Experimental evidence for intramolecular attractive nonbonded C-F ··· H-C interactions in 2',3'-dideoxy-4'-(fluoromethyl)nucleosides - Through- space J(CF) and J(HF) NMR coupling constants, correlation with empirical parameters of solvent polarity and single-crystal X-ray structures

Article abstract of DOI:10.1002/(sici)1099-0690(199901)1999:1<187::aid-ejoc187>3.0.co;2-r

A collection of 5'-O-benzyl-2',3'-dideoxy-4'-(fluoromethyl)nucleosides carrying both purinic and pyrimidinic nucleobases (uracil, 5-Br-uracil, 5- O₂N-uracil, 6-Cl-purine and inosine) were synthesized in both the α and the β form. Through-space-transmitted ⁶J(CF) NMR coupling constants between F and C-6 (pyrimidinic base) or C-8 (purinic base) were observed for all of the α anomers of the compounds examined, whilst the corresponding ⁷J(HF) coupling constants were resolved only for the 5-substituted uracil derivatives. The absolute values of all the through-space couplings were found to decrease monotonically with increasing solvent polarity (CDCl₃, MeOD, [D₆]acetone, [D₆]DMSO). This trend suggests that the through-space interaction is mediated by an intramolecular (sp³)C-F ··· H-C(sp²) hydrogen bond. The possibility of any relevant solvent-induced conformational change influencing the F/base mutual spatial relationship in the molecules investigated was ruled out by heteronuclear steady-state ¹H{¹⁹F}-NOE experiments. A linear correlation was observed between ⁶J(CF) and ⁷J(HF) coupling constants and the Kamlet-Taft's hydrogen bond basicity parameter β. The crystal structures of the α and β anomers of the 5-nitrouracil nucleoside show evidence that the H-6 of the nucleobase forms hydrogen-bond- like interactions involving the O-benzyl oxygen atom in the β anomer, and that in the case of the α anomer this is replaced by the F atom of the fluoromethyl group.

Full text of DOI:10.1002/(sici)1099-0690(199901)1999:1<187::aid-ejoc187>3.0.co;2-r

FULL PAPER
Experimental Evidence for Intramolecular Attractive Nonbonded C–F...H–C
Interactions in 2Ј,3Ј-Dideoxy-4Ј-(fluoromethyl)nucleosides – Through-Space
JCF and JHF NMR Coupling Constants, Correlation with Empirical
Parameters of Solvent Polarity and Single-Crystal X-ray Structures
Andrea Mele,*[a] Barbara Vergani,[a] Fiorenza Viani,[b] Stefano Valdo Meille,[a]
Alessandra Farina,[a] and Pierfrancesco Bravo[b]
Keywords: Fluorinated dideoxynucleosides / Through-space interactions / Hydrogen bonds / Long-range coupling
constants
A collection of 5Ј-O-benzyl-2Ј,3Ј-dideoxy-4Ј-(fluoromethyl)-
nucleosides carrying both purinic and pyrimidinic
nucleobases (uracil, 5-Br-uracil, 5-O2N-uracil, 6-Cl-purine
and inosine) were synthesized in both the α and the β form.
Through-space-transmitted JCF NMR coupling constants
between F and C-6 (pyrimidinic base) or C-8 (purinic base) was observed between JCF and JHF coupling constants and
an intramolecular (sp3)C–F...H–C(sp2) hydrogen bond. The
possibility of any relevant solvent-induced conformational
change influencing the F/base mutual spatial relationship in
the molecules investigated was ruled out by heteronuclear
6
19
steady-state 1H{ F}-NOE experiments. A linear correlation
6
7
were observed for all of the α anomers of the compounds
examined, whilst the corresponding JHF coupling constants
the Kamlet–Taft’s hydrogen bond basicity parameter β. The
crystal structures of the α and β anomers of the 5-nitrouracil
nucleoside show evidence that the H-6 of the nucleobase
forms hydrogen-bond-like interactions involving the O-
benzyl oxygen atom in the β anomer, and that in the case of
7
were resolved only for the 5-substituted uracil derivatives.
The absolute values of all the through-space couplings were
found to decrease monotonically with increasing solvent
polarity (CDCl3, MeOD, [D6]acetone, [D6]DMSO). This trend
suggests that the through-space interaction is mediated by
the
α anomer this is replaced by the F atom of the
fluoromethyl group.
Considerable efforts have been made in the last few years CϪF...HϪD (D ϭ donor ϭ O, N) contacts, consistent with
to develop new nucleoside analogues likely to exhibit im- inter- or intramolecular hydrogen bonds,[7] an extensive
proved activity, or decreased toxicity, with respect to 3Ј- Cambridge Structural Database search showed that the
azido-3Ј-deoxythimidine (AZT), the first anti-HIV drug.[1] statistical occurrence of CϪF...HϪD contacts with F...H dis-
In this context, we have developed a synthetic strategy for tances below the sum of their van der Waals radii is very
obtaining new fluoro-substituted nucleoside analogues by low with respect to the more classical DϪH...A interactions
exploiting the sulphoxide-mediated route to optically active (D ϭ donor, O or N, typically; A ϭ acceptor atom, e.g.
fluoro-substituted compounds.[2,3] The structure-activity carbonyl oxygen atom etc.).[8] In a recent NMR and X-ray
correlation for nucleoside analogues is difficult to establish study on the protonation of fluoro cryptands,[9] the authors
due to the number of alternative metabolic pathways avail- were unable to decide for or against intramolecular
able for activation and the number of different routes for CF...HNϩ hydrogen bond, concluding that “should a
biological activity;[4] thus, many studies have been devoted CF...HNϩ hydrogen bonding interaction exist, it is certainly
to establishing the conformational preferences of nucleo- going to be very weak in nature”. (ii) Conversely, other
sides, either active or inactive against HIV.[5] The role of scientists, mainly crystallographers, introduced a broader
fluorine substitution in bioorganic substrates is often dis- concept of hydrogen bond including, along with the tra-
cussed in terms of the ability of fluorine to act as a hydro- ditional “strong” O...HϪO and O...HϪN interactions,
gen or hydroxy mimic.[6] Furthermore, the ability of car- “weak” and unusual interactions, like O...HϪC.[10] Al-
bon-bound fluorine (CϪF) to establish hydrogen bonds though the large difference in energy range between the two
with HϪD groups (D ϭ donor) is often invoked, but still sets of nonbonded interactions (2Ϫ20 kJ molϪ1 for the
not unequivocally accepted, as factor playing a role in the “weak” hydrogen bonds compared to the 20Ϫ40 kJ molϪ1
binding of fluoro compounds to receptors. The terms of generally accepted for the “strong” ones), it was demon-
this controversial debate can be summarized as follows: (i) strated that O...HϪC interactions possess the directional
although some crystallographic studies revealed short features distinguishing hydrogen bonds and that, even if
characterized by contacts longer than the sum of O and C
[a]
van der Waals radii, they play a role in driving crystal pack-
ing, especially when stronger O...HϪO or O...HϪN interac-
tions are not available. More importantly, the relevance of
these interactions is enhanced by increasing the CϪH acid-
ity and by the presence of cooperative effects.[11Ϫ13]
Dipartimento di Chimica del Politecnico and
CNR-Centro di Studio sulle Sostanze Organiche Naturali,
Via Mancinelli, 7, I-20131 Milano, Italy
Fax: (internat.) ϩ 39/2-23993080
[b]
E-mail: mele@dept.chem.polimi.it
Eur. J. Org. Chem. 1999, 187Ϫ196
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1434Ϫ193X/99/0101Ϫ0187 $ 17.50ϩ.50/0
187

A. Mele, B. Vergani, F. Viani, S. V. Meille, A. Farina, P. Bravo
FULL PAPER
In a previous solution NMR study on a collection of methane or 1,2-dichloroethane, to the silylated bases and
2Ј,3Ј-dideoxy-4Ј-(fluoroalkyl)nucleosides[14] (fluoroalkyl ϭ subsequently pouring trimethylsilyl triflate into the re-
RF ϭ CH2F, CHF2 and CF3; bases: thymine and 5-F-uracil; sulting slurry as a catalyst (Schemes 3 and 4).
see Scheme 1 for atom numbering of pyrimidinic bases),
we showed the existence of through-space scalar coupling
between the F of the remote RF moiety and the H-6/C-6
of the nucleobase in the α anomer of all the compounds
examined, the only exception being the CF3 derivatives
(Scheme 2).
Scheme 3. Reagents and conditions: (i) 2,4-bis(trimethylsilyloxy)-
uracil (R ϭ H), 5-bromo-2,4-bis(trimethylsilyloxy)uracil (R ϭ Br),
5-nitro-2,4-bis(trimethylsilyloxy)uracil (R
TMSOTf; room temp.
ϭ NO2); (CH2Cl)2;
The inosyl derivative 5 was synthesized by dissolving the
6-chloropurin-9-yl intermediate 4 in methanol and adding
dropwise an aqueous solution of sodium hydroxide. The
mixture was refluxed for 2 h; the final compound 5 was
obtained in a yield of 70% (Scheme 4).
