CAL-B-catalyzed acylation of nucleosides and role of the sugar conformation: An improved understanding of the enzyme-substrate recognition

Article abstract of DOI:10.1002/ejoc.201200609

We demonstrate that the sugar ring conformation of nucleosides plays a critical role during Candida antarctica lipase B (CAL-B) catalyzed acylation. Specifically, the North (N), but not the South (S) nucleoside sugar ring conformation is preferred for efficient binding at the catalytic site. In this study, we used nuclear magnetic resonance (NMR) spectroscopy experiments to establish the sugar ring conformation of nucleosides and performed molecular modeling studies to support the observations. The ribo-and 2'-substituted (OMe, F) nucleosides displaying the N-conformation undergo rapid and facile acylation compared to the 2'-deoxynucleosides with the S-conformation. This study improves our understanding of the critical role that sugar conformation plays in enzyme-substrate recognition during biotransformations using CAL-B. To the best of our knowledge, this is the first experimental report offering a rationale for the observed selectivity during acylation of nucleosides containing the N-sugar conformation.

Full text of DOI:10.1002/ejoc.201200609

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
FULL PAPER
DOI: 10.1002/ejoc.201200609
CAL-B-Catalyzed Acylation of Nucleosides and Role of the Sugar
Conformation: An Improved Understanding of the Enzyme-Substrate
Recognition
Saúl Martínez-Montero,[a] Susana Fernández,[a] Yogesh S. Sanghvi,[b] Vicente Gotor,*[a] and
Miguel Ferrero*[a]
Keywords: Nucleosides / Enzymes / Acylation / Sugar conformation / Molecular modeling
We demonstrate that the sugar ring conformation of nucleo-
sides plays a critical role during Candida antarctica lipase B
(CAL-B) catalyzed acylation. Specifically, the North (N), but
not the South (S) nucleoside sugar ring conformation is pre-
ferred for efficient binding at the catalytic site. In this study,
we used nuclear magnetic resonance (NMR) spectroscopy
experiments to establish the sugar ring conformation of nu-
cleosides and performed molecular modeling studies to sup-
port the observations. The ribo- and 2Ј-substituted (OMe, F)
nucleosides displaying the N-conformation undergo rapid
and facile acylation compared to the 2Ј-deoxynucleosides
with the S-conformation. This study improves our under-
standing of the critical role that sugar conformation plays in
enzyme–substrate recognition during biotransformations
using CAL-B. To the best of our knowledge, this is the first
experimental report offering a rationale for the observed
selectivity during acylation of nucleosides containing the N-
sugar conformation.
Introduction
The furanose ring of the nucleosides in solution can
adopt two major antipodal conformations classified as
North (N, 2Ј-exo) and South (S, 3Ј-exo) described in the
pseudorotation cycle (Scheme 1).[1] In solution, the ribo-
sugar is invariably in a rigid N-type conformation, whereas
the 2Ј-deoxy-sugar is flexible, adopting the S-conforma-
tion.[2] The preference for N- or S-conformation is deter-
mined by the interplay of important interactions resulting
from anomeric and gauche effects exerted by the 2Ј-substit-
uents on the furanose ring.[3] It has been demonstrated that
the biological activity of certain nucleosides is influenced
by the sugar conformation which plays a critical role in re-
cognition and binding to enzymes.[4] Additionally, the 2Ј-
sugar modifications (OMe, F and OCH2CH2OMe) are ex-
tensively used to modulate the conformational properties of
oligonucleotides.[5]
Scheme 1.
In recent studies, we observed that enzymatic acylation
at the 5Ј-OH group of ribonucleosides was regioselective;
whereas the acylation of 2Ј-deoxynucleosides often fur-
nished a mixture of products.[6] Since 2Ј-deoxynucleosides
are preferentially in the S-conformation and the ribonucleo-
sides are in the N-conformation, we postulated that the
sugar pucker may play a key role in the directing enzyme
selectivity during acylation.
Because of the general interest in biocatalysis[7] and to
improve our understanding of substrate recognition by en-
zymes, we selected Candida antarctica lipase B (CAL-B) for
this study. The easy access, low cost, commercial availabil-
ity, possibility for reuse and amenability to a wide range of
[a] Departamento de Química Orgánica e Inorgánica and Instituto substrates were key attributes in the selection of CAL-B.[6]
Universitario de Biotecnología de Asturias, Universidad de
Due to our continued interest in exploring the potential
Oviedo,
of CAL-B mediated acylation reactions,[6] we selected 15
33006 Oviedo (Asturias), Spain
Fax: +34-98-510-3448
nucleosides (as 1 and 5) possessing both the N- and S-con-
formations of the sugar residue for study (Table 1). Further-
more, these nucleosides are important building blocks for
the assembly of therapeutic oligonucleotides.[5] Having ac-
cess to their 5Ј-O-acylated derivatives will be of significant
value, because conventional chemical synthesis would re-
E-mail: mferrero@uniovi.es
vgs@uniovi.es
Homepage: http://www.unioviedo.es/bioorganica/NatProd.html
[b] Rasayan Inc.,
2802 Crystal Ridge Road, Encinitas, CA 92024-6615, USA
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201200609.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
V. Gotor, M. Ferrero et al.
FULL PAPER
Table 1. Relationship between N/S conformation and CAL-B-catalyzed acylation of various nucleosides.
Entry
Nucleoside[c,d]
Solubility[e] J1Ј2Ј cis/trans [Hz]
N [%][a]
T
Conversion[b] Acylated nucleosides[b,c] [%]
[mg/mL]
THF
DMSO THF DMSO [h]
[%]
2
3
4
1
2
3
4
dU (1aa)
U (1ab)
2Ј-O-Me-U (1ac)
2Ј-F-U (1ad)
5.3
3.0
16.0
13.3
6.8 t
3.7
3.0
6.8 t
5.4
5.2
28
63
70
80
28
46
48
80
2
4
1
99
92
97
Ͼ 99
Ͼ 99
4
4
3
Ͼ 99
Ͼ 99
Ͼ 99
2.0
2.0
0.5
5
6
7
T (1ba)
2Ј-O-Me-T (1bc)
LNA-T (5)
4.4
10
16
6.6 t
3.2
0
6.8 t
5.4
0
32
68
100
28
46
100
2.5
1.5
1.5
99
99
Ͼ 99
89
Ͼ 99
Ͼ 99
6
5
5
1
1
8
9
10
dCBz (1ca)
2.1
6.2 t
0
0
6.2 t
2.4
0
41
100
100
41
76
100
19
3
2
99
99
99
79
Ͼ 99
Ͼ 99
16
2Ј-O-Me-CBz (1cc) not available
2Ј-F-CBz (1cd)
15
11
12
13
dABz (1da)
2Ј-O-Me-ABz (1dc)
2Ј-F-ABz (1dd)
5.4
15
4.6
7.4/6.2
5.4
3.8
6.8 t
5.4
2.4
28
46
62
28
46
76
2
2
2.5
Ͼ 99
60
97
95
97
98
4
3
2
14
15
dGIbu (1ea)
0.4
0.6
6.8 t
5.6
7.6/6.4
6.0
28
44
23
40
24
7.5
97
99
99
Ͼ 99
2Ј-O-Me-GIbu (1ec)
[a] N (%) = 100 – S (%); ribonucleosides: S (%) = 10J1Ј2Ј; deoxyribonucleosides: S (%) = 100[J1Ј2Ј(cis) + J1Ј2Ј(trans)-7.1]/9; from ref.[8d]
[b] Monitored by HPLC. [c] See Scheme 2 for details. [d] Note on compound numbering: the compounds are identified with a number
and two letters. The first lower case letter refers to the base on the nucleoside (a, U; b, T; c, CBz; d, ABz; e, GIbu), and the second lower
case letter refers to the 2Ј-substituent X on the sugar (a, H; b, OH; c, OMe; d, F). [e] In THF.
