High-quality oligo-RNA synthesis using the new 2′-O-TEM protecting group by selectively quenching the addition of p-tolyl vinyl sulphone to exocyclic amino functions

Article abstract of DOI:10.1139/V07-025

During the F^--promoted deprotection of the oligo-RNA, synthesized using our 2′-O-(4-tolylsulfonyl)ethoxymethyl (2′-O-TEM) group [Org. Biomol. Chem. 5, 333 (2007)], p-tolyl vinyl sulphone (TVS) is formed as a by-product. The TVS formed has been

Full text of DOI:10.1139/V07-025

293
High-quality oligo-RNA synthesis using the new 2′-
O-TEM protecting group by selectively quenching
the addition of p-tolyl vinyl sulphone to exocyclic
amino functions1
Chuanzheng Zhou, Wimal Pathmasiri, Dmytro Honcharenko,
Subhrangsu Chatterjee, Jharna Barman, and Jyoti Chattopadhyaya
Abstract: During the F–-promoted deprotection of the oligo–RNA, synthesized using our 2′-O-(4-
tolylsulfonyl)ethoxymethyl (2′-O-TEM) group [Org. Biomol. Chem. 5, 333 (2007)], p-tolyl vinyl sulphone (TVS) is
formed as a by-product. The TVS formed has been shown to react with the exocyclic amino functions of adenosine
(A), guanosine (G), and cytidine (C) of the fully deprotected oligo–RNA to give undesirable adducts, which are then
1
purified by HPLC and unambiguously characterized by H, 13C Heteronuclear Multiple Bond Correlation (HMBC)
NMR and mass spectroscopic analysis. The relative nucleophilic reactivities of the nucleobases toward TVS have been
found to be the following: N6–A > N4–C > N2–G > > N3–U. This reactivity of TVS toward RNA nucleobases to give
various Michael adducts could, however, be suppressed by using various amines as scavengers. Among all these
amines, morpholine and piperidine are the most efficient scavenger for TVS, which gave highly pure oligo–RNA even
in the crude form and can be used directly in RNA chemical biology studies.
Key words: RNA synthesis, RNA alkylation, p-tolyl vinyl sulphone, Michael addition.
Résumé : Au cours de la déprotection sous l’influence de l’ion fluorure de l’oligo–ARN synthétisé à l’aide de notre
groupe protecteur 2′-O-(4-tolylsulfonyl)éthoxyméthyl (2′-O-TEM) [Org. Biomol. Chem. 5, 333 (2007)], il y a formation
de la p-tolyl vinyl sulfone (TVS) comme sous-produit. Il a été montré que la TVS réagit avec les fonctions aminées
exocycliques de l’adénosine (A), de la guanosine (G) et de la cytidine (C) de l’oligo-ARN complètement déprotégé
pour conduire à la formation de produits indésirables qui ont été purifiés et caractérisés sans ambiguïté par spectro-
1
métrie de masse et par corrélation des liaisons multiples hétéronucléaires (CLMH) des spectres RMN du H et du 13C.
On a établi que les réactivités nucléophiles relatives des nucléobases vis-à-vis de la TVS sont dans l’ordre suivant: N6–
A > N4–C > N2–G > > N3–U. En utilizant diverses amines comme pièges, il serait toutefois possible d’éliminer cette
réactivité de la TVS vis-à-vis des nucléobases pour conduire à divers adduits de Michael. Parmi ces amines, la mor-
pholine et la pipéridine sont les pièges les plus efficaces pour la TVS; leur utilization conduit, même dans le produit
brut, à la formation d’oligo–ARN très pur qui peut être utilizé dans des études de chimie biologique de l’ARN
Mots-clés : synthèse d’ARN, alkylation d’ARN, p-tolyl vinyl sulfone, addition de Michael.
[Traduit par la Rédaction] Zhou et al. 301
Introduction
ology for the chemical synthesis of RNA (3–7). Recently,
we have developed a new RNA synthesis strategy based on
protecting 2′-OH with a 2-(4-tolylsulfonyl)ethoxymethyl (2′-
O-TEM) (8) group and showed its successful use in the
solid-phase synthesis of at least 12 different oligo–RNAs of
varying sizes (14–38 mer long) and sequences (8). The main
advantages of this 2′-O-TEM group based RNA synthesis
strategy (8) are (i) the low coupling time (120s), (ii) high
The discovery of RNA interference (RNAi) (1) and espe-
cially the finding that small interfering RNA (siRNA) (2)
can induce degradation of the target RNA have highlighted
the need for facile chemical synthesis of oligo–RNA se-
quences (e.g., a synthetic 21 mer siRNA) (2) and have
thereby stimulated work on the development of new method-
Received 5 February 2007. Accepted 13 March 2007. Published on the NRC Research Press Web site at canjchem.nrc.ca on
27 April 2007.
In honor of the career contributions of Professor Kelvin Kenneth Ogilvie and for his significant contributions in the field of
nucleoside/nucleic acid chemistry and automated DNA and RNA synthesis.
C. Zhou, W. Pathmasiri, D. Honcharenko, S. Chatterjee, J. Barman, and J. Chattopadhyaya.2 Department of Bioorganic
Chemistry, Box 581, Biomedical Center, Uppsala University, S-751 23 Uppsala, Sweden.
1This article is part of a Special Issue dedicated to Professor Kelvin K. Ogilvie.
2Corresponding author (e-mail: jyoti@boc.uu.se).
