Synthesis of (S)-isoprenoid thiodiphosphates as substrates and inhibitors

Article abstract of DOI:10.1021/jo010505n

Thiolo thiophosphate analogues of isopentenyl diphosphate (IPP), dimethylallyl diphosphate (DMAPP), geranyl diphosphate (GPP), farnesyl diphosphate (FPP), and geranylgeranyl diphosphate (GGPP) were synthesized. Inorganic thiopyrophosphate (SPP_i) was prepared from trimethyl phosphate in four steps. The tris(tetra-n-butylammonium) salt was then used to convert isopentenyl tosylate to (S)-isopentenyl thiodiphosphate (ISPP). (S)-Dimethylallyl (DMASPP), (S)-geranyl (GSPP), (S)-farnesyl (FSPP), and (S)-geranylgeranyl thiodiphosphate (GGSPP) were prepared from the corresponding bromides in a similar manner. ISPP and GSPP were substrates for avian farnesyl diphosphate synthase (FPPase). Incubation of the enzyme with ISPP and GPP gave FSPP, whereas incubation with IPP and GSPP gave FPP. GSPP was a substantially less reactive than GPP in the chain elongation reaction and was an excellent competitive inhibitor, K_I^GSPP=24.8 μM, of the enzyme. Thus, when ISPP and DMAPP were incubated with FPPase, GSPP accumulated and was only slowly converted to FSPP.

Full text of DOI:10.1021/jo010505n

J . Org. Chem. 2001, 66, 6705-6710
6705
Syn th esis of (S)-Isop r en oid Th iod ip h osp h a tes a s Su bstr a tes a n d
In h ibitor s
Richard M. Phan and C. Dale Poulter*
Department of Chemistry, University of Utah, Salt Lake City, Utah 84112
poulter@chemistry.utah.edu
Received May 18, 2001
Thiolo thiophosphate analogues of isopentenyl diphosphate (IPP), dimethylallyl diphosphate
(DMAPP), geranyl diphosphate (GPP), farnesyl diphosphate (FPP), and geranylgeranyl diphosphate
(GGPP) were synthesized. Inorganic thiopyrophosphate (SPPi) was prepared from trimethyl
phosphate in four steps. The tris(tetra-n-butylammonium) salt was then used to convert isopentenyl
tosylate to (S)-isopentenyl thiodiphosphate (ISPP). (S)-Dimethylallyl (DMASPP), (S)-geranyl (GSPP),
(S)-farnesyl (FSPP), and (S)-geranylgeranyl thiodiphosphate (GGSPP) were prepared from the
corresponding bromides in a similar manner. ISPP and GSPP were substrates for avian farnesyl
diphosphate synthase (FPPase). Incubation of the enzyme with ISPP and GPP gave FSPP, whereas
incubation with IPP and GSPP gave FPP. GSPP was a substantially less reactive than GPP in the
chain elongation reaction and was an excellent competitive inhibitor, KIGSPP ) 24.8 µM, of the
enzyme. Thus, when ISPP and DMAPP were incubated with FPPase, GSPP accumulated and was
only slowly converted to FSPP.
In tr od u ction
Prenyl transfers are nucleophilic substitution reac-
tions. Mechanistic studies indicate that the reactions are
associative for strong nucleophiles, such as the zinc
thiolate in protein prenylation, but proceed through a
“late” transition state with substantial development of
positive charge in the allylic moiety.8 For weak nucleo-
philes, such as a carbon-carbon double bond, the reaction
appears to occur by a dissociative mechanism.9 We sought
to obtain direct evidence for the dissociative reaction by
detecting recombination of the allylic cation and PPi
through positional isotope exchange of oxygen atoms in
the pyrophosphate leaving group10 and by an enzyme-
catalyzed thiono f thiolo rearrangement of thiodiphos-
phate analogues.11 Thiono f thiolo isomerization was a
particularly attractive technique for detecting ion-pair
recombination. Model studies showed that (O)-dimethyl-
allyl-(O,O)-dimethyl thiophosphate readily underwent
ion-pair return accompanied by both allylic and thiono
f thiolo rearrangement during solvolysis12 and that the
allylic thiolo derivatives were ∼106-fold less reactive than
their thiono counterparts. However, neither positional
isotope exchange nor thiono to thiolo rearrangements
were detected in these experiments. Recently published
X-ray structures for binary complexes of FPPase and
allylic diphosphate substrates show that the negatively
charged nonbridging oxygen atoms are coordinated to
Isoprenoid molecules are constructed from simple five-
carbon building blocks through a series of fundamental
“building reactions” in the central part of the pathway
mediated by prenyltransferases. These enzymes catalyze
the alkylation of electron-rich functional groups by allylic
isoprenoid diphosphates. Some common prenyl transfer
reactions include the chain elongation of dimethylallyl
diphosphate (DMAPP) to farnesyl diphosphate (FPP)1
by sequential addition of two molecules of isopentenyl
diphosphate (IPP),2 the joining of two molecules of
farnesyl diphosphate (FPP) to give squalene,3 the post-
translational modification of proteins by attachment of
farnesyl and geranylgeranyl groups,4 and the modifica-
tion of adenosine residues in tRNA with a dimethylallyl
moiety.5 Isoprenoid cyclases catalyze intramolecular ver-
sions of prenyl-transfer reactions by electrophilic alky-
lation of distal double bonds to create a variety of cyclic
isoprenoids.6 The intermediates produced by prenyl
transfer and cyclization reactions are ultimately trans-
formed into over 30 000 different metabolites7 that
perform a variety of functions in their respective hosts.