quire multi-step protocols to prepare these products. gauche effects.[3] The observed sugar conformation for the
Herein, we describe: (i) NMR-based conformational analy- majority of the nucleosides in THF was consistent with the
sis of nucleosides 1 and 5; (ii) CAL-B-catalyzed acylation data in DMSO. Based on this data, we decided to proceed
of nucleosides 1 and 5; (iii) molecular modeling of these with the use of THF as a suitable solvent for our studies.
Next, we embarked on the CAL-B-catalyzed acylation of
monomers 1 (Scheme 2) in order to study the relationship
between the regioselectivity and the nucleoside conforma-
tion. All reactions were monitored by HPLC to determine
the ratio of acylated products. For acylation studies, we
elected the levulinyl protecting group[9] for a variety of
reasons. First, due to its utility as an orthogonal protecting
group during the solution-phase synthesis of oligonu-
cleotides.[10] Second, our prior experience with the lipase-
mediated levulinylation of nucleosides,[6] and third, it is an
inherently small, atom-efficient and inexpensive protecting
nucleosides and their binding with CAL-B to substantiate
the results.
Results and Discussion
Although several protocols are reported for the calcula-
tion of percentage of N- and S-conformation,[8] we used
the most widely accepted method described by Altona and
Sundaralingam.[8c,8d] In this protocol the conformation of
nucleoside analogues in solution is obtained from coupling
1
constants between the ribose ring protons in the H NMR
spectrum. The equations for calculating the percentage of
S-conformation[8e] in ribonucleosides and 2Ј-deoxynucleos-
ides are shown in Equations (1) and (2), respectively.
Ribonucleosides: S (%) = 10J1Ј2Ј
(1)
2Ј-Deoxynucleosides: S (%) = 100[J1Ј2Ј(cis) + J1Ј2Ј(trans)-7.1]/9 (2)
Most of the N/S conformational ratios of nucleosides are
reported in D2O or DMSO. However, the proposed enzy-
matic study required the use of THF as the preferred sol-
vent. Therefore, we undertook the conformational study of
nucleosides in THF and also in DMSO as a solvent to com-
pare our data with the reported values. The percentage of
N-conformation calculated from the coupling constants in
[D8]THF and [D6]DMSO of the nucleosides is summarized
in Table 1. In all cases, the percentage of N-population in-
creases as the 2Ј-substituent of the nucleoside became more
electronegative (e.g. 2Ј-H Ͻ 2Ј-OH Ͻ 2Ј-F), this was pri-
Scheme 2. CAL-B-catalyzed acylation of nucleosides with aceto-
marily due to the increased strength of the 2Ј-substituent noxime levulinate in THF. See Table 1 for experimental results.
2
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
CAL-B-Catalyzed Acylation of Nucleosides
group for scale-up work. The acylation is carried out with tries 11–13). An increase in the ratio of N-conformer en-
3 equiv. of acetonoxime levulinate[6i] by using THF as sol- hanced the preference of CAL-B for acylation at the 5Ј-OH
vent and CAL-B as catalyst.
group. In the case of 2Ј-O-Me-ABz (1dc), the reaction
Enzymatic levulinylation of 2Ј-deoxyuridine (1aa) (see stalled after 60% conversion, and prolonged reaction times
footnote [d] in Table 1), with predominantly the S-type did not improve the conversion (Table 1, Entry 12). A dif-
sugar conformation in THF, furnished a mixture of three ferent behavior was observed in the enzymatic reactions of
acylated products 2aa, 3aa, and 4aa (Table 1, Entry 1). N2-isobutyryl-2Ј-deoxyguanosine (1ea) and N2-isobutyryl-
During acylation of various 2Ј-deoxynucleosides, CAL-B 2Ј-O-methylguanosine (1ec). CAL-B showed excellent re-
exhibited high selectivity towards the 5Ј-hydroxy group dur- gioselectivity during acylation of both compounds (1ea and
ing the acylation process, whereas the other acyl derivatives 1ec), although longer reaction times were required to drive
were obtained as minor compounds (Table 1, Entries 1, 5, the reaction to completion (Table 1, Entries 14, 15), likely
8 and 11). Analysis of the conformational preference of 2Ј- due to the low solubility of starting nucleosides in THF.
deoxy-2Ј-substituted uridine derivatives revealed an in- Among the 2Ј-deoxynucleosides studied, N2-isobutyryl-2Ј-
crease in the population of N-type conformation as the 2Ј- deoxyguanosine (1ea) was unique due to its analogous be-
substituent became more electronegative.[3,11] Thus, we per- havior to those of the ribonucleosides in which CAL-B
formed the enzymatic reaction with 2Ј-hydroxy, 2Ј-O- demonstrated total selectivity toward the 5Ј-OH group.
methyl, and 2Ј-deoxy-2Ј-fluoro derivatives (1ab, 1ac, and
In order to further establish the conformational prefer-
1ad, respectively) (Table 1, Entries 2–4). The acylation of ence of CAL-B, we investigated the lipase acylation process
nucleosides 1ac and 1ad was found to be facile and selective. with a nucleoside locked in the N-conformation. The locked
Interestingly, the acylation was highly dependent on the nucleic acid analogue of thymin-1-yl (LNA-T, 5, Scheme 3),
preference for N-type conformation. Levulinylation of the a synthetic structural analogue of thymidine with fixed N-
2Ј-deoxy-2Ј-fluoro derivative (1ad), with the N/S equilib- conformation, was an ideal molecule for this enzymatic
rium strongly biased toward N-type conformation in THF study.
was complete in 30 min, yielding exclusively the 5Ј-O-acyl-
ated compound 2ad (Table 1, Entry 4). The longer reaction
time necessary to complete acylation of uridine (1ab) was
probably due to the low solubility of this nucleoside in THF
(Table 1, Entry 2).
