Proc. Indian Acad. Sci. (Chem. Sci.), Vol. 106, No. 5, October 1994, pp. 1231-1251.
9 Printed in India.
Synthesis and applications of phosphatidylinositols and their analogues I
M S S H A S H I D H A R
Division of Organic Chemistry (Synthesis), National Chemical Laboratory, Pune 411008, India
Aktraet. There is an upsurge of interest in the chemistry and biochemistry of phosphoino- tides, their analogues and the related enzymes due to their involvement in the inositol phosphates mediated cellular signal transduction pathways. The present review deals with the recent developments in the synthesis and applications of phosphatidylinositol and its derivatives.
Keywords. Phospholipid; phosphatidylinositol; inositol phosphates; phospholipase C;
signal transduction; second messenger.
Introduction
Lipids constitute a m a j o r class of biologically i m p o r t a n t molecules. There is an increasing awareness o f the active roles played b y m e m b r a n e lipids such as p h o s - phatidylcholine, p h o s p h a t i d y l e t h a n o l a m i n e , phosphatidylglycerol, a n d p h o s p h a t i - dylinositol (Ptdlns) in the structure and function of cells. P t d l n s has generated a lot of interest due to its i n v o l v e m e n t in the myo-inositol m e d i a t e d cellular signal t r a n s d u c t i o n p a t h w a y . This article intends to c o v e r s o m e o f the recent d e v e l o p m e n t s in the synthesis a n d applications o f P t d l n s a n d their analogues.
Results and discussions
Phosphatidylinositols
P t d I n s exist as a m i n o r c o m p o n e n t in b i o - m e m b r a n e s ( 5 - 1 0 % of total phospholipids).
T h e first d e m o n s t r a t i o n o f lipid b o u n d myo-inositol in m a m m a l s was m a d e by Folch
1 NCL commun, no. 6014
Abbreviations: PtdIns, phosphatidylinositol; AMP, adinosinemonophosphate; PtdIns (3) P, phosphatidy- linositol-3-phosphate; PtdIns (4)P, phosphatidylinositol-4-phosphate; PtdIns(3,4)P2, phosphatidylinositol- 3,4-bisphosphate; PtdIns(4,5)P2, phosphatidylinositol-4,5-bisphosphate; PtdIns(3,4,5)Pa, phosphatidy- linositol-3,4,5-triphosphate; Ins(1)P, Myo-inositol-l-phosphate; Ins(1,2-cyc)P, Myo-inositol-l,2-cyclic- phosphate; Ins (1,4, 5) P3, Myo-inositol- 1,4,5-triphosphate; DAG, diacylglycerol; TBDMS, t-butyldimethyl- silyl; MSNT, l-(mesitylene-2-sulphonyl)-3-nitro-l,2,4-triazole; NPCL, 5,5-dimethyl-2-oxo-2-chloro-l,3,2- dioxaphophorinan; PLA2, phospholipase A2; PLD, phospholipase D; PI-PLC, phosphatidylinositol- specific phospholipase C; PI-4-kinase, phosphatidylinositol-4-kinase; P, PO 3 H2; Bn, benzyl; Pro, propenyl;
DAG, diacylglycerol; A, allyl; MOM, methoxymethyl; Ac, acetyl; THP, tetrahydropyranyl; PMB, p-methoxybenzyl; nD, D-enantiomer of the compound n; nL, L-enantiomer of the compound n; nHL, racemate n)
1231
1232 M S Shashidhar
and Wooley who discovered myo-inositol in brain lipids (Folch and Wooley 1942).
Several phosphorylated (on the myo-inositol ring) derivatives of Ptdlns are known, most important of them being Ptdlns (4, 5)P2 (see below). Ptdlns (3)P was identified in transformed fibroblasts (Whitman et al 1988). Other phosphoinositides containing the 3-phosphate, such as, Ptdlns(3,4)P2 and Ptdlns(3,4,5)Pa have been found in many cell types (Majerus et al 1990). Glycosyl phosphatidylinositols which were recently discovered, are a class of glycolipids that anchor proteins, polysaccharides or small oligosaccharides to cell. membranes through covalent linkages. They have been found in a wide variety of cells and tissues. Myo-inositol as well as its di-mannosides are also present (Anderson and Roberto 1930; Anderson et al 1938) in the mycobacterium phospholipids. This lipid consists of a tri-substituted myo~ in which the 2- and 6-hydroxy groups are linked to carbohydrates and the 1-hydroxy group is linked to phosphatidic acid (Lee and Ballou 1965).
Phosphatidylinositols: Importance
Cellular signal transduction mechanisms translate external signals into internal signals through second messengers. Two major signal transduction pathways are now known, one employs cyclic AMP and the other employs Ins(l,4, 5)Pa and DAG.
In the past few years it has been shown that an extracellular signal (such as a drug or a hormone) can activate PI-PLC which hydrolyzes the membrane lipid PtdIns (4, 5) P2 to Ins (1,4, 5)P3 and DAG, both of which serve as second messengers. Ins (1,4, 5)Pa is released into the cytosol, whereas DAG remains in the cell membrane, Consequent to this a series of complex reactions occur, resulting in the mobilization of calcium ions from endoplasmic reticuhim to the cytosol. DAG stimulates protein kinase C which catalyzes the transfer of phosphate groups from ATP to other proteins, which in turn alters the protein function. 1,2-Diacylglycerol also serves as a messenger by providing the substrate for icosanoid production. Both Ins (1,4,5)P3 and DAG are generated rapidly in low concentrations and are quickly removed, properties characteristic of molecules that signal cells to carry out designated functions. The supply of PtdIns (4,5) P2 in the cell membrane is maintained by sequential phosphory- lation of the more abundant PtdIns by specific 4- and 5-kinases. Evidence has also begun to accumulate to show that phosphoinositides might be involved in vesicular traffic in yeast and some mammalian cells (Skinner 1993; Hay and Martin 1993;
Cleves et al 1991) However, the exact role played by phosphoinositides in vesicular traffic has not been clearly demonstrated.
P I - P L C enzymes which cleave phosphatidylinositols exist in a wide variety of tissues and organisms. Extracellular P I - P L C s have been isolated from the culture media of several microorganisms. Intracellular/membrane bound PI-PLCs are prevalent in mammalian cells. The extracellular P I - P L C s are water-soluble and are relatively specific for Ptdlns and glycosyl Ptdlns. PI-PLCs specific for Ptdlns(4,5)P2 are calcium ion dependent for their activity, whereas bacterial enzymes have no metal ion dependence and do not cleave phosphorylated forms of Ptdlns. Calcium ion indepen- dent P I - P L C s specific for glycosyl Ptdlns have been purified from Trypanasoma brucei and rat liver. The enzyme from rat liver has been implicated in insulin action.