* To whom correspondence should be addressed. Phone: (801) 581-
6685. Fax: (801) 581-4391.
(1) Abbreviations used: BHDA, bicyclo[2.2.1]hept-5-ene-2,3-dicar-
boxylic acid; BME, 2-mercaptoethanol; BSA, bovine serum albumin;
DMAPP, dimethylallyl diphosphate; DMASPP, (S)-dimethylallyl thio-
diphosphate; FPP, farnesyl diphosphate; FPPase, farnesyl diphosphate
synthase; FSPP, (S)-farnesyl thiodiphosphate; GGPP, geranylgeranyl
diphosphate; GPP, geranyl diphosphate; GGSPP, (S)-geranylgeranyl
thiodiphosphate; GSPP, (S)-geranyl thiodiphosphate; IPP, isopentenyl
diphosphate; ISPP, (S)-isopentenyl thiodiphosphate; PPi, inorganic
pyrophosphate salt; SPPi, inorganic thiopyrophosphate salt.
(2) Poulter, C. D.; Rilling, H. C. Prenyltransferase. The Mechanism
of the Reaction. Biochemistry 1976, 15, 1079-1083.
(7) Dictionary of Organic Compounds; Chapman and Hall: London,
1996.
(8) (a) Dolence, J . M.; Poulter, C. D. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 5008-5011. (b) Harris, C. M.; Poulter, C. D. Recent Studies
of the Mechanism of Protein Prenylation. Nat. Prod. Rep. 2000, 17,
131-214.
(9) (a) Davisson, V. J .; Neal, T. R.; Poulter, C. D. J . Am. Chem. Soc.
1993, 115, 1235-1245. (b) Davisson, V. J .; Poulter, C. D. J . Am. Chem.
Soc. 1993, 115, 1245-1260.
(3) J arstfer, M. B.; Blagg, B. S. J .; Rogers, D. H.; Poulter, C. D. J .
Am. Chem. Soc. 1996, 118, 13089-13090.
(10) Mash, E. A.; Gurria, G. M.; Poulter, C. D. J . Am. Chem. Soc.
1981, 103, 3927-3929.
(4) Mayer, M. P.; Prestwich, G. D.; Dolence, J . M.; Bond, P. D.; Wu,
H.; Poulter, C. D. Gene 1993, 132, 41-47.
(11) Yi, E. C. Ph.D. Thesis, University of Utah, 1999.
(12) Poulter, C. D.; Mautz, D. S. J . Am. Chem. Soc. 1991, 113, 4895-
4903.
(5) Caillet, J .; Droogmans, L. J . Bacteriol. 1988, 170, 4147-4152.
(6) Cane, D. E. Acc. Chem. Res. 1985, 18, 220-226.
10.1021/jo010505n CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/13/2001

6706 J . Org. Chem., Vol. 66, No. 20, 2001
Phan and Poulter
Sch em e 1. Syn th esis of
positively charged groups in the active site.13 Presumably,
these interactions are sufficiently strong to prevent bond
rotations in the enzyme-bound pyrophosphate group in
the allylic cation-PPi complex.
Tr is(tetr a -n -bu tyla m m on iu m ) Th iop yr op h osp h a te
Several groups have synthesized prenyltransferase
inhibitors by replacing the diphosphate leaving group
with less reactive or unreactive “diphosphate” mimics.
These include phosphonophosphates,14 phosphonophos-
phinates,15 and bisphosphonates.16 Typically, these com-
pounds are modest to poor prenyltransferase inhibitors.
Although the structures of the analogues do not precisely
duplicate those of the normal substrates, these differ-
ences do not appear to adequately account for their
typically poor binding properties. Diphosphate esters are
trianions at physiological pH. One possible reason for the
loss in binding affinity is the increase in the pKa’s of the
analogues as the bridging oxygen atoms in the diphos-
phate group are replaced by less electron-withdrawing
carbons.17 These considerations led us to design a new
class of prenyltransferase inhibitors with structures and
ionization properties very similar to their normal sub-
strates by replacing the diphosphate oxygen atom at-
tached to the isoprenoid moiety with sulfur.18
a
(a) nBu4NOH; (b) (MeO)2P(S)Cl; (c) TMSI; (d) nBu4NOH/H2O.
Sch em e 2. Syn th esis of (S)-Alk yl
Th iod ip h osp h a tes
Resu lts a n d Discu ssion
Syn th esis of Tr is(tetr a -n -bu tyla m m on iu m ) Th io-
p yr op h osp h a te. We based our syntheses of (S)-isopen-
tenyl thiodiphosphate (ISPP), (S)-dimethylallyl (DMASPP),
(S)-geranyl (GSPP), (S)-farnesyl (FSPP), and (S)-geran-
ylgeranyl thiodiphosphate (GGSPP) on the general pro-
cedure for isoprenoid diphosphates developed by Davis-
son et al.19 This method uses the tris(tetra-n-butyl-
ammonium) salt of inorganic pyrophosphate to convert
activated isoprenoid compounds to the corresponding
diphosphate esters by nucleophilic displacement. We
reasoned that the sulfur atom in inorganic thiodiphos-
phate would be sufficiently more nucleophilic than the
oxygen atoms to produce the (S)-alkyl thiodiphosphates
regioselectivity.
