Preparation of chiral α-monofluoroalkylphosphonic acids and their evaluation as inhibitors of protein tyrosine phosphatase 1B

Article abstract of DOI:10.1039/a908086d

Enantiomerically pure α-monofluoroalkylphosphonic acids 4-9 were synthesized by diastereoselective electrophilic fluorination of α-carbanions of asymmetric phosphonamidates bearing (-)-ephedrine as a chiral auxiliary. The diastereomeric excess of the fluo

Full text of DOI:10.1039/a908086d

View Article Online / Journal Homepage / Table of Contents for this issue
Preparation of chiral ꢀ-monofluoroalkylphosphonic acids and their
evaluation as inhibitors of protein tyrosine phosphatase 1B
Christopher C. Kotoris,a Wendy Wen,a Alan Loughb and Scott D. Taylor*ac
1
a Department of Chemistry, University of Toronto, Mississauga Campus,
3359 Mississauga Rd. North, Mississauga, Ontario, Canada, L5L 1C6
b Department of Chemistry, University of Toronto, St. George Campus, 80 St. George St.,
Toronto, Ontario, Canada, M5S 1A1
c Current address: Department of Chemistry, University of Waterloo,
200 University Avenue West, Waterloo, Ontario, Canada, N2L 3G1
Received (in Cambridge, UK) 8th October 1999, Accepted 17th February 2000
Published on the Web 29th March 2000
Enantiomerically pure α-monofluoroalkylphosphonic acids 4–9 were synthesized by diastereoselective electrophilic
fluorination of α-carbanions of asymmetric phosphonamidates bearing (Ϫ)-ephedrine as a chiral auxiliary. The
diastereomeric excess of the fluorination reaction was highly dependent on the nature of the base and counterion
with de’s ranging from 2–72%. Diastereomerically pure α-fluorophosphonamidates were obtained by column
chromatography. The absolute stereochemistry of the fluorinated phosphonamidates was established by X-ray
crystallography. Removal of the ephedrine auxiliary using MeOH–TFA followed by treatment with TMSBr afforded
α-monofluoroalkylphosphonic acids 4–9 in modest to good yields. 19F-NMR analysis of the chiral phosphonic acids
4–9 in the presence of the chiral base quinidine indicated that the phosphonic acids were obtained in greater than
97% ee. Inhibition studies with 4–9 and protein tyrosine phosphatase 1B (PTP1B) revealed that the R-enantiomers
were approximately 10-fold more potent inhibitors than the corresponding S-enantiomers, but 10-fold less potent
than their α,α-difluoro analogues. Possible reasons for these differences are discussed.
Introduction
One of the most effective tactics for obtaining inhibitors of
enzymes that bind or hydrolyze phosphate groups is to replace
the phosphate group of the substrate with a non-hydrolyzable
phosphate mimetic. The α-fluoroalkylphosphonic acid moiety
is considered to be a good phosphate mimetic1 and there are
numerous reports in the literature describing inhibitors of vari-
ous enzymes that bind or hydrolyze phosphate esters that bear
this moiety.2a–c Our interest in the synthesis of α-fluoroalkyl-
phosphonates stems from our involvement in the design and
synthesis of inhibitors of a class of enzymes known as
protein tyrosine phosphatases (PTPs).3–5 PTPs are enzymes
that catalyze the removal of phosphate groups from phospho-
tyrosine residues in peptides and proteins.6 They are key regu-
lators in a wide variety of crucial kinase-dependent signal
transduction pathways. One PTP that has attracted consider-
able attention from both academia and the pharmaceutical
industry is PTP1B.6 Recently, it has been shown that PTP1B
plays a major role in modulating both insulin sensitivity and
fuel metabolism.7 Consequently, inhibitors of this enzyme may
be useful as therapeutics for the treatment of type II diabetes
and obesity.7 Among the most potent inhibitors of PTP1B
are those containing the α,α-difluoromethylenephosphonic
acid (DFMP) moiety.3,5,8–11 Certain peptides bearing difluoro-
methylenephosphonylphenylalanine (F2Pmp, 1) are nanomolar
inhibitors of PTP1B and can bind up to 2 × 103 times better
than the analogous peptide bearing methylenephosphonyl-
phenylalanine.8,9 The high affinity of the DFMP group for
PTP1B is such that even comparatively simple non-peptidyl
compounds bearing this moiety, such as 2 and 3, are relatively
good inhibitors of this enzyme while their non-fluorinated
analogues are very poor inhibitors.3,11 Why are α-fluoroalkyl-
phosphonates dramatically better inhibitors than their non-
fluorinated counterparts? Two possible explanations have been
put forward.9 One is that the PTPs require the dianionic form
of the phosphonates for binding. At pH 6–7, the non-
fluorinated phosphonates exist only partially as the dianions
while the fluorinated derivatives are almost completely ionized9
and therefore bind more tightly. However, it has been shown that
both peptidyl and non-peptidyl compounds bearing the DFMP
group inhibit PTP1B in a pH-independent manner in the pH
range spanning the pKa2 of the phosphonate moiety.3,9 This
indicates that the dianionic and monoanionic forms of these
inhibitors bind equally well and that the enhanced inhibition of
the DFMP-bearing inhibitors compared to their non-fluoro
analogues is probably not due to pKa differences. The other
explanation is that the fluorines increase the affinity by
H-bonding with specific residues in the active site. This explan-
ation is supported by data obtained from the crystal structure
of 2 complexed with PTP1B which was recently reported by
Burke and co-workers.10 On the basis of this X-ray structure and
molecular dynamics calculations, Burke has suggested that an
unusually strong fluorine hydrogen bond exists between the
pro-R fluorine of 2 and the amido group of Phe-182.10 Burke
has proposed, based on molecular dynamics calculations, that
the pro-R fluorine of 2 contributes an average of Ϫ4.6 kcal
molϪ1 more interaction energy than the pro-S fluorine when
binding to the enzyme.10 A similar interaction between PTP1B
DOI: 10.1039/a908086d
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281
This journal is © The Royal Society of Chemistry 2000
1271

View Article Online
Table 1 Effect of base and counter-ion on the electrophilic fluorin-
ation of phosphonamide 12 with NFSI
and the pro-R fluorine of an F2Pmp-bearing peptide was noted
in a recent molecular dynamics study by Tracey and Glover.12
These results suggest that enantiomerically pure (R)-α-mono-
fluoroalkylphosphonic acids may be as effective inhibitors of
PTP1B as their difluoro analogues. To test this possibility and
to gain further insight into the role of the fluorines in PTP1B
inhibition, we report here the synthesis of enantiomerically
pure α-monofluorophosphonic acids 4–9 and their evaluation
as inhibitors of human PTP1B.
Base
Yielda of 13 (%)
Deb,c (%)
n-BuLi
68d
81d
69d
71d
68d
LiHMDS
NaHMDS
KHMDS
70d, 16e
18d
37d
a Isolated yields. b De’s were determined by obtaining the 19F-NMR of
the crude residue after aqueous workup of the reaction before chrom-
atography. c Absolute stereochemistry not determined. d Performed
using 0.95 equiv. base and 1.1 equiv. NFSI as described in the Experi-
mental section. e Reverse addition: 1.1 equiv. of base was added to 12
(1 equiv.) and NFSI (1.1 equiv.) over a period of 1 h at Ϫ78 ЊC. After
stirring an additional 2 h at Ϫ78 ЊC the reaction was warmed to rt and
quenched as described in the Experimental section.
Results and discussion
To our knowledge, only two reports describing the synthesis
of enantiomerically pure α-monofluoroalkylphosphonic acids
have appeared in the literature.13,14 In both instances, the
α-monofluoroalkylphosphonic acids were obtained via stereo-
selective hydrogenation of vinyl phosphonates ((EtO) POCF᎐
᎐
2
CRRЈ).13,14 However, this procedure would not be applicable to
benzylic α-monofluoroalkylphosphonates such as 4–9. Indeed,
α-monofluoroalkylphosphonates can be prepared in modest
yield by electrophilic fluorination of α-carbanions of achiral
alkyl phosphonate esters with NFSI.28 More recently, we have
shown that racemic benzylic α-monofluoroalkylphosphonic
acids can be readily prepared in good yield via electrophilic
fluorination using NFSI.29 Consequently, we reasoned that the
approach outlined in Scheme 1 may be a viable method for the
preparation of chiral benzylic α-monofluoroalkylphosphonic
acids.
We examined trans-(R,R)-1,2-bis(N-methylamino)cyclo-
hexane (10) and (Ϫ)-ephedrine as chiral auxiliaries since these
auxiliaries have been shown to be effective for the asymmetric
synthesis of α-alkylphosphonic acids and were readily obtain-
able.17,18 The naphthyl derivatives were used as model systems.
Chiral phosphonamide 12 was prepared by reacting the phos-
phonic acid 11 with oxalyl chloride–cat. DMF followed by reac-
tion of the crude phosphoryl chloride with 10 in the presence
of two equivalents of triethylamine (Scheme 2). Our previous
a general procedure for the synthesis of enantiomerically pure
α-monofluoroalkylphosphonic acids has never been reported.
Shibuya and co-workers have attempted to prepare enantio-
merically enriched benzylic α-monofluoroalkylphosphonates
by DAST fluorination of optically active benzylic α-hydroxy-
phosphonate derivatives. However these reactions proceeded
with very low ee (3–5%).15
The general approach we have examined for the preparation
of enantiomerically enriched benzylic α-monofluoroalkyl-
phosphonic acids is outlined in Scheme 1. In this approach, the
Scheme 1
fluorine is introduced in a diastereoselective manner by electro-
philic fluorination of α-carbanions of asymmetric phosphon-
amides or phosphonamidates using a chiral amino alcohol
or diamine as an auxiliary. Removal of the chiral auxiliary
would yield the α-monofluoroalkylphosphonic acids. This
approach has been employed by a number of other groups
for the preparation of enantiomerically enriched α-alkyl-
phosphonic acids using alkyl halides as electrophiles.16–19
Differding and Lang have shown that non-racemic α-mono-
fluorocarbonyl compounds can be obtained in modest ee’s (10–
70%) by fluorination of prochiral metal enolates using optically
pure N-fluorosultams as enantioselective electrophilic fluorinat-
ing agents.20,21 Later, Davis and co-workers22–26 prepared a wide
variety of chiral α-monofluorocarbonyl compounds in good to
excellent de’s by fluorination of chiral enolates bearing Evan’s
oxazolidinone27 as a chiral auxiliary and using N-fluoro-
benzenesulfonimide (NFSI) as electrophilic fluorinating agent.
In addition, Differding and co-workers have shown that racemic
Scheme 2
studies29 and those of Differding28 on the electrophilic fluorin-
ation of phosphonates with NFSI indicated that the yields of
fluorinated products were dependent on the nature of the base
and counter-ion. Consequently, electrophilic fluorination of 12
with NFSI was examined with different bases and counter-ions.
