Free-radical ring closure to conformationally locked α-l-carba-LNAs and synthesis of their oligos: Nuclease stability, target RNA specificity, and elicitation of RNase H

pubs.acs.org/joc

Free-Radical Ring Closure to Conformationally Locked

r-^L-Carba-LNAs and Synthesis of Their Oligos: Nuclease Stability,

Target RNA Specificity, and Elicitation of RNase H

Qing Li, Fengfeng Yuan, Chuanzheng Zhou, Oleksandr Plashkevych, and

Jyoti Chattopadhyaya*

Program of Bioorganic Chemistry, Institution of Cell and Molecular Biology, Box 581, Biomedical Center,

Uppsala University, SE-751 23 Uppsala, Sweden

jyoti@boc.uu.se

Received May 7, 2010

A new class of conformationally constrained nucleosides, R-L-ribo-carbocyclic LNA thymidine (R-L-

carba-LNA-T, LNA is an abbreviation of locked nucleic acid) analogues and a novel “double-

locked” R-^L-ribo-configured tetracyclic thymidine (6,7⁰-methylene-bridged-R-L-carba-LNA-T) in

which both the sugar puckering and glycosidic torsion are simultaneously constrained, have been

synthesized through a key step involving 5-exo free-radical intramolecular cyclization. These R-^L-

carba-LNA analogues have been subsequently transformed to corresponding phosphoramidites and

incorporated into isosequential antisense oligonucleotides (AONs), which have then been examined

for the thermal denaturation of their duplexes, nuclease stability, and RNase H recruitment

capabilities. Introduction of a single 6⁰,7⁰-substituted R-L-carba-LNA-T modification in the AON

strand of AON/RNA heteroduplex led to T_mreduction by 2-3 °C as compared to the native

heteroduplex, whereas the parent 2⁰-oxa-R-L-LNA-T modification at the identical position in the

AON strand has been found to lead to an increase in the T_mby 3-5 °C. This suggests that the 6⁰and 7⁰

substitutions lead to much reduced thermal stability for the modified heteroduplex, especially the

hydrophobic 7⁰-methyl on R-L-carba-LNA, which is located in the major groove of the duplex. All of

the AONs incorporating 6⁰,7⁰-substituted R-L-carba-LNA-T have, however, showed considerably

improved nuclease stability toward 3⁰-exonuclease (SVPDE) and in human blood serum compared to

the 2⁰-oxa-R-L-LNA-T incorporated AONs. The hybrid duplexes that are formed by 6⁰,7⁰-substituted

R-^L-carba-LNA-T-modified AONs with complementary RNA have been found to recruit RNase H

with higher efficiency than those of the β-^D-LNA-T or β-D-carba-LNA-T-modified counterparts.

These greatly improved nuclease resistances and efficient RNase H recruitment capabilities elevate

the R-L-carba-LNA-modified nucleotides into a new class of locked nucleic acids for potential RNA

targeting therapeutics.

6122 J. Org. Chem. 2010, 75, 6122–6140

Published on Web 08/26/2010

DOI: 10.1021/jo100900v

2010 American Chemical Society

r

Li et al.

JOCArticle

Introduction

antisense properties, such as RNA affinity, nuclease resis-

tance, and RNase H elicitation, as compared to that of the

native counterpart. Some of these locked nucleotides have

been tested for siRNA action and have shown remarkable

silencing efficiency.^25,26

Among the varieties of gene silencing technologies, the

antisense strategy¹and RNA interference²are considered to

be extremely potent in terms of ensuring complete target

mRNA breakdown. During the last couple of decades,

exploration and examination of novel chemically modified

oligonucleotides that act as potent and selective therapeutic

agents have gained momentum and have led to the develop-

ment of analogues which have good pharmacokinetics and

minimum toxicity. In general, there are three main types of

possible chemical modifications of native nucleotides, and

each involves modification of either sugar, phosphodiester

linkage, or nucleobase.³

Recently,carba-LNA,carba-ENA,C6⁰,C7⁰-substituted carba-

LNAs, and C6⁰,C8⁰-substituted carba-ENAs have been synthe-

sized by our group through a key step involving intramolecular

free-radical ring closure.^27-32The striking feature of carba-LNA

and carba-ENA derivatives is that they render much better

nuclease resistance for modified AONs than LNA.³²The carbo-

cyclic ring of carba-LNA and carba-ENA also provides an

effective handle for engineering new types of modifications in the

minor groove, which can significantly modulate important anti-

sense properties such as target RNA affinity, nuclease resistance

without impairing their RNase H recruitment capabilities.

Another intriguing class of conformationally constrained

nucleoside, R-^L-LNA, is a diastereomer of LNA^7,8(referred to

as β-D-LNA further in the text in order to distinguish from

R-^L-LNA). The R-L-LNA has also been synthesized by Wengel

et al.^33-40Thermal stability of R-L-LNA-modified AON with

complementary nucleic acid targets was found to be comparable

to that of LNA-containing counterparts. Moreover, it has been

shown that the R-^L-LNA modification renders better nuclease

resistance compared to LNA modification.³⁶In vitro and in vivo

experiments showed that R-^L-LNA-modified AONs retain the

ability of RNase H elicitation and exhibit even higher gene

knockdown efficacy than LNA-modified counterparts.^41,42

In recent years, a variety of conformationally constrained

nucleotides^4-6have been synthesized and incorporated into

antisense oligonucleotides (AONs) and/or siRNAs for in vitro

biological studies. One particular class of compounds, so-called

locked nucleic acid (LNA,^7,8also called BNA,^9,10BNA is the

abbreviation of bridged nucleic acid) containing a methylene

linkage between the 2⁰-oxygen and 4⁰-carbon of the ribose ring,

has attracted extensive interest due to their remarkable hybridi-

zation properties.⁸The excellent thermodynamic features of

LNA stimulated synthesis of a number of other North-type con-

formationally constrained nucleoside analogues, such as amino-

LNA,¹¹β-bicyclonucleoside,¹²C6⁰-substituted LNA,^13,14

BNA^{COC 15}BNA^{NC 16,17}2⁰-O,4⁰-C-ethylene-bridged nucleic

, ,

acid(ENA),^18,19aza-ENA,^20-22carba-ENA^23,24etc. All of these

locked nucleoside-modified AONs have shown much better

(1) Crooke, S. T. Annu. Rev. Med. 2004, 55, 61–95.

(2) Novina, C. D.; Sharp, P. A. Nature 2004, 430, 161–164.

(3) Freier, S. M.; Altmann, K. H. Nucleic Acids Res. 1997, 25, 4429–4435.

(4) Zhou, C.; Chattopadhyaya, J. Curr. Opin. Drug Discovery Dev. 2009,

12, 876–898.

(25) Nauwelaerts, K.; Fisher, M.; Froeyen, M.; Lescrinier, E.; Van

Aerschot, A.; Xu, D.; Delong, R.; Kang, H.; Juliano, R. L.; Herdewijn, P.

J. Am. Chem. Soc. 2007, 129, 9340–9348.

(26) Bramsen, J. B.; Laursen, M. B.; Nielsen, A. F.; Hansen, T. B.; Bus, C.;

Langkjær, N.; Babu, B. R.; Højland, T.; Abramov, M.; Van Aerschot, A.; Odadzic,

D.; Smicius, R.; Haas, J.; Andree, C.; Barman, J.; Wenska, M.; Srivastava, P.;

Zhou, C.; 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, 2867–2881.

(27) Srivastava, P.; Barman, J.; Pathmasiri, W.; Plashkevych, O.;

Wenska, M.; Chattopadhyaya, J. J. Am. Chem. Soc. 2007, 129, 8362–8379.

(28) Zhou, C.; Liu, Y.; Andaloussi, M.; Badgujar, N.; Plashkevych, O.;

Chattopadhyaya, J. J. Org. Chem. 2009, 74, 118–134.

(29) Xu, J.; Liu, Y.; Dupouy, C.; Chattopadhyaya, J. J. Org. Chem. 2009,

74, 6534–6554.

(5) Herdewijn, P. Liebigs Ann. 1996, 1337–1348.

(6) Mathe, C; Perigaud, C. Eur. J. Org. Chem. 2008, 1489–1505.

(7) Singh, S. K.; Nielsen, P.; Koshkin, A. A.; Wengel, J. Chem. Commun.

1997, 1643–1644.

(8) Koshkin, A. A.; Singh, S. K.; Nielsen, P.; Rajwanshi, V. K.; Kumar, R.;

Meldgaard, M.; Olsen, C. E.; Wengel, J. Tetrahedron 1998, 54, 3607–3630.

(9) Obika, S.; Nanbu, D.; Hari, Y.; Morio, K.; In, Y.; Ishida, T.;

Imanishi, T. Tetrahedron Lett. 1997, 38, 8735–8738.

(10) Obika, S.; Nanbu, D.; Hari, Y.; Andoh, J.; Morio, K.; Doi, T.;

Imanishi, T. Tetrahedron Lett. 1998, 39, 5401–5404.

(11) Singh, S.K.;Kumar, R.;Wengel, J.J. Org. Chem. 1998,63, 10035–10039.

(12) Wang, G.; Girardet, J. L.; Gunic, E. Tetrahedron 1999, 55, 7707–7724.

(13) Enderlin, G.; Nielsen, P. J. Org. Chem. 2008, 73, 6891–6894.

(14) 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, 1569–1581.

(15) Hari, Y.; Obika, S.; Ohnishi, R.; Eguchi, K.; Osaki, T.; Ohishi, H.;

Imanishi, T. Bioorg. Med. Chem. 2006, 14, 1029–1038.

(16) Rahman, S. M. A.; Seki, S.; Obika, S.; Yoshikawa, H.; Miyashita,

K.; Imanishi, T. J. Am. Chem. Soc. 2008, 130, 4886–4896.

(17) Prakash, T. P.; Siwkowski, A.; Allerson, C. A.; Migawa, M. T.; Lee,

S.; Gaus, H. J.; Black, C.; Seth, P. P.; Swayze, E. E.; Bhat, B. J. Med. Chem.

2010, 53, 1636–1650.

(18) Morita, K.; Takagi, M.; Hasegawa, C.; Kaneko, M.; Tsutsumi, S.;

Sone, J.; Ishikawa, T.; Imanishi, T.; Koizumi, M. Bioorg. Med. Chem. 2003,

11, 2211–2226.

(19) Morita, K.; Hasegawa, C.; Kaneko, M.; Tsutsumi, S.; Sone, J.;

Ishikawa, T.; Imanishi, T.; Koizumi, M. Bioorg. Med. Chem. Lett. 2002,

12, 73–76.

(30) Zhou, C.; Plashkevych, O.; Chattopadhyaya, J. Org. Biomol. Chem.

2008, 6, 4627–4633.

(31) Zhou, C.; Plashkevych, O.; Chattopadhyaya, J. J. Org. Chem. 2009,

74, 3248–3265.

(32) Zhou, C.; Chattopadhyaya, J. J. Org. Chem. 2010, 75, 2341–2349.

(33) Rajwanshi, V. K.; Hakansson, A. E.; Dahl, B. M.; Wengel, J. Chem.

Commun. 1999, 1395–1396.

(34) Rajwanshi, V. K.; Hakansson, A. E.; Kumar, R.; Wengel, J. Chem.

Commun. 1999, 2073–2074.

(35) Hakansson, A. E.; Koshkin, A. A.; Sørensen, M. D.; Wengel, J.

J. Org. Chem. 2000, 65, 5161–5166.

(36) Sørensen, M. D.; Kværnø, L.; Bryld, T.; Hakansson, A. E.;

Verbeure, B.; Gaubert, G.; Herdewijn, P.; Wengel, J. J. Am. Chem. Soc.

2002, 124, 2164–2176.

(37) Rajwanshi, V. K.; Hakansson, A. E.; Sørensen, M. D.; Pitsch, S.;

Singh, S. K.; Kumar, R.; Nielsen, P.; Wengel, J. Angew. Chem., Int. Ed. 2000,

39, 1656–1659.

(38) Petersen, M.; Hakansson, A. E.; Wengel, J.; Jacobsen, J. P. J. Am.

Chem. Soc. 2001, 123, 7431–7432.

(20) Varghese, O. P.; Barman, J.; Pathmairi, W.; Plashkevych, O.;

Honcharenko, D.; Chattopadhyaya, J. J. Am. Chem. Soc. 2006, 128, 15173–

15187.

(21) Honcharenko, D.; Barman, J.; Varghese, O. P.; Chattopadhyaya, J.

Biochemistry 2007, 46, 5635–5646.

(39) Kumar, T. S.; Madsen, A. S.; Wengel, J.; Hrdlicka, P. J. J. Org.

Chem. 2006, 71, 4188–4201.

(40) Kumar, T. S.; Madsen, A. S.; Østergaard, M. E.; Sau, S. P.; Wengel,

J.; Hrdlicka, P. J. J. Org. Chem. 2009, 74, 1070–1081.

(41) Fluiter, K.; Frieden, M.; Vreijling, J.; Rosenbohm, C.; De Wissel,

M. B.; Christensen, S. M.; Koch, T.; Ørum, H.; Baas, F. ChemBioChem 2005,

6, 1104–1109.

(22) Honcharenko, D.; Zhou, C.; Chattopadhyaya, J. J. Org. Chem. 2008,

73, 2829–2842.

(23) Albæk, N.; Petersen, M.; Nielsen, P. J. Org. Chem. 2006, 71, 7731–7740.

(24) Kumar, S.; Hansen, M. H.; Albæk, N.; Steffansen, S. I.; Petersen, M.;

Nielsen, P. J. Org. Chem. 2009, 74, 6756–6769.

(42) Frieden, M.; Christensen, S. M.; Mikkelsen, N. D.; Rosenbohhm, C.;

Thrue, C. A.; Westergaard, M.; Hansen, H. F.; Ørum, H.; Koch, T. Nucleic

Acids Res. 2003, 31, 6363–6372.

J. Org. Chem. Vol. 75, No. 18, 2010 6123

Li et al.

JOCArticle

FIGURE 1. Structures of β-D-LNA (A),^7-9β-D-carba-LNA (B),^27-29R-L-LNA (C),^33-376⁰-Me-R-L-LNA (D),⁴⁴2⁰-amino-R-L-LNA (E)³⁹and

derivatives (F-H),⁴⁰tetracyclic 2⁰-amino-R-L-LNA (I),³⁹7⁰-methylene-R-L-carba-LNA (J): This nucleoside was reported in a recent patent,⁴³

but it was not characterized, not even partially, by either ¹H or ¹³C NMR. Instead, only ³¹P NMR of the corresponding phosphoramidite was

supplied, which in fact could be the same for any nucleoside phosphoramidite). 6⁰,7⁰-Substituted R-L-carba-LNA derivatives (K-M), along

with a novel double-locked R-^L-carba-LNA analogue: 6,7⁰-methylene-bridged R-L-carba-LNA-T (N).