To explore the influence of the nucleobase during the
enzymatic process, we carried out CAL-B-catalyzed acyl-
Scheme 3. Enzymatic acylation of LNA-T with acetonoxime levul-
ation of thymidine (1ba) and 2Ј-methoxy-5-methyluridine
(1bc). The results compiled in Entries 5 and 6 of Table 1
indicate that CAL-B selectively acylated the nucleoside with
a preference for the N-conformation. In these experiments,
the 5Ј-O-regioselectivity for the acylation of thymidine was
modest (89%), whereas acylation of 2Ј-methoxy-derivative
afforded high regioselectivity (99%) toward the primary hy-
droxy group. Slightly shorter reaction times were observed
with uracil-1-yl than thymin-1-yl monomers, probably due
to the higher solubility of the former (Table 1, Entries 1 vs.
5 and 3 vs. 6).
Acylation of cytidine analogues including 2Ј-deoxycytid-
ine (1ca), 2Ј-O-methylcytidine (1cc) and N4-benzoyl-2Ј-de-
oxy-2Ј-fluorocytidine (1cd) derivatives furnished the ex-
pected results. Total selectivity was observed with the 2Ј-O-
Me and 2Ј-F nucleosides, both substrates have exclusively
an N-conformation in THF (Table 1, Entries 9 and 10). In
contrast, 2Ј-deoxycytidine, in which the S-type puckered ge-
ometry was favored, reacted with much lower regioselectiv-
ity furnishing a mixture of mono- and diacyl products
(Table 1, Entry 8).
inate catalyzed by CAL-B.
The excellent selectivity observed during acylation of
LNA-T (5) with CAL-B further demonstrated that the en-
zyme was able to accommodate a constrained structure in
the active site displaying the conformational preference
(Table 1, Entry 7). The protected nucleoside 6 is an impor-
tant building block for LNA-based therapeutic oligonucleo-
tides showing improved duplex stability and resistance to
nuclease degradation.[12]
Remarkably, a simple replacement of the 2Ј-H atom in
1ca with a fluorine atom in cytidine analogue 1cd (Table 1,
Entries 8 and 10) dramatically influenced the population of
N-conformation (41% for 1ca and 100% for 1cd) impacting
the selectivity during acylation with CAL-B. This data col-
lectively (Table 1) supports our hypothesis that the sugar
pucker plays a critical role in how nucleosides are recog-
nized and plausibly fit into the enzyme active site. These
results encouraged us to perform molecular modeling of the
CAL-B binding site with nucleosides hoping to better
understand the observed experimental data in which it was
apparent that the 2Ј-substituent clearly impacts acylation
profile.
It was also of interest to evaluate the relationship of re-
gioselectivity to conformation of purine nucleosides. Thus,
we investigated the acylation of several adenosine and gua-
nosine derivatives. The acylation reaction catalyzed by
CAL-B on 2Ј-deoxyadenosine (1da), 2Ј-O-Me-adenosine
(1dc) and N6-benzoyl-2Јdeoxy-2Ј-fluoroadenosine (1dd) de-
Molecular Modeling
CAL-B is one of the most widely used industrial lipases
rivatives took place with high regioselectivity (Table 1, En- due to its high enantioselectivity, acceptance of a wide
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
V. Gotor, M. Ferrero et al.
FULL PAPER
range of substrates, thermal stability, and stability in or- levulinylation at the 5Ј-position since it is the major regio-
ganic solvents. In contrast to most lipases, CAL-B has no isomer obtained in all cases. When geometry optimization
lid covering the entrance to the active site and shows no was carried out with 2Ј-deoxyuridine (1aa) (Figure 1), the
interfacial activation. It is composed of 317 amino acids tetrahedral intermediate for the levulinylation placed both
with a molecular weight of 33.5 kDa, and its X-ray struc- the sugar and the base rings inside the active site.
ture has been well studied.[13] The X-ray structure of CAL-
B (1LBS[13]) was obtained from the Protein Data Bank.[14]
The catalytic triad is formed by residues Asp187-His224-
Ser105 and the oxyanion hole is formed by the main chain
NH groups of residues Gln106, Thr40, and the hydroxy
group of the latter.
The starting point for modeling was the tetrahedral inter-
mediate (Scheme 4)[15] which mimics the transition state for
the levulinylation of nucleosides. Geometry optimization of
this simple intermediate formed all six essential hydrogen
bonds for catalysis (d1–d6 in Scheme 4).[16] To model each
of the individual transition states, we added the quaternary
carbon atom at the 5Ј-position corresponding to levulinyl
and its substituents (2Ј-deoxy-, ribo-, 2Ј-methoxy-, and 2Ј-
fluoro nucleosides; first the sugar and later the base) se-
quentially until we obtained the intermediate shown in
Scheme 4. We then systematically searched for catalytically
productive conformations. We defined the catalytically pro-
ductive conformations as those that (a) contained all six
key catalytic hydrogen bonds, (b) avoided steric clashes be-
tween the intermediate and the lipase, and (c) where access-
ible within the intermediate.
Figure 1. Best conformations of nucleoside derivative intermediates
(1aa–1ad) of the uracil family in CAL-B. The tetrahedral interme-
diates formed in all cases show the sugar conformation as North,
independently of the preferred conformation for the sugar in solu-
tion (closest to North or South). The effective binding of the nucle-
oside sugar moiety appears to be the key factor responsible for the
high regioselectivity. See Table 1 for details and discussion in the
text. The image, to ensure a clear view of relevant interactions,
displays only the intermediates (omitting all the rest of the atoms
from the enzyme) within the lipase catalytic site.
This binding site appears to be the most productive con-
formation where all six catalytically essential H-bonds (d1–
Scheme 4. Tetrahedral intermediate for 5Ј-O-levulinylation of β-d-
nucleosides. Intermediates of the transition states for the levulinyl-
ation of nucleoside derivatives 1 and 5. Key hydrogen bonds be-
tween CAL-B and the tetrahedral intermediate are: (i) one from
Nδ of His224 to the carboxylate group of Asp187 (d1), (ii) two
from Nε of His224 to the oxygen atom of Ser105 (d2) and the 5Ј-
O-group of the nucleoside (d3), and (iii) three from the oxyanion
to Gln106 (d4) and to Thr40 (d5 and d6). For computer modeling
the different groups were introduced in a stepwise fashion.
d6, Scheme 4) were involved in the substrate recognition. It
is noteworthy that the sugar conformation was North, quite
different from the major conformer of 2Ј-deoxyuridine in
solution (Table 1, Entry 1). Similar results were observed
with the tetrahedral intermediates for the levulinylation of
uridine (1ab), 2Ј-O-methyluridine (1ac) and 2Ј-fluorourid-
ine (1ad). When these structures were built and minimized,
the productive conformations obtained showed that the
sugar puckering was North, closer to their major conforma-
tion in THF solution (Table 1, Entries 2–4).
