Evidence for a specific phosphatidylinositol 4- phosphate phosphatase in human erythrocyte membranes

Transcription

1 Evidene for a speifi phosphatidylinositol 4- phosphate phosphatase in human erythroyte membranes S. E. Mak and F. B. St. C. Palmer Department of Biohemistry, Dalhousie University, Halifax, Nova Sotia, Canada B3H 4H7 Abstrat Human erythroyte membranes exhibit a speifi phosphatidylinositol 4-phosphate phosphohydrolase (PtdIns4P phosphatase) ativity whih hydrolyzes PtdIns4P and lysoptdins4p but not phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) or lysoptdins(4,5)p2. Phosphatidi aid, lysophosphatidi aid, glyerophosphoinositol 4-phosphate, glyerophosphoinositol4,5-bisphosphate, inositol mono, bis, and tris phosphates and several other sugar and nuleoside phosphates are not hydrolyzed. The PtdIns4P phosphatase ativity is not affeted by Ca2+ or Mg+ ions nor inhibited by EDTA. Maximum in vitro ativity requires non-ioni (Triton X-1 00) detergents. The phosphatase is very stable in isolated membranes at low temperatures but is rapidly inativated above 35'C. This ritial inativation temperature is lowered to 20-25'C by solubilizing the membranes with non-ioni detergents. Arrhenius plots of the ativity show an infletion at these ritial temperatures, suggesting a temperaturedependent hange in the environment or onformation of the enzyme. Sulfhydryl-reating reagents are potent inhibitors. Dithioerythritol stimulates only when the membranes are solubilized with non-ioni detergent. The loation of ation-independent PtdIns4P phosphatase ativity in the membrane and of Mg2+-dependent PtdIns(4,5)P2 phosphatase ativity in the ytosol was also observed for monkey, rabbit, rat, and dog erythroytes. Both ativities are loated in the ytosol of sheep erythroytes.-mak, S. E., and F. B. St. C. Palmer. Evidene for a speifi phosphatidylinositol 4-phosphate phosphatase in human erythroyte membranes. J. Lipid Res : Supplementary key words phospholipids mammalian erythroytes erythroyte ytosol lipid enzymology membrane solubilization non-ioni detergents thermal inativation phosphatidylinositol 4,5- bisphosphate phosphatase phospholipid phosphatase ation requirements The monoesterified phosphate groups of phosphatidylinositol 4-phosphate (PtdIns4P) and phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P2) exhibit a very rapid metaboli turnover in all tissues and miroorganisms that have been studied (1). This ours via an ATP-dependent phosphorylation-dephosphorylation yle. Strutural analysis of the naturally ourring polyphosphoinositides and of the PtdIns4P produed during both synthesis and dephosphorylation of PtdIns(4,5)P2 by ex- trats of bovine brain and protozoa indiates that phosphatidylinositol (PtdIns) is phosphorylated first in the 4 then the 5 position and that PtdIns(4,5)P2 is sequentially dephosphorylated in the reverse order (2-4). The phosphorylation steps are atalyzed by speifi PtdIns and PtdIns4P kinases (1). Most studies have onluded that a single phosphatase removes both phosphates from Ptdlns(4,5)P2. Crude and partially purified preparations from brain, kidney and iris musle onvert PtdIns(4,5)P to PtdIns (5-9). Moreover, PtdIns(4,5)P2 and PtdIns4P appear to ompete for the same phosphatase in kidney extrats (10). Finally, a homogeneous preparation of PtdIns(4,5)P2 phosphatase from rat brain (11) hydrolyzes PtdIns4P as well as PtdIns(4,5)P2. However, the possibility of separate PtdIns(4,5)P2 and PtdIns4P phosphatases must still be onsidered sine highly speifi PtdIns(4,5)P2 phosphatases have now been partially purified from both protozoa and human erythroytes (12, 13). Neither enzyme hydrolyzes PtdIns4P in vitro. There is a seond ylial pathway whih, in some tissues, responds to reeptor-mediated influx of Ca" ions (1). The polyphosphoinositides are hydrolyzed by a alium-ativated phosphodiesterase to diaylglyerol and the orresponding inositol bis and trisphosphates. Replaement ours by de novo synthesis of PtdIns followed by phosphorylation. In erythroytes, the Ptdlns and PtdIns4P kinases are present on the ytosoli surfae of