C. Nucleic Acids Res. 2010, 38, 6729. (16) Kumar, R.; Singh, S. K.; Koshkin, A. A.; Rajwanshi, V. K.; Meldgaard, M.; Wengel, J. Bioorg. Med. Chem. Lett. 1998, 8, 2219. (17) Singh, S. K.; Kumar, R.; Wengel, J. J. Org. Chem. 1998, 63, 10035. (18) Sorensen, M. D.; Petersen, M.; Wengel, J. Chem. Commun. 2003, 17, 2130. (19) Prakash, T. P.; Siwkowski, A.; Allerson, C. R.; Migawa, M. T.; Lee, S.; Gaus, H. J.; Black, C.; Seth, P. P.; Swayze, E. E.; Bhat, B. J. Med. Chem. 2010, 53, 1636. (20) Meldgaard, M.; Hansen, F. G.; Wengel, J. J. Org. Chem. 2004, 69, 6310. (21) Enderlin, G.; Nielsen, P. J. Org. Chem. 2008, 73, 6891. (22) Seth, P. P.; Vasquez, G.; Allerson, C. A.; Berdeja, A.; Gaus, H.; Kinberger, G. A.; Prakash, T. P.; Migawa, M. T.; Bhat, B.; Swayze, E. E. J. Org. Chem. 2010, 75, 1569. (23) Sorensen, M. D.; Kvaerno, L.; Bryld, T.; Hakansson, A. E.; Verbeure, B.; Gaubert, G.; Herdewijn, P.; Wengel, J. J. Am. Chem. Soc. 2002, 124, 2164. (24) Kumar, T. S.; Madsen, A. S.; Wengel, J.; Hrdlicka, P. J. J. Org. Chem. 2006, 71, 4188. (25) Kumar, T. S.; Madsen, A. S.; Ostergaard, M. E.; Sau, S. P.; Wengel, J.; Hrdlicka, P. J. J. Org. Chem. 2009, 74, 1070. (26) Wang, G. Y.; Gunic, E.; Girardet, J. L.; Stoisavljevic, V. Bioorg. Med. Chem. Lett. 1999, 9, 1147.

Chuanzheng Zhou was born in Hubei, China, in 1978. He earned his bachelor degree in chemistry from Nankai University in 2001 followed by a master's degree in organic chemistry in 2004 under the supervision of Prof. Zhen Xi. From October 2004 to September 2010 Chuanzheng Zhou worked under the supervision of Prof. Jyoti Chattopadhyaya at the Chemical Biology Program, Biomedical Center, Uppsala University on nucleic acids chemistry and structure leading to a Ph.D. degree in bioorganic chemistry in 2010. At present he is a postdoctoral fellow with Prof. Marc Greenberg at the Johns Hopkins University.

Jyoti Chattopadhyaya obtained his Ph.D. from National Chemical laboratory and Pune University in 1974 followed by a posdoctoral training with Prof. Collin B. Reese, FRS, in London University at King’s College (1974−1979). He has been holding a full-Chair of Bioorganic Chemistry since 1982 at Uppsala University. His research interests include chemical synthesis of DNA and RNA and their derivatives in order to modulate the biophysical, structural, and biological properties, stereoelectronic effects in nucleotides and how they assist in the preorganization of DNA and RNA, synthesis of nucleos(t)ide and nucleotide analogues as potential therapeutics, carbohydrate and heterocyclic chemistry as well as solution structures of nucleic acids by NMR spectroscopy. He has supervised 30 Ph.D. students and is the author of more than 400 scientific publications 3830