Scheme 1. Molecular formulae and structures of compounds 1Ϫ5
Scheme 4. Reagents and conditions: (i) 6-chloro-9-(trimethylsilyl)-
purine (Y ϭ Cl); CH2Cl2; TMSOTf; room temp.; (ii) NaOH,
CH3OH, H2O; ∆T
7
6
Scheme 2. Through-space JHF and JCF couplings constants
All the observed through-space couplings were found to
exhibit a strong solvent effect, and without exception this
led to a decrease in the absolute value of J with increasing
hydrogen bond acceptor capabilities of the solvent. Accord-
ingly, we proposed that through-space heteronuclear J
coupling constants could be exploited in the investigation
of nonbonded CϪF...HϪC interactions which, in turn, can
be regarded as the analogues of the putative O...HϪC hy-
drogen bonds mentioned above. In the present work we re-
port on the synthesis, NMR and X-ray investigations car-
ried out on a selection of 5Ј-O-benzyl-2Ј,3Ј-dideoxy-4Ј-
(fluoromethyl)nucleosides carrying both purinic and pyrim-
idinic nucleobases with the purpose of obtaining more gen-
eral and quantitative insight into the nonbonded
C(sp3)ϪF...HϪC(sp2) interaction which gives rise to the
through-space transmission of spin-spin coupling.
1H- and 13C-NMR Study
7
6
The notation JHF and JCF will be used throughout this
paper to indicate the through-space couplings between F
and the nuclei (H or C) on the purinic or pyrimidinic base,
although this notation refers to a hypothetical through-
bonds pathway. The majority of the compounds investi-
gated was analysed as a 1:1 mixture of the α and β anomers.
The reported NMR data (1H- and 13C-chemical shifts, het-
eronuclear HϪF and CϪF coupling constants) are related
to both the α and β anomers of compounds 1Ϫ5, although
only α nucleosides exhibit the F atom and the nucleobase
in a syn relationship, which is mandatory for the occurrence
of any through-space interaction between the remote F and
a suitable H/C nucleus on the base. The values of the nor-
mally observed through-bonds heteronuclear coupling con-
Results and Discussion
Synthesis
2
1
2
3
stants (namely JHF, JCF, JCF, JCF) in the β isomers, and
their variation as a function of the solvent polarity, were
used as a reference for the qualitative estimation of solvent
The molecular structure and atom numbering of the nu- effects on couplings. As through-space F...H interactions
cleosides 1Ϫ5 are displayed in Scheme 1. The 5-O-benzyl- cannot take place for the β nucleosides in our model com-
protected nucleoside analogues 1؊4 were obtained as a pounds (for obvious geometrical reasons) the solvent effects
mixture of β and α anomers in good yields (Ն 80% in every on the coupling constants of the latter represent a non-
case) at room temp. and in a few minutes, by the addition specific and constant contribution that will be taken into
of a solution of acetylated lactol 6, dissolved in dichloro- account when discussing the possible solute-solvent inter-
188
Eur. J. Org. Chem. 1999, 187Ϫ196

Intramolecular Attractive Nonbonded C–F···H–C Interactions in 2Ј,3Ј-Dideoxy-4Ј-(fluoromethyl)nucleosides
FULL PAPER
Table 1. Heteronuclear HϪF coupling constants (Hz) of compounds 1؊5 (α and β anomers) in different solvents;[a] see Scheme 1 for
atom numbering
CDCl3
α anomer β anomer
MeOD
α anomer β anomer
[D6]acetone
α anomer β anomer
[D6]DMSO
α anomer β anomer
compd.
J(FϪH)
1
2J(FϪH-6Јa)
2J(FϪH-6Јb)
4J(FϪH-5Јa)
4J(FϪH-5Јb)
7J(FϪH-6)
45.1
47.1
1.7
47.2
47.2
1.3
48.0
47.5
47.3
1.5
47.9
47.4
1.8
47.5
47.5
1.6
47.4
47.4
1.8
47.4
47.4
1.4
47.1
[b]
[b]
2.3
1.7
1.7
2.1
1.6
2.2
1.1
1.8
Ϫ
1.4
Ϫ
1.1
Ϫ
Ϫ
Ϫ
2
3
4
5
2J(FϪH-6Јa)
2J(FϪH-6Јb)
4J(FϪH-5Јa)
4J(FϪH-5Јb)
7J(FϪH-6)
48.3
47.0
1.6
47.1
47.1
1.0
1.6
Ϫ
46.9
46.9
0.8
1.6
Ϫ
48.1
47.4
47.4
1.3
1.6
Ϫ
48.1
47.4
47.4
1.5
1.6
Ϫ
47.5
47.1
1.1
1.6
Ϫ
47.8
47.0
1.8
47.4
47.4
1.4
47.1
47.3
[b]
[b]
[b]
[b]
2.1
1.8
1.4
2.7
2.2
2.0
1.4
Ϫ
2J(FϪH-6Јa)
2J(FϪH-6Јb)
4J(FϪH-5Јa)
4J(FϪH-5Јb)
7J(FϪH-6)
48.4
47.9
47.4
47.2
1.0
48.3
[b]
[b]
[b]
[b]
[b]
[b]
47.3
47.3
46.6
[b]
[b]
[b]
1.3
1.8
Ϫ
[b]
[b]
[b]
1.8
3.1
2.6
Ϫ
2.4
2.2
2J(FϪH-6Јa)
2J(FϪH-6Јb)
4J(FϪH-5Јa)
4J(FϪH-5Јb)
LW(H-8)[c]
47.8
47.5
47.1
1.8
47.8
47.8
47.4
2.2
47.5
47.5
2.0
47.3
47.3
47.3
1.8
1.9
[b]
[b]
46.8
47.4
47.3
[b]
[b]
[b]
1.6
1.9
[b]
[b]
[b]
[b]
1.9
2.1
1.8
[b]
1.53
47.8
0.98
47.3
46.9
2.8
.
2.21
47.7
1.47
47.7
47.7
2.1
1.51
47.3
47.6
1.8
1.9
1.76
1.54
2J(FϪH-6Јa)
2J(FϪH-6Јb)
4J(FϪH-5Јa)
4J(FϪH-5Јb)
LW(H-8)[c]
47.8
47.5
47.2
1.9
[d]
[d]
[d]
[d]
[d]
47.3
47.4
47.7
[b]
[b]
[b]
[b]
[b]
[b]
1.8
1.6
1.9
1.32
1.8
1.0
1.5
1.35
1.41
[a]
The diastereotopic pairs H-6Ј and H-5Ј were not assigned stereospecifically. In all the cases the more deshielded proton was labelled
as “a”. Ϫ [b] Not determined due to peak overlap. Ϫ [c] As J(FϪH-8) was not resolved, we report here the value (Hz) of the linewith at
7
half height. Ϫ [d] Data not available in this solvent.
ference with the through-space transmission of HF and CF
spin information in the α anomers.
Tables 1, 2 and 3 show that a clear and monotonic trend
for the variation of HF and CF coupling constants with the
The NMR experiments were carried out in CDCl3, [D6]- nature of the solvent can only be found in the case of the
7
6
acetone, MeOD and [D6]DMSO. The choice of the NMR through-space coupling constants JHF and JCF, and that
solvents was based upon their polarity and hydrogen bond solvent-induced values of ∆J on the latter are much larger
profile. CDCl3 can be considered, as a first approximation, than those found, if any, for the other heteronuclear
as apolar and without any specific hydrogen bond proper- through-bond-transmitted coupling constants. For com-
7
6
ties, acetone and DMSO are hydrogen bond acceptors, pound 2, for example, the relative changes in JHF and JCF
whilst MeOD is amphiprotic, e.g. both hydrogen bond do- (defined as ∆J/Jmax) are 48 and 41%, respectively. For com-
nor and acceptor.[15]
parison, the same quantities for the one-bond JCF of 2α
1
3
Table 1 gives a selection of H-NMR data for the com- and 2β are 1.8 and 2.8%, respectively, and 10% for JC-3ЈϪF
pounds 1Ϫ5 in the solvents mentioned above. Only those of 2β. The apparent anomalies in the data for MeOD will
signals that are affected by heteronuclear FϪH spin-spin be discussed later. When dealing with the solvent depen-
couplings are considered. The α anomer of compounds 1, 2 dence of the NMR coupling constants, it is generally as-
and 3 show a large through-space FϪH coupling constant, sumed that any variation in J can be ascribed to two main
whilst the corresponding values for 4 and 5 could not be factors:[16] (i) electronic changes due to mutual solute-sol-
obtained due to the intrinsic large width of the signal re- vent interaction; (ii) solvent-induced conformational tran-
lated to H-8.
sitions, such as internal rotation and ring interconversions.
13C-NMR data for compounds 1Ϫ5 in all of the solvents The necessity of a clear distinction between these two differ-
are summarized in Tables 2 and 3. As expected, C-6 of com- ent sources of solvent dependence restricted the investi-
pounds 1α, 2α and 3α, and C-8 of 4α and 5α show a quite gation to purely rigid molecular structures, such as unsatu-
6
large JCF value between the fluorine nucleus and C-6/8. rated and aromatic compounds,[17,18] or model compounds
The absence of a CϪF coupling constant on C-1Ј supports with few rotational barriers.[19,20] The compounds con-
the hypothesis of a through-space pathway for the trans- sidered in the present study are highly flexible and, there-
mission of the spin information between these nuclei. As an fore, do not belong to either class of molecules mentioned
example, we show in Figures 1 and 2 an expansion of the above. The issue of the large solvent dependency of the
1H- and 13C-NMR spectrum of compound 1, showing the long-range couplings in 1Ϫ5 must be correctly addressed at
difference between the anomers α and β and the decrease this stage. We carried out a set of heteronuclear steady-state
in through J-space with increasing solvent polarity.
