The economical synthesis of [2'-(13)C, 1,3-(15)N2]uridine; preliminary conformational studies by solid state NMR.

Article abstract of DOI:10.1039/b301275a

The synthesis of [2'-(13)C, 1,3-(15)N2]uridine 11 was achieved as follows. An epimeric mixture of D-[1-(13)C]ribose 3 and D-[1-(13)C]arabinose 4 was obtained in excellent yield by condensation of K13CN with D-erythrose 2 using a modification of the Kiliani-Fischer synthesis. Efficient separation of the two aldose epimers was pivotally achieved by a novel ion-exchange (Sm3+) chromatography method. D-[2-(13)C]Ribose 5 was obtained from D-[1-(13)C]arabinose 4 using a Ni(II) diamine complex (nickel chloride plus TEMED). Combination of these procedures in a general cycling manner can lead to the very efficient preparation of specifically labelled 13C-monosaccharides of particular chirality. 15N-labelling was introduced in the preparation of [2'-(13)C, 1,3-(15)N2]uridine 11 via [15N2]urea. Cross polarisation magic angle spinning (CP-MAS) solid-state NMR experiments using rotational echo double resonance (REDOR) were carried out on crystals of the labelled uridine to show that the inter-atomic distance between C-2' and N-1 is closely similar to that calculated from X-ray crystallographic data. The REDOR method will be used now to determine the conformation of bound substrates in the bacterial nucleoside transporters NupC and NupG.

Full text of DOI:10.1039/b301275a

View Article Online / Journal Homepage / Table of Contents for this issue
The economical synthesis of [2Ј-13C, 1,3-15N2]uridine;
preliminary conformational studies by solid state NMR
Simon G. Patching,a David A. Middleton,b Peter J. F. Hendersona and Richard B. Herbert*a
a Astbury Centre for Structural Molecular Biology, University of Leeds, Leeds, UK LS2 9JT.
E-mail: r.b.herbert@chem.leeds.ac.uk
b Department of Biomolecular Sciences, UMIST, PO Box 88, Manchester, UK M60 1QD
Received 13th February 2003, Accepted 23rd April 2003
First published as an Advance Article on the web 9th May 2003
The synthesis of [2Ј-13C, 1,3-15N2]uridine 11 was achieved as follows. An epimeric mixture of -[1-13C]ribose 3
and -[1-13C]arabinose 4 was obtained in excellent yield by condensation of K13CN with -erythrose 2 using a
modification of the Kiliani–Fischer synthesis. Efficient separation of the two aldose epimers was pivotally achieved
by a novel ion-exchange (Sm3ϩ) chromatography method. -[2-13C]Ribose 5 was obtained from -[1-13C]arabinose 4
using a Ni() diamine complex (nickel chloride plus TEMED). Combination of these procedures in a general cycling
manner can lead to the very efficient preparation of specifically labelled 13C-monosaccharides of particular chirality.
15N-labelling was introduced in the preparation of [2Ј-13C, 1,3-15N2]uridine 11 via [15N2]urea. Cross polarisation magic
angle spinning (CP-MAS) solid-state NMR experiments using rotational echo double resonance (REDOR) were
carried out on crystals of the labelled uridine to show that the inter-atomic distance between C-2Ј and N-1 is closely
similar to that calculated from X-ray crystallographic data. The REDOR method will be used now to determine the
conformation of bound substrates in the bacterial nucleoside transporters NupC and NupG.
dual-isotope labelling of a nucleoside substrate uridine 11. We
report also on the application of CP-MAS NMR spectroscopy
to confirm an inter-atomic distance in uridine 11 in the crystal-
line state and thereby to validate the REDOR method used.
Introduction
All living cells utilise membrane transport proteins to assimilate
nutrients and to expel waste products through cell membranes.1
The concentrative (Hϩ-driven) nucleoside transport protein
NupC from Escherichia coli is one such, important, trans-
port protein.2 E. coli also synthesises another concentrative
nucleoside transporter, NupG,3 the gene sequence of which is,
curiously, unrelated to that of NupC. It has been deduced ten-
tatively from structure–activity relationships with the substrate
adenosine 1 and other nucleosides that the conformation of
compounds on binding to NupC is required to be anti whereas
on binding to NupG the more inherently crowded syn-con-
formation can be accommodated (Fig. 1).4 This appears to be an
important distinction between the two transport proteins rele-
Results and discussion
Synthesis of [2Ј-13C, 1,3-15N2]uridine
vant to the structures of the binding sites in the proteins and
worthy of further exploration. Uridine and adenosine have very
similar binding and transport characteristics. Complexities with
the (NMR) analysis of protein-bound adenosine has directed
us first to synthesise [2Ј-13C, 1,3-15N2]uridine 11, in order to
study substrate binding conformations in NupC and NupG
using the solid state NMR method of cross polarisation magic-
angle spinning (CP-MAS) NMR. This method enables study of
the transporters in their natural states within native mem-
branes.4,5 In collaboration with Professor Watts, Oxford and
others we are pioneering the application of solid state NMR
spectroscopy to the solution of structure and mechanism of
action of membrane transport proteins.4–6 Implementation
of such studies is critically dependent on the availability of
appropriately labelled nucleoside derivatives. We demonstrate
here innovative and efficient procedures for accomplishing
Model building on uridine with reference to known X-ray
crystallographic data7 led to the conclusion that the appropri-
ate NMR experiments (below) were best carried out with a
13C-label at C-2Ј and a 15N-label at N-3 (Fig. 2). It was simpler
to synthesise a compound with a 15N-label also at N-1; this label
could provide a harmless stowaway or internal standard in the
NMR experiments, since the distance between N-1 and C-2Ј in
the sugar, unlike C-2Ј/N-3, does not change on rotation about
the glycosidic bond in uridine (Fig. 2).
