Inhibition of myo-inositol monophosphatase isoforms by aromatic phosphonates

Article abstract of DOI:10.1016/S0968-0896(98)00148-5

α-Hydroxyphosphonates are moderately potent (K(i)=6-600μM) inhibitors of the enzyme myo-inositol monophosphatase (McLeod et al., Med. Chem. Res. 1992, 2, 96). Hydroxy-[4-(5,6,7,8-tetrahydronaphtyl-1-oxy)phenyl]methyl phosphonate (3) was resynthesized and its inhibitory potency towards the recombinant bovine brain enzyme confirmed (K(i)=20μM). Similar aromatic difluoro-, keto-, and ketodifluorophosphonates (5, 7, 9) were inactive. Compound 3 was 15-fold less active on the human as compared to the bovine enzyme. Molecular modeling suggested that the hydrophobic part of the inhibitor interacts with amino acid side chains that are located at the interface between the enzyme subunits in an area (amino acids 175-185) with low similarity between the two isozymes. Phe-183 in the human enzyme was replaced with leucine, the corresponding residue in the bovine isoform. The three isozymes (human wild-type, bovine wild-type and human F183L) had similar kinetic properties, except that the bovine enzyme was less effectively inhibited by high concentrations of the activator Mg²⁺. The F183L mutant enzyme had a twofold increased affinity for compound 3 as compared to the human wild-type form. We conclude that residue 183 contributes to the binding of aromatic hydroxyphosphonates to IMPase, but it is not the only determining factor for inhibitor specificity with respect to different isozymes. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(98)00148-5

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 1865±1874
Inhibition of myo-Inositol Monophosphatase Isoforms by
Aromatic Phosphonates
Axel J. Ganzhorn,* Jan Ho¯ack,y Patricia D. Pelton,{ Francoise Strasser,x
,
Marie-Christine Chanal and Serge R. Piettrek
Marion Merrell Research Institute, 16 rue d' Ankara, 76080 Strasbourg, France{
Received 11 August 1997; accepted 2 June 1998
AbstractÐꢀ-Hydroxyphosphonates are moderately potent (Ki=6±600 mM) inhibitors of the enzyme myo-inositol
monophosphatase (McLeod et al., Med. Chem. Res. 1992, 2, 96). Hydroxy-[4-(5,6,7,8-tetrahydronaphtyl-1-oxy)phe-
nyl]methyl phosphonate (3) was resynthesized and its inhibitory potency towards the recombinant bovine brain enzyme
con®rmed (Ki=20 mM). Similar aromatic di¯uoro-, keto-, and ketodi¯uorophosphonates (5, 7, 9) were inactive. Com-
pound 3 was 15-fold less active on the human as compared to the bovine enzyme. Molecular modeling suggested that
the hydrophobic part of the inhibitor interacts with amino acid side chains that are located at the interface between the
enzyme subunits in an area (amino acids 175±185) with low similarity between the two isozymes. Phe-183 in the human
enzyme was replaced with leucine, the corresponding residue in the bovine isoform. The three isozymes (human wild-
type, bovine wild-type and human F183L) had similar kinetic properties, except that the bovine enzyme was less
e€ectively inhibited by high concentrations of the activator Mg2+. The F183L mutant enzyme had a twofold increased
anity for compound 3 as compared to the human wild-type form. We conclude that residue 183 contributes to the
binding of aromatic hydroxyphosphonates to IMPase, but it is not the only determining factor for inhibitor speci®city
with respect to di€erent isozymes. # 1998 Elsevier Science Ltd. All rights reserved.
Introduction
of intracellular inositol pools and reduced resynthesis of
membrane inositol lipids. Hyperactive cells, as they may
occur in mania, would subsequently be forced to slow
down, provided their stimulation is linked to the inositol
lipid second messenger system and they have limited
access to dietary inositol.6
myo-Inositol monophosphatase (E.C. 3.1.3.25, IMPase),
which hydrolyzes inositol monophosphate isomers to
produce free inositol, is an important enzyme within the
inositol lipid signaling system. Due to its uncompetitive
inhibition by lithium, IMPase was proposed as the tar-
get of lithium therapy in bipolar disorder.1±5 According
to the hypothesis, inhibition would lead to a depletion
A variety of studies carried out over the last few years
have provided important and exhaustive information
regarding the structure, mechanism and kinetic proper-
ties of myo-inositol monophosphatase. Both the bovine
and human isoforms have been cloned and produced in
large quantities in bacterial expression systems.7±9 The
three-dimensional structure of the enzyme shows a large
hydrophilic active site with several acidic residues that
are involved in substrate and metal ion binding.10±12
The enzyme is activated by Mg2+ in a nonlinear fash-
ion,13 which is explained by the requirement for more
than one metal ion for substrate binding and cata-
lysis.9,14±16 Catalytic mechanisms were proposed with
roles for two Mg2+ ions in binding the phosphate moiety
of the substrate, activating a water nucleophile and
Key words: myo-Inositol monophosphatase; phosphonate
inhibitors; isozymes; mutagenesis.
*Corresponding author: Tel: (33) 3 88608790; Fax: (33) 3
8845075
yCurrent address: Astra Hassle AB, S-431 83 Molndal, Sweden.
{Current address: Department of Cell Biology, Neurobiology &
Anatomy, University of Cincinnati, P.O. 67052, Cincinnati,
OH 45267-0251, U.S.A.
xCurrent address: Department of Biochemistry, Allegheny
University, Philadelphia, PA 19102, U.S.A.
kCurrent address: Faculte des Sciences, Universite de Rouen,
76821 Mont Saint Aigan, France.
{Now Synthelabo Biomoleculaire, 16 rue d' Ankara, 76080
Strasbourg, France.
0968-0896/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00148-5

1866
A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
facilitating ester hydrolysis.12,16,17 More recently, struc-
tural and kinetic evidence was provided, which clearly
points to a third metal ion contributing to enzyme cata-
lysis.18,19 Lithium inhibits the phosphatase reaction
uncompetitively by binding to an enzyme phosphate
complex and preventing the dissociation of inorganic
phosphate.1,20 Even though widely used in the treatment
of bipolar disorder, lithium has an unfavorable side
e€ect pro®le and a very narrow therapeutic window.