The two types of enzymes produce different products when acting on analogous Ptdlns substrates. The mammalian enzyme usually produces a mixture of Ins(1,2- cyc)P and Ins(l)P, depending on the pH, enzyme sub-type and other conditions
Phosphatidylinositols and analogues 1233 (Dawson et al 1971; Kim et al 1989) whereas the bacterial enzyme produces Ins (1,2- cyc) P exclusively (Ferguson et al 1985; Volwerk et al 1990).
PLA, --~ __OAR 1 0
o
PLD " - ~ R2~'-..(~ ~ .~ H
~ 1 7 6 ~
HO~y~#O -- P-- ~..
.... o .
'~HO"'~ ~ H "'"OH 1
-4-KINAS--E --/ PI - PLC
PI
Scheme 1.
R,
= c~( CH.z)~6 CO
--R 2= CH3(CH2) 4 (CH=CHCH2) 4 CH2CO--
The developments outlined above have revived interest in the chemistry and biochemistry of phosphoinositides and related enzymes. Scheme 1 shows some of the important enzymes for which Ptdlns serves as a substrate. Studies toward the delineation of the roles of various enzymes involved in the myo-inositol mediated signal transduction processes has led to the design and synthesis of a number of myo-inositol phospholipids and their analogues. Such analogues are also of medical interest because of their potential as pharmacological agents.
Phosphatidylinositols: Synthesis
Myo-inositol, which forms the head group of Ptdlns is a meso isomer of hexahydroxy cyclohexane, In myo-inositol five of the hydroxy groups are equatorial and only one is axial. Thus, for all practical purposes the ring is regarded as rigid (flipping of the ring would result in one equatorial and five axial hydroxy groups). However, heavy substitution on the myo-inositol ring might distort the chair conformation of the cyclohexane ring. Different representations of the myo-inositol ring and the most widely used numbering of the ring carbons are shown in scheme 2 using myo- inositol-l-phosphate (Ins(1)P) as an example (for an account on the implications of the stereochemistry of myo-inositol phosphate see Parthasarthy and Eisenberg 1986). All naturally occurring analogues contain myo-inositol head group substituted at the D-1 position an.d the glycerol moiety acylated at sn-1 and sn-2 positions. No phosphoinositides of the L-series are known to occur in nature.
The key step in the chemical synthesis of phospholipids is phosphorylation, leading to the formation of the phosphodiester bond. Phospholipids have been synthesized using various phosphodiesters, phosphotriesters, phosphites and H-phosphonates
1234 M S Shashidhar
HO Q 0 OH
0 ~ , 2 ~ "a/~OH
HO OH
HO---~\ 1 4 \~OH 4 ~ r l / t
H O ' @ ' y '''OH ~ k-"~OHoH 5 H
OH OH
Myo- Inositol
Scheme 2.
D-lns(I)P D-ins (3) P
o r o r
L- Cns(3)P L-Cns (I) P
(Lindh and Stawinski 1989). Synthesis of phosphoinositides in addition requires a suitably protected myo-inositol derivative. Several methodologies for the synthesis of enantiomerically pure myo-inositol derivatives have been developed in the recent past (Cosgrove 1980; Billington 1989; Potter 1990; Bruzik and Tsai 1992). The most extensively used hydroxy protecting group during the synthesis of phosphoinositols is the benzyl group, which can be cleaved by hydrogenolysis or using ethanethiol.
However, when the use of such conditions are not compatible with.the structure of the desired product, ketals, acetals, esters or orthoesters have been used. Many of the O-alkylated derivatives of myo-inositol are prepared taking advantage of the ease of alkylation of the equatorial 1 (or 3-)-hydroxy group over the axial 2-hydroxy group.
Phosphatidylinositols as substrates
Dipalmitoyl PtdIns(4,5)P 2 was synthesized (Dreef et al 1988) starting from the optically active 4,5-di-O-allyl-3,6-di-O- benzyl-D-myo-inositol 1 (scheme 3).
OH OBn
B n O ~ O H ~ ~ B n O . O H
AO '@ " ~ " ' O Bn Pro 0": "'O Bn
I I
AO OPro
I 2
3
R =
CHs(CH2)t4 CO--
Scheme 3.
Ptd lns (4, 5 ) P2
Phosphatid ylinositols and analogues 1235 The key intermediate 2 was prepared from 1 by temporary protection of the 1-hydroxy group as t-butyldimethylsilyl (TBDMS) ether. The phosphoramidite 3 was then coupled to 2 and the phosphite obtained was oxidized with t-butylhydroperoxide.
The propenyl ethers were then removed by acid hydrolysis and the free hydroxy groups generated at the 4- and 5-positions were phosphorylated by P (III) approach.
All the benzyl protecting groups were cleaved by hydrogenolysis to obtain dipalmitoyl Ptdlns (4, 5) P2. This method however cannot be employed for the synthesis of Ptdlns containing unsaturated fatty acids.
OBn B n O ~ O H
B n(~""~""O Bn OBn
F
OROR' X R30 '~ H R'O~ ~ .O--IpI--o -'I
R' O ' ~ O R ' OR;' OR ~
4 D
Scheme 4.
R 3 = CH 3(cH2)I4CO,X = O,S
~ 6 R' B n 5 , = , R~' = CH3
_ ~ R I R z : H
Dipalmitoyl Ptdlns as well as the corresponding phosphorothioate analogue were synthesized (Salamonczyk and Bruzik 1990) from D-2,3,4,5,6-penta-O-benzyl-myo- inositol (4D), 1,2-dipalmitoyl-sn-glycerol and chloro-N,N-diisopropylaminome- thoxyphosphine (scheme 4). Protected diastereomeric phosphorothioates 5 (X=S) could be separated by column chromatography. The benzyl protecting groups of the individual diastereomers were then removed with borontrifluoride etherate and ethanethiol. The diastereomeric lipids 6 were characterized by N M R spectroscopy and the cleavage by phospholipase A 2 and C.
Preparation of all the four stereoisomers of dihexadecanoyl Ptdlns (scheme 5) was reported by Young and co-workers (Young et al 1990). The enantiomeric penta-protected myo-inositols 7D and 7L were coupled with phenyl esters of di-O-hexadecanoyl glyceryl phosphates (sn-3 as well as sn-1) in the presence of MSNT.
The phenyl phosphate moiety was first cleaved by hydrogenation and the acid so generated was used for the hydrolysis of the remaining acid sensitive protecting groups. All the four synthetic Ptdlns 8D, 8L, 9D and 9L were tested as substrates for Ptdlns 4-kinase from human erythrocytes.