Can. J. Chem. 85: 293–301 (2007)
doi:10.1139/V07-025
© 2007 NRC Canada

294
Can. J. Chem. Vol. 85, 2007
Scheme 1. Mechanism of deprotection of the 2′-O-TEM group and the adduct formation with tolyl vinyl sulfone (1).
stepwise coupling yield (97%–99%), leading to (iii) the
crude RNA, obtained after a simple deprotection step. The
crude form RNA was sufficiently pure for chemical biology
research without the need for any additional purification
steps.
internucleotidic phosphate backbone are not amenable to
adduct formation with TVS. We anticipated that the possible
nucleobases that could potentially be involved in the adduct
formation with TVS are 9-guaninyl, 9-adeninyl, and 1-
cytosinyl moieties in oligo–RNA. Keeping this in mind, four
The 2′-O-TEM group is stable upon ammonia treatment
and can be quickly cleaved by 1 mol/L tetrabutylammonium
fluoride (TBAF) in THF through F–-promoted β elimination
(Scheme 1). But the by-product p-tolyl vinyl sulphone
(TVS) (1), which is generated during deprotection, attacks
the nucleophilic sites of nucleobases of the oligo–RNA to
form some undesirable side products that somewhat reduce
the purity of the oligo–RNA. To obtain clean oligo–RNA at
the end of the synthesis, we had earlier employed a mixture
of 10% (v/v) n-propylamine and 1% (v/v) bis(2-
mercaptoethyl)ether in 1 mol/L TBAF–THF as a scavenger
for TVS during the deprotection of the 2′-O-TEM group.
Clearly, we wanted to find other neutral alternative(s) to this
scavenger because bis(2-mercaptoethyl)ether has a foul
odour and it can also deactivate many enzymes, even in trace
amounts (ppm).
dimeric ribonucleotides, 5′r(U[2′-o-TEM] pU[2′-o-TEM]
) (2),
5′r(G[2′-o-TEM] pU[2′-o-TEM]) (3), 5′r(A[2′-o-TEM] pU[2′-o-TEM]) (4),
5′r(C[2′-o-TEM] pU[2′-o-TEM]) (5), (Scheme 2), were synthesized.
They were treated with 1 mol/L TBAF–THF for 4 h at RT,
then the reaction mixtures were analyzed by reverse phase
HPLC (RP-HPLC). The RP-HPLC profiles are shown in
Fig. 1.
5′
As expected, r(U[2′-o-TEM] pU[2′-o-TEM]) (2) upon treatment
with 1 mol/L TBAF–THF gave a single product (Fig. 1A,
RT = 5.89 min), which proved to be the desired product
5r(UpU) (6) by MALDI-TOF MS (m/z calcd.: 551.10 [MH]+;
found: 551.11). Treatment of 5′r(G[2′-o-TEM] pU[2′-o-TEM]) (3)
with 1 mol/L TBAF–THF also gave the desired product
5′r(GpU) (7). (Fig. 1B, RT = 6.56 min, m/z calcd.: 590.12
[MH]+; found: 590.11). We therefore could safely exclude 1-
uracilyl and 9-guaninyl aglycons as substrates for adduct
formation with TVS.
The RP-HPLC analysis of the reaction of 5′r(C[2′-o-TEM]
pU[2′-o-TEM]) (5) is shown in Fig. 1C. The product with RT =
4.75 min was proved by MS to be the fully deprotected
product 5′r(CpU) (9) (m/z calcd.: 550.12 [M]–; found:
550.12). However, the peak appearing at 18.43 min (Fig. 1C)
was the major product, which was analyzed by MS (m/z
found: 730.32 [M]–). This suggested that this product was
formed by the reaction of one molecule of PVS with
5′r(CpU) (9). This major product was separated and its base
composition was analyzed by digesting it with mixture of
phosphodiesterase I and shrimp alkaline phosphatase (see
experimental section). The digestion mixture was applied to
RP-HPLC analysis. As shown in the Fig. 1C inset, the diges-
tion gave two products. One appears at 4.06 min, corre-
sponding to uridine. The second peak appears at 17.01 min.
It was probably formed by the addition of TVS to cytidine
and clearly needed further characterization (represented by
X in Fig. 1C).
Earlier work (8) has excluded N3 of 1-uracilyl as a reac-
tive centre for the side product formation. To this point, the
potential nucleophilic sites have been known to be N1 of 9-
adeninyl (9, 10), N7 and N1 of 9-guaninyl (11, 12), and N3
of 1-cytosinyl (13, 14). In this study, through isolation of the
pure adducts formed with TVS and by identifying them un-
ambiguously by MS and NMR, we have unexpectedly found
that N2 of 9-guaninyl, N4 of 1-cytosinyl, and especially N6
of 9-adeninyl are the reactive nucleophilic sites for adduct
formation with TVS, and more powerful scavengers were
therefore developed.
Results and discussion
Deprotection of the 2′-O-TEM group from dimeric
RNAs
It was known (8) that when deprotecting with 1 mol/L
TBAF–THF, pure poly-U can be easily obtained as the sole
product, hence the lactam function of 1-uracilyl and the
© 2007 NRC Canada

Zhou et al.
295
Scheme 2. 10: N4-(4-tolylsulfonylethyl)cytidine (CN4-TSE); 11: N6-(4-tolylsulfonylethyl)adenosine (AN6-TSE); 12: N3-(4-tolylsulfonyl-
ethyl)uridine (UN3-TSE); 13: N2-(4-tolylsulfonylethyl)guanosine (GN2-TSE); 14: N3-(4-nitrophenylsulfonylethyl)thymidine (TN3-NSE); 15:
N1-(4-nitrophenylsulfonylethyl)guanosine (GN1-NSE).
Dimer 5′r(A[2′-o-TEM] pU[2′-o-TEM]) (4) was treated with
1 mol/L TBAF–THF and gave a very small amount of the
desired product 5′r(ApU) (8) (Fig. 1D, RT = 8.54 min, m/z
calcd.: 572.41[M]–; found: 572.31). The predominant reac-
tion product (m/z found: 754.34 [M]–) appears at 21.00 min.
Base composition analysis of this major product by enzy-
matic digestion, as described earlier, gave uridine (Fig. 1D
inset, RT = 3.95 min) and another product (Fig. 1D, repre-
sented by Y, RT = 21.83 min). This result suggested that both
the cytidine and adenosine were reactive toward adduct for-
mation with TVS.