To qualitatively explain the behavior of CAL-B-catalyzed
acylations for each nucleoside derivative 1 and 5, 2Ј-deoxy-
In order to see if there is any secondary influence due to
or 2Ј-substituted nucleosides derived from uracil, thymine, the nature of the base, purine or pyrimidine, we also studied
cytidine, adenine and guanine, we focused on uracil ana- the intermediates in the adenosine and guanosine deriva-
logues as a representative example for this study (for more tives. As shown in Figure 2, the purine bases that are larger
results, see Supporting Information: Figures S2 and S3). in size relative to the pyrimidine nucleosides, point out
Thus, we modeled the key intermediates for nucleoside towards the solvent. As a result, the base protecting groups
4
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
CAL-B-Catalyzed Acylation of Nucleosides
are situated outside of the catalytic site. Therefore, in these for a shorter period. This may be the trigger point for CAL-
cases the sugar interaction is the primary driver for the ob- B to acylate 1ea without a loss of 5Ј-O-selectivity. We postu-
served selectivity during acylation.
late that the speed of acylation of 2Ј-deoxyguanosine is sig-
nificantly slower perhaps due to the reasonable energy of
activation required for interconversion of the S- to N-type
sugar in purine nucleosides (ca. 5 kcal/mol). We believe this
study offers, for the first time, insight into the design ele-
ments and the ability to predict how a nucleoside may un-
dergo acylation by simply reviewing the sugar conformation
in the solvent of choice.
Conclusions
Herein, we report the excellent 5Ј-O-regioselectivity dem-
onstrated by CAL-B during levulinylation of various nu-
cleoside analogues. Among the 15 nucleosides tested, ana-
logues that presented the sugar with the N-conformation
were acylated at a faster speed than those possessing the S-
conformation. The experimental data were clearly sup-
ported by molecular modeling that confirmed the critical
role and influence of sugar conformation on the speed and
selectivity of the acylation. The molecular modeling studies
confirmed that nucleosides are well recognized and accom-
modated effectively in the binding site of CAL-B en route
to the observed regioselectivity. The modeling studies also
confirmed that the base moiety (purine or pyrimidine) does
not have any significant influence either on the rate or selec-
tivity of the acylation reaction. The modeling data indicated
that the N-conformation is possibly the key catalytic inter-
mediate responsible for the observed selectivity. We specu-
late that the conformational flexibility of 2Ј-deoxynucleos-
ides is the key factor in maintaining the regioselectivity dur-
ing the acylation step. Therefore, a nucleoside analogue
containing the S-conformation may undergo slow, but se-
lective, acylation as long as the sugar is able to transiently
adopt the N-conformation ideal for favorable binding at the
catalytic site of CAL-B. Collectively, we believe this study
significantly improves our understanding of CAL-B-medi-
ated acylation of nucleosides and increases our knowledge
of enzyme–substrate recognition. This study further rein-
forces the utilization of biocatalysis as an important tool
for organic chemists working to manufacture nucleoside-
based therapeutic products.
Figure 2. Best conformations of nucleoside derivative intermediates
with pyrimidine (1aa, 1ba) and purine (1da, 1ea) bases in CAL-B.
The tetrahedral intermediates formed in all cases show the sugar
conformation as North, independent of the preferred conformation
for the sugar in solution (closest to North or South). Purine bases
are larger in size relative to the pyrimidine bases and are pointing
out towards the solvent; a portion of the base protecting groups is
placed outside of the catalytic site, showing no dependence of base
in the catalytic intermediate. The image displays the intermediates
in a stick representation, and the rest of the atoms from the enzyme
are in a surface representation. NOTE: In order to appreciate the
details of various interactions, full page views are available in the
Supporting Information.
In addition to the sugar conformation, the solubility of
nucleosides in THF may also influence the outcome of the
acylation reactions. For example, the preferred N-confor-
mation and relatively higher solubility of 2Ј-deoxy-2Ј-
fluorouridine (1ad) in THF may have contributed to the
fastest rate of acylation compared to the other 14 nucleo-
sides studied in this work. Similarly, the poor solubility of
2Ј-deoxyguanosine (1ea) in THF likely explains the longest
reaction time (24 h) for acylation among the nucleosides
studied. More importantly, these results indicate that both
types of nucleosides with the N- or S-conformation will un-
dergo selective acylation events but at significantly different
speeds. The N/S conformation ratio results from a dynamic
process where the energy barrier between the two confor-
mations is around 1–2 kcal/mol. Therefore, the nucleosides
that are predisposed to assuming the N-conformation poss-
ibly undergo faster acylation, clearly demonstrated by 2Ј-
deoxy-2Ј-fluorouridine (1ad). On the other hand, the sugar
in 2Ј-deoxyguanosine (1ea) remains in the S-conformation
Experimental Section
General: Candida antarctica lipase B (CAL-B, Novozym 435, 7300
PLU/g) was a gift from Novo Nordisk Co. Conversions and re-
gioselectivity were determined by HPLC using Spherisorb W 5 μm
(0.46ϫ25 cm), Lichrosorb 10 μm (0.46ϫ15 cm) or Mediterranea
Sea18 (0.45ϫ25 cm) columns at 30 °C. Specific experimental con-
ditions used are indicated below together with the experimental
data for each 5Ј-O-acylated nucleoside.
General Procedure for the Enzymatic Acylation of Nucleosides (1
and 5): In a standard procedure, THF was added to an Erlenmeyer
flask that contained the appropriate nucleoside derivative
(0.1 mmol), acetonoxime levulinate[6i] (0.3 mmol) and CAL-B (1:1;
w/w) under nitrogen. The reaction mixture was shaken at 250 rpm
the majority of the time but does exist in N-conformation by using a Heidolph apparatus and the reaction progress monitored
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
V. Gotor, M. Ferrero et al.
FULL PAPER
by HPLC. The immobilized enzyme was filtered off and washed
with CH2Cl2, and the combined solvents were distilled under vac-
uum. The crude product was purified by flash chromatography
(gradient eluents 2–5% MeOH/CH2Cl2) furnishing 5Ј-O-levulinyl-
ated nucleosides 2 and 6. The isolated yield for each product is
shown in parentheses (e.g. 86% for 2aa) along with their analytical
data. Syntheses of compounds 2ab,[6b] 2ba,[6h] 3ba,[6h] 4ba,[6h]
2ca,[6h] 3ca,[6g] 4ca,[6h] 2da,[6h] 3da,[6g] 4da,[6h] 3dc,[9a] 2ea,[6h] 4ea,[6h]
and 5[12f] have been described previously.