The striking biochemical features of R-L-LNA prompted us and

others^43,44to synthesize R-L-ribo-carba-LNA, in which a methy-

lene group replaces the 2⁰-oxygen of R-L-LNA yielding a 2⁰,4⁰-

carbocyclic locked ring. Unlike the 2⁰,4⁰-carbocycle locked ring

in carba-LNA (referred to as β-^D-carba-LNA further in the text

in order to distinguish from R-^L-carba-LNA), which is located in

the minor groove of AON/RNA duplexes, the 2⁰,4⁰-carbocyclic

ring in R-^L-carba-LNA is located in the major groove of AON/

RNA duplexes (see Figure SII 41 and 42 in Supporting Infor-

mation). Hence C6⁰/7⁰-substitued R-L-carba-LNAs are also very

good models to study how different substitutions in the major

groove modulate biophysical and biochemical properties of

AON/RNA duplexes. Herein, we report a viable synthetic route

to 6⁰,7⁰-substituted R-L-carba-LNA thymidine derivatives

(Figure 1K-M) along with a so-called “double-locked” tetra-

cyclic R-^L-carba-LNA thymidine (Figure 1N) formed by the

intramolecular addition of methylene free radical to thymine

nucleobase. In order to evaluate R-^L-carba-LNA nucleosides’

potential for therapeutic applications, all of these modified

nucleotides have subsequently been introduced into AONs to

study thermal stabilities of their duplexes with complementary

RNA or DNA, as well as the 3⁰-exonuclease resistances, human

blood serum stabilities, and RNase H recruitment capabilities in

comparison with that of R-^L-LNA and β-D-carba-LNA-mod-

ified counterparts.

known nucleoside 3,5-di-O-benzyl-4-hydroxylmethyl-1,2-O-

isopropylidene-β-^L-ribofuranose 1,⁴⁵which was oxidized to

the corresponding aldehyde 2 through Swern oxidation. The

crude aldehyde was subjected to Grignard reaction with

vinylmagnesium bromide to afford two pure diastereomers

3a (6⁰R-OH) and 3b (6⁰S-OH) in moderate yields (32% for

3a, 29% for 3b in two steps). The configuration at C6⁰could

not be confirmed by nuclear Overhauser effect (NOE) ex-

periment at this stage. This could, however, be confirmed

after the free-radical cyclization step (see section 2 for

discussion on NMR characterization), which subsequently

showed the C6⁰-R and -S stereochemistry for 3a and 3b,

respectively. Acetylation of 3a with a mixture of acetic anhy-

dride, acetic acid, and triflic acid gave the corresponding

triacetate 4a as an anomeric mixture. The crude triacetate 4a

46

€

was then subjected to a modified Vorbruggen reaction

involving in situ silylation of the thymine and subsequent

trimethylsilyl triflate mediated coupling to give exclusive R-

L-ribofuranosyl thymine derivative 5a. Deacetylation of 5a

using 30% methylamine in ethanol at room temperature

overnight gave compound 6a quantitatively, which was

treated with phenyl chlorothioformate to afford the key

free-radical precursor 7a (6⁰-R) in high yield (86%) without

any formation of 6⁰-O-phenoxythiocarbonyl (PTC) product

even when 2.5 equiv of phenyl chlorothioformate was used.

We have also similarly synthesized the second free-radical

precursor 7b (6⁰-S) from 3b (6⁰-S) (Scheme 1).

Results and Discussion

The free-radical cyclization was carried out in refluxing

anhydrous toluene with Bu₃SnH, using AIBN as the initiator,

which was added dropwise in order to avoid the side reactions.

Cyclization of 7a (6⁰-R) took place with high stereoselectivity

1. Synthesis of Diastereomerically Pure 6⁰,7⁰-Substituted

r-^L-Carba-LNAs and 6,7⁰-Methylene-Bridged r-L-Carba-

LNA-T. The synthetic route to the 6⁰,7⁰-substituted R-L-

carba-LNA thymidine phosphoramidites 16a, 16b, and 16c

is shown in Schemes 1-3. The synthesis started from the

(45) Nacro, K.; Lee, J.; Barchi, J. J.; Lewin, N. E.; Blumberg, P. M.;

Marquez, V. E. Tetrahedron 2002, 58, 5335–5345.

€

(46) Vorbruggen, H; Krolikiewicz, K.; Bennua, B. Chem. Ber. 1981, 114,

1234–1255.

(43) Seth, P. P.; Swayze, E. E. Patent WO2009/067647 A1.

(44) Seth, P. P.; Swayze, E. E. Patent WO2010/036698 A1.

6124 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

JOCArticle

SCHEME 1

SCHEME 2

to give R-L-carba-LNA nucleoside 8a (6⁰R-OH,7⁰S-CH₃)

in good yield (57%, Scheme 2). Under an identical condi-

tion, cyclization of 7b (6⁰-S), on the other hand, gave R-L-

carba-LNA nucleoside 8b (6⁰S-OH,7⁰S-CH₃) in 43% yield

along with an unexpected product 8c in a 14% yield

(Scheme 2). The mechanism of the free-radical cyclization

reaction for the formation of 8a, 8b, and 8c will be discu-

ssed in section 3.

In an effort to remove the 6⁰S-OH in compound 8b by

radical deoxygenation strategy, we attempted to esterify the

6⁰-OH using phenyl chlorothioformate but failed. This sug-

gested that 6⁰S-OH is much more inert than the 2⁰-OH group

in the pentose sugar moiety. Transformation of OH to

(methylthio)thiocarbonate has been proven to be an efficient

solution to remove an inert hydroxyl group by radical deoxy-

genation.⁴⁷Thus, 8b was treated with CS₂and MeI along

with NaH as a base at 0 °C for 4 h, giving the radical precur-

sor 9 in 64% yield. Then, compound 9 was subjected to the

standard Barton-McCombie deoxygenation⁴⁸in the pre-

sence of Bu₃SnH and AIBN to furnish 7⁰S-CH₃-R-L-carba-

LNA nucleoside 10 in 42% yield plus an unexpected bicyclo-

[2.2.1]-2⁰,6⁰-methylene-bridged hexopyranosyl nucleoside 11

in 7% yield (Scheme 2). The radical rearrangement process⁴⁹

that leads to 11 will be discussed in section 3.

In order to convert 8a/b to the corresponding phosphor-

amidites for solid-supported DNA synthesis, the 6⁰-OH

(48) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans. 1

1975, 16, 1574–1585.

(49) Zhou, C.; Plashkevych, O.; Liu, Y.; Badgujar, N.; Chattopadhyaya,

J. Heterocycles 2009, 78, 1715–1728.

(47) Mereyala, H. B.; Pola, P. Synth. Commun. 2002, 32, 2453–2458.

J. Org. Chem. Vol. 75, No. 18, 2010 6125

Li et al.

JOCArticle

SCHEME 3

group of 8a and 8b must be protected (Scheme 3). Thus

compounds 8a and 8b were treated with p-toluoyl chloride in

dry pyridine to give compounds 12a and 12b in 73 and 71%

yield, respectively. Then compounds 12a, 12b, and 10 were

subjected to debenzylation using 20% Pd(OH)₂/C and ammo-

nium formate followed by selectively protecting the 5⁰-OH with

DMTr, giving 13a, 13b, and13c, respectively. It should be noted

that debenzylation of 12b should be limited to 20 min, and a pro-

longation of reaction time led to cleavage of a 6⁰-O-(p-toluoyl)

group by the attack of the primary 5⁰-OH group.²⁸In order to

invert the configuration of 3⁰-OH, compounds 13a, 13b, and13c

were oxidized with Dess-Martin periodinane^50-52followed by

the reduction with sodium borohydride in ethanol to give 15a,

15b, and 15c, respectively, in 57-71% yields. The complete

inversion of 3⁰-OH is facilitated by the carbocyclic ring and 7⁰-

Me on the top side of the furan sugar, which leads to significant

steric hindrance, thus promoting exclusive attack of the hydride

from the bottom of the furanose sugar ring. Compounds 15a,

15b, and 15c were then phosphitylated with 2-cyanoethyl-N,N-

diisopropylphosphoraminochloridite under standard condi-

tions,^27-29giving phosphoramidites 16a, 16b, and 16c, respec-

tively, as a diastereomeric mixture in 63-71% yield.

The unexpected product 8c obtained during free-radical

cyclization of 7b was also transformed to corresponding phos-

phoramidite 22 (Scheme 4). Thus, 8c was treated with CS₂,

MeI, and NaH, leading to the successful acquisition of radical

precursor 17 in 69% yield, followed by the Barton-McCombie

deoxygenation⁴⁸affording product 18 in high yield (95%).

Debenzylation of 18 using 20% Pd(OH)₂/C and ammonium

formate, followed by 5⁰-O-dimethoxytritylation smoothly gave

19 (72% in two steps). Since oxidation of 19 with Dess-Martin

periodinane failed to give the corresponding ketone, another

mild oxidizing reagent known as TPAP (tetra-n-propylammo-

nium perruthenate) was used.⁵³Therefore, compound 19 was

treated with catalytic amount of TPAP along with the stoichio-

metric oxidant NMO (N-methylmorpholine-N-oxide) at room

temperature to afford 20, which was directly subjected to the

reduction with sodium borohydride at -20 °C, giving products

21 (22% in two steps) plus recovery of starting material 19

(19% in two steps). Subsequently, the phosphitylation of 21

was performed with 2-cyanoethyl N,N-diisopropylphosphor-

aminochloridite under standard conditions^27-29to furnish 22

(73% yield) as a diastereomeric mixture.

2. NMR Characterization of Key Intermediates Involved in

the Synthesis of r-^L-Carba-LNA Analogues. All of the key

carbocyclic nucleoside intermediates have been characterized

by ¹H, ¹³C, COSY, ¹H-¹³C HMQC, long-range ¹H-¹³Ccorre-

lation (HMBC) NMR experiments as well as by mass

spectroscopy (see Supporting Information).

The formation of bicyclic systems in compounds 8a, 8b,

11, and the tetracyclic system in compound 8c was confirmed

by HMBC and COSY experiments (discussed in details in

part II of Supporting Information).

The orientation of substituents in the carbocyclic moiety

of compound 8a and 8b was determined by 1D NOE experi-

ments (Figure 2) as well as from the vicinal coupling con-

stants evaluation. For compound 8a, irradiation of H6 of

0

thymine led to NOE enhancement for H7⁰(2.6%) (d_H6-H7

≈

0

˚

0

2.5 A) and for 6 -OH (0.9%) (d_{H6-6 -OH}≈ 3.2 A), but none

for 7⁰-CH₃and H6⁰(d_{H6-7 -CH}≈ 4.2 A, d_H6-H6≈ 4.6 A),

˚

0

3

whereas irradiation of 7⁰-CH₃led to NOE enhancement for

0

H6⁰(1.7%) (d_{H6 -7 -CH}≈ 2.5 A) and H3 (1.7%) (d_{H3 -7 -CH}

≈

˚

0

3

0

˚

0

2.4 A), but none for 6 -OH (d_{7 -CH -6 -OH}≈ 3.9 A), strongly

3

suggesting that C6⁰is in 6⁰R configuration, and C7⁰is in 7⁰S

configuration. In addition, the trans disposition of H6⁰and

H7⁰was also in agreement with the small coupling constant

(50) Bose, D. S.; Narsaiah, A. V. Synth. Commun. 1999, 29, 937–941.

(51) Chaudhari, S. S.; Akamanchi, K. G. Tetrahedron Lett. 1998, 39,

3909–3912.

(53) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis

1994, 7, 639–666.

(52) Chaudhari, S. S.; Akamanchi, K. G. Synthesis 1998, 5, 760–764.

6126 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

JOCArticle

SCHEME 4

for ³J_{6 ,7}(3.6 Hz, hence dihedral angle H6 -C6 -C7 -H7 ≈

232° according to Karplus equation). As for compound

8b, irradiation of H6 of the thymine group led to NOE

In addition, selective irradiation of H8⁰⁰of 19 (having the

same carbon skeleton as 8c) led to strong ₀NOE enhancement

0

˚

00

0

00

for H5 (5.2%) (d_H5-H8≈2.3 A) andforH2 (1.2%) (d_{H2 -H8}

≈

enhancement for H7⁰(1.4%) (d_H6-H7≈ 2.8 A) and H6

2.8 A), but none for 5-CH₃group (d_{5-CH -H8}≈ 3.8 A),

0

˚

0

00

3

0

˚

0

(2.7%) (d_H6-H7≈ 2.5 A), but none for 7 -CH₃and 6 -OH

suggesting that C5 is in 5S configuration.

The configuration of 3⁰-OH in compounds 15a/b/c and 21

was determined by 1D NOE experiment. Irradiation of H3⁰

showed 1.3, 0.6, and₀0.5% NOE enhancements for H1⁰

˚

0

(d_{H6-7 -CH}≈ 4.3 A, d_{H6-6 -OH}≈ 4.4 A), whereas irradia-

3

tion of 7⁰-CH₃leads to NOE enhancement for H3⁰(1.6%)

0

˚

0

(d_{H3 -7 -CH}≈ 2.4 A), but none for H6 (d_{H6 -7 -CH}≈ 3.8 A),

3

0

suggesting that C6⁰is in 6⁰S configuration, and C7⁰is in 7⁰S

(d_{H1 -H3}≈ 2.3 A), H2 (d_{H2 -H3}≈ 2.7 A), and H5 , respecti-

vely, in compound 15a (Figure SII.12 in Supporting Infor-

mation), 3.2 and 1.9% NOE enhancements for H1⁰and H5⁰,

respectively, in compound 21 (Figure SII.11). Irradiation of H1⁰

˚

0

configuration. Furthermore, the large coupling constant for

3

J_{6 ,7}(8.5 Hz) corresponding to the dihedral angle of H6⁰-

C6⁰-C7⁰-H7⁰≈ 26° also suggested a cis disposition of H6⁰

and H7⁰in 8b.

0

led to 1.7 and 1.2% enhancement for H2⁰(d_{H1 -H2}≈2.3 A) and

˚

0

˚

0

The stereochemistry of compound 11 was also determined

by 1D NOE experiment (Figure 2). Therefore, selective irra-

diation of H4⁰led to distinct NOE enhancement for H3⁰

H3⁰(d_{H1 -H3}≈ 2.5 A), respectively, in 15b (Figure SII.13) and

2.5 and 2.0% NOE enhancement for H2⁰and H3⁰, respectively,

in compound 15c (Figure SII.14). These observations thus

unequivocally indicated that 3⁰-OH’s are in pseudoequatorial

positions, while H3⁰and H1⁰are in pseudoaxial positions in

compounds 15a/b/c and 21.

