Multiple metabolic pools of phosphoinositides and phosphatidate in human erythrocytes incubated in a medium that permits rapid transmembrane exchange of phosphate

Mass measurement of inositol phosphates

This review summarises the methods available for the mass measurement of inositol phosphates, i.e., use of radioactive inositol lipid precursors, optical techniques, gas chromatography, mass spectrometry, nuclear magnetic resonance, fast atom bombardment and assays specific for Ins(1,4,5)P3. Examples of the use of each method, its sensitivity, advantages and drawbacks are given.

Functional heterogeneity of polyphosphoinositides in human erythrocytes

After labelling of erythrocytes with [32P]P1 for 23 h, the specific radioactivities of the phosphomonoester groups of PtdIns4P and of PtdIns(4,5)P2 approached equilibrium values which were close to that of the gamma-phosphate of ATP (78-85%), showing that almost all of these phosphate groups were metabolically active. Phosphoinositidase C (PIC) activation, using Ca2+ and the ionophore A23187, of 32P-prelabelled erythrocytes was used to investigate a possible functional heterogeneity of the phosphoinositides. Hydrolysis of PtdIns(4,5)P2, measured from its radioactivity, decreased as function of the time of prelabelling up to a constant value equal to that measured from its content. In contrast, hydrolysis of PtdIns4P, determined both from radioactivity and from content, was always the same. These data suggest that newly labelled molecules of PtdIns(4,5)P2, initially accessible to PIC, then moved towards a PIC-resistant pool. This was further confirmed by measuring the fraction of labelled PtdIns(4,5)P2 molecules accessible to PIC after a prelabelling period of 5 min and different times of reincubation. Hydrolysis by PIC was also measured in erythrocytes in which the phosphoinositide content had been modified by activation (Mg2+-enriched cells) or inhibition (ATP-depleted cells) of the phosphoinositide kinases. The sizes of the PIC-resistant pools of polyphosphoinositides were not affected by these treatments, indicating that the kinases (and the phosphatases) act on the PIC-sensitive pools. This was also shown by the decrease in the production of Ins(1,4,5)P3 upon PIC activation in ATP-depleted erythrocytes. A model is presented in which the PIC-sensitive pools of polyphosphoinositides are those which are accessible to the kinases and the phosphatases and are rapidly turned over.

Association of profilin with filament-free regions of human leukocyte and platelet membranes and reversible membrane binding during platelet activation

Profilin is a conserved, widely distributed actin monomer binding protein found in eukaryotic cells. Mammalian profilin reversibly sequesters actin monomers in a high affinity profilactin complex. In vitro, the complex is dissociated in response to treatment with the polyphosphoinositides, phosphatidylinositol monophosphate, and phosphatidylinositol 4,5-bisphosphate. Here, we demonstrate the ultrastructural immunolocalization of profilin in human leukocytes and platelets. In both cell types, a significant fraction of profilin is found associated with regions of cell membrane devoid of actin filaments and other discernible structures. After platelet activation, the membrane association of profilin reversibly increases. This study represents the first direct evidence for an interaction between profilin and phospholipids in vivo.

Analysis of the regulation of phosphatidylinositol-4,5-bisphosphate synthesis by arachidonic acid in exocrine pancreas

In pancreatic acinar cells prelabeled with either 32Pi or myo-[3H]inositol, arachidonic acid (10-50 microM) rapidly decreased the steady-state levels of [32P]phosphatidylinositol 4',5'-bisphosphate [( 32P]PtdIns4,5P2) and inhibited carbachol-stimulated accumulation of [3H]inositol trisphosphate [( 3H]InsP3). Both actions of arachidonic acid were rapidly reversed by bovine serum albumin (BSA). Indomethacin and nordihydoguaiaretic acid failed to block the inhibitory effects of arachidonic acid on [32P]PtdIns4,5P2 levels. Arachidonic acid (10-50 microM) also caused a prompt depletion of cellular ATP which was rapidly reversed by BSA. The ATP-depleting action of arachidonate paralleled in terms of concentration dependence and time course its inhibitory effects on [32P]PtdIns4,5P2 and [3H]InsP3 levels. Exposure of acinar cells to 50 microM arachidonic acid produced an increase in oxygen consumption which exceeded that elicited by either carbachol or ionomycin. Arachidonic acid (10-50 microM) also caused a concentration-dependent rise in cytosolic Ca2+, which was partially obtunded by Ca2+ deprivation. A proposed mechanism involving arachidonic acid as a negative feedback regulator of polyphosphoinositide turnover in exocrine pancreas is discussed.