The de of the reaction was determined by examining the crude
reaction mixture by 19F-NMR. Fluorination of 12 to give the
fluorinated phosphonamides 13 was achieved by treating 12
with 0.95 equiv. of base at Ϫ78 ЊC for 2 h followed by the
addition of 1.1 equiv. NFSI (Scheme 2). The yields and de’s of
these fluorination reactions are given in Table 1. The yield of
the fluorination reaction was not highly dependent on the
nature of the base or counter-ion with good yields being
1272
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281

View Article Online
Table 2 Effect of base and counter-ion on the electrophilic fluorin-
ation of trans isomer 19 with NFSI
Table 3 Effect of base and counter-ion on the electrophilic fluorin-
ation of cis isomer 16 with NFSI
Yielda of 28
and 31 (%)
Yielda of
22 and 25 (%)
Base
Deb (%)
Base
Deb (%)
LiHMDS
NaHMDS
KHMDS
LDA
n-BuLi
t-BuLi
46c
2 (R)
LiHMDS
NaHMDS
KHMDS
LDA
n-BuLi
t-BuLi
29c
2 (S)
75,c 68,d 62e
54,c 50,d 72e (R)
40 (R)
54,c 47d
41c
58,c 46d (S)
36 (S)
8 (S)
62c
66c
38c
31c
3 (S)
18 (S)
16 (S)
44c
33c
24 (S)
20 (S)
24c
a Isolated yields. b De’s were determined by obtaining the 19F-NMR of
the crude residue after aqueous workup of the reaction before chrom-
atography. c Performed using 0.95 equiv. base and 1.1 equiv. NFSI as
described in the Experimental section. d Same procedure as for b except
1.1 equiv. of base was used. e Reverse addition: 1.1 equiv. of base was
added to 19 (1 equiv.) and NFSI (1.1 equiv.) over a period of 1 h at
Ϫ78 ЊC. After stirring an additional 2 h at Ϫ78 ЊC the reaction was
warmed to rt and quenched as described in the Experimental section.
a Isolated yields. b De’s were determined by obtaining the 19F-NMR of
the crude residue after aqueous workup of the reaction before chrom-
atography. c Performed using 0.95 equiv. base and 1.1 equiv. NFSI as
described in the Experimental section. d Reverse addition: 1.1. equiv. of
base was added to 16 (1 equiv.) and NFSI (1.1. equiv.) over a period of 1
h at Ϫ78 ЊC. After stirring an additional 2 h at Ϫ78 ЊC the reaction was
warmed to rt and quenched as described in the Experimental section.
slight decrease in de and yield (Table 2). However, reverse
addition of NFSI and NaHMDS (1.1 equiv. of NaHMDS
added to a solution of the phosphonamidate and 1.1 equiv.
NFSI at Ϫ78 ЊC over a period of 1 h) resulted in a increase in
the de to 72% with 19 but a slight decrease in the yield (Table 2);
this was not the case with the cis-isomer 16 which resulted in a
decrease in de and yield (Table 3).
obtained in all cases. However, the de of the reaction was highly
dependent on the nature of the cation but not the base itself.
Bases with lithium counter-ions (n-BuLi and LiHMDS) gave
moderately good de’s (68–70%) while NaHMDS and KHMDS
gave low de’s. 19F-NMR analysis of the products obtained
using KHMDS as base revealed preferential formation of
the stereoisomer opposite from that obtained when using the
lithium and sodium bases. Reverse addition of NFSI and
LiHMDS (1.1 equiv. of LiHMDS added to a solution of the
phosphonamide and 1.1 equiv. NFSI at Ϫ78 ЊC over a period
of 1 h) resulted in a large decrease in the de (16%) and preferen-
tial formation of the stereoisomer opposite from that obtained
when adding the base first. The de’s obtained with the lithium
bases are comparable to those obtained by Hanessian and co-
workers (68–80%) on the asymmetric α-alkylation of benzyl
phosphonamides bearing 10 as auxiliary using alkyl halides as
electrophiles and n-BuLi as base.18 Unfortunately, separation
of the diastereomeric products 13 proved to be exceedingly
difficult. Even after several recrystallizations, diastereo-
merically pure 13 was not obtained. Attempts to separate the
two diastereomers by silica gel column chromatography were
also unsuccessful.
Reaction of the phosphonic acid dichloride derived from 11
with (Ϫ)-ephedrine yielded two diastereomeric phosphon-
amidates 16 and 19 which were readily separated by silica gel
chromatography (Scheme 3). Diastereomer 16 was designated
as the cis-isomer (phosphoryl and 3-methyl group cis) while
diastereomer 19 was designated as the trans-isomer (phosphoryl
and 3-methyl group trans).30 The fluorination of 16 and 19
(Scheme 3) was first attempted by subjecting them to the same
fluorination conditions as described above for 12. The results of
these fluorination reactions are shown in Tables 2 and 3. For
both isomers, the yield and the de of the reaction were highly
dependent on the base and cation. NaHMDS gave the highest
yields (54% for cis-isomer 16 and 75% for trans-isomer 19).
Lithium bases gave very low de’s (2–24%) for both 16 and 19.
Modest de’s were obtained with both 16 and 19 using NaHMDS
(58% and 54%) and KHMDS (36% and 40%). Although the de’s
of the reactions were modest, the diastereomeric fluorinated
phosphonamidates exhibited large differences in mobility on
silica gel, (Rf differences ranged from 0.2–0.25) and could be
readily separated by silica gel flash chromatography. We also
found that, after chromatographic separation, the fluorinated
cis-isomers 22 and 25 could be readily recrystallized and their
absolute stereochemistry determined by X-ray crystallography.
Attempts to obtain X-ray quality crystals of the trans-isomers
28 and 31 were unsuccessful. In an attempt to improve the de
of the reactions, the reaction with NaHMDS was performed
under a variety of different conditions. Fluorination with the
trans-isomer 19 using 1.1 equiv. of NaHMDS resulted in a
The m-(phenyl)benzyl phosphonamidates, 17 and 20, and
benzyl phosphonamidates, 18 and 21, were also prepared
(Scheme 3). These phosphonamidates were fluorinated using
0.95 equiv. NaHMDS and 1.1 equiv. NFSI to give the fluorin-
ated cis-isomers 23, 24, 26 and 27 and the fluorinated trans-
isomers 29, 30, 32 and 33 (Scheme 3). For each fluorination
reaction, the fluorinated diastereomeric products could be
readily separated from one another by silica gel flash chrom-
atography. The fluorination reaction of the m-(phenyl)benzyl
phosphonamidates (17 and 20) proceeded in overall good yields
(68%, cis-isomer; 85%, trans-isomer) but with low de’s (25%,
cis-isomer; 26%, trans-isomer). The fluorination reaction of the
benzyl phosphonamidates (18 and 21) also proceeded in overall
modest yields (44%, cis-isomer; 47%, trans-isomer) and low
de’s (33% cis-isomer; 29% trans-isomer). cis-Isomers 23, 24, 26
and 27 were also found to be readily recrystallized and their
structure and absolute stereochemistry were determined by
X-ray crystallography.
The next step was to develop a racemization-free procedure
for the removal of the ephedrine auxiliary from fluorinated
isomers 22–33 to obtain the desired enantiomerically pure free
acids 4–9. Since the absolute stereochemistry of all the fluorin-
ated cis-isomers 22–27 was known, our initial studies were per-
formed using these phosphonamidates. Sting and Steglich have
reported that racemization-free hydrolysis of chiral ephedrine
α-alkylphosphonamidates can be accomplished by subjecting
them to concentrated HCl at 110 ЊC for 20 h.17 However, apply-
ing these conditions to 22 and 25 resulted in a mixture of com-
pounds and we were unable to purify 4 and 5 from the mixture.
Nevertheless, we found that the auxiliary could be removed
without racemization using a modified version of a procedure
developed by Calvo (Scheme 3).31 This involved treating 22–27
with 10% TFA–MeOH, followed by reaction with TMSBr (10
equiv.). The contaminating ammonium salts were removed
using anhydrous ether followed by filtration. Hydrolysis of the
TMS ester in methanol–H2O gave the free acids 4–9 in modest
to good yields (39–70%). We found that the enantiomeric purity
of 4–9 could be readily determined by preparing solutions of
4–9 and 1 equiv. of the chiral base quinidine in CDCl3 followed
by 19F-NMR analysis of the salt solutions. For example, the
19F-NMR spectra of the salts derived from phosphonic acids 4
and 5 consist of a single set of doublet of doublets with each set
having a chemical shift significantly different from the other set
(Fig. 1a and 1b). The quinidine salts of a racemic mixture of
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281
1273

View Article Online
Scheme 3
4 and 5 appear as two distinct sets of doublet of doublets
the base and counter-ion. When HMDS bases were used,
(Fig. 1c). These results indicate that removal of the ephedrine
auxiliary proceeded without racemization and both 4 and 5
were obtained in high enantiomeric purity (>97% ee). Similar
results were obtained with acids 6–9. The ephedrine auxiliary
was also removed from the fluorinated trans-isomers 28–33 to
give acids 4–9 without racemization and in modest to good
yields using the procedure described above (Scheme 3).
Since the absolute configuration of the phosphonic acids
derived from cis-isomers 22–27 was known, it was then possible
to determine the absolute configuration of the phosphonic
acids derived from the trans-isomers 28–33 by analysis of their
19F-NMR spectrum in the presence of one equiv. of quinidine.
From the results of these studies, the absolute stereochemistry
of the trans-fluorophosphonamidates 28–33 was also deter-
mined. The fluorination of the trans-isomers 19–21 using
NaHMDS as base resulted in preferential formation of the R-
enantiomers (28–30). The fluorination studies with the
model trans-phosphonamidate 19 using a variety of different
bases (Table 2) demonstrated that the stereochemical outcome
of the fluorination reaction is dependent upon the nature of
the reaction proceeded with preferential formation of the
R-enantiomer 28; this preference was almost negligible when
lithium was the counter-ion. When other lithium bases were
used, the S-enantiomer 31 was formed preferentially although
the de’s were very low (3–18%, Table 2). In contrast to the
fluorination reactions with the trans-isomers, the fluorination
reactions with the cis-isomers 16–18 using NaHMDS
proceeded with preferential formation of the S-enantiomers.
Studies with the model cis-isomer 16 showed that this was the
case regardless of the base and counter-ion, although with the
lithium bases this preference was small to almost negligible
(Table 3). The change or increase in stereoselectivity between
lithium and sodium or potassium bases may be due to a number
of factors such as differences in the degree of complexation of
the cations with the phosphoryl and other heteroatoms of the
phosphonamidate, differences in the degree of complexation of
the cation with the solvent, differences in size of the cation, etc.
Further studies are necessary to ascertain the origin of the
effect of cation on diastereoselectivity.
Inhibition studies with phosphonic acids 2–9 and PTP1B
1274
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281

View Article Online
Table 4 Inhibition studies with phosphonic acids 2–7 and PTP1B
Phosphonic
acid
IC50/µMa
71
675 40
7500 1000
Ki/µMa
34
345 24
ND
2
4
5
3
6
7
6
4
33
4
16
3
315 20
3500 500
158 12
ND
a Errors are reported as standard deviations.
Waals interactions with the phenyl ring of Phe-182.10 Never-
theless, it is possible that this interaction is necessary in order
for the difluoro inhibitors to be correctly positioned in the
active site so that optimal interactions can be attained between
the enzyme and other functionalities on the inhibitor. This may
include the formation of an optimal fluorine H-bond between
the pro-R fluorine and the N–H of Phe-182, hydrophobic inter-
actions between the aryl rings of the inhibitors and the side
chains of certain amino acids such as Tyr-46, Phe-182, Ala-217
and Ile-219, and electrostatic binding of the phosphate group
to the positively charged phosphate binding site.10 It is also
possible that the strength of a FC–F ؒ ؒ ؒ H–N hydrogen bond
may be greater than a HC–F ؒ ؒ ؒ H–N hydrogen bond. How-
ever, we are not aware of any experimental studies supporting
such an argument.