34.69 ppm, see Fig. 2 and Table 1) (12) suggested that the
TVS is attached to the N atom, and the potential site could
be either N4 or N3 of cytidine. The assignment of the cova-
1
lent binding site of TVS could easily be achieved by H, 13
C
Heteronuclear Multiple Bond Correlation (HMBC) NMR ex-
periment (12). As shown in Fig. 2A, the β-H shows a strong
cross peak with C4 (shown by red line) but no cross-
correlation with the C2, which strongly supports the expla-
nation that the TVS has been formed as an adduct to N4 of
cytidine to give N4-(4-tolylsulfonylethyl)cytidine (CN4-TSE
10).
,
The reaction between adenosine and TVS gave a major
product with [MH]+ at m/z 450.18, which also indicated that
a unimolecular adduct had formed between TVS and
adenosine. Two unambiguous NMR proofs were used to
confirm that TVS is indeed covalently bonded to the N6 of
adenosine to give product N6-(4-tolylsulfonylethyl)adenosine
Characterization of TVS adducts with aglycones of RNA
To determine the chemical nature of the adducts (X and Y
in the HPLC elution profile in Figs. 1C and 1D), either
cytidine or adenosine was treated with 3 equiv. of TVS at
RT in 1 mol/L TBAF–THF for 24 h. The major product of
the reaction between cytidine and TVS was separated by sil-
ica gel chromatography. This compound showed [MH]+ at
m/z 426.17, which suggested that it was an unimolecular
1
(AN6-TSE, 11): (i) in the H,13C HMBC spectrum (Fig. 2B),
the HN6 shows cross coupling with C5, C6, and β-C, and
(ii) more direct proof is the fact that the HN6 of adenosine
1
1
adduct of TVS and cytidine. The chemical shift of β-C (δ
,
shows a strong cross peak with β-H (Fig. S2.1) in the H, H
β-C
© 2007 NRC Canada

296
Can. J. Chem. Vol. 85, 2007
Fig. S4.1) in the 1H NMR spectrum of 13; (ii) the δβ-C and δ
Fig. 1. RP-HPLC analysis of the reaction products upon treat-
ment of compounds 2–5 with 1 mol/L TBAF–THF at RT for
4 h. The products (RT = 18.43 min in C and RT = 21.00 min in
B were digested respectively by mixture of phosphodiesterase
and alkaline phosphatase. HPLC analysis of the digestion reac-
tion is shown in the inset. RP-HPLC conditions: 0′→10′→35′, A
→A–B 9:1, v/v→A–B 2:8, RT, flow rate 1 mL/min. Buffer A:
0.1 mol/L TEAA in H2O–CH3CN (v/v, 95:5). Buffer B:
0.1 mol/L TEAA in H2O–CH3CN (v/v 50:50).
β-H
(see Table 1) in 13 are much different than in compound
N1-(4-nitrophenylsulfonylethyl)guanosine (GN1-NSE, 15,
Scheme 2) (12); (iii) in the 1H,13C HMBC spectrum
(Fig. 2D), β-H shows a strong cross peak with C2 but none
with C5, C6, or C8 in 13.3
The Michael addition between the α,β-unsaturated com-
pound, such as acrylonitrile with DNA or RNA under differ-
ent conditions, leads to cyanoethylation of nucleobases,
which has been studied extensively. Because of the natural
electronic structure of nucleobases (18), the cyanoethylation
generally occurs at N3 of uridine (19) and cytidine (9), N1 of
adenosine (9, 20), N7, and N1 of guanosine (9, 12). Though
cyanoethylation at N6 of adenosine (20) was also observed,
it was in fact obtained by Dimroth rearrangement (21) of N1
of cyanoethylated adenosine in the presence of a strong base
such as aqueous sodium hydroxide. Alkylation of nucleo-
sides with other reagents such as diazomethane and dimethyl
sulfate also occurs at the ring nitrogen atom such as N7 and
N1 of purine (10, 11) and N3 of pyrimidine (11, 14).
The present unexpected findings may have important ap-
plications in directly alkylating exocyclic amines of nucleo-
sides as probes to determine the structure and (or) function
of a nucleic acid (22, 23). Many methods have been devel-
oped for the alkylation of exocyclic amino groups (24–26),
but these methods (24–26) are multi-step reactions and are
time-consuming and laborious. To our knowledge, no direct
alkylation method has been developed until now. The present
finding that TVS can indeed be added selectively to an
exocyclic amino of cytidine, adenosine, and guanosine pro-
vides an easier way to the direct alkylation of exocyclic
amino groups of nucleobases if they exist as a part of a non-
hydrogen-bonded loop or hairpin structure in the RNA scaf-
folds. Moreover, the phenyl moiety of TVS is relatively easy
to derivatize for fluorescence detection purposes. Alterna-
tively, by attaching other functional groups to the phenyl
moiety, such as some specific cross-linking agents, we can
conveniently introduce various functional groups to the non-
hydrogen-bonded exocyclic amino of the nucleobases to
map the folding in the RNA scaffolds by estimating spatial
proximities.
COSY NMR spectrum, thereby confirming that the N6 of
adenosine has undergone a conjugate Michael addition to
TVS.3
Co-injection of compound 10 with X on RP-HPLC has
also shown that they are the same compound. In the same
way, Y was shown to be compound 11 (Fig. S5.1 and S5.2).3
It has been reported that thymidine and guanosine can re-
act with with acrylonitrile or phenylvinyl sulfone in the pres-
ence of a base such as tetra-n-butylammonium hydroxide
(12), TBAF–THF (15), or ammonia (16, 17). These α,β-un-
saturated compounds were adducted to the N3 of thymidine
(12, 15–17) and the N1 of guanosine (12), respectively. Here,
we have also studied the reaction between uridine or
guanosine with TVS in 1 mol/L TBAF–THF. As reported
(12), TVS is adducted to the N3 of uridine (NMR spectra
shown in Figs. 2C and S3) to give the product N3-(4-
tolylsulfonylethyl)uridine (UN3-TSE, 12).3 For guanosine, the
predominant product was proved to be N2-(4-tolyl-
sulfonylethyl)guanosine (GN2-TSE, 13) in an unambiguous
manner by NMR and not GN1-TSE, as earlier believed (12).