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

(60) Zhou, C.; Chattopadhyaya, J. Arkivoc 2009, iii, 171. (61) Liu, Y.; Xu, J. F.; Karimiahmadabadi, M.; Zhou, C. Z.; Chattopadhyaya, J. J. Org. Chem. 2010, 75, 7112. (62) Rajwanshi, V. K.; Hakansson, A. E.; Dahl, B. M.; Wengel, J. Chem. Commun. 1999, 15, 1395. (63) Rajwanshi, V. K.; Hakansson, A. E.; Kumar, R.; Wengel, J. Chem. Commun. 1999, 20, 2073. (64) Rajwanshi, V. K.; Hakansson, A. E.; Sorensen, M. D.; Pitsch, S.; Singh, S. K.; Kumar, R.; Nielsen, P.; Wengel, J. Angew. Chem., Int. Ed. 2000, 39, 1656. (65) Li, Q.; Yuan, F. F.; Zhou, C. Z.; Plashkevych, O.; Chattopadhyaya, J. J. Org. Chem. 2010, 75, 6122. (66) Seth, P. P.; Allerson, C. R.; Berdeja, A.; Swayze, E. E. Bioorg. Med. Chem. Lett. 2011, 21, 588. (67) Beckwith, A. L. J. Tetrahedron 1981, 37, 3073. (68) Rajanbabu, T. V. Acc. Chem. Res. 1991, 24, 139. (69) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1972, 94, 8205. (70) Plashkevych, O.; Chatterjee, S.; Honcharenko, D.; Pathmasiri, W.; Chattopadhyaya, J. J. Org. Chem. 2007, 72, 4716. (71) Petersen, M.; Bondensgaard, K.; Wengel, J.; Jacobsen, J. P. J. Am. Chem. Soc. 2002, 124, 5974. (72) Nielsen, J. T.; Stein, P. C.; Petersen, M. Nucleic Acids Res. 2003, 31, 5858. (73) Tereshko, V.; Gryaznov, S.; Egli, M. J. Am. Chem. Soc. 1998, 120, 269. (74) Egli, M.; Minasov, G.; Teplova, M.; Kumar, R.; Wengel, J. Chem. Commun. 2001, 7, 651. (75) Seth, P. P.; Yu, J.; Allerson, C. R.; Berdeja, A.; Swayze, E. E. Bioorg. Med. Chem. Lett. 2011, 21, 1122. (76) Gyi, J. I.; Lane, A. N.; Conn, G. L.; Brown, T. Biochemistry 1998, 37, 73. (77) Monia, B. P.; Johnston, J. F.; Sasmor, H.; Cummins, L. L. J. Biol. Chem. 1996, 271, 14533. (78) Crooke, S. T. Annu. Rev. Med. 2004, 55, 61. (79) Shaw, J. P.; Kent, K.; Bird, J.; Fishback, J.; Froehler, B. Nucleic Acids Res. 1991, 19, 747. (80) Kurreck, J.; Wyszko, E.; Gillen, C.; Erdmann, V. A. Nucleic Acids Res. 2002, 30, 1911. (81) Zhou, C.; Chattopadhyaya, J. J. Org. Chem. 2010, 75, 2341. (82) Wu, H. J.; Lima, W. F.; Crooke, S. T. J. Biol. Chem. 1999, 274, 28270. (83) Nowotny, M.; Cerritelli, S. M.; Ghirlando, R.; Gaidamakov, S. A.; Crouch, R. J.; Yang, W. EMBO J. 2008, 27, 1172. (84) Hu, J.; Matsui, M.; Gagnon, K. T.; Schwartz, J. C.; Gabillet, S.; Arar, K.; Wu, J.; Bezprozvanny, I.; Corey, D. R. Nat. Biotechnol. 2009, 27, 478. (85) Gagnon, K. T.; Pendergraff, H. M.; Deleavey, G. F.; Swayze, E. E.; Potier, P.; Randolph, J.; Roesch, E. B.; Chattopadhyaya, J.; Damha, M. J.; Bennett, C. F.; Montaillier, C.; Lemaitre, M.; Corey, D. R. Biochemistry 2010, 49, 10166. (86) Hu, J. X.; Gagnon, K. T.; Liu, J.; Watts, J. K.; Syeda-Nawaz, J.; Bennett, C. F.; Swayze, E. E.; Randolph, J.; Chattopadhyaya, J.; Corey, D. R. Biol. Chem. 2011, 392, 315. (87) Watts, J. K.; Corey, D. R. Bioorg. Med. Chem. Lett. 2010, 20, 3203. (88) Watts, J. K.; Deleavey, G. F.; Damha, M. J. Drug Discovery Today 2008, 13, 842. (89) Bramsen, J. B.; Laursen, M. B.; Nielsen, A. F.; Hansen, T. B.; Bus, C.; Langkjaer, N.; Babu, B. R.; Hojland, T.; Abramov, M.; Van Aerschot, A.; Odadzic, D.; Smicius, R.; Haas, J.; Andree, C.; Barman, J.; Wenska, M.; Srivastava, P.; Zhou, C. Z.; Honcharenko, D.; Hess, S.; Muller, E.; Bobkov, G. V.; Mikhailov, S. N.; Fava, E.; Meyer, T. F.; Chattopadhyaya, J.; Zerial, M.; Engels, J. W.; Herdewijn, P.; Wengel, J.; Kjems, J. Nucleic Acids Res. 2009, 37, 2867. (90) Bramsen, J. B.; Pakula, M. M.; Hansen, T. B.; Bus, C.; Langkjaer, N.; Odadzic, D.; Smicius, R.; Wengel, S. L.; Chattopadhyaya, J.; Engels, J. W.; Herdewijn, P.; Wengel, J.; Kjems, J. Nucleic Acids Res. 2010, 38, 5761.