S-Enantiomers 5 and 7 are approximately 105-fold poorer
inhibitors than their difluoro compounds 2 and 3. Thus, the
substitution of the pro-R fluorine atoms in 2 and 3 with a
hydrogen atom results in an approximately 105-fold decrease in
affinity for PTP1B. On the basis of the X-ray structure of the
PTP1B–2 complex and molecular dynamics calculations, Burke
and co-workers have suggested that the H-bond between the
pro-R fluorine of 2 and the backbone N–H of Phe-182 may
contribute as much as Ϫ4.6 kcal molϪ1 to the binding process.10
However, the 105-fold difference in affinity between 2 and 5
reported here corresponds to a difference of Ϫ2.75 kcal molϪ1
in binding energy which suggests that Ϫ4.6 kcal molϪ1 may be
an overestimation of the strength of this H-bond. The subject
of fluorine hydrogen bonds involving C–F is a subject of much
controversy.32–34 Nevertheless, there is little doubt now that such
H-bonds can form in certain instances35 although the optimal
strength of such bonds is still unknown. To our knowledge,
studies estimating the optimal strength of C–F ؒ ؒ ؒ H–N
hydrogen bonds have not been reported. However, ab initio
calculations by O’Hagan and co-workers predict that the
strength of an optimal C–F ؒ ؒ ؒ H–O hydrogen bond in a
Fig. 1 19F-NMR spectra of (a) quinidine salt of phosphonic acid 4,
(b) quinidine salt of phosphonic acid 5, (c) quinidine salt of a racemic
mixture of phosphonic acids 4 and 5.
HO–H ؒ ؒ ؒ F–CH3 complex to be 2.4 kcal mol .
Ϫ1 32 This value is
close to the value (Ϫ2.7 kcal molϪ1) that we have obtained for
the difference in binding energy between 2 and 5.
were performed at pH 6.5 in BIS-TRIS buffer–5% DMSO. The
benzylphosphonic acids 8 and 9 were not inhibitors of the
enzyme even at concentrations as high as 1.0 mM. The results
of the inhibition studies with 4–7 and their difluoro analogues
are given in Table 4. Compounds 4–7 were found to be competi-
tive inhibitors of PTP1B. The R-enantiomers 4 and 6 are not as
effective inhibitors as their difluoro analogues 2 and 3, being
9.5-fold less potent inhibitors than compounds 2 and 3. Thus,
substitution of the pro-S fluorine with hydrogen in 2 or 3 results
in a 9.5-fold decrease in affinity for the enzyme. These results
indicate that the pro-S fluorine has some role in enhancing the
affinity of the difluoro inhibitors and support the argument3,9
that the fluorines do not enhance binding by reducing the pKa
of the phosphonic acid moiety. It is possible that this may be
due to a direct contribution of the pro-S fluorine to binding
involving specific interactions of the pro-S fluorine with resi-
dues in the active site. However, Burke’s analysis of the PTP1B–
2 complex did not uncover any significant role that the pro-S
fluorine may have in binding besides the formation of van der
In conclusion, α-monofluoroalkylphosphonic acids 4–9 were
synthesized in high enantiomeric purity. This was accomplished
via the electrophilic fluorination of α-carbanions of asymmetric
phosphonamidates bearing (Ϫ)-ephedrine as a chiral auxiliary,
followed by chromatographic separation of the resulting
diastereomers and removal of the auxiliary. The synthetic
methodology outlined here should be applicable to the syn-
thesis of chiral α-monofluoroalkylphosphonic acid inhibitors
of other enzymes that bind or hydrolyze phosphate esters. We
are currently investigating other chiral auxiliaries in attempts
to improve the de’s of the fluorination reactions. Inhibition
studies with phosphonic acids 4–7 and PTP1B indicated that
the pro-S fluorine in difluoro inhibitors 2 and 3 is essential for
good inhibition. Moreover, our inhibition studies demonstrate
that the pro-R fluorine in the difluoro inhibitors contributes
significantly more than the pro-S fluorine towards PTP1B affin-
ity, however, the contribution of the pro-R fluorine to binding
may have been overestimated by previous workers.10
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281
1275

View Article Online
2925 (br), 2285 (br), 1597, 1506, 1407, 1281, 1166, 1003, 950,
Experimental
General
826 and 744; δH(200 MHz; CD3OD) 7.82–7.78 (4 H, br m),
7.46–7.44 (3 H, br m) and 3.29 (2 H, d, JHP 20.5, CH2P); δP(80
MHz; CD3OD) 26.3 (1P, s); δC(100 MHz; CD3OD) 134.9 (d),
133.6 (d), 131.8 (d), 129.4 (d), 129.2 (d), 128.8 (d), 128.5 (d),
127.0, 126.6 and 36.0 (d, JCP 134.8, CH2P); m/z (ES) 221 (100%).
Unless otherwise noted, all reagents for syntheses were
obtained from commercial suppliers (Aldrich, Milwaukee,
Wisconsin, USA or Lancaster Synthesis Inc. Windham, New
Hampshire, USA) and were used without further purification.
Tetrahydrofuran (THF) and diethyl ether (ether) were distilled
from sodium–benzophenone ketyl under argon. Dichloro-
methane and toluene were distilled from calcium hydride under
argon. DMF was distilled under reduced pressure from calcium
hydride and stored over 4 Å sieves under argon. Reactions
involving moisture-sensitive reagents were executed under an
inert atmosphere of dry argon or nitrogen. Flash chroma-
tography was performed using silica gel 60 (Toronto Research
Chemicals, 230–400 mesh ASTM). 1H-, 31P- and 19F-NMR
spectra were recorded on a Varian 200-Gemini NMR spec-
trometer. Unless stated otherwise, all 13C spectra were recorded
on a Varian-400 instrument. For 1H-NMR spectra run in
CDCl3, chemical shifts (δ) are reported in parts per million
relative to the internal standard tetramethylsilane (TMS). For
1H spectra run in CD3OD, chemical shifts (δ) are reported in
parts per million relative to the residual CH3 peak at δ 3.30
ppm. For 13C spectra run in CDCl3, chemical shifts are reported
in parts per million relative to the CDCl3 residual carbons
(δ 77.0 for the central peak). For 13C spectra run in CD3OD,
chemical shifts are reported in parts per million relative to the
CD3OD carbon (δ 49.0 for the central peak). All 31P-NMR
spectra were proton decoupled and chemical shifts are reported
in parts per million relative to 85% phosphoric acid (external).
For 19F-NMR, chemical shifts are reported in parts per million
relative to trifluoroacetic acid (external). All NMR couplings
are given in Hz. Electron impact (EI) were obtained on a
Micromass 70-S-250 mass spectrometer. Electrospray mass
spectra were obtained using a Micromass Platform mass
spectrometer. IR spectra were recorded on an Avatar (Nicolet)
360 FTIR spectrophotometer. All melting points were taken on
a Fisher–Johns melting point apparatus, and are uncorrected.
Analytical HPLC was performed using a Waters LC 4000
System equipped with a Vydac 218TP54 analytical C-18
reversed phase column and a Waters 486 tunable absorbance
detector set at 254 nM. Specific rotations were determined
using a Perkin-Elmer 243B polarimeter. [α]D values are given in
10Ϫ1 deg cm2 gϪ1. Buffer chemicals and bovine serum albumin
(BSA) were obtained from Sigma Chemical Company. Enzyme
assay solutions were prepared with deionized/distilled water.
Fluorescein diphosphate (FDP) and human PTP1B were a gift
from Merck-Frosst Canada Inc (Montreal, Canada).
Biphenyl-3-ylmethylphosphonic acid 14. White solid (79%
yield); mp 172–174 ЊC (from H2O) (Found: C, 62.7; H, 5.2.
C13H13O3P requires C, 62.9; H, 5.3%); νmax(KBr)/cmϪ1 2921 (br),
2312 (br), 1601, 1574, 1478, 1422, 1166, 1124, 1027, 939, 775
and699;δH(200MHz;CD3OD)7.61–7.30(9H, m)and3.18(2H,
d, JHP 22.0, CH2P); δP(80 MHz; CD3OD) 26.3 (1P, s); δC(100
MHz; CD3OD) 142.5 (d), 142.2, 134.8 (d), 129.9, 129.8, 129.6
(d), 128.3, 128.0, 127.8, 126.2 (d) and 35.8 (d, JCP 134.7, CH2P);
m/z (ES) 247 (100%).
Benzylphosphonic acid 15. White solid (99% yield); mp 171–
172 ЊC (from H2O) (lit.,38 169–171 ЊC); νmax(KBr)/cmϪ1 2865,
2359, 1496, 1457, 1262, 1217, 1075, 992, 943, 783 and 693;
δH(200 MHz; CD3OD) 7.32–7.25 (5 H, m), 5.19 (2 H, s) and 3.11
(2 H, d, JHP 22.0, CH2P); δP(80 MHz; CD3OD) 26.2 (1 P, s);
δC(100 MHz; CD3OD) 134.4 (d), 131.0 (d), 129.4, 127.6 and
36.0 (d, JCP 135.4, CH2P); m/z (ES) 171 (100%).
(3aR,7aR)-2-(2-Naphthylmethyl)-2,3,3a,4,5,6,7,7a-octahydro-
1,3-dimethyl-1,3,2ꢁ5-benzodiazaphosphol-2(1H)-one 12
To a suspension of phosphonic acid 11 (2.8 g, 12.7 mmol) in
anhydrous CH2Cl2 (25 cm3) and a catalytic quantity of DMF
was added oxalyl chloride (4.83 g, 38.1 mmol) over a period of
several minutes. The reaction was stirred for 4 h during which
time the reaction mixture became a clear pale yellow solution.
The solution was concentrated by rotary evaporation and
placed under high vacuum overnight during which time
the solution solidified yielding the crude phosphonic acid
dichloride as a pale yellow solid. To a solution of the crude
phosphonic acid dichloride in anhydrous CH2Cl2 (30 cm3) was
added a solution of trans-(1R,2R)-N,NЈ-dimethyl-1,2-diamino-
cyclohexane 1019 (1.80 g, 12.7 mmol) and Et3N (3.57 cm3, 25.6
mmol) in anhydrous CH2Cl2 (30 cm3) dropwise over a period of
30 minutes under an atmosphere of nitrogen. After stirring for
12 h, the reaction mixture was filtered over a short pad of Celite
and washed with EtOAc (100 cm3). The crude residue was puri-
fied by silica gel column chromatography using MeOH–EtOAc
(2:98) as eluent to give pure 12 as a white solid (2.58 g, 62%).
Mp 106–108 ЊC; νmax(KBr)/cmϪ1 3058, 2933, 2814, 2657, 1980,
1782, 1629, 1599, 1503, 1467, 1337, 1315, 1057 and 942; δH(200
MHz; CDCl3) 7.83–7.71 (4 H, m), 7.50–7.40 (3 H, m), 3.58–3.14
(2 H, m, CH2P), 2.72–2.66 (1 H, m, CHN), 2.60 (3 H, d, J 10.3,
NCH3), 2.42 (3 H, d, J 11.7, NCH3), 2.06–1.65 (5 H, m) and
1.29–0.89 (4 H, m); δP(80 MHz; CDCl3) 75.5 (1 P, s); δC(100
MHz; CDCl3) 133.4 (br s), 132.2 (br s), 130.7 (d), 128.5 (br m),
127.6 (br s), 127.4, 126.0, 125.4, 64.6 (d, JCP 5.5), 64.0 (d, JCP
7.3), 35.1 (d, JCP 107.0, CH2P), 29.5, 28.5 (d, JCP 9.2), 28.1 (m)
and 24.2 (d, JCP 7.3); m/z (EI) 328 (Mϩ, 8%), 187 (35), 84 (100).