The synthesis of tris(tetra-n-butylammonium) thio-
pyrophosphate (SPPi) is shown in Scheme 1. Trimethyl
phosphate (1) was selectively monodemethylated with
tetrabutylammonium hydroxide. Monodemethylation was
also seen for sodium hydroxide, ammonium hydroxide,
and trimethylamine; however, the sodium, ammonium,
and tetramethylammonium SPPi salts were not suf-
ficiently soluble in organic solvents and had to be
converted to the tetra-n-butylammonium form by ion-
exchange chromatography before proceeding. The syn-
thesis of anhydride 3 from methyl phosphate (2) and
a
(a) Tris-(tetra-n-butylammonium)thiopyrophosphate, CH3CN.
+
(b) Dowex AG 50W-X8 (NH4 form).
dimethyl thiophosphorochloridate was touchy. The best
yields were achieved when the progress of the reaction
was carefully monitored at -35 °C. Tetramethylthio-
phosphate 3 was unstable upon extended exposure to 2
or dimethyl thiophosphorochloridate. Thus, the cold
reaction mixture was passed through a silica column to
separate 3 from the other components in the reaction
mixture. The anhydride purified in this manner can be
stored at -20 °C.
The methyl groups were removed from anhydride 3 by
treatment with trimethylsilyl iodide (TMSI) at -35 °C.
The mechanism for dealkylation, displacement of iodide
from TMSI by the phosphoryl oxygen or thiophosphoryl
sulfur followed by a second displacement at the methyl
groups by iodide,20 predicts formation of thiolo-TMS
diphosphate 4. The TMS derivative obtained from the
reaction could not be chromatographed and was treated
with tetra-n-butylammonium hydroxide without purifica-
tion to give [(n-Bu)4N]3(SPPi). The salt is stable for up to
a year when stored as a lyophilized powder at -80 °C.
Attempts to remove the methyl groups with trimethyl-
silyl bromide (TMSBr) or TMSI generated in situ by
treatment TMSBr with NaI were not successful.
(13) Tarshis, L. C.; Proteau, P. J .; Kellogg, B. A.; Sacchettini, J . C.;
Poulter, C. D. Biochemistry 1996, 35, 1518-15023.
(14) (a) Corey, E. J .; Volante, R. R. J . Am. Chem. Soc. 1976, 98,
1291-1293. (b) Cane, D. E.; Yang, G.; Xue, Q.; Shim, J . H. Biochemistry
1995, 34, 2471-2479.
(15) McClard, R. W.; Fujita, T. S.; Stremler, K. E.; Poulter, C. D. J .
Am. Chem. Soc. 1987, 109, 5544-5545.
(16) Stremler, K. E.; Poulter, C. D. J . Am. Chem. Soc. 1987, 109,
5542-5544.
(17) (a) Blackburn, G. M.; England, D. A.; Kolkmann, F. J . Chem.
Soc., Chem. Commun. 1981, 930-932. (b) Blackburn, G. M.; Kent, D.
E.; Kolkmann, F. J . Chem. Soc., Chem. Commun. 1981, 1188-1190.
(c) Blackburn, G. M.; Eckstein, F.; Kent, D. E.; Perree, T. D. Nucleosides
Nucleotides 1985, 4, 165-167.
(18) Phan, R. M.; Poulter, C. D. Organic Lett. 2000, 2, 2287-2289.
(19) (a) Davisson, V. J .; Woodside, A. B.; Poulter, C. D. Method
Enzymol. 1984, 110, 130-144. (b) Davisson, V. J .; Woodside, A. B.;
Neal, T. R.; Stremler, K. E.; Poulter, C. D. J . Org. Chem. 1986, 51,
4768-4778.
Syn th esis of (S)-Alk yl Th iod ip h osp h a tes. (S)-Alkyl
isopentenyl and allylic thiodiphosphates were obtained
by the procedure of Davisson et al.19a with minor modi-
fications. As illustrated in Scheme 2, ISPP was prepared
from isopentenyl tosylate (5), while DMASPP, GSPP,
(20) (a) Blackburn, G. M.; Ingleson, K. J . Chem. Soc., Chem.
Commun. 1978, 870-871. (b) Zygmunt, J .; Kafarski, P.; Matalerz, P.
Synthesis 1978, 609.

Synthesis of (S)-Isoprenoid Thiodiphosphates
J . Org. Chem., Vol. 66, No. 20, 2001 6707
FSPP, and GGSPP were prepared from the corresponding
allylic bromides. Previously, attempts to synthesize IPP
from isopentenyl chloride or bromide gave poor yields
because of a competing elimination that was not seen
when the displacement reaction was conducted with
tosylate 5.19b DMASPP, GSPP, and FSPP were synthe-
sized from the corresponding commercially available
bromides. GGSPP was prepared from geranylgeranyl
bromide19b obtained from geranylgeraniol.
Typically, the activated isoprenoid derivatives were
slowly added to acetonitrile solutions containing 2-3
equiv of SPPi cooled to -35 °C. After 30-45 min, the
reaction was complete, and the solvent was removed at
reduced pressure. Although sulfur is substantially more
nucleophilic than oxygen, excess SPPi was used, and the
reaction was run at -30 °C to ensure high regioselectivity
for displacement by sulfur. When the order of addition
of the reactants was reversed, small amounts diphos-
phate esters were formed from displacement by oxygen.