Phosphoinositides 8D and 9D were phosphorylated by the kinase at a rate comparable to that of its natural substrate but 8L and 9L were not phosphorylated by the same enzyme. This indicates that the kinase from human erythrocytes recognizes only the head group of Ptdlns during phosphorylation and that the structure and configuration of the diacylglycerol are inconsequential. Perhaps, the hydrocarbon portion of the phospholipid only serves as an anchor in the bio-membrane. Similar preferences have been reported for the cleavage of Ptdlns by P I - P L C (see below).
Thus, the properties of Ptdlns 4-kinase seems to be similar to those of P I - P L C as regards the recognition of its substrate.
1236 M S Shashidhar --/---0
o_/_
ZD
l,
OH
o c , 3
_TL -~R
H I R 0 OH
J_.__ II O. ~ .OH
H O " ' " ~ " J ' O H OH
_8D (Sn-3) 9 L ( S n - I )
Z D .--~---~ g_ D ( Sn-1 )
Scheme 5.
7 L - - ~ " - - " _8L(Sn-5)
A chromogenic water soluble phosphodiester derivative of myo-inositol was synthesized and tested as a substrate for bacterial P I - P L C (Shashidhar et al 1991a).
Racemic myo-inositol-l-(4-nitrophenyl phosphate) I~..QDL) was prepared from myo- inositol (scheme 6). Since 4-nitrophenylphosphates are quite sensitive to strong acids and bases, hydroxy protecting groups which require such conditions for the deprotection could not be used. Also, protecting groups that require removal by hydrogenolysis could not be used because of simultaneous reduction of the aromatic nitro group, Thus iso-propylidene and 4-methoxytetrahydropyran-4-yl groups, which can be cleaved under mild acid conditions were used for the protection of the myo-inositol hydroxy groups. The penta-protected myo-inositol "]DE was phosphory- lated with 4-nitrophenylphosphorodichloridate in pyridine and all the protecting groups were hydrolyzed by mild acid to generate the nitrophenyl phosphate 10DE
The myo-inositol derivative 1ODL was cleaved by bacterial P I - P L C to Ins(1,2- cyc)P and 4-nitrophenol (as shown by 31P-NMR). The continued incubation of the reaction mixture resulted in the formation of Ins (1)P. Thus, the mode of action of P I - P L C (Volwerk et al 1990) on synthetic substrates is identical to that on its natural substrate. The nitrophenyl derivative 1._00DL could also be used to determine the effect of pH on the activity of PI-PLC. Phospholipase C from B. cereus showed high activity at pH 6-7 and then its activity decreased with increase in pH. The activity dropped essentially to zero at pH 9. Similar results were obtained with a fluorescent substrate, racemic myo-inositol 1-naphthylphosphate, which was prepared
Phosphatidylinositols and analogues 1237
_TD Z L
OH O 0 OH
II II
~OH
.... L . L oH
y" OH
OH OH
PNP = p - Nitrophenyl
lo D lo
LI PI - PLC
I
PI-PLC0 II
H O ~ ; OH
H O , " ' ~ %O H OH 11D
Sekeme 6.
NO REACTION
using 7DL and naphthylphosphorodichloridate. The same fluorescent substrate was also synthesized from racemic penta-O-benzyl myo-inositol (Shashidhar et al 1991b).
Availability of such substrates for the continuous assay of PI- PLC might provide an alternative to the use of expensive radiolabelled substrates that can only be used in discontinuous assays.
Leigh and co-workers (Leigh et a11992) examined the enantioselectivity of bacterial P I - P L C with respect to the chirality of the myo-inositol moiety using enentiomeric myo-inositol-l-(4-nitrophenyl phosphate). They synthesized (scheme 6) 1OD and 1OL from 7D and __TL respectively and studied their cleavage by P I - P L C under various conditions. Use of these enantiomeric myo-inositol derivatives as substrates showed P I - P L C is selective in the binding as well as cleavage of its substrates. The D-isomer was found to be a reasonably good substrate (V=. x of 650 + / - 5 0 / ~ m o l m i n -1 (mg of protein) -1 and K u - - 5 + / - 1 mM). The specific activity for the D-isomer was about 35% that of PtdIns (solubilized in sodium deoxycholate) for which the apparent Vm. x is about 1800/~molmin -1 (rag of protein) -1 (Koke et al 1991). The observed activity for the nitrophenyl derivative is 1 to 2 orders of magnitude greater than that of the thiophosphate 3ODL (Vm~ = 16.9 + / - 0.6/~mol min- 1 (mg of protein)- 1) and 4 orders of magnitude greater than that of the naphthyl derivative (specific enzyme activity of about 0.044 #mol min- t (mg of protein)- t at a substrate concentration of
1238 M S Shashidhar
about 0-8 mM, Shashidhar et al 1991b). The L-isomer I(1.0L) on the other hand was neither a substrate for P I - P L C nor an inhibitor, even when used at a five-fold molar excess over the D-isomer l([9_D ). This indicates that the L-isomer does not bind to the active site and hence shows that the binding pocket of this enzyme is highly sensitive to the stereochemistry of the myo-inositol ring.
O
MoMo"
1 2 D L 1.~3D
s.
oL --'o OH
o,:..sH O ~ 0 H 04~..7~0 "~P~" 0 ~
.I J.. ..+.o.,
14 (Sp)
+Invermbn H 0 " " " ~ " " 0 H
kOR I HO"~""O~ H at p
=
OH OH
14 ( R p + Sp) 1 5 - exo (Ps)
R I : CH 3 ( CH2)14C0 Scheme 7.
A study of the stereochemical course of the cleavage of Ptdlns by P I - P L C from B. cereus as well as guinea pig uterus, using the phosphorothioate analogue of dipalmitoyl Ptdlns as a substrate was reported by Lin et al (1990). The synthesis (scheme 7) started from the racemic dicyclohexylidene derivative 12DL of myo-inositol which was resolved (as the corresponding camphanate esters), and converted to the methoxymethyl ether 13D. The alcohol 13D was coupled to 1,2-dipalmitoyl-sn- glycerol by phosphorus (III) approach, followed by deprotection. The mixture of diastereomers 14 (Rp and Se) obtained was used as a substrate for PI-PLC. Using 31P-NMR spectroscopy it was shown that P I - P L C cleaves only the R e isomer with inversion of configuration at phosphorus, and not the Se isomer. In experiments with the bacterial enzyme, formation of only the exo-cyclic phosphate 15 was observed, whereas, when the mammalian enzyme was used, formation of both the exo-cyclic phosphate 15 as well as Ins(1)P could be observed. Hence it was concluded that P I - P L C cleaves Ptdlns to produce cyclic and non-cyclic phosphates of inositol by two parallel reactions without the involvement of a covalent enzyme substrate complex (however, see below).