Consequently, the design and evaluation of novel and
selective IMPase inhibitors, both as pharmacological
tools and leads for drug development, has received
much attention. Most of the active compounds described
so far were phosphate, phosphonate or bisphosphonate
derivatives.21±28 Non-hydrolysable bisphosphonic acids,
for example, are potent inhibitors with anities in the
nanomolar range,27 but are highly charged and there-
fore have low bioavailability in the brain.29 A promising
novel concept of inhibiting myo-inositol monophos-
phatase and other metal containing enzymes based on
the chelating properties of the tropolone ring was
recently discovered, but the in vivo properties of such
compounds still need to be evaluated.30,31
barrier. We have extended their series to include
di¯uoro-, keto-, and ketodi¯uorophosphonates and
report here the e€ect of these compounds on both
bovine and human brain IMPase.
Results and Discussion
Chemistry
Aldehyde 1 was readily obtained from 1-hydroxy-
5,6,7,8-tetrahydronaphtalene and p-¯uorobenzaldehyde
under basic conditions (Scheme 1).24 Reaction with
dimethyl or di-tert-butyl phosphite in the presence of a
base gave the expected adducts 2a and 2b, respectively.
Dimethyl phosphonate 2a was transformed into the
corresponding acid 3.24 Oxidation of di-tert-butyl phos-
phonate 2b was performed using pyridinium dichromate
(PDC) and furnished the oxophosphonate 4 in good
yield. This ester was hydrolyzed to the acid by treatment
with tri¯uoroacetic acid.
An analogous strategy was followed for b-hydroxy and
b-oxo-a,a-di¯uorophosphonates 7 and 9. Thus reaction
of aldehyde 1 with diethyl a-lithio-a,a-di¯uoromethyl-
phosphonate at low temperature led to the formation of
the expected adduct 6 in fair yield. Sequentially treating
6 with trimethylsilyl bromide (TMSBr) and water liber-
McLeod et al.24 have synthesized a series of hydroxy-
phosphonate derivatives, which inhibit the enzyme from
bovine brain in the low micromolar range, but have
reported no data on the human brain enzyme. These
may be interesting compounds since they are not as
highly charged as the bisphosphonate series and there-
fore have a better chance to cross the blood±brain
ated acid
7 in quantitative yield. b-Hydroxy-a,a-
di¯uoromethylenephosphonate 6 was also oxidized by
PDC to a€ord the corresponding oxophosphonate 8 in
excellent yield. Deprotection as above (TMSBr/H2O)
Scheme 1. Synthesis of phosphonates 3 and 5 and di¯uorophosphonates 7 and 9.

A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
1867
led to the isolation of b-oxo-a,a-di¯uorophosphonic
acid 9.32
enzyme, we also puri®ed the enzyme from human brain.
Tissue and recombinant enzyme had very similar kinetic
properties and were both inhibited equally well by
lithium (Ki=0.2 mM), but neither one by compound 3.
Inhibition of bovine and human IMPase
Enzyme activity was measured in a standard assay,
where the production of inorganic phosphate is detected
through its complexation with ammonium molybdate.33
With dl-inositol 1-phosphate (Ins(1)P) as the enzyme
substrate, neither 3, 5, 7, nor 9, at a concentration of
0.5 mM, inhibited human recombinant myo-inositol
monophosphatase (IMPase). Bovine IMPase, on the
other hand, was inhibited by 3 with an IC50 value of
60 mM. Assuming competitive inhibition and based on a
Km value of 0.1 mM,9 this corresponds to a Ki of 20 mM,
in agreement with the original result (Ki=32 mM) by
MacLeod et al.24 The bovine isoform was not inhibited
by 0.5 mM of 5, 7, or 9. Inhibition by 3 was also mea-
sured in a continuous photometric assay.13 Again, the
IC50 was 60 mM for the bovine enzyme, but less than
20% inhibition was found at 0.5 mM compound with
the human isozyme. In order to rule out a possible pro-
blem related to the bacterial expression of the human
Structural interpretation
Previous comparisons of bovine and human IMPase did
not indicate any major di€erences regarding their a-
nities for substrates, substrate analogues and other
inhibitors including a bisphosphonate derivative.8,29 It
therefore appears that the structural di€erence between
the two isozymes that accounts for the di€erence in
inhibition by 3, is not located within the substrate/
phosphate binding pocket. However, it cannot be too
distant from the active site, since 3 is a competitive
inhibitor versus inositol 1-phosphate.24
A molecular model for the bovine IMPase structure was
generated based on the 3-dimensional structure of the
human enzyme10 by amino acid replacement followed
by energy minimization and inhibitor 3 was tentatively
docked into the active site (Fig. 1). For positioning the
Figure 1. Proposed mode of binding of 3 to the active site of bovine myo-inositol monophosphatase. The model of bovine IMPase
based on the X-ray structure of the human enzyme10 is shown with two metal ions bound. A gadolinium ion (blue sphere) occupies
metal binding site M1, whereas binding site M2 (magenta sphere) is presumably occupied by lithium, invisible though in the X-ray
structure.11 M1 and M2 bind two Mg2+ ions in catalytically active complexes.16 Compound 3 was placed in the active site in such a way
that the phosphonate moiety takes the place of the sulfate ion, which corresponds to the phosphate binding site in enzyme±substrate
complexes.11 In this model, one phosphonate oxygen interacts with both metal ion binding sites and the a-hydroxy group is in contact
with M2 (see also Scheme 2B). The hydrophobic part of the inhibitor occupies a pocket that extends from the metal binding sites
towards the interface between the two subunits.

1868
A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
hydroxyphosphonate moiety we took into account the
information from the published X-ray structures of
enzyme complexes with sulfate,10 inorganic phosphate12
and inositol 1-phosphate.11 These as well as kinetic and
spectroscopic evidence9,14,20,34,35 indicate strong inter-
actions between at least two metal ions and the phos-
phate moiety as main contributors to substrate anity.
Metal ion M1 binds to one of the phosphate oxygens
and to two acidic protein residues, Glu-70 and Asp-90.