70–72 °C. IR (KBr): ν = 3392, 3055, 1743, 1691 cm–1. [α]20 = –33
˜
D
(c = 0.35, CH2Cl2). 1H NMR (CDCl3, 300.13 MHz): δ = 1.81 (s, 3
2
H), 2.15 (s, 3 H), 2.59 (m, 2 H), 2.80 (m, 2 H), 3.79 (d, JHH
=
2
8.1 Hz, 1 H), 4.01 (d, JHH = 7.9 Hz, 1 H), 4.08 (s, 1 H), 4.39 (d,
2JHH = 12.7 Hz, 1 H), 4.43 (s, 1 H), 4.53 (d, JHH = 12.7 Hz, 1 H),
2
5.52 (s, 1 H), 7.44 (s, 1 H), 10.04 (br. s, 1 H) ppm. 13C NMR
(CDCl3, 75.5 MHz): δ = 12.5, 27.6, 29.6, 37.7, 62.8, 70.0, 72.0, 79.1,
86.6, 86.9, 109.9, 134.7, 150.0, 164.4, 172.7, 207.0 ppm. HRMS
(ESI+): calcd. for C16H20N2NaO8 [M + Na]+ 391.1112, found
391.1106. HPLC: Mediterranea, mixture 12% MeCN/H2O for
7 min and then gradient to reach 40% MeCN at 20 min,
0.8 mLmin–1, tR = 18.7 min.
5Ј-O-Levulinyl-2Ј-deoxyuridine (2aa): White solid (86%). Rf (10%
MeOH/CH2Cl2) = 0.47. M.p. 69–71 °C. [α]2D0 = +15 (c = 0.5,
1
CH2Cl2). H NMR (CDCl3, 300.13 MHz): δ = 2.18 (m, 4 H), 2.62
3
N6-Benzoyl-5Ј-O-levulinyl-2Ј-O-methylcytidine (2cc): White solid
(94%). Rf (10% MeOH/CH2Cl2) = 0.60. M.p. 142–144 °C. IR
(m, 3 H), 2.75 (m, 2 H), 4.31 (m, 5 H), 5.79 (d, JHH = 7.7 Hz, 1
3
H), 6.26 (m, 1 H), 7.63 (d, JHH = 7.8 Hz, 1 H), 10.14 (br. s, 1 H)
ppm. 13C NMR (CDCl3, 75.5 MHz): δ = 27.8, 29.7, 38.0, 40.4,
64.0, 71.2, 84.4, 85.5, 102.6, 140.2, 150.6, 164.1, 172.7, 207.4 ppm.
HRMS (ESI+): calcd. for C14H18N2NaO7 [M + Na]+ 349.1006,
found 349.0997. HPLC: Spherisorb W, mixture 5% MeOH/CH2Cl2
for 9.5 min and then gradient to reach 20% MeOH at 16 min,
0.8 mLmin–1, tR = 8.0 min.
(KBr): ν = 3056, 2986, 1743, 1703, 1667 cm–1. [α]20 = +36 (c =
˜
D
1
0.35, CH2Cl2). H NMR (CDCl3, 300.13 MHz): δ = 2.20 (s, 3 H),
2.62 (m, 2 H), 2.83 (m, 2 H), 3.71 (s, 3 H), 3.87 (d, JHH = 5.0 Hz,
1 H), 4.12 (dd, JHH = 5.0, JHH 9.2 Hz, 1 H), 4.20 (m, 1 H), 4.50
3
3
3
3
(m, 2 H), 5.96 (s, 1 H), 7.50 (t, JHH = 7.5 Hz, 2 H), 7.58 (m, 2 H),
7.91 (m, 2 H), 8.28 (d, JHH = 7.4 Hz, 1 H) ppm. 13C NMR
3
(CDCl3, 75.5 MHz): δ = 27.9, 29.8, 37.9, 58.9, 62.2, 68.2, 81.4, 83.3,
89.0, 96.6, 127.7, 129.0, 132.0, 144.3, 154.7, 162.7, 166.7, 172.5,
206.6 ppm. HRMS (ESI+): calcd. for C22H26N3O8 [M + H]+
460.1714, found 460.1724. HPLC: Lichrosorb, mixture 3% MeOH/
CH2Cl2, for 7 min and then gradient to reach 8% MeOH at 12 min,
0.8 mLmin–1, tR = 7.6 min.
5Ј-O-Levulinyl-2Ј-O-methyluridine (2ac): White solid (92%). Rf
(10% MeOH/CH Cl ) = 0.50. M.p. 114–116 °C. IR (KBr): ν =
˜
2
2
3226, 3107, 3057, 2953, 1746, 1709, 1677 cm–1. [α]2D0 = +64 (c = 1.0,
1
CH2Cl2). H NMR (CDCl3, 300.13 MHz): δ = 2.21 (s, 3 H), 2.58
(t, 3JHH = 6.2 Hz, 2 H), 2.82 (t, 3JHH = 6.2 Hz, 2 H), 3.65 (s, 3 H),
3
3
3.86 (dd, JHH = 2.1, JHH 4.8 Hz, 1 H), 4.15 (m, 2 H), 4.46 (m, 2
3
3
N6-Benzoyl-2Ј-deoxy-2Ј-fluoro-5Ј-O-levulinylcytidine (2cd): White
solid (92%). Rf (10% MeOH/CH2Cl2) = –0.46. M.p. 150–152 °C.
H), 5.80 (d, JHH = 8.1 Hz, 1 H), 5.92 (d, JHH = 2.0 Hz, 1 H),
7.73 (d, JHH = 8.1 Hz, 1 H), 9.28 (br. s, 1 H) ppm. 13C NMR
3
IR (KBr): ν = 3336, 3290, 3126, 3057, 2987, 1738, 1702, 1656 cm–1.
(CDCl3, 75.5 MHz): δ = 27.8, 29.7, 37.9, 58.8, 62.5, 68.6, 81.6, 83.2,
87.9, 102.5, 139.5, 150.1, 163.2, 172.4, 206.5 ppm. HRMS (ESI+):
calcd. for C15H20N2NaO8 [M + Na]+ 379.1112, found 379.1083.
HPLC: Lichrosorb, mixture 4% MeOH/CH2Cl2, 0.75 mLmin–1, tR
= 6.3 min.
˜
[α]2D0 = +78 (c = 0.5, CH2Cl2). H NMR (CDCl3, 300.13 MHz): δ
1
= 2.18 (s, 3 H), 2.60 (t, 3JHH = 6.2 Hz, 2 H), 2.80 (t, 3JHH = 6.2 Hz,
2 H), 4.25–4.57 (m, 4 H), 5.19 (dd, 3JHH = 3.7, 2JHF 52.4 Hz, 1 H),
5.95 (d, JHF = 17.6 Hz, 1 H), 7.48 (t, JHH = 7.5 Hz, 2 H), 7.58
3
3
3
3
(m, 2 H), 7.93 (d, JHH = 7.3 Hz, 2 H), 8.17 (d, JHH = 7.6 Hz, 1
2Ј-Deoxy-2Ј-fluoro-5Ј-O-levulinyluridine (2ad): White foam (95%).