Thus, (i) the imino proton is present (δN1-H = 10.94 ppm, see
Relative nucleophilic reactivities of the exocyclic amino
groups of RNA toward TVS
Earlier we showed that in dimeric RNAs neither uridine
nor guanosine participates in the adduct formation with
TVS. Subsequently we showed that all of the four mono-
meric ribonucleosides are found to react with TVS. This ap-
parent contradiction may be explained by comparing the
nucleophilic reactivities of different ribonucleoside aglycons
in an oligo–RNA toward TVS. We therefore chose a
5′
tetrameric RNA, r(UpApCpG) (16, Scheme 3), for a model
study; 0.5 OD (12.5 µmol) of 16 was treated with 0, 1/4, 1/2,
1, 2, 4, and 20 equiv. of TVS, respectively, in 1 mol/L
TBAF–THF (total reaction volume 40 µL) at RT for 24 h.
The aliquots of the reactions were analyzed by RP-HPLC
(Fig. 3). The maldi-tof MS spectra showed that each of the
3 Supplementary data for this article are available on the journal Web site (http://canjchem.nrc.ca) or may be purchased from the Depository
of Unpublished Data, Document Delivery, CISTI, National Research Council Canada, Ottawa, ON K1A 0R6, Canada. DUD 5155. For more
information on obtaining material refer to http://cisti-icist.nrc-cnrc.gc.ca/irm/unpub_e.shtml.
© 2007 NRC Canada

Zhou et al.
297
1
Fig. 2. H, 13C HMBC NMR spectrum of compounds 10 (panel A, crosspeaks at δ 3.81), 11 (panel B, crosspeaks at δ 6.9), 12
(panel C, crosspeaks at δ 4.36), and 13 (panel D, crosspeaks at δ 3.51, 3.58).
adenosine and N4 of cytidine, or all the exocyclic amino
functions of ribonucleobases is possible. Since all the
exocyclic amino functions participate in H-bond formation
in the tertiary structure of nucleic acid, their selective
alkylation by TVS gives the possibility of identifying those
nucleobases that are not hydrogen bonded, and thus TVS
can be used as a chemical probe for the structure and func-
tion of the nucleic acid.
Table 1. Comparison of the chemical shifts of β-H and β-C of
compounds 10–15.
Compounds
δβ-C (ppm)
δβ-H (ppm)
10
11
12
34.69
34.65
34.95
35.66
34.60
34.70
3.80
3.89
4.25
3.61, 3.50
4.12
13
14a
15a
4.31
Morpholine and piperidine as scavenger for TVS to
avoid adduct formation
aThe NMR data are taken from Mag and Engels (12).
In the 2′-O-TEM group based solid-phase RNA synthesis
(8), a mixture of 10% n-propylamine and 1% bis(2-
mercaptoethyl) ether was used as the scavenger for TVS.
Two disadvantages of employing bis(2-mercaptoethyl)ether
as the deprotecting agent are its foul odour and its ability, in
trace amounts, to possibly cause enzyme deactivatation in
biological assays. In this study we wanted to use an appro-
priate amine as the sole scavenger to examine if the quality
of the crude RNA could be further improved.
three products 17 (m/z found: 1406.23 [MH]+), 18 (m/z
found: 1406.23 [MH]+), and 19 (m/z found: 1406.26 [MH]+)
has only one molecule of TVS attached, thereby suggesting
that the products 17, 18, and 19 are isomeric in nature. Fur-
thermore, base composition analysis of these products by en-
5′
zyme digestion showed that 17 is r(UpApCpGN2-TSE), 18 is
5′r(UpApCN4-TSE pG), and 19 is 5′r(UpAN6-TSE pCpG) (see
Fig. 4).
The yields of the three products in the reactions between
5′r(UpApCpG) (16) and different stoichiometries of TVS are
listed in Table 2. In the presence of a limited amount of
TVS, (1/4, 1/2 equiv.), only N6 of adenosine was alkylated
The Michael addition of amine to TVS (27–29) has shown
that primary amines are notably less reactive than secondary
amines because the former binds to two moles of solvent
through H-bonds, compared with the latter that binds only to
one mole of the solvent. It was also shown (27–29) that
large differences in the basicity of amines has relatively little
effect on the reactivity. Thus morpholine (pKa = 8.4) is about
one seventh as reactive as piperidine (pKa = 11.1), but the
former is > 100 times less basic. In our case, the ideal amine
should have better nucleophilicity toward TVS than the
exocyclic amino groups of the nucleobases of RNA, and at
the same time the basicity of the amine cannot compromise
any oligo–RNA chain cleavage. Keeping this in mind, sev-
eral types of amines with very different pKa values were se-
5′
to give r(UpAN6-TSE pCpG) (19). With the increase in con-
centration of TVS (1 or 2 equiv.), N4 of cytidine was also
alkylated to give 5′r(UpApCN4-TSE pG) (18). By increasing
the amount of TVS further to 20 equiv., a small amount of
product by the alkylation of N2 of guanosine was also ob-
5′
served to give r(UpApCpGN2-TSE) (17). This proves that the
alkylation tendency of ribonucleotide by TVS in oligo–RNA
in 1 mol/L TBAF–THF follows the sequence, N6 – A > N4 –
C > N2 – G > > N3 – U. By controlling the concentration of
TVS, selective alkylation of N6 of adenosine, or both N6 of
© 2007 NRC Canada

298
Can. J. Chem. Vol. 85, 2007
Scheme 3.
Fig. 4. The compounds 17, 18, and 19 were digested by a mix-
ture of phosphodiesterase and alkaline phosphatase. The HPLC
profiles of digestion mixture are shown in (A) 17, (B) 18, and
(C) 19. RP-HPLC conditions: 0′→10′→42′, A→A–B 9:1, v/v →B,
RT, flow rate 1 mL/min. Buffer A: 0.1 mol/L TEAA in H2O–
CH3CN (v/v, 95:5). Buffer B: 0.1 mol/L TEAA in H2O–CH3CN
(v/v 50:50). Abbreviations are A = adenosine, G = guanosine, C =
cytidine, U = uridine, AN6-TSE = N6-(4-tolylsulfonylethyl)ade-
Fig. 3. RP-HPLC analysis of the reactions between
5′r(UpApCpG) (16) (0.5 OD, 12.5 µmol) with 0, 1/4, 1/2, 1, 2, 4,
and 20 equiv. of TVS in 1 mol/L TBAF–THF at RT for 24 h.