0

˚

0

(3.8%) (d_{H3 -H4}≈ 2.3 A), and irradiation of 7 -CH₃led to

NOE enhancement for H3⁰(1.2%) (d_{H3 -7 -CH}≈ 2.4 A) and

˚

0

3

H4⁰(1.4%) (d_{H4 -7 -CH}≈ 2.4 A), respectively, which sug-

˚

0

3

gested that H3⁰and H4⁰are cis oriented and they are on the

3. Mechanism of the Free-Radical Cyclization Reaction and

Radical Rearrangement. It is known^54,55that the 5-hexenyl

intramolecular ring-closure reaction undergoes preferentially

1,5-ring closure (exo mode) over 1,6-ring closure (endo mode),

yielding the thermodynamically less stable 5-exo-cyclization

product.^53,54The previous studies of 5-hexenyl and 6-heptenyl

cyclization of the substituted β-^D-carba-LNA and β-D-carba-

ENA analogues in our lab have also suggested that the forma-

tion of bicyclic product is going predominantly via exo mode

ring closure over endo mode.^27-29After treatment of 7a with

Bu₃SnH and AIBN, the C2⁰radical is supposed to be generated

(TS1 in Scheme 5A), which should be capable of attacking the

CdC double bond from both the “top” and “bottom” faces by

the 5-exo cyclization pathway, resulting into two plausible

intermediates, TS2 and TS3 (see Scheme 5A). The optimized

structures have shown that in the TS2 state, the thymine moiety,

same face as that of 7⁰-CH₃. The observation that irradiation

of H1⁰led to NOE enhancement for H5⁰,5⁰⁰(1.1%) (d_{H1 -H5 ,5}

≈

0

00

0

˚

0

⁰00

2.8 A) but nonefor H3 (d_{H1 -H3}≈ 3.7 A) and H4 (d_{H1 -H4}

≈

0

˚

4.2 A) suggested that H5 , 5 are located on the face close to

H1⁰. Hence, both and C3⁰and C4⁰are in R configuration for

compound 11 (Figure 2).

For compound 8c, in which four new chiral centers have

been formed during a single free-radical cyclization step, the

configuration of every chiral center was also well determined

by 1D NOE experiments (see Figure 2). Selective irradiation

of H6 in the thymine moiety led to strong₀NOE enhancement

for H6⁰(4.9%) (d_H6-H6≈ 2.3 A) and H8 (2.4%) (d_H6-H8

≈

˚

⁰0

0

˚

0

2.4 A), but none for 6 -OH and H7 (d_{H6-6 -OH}≈ 4.1 A,

0

˚

0

d_H6-H7≈ 3.8 A), unequivocally suggesting that C6 is in 6S

configuration, C7⁰is in 7⁰S configuration, and C6 is in 6S

configuration, respectively. The trans disposition of H6⁰and

3

H7⁰was also consistent with the coupling constant ( J_{6 ,7}

0

=

(54) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron 1985, 41, 3925–3941.

(55) Beckwith, A. L. J. Tetrahedron 1981, 37, 3073–3100.

2.0 Hz, with dihedral angle H6⁰-C6⁰-C7⁰-H7⁰≈ 116°).

J. Org. Chem. Vol. 75, No. 18, 2010 6127

Li et al.

JOCArticle

FIGURE 2. Key NOE contacts of carbocyclic compound 8a, 8b, 8c/19, and 11; R = OH for 8c, R = H for 19 (see text for discussions), showing

6⁰R and 7⁰S configuration for compound 8a, 6⁰S and 7⁰S configuration for compound 8b, 5S and 6S configuration for compound 8c/19, and 3⁰R

and 4⁰R configuration for compound 11.

•

developing 7⁰-CH₂radical and 6⁰-OH are all occupying the

axial positions. The steric hindrance between them makes TS2

much more unstable than TS3 because, in the TS3 state the

(eq) is favored. On the other hand, in TS4, the trans orienta-

•

tion of 6⁰-OH (eq) and 7⁰-CH₂radical (ax) is favored, but

•

1,3-diaxial dispostion of thymine (ax) and 7⁰-CH₂radical

•

thymine moiety, developing 7⁰-CH₂radical and 6⁰-OH are

(ax) is unfavored. Taking together, cyclization of 7b through

both TS5 and TS6 is possible to give product 8b and 8c with

TS5 predominating since the products 8b and 8c were

obtained in a ratio of 3:1. The formation of minor product

8c starting from TS6 can be easily understood as follows:

occupying the axial, equatorial, and axial positions, respectively

(Scheme 5A). This comparison may explain why the exclusive

formation of cyclic product 8a (6⁰R,7⁰S) has been observed.

Similarly, cyclization of 7b can proceed through inter-

mediates TS5 and TS6 (Scheme 5B). In TS5, the cis orienta-

•

First, in TS6, the primary -CH₂is not stable and will be

•

tion of 6⁰-OH (eq) and 7⁰-CH₂radical (eq) is unfavored

intramolecularly trapped by the double bond of the thymine

moietybeforebeing quenchedbyBu₃SnH to give intermediate

TS7. Then, the newly formed radical center at C5 of the

because of steric hindrance, but the orientation between

•

thymine moiety (ax) and 6⁰-OH (eq) as well as 7⁰-CH₂radical

6128 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

SCHEME 5. Mechanism of the Formation of 8a, 8b, 8c, and 11 by Intramolecular Free-Radical Cyclization or Rearrangement

JOCArticle

J. Org. Chem. Vol. 75, No. 18, 2010 6129

Li et al.

JOCArticle

TABLE 1. Thermal Denaturation of Duplexes of Native, R-L-Carba-LNA Derivatives and R-L-LNA-Modified AONs with Complementary RNA or DNA^a

^aMolecular weights of all antisense sequences are conformed by MALDI-TOF mass spectrum (see Table SII. 2 in Supporting Information). A =

native adeninyl, C = cytosinyl, T = thyminyl, ‘T ’ indicates R-L-carba-LNA or R-L-LNA-modified thymidime monomer with specified structure. T_m

values measured as the maximum of the first derivative of the melting curve (A_260nmvs T) in medium salt buffer (60 mM Tris-HCl at pH 7.5, 60 mM KCl,

0.8 mM MgCl₂) with temperature range of 20 to 65 °C using 1 μM concentrations of the two complementary strands. The value of T_mgiven is the average

of two or three independent measurements. If the error of the first two measurements exceeded (0.3 °C, the third measurement was carried out to

confirm if the error is indeed within (0.3 °C. ^bΔT_mvalues were obtained by comparing the T_mvalues of AONs 2-21 with that of native AON 1. The

ΔT_m(ave) value obtained here is the average value for four AONs incorporated with the same compound at four different modification sites. ^cThe RNA

selectivity ΔΔT_mwas calculated by this equation: ΔΔT_m= ΔT_m(ave) of AON/RNA - ΔT_m(ave) of AON/DNA. ^dT_mof AON1/RNA and AON1/DNA

are adopted for T_mcomparison; therefore, their ΔT_mvalues are set to 0.

thymine moiety was reduced by Bu₃SnH from the less

sterically hindered face, giving chiral C5 in S-configuration.

Hence, tetracyclic nucleoside 8c was obtained through a radical

cyclization process, which is different from that of formation of

tetracyclic 2⁰-amino-R-L-LNA³⁹because the latter was obtained

through aza-Michael addition of the 2⁰-amino to C6 position of

thymine. Moreover, the formed 2⁰N-C6 bond in the latter was

not stable, and decomposition under basic conditions was

observed. Instead, the carbocyclic compound 8c was found to

be a much more stable tetracyclic nucleoside.

In our previous study, a radical deoxygenation of C6⁰-

OH of β-^D-carba-LNA led to an unusual 2⁰,6⁰-methylene-

bridged hexopyranosyl nucleoside,⁴⁹which was obtained

through a radical rearrangement of the generated 6⁰C^•to 4⁰C^•

radicals. The formation of compound 11 during radical deoxy-

genation of the 6⁰-OH of R-L-carba-LNA nucleoside 9 was

supposed to have gone through a similar mechanism. Thus after

treatment of compound 9 with Bu₃SnH and AIBN in refluxing

toluene, 6⁰C^•was putatively formed (Scheme 5C), which can be

reduced by Bu₃SnH directly to give product 10. On the other

hand, the 6⁰C^•could also lead to scission of the C4⁰-O4⁰bond to

give a new C6⁰dC4⁰double bond. As a result, the 4⁰C^•radical

was transformed to 4⁰O^•(TS8). Then the 4⁰O^•radical attacked

the C6⁰dC4⁰again, resulting in formation of O4⁰-C6⁰bond and

the rearrangement of 4⁰O^•radical to 4⁰C^•, which was reacted

with Bu₃SnH by the less hindered face to furnish hexopyranosyl

nucleoside 11.

4. Synthesis and Purification of r-^L-Carba-LNA Deriva-

tives and r-^L-LNA-Modified AONs. The phosphoramidites

16a/b/c and 22 as well as R-^L-LNA monomer^35,36were incor-

porated as monosubstitutions, but at four different sites in a

15-mer DNA sequence on an automated RNA/DNA synthe-

sizer. With exception of specially designed fast deprotecting

DNA monomer blocks, the standard DNA synthesis reagents

and cycles were utilized to synthesize the DNA oligos targeted

to coding region of SV 40 large T antigen.^56,57The sequences,

modification site, and structures of modifications are shown

in Table 1. All of the modified building blocks gave modest

(56) Graessmann, M.; Michaels, G.; Berg, B.; Graessmann, A. Nucleic

Acids Res. 1991, 19, 53–59.

(57) Wagner, R. W.; Matteucci, M. D.; Lewis, J. G.; Gutierrez, A. J.;

Moulds, C.; Froehler, B. C. Science 1993, 260, 1510–1513.

6130 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

JOCArticle

total coupling yield (20 to 45%). The reduction of coupling

yield probably originates from close proximity of the carbo-

cyclic or heterocyclic ring fused on the upper face of the sugar

moiety as well as the 6⁰and 7⁰substituents to the 3⁰-phosphate,

thereby interfering with the coupling and reducing coupl-

ing efficiency to some extent. Cleavage from the solid phase

support and deprotection steps were carried out by treating

the solid support with 33% aqueous ammonia at rt for 12 h (for

AONs 10-21) or at 55 °C for 72 h (deprotection of Tol group

used for protection 6⁰-OH in AONs 2-9 needs a longer

deprotection time), followed by purification through 20%

denatured PAGE, and confirmation of structural integrity

using MALDI-TOF mass spectroscopy (see Table SII.3 in

Supporting Information).

5. Thermal Denaturation Studies. The T_mvalues of du-

plexes formed by AONs 2-21 with the complementary RNA

or DNA have been measured and compared with that of the

native counterpart (Table 1). Just as in previous reports,^33-36

one R-L-LNA (type IV) modification resulted in roughly 4.4 °C

increase in T_mfor the AON/RNA hybrid. On the contrary, 7⁰R-

Me-R-^L-carba-LNA (type III) led to T_mdecrease around 3 °C/

modification. Type I and type II modifications, just like the type

III modification, also led to T_mdecrease by around 3 °C/

modification. This observation indicates that 6⁰-OH in 6⁰S-

OH-7⁰R-Me-R-L-carba-LNA (type I) and 6⁰R-OH-7⁰R-Me-R-

L-carba-LNA (type II) exert no obvious effect on T_mfor the

AON/RNA hybrid regardless of their chirality. Hence, the drop

in T_mcaused by types I, II, and III modification could be due to

steric clash of the hydrophobic 7⁰R-methyl group, which points

located in the minor groove lead to stabilization of the duplexes.

We have also found that all the R-^L-LNA and R-L-carba-LNA

derivatives are RNA selective since ΔΔT_m(ΔΔT_m= ΔT_m(ave)

of AON/RNA - ΔT_m(ave) of AON/DNA) values were found

in the range of 1.6-2.8 °C (Table 1).

Incorporation of the hyperconstrained 6,7⁰-methylene-

bridged-R-^L-carba-LNA thymidine (type V) into 15-mer oligo-

nucleotides led to dramatic decrease in thermal affinity toward

both complementary RNA and DNA (T_mdropped 7-14 °C

with RNA, and dropped 7-18 °C with DNA; see AON 18-21

in Table 1). This result was quite similar to the previous

observations that AONs modified by 5S,6R-configured tetra-

cyclic “locked LNA” also exhibit very low affinity toward

complementary RNA and DNA.³⁹It was speculated that the

loss of aromaticity in nucleobase resulted in the increased steric

bulk of the nucleobase moiety (from planar to tetrahedral

geometry at C5 and C6 position), thereby destabilizing the

duplex owing to perturbation of the base stacking of modified

nucleic acid with neighboring base pair to some extent due to

energetically unfavorable intrastrand interaction. Alterna-

tively, as a consequence of constrained glycosidic torsion angle

of type V, the nucleobase participating in hydrogen bonding

might not be disposed optimally for efficient Watson-Crick

base pairing, therefore, also impairing the duplexes stability to

some extent.

6. Circular Dichroism Analysis. 6.1. CD of AON-RNA

Duplex: CD spectra were recorded to evaluate the overall

conformation of single modified AON/DNA and AON/

RNA duplexes compared to the native DNA/DNA, RNA/

RNA, and DNA/RNA duplexes (see Figure 3). As shown in

Figure 3A, all of the single modified AON/RNA hybrids

exhibited CD profiles intermediate between the native

A-type RNA/RNA and B-type DNA/DNA duplexes, but

resemble the natural AON 1/RNA hybrid, which suggested

that the single modification with types I, II, III, IV, and V in

AONs 4/8/12/16/20 did not cause much conformational

perturbation for AON/RNA duplex compared to natural

AON 1/RNA duplex. Hence, types I, II, III, IV, and V

modified AON/RNAs could be good substrates for RNase

H digestion. Especially, we found that the hyperconstrained

tetracyclic 6,7⁰-methylene-bridged R-L-carba-LNA-Thy

(type V) modification led to less conformational perturba-

tion for AON/RNA hybrid than type I, II, III, and IV

modifications. As discussed in section 5, the conformation-

ally constrained glycosidic torsion angle of type V modifica-

tion could however impair its ability to form efficient

Watson-Crick base pair with adenine in the opposite strand,

thus this type V modified AON strand might perturb the

driving force for conformation change necessary in the

modified AON strand for forming thermodynamically stable

duplex with the target opposite RNA strand.