1H{19F}-NOE experiments in order to investigate possible
Eur. J. Org. Chem. 1999, 187Ϫ196
189

A. Mele, B. Vergani, F. Viani, S. V. Meille, A. Farina, P. Bravo
FULL PAPER
Figure 2. Expansion of 13C{1H}-NMR spectrum of compound 1
in different solvents showing the signals assigned to C-6; anomer
β: singlet; anomer α: doublet due to through-space JCF; from top
to bottom: CDCl3, [D4]methanol, [D6]acetone and [D6]DMSO; the
spectra are resolution-enhanced by gaussian multiplication prior to
Fourier transform
Figure 1. Expansion of 1H-NMR spectrum of compound 1 in differ-
ent solvents showing the signals assigned to H-6; anomer β: dou-
blet; anomer α: doublet of doublets due to through-space JHF; from
top to bottom: CDCl3, [D4]methanol, [D6]acetone and [D6]DMSO
changes, caused by the solvent, in the mutual position of
the RF moiety with respect to the nucleobase. The results, of spin information from F to the amide moiety. The au-
expressed as % enhancement of H-6/8, are summarized in thors corroborated this view by reporting on a marked de-
4
6
Table 4. It is well known that steady-state NOE does not crease of through-space JNF and JHF with increasing sol-
allow the quantitative measurement of internuclear dis- vent polarity, as a result of the decreased persistence of in-
tances,[21] hence the data in Table 4 emphasize that positive, tramolecular hydrogen bond. More recently, through-space
detectable and significant NOEs between F and H-6/8 in HF coupling constants in urolobin difluoroboron com-
the α anomers of all the investigated compounds can be plexes were reported, giving direct evidence of F...HϪN hy-
measured, independent of solvent polarity. Indeed, no dra- drogen bond.[24] The data reported in Tables 1, 2 and 3 fit
matic change in the conformational preferences of the nu- well with the model of through-space couplings mediated
cleosides (at least as far as RF and the base are concerned) via an intramolecular hydrogen bond. Solvents like acetone
can be observed experimentally. This rules out any possible and DMSO diminish but do not destroy the intramolecular
significant contribution of solvent-induced conformational CϪF...HϪC interaction, by participating via a solute/sol-
transitions to the large solvent effect detected in both the vent hydrogen bond. The efficiency of the H-relayed trans-
7
6
through-space JHF and JCF. The latter, which monoton- mission of spin-spin couplings is hence lowered, affording
7
6
ically decreases with solvent as shown in Tables 1, 2 and lower values of JHF and JCF accordingly. We also at-
3, can thus be consistently correlated with the increasing tempted to make this qualitative view more quantitative by
hydrogen bond acceptor properties of the medium.
searching for a linear correlation between the through-space
The physical basis for the through-space transmission of J and some empirical solvent hydrogen bond donor indi-
spin-spin couplings has been extensively reviewed.[22] The cator. The NMR parameters of compound 3 were therefore
main factor affecting the through-space route is the overlap- measured in three more solvents, [D3]acetonitrile, [D8]te-
ping of F lone pairs with the σ*(CH) molecular orbital. trahydrofuran and [D9]hexamethylphosphoric triamide
Theoretical and experimental studies on 2-fluoro-N-methyl- (HMPT), in order to provide a minimum number of data
benzamide[23] showed that the intramolecular CϪF...HϪN points for a linear correlation with a good confidence level,
7
6
hydrogen bond is an efficient pathway for the transmission and the observed values of JHF and JCF plotted against
190
Eur. J. Org. Chem. 1999, 187Ϫ196

Intramolecular Attractive Nonbonded C–F···H–C Interactions in 2Ј,3Ј-Dideoxy-4Ј-(fluoromethyl)nucleosides
FULL PAPER
Table 2. Selection of 13C{1H}-NMR data (chemical shift δ in ppm from internal TMS and heteronulcear JCF in Hz) of compounds 1Ϫ3
(α and β anomer) in different solvents; see Scheme 1 for atom numbering; chemical shifts are reported with two decimal figures only
when required for peak assignment
CDCl3
β anomer
MeOD
α anomer β anomer
[D6]acetone
α anomer β anomer
[D6]DMSO
α anomer β anomer
compd.
C atom
α anomer
δ
J
δ
J
δ
J
δ
J
δ
J
δ
J
δ
J
δ
J
1
C6
139.5 6.5
140.2 Ϫ
141.8 5.0
142.4 Ϫ
141.4 4.2
141.8 Ϫ
140.2 3.0
140.3 Ϫ
C6Ј
C4Ј
C5Ј
C3Ј
C6
85.6 174.1 85.0 177.7 86.2 173.0 86.3 174.6 86.6 172.4 86.6 174.5 84.5 171.5 84.7 172.6
85.7 17.1 85.6 18.2 87.3 17.2 87.3 17.7 86.9 17.1 86.8 17.6 84.8 17.1 84.9 17.4
72.1 6.6
28.1 4.2
139.3 7.6
71.9 4.8
27.7 3.4
140.1 Ϫ
73.1 6.1
28.7 4.3
141.3 6.8
72.8 5.5
28.5 3.8
141.9 Ϫ
73.4 6.0
29.2 4.3
141.2 5.8
73.2 5.4
29.0 4.0
141.7 Ϫ
71.4 5.5
27.5 4.2
139.9 4.5
71.1 5.3
27.2 4.2
140.0 Ϫ
2
3
C6Ј
C4Ј
C5Ј
C3Ј
C6
C6Ј
C4Ј
C5Ј
C3Ј
85.4 174.1 84.9 177.2 86.2 172.8 86.2 174.4 85.1 171.0 86.6 174.2 84.5 171.0 84.7 172.3
86.3 17.3 86.1 18.4 87.8 16.8 87.7 17.6 87.7 16.6 87.5 17.4 85.3 16.8 85.4 17.1
71.9 6.6
27.7 4.2
144.3 9.1
84.82 174.6 84.77 177.8 85.7 172.8 85.9 174.8 86.2 172.5 86.4 174.0 84.21 171.7 84.29 172.5
88.2 17.0 88.0 18.2 89.4 16.7 89.3 17.7 89.4 16.7 89.2 17.9 87.1 16.7 87.0 17.6
71.6 6.4
27.03 4.1
71.3 5.2
27.4 3.5
145.3 Ϫ
73.0 6.3
28.3 4.5
146.0 7.3
72.4 5.5
28.2 3.7
146.9 Ϫ
73.5 6.0
28.9 4.2
145.9 6.7
72.9 5.5
28.7 3.9
146.6 Ϫ
71.5 5.5
27.1 4.2
144.5 5.9
70.8 5.3
26.8 3.9
145.1 Ϫ
70.5 5.6
26.96 3.5
72.7 6.3
27.6 4.4
71.8 5.8
27.7 3.5
73.2 6.2
28.2 <0.5 28.1 <0.5 26.3 4.4
72.2 5.3
71.1 5.9
70.1 5.3
26.1 3.8
Table 3. Selection of 13C{1H}-NMR data (chemical shift δ in ppm from internal TMS and heteronulcear JCF in Hz) of compounds 4Ϫ5
(α and β anomer) in different solvents; see Scheme 1 for atom numbering; chemical shifts are reported with two decimal figures only
when required for peak assignment
CDCl3
β anomer
MeOD
α anomer β anomer
[D6]acetone
α anomer β anomer
[D6]DMSO
α anomer β anomer
compd.
C atom
α anomer
δ
J
δ
J
δ
J
δ
J
δ
J
δ
J
δ
J
δ
J
4
C8
143.6 5.7
144.0 Ϫ
146.4 4.2
146.8 Ϫ
146.3 3.2
146.5 Ϫ
145.69 1.8
145.75 Ϫ
C4Ј
C6Ј
C5Ј
C3Ј
C8
87.0 17.4 86.9 17.9 88.5 17.4 88.4 17.1 88.3 17.2 88.2 17.2 86.27 17.1 86.29 17.1
84.9 175.3 85.3 176.1 85.7 174.0 86.2 174.0 86.1 172.6 86.5 172.6 84.2 172.1 84.7 172.1
71.9 5.9
27.9 4.3
138.4 4.5
71.5 5.3
28.1 3.5
138.6 Ϫ
72.9 5.5
28.7 4.2
139.9 4.1
72.3 5.3
28.9 4.2
140.3 Ϫ
73.2 5.6
29.2 4.6
139.3 3.2
72.9 5.3
29.4 4.0
139.5 Ϫ
71.3 5.0
27.6 4.2
70.8 5.3
27.7 4.2
[a]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
5
138.3
C4Ј
C6Ј
C5Ј
C3Ј
86.56 17.8 86.52 17.8 88.0 17.4 88.0 17.4 87.9 17.5 87.8 17.8 85.3 17.4
84.9 175.2 85.3 175.5 85.8 173.7 86.2 174.2 86.3 173.6 86.7 173.6 84.2 172.4
71.9 5.7
28.1 4.1
71.6 5.4
28.3 3.5
73.0 5.6
28.8 4.3
72.5 5.5
29.1 3.7
73.5 5.4
29.6 4.5
73.3 5.4
29.7 4.1
71.3 4.8
27.7 4.1
[a] Broad singlet. Ϫ [b] Data not available in this solvent.