Platelet-derived-growth-factor-stimulated heterogeneous polyphosphoinositide metabolism and phosphate uptake in C3H fibroblasts

Pig platelet-derived growth factor (PDGF) increased the rate of [32P]Pi uptake by murine fibroblasts, resulting in a 3-9-fold elevation of the specific radioactivity of ATP, PtdInsP, PtdInsP2, PtdIns and phosphatidic acid. The specific radioactivity was 10-60-fold higher in ATP than in the four phospholipids. These substances are therefore not in metabolic equilibrium, which complicates determination of inositol phospholipid turnover.

The separation of [32P]inositol phosphates by ion-pair chromatography: optimization of the method and biological applications

We have developed an ion-pair reverse-phase HPLC method to measure inositol phosphates in 32P-labeled cells. The different chromatographic parameters were analyzed to optimize the resolution of the 32P-labeled metabolites. Analysis of inositol phosphates in biological samples was improved by a single charcoal pretreatment which eliminated interfering nucleotides without removing inositol phosphates. The kinetics of production of inositol phosphates in calcium-activated erythrocytes, vasopressin-stimulated hepatocytes, and thrombin-activated platelets were analyzed. Original data on the activation of phosphoinositide phospholipase C were obtained in intact erythrocytes by direct measurement of inositol (1,4,5)P3. Data from agonist-stimulated hepatocytes and platelets were consistent with those from previous studies. In conclusion, this technique offers many advantages over the methodologies currently employed involving anion-exchange chromatography and [3H]inositol labeling: (i) 32P labeling is less expensive and more efficient than 3H labeling and can be used with all types of cells without permeabilization treatments and (ii) ion-pair HPLC gives good resolution of inositol phosphates from nucleotides with shorter retention times, and long reequilibration periods are not required.

Determination of the steady-state turnover rates of the metabolically active pools of phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate in human erythrocytes

When intact human erythrocytes are incubated at metabolic steady state in a chloride-free medium containing [32P]Pi, there is rapid labelling of the gamma-phosphate of ATP, followed by a slower labelling of the monoester phosphate groups of phosphatidylinositol 4-phosphate (PtdIns4P) and phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P2] [King, Stephens, Hawkins, Guy & Michell (1987) Biochem. J. 244, 209-217]. We have analysed the early kinetics of the labelling of these phosphate groups, in order to determine: (a) the steady-state rates of the interconversions of phosphatidylinositol, PtdIns4P and PtdIns(4,5)P2; and (b) the fractions of the total cellular complement of PtdIns4P and PtdIns(4,5)P2 that participate in this steady-state turnover. The experimental data most closely fit a pattern of PtdIns4P and PtdIns(4,5)P2 turnover in which one-quarter of the total cellular complement of each lipid is in the metabolic pool that participates in rapid metabolic turnover, with rate constants of 0.028 min-1 for the interconversion of PtdIns and PtdIns4P, and of 0.010 min-1 for the PtdIns4P/PtdIns(4,5)P2 cycle. These rate constants represent metabolic fluxes of approx. 2.1 nmol of lipid/h per ml of packed erythrocytes between PtdIns and PtdIns4P and of approx. 5.7 nmol/h per ml of cells between PtdIns4P and PtdIns(4,5)P2.

Increased phosphoinositide breakdown by phospholipase C in erythrocyte membranes from patients with cystic fibrosis

Phosphoinositide phospholipase C activity has been measured in erythrocyte membranes from age-matched control and CF subjects. Inositol phospholipids were labelled with [3H]myo-inositol and control experiments demonstrated that the [3H]-labelled products released by incubation of membranes with Ca2+ were derived specifically from erythrocytes (a) by purification of erythrocytes on cellulose columns, (b) by demonstration that the phospholipase C activity was inhibited by 10 mmol/l neomycin but not by 1 mmol/l p-methylsulphonylfluoride. The [3H]-labelled products were shown to be inositol phosphates by their elution from anion-exchange columns. Membranes from CF patients showed increased phospholipase C activity compared to controls which did not correlate with the degree of [3H]inositol labelling of the membranes, with pancreatic function as assessed by serum immunoreactive trypsin or with medications taken by the patients.