The residue was dissolved in water and passed through
an ion-exchange column replace the tetra-n-butylammo-
nium cations with ammonium. This step was taken
because the tetrabutylammonium salts streaked when
chromatographed on cellulose and exchange of tetrabu-
tylammonium for ammonium facilitated the final puri-
fication step by chromatography on cellulose.
F igu r e 1. Products from incubation of avian FPPase and [14C]-
IPP with GSPP or GPP: lanes 1 and 2, incubations with GPP;
lanes 3 and 4, incubations with GSPP; lane 5, [14C]IPP; lane
6, [14C]FPP.
F igu r e 2. Products from incubation of avian FPPase and [14C]-
DMAPP with IPP or ISPP: lanes 1 and 2, incubations with
IPP; lanes 3 and 4, incubations with ISPP; lane 5, [14C]DMPP;
lane 6, [14C]FPP.
The ammonium salts of allylic isoprenoid diphosphates
decompose upon prolonged lyophilization because the
tris-ammonium forms are converted to less stable di- or
mono-ammonium species due to removal of NH3 under
vacuum. Decomposition is less of a problem for the
ammonium salts of the corresponding allylic thiolo-
diphosphates because of the increased kinetic stability
of the thiolo linkage. However, we always store all of
these materials as lyophilized powders at -80 °C.
The thiolo structures for the thiodiphosphates were
confirmed from their NMR spectra. The 1H-decoupled 31P
spectra have characteristic AB patterns with doublets
(J ∼ 29 Hz) at approximately 8 and -8 ppm for PR and
Pâ, respectively. In 1H-coupled 31P spectra, the doublet
at 8 ppm was further split into triplets (J ∼ 8 Hz) by the
with GPP, while only 4% of the radioactivity comigrated
with FPP when GSPP was the allylic substrate (see
Figure 1).
In the normal chain elongation reaction catalyzed by
FPPase, DMAPP condenses with IPP to form GPP, and
the newly formed allylic product is then the substrate
for a second condensation with IPP to give FPP. The
enzyme-catalyzed condensation is an electrophilic alky-
lation of the double bond in IPP by the allylic carbocation
generated from the allylic substrate.9 Thus, one would
anticipate that ISPP would be a more efficient substrate
for FPPase than GSPP. This assumption was verified by
the results shown in Figure 2, where FPPase was
incubated with [14C] DMAPP and a 6-fold excess of ISPP
at 37 °C for 20 h. A control experiment was run in
parallel under similar conditions except that IPP replaced
ISPP and the incubation time was 10 min. As described
above for GSPP, a portion of each sample was co-spotted
with cold GPP and analyzed by TLC. When ISPP was
the homoallylic substrate, new spots were seen at Rf )
0.23 and 0.16, corresponding to FSPP (8%) and GSPP
(92%), respectively, along with a residual spot at the
origin that is presumably an impurity in the [14C]-
DMAPP. As anticipated, in the incubation with IPP, [14C]-
DMAPP was converted to FPP within 10 min.
The identity of the products was verified by HPLC-
MS. Negative-ion HPLC of a reaction mixture from an
incubation of FPPase with ISPP and DMAPP on a
Nucleodex â-OH column gave peaks with retention times
of 8.26 (m/z 245, M - H+ for DMAPP), 9.53 (m/z 261,
M - H+ for ISPP), and 11.56 min (m/z 329, M - H+ for
GSPP). Analysis of the products from a similar incubation
with IPP and DMAPP gave a small peak at 11.39 (m/z
313, M - H+) for GPP and a large peak at 14.13 min
(m/z 381, M - H+) for FPP. These results confirm that
ISPP condenses with DMAPP in the first round of chain
1
C(1) methylene protons. The H NMR spectra were also
consistent with the thiolo structure with the bridging
sulfur attached to C(1) of the isoprenoid moiety. The C(1)
protons in allylic thiodiphosphates DMASPP, GSPP,
FSPP, and GGSPP appear as triplets at ∼3.5 ppm as a
result of similar coupling constants to the proton at C(2)
and PR and as a doublet of triplets at 3.0 ppm in ISPP.
The C(1) methylene protons appear at ∼4.5 ppm in the
normal allylic diphosphate esters and at ∼4.0 ppm in
IPP.
Stu d ies w ith F a r n esyl Dip h osp h a te Syn th a se. We
conducted a series of experiments with recombinant
avian FPPase to study the effects of replacing the
bridging oxygen atom in the normal substrates for chain
elongation. Avian FPPase was incubated with GSPP in
2-fold molar excess over [14C]IPP for 20 h at 37 °C. A
control reaction was conducted at the same time where
GSPP was replaced by GPP in a 10-min incubation. A
10 µL portion of each sample was spotted on a silica
gel TLC plate, the plate was developed with CHCl3/
methanol/water/acetic acid, and the radioactive materials
were visualized by phosphorimaging. The incubation of
GSPP and [14C]IPP gave single new spot whose mobility
was similar to FPP formed in the control reaction. All of
the [14C]IPP was consumed in the 10-minute incubation

6708 J . Org. Chem., Vol. 66, No. 20, 2001
Phan and Poulter
Exp er im en ta l Section
Gen er a l Meth od s. All experiments were performed under
an inert atmosphere unless otherwise indicated. NMR spectra
1
were collected at 23 °C. H chemical shifts were referenced to
internal CDCl3 at 7.26 or D2O at 4.8. Coupling constants are
reported in hertz. 13C NMR chemical shift was referenced to
internal CDCl3 at 77.23. 13C NMR spectra for [(n-Bu)4N]3SPPi
and the isoprenoid thiodiphosphates were referenced to ex-
ternal methanol (10%) at 49.15. 31P NMR spectra were
referenced to external phosphoric acid (5%) in D2O. A DB-5
capillary column was used for GC-MS. FAB, CI, or EI mass
spectroscopy were used to obtain high-resolution mass spectra.