Table 4. 1H{19F} NOE (%) observed on the signal of H-6 (com- gen bond. The quantitative correlation obtained on the mo-
pounds 1, 2 and 3, anomer α only) and H-8 (compounds 4 and 5,
del compound 3 may shed some light on the solvent effect
anomer α only) in different solvents; see Scheme 1 for atom num-
experienced by all the compounds examined, including 4
and 5, whose 7JHF are not resolved. Moreover, the inductive
electronic effect (ϩI or ϪI) of the Y substituent on the nu-
cleobase is likely to affect the electron-withdrawing charac-
ter of C-6/8, hence affecting the hydrogen bond donor capa-
bilities of H-6/8. This would represent a further proof, al-
beit indirect, of the attractive interaction between F and H-
6/8. Indeed, the HammettЈs constants for inductive sub-
stituent effect σI show a reasonable linear correlation with
the values of SCC (Substituent Coupling Constant) in
CDCl3.[29] The regression indicates that the electron-with-
drawing properties of the functional groups on the nucleo-
base do affect the entity of the observed couplings, in good
agreement with the model of F...H interaction proposed
hitherto. The NMR data we have reported and discussed
so far and the regression analyses permit us to conclude
that the intramolecular hydrogen bond mechanism provides
a relevant contribution to the through-space HF and CF
bering
compd. CDCl3
MeOD
[D6]acetone
[D6]DMSO
1
2
3
4
5
2.8
4.1
4.2
0.8
0.9
3.4
3.5
5.7
0.6
0.8
2.0
3.3
5.1
1.1
0.6
1.0
2.2
3.6
0.5
1.0
the KamletϪTaftЈs solvatochromic parameter β, taken as
empirical hydrogen bond acceptor scale.[25Ϫ27] The graph is
reported in Figure 3 and it is based on the data summarized
in Table 5.[28]
This linear relationship provides further quantitative evi-
dence that the through-space CF and HF coupling con-
stants are mediated by intramolecular CϪF...HϪC hydro-
Eur. J. Org. Chem. 1999, 187Ϫ196
191

A. Mele, B. Vergani, F. Viani, S. V. Meille, A. Farina, P. Bravo
FULL PAPER
Table 7. Significant torsion angles [°] in the solid-state conforma-
tion of 3α and 3β
3α
3β
C6ϪN1ϪC1ЈϪO1Ј
C1ЈϪO1ЈϪC4ЈϪC6Ј
C1ЈϪO1ЈϪC4ЈϪC5Ј
O1ЈϪC4ЈϪC5ЈϪO5Ј
C7ЈϪO5ЈϪC5ЈϪC4Ј
C5ЈϪO5ЈϪC7ЈϪC1Љ
O5ЈϪC7ЈϪC1ЉϪC2Љ
O1ЈϪC4ЈϪC6ЈϪF6Ј
O1ЈϪC4ЈϪC6ЈϪF6ЈA
O1ЈϪC4ЈϪC6ЈϪF6ЈB
2.1(10)
47.1(8)
Ϫ105.1(10)
132.2(8)
Ϫ105.0(6)
135.9(5)
Ϫ56.9(7)
178.0(5)
Ϫ66.6(7)
Ϫ36.1(9)
Ϫ174.0(8)
173.4(11)
Ϫ168.2(13)
Ϫ103.2(16)
Ϫ50.7(13)
51.3(8)
Ϫ73.4(9)
X-ray Crystallography
Views of the molecular conformations of 3α and 3β are
presented in Figure 4. Bond lengths and angles are normal,
considering the associated standard errors, and do not re-
quire comments. Rather our focus will be on the compari-
son between the conformations, the hydrogen bonding and
the packing patterns of the two isomers.
Figure 3. Plot of experimental through-space spin-spin coupling
constants of compound 3 vs. KamletϪTaftЈs parameter β
7
6
Table 5. J(FϪH-6) and J(FϪC-6) of compound 3 in the extended set
The furanoside is an envelope with apical C3Ј in 3α and
a twist boat in 3β. In both cases deviations from planarity
of solvents and empirical solvent parameters for linear correlation
˚
are modest: C3Ј is 0.156 A away from the ring atomsЈ mean
solvent[a]
KamletϪTaftЈs β[b] rel. permitt.[c] JHF
6JCF
7
[d]
[d]
least-squares plane in the α anomer, while deviations of ring
˚
atoms from the plane are less than 0.09 A for the β isomer.
chloroform 0.00
acetonitrile 0.31
4.81
3.10
2.86
2.43
2.35
2.62
2.21
1.61
9.10
7.33
6.75
6.46
7.33
5.87
4.70
The differences mentioned for the furanoside should hardly
affect the overall conformation of the two molecules, given
the flexibility of five-membered rings. The benzyl ether side
chain bound to C5Ј is fully extended in 3α while the
C5ЈϪO5ЈϪC7ЈϪC1Љ and the O1ЈϪC4ЈϪC5ЈϪO5Ј torsion
angles are gauche in the more compact β isomer (see Figure
4 and Table 7) bringing the O5Ј atom in proximity to the
base residue. Other related (substantial) differences involve
the C6ϪN1ϪC1ЈϪO1Ј torsion angle linking the base to the
sugar residue: It is 2° in 3α and 47°, i.e. nearly gauche, in
3β. As a consequence, the hydrogen atom on C1Ј is found
close to the O2 atom while the hydrogen atom on C6, in
35.94
20.56
7.58
acetone
THF
0.48
0,55
(0.62)
0,76
1.05
methanol
DMSO
HMPT
33.66
46.45
29.6
[a] Literature data are referred to nondeuterated solvents. Ϫ [b] Val-
ues taken from ref.[15], p. 378. The value referred to methanol is in
parentheses in the original reference and is reported by the author
[c]
as “relatively less certain”. Ϫ
Values taken from ref.[15], pp.
[d]
408Ϫ411. Ϫ Coupling constants in Hz. Sine bell multiplication
of the FID were applied prior to Fourier transform.
˚
Table 6. Values of substituent coupling constant (SCC, see text for
definition) of 5-Y-uracil nucleosides and corresponding values of
HammettЈs σI[a]
both moieties at ca. 2.3 A to O5A of the nitro group, inter-
acts with the furanoside O1Ј atom, at distances of 2.26 and
˚
2.54 A in the α and β anomer, respectively. The intramolec-
ular environment of H6 is completed by another relatively
short interaction involving the appropriately located O5Ј in
3β and F6Ј in 3α. Further geometrical details of the “cage”
around the H6 atom are given in Table 8. The confor-
mational arrangement of both molecules suggest an attract-
ive nature of all the mentioned interactions involving H6.
More specifically, in the β anomer, the H6...O5Ј distance
(2.36A) and the C6ϪH6 O5Ј angle (162°) are values close
to the optimized geometry for this kind of hydrogen bond.
In the α-anomer the eclipsing of C6 with O1Ј leads to the
substituent Y
HammettЈs σI
SCC(HF)
SCC(CF)
H[b]
Me[c]
F[c]
0.00
Ϫ0.04
0.50
0.44
0.65
0.0
0.0
Ϫ0.3
1.0
Ϫ0.1
1.3
Br[b]
NO2
0.9
1.3
1.0
2.4
[b]
[a]
[b]
Values taken from ref.[31]
Ϫ
Values taken from ref.[14]
Ϫ
...
˚
[c] This work.
spin-spin coupling, although other sources of transmission optimization of the interaction of H6 with O1Ј rather than
of spin information may also be operating. On the other with F6Ј, which however does not display disorder, contrary
hand, the present study demonstrated the existence of a sig- to what we find in the β anomer, where the fluorine atom
nificant interaction in solution between F and H in the title is distributed with a 0.5 occupancy factor over two posi-
class of compounds which is likely to be described as tions. The relative conformational freedom of the fluoro-
(sp3)CϪF...HϪC(sp2) hydrogen bond.
methyl group leads in the present instances to gauche ar-
192
Eur. J. Org. Chem. 1999, 187Ϫ196

Intramolecular Attractive Nonbonded C–F···H–C Interactions in 2Ј,3Ј-Dideoxy-4Ј-(fluoromethyl)nucleosides
FULL PAPER
˚
rangements to the O1Ј atom: In the case of the α anomer Table 9. Intermolecular distances [A] in 3α and 3β; notation: A ϭ
acceptor, D ϭ donor
the rotational isomeric state giving the best compromise ge-
...
...
˚
ometry [H6 F6Ј 2.81 A, for C6ϪH6 F6Ј ϭ 141°] results,
with deviations from the exact gauche conformation of the
F6ЈϪC6ЈϪC4ЈϪO1Ј angle bringing F6Ј closer to H6. All
these features support the idea of a weakly attractive nature
of the H6...F6Ј interaction, and this is also consistent with
the observed distance, which is somewhat larger than the
3α
3β
HиииA DиииA DϪHиииA HиииA DиииA DϪHиииA
N3ϪH3иииO4
2.00 2.83 161
2.06 2.88 160
2.61 3.14 121
N3ϪH3иииO5B 2.69 3.26 125
C1ЈϪH1ЈиииO2 2.70 3.22 114
C1ЈϪH1ЈиииO4 2.63 3.39 135
sum of the van der Waals radii. What has been noted for C2ЈϪH2ЈиииF6ЈB
2.68 3.47 138
2.65 3.56 156
O...HϪC bonds[10] should be even more true for F...HϪC
bonds: The energy of these primarily electrostatic interac-
tions should decrease much more slowly with distance than
for van der Waals interactions. Therefore, even if the dis-
tances H...F are long and cannot be expected to be the main
factor in determining conformational preferences or crystal
packing, they should be considered carefully. In Table 9 the
C2ЈϪH2ЈиииO5A 2.64 3.56 157
C2ЈϪH2ЈиииO5B 2.86 3.55 128
C2ЈϪH2ЈиииO2
C3ЈϪH3ЈиииF6Ј 2.78 3.52 134
C3ЈϪH3ЈиииO5B 2.71 3.48 137
C5ЈϪH5ЈиииN5 2.80 3.50 129
C5ЈϪH5ЈиииO2
C5ЈϪH5ЈиииO5B 2.87 3.43 118
C6ЈϪH6ЈиииO5B
C5ЉϪH5ЉиииF6ЈA
2.83 3.55 132
2.57 3.51 161
2.72 3.16 110
2.69 3.60 169
3.19
more relevant intermolecular hydrogen bonds and a num- C5ЉϪH5ЉиииO5B
O5BиииO4
3.03
2.95
3.06
ber of rather short contacts involving the fluorine, nitrogen
N1иииO2
and oxygen atoms are listed for both crystals. Interestingly,
both are characterized by the hydrogen bond involving H3
and the O4 atom with closely similar geometrical features,
and by an interaction involving the O5B of the nitro group
to a hydrogen atom. In addition a number of O...HϪC and
F...HϪC interactions are apparent with geometrical features
similar to those identified intramolecularly for the α and β
anomers, confirming suggestions that these interactions are
also likely to play a significant cooperative role in the pro-
cess of crystal formation and stabilization.