It may also ligate and activate a water nucleophile for
in-line attack on the phosphoester16 (Scheme 2A). Metal
ion M2 interacts with two phosphate oxygens including
the ester oxygen, as well as three acidic residues, Asp-90,
Asp-93, and Asp-220. It should be noted that alternative
hydrolytic mechanisms have been proposed with the
water nucleophile being activated by M2 or even a third
metal ion M3 (not shown in Scheme 2A) for a non-
inline attack on the phosphoester.17,18 We assumed a
similar arrangement in our model for hydroxyphos-
phonate binding with one of the phosphonate oxygens
binding to both metal ions and an additional interaction
between the a-hydroxy group and metal ion M2
(Scheme 2B). In enzyme±substrate complexes, the 2-
and 4-OH groups of the inositol ring bind to enzyme
residues in the substrate binding pocket, in particular
Asp-93 and Glu-213 (Scheme 2A). In contrast, the aro-
matic part of 3 points away from this site and into a
region, which extends towards the interface between the
two enzyme subunits (Fig. 1, Scheme 2B). Other known
inhibitors of IMPase most likely exploit interactions
similar to those in Scheme 2: chelation of metal ions,
hydrogen bonds in the inositol binding site and addi-
tional, most likely hydrophobic, interactions in a large
empty pocket between the active site and the subunit
interface. Inositol phosphate derivatives, where the 6-
OH group of the inositol ring had been replaced with
long aliphatic/aromatic side chains are excellent inhibitors
of myo-inositol monophosphatase with Ki values as
low as 70 nM.22 It is reasonable to assume that these
side chains occupy the same pocket as the tetra-
hydronaphtyloxyphenyl group of inhibitor 3. A similar
mode of binding may be proposed for a series of
hydroxymethylenebisphosphonic acid derivatives.26 In
contrast, 4-hydroxyphenoxymethylenebisphosphonates,
in particular the thoroughly studied compound
L690330,27,29 still have the phosphoester oxygen,
instead of the a-hydroxy group. According to Scheme 2,
these compounds may rather interact with acidic resi-
dues in the inositol binding site such as Glu-213, pre-
sumably through hydrogen bonding involving the
hydroxyl group in para position of the phenyl ring.
Following a di€erent concept, inositol phosphate ana-
logues have been designed, where the 6-OH group of the
inositol ring had been replaced with an ethylene glycol
side chain, providing an additional chelating interaction
with metal ion M2.23 A totally unrelated class of inhi-
bitors, namely hydroxytropolone derivatives, have
recently been described.30,31 These appear to be very
ecient chelators of the Mg2+ ions in the active site of
IMPase, with the three hydroxyl groups of the tropolone
ring superimposing on the water and two phosphate
oxygens in Scheme 2A.
Human and bovine IMPase are 85% identical with
respect to their primary sequence.8 However, the per-
centage of identical amino acids is considerably reduced
between residues 174 and 187:
bovine 170 ETVRI I LSNI ERLLCLPI HGI 190
human 170 ETVRMVLSNMEKLFCI PVHGI 190
The X-ray structure of the human enzyme10 reveals that
this stretch corresponds to the a-helix which is the major
contact between the two enzyme subunits. This helix in
Scheme 2. Binding of d-inositol 1-phosphate (A) and proposed binding of 3 (B) to the active site of IMPase. Some interactions are not
shown for reasons of clarity.

A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
1869
one subunit is adjacent to the active site of the other
subunit. Our docking result suggests that the tetra-
hydronaphtyl part of 3 may be located closely to the
helix mentioned above, and in particular could make a
direct hydrophobic contact with residue 183, which is a
leucine in the bovine, but a phenylalanine in the human
enzyme (Fig. 1). The di€erence in size of these two resi-
dues could be an explanation for the species dependent
inhibition by 3. Alternatively, the di€erences at the
dimer interface might slightly modify the shape of the
pocket available for compound 3.
sium, because at 1 mM, the metal ion is already inhibi-
tory with human IMPase. Since inhibition is
uncompetitive with respect to Ins(1)P13, this causes the
substrate Km to drop. In fact, at 0.5 mM Mg2+, the Km
of the human enzyme for Ins(1)P increases by two±
threefold.
A di€erent assay system with 4-nitrophenyl phosphate
as the substrate was used to compare the inhibition of
the di€erent isozymes by 3. 4-Nitrophenyl phosphate
was previously characterized as
a
substrate of
IMPase.9,33 The assay is more precise and less suscep-
tible to interference with higher inhibitor concentrations
than the discontinuous measurement using molybdate
complexation of inorganic phosphate or the multi-
component coupled systems. The Km values for 4-nitro-
phenyl phosphate and the IC50 values for 3 are summarized
in Table 2. The apparent anities for the inhibitor are
about threefold higher than in the assay systems with
inositol 1-phosphate. This is due to the much higher
concentrations of Mg2+ that have to be used with the
less ecient substrate.13 (For example, the Ki for human
IMPase with 4-nitrophenyl phosphate as the substrate
increases from 75 mM at 5 mM MgCl2 to about 200 mM
at 1 mM Mg2+.) As a consequence, inhibition of the
human enzyme now becomes more easily detectable, but
the IC50 value is still about 15-fold higher as compared
to the bovine isoform.
Mutagenesis
Phe-183 in the human enzyme was replaced with leu-
cine. A summary of all kinetic constants for human and
bovine inositol monophosphatase, as well as the F183L
mutant is given in Table 1. The mutant enzyme has the
same kinetic characteristics as wild-type human IMPase.
For example, it is activated by low, but inhibited by
high concentrations of Mg2+. The activation phase is
non-linear with a Hill coecient of 2 and the optimal
concentration of Mg2+ is about 0.5 mM. In agreement
with a previous report,8 the bovine enzyme has a sig-
ni®cantly higher Ki value for the inhibition by Mg2+
,
which is also re¯ected by a shift of the optimal Mg2+
concentration from 0.5 mM for the human to about
1 mM for the bovine isoform. The di€erence in Km for
Ins(1)P is also signi®cant, but this is probably not due to
a higher anity of the human isozyme for the substrate.
Rather it is a consequence of its lower Ki for magne-
In order to correctly compare the anity of 3 for the
di€erent enzymes, a complete inhibition study was car-
ried out to determine Ki values (Table 2). Inhibition by 3
versus 4-nitrophenyl phosphate was competitive with
bovine and human IMPase, as well as F183L (data not
shown). The Ki with the bovine enzyme was 15-fold
Table 1. Kinetic and inhibition constants of IMPase isozymes
at pH 7.5 and 37 ꢀC
Constanta
Human
Isozyme F183L
Bovine
1 b
kcat(s
)
22
31
26
Table 2. Inhibition of IMPase isozymes by compound 3 at pH
7.5 and 37 ꢀC with 4-nitrophenyl phosphate as the substrate
A0.5(mM)b
opt(mM)c
0.16
0.5
1.8
0.03
2.1
0.12
0.5
2.0
0.07
1.5
0.2
1.0
1.8
0.1
5.2
A
nb
Isozyme
Km,a mM
IC50,b mM
Ki,c mM
Km(mM)d
Ki,Mg(mM)b
Bovine
Human
F183L
F183A
2.0‹0.2
1.4‹0.1
2.0‹0.2
1.8‹0.2
20‹2
310‹40
75‹10
150‹10
5.0‹0.4
75‹9
36‹2
aStandard errors were below 25% for all constants.
bkcat (turnover number), A0.5 (the concentration of magnesium
giving half-maximal velocity, n (Hill coecient), and Ki,Mg
(Mg2+ inhibition constant) were obtained by varying the con-
centration of Mg2+ over a wide range at 0.5 mM Ins(1)P.