H) ppm. 13C NMR (CDCl3, 75.5 MHz): δ = 27.8, 29.7, 37.9, 62.2,
R (10% MeOH/CH Cl ) = 0.47. IR (KBr): ν = 3414, 3211, 3057,
˜
f
2
2
3
2
2953, 1695, 1496 cm–1. [α]2D0 = +14 (c = 0.5, CH2Cl2). 1H NMR
68.5, 80.5, 90.5 (d, JCF = 34.6 Hz), 93.4 (d, JCF = 187.6 Hz),
97.0, 127.7, 128.9, 132.8, 133.2, 144.8, 154.9, 163.0, 166.8, 172.6,
207.1 ppm. HRMS (ESI+): calcd. for C21H23FN3O7 [M + H]+
448.1515, found 448.1497. HPLC: Spherisorb W, mixture 3%
MeOH/CH2Cl2, for 7 min and then gradient to reach 8% MeOH
at 12 min, 0.75 mLmin–1, tR = 11.0 min.
(CDCl3, 300.13 MHz): δ = 2.19 (s, 3 H, CH3-Lev), 2.57 (m, 2 H),
2
2.78 (m, 2 H), 4.18–4.49 (m, 4 H), 5.13 (d, JHF = 52.9 Hz, 1 H),
3
3
5.77 (d, JHH = 8.0 Hz, 1 H), 5.89 (d, JHF = 18.7 Hz, 1 H), 7.62
(d, 3JHH = 8.1 Hz, 1 H), 10.13 (br. s, 1 H) ppm. 13C NMR (CDCl3,
75.5 MHz): δ = 27.6, 29.8, 37.8, 62.6, 68.9, 80.5, 89.4 (d, JCF
3
=
2
N6-Benzoyl-5Ј-O-levulinyl-2Ј-O-methyladenosine (2dc): White foam
35.1 Hz), 93.3 (d, JCF = 187.1 Hz), 102.5, 141.0, 150.3, 164.0,
172.7, 207.7 ppm. HRMS (ESI+): calcd. for C14H17FN2NaO7 [M
+ Na]+ 367.0912, found 367.0905. HPLC: Spherisorb W mixture
3% MeOH/CH2Cl2 for 5.0 min and then gradient to reach 6%
(55%). R (10% MeOH/CH Cl ) = 0.52. IR (KBr): ν = 3405, 2922,
˜
f
2
2
1712, 1613 cm–1. [α]2D0 = –10 (c = 0.5, CH2Cl2). H NMR (CDCl3,
300.13 MHz): δ = 2.14 (s, 3 H), 2.68 (m, 2 H), 2.77 (m, 2 H), 3.59
(s, 3 H), 4.36–4.49 (m, 4 H), 4.44 (m, 2 H), 6.31 (m, 1 H), 7.49 (t,
1
MeOH at 10 min and 9% MeOH at 12 min, 0.8 mLmin–1, tR
=
3JHH = 7.7 Hz, 2 H), 7.58 (t, JHH = 7.6 Hz, 1 H), 8.06 (d, JHH
=
C
9.2 min.
3
3
7.5 Hz, 2 H), 8.40 (s, 1 H), 8.75 (s, 1 H), 9.87 (br. s, 1 H) ppm. 13
5Ј-O-Levulinyl-2Ј-methoxy-5-methyluridine (2bc): White foam
NMR (CDCl3, 75.5 MHz): δ = 27.7, 29.7, 37.8, 58.9, 63.0, 69.4,
82.3, 83.1, 123.6, 128.3, 128.6, 132.7, 133.4, 141.7, 149.9, 151.5,
152.3, 165.4, 172.6, 176.7, 206.1 ppm. HRMS (ESI+): calcd. for
C23H26N5O7 [M + H]+ 484.1827, found 484.1837. HPLC: Lichro-
sorb, mixture 5% MeOH/CH2Cl2, 0.8 mLmin–1, tR = 8.2 min.
(95%). R (10% MeOH/CH Cl ) = 0.54. IR (KBr): ν = 3392, 3202,
˜
f
2
2
3066, 2932, 1742, 1703, 1681 cm–1. [α]2D0 = +20 (c = 1.0, CH2Cl2).
1H NMR (CDCl3, 300.13 MHz): δ = 1.92 (s, 3 H), 2.19 (s, 3 H),
3
2.61 (m, 2 H), 2.80 (m, 2 H), 3.59 (s, 3 H), 3.83 (dd, JHH = 2.3,
3
3
3JHH 4.9 Hz, 1 H), 4.12 (m, 2 H), 4.36 (dd, JHH = 2.3, JHH
3
3
12.6 Hz, 1 H), 4.50 (dd, JHH = 3.5, JHH 12.6 Hz, 1 H), 5.92 (m,
3JHH = 2.3 Hz, 1 H), 7.41 (s, 1 H), 9.87 (br. s, 1 H) ppm. 13C NMR
(CDCl3, 75.5 MHz): δ = 12.3, 27.8, 29.7, 37.8, 58.7, 62.8, 68.8, 81.4,
83.0, 87.8, 111.0, 135.3, 150.3,164.1, 172.4, 206.6 ppm. HRMS
(ESI+): calcd. for C16H22N2NaO8 [M + Na]+ 393.1268, found
393.1252. HPLC: Spherisorb W, mixture 7% MeOH/CH2Cl2,
0.8 mLmin–1, tR = 4.6 min.
N6-Benzoyl-2Ј-deoxy-2Ј-fluoro-5Ј-O-levulinyladenosine (2dd): White
solid (91%). Rf (10% MeOH/CH2Cl2) = 0.49. M.p. 63–65 °C. IR
(KBr): ν = 3396, 3055, 2987, 1713, 1612 cm–1. [α]20 = –5 (c = 0.35,
˜
D
1
CH2Cl2). H NMR (CDCl3, 300.13 MHz): δ = 2.14 (s, 3 H), 2.53
(m, 2 H), 2.73 (m, 2 H), 4.32 (m, 1 H), 4.44 (m, 2 H), 4.82 (ddd,
3
3
3
3JHH = 4.3, JHH 7.9, JHF 20.3 Hz, 1 H), 5.54 (dd, JHH = 4.1,
2JHF 52.5 Hz, 1 H), 6.30 (d, JHF = 17.7 Hz, 1 H), 7.49 (t, JHH
=
3
3
3
3
5Ј-O-Levulinyl-2Ј-O,4Ј-C-methylene-
D
-ribofuranosyl-5-methylurid- 7.5 Hz, 2 H), 7.59 (t, JHH = 7.2 Hz, 1 H), 8.02 (d, JHH = 7.5 Hz,
ine (6): White solid (89%). Rf (10% MeOH/CH2Cl2): 0.44; M.p.