C2иииO2
N5иииO4
3.17
2.87
3.16
F6ЈBиииO2
F6ЈAиииO5B
˚
Table 8. Intramolecular distances [A] and angles [°] in 3α and 3β;
notation: A ϭ acceptor, D ϭ donor
3α
3β
HиииA DиииA DϪHиииA HиииA DиииA DϪHиииA
C6ϪH6иииO1Ј
2.26 2.65 104
2.54 2.77 95
2.33 2.66 101
C6ϪH6иииO5A 2.34 2.67 100
C6ϪH6иииF6Ј
C6ϪH6иииO5Ј
C1ЈϪH1ЈиииO2 2.63 2.71 84
C2ЈϪH2ЈиииO2 2.78 2.97 92
2.81 3.58 141
2.36 3.26 162
2.35 2.76 104
Conclusions
The α anomers of the class of compounds discussed in
the present paper represent an example of systems preor-
ganized in a conformation such that F and the nucleobase
are in an ideal relationship for an attractive interaction (hy-
drogen bond) to take place. Only in preorganized environ-
ments, as already observed for fluorocarbons[30], do intrin-
sically weak interactions like CϪF...HϪC become relevant.
The linear correlations of through-space J couplings with
the hydrogen bond basicity parameter β and the effect of
the Y substituent support the conclusion that through-
Figure 4. ORTEP view of 3α and 3β showing the atomic labelling
scheme
space couplings reflect not only the geometrical vicinity of fluorine-involving hydrogen bonds. The results of the X-ray
the interacting nuclei, but the existence of an attractive in- investigation on both anomers of compound 3 indicate that,
teraction between them as well. This opens up the possibil- in the presence of appropriate cooperative effects, the hy-
ity to exploit these NMR parameters, easily accessible from drogen bond acceptor profile is not dramatically upset upon
routine spectra, for the investigation of unusual types of replacement of O with F. It is clear that the energy involved
Eur. J. Org. Chem. 1999, 187Ϫ196
193

A. Mele, B. Vergani, F. Viani, S. V. Meille, A. Farina, P. Bravo
FULL PAPER
in CϪF...HϪC hydrogen bonds is low since it involves dis-
tances larger than the sum of van der Waals radii.
General Procedure for the Glycosylation Reaction: The purinic/pyri-
midinic base (2.5 mmol) was suspended in hexamethyldisilazane
(HMDS, 35.4 mmol) and refluxed in the presence of catalytic
amounts (0.2 mmol) of ammonium sulfate for 4 h under nitrogen.
Volatiles were distilled off under reduced pressure and the resulting
residue dried under reduced pressure for 1 h at room temp. A solu-
tion of acetylated lactol 6 (1.0 mmol) in CH2Cl2 or (ClCH2)2 (25
mL) was added to the silylated base followed by the addition of
trimethylsilyl triflate (TMSOTf, 2.0 mmol) to the resulting slurry.
The clear solution was stirred at room temp. for 15 min, diluted
with CHCl3 and washed with satd. NaHCO3 solution. The aqueous
portions were extracted with CHCl3 and the combined organic lay-
ers were dried with anhydrous sodium sulfate. The solvent was re-
moved under reduced pressure to give a crude product that was
purified by flash chromatography.
Experimental Section
1H and 13C NMR: Bruker ARX 400 spectrometer (400/100 MHz).
Ϫ Heteronuclear 1H{19F} NOE: Bruker AC 250 (250 MHz op-
erating frequency on proton) equipped with a supplementary BM1
broadband modulator operating at 235 MHz. Chemical shifts are
given in ppm (δ) relative to internal tetramethylsilane (TMS). All
the samples have been accurately degassed by carefully bubbling
N2. The coupling constants were determined on the resolution en-
hanced spectra (typically LB ϭ Ϫ2.5; GB ϭ 0.28). Ϫ MS: Shim-
adzu MALDI II equipped with a pulsed N2 laser (λ ϭ 337 nm), α-
cyanohydroxycinnamic acid was used as matrix for all the com-
pounds, time-delayed extraction was optimized for m/z of the pro-
tonated (or cationized) molecule for all compounds. Ϫ X-ray crys-
tallography: See Table 10. Ϫ Flash chromatographies: Silica gel 60
(60Ϫ200 µm, Merck). Ϫ All reactions were monitored by TLC:
Analytical Merck silica gel 60F254 TLC plates. Ϫ Combustion mi-
croanalyses: Redox SNC, Cologno Monzese (Milano). Ϫ All the
reactions run in organic solvents were carried out in flame-dried
flasks under nitrogen. Dichloromethane and 1,2-dichloroethane
were distilled from calcium hydride prior to use. Commercially
available reagent-grade reagents were employed without purifi-
cation. The synthesis of 1-O-acetyl-5-O-benzyl-2,3-dideoxy-4-C-
(fluoromethyl)-β/α--glycero-pentafuranose (6) has already been
described.[3]
1-[5-O-Benzyl-2,3-dideoxy-4-C-(fluoromethyl)-β/α--glycero-penta-
furanosyl]uracil (1): Starting from 6 and 2,4-bis(trimethylsilyloxy)-
uracil in (ClCH2)2, a 1:1 β/α mixture (1H-NMR ratio) of monofluo-
robenzyl-protected uridine analogues 1 was obtained (85% yield)
and purified, but not resolved, by flash chromatography (n-hexane/
ethyl acetate, 7:3) to give the product mixture as a white solid, Rf ϭ
20
0.35. Ϫ M. p. 82°C (DMSO/CH3COCH3, 1:1). Ϫ [α]D ϭ Ϫ2.19
1
(c ϭ 0.7, CH3COCH3). Ϫ NMR of α anomer: H NMR (CDCl3):
δ ϭ 1.93Ϫ2.53 (m, 8 H, H-2Ј and H-3Ј α/β), 3.40 (dd, J ϭ 9.9 and
2.3 Hz, 1 H, H-5Јb), 3.44 (dd, J ϭ 9.9 and 1.7 Hz, 1 H, H-5Јa),
4.50 (dd, J ϭ 47.0 and 10.1 Hz, 1 H, H-6Јb), 4.61 (dd, J ϭ 45.1
and 10.1 Hz, 1 H, H-6Јa), 4.54Ϫ4.60 (m, 4 H, H-7Ј α/β), 5.74 (d,
J ϭ 8.1 Hz, 1 H, H-5), 6.23Ϫ6.16 (m, 2 H, H-1Ј α/β), 7.27Ϫ7.41
(m, 10 H, Harom α/β), 7.58 (dd, J ϭ 8.1 and 1.8 Hz, 1 H, H-6), 9.06
(s, 1 H, NH). Ϫ 13C NMR (CDCl3): δ ϭ 28.1 (d, J ϭ 4.2 Hz, C-
3Ј), 32.1 (s, C-2Ј), 72.1 (d, J ϭ 6.6 Hz, C-5Ј), 73.7 (s, C-7Ј), 85.7
(d, J ϭ 17.1 Hz, C-4Ј), 85.60 (d, J ϭ 174.1 Hz, C-6Ј), 86.2 (s, C-
1Ј), 102.6 (s, C-5), 127.9, 128.00, and 128.5 (s, Carom), 137.4 (s, C-
1Љ), 139.5 (d, J ϭ 6.5 Hz, C-6), 150.4 (s, C-2), 163.1 (s, C-4). Ϫ
NMR of β anomer: 1H NMR (CDCl3): δ ϭ 1.93Ϫ2.53 (m, 8 H,
H-2Ј and H-3Ј α/β), 3.60 (dd, J ϭ 10.0 and 1.7 Hz, 1 H, H-5Јb),
3.73 (dd, J ϭ 10.0 and 1.3 Hz, 1 H, H-5Јa), 4.30 (AB part of ABX,
J ϭ 47.2 and 10.0 Hz, 2 H, H-6Ј), 4.54Ϫ4.60 (m, 4 H, H-7Ј α/β),
5.44 (dd, J ϭ 8.1 and 1.0 Hz, 1 H, H-5), 6.23Ϫ6.16 (m, 2 H, H-1Ј
α/β), 7.27Ϫ7.41 (m, 10 H, Harom α/β), 7.75 (d, J ϭ 8.1 Hz, 1 H, H-
6), 9.01 (s, 1 H, NH). Ϫ 13C NMR (CDCl3): δ ϭ 27.7 (d, J ϭ 3.4
Hz, C-3Ј), 32.7 (s, C-2Ј), 71.9 (d, J ϭ 4.8 Hz, C-5Ј), 73.9 (s, C-7Ј),
85.6 (d, J ϭ 18.2 Hz, C-4Ј), 85.0 (d, J ϭ 177.7 Hz, C-6Ј), 86.33 (s,
C-1Ј), 105.1 (s, C-5), 127.94, 128.3, and 128.7 (s, Carom), 137.2 (s,
C-1Љ), 140.2 (s, C-6), 150.35 (s, C-2), 163.20 (s, C-4). Ϫ MS
(MALDI): m/z ϭ 357.0 [M ϩ Na]ϩ. Ϫ IR (neat): ν˜ ϭ 3388.23
Table 10. Crystal data, collection and refinement parameters for 3α
and 3β
3α
3β
crystal data
formula
C17H18FN3O6
379.34
C17H18FN3O6
379.34
M [g molϪ1
]
˚
a [A]
5.9830(10)
8.222(2)
35.411(6)
1741.9(6)
1.446
5.9620(10)
14.342(2)
20.051(3)
1714.5(5)
1.470
˚
b [A]
˚
c [A]
3
˚
V [A ]
ρcalcd. [g cmϪ1
Z
]
4
4
crystal system
space group
F(000)
orthorhombic
P212121
792
orthorhombic
P212121
792
µ (Cu-Kα) [mmϪ1
]
1.006
1.022
data collection
radiation
(λ ϭ 1.54178 (Cu-Kα),
(λ ϭ 1.54178 (Cu-
Kα),
graphite monochromator graphite mono-
chromator
Siemens P4
0.8 ϫ 0.4 ϫ 0.02
25
diffractometer
crystal size [mm]
temperature [°C]
Siemens P4
0.4 ϫ 0.08 ϫ 0.06
25
cmϪ1
, 2960.06, 1691.03, 1486.92, 1452.84, 1377.54, 1283.29,
1222.20, 1053.25, 817.31, 766.11. Ϫ C17H19FN2O4 (334): calcd. C
data collection mode
scan range (2 θ) [°]
measured reflections
unique reflections
observed reflections
I Ն
θϪ2 θ scan
5.00/133.90
2439
θϪ2 θ scan
7.58/136.00
2581
61.07, H 5.73, N 8.38; found C 61.10, H 5.72, N 8.35.