The Kiliani–Fischer synthesis is traditionally used to convert
an aldose into the next highest homologue, but the yields can be
Fig. 2 Through-space internuclear distances between isotopic labels in
extreme conformations of [2Ј-13C, 1,3-15N2]uridine 11.
Fig. 1 Conformations of adenosine 1.
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 0 5 7 – 2 0 6 2
2057

View Article Online
Scheme 1 i. K13CN (aq)/pH 7.2–7.5; ii. H2–5% Pd–BaSO4 (aq)/pH 4.2; iii. Dowex-Hϩ (aq); iv. Separation on a cation-exchange column in the
Sm3ϩ form; v. NiCl2ؒ6H2O–TEMED–MeOH/60 ЊC; vi. Separation on a cation-exchange column in the Sm3ϩ form; vii. MeOH–Dowex-Hϩ;
viii. Ac2O–Pyridine; ix. Ac2O–AcOH–H2SO4; x. Polyphosphoric acid/85 ЊC; xi. N,O-Bis(trimethylsilyl)acetamide–MeCN; xii. SnCl4–MeCN;
xiii. NaOMe–MeOH.
low.8 An improved version has as its key the stabilisation of the
intermediate aldonitrile in moderately acidic solution prior to
palladium-catalysed hydrogenation to an intermediate imine.8
This method, starting with -erythrose and K13CN, was
adapted to synthesise a mixture of -[1-13C]ribose 3 and its C-2
epimer -[1-13C]arabinose 4 (Scheme 1).
-[2-13C]ribose 5. Reaction of the -[1-13C]arabinose 4 with
molybdenyl acetylacetonate gave an unacceptable equilibrium
mixture of 4 and 5 in 2 : 1 ratio (as in ref. 10) and 71% recovery
after Sm3ϩ chromatography (i.e. 24% yield of -ribose). On the
other hand a nickel diamine complex (nickel() chloride plus
TEMED) in methanol gave a rapid and favourable rearrange-
ment and equilibration of the two epimers. After silica (par-
tially removes nickel) and Sm3ϩ chromatography, a ratio of
28 : 72 (as in ref. 11) in favour of -[2-13C]ribose 5 was obtained
(recovery after chromatography: 87%; -ribose epimerisation
and after chromatography: an acceptable 63% yield).
Separation of the two sugars using established published
procedures for aldoses that involve ion exchange columns in
Ca2ϩ or Ba2ϩ forms was problematic and indeed unacceptably
inefficient in terms of recovery for costly labelled compounds.
However, the effective analytical separation of -ribose and
-arabinose has been observed on cation-exchange TLC plates
in the La3ϩ or Sm3ϩ forms.9 We adapted this procedure to pre-
parative chromatography and immediately found that -ribose
and -arabinose were efficiently and completely separated on
an ion-exchange column in the Sm3ϩ form. High loading was
possible, with -arabinose eluting essentially in the void volume
and -ribose showing complex formation but still keeping to a
practical elution time and volume. The Sm3ϩ column can be
used repeatedly without recharging or cleaning and appears to
be indefinitely stable. We recommend this type of column for
wider application in the separation of isomeric carbohydrates
and polyols.
Overall, the route to -[1-13C]ribose 3 and also -[1-13C]-
arabinose 4 permits the economical and efficient preparation
of monosaccharides labelled with 13C at C-1, or alternatively at
C-2 by rearrangement and equilibration with nickel (or molyb-
denate). The advantage of our approach if used in a cyclical
way is high efficiency and thereby the consumption of an
unwanted epimer from the initial synthetic step. The rapid and
clean separation on the Sm3ϩ column is an essential component
of this approach.
[1,3-15N2]Uracil 8 was prepared from [15N2]urea 7 by adap-
tation of published methods12 and protected by silylation
(Scheme 1). The -[2-13C]ribose 5 was converted via its methyl
glycoside into the tetraacetate 6 (Scheme 1). The sugar and
basic moieties were fused together through SnCl4 catalysis13 to
give exclusively the β-anomer 10; deprotection gave [2Ј-13C, 1,3-
15N2]uridine 11 (Scheme 1) in 6.1% and 35.2% overall yield from
The epimeric products -[1-13C]ribose 3 and -[1-13C]arab-
inose 4 were obtained pure in a 0.8 : 1 ratio, respectively with an
acceptable overall yield of 49% from K13CN; solution NMR
analysis confirmed the expected exclusive labelling of C-1.
The proposed solid state NMR experiments require -[2-13C]-
ribose 5 and not -[1-13C]ribose 3. We achieved labelling of C-2,
however, as planned by a rearrangement involving a switch of
labelling from C-1 to C-2: there are a number of reactions of
aldoses that involve transition metal complexes, which result in
C-2 epimerisation and those involving molybdenate10 and
nickel()11 have, importantly, been shown to proceed with a
stereospecific 1,2-rearrangement in the carbon skeleton of the
aldose. This we exploited, in a novel labelling strategy, using the
C-1,2-rearrangement of -[1-13C]arabinose 4 to give the desired
1
K13CN and [15N2]urea, respectively. Solution H, 13C and 15N
NMR confirmed the labelling pattern in 11.