Each reaction has the same total volume of 40 µL. RP-HPLC
conditions: 0′→10′→42′, A→A–B 9:1 →B, RT, flow rate
1 mL/min. Buffer A: 0.1 mol/L TEAA in H2O–CH3CN (v/v,
95:5). Buffer B: 0.1 mol/L TEAA in H2O–CH3CN (v/v 50:50).
nosine, CN4-TSE = N4-(4-tolylsulfonylethyl)cytidine, GN2-TSE
N2-(4-tolylsulfonylethyl)guanosine.
=
lected to test their abilities as scavengers of TVS in 1 mol/L
TBAF–THF solution. The 5′r(A[2′-o-TEM] pU[2′-o-TEM] pC[2′-o-
TEM] pG[2′-o-TEM] (20, Scheme 3) was used as the model com-
pound and was treated with 1 mol/L TBAF–THF containing
10% (v/v) of different amines at RT for 24 h, and the reac-
tion mixtures were analyzed by RP-HPLC (Fig. S7).3 The
percentage of chain cleavage product and side products
formed through the reaction of TVS to oligo–RNA were ob-
tained by integrating the HPLC profile (see Fig. S7) and
they are given in Table 3.3
Table 3 shows that a primary or secondary amine, such as
n-propylamine, piperidine, or morpholine, when used alone
can scavenge TVS more efficiently than the standard
deprotection condition (8) [a mixture of n-propylamine and
bis(2-mercaptoethyl)ether] employed earlier for the
deprotection of the 2′-O-TEM group. Among these amines,
the morpholine and piperidine are especially attractive.
© 2007 NRC Canada

Zhou et al.
299
Table 2. The yields (%) of three isomeric products 17, 18, and 19 in the reactions between
5′r(UpApCpG) (16) and different stoichiometries of TVS (1).
Stoichiometry of
5′r(UpApCpGN2-TSE
)
5′r(UpApCN4-TSE pG)
5′r(UpAN6-TSE pCpG)
TVS (1)
(17)
(18)
(19)
0
0
0
1.0
1.9
5.3
13.2
0
0
0
0
0
0
0
0.8
4.5
1/4
1/2
1
2
4
1.4
1.5
3.4
5.1
15.1
42.0
20
Table 3. Deprotectiona of 20 with 1 mol/L TBAF–THF containing different amines.
Oligo–RNA chain
cleavage (%)
TVS adduct with
oligo–RNA (%)
c
Aminea
pKa
Standard conditionb
n-Propylamine
Piperidine
—
9.1
1.6
8.3
0
17.0
8.1
0
10.71
11.12
8.33
Morpholine
7.2
aDeprotection reagents: a solution of 10% (v/v) of each amine in 1 mol/L TBAF–THF, RT, 24 h.
bA mixture of 10% n-propylamine and 1% bis(2-mercaptoethyl)ether in 1 mol/L TBAF–THF (8).
cThe pKa values are taken from Perrin (30).
Fig. 5. 20% denatured PAGE picture of crude RNAs obtained
Morpholine, benefiting form its lower pKa (8.4), does not
cause any chain cleavage but allows the addition of TVS to
oligo–RNA to some extent (7.2%) in 1 mol/L TBAF–THF
solution. On the other hand, piperidine can completely
quench the reaction between TVS and oligo–RNA, but it
leads to a small amount (8.3%) of chain cleavage (Table 3).
To examine the potential of morpholine and piperidine as
candidates as scavengers of the undesirable byproduct TVS,
we have deprotected two 21-mer RNAs using 1 mol/L
TBAF–THF, containing either a mixture of n-propylamine
and bis(2-mercaptoethyl)ether (8), piperidine, or morpho-
line. Thus, two 21-mer RNAs were synthesized on a solid
support with the 2′-O-TEM group using our published proce-
dure (8) (sequence I: 5′-GAC GUA AAC GGC CUC AAG
UUC; sequence II: 5′-ACU UGU GGC CGU UUA CGU
CGC). After treatment with 25% NH3–MeOH for 20 h at RT
and then for 4 h at 40 °C, they were divided into three parts
and incubated with (a) the standard deprotecting reagent (8)
(1 mol/L TBAF–THF containing 10% n-propylamine and
1% bis(2-mercaptoethyl) ether), (b) 1 mol/L TBAF–THF
containing 10% of piperidine, or (c) 1 mol/L TBAF–THF
containing 10% of morpholine at RT for 20 h. After evapo-
ration of the volatile materials, the products were first
passed through a Sep-Pak® Plus (C18, reverse phase) (31)
column to remove hydrophilic matter such as excess of F–,
and then NAPTM-10 (Sephadex G25) column (32) to sepa-
rate the smaller molecules from the larger ones on the basis
of size (thus, the full-size oligo–RNA elutes from the trun-
cated sequences and hydrophobic protecting group residues)
to give the crude products, which are of high purity and can
be used directly for various biological experiments. These
crude products were analyzed by 20% denatured PAGE and
gel picture obtained by UV–visualization (Fig. 5).
from the deprotection of RNA sequence I (5′-GAC GUA AAC
GGC CUC AAG UUC) and RNA sequence II (5′-ACU UGU
GGC CGU UUA CGU CGC) with different 2′-O-TEM
deprotecting conditions. The 2′-O-TEM group was deprotected
with 1 mol/L TBAF–THF containing different scavengers at RT
for 20 h. Lanes 1 and 4: 10% n-propylamine and 1% bis(2-
mercaptoethyl)ether. Lanes 2 and 5: 10% piperidine. Lanes 3 and
6: 10% morphholine. Marker, upper band: xylene cyanol blue;
bottom band: bromophenol blue. Gel picture obtained by UV–vis
(254 nm) showing all UV absorbing compounds both nucleic ac-
ids (full product and n-1 failure sequences) as well as the
cleaved protecting groups. Note that the noncharged TVS (1)
(λmax = 233 nm) also runs on the gel and is visible by 254 nm
light shadowing if in high concentration (Fig. S8).3 The bands
between dashed lines correspond to TVS. Lanes 7 and 8 were
run together as a control. Lane 7 product obtained by incubating
compound TVS with 1 mol/L TBAF–TMF. Lane 8 product ob-
tained by incubating compound TVS with 1 mol/L TBAF–TMF
containing 10% morpholine.