Data were ®tted by the Hill equation, modi®ed to account
for inhibition at higher concentrations of magnesium.13 v=
kcat[E][Mg2+]n/(A0.5n+[Mg2+]n(1+[Mg2+]/Ki)). [E], the total
enzyme concentration, was determined from the UV spec-
trum.40
n.d.d
aThe substrate concentration was varied over a ninefold range
up to 10 mM at 5 mM MgCl2.
bIC50 values were determined by varying the inhibitor con-
centration between 0 and 200 mM at 5 mM MgCl2, and 2 mM
4-nitrophenyl phosphate.
cKi values were determined by varying the substrate over a
ninefold range up to 10 mM at 5 mM MgCl2 and di€erent, ®xed
concentrations of inhibitor (0, 25, 50 and 100 mM with bovine
IMPase; 0, 100, 200, and 400 mM with human IMPase; 0, 50,
100, and 200 mM with F183L). Data were ®tted with the pro-
gram COMP.41
cOptimal Mg2+ concentration, estimated from the Mg2+ acti-
vation/inhibition curve.
dMichaelis constant for Ins(1)P. The substrate concentration
was varied over a ninefold range up to 0.5 mM at 1 mM Mg2+
initial velocities were ®tted by the Michaelis±Menten equation.
,
dNot determined.

1870
A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
lower as compared to human IMPase. The F183L
mutation increases the anity of the human enzyme for
3 by about twofold. This result shows that our original
hypothesis was correct and residue 183 is somehow
involved in the binding of 3. However, it is only in part
responsible for the di€erence between human and
bovine IMPase regarding the inhibition by a-hydroxy-
phosphonates. Other amino acids must be involved
either through a direct interaction with the inhibitor or
indirectly by increasing its interaction with F183. It may
also be possible that the size of Phe-183 interferes with
the binding of 3 in the human enzyme. However, repla-
cing phenylalanine 183 with alanine was less e€ective
than the leucine mutation (Table 2). An attempt to
produce more `bovine-like' double mutants by introdu-
cing additional mutations at positions 175 and 179 was
equally unsuccessful (data not shown). This is not too
surprising, since the distance between these two residues
and the phosphate binding site in the published X-ray
structure is about 19 and 24 A, as compared to about
12 A for Phe-183 (Fig. 1).
T7 RNA polymerase plus ampicillin resistance and used
for expression of the mutant protein was a gift from
Stan Tabor (Harvard University). All oligonucleotides
were synthesized with an Applied Biosystems Model
381A DNA synthesizer using phosphoramidate chem-
istry. The Applied Biosystems automated DNA sequen-
cer Model 373A and the corresponding Taq di-deoxy
terminator sequencing kit were used for sequencing and
the analysis was aided by the University of Wisconsin
Genetic Computer Group program. The components
for the growth and propagation of all the E. coli strains
were obtained from Difco (Detroit, USA) or Sigma.
Isopropyl-b-d-thiogalactopyranoside (IPTG) was pur-
chased from Clonetech (Ozyme). The competent strain
of E. coli JM109 was supplied by Stratagene. The strain
for expression of the proteins was BL21(DE3)pLys S
and was donated by Raymond Leppick. Unless other-
wise stated, chemical reagents were obtained from com-
mercial sources and used without further puri®cation.
Tetrahydrofuran (THF) was distilled under nitrogen
from sodium/benzophenone immediately prior to use.
Di-iso-propylamine and triethylamine were distilled from
calcium hydride and stored under nitrogen over 3 A
molecular sieves. Reactions involving LDA and TMSBr
were conducted under an inert atmosphere. Drying of
the organic extract was carried out using MgSO4.
Chromatography was performed using Merck 60 230-
400 mesh silica gel according to the procedure published
by Still (abbreviation for the elution solvents are the
following: H=heptane, C=methylene chloride,
A=ethyl acetate, E=diethyl ether).38 Unless otherwise
stated, 1H NMR, 19F NMR, proton-decoupled 31P
NMR, and proton-decoupled 13C NMR spectra were
recorded in deuterated chloroform at 200, 188, 81, and
50 MHz, respectively; chemical shifts are expressed in
ppm down®eld from internal or external tetra-
methylsilane, hexa¯uorobenzene, phosphoric acid and
deuterated chloroform, respectively.
Conclusion
The a-hydroxyphosphonate 3 is a moderately potent
inhibitor of bovine myo-inositol monophosphatase, but
much less ecient on the human isozyme. This is a
somewhat surprising result, since such di€erences are
not found with other substrates or substrate analogues.
We propose that the hydrophobic side chain of 3 points
towards the interface between the two enzyme subunits
and that residue 183 at this interface is in part respon-
sible for the species selectivity. A F183L mutant has an
increased anity for the inhibitor as compared to wild-
type human IMPase. The di€erence is about twofold
and consistent with an increase in hydrophobic binding
due to one additional methylene unit.36,37 It therefore
appears that a leucine in this position contributes to the
anity of 3 for bovine IMPase through a favourable
hydrophobic interaction with the tetrahydronaphtyl
moiety of the inhibitor.