Bruzik and co-workers (Bruzik et al 1992) described the synthesis of diastereomers of oxygen isotope labelled P-chiral dipalmitoyl Ptdlns (scheme 8) and studied the
Phosphatidylinositols and analogues 1239
16
~ 0 F-- ORJ
RI 0 u~L.~m H
MO MO'" y .... OMOM OMOM R t -- CH 3 (CH2) 4 CO
J \
/ \
/ \
/ \
OH D, AG OH DAG
HO 0 . . . X H O ~ P ... 9
HO""~"~
H O " " ""OHI
OH OH
1"/ (Sp) , X=S e= "to I__88 (Rp) , X : S
19 (Rp), X=O 20 (Sp) ,X=O
Sr 8.
stereochemical course of cleavage of Ptdlns by PI-PLC from B. cereus as well as bovine brain, using these synthetic substrates. The phosphite 16 was obtained from a diastereomerically pure D-camphor derivative of myo-inositol (Bruzik and Tsai 1992). Treatment of 1__66 with sulphur followed by chromatography and deprotection yielded isomers 17 (Sp) and 18 (Rp). Similarly, oxidation of 16 with iodine in presence of H2170 yielded--diastereomeric [160, 170] dipalmitoyl Ptdlns 19 (Rp) and 20 (Sp).
The configurations of the thiophospholipids were assigned based on the analysis of the alp NMR spectra and the stereospecificity of their cleavage by phospholipase A2 while those of 19 and 20 were assigned by chemical correlation with isomeric phosphorothioate derivative of dipalmitoyl Ptdlns via desulphurization.
When the synthetic diastereomeric lipids were used as substrates for PI-PLC and the products analyzed by 31p NMR spectroscopy, it was revealed that bovine brain PI-PLC converts Ptdlns to lns(1,2-cyc)P with inversion of configuration at phosphorus. In order to determine the steric course of formation of Ins(1)P by PI-PLC, diastereomeric 19 and 20 were hydrolyzed enzymatically to [160, 170, 180]
Ins (1)P with H2180. Ins (1)P so obtained was cyclized chemically and the resulting cyclic phosphate was analyzed by 3tp NMR. Results of such experiments indicated
1240 M S Shashidhar
that the conversion of Ptdlns to Ins (1)P by bacterial as well as mammalian PI-PLCs proceeds with retention of configuration at phosphorus. However, the conversion of Ins (1,2-cyc)P to Ins (1)P could be observed directly (by 31p N MR) only in the case of the bacterial enzyme and not when the mammalian enzyme was used. An earlier mechanistic study (Volwerk et al 1990) of the bacterial P I - P L C with the aid of 31P-NMR, using the natural Ptdlns had shown clearly that this enzyme converts Ptdlns 21 to DAG and D-Ins(l)P 23 in a sequential manner (scheme 9).
RO~]J "I II
OH O RO H O
.o .ro_i_o j
P I - PLC ~ H O ~ . ~ O PI--PLCoo@ ~
...
L.../J ...
O H F , , , - ..,,,t. j .... s , , , ,T -,o. .o,'-y--,,o.
OH OH OH
21 22 23
Scheme 9.
The first step involved an intramolecular regiospecific phosphotransfer reaction to form DAG and D-Ins(1,2-cyc)P, 22. The cyclic phosphate so generated is then hydrolyzed stereospecifically at a much slower rate by the same enzyme to D-Ins (1) P.
This is further supported by the fact that 2-O-methyl derivative 58 (R=CH3) of Ptdlns (scheme 20) is an inhibitor of bacterial P I - P L C (Garigapati and Roberts 1993a;
Lewis et al 1993). Thus, for the mammalian enzyme an alternative mechanism in which Ins(1,2-cyc)P and Ins(1)P are formed by independent pathways (perhaps at two different active sites), with the intermediacy of a covalent enzyme-Ptdlns complex (for the formation of Ins (1)P) cannot be ruled out. However, this is unlikely consi- dering the fact that the 2-deoxy analogue 26 (scheme 10) of Ptdlns is not a substrate for a mammalian P I - P L C isolated from human melanoma cell line (Seitz et al 1992), (see below).
When a mixture of diastereomers of Ptdlns (having the D-configuration at myo-inositol and racemic at sn-2 position of 1,2-dipalmitoylglycerol) was used as a substrate for P I - P L C (bacterial as well as mammalian) both the diastereomers were cleaved completely. This indicates that the diacylglycerol moiety is not involved in recognition or binding by the enzyme. This is in contrast to phospholipases A 2 and C which are specific to the L-isomer (Kuipers et al 1990). The stereospecificity of phospholipases D (of plant origin) on the other hand is known to depend on the source of the enzyme (Bugaut et al 1985).
Seitz and others reported the synthesis and enzymatic properties of a 2-deoxy analogue of Ptdlns (Seitz et al 1992). The synthesis (scheme 10) began with racemic tetra-benzyl myo-inositol 2__44DL. The diol 2._44DL was selectively acetylated at 1-position and deoxygenated by thionocarbonate formation and radical reduction. The 2-deoxy myo-inositol derivative was resolved as camphanate ester, and the D-isomer 25D was converted to the Ptdlns analogue 26. The 2-deoxy Ptdlns 26 was evaluated as both a substrate as well as an inhibitor of P I - P L C isolated from human melanoma cell
Phosphatidylinositols and analo#ues 1241
OH H 0 ~O~--H
H ! II _ _ J
.oo o HO
B n O " y ' " O B n B n O " " y " " O B n H O " ' - y ' " O H ... J,,. /I, ... OH
OBn OBn OH
2_4 DL 2_5 D 2_6
R : CH3(CH2)14CO Scheme 10.
line, which has preference for Ptdlns as a substrate. It was revealed that 26 is not a substrate but a weak inhibitor of PI-PLC. Thus this data is consistent with the mechanism in which the cyclic phosphate of myo-inositol is an intermediate, during the cleavage of Ptdlns by PI-PLC, to Ins(l) P. The 2-deoxy Ptdlns would be expected to serve as substrate, if the hydrolysis by P I - P L C proceeded through direct displace- ment by water, on phosphorus to generate Ins(l)P and DAG.
A few short chain Ptdlns were synthesized and used as substrates for bacterial P I - P L C (Garigapati and Roberts 1993b, Lewis et al 1993). The synthesis of these Ptdlns (scheme 11) involved the coupling of the D-enantiomer (or racemic) of 2,3,4,5,6-penta-O-benzyl-myo-inositol 4D (or 4DE) with the H-phosphonates 27 in the presence of NPCL and oxidation of the resulting phosphite with iodine. The benzyl groups were then removed by hydrogenation. Cyclohexylidene or iso- propylidene protected myo-inositols could not be used for the synthesis of short chain phosphoinositides as the final deprotection step requires the use of acidic conditions to which short chain ester group at the sn-2 position is not stable. The diglycerides required for this synthesis were prepared enzymatically from the corresponding phosphatidylcholine and bacterial phospholipase C.