6.2. CD of AON-DNA Duplex: The duplexes of AONs 4/

8/12/16 with complementary DNA also showed relatively

similar global helical conformation to the native DNA/DNA

homoduplex (shown in Figure 3B). However, the CD profile

for the duplex of type V modified AON 20 with complemen-

tary DNA was found to obviously shift to shorter wave-

length (blue shift) with more intense positive Cotton peak

around 276 nm compared with the native AON 1/DNA

duplex. Hence, it seems a single type V modification can

significantly perturb the global conformation of AON/DNA

duplex.

toward vicinal 3⁰-phosphate (d_{(7 S-methyl)-3 P}≈ 4.1 A), thereby

impairing the AON/RNA thermal stability by perturbing the

hydration pattern^58-63and other stereoelectronic interations in

the major groove of the DNA/RNA duplex. Similarly declined

T_mresults were also observed in the 2⁰-N-ethyl and acetyl

functionalized 2⁰-amino-R-L-LNA-modified AONs, which

showed 3-10 °C decrease (sequence dependent) in the thermal

stability toward complementary RNA,⁴⁰once again suggesting

that the introduction of a bulky hydrophobic group in position

pointing at 3⁰-phosphate in the major groove of the AON/RNA

hybrid can cause substantial negative effects on the duplexes’

stability.

˚

0

It is noteworthy that the 6⁰,7⁰-substituted β-D-carba-LNA

derivatives were found in our previous studies^27-29to lead to

increased T_mby 2-4 °C depending on different substitutions at

6⁰and 7⁰positions. This should be compared to the observation

made in the present study showing a T_mdrop of 2-3 °C for

6⁰,7⁰-substituted R-L-carba-LNA. This significant difference in

ꢀ

thermal stabilities of β-^D-carba-LNA vis-a-visR-L-carba-LNA-

modified duplexes hints that the substitutions on the carbocyc-

lic ring of R-^L-carba-LNA located in the major groove of DNA/

RNA duplex have significantly destabilized the duplex, while

the modifications of the carbocyclic ring of β-^D-carba-LNA

(58) Teplova, M.; Minasov, G.; Teresshko, V.; Inamati, G. B.; Cook,

P. D.; Manoharan, M.; Egli, M. Nat. Struct. Biol. 1999, 6, 535–539.

(59) Egli, M.; Minasov, G.; Teplova, M.; Kumar, R.; Wengel, J. Chem.

Commun. 2001, 651–652.

(60) Schneider, B.; Patel, K.; Berman, H. M. Biophys. J. 1998, 75, 2422–

2434.

(61) Tereshko, V.; Gryaznov, S.; Egli, M. J. Am. Chem. Soc. 1998, 120,

269–283.

(62) Auffinger, P.; Westhof, E. Angew. Chem., Int. Ed. 2001, 40, 4648–

4650.

(63) Ball, P. Chem. Rev. 2008, 108, 74–108.

J. Org. Chem. Vol. 75, No. 18, 2010 6131

Li et al.

JOCArticle

The native AON was completely degraded in ∼10 min under

present conditions. All of the other modified AONs containing

R-^L-carba-LNA derivatives showed considerably improved 3⁰-

exonuclease resistance to a variable extent. Because the AONs

were singly modified at position T13, the phosphate P14 (see the

structure in Figure 4A for phosphate (P) numbering) was consi-

derably resistant to 3⁰-exonuclease cleavage. Moreover, T13

modification also improves the stability of P13 phosphate to

some extent. Therefore, two bands referring to 14-mer and 13-mer

oligos could be observed on the PAGE pictures (Figure SII.44).

However, once the P13 was cleaved, AONs were degraded to

the monomer blocks quickly, and hence no bands referring

to 12-mer to dimer oligos could be observed.

Total percentages of intergrated 14-mer and 13-mer

AONs were plotted against time points to give SVPDE

digestion curve for each selected AON in Figure 4B, and

pseudo-first-order reaction rates (Figure 4B and Table SII.2)

were obtained by fitting the curves to single-exponential

decay functions. A comparison of digestion rates of AONs

with different types of modifications exhibited the following

results and implications:

7.1. Relative 3⁰-Exonucleolytic Stabilities of r-L-Carba-

LNA-Modified AONs. The stabilities of AONs 2, 6, 10, 14,

18 toward SVPDE incubation decreased in following order:

6,7⁰-methylene-bridged R-L-carba-LNA-T (type V) modified

AON 18 (k = 0.0040 ( 0.0012 min^-1) > 7⁰R-methyl-R-L-

carba-LNA (type III) modified AON 10 (k = 0.0068 (

0.0009 min^-1) ≈ 6⁰R-hydroxyl-7⁰S-methyl-R-L-carba-LNA

(type II) modified AON 6 (k = 0.0076 ( 0.0018 min^-1) >

6⁰S-hydroxyl-7⁰S-methyl-R-L-carba-LNA (type I) modified

AON 2 (k = 0.0125 ( 0.0007 min^-1) > R-L-LNA (type IV)

modified AON 14 (k = 0.0777 ( 0.0072 min^-1) > β-D-LNA

(type VI) incorporated AON (k = 0.5331 ( 0.1800 min^-1).

Hence, all R-^L-carba-LNA analogue modified AONs have

been found to be 3⁰-exonucleolytically more stable than

parent R-^L-LNA and β-D-LNA-modified counterparts. An-

other strategy, namely, phosphorothioate backbone modifi-

cation, is also popular to produce AONs with high

nucleolytic stability, but this type of modification, on the

other hand, can lead to extensive cellular toxicity and side

effects.¹Hence, R-L-carba-LNA-modified AONs with phos-

phodiester linkage probably will show better pharmacologic

properties than AONs containing both R-^L(or β-D)-LNA

and phosphorothioate backbone modifications.

7.2. Effect of 7⁰-Me Substitution on 3⁰-Exonucleolytic

Stability. We have found that 7⁰R-methyl-R-L-carba-LNA

(type III) modified AON 10 is about 10 times more stable

than R-^L-LNA (type IV) modified AON 14, which suggests

that the replacement of the 2⁰-O- with hydrophobic methyl-

methylene function in R-^L-LNA can render significantly

positive effects on the nuclease resistance.

7.3. Effect of 6⁰R-OH Substitution (Pointing away from 3⁰-

Phosphate) on 3⁰-Exonucleolytic Stability. C6⁰-OH substitu-

tion on the R-^L-carba-LNA can modulate the stability of

modified AONs to different extent depending on the stere-

ochemical orientation of the hydroxyl group. 6⁰R-Hydroxyl

substitution led to virtually no effect since 6⁰R-hydroxyl-7⁰S-

methyl-R-^L-carba-LNA (type II) modified AON 6 have

shown very similar overall stability with the 7⁰R-methyl-R-

L-carba-LNA (type III) modified AON 10. We have also

found the 6⁰R-OH substitution can improve the nuclease

resistance for the vicinal 5⁰-phosphate as can be seen from the

FIGURE 3. CD spectra of duplexes formed by AON 1/4/8/12/16/

20 with complementary RNA (A) or DNA (B). Typical B-type

(DNA/DNA duplex) and A-type (RNA/RNA duplex) spectra are

also presented for comparison. Native DNA/DNA duplex was

formed by the native AON 1 with complementary DNA, while

native RNA/RNA duplex was formed by 3⁰-r(CUU CUU UUU

UAC UUC) with complementary RNA. The CD spectra were

measured in the same buffer as used for UV melting experiment

(Table 1) at 20 °C. The total strand concentration is 10 μM.

7. Comparison of 3⁰-Exonuclease Stabilities of AONs

Containing r-^L-Carba-LNA Derivatives with the Parent r-L-

LNA. It is known that 3⁰-exonuclease activity is predomi-

nantly responsible for enzymatic degradation of AON in serum-

containing medium or in various eukaryotic cell lines, and

modifications located at 3⁰-terminus can significantly contribute

to the nuclease resistance of an oligonucleotide.⁶⁴The stabilities

of the newly synthesized AONs (AON 2, 6, 10, 14, 18) with a

single modification (types I, II, III, IV, V) at position T13 (posi-

tion 3 from 3⁰-end) toward 3⁰-exonuclease (phosphodiesterase I

from Crotalus adamateas venom [SVPDE] used in this study)

were measured. Thus the selected AONs were labeled at the

5⁰-end with ³²P-ATP/kinase and then incubated with SVPDE

[SVPDE 6.7 ng/μL, AON 3 μM, 100 mM Tris-HCl (pH 8.0), 15

mM MgCl₂, total volume 30 μL] at 21 °C. Aliquots were taken

out at appropriate time intervals and analyzed by 20% denatur-

ing PAGE. The gel pictures were obtained upon autoradiogra-

phy and are shown in Figure SII.44 in Supporting Information.

(64) Maier, M. A.; Leeds, J. M.; Balow, G.; Springer, R. H.; Bharadwaj,

R.; Manoharan, M. Biochemistry 2002, 41, 1323–1327.

6132 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

JOCArticle

FIGURE 4. (A) Molecular structure of T13 modified AONs. The full sequence is 3⁰-d (CTT¹³CTT TTT TAC TTC-³²P)-5⁰. (B) Amount of

remaining initial oligonucleotide (taken 14-mer and 13-mer together in the calculation of percentage remaining) during 3⁰-exonuclease

(SVPDE) promoted digestion. The digestion conditions were used as follows: AON 3 μM (5⁰-end ³²P-labeled with specific activity 80 000 cpm),

100 mM Tris-HCl (pH 8.0), 15 mM MgCl₂, SVPDE 6.7 ng/μL, reaction temperature 21 °C, total reaction volume 30 μL.

band corresponding to 13-mer oligo which was observed in

PAGE for type II modified AON 6 while the 13-mer band

was not present for type I modified AON 2 and type III mo-

dified AON 10 (Figure SII.44 in Supporting Information).

Effectively the 6⁰R-OH protects the 5⁰-phosphate by inter-

fering with binding of SVPDE to this phosphate.³²Indeed

the 6⁰R-OH in 6⁰R-hydroxyl-7⁰S-methyl-R-L-carba-LNA

(type II) has been found located relatively close to the 5⁰-

upon incorporation into modified AON 2, to significantly less

3⁰-exonucleolytic stable duplex than the duplex incorporat-

ing the 7⁰R-methyl-R-L-carba-LNA (type III) modification

(AON 10).

7.5. Effect of Tetracyclic System on 3⁰-Exonucleolytic

Stability. 6,7⁰-Methylene-bridged R-L-carba-LNA-T (type V)

modified AON 18 showed better 3⁰-exonucleolytic resis-

tance than type I, II, and III modified AONs. It can be

concluded that the extra six-membered ring of 6,7⁰-methy-

lene-bridged R-^L-carba-LNA-T (type V), which lies above

the pentose sugar, gave slightly higher nuclease resistance

compared to other R-^L-carba-LNA derivatives (types I, II,

and III).

Comparison of 3⁰-exonucleolytic resistance of R-L-carba-

LNA-modified AONs with that of β-^D-LNA and β-D-

carba-LNA-modified counterparts has also been carried out.

Corresponding PAGE, 3⁰-exonuclease (SVPDE) promoted

digestion plot and discussion can be found in part II of

Supporting Information.

˚

0

phosphate (d_{5 R-OH-5 P}≈ 4.0 A).

7.4. Effect of 6⁰S-OH Substitution (Pointing at 3⁰-Phos-

phate) on 3⁰-Exonucleolytic Stability. Previous study³²has

shown that in β-^D-carba-LNA 6⁰-OH substitution can re-

markably reduce 3⁰-exonucleolytic stability of modified AONs

if this OH points at the vicinal 3⁰-phosphate probably because

of its assistance in the departure of 3⁰-oxyanion during SVPDE

mediated 3⁰O-P bond scission.³²Similar conclusions can be

reached in the present investigation as 6⁰S-hydroxyl substitu-

tion [OH group is also pointing at the vicinal 3⁰-phosphate]

in 6⁰S-hydroxyl-7⁰S-methyl-R-L-carba-LNA (type I) leads,

J. Org. Chem. Vol. 75, No. 18, 2010 6133

Li et al.

JOCArticle

8. Relationship between 3⁰-Exonuclease Stability and Sol-

vation Free Energy for r-^L-Carba-LNA Derivatives and Par-

ent r-^L-LNA. The solvation properties of modified AON

affect the relative access of the water molecule to the scissile

phosphate, therefore giving variable extent of 3⁰-exonuclease

resistance. In order to check the relationship between solva-

tion capability and nuclease stability of AONs with different

modifications, we have analyzed the solvation free energy of

R-^L-carba-LNA derivatives (types I, II, III) as well as that of

the parent R-^L-LNA (type IV) (Table SII.1) employing

Baron and Cossi’s implementation of polarizable conductor

CPCM model⁶⁵of solvation on the ab initio optimized

(HF,6-31G** basis set) gas phase molecular geometries.

The solvation free energies obtained have clearly demon-

strated relative hydrophilic nature of different 2⁰,4⁰-con-

strained modifications in R-^L-carba-LNA and R-L-LNA.

The energy of solvation of R-^L-carba-LNA derivatives

(types I, II, III) and R-^L-LNA (type IV) decreases in the

following order: 7⁰R-methyl-R-L-carba-LNA (type III,

-3.04 kcal mol^-1) > 6⁰R-hydroxyl-7⁰S-methyl-R-L-carba-

LNA (type II, -4.30 kcal mol^-1) > 6⁰S-hydroxyl-7⁰S-meth-

yl-R-^L-carba-LNA (type I, -4.35 kcal mol^-1) > R-L-LNA

(type IV, -8.51 kcal mol^-1). This trend shows the R-L-carba-

LNA derivatives are not as well solvated as the parent R-^L-

LNA, which suggests the scissile phosphate of R-^L-LNA is

relatively more solvated as compared to that of R-^L-carba-

LNA derivatives. The magnitudes of 3⁰-exonuclease stability

of these R-^L-2⁰,4⁰-constrained nucleoside modified AONs

have also shown the following decreasing trend: 7⁰R-meth-

yl-R-^L-carba-LNA (type III, k = 0.0068 ( 0.0009 min^-1) ≈

6⁰R-hydroxyl-7⁰S-methyl-R-L-carba-LNA (type II, k =

0.0076 ( 0.0018 min^-1) > 6⁰S-hydroxyl-7⁰S-methyl-R-L-

carba-LNA (type I, k = 0.0125 ( 0.0007 min^-1) . R-L-

LNA (type IV, k = 0.0777 ( 0.0072 min^-1), which is fully

consistent with the results from solvation free energy calcu-

lation. This corroborated our original hypothesis that the

hydration around a scissile phosphate was most probably

critical for the nuclease-promoted hydrolysis.²⁷

9. Stability of Functionalized r-L-Carba-LNA-Modified

AONs in Human Blood Serum. The newly synthesized AONs

with a single modification types I, II, III, IV, and V at

position T13 (position 3 from 3⁰-end) were assayed for

stability in human blood serum. The AONs (5⁰-end ³²P-

labeled) were incubated with human blood serum (male, type AB)

for up to 48 h at 21 °C, and aliquots were taken out at

regular time intervals and then analyzed by 20% denaturing

PAGE. The gel pictures obtained by autoradiography are

shown in Figure SII.43. Due to the presence of alkaline

phosphatase in blood serum that gradually removes the

5⁰-end ³²P label, it was impossible to obtain accurate degra-

dation rate for each AON by quantifying the gel picture.