Chemistry
4-(5,6,7,8-Tetrahydronaphtyloxy)benzaldehyde (1). To a
slurry of sodium hydride (400 mg, 10 mmol, 60% grade
in mineral oil, prewashed under argon with pentane
(3Â3 mL) in dimethylformamide (DMF) (25 mL) was
added at room temperature 1-hydroxy-5,6,7,8-tetra-
hydronaphtalene (1.24 g, 10 mmol) in portions over a
5 min period of time. The resultant mixture was stirred
for 30 min and 4-¯uorobenzaldehyde (1.24 g, 10 mmol)
was added dropwise. The solution was heated at 70 ꢀC
for 2 h and cooled down. DMF was evaporated under
reduced pressure and the residue was treated with water
and extracted with methylene chloride (CH2Cl2)
(3Â20 mL). Drying of the combined organic extracts
and evaporation of the volatiles left a crude product,
which was chromatographed and eluted with H/C (75/
Experimental
Materials
dl-Inositol 1-phosphate was purchased from Bachem
and 4-nitrophenyl phosphate from Sigma. All reagents
for the polymerase chain reaction were obtained from
Perkin Elmer. The PCR products were puri®ed using
the Gene-Clean kit (Bio 101). Restriction enzymes,
other DNA modifying enzymes and corresponding buf-
fers plus the vector pUC19 were from New England
Biolabs. The vector, pT7.6, containing the promotor for

A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
1871
25) to give the desired product as a colorless oil (1.83 g,
67%). 1H NMR: d 1.70±1.88 (m, 4H), 2.55±2.62 (m,
2H), 2.80±2.88 (m, 2H), 6.78±7.20 (m, 5H), 7.78±7.88
(m, 2H), 9.90 (s, 1H).
1 h of additional stirring, glacial acetic acid (229 mg,
3.8 mmol) was added and the solution was warmed to
room temperature. It was then poured into brine
(40 mL) and extracted with CH2Cl2 (3Â20 mL). The
combined organic extracts were dried and evaporated to
give a crude reddish oil, which was chromatographed.
Elution H/E (65/35, then 7/3) furnished 1.12 g of pro-
duct 6, recrystallized from C/H (1.01 g, 53%). Melting
Dimethyl 1-[4-(5,6,7,8-tetrahydronaphtyloxy)phenyl]-1-
hydroxymethylphosphonate (2a). A mixture of aldehyde
1
(756 mg, 3 mmol), dimethylphosphite (363 mg,
point 134±135 ꢀC. H NMR: d 1.28±1.39 (m, 6H), 1.70±
1
3.3 mmol) and triethylamine (303 mg, 3.3 mmol) in
acetonitrile (6 mL) was re¯uxed for 18 h. The volatiles
were evaporated and the residue was treated with water.
Extraction of the resultant aqueous phase with CH2Cl2
(3Â20 mL), drying of the organic phases and evapora-
tion a€orded a crude yellow oil, which was subjected to
chomatography. Elution with ethyl acetate delivered
1.82 (m, 4H), 2.6±2.7 (m, 2H), 2.78±2.88 (m, 2H), 4.14±
4.34 (m, 5H), 4.98±5.20 (dm, 1H, J=20.3 Hz), 6.72 (d,
1H, J=8.2 Hz), 6.88±6.96 (m, 3H), 7.08 (t, 1H,
J=7.8 Hz), 7.44 (d, 2H, J=8.6 Hz); 19F NMR: d 36.43
(ddd, 1F, J=20.3, 105.6, 303 Hz), 47.30 (ddd, 1F,
J=6.6, 100.6, 303 Hz). 31P NMR: d 7.59 (dd, 1P,
J=100.1, 105.1 Hz). MS (C.I./NH3): 458 (M+NH4+),
441 (M+H+).
1
350 mg of colorless, oily product 2a (32%). H NMR: d
1.72±1.79 (m, 4H), 2.6±2.7 (m, 2H), 2.78±2.88 (m, 2H),
3.68 (d, 3H, J=8.4 Hz), 3.73 (d, 3H, J=8.4 Hz), 4.23
(dd, 1H, J=5.6, 9.5 Hz), 5.01 (dd, 1H, J=5.6, 10.4 Hz),
6.71 (d, 1H, J=8.0 Hz), 6.87±6.93 (m, 3H), 7.06 (t, 1H,
J=7.8 Hz), 7.38±7.48 (m, 2H); 31P NMR: d 24.26 (s,
1P). M.S. (C.I./NH3): 380 (M+NH4+), 363 (M+H+).
Diethyl 1-[4-(5,6,7,8-tetrahydronaphtyloxy)phenyl]-2-
oxo-1,1-di¯uoroethylphosphonate (8). A mixture of 2-
hydroxy-1,1-di¯uorophosphonate 2b (440 mg, 1 mmol),
pyridinium dichromate (PDC, 752 mg, 2 mmol) and 3 A
molecular sieves (440 mg, beads, 8±12 mesh) in CH2Cl2
(10 mL) was stirred for 48 h at room temperature. Di-
ethyl ether (20 mL) was then added while stirring and
the mixture was ®ltered through a plug of celite. Evap-
oration of the volatiles and chromatography of the
residual orange oil (elution: H/E (2/3)) led to the isola-
tion of 388 mg of colorless, oily product (88%). 1H
NMR: d 1.28±1.37 (t, 6H, J=6.6 Hz), 1.70±1.85 (m,
4H), 2.52±2.62 (m, 2H), 2.78±2.88 (m, 2H), 4.24±4.44
(m, 4H), 6.82 (d, 1H, J=7.8 Hz), 6.88±6.96 (m, 2H),
6.99 (d, 1H, J=6.8 Hz), 7.14 (t, 1H, J=7.7 Hz), 8.11 (d,
Di-tert-butyl 1-[4-(5,6,7,8-tetrahydronaphtyloxy)phenyl]-
1-hydroxymethylphosphonate (2b). Di-tert-butyl phos-
phite (776 mg, 4 mmol) was added dropwise to a cooled
( 78 ꢀC) THF/hexanes solution of LDA (4 mmol), pre-
pared at 0 ꢀC (30 min of stirring) from diisopropylamine
(405 mg, 4 mmol) in THF (6 mL) and n-butyllithium
(2.5 mL of a 1.6 N solution in hexanes, 4 mmol). The
resultant milky mixture was stirred for 45 min, after
which a THF (2 mL) solution of aldehyde 1 (1009 mg,
4 mmol) was added drop by drop. After stirring for an
additional 20 min at the same temperature, the mixture
was warmed up to room temperature and stirring was
continued for 1 h. The solution was poured into brine
(40 mL) extracted with diethyl ether (3Â20 mL) and the
combined extracts were dried. Evaporation and chroma-
tography (elution: H/E (1/4)) gave 1.38 g of product 2b,
which was recrystallized from C/H (1.34 g in 3 crops,
75%). Melting point 150±151 ꢀC. 1H NMR: d 1.41 (s,
9H), 1.45 (s, 9H), 1.72±1.79 (m, 4H), 2.6±2.7 (m, 2H),
2.78±2.88 (m, 2H), 4.41 (dd, 1H, J=5.0, 11.8 Hz), 4.81
(dd, 1H, J=5.0, 9.5 Hz), 6.68 (d, 1H, J=7.3 Hz), 6.87±
6.93 (m, 3H), 7.04 (t, 1H, J=7.8 Hz), 7.38±7.48 (m, 2H);
31P NMR: d 14.45 (s, 1P). M.S. (C.I./NH3): 464
(M+NH4+), 447 (M+H+).