Oq_H 0
L--O--~--OH H
I27
R= CH3(CH2)nCO- ~ n : 4 ~ 5
Scheme 11.
4 D
o. o74-.
OH 28
It was revealed by using synthetic short chain Ptdlns 28 that the bacterial P I - P L C shows a five- to six-fold interfacial activation when its substrate is present at the interface as against a monomer in solution. The use of enantiomeric substrates (with respect to the myo-inositol ring) showed that the lipid head group must have the D-configuration to be a substrate for PI-PLC. The L-isomer is neither a substrate nor an inhibitor. The preference of bacterial P I - P L C for the D-enantiomer for its
1242 M S Shashidhar
phosphotransferase activity leading to the formation of Ins(1,2-cyc) phosphate had earlier been demonstrated using a water-soluble substrate, in absence of any added detergents (Leigh et al 1992). The same enzyme is also known to be stereospecific during its cyclicphosphodiesterase activity i.e. it hydrolyzes only the D-enantiomer of Ins (1, 2-cyc) P to D-Ins (1) P (Volwerk et al 1990).
Hendrickson and co-workers synthesized (scheme 12) thiophosphate analogue 3ODE (Hendrickson et al 1991a) as well as a pyreno derivative 3._[DE of Ins(1)P 4 (Hendrickson et al 1991b) and used them for the kinetic study of PI-PLC from Bacillus cereus (Hendrickson et al 1992). Both the derivatives were prepared from a common racemic penta-protected myo-ifiositol 29DE (R=H). Michaelis-Arbuzov reaction of the dimethylphosphite 29DE (R=P(OCHa)2) with alkylthiophthalimides gave thiophosphates 3ODE. The coupling of the dicyclohexylidene derivative 29DE (R=H) with 4-(1-pyreno) butanol by P(III) approach yielded the fluorescent substrate 31DE.
OH 0
.... L
J
... OHH O , " ' y ~ * O H OH
C) o.
~o~o.,'-.F-',,., ?
2~DL R=H~ P(OCH3) 2
),
/ ",,
OH 0
30 DL R = CHs(CH2)n
n = 7 , il~ 15 Scheme 12.
31DL
Both the PtdIns analogues 3ODL and 3_llDL were tested as substrates for bacterial PI-PLC. The activity of the synthetic substrates 30DE (thio) and 31DE (pyreno) were 1~ and 4 ~ respectively compared to that of the natural substrate. It is interesting to note that thiophospho analogues have also been synthesized as potenial inhibitors of PI-PLC (Alisi et al 1992).
P hosphatid ylinositols and analogues 1243 Synthesis of a few lyso-PtdIns 33, the corresponding ether analogues (glyceryl ether derivatives) and n-alkyl phosphoinositols 34 (scheme 13) were reported recently (Filtuh and Eibl 1992). All the lyso-lipids were prepared from pentaacetyl-myo-inositol 32D by P(III) approach on 5-10g scale in greater than 99% optical purity. The corresponding L-isomers were also synthesized.
OH ~1 OAc
HO~.~r~OPl _OR AcO , ~ H
.... L . J .... OH
H O " " ~ " " O H A cO~*''~"'OAc
OH OAc
o. 0."0~
34. 3__22 D 33
R = alkyl R = CH3(CH2) n CO--
Scheme 13.
Natural PtdIns have several unsaturations in the fatty acid chain and hence are prone to oxidation. It is often convenient to have Ptdlns with saturated side chains as these compounds are stable to oxidation and hence can be stored for extended periods of time. Since enzymes like PI-PLC recognize only the myo-inositol head group, saturated PtdIns can be used as substrates for studies which involve such enzymes. DilaurylPtdIns, dimyristoylPtdIns, dipalmitoylPtdIns and dioleoylPtdIns (36) were synthesized (scheme 14) from purified yeast Ptdlns (Kume et al 1992). First, all the five hydroxy groups of the myo-inositol ring (in yeast Ptdlns) were protected
Yeast
Scheme 14.
o .o_E
.... K . ~o,,. OH
THPO" "~ '"OTH P OTHP
3~
I
0 ( ~ 0 RO H
H 0 - - ~ - - 0
....
L J,.. OH
H O"'y%OH OH
36
R : CH3(CH2)nCO J n= 10~12,14
R = Oleoyl
1244 M S Shashidhar
as tetrahydropyranyl (THP) ethers (35) and the fatty acids were hydrolyzed by treatment with methanolic potassium hydroxide. The free hydroxyls of the glycerol backbone (in 35) were reacylated with acyl anhydrides followed by the hydrolysis of the T H P ethers to obtain PtdIns (30 with well-defined fatty acid chains. The synthetic PtdIns were obtained in 32-43% yield.
An enzymatic synthesis (scheme 15) of pyrene labelled (on the fatty acid chain) phosphoinositides and their aggregation states in organic solvents and the lateral mobility in lipid bilayers was investigated by OadeUa and co-workers (Gaddla et al 1990). Yeast Ptdlns was acetylated to protect the five hydroxy groups of the myo.
inositol moiety and the sn-2 fatty acyl chain was cleaved by treatment with Crotalus adamanteus phospholipase A2 (37). Reacylation of the free hydroxy group so generated with the pyrene-labelled fatty acid followed by deacetylation of the acetates gave the pyrene-labelled PtdIns 38. A similar sequence of reactions with PtdIns(4, 5)P2 was not successful, as acetylated Ptdlns(4, 5)P2 was not a substrate for phospholipase A2. However, pyrene-labelled Ptdlns could be phosphorylated with Ptdlns kinase and PtdIns(4)P kinase sequentially, to obtain the phosphorylated derivatives of pyrene-labelled Ptdlns (39 and 40). This perhaps suggests that the fatty acid composition of the fatty acid side chains have no effect on the activity of these kinases. The aggregation behaviour of these lipids when examined by fluorescence techniques showed that phosphorylated forms of PtdIns form micellar structures in organic solvents. When examined in phosphatidylcholine bilayers, it was revealed that the collision frequency of PtdIns decreased with increase in the degree of phosphorylation.
F O R t r--OR I
oAc o HO,'F'H OH O R'O--4--H
,oo, k o- -oJ
.,,,L
.;L OH..,,L.
],,o OHA c O""y'"OAc HO""~"~
OA: OH
37 38
R = Pyrenyl decanoyl z
4- K i nose
z F "OR1 r-OR1
HO 0 R O="~H OH 0 RZo~L=H
- - - 5-Kino,, H O , ~ O - - ~ - - u H O ' - J = ( ~ (3....L.v.j,,.,,OH OH
OH
40 39
Scheme 15.