As shown in Fig. 5, all three deprotecting conditions give
highly pure crude RNA. Use of morpholine or piperidine,
© 2007 NRC Canada

300
Can. J. Chem. Vol. 85, 2007
however, gives only a very minor improvement of RNA pu-
rity (-84%) compared with that of the standard condition (8)
(-82%) [a mixture of n-propylamine and bis(2-
mercaptoethyl)ether], which is in contrast to the result ob-
tained from the deprotection of the tetrameric RNA (20) (see
Table 3) with the above three reagents.
completion of the synthesis, the solid supports were treated
with 25% NH3–MeOH at RT for 20 h and then at 40 °C for
4 h. After filtration, each sample was separated into three
parts and the volatile materials were evaporated. The resi-
dues were co-evaporated with dry methanol twice and dried
in vacuum. Then they were treated with 1 mol/L TBAF–
THF (dried over 4 Å molecular sieves overnight) containing
different scavengers at RT for 20 h. After evaporation of the
volatile matter, the residues were diluted with double-
distilled water and applied to a NAP-10 column (Amersham
Pharmacia Biotech, Sweden). Loading volume is 1 mL and
elution was done with 1.2 mL of double-distilled water. The
collected fractions were subjected to Sep-Pak® Plus C18
cartridges (Waters) according to manufacturer instructions,
to give the crude products. The crude RNAs were subjected
to 20% denatured PAGE analysis without any purification
step.
Conclusion
In this study, we found that N6 of adenosine, N4 of
cytidine, and N2 of guanosine, just as N3 of thymidine (12),
can be alkylated by the α,β-unsaturated compound TVS in
1 mol/L TBAF–THF solution through Michael addition. In
oligo–RNA, their nucleophilicities toward TVS are different:
N6 – A > N4 – C > N2 – G >> N3 – U. Hence, by controlling
the concentration of TVS, selective alkylation of these
nucleophilic sites can be easily achieved. This discovery not
only provides a new approach for the alkylation of non-
hydrogen-bonded exocyclic amino groups of nucleobases in
RNA, but also provides a new strategy for the potential ap-
plication of the reagent TVS as a chemical probe for detect-
ing the structure and (or) the function of nucleic acid. To
avoid adduct formation and chain cleavage during
deprotection of 2′-O-TEM group with 1 mol/L TBAF–THF,
both morpholine and piperidine have been proved to be good
scavengers of TVS, which can give high quality crude RNA,
bypassing the practical problems of employing the mixture
of n-propylamine and bis(2-mercaptoethyl)ether (8).
General procedure of reaction between ribonucleosides
with p-tolyl vinyl sulphone (1)
Ribonucleoside (1 mmol) was dissolved in dry 1 mol/L
TBAF–THF (5ml) and p-tolyl vinyl sulphone 1 (3 mmol for
adenosine, cytidine, uridine and 1 mmol for guaonsine) was
added. The reaction was stirred at RT for 24 h and analyzed
by HPLC. Then the reaction mixture was evaporated to dry-
ness. The residue was diluted with dichloromethane (DCM)
and washed with satd. NaHCO3 and dried over MgSO4. The
major product was obtained by short column chromatogra-
phy.
Experimental section
N4-(4-tolylsulfonylethyl) cytidine (10) (CN4-TSE
)
HPLC analysis of the reaction mixture showed that 38.6%
of the cytidine was transformed into product 10 and 38.3%
of the cytidine was unreacted. CC (DCM, MeOH from 2%
General experimental methods
Chromatographic separations were performed on Merck
G60 silica gel. Thin layer chromatography (TLC) was per-
formed on Merck pre-coated silica gel 60 F254 glass-backed
1
to 10%). H NMR (270 MHz, CDCl3 + CD3OD) δ: 2.45 (s,
1
3H, PhMe), 3.46 (t, J = 5.8 Hz, 2H, SO2CH2), 3.75–3.96 (m,
4H, SO2CH2CH2, 5′,5′′-H), 4.09 (broad, 1H, 4′-H), 4.18
(broad, 2H, 2′3′-H), 5.69 (d, J = 1.5 Hz, 1H, 1′-H), 5.76 (d, J
= 7.4 Hz, 1H, 5-H), 7.37 (d, J = 7.7 Hz, 2H, phH), 7.77 (d, J
= 8.3 Hz, 2H, phH), 7.87 (d, J = 7.4 Hz, 1H, 6-H). 13C NMR
(67.9 MHz, CDCl3 + CD3OD) δ: 23.94 (PhMe), 34.69
(SO2CH2CH2) 54.79 (SO2CH2CH2), 60.88 (5′-C), 69.31 (3′-
C), 74.94 (2′-C), 84.88 (4′-C), 92.53 (1′-C), 95.81 (5-C),
127.98 (ph), 130.24 (ph), 136.10 (ph), 141.11 (C6), 145.47
(ph), 157.21 (2-C), 164,05 (C4). MALDI-TOF MS: calcd.:
426.13 [MH]+; found: 426.17.
plates. H NMR spectra were recorded at 270.1 MHz and
500 MHz, respectively. 13C NMR spectra were recorded at
67.9 MHz, 125.7 MHz, and 150.9 MHz, respectively. Chem-
ical shifts are reported in ppm (δ scale). Reversed-phase
(RP) HPLC: Kromasil 100, C18, 5µ, 100*4.6 mm. Flow
1 mL/min at RT, UV detector with detecting wavelength of
260 nm. Buffer A: 0.1 mol/L TEAA in H2O–CH3CN (95:5).