Formation of [3H]inositol metabolites in rat hippocampal formation slices prelabelled with [3H]inositol and stimulated with carbachol

Rat hippocampal formation slices were prelabelled with [3H]inositol and stimulated with carbachol for times between 7 s and 3 min. The [3H]inositol metabolites in an acid extract of the slices were resolved with anion-exchange HPLC. Carbachol dramatically increased the concentration of [3H]inositol monophosphate, [3H]inositol bisphosphate (two isomers), [3H]inositol 1,3,4-trisphosphate, [3H]inositol 1,4,5-trisphosphate, and [3H]inositol 1,3,4,5-tetrakisphosphate. The levels of [3H]inositol 1,4,5-trisphosphate rose most rapidly; they were maximally elevated after only 7 s and declined toward control levels in 1 min followed by a more sustained elevation in levels for up to 3 min. When [3H]inositol 1,4,5-trisphosphate was incubated with hippocampal formation homogenates in an ATP-containing buffer it was very rapidly metabolised. After 5 min [3H]inositol 1,4-bisphosphate, [3H]inositol 1,3,4-trisphosphate, and [3H]inositol 1,3,4,5-tetrakisphosphate could be detected in the homogenates. Under similar experimental conditions [3H]inositol 1,3,4,5-tetrakisphosphate is metabolised to [3H]inositol 1,3,4-trisphosphate and an inositol bisphosphate isomer that is not [3H]inositol 1,4-bisphosphate. We conclude that like other tissues the primary event in the hippocampus following carbachol stimulation is the activation of phosphatidylinositol 4,5-bisphosphate selective phospholipase C.

Inositol lipid turnover and compartmentation in canine trachealis smooth muscle

We established conditions for the study of metabolism and compartmentation of inositol phospholipids in canine trachealis muscle. Unstimulated muscle was incubated with myo-[3H]inositol for 30 min at 37 degrees C which resulted in labeling of the tissue free myo-inositol pool, whereas only a small amount of radioactivity was incorporated into inositol phospholipids or inositol phosphates. After addition of 5.5 microM carbachol, phosphatidylinositol (PI), phosphatidylinositol-4-phosphate (PIP), and phosphatidylinositol-4,5-bisphosphate (PIP2), specific radioactivities increased exponentially, reaching apparent constant values in 180-240 min. Initial rates of increases in PI, PIP, and PIP2 specific radioactivities were 39, 32, and 66 times that measured in unstimulated muscle. Metabolic flux rates (nmol.100 nmol total lipid Pi-1.min-1) during development of force averaged 0.42 +/- 0.09 and during force maintenance averaged 0.14 +/- 0.01. Fractions of total PI, PIP, and PIP2 pools that were linked to muscarinic cholinergic activation were estimated to be 0.97, 0.85, and 0.65, respectively. Initial rates of increase in specific radioactivities and specific radioactivities during carbachol activation were similar in PI, PIP, and PIP2 fast active compartments, suggesting metabolic flux from PI to PIP to PIP2 was in near chemical equilibrium. Turnover times for PI, PIP, and PIP2 fast active compartments were estimated to be 21, 1.6, and 4.0 min, respectively.

1D-myo-inositol 1,4,5-trisphosphate dephosphorylation by rat enterocytes involves an intracellular 5-phosphatase and non-specific phosphatase activity at the cell surface

We studied the dephosphorylation of Ins(1,4,5)P3 (inositol 1,4,5-trisphosphate) by permeabilized rat intestinal epithelial cells incubated in a medium resembling intracellular ionic strength and pH. Saponin-permeabilized cells rapidly dephosphorylated Ins(1,4,5)P3 to a mixture of three InsP2 (inositol bisphosphate) isomers, namely Ins(1,4)P2, Ins(1,5)P2 and Ins(4,5)P2. These products were identified by h.p.l.c. analysis after dephosphorylation of both 3H- and 32P-labelled Ins(1,4,5)P3. Ins(1,4)P2 accumulated to about half of the concentration attained by Ins(1,5)P2 and Ins(4,5)P2. Ins(1,4,5)P3 dephosphorylation was inhibited, by up to 75%, by 10 mM-glucose 6-phosphate. In these conditions Ins(1,4)P2 became the predominant product, indicating that glucose 6-phosphate inhibited non-specific dephosphorylation of Ins(1,4,5)P3, at least at the 1- and 4-phosphate groups. Ins(1,4)P2 was further dephosphorylated, and the major InsP (inositol monophosphate) product was Ins4P. Most of the glucose 6-phosphate-inhibitable Ins(1,4,5)P3 phosphatase activity was exposed on the cell surface. The glucose 6-phosphate-insensitive Ins(1,4,5)P3 5-phosphatase activity was not detected until the cells were permeabilized with saponin. This intracellular 5-phosphatase activity was: (i) predominantly associated with the particulate portion of the cell; (ii) strongly inhibited by 10 mM-2,3-bisphosphoglycerate; (iii) insensitive to 50 mM-Li+. Therefore the Ins(1,4,5)P3 5-phosphatase activity in enterocytes appears similar to the 5-phosphatase that has been characterized in a number of cell types.