A distance measurement on crystals of uridine by using CP-MAS
solid state NMR with the application of REDOR
In a simple CP-MAS NMR experiment, dipolar (through
space) coupling information between nuclei, e.g. heteronuclei
13C and 15N, is inevitably lost. This is a significant loss since it is
from these couplings that crucial internuclear distances within
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 0 5 7 – 2 0 6 2
2058

View Article Online
a substrate–protein complex may be measured. In the case of
[2Ј-13C, 1,3-15N2]uridine 11 the distance between the labels on
C-2Ј and N-3 define the conformational relationship in the
nucleoside between the sugar moiety and the base (Fig. 2).
Ultimately, the goal of this work is to solve nucleoside struc-
tures within a protein. Initially, however, it was important to
establish the accuracy of the measurements of inter-atomic
distances in the nucleoside. Rotational echo double resonance
(REDOR) is an effective technique whereby dipolar coupling
information between heteronuclei may be restored and accurate
distances measured. This involves two experiments. First, the
intensity (S) is measured from a “full-echo” spectrum of one
nucleus (e.g. 13C) in the absence of dipolar coupling to the other
nucleus or nuclei (e.g. 15N). The experiment is then repeated
in the presence of dipolar coupling, and the intensity (S0) is
measured from the resulting “dephased-echo” spectrum. In
practice, a series of S and S0 values are obtained after allowing
the dipolar interaction to evolve for different time intervals.
Numerical simulations of the dephasing profile of S/S0 values
at the different dephasing intervals yield the internuclear
distance or distances.14,15
Fig. 4 The dephasing profile for a 13C-observed, 15N-dephased
REDOR experiment to measure the intramolecular distance between
C-2Ј and N-1 in crystals of [2Ј-13C, 1,3-15N2]uridine 11. REDOR solid
state NMR spectra were obtained for crystals of 11 as shown in Fig. 3
using a range of echo times of up to 7 ms. The intensity of the signal for
C-2Ј in each dephased-echo spectrum (S) was divided by that in the
full-echo spectrum (S0) to give the dephasing profile that was produced
by recovering the dipolar coupling between C-2Ј and N-1 in 11 (᭹). The
computer-simulated dephasing profile for an internuclear distance of
2.5 Å between a 13C–15N spin pair is also shown (––).
Crystals of [2Ј-13C, 1,3-15N2]uridine 11 were analysed by
both 13C and 15N CP-MAS NMR to reveal two overlapping
resonances for each of the labelled positions (Fig. 3). This is
consistent with the known crystal structure of uridine where
there are two molecules in the asymmetric unit.7
A
13C-
observed, 15N-dephased REDOR experiment was performed to
measure the distance between C-2Ј in the ribose unit and N-1 in
the base moiety of 11. The distance measured was 2.5 Å (Fig. 4,
cf. Fig. 2). This matches the 2.4 Å that was calculated16 from
the X-ray crystal structure coordinates for uridine.7
A
15N-
observed, 13C-dephased REDOR experiment on the same
material assisted the assignment of the two distinctly separate
doublet peaks in the 15N spectrum to the two 15N labels (Fig. 5).
Dephasing of the doublet at 149 ppm was greater than dephas-
ing of the doublet at 158 ppm. Hence, the upfield doublet is
ascribed to N-1, which is only 2.4 Å from C-2Ј, and, crucially,
the downfield doublet to N-3, which is further away from
C-2Ј. This latter is expected to provide the inter-atomic distance
from which the uridine conformation in the protein can be
determined.
These results establish, in our hands, the REDOR method
and, supported by a good supply of labelled uridine 11, give
us the necessary confidence in further NMR experiments to
determine the conformation of bound substrates in the
Fig. 5 Proton decoupled 15N CP-MAS NMR spectra obtained with
crystals of [2Ј-13C, 1,3-15N2]uridine 11 in a 15N-observed, 13C-dephased
REDOR experiment. The full-echo spectrum (A) and dephased-echo
spectrum (B) were both obtained over 2048 acquisitions for an echo
time of 2.2 ms. All other conditions were as stated in Fig. 3.
bacterial nucleoside transporters NupC and NupG and thus to
provide important information on the substrate binding sites of
these two membrane transport proteins. Further importance is
attached to these bacterial transporters in that NupC shares
27% amino acid sequence identity to, and the same natural
substrate specificity with, the human nucleoside transporter
hCNT1.17 Information gained in the accessible bacterial system
is of relevance to the human case, which is a much more
difficult one to study.
Fig. 3 Proton decoupled 13C CP-MAS NMR spectra obtained in a
13C-observed, 15N-dephased REDOR experiment to measure the
intramolecular distance between C-2Ј and N-1 in crystals of [2Ј-13C,
1,3-15N2]uridine 11. The full-echo spectrum (A) was obtained for ca. 4
mg crystals of 11 in a normal cross polarisation (CP) experiment over
1
2951 acquisitions at 100.62 MHz for 13C (400.14 MHz for H) with a
MAS rate of 4228 Hz and using a proton field strength of 65 kHz for
CP and for decoupling and a CP contact time of 2 ms. An expansion of
the signal for C-2Ј is shown in B. An expansion of the dephased-echo
spectrum is shown in C; this was obtained as for A/B but with
rotationally synchronous dephasing π-pulses applied at the resonance
frequency of 15N (at 40.54 MHz) for an echo time of 4.5 ms.