the membrane (14). With some exeptions, mammalian erythroyte membranes also ontain the alium-dependent phosphodi- esterase (15). A phosphatase in human erythroyte membranes onverts inositol 1,4,5-trisphosphate (16) to inositol Abbreviations: PtdIns, 1-(3-sn-phosphatidyl)-~-myo-inositol; Ptd- Ins4P, 1-(3-sn-phosphatidyl)-~-myo-inositol 4-phosphate; GroPIns4P, glyerophosphoryl-~myo-inositol 4-phosphate; PtdIns(4,5)P2, 1-(3-mphosphatidyl)-~-myo-inositol 4,5-bisphosphate; GroPlns(4,5)Pt, glyerophosphoryl-d-myo-inositol 4,5-bisphosphate; Pipes, piperazine- N,N'-bis(2-ethanesulfoni aid); SDS, sodium dodeyl sulfate; CTAB, etyltrimethylammonium bromide; EGTA, ethyleneglyol-bis-(fi-aminoethyl ether)n,n'-tetraaeti aid. Journal of Lipid Researh Volume 25,

2 1,4-bisphosphate whih is the end produt of both PtdIns(4,5)P2 and PtdIns4P degradation by this route. In some speies, the diaylglyerol may be phosphorylated to give phosphatidi aid, but erythroytes lak the enzymes needed for PtdIns resynthesis. This inability to replae lost phosphoinositide, and an intraellular alium onentration too low to support diesterase ativity, make it unlikely that the diesterase-initiated ataboli pathway is physiologially signifiant in healthy erythroytes (1 7). Therefore, only the phosphorylation-dephosphorylation yle an aount for the rapid labeling of the monoesterified phosphate groups of PtdIns4P and PtdIns(4,5)P2 in erythroytes (1 8). An undoumented report suggests that swine erythroytes an hydrolyze exogenous PtdIns4P (1 9). However, of the two phosphatase ativities required for this yle, only the human erythroyte ytosoli PtdIns-(4,5)P-speifi phosphatase has been demonstrated diretly (1 3). The present report desribes the harateristis of a PtdIns4P-speifi phosphatase ativity in human erythroyte membranes and provides evidene of a similar ativity in erythroytes from several other mammalian speies. MATERIALS AND METHODS Lipids and substrates PtdIns(4,5)P2 ontaining less than 5% PtdIns4P was prepared from bovine brain (20) and purified on immobilized neomyin (2 1). Most PtdIns4P was prepared from the rude PtdIns(4,5)P2 using the partially purified PtdIns(4,5)P2 phosphatase from Crithidia fasiulata (4). Some naturally ourring PtdIns4P was isolated from extrats of bovine brain by neomyin hromatography (2 1). Rat liver PtdIns was also isolated by hromatography on immobilized neomyin (21). Other lipids were from Serdary Researh Laboratories, London, Canada. Lipids were stored in hloroform at -20 C. The extraellular PtdIns-speifi phosphodiesterase of Baillus thuringiensis (22) was used to prepare myoinositol 1,2-yli phosphate from the PtdIns. Myoinositol 1,4-bisphosphate and myoinositol 1,4,5-trisphosphate were prepared from the purified PtdIns4P and PtdIns(4,5)P2 using the polyphosphoinositide phosphodiesterase present in rude extrats of Crithidiafasiulatu (23). Diesterase reations were done in 25 mm Pipes (ph 7.2). This ontained 0.1% (w/v) sodium deoxyholate for the PtdIns diesterase, and was saturated with diethyl ether in the presene of 0.2 mm CaCI2 and 0.1 M NaCl for the polyphosphoinositide diesterase. Reations were stopped by adding an equal volume of 10% (w/v) trihloroaeti aid, whih preipitated the lipid, protein, and Pipes. The trihloroaeti aid was extrated from the supernatants with diethyl ether. LysoPtdIns4P and lysoptdins(4,5)p2 were prepared from the purified PtdIns4P and PtdIns(4,5)P2, respetively, using bee venom phospholipase A2 (Sigma Chemial Co.) in 50 mm triethanolamine at ph 8.5. After aidifying the reation mixture with a few drops of formi aid and adding 4 volumes of hloroform-methanol2: 1 (v/v), the lysolipids were reovered in the upper aqueous phase. Glyerophosphoinositol 4-phosphate (GroPIns4P) and glyerophosphoinositol4,5-bisphosphate (GroPIns(4,5)P2) were prepared by mild alkaline hydrolysis of PtdIns4P and PtdIns(4,5)P2 (24). Triton X-100 (p-tert-otylphenoxypolyoxyethylene), Tween 60 (polyoxyethylene sorbitan monostearate), and Tween 80 (polyoxyethylene sorbitan monooleate) were from the Sigma Chemial Co. Cutsum (isootylphenoxypolyoxyethylene) and