2 H), 8.31 (s, 1 H), 8.73 (s, 1 H), 9.45 (br. s, 1 H) ppm. 13C NMR
6
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
CAL-B-Catalyzed Acylation of Nucleosides
2005, 127, 15145–15150; b) V. E. Marquez, T. Ben-Kasus, J. J.
Barchi Jr., K. M. Green, M. C. Nicklaus, R. Agbaria, J. Am.
Chem. Soc. 2004, 126, 543–549.
a) Antisense Drug Technology Principles: Strategies and Appli-
cations (Ed.: S. T. Crooke), CRC Press, Boca Raton, FL, 2008;
b) Y. S. Sanghvi, “A Status Update of Modified Oligonucleo-
tides for Chemotherapeutics Applications” in Current Proto-
cols in Nucleic Acid Chemistry, John Wiley & Sons, 4.1.1–
4.1.22, September 2011.
a) S. Martínez-Montero, S. Fernández, Y. S. Sanghvi, J. Chat-
topadhyaya, M. Ganesan, N. G. Ramesh, V. Gotor, M. Ferrero,
J. Org. Chem. 2012, 77, 4671–4678; b) S. Martínez-Montero, S.
Fernández, Y. S. Sanghvi, V. Gotor, M. Ferrero, J. Org. Chem.
2010, 75, 6605–6613; c) S. Martínez-Montero, S. Fernández, T.
Rodríguez-Pérez, Y. S. Sanghvi, K. Wen, V. Gotor, M. Ferrero,
Eur. J. Org. Chem. 2009, 3265–3271; d) J. García, A. Díaz-
Rodríguez, S. Fernández, Y. S. Sanghvi, M. Ferrero, V. Gotor,
J. Org. Chem. 2006, 71, 9765–9771; e) A. Díaz-Rodríguez, S.
Fernández, Y. S. Sanghvi, M. Ferrero, V. Gotor, Org. Process
Res. Dev. 2006, 10, 581–587; f) J. García, S. Fernández, M.
Ferrero, Y. S. Sanghvi, V. Gotor, Org. Lett. 2004, 6, 3759–3762;
g) J. García, S. Fernández, M. Ferrero, Y. S. Sanghvi, V. Gotor,
Tetrahedron: Asymmetry 2003, 14, 3533–3540; h) J. García, S.
Fernández, M. Ferrero, Y. S. Sanghvi, V. Gotor, J. Org. Chem.
2002, 67, 4513–4519; i) For synthesis of acetonoxime levulinate
see: I. Lavandera, J. García, S. Fernández, M. Ferrero, V. Go-
tor, Y. S. Sanghvi, “Enzymatic Regioselective Levulinylation of
2Ј-Deoxyribonucleosides and 2Ј-O-Methylribonucleosides” in
Current Protocols in Nucleic Acid Chemistry (Eds.: M. Egli, P.
Herdewijn, A. Matsuda, Y. S. Sanghvi), John Wiley & Sons,
New York, 2011, chapter 2.11, pp. 2.11.1–2.11.36.
(CDCl3, 75.5 MHz): δ = 27.7, 29.7, 37.8, 62.6, 69.4, 80.9, 87.3 (d,
2
3JCF = 34.1 Hz), 93.3 (d, JCF = 188.0 Hz), 123.4, 128.0, 128.8,
132.9, 133.4, 141.9, 149.8, 151.1, 153.3, 165.0, 172.6, 207.2 ppm.
HRMS (ESI+): calcd. for C22H23FN5O6 [M + H]+ 472.1627, found
472.1638. HPLC: Mediterranea, mixture 12% MeCN/H2O for
7 min and then gradient to reach 40% MeCN at 20 min,
0.8 mLmin–1, tR = 19.6 min.
[5]
[6]
N2-Isobutyryl-5Ј-O-levulinyl-2Ј-O-methylguanosine (2ec): White so-
lid (88%). Rf (10% MeOH/CH2Cl2) = 0.42. M.p. 52–54 °C. IR
(KBr): ν = 3194, 3055, 2984, 2937, 1716, 1682 cm–1. [α]20 = –20 (c
˜
D
1
= 1.2, CH2Cl2). H NMR ([D4]MeOH, 300.13 MHz): δ = 1.23 (d,
3JHH = 6.8 Hz, 6 H), 2.15 (s, 3 H), 2.58 (m, 2 H), 2.80 (m, 3 H),
3
3.54 (s, 3 H), 4.20 (m, 1 H), 4.29–4.48 (m, 4 H), 6.05 (d, JHH
=
4.1 Hz, 1 H), 8.17 (s, 1 H) ppm. 13C NMR ([D4]MeOH, 75.5 MHz):
δ = 17.9, 27.4, 28.2, 35.5, 37.2, 57.7, 63.2, 69.3, 82.2, 83.2, 86.6,
120.2, 138.1, 148.4, 148.9, 156.0, 172.8, 180.4, 208.0 ppm. HRMS
(ESI+): calcd. for C20H28N5O8 [M + H]+ 466.1932, found 466.1938.
HPLC: Spherisorb W, mixture 8% MeOH/CH2Cl2 for 5.0 min and
then gradient to reach 15% MeOH at 15 min, 0.8 mLmin–1, tR
=
7.3 min.
Molecular Modeling: All minimizations were performed with an
iMac by using the molecular modeling package Molecular Op-
erating Environment 2011.10 (Chemical Computing Group,
Inc.).[17] The Amber99 force field[18] and the corresponding dictio-
nary charges were used as implemented in the program version
2011.10. Due to the lack of parameters for THF compatible with
the force field used, we opted for a continuum solvation method,
an approach that has already afforded good results in similar stud-
ies.[19] In particular, a distance-dependent relative dielectric con-
stant of 2 was selected as a good compromise between the dielectric
constant of THF[20] and that of a globular protein matrix.[21] The
convergence criterion of the energy minimization was set to an rms
gradient value of 0.00001 kcal/mol. High quality pictures of repre-
sentative structures in Figure 1 and 2 were created by using Py-
MOL 0.99.[22]
[7]
[8]
a) N. Li, T. J. Smith, M.-H. Zong, Biotechnol. Adv. 2010, 28,
348–366; b) M. Ferrero, V. Gotor, Chem. Rev. 2000, 100, 4319–
4347; c) M. Ferrero, V. Gotor, Monatsh. Chem. 2000, 131, 585–
616.
a) S. K. Mahto, C. S. Chow, Bioorg. Med. Chem. 2008, 16,
8795–8800; b) S. Obika, K. Morio, D. Nanbu, T. Imanishi,
Chem. Commun. 1997, 1643–1644; c) F. A. A. M. de Leeuw, C.
Altona, J. Chem. Soc. Perkin Trans. 2 1982, 375–384; d) C.