Experimental
General
Starting materials, reagents and reaction solvents were pur-
chased from Aldrich (Gillingham, UK) or Lancaster synthesis
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 0 5 7 – 2 0 6 2
2059

View Article Online
Ltd (Morecambe, UK). Ion-exchange resin was purchased
from Aldrich. 13C- and 15N-labelled starting materials and
deuterated DMSO were purchased from Cambridge Isotopes,
Inc. (Andover, MA, USA). Deuterated chloroform was pur-
chased from Apollo Scientific Ltd (Whaley Bridge, Derbyshire,
UK) and deuterium oxide was purchased from Fluorochem
Ltd (Old Glossop, Derbyshire, UK). General solvents were
purchased from Fisher Scientific (Loughborough, UK), and
solvents were dried using standard procedures. Water-sensitive
materials were stored under vacuum over P2O5. Column chrom-
atography was performed using Flash 60 (230–400 mesh)
silica gel and normal phase TLC was performed on Merck silica
gel 60 F254 coated aluminium sheets. Compounds were visual-
ised on TLC plates under UV light (254 nm) or by treating with
a 5% solution of sulfuric acid in ethanol followed by charring in
an oven. Organic phases were dried using anhydrous Na2SO4
and freeze-drying was performed using a Chemlab apparatus.
Solution state NMR spectra were recorded using a Bruker
Avance 250 MHz instrument at the field strength indicated.
Chemical shifts were reported as δ values downfield from the
internal standard tetramethylsilane, which had a δ value of
zero, and coupling constants (J) were reported in Hz. Electron
ionisation (EI) mass spectra were recorded using a VT
Autospec instrument and electrospray (ES) mass spectra were
recorded using a Micromass quadrupole time-of-flight instru-
ment. Solid state NMR spectra were recorded using a Bruker
Avance 400 MHz instrument with a Bruker three channel probe
using the conditions as indicated. Samples were contained
within a 4 mm zirconia rotor.
Fifty fractions were collected (1–10, ca. 3 cm3 each and 11–50,
8–10 cm3 each), and these were analysed by TLC (90 : 10
acetone–water) to reveal nothing in fractions 1–3, unreacted
imine in fraction 4, exclusively -arabinose in fractions 5–10,
-ribose and a trace of -arabinose in fraction 11 and
exclusively -ribose from fraction 12 onwards. Nothing was
detected in the effluent collected after fraction 50. A similar
separation was achieved with the second half of the product
mixture. The fractions that were shown to contain exclusively
-ribose or -arabinose were independently combined. The
water was removed initially by evaporation under vacuum and
then by freeze-drying and finally under vacuum over P2O5 to
leave both 3 (280 mg, 22%) and 4 (336 mg, 27%) as colourless
oils with partial crystallisation; in the following, f = furanose
and p = pyranose forms; -[1-13C]ribose 3: δC(62.9 MHz; D2O)
101.2 (C-1βf, 11.4%)*, 96.5 (C-1αf, 5.8%)*, 94.1 (C-1βp,
62.4%)*, 93.8 (C-1αp, 20.4%)*, 88.2 (C-4αf ), 82.7 (C-4βf ), 75.5
(C-2βf ), 71.3 (C-2βp), 70.7 (C-3βf ), 70.3 (C-2αp and C-3αf ),
69.5 (C-3αp), 69.3 (C-3βp), 67.6 (C-4αp), 67.4 (C-4βp), 63.2
(C-5αp and C-5βp) and 62.8 (C-5αf and C-5βf ); -[1-13C]-
arabinose 4: δC(62.9 MHz; D2O) 101.3 (C-1αf, 3.6%)*, 97.1
(C-1αp, 61.3%)*, 95.4 (C-1βf, 1.4%)*, 92.9 (C-1βp, 32.9%)*,
72.7 (C-4αp), 72.1 (C-2αp), 70.0 (C-4βp), 68.9 (C-2βp), 68.8
(C-3αp), 68.7 (C-3βp), 66.7 (C-5αp) and 62.7 (C-5βp).
Preparation of a samarium (Sm3؉) cation-exchange column and
separation of D-ribose and D-arabinose
Dowex 8 × 50W, 200–400, Hϩ cation-exchange resin (20 g)
was rinsed three times with distilled water (120 cm3, ca. 0.6 vol).
The resin was stirred for ca. 20 minutes at room temperature
with an aqueous solution of samarium() chloride hexahydrate
(12 cm3, 25%, ca. 0.6 vol) and then poured into a chromato-
graphic column (42 cm × 2.2 cm). The resultant resin bed (6.5 ×
2.0 cm) was washed under gravity elution with an aqueous
solution of samarium() chloride hexahydrate (28 cm3, 25%,
ca. 1.4 vol) and then distilled water (800 cm3). A test mixture of
unlabelled -ribose (200 mg) and -arabinose (200 mg) was
separated on the column as described above leading to the
recovery of -arabinose (200 mg, 100%) and -ribose (197 mg,
98.5%).
In the following 13C NMR data, a resonance with 13C
enrichment is indicated by *.