Brij 35 (polyoxyethylene dodeyl ether) were from Fisher Sientifi Ltd., Montreal, P.Q. Ammonyx LO (dodeyldimethylamine oxide) and Miranol H2M (disodium 1 -arboxymethyl- 1-(2-arboxymethoxylethyl)-2-undeylimidazolinium hydroxide) were gifts of the Onyx Oil and Chemial Co., Jersey City, NJ and the Miranol Chemial Go., Irvington, NJ, respetively. Erythroyte preparations The erythroyte fration from human blood, no longer suitable for transfusion, was obtained from the Red Cross Transfusion Servie. It had been olleted in aid-itratedextrose and stored at 4OC for several days. Fresh blood from volunteers was drawn by venipunture into heparinized ontainers. Animal blood was olleted in heparinized ontainers by venipunture (large animals) or ardia punture (small animals). Erythroytes were sedimented at 750 g for 10 min and washed twie with ieold 150 mm NaCl ontaining 1 mm EDTA. The supernatant, buffy oat, and the top layer of erythroytes (approximately 20% of the erythroytes) were removed eah time. Contamination of these preparations was white ells and 7-15 platelets per lo4 erythroytes as determined in a Coulter Counter. The washed ells were hemolyzed in 20 volumes of 10 mm Tris (ph 7.4)- 1 mm EDTA. The membranes were sedimented at 17,500 g for 15 min, washed three times with the hemolysis buffer, and stored at 4OC. Cytosol proteins in the supernatant were obtained free from hemoglobin by repeated ammonium sulfate preipitation at 4OC (1 3). The preipitated ytosol proteins were suspended in 50 mm imidazole buffer (ph 7.0) ontaining 0.5 mm EGTA and dialyzed against the same buffer. The dialysate was entrifuged to remove any insoluble material and stored at 4OC. Analyses Total phosphorus was measured by the method of Bartlett (25) after digestion of the lipid samples in perhlori 76 Journal of Lipid Researh Volume 25, 1984

3 aid. Protein was measured by the Lowry proedure as modified by Peterson (26). Results were alulated using the hyperboli approximation from two standards (27). Phosphatase assays The assay proedures for PtdIns4P and PtdIns(4,5)P2 phosphatases were adapted from those desribed before (12, 13). All reation mixtures ontained 50 mm Pipes (ph 7.2), 1.0 mm EGTA, 1 mm substrate, and (w/v) Triton X-100 in a total volume of 0.15 ml. The requirements for CTAB and Mg2+ varied with the tissue and the substrate. Assays of erythroyte ytosol PtdIns(4,5)P2 phosphatase ontained 0.4 mm MgC12 and 2.5 mm CTAB (13). The erythroyte PtdIns4P phosphatase assays ontained 1.5 mm CTAB but no Mg2+. Reation mixtures were preinubated without enzyme at 37 C for 2 min and the enzymi reations were started by adding the tissue preparations (0.03 ml). In some later experiments, Triton X-100 was added to the tissue preparations at 0 C on an ie bath 10 min before their use in the assay. Reations were stopped after 5-30 min by adding 0.35 ml of 5% (w/v) SDS-50 mm EDTA and the inorgani phosphate measured by the automated method of Hegyvary, Kang, and Bandi (28). Ativity with other phospholipids and with nonlipid phosphate esters was measured in the same way. Nonspeifi phosphatase ativity was measurdd spetrophotometrially by observing the hydrolysis of p-nitrophenylphosphate at ph 4.8 in 25 mm itrate buffer, at ph 7.7 in 25 mm Tris/hloride, and at ph 10.5 in 50 mm glyine (29). Hydrolysis of PtdIns4P RESULTS Preliminary experiments indiated that the ability to release inorgani phosphate from purified PtdIns4P was loalized in the membrane fration of human erythroytes. Washed, hemoglobin-free membranes were used to determine optimal assay onditions. Without nonioni detergent, erythroyte membrane PtdIns4P phosphatase ativity was low (Fig. 1) but ould be inreased almost 3- fold by Triton X-100. Maximum ativity was obtained at detergent onentrations of (w/v)or higher. Cationi detergent (CTAB) had little effet at low CTAB/ PtdIns4P ratios but inhibited the reation at ratios of 2 or greater. Ativity was marginally inreased at CTAB/ PtdIns4P ratios of 1.O Therefore CTAB (1.5 mm) was inluded in the standard assay. The assay measured release of inorgani phosphate from PtdIns4P. Thin-layer hromatography (23) of the reation mixture onfirmed that PtdIns was the only other produt. The reation was presumed to be enzymi sine PtdIns4P was not hydrolyzed by membranes that had - 'Z 30 0 L a m E T 20.