Adsorption of cations to phosphatidylinositol 4,5-bisphosphate

We investigated the binding of physiologically and pharmacologically relevant ions to the phosphoinositides by making 31P NMR, electrophoretic mobility, surface potential, and calcium activity measurements. We studied the binding of protons to phosphatidylinositol 4,5-bisphosphate (PIP2) by measuring the effect of pH on the chemical shifts of the 31P NMR signals from the two monoester phosphate groups of PIP2. We studied the binding of potassium, calcium, magnesium, spermine, and gentamicin ions to the phosphoinositides by measuring the effect of these cations on the electrophoretic mobility of multilamellar vesicles formed from mixtures of phosphatidylcholine (PC) and either phosphatidylinositol, phosphatidylinositol 4-phosphate, or PIP2; the adsorption of these cations depends on the surface potential of the membrane and can be described qualitatively by combining the Gouy-Chapman theory with Langmuir adsorption isotherms. Monovalent anionic phospholipids, such as phosphatidylserine and phosphatidylinositol, produce a negative electrostatic potential at the cytoplasmic surface of plasma membranes of erythrocytes, platelets, and other cells. When the electrostatic potential at the surface of a PC/PIP2 bilayer membrane is -30 mV and the aqueous phase contains 0.1 M KCl at pH 7.0, PIP2 binds about one hydrogen and one potassium ion and has a net charge of about -3. Our mobility, surface potential, and electrode measurements suggest that a negligible fraction of the PIP2 molecules in a cell bind calcium ions, but a significant fraction may bind magnesium and spermine ions.

Dephosphorylation of 1D-myo-inositol 1,4-bisphosphate in rat liver

Dephosphorylation of 1D-myo-inositol 1,4-bisphosphate [Ins(1,4)P2] in rat liver is catalysed by a cytosolic phosphatase that removes the 1-phosphate group. The Km for Ins(1,4)P2 is approx. 17 microM. Li+ (100 mM) causes 50% inhibition of Ins(1,4)P2 phosphatase activity when activity is measured at the very low substrate concentration of 10 nM, but on raising the substrate concentration to 100 microM there is a greater than 10-fold increase in sensitivity to Li+, suggesting that Li+ acts mainly, but not entirely, as an uncompetitive inhibitor of Ins(1,4)P2 phosphatase. In addition, rat liver cytosol shows Li+-sensitive phosphatase activity against 1D-myo-inositol 1-,3- and 4-monophosphates. The Ins(1,4)P2 1-phosphatase and inositol monophosphatase activities all share an apparent Mr of 47 x 10(3), as determined by gel-filtration chromatography. However, the Ins(1,4)P2 1-phosphatase is more sensitive to inactivation by heat, and can be separated from inositol monophosphatase activity by anion-exchange chromatography. We conclude that rat liver cytosol contains an Ins(1,4)P2 1-phosphatase that is distinct from, but in many ways similar to, inositol monophosphatase.

Erythrocytes attached to a wheat germ agglutinin coated surface display an altered phospholipid metabolism