P hosphatid ylinositols and analogues 1245 Phosphatidylinositol analogues as inhibitors
Synthetic Ptdlns derivatives that are not degraded by P I - P L C are useful as inhibitors in enzymatic studies and as ligands for the isolation and purification of Ptdlns binding proteins. They are also of medicinal interest as pharmacological agents. Since P I - P L C cleaves the P - O bond (connected to DAG, see scheme 1) in Ptdlns, replacing this P - O bond by a P - C bond (which is non-hydrolyzable) generates molecules that are inhibitors of PI-PLC. This approach was followed (Shashidhar et a11990a) for the synthesis of phosphonate analogues 41-44 of Ptdlns (scheme 16). Racemic penta- benzyl-myo-inositol ~DL) was condensed with phosphonic acids in the presence of trichloroacetonitrile and the benzyl protecting groups were cleaved by hydrogenolysis. All the phosphonate derivatives were stable and there was no evidence for the migration of the phosphonate group to other hydroxyl groups of the myo-inositol ring.
o . o
-O.o. o.
OH
Scheme 16.
41" n = 2 , R--CH 3 4._.22 " n = 9 , R=COOH
4:3 " n = 3 , R--CH3(CH2)14CO0- 44" n = 3 , R= H2N (CH2)11 O -
All the non-hydrolyzable analogues 41-44 inhibited the activity of P I - P L C from B. cereus in a mixed micellar assay system containing sodium deoxycholate (Shashidhar et a11990b). The phosphonate analogue 44 (scheme 16) of Ptdlns with an amino group at the end of aliphatic chain was used as a ligand for the affinity chromatographic purification of P I - P L C from B. cereus (Shashidhar et a11990c).
This amino derivative on coupling to cyanogen bromide activated Sepharose resulted in an affinity matrix that was specific for B. cereus PI-PLC. Competition experiments with myo-inositol in the elution medium indicated that the specific binding of the enzyme to the affinity matrix most likely involved the enzyme active site. Although the myo-inositol derivative 44 was not a very good inhibitor of P I - P L C (it exhibited an ICso value of 10 mM at a substrate concentration of 2 mM, when used in a mixed micellar assay) it is likely that the phosphonate coupled to the Sepharose matrix allowed more efficient binding of the enzyme as compared to the phosphonate analogue present in a mixed micellar medium.
Synthesis of some racemic Ptdlns and Ptdlns (4,5)P2 analogues in which oxygen atom at the C-1 position of the myo-inositol ring is replaced by a carbon atom were synthesized (James et a11990) in order to see the behavior of P I - P L C towards these homologated myo-inositol esters. The inosose 45DL (scheme 17) was prepared starting from myo-inositol through the intermediacy of the 1,2-cyclohexylidene derivative.
Wittig reaction of the inosose 4...55DL with methyl(triphenyl) phosphonium bromide followed by reduction with borane methylsulphide complex gave a mixture of two
1246 M S Shashidhar
OBn OH 0
B n O ' ~ , , O ~ ----e. H O ~ r ~ C H 2 0 C (CHz)I4CH3
r
OR ~ ~)R 2
4._55 D L : R t = R Z = B n 4...66 " RZ= H
4"/ : RI=RZ=PMB 4._88 " R z = PO3H2 4._99 " R ~' = S0-5 Na §
Scheme 17.
homologated myo-inositol derivatives, which were converted to the corresponding palmitates 46. Phosphorylated and sulphated derivatives 48 and 49 were synthesized by a similar route by the temporary protection of the 4- and 5-hydroxy groups as 4-methoxybenzyl ethers, followed by phosphorylation with bis(benzyloxy)(diiso- propylamino) phosphine in presence of 1H-tetrazole or sulphation with sulphurtri- oxide in pyridine. Among the compounds synthesized, analogues of PtdIos(4,5)P2 inhibited the activity of P I - P L C in vitro but were inactive in the assay with intact cells (Kaufmann et al 1991).
_4D - - - ~ - - - ~
Scheme l&
o . o ~
H O ~ / , O - - P--O ..--~
H O , J ~ H ,,,,0 H CH3
5._00 9 R = CHs( C H P ) 1 4 C O -
Uncharged methylphosphonate analogues 50
(Rp
and S~) of PtdIns (scheme 18), which may act as P I - P L C inhibitors were synthesized from enantiomeric 2,3,4,5,6-penta-O-benzyl-myo-inositol 4D and 1,2-dipalmitoyl-sn-glycerol (Dreef et al 1991). The two alcohols were coupled using a bifunctional phosphonylating agent, bis[6-(trifluoromethyl) benzotriazol-l-yl] methylphosphonate. Rp and Sp isomers could be separated before the final deprotection of myo-inositol hydroxy groups.Alisi and co-workers prepared several analogues 5(~-5__5,5 scheme 19) of PtdIns with thiophosphoester bonds, starting from racemic 1,1'-dithiobis(2,3-propanediol) and racemic penta-benzyl-myo-inositol _4DL (Alisi et al 1992). These analogues were expected to function as novel P I - P L C inhibitors. They also synthesized a racemic dithio analogue 57 (scheme 19) of PtdIns and determined its cytotoxic activity and its effect on mammalian P I - P L C (Alisi et al 1993). Racemic octadecylphosphodi- thionyl-l-myo-inositol was prepared by the condensation of penta-benzyl-myo- inositol 4DL and octadecanethiol by P(III) approach followed by sulphurization with sulphur. The product obtained was a mixture of four isomers, due to the presence of two chiral centres at C-1 of the myo-inositol ring and phosphorus. The mixture of diastereomers caused 40% cellular death in vivo (in K562 erythroleukemia cell line)
Phosphatid ylinositols and analogues 1247
o .
H O , , ' y % , O H OH OH
5"1 , R I = C H s ( C H 2 ) . n , 'n = 0 , 4 , 8 , 1 0 , 1 2 , 1 4 R 2== H
5._.22 , R' = R z = C H 3 ( C H 2 ) ~ C O 0 - , rt = 6 , 8 , 14 5...33, R D == C H s ( C H 2 ) 1 6 C O 0 - ; R 2 = C H 3 C 0 0 - 54. , Fr I = R 2 = C H 3 ( C H 2 ) . n O - ; 'n = 7 , 17 555 , R I = C H 3 ( C H 2 } n O - , ~ =" 7 , 17 ; R 2 = C H 3 0
OH $
R S ~ p ~ O C H 3 II
_(, N,,r_ I % . o.