Visual comparison of the gel pictures have shown that 6⁰S-

hydroxyl-7⁰S-methyl-R-L-carba-LNA (type I) modified

AON 2 and 7⁰R-methyl-R-L-carba-LNA (type III) modified

AON 10 can be sustained in blood serum for more than

48 h. These modifications result in more stable AONs com-

pared to 6⁰R-hydroxyl-7⁰S-methyl-R-L-carba-LNA (type

II) modified AON 6 and 6,7⁰-methylene-bridged R-L-carba-

LNA-T (type V) modified AON 18, which is sustained for

36 h in blood serum. The R-^L-LNA (type IV) modified AON

14 exhibited the least stability in human blood serum. The

order of relative stabilities of AONs in human blood serum

has been found to be similar to that observed upon treat-

ment by 3⁰-exonucleases.

10. RNase H Digestion Studies of Functionalized r-^L-

Carba-LNA-Modified AON/RNA Duplexes. It has been

reported that RNase H binds in the minor groove of the

DNA/RNA hybrid.^66,67Hence R-L-carba-LNA modifica-

tions are supposed to have relatively minor effect on the

RNase H recruitment than β-^D-carba-LNA modification

because the 2⁰,4⁰-carbocyclic ring in the former is located in

the major groove but the carbocyclic ring in the latter is

located in the minor groove of the AON/RNA hybrid.

Hence, the RNase H recruitment study of four newly

synthesized R-^L-carba-LNA derivative modified AONs

(containing a single type I, II, III, or V modification) as well

as the parent R-^L-LNA-modified AON (type IV) has been

carried out and compared with the native AON 1 as well as

with the β-^D-carba-LNA-modified AONs.^27-29

The gel pictures in Figure SII.37-39 have shown that all

the modified AONs 2-21/RNA hybrids are good substrates

for RNase H. For type I, II, III, and IV modified AON/RNA

duplexes, the cleavage patterns have been found to be very

similar and independent of the nature of the modification

but dependent on the site of modification. As shown in

Figure 5A, the cleavage activity of RNase H was suppressed

within a stretch of a 5 base pairs region that starts from the

modification site toward the 3⁰-end in the RNA strand. In

addition, the original preferred A8 cleavage site also shifted

to the edges of suppressed region if A8 is included within this

5 bp suppressed area. Hence, the observed RNase H cleavage

patterns for type I, II, III, and IV modified AON/RNA

duplexes were found to be very closely similar to the previous

studies of β-^D-carba-LNA,^27-29β-D-carba-ENA,^27,28and

β-^D-aza-ENA^20,21modified AON/RNA hybrids, suggesting

that RNase H does not make any distinction on whether the

2⁰,4⁰-carbocylcic ring is located in the minor or the major

groove, it renders very similar cleavage pattern.

However, it is interesting to note that tetracyclic 6,7⁰-

methylene-bridged R-^L-carba-LNA (type V) modified

AON 19/RNA and AON 20/RNA hybrids showed different

cleavage patterns compared to the type I, II, III, and IV

modified AON/RNA duplexes at the same position: a stretch

of 3 bp region was observed for RNase H cleavage suppres-

sion in AON 19/RNA duplex, whereas a stretch of 5 bp

suppressed region was also observed in AON 20/RNA

duplex but with different preferred cleavage site and sup-

pressed stretch region (see Figure 5B). This result hints that

type V modification led to relatively less conformational

perturbation for AON/RNA hybrid than type I, II, III, and IV

modifications, which is consistent with the result obtained by

CD study.

All the R-^L-carba-LNA derivative modified AON/RNA

duplexes were degraded by RNase H with comparable

cleavage rates as digestion of native AON/RNA hybrid

(Figure 6). Remarkably, the cleavage rates of types III

(7⁰R-methyl-R-L-carba-LNA) and IV (R-L-LNA) modified

AON/RNA duplex were even 2 times higher than cleavage of

(66) Nowotny, M.; Gaidamakov, S. A.; Crouch, R. J.; Yang, W. Cell

2005, 121, 1005–1016.

(67) Nowotny, M.; Cerritelli, S. M.; Ghirlando, R.; Gaidamakov, S. A.;

Crouch, R. J.; Yang, W. EMBO J. 2008, 27, 1172–1181.

(65) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995–2001.

6134 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

JOCArticle

FIGURE 5. Escherichia coli RNase H1 promoted cleavage pattern of AONs 1-21/RNA heteroduplexes. Vertical arrows show the RNase H

cleavage sites, with the relative length of the arrow indicating the extent of the cleavage. The square boxes show the stretch of the modification,

which is resistant to RNase H1 cleavage, thereby giving footprints.

the native counterpart. Given that β-D-carba-LNA-modified

AON/RNA duplexes generally showed less RNase H diges-

tion efficiency than the native counterpart,^28,29we could

arrive at the conclusion that R-^L-carba-LNA modifications

led to much less effect on RNase H elicitation than β-^D-

carba-LNA modification. This is because of the fact that the

stereochemical location of the hydrophobic 2⁰,4⁰-carbocyclic

ring in β-^D-carba-LNA is in the minor groove, which retards

the RNase H binding in the minor groove, but, in contrast,

that is not the case for the R-^L-carba-LNA since its 2⁰,4⁰-

carbocyclic ring is located in the major groove.

2. The thermal denaturation study of duplexes formed by

AON containing a single R-^L-carba-LNA derivatives

with complementary RNA revealed that T_mof 6⁰,7⁰-

substituted R-^L-carba-LNA-modified AON/RNA is

about 3 °C lower than that of the native AON/RNA

and 7 °C lower than that of the parent R-^L-LNA-modified

counterpart. This suggests that the hydration perturba-

tion in the major groove caused by the substitution of R-^L-

carba-LNA results in more negative effects on duplex

thermal stability compared to the substitution of β-^D-

carba-LNA derivatives in minor groove.

3. Substituted R-^L-carba-LNA-modified AONs have

been found in presence of both snake venom phospho-

diesterase and human blood serum to be more stable

than the parent R-^L-LNA and β-D-LNA-modified

counterparts.

4. R-^L-Carba-LNA derivatives and R-L-LNA-modified

AON/RNA duplexes have shown similar RNase H

digestion pattern as the β-^D-carba-LNA-modified

counterparts, but better digestion efficiency. Especially,

the cleavage rates of types III (7⁰R-methyl-R-L-carba-

LNA) and IV (R-^L-LNA) modified AON/RNA duplex

were, in some sequences, even 2 times faster than

cleavage rates of the native counterpart, suggesting the

R-^L-carba-LNA modification in which the 2⁰,4⁰-carbo-

cyclic ring is located in the major groove can lead to a

positive effect on the RNase H digestion efficiency of

AON/RNA.

Conclusion

1. Herein, we report convenient synthetic routes toward

the R-^L-carba-LNA thymidine derivatives. The synth-

esis of these novel R-^L-ribo-configured conformatio-

nally constrained carbocyclic analogues has been

achieved by employing free-radical scavenging by

a tethered olefin in an intramolecular reaction as a

key step. Various NMR experiments, including ¹H,

¹³C, ¹H-¹H COSY, one-bond ¹H-¹³C correlation

(HMQC), and long-range H-¹³C HMBC have been

1

carried out to characterize all synthesized compounds

unambiguously. The relative chirality of substituents

in the key intermediates was determined by 1D NOE

and was also corroborated by the ³J_HHcoupling con-

stants obtained from H-H homo or double decoupling

experiments.

J. Org. Chem. Vol. 75, No. 18, 2010 6135

Li et al.

JOCArticle

FIGURE 6. Bar plots of the observed cleavage rates of the RNase H1 promoted degradation of RNA in AON 1-21/RNA hybrid duplexes.

Implication. The obvious merit of introduction of R-L-carba-

LNA derivatives into AONs is that they can improve nuclease

resistance while rendering positive effect on the RNase H

digestion. Unfortunately, all four newly synthesized R-^L-carba-

LNA derivatives in the present study led to lower RNA affinity

(around -3 °C/modification compared to native counterpart).

As discussed in section 5, the decrease in the RNA affinity pro-

bably originates from the steric clash with the hydrophobic 7⁰R-

methyl group in R-^L-carba-LNA, which points toward vicinal 3⁰-

phosphate, thereby impairing the AON/RNA thermal stability

by perturbing the hydration pattern in the major groove of the

DNA/RNA duplex. Work is in progress to replace 7⁰R-methyl

group with less steric and less hydrophobic hydrogen atom as in

parent R-^L-carba-LNA [structure (O) in Figure 1] which upon

introduction into AONs should hopefully lead to a more

effective antisense therapeutic candidate, thereby will further

add to the value of R-^L-carba-LNA as a modified nucleoside in

the AON or siRNA. We anticipate that this parent R-^L-carba-

LNA should have (1) higher target RNA affinity, just like R-^L-

LNA, (2) significantly increased enzymatic stability compared

to R-^L-LNA, and (3) better RNase H elicitation efficiency than

R-^L-LNA, β-D-LNA, and β-D-carba-LNA derivatives.

O-isopropylidene-β-^L-ribofuranose (3b). Oxalyl chloride (2.84 mL,

32 mmol) was added to the precooled dichloromethane

(92 mL) at -78 °C. Then DMSO (3.90 mL, 54 mmol) in dichlor-

omethane (8 mL) was added dropwise to the solution over 30 min.

After stirring for 20 min, a solution of 1 (5.20 g, 13.0 mmol) in

dichloromethane(24mL) was addeddropwiseover about20 min,

and the mixture was kept stirring at -78 °C for another 30 min.

DIPEA (16 mL) was added to this cooled mixture. The reaction

solutionwasthenallowedtowarmtoroomtemperature andstirred

for 1 h whereupon water (10 mL) was added. The organic layer was

separated, washed with water (50 mL) and brine (50 mL), dried

over MgSO₄, and concentrated under reduced pressure to give the

corresponding aldehyde 2 as a yellowish oil. This oil was dissolved

in THF (130 mL) under nitrogen and cooled to 0 °C. Vinyl magne-

sium bromide (1.0 M solution in THF, 26 mL, 26 mmol) was added,

and then the reaction was allowed to warm to room temprature and

kept stirring at this temperature overnight. The reaction was slowly

quenched with saturated ammonia chloride solution (10 mL). Then

the mixture was partitioned between ethyl acetate (150 mL) and

water (100 mL), the organic layer was then washed with saturated

NaHCO₃and brine, dried over MgSO₄, and concentrated under

reduced pressure. The crude material was purified by column chro-

matography on silica gel (0-20% ethyl acetate in cyclohexane, v/v)

to give 3a (1.80 g, 32%) and 3b (1.61 g, 29%) as colorless oil. 3a: ¹H

NMR (500 MHz, CDCl₃þ DABCO) δ7.3-7.2 (10H, m, Bn), 5.73

(1H, m, H7), 5.69 (1H, d, J_{H1, H2}= 3.5 Hz, H1), 5.28 (1H, d,

Experimental Section

J=17.5 Hz, H8), 5.13 (1H, d, J=10.5 Hz, H8⁰), 4.60 (1H, d, J_gem

11.5 Hz, CH₂Bn), 4.57 (1H, d, J_gem=11.5 Hz, CH₂Bn), 4.53 (1H, d,

=

3,5-Di-O-benzyl-4-C-(1R-hydroxylallyl)-1,2-O-isopropylidene-β-

L-ribofuranose (3a) and 3,5-Di-O-benzyl-4-C-(1S-hydroxylallyl)-1,2-

6136 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

JOCArticle

J_gem=11.5 Hz, CH₂Bn), 4.52 (1H, d, J_gem=11.5 Hz, CH₂Bn),

4.48 (1H, m, H6), 4.45 (1H, m, H2), 4.05 (1H, d, J_{H2, H3}= 5.0 Hz,

H3), 3.99 (1H, d, J_gem= 11.0 Hz, H5), 3.74 (1H, d, J_gem= 11.0 Hz,

H5⁰), 2.74 (s, DABCO), 1.43 (3H, s, CH₃), 1.23 (3H, s, CH₃); ¹³C

NMR (125 MHz, CDCl₃þ DABCO) δ 137.2 (Bn), 136.6 (Bn),

133.7 (C7), 127.3-126.5 (aromatic), 116.0 (C8), 112.4 (isopropyl),

103.3 (C1), 87.7 (C4), 77.8 (C2), 76.3 (C3), 72.6 (C6), 71.6 (CH₂Bn),

71.3 (CH₂Bn), 69.2 (C5), 45.9 (DABCO), 25.7 (CH₃), 25.1 (CH₃);

MALDI-TOF m/z [M þ Na]^þ, found 449.199, calcd 449.194. 3b:

¹H NMR (500 MHz, CDCl₃) δ 7.3-7.2 (10H, m, Bn), 5.84 (1H, m,

H7), 5.59 (1H, d, J_{H1, H2}= 3.0 Hz, H1), 5.31 (1H, d, J = 17.5 Hz,

H8), 5.15 (1H, d, J=11.0 Hz, H8⁰), 4.68 (1H, d, J_gem= 12.0 Hz,

δ 7.3-7.2 (Bn), 7.03 (1H, s, H6), 5.69 (1H, m, H7⁰), 5.36 (1H, d,

0

J = 16.5 Hz, H8⁰), 5.34 (1H, d, J_{H1 , H2}= 5.4 Hz, H1 ), 5.15 (1H,

d, J = 10.5 Hz, H8⁰⁰), 4.61 (1H, d, J_gem= 11.0 Hz, CH₂Bn), 4.49

(1H, m, H2⁰), 4.49 (1H, d, J_gem= 11.0 Hz, CH₂Bn), 4.46 (1H, m,

0

H3⁰), 4.42 (1H, d, J_gem= 11.0 Hz, CH₂Bn), 4.41 (1H, d, J_gem

=

0

11.0 Hz, CH₂Bn), 4.24 (1H, d, J_{H6 , H7}= 5.5 Hz, H6 ), 3.77 (1H,

d, J_gem= 10.0 Hz, H5⁰), 3.76 (1H, s, 2⁰-OH), 3.42 (1H, s,

6⁰-OH), 3.37 (1H, d, J_gem= 10.0 Hz, H5⁰⁰), 1.83 (3H, s, 5-CH₃);

¹³C NMR (125 MHz, CDCl₃) δ 162.5 (C4), 149.6 (C2), 138.6

(C6), 136.4 (aromatic), 135.8 (aromatic), 133.8 (C7⁰), 127.6-

126.9 (aromatic), 116.9 (C8⁰), 110.0 (C5), 97.4 (C1⁰), 89.2 (C4⁰),

74.2 (C3⁰), 73.2 (C6⁰), 72.8 (CH₂Bn), 72.1 (CH₂Bn), 71.9 (C2⁰),

70.6 (C5⁰), 11.2 (5-CH₃); MALDI-TOF m/z [M þ Na]^þ, found

517.200, calcd 517.195.