Altona, M. Sundaralingam, J. Am. Chem. Soc. 1973, 95, 2333–
2344; e) Altona and Sundaralingam protocol utilizing the ob-
served coupling-constants have been widely used to support the
two-state (N/S) conformational model for nucleosides. Briefly,
the experimental NMR coupling constants represent time-
averaged couplings related to the couplings of the individual
conformers. It was shown that the percentage of S-type con-
former in solution of various nucleosides can be calculated by
using one of the two Equations (1) or (2). The interpretation
of NMR coupling constants was supported by X-ray studies
and pseudorotational model of five-membered rings.[8d]
a) For comprehensive literature, see: Current Protocols in Nu-
cleic Acid Chemistry (Eds.: M. Egli, P. Herdewijn, A. Matsuda,
Y. S. Sanghvi), John Wiley & Sons, New York, 2011, chapter
2; b) Protective Groups in Organic Synthesis 4th ed. (Eds.:
P. G. M. Wuts, T. W. Green), John Wiley & Sons, New York,
2007; c) Protecting Groups, 3rd ed. (Ed.: P. J. Kocienski), Georg
Thieme Verlag, Stuttgart, Germany, 2005.
A. Grajkowski, J. Cieslak, A. Gapeev, C. Schindler, S. L. Be-
aucage, Bioconjugate Chem. 2010, 21, 2147–2152.
W. Guschlbauer, K. Jankowski, Nucleic Acids Res. 1980, 8,
1421–1433.
a) C. B. Hughesman, R. F. B. Turner, C. A. Haynes, Biochemis-
try 2011, 50, 5354–5368; b) D. A. Di Giusto, G. C. King, Nu-
cleic Acids Res. 2004, 32, e32; c) B. Vester, J. Wengel, Biochem-
istry 2004, 43, 13233–13241; d) S. K. Singh, P. Nielsen, A. A.
Koshkin, J. Wengel, Chem. Commun. 1998, 455–456; e) A. A.
Koshkin, S. K. Singh, P. Nielsen, V. K. Rajwanshi, R. Kumar,
M. Meldgaard, C. E. Olsen, J. Wengel, Tetrahedron 1998, 54,
3607–3630; f) A. A. Koshkin, J. Fensholdt, H. M. Pfundheller,
C. Lomholt, J. Org. Chem. 2001, 66, 8504–8521.
Supporting Information (see footnote on the first page of this arti-
cle): 1H, 13C and DEPT NMR spectra of 5Ј-O-levulinyl analogs for
1–5, details of molecular modeling experiments, calculation of the
percentage of N/S conformation by NMR spectroscopy exemplified
for 1ab and full page size version of Figure 2.
Acknowledgments
[9]
Financial support of this work by the Spanish Ministerio de Educa-
ción y Ciencia (MEC) (Project MEC-CTQ-2007-61126) and the
Plan de Ciencia Tecnología e Innovación (PCTI) del Principado
de Asturias (2006–2009)/Fondo Europeo de Desarrollo Regional
(FEDER) (2007–2013) (Project EQUIP09-07) is gratefully ac-
knowledged. S. M.-M. is thankful to MEC for a predoctoral
Formación de Profesorado Universitario (FPU) fellowship.
[10]
[11]
[12]
[1] C. Altona, M. Sundaralingam, J. Am. Chem. Soc. 1972, 94,
8205–8212.
[2] W. Saenger (Ed.), Principles of Nucleic Acid Structure,
Springer-Verlag, New York, 1984.
[3] C. Thibaudeau, P. Acharya, J. Chattopadhyaya, Stereoelec-
tronic Effects in Nucleosides and Nucleotides and Their Struc-
tural Implications, 2nd ed., Uppsala University Press, Uppsala,
Sweden, 2005.
[4] a) V. E. Marquez, Y. Choi, M. J. Comin, P. Russ, C. George,
M. Huleihel, T. Ben-Kasus, R. Agbaria, J. Am. Chem. Soc.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
V. Gotor, M. Ferrero et al.
FULL PAPER
[13] J. Uppenberg, N. Öhrner, M. Norin, K. Hult, G. J. Kleywegt,
S. Patkar, V. Waagen, T. Anthonsen, T. A. Jones, Biochemistry
1995, 34, 16838–16851.
10, 1830–1838; b) V. Léonard, L. Fransson, S. Lamare, K.
Hult, M. Graber, ChemBioChem 2007, 8, 662–667; c) M. Cam-
menberg, K. Hult, S. Park, ChemBioChem 2006, 7, 1745–1749;
d) S. Raza, L. Fransson, K. Hult, Protein Sci. 2001, 10, 329–
338.
The dielectric constant of THF at 20 °C is 7.58; see J. G.
Speight, Perry’s Standard Tables and Formulas for Chemical En-
gineers, McGraw-Hill, New York, 2003.
[14] http://www.rcsb.org/pdb/.
[15] I. Lavandera, S. Fernández, J. Magdalena, M. Ferrero, R. J.
Kazlauskas, V. Gotor, ChemBioChem 2005, 6, 1381–1390.
[16] To identify hydrogen bonds, a donor–acceptor separation of
less than 3.20 Å and a donor–hydrogen–acceptor angle of 120°
or larger are required.
[17] Molecular Operating Environment 2011.10, Chemical Comput-
ing Group, Inc., Montréal, Québec, Canada, 2011, http://
www.chemcomp.com.
[20]
[21]
M. Laberge, Biochim. Biophys. Acta 1998, 1386, 305–330.
[22] W. L. DeLano, The PyMOL Molecular Graphics System, DeL-
ano Scientific LLC, San Carlos, CA, 2012, http://www.pymol.
org.
[18] J. Wang, P. Cieplak, P. A. Kollman, J. Comput. Chem. 2000,
Received: May 9, 2012
Published Online: ᭿
21, 1049–1074.
[19] a) E. García-Urdiales, N. Ríos-Lombardía, J. Mangas-
Sánchez, V. Gotor-Fernández, V. Gotor, ChemBioChem 2009,
8
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O20609
/KAP1
Date: 22-08-12 16:27:21
Pages: 9
CAL-B-Catalyzed Acylation of Nucleosides
Nucleoside Sugar Conformation
S. Martínez-Montero, S. Fernández,
Y. S. Sanghvi, V. Gotor,*
M. Ferrero* ....................................... 1–9
CAL-B-Catalyzed Acylation of Nucleos-
ides and Role of the Sugar Conformation:
An Improved Understanding of the En-
zyme-Substrate Recognition
The sugar conformation of nucleosides
plays a critical role in Candida antarctica
lipase B catalyzed acylation. Specifically,
the North nucleoside sugar ring confor-
mation is preferred for efficient binding at
the catalytic site. Nuclear magnetic reson-
ance spectroscopy was used to establish the
sugar conformation, which was confirmed
by molecular modeling studies.
Keywords: Nucleosides / Enzymes / Acyl-
ation / Sugar conformation / Molecular
modeling
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
9