Erythrocytes were bound to a lectin-coated surface; the multivalent attachment to this surface resulted in a severe deformation of the cells and an alteration in the cellular phospholipid metabolism. Human erythrocytes were allowed to bind for 20 min at 20 degrees C to polystyrene beads coated with wheat germ agglutinin (WGA beads). The bound erythrocytes were then lysed to produce stroma bound to WGA beads. Control stroma and stroma-WGA beads were incubated at 37 degrees C with gamma-32P-ATP to examine the phospholipid labeling patterns. The control stroma incorporated 32P-label into phosphatidylinositol-4-phosphate and phosphatidylinositol-4,5-bisphosphate, in agreement with earlier studies. However, the stroma-WGA beads showed incorporation of 32P-label into phosphatidic acid in addition to that in the phosphoinositides. The quantity of 32P-phosphatidic acid produced during the 20-min assay was 3.23 +/- 0.84 (n = 7) picomoles/micrograms stromal cholesterol; the amount synthesized, however, was dependent on the procedure used to prepare the stroma-WGA beads. If the erythrocytes were bound to the WGA beads at 0 degrees C instead of 20 degrees C, the quantity of 32P-phosphatidic acid produced during the subsequent 37 degrees C assay with gamma-32P-ATP was decreased 4.2 fold; the phosphoinositide labeling pattern was unchanged. In addition, when the time for binding of intact erythrocytes to the WGA beads was varied from 1 to 20 minutes, there was a time-dependent increase in the amount of 32P-phosphatidic acid produced. This induction of phosphatidic acid synthesis could not be duplicated with fluid phase WGA. Therefore, the multivalent binding of intact erythrocytes to WGA beads causes an alteration in phospholipid metabolism.

Endogenous calcium in sickle cells does not activate polyphosphoinositide phospholipase C

Sickle-cell-anaemia erythrocytes (SS cells) are known to have a high Ca2+ content (particularly the dense cell fraction) and to take up Ca2+ on deoxygenation. It has been reported that this high Ca2+ was responsible for the activation of the Ca2+-dependent K+ loss, and of the Ca2+-sensitive polyphosphoinositide phospholipase C (PIC) in dense SS cells. We found that, either in the total population of SS cells or in the light or dense fractions, the content of phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P2] was not changed, whereas that of phosphatidylinositol 4-phosphate was increased and that of phosphatidic acid (PtdOH) was decreased compared with normal (AA) erythrocytes. Deoxygenation-induced Ca2+ entry into SS cells did not change the concentration or, in 32P-prelabelled cells, the radioactivity of polyphosphoinositides and PtdOH. It also failed to induce the formation of inositol 1,4,5-trisphosphate, the product of PtdIns(4,5)P2 hydrolysis by PIC, which was measured by an original method using ion-pair reverse-phase h.p.l.c. Thus there was no evidence of an endogenous Ca2+ effect on the PIC activity in SS cells, in agreement with the demonstration that the excess Ca2+ in SS cells is compartmentalized into internal vesicles and unavailable as free Ca2+. The 32P incorporation in polyphosphoinositides and PtdOH was markedly higher in SS than in AA cells, but this increase was the same in both dense and light SS cells. The increase in the turnover of these phospholipids in SS cells is consistent either with an activation of the lipid kinases and phosphatases or with perturbation in the metabolic compartmentation of these lipids.

Inositol lipids: receptor-stimulated hydrolysis and cellular lipid pools

Our current knowledge of the process by which receptors stimulate the hydrolysis of phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) has its origin in the discovery by Hokin & Hokin (J. biol. Chem. 263, 967 (1953] that some pancreatic secretagogues not only elicit exocrine secretion but also stimulate the metabolism of membrane phospholipids. Despite the recent elucidation of many aspects of this widespread signalling system, there is still little information on the control of the supply of its substrate, PtdIns(4,5)P2. In particular, some studies have suggested that inositol-lipid-mediated signalling involves much or all of the inositol lipid complement of the stimulated cells, whereas other observations have equally clearly implicated the receptor-activated hydrolysis of an inositol phospholipid pool that comprises only a small fraction of the total cellular complement of these lipids. These studies, which have largely employed radiochemical analyses using single isotopes, are briefly reviewed. In addition, we report the first information obtained by a new procedure for analysing the metabolic characteristics of the inositol lipids that are broken down during stimulation. This technique employs cells that are doubly labelled in the inositol moiety of their lipids (to isotopic equilibrium with 14C and only briefly with 3H) to search for functional metabolic heterogeneity among the inositol lipids of stimulated cells. Using this method, we have found that the inositol phosphates liberated in stimulated cells during brief stimulation of V1a-vasopressin receptors or prostaglandin F2 alpha receptors come from phospholipid that has a turnover rate typical of the bulk of the cellular inositol lipids.