Found: [Mϩ], 328.1702. C19H25OPN2 requires m/z 328.1705.
General procedure for the preparation of phosphonic acids 11, 14
and 15
Trimethylsilyl bromide (TMSBr, 5 equiv.) was added to a solu-
tion of one of naphthalen-2-ylmethylphosphonic acid methyl
ester,36 biphenyl-3-ylmethylphosphonic acid methyl ester29
or benzylphosphonic acid diethyl ester (Aldrich, Milwaukee,
Wisconsin, USA) in anhydrous CH2Cl2 (approximately 5–10
cm3 CH2Cl2/mmol of ester). For the methyl esters, the solution
was stirred under an atmosphere of N2 at room temperature for
12 h. For the ethyl ester, the solution was refluxed under an
atmosphere of N2 for 24 h. The mixture was then concentrated
and placed under high vacuum for 2 h. The product was then
re-dissolved in CH2Cl2 and water (approximately 5–10 cm3
H2O/mmol of phosphonate) was added. The mixture was
stirredvigorouslyfor30minduringwhichtimeawhiteprecipitate
formed. The precipitate was then filtered, washed extensively
with CH2Cl2 and placed under high vacuum until dry.
General procedure for the fluorination of 12 to give diastereo-
meric (3aR,7aR)-2-[1-fluoro(2-naphthyl)methyl]-2,3,3a,4,5,6,
7,7a-octahydro-1,3-dimethyl-1,3,2ꢁ5-benzodiazaphosphol-
2(1H)-one, 13
To a solution of phosphonamide 12 in anhydrous THF
(approximately 5–10 cm3 THF/mmol of phosphonamide) at
Ϫ78 ЊC was added base (0.95 equiv.) over a period of 2 minutes.
The resulting orange solution was stirred for 2 h at Ϫ78 ЊC. A
solution of NFSI (1.1 equiv.) in anhydrous THF (approxi-
mately 2–4 cm3 THF/mmol NFSI) was added over a period of
2 minutes, during which time the solution turned from orange
to yellow–brown. After addition, the solution was stirred for 2 h
at Ϫ78 ЊC, during which time a precipitate sometimes formed.
2-naphthylmethylphosphonic acid 11. White solid (97% yield);
mp 225–227 ЊC (from H2O) (lit.,37 229–230 ЊC); νmax(KBr)/cmϪ1
1276
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281

View Article Online
The reaction was quenched with 10% NH4Cl and the resulting
solution (precipitate dissolves) was extracted with CHCl3.
The combined organic layers were washed with brine, dried
(MgSO4) and concentrated in vacuo to give a yellow residue.
A reverse addition procedure was also examined in which
1.1 equiv. of LiHMDS was added to 12 (1 equiv.) and NFSI
(1.1 equiv.) over a period of 1 h at Ϫ78 ЊC. After stirring an
additional 2 h at Ϫ78 ЊC the reaction was warmed to rt and
quenched as described above. Column chromatography of the
crude residue gave 13 as a mixture of two diastereomers. For
yields and de’s see Table 1. νmax(KBr)/cmϪ1 3058, 2936, 2863,
2817, 1663, 1601, 1506, 1446, 1368, 1260, 1211, 1170, 1059 and
922; δH(200 MHz; CD3OD) 7.96–7.86 (4 H, m), 7.64–7.47 (3 H,
m), 6.18 (1 H, dd, JHP 404 and JHF 45.8, CHF), 5.89 (1 H, dd,
JHP 3.7 and JHF 45.4, CHF), 2.82–2.42 (5 H, m), 2.12–1.71
(7 H, m) and 1.32–0.99 (4 H, m); δP(80 MHz; CD3OD) 37.7 (d,
JPF 80.9) and 33.7 (d, JPF 82.4); δF(188 MHz; CD3OD) Ϫ113.4
(1 F, dd, JFH 45.8 and JFP 82.4) and Ϫ120.9 (1 F, dd, JFH 45.4
and JFP 80.9); δC(100 MHz; CD3OD) 134.9 (d), 134.4, 132.9 (br
dd), 129.1 (br m), 128.0 (br m), 126.6 (br dd), 126.0 (br t), 124.9
(br m), 98.0–87.5 (m, CHF), 66.3 (d, JCP 6.4), 66.1 (d, JCP 4.6),
65.7 (br s), 65.5 (d, JCP 2.8), 64.9 (d, JCP 7.3), 30.8, 29.8–29.3 (m)
and 25.3 (br m); m/z (EI) 346 (Mϩ, 6%), 328 (10), 279 (6), 187
(100) and 153 (12). Found: [Mϩ], 346.1612. C19H24OPN2F
requires m/z 346.1610.
(c 0.0239 in MeOH) (Found: C, 71.7; H, 6.3; N, 4.0. C21H22-
NO2P requires C, 71.8; H, 6.3; N, 4.0%); νmax(KBr)/cmϪ1 3058,
2960, 2925, 1508, 1452, 1243, 1234, 1189, 980, 856, 827, 752 and
701; δH(200 MHz; CDCl3) 7.82–7.78 (4 H, br m), 7.50–7.38
(3 H, br m), 7.22 (3 H, br s), 7.08 (2 H, br s), 5.70 (1 H, d, J 5.9,
HCO), 3.59 (2 H, d, JHP 20.5, CH2P), 3.54–3.44 (1 H, m, HCN),
2.70 (3 H, d, J 8.8, NCH3) and 0.10 (3 H, d, J 7.3, CCH3); δP(80
MHz; CDCl3) 36.8 (1 P, s); δC(100 MHz; CDCl3) 135.4 (d),
133.0 (d), 131.9 (d), 129.7 (d), 128.3 (d), 127.9 (br m), 127.8,
127.6, 127.3 (br d), 127.1, 126.0, 125.4, 125.3, 79.8 (PhCO), 60.6
(d, JCP 8.1, CH3CN), 35.4 (d, JCP 122.3, CH2P), 29.8 (d, JCP 6.6,
NCH3) and 13.3 (CH3C); m/z (EI) 351 (Mϩ, 50%), 294 (17),
210 (41), 141 (62), 104 (100), 84 (20). Found: [M]ϩ, 351.1378.
C21H22O2PN requires m/z 351.1388.
(2R,4S,5R)-2-(m-Phenylbenzyl)-3,4-dimethyl-2-oxo-5-phenyl-
1,3,2ꢁ5-oxazaphospholidine (cis-isomer) 17. Colourless oil
(24%); [α]D22 Ϫ47.1 (c 0.0133 in MeOH); νmax(KBr)/cmϪ1 3029,
2972, 2905, 1599, 1479, 1454, 1297, 1060, 974, 886 and 701;
δH(200 MHz; CDCl3) 7.60–7.23 (9 H, m), 4.78 (1 H, t, J 5.9,
HCO), 3.51 (2 H, d, JHP 20.5, CH2P), 3.22–3.09 (1 H, m, HCN),
2.70 (3 H, d, J 10.3, NCH3) and 0.70 (3 H, d, J 7.4, CCH3);
δP(80 MHz; CDCl3) 38.2 (1 P, s); δC(100 MHz; CDCl3) 141.4 (d),
140.5, 136.1 (d), 133.0 (d), 129.0 (d), 128.8, 128.5 (d), 128.3 (d),
128.2, 128.1, 127.4, 127.0, 126.1, 125.6 (d), 82.5 (PhCO), 58.4
(d, JCP 9.5, CH3CN), 34.6 (d, JCP 119.3, CH2P), 28.5 (d, JCP 5.9,
NCH3) and 14.4 (CH3C); m/z (EI) 377 (Mϩ, 60%), 362 (19), 210
(28), 167 (17), 104 (100). Found: [M]ϩ, 377.1545. C23H24O2PN
requires m/z 377.1545.
General procedure for the preparation of phosphonamidates
16–21
A mixture of (1R,2S)-(Ϫ)-ephedrine (1 equiv.) and Et3N
(2 equiv.) dissolved in anhydrous toluene (approximately 1.2
cm3 toluene/mmol (Ϫ)-ephedrine) was added dropwise to a
suspension of the crude phosphonic acid dichloride (1 equiv.,
prepared from phosphonic acids 11, 14 and 15 using the
procedure described above during the preparation of 12) in
anhydrous toluene (approximately 2 cm3 toluene/mmol phos-
phonic acid dichloride) at Ϫ40 ЊC, under an atmosphere of
nitrogen. After stirring for 1 h at Ϫ40 ЊC, the reaction mixture
was stirred at room temperature overnight. The resulting yellow
slurry (containing a white precipitate) was diluted with EtOAc
(100 cm3), washed with brine, dried (MgSO4), and concentrated
in vacuo. Separation and purification of the trans- and cis-
phosphonamidates 16–21 were achieved by silica gel column
chromatography using MeOH–EtOAc (2:98) as eluent for 16,
17, 19 and 20 and MeOH–EtOAc (3:97) as eluent for 18 and
21. The cis-isomers 16–18 exhibited larger Rf values on silica
TLC plates and eluted before the trans-isomers 19–21 during
column chromatography.
(2S,4S,5R)-2-(m-Phenylbenzyl)-3,4-dimethyl-2-oxo-5-phenyl-
1,3,2ꢁ5-oxazaphospholidine (trans-isomer) 20. White solid
(56%); mp 120–122 ЊC (from pentane–CH2Cl2); [α]D22 Ϫ58.7 (c
0.0267 in MeOH) (Found: C, 73.4; H, 6.4; N, 3.7. C23H24NO2P
requires C, 73.2; H, 6.4; N, 3.7%); νmax(KBr)/cmϪ1 3021, 2971,
2909, 1597, 1479, 1452, 1252, 1213, 1064, 972, 887 and 703;
δH(200 MHz; CDCl3) 7.59–7.11 (9 H, m), 5.69 (1 H, d, J 5.9,
HCO), 3.51 (2 H, d, JCP 19.0, CH2P), 3.55–3.45 (1 H, m, HCN),
2.74 (3 H, d, J 8.8, NCH3) and 0.08 (3 H, d, J 7.3, CCH3); δP(80
MHz; CDCl3) 37.0 (1 P, s); δC(100 MHz; CDCl3) 141.2 (d),
140.3, 135.4 (d), 132.8 (d), 128.7 (m), 128.6, 127.9, 127.7, 127.2,
126.7, 125.4, 79.8 (PhCO), 60.9 (d, JCP 7.4, CH3CN), 35.0 (d,
JCP 122.3, CH2P), 29.7 (d, JCP 6.6, NCH3) and 13.2 (CH3C); m/z
(EI) 377 (Mϩ, 100%), 362 (30), 210 (29), 167 (25), 104 (85).
Found: [M]ϩ, 377.1549. C23H24O2PN requires m/z 377.1545.