Synthesis of D-[1-13C]ribose 3 and D-[1-13C]arabinose 4
(adapted with essential modifications from ref. 8)
With rapid stirring at room temperature, the pH of an aqueous
solution (50 cm3) of potassium [13C]cyanide (0.55 g, 8.33 mmol)
was lowered to 7.2 by dropwise addition of aqueous acetic acid
(3 M). An aqueous solution (30 cm3) of -erythrose 2 (1.00 g,
8.33 mmol) was slowly added, with addition of acetic acid
(3 M) or sodium hydroxide (1 M) as appropriate to maintain a
pH of 7.2–7.5 during addition of -erythrose and subsequent
reaction, which was allowed to proceed for 1.5 h. The reaction
mixture was analysed by TLC (90 : 10 acetone–water) to reveal
a complete conversion of -erythrose (Rf 0.67) into the aldoni-
triles (Rf 0.81). The pH of the aldonitrile solution was lowered
to 4.7 by dropwise addition of acetic acid (3 M) and the solu-
tion was stored at 4 ЊC awaiting reduction (ca. 1 h later). The
pH of the aldonitrile solution was lowered to 4.2 by dropwise
addition of acetic acid (3 M). Following addition of pre-
reduced 5% palladium on barium sulfate (516 mg = 25.8 mg,
0.24 mmol Pd), the flask was evacuated three times and flushed
with hydrogen. Vigorous stirring under an atmosphere of
hydrogen was then allowed to proceed at room temperature for
ca. 15 h. The reaction mixture was analysed by TLC (90 : 10
acetone–water) to reveal a complete conversion of the aldo-
nitriles (Rf 0.80) into the imines (Rf 0.46). The catalyst was
removed over Celite and the pH of the resultant filtrate was
lowered from 4.4 to ca. 3.0 by repeated batch-wise treatment
with Dowex 8 × 50W, 200–400, Hϩ cation-exchange resin
(4 × 10 g). The final filtrate was analysed by TLC (90 : 10 acet-
one–water) alongside the authentic compounds to reveal
-ribose (Rf 0.50) and -arabinose (Rf 0.38) with no imine
detected. The water was removed by freeze-drying to leave the
epimeric mixture of 3 and 4 as a pale yellow oil (1.07 g) which
was stored at 4 ЊC. The epimeric sugars were separated in two
portions on a cation-exchange column prepared in the samar-
ium (Sm3ϩ) form (see below). The oil was dissolved in distilled
water (ca. 2 cm3) and half of the resultant syrup was applied to
the column, which was eluted under gravity with distilled water.
Molybdate catalysed epimerisation of D-arabinose
(adapted from ref. 10)
A solution of -arabinose (250 mg, 1.67 mmol) and molyb-
denyl acetylacetonate (27 mg, 0.083 mmol) in DMF (16.8 cm3)
was stirred at 50 ЊC for 24 h. TLC analysis (90 : 10 acetone–
water) alongside the authentic compounds revealed both
-arabinose (Rf 0.49) and -ribose (Rf 0.41) with the former
epimer predominating. The reaction mixture was cooled
(ice-bath), diluted with water (40 cm3) and then extracted with
dichloromethane (2 × 80 cm3). The aqueous layer was stirred
for 1 h at room temperature with an excess of Dowex 8 × 50W,
200–400, Hϩ cation-exchange resin (4 g). The resin was then
removed by filtration and washed with water. The combined
filtrate and washings were stirred for 1 h at room temperature
with an excess of Dowex 1 × 8–50, 20–50 anion-exchange resin
Ϫ
(4 g) prepared in the HCO3 form. The resin was removed by
filtration, washed with water and the combined filtrate and
washings were evaporated to leave a pale yellow oil (197 mg).
TLC analysis alongside the authentic compounds revealed that
the proportion of product epimers was unchanged. A portion
of the oil (60 mg) was passed down a cation-exchange column
prepared in the samarium form (see above) to give a clean
separation of the two epimers. The water was removed from
the combined -arabinose and -ribose-containing fractions by
evaporation and then by freeze-drying to leave -arabinose
(36 mg, 47.3%) and -ribose (18 mg, 23.6%) as a colourless oil
with partial crystallisation; δC(62.9 MHz; D2O) as for authentic
-ribose.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 0 5 7 – 2 0 6 2
2060

View Article Online
Synthesis of D-[2-13C]ribose 5 (adapted from ref. 11)
5.3, NH-1), 7.40 (1H, m, H-6) and 5.46 (1H, m, H-5);
(unlabelled: 11.06 (1H, br s, NH-3), 10.86 (1H, br s, NH-1),
7.41 (1H, dd, J5,6 7.6, J6,N-1 5.6, H-6) and 5.46 (1H, dd,
J5,6 7.6, J5,N-1/3 1.6, H-5)); δC(62.9 MHz; DMSO-d6) 164.7 (C-4,
d, J4,N-3 11.4), 151.9 (C-2, dd, J2,N-1 18.6, J2,N-3 16.3), 142.6 (C-6,
d, J6,N-1 11.0) and 100.6 (C-5, d, J5,N-3/1 6.2); m/z (EI) 114 (100%,
Mϩ).
A solution of -[1-13C]arabinose 4 (300 mg, 2.00 mmol) in
methanol (25 cm3) was stirred at 60 ЊC. Nickel() chloride
hexahydrate (474 mg, 2.00 mmol) and TEMED (0.6 cm3, 4.00
mmol) were sequentially added, to form pale green and dark
green solutions, respectively. TLC analysis (90 : 10 acetone–
water) after 5 min alongside the authentic compounds revealed
-arabinose (Rf 0.37) and -ribose (Rf 0.47), with the latter
epimer predominating, and a faint area of green material at the
baseline. After allowing it to cool to room temperature the reac-
tion mixture was diluted with distilled water (90 cm3) and then
maintained at a pH of 6.5 for 1 h, by addition of dilute sulfuric
acid (0.5 M) as appropriate, whilst stirring at room temperature.