CH₂Bn), 4.56 (1H, d, J_gem= 12.0 Hz, CH₂Bn), 4.53 (1H, d, J_gem

=

12.0 Hz, CH₂Bn), 4.52 (1H, d, J_gem= 12.0 Hz, CH₂Bn), 4.46 (1H, t,

H2), 4.35 (1H, m, H6), 4.06 (1H, d, J_gem= 10.5 MHz, H5), 3.99

(1H, d, J_{H2, H3}= 5.5 Hz, H3), 3.81 (1H, d, J_gem= 10.5 MHz, H5⁰),

2.79 (1H, d, J= 6.0 Hz 6-OH), 1.45 (3H, s, CH₃),1.25(3H, s, CH₃);

¹³C NMR (125 MHz, CDCl₃) δ 136.9 (Bn), 136.7 (Bn), 135.0 (C7),

127.3-126.6 (aromatic), 115.9 (C8), 112.6 (isopropyl), 103.0 (C1),

86.1 (C4), 78.7 (C2), 76.3 (C3), 73.8 (C6), 72.8 (CH₂Bn), 71.4

(CH₂Bn), 70.4 (C5), 25.9 (CH₃), 25.3 (CH₃); MALDI-TOF m/z

[M þ Na]^þ, found 449.198, calcd 449.194.

1-[3,5-Di-O-benzyl-4-C-(1R-hydroxylallyl)-6-hydroxyl-2-O-

phenoxylthiocarbonyl-r-^L-ribofuranosyl]thymine (7a). Com-

pound 6a (872 mg, 1.70 mmol) was coevaporated twice with

anhydrous pyridine and dissolved in the same solvent (38 mL).

The solution was cooled by ice bath, and then phenyl chlor-

othionoformate (0.29 mL, 2.10 mmol) was added dropwise,

while temperature was maintained at 0 °C during addition.

After 3 h of stirring at room temperature, pyridine was

recovered under reduced pressure. The residue was dissolved

in dichloromethane (50 mL) and washed with saturated solu-

tion of NaHCO₃(20 mL). The organic layer was separated,

dried over MgSO₄, and concentrated in vacuo. Crude product

was chromatographed over silica gel (10-25% ethyl acetate in

cyclohexane, v/v) to afford 7a (920 mg, 86%) as a white foam:

¹H NMR (500 MHz, CDCl₃) δ 8.61 (1H, br s, NH), 7.30-7.17

(13H, m, aromatic), 6.86 (2H, d, J = 8.0 Hz, aromatic), 5.94

1-[4-C-(1R-Acetoxyallyl)-2,6-di-O-acetyl-3,5-di-O-benzyl-r-

L-ribofuranosyl]thymine (5a). Acetic anhydride (2.44 mL, 25

mmol) and acetic acid (85 mL) were added to 3a (1.80 g, 4.2

mmol). The mixture was cooled with ice bath, and triflic acid

(0.02 mL, 0.14 mmol) was added. After stirring for 2 h at room

temperature, the reaction was quenched with cold saturated

NaHCO₃solution (20 mL) and extracted with dichloro-

methane (50 mL ꢀ 3), the organic layer was separated, dried

over MgSO₄, and evaporated to dryness to furnish crude pro-

duct 4a, which was coevaporated with anhydrous CH₃CN

twice and dissolved in the same solvent. Thymine (720 mg,

5.70 mmol) and N,O-bis(trimethylsilyl)acetamide (2.9 mL, 11.8

mmol) were added to the solution and refluxed for 45 min until

the suspension became a clear solution. Then reaction mixture

was cooled to 0 °C, and TMSOTf (1.1 mL, 6.10 mmol) was

added dropwise, then the reaction was refluxed overnight.

CH₃CN was evaporated under reduced pressure. To the residue

was added saturated ammonia chloride solution (10 mL), and

the mixture was extracted with dichloromethane (50 mL). The

organic layer was dried over MgSO₄, evaporated, and chro-

matographed over silica gel (0-1% methanol in dichlorome-

thane, v/v) to give 5a as a white foam (1.84 g, 76%): ¹H NMR

(1H, app t, H2⁰), 5.81 (1H, d, J_{H1 , H2}= 6.0 Hz, H1⁰), 5.70

0

(1H, m, H7⁰), 5.46 (1H, d, J = 17.0 Hz, H8⁰), 5.17 (1H, d, J =

0

10.5 Hz, H8⁰⁰), 4.63 (1H, d, J_{H2 , H3}=6.5 Hz, H3 ), 4.59 (1H, d,

0

J_{H6 , H7}= 5.5 Hz, H6 ), 4.56 (1H, d, J_gem= 12.0 Hz, CH₂Bn),

4.55 (1H, d, J_gem= 12.0 Hz, CH₂Bn), 4.45 (1H, d, J_gem= 12.0

Hz, CH₂Bn), 4.44 (1H, d, J_gem=12.0 Hz, CH₂Bn), 3.73 (1H, d,

_gem= 11.0 Hz, H5⁰), 3.52 (1H, d, J_gem= 11.0 Hz, H5⁰⁰), 3.29

J

(1H, bs, 6⁰-OH), 1.85 (3H, s, 5-CH₃); ¹³C NMR (125 MHz,

CDCl₃) δ 193.2 (C=S), 152.3 (C4), 149.4 (C2), 137.5 (C6),

136.6 (aromatic), 136.2 (aromatic), 133.8 (C7⁰), 128.6-125.8

(aromatic), 120.8 (aromatic), 116.7 (C8⁰), 110.4 (C5), 89.5

(C4⁰), 89.3 (C1⁰), 80.0 (C2⁰), 73.8 (C3⁰), 73.7 (CH₂Bn), 72.9

(CH₂Bn), 71.9 (C6⁰), 69.0 (C5⁰), 11.4 (5-CH₃); MALDI-TOF

m/z [M þ Na]^þ, found 653.197, calcd 653.193.

(600 MHz, CDCl₃þ DABCO) δ 7.3-7.1 (Bn), 6.95 (1H, s, H6),

(1S,3R,4S,5S,6R,7S)-7-Benzyloxy-1-benzyloxymethyl-6-hy-

droxyl-5-methyl-3-(thymin-1-yl)-2-oxabicyclo[2.2.1]heptane (8a).

Compound 7a (974 mg, 1.55 mmol) was dissolved in 140 mL of

anhydrous toluene that was purged with N₂for half an hour. The

mixure was heated to reflux and Bu₃SnH (0.41 mL in 12 mL of

anhydrous toluene, 1.55 mmol) and AIBN (293 mg in 24 mL

anhydrous toluene, 1.55 mmol) were added dropwise in four

portions over 4 h. Then the refluxing was continued for another

1 h. Solvent was evaporated, and the residue was purified by

column chromatography on silica gel (20-40% ethyl acetate in

cyclohexane, v/v) to give 8a (420 mg, 57%) as a white solid: ¹H

NMR (600 MHz, CDCl₃) δ 8.63 (1H, br s, NH), 8.16 (1H, ₀s, H6),

0

6.00 (1H, d, J_{H1 , H2}=5.4 Hz, H1 ), 5.70 (1H, d, J_{H6 , H7}= 4.8

Hz, H6⁰), 5.62 (1H, m, H7⁰), 5.26 (1H, d, J = 16.8 Hz, H8⁰), 5.16

(1H, m, H2⁰), 5.16 (1H, d, J = 10.8 Hz, H8⁰⁰), 4.55 (1H, d,

J_gem= 12.6 Hz, CH₂Bn), 4.44 (1H, d, J_gem= 12.6 Hz, CH₂Bn),

4.43 (1H, d, J_gem= 12.6 Hz, CH₂Bn), 4.34 (1H, d, J_gem= 12.6

Hz, CH₂Bn), 4.32 (1H, m, H3⁰), 3.69 (1H, d, J_gem= 10.8 Hz,

H5⁰), 3.50 (1H, d, J_gem= 10.8 Hz, H5⁰⁰), 2.79 (s, DABCO), 2.06

(3H, s, acetyl-CH₃), 1.94 (3H, s, acetyl-CH₃), 1.82 (3H, s,

5-CH₃); ¹³C NMR (150 MHz, CDCl₃þ DABCO) δ 169.3

(C=O, acetyl), 167.9 (C=O, acetyl), 162.8 (C4), 149.5 (C2),

136.5 (aromatic), 135.9 (aromatic), 134.3 (C6), 130.4 (C7⁰),

127.5-126.8 (aromatic), 118.3 (C8⁰), 110.5 (C5), 86.8 (C4⁰),

85.3 (C1⁰), 74.9 (C3⁰), 73.5 (C6⁰), 73.4 (CH₂Bn), 72.8 (C2⁰), 72.7

(CH₂Bn), 67.9 (C5⁰), 45.4 (DABCO), 20.2 (acetyl-CH₃), 19.6

(acetyl-CH₃), 11.6 (5-CH₃); MALDI-TOF m/z [M þ Na]^þ,

found 601.217, calcd 601.216.

0

7.30-7.18 (10H, m, Bn), 6.16(1H, d, J_{H1 , H2}=2.4Hz, H1 ), 4.63

(1H, d, J_gem= 12.0 Hz, CH₂Bn), 4.62 (1H, d, J_gem= 12.0 Hz,

CH₂Bn), 4.45 (1H, d, J_gem= 12.0 Hz, CH₂Bn), 4.44 (1H, d,

J

_gem= 12.0 Hz, CH₂Bn), 3.86 (1H, d, J_{H2 , H3}= 1.5 Hz, H3⁰),

0 0

3.81 (1H, m, H6⁰), 3.81 (1H, d, J_gem= 10.2 Hz, H5⁰), 3.74 (1H, d,

J

_gem= 10.2 Hz, H5⁰⁰), 2.83 (1H, s, 6⁰-OH), 2.63 (1H, m, H2⁰),

1-[3,5-Di-O-benzyl-2,6-dihydroxyl-4-C-(1R-hydroxylallyl)-r-

L-ribofuranosyl]thymine (6a). Compound 5a (1.0 g, 1.70 mmol)

was treated with 30% methylamine solution in ethanol (60 mL)

at room temperature overnight. Then the reaction mixture was

evaporated to dryness and purified by column chromatography

on silica gel (0-3% methanol in dichloromethane, v/v) to give

6a (872 mg, 100%) as a white solid: ¹H NMR (500 MHz, CDCl₃)

1.86 (3H, s, 5-CH₃), 1.56 (1H, m, H7⁰), 0.95 (3H, d, J_{7 -CH3, H7}

0

=

7.8 Hz 7⁰-CH₃); ¹³C NMR (125 MHz, CDCl₃) δ 163.1 (C4), 149.3

(C2), 137.8 (C6), 136.2 (aromatic), 136.1 (aromatic), 127.6-126.6

(aromatic), 107.6 (C5), 88.6 (C1⁰), 88.1 (C4⁰), 80.4 (C3⁰), 78.9

(C6⁰), 73.1 (CH₂Bn), 70.8 (CH₂Bn), 68.1 (C5⁰), 47.8 (C2⁰), 32.3

(C7⁰), 25.9 (cyclohexane, coming from chromatography elution

J. Org. Chem. Vol. 75, No. 18, 2010 6137

Li et al.

JOCArticle

solvent), 17.8 (7⁰-CH₃), 11.4 (5-CH₃); MALDI-TOF m/z

4.49 (1H, d, J_gem= 12.0 Hz, CH₂Bn), 4.48 (1H, d, J_gem= 12.0 Hz,

CH₂Bn), 4.13 (1H, s, H3⁰), 3.70 (1H, d, J_gem= 9.5 Hz, H5⁰),

3.62 (1H, d, J_gem= 9.5 Hz, H5⁰⁰), 2.79 (1H, s, H2⁰), 2.36 (3H, s,

Tol-CH₃), 1.65 (1H, ₀m, H7⁰), 1.46 (3H, s, 5-CH₃), 1.20 (3H, d,

[M þ Na]^þ, found 501.203, calcd 501.200.

(1S,3R,4S,5S,6S,7S)-7-Benzyloxy-1-benzyloxymethyl-6-hy-

droxyl-5-methyl-3-(thymin-1-yl)-2-oxabicyclo[2.2.1]heptane (8b)

and (1R,6S,7S,9R,10S,11S,12S,13S)-12-Benzyloxy-11- benzyloxy-

methyl-2,4-diaza-10-hydroxyl-6-methyltetracyclo[9,2,1,0^2,70^9,13]-tet-

radecan-3,5-dione (8c). Compound 7b (3.0 g, 4.76 mmol) was

dissolved in 410 mL of anhydrous toluene that was purged with

N₂for 30 min prior to use. The mixture was heated to reflux and

Bu₃SnH (1.26 mL in 44 mL anhydrous toluene, 4.76 mmol) and

AIBN (899 mg in 92 mL of anhydrous toluene, 4.76 mmol) were

added dropwise in four portions over 4 h. Then the refluxing was

continued for another 1 h. Solvent was evaporated, and the residue

was purified by column chromatography on silica gel (20-40%

ethyl acetate in cyclohexane, v/v, then 0-0.5% methanol in di-

chloromethane, v/v) to give 8b (986 mg, 43%) and 8c (310 mg, 14%)

as white solids. 8b:¹H NMR (500 MHz, CDCl₃) δ8.74(1H, s, NH),

J_{7 -CH3, H7}=7.0Hz7 -CH₃);¹³CNMR(125MHz, CDCl₃) δ164.9

(Tol-CdO), 162.9 (C4), 149.0 (C2), 143.3 (aromatic), 136.5

(aromatic), 136.3 (aromatic), 136.2 (C6), 128.6-125.8 (aromatic),

107.9 (C5), 88.7 (C1⁰), 88.3 (C4⁰), 78.9 (C3⁰), 78.0 (C6⁰), 72.6

(CH₂Bn), 71.1 (CH₂Bn), 62.4 (C5⁰), 48.2 (C2⁰), 31.9 (C7⁰), 20.7

(Tol-CH₃), 17.4 (7⁰-CH₃), 11.0 (5-CH₃); MALDI-TOF m/z

[M þ H]^þ, found 597.262, calcd 597.260.