(2R,4S,5R)-2-Benzyl-3,4-dimethyl-2-oxo-5-phenyl-1,3,2ꢁ5-
oxazaphospholidine (cis-isomer) 18. White solid (20%); mp 133–
134 ЊC (from pentane–CH2Cl2); [α]D22 Ϫ40.2 (c 0.0438 in MeOH)
(Found: C, 67.4; H, 6.7; N, 4.6. C17H20NO2P requires C, 67.8;
H, 6.7; N, 4.65%); νmax(KBr)/cmϪ1 3029, 2943, 2874, 1954, 1881,
1602, 1494, 1452, 1240, 1059, 982, 897 and 699; δH(200 MHz;
CDCl3) 7.36–7.23 (10 H, m), 4.67 (1 H, t, J 5.9, HCO), 3.44
(2 H, d, JHP 19.0, CH2P), 3.17–3.03 (1 H, m, HCN), 2.68 (3 H,
d, J 10.3, NCH3) and 0.68 (3 H, d, J 5.9, CCH3); δP(80 MHz;
CDCl3) 38.4 (1 P, s); δC(100 MHz; CDCl3) 136.5 (br s), 132.8
(d), 129.7 (d), 128.6, 128.3, 128.1, 126.9 (br s), 126.3, 82.5
(PhCO), 58.6 (d, JCP 9.1, CH3CN), 34.8 (d, JCP 119.8, CH2P),
28.5 (d, JCP 5.5, NCH3) and 14.4 (CH3C); m/z (EI) 301 (Mϩ,
26%), 147 (24), 104 (100), 91 (39), 84 (47). Found: [M]ϩ,
301.1233. C17H20O2PN requires m/z 301.1232.
(2R,4S,5R)-2-(2-Naphthylmethyl)-3,4-dimethyl-2-oxo-5-
phenyl-1,3,2ꢁ5-oxazaphospholidine (cis-isomer) 16. White solid
(24%); mp 180–182 ЊC (from pentane–CH2Cl2); [α]D22 Ϫ73.2 (c
0.0284 in MeOH) (Found: C, 71.5; H, 6.4; N, 4.0. C21H22NO2P
requires C, 71.8; H, 6.3; N, 4.0%); νmax(KBr)/cmϪ1 3050, 2976,
2907, 1599, 1453, 1264, 1246, 1197, 986, 884, 862, 748 and 702;
δH(200 MHz; CDCl3) 7.82–7.74 (4 H, br m), 7.49–7.40 (3 H, br
m), 7.30–7.19 (5 H, br m), 4.74 (1 H, br t, J 5.9, HCO), 3.60
(2 H, d, JHP 20.5, CH2P), 3.10–3.00 (1 H, m, HCN), 2.67 (3 H,
d, J 8.8, NCH3) and 0.64 (3 H, d, J 5.9, CCH3); δP(80 MHz;
CDCl3) 38.4 (1 P, s); δC(100 MHz; CDCl3) 135.9 (d), 133.1 (d),
131.8 (d), 129.7 (d), 127.9 (br m), 127.8, 127.5 (d), 127.4, 127.2,
126.0, 125.9, 125.6 (br d), 82.2 (PhCO), 58.1 (d, JCP 9.6,
CH3CN), 34.4 (d, JCP 119.3, CH2P), 28.3 (d, JCP 5.8, NCH3) and
14.2 (CH3C); m/z (EI) 351 (Mϩ, 70%), 294 (21), 210 (46), 141
(64), 104 (100), 84 (85). Found: [M]ϩ, 351.1384. C21H22O2PN
requires m/z 351.1388.
(2S,4S,5R)-2-Benzyl-3,4-dimethyl-2-oxo-5-phenyl-1,3,2ꢁ5-
oxazaphospholidine (trans-isomer) 21. White solid (41%); mp
169–171 ЊC (from pentane–CH2Cl2); [α]D22 Ϫ6.2 (c 0.0488 in
MeOH) (Found: C, 67.9; H, 6.8; N, 4.7. C17H20NO2P requires
C, 67.8; H, 6.7; N, 4.65); νmax(KBr)/cmϪ1 3031, 2925, 2818,
1966, 1901, 1818, 1601, 1494, 1455, 1239, 1064, 980 and 701;
δH(200 MHz; CDCl3) 7.30–7.09 (10 H, m), 5.69 (1 H, d, J 5.8,
HCO), 3.58–3.42 (1 H, m, HCN), 3.42 (2 H, d, JHP 22.0, CH2P),
(2S,4S,5R)-2-(2-Naphthylmethyl)-3,4-dimethyl-2-oxo-5-
phenyl-1,3,2ꢁ5-oxazaphospholidine (trans-isomer) 19. White
solid (54%); mp 132–134 ЊC (from pentane–CH2Cl2); [α]D22 ϩ1.1
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281
1277

View Article Online
2.69 (3 H, d, J 8.7, NCH3) and 0.15 (3 H, d, J 7.3, CCH3); δP(80
MHz; CDCl3) 37.0 (1 P, s); δC(100 MHz; CDCl3) 136.1 (d),
132.7 (d), 130.1 (d), 128.6 (d), 128.3, 128.0, 126.9 (d), 125.8,
80.2 (PhCO), 61.3 (d, JCP 7.4, CH3CN), 35.8 (d, JCP 123.6,
CH2P), 30.2 (d, JCP 5.4, NCH3) and 13.7 (CH3C); m/z (EI) 301
(Mϩ, 19%), 147 (46), 104 (79), 97 (100), 91 (46), 84 (93). Found:
[M]ϩ, 301.1239. C17H20O2PN requires m/z 301.1232.
(2R,4S,5R)-2-[(1S)-ꢀ-Fluoro(m-phenyl)benzyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 23. White solid
(43%); mp 126–128 ЊC (from pentane–CH2Cl2); [α]D22 Ϫ85.3 (c
0.0252 in MeOH) (Found: C, 69.4; H, 5.8; N, 3.5. C23H23FNO2P
requires C, 69.9; H, 5.9; N, 3.55%); νmax(KBr)/cmϪ1 3030, 2965,
2919, 2840, 1957, 1887, 1596, 1453, 1263, 1216, 1185, 1040, 992,
856 and 703; δH(200 MHz; CDCl3) 7.83 (1 H, s), 7.66–7.30 (8 H,
m), 6.12 (1 H, dd, JHP 8.8 and JHF 45.8, CHF), 5.69 (1 H, d,
J 5.9, HCO), 3.80–3.66 (1 H, m, HCN), 2.31 (3 H, d, J 10.2,
NCH3) and 0.81 (3 H, d, J 5.9, CCH3); δP(80 MHz; CDCl3) 28.9
(1 P, d, JPF 76.3); δF(188 MHz; CDCl3) Ϫ125.8 (1 F, dd, JFH 45.8
and JFP 76.3); δC(100 MHz; CDCl3) 141.4, 140.4, 135.8 (d),
134.1 (d), 128.9, 128.8, 128.4, 128.2, 127.6, 127.2, 127.1, 125.8,
124.3 (dd), 124.0 (dd), 89.7 (dd, JCP 149.8 and JCF 189.3, CHF),
82.5 (PhCO), 58.7 (d, JCP 9.6, CH3CN), 29.5 (d, JCP 4.4, NCH3)
and 14.1 (CH3C); m/z (EI) 395 (Mϩ, 45%), 210 (71), 185 (60),
146 (42), 104 (100). Found: [M]ϩ, 395.1464. C23H23O2PNF
requires m/z 395.1460.
General procedure for the preparation of fluorophosphon-
amidates 22–33
The same procedure as that described for the fluorination of 12
was used except phosphonamidates 16–21 were used as sub-
strates. For the fluorination of phosphonamidates 17, 18, 20
and 21, NaHMDS was used as base. The fluorination of 16 and
19 was examined in detail using a variety of bases, counter-ions
and conditions. For the conditions and results of these studies
see Tables 2 and 3. Separation and purification of the fluorin-
ated cis-isomers 22–27 was achieved by silica gel column
chromatography using EtOAc–hexane (1:1) as eluent. In the
case of the cis-isomers 22–27, the S-enantiomers (stereo-
chemistry at the α-carbon) exhibited larger Rf values on silica
TLC plates and eluted before the R-enantiomers during column
chromatography. Separation and purification of the fluorinated
trans-isomers 28–33 was achieved by silica gel column chrom-
atography using EtOAc–hexane (7:3) as eluent. In the case of
the trans-isomers 28–33, the R-enantiomers (stereochemistry
at the α-carbon) exhibited larger Rf values on silica TLC
plates and eluted before the S-enantiomers during column
chromatography.
(2R,4S,5R)-2-[(1R)-ꢀ-Fluoro(m-phenyl)benzyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 26. White solid
(25%); mp 147–149 ЊC (from pentane–CH2Cl2); [α]D22 ϩ28.5 (c
0.0215 in MeOH) (Found: C, 69.6; H, 5.9; N, 3.5. C23H23FNO2P
requires C, 69.9; H, 5.9; N, 3.55%); νmax(KBr)/cmϪ1 3036, 2970,
2866, 1950, 1886, 1597, 1476, 1258, 1212, 1044, 960, 862 and
700; δH(200 MHz; CDCl3) 7.75 (1 H, s), 7.66–7.34 (8 H, m), 6.06
(1 H, dd, JHP 4.4 and JHF 45.0, CHF), 5.44 (1 H, t, J 6.6, HCO),
3.83–3.72 (1 H, m, HCN), 2.82 (3 H, d, J 10.3, NCH3) and 0.81
(3 H, d, J 7.3, CCH3); δP(80 MHz; CDCl3) 29.6 (1 P, d, JPF 86.2);
δF(188 MHz; CDCl3) Ϫ120.6 (1 F, dd, JFH 45.0 and JFP 86.2);
δC(100 MHz; CDCl3) 141.5 (br d), 140.5, 135.8 (d), 133.9 (dd),
129.0, 128.7, 128.4, 128.3, 127.7, 127.5, 127.2, 126.4, 125.4 (dd),
89.6 (dd, JCP 150.8 and JCF 186.0, CHF), 83.0 (PhCO), 58.5 (d,
JCP 10.2, CH3CN), 28.4 (d, JCP 6.6, NCH3) and 14.9 (CH3C);
m/z (EI) 395 (Mϩ, 45%), 210 (73), 185 (63), 146 (42), 104 (100).
Found: [M]ϩ, 395.1460. C23H23O2PNF requires m/z 395.1450.
(2R,4S,5R)-2-[(1S)-Fluoro(2-naphthyl)methyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 22. White solid (see
Table 3 for yields); mp 169–171 ЊC (from pentane–CH2Cl2); [α]D22
Ϫ102.5 (c 0.0268 in MeOH) (Found: C, 67.7; H, 5.8; N, 3.8.
C21H21FNO2P requires C, 68.3; H, 5.7; N, 3.8%); νmax(KBr)/
cmϪ1 3059, 2980, 2933, 1505, 1454, 1260, 1191, 1060, 971, 913,
890, 827, 757 and 701; δH(200 MHz; CDCl3) 8.04 (1 H, s), 7.93–
7.84 (3 H, br m), 7.71 (1 H, d, J 8.8), 7.52 (2 H, s), 7.37 (5 H, s),
6.21 (1 H, dd, JHP 8.8 and JHF 45.7, CHF), 5.72 (1 H, d, J 5.9,
HCO), 3.77–3.69 (1 H, m, HCN), 2.23 (3 H, d, J 8.8, NCH3)
and 0.78 (3 H, d, J 5.9, CCH3); δP(80 MHz; CDCl3) 28.9 (1 P, d,
JPF 76.3); δF(188 MHz; CDCl3) Ϫ125.1 (1 F, dd, JFH 45.7 and
JFP 76.3); δC(100 MHz; CDCl3) 135.8 (d), 133.1, 132.8 (d), 130.9
(d), 128.3, 128.2, 128.2, 128.0, 127.7, 126.4, 125.8, 124.4 (dd),
123.0 (dd), 89.8 (dd, JCP 149.4 and JCF 188.9, CHF), 82.5
(PhCO), 58.6 (d, JCP 9.5, CH3CN), 29.5 (d, JCP 4.3, NCH3)
and 14.0 (CH3C); m/z (EI) 369 (Mϩ, 66%), 210 (42), 159 (100),
104 (48). Found: [M]ϩ, 369.1298. C21H21O2PNF requires m/z
369.1294.