Rates of production and consumption of phosphatidic acid upon thrombin stimulation of human platelets

Human platelets were labelled with [32P]Pi and [3H]glycerol before gel filtration. In unstimulated cells, the specific 32P radioactivity in phosphatidic acid (PtdOH) was similar to that of phosphatidylinositol (PtdIns) but only 4% of that of the gamma-phosphate of ATP. Upon 3 min of stimulation with 0.5 U/ml of thrombin, there was a 20-fold increase in specific 32P radioactivity of PtdOH which approached that of the ATP gamma-phosphate. Based on constant rates of synthesis and removal, this thrombin-induced increase in specific 32P radioactivity in PtdOH allowed us to calculate the flux of phosphate through PtdOH upon stimulation. Synthesis and removal occurred at rates of 107 and 52 nmol min-1/10(11) cells, respectively. The specific [3H]glycerol radioactivity was similar in PtdIns, phosphatidylinositol 4-phosphate and phosphatidylinositol 4,5-bisphosphate in unstimulated platelets. In PtdOH, it was 50% of that of the inositol phospholipids. Thrombin stimulation induced no changes in the specific 3H radioactivity of the inositol phospholipids whereas specific [3H]PtdOH increased to the level of these lipids. It is concluded that PtdIns, PtdInsP and PtdInsP2 exist in a metabolic homogenous pool in human platelets.

Phorbol ester inhibits phosphatidylserine synthesis in human promyelocytic leukaemia HL60 cells. Possible involvement of free radicals and correlation with phosphorylation of nuclear protein 1b

Treatment of human promyelocytic leukaemia HL60 cells in conditioned medium with 12-O-tetradecanoylphorbol 13-acetate (TPA) for 4 h resulted in 25-30% inhibition of labelling of phosphatidylserine (PS) with [U-14C]serine. PS labelling was 40% lower, and no inhibitory TPA effect was observed when the experiments were performed in fresh medium. Cycloheximide or puromycin also inhibited PS labelling by 38-44%; their inhibitory effects were non-additive with that of TPA and occurred only in conditioned medium. Catalase (CAT) and superoxide dismutase (SOD), both free-radical scavengers, and H7, a protein kinase C inhibitor, reversed to various extents the inhibitory effect of TPA on PS synthesis. On the other hand, chlorobenzoic acid, a free-radical-generating agent, also inhibited PS synthesis by 22% after 4 h treatment when conditioned medium was used. When ethanolamine was added to cells in conditioned medium to quench PS formation through the exchange of free serine with the ethanolamine moiety of phosphatidylethanolamine (PE), PS labelling was decreased by 33% and the inhibitory TPA effect was significantly decreased. On the other hand, ethanolamine had marginal quenching effect on PS labelling when added to cells in fresh medium. TPA increased the phosphorylation of various proteins in the cells, including protein lb (Mr 80,000; pI 5.5) shown to be localized mainly in the nuclear fraction. Chlorobenzoic acid selectively stimulated the phosphorylation of protein lb, whereas CAT and SOD specifically attenuated the TPA-stimulated phosphorylation of this protein. All these agents affected phosphorylation of protein lb only if conditioned medium was used. The findings suggested that net synthesis of PS through the base-exchange mechanism was stimulated in HL60 cells by cell products present in the conditioned medium. TPA inhibited this stimulated PS synthesis by a mechanism which appeared to involve active oxygen species and protein synthesis and might be related to the phosphorylation of protein lb.

The pathway of myo-inositol 1,3,4-trisphosphate dephosphorylation in liver

We studied the dephosphorylation pathway for Ins(1,3,4)P3 (inositol 1,3,4-trisphosphate) by liver homogenates and soluble and particulate subfractions incubated in media resembling physiological ionic strength and pH. Ins(1,3,4)P3 was dephosphorylated to two InsP2 (inositol bisphosphate) isomers, one of which is Ins(3,4)P2 [Shears, Parry, Tang, Irvine, Michell & Kirk (1987) Biochem. J. 246, 139-147]. The second InsP2 is the 1,3 isomer. Ins(3,4)P2 is dephosphorylated to inositol 3-phosphate by an enzyme activity located in both soluble and particulate fractions. The phosphatase(s) that attacks Ins(1,3)P2 was largely soluble, but we have not determined which phosphate(s) is removed. When the initial substrate concentration was 1 nM, the rate of dephosphorylation of Ins(1,4)P2 greater than Ins(1,3)P2 greater than Ins(3,4)P2. None of these bisphosphates was phosphorylated when incubated with liver homogenates and 5 mM-ATP, but their rates of dephosphorylation were then decreased.