- E - 0 E - 10 x u CTAB I Ptdlns 4P Molar Ratio Fig. 1. Dependene of human erythroyte membrane PtdIns4P phosphatase on the CTAB/PtdIns4P molar ratio. The standard assay proedure was used with (0) and without (0) (w/v) Triton X- 100 in the reation mixture whih ontained 50 mm Pipes (ph 7.2), 1 mm PtdIns4P, CTAB (as indiated), and 0.63 mg. ml-' membrane protein. been heated to 100 C for 10 min or exposed to trypsin. The ativity was optimal over a broad ph range from 6.5 to 8 and at substrate onentrations above 0.5 mm. The analytial proedure was not sensitive enough to estimate initial reation rates at very low substrate onentrations but substrate saturation urves suggested a K,,, of less than 0.1 mm. Substrate inhibition was not evident at up to 2 mm PtdIns4P. With 1.0 mm PtdIns4P and 1.5 mm CTAB, the release of inorgani phosphate was linear for 5-10 min at 37 C and suffiient inorgani phosphate was released for aurate measurement. A variety of detergents was assessed for the ability to stimulate PtdIns4P phosphatase ativity at the optimum CTAB/PtdIns4P ratio of 1.5 (Table 1.). Cutsum is struturally very similar to Triton X-100 and gave omparable stimulation. Further inreases in ativity ourred at onentrations greater than (w/v), but sensitivity was redued by a dramati inrease in the release of inorgani phosphate in the enzyme ontrols (substrate added after the inubation). The effet was independent of PtdIns4P and CTAB and proportional to the quantity of membranes present. It must represent a detergentdependent release of phosphate from some membrane omponent, the nature of whih was not investigated further. Otylgluoside stimulated PtdIns4P phosphate ativity, but at onentrations above 15 mm it also aused a release of substantial amounts of phosphate from the membranes. For this reason neither Cutsum nor otylgluoside was deemed suitable and no other detergent Mnk mid Palmer Phosphatidylinositol 4-phosphate phosphatase in erythroyte membranes 77

4 TABLE 1. None Effet of detergens on human erythroyte Ptdlns4P phosphatase Detergent" Conentration Ativityb Triton X-I 00 (ontrol) Cutsum 0.05% 0.1% 0.4% 1.O% 2.0% 3.0% Otylgluoside 7.5 mm 15.0 mm 20.0 mm 22.0 mm 25.0 mm 30.0 mm Brg % 0.1% 0.4% 1.O% Miranol H2M Tween 60 Tween 80 Ammonyx LO Deoxyholate Detergents were substituted for Triton X-100 in the standard assay system. Ativities are expressed as nmol * min-' mg protein-'. was as effetive as Triton X-100. Sodium deoxyholate inhibited the reation. The erythroyte membrane PtdIns4P phosphatase was not stimulated by magnesium ions even though the erythroytes had been prepared in the presene of EDTA (Table 2.). Adding 10 mm EDTA to the reation mixture had no effet, nor did alium or lithium ions up to 2 mm. The membranes did exhibit a small Mg2+-dependent TABLE 2. Polyphosphoinositide phosphatases of human erythroyte membranesa Ativity PtdIns4P Ptdlns(4.5)Pz Addition Phosphatase Phosphatase None 23 MgC12 (0.5 mm) 25 EDTA (10 mm) 23 CaC12 (2 mm) ' Membranes were prepared from blood olleted in aid-itratedextrose and stored at 4OC for 3 days. PtdIns(4,5)P2 phosphatase ativity, usually less than 5% of the ativity with PtdIns4P. Some membrane preparations exhibited a very low EDTA-independent PtdIns(4,5)P2 phosphatase ativity whih did not exeed 2% of the PtdIns4P phosphatase value. The possibility that a signifiant proportion of the PtdIns4P phosphatase was ontributed by the small number of white blood ells and platelets in the erythroyte preparations was exluded by analyzing membranes from heavily ontaminated preparations. Both whole blood (20 white ells and 425 platelets per lo4 erythroytes) and washed blood ells from whih the buffy oat had not been removed (13 white ells and 230 plateiets per 1 O4 erythroytes) were hemolyzed. In spite of the fold greater ontamination with white ells and platelets, membranes reovered from