(2R,4S,5R)-2-[(1S)-ꢀ-Fluorobenzyl]-3,4-dimethyl-2-oxo-5-
phenyl-1,3,2ꢁ5-oxazaphospholidine 24. White solid (29%); mp
129–131 ЊC (from pentane–CH2Cl2); [α]D22 Ϫ93.9 (c 0.0411 in
MeOH) (Found: C, 63.4; H, 6.0; N, 4.4. C17H19FNO2P requires
C, 63.9; H, 6.0; N, 4.4%); νmax(KBr)/cmϪ1 3086, 2974, 2867,
1495, 1452, 1324, 1265, 1218, 1187, 987, 956 and 700; δH(200
MHz; CDCl3) 7.53–7.26 (10 H, m), 5.98 (1 H, dd, JHP 4.4 and
JHF 45.4, CHF), 5.38 (1 H, t, J 6.6, HCO), 3.79–3.67 (1 H, m,
HCN), 2.81 (3 H, d, J 9.5, NCH3) and 0.80 (3H, d, J 5.8,
CCH3); δP(80 MHz; CDCl3) 29.5 (1 P, d, JPF 86.2); δF(188 MHz;
CDCl3) Ϫ120.4 (1 F, dd, JFH 45.4 and JFP 86.2); δC(100 MHz;
CDCl3) 136.1 (d), 133.8 (d), 128.9 (br d), 128.5 (br s), 128.4,
126.8 (d), 126.6 (d), 126.5, 90.0 (dd, JCP 148.2 and JCF 185.8,
CHF), 82.9 (PhCO), 58.5 (d, JCP 9.2, CH3CN), 28.5 (d, JCP 5.5,
NCH3) and 14.9 (d, JCP 2.7, CH3C); m/z (EI) 319 (Mϩ, 18%),
210 (58), 192 (29), 104 (100), 84 (98). Found: [M]ϩ, 319.1145.
C17H19O2PNF requires m/z 319.1137.
(2R,4S,5R)-2-[(1R)-Fluoro(2-naphthyl)methyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 25. White solid (see
Table 3 for yields); mp 135–137 ЊC (from pentane–CH2Cl2); [α]D22
ϩ40.8 (c 0.0135 in MeOH) (Found: C, 68.5; H, 5.7; N, 3.9.
C21H21FNO2P requires C, 68.3; H, 5.7; N, 3.8%); νmax(KBr)/
cmϪ1 3060, 2968, 2876, 1506, 1456, 1297, 1214, 1190, 1011, 976,
869, 748 and 701; δH(200 MHz; CDCl3) 7.99 (1 H, s), 7.89–7.81
(3 H, br m), 7.64 (1 H, d, J 7.3), 7.57–7.44 (2 H, br m), 7.32
(5 H, s), 6.15 (1 H, dd, JHP 4.4 and JHF 45.7, CHF), 5.45 (1 H, br
t, J 6.6, HCO), 3.77–3.73 (1 H, m, HCN), 2.82 (3 H, d, J 10.2,
NCH3) and 0.79 (3 H, d, J 5.8, CCH3); δP(80 MHz; CDCl3) 29.5
(1 P, d, JPF 85.5); δF(188 MHz; CDCl3) Ϫ119.8 (1 F, dd, JFH 45.7
and JFP 85.5); δC(100 MHz; CDCl3) 135.8 (d), 133.4, 132.9 (d),
130.8 (dd), 128.4 (br d), 128.3, 128.3, 128.2, 127.7, 126.6, 126.4,
126.4, 126.1 (t), 123.8 (dd), 89.8 (dd, JCP 147.9 and JCF 186.0,
CHF), 82.9 (PhCO), 58.5 (d, JCP 10.2, CH3CN), 28.5 (d, JCP
6.6, NCH3) and 14.9 (d, JCP 3.6, CH3C); m/z (EI) 369 (Mϩ,
31%), 210 (27), 159 (100), 104 (60). Found: [M]ϩ, 369.1291.
C21H21O2PNF requires m/z 369.1294.
(2R,4S,5R)-2-[(1R)-ꢀ-Fluorobenzyl]-3,4-dimethyl-2-oxo-5-
phenyl-1,3,2ꢁ5-oxazaphospholidine 27. White solid (15%); mp
114–116 ЊC (from pentane–CH2Cl2); [α]D22 ϩ26.3 (c 0.0404 in
MeOH) (Found: C, 63.6; H, 6.05; N, 4.4. C17H19FNO2P
requires C, 63.9; H, 6.0; N, 4.4%); νmax(KBr)/cmϪ1 3442, 3059,
2983, 1644, 1494, 1453, 1336, 1263, 1218, 1195, 1058, 958 and
701; δH(200 MHz; CDCl3) 7.57–7.27 (10 H, m), 6.01 (1 H, dd,
JHP 8.0 and JHF 45.8, CHF), 5.61 (1 H, br t, J 2.9, HCO), 3.74–
3.60 (1 H, m, HCN), 2.25 (3 H, d, J 10.2, NCH3) and 0.76 (3 H,
d, J 5.9, CCH3); δP(80 MHz; CDCl3) 28.9 (1 P, d, JPF 79.4);
δF(188 MHz; CDCl3) Ϫ125.8 (1 F, dd, JFH 45.8 and JFP 79.4);
δC(100 MHz; CDCl3) 136.2 (br d), 133.9 (d), 128.5 (br s), 126.1
(br s), 125.6 (br s), 90.1 (dd, JCP 149.2 and JCF 188.5, CHF),
82.6 (PhCO), 58.9 (d, JCP 9.2, CH3CN), 29.4 (d, JCP 4.6, NCH3)
and 14.2 (d, JCP 2.7, CH3C); m/z (EI) 319 (Mϩ, 19%), 210 (59),
1278
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281

View Article Online
192 (27), 104 (100). Found: [M]ϩ, 319.1126. C17H19O2PNF
requires m/z 319.1137.
J 5.8, HCO), 3.90–3.83 (1 H, m, HCN), 2.98 (3 H, d, J 8.7,
NCH3) and 0.84 (3 H, d, J 7.3, CCH3); δP(80 MHz; CDCl3) 28.3
(1 P, d, JPF 88.5); δF(188 MHz; CDCl3) Ϫ115.6 (1 F, dd, JFH 45.8
and JFP 88.5); δC(100 MHz; CDCl3) 141.4 (d), 140.3, 135.3 (d),
134.1 (dd), 128.9, 128.6, 128.2, 128.1, 127.5, 127.4, 127.0, 125.5,
125.4, 87.7 (dd, JCP 150.5 and JCF 187.8, CHF), 80.6 (PhCO),
61.0 (d, JCP 8.8, CH3CN), 29.4 (d, JCP 6.6, NCH3) and 13.9
(CH3C); m/z (EI) 395 (Mϩ, 52%), 210 (73), 185 (65), 146 (40),
104 (100). Found: [M]ϩ, 395.1445. C23H23O2PNF requires m/z
395.1450.
(2S,4S,5R)-2-[(1R)-Fluoro(2-naphthyl)methyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 28. White solid (see
Table 2 for yields); mp 125–127 ЊC (from pentane–CH2Cl2); [α]D22
ϩ238.6 (c 0.0331 in MeOH) (Found: C, 68.4; H, 6.1; N, 3.7.
C21H21FNO2P requires C, 68.3; H, 5.7; N, 3.8%); νmax(KBr)/
cmϪ1 3059, 2986, 2966, 1507, 1455, 1259, 1178, 1065, 1024, 965,
862, 829, 756 and 704; δH(200 MHz; CDCl3) 7.99 (1 H, s), 7.93–
7.84 (3 H, br m), 7.66 (1 H, d, J 7.3), 7.55–7.47 (2 H, br m), 7.39
(5 H, s), 6.22 (1 H, dd, JHP 8.1 and JHF 45.8, CHF), 5.84 (1 H, d,
J 5.8, HCO), 3.74–3.61 (1 H, m, HCN), 2.33 (3 H, d, J 8.8,
NCH3) and 0.90 (3 H, d, J 7.3, CCH3); δP(80 MHz; CDCl3) 28.0
(1 P, d, JPF 80.9); δF(188 MHz; CDCl3) Ϫ119.7 (1 F, dd, JFH 45.8
and JFP 80.9); δC(100 MHz; CDCl3) 135.4 (d), 133.1, 132.8 (d),
130.8 (d), 128.2, 128.2, 128.1, 128.0, 127.6, 126.4, 126.3, 125.7,
124.8 (dd), 123.1 (dd), 88.0 (dd, JCP 152.0 and JCF 192.3, CHF),
80.5 (PhCO), 61.1 (d, JCP 9.5, CH3CN), 30.8 (d, JCP 5.1, NCH3)
and 14.8 (d, JCP 3.7, CH3C); m/z (EI) 369 (Mϩ, 61%), 210 (40),
159 (100), 104 (43). Found: [M]ϩ, 369.1301. C21H21O2PNF
requires m/z 369.1294.
(2S,4S,5R)-2-[(1R)-ꢀ-Fluorobenzyl]-3,4-dimethyl-2-oxo-5-
phenyl-1,3,2ꢁ5-oxazaphospholidine 30. Colorless oil (30%); [α]D22
ϩ26.3 (c 0.0476 in MeOH) νmax(KBr)/cmϪ1 3440, 3063, 2914,
1633, 1495, 1453, 1333, 1257, 1214, 1181, 972 and 699; δH(200
MHz; CDCl3) 7.56–7.26 (10 H, m), 6.04 (1 H, dd, JHP 8.0 and
JHF 45.8, CHF), 5.83 (1 H, d, J 5.9, HCO), 3.74–3.61 (1 H, m,
HCN), 2.38 (3 H, d, J 8.6, NCH3) and 0.86 (3H, d, J 6.6,
CCH3); δP(80 MHz; CDCl3) 27.9 (1 P, d, JPF 79.3); δF(188 MHz;
CDCl3) Ϫ120.3 (1 F, dd, JFH 45.8 and JFP 79.3); δC(100 MHz;
CDCl3) 135.8 (d), 133.8 (d), 128.4 (br s), 125.9 (br s), 88.3 (dd,
JCP 151.9 and JCF 190.4, CHF), 80.7 (PhCO), 61.4 (d, JCP 9.2,
CH3CN), 30.9 (d, JCP 5.5, NCH3) and 14.9 (d, JCP 3.7, CH3C);
m/z (EI) 319 (Mϩ, 31%), 210 (69), 192 (33), 104 (100). Found:
[M]ϩ, 319.1137. C17H19O2PNF requires m/z 319.1137.