0

(1S,3R,4S,5S,6R,7S)-1-(4,4⁰-Dimethoxytrityloxymethyl)-7-hy-

droxyl-5-methyl-6-(4-methylbenzoate)-3-(thymin-1-yl)-2-oxabicy-

clo[2.2.1]heptane (13a). To a solution of compound 12a (252 mg,

0.420 mmol) in anhydrous methanol (10 mL) were added 20%

Pd(OH)₂/C (315 mg) and ammonium formate (1.60 g, 25.5

mmol), and the mixture was refluxed for 2 h. The suspension

was filtered over Celite, and the organic phase was evaporated in

vacuo to give crude 12a⁰, which was coevaporated twice with

anhydrous pyridine and dissolved in the same solvent (6 mL).

4,4⁰-Dimethoxytrityl chloride (282 mg, 0.84 mmol) was added,

and the mixture was stirred overnight at room temperature. Then

solvent was removed, and the residue was diluted with dichloro-

methane (10 mL), washed with aqueous saturated NaHCO₃

solution (10 mL), and dried over MgSO₄. After evaporation of

solvent, the residue was subjected to column chromatography on

silica gel (0.3-1.2% methanol in dichloromethane containing

1% pyridine, v/v) to afford 13a (225 mg, 75%): ¹H NMR (500

MHz, CDCl₃þ DABCO) δ 7.73 (1H, s, H6), 7.71 (2H, m,

0

7.36 (1H, s, H6), 7.30-7.23 (Bn), 5.95 (1H, d, J_{H1 , H2}= 2.5 Hz,

H1⁰),4.63(1H, d,J_gem= 11.5 Hz, CH₂Bn), 4.62 (1H, d, J_gem=11.5

Hz, CH₂Bn), 4.48 (1H, d, J_gem= 11.5 Hz, CH₂Bn), 4.47 (1H, d,

0

_gem= 11.5 Hz, CH₂Bn), 4.15 (1H, s, J_{H2 , H3}= 1.5 Hz, H3 ), 3.97

0

J

(1H, dd, J_{H6 , H7}= 8.5 Hz, J_{H6 , 6 -OH}= 3.0 Hz, H6⁰), 3.90 (1H, d,

0

J

_gem= 11.0 Hz, H5⁰), 3.81 (1H, d, J_gem= 11.0 Hz, H5⁰⁰), 2.68 (1H,

d, J_{H6 , 6 -OH}= 3.0 Hz, 6⁰-OH), 2.66 (1H, m, H2⁰), 1.92 (1H, m,

0

H7⁰), 1.86 (3H, s, 5-CH₃), 0.89 (3H, d, J_{7 -CH3, H7}=7.5Hz7 -CH₃);

¹³C NMR (125 MHz, CDCl₃): δ 162.9 (C4), 148.8 (C2), 136.4

(aromatic), 136.3 (aromatic), 134.9 (C6), 127.6-126.7 (aromatic),

108.3 (C5), 88.1 (C4⁰), 87.9 (C1⁰), 80.3 (C3⁰), 72.9 (C6⁰), 71.3

(CH₂Bn), 71.3 (CH₂Bn), 66.3 (C5⁰), 46.3 (C2⁰), 29.1 (C7⁰), 25.9

(cyclohexane, coming from chromatography solvent), 11.7

(7⁰-CH₃), 11.4 (5-CH₃); MALDI-TOF m/z [M þ Na]^þ, found

501.202, calcd 501.200. 8c:¹H NMR (500 MHz, CDCl₃) δ7.77 (1H,

0

aromatic), 7.33-7.08 (11H, m, aromatic), 6.68 (4H, m, aromatic),

0

6.14 (1H, d, J_{H1 , H2}= 2.5 Hz, H1 ), 4.89 (1H, d, J_{H6 , H7}= 4.0

Hz, H6⁰), 4.47 (1H, d, J_{H2 , H3}= 1.0 Hz, H3⁰), 3.62 (3H, s,

CH₃O), 3.59 (3H, s, CH₃O), 3.48 (1H, d, J_gem= 10.0 Hz, H5⁰),

3.37 (1H, d, J_gem= 10.0 Hz, H5⁰⁰), 2.76 (s, DABCO), 2.72 (1H, s,

0

s, NH), 7.28-7.18 (10H, m, Bn), 6.36 (1H, d, J_{H1 , H2}= 2.0 Hz,

H1⁰),4.57(1H, d,J_gem= 12.0 Hz, CH₂Bn), 4.56 (1H, d, J_gem=12.0

Hz, CH₂Bn), 4.49 (1H, d, J_gem= 12.0 Hz, CH₂Bn), 4.41 (1H, d,

H2⁰), 2.36 (3H, s, Tol-CH₃), 1.63 (1H, m, H7⁰), 1.41 (1H, s,

5-CH₃), 1.26(3H, d, J_{7 -CH3, H7}=7.5 Hz7 -CH₃); ¹³C NMR(125

MHz, CDCl₃þ DABCO) δ 164.9 (Tol-CdO), 163.2 (C4), 157.5

(aromatic), 149.4 (C2), 143.4 (aromatic), 136.1 (C6), 134.5

(aromatic), 134.1 (aromatic), 129.0-124.3 (aromatic), 112.2

(aromatic), 107.9 (C5), 88.8 (C1⁰), 85.7 (C4⁰), 79.0 (OCPh₃),

78.9 (C6⁰), 73.3 (C3⁰), 56.9 (C5⁰), 54.1 (CH₃O), 50.8 (C2⁰), 45.7

(DABCO), 32.1 (C7⁰), 20.7 (Tol-CH₃), 17.6 (7⁰-CH₃), 10.9 (5-

CH₃); MALDI-TOF m/z [M þ Na]^þ, found 741.283, calcd

741.279.

0

J

_gem= 12.0 Hz, CH₂Bn), 4.04 (1H, d, J_{H2 , H3}= 2.0 Hz, H3 ), 3.84

(1H, d, J_gem= 10.5 Hz, H5⁰), 3.75 (1H, app t, H6⁰), 3.74 (1H, d,

_gem= 10.5 Hz, H5⁰⁰), 3.29 (1H, ddd, J_{H5, H6}= 10.5 Hz, J_{H8 , H6}

=

0

J

00

0

4.0 Hz, J_{H8 , H6}= 12.0 Hz, H6), 3.09 (1H, d, J_{H6 , 6 -OH}= 2.0 Hz,

6⁰-OH), 2.29 (1H, m, J_{H5, H6}= 10.5 Hz, J_{H5, 5-CH3}= 7.0 Hz, H5),

0

00

2.19 (1H, app tt, J_{H8 , H6}= 4.0 Hz, J_{H8 , H7}= 3.5 Hz, J_{H8 , H8}=

0

13.5 Hz, H8⁰), 2.15 (1H, m, H2⁰), 1.95 (1H, m, J_{H8 , H7}= 3.5 Hz,

0

00

0

00

J_{H8 , H7}= 4.0 Hz, J_{H6 , H7}=2.0 Hz, H7 ), 1.35 (1H, m, J_{H8 , H6}

=

12.0 Hz, J_{H8 , H7}= 4.0 Hz, J_{H8 , H8}= 13.5 Hz, H8 ), 1.18 (1H, d,

00

0

00

J

_{H5, 5-CH3}= 7.0 Hz, 5-CH₃); ¹³C NMR (125 MHz, CDCl₃) δ 169.6

(1S,3R,4S,5S,6R,7R)-1-(4,4⁰-Dimethoxytrityloxymethyl)-

7-hydroxyl-5-methyl-6-(4-methylbenzoate)-3-(thymin-1-yl)-2-

oxabicyclo[2.2.1]heptane (15a). Compound 13a (115 mg, 0.160

mmol) was dissolved in anhydrous dichloromethane (4 mL),

and 15% Dess-Martin periodinane in CH₂Cl₂(0.67 mL, 0.320

mmol) was added dropwise under nitrogen. After stirring at

room temperature for 3 h, the reaction was quenched by

aqueous saturated Na₂S₂O₃(1 mL) and NaHCO₃solution

(1 mL), extracted with dichloromethane (10 mL ꢀ 3). The organic

layer was then dried over MgSO₄and concentrated to dryness to

give 14a. The obtained crude product 14a was dissolved in 95%

ethanol (3 mL), and NaBH₄(12 mg, 0.320 mmol) was added in

portions. The mixture was allowed to stir at room temperature

for 1 h. Then the solvent was removed, and the residue was

extracted with dichloromethane. The organic layer was washed

with aqueous saturated NaHCO₃solution, dried over MgSO₄,

and concentrated under reduced pressure. The residue obtained

was subjected to short column chromatography on silica gel

(0.3-1.2% methanol in dichloromethane containing 1% pyr-

(C4), 152.7 (C2), 136.8 (aromatic), 136.3 (aromatic), 127.6-126.6

(aromatic), 84.5 (C4⁰), 81.1 (C3⁰), 80.3 (C1⁰), 74.4 (C6⁰), 72.9

(CH₂Bn), 71.3 (CH₂Bn), 66.6 (C5⁰), 47.8 (C6), 40.2 (C5), 39.0

(C2⁰), 36.2 (C7⁰), 30.2 (C8⁰), 9.8 (5-CH₃); MALDI-TOF m/z [M þ

Na]^þ, found 501.201, calcd 501.200.

(1S,3R,4S,5S,6R,7S)-7-Benzyloxy-1-benzyloxymethyl-5-methyl-

6-(4-methylbenzoate)-3-(thymin-1-yl)-2-oxabicyclo[2.2.1]heptane

(12a). Compound 8a (285 mg, 0.60 mmol) was coevaporated

with anhydrous pyridine twice and dissolved in the same solvent

(10 mL). The mixture was cooled with an ice bath, and 4-methyl

benzoyl chloride (0.12 mL, 0.90 mmol) was added dropwise to

this precooled solution. The mixture was allowed to stir at room

temperature for 6 h. Pyridine was recovered under reduced pressure,

and the residue was dissolved in dichloromethane (10 mL). The

obtained solution was washed with saturated NaHCO₃aqueous

solution (10 mL), dried over MgSO₄, and concentrated under

reduced pressure to give crude product, which was subjected to

short column chromatography on silica gel (15-30% ethyl acetate

in cyclohexane, v/v) to afford 12a (260 mg, 73%): ¹H NMR (500

MHz, CDCl₃) δ 8.69 (1H, br s, NH), 7.82 (1H, s, H6), 7.80 (2H, m,

1

idine, v/v) to furnish 15a (66 mg, 57%): H NMR (500 MHz,

CDCl₃þ DABCO) δ 7.78 (2H, m, aromatic), 7.72 (1H, s, H6),

0

aromatic), 7.29-7.10 (12H, m, aromatic), 6.06 (1H, d, J_{H1 , H2}

=

7.34-7.09 (11H, m, aromatic), 6.71 (4H, m, aromatic), 5.84

0

2.5 Hz, H1⁰), 4.91 (1H, d, J_{H6 , H7}= 4.0 Hz, H6⁰), 4.67 (1H, d,

(1H,d, J_{H1 , H2}=2.5Hz,H1 ),5.25(1H,d,J_{H6 , H7}=4.0Hz,H6 ),

0

J_gem= 12.0 Hz, CH₂Bn), 4.60 (1H, d, J_gem= 12.0 Hz, CH₂Bn),

4.49 (1H, s, H3⁰), 3.66 (3H, s, CH₃O), 3.64 (3H, s, CH₃O), 3.49

6138 J. Org. Chem. Vol. 75, No. 18, 2010

Li et al.

JOCArticle

(2H, dd, J_gem= 10.5 Hz, 10.5 Hz, H5⁰, 5⁰⁰), 2.75 (s, DABCO),

2.68 (1H, s, H2⁰), 2.36 (3H, s, Tol-CH₃), 1.77 (1H, m, H7⁰), 1.51

(1H, d, J_gem= 11.5 Hz, CH₂Bn), 4.27 (1H, dd, J_{H2 , H3}= 4.0

0

Hz, J_{H3 , H4}= 10.0 Hz, H3⁰), 3.69 (2H, d, H5⁰, 5⁰⁰), 2.94 (1H, d,

0

(1H, s, 5-CH₃), 1.39 (3H, d, J_{7 -CH3, H7}= 7.0 Hz 7⁰-CH₃); ¹³C

NMR (125 MHz, CDCl₃þ DABCO) δ 165.1 (Tol-C=O), 163.2

(C4), 157.6 (aromatic), 149.5 (C2), 143.3 (aromatic), 135.5 (C6),

134.3 (aromatic), 134.2 (aromatic), 129.0-124.3 (aromatic),

112.2 (aromatic), 108.0 (C5), 87.7 (C4⁰), 86.2 (C1⁰), 85.8

(OCPh₃), 81.7 (C6⁰), 77.2 (C3⁰), 58.9 (C5⁰), 54.1 (CH₃O), 48.9

(C2⁰), 45.6 (DABCO), 34.3 (C7⁰), 20.7 (Tol-CH₃), 17.4 (7⁰-CH₃),

11.0 (5-CH₃); MALDI-TOF m/z [M þ Na]^þ, found 741.281,

calcd 741.279.

0

J_{H2 , H3}=4.0 Hz, H2 ), 2.52 (1H, m, J_{H3 , H4}=10.0 Hz, J_{H4 , H6}

0

=

1.5 Hz, H4⁰), 2.06 (1H, m, J_{H6 , H7}= 4.0 Hz, J_{H7 , 7 -CH3}= 7.0 Hz,

0

H7⁰), 1.88 (3H, s, 5⁰-CH₃), 0.92 (3H, d, J_{7 -CH3, H7}= 7.0 Hz 7⁰-

CH₃); ¹³C NMR (125 MHz, CDCl₃) δ163.1 (C4), 148.8 (C2), 137.5

(aromatic), 137.0 (aromatic), 135.1 (C6), 127.4-126.5 (aromatic),

108.0 (C5), 83.1 (C1⁰), 81.6 (C6⁰), 73.6 (C3⁰), 72.5 (CH₂Bn), 70.5

(CH₂Bn), 64.9 (C5⁰), 47.7 (C2⁰), 40.8 (C4⁰), 36.4 (C7⁰), 11.7 (5-

CH₃), 8.9 (7⁰-CH₃); MALDI-TOF m/z [M þ H]^þ, found 463.219,

calcd 463.223.

0

(1S,3R,4S,5S,6S,7S)-7-Benzyloxy-1-benzyloxymethyl-5-methyl-

6-((methylthio)thiocarbonyl)oxy-3-(thymin-1-yl)-2-oxabicyclo[2.2.1]-

heptane (9). To a solution of compound 8b (190 mg, 0.40 mmol) in

dry THF (6 mL) was added 60% NaH (48 mg, 1.19 mmol) at 0 °C,

After stirring at rt for 1 h, CS₂(0.24 mL, 3.97 mmol) was added to the

suspension at 0 °C, and the resulting mixture was added to stirred at

rt for another 1 h. The solution became clear and was cooled by ice.