(27) Morita, K.; Takagi, M.; Hasegawa, C.; Kaneko, M.; Tsutsumi, S.; Sone, J.; Ishikawa, T.; Imanishi, T.; Koizumi, M. Biorg. Med. Chem. 2003, 11, 2211. (28) Varghese, O. P.; Barman, J.; Pathmasiri, W.; Plashkevych, O.; Honcharenko, D.; Chattopadhyaya, J. J. Am. Chem. Soc. 2006, 128, 15173. (29) Honcharenko, D.; Zhou, C. Z.; Chattopadhyaya, J. J. Org. Chem. 2008, 73, 2829. (30) Wenska, M.; Honcharenko, D.; Pathmasiri, W.; Chattopadhyaya, J. Heterocycles 2007, 73, 303. (31) Honcharenko, D.; Barman, J.; Varghese, O. P.; Chattopadhyaya, J. Biochemistry 2007, 46, 5635. (32) Rahman, S. M. A.; Seki, S.; Obika, S.; Yoshikawa, H.; Miyashita, K.; Imanishi, T. J. Am. Chem. Soc. 2008, 130, 4886. (33) Mitsuoka, Y.; Kodama, T.; Ohnishi, R.; Hari, Y.; Imanishi, T.; Obika, S. Nucleic Acids Res. 2009, 37, 1225. (34) Srivastava, P.; Barman, J.; Pathmasiri, W.; Plashkevych, O.; Wenska, M.; Chattopadhyaya, J. J. Am. Chem. Soc. 2007, 129, 8362. (35) Zhou, C. Z.; Plashkevych, O.; Chattopadhyaya, J. Org. Biomol. Chem. 2008, 6, 4627. (36) Zhou, C.; Liu, Y.; Andaloussi, M.; Badgujar, N.; Plashkevych, O.; Chattopadhyaya, J. J. Org. Chem. 2009, 74, 118. (37) Xu, J. F.; Dupouy, C.; Liu, Y.; Chattopadhyaya, J. J. Org. Chem. 2009, 74, 6534. (38) Albaek, N.; Petersen, M.; Nielsen, P. J. Org. Chem. 2006, 71, 7731. (39) Kumar, S.; Hansen, M. H.; Albaek, N.; Steffansen, S. I.; Petersen, M.; Nielsen, P. J. Org. Chem. 2009, 74, 6756. (40) Seth, P. P.; Allerson, C. R.; Berdeja, A.; Siwkowski, A.; Pallan, P. S.; Gaus, H.; Prakash, T. P.; Watt, A. T.; Egli, M.; Swayze, E. E. J. Am. Chem. Soc. 2010, 132, 14942. (41) Lebreton, J.; Escudier, J. M.; Arzel, L.; Len, C. Chem. Rev. 2010, 110, 3371. (42) Lamb, R. C.; Toney, M. K.; Ayers, P. W. J. Am. Chem. Soc. 1963, 85, 3483. (43) Giese, B. Radicals in Organic Synthesis: Formation of CarbonCarbon Bonds; Pergamon Press: Oxford, 1986. (44) Wu, J. C.; Xi, Z.; Gioeli, C.; Chattopadhyaya, J. Tetrahedron 1991, 47, 2237. (45) Xi, Z.; Agback, P.; Plavec, J.; Sandstrom, A.; Chattopadhyaya, J. Tetrahedron 1992, 48, 349. (46) Xi, Z.; Glemarec, C.; Chattopadhyaya, J. Tetrahedron 1993, 49, 7525. (47) Xi, Z.; Rong, J. H.; Chattopadhyaya, J. Tetrahedron 1994, 50, 5255. (48) Xi, Z.; Agback, P.; Sandstrom, A.; Chattopadhyaya, J. Tetrahedron 1991, 47, 9675. (49) Koole, L. H.; Wu, J. C.; Neidle, S.; Chattopadhyaya, J. J. Am. Chem. Soc. 1992, 114, 2687. (50) Velazquez, S.; Jimeno, M. L.; Huss, S.; Balzarini, J.; Camarasa, M. J. J. Org. Chem. 1994, 59, 7661. (51) de Oliveira, R. B.; Alves, R. J.; de Souza, J. D.; Prado, M. A. F. J. Braz. Chem. Soc. 2003, 14, 442. (52) Kumamoto, H.; Ogamino, J.; Tanaka, H.; Suzuki, H.; Haraguchi, K.; Miyasaka, T.; Yokomatsu, T.; Shibuya, S. Tetrahedron 2001, 57, 3331. (53) Marcocontelles, J.; Ruiz, P.; Martinez, L.; Martinezgrau, A. Tetrahedron 1993, 49, 6669. (54) Upadhayaya, R.; Deshpande, S. G.; Li, Q.; Kardile, R. A.; Sayyed, A. Y.; Kshirsagar, E. K.; Salunke, R. V.; Dixit, S. S.; Zhou, C.; Foldesi, A.; Chattopadhyaya, J. J. Org. Chem. 2011, 76, 4408. (55) Le Clezio, I.; Escudier, J. M.; Vigroux, A. Org. Lett. 2003, 5, 161. (56) Dupouy, C.; Lavedan, P.; Escudier, J.-M. Eur. J. Org. Chem. 2008, 7, 1285. (57) Le Clezio, I.; Gornitzka, H.; Escudier, J. M.; Vigroux, A. J. Org. Chem. 2005, 70, 1620. (58) Zhou, C. Z.; Plashkevych, O.; Chattopadhyaya, J. J. Org. Chem. 2009, 74, 3248. (59) Zhou, C.; Chattopadhyaya, J. Heterocycles 2009, 78, 1715. 3831