1-[5-O-Benzyl-2,3-dideoxy-4-C-(fluoromethyl)-β/α--glycero-penta-
furanosyl]-5-bromouracil (2): Starting from 6 and 5-bromo-2,4-bis-
(trimethylsilyloxy)uracil in (ClCH2)2, a 1:1 β/α mixture (1H-NMR
ratio) of monofluorobenzyl-protected uridine analogues 2 was ob-
tained (82% yield) and purified, but not resolved, by flash chroma-
tography (n-hexane/ethyl acetate, 3:2) to give the product mixture
2034
2125
1299
1316
2 σ(I)
2 σ(I)
absorption correction
min./max. transmission
ψ scan
ψ scan
0.238/0.367
0.239/0.366
structure solution and
refinement
20
structure solution program SIR92[32]
SIR92[32]
as an oil, Rf ϭ 0.35. Ϫ [α]D ϭ Ϫ4.73 (c ϭ 0.9, CH3COCH3). Ϫ
refinement method
full-matrix least squares on full-matrix least
NMR of α anomer: 1H NMR (CDCl3): δ ϭ 1.87Ϫ2.58 (m, 8 H,
H-2Ј and H-3Ј α/β), 3.40 (dd, J ϭ 9.9 and 2.3 Hz, 1 H, H-5Јb),
3.44 (dd, J ϭ 9.9 and 1.7 Hz, 1 H, H-5Јa), 4.516 (d, J ϭ 12.2 Hz,
1 H, H-7Јb), 4.518 (dd, J ϭ 47.0 and 10.1 Hz, 1 H, H-6Јb), 4.57
(d, 1 H, J ϭ 12.2 Hz, H-7Јa ), 4.66 (dd, J ϭ 48.3 and 10.1 Hz, 1
H, H-6Јa), 6.14 (dd, 1 H, J ϭ 5.8 and 6.9 Hz, H-1Ј), 7.27Ϫ7.41
(m, 10 H, Harom α/β), 7.92 (d, J ϭ 2.7 Hz, 1 H, H-6), 9.52 (br. s, 1
F2
squares on F2
256
refined parameters
R1
wR2
245
0.0794
0.0579
0.2173
0.1379
goodness-of-fit
1.016
1.019
Ϫ3
˚
∆ρ (max/min) [e A
]
0.211/Ϫ0.249
0.291/Ϫ0.197
structure refinement
program
SHELXL97[33]
SHELXL97[33]
194
Eur. J. Org. Chem. 1999, 187Ϫ196

Intramolecular Attractive Nonbonded C–F···H–C Interactions in 2Ј,3Ј-Dideoxy-4Ј-(fluoromethyl)nucleosides
FULL PAPER
H, NH). Ϫ 13C NMR (CDCl3): δ ϭ 27.7 (d, J ϭ 4.2 Hz, C-3Ј),
32.3 (s, C-2Ј), 71.9 (d, J ϭ 6.6 Hz, C-5Ј), 73.7 (s, C-7Ј), 85.4 (d,
J ϭ 174.1 Hz, C-6Ј), 86.3 (d, J ϭ 17.3 Hz, C-4Ј), 86.7 (s, C-1Ј),
96.8 (s, C-5), 127.7, 127.9, and 128.5 (s, Carom), 137.3 (s, C-1Љ),
139.3 (d, J ϭ 7.5 Hz, C-6), 149.8 (s, C-2), 159.0 (s, C-4). Ϫ NMR
of β anomer: 1H NMR (CDCl3): δ ϭ 1.87Ϫ2.58 (m, 8 H, H-2Ј and
H-3Ј α/β), 3.53 (dd, J ϭ 10.1 and 1.6 Hz, 1 H, H-5Јb), 3.71 (dd,
J ϭ 10.1 and 1.0 Hz, 1 H, H-5Јa), 4.26 (dd, J ϭ 47.1 and 9.8 Hz,
1 H, H-6Јb), 4.29 (dd, J ϭ 47.1 and 9.8 Hz, 1 H, H-6Јa), 4.61 (d,
1 H, J ϭ 12.1 Hz, H-7Јb ), 4.70 (d, J ϭ 12.1 Hz, 1 H, H-6Јa), 6.17
(dd, 1 H, J ϭ 5.9 and 6.2 Hz, H-1Ј), 7.27Ϫ7.41 (m, 10 H, Harom
α/β), 8.32 (s, 1 H, H-6), 9.50 (br. s, 1 H, NH). Ϫ 13C NMR
(CDCl3): δ ϭ 27.4 (d, J ϭ 3.5 Hz, C-3Ј), 32.6 (s, C-2Ј), 71.3 (d,
obtained (80% yield) and purified, but not resolved, by flash chro-
matography (n-hexane/ethyl acetate, 3:2) to give the product mix-
20
ture as an oil, Rf ϭ 0.35. Ϫ M. p. ϭ 85°C (CHCl3). Ϫ [α]D
ϭ
Ϫ10.28 (c ϭ 0.4, CHCl3/CH3COCH3 1:1). Ϫ NMR of α anomer:
1H NMR (CDCl3): δ ϭ 2.15Ϫ2.75 (m, 8 H, H-3Ј and H-2Ј α/β),
3.52 (m, 2 H, H-5Ј), 4.52 (dd, J ϭ 46.8 and 9.9 Hz, 1 H, H-6Јb),
4.59 (m, 2 H, H-7), 4.61 (dd, J ϭ 47.8 and 9.9 Hz, 1 H, H-6Јa),
6.43 (dd, J ϭ 4.6 and 6.5 Hz, 1 H, H-1Ј), 7.10Ϫ7.40 (m, 5 H,
Harom), 8.38 (s, 1 H, H-8), 8.73 (s, 1 H, H-2). Ϫ 13C NMR (CDCl3):
δ ϭ 27.9 (d, J ϭ 4.3 Hz, C-3Ј), 32.3 (s, C-2Ј), 71.9 (d, J ϭ 5.9 Hz,
C-5Ј), 73.8 (s, C-7Ј), 84.9 (d, J ϭ 175.3 Hz, C-6Ј), 86.38 (s, C-1Ј),
87.00 (d, J ϭ 17.4 Hz, C-4Ј), 127.66, 127.97, and 128.51 (s, Carom),
132.21 (s, C-5), 137.52 (s, C-1Љ), 143.64 (d, J ϭ 5.7 Hz, C-8), 151.03
J ϭ 5.5 Hz, C-5Ј), 73.8 (s, C-7Ј), 84.9 (d, J ϭ 177.2 Hz, C-6Ј), 86.1 or 151.09 (s, C-6), 151.85 (s, C-2). Ϫ NMR of β anomer: 1H NMR
(d, J ϭ 18.4 Hz, C-4Ј), 86.8 (s, C-1Ј), 96.6 (s, C-5), 127.8, 128.1
and 128.6 (s, Carom), 137.0 (s, C-1Љ), 140.1 (s, C-6), 149.77 (s, C-2),
159.1 (s, C-4). Ϫ MS (MALDI): m/z ϭ 436.6 [M ϩ Na]ϩ, 452.5 [M
ϩ K]ϩ. Ϫ IR (neat): ν˜ ϭ 3431.40 cmϪ1, 2959.01, 1695.72, 1453.08,
1271.46, 1054.69, 1028.07, 1008.09, 757.98. Ϫ C17H18BrFN2O4
(412): calcd. C 49.41, H 4.39, N 6.78; found C 49.40, H 4.42, N
6.75.