Analytical HPLC separations were performed with a 4 × 200
mm Nucleodex â-OH column from Machery-Nagel Inc.23
Radioactivity was measured in Cytoscint scintillation cocktail.
Ma ter ia ls. THF was freshly distilled over Na and ben-
zophenone. Acetonitrile and CH2Cl2 were freshly distilled over
CaH2. All other solvents were HPLC grade and used without
further purification. [C14-IPP] was purchased from Amersham.
[C14-DMAPP] and [C14-FPP] were from American Radiolabeled
Chemicals, Inc. BME was from Fisher. All other chemical
reagents were purchased from Aldrich. Tris(tetra-n-butylam-
monium) salts of PPi, isopentenyl tosylate, farnesyl bromide,
IPP, DMAPP, and GPP were prepared as described by Davis-
son et al.19 TsCl was recrystallized from CHCl3/hexanes and
dried under vacuum for 24 h. Isopentenyl alcohol was distilled
under nitrogen prior to use. BHDA was prepared from the
corresponding anhydride (Fluka).24 Silica flash chromatogra-
phy was performed with Merck silica gel (200-245 mesh).
Thin-layer chromatography was on Merck silica gel 60 Å F254
TLC plates. Silica gel was visualized with phosphomolybdic
acid. Cellulose for flash chromatography was purchased from
Whatman and used according to method of Woodside.25 Cel-
lulose TLC plates were developed with a sulfosalicylic acid-
ferric chloride spray.19a
Tetr a -n -bu tyla m m on iu m Dim eth yl P h osp h a te (2). To
6 mL of neat trimethyl phosphate (51.3 mmol) was added 33.6
mL (51.3 mmol) of 40% (w/w) tetra-n-butylammonium hydrox-
ide. The mixture was heated at reflux at 100 °C for 24 h.
Methanol was removed by rotary evaporation, and the residue
was lyophilized for 24 h. Traces of water were removed by the
addition of benzene, followed by its removal at reduced
pressure, to give 18 g (95%) of a white solid: 1H NMR (D2O)
1.02 (12H, t, J H,H ) 7.6), 1.44 (8H, m), 1.74 (8H, m), 3.25 (8H,
t, J H,H ) 8.6), 3.54 (6H, d, J H,P ) 10.8); 13C NMR (D2O) 12.73
(s), 18.93 (s), 22.88 (s), 52 (d, J C,P ) 6), 57.78 (s); 31P NMR
(D2O) 2.83 (s); HRMS (FAB-) [M - 1]- calcd for C2H7PO4
125.00090, found 125.00037.
Tetr a m eth yl Th iod ip h osp h a te (3). Tetra-n-butylammo-
nium dimethyl phosphate (2) (9.8 g, 26.7 mmol) was dissolved
in 7 mL of acetonitrile. The mixture was cooled to -35 °C
before dropwise addition of 3 mL of dimethyl chlorothiophos-
phate (24.7 mmol). The mixture was allowed to stir at -35 °C
for 30 min, immediately loaded onto a silica gel column (250
mL), and eluted with CH2Cl2. Solvent was removed in vacuo
to give 1.85 g (30%) of a yellow oil: 1H NMR (CDCl3) 3.86 (d,
6 H, J H,P ) 0.73), 3.9 (d, 6H, J H,P ) 0.9); 13C NMR (CDCl3) δ
55.1 (d, J C,P ) 6), δ 55.4 (d, J C,P ) 5.5); 31P NMR (CDCl3) δ
-13.8 (d, J P,P ) 20, P(O)), δ 55.6 (d, J P,P ) 20, P(S)); HRMS
(EI) [M]+ calcd for C10H12SP2O6 249.9830, found 249.9841.
Tr is(t et r a -n -b u t yla m m on iu m )t h iod ip h osp h a t e ([(n -
Bu )4N]3SP P i). Iodotrimethylsilane (3.41 g, 33.84 mmol) was
added dropwise to 1.5 g of neat tetramethyl thiodiphosphate
(6 mmol) at -35 °C, and the mixture was slowly allowed to
warm to room temperature over 6 h. The solvent was removed
in vacuo to give tris(trimethylsilane) thiodiphosphate (4) as a
brown semisolid: 31P NMR (CD3CN in 1 M EDTA) -32.64 (d,
F igu r e 3. Double-reciprocal plot of v-1 versus [GPP]-1 at
different fixed concentrations of GSPP: [GSPP] ) 0 (O),
[GSPP] ) 5 µM (b), [GSPP] ) 10 µM (0), and [GSPP] ) 20
µM (9).
elongation, but the product of the reaction, GSPP, is a
poor substrate for the second round.