2H, J=9.0 Hz); 19F NMR:
d 51.95 ppm (d, 1F,
J=96.4 Hz); 31P NMR: d 4.34 (t, 1P, J=96.4 Hz). M.S.
(C.I./NH3): 439 (M+NH4+), 457 (M+H+).
Ditert-butyl 1-[4-(5,6,7,8-tetrahydronaphtyloxy)phenyl]-
1-oxo-methylphosphonate (4). The procedure hereabove
was used starting from di-tert-butyl 1-hydroxyphos-
phonate 2b (981 mg, 2.2 mmol). Chromatography and
elution (H/E (3/7) gave 802 mg of light-yellow, oily
product (82%). 1H NMR: d 1.54 (s, 18H), 1.72±1.79 (m,
4H), 2.53±2.63 (m, 2H), 2.78±2.88 (m, 2H), 6.82 (d, 1H,
J=7.8 Hz), 6.87±6.94 (m, 2H), 6.98 (d, 1H, J=7.8 Hz),
7.13 (t, 1H, J=7.8 Hz), 8.22±8.32 (m, 2H); 31P NMR: d±
7.0 (s, 1P).
Diethyl 1-[4-(5,6,7,8-tetrahydronaphtyloxy)phenyl]-2-
hydroxy-1,1-di¯uoroethylphosphonate (6). To a THF/
hexanes solution of LDA (3.8 mmol), prepared as above
and cooled at 78 ꢀC, was added dropwise a THF
(2 mL) solution of diethyl di¯uoromethylphosphonate
(716 mg, 3.8 mmol) and stirring was continued for
30 min. To the resultant orange solution was added a
THF solution of aldehyde 1 (960 mg, 3.8 mmol). After
1-[4-(5,6,7,8-Tetrahydronaphtyloxy)phenyl]-1-oxomethyl-
phosphonic acid (5). Di-tert-butyl 1-oxophosphonate 4
(223 mg, 0.5 mmol) was dissolved in tri¯uoroacetic acid
(5 mL) and the solution was stirred for 2 h at room
temperature. Removal of the solvent under reduced
pressure left 148 mg of pure product (1H and 31P NMR
spectroscopy) (83%). Melting point 164±165 ꢀC. 1H
NMR (CDCl3/DMSO-d6): d 1.68±1.80 (m, 4H), 2.50±

1872
A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
2.60 (m, 2H), 2.75±2.85 (m, 2H), 6.78 (d, 1H,
J=7.8 Hz), 6.82±6.92 (m, 2H), 6.97 (d, 1H, J=7.6 Hz),
7.11 (t, 1H, J=7.7 Hz), 8.10±8.32 (m, 2H); 31P NMR
(CDCl3/DMSO-d6): d 1.0 (br. s, 1P). MS (C.I./NH3):
350 (M+NH4+). Anal. calcd for C17H17O5P: C: 61.45;
H: 5.16. Found: C: 60.96; H: 5.13.
after recrystallization from chloroform/heptane. Melt-
ing point 186±187 ꢀC. 1H NMR (CDCl3/DMSO-d6): d
1.70±1.82 (m, 4H), 2.52±2.62 (m, 2H), 2.78±2.88 (m,
2H), 6.80 (d, 1H, J=7.8 Hz), 6.86±6.93 (m, 2H), 6.93 (d,
1H, J=7.2 Hz), 7.13 (t, 1H, J=7.7 Hz), 8.16 (d, 2H,
J=8.9 Hz); 19F NMR (CDCl3/DMSO-d6): d 50.52 ppm
(d, 1F, J=92.4 Hz); 31P NMR (CDCl3/DMSO-d6): d
2.62 (t, 1P, J=93.1 Hz). MS (C.I./NH3): 439
(M+NH4+), 457 (M+H+). Anal. calcd for
C18H17F2O5P: C: 56.55; H: 4.48. Found: C: 56.45; H:
4.38.
Dealkylation using TMSBr/H2O sequence. Trimethyl-
silyl bromide (3.3 equiv for 2a and 6; 2.2 equiv for 8)
was added at room temperature to a solution of starting
dialkyl phosphonate (1 equiv) in acetonitrile (4 mL per
mmol of starting material). Stirring was continued for
24 h and the solution was then concentrated by evap-
oration of the volatiles under reduced pressure. The
residue was taken up with toluene (3 mL) and evapo-
rated; this was repeated twice to ensure the complete
removal of any HBr. The residual oil was dissolved in
acetonitrile (3 mL) and water was (3 mL) added. After
16 h of stirring, the solution was evaporated to furnish
the pure product.
Molecular modeling
Modeling was done using Sybyl 5.32. (Tripos Ass., St.
Louis, MO 62177). A model for the bovine IMPase
structure was generated by amino acid replacement in
the X-ray structure of the human enzyme, followed by
energy optimization with the Kollman all atom para-
meter set for 200 steps using a conjugate gradient mini-
mizer. Docking of 3 in the active site was done manually
and aimed at optimizing the interactions of the phos-
phonate group with the bimetallic center in the enzyme
and at avoiding unfavorable steric contacts. It is
obvious that the resulting complex is tentative but useful
for giving a possible explanation for the species di€er-
ences which could be tested experimentally.
1-[4-(5,6,7,8-Tetrahydronaphtyloxy)phenyl]-1-hydroxy-
methylphosphonic acid (3). The reaction was carried out
on a 0.5 mmol scale; 167 mg (100%) of creamy powder.