Buffer B: 0.1 mol/L TEAA in H2O–CH3CN (50:50).
Polyacrylamide gel electrophoresis (PAGE): 20% acrylamide
(acrylamide/bisacrylamide 29:1), 7 mol/L urea, TBE buffer.
Oligoribonucleotide synthesis
N6-(4-tolylsulfonylethyl) adenosine (11) (AN6-TSE
)
Oligo–RNAs 2–5 and 20 were synthesized through our
published procedure (8). After the completion of the auto-
mated solid-phase synthesis, the solid supports were treated
with 25% NH3–MeOH at RT overnight, filtered, and purified
by HPLC. Oligo–RNA 14 was synthesized by a TBDMS ap-
proach according to reported procedure (33) and purified by
HPLC. Their structural integrity were verified by MALDI-
TOF MS: (2) calcd. 974.92 [MH]+, found 975.29; (3) calcd.
1013.94 [MH]+, found 1014.24; (4) calcd. 997.95 [MH]+,
found 998.25. (5) cal. 973.94 [MH]+, found 974.29; (20)
calcd. 2071.85 [MH]+, found 2070.15; (14) calcd. 1224.21
[MH]+, found 1224.19.
HPLC analysis of the reaction mixture showed that 35.6%
of the adenosine was transformed into product 11 and 56.8%
of the adenosine was unreacted. CC (DCM, Acetone from
1
20% to 50%). H NMR (270 MHz, CDCl3 + CD3OD) δ:
2.36 (s, 3H, PhMe), 3.50 (t, 2H, SO2CH2), 3.69 (t, 1H, 5′-
H), 3.84–4.01 (m, 3H, 5′′-H, SO2CH2CH2), 4.26 (broad, 1H,
4′-H), 4.44 (broad, 1H, 3′-H), 5.50 (broad, 1H, 2′-OH), 5.84
(d, J = 6.7 Hz, 1H, 1′-H), 6.62 (broad, 1H, 5′-OH), 6.97 (t,
J = 6.1 Hz, 1H, N6H), 7.25 (d, J = 7.9 Hz, 2H, phH), 7.74
(d, J = 8.2 Hz, 2H, phH), 7.90 (s, 1H, 8-H), 8.03 (s, 1H, 2-
H). 13C NMR (67.9 MHz, CDCl3 + CD3OD) δ: 21.60
(PhMe), 34.65 (SO2CH2CH2) 54.91 (SO2CH2CH2), 62.99
(5′-C), 72.37 (3′-C), 74.01 (2′-C), 87.45 (4′-C), 90.86 (1′-
The 21-mer RNAs, sequence I and sequence II, were syn-
thesized according to our published procedure (8). After
© 2007 NRC Canada

Zhou et al.
301
C), 120.46 (5-C), 127.90 (ph), 130.00 (ph), 135.95 (ph),
140.67 (8-C), 145.07 (ph), 147.49 (4-C), 152.06 (2-C),
154.08 (6-C). MALDI-TOF MS: calcd.: 450.14 [MH]+;
found: 450.19.
Forskning), and the EU-FP6 funded RIGHT project (Project
no. LSHB-CT-2004–005276) is gratefully acknowledged.
References
1. A. Fire, S. Xu, M.K. Montgomery, S.A. Kostas, S.E. Driver,
and C.C. Mello, Nature (London), 391, 808 (1998).
2. S.M. Elbashir, J. Harborth, W. Lendechel, A. Yalcin, K.
Weber, and T. Tuschl. Nature (London), 411, 494 (2001).
3. C.B. Reese. Org. Biomol. Chem. 3, 3851 (2005).
4. S. Pitsch, P.A. Weiss, L. Jenny, A. Stutz, X. Wu, and H.C.
Acta. Helv. Chim. Acta, 84, 3773 (2001).
N3-(4-tolylsulfonylethyl) uridine (12) (UN3-TSE
)
HPLC analysis of the reaction mixture showed that 57.2%
of the uridine was transformed to product 12 and 34.1% of
the uridine was unreacted. CC (DCM, MeOH from 5% to
1
7%). H NMR (270 MHz, CDCl3 + CD3OD) δ: 2.39 (s, 3H,
PhMe), 3.42 (t, J = 7.0 Hz, 2H, SO2CH2), 3.76 (dd, J =
12.2 Hz, 2H, 5′,5′′-H), 3.99 (d, J = 1.5 Hz, 1H, 4′-H), 4.10–
4.21 (m, 4H, 2′, 3′, -H, SO2CH2CH2), 5.65 (d, J = 8.1 Hz,
1H, 5-H), 5.72 (d, J = 2.3 Hz, 1H, 1-H), 7.32 (d, J = 7.8 Hz,
2H, phH), 7.74 (d, J = 8.2 Hz, 2H, phH), 7.85 (d, J =
8.1 Hz, 1H, 6-H). 13C NMR (67.9 MHz, CDCl3 + CD3OD)
δ: 23.88 (PhMe), 34.95 (SO2CH2CH2) 52.56 (SO2CH2CH2),
60.95 (5′-C), 69.41 (3′-C), 74.73 (2′-C), 85.16 (4′-C), 91.13
(1′-C), 101.40 (5-C), 128.04 (ph), 130.11 (ph), 135.67 (ph),
139.81 (6-C), 145.36 (ph), 150.92 (2-C), 162.60 (4-C).
MALDI-TOF MS: calcd.: 427.11 [MH]+; found: 427.15.
5. S.A. Caringe, F.E. Wincott, and M.H. Caruthers. J. Am. Chem.
Soc. 120, 11820 (1998).
6. A. Semenyuk, A. Foldesi, T. Johansson, C. Estmer-Nilsson, P.
Blomgren, M. Brannvall, L.A. Kirsebom, and M.
Kwiatkowski. J. Am. Chem. Soc. 128, 12356 (2006).
7. T. Ohgi, Y. Masutomi, K. Ishiyama, H. Kitagawa, Y. Shiba,
and J. Yano. Org. Lett. 7, 3477 (2005).
8. C.Z. Zhou, D. Honcharenko, and J. Chattopadhyaya. Org.
Biomol. Chem. 5, 333 (2007).