Kin etic Stu d ies w ith GSP P . In a preliminary com-
munication of this work, we reported that GSPP reacts
much more slowly with IPP than does GPP.18 This
experiment was performed by the incubation of [14C]IPP
and GSPP or GPP with recombinant FPPase, and the
initial rate of the reactions were determined by the
standard acid lability assay.21 Under conditions where
the condensation of [14C]IPP and GPP gave a predictable
amount of product, no detectable formation of product
was seen for [14C]IPP and GSPP. When the concentration
of FPPase was increased 100-fold and the time of the
incubation was lengthened 10-fold, only slight amount
of product was observed. When compared to the normal
assay conditions for this enzyme, the reaction of IPP with
GSPP is estimated to be 105 times slower than for GPP.
Additional kinetic measurements revealed that GSPP
was an efficient competitive inhibitor of FPPase with
KIGSPP ) 24.8 µM, as illustrated by the double-reciprocal
plot of initial velocity versus [GPP]-1 at different fixed
concentrations of GSPP (see Figure 3), in comparison,
GPP
KMapp
) 4.5 µM under similar conditions.22 The steady-
state kinetic measurements show that GSPP binds to the
allylic site in FPPase with excellent selectivity and
affinity.
Con clu sion
In summary, the thiolo isomers of isoprenoid thio-
diphosphate esters were synthesized by treating allylic
bromides and homoallylic tosylates with tris(tetra-n-
butylammonium)thiopyrophosphate. The thiolo analogue
of geranyl diphosphate was substrate for condensation
with IPP catalyzed by FPPase; however, GSPP was ∼105-
fold less reactive than GPP itself. ISPP was also a
substrate for the chain elongation reaction. Condensation
with DMAPP produced GSPP, which accumulated be-
cause of its poor reactivity. The direct displacement
protocol using the tris(tetra-n-butylammonium) salt of
SPPi should be generally applicable to a variety of
thiodiphosphate analogues where the center undergoing
substitution is not sterically hindered.
(23) Lange, B. M.; Ketchum, R. E. B.; Croteau, R. Plant Physiol., in
press.
(24) (a) Child, J . J .; Knapp, C.; Eveleigh, D. E. Mycologia 1973, 65,
1087-1086. (b) Mallette, M. F. J . Bacteriology 1967, 94, 283-290.
(25) Woodside, A. B.; Huang, Z.; Poulter, C. D. Org. Synth. 1987,
66, 211-219.
(21) Rilling, H. C. Methods Enzymol. 1985, 110, 145-152.
GPP
(22) IPP is a substrate inhibitor, and the apparent value for KM
GPP
depends on the concentration of IPP. In this case, KMapp
was
measured at [IPP] ) 10 mM. See: Laskovics, F. M.; Krafcik, J . M.;
Poulter, C. D. J . Biol. Chem. 1979, 254, 9458-9463.

Synthesis of (S)-Isoprenoid Thiodiphosphates
J . Org. Chem., Vol. 66, No. 20, 2001 6709
J P,P ) 16.5, P(O)), 29.94 (d, J P,P ) 16.5, P(S)). TMS derivative
4 was dissolved in 20 mL of acetonitrile. The solution was
titrated with 40% (w/w) tetra-n-butylammonium hydroxide to
pH 7.2 and acetonitrile was then removed at reduce pressure.
Benzene (10 mL) was added to give a heterogeneous mixture
composed of three layers. The bottom layer was separated,
concentrated in vacuo, and lyophilized to give 5.36 g (96%) of
a yellow gelatinous residue, which was used directly in next
step: 31P NMR (CDCl3), -9.2 (d, J P,P ) 30, P(O)), 36.8 (d,
NMR (D2O) 8.0 (d, P(S), J P,P ) 29 Hz), -10.5 (d, P(O), J P,P )
29 Hz); HRMS (FAB-) [M - 1]- calcd for C5H11SP2O6 329.03776,
found 329.03612.
(S)-(E,E)-3,7,11-Tr im eth yl-2,6,10-dodeca tr ien -1-yl Th io-
d ip h osp h a te (F SP P ). Farnesyl bromide (0.347 g, 1.22 mmol)
was added dropwise to a cold solution of tris(tetra-n-butyl-
amonium) thiodiphosphate (2.5 g, 2.68 mmol) in 15 mL of
acetonitrile. The reaction mixture was allowed to stir for 30
min. After the workup, the resulting tetrabutylamonium salt
was converted to ammonium form with 55 equiv of resin and
lyophilized. The resulting powder was dissolved in a limited
amount of elution buffer and chromatographed on cellulose
with a flow rate 2 mL/min to yield 0.46 g (84%) white-yellowish
solid: Rf 0.4; 1H NMR (D2O) 5.4 (t, 1H,J H,H ) 7.69 Hz), 5.2 (b,
1
J P,P ) 30, P(S)); H NMR (CDCl3), 1.02 (12H, t, J H,H ) 7.1),
1.44 (8H, m), 1.74 (8H, m), 3.25 (8H, t, J H,H ) 8.2); HRMS
(FAB-) [M - 1]- calcd for H4SP2O6 192.91256, found 192.91181.