Melting point 153±154 ꢀC. H NMR (CD3CN/DMSO-
1
d6): d 1.70±1.80 (m, 4H), 2.54±2.64 (m, 2H), 2.72±2.82
(m, 2H), 4.75 (d, 1H, J=12.5 Hz), 6.68 (d, 1H,
J=7.2 Hz), 6.82 (d, 2H, J=8.5 Hz), 6.90 (d, 1H,
J=7.7 Hz), 7.07 (t, 1H, J=7.7 Hz), 7.35±7.45 (m, 2H);
13C NMR (CD3CN/DMSO-d6): d 23.23, 23.42, 23.99,
29.99, 70.94 (d, J=161 Hz), 117.21, 118.34, 125.87,
127.15, 129.83, 129.93, 134.29, 140.43, 154.85, 158.04;
31P NMR (CD3CN/DMSO-d6): d 22.30 (br, s, 1P).
Anal. calcd for C17H19O5P: C: 61.08; H: 5.73. Found: C:
60.25; H: 5.52.
Enzyme assays
For the determination of IC50 values and with inositol
1-phosphate as the substrate, release of inorganic phos-
phate was measured in a discontinuous assay, based on
the method by Attwood et al.33 The enzyme (5±10 pmol)
was incubated at 37 ꢀC in 50 mM Tris±Cl, pH 7.5, in the
presence of 0.2 mM (2ÂKm) inositol 1-phosphate,
0.5 mM (with human IMPase) or 1 mM (bovine
IMPase) MgCl2, and di€erent concentrations of inhi-
bitor. The total assay volume was 0.5 mL. Aliquots of
100 mL were removed at di€erent time points and added
to 100 mL of a phosphate reagent in a microtiter plate.
The absorbance at 340 nm was determined in a plate
reader (Dynatech MR 5000) and the absorbance change
per minute calculated by linear regression. Alternatively,
in particular for the more detailed kinetic studies
(Table 1), continuous, coupled assay systems were
used.9,13 With 4-nitrophenyl phosphate, the formation
of nitrophenolate was monitored at 400 nm
(e=16.7Â103 M 1 cm 1).
1-[4-(5,6,7,8-Tetrahydronaphtyloxy)phenyl]-2-hydroxy-
1,1-di¯uoroethylphosphonic acid (7). The reaction was
carried out on a 0.5 mmol scale; 199 mg (100%) of white
solid. Melting point 165±166 ꢀC. 1H NMR (CD3CN/
DMSO-d6): d 1.70±1.80 (m, 4H), 2.55±2.65 (m, 2H),
2.74±2.84 (m, 2H), 5.05 (dd, 1H, J=5.0, 22.7 Hz), 6.71
(d, 1H, J=7.9 Hz), 6.84 (d, 2H, J=8.7 Hz), 6.93 (d, 1H,
J=7.5 Hz), 7.10 (t, 1H, J=7.8 Hz), 7.40 (d, 2H,
J=8.5 Hz); 13C NMR (CD3CN/DMSO-d6): d 23.23,
23.44, 23.99, 29.99, 73.23 (m), 117.06, 118.31, 126.10,
127.21, 130.15, 130.84, 131.74, 140.55, 154.66, 158.94;
19F NMR (CD3CN/DMSO-d6): d 34.35 (ddd, 1F,
J=22.7, 101.0, 297.4 Hz), 47.30 (ddd, 1F, J=4.2, 96.4,
297.4 Hz); 31P NMR (CD3CN/DMSO-d6): d 7.64 (t, 1P,
J=99.0 Hz). Anal. calcd for C18H19F2O5P.0.75H2O: C:
54.34; H: 5.19. Found: C: 54.71; H: 5.14.
Mutagenesis
The wild-type clone of human myo-inositol monophos-
phatase was obtained by RT-PCR as previously
described for the bovine enzyme.9 Oligonucleotides were
based on the published sequence for the human isoform.8
1-[4-(5,6,7,8-Tetrahydronaphtyloxy)phenyl]-2-oxo-1,1-di-
¯uoroethylphosphonic acid (9). The reaction was carried
out on a 0.6 mmol scale; 128 mg (55%) of white solid

A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
1873
The mutant clones were prepared by PCR using
the technique of overlap extension39 and the GeneAmp
reagents (Perkin Elmer) according to the manufacturer's
instructions. For F183L the following oligonucleotide
primers were used: P1 (50-CTAATATGGAAAAGCT-
TCTTTGCATTCCTGTT-30) and P2 (50-GAACAG-
GAATGCAAAGAAGCTTTTCCATATTAG-30) are
complementary and contain the mutant sequence as
noted in bold print. P3 (50-GAGCTCCTGCAGA-
GGAGGAATATAATATGGC-TGATCCTT-30) and
P4 (50-GTCGACTCTAGATTAATCTTCGTCGTC-30)
¯ank the entire coding sequence of the human myo-
inositol monophosphatase cDNA plus restriction sites
for subcloning purposes (outlined in bold print). P3 also
contains a ribosome binding site with spacer for pro-
karyotic expression (underlined) and the start codon,
ATG. The same strategy was applied for the other
mutants: P3 and P4 were kept the same and P1 and P2
designed according to the desired target sequence. The
full length mutated DNA was digested with Pst I and
Xba I and subcloned ®rst into the vector pUC19 for
sequencing and subsequently into the expression vector,
pT7.6. The proteins were expressed in E. coli strain
BL21(DE3)pLysS and puri®ed to apparent homo-
geneity as previously described for the bovine enzyme.9
The concentrations of the puri®ed enzymes were deter-
9. Strasser, F.; Pelton, P.D.; Ganzhorn, A. J. Biochem. J. 1995,
307, 585.
10. Bone, R.; Springer, J. P.; Atack, J. R. Proc. Natl. Acad.
Sci. U.S.A. 1992, 89, 10031.
11. Bone, R.; Frank, L.; Springer, J. P.; Pollack, S. J.;
Osborne, S.; Atack, J. R.; Knowles, M. R.; McAllister, G.;
Ragan, C. I.; Broughton, H. B.; Baker, R.; Fletcher, S. R.
Biochemistry 1994, 33, 9460.
12. Bone, R.; Frank, L.; Springer, J. P.; Atack, J. R. Biochem-
istry 1994, 33, 9468.
13. Ganzhorn, A. J.; Chanal, M. C. Biochemistry 1990, 29,
6065.
14. Greasley, P. J.; Gore, M. G. FEBS Lett. 1993, 331, 114.
15. Cole, A. G.; Gani, D. J. Chem. Soc., Chem. Commun. 1994,
1139.
16. Pollack, S. J.; Atack, J. R.; Knowles, M. R.; McAllister,
G.; Ragan, C. I.; Baker, R.; Fletcher, S. R.; Iversen, L. L.;
Broughton, H. B. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
5766.