Methyl iodide (0.16 mL, 2.60 mmol) was added at 0 °C, and the

mixture was stirred at rt for 2 h. The reaction was quenched by chilled

water (2 mL) and extracted with ethyl acetate (10 mL) three times.

The organic layer was dried over MgSO₄and evaporated to dryness.

The residue was chromatographed over silica gel (15-35% ethyl

General Procedure for Phosphoramidite Synthesis. To a solu-

tion of substrate (1 equiv) in dry dichloromethane were added

DIPEA (6 equiv) and 2-cyanoethyl N,N-diisopropyl phosphor-

amidochloridite (3 equiv) dropwise in an ice bath. The reaction

was allowed to warm to rt and stirred at this temperature for 3 h.

After being quenched with methanol, the mixture was diluted by

dichloromethane, washed with saturated NaHCO₃solution,

dried over MgSO₄, and concentrated. The residue was subjected

to short column chromatography on silica gel to give phosphor-

amidite, which was first precipitated in n-hexane and then dried

over P₂O₅on vacuum for 3 days before it was used for DNA

synthesis.

1

acetate in cyclohexane) to give compound 9 (145 mg, 64%): H

Oligonucleotide Synthesis and Purification. All AONs were

synthesized using an automated DNA/RNA synthesizer based

on phosphoramidite chemistry. For native A, G, and C building

block, fast deprotecting phosphoramidites (Ac for C, iPr-Pac

for G, Pac for A) were used. Standard DNA synthesis reagents

and cycle were used except that 0.25 M 5-[3,5-bis(trifluoro-

methyl)phenyl]-1H-tetrazole (activator 42) was used as the

activator and Tac₂O as the cap A. For incorporation of mod-

ified nucleotides, extended coupling time (10 min comparing to

25 s for native nucleotides) was used. For AONs 2-9, the

deprotections were carried out in 33% aqueous NH₃for 72 h

at 55 °C (the longer deprotection times were used to ensure

complete removal of the Tol group). Other oligoes (AONs 1 and

10-21) were deprotected at room temperature by treatment

with 33% aqueous NH₃for 12 h. After deprotection, all crude

oligos were purified by denaturing PAGE (20% polyacrylamide

with 7 M urea), extracted with 0.3 M NaOAc, and desalted with

a C-18 reverse phase cartridge to give AONs in >99% purity,

and correct masses have been obtained by MALDI-TOF mass

spectroscopy for each of them.

RNA was also synthesized by a solid-supported phosphoram-

idite approach based on the 2⁰-O-TEM strategy.^68,69

UV Melting Experiments. Determination of the T_mof the

AON/RNA hybrids or AON/DNA duplex was carried out in

the following buffer: 60 mM Tris-HCl (pH 7.5), 60 mM KCl, 0.8

mM MgCl₂. Absorbance was monitored at 260 nm in the

temperature range from 20 to 65 °C using a UV spectrophot-

ometer equipped with a Peltier temperature programmer with

the heating rate of 1 °C/min. Prior to measurements, the samples

(1 μM of AON and 1 μM complementary DNA or RNA

mixture) were preannealed by heating to 80 °C for 5 min

followed by slow cooling to 21 °C and 30 min equilibration at

this temperature. The value of T_mis the average of two or three

independent measurements. If error of the first two measure-

ments is >(0.3 °C, the third measurement was carried out to

check if the error is indeed within (0.3 °C; otherwise, it is

repeated.

NMR (600 MHz, CDCl₃) δ 8.77 (1H, s, NH), 7.52 (1H, s, H6),

0

7.38-7.28 (10H, m, aromatic), 6.10 (1H, d, J_{H1 , H2}= 1.8 Hz, H1 ),

5.89 (1H, d, J_{H6 ,H7}= 8.4 Hz, H6⁰), 4.74 (1H, d, J_gem= 11.4 Hz,

CH₂Bn), 4.69 (1H, d, J_gem= 11.4 Hz, CH₂Bn), 4.60 (1H, d, J_gem=

0

11.4 Hz, CH₂Bn), 4.47 (1H, d, J_gem=11.4 Hz, CH₂Bn), 4.21 (1H, d,

0

J_{H2 , H3}=1.2Hz, H3 ), 3.84(1H, d, J_gem=11.4Hz, H5⁰), 3.76 (1H,

0

d, J_gem= 11.4 Hz, H5⁰⁰), 2.87 (1H, m, H2⁰), 2.48 (3H, s, SCH₃), 2.29

(1H, m, H7⁰), 1.99 (3H, s, 5-CH₃), 0.90 (3H, d, J_{7 -CH3, H7}= 7.2 Hz

7⁰-CH₃); ¹³C NMR (150 MHz, CDCl₃) δ 213.8 (C=S), 162.8 (C4),

148.9 (C2), 136.7 (aromatic), 136.0 (aromatic), 134.7 (C6), 127.6-

126.7 (aromatic), 108.9 (C5), 88.3 (C4⁰), 87.8 (C1⁰), 81.1 (C3⁰), 78.7

(C6⁰), 72.8 (CH₂Bn), 71.4 (CH₂Bn), 65.3 (C5⁰), 46.4 (C2⁰), 29.6

(C7⁰), 18.1 (CH₃S), 12.1 (7⁰-CH₃), 11.7 (5-CH₃); MALDI-TOF m/z

[M þ Na]^þ, found 591.163, calcd 591.160.

0

(1S,3R,4S,5S,7S)-7-Benzyloxy-1-benzyloxymethyl-5-methyl-

3-(thymin-1-yl)-2-oxabicyclo[2.2.1]heptane (10) and (1R,3R,4S,

5S,6S,7S)-5-Benzyloxy-6-benzyloxymethyl-7-methyl-3-(thymin-

1-yl)-2-oxabicyclo[2.2.1]heptane (11). Compound 9 (130 mg,

0.230 mmol) was dissolved in dry toluene (6 mL) and purged

with dry nitrogen for 30 min. AIBN (8.5 mg, 0.050 mmol) and n-

Bu₃SnH (0.19 mL, 0.680 mmol) were added, and the reaction

mixture was refluxed for 1.5 h. The mixture was cooled to rt.

After evaporation of the solvent, the residue was chromato-

graphed over silica gel (15-40% ethyl acetate in cyclohexane) to

1

give compounds 10 (45 mg, 42%) and 11 (9 mg, 7%). 10: H

NMR (500 MHz, CDCl₃) δ 8.79 (1H, s, NH), 7.56 (1H, s, H6),

0

7.39-7.28 (10H, m, aromatic), 6.09 (1H, d, J_{H1 , H2}= 2.0 Hz,

=

H1⁰), 4.74 (1H, d, J_gem= 10.0 Hz, CH₂Bn), 4.65 (1H, d, J_gem

10.0 Hz, CH₂Bn), 4.59 (1H, d, J_gem= 10.0 Hz, CH₂Bn), 4.58

(1H, d, J_gem= 10.0 Hz, CH₂Bn), 4.06 (1H, s, H3⁰), 3.75 (1H, d,

J

_gem= 8.5 Hz, H5⁰), 3.72 (1H, d, J_gem= 8.5 Hz, H5⁰⁰), 2.78 (1H,

s, H2⁰), 2.02 (1H, dd, J_{H6 , H7}= 8.0 Hz, J_{H6 , H6}= 11.5 Hz,

0

00

H6⁰), 1.96 (3H, s, 5-CH₃), 1.70 (1H, m, H7⁰), 1.54 (1H, dd,

J_{H6 , H7}= 4.0 Hz, J_{H6 , H6}= 11.5 Hz, H6⁰⁰), 1.00 (3H, d,

0

00

J_{7 -CH3, H7}= 6.0 Hz, 7⁰-CH₃); ¹³C NMR (125 MHz, CDCl₃) δ

162.9 (C4), 148.9 (C2), 136.9 (aromatic), 136.6 (aromatic), 135.0

(C6), 127.4-126.6 (aromatic), 108.0 (C5), 88.6 (C1⁰), 88.2 (C4⁰),

81.3 (C3⁰), 72.7 (CH₂Bn), 71.0 (CH₂Bn), 66.8 (C5⁰), 46.9 (C2⁰),

37.4 (C6⁰), 22.7 (C7⁰), 19.8 (7⁰-CH₃), 11.7 (5-CH₃); MALDI-

TOF m/z [M þ Na]^þ, found 485.208, calcd 485.205. 11: ¹H

NMR (500 MHz, CDCl₃) δ 8.74 (1H, s, NH), 7.41 (1H, s, H6),

7.31-7.19 (10H, m, aromatic), 5.80 (1H, s, H1⁰), 4.80 (1H, d,

0

CD Spectroscopy. CD spectra were recorded from 300 to 200

nm in 0.2 cm path length cuvettes. Spectra were obtained with an

AON/RNA or AON/DNA duplex concentration of 10 μM in

(68) Zhou, C; Honcharenko, D.; Chattopadhyaya, J. Org. Biomol. Chem.

2007, 5, 333–343.

(69) Zhou, C.; Pathmasiri, W.; Honcharenko, D.; Chatterjee, S.; Barman,

J.; Chattoapdhyaya, J. Can. J. Chem. 2007, 85, 293–301.

J

_gem= 11.5 Hz, CH₂Bn), 4.51 (1H, d, J_gem= 11.5 Hz, CH₂Bn),

4.45 (1H, d, J_gem= 11.5 Hz, CH₂Bn), 4.41 (1H, s, H6⁰), 4.35

J. Org. Chem. Vol. 75, No. 18, 2010 6139

Li et al.

JOCArticle

60 mM Tris-HCl (pH 7.5), 60 Mm KCl, 0.8 mM MgCl₂. All

spectra were measured at 20 °C, and each spectrum is an average

of five experiments from which the CD value of the buffer was

subtracted.

10 mM MgCl₂, 0.1 mM EDTA, and 0.1 mM DTT at 21 °C in the

presence of 0.01 U E. coli RNase H (obtained from USB

Corporation, Cleveland, OH). Prior to the addition of the

enzyme, reaction components were preannealed in the reaction

buffer by heating at 80 °C for 5 min followed by slow cooling to

21 °C and 30 min equilibration at this temperature. Total

reaction volume was 30 μL. Aliquots of 3 μL were removed

after 5, 10, 15, 30, and 60 min, and the reactions were terminated

by mixing with stop solution [containing 0.05 M EDTA, 0.05%

(w/v) bromophenol blue, and 0.05% (w/v) xylene cyanole in

80% formamide]. The samples were subjected to 20% 7 M urea

PAGE and visualized by autoradiography. Pseudo-first-order

reaction rates could be obtained by fitting the digestion curves to

single-exponential decay functions.

³²P Labeling of Oligonucleotides. The oligoribonucleotides

and oligodeoxyribonucleotides were 5⁰-end-labeled with ³²P

using T4 polynucleotide kinase, [γ-³²P]ATP, and the standard

protocol. Labeled AONs and the target RNA were purified by

QIAquick Nucleotide Removal Kit, and specific activities were

measured using a Beckman LS 3801 counter.

SVPDE Degradation Studies. Stability of the AONs toward

3⁰-exonucleases was tested using phosphodiesterase I from

Crotalus adamanteus (obtained from USB Corporation, Cleve-

land, OH). All reactions were performed at 3 μM DNA con-

centration (5⁰-end ³²P-labeled with specific activity 80 000 cpm)

in 100 mM Tris-HCl (pH 8.0) and 15 mM MgCl₂at 21 °C. An

exonuclease concentration of 6.7 ng/μL was used for digestion

of oligonucleotides. Total reaction volume was 30 μL. Aliquots

(3 μL) were taken at proper time points and quenched by

addition of stop solution (4 μL) [containing 0.05 M EDTA,

0.05% (w/v) bromophenol blue, and 0.05% (w/v) xylene cya-

nole in 80% formamide]. Reaction progress was monitored by

20% denaturing (7 M urea) PAGE and autoradiography.

Stability Studies in Human Blood Serum. AONs at 2 μM

concentration (5⁰-end ³²P-labeled with specific activity 80 000

cpm) were incubated in 10 μL of human blood serum (male AB,

obtained from Sigma-Aldrich) at 21 °C (total reaction volume

was 36 μL). Aliquots (3 μL) were taken at proper time points and

quenched with 4 μL of stop solution [containing 0.05 M EDTA,

0.05% (w/v) bromophenol blue, and 0.05% (w/v) xylene cya-

nole in 80% formamide], resolved in 20% polyacrylamide

denaturing (7 M urea) gel electrophoresis and visualized by

autoradiography.

Acknowledgment. Generous financial support from the

Swedish Natural Science Research Council (Vetenskapsradet),

the Swedish Foundation for Strategic Research (Stiftelsen

€

for Strategisk Forskning), and the EU-FP6 funded RIGHT

project (Project No. LSHB-CT-2004-005276) is gratefully

acknowledged.

Supporting Information Available: ¹H, ¹³C, ³¹P, DEPT,

COSY, HMQC, HMBC of R-^L-carba-LNA derivatives (part I);

1D NOE spectra, ¹H homo or double decoupling spectra of key

intermediates 5a, 5b, 8a, 8b, 8c, 10, 11, 19, 21, 15a, 15b, and 15c

(part II); autoradiograms of 20% denaturing PAGE as well as

degradation curves, showing the cleavage kinetics of target

RNA in AON/RNA hybrids by E. coli RNase H1; synthesis

and NMR characterization of intermediates 5b, 6b, 7b, 12b, 13b,

15b, 13c, 15c, 17, 18, 19, 21 as well as the characterization for

final amidites 16a, 16b, 16c, and 22 (part II); chemical models of

AON/RNA hybrids containing one R-^L-carba-LNA or β-D-

carba-LNA modification (part II). This material is available

free of charge via the Internet at http://pubs.acs.org.

RNase H Digestion Assay. Target 0.1 μM RNA (specific

activity 80 000 cpm) and AON (2 μM) were incubated in

a buffer containing 20 mM Tris-HCl (pH 7.5), 20 mM KCl,

6140 J. Org. Chem. Vol. 75, No. 18, 2010

Article Doi

Free-radical ring closure to conformationally locked α-l-carba-LNAs and synthesis of their oligos: Nuclease stability, target RNA specificity, and elicitation of RNase H

DOI: 10.1021/jo100900v

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jo100900v

Full text of DOI:10.1021/jo100900v

Products guided by the article

R&D Labs maybe for 1242670-98-9

Relevant to this article

Hot Product