Gen er a l P r oced u r e for P r ep a r a tion of Th iod ip h os-
p h a tes. Two equivalents of tris(tetra-n-butylamonium)thio-
diphosphate in acetonitrile was placed in a dried flask, and
the solution was cooled to -35 °C. To the cold, well-stirred
solution was slowly added 1 equiv of allylic bromide or
homoallylic tosylate. The mixture was allowed to stir for 30-
45 min at -35 °C. The solvent was then removed at reduced
pressure to give an opaque semisolid residue. This residue was
dissolved in a minimum amount of water and passed through
a column of 40-60 equiv of DOWEX AG 50W-X8 (100-200
2H), 3.5 (t,2H, J H,P ) 8.3 Hz), 2.1 (m, 8H), δ1.7 (d, 6H, J H,H
)
6.95 Hz), 1.6 (s, 6H); 13C NMR (D2O) 16.0 (s), 17.7 (s), 25.7 (s),
26.6 (s), 26.8 (s), 28.3 (s), 39.8 (d), 120.2 (d), 124.5 (s), 124.9
(s), 132.0 (s), 135.9 (s), 140.9 (s); 31P NMR (D2O) 7.6 (d, J P,P
)
29.31 Hz); -7.4 (d, J P,P ) 29.31 Hz); HRMS (FAB-) [M - 1]-
calcd for C15H27SP2O6 397.10036, found 397.10170.
(S)-(E,E,E)-3,7,11,15-Tetr a m eth ylh exa d eca 2,6,10,14,-
tetr a en -1-yl Th iod ip h osp h a te (GGSP P ). Geranylgeranyl
bromide (0.34 g, 0.91 mmol) was added dropwise to a cold
solution of tris(tetra-n-butylamonium)thiodiphosphate (1.56 g,
1.71 mmol) in 5 mL of acetonitrile. The mixture was allowed
to stir for 30 min. After the workup, the resulting tetrabutyl-
amonium salt was converted to ammonium form with 60 equiv
of resin and lyophilized. The resulting powder was dissolved
in a limited amount of elution buffer and chromatographed
on cellulose with a flow rate 2 mL/min to yield 0.32 g (65%) of
white-yellowish solid: Rf 0.47; 1H NMR (D2O) 5.4 (t, 1H,
J H,H ) 8.3 Hz), 5.2 (b, 3H), 3.5 (t,2H, J H,P ) 8.79 Hz), 2.1 (m,
12H), 1.7 (d, 6H, J H,H ) 12.7 Hz), 1.6 (d, 9H, J H,H ) 5.3 Hz);
13C NMR (D2O) 15.8 (s), 17.5 (s), 23.5 (s), 25.5 (s), 26.7 (s),
27.9 (s), 31.6 (s), 39.7 (s), 124.4 (d), 124.6 (s), 125.3 (s), 131.0
(s), 133. 8 (s), 134.8 (s), 135.7 (s), 139.7 (s); 31P NMR (D2O) 7.6
(d, PS, J P,P ) 27.5 Hz), -7.1 (d, J P,P ) 28.8 Hz); high resolution
(FAB-) mass spectrum [M - 1]- calcd for C20H35SP2O6
465.16296, found 465.16190.
Kin etic Stu d ies w ith F P P a se. The acid lability assay was
used to measure initial velocities of coupling reaction of [14C]-
IPP with GPP or GSPP.21 Each standard assay contained 70
ng of purified enzyme, 20 µM [14C]IPP (10 µCi/µmol), 50 µM
GPP or GSPP, with a buffer solution containing 40 mM BHDA
at pH 7.3, 20 mM BME, 2 mM MgCl2, 2 mg/mL BSA in a total
volume of 200 µL. The assay mixture was incubated at 37 °C
for 10 min. The assays were quenched with 200 µL of
methanol/HCl (4:1, v/v), incubated for an addition of 10 min,
and extracted with 1 mL of ligroine. The radioactivity in a 0.5
+
mesh) cation-exchange resin (NH4 form). The column was
eluted with two volumes of ion-exchange buffer (25 mM NH4-
HCO3 and 2% (v/v) of 2-propanol in water), and the eluent was
lyophilized to give a light yellow powder. The resulting powder
was dissolved in a limited amount of water and purified by
chromatography on cellulose with elution by 50 mM am-
monium bicarbonate in 1:2:1 (v/v/v) acetonitrile/2-propanol/
water. The desired fractions were pooled, concentrated by
rotatory evaporation, and lyophilized to give a yellowish
powder that was then stored at -80 °C.
(S)-3-Meth yl-3-bu ten e-1yl Th iodiph osph ate (ISP P ). Iso-
pentenyl tosylate (0.11 g, 0.45 mmol) was treated with tris-
(tetra-n-butylamonium) thiodiphosphate (0.97 g, 1 mmol) in
acetonitrile (5 mL) for 45 min. The resulting tetrabutylamo-
nium salt was converted to ammonium form with 50 equiv of
resin. After the lyophilization, the resulting powder was
dissolved in 1 mL of water. The mixture was passing through
the flash cellulose column to yield 0.126 g (57%) of white-
1
yellowish solid: Rf 0.33; H NMR (D2O) 4.85 (d, 2H, J H,H ) 5
Hz), 3 (m,2H), 2.46 (dt, 2H, J H,H ) 7 Hz), 1.77 (s, 3H); 13C NMR
(D2O) 21.6 (s), 28.4 (d), 38.3 (d), 111.3 (s), 145.8 (s); 31P NMR
(D2O) 7 (d, PS, J P,P ) 29 Hz), -7.9 (d, J P,P ) 29 Hz); HRMS
(FAB-) [M - 1]- calcd for C5H11SP2O6 260.97516, found
260.97402.
(S)-3-Meth yl-2-bu ten -1-yl Th iodiph osph ate (DMASP P ).