The solution was then stirred for 30 minutes at room temper-
ature with an excess of Dowex 8 × 50W, 200–400, Hϩ cation-
exchange resin (2.5 g). The resin was removed by filtration and
washed with water. The combined filtrate and washings were
stirred for 1½ h at room temperature with an excess of Dowex
1 × 8–50, 20–50 anion-exchange resin (2.5 g) prepared in the
HCO3Ϫ form. The resin was removed by filtration, washed with
water, and the combined filtrate and washings were evaporated
to leave a pale green oil. Most of the remaining nickel-contain-
ing material (green) was removed on a silica column (100%
acetone) to give the epimeric mixture as a cloudy oil. The sugars
were separated in three portions on a cation-exchange column
prepared in the samarium form (see above), which also removes
any remaining nickel, followed by removal of the water by
evaporation and then by freeze drying to give 4 (73 mg, 24%)
and 5 (189 mg, 63%) as a colourless oil with partial crystallis-
ation; δC(ribose) (250 MHz; D2O) 75.5 (C-2βf, 11.8%)*, 71.3
(C-2βp, 64.4%)*, 71.2 (C-2αf, 6.0%)* and 70.3 (C-2αp, 17.8%)*;
m/z (EI) 152 (Mϩ ϩ H), 120 (5%, Mϩ Ϫ CH2OH), 74 (90), 61.0
(100); m/z (ES) 174.0418 (Mϩ ϩ Na, required for C413CH10-
O5Na 174.0426).
Synthesis of 2Ј,3Ј,5Ј-tri-O-acetyl-[2Ј-13C, 1,3-15N2]uridine 10
(adapted from ref. 13)
A suspension of [1,3-15N2]uracil 8 (196 mg, 1.72 mmol) in
anhydrous acetonitrile (3.7 cm3) was stirred at room temper-
ature with N,O-bis(trimethylsilyl)acetamide (937 µl, 769 mg,
3.78 mmol). After 0.5 h the uracil had not completely dissolved,
so further N,O-bis(trimethylsilyl)acetamide (500 µl) was added.
After ca. 1 h the uracil had completely dissolved to give a pale
yellow solution. A solution of 1,2,3,5-tetra-O-acetyl--[2-13C]-
ribofuranoside 6 (306 mg, 0.96 mmol) in anhydrous acetonitrile
(9.1 cm3) was added to the solution of silylated uracil. The
combined solution was cooled (ice-bath) and tin() chloride
(244 µl, 537 mg, 2.06 mmol) was added dropwise with stirring,
which was continued at room temperature. After 18.5 h, further
tin() chloride (300 µl) was added. After a total reaction time
of 22.5 h the reaction mixture was cooled (ice-bath) and a solu-
tion of sodium hydrogen carbonate (1.0 g) in water (3.4 cm3)
was added dropwise with continued stirring. The resultant
white precipitate was removed by filtration and washed with
acetonitrile. The combined filtrates were evaporated to leave a
pale yellow oil. The protected nucleoside was separated from
unreacted starting materials by gradient elution on a silica
column (99 : 1–90 : 10 chloroform–methanol) and then further
purified on a second silica column (99 : 1 chloroform–
methanol) to give 10 (222 mg, 62.0%) as a pale yellow oil;
δH(250 MHz; DMSO-d6) 9.55 (1H, d, JNH-3,N-3 91.1, NH-3), 7.44
(1H, dd, J5,6 8.2, J5,NH-1/3 2.0, H-5), 6.06 (1H, t, J 4.3, H-1Ј), 5.82
(1H, m, J 2.2, 2.3, 2.6, 2.7, H-6), 5.35 (1H, dt, J1Ј,2Ј = J2Ј,3Ј 5.7,
J2Ј,C-2Ј 156.5, H-2Ј), 5.35 (1H, m, H-3Ј), 4.36 (3H, m, over-
lapping H-4Ј, H-5Јa and H-5Јb) and 2.16, 2.14, 2.11 (9H, 3 × s,
3 × CH3); (unlabelled: 9.55 (1H, br s, NH-3), 7.42 (1H, d, J5,6
8.2, H-5), 6.06 (1H, d, J1Ј,2Ј 4.9, H-1Ј), 5.82 (1H, d, J5,6 8.1, H-6),
5.36 (2H, m, overlapping H-2Ј and H-3Ј), 4.36 (3H, m, over-
lapping H-4Ј, H-5Јa and H-5Јb) and 2.16, 2.14, 2.11 (9H, 3 × s,
Synthesis of 1,2,3,5-tetra-O-acetyl-D-[2–13C]ribofuranoside 6
This compound was prepared from 5 over three steps via
1-O-methyl--[2-13C]ribofuranoside and 1-O-methyl-2,3,5-tri-
O-acetyl--[2-13C]ribofuranoside using the Guthrie–Smith18
method to give 6 (306 mg, 59.5%) as a pale yellow oil; δH(250
MHz; CDCl3) 6.43 (dd, J1α,2α 3.8, J1α,C-2α 2.6, H-1α), 6.17 (br dd,
J1β,2β < 1, J1β,C-2β ca. 2, H-1β), 5.36 (ddd, J2β,3β = J3β,4β 4.7, J3β,C-2β
1.9, H-3β), 5.34 (ddd, J1β,2β < 1, J2β,3β 4.7, J2β,C-2β 162.7, H-2β),
5.28 (ddd, J2α,3α 6.8, J3α,4α 3.0, J3α,C-2α 1.6, H-3α) and 5.24 (ddd,
J1α,2α 3.8, J2α,3α 6.8, J2α,C-2α152.3, H-2α), 4.45 (dd, J4β,5βa 3.5, J 6.1,
H-4β), 4.34 (dd, J4β,5βa 3.5, J5βa,5βb 11.6, H-5βa) (H-4α over-
lapping), 4.15 (dd, J4β,5βb 5.0, J5βa,5βb 11.6, H-5βb) (H-5α a and b
overlapping), 2.14, 2.13, 2.12, 2.09 and 2.14, 2.11, 2.10, 2.08
(12H, 8 × s, 4 × CH3, α and β, respectively); δC(62.9 MHz;
3 × CH ); δ (62.9 MHz; DMSO-d ) 170.2, 169.7 (× 2) (3 × C᎐
᎐
3
C
6
O), 162.9 (C-4, d, J4,N-3 9.5), 150.3 (C-2, t, J2,N-1/3 18.7), 139.3
(C-6, d, J6,N-1 12.6), 103.5 (C-5, d, J5,N-3/1 6.9), 87.4 (C-1Ј, td,
J1Ј,2Ј 21.9, J1Ј,N-1 13.6), 79.5 (C-4Ј, d, J2Ј,4Ј 1.1), 72.7 (C-2Ј)*, 70.7
(C-3Ј, d, J2Ј,3Ј 19.3), 63.2 (C-5Ј) and 20.8, 20.5, 20.4 (3 × CH3);
Rf (90 : 10 chloroform–methanol) 0.57.