(2S,4S,5R)-2-[(1S)-Fluoro(2-naphthyl)methyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 31. White solid (see
Table 2 for yields). mp 181–183 ЊC (from pentane–CH2Cl2);
[α]D22 Ϫ215.3 (c 0.0273 in MeOH) (Found: C, 68.3; H, 6.1; N,
3.8. C21H21FNO2P requires C, 68.3; H, 5.7; N, 3.8%); νmax(KBr)/
cmϪ1 3061, 2985, 2969, 1507, 1454, 1251, 1185, 1062, 973, 955,
876, 857, 752 and 702; δH(200 MHz; CDCl3) 8.06 (1 H, s), 7.95–
7.84 (3 H, br m), 7.73 (1 H, d, J 7.4), 7.54–7.48 (2 H, br m),
7.42–7.27 (5 H, m), 6.04 (1 H, dd, JHP 4.4 and JHF 45.8, CHF),
5.82 (1 H, d, J 5.8, HCO), 3.82–3.75 (1 H, m, HCN), 2.89 (3 H,
d, J 8.8, NCH3) and 0.79 (3 H, d, J 7.3, CCH3); δP(80 MHz;
CDCl3) 28.5 (1 P, d, JPF 87.0); δF(188 MHz; CDCl3) Ϫ114.9
(1 F, dd, JFH 45.8 and JFP 87.0); δC(100 MHz; CDCl3) 135.4 (d),
133.4, 132.9 (d), 130.9 (dd), 128.4 (br d), 128.3, 128.1, 128.1,
127.7, 126.5, 126.4, 126.3 (m), 125.6, 123.9 (dd), 87.9 (dd, JCP
151.6 and JCF 189.7, CHF), 80.8 (PhCO), 61.1 (d, JCP 8.8,
CH3CN), 29.5 (d, JCP 6.6, NCH3) and 14.1 (d, JCP 2.9, CH3C);
m/z (EI) 369 (Mϩ, 66%), 210 (38), 159 (100), 104 (42). Found:
[M]ϩ, 369.1297. C21H21O2PNF requires m/z 369.1294.
(2S,4S,5R)-2-[(1S)-ꢀ-Fluorobenzyl]-3,4-dimethyl-2-oxo-5-
phenyl-1,3,2ꢁ5-oxazaphospholidine 33. White solid (17%); mp
165–167 ЊC (from pentane–CH2Cl2); [α]D22 Ϫ108.3 (c 0.0263 in
MeOH) (Found: C, 63.5; H, 6.0; N, 4.4. C17H19FNO2P requires
C, 63.9; H, 6.0; N, 4.4%); νmax(KBr)/cmϪ1 3383, 3059, 2971,
1603, 1494, 1455, 1334, 1259, 1184, 1067, 977 and 699; δH(200
MHz; CDCl3) 7.62–7.11 (10 H, m), 5.87 (1 H, dd, JHP 4.4 and
JHF 45.5, CHF), 5.81 (1 H, d, J 5.8, HCO), 3.83–3.70 (1 H, m,
HCN), 2.86 (3 H, d, J 8.8, NCH3) and 0.76 (3 H, d, J 7.3,
CCH3); δP(80 MHz; CDCl3) 28.3 (1 P, d, JPF 87.0); δF(188 MHz;
CDCl3) Ϫ115.2 (1 F, dd, JFH 45.5 and JFP 87.0); δC(100 MHz;
CDCl3) 135.7 (d), 133.7 (d), 129.9, 128.6, 128.4, 128.3, 126.9 (br
t), 125.8, 88.2 (dd, JCP 151.9 and JCF 189.4, CHF), 80.9 (PhCO),
61.5 (d, JCP 8.3, CH3CN), 29.7 (d, JCP 6.4, NCH3) and 14.3 (d,
JCP 2.7, CH3C); m/z (EI) 319 (Mϩ, 13%), 210 (43), 147 (100),
104 (72), 84 (38). Found: [M]ϩ, 319.1147. C17H19O2PNF
requires m/z 319.1137.
(2S,4S,5R)-2-[(1R)-ꢀ-Fluoro(m-phenyl)benzyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 29. White hygro-
scopic solid (53%); [α]D22 ϩ27.1 (c 0.0240 in MeOH) (Found:
C, 69.4; H, 6.05; N, 3.5. C23H23FNO2P requires C, 69.9; H, 5.9;
N, 3.55%); νmax(KBr)/cmϪ1 3031, 2970, 2933, 2829, 1954, 1888,
1732, 1599, 1478, 1258, 1184, 1064, 971, 860 and 701; δH(200
MHz; CDCl3) 7.85 (1 H, s), 7.71–7.34 (8 H, m), 6.19 (1 H, dd,
JHP 8.1 and JHF 45.8, CHF), 5.93 (1 H, d, J 5.9, HCO), 3.80–
3.74 (1 H, m, HCN), 2.50 (3 H, d, J 8.8, NCH3) and 0.95 (3 H,
d, J 5.9, CCH3); δP(80 MHz; CDCl3) 27.9 (1 P, d, JPF 82.4);
δF(188 MHz; CDCl3) Ϫ120.3 (1 F, dd, JFH 45.8 and JFP 82.4);
δC(100 MHz; CDCl3) 141.3 (br d), 140.3, 135.5 (d), 134.1 (d),
128.9, 128.7, 128.3, 128.2, 127.5, 127.3, 127.0, 125.8, 124.6 (dd),
124.3 (dd), 87.9 (dd, JCP 151.9 and JCF 195.2, CHF), 80.6
(PhCO), 61.2 (d, JCP 9.5, CH3CN), 30.9 (d, JCP 5.9, NCH3) and
14.9 (CH3C); m/z (EI) 395 (Mϩ, 49%), 210 (69), 185 (63), 146
(37), 104 (100). Found: [M]ϩ, 395.1464. C23H23O2PNF requires
m/z 395.1454.
General procedure for synthesis of chiral phosphonic acids 4–9
A solution of the oxazaphospholone (22–33) in 10% TFA–
MeOH (approximately 1 cm3 10% TFA–MeOH/0.1 mmol of
22–33) was stirred at 25 ЊC overnight. The solution was then
concentrated in vacuo, re-dissolved in benzene and re-
concentrated a total of 3 times prior to being placed under
high vacuum for 2 h. The resulting product was dissolved in
anhydrous CH2Cl2 (approximately 3–5 cm3/1 mmol of 22–33)
and stirred at 25 ЊC for 24 h in the presence of TMSBr (10
equiv.). The solution was concentrated in vacuo and placed
under high vacuum for several hours. Anhydrous ether was then
added followed by filtration and concentration of the filtrate
(if a precipitate formed during the rotary evaporation, more
anhydrous ether was added and the suspension re-filtered). The
filtrate was dissolved in MeOH, containing 1 drop of water, and
stirred for 15 min. The solution was then concentrated in vacuo
and the resulting product was washed extensively with benzene
and CH2Cl2 to yield pure product. In addition to spectroscopic
analysis of phosphonic acids 4–9 (see below), their purity was
also examined by analytical reversed phase HPLC (solvent A:
acetonitrile; solvent B: water with 0.1% TFA modifier) using the
following gradient: 0 min: 25% A; 5 min: 25% A; 10 min: 100%
A; 15 min: 100% A; 20 min: 75% A; 30 minutes: 25% A. All of
the phosphonic acids were indicated to be pure by the criterion
of analytical HPLC. The ee’s of the phosphonic acids were
(2S,4S,5R)-2-[(1S)-ꢀ-Fluoro(m-phenyl)benzyl]-3,4-dimethyl-
2-oxo-5-phenyl-1,3,2ꢁ5-oxazaphospholidine 32. White solid
(32%); mp 131–133 ЊC (from pentane–CH2Cl2); [α]2D2 Ϫ234.6 (c
0.0251 in MeOH) (Found: C, 70.1; H, 5.9; N, 3.7. C23H23FNO2P
requires C, 69.9; H, 5.9; N, 3.55%); νmax(KBr)/cmϪ1 3032, 2970,
2937, 2824, 1947, 1878, 1737, 1597, 1478, 1255, 1114, 1064, 954,
859 and 699; δH(200 MHz; CDCl3) 7.92 (1 H, s), 7.72–7.34 (8 H,
m), 6.15 (1 H, dd, JHF 45.6 and JHP 4.4, CHF), 5.90 (1 H, d,
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281
1279

View Article Online
Table 5 Crystallographic data for compounds 22–24
22
23
24
Chemical formula
Formula weight
C21H21FNO2P
369.36
C23H23FNO2P
395.39
C17H19FNO2P
319.30
Crystal system
Space group
Absorption coefficient
orthorhombic
P2(1)2(1)2(1)
0.172 mmϪ1
orthorhombic
P2(1)2(1)2
0.167 mmϪ1
orthorhombic
P2(1)2(1)2(1)
0.191 mmϪ1
Unit cell dimensions
a/Å
5.7212(4)
11.8451(5)
5.7208(2)
b/Å
c/Å
11.5757(16)
28.177(4)
27.3815(19)
6.0985(8)
11.2412(8)
24.5839(15)
α/Њ
β/Њ
90
90
90
90
90
90
γ/Њ
90
1866.1(4)
100.0(1)
4
90
1978.0(3)
100.0(1)
4
90
Unit cell volume/Å3
Temperature/K
Z
1580.96(16)
100.0(1)
4
Reflections collected
Independent reflections
R indices (all data)
R1
6987
9348
7844
1586 [R(int) = 0.143]
3440 [R(int) = 0.090]
3556 [R(int) = 0.070]
0.0887
0.1171
0.1102
0.1080
0.0717
0.1000
wR2
Final R indices
R1
0.0521
0.1049
1.004
0.0552
0.0949
0.952
0.0452
0.0924
0.949
wR2
Goodness-of-fit on F2
determined to be >97% using the 19F-NMR method described
below.
1020, 946, 783, 732, 697, 631 and 542; δH(200 MHz; CD3OD)
7.51–7.37 (5 H, m) and 5.68 (1 H, dd, JHP 8.1 and JHF 44.7,
CHF); δP(80 MHz; CD3OD) 15.2 (1 P, d, JPF 85.5); δF(188
MHz; CD3OD) Ϫ120.5 (1 F, dd, JFH 44.7 and JFP 85.5); δC(100
MHz; CD3OD) 135.8 (d), 129.9, 129.3, 128.2 (br t) and 91.7
(dd, JCP 166.6 and JCF 179.4, CHF); m/z (ES) 189 (100%).
(R)-Fluoro(2-naphthyl)methylphosphonic acid 4. Obtained as
a white solid in 57% yield from 25 and 62% yield from 28; mp
180–181 ЊC (decomp.); HPLC retention time = 3.65 min; [α]D22
ϩ45.7 (c 0.0433 in MeOH); νmax(KBr)/cmϪ1 2790 (br), 2265 (br),
1602, 1509, 1369, 1236, 1209, 1047, 940, 817 and 742; δH(200
MHz; CD3OD) 7.97 (1 H, s), 7.89–7.85 (3 H, br m), 7.63 (1 H,
d, J 8.8), 7.50–7.46 (2 H, br m) and 5.88 (1 H, dd, JHP 8.8 and
JHF 45.8, CHF); δP(80 MHz; CD3OD) 15.4 (1 P, d, JPF 83.9);
δF(188 MHz; CD3OD) Ϫ120.2 (1 F, dd, JFH 45.8 and JFP 83.9);
δC(100 MHz; CD3OD) 134.9, 134.4 (d), 133.1 (dd), 129.1, 129.0
(br d), 128.7, 127.6, 127.6 (m), 127.4, 125.5 (br dd) and 91.7 (dd,
JCP 166.2 and JCF 180.1, CHF); m/z (ES) 239 (100%).
(S)-Fluoro(phenyl)methylphosphonic acid 9. Obtained as a
white solid in 46% yield from 24 and 41% yield from 33. HPLC
chromatogram, mp and all spectral data were identical to those
reported for 8. Optical rotation was approximately the same as
that obtained for 8 but of opposite sign.
Determination of enantiomeric excess of phosphonic acids 4–9
A mixture of the phosphonic acid (4–9, 10–15 mg) and quin-
idine (1 equiv.) was dissolved in CDCl3 (approximately 1 cm3),
shaken until a homogeneous solution was obtained, and the
19F-NMR was recorded. For each of the phosphonic acids 4–9,
only a single set of doublet of doublets was evident in the
spectra indicating that 4–9 were obtained in >97% ee.