Dimethylallyl bromide (0.3 g, 2 mmol) was added dropwise to
a cold solution of tris(tetra-n-butylamonium) thiodiphosphate
(3.64 g, 3.91 mmol) in 15 mL of acetonitrile,. The mixture was
allowed to stir for 30 min. After the workup, the resulting
tetrabutylamonium salt was converted to the ammonium form
with 40 equiv of resin and lyophilized. The resulting powder
was dissolved in a limited amount of elution buffer and
chromatographed on cellulose with a flow rate 2 mL/min to
yield 0.5 g (89%) of white-yellowish solid: Rf 0.3; 1H NMR
(D2O) 5.3 (t, 1H, J H,H ) 7.8 Hz), 3.5 (t, 2H, J H,H ) 8.8 Hz), 1.8
(d, 6H, J H,H ) 8 Hz); 13C NMR (D2O) 17.3 (s), 25.1 (s), 28.3 (s),
120.3 (d), 137.8 (s); 31P NMR (D2O) 8.6 (d, PS, J P,P) 28 Hz),
-9.0 (d, J P,P ) 28 Hz); HRMS (FAB-) [M - 1]- C5H11SP2O6
260.9755, found 260.9740.
(S)-(E)-3,7-Dim eth yl-2,6-octadien -1-yl Th iodiph osph ate
(GSP P ). Geranyl bromide (0.438 g, 2 mmol) was added
dropwise to a cold solution of tris(tetra-n-butylamonium)
thiodiphosphate (4.30 g, 4.62 mmol) in 20 mL of acetonitrile.
The mixture was allowed to stir for 30 min. After the workup,
the resulting tetrabutylamonium salt was converted to am-
monium form with 40 equiv of resin and lyophilized. The
residual powder was dissolved in a limited amount of elution
buffer and chromatographed on cellulose with a flow rate 2
mL/min to yield 0.61 g (80%) of white-yellowish solid: Rf 0.38;
1H NMR (D2O) 5.4 (t, 1H, J H,H ) 9 Hz), 5.15 (b, 1H), 3.5 (t,
2H, J H,P ) 8.79 Hz), 2.1 (m, 4H), 1.7 (d, 6H, J H,H ) 5.13 Hz),
1.6 (s, 3H); 13C NMR (D2O) 15.7 (s), 17.5 (s), 25.3 (s), 26.1(s),
28.3 (s), 39.2 (s), 120.3 (d), 124.5 (s), 133.6 (s), 141.0 (s); 31P
mL sample of the organic layer was measured by liquid
GSPP
scintillation spectrometry. The IC50
was measured with 4.5
µM GPP, 10 µM [14C]IPP (10 µCi/µmol), and varied concentra-
tions of the thiodiphosphate (0.5, 1, 5, 10, 50, 100 µM). KIGSPP
was measured with 10 µM [14C]IPP (10 µCi/µmol) and varied
concentrations of GPP (3, 4, 5, 6 µM) and GSPP (0, 5, 10, 20
µM).
P r od u ct Stu d ies. Incubations with GSPP were in a total
volume of 50 µL, which included 25 µL of buffer solution
(containing 40 mM BHDA at pH 7.3, 20 mM BME, 2 mM
MgCl2, 2 mg/mL BSA), 3 µL of 1mM GSPP (50 µM) or 3 µL of
1mM GPP (50 µM), 1 µL of 1mM [14C]IPP (20 µM, 55 µCi/µmol),
14.5 µL of water and 7 µL of FPPase (6.87 mg/mL, 20 µM).
Similar conditions were used for ISPP and included 25 µL of
buffer, 6 µL of 1mM ISPP (120 µM) or 6 µL of 1mM IPP (120
µM), 1 µL of 1mM [14C]DMAPP (20 µM, 55 µCi/µmol), 11.5 µL
of water and 7 µL of FPPase (6.87 mg/mL, 20 µM). Samples
containing GSPP and ISPP were incubated at 37 °C for 20 h,
while those containing GPP and IPP were incubated at 37 °C
for 10 min. A 10 µL portion of each sample was spotted on a
silica gel 60 F254 TLC plate and developed with 125:75:20:10
(v/v/v/v) CHCl3/methanol/water/acetic acid. The plates were
wrapped with plastic wrap and imaged for 20 h.
HP LC-MS for P r od u cts fr om ISP P . A mixture of 32 µL
of 500 mM BHDA buffer, 5.5 µL of 72 mM MgCl2, 40 µL of 10
mg/mL BSA, 2.8 µL of 1.4 M BME, 56 µL water, 6.5 µL of

6710 J . Org. Chem., Vol. 66, No. 20, 2001
Phan and Poulter
174.89 mM ISPP (2.98 mmol), 17 µL of 153 mM DMAPP (0.99
mmol), and 40 µL of FPPase (24.9 mg/mL, 30 nmol) was placed
in a 1.5 mL Eppendorf tube. A control reaction with IPP was
run in parallel. The two mixtures were incubated at 37 °C for
20 h. The solutions were placed in the 10 µm centrifuge filters,
and centrifuged at 10 000 rpm for 20 min. The filters were
rinsed with 100 µL of 25 mM ammonium bicarbonate solution,
and the combined filtrates were lyophilized. The residues were
dissolved with 200 µL of 1:1 (v/v) water/acetonitrile, and a 5
µL portion was analyzed by HPLC-MS.
Ack n ow led gm en t. This work was supported by
NIH Grant GM 21328. R.M.P. is an NIH Predoctoral
Trainee, GM 08573.
Su p p or tin g In for m a tion Ava ila ble: 1H and 31P NMR
spectra for obtained compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.
J O010505N