5 6 , R = CH3{CH2)17 57
Scheme 19.
and inhibited the activity of P I - P L C from human platelets to the extent of 50~, at a concentration of 10-4 M. These observations seem to suggest a relationship between cytotoxic activity and enzyme inhibition. However, the fact that cytotoxic activity could be due to other activities of the compound could not be ruled out.
1 - - - - OR 2
o
R 2 0 " ~ HOR' ~--o--]
H O ~ O - - I ,* %
24. D L 9 > H O , , , ~ " ~ % , , O H OH OH
5._88, R ' = C H 3 or C H 3 ( C H 2 } 6
R 2 = C H 3 ( C H 2 ) s C O - Scheme 20.
2-O-alkyl PtdIns were prepared and tested as inhibitors of P I - P L C (Garigapati and Roberts 1993b; Lewis et al 1993). The racemic diol 24DL (scheme 20) was converted to 2-O-alkyl-3,4,5,6-tetra-O-benzyl-myo-inositol, and phosphorylated by H-phosphonate methodology. All the benzyl ethers were cleaved by hydrogenolysis to obtain 2-O-alkyl-PtdIns 58. Diheptanoyl (2-O-methyl) PtdIns 58 (RI:CH3, R2--CH3(CH2)sCO) when present at a concentration of 0.75mM inhibited the cleavage of 4.5 mM diheptanoyl PtdIns to the extent of about 50%. This result indicates
1248 M S Shashidhar
that the presence of the axial 2-hydroxy group is essential for the enzyme activity but is not required for the binding of the substrate to the enzyme. Thus, this class of compounds forms the mechanism based inhibitors for PI-PLC.
Glycosyl phosphatidylinositols and ceramide analogues
GPI anchors are reported to be involved in the signal transduction of insulin (Ferguson and Williams 1988) and nerve growth factor (Chan et al 1989). This has generated an interest in the chemical synthesis of GPI and its analogues. The synthetic studies related to GPI anchors involve extensive use of carbohydrate chemistry which is out of scope of the present review. Accordingly only a brief mention of such reports is included here.
OR OH ~1 0 - - H O,~,L.~%,,O OH
OH
Scheme 21.
OH 0
OH
5"9 OH
Murakata and Ogawa reported the synthesis of glycobiosyl Ptdlns 59 (scheme 21) which is a part of the GPI anchor (Murakata and Ogawa 1991). The same authors later reported the first stereoselective total synthesis of the GPI anchor of Trypanasoma brucei (Murakata and Ogawa 1992). StereocontroUed glycosylation techniques were used to construct the oligosaccharide moiety and highly efficient H-phosphonate approach was used to build the two phosphodiesters.
Elie and co-workers reported the synthesis of a mycobacterial phospholipid fragment 63, which involved the iodonium mediated mannosylation of myo-inositol (Elie et al 1989, 1992). The axial hydroxy group of the penta protected myo-inositol 60 or 61 (scheme 22) was first coupled with a D-mannopyranoside derivative using N-iodosuccinimide and trifluoromethanesulphonic acid. Unexpectedly, the mixture of products contained 1,2-trans as well as the 1,2-cis linked diastereomeric dimers.
For the synthesis of 63, the p-methoxybenzyl group was cleaved by oxidation with DDQ, and the diastereomers were separated by column chromatography. A second mannopyranosyl unit was introduced at the 6-position of the myo-inositol ring as earlier and the allyl ether was cleaved selectively. Phosphorylation of the free 1-hydroxy group using H-phosphonate methodology followed by deprotection of the hydroxy groups gave the mycobacterial lipid fragment 63.
Cottaz and co-workers report the synthesis (scheme 23) of glucopyranosyl myo-inositol derivatives 65D and 65L (Cottaz et al 1993). They coupled the enantiomeric myo-inositol derivative 64D (or 64L) with a protected 2-azidoglucosyl fluoride in the
Phosphatidylinositols and analogues 1249
8nO~OA
OH_ _ 60, R = PM B BnO,.,~'~y'*",,OR ~ %,,
$
OBnOH OH
HO'~o
o
H O , ~ V ' % O OH
o. ~o__k.,~o"
6:3 I ~ ~ OH
HO OH
Scheme 22.
OBn B n O ~ O P M B
n
O"'~T%"O'~ ~H
B
OBn 64 D Scheme 23.
6'1, R=Bn
$
.o; x-~-i-~
OHo. o
o .
62
OH 0 RO . , ~ H ~
presence of zirconocene dichloride and silver perchlorate. The 4-methoxybenzyl ether was then cleavedselectively. The =-configuration of the anomeric linkage of the pseudo-disaccharide was assigned based on the t H - N M R spectrum. H-phosphonate method was used for the phosphorylation and deprotection of all the hydroxy groups as well as the reduction of the azide to the corresponding amine was carried out in a single step by hydrogenolysis to obtain the phospholipid 65D (or 65L)
D-erythro-ceramide-l-phosphoinositol and its derivatives are known to occur in some bacteria, fungi, yeast and plants, but their functions are not known. The derivatives of D-erythro-ceramide-l-phosphoinositol usually contain one or more monosaccharide residues attached at 2- or 6-position of the myo-inositol ring.
Frantova and co-workers (Frantova et al 1992) synthesized (scheme 24) a diastereo- meric mixture of the ceramide phosphoinositol 66 from penta acetyl-myo-inositol 32DL and racemic 3-benzoyl ceramide. The condensation was carried out using the bifunctional phosphitylating agent 2-cyanoethyl N,N,N',N'-tetra-iso-propylpho- sphoramidite. Kratzer and co-workers (Kratzer et al 1993) recently reported the synthesis of 69 from the penta-protected inositol derivative 67 and the protected azido sphingosine 68.
1250 M S Shashidhar
3_.22 O L "--~
OH 0 OH
HO.~O--P--O~
II.... L v J .... I o l d , H 0," y ' % O H OH HNR
OH
66, R! = C17H55 CO- , R 2 = Ct5H31 OR
R O ~ , OH R O " * y " " O R
OR
6 . 7 , R = M O M
Sel,em, e ?.,4.
N5
" R2
OTBDMS
---@
~)H 0 NHR'
/ II = 2
HO~,,"'K",,.~O~P--O.._ ~ A ~ / R
.L .i,.. , v y - . . .
H 0'~ ~ " " 0 H OH OH OH
6._.99, R I = CH3(CH2)troCO- R2--- C H3(C H2)12-
Conclusions
The present survey of the recent literature shows that several methodologies for the synthesis of Ptdlns and their analogues have been developed and that the availability of synthetic phosphoinositides and their suitably tailored analogues is essential to understand the functioning of the enzymes associated with the Ptdlns cycle. We hope that the wealth of information made available in the recent past will help to understand the complexities of the Ptdlns cycle and the Ins(l,4,5)P3 mediated signal transduction pathway and aid in the development of drugs/cures for many diseases.