9. J.J. Solomon, I.L. Cote, M. Wortman, K. Decker, and A. Segal.
Chem. Biol. Interactions, 51, 167 (1984).
10. J.W. Jones and R.K. Robins. J. Am. Chem. Soc. 85, 193 (1963).
11. P.D. Lawley and P. Brookes. Biochem. J. 127 (1963).
12. M. Mag and J.W. Engels. Nucleic Acids Res. 16, 3525 (1988).
13. P. Brookes, and P.D. Lawley. J. Chem. Soc. 1348 (1962).
14. R.L.C. Brimacombe, B.E. Griffin, J.A. Haines, W.J. Haslam,
and C.B. Reese. Biochemistry, 4, 2452 (1965).
15. K.K. Ogilvie and S.L. Beaucage. Nucleic Acids Res. 7, 805
(1979).
16. T. Umemoto and T. Wada. Tetrahedron Lett. 46, 4521 (2005).
17. D.C. Capaldi, H. Gaus, A.H. Krotz, J. Arnold, R.L. Carty,
M.N. Moore, A.N. Scozzari, K. Lowery, D.L. Cole, and V.T.
Ravikumar. Org. Process Res. Dev. 832 (2003).
N2-(4-tolylsulfonylethyl) guanosine (13) (GN2-TSE
)
HPLC analysis of the reaction mixture showed that 50.3%
of the guanosine was transformed to product 13 and 23.8%
of the guanosine was unreacted. The product 13 was too po-
lar to purify by silica gel column chromatography, hence
preparative HPLC was used to purify a small amount of
1
product 13 for NMR and MS analysis. H NMR (270 MHz,
DMSO-d6) δ: 2.38 (s, 3H, PhMe), 3.51–3.60 (m, 6H,
SO2CH2, SO2CH2CH2, 5′,5′′-H), 3.99 (t, J = 1.5 Hz, 1H, 4′-
H), 4.11 (t, 1H, 3′-H), 4.47 (t, J = 5.3 Hz, 1H, 2′-H), 5.68
(d, J = 5.8 Hz, 1H, 1′-H), 6.88 (br., 1H, N2H), 7.42 (d, J =
7.9 Hz, 2H, phH), 7.77 (d, J = 8.2 Hz, 2H, phH), 7.93 (s,
1H, 8-H), 10.94 (br., 1H, N1H). 13C NMR (67.9 MHz,
DMSO-d6) δ: 21.64 (PhMe), 34.95 (SO2CH2CH2) 54.11
(SO2CH2CH2), 62.11 (5′-C), 70.99 (3′-C), 74.27 (2′-C),
85.87 (4′-C), 87.21 (1′-C), 117.70 (5-C), 128.17 (ph), 130.43
(ph), 136.46 (8-C), 136.70 (ph), 144.96 (ph), 151.06 (4-C),
152.89 (2-C), 157.67 (6-C). MALDI-TOF MS: calcd.:
466.23 [MH]+; found: 466.13.
18. B. J. Pullman. Chem. Soc. B, 1621 (1959).
19. J. Ofengand. Methods Enzymol. 20, 162 (1971).
20. A. Segal, U. Mate, and J.J. Solomon. Chem. Biol. Interactions,
28, 333 (1979).
21. J.D. Engel. Biochem. Biophys. Res. Comm. 64, 581 (1975).
22. R. Micura, W. Pils, C. Höbartner, K. Grubmayr, M.O. Ebert,
and B. Jaun. Nucleic Acids Res. 29, 3997 (2001).
23. H.M. Kong, L.F. Lin, N. Porter, S. Stickel, D.P.J. Byrd, and
R.J. Roberts. Nucleic Acids Res. 28, 3216 (2000).
24. O. Gorchs, M. Mernandez, L. Garriga, E. Pedroso, A.
Grandas, and J. Farras. Org. Lett. 4, 1827 (2002).
25. C. Höbartner, C. Kreutz, E. Flecker, E. Ottenschlager, W. Pils,
K. Grubmayr, and R. Micura. Monatsh. Chem. 134, 851 (2003).
26. M. Beier, and W. Pfleiderer. Helv. Chim. Acta, 86, 2533 (2003).
27. S.T. McDowell, C.J.M. Stirling. J. Chem. Soc. B, 343 (1967).
28. S.T. McDowell and C.J.M. Stirling. J. Chem. Soc. B, 348
(1967).
Base composition analysis
Shrimp alkaline phosphatase and phosphodiesterase I
(Crotalus adamanteus venom) were purchased from
Amersham Pharmacia Biotech, Sweden. A reaction mixture
containing 20 mmol/L Tris-HCl (pH 8.0), 10 mmol/L MgCl2,
shrimp alkaline phosphatase (0.5 unit), and phosphodiester-
ase I (Crotalus adamanteus venom) (0.4 unit) was added to
oligoribonucleotide (0.1 OD unit at 260 nm) with a total re-
action volume of 30 µL. The reaction mixture was incubated
at 37 °C for 12 h and subjected to HPLC analysis directly.
29. S.T. McDowell and C.J.M. Stirling. J. Chem. Soc. B, 351
(1967).
30. D.D. Perrin. Dissociation constants of organic bases in aque-
ous solution. Supplement. Butterworths, London, UK. 1972.
31. F. Wincott, A. DiRenzo, C. Shaffer, S. Grimm, D. Tracz, C.
Workman, D. Sweedler, C. Gonzalez, S. Scaringe, and N.
Usman. Nucleic Acids Res. 23, 2677 (1995).
32. Handbook of gel filtration, principles and methods. Edited by
Acknowledgements
Amersham Biosciences.
33. J. Milecki, E. Zamaratski, T.V. Maltseva, A. Földesi, R.
Adamiak, and J. Chattopadhyaya. Tetrahedron Lett. 55, 6603
(1999).
Generous financial support from the Swedish Natural Sci-
ence Research Council (Vetenskapsr et), the Swedish Foun-
dation for Strategic Research (Stiftelsenf Strategisk
© 2007 NRC Canada