(S)-Fluoro(2-naphthyl)methylphosphonic acid 5. Obtained as
a white solid 58% yield from 22 and 65% yield from 31. HPLC
chromatogram, mp, and all spectral data were identical to those
reported for 4. Optical rotation was approximately the same as
that obtained for 4 but of opposite sign.
Kinetic studies with PTP1B
(R)-ꢀ-Fluoro(m-phenyl)benzylphosphonic acid 6. Obtained
as a white solid in 64% yield from 26 and 70% yield from
29; HPLC retention time = 3.35 min; [α]D22 ϩ31.1 (c 0.0211
in MeOH); mp 175 ЊC (decomp.); νmax(KBr)/cmϪ1 3035 (br),
2333 (br), 1598, 1478, 1253, 1183, 1025, 952, 753 and 699;
δH(200 MHz; CD3OD) 7.76 (1 H, s), 7.63–7.30 (8 H, br m)
and 5.78 (1 H, dd, JHP 8.8 and JHF 44.0, CHF); δP(80 MHz;
CD3OD) 15.1 (1 P, d, JPF 85.4); δF(188 MHz; CD3OD) Ϫ120.6
(1 F, dd, JFH 44.0 and JFP 85.4); δC(100 MHz; CD3OD) 142.5,
(d), 141.9, 136.4 (d), 129.9, 129.9, 128.6, 128.4, 128.0, 127.0 (t),
126.7 (t) and 91.5 (dd, JCP 165.5 and JCF 180.1, CHF); m/z (ES)
265 (100%).
Rates of PTP1B-catalyzed dephosphorylation in the presence
or absence of inhibitors were determined using FDP as sub-
strate in assay buffer containing 100 mM BIS-TRIS, 4 mM
EDTA, 5 mM DTT (dithiothreitol) and 0.2 mg mLϪ1 BSA, pH
6.5 at 25 ЊC as previously described3 except 5% DMSO was
present in the reaction mixture. IC50 determinations were
determined at nine or ten different inhibitor concentrations
with FDP at Km concentration (20 µM). Ki’s for 2, 3, 4, and 6
were determined by measuring the initial rate (v) using various
FDP concentrations (20, 25, 35, 50 and 100 µM FDP) at
various fixed concentrations of inhibitor, [(0, 20, 40, 60, 80 and
100 µM) 2, (0, 20, 40, 50, 60 and 80 µM) 3, (0, 300, 500, 750,
1000 and 1500 µM) 4, (0, 100, 200, 300, 500 and 750) 6] as
previously described.3 All assays were performed in duplicate.
(S)-ꢀ-Fluoro(m-phenyl)benzylphosphonic acid 7. Obtained as
a white solid in 69% yield from 23 and 67% yield from 32.
HPLC chromatogram, mp and all spectral data were identical
to those reported for 6. Optical rotation was approximately the
same as that obtained for 6 but of opposite sign.
X-Ray structural characterization†
A summary of selected crystallographic data is given in Tables
5 and 6. Data were collected on a Nonius KappaCCD diffract-
(R)-Fluoro(phenyl)methylphosphonic acid 8. Obtained as a
white solid in 39% yield from 27 and 47% yield from 30; HPLC
retention time = 3.57 min; [α]D22 ϩ36.9 (c 0.0489 in MeOH); mp
114–115 ЊC; νmax(KBr)/cmϪ1 3421, 2355, 1634, 1494, 1454, 1196,
† CCDC reference number 207/399. See http://www.rsc.org/suppdata/
p1/a9/a908086d for crystallographic files in .cif format.
1280
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281

View Article Online
Table 6 Crystallographic data for compounds 25–27
25
26
27
Chemical formula
Formula weight
C21H21FNO2P
369.36
C23H23FNO2P
395.39
C17H19FNO2P
319.30
Crystal system
Space group
Absorption coefficient
monoclinic
P2(1)
orthorhombic
P2(1)2(1)2
0.162 mmϪ1
monoclinic
P2(1)
0.171 mmϪ1
0.189 mmϪ1
Unit cell dimensions
a/Å
b/Å
c/Å
6.6666(4)
12.1266(5)
11.6697
90
12.272(2)(5)
24.9928(3)
6.6447(7)
90
9.9593(7)
9.9593(7)
9.9593(7)
90
α/Њ
β/Њ
γ/Њ
96.743(3)
90
936.89(8)
200.0(1)
90
90
2038.1(2)
100.0(1)
92.346(3)
90
798.51(9)
100.0(1)
2
Unit cell volume/Å3
Temperature/K
Z
2
4
Reflections collected
Independent reflections
R indices (all data)
R1
6432
12421
4608 [R(int) = 0.051]
10240
3541 [R(int) = 0.055]
3690 [R(int) = 0.032]
0.0393
0.0800
0.0473
0.0796
0.0416
0.0937
wR2
Final R indices
R1
0.0328
0.0771
1.050
0.0354
0.0766
1.052
0.0366
0.0915
1.034
wR2
Goodness-of-fit on F2
11 H. K. Kole, M. S. Smyth, P. L. Russ and T. R. Burke, Biochem. J.,
1995, 311, 1025.
12 N. R. Glover and A. S. Tracey, Biochemistry, 1999, 38, 5256.
13 G. M. Blackburn and A. Rashid, J. Chem. Soc., Chem. Commun.,
1988, 317.
ometer using graphite monochromated Mo-Kα radiation
(λ = 0.71073 Å). A combination of 1Њ phi and omega (with
kappa offsets) scans was used to collect sufficient data. The
data frames were integrated and scaled using the Denzo-SMN
package.39 The structures were solved and refined using the
SHELXTL/PC V5.140 package. Refinement was by full-matrix
least-squares on F2 using all data (negative intensities included).
14 A. S. Campbell and G. R. J. Thatcher, Tetrahedron Lett., 1991, 32,
2207.
15 T. Yokomatsu, T. Yamagishi, K. Matsumoto and S. Shibuya,
Tetrahedron, 1996, 52, 11725.
16 Y. Bennani and S. Hanessian, Chem. Rev., 1997, 97, 3174.
17 M. Sting and W. Steglich, Synthesis, 1990, 132.
18 Y. Bennani and S. Hanessian, Tetrahedron, 1996, 52, 13837 and
references therein.
Acknowledgements
This work was supported by a Natural Sciences and Engineer-
ing Research Council (NSERC) of Canada operating grant to
S. D. T. We also thank NSERC for a summer research scholar-
ship for W. W. and the University of Toronto for a University
of Toronto Open Fellowship for C. C. K. We also thank
Merck-Frosst Canada for gifts of PTP1B and FDP.
19 S. E. Denmark and C.-T. Chen J. Am. Chem. Soc., 1995, 117, 11879
and references therein.
20 E. Differding and R. W. Lang, Tetrahedron Lett., 1988, 29,
6087.
21 E. Differding and R. W. Lang, Helv. Chim. Acta, 1989, 72,
1248.
22 F. A. Davis and W. Han, Tetrahedron Lett., 1992, 33, 1153.
23 F. A. Davis and H. Qi, Tetrahedron Lett., 1996, 37, 4345.
24 F. A. Davis, P. V. N. Kasu, G. Sundarababu and H. Qi, J. Org.
Chem., 1997, 62, 7546.
25 F. A. Davis and P. V. N. Kasu, Tetrahedron Lett., 1998, 39, 6135.
26 F. A. Davis and W. Han, Tetrahedron Lett., 1991, 32, 1631.
27 D. A. Evans, T. C. Britton, J. A. Ellman and R. L. Dorrow, J. Am.
Chem. Soc., 1990, 112, 4011.
28 E. Differding, G. M. Ruegg and R. W. Lang, Synlett, 1991, 395.
29 S. D. Taylor, C. C. Kotoris, A. N. Dinaut and M. J. Chen,
Tetrahedron, 1998, 54, 1691.
30 D. B. Cooper, C. R. Hall, J. M. Harrison and T. D. Inch, J. Chem.
Soc., Perkin Trans. 1, 1977, 1969.
31 K. C. Calvo, J. Am. Chem. Soc., 1985, 107, 3690.
32 D. O. O’Hagan and H. S. Rzepa, Chem. Commun., 1997, 645.
33 J. D. Dunitz and R. Taylor, Chem. Eur. J., 1997, 3, 89.
34 J. A. K. Howard, V. J. Hoy, D. O. O’Hagan and G. T. Smith,
Tetrahedron, 1996, 52, 12613.
35 T. J. Barbarich, C. D. Rithner, S. M. Miller, O. P. Anderson and
S. H. Strauss, J. Am. Chem. Soc., 1999, 121, 4280.
36 S. D Taylor, A. N. Dinaut, A. Thadini and Z. Huang, Tetrahedron
Lett., 1996, 37, 8089.
37 L. Arbusow, Zh. Obshch. Khim., 1950, 20, 1249.
38 R. Rabinowitz, J. Org. Chem., 1963, 28, 2975.
39 Z. Otwinowski and W. Minor, Methods Enzymol., 1997, 276,
307.
References
1 G. M. Blackburn and D. E. Kent, J. Chem. Soc., Chem. Commun.,
1981, 511.
2 For examples see: (a) G. M. Blackburn, D. E. Kent and F. Kolkman,
J. Chem. Soc., Perkin Trans 1, 1986, 913; (b) M. F. Gordeev, D. V.
Patel, P. L. Barker and E. M. Gordon, Tetrahedron Lett., 1994, 35,
7585; (c) J. Nieschalk, A. S. Batsanov, D. O’Hagan and J. A. K.
Howard, Tetrahedron, 1996, 52, 165.
3 S. D. Taylor, C. C. Kotoris, A. N. Dinaut, Q. Wang, C.
Ramachandran and Z. Haung, Bioorg. Med. Chem., 1998, 6, 1457.
4 C. C. Kotoris, M.-J. Chen and S. D. Taylor, Bioorg. Med. Chem.
Lett., 1998, 8, 3275.
5 Q. Wang, Z. Huang, C. Ramachandran, A. N. Dinaut and S. D.
Taylor, Bioorg. Med. Chem. Lett., 1998, 8, 345.
6 Z.-Y. Zhang, Crit. Rev. Biochem. Mol. Biol., 1998, 33, 1.
7 M. Elchebley, P. Payette, E. Michaliszyn, W. Cromlish, S. Collins,
A. L. Loy, D. Mormandin, A. Cheng, J. Himms-Hagen, C.-C. Chan,
C. Ramachandran, M. J. Gresser, M. L. Tremblay and B. P.
Kennedy, Science, 1999, 283, 1544.
8 T. R. Burke, H. K. Kole and P. P. Roller, Biochem. Biophys. Res.
Commun., 1994, 204, 129.
9 L. Chen, L. Wu, A. Otaka, M. S. Smyth, P. P. Roller, T. R. Burke,
J. den Hertog and Z.-Y. Zhang, Biochem. Biophys. Res. Commun.,
1995, 216, 976.
10 T. R. Burke, B. Ye, X. Yan, S. Wang, Z. Jia, L. Chen, Z.-Y. Zhang
and D. Barford, Biochemistry, 1996, 35, 15989.
40 G. M. Sheldrick, SHELXTL/PC V5.1, Bruker Analytical X-ray
Systems, Madison, Wisconsin, U.S.A. 1997.
J. Chem. Soc., Perkin Trans. 1, 2000, 1271–1281
1281