(CDCl3): δ ϭ 2.15Ϫ2.75 (m, 8 H, H-3Ј and H-2Ј α/β), 3.55 (dd,
J ϭ 10.0 and 1.9 Hz, 1 H, H-5Јb), 3.63 (dd, J ϭ 10.0 and 1.8 Hz,
1 H, H-5Јa), 4.40 (dd, J ϭ 47.1 and 9.6 Hz, 1 H, H-6Јb), 4.44 (dd,
J ϭ 47.5 and 9.6 Hz, 1 H, H-6Јa), 4.54 (s, 2 H, H-7Ј), 6.44 (t, J ϭ
5.5 Hz, 1 H, H-1Ј), 7.22Ϫ7.40 (m, 5 H, Harom), 8.42 (s, 1 H, H-8),
8.70 (s, 1 H, H-2). Ϫ 13C NMR (CDCl3): δ ϭ 28.1 (d, J ϭ 3.5 Hz,
C-3Ј), 32.2 (d, J ϭ 1.3 Hz, C-2Ј), 71.5 (d, J ϭ 5.3 Hz, C-5Ј), 73.7
(s, C-7Ј), 85.3 (d, J ϭ 176.1 Hz, C-6Ј), 86.49 (s, C-1Ј), 86.91 (d,
J ϭ 17.9 Hz, C-4Ј), 127.78, 128.06, and 128.54 (s, Carom), 132.21
(s, C-5), 143.9 (s, C-8), 151.03 or 151.09 (s, C-6), 151.83 (s, C-2). Ϫ
MS (MALDI): m/z ϭ 399.1 [M ϩ Na]ϩ; 415.3 [M ϩ K]ϩ. Ϫ IR
(nujol): ν˜ ϭ 1592.81 cmϪ1, 1560.02, 1339.99, 1203.00, 1128.68,
1055.70, 1030.82, 742.94, 635.02. ϪC18H18ClFN4O2 (377): calcd. C
57.37, H 4.81, N 14.87; found C 57.40, H 4.82, N 14.85.
1-[5-O-Benzyl-2,3-dideoxy-4-C-(fluoromethyl)-β/α--glycero-penta-
furanosyl]-5-nitrouracil (3): Starting from 6 and 5-nitro-2,4-bis(tri-
methylsilyloxy)uracil in (ClCH2)2, a 1:1 β/α mixture (1H-NMR ra-
tio) of monofluorobenzyl-protected nitrouridine analogues 3 was
obtained (85% yield) and purified, but not resolved, by flash chro-
matography (chloroform/methanol, 95:5) to give the product mix-
1
ture as a white solid, Rf ϭ 0.35. Ϫ NMR of α anomer: H NMR
(CDCl3): δ ϭ 1.98Ϫ2.74 (m, 8 H, H-2Ј and H-3Ј α/β), 3.45 (m, 2
H, H-5Ј), 4.53 (dd, J ϭ 47.3 and 10.4 Hz, 1 H, H-6Јb), 4.54 (d,
J ϭ 12.0 Hz, 1 H, H-7Јb), 4.58 (d, J ϭ 12.0 Hz, 1 H, H-7Јa ), 4.73
(dd, J ϭ 48.4 and 10.4 Hz, 1 H, H-6Јa), 6.10 (dd, 1 H, J ϭ 6.2 and
4.3 Hz, H-1Ј), 7.27Ϫ7.40 (m, 10 H, Harom α/β), 8.41 (br. s, 1 H,
NH), 9.17 (d, J ϭ 3.1 Hz, 1 H, H-6). Ϫ 13C NMR (CDCl3): δ ϭ
27.0 (d, J ϭ 4.4 Hz, C-3Ј), 33.2 (d, J ϭ 1.1 Hz, C-2Ј), 71.6 (d, J ϭ
6.4 Hz, C-5Ј), 73.9 (s,C-7Ј), 84.82 (d, J ϭ 174.6 Hz, C-6Ј), 88.2 (d,
J ϭ 17.0 Hz, C-4Ј), 88.5 (s, C-1Ј), 125.7 (br. s, C-5), 127.7, 128.2,
and 128.6 (s, Carom), 137.2 (s, C-1Љ), 144.3 (d, J ϭ 9.1 Hz, C-6),
148.17 (s, C-2), 153.65 (s, C-4). Ϫ NMR of β anomer: 1H NMR
(CDCl3): δ ϭ 1.98Ϫ2.74 (m, 8 H, H-2Ј and H-3Ј α/β), 3.54 (dd,
J ϭ 10.1 and 1.3 Hz,1 H, H-5Јb), 3.75 (d, J ϭ 10.1 Hz, H-5Јa),
4.27 (dd, J ϭ 46.9 and 9.7 Hz, 1 H, H-6Јb), 4.31 (dd, J ϭ 47.2 and
9.7 Hz, 1 H, H-6Јa), 4.61 (d, J ϭ 12.4 Hz, 1 H, H-7Јb), 4.77 (d,
J ϭ 12.4 Hz, 1 H, H-7Јa ), 6.15 (dd, 1 H, J ϭ 6.6 and 4.4 Hz, H-
1Ј), 7.27Ϫ7.40 (m, 10 H, Harom α/β), 8.46 (br. s, 1 H, NH), 9.64 (s,
1 H, H-6). Ϫ 13C NMR (CDCl3): δ ϭ 26.06 (d, J ϭ 3.5, C-3Ј Hz),
33.55 (d, J ϭ 1.3, C-2Ј), 70.5 (d, J ϭ 5.6 Hz, C-5Ј), 74.0 (s, C-7Ј),
84.77 (d, J ϭ 177.7 Hz, C-6Ј), 88.0 (d, J ϭ 18.2 Hz, C-4Ј), 88.6 (s,
C-1Ј), 125.6 (br. s, C-5), 128.0, 128.3, and 128.7 (s, Carom), 136.9 (s,
C-1Љ), 145.3 (s, C-6), 149.14 (s, C-2), 153.73 (s, C-4). Ϫ MS
(MALDI): m/z ϭ 402.6 [M ϩ Na]ϩ. Ϫ IR (nujol): ν˜ ϭ 3450.84
Synthesis of the Inosyl Derivative 5: The 6-chloropurin-9-yl deriva-
tive 4 (1.0 mmol, 372 mg) was added to a solution of sodium hy-
droxide (100 mmol, 2.4 g) in methanol (24 mL) and water (2.0 mL).
The solution was heated at reflux for 2 h, then diluted with water,
the pH was adjusted to 3 by adding diluted acetic acid (ca. 0.1 ).
The organic compounds were extracted with ethyl acetate (3 ϫ 10
mL), dried with anhydrous sodium sulfate, filtered and concen-
trated under reduced pressure to give a residue which was flash-
chromatographed in ethyl acetate/2-propanol/methanol (50:5:1). A
1:1 mixture of the α/β anomers 5 was obtained (250 mg, 70% yield)
as a pale yellow solid, Rf ϭ 0.35. Ϫ M. p. 140°C (diisopropyl
20
ether). Ϫ [α]D ϭ Ϫ11.93 (c ϭ 0.5, CH3COCH3). Ϫ NMR of α
anomer: 1H NMR ([D6]acetone): δ ϭ 2.14Ϫ2.89 (m, 8 H, H-2Јand
H-3Ј α/β) 3.62 (m, 2 H, H-5Ј), 4.56 (dd, J ϭ 47.7 and 9.8 Hz, H-
6Јb), 4.62 (dd, J ϭ 47.7 and 9.8 Hz, H-6Јa), 4.62 (m, 2 H, H-7Ј),
6.37Ϫ6.43 (m, 2 H, H-1Ј α/β), 7.10Ϫ7.45 (m, 5 H, Harom), 8.09 (s,
1 H, H-8), 8.11 (s, 1 H, H-2). Ϫ 13C NMR ([D6]acetone): δ ϭ 29.6
(d, J ϭ 4.5 Hz, C-3Ј), 33.17 (s, C-2Ј), 73.5 (d, J ϭ 5.4 Hz, C-5Ј),
74.9 (s, C-6Ј), 86.3 (d, J ϭ 173.6 Hz, C-6Ј), 87.1 (s, C-1Ј), 87.9 (d,
J ϭ 17.5 Hz, C-4Ј), 126.85 (s, C-5), 129.07, 129.11, and 129.88 (s,
Carom), 139.3 (d, J ϭ 3.2 Hz, C-8), 140.1 (s, C-1Љ), 146.9 (s, C-2),
158.08 (s, C-6). Ϫ NMR of β anomer: 1H NMR ([D6]acetone): δ ϭ
2.10Ϫ2.50 (m, 8 H, H2-2Ј, H2-3Ј, α/β) 3.65 (dd, J ϭ 9.8 and 1.7
Hz, 1 H, H-5Јb), 3.71 (dd, J ϭ 9.8 and 2.1 Hz, 1 H, H-5Јa), 4.49
(d, J ϭ 47.7 Hz, 2 H, H2-6Ј), 4.57 (s, 2 H, H2-7Ј), 6.39 (t, J ϭ 6.0
Hz, 1 H, H-1Ј), 7.20Ϫ7.50 (m, 5 H, Harom), 8.07 (s, 1 H, H-8), 8.13
(s, 1 H, H-2). Ϫ 13C NMR ([D6]acetone): δ ϭ 29.7 (d, J ϭ 4.1 Hz,
C-3Ј), 33.14 (s, C-2Ј), 73.3 (d, J ϭ 5.4 Hz, C-5Ј), 74.7 (s, C-6Ј), 86.7
(d, J ϭ 173.6 Hz, C-6Ј), 87.2 (s, C-1Ј), 87.8 (d, J ϭ 17.8 Hz, C-4Ј),
126.80 (s, C-5), 129.09, 129.18, and 129.85 (s, Carom), 139.5 (s, C-
8), 139.9 (s, C-1Љ), 146.8 (s, C-2), 158.06 (s, C-6). Ϫ MS (MALDI):
m/z ϭ 381.6 [M ϩ Na]ϩ, 397.5 [M ϩ K]ϩ. Ϫ IR (nujol): ν˜ ϭ
cmϪ1
, 3174.60, 3073.65, 2834.79, 2364.67, 1729.29, 1703.33,
1605.07, 1512.68, 1458.62, 1342.93, 1331.04, 1273.37, 1101.52,
820.69, 743.59. Ϫ C17H18FN3O6 (379): calcd. C 53.83, H 4.78, N
11.08; found C 53.85, H 4.79, N 11.06. Ϫ Crystallization in ethanol
afforded a mixture of α and β anomers that were carefully sepa-
20
rated by pins in pure form. Ϫ β Anomer: [α]D ϭ ϩ21.00 (c ϭ
20
0.5, CHCl3); m. p. 196°C (CH3CH2OH). Ϫ α Anomer: [α]D
Ϫ19.71 (c ϭ 0.4, CHCl3); m. p. 76°C (CH3CH2OH).