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832

Chemical Reviews

Review

(91) Dutta, S.; Bhaduri, N.; Rastogi, N.; Chandel, S. G.; Vandavasi, J. K.; Upadhayaya, R. S.; Chattopadhyaya, J. Med. Chem. Commun. 2011, 2, 206. (92) Dutta, S.; Bhaduri, N.; Upadhayaya, R. S.; Rastogi, N.; Chandel, S. G.; Vandavasi, J. K.; Plashkevych, O.; Kardile, R. A.; Chattopadhyaya, J. Med. Chem. Commun. 2011, 2, 1110. (93) Zhou, C. Z.; Honcharenko, D.; Chattopadhyaya, J. Org. Biomol. Chem. 2007, 5, 333. (94) Zhou, C. Z.; Pathmasiri, W.; Honcharenko, D.; Chatterjee, S.; Barman, J.; Chattopadhyaya, J. Can. J. Chem. 2007, 85, 293. (95) Manoharan, M. Biochim. Biophys. Acta 1999, 1489, 117. (96) Thrane, H.; Fensholdt, J.; Regner, M.; Wengel, J. Tetrahedron 1995, 51, 10389. (97) Nielsen, P.; Dreiøe, L. H.; Wengel, J. Biorg. Med. Chem. 1995, 3, 19. (98) Hendrix, C.; Rosemeyer, H.; Verheggen, I.; Seela, F.; VanAerschot, A.; Herdewijn, P. Chem.Eur. J. 1997, 3, 110. (99) Allart, B.; Khan, K.; Rosemeyer, H.; Schepers, G.; Hendrix, C.; Rothenbacher, K.; Seela, F.; Van Aerschot, A.; Herdewijn, P. Chem. Eur. J. 1999, 5, 2424. (100) Jackson, A. L.; Burchard, J.; Schelter, J.; Chau, B. N.; Cleary, M.; Lim, L.; Linsley, P. S. RNA 2006, 12, 1179. (101) Jackson, A. L.; Burchard, J.; Leake, D.; Reynolds, A.; Schelter, J.; Guo, J.; Johnson, J. M.; Lim, L.; Karpilow, J.; Nichols, K.; Marshall, W.; Khvorova, A.; Linsley, P. S. RNA 2006, 12, 1197. (102) Turner, J. J.; Jones, S. W.; Moschos, S. A.; Lindsay, M. A.; Gait, M. J. Mol. Biosyst. 2007, 3, 43. (103) Haupenthal, J.; Baehr, C.; Kiermayer, S.; Zeuzem, S.; Piiper, A. Biochem. Pharmacol. 2006, 71, 702. (104) Hong, J. M.; Huang, Y. Y.; Li, J.; Yi, F.; Zheng, J.; Huang, H. A.; Wei, N.; Shan, Y. Q.; An, M. R.; Zhang, H. Y.; Ji, J. G.; Zhang, P. Z.; Xi, Z.; Du, Q. A.; Liang, Z. C. FASEB J. 2010, 24, 4844. (105) Raemdonck, K.; Remaut, K.; Lucas, B.; Sanders, N. N.; Demeester, J.; De Smedt, S. C. Biochemistry 2006, 45, 10614. (106) Bartlett, D. W.; Davis, M. E. Biotechnol. Bioeng. 2007, 97, 909. (107) Lee, S.; Yang, S. C.; Kao, C.-Y.; Pierce, R. H.; Murthy, N. Nucleic Acids Res. 2009, 37, e145.

3832

dx.doi.org/10.1021/cr100306q | Chem. Rev. 2012, 112, 3808−3832