Synthesis of [2Ј-13C, 1,3-15N2]uridine 11
CDCl ) 170.9, 170.1, 169.8 (d, J < 5), 169.4 (d, J < 5) (4 × C᎐O,
᎐
3
Sodium methoxide anhydrous powder (13 mg, 0.24 mmol) was
added to a stirred solution of 2Ј,3Ј,5Ј-tri-O-acetyl-[2Ј-13C, 1,3-
15N2]uridine 10 (150 mg, 0.40 mmol) in anhydrous methanol
(5.3 cm3); stirring was continued at room temperature for
ca. 30 h. The reaction mixture was neutralised by stirring at
room temperature with Amberlite IR-120 (strongly acidic)
cation-exchange resin. The resin was removed by filtration,
washed with anhydrous methanol and the combined filtrates
were evaporated to leave the deprotected nucleoside as a pale
yellow oil (97 mg, 98%). Crystallisation was achieved by cooling
the oil in an ice bath and adding a little ice-cold methanol.
Scratching with a glass rod induced the formation of pale
crystals; these were collected by filtration to give 11 (1st crop
16.3 mg, 2nd crop 14.3 mg, total 31.0%) as pale cream plates;
δH(250 MHz; DMSO-d6) 11.30 (1H, br undefined d, NH-3), 7.87
(1H, dd, J5,6 8.1, J5,NH-1/3 2.1, H-5), 5.76 (1H, m, H-1Ј), 5.63 (1H,
ddd, J5,6 8.1, J6,N-1 4.7, J6,N-3 2.4, H-6), 5.38 (1H, unresolved
m, OH-2Ј), 5.10 (2H, m, overlapping OH-3Ј and OH-5Ј),
4.00 (1H, dm, J2Ј,C-2Ј 147.8, H-2Ј), 3.94 (1H, m, H-3Ј), 3.82 (1H,
β), 98.5 (C-1β, t, J1β,2β 23.7), 94.4 (C-1α, t, J1α,2α 22.8), 82.0
(C-4α), 79.6 (C-4β), 74.5 (C-2β)*, 70.6 (C-3β, d, J2β,3β 20.4), 70.4
(C-2α)*, 70.1 (C-3α), 64.0 (C-5β, d, J2β,5β < 2), 63.7 (C-5α) and
21.4, 21.1, 20.9, 20.8 (4 × CH3, β) (4 × CH3, α, overlapping);
m/z (ES) 342 (100%, Mϩ ϩ Na), 661 (90, 2Mϩ ϩ Na).
Synthesis of [1,3-15N2]uracil 812
A mixture of [15N2]urea 7 (500 mg, 8.06 mmol), propiolic acid
(496 µl, 564 mg, 8.06 mmol) and polyphosphoric acid (12.0 g,
5.8 cm3) was heated at 85 ЊC (air condenser) with gentle stirring
for 4 h to give a yellow syrup. The mixture was cooled (ice-bath)
and water (24 cm3) was slowly added. The flask was swirled
to achieve complete dissolution of the syrup, on which a pale
solid began to form. The flask was kept at 4 ЊC for ca. 24 h, then
the solid was collected by filtration, washed with a little water
and dried under vacuum over P2O5 to give 8 (532 mg, 57.9%) as
a cream coloured solid; δH(250 MHz; DMSO-d6) 11.06 (1H,
d, JNH-3,N-3 89.8, NH-3), 10.90 (1H, dd, JNH-1,N-1 92.5, JNH-1,N-3/6
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 0 5 7 – 2 0 6 2
2061

View Article Online
4 S. G. Patching, PhD thesis, Chemical and Solid State NMR
Approaches to Determine Structure-Activity Relationships for
Substrates Bound to the Nucleoside Transport Proteins, NupC and
NupG, of Escherichia coli, University of Leeds, 2002.
5 P. J. R. Spooner, N. G. Rutherford, A. Watts and P. J. F. Henderson,
Proc. Natl. Acad. Sci. U. S. A., 1994, 91, 3877.
6 (a) P. J. R. Spooner, W. J. O’ Reilly, S. W. Homans, N. G. Rutherford,
P. J. F. Henderson and A. Watts, Biophys. J., 1998, 75, 2794; (b) A. N.
Appleyard, R. B. Herbert, P. J. F. Henderson, A. Watts and P. J. R.