these preparations had the same speifi ativity as membranes from the purified erythroytes. The speifi ativity of human erythroyte membrane preparations in the standard assay system was 19.4 k 3.7(10) nmol*min-'.mg protein-' (mean f SD for ten determinations). The ativity was quite stable; membrane suspensions stored at 4OC retained 60-70% of their initial ativity after 3 months. The variability of these measurements was a harateristi of the assay rather than the membrane preparations sine a similar SD was observed in repeated assays of the same sample of membranes (15.5 k 3.2 nmol min-' * mg-', eight determinations). Substrate speifiity Among potential lipid substrates (those having monoesterified phosphate groups), appreiable phosphatase ativity was observed only with PtdIns4P and IysoPtd- Ins4P (Table 3). Neither phosphatidi aid nor lyso- TABLE 3. Substrate speifiity of human erythroyte membrane PtdIns4P phosphatase Substrate Ptdlns4P (1.5 mm CTAB) LysoPtdlns4P (1.0 mm CTAB) GroPlns4P Inositol 1,4-bisphosphate Ativitya N D ~ ND Ptdlns(4,5)P2 (2.0 mm CTAB, 0.5 mm Mg2+) 2.7 LysoPtdIns(4,5)P2 (2.0 mm CTAB, 0.5 mm Mg2+) 1.5 GroPlns(4,5)Pz (0.5 mm Mg") 0.3 lnositol 1,4,5-trisphosphate (0.5 mm Mg2+) 0.1 Phosphatidi aid Lysophosphatidi aid ND ND,.. Standard was assay used with 1 mm substrate. Assays were done at several CTAB onentrations with and without 0.5 mm Mg2+. Maximum ativity was obtained at the onditions listed. a Ativity is expressed as nmol. min-'. mg protein-'. * ND is not detetable at 0.05 nmol. min-'. mg protein-'. 78 Journal of Lipid Researh Volume 25, 1984

5 phosphatidi aid was hydrolyzed under a variety of inubation onditions. The ativity with PtdIns(4,5)Pz and lysoptdins(4,5)p2 was Mg'+-dependent and was at least an order of magnitude smaller than the rate with PtdIns4P. Neither GroPIns4P nor inositol 1,4-bisphosphate was hydrolysed. The very low Mg2+dependent ativity with inositol 1,4,5-trisphosphate and GroPIns(4,5)Pz was most probably due to the inositol 1,4,5-trisphosphate phosphatase present in human erythroyte membranes (16). Other speifi or nonspeifi phosphatases might be responsible for the observed ativity. Alkaline phosphatases of both animal (bovine intestine) and baterial (E. oli) origin and aid phosphatase are able to hydrolyze PtdIns4P and PtdIns(4,5)Pz although the reation rate is orders of magnitude smaller than with p-nitrophenyl phosphate or glyerol phosphate' (5). The erythroyte membranes were therefore tested for ativity with a number of phosphate esters. Eah potential substrate was assayed with membranes alone, with Triton X-100, with 1.0 mm CTAB, and with both Triton X-100 and CTAB. In eah ase the assays were done with and without 0.5 mm MgCl2. There was no detetable ativity with gluose 6-phosphate, frutose 6-phosphate, glyerol 2- phosphate, glyeraldehyde 3-phosphate, 2,3-bisphosphoglyerate, phosphoglyolate, ATP, ADP, AMP, inositol 1,2-yli phosphate, and inositol 2-phosphate. The lower limit of detetion in these experiments was nmol - min-' mg protein-', about 1 % of the ativity with PtdIns4P. Frutose 1,6-bisphosphate was hydrolyzed very slowly by the membranes (0.5 nmol min-' mg protein-'). p-nitrophenyl phosphate was also slowly hydrolyzed, the highest rate being at aid ph. The observed ativities were 2.2, 0.7, and 0.2 nmol*min-'*mg protein-' at ph 4.8, 7.5, and 10.5, respetively. The hydrolysis of both frutose 1,Bbisphosphate and p-nitrophenylphosphate was not inhibited by EDTA. Inhibitors Preinubation of human erythroyte membranes at 37OC for 30 min with fluoride (5 mm), o-phenanthroline (5 mm), or dithioerythritol (5 mm) had no effet on the PtdIns4P phosphatase. The sulfhydryl reating reagents p-hloromeruribenzoate and p-hloromeruriphenylsulfonate were potent inhibitors. Inhibition was 70-80% at 0.1 mm and % at 0.5 mm. Total inhibition by N- ethylmaleimide (1 O mm) onfirmed the dependene on sulfhydryl groups. Enzyme stability Attempts to follow the reation to ompletion indiated that the extent to whih PtdIns4P was hydrolyzed de- ' Mak, S. E., and F. B. St. C. Palmer. Unpublished data pended