ϭ
9-[5-O-Benzyl-2,3-dideoxy-4-C-(fluoromethyl)-β/α--glycero-penta-
furanosyl]-6-chloropurine (4): Starting from 6 and 6-chloro-9-(tri-
methylsilyl)purine in CH2Cl2, a 1:1 β/α mixture (1H-NMR ratio) 1344.05, 1204.43, 1119.29, 1025.61, 697.37. Ϫ C18H19FN4O3 (358):
of monofluorobenzyl-protected 6-chloropurine analogues 4 was calcd. C 60.33, H 5.34, N 15.63; found C 60.30, H 5.32, N 15.65.
3449.92 cmϪ1
, 3049.48, 2864.59, 2138.74, 1705.07, 1586.42,
Eur. J. Org. Chem. 1999, 187Ϫ196
195

A. Mele, B. Vergani, F. Viani, S. V. Meille, A. Farina, P. Bravo
FULL PAPER
[16]
[17]
[18]
M. Barfield, M. D. Johnston, Jr., Chem. Rev. 1972, 73, 53Ϫ73.
A. M. Ihrig, S. L. Smith, J. Am. Chem. Soc. 1972, 12, 34Ϫ41.
S. Suntioinen, R. Laatikainen, Magn. Reson. Chem. 1992, 30,
415Ϫ419.
Acknowledgments
The authors wish to thank Dr. Giovanni Fronza and Alberto Ar-
none for helpful discussions. Mr. Emanuele Canu, Shimadzu Italia,
and Mr. Walter Panzeri are kindly acknowledged for technical as-
sistence in MALDI mass spectrometry.
[19]
[20]
S. Ng, J. Magn. Reson. 1978, 30, 139Ϫ140.
T. Shaeffer, G. H. Penner, R. Sebastian, J. Peeling, C. Beaulieu,
Can. J. Chem. 1991, 69, 1047Ϫ1053.
[21]
D. Neuhaus, M. Williamson, The Nuclear Overhauser Effect in
Structural and Conformational Analysis, VCH, Weinheim, 1990.
[22] [22a] J. Hilton, H. Sutcliffe, Progr. Nucl. Magn. Reson. Spectrosc.
[1]
[22b]
L. G. Jeong, H. O. Kim, J. W. Beach, W. K. Chung, M. W.
1975, 10, 27Ϫ39. Ϫ
R. H. Contreras, J. C. Facelli, Ann.
Chun, C. K. Chu, Anti-AIDS Drug Development Ϫ Challenges,
Strategies and Prospects (Eds.: P. Mohan, M. Baba), Harwood
Academic Publishers, Chur, 1995, p. 39Ϫ63.
Rep. NMR Spectrosc. 1993, 27, 255Ϫ356.
[23]
[24]
[25]
[26]
I. D. Rae, J. A. Weingold, R. H. Contreras, R. B. Biekofsky,
Magn. Reson. Chem. 1993, 31, 836Ϫ840.
[2]
P. Bravo, A. Mele, G. Salani, F. Viani, Bioorg. Med. Chem. Lett.
A. Gossauer, F. Fehr, F. Nydegger, H. Stöckli-Evans, J. Am.
Chem. Soc. 1997, 119, 1599Ϫ1608.
1994, 4, 1577Ϫ1580.
[3]
P. Bravo, A. Mele, G. Salani, F. Viani, P. La Colla, Gazz. Chim.
R. W. Taft, M. J. Kamlet, Magn. Reson. Chem. 1980, 14,
485Ϫ493.
Ital. 1995, 125, 295Ϫ303.
M. Baba, J. Balzarini, R. Pauwels, E. De Clercq, ref.[1], p.
C. Beeson, N. Pham, G. Shipps, Jr., T. A. Dix , J. Am. Chem.
Soc. 1993, 115, 6803Ϫ6812.
[4]
239Ϫ267.
[5]
[27]
[28]
P. Van Roey, C. K. Chu, Nucleosides and Nucleotides as Anti-
Parameters taken from ref.[15], p. 378.
tumor and Antiviral Agents (Eds.: C. K. Chu, D. C. Baker),
Plenum Press, New York, 1993, p. 285Ϫ302.
The pairwise correlation was reasonably linear, with satisfying
r2 factors (7 points, 0.902 and 0.921 for JHF and JCF, respec-
tively). The linearity improved dramatically if the point related
to [D4]methanol was excluded. This is not unexpected, as the
value of KamletϪTaftЈs β for that solvent is less certain than
for the others. Moreover, methanol is the only solvent, within
the chosen set, with hydrogen bond donor properties, which is
likely to introduce spurious effects which are not taken into
account for by the β parameter. The main results of the re-
7
6
[6]
D. OЈHagan, H. S. Rzepa, J. Chem. Soc., Chem. Commun.
1997, 645Ϫ652.
[7a] P. Murray-Rust, W. C. Stallings, C. T. Monti, R. K. Preston,
[7]
J. P. Glusker, J. Am. Chem. Soc. 1983, 105, 3206Ϫ3214. Ϫ
[7b]
L. H. Takahashi, R. Radhakrishnan, R. E. Rosenfield, Jr.,
E. F Meyer, D. A. Trainor, J. Am. Chem. Soc. 1989, 111,
[7c]
3368Ϫ3374. Ϫ
M. Shibakami, A. Sekiya J. Chem. Soc.,
[7d]
7
Chem. Commun. 1992, 1742Ϫ1743. Ϫ
L. Shimoni, H. L.
gression analysis are as follows: (i) JHF ϭ (Ϫ1.42 ± 0.13) β ϩ
6
Carrel, J. P. Glusker, M. M. Coombs J. Am. Chem. Soc. 1994,
116, 8162Ϫ8168. Ϫ [7e] A. Zapata, B. Bernet, A. Vasella, Helv.
Chim. Acta 1996, 79, 1169Ϫ1191. Ϫ [7f] S. J. Borwick, J. A. K.
(3.17 ± 0.08), r2 ϭ 0.966, n ϭ 6 points; (ii) JCF ϭ (Ϫ4.14 ±
0.22) β ϩ (8.93 ± 0.14), r2 ϭ 0.989, n ϭ 6 points.
SCC is defined as the difference between the through-space J
observed in the 5-Y-uracil derivative and that of the uracil nu-
cleoside (Y ϭ H) taken as reference: SCC ϭ J(Y) Ϫ J(H). The
data for the correlation are reported in Table 6. The equation
obtained for HF and CF coupling constants are, respectively:
SCS(HF) ϭ (2.21 ± 0.14) σI Ϫ (0.10 ± 0.6), r2 ϭ 0.987 and
SCS(CF) ϭ (3.17 ± 0.54) σI Ϫ (0.06 ± 0.22), r2 ϭ 0.921. The
results of these correlations can only be interpreted in a se-
miquantitative way, since the substituent set is mainly made up
of functional groups with ϪI effect.
[29]
Howard, C. W. Lehmann, D. OЈHagan, Acta. Crystallogr. 1997,
[7g]
C53, 124Ϫ126. Ϫ
M. Pham, M. Gdaniec, T. Polonski, J.
Org. Chem. 1998, 63, 3731Ϫ3734.
[8] [8a]
J. D. Dunitz, R. Taylor, Chem. Eur. J. 1997, 3, 89Ϫ98. Ϫ
[8b] J. A. K. Howard, V. J. Hoy, D. OЈHagan, G. T. Smith, Tetra-
hedron 1996, 38, 12613Ϫ12622.
[9] [9a]
H. Plenio, R. Diodone, Chem. Ber. 1997, 130, 633Ϫ640. Ϫ
H. Plenio, J. Hermann, R. Diodone, Inorg. Chem. 1997,
[9b]
36, 5722Ϫ5729.
[10]
[30]
[31]
G. R. Desiraju, Angew. Chem. Int. Ed. Engl. 1995, 34,
2311Ϫ2327.
H. Plenio, Chem. Rev. 1997, 97, 3363Ϫ3384.
J. Bromilow, R. T. C. Brownlee, V. O. Lopez, R. W. Taft, J. Org.
Chem. 1979, 44, 4766Ϫ4770.
[11]
[12]
[13]
[14]
G. R. Desiraju, J. Chem. Soc., Chem. Commun. 1989, 179Ϫ180.
G. R. Desiraju, J. Chem. Soc., Chem. Commun. 1990, 454Ϫ455.
T. Steiner, J. Chem. Soc., Chem. Commun. 1994, 2341Ϫ2342.
A. Mele, G. Salani, F. Viani, P. Bravo, Magn. Reson. Chem.
1997, 35, 168Ϫ174.
[32]
[33]
A. Altomare, G. Cascarano, G. Giacovazzo, A. Guagliardi, J.
Appl. Crystallogr. 1993, 26, 343Ϫ349.
G. M. Sheldrick, SHELXL97, Universität Göttingen, 1997.
Received July 20, 1998
[15]
C. Reichardt, Solvents and Solvent Effects in Organic Chemistry,
2nd ed., VCH, Weinheim, 1990, p. 13Ϫ17.
[O98339]
196
Eur. J. Org. Chem. 1999, 187Ϫ196