17. Wilkie, J.; Cole, A.G.; Gani, D. J. Chem. Soc., Perkin
Trans. 1995, 1, 2709.
18. Ganzhorn, A. J.; Lepage, P.; Pelton, P. D.; Strasser, F.;
Vincendon, P.; Rondeau, J.-M. Biochemistry 1996, 35, 10957.
19. Ganzhorn A. J.; Rondeau, J.-M. Protein Engng 1997, 10
(Suppl.), 61.
20. Leech, A. P.; Baker, G. R.; Shute, J. K.; Cohen, M. A.;
Gani, D. Eur. J. Biochem. 1993, 212, 693.
mined from the UV spectra using an extinction coe-
cient e=26 200 M 1 cm
1 40
21. Atack, J. R.; Fletcher, S. R. Drugs Fut. 1994, 19, 857.
.
Speci®c activities were
22. Baker, R.; Carrick, C.; Leeson, P. D.; Lennon, I. C.;
Liverton, N. J. J. Chem. Soc., Chem. Commun. 1991, 298.
15 U/mg for the wild-type human enzyme, 24 U/mg for
F183L, 19 U/mg for F183A, 19 U/mg for F183L/M179I,
18 U/mg for F183L/V175I, and 24 U/mg for the bovine
enzyme. They were determined in a standard assay with
4 mM 2-glycerophosphate and 2 mM MgCl2 at pH 7.5
and 37 ꢀC.33
23. Schulz, J.; Wilkie, J.; Lightfoot, P.; Rutherford, T.; Gani,
D. J. Chem. Soc., Chem. Commun. 1995, 2353.
24. MacLeod, A. M.; Baker, R.; Hudson, M.; James, K.; Roe,
M. B.; Knowles, M.; MacAllister, G. Med. Chem. Res. 1992, 2,
96.
25. Kulagowski, J. J.; Baker, R.; Fletcher, S. R. J. Chem. Soc.,
Chem. Commun. 1991, 1649.
References and Notes
1. Hallcher, L. M.; Sherman, W. R. J. Biol. Chem. 1980, 255,
10896.
26. Fletcher, S. R.; Baker, R.; Leeson, P. D.; Teall, M.;
Harley, E. A.; Ragan, C. I. Bioorg. Med. Chem. Lett. 1992, 2,
627.
2. Berridge, M. J.; Downes, C. P.; Hanley, M. R. Cell 1989, 59,
411.
27. Fletcher, S. R.; Baker, R.; Ladduwahetty, T.; Sharpe, A.;
Teall, M.; Atack, J. R. Bioorg. Med. Chem. Lett. 1993, 3,
141.
3. Nahorski, S. R.; Ragan, C. I.; Challiss, R. A. J. Trends
Pharmacol. Sci. 1991, 12, 297.
4. Gani, D.; Downes, C. P.; Batty, I.; Bramham, J. Biochim.
Biophys. Acta 1993, 1177, 253.
28. van Steijn, A. M. P.; Willems, H. A. M.; de Boer, T.;
Geurts, J. L. T.; van Boeckel, C. A. A. Bioorg. Med. Chem.
Lett. 1995, 5, 469.
5. Belmaker, R.H.; Bersudsky, Y.; Benjamin, J.; Agam, G.;
Levine, J.; Kofman, O. In Depression and Mania: from Neuro-
biology to Treatment; Gessa, G.; Fratta, W.; Pani, L.; Serra, G.,
Eds.; Raven: New York, 1995; pp 67±84.
29. Atack, J. R.; Cook, S. M.; Watt, A. P.; Fletcher, S. R.;
Ragan, C. I. J. Neurochem. 1993, 60, 652.
30. Piettre, S. R.; Ganzhorn, A. J.; Ho¯ack, J.; Islam, K.;
Hornsperger, J.-M. J. Am. Chem. Soc. 1997, 119, 3201.
6. Spector, R; Lorenzo, A. V. Am. J. Physiol. 1975, 228, 1510.
7. Diehl, R. E.; Whiting, P.; Potter, J.; Gee, N. S.; Ragan, C. I.;
Linemeyer, D.; Schoepfer, R.; Bennett, C. D.; Dixon, R. A. F.
J. Biol. Chem. 1990, 265, 5946.
31. Piettre, S. R.; Andre, C.; Chanal, M-C.; Ducep, J-B.;
Lesur, B.; Piriou, F.; Raboisson, P.; Rondeau, J-M.; Schelcher,
C.; Zimmermann, P.; Ganzhorn, A. J. J. Med. Chem. 1997, 40,
4208.
8. McAllister, G.; Whiting, P.; Hammond, E. A.; Knowles,
M. R.; Atack, J. R.; Bailey, F. J.; Maigetter, R.; Ragan, C. I.
Biochem. J. 1992, 284, 749.
32. The corresponding a-¯uorophosphonate i and a,a,b-tri-
¯uorophosphonate ii were also prepared by treating 2a and 6

1874
A. J. Ganzhorn et al./Bioorg. Med. Chem. 6 (1998) 1865±1874
with diethylaminosulfur tri¯uoride (DAST) at room tempera-
ture; however the products could not be converted to the cor-
responding acids. No reaction was observed between
compound 8 and DAST.
35. Saudek, V.; Vincendon, P.; Do, Q.-T.; Atkinson, R. A.;
Sklenar, V.; Pelton, P. D.; Piriou, F.; Ganzhorn, A. J. Eur. J.
Biochem. 1996, 240, 288.
36. Hansch, C.; Schae€er, J.; Kerley, R. J. Biol. Chem. 1972,
247, 4703.
37. Ganzhorn, A. J.; Green, D. W.; Hershey, A. D.; Gould,
R. M.; Plapp, B. V. J. Biol. Chem. 1987, 262, 3754.
38. Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43, 2923.
39. Ho, S. N.; Hunt, H. D.; Horton, R. M.; Pullen, J. K.;
Pease, L. R. Gene 1989, 77, 51.
33. Attwood, P. V.; Ducep, J.-B.; Chanal, M.-C. Biochem. J.
1988, 253, 387.
40. Ganzhorn, A. J.; Vincendon, P.; Pelton, J. T. Biochim.
Biophys. Acta 1993, 1161, 303.
34. Greasley, P. J.; Hunt, L. G.; Gore, M. G. Eur. J. Biochem.
1994, 222, 453.
41. Cleland, W. W. Meth. Enzymol. 1979, 63, 103.