References
Alisi M A, Brufani M, Filocamo L, Gostoli G, Cellai L, IanelE M A, Melino G, Cesta M C and Lappa S 1993 Bioorg. Med. Chem. Lett. 3 1931
Alisi M A, Brufani M, Filocamo L, Gostoli G, Lappa S, Maiorana S, Cesta M C, Ferrari E and Pagella P 1992a Tetrahedron Lett. 33 3891
Alisi M A, Brufani M, Filocamo L, Gostoli G, Maiorana S, Cesta M C, Ferrari E, Lappa S and Pagella P 1992b Tetrahedron Lett. 33 7793
Anderson R J, Lothrop W C and Creighton M M 1938 J. Biol. Chem. 125 299 Anderson R J and Roberto E G 1930 J. Am. Chem. Soc. 52 5023
Billington D C 1989 Chem. Soc. Rev. 18 83
Bruzik K S, Morocho A M, Jhon D Y, Rhee S G and Tsai M D 1992 Biochemistry 29 2747 Bruzik K S and Tsai M D 1992 J. Am. Chem. Soc. 114 6361
Bugaut M, Kuksis A and Myher J J 1985 Bioehim. Biophys. Acta. 835 304
Phosphatidylinositols and analogues 1251
Chan B L, Chao M V and Saltiel A R 1989 Proc. Natl. Acad. Sci. USA 86 1756 Cleves A, McGee T and Bankaitis V 1991 Trends Cell Biol. 1 30
Cosgrove D J 1980 lnositolphosphates. Their chemistry, biochemistry and physiology (New York: Elsevier) Cottaz S, Brimacombe J S and Ferguson M A J 1993 J. Chem. Soc., Perkin Trans. I 2945
Dawson R M C, Freinkel N, Jungalwala F B and Clarke N 1971 Biochem J. 122 605 Dreef C E, Douwes M, Elie C J J, van der Marel G A and van Boom J H 1991 Synthesis 443 DreefC E, Elie C J J, Hoogerhout P, van der Marel G A and van Boom J H 1988 Tetrahedron Lett. 29 6513 Elie C J J, Dreef C E, Verduyn R, van der Marel G A and van Boom J H 1989 Tetrahedron 45 3477 Eiie C J J, Verduyn R, Dreef C E, van der Marel G A and van Boom J H 1992 J. Carbohyd. Chem. 11 715 Ferguson M A J, Low M G and Cross G A M 1985 J. Biol. Chem. 260 14547
Ferguson M A J and Williams A F 1988 Annu. Rev. Biochem. 57 285 Filtuh E and Eibl H 1992 Chem. Phys. Lipids 60 253
Foich J and Wooley D W 1942 J. Biol. Chem. 142 963
Frantova A Y, Stepanov A E, Bushnev A S, Zvonkova E N and Shvets V 11992 Tetrahedron Lett. 33 3539 Gadella T W A Jr, Moritz A, Westerman J and Wirtz K W A 1990 Biochemistry 29 3389
Garigapati R and Roberts M F 1993 Tetrahedron Lett. 34(a) 769, (b) 5579 Hay J C and Martin T F J 1993 Nature (London) 366 572
Hendrickson H S, Hendrickson E K, Johnson J L, Khan T H and Chial H J 1992 Biochemistry 12169 Hendrickson E K, Johnson J L and Hendrickson H S 1991 Bioorg. Med. Chem. Lett. (a) 6t5, (b) 619 James D, Massy R and Wyss P 1990 Heir. Chim. Acta. 73 1037
Kaufman F, Massy D J R, Pirson W and Wyss P 1991 in Inositol phosphates and derivatives: Synthesis, biochemistry and therapeutic potential (ed.) A B Reitz (ACS Symposium series 463) p. 202
Kim J W, Ryu S H and Rhee S G 1989 Biochem. Biophys. Res. Commun. 163 177
Koke J A, Yang M, Henner D J, Volwerk J J and Grimth O H 1991 Protein Expression Purification 2 51 Kratzer B, Mayer T G and Schmidt R R 1993 Tetrahedron Lett. 34 6881
Kuipers O P, Dekker N, Verheij H M and de Haas G-H 1990 Biochemistry 29 6994
Kume T, Taguchi R, Tomita M, Tokuyama S, Morizawa K, Nakachi N, Hirano I and Ikezawa H 1992 Chem. Pharm. Bull. 40 2133
Lee Y C and Ballou C E 1965 Biochemistry 4 1395
Leigh A J, Volwerk J J, Griffith O H and Keana J F W 1992 Biochemistry 31 8978 Lewis K A, Garigapati V R, Zhou C and Roberts M F 1993 Biochemistry 32 8836 Lin G, Bennet F C and Tsai M D 1990 Biochemistry 29 2747
Lindh I and Stawinski 1989 J. Org. Chem. 54 1338, and references cited therein
Majerus P W, Ross T S, Cunningham T W, Caldwell K K, Jefferson A B and Bansal U S 1990 Cell63 459 Murakata C and Ogawa T 1991 Tetrahedron Lett. 32 101
Murakata C and Ogawa T 1992 Carbohyd. Res. 235 95 Parthasarathy R and Eisenberg F Jr 1986 Biochem. J. 235 313
Potter B V L 1990 In Comprehensive medical chemistry (ed.) C Hansch (New York: Pergamon Press) vol. 3 Salamonczyk G M and Bruzik K S 1990 Tetrahedron Lett. 31 2015
Seitz S P, Kaltenbach R F I I I , Vreekamp R H, Calabrese J C and Perella F W 1992 Bioor 0. Med. Chem.
Lett. 2 171
Skinner-H B 1993 EMBO J. 12 4775
Shashidhar M S, Keana J F W, Volwerk J J and Griffith O H 1990a Chem. Phys. Lipids 53 103 Shashidhar M S, Keana J F W, Volwerk J J and Griflith O H 1990b Chem. Phys. Lipids 56 159 Shashidhar M S, Volwerk J J, Keana J F W and Griffith O H 1990c Biochim. Biophys. Acta 1042 410 Shashidhar M S, Volwerk J J, Keana J F W and Grifflth O H 1991a Chem. Phys. Lipids 60 101 Shashidhar M S, Volwerk J J, Keana J F W and Griffith O H 1991b Anal. Biochem. 198 10 Volwerk J J, Shashidhar M S, Kuppe A and Griffith O H 1990 Biochenistry 29 8056 Whitman M, Downes C P, Keeler M, Keller T and Cantley L 1988 Nature (London) 332 644
Young R C, Downes C P, Eggleston D S, Jones M, Macphee C H, Rana K K and Ward J G 1990 J.
Med. Chem. 33 641