Spooner, Biochim. Biophys. Acta, 2000, 1509, 55; (c) H. Venter,
A. E. Ashcroft, J. N. Keen, P. J. F. Henderson and R. B. Herbert,
Biochem. J., 2002, 363, 243; (d ) J. A. Watts, A. Watts and D. A.
Middleton, J. Biol. Chem., 2001, 276, 43197; (e) D. A. Middleton,
Z. Ahmed, C. Glaubitz and A. Watts, J. Magn. Reson., 2000, 147,
366; ( f ) D. A. Middleton, S Rankin, M. Esmann and A. Watts,
Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 13602.
m, H-4Ј) and 3.56 (2H, m, H-5Јa and H-5Јb); (unlabelled: 11.33
(1H, br s, NH-3), 7.92 (1H, d, J5,6 8.1, H-5), 5.78 (1H, d,
J1Ј,2Ј 5.3, H-1Ј), 5.66 (1H, d, J5,6 8.1, H-6), 5.42 (1H, d, J2Ј,2Ј-OH
5.6, OH-2Ј), 5.14 (2H, m, overlapping OH-3Ј and OH-5Ј), 4.03
(1H, q, J1Ј,2Ј = J2Ј,3Ј 5.3, H-2Ј), 3.98 (1H, q, J 4.6 and 4.3, H-3Ј),
3.85 (1H, q, J 3.1 and 3.5, H-4Ј) and 3.60 (2H, m, H-5Јa and
H-5Јb); δC(62.9 MHz; DMSO-d6) 163.5 (C-4), 151.1 (C-2),
141.1 (C-6), 102.1 (C-5), 88.0 (C-1Ј), 85.2 (C-4Ј), 73.9 (C-2Ј)*,
70.2 (C-3Ј) and 61.2 (C-5Ј); m/z (EI) 247 (19%, Mϩ), 229
(39, Mϩ Ϫ OH ϩ H), 134 (5, Mϩ Ϫ base ϩ H), 114 (100,
Mϩ Ϫ sugar ϩ H); m/z (ES) 270.0568 (Mϩ ϩ Na, required
for C813CH1215N2O6Na 270.0567); Rf (90 : 10 chloroform–
methanol) 0.05.
7 E. A. Green, R. D. Rosenstein, R. Shiono, D. J. Abraham, B. L. Trus
and R. E. Marsh, Acta Crystallogr., Sect. B: Struct. Crystallogr.
Cryst. Chem., 1975, 31, 102.
Acknowledgements
8 A. S. Serianni, H. A. Nunez and R. Barker, Carbohydr. Res., 1979,
72, 71.
9 Y. Israeli, C. Lhermet, J. Morel and N. Morel-Desrosiers, Carbohydr.
Res., 1996, 289, 1.
10 M. L. Hayes, N. J. Pennings, A. S. Serianni and R. Barker, J. Am.
Chem. Soc., 1982, 104, 6764.
11 T. Tanase, F. Shimizu, M. Kuse, S. Yano, M. Hiddai and
S. Yoshikawa, Inorg. Chem., 1988, 27, 4085.
12 K. L. Anderson-Altmann, C. G. Phung, S. Marromoustakos,
Z. W. Zheng, J. C. Facelli, C. D. Poulter and D. M. Grant, J. Phys.
Chem., 1995, 99, 10454.
13 (a) K. L. Anderson-Altmann, C. G. Phung, S. Marromoustakos,
Z. Zhang, J. C. Facelli, C. D. H. Vorbrüggen and G. Höfle,
Chem. Ber., 1981, 114, 1256; (b) M. Saneyoshi and E. Satoh,
Chem. Pharm. Bull., 1979, 27, 2518; (c) U. Niedballa and
H. Vorbrüggen, J. Org. Chem., 1976, 41, 2084.
We are grateful to the Engineering and Physical Sciences
Research Council (EPSRC) for a studentship (S. G. P.) and to
the Wellcome Trust, the Biotechnology and Biological Sciences
Research Council (BBSRC), SmithKline Beecham Pharm-
aceuticals and Glaxo Wellcome (now GlaxoSmithKline), the
Royal Society and the University of Leeds for financial support
of this research. This work was carried out as part of the
BBSRC-funded North of England Structural Biology Centre.
We thank Dr A. E. Ashcroft for mass spectra, Dr K. Seshadri
for assistance in the calculation of distances from crystal struc-
ture coordinates and Professor S. A. Baldwin for discussions on
nucleoside transporters.
14 T. Gullion and J. Schaefer, J. Magn. Reson., 1989, 81, 196.
15 J. M. Goetz and J. Schaefer, J. Magn. Reson., 1997, 127, 147.
16 The fractional cell coordinates were transformed into orthogonal
cartesian coordinates by means of a suitable transformation matrix
using the cell parameters a, b, c and α, β, γ. A Fortran code was used
to accomplish this.
17 M. W. L. Ritzel, S. Y. M. Yao, M-Y. Huang, J. F. Elliot, C. E. Cass
and J. D. Young, Am. J. Physiol., 1997, 272, C707.
18 R. D. Guthrie and S. C. Smith, Chem. Ind. (London), 1968, 547.
References
1 P. J. F. Henderson, Function and Structure of Membrane Transport
Proteins in The Transporter Factsbook, ed. J. K. Griffith and
C. E. Sansom, Academic Press, London, 1998.
2 J. E. Craig, Y. Zhang and M. P. Gallagher, Mol. Microbiol., 1994, 11,
1159.
3 S. E. Westh-Hansen, N. Jensen and A. Munch-Peterson, Eur. J.
Biochem., 1987, 168, 385.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 0 5 7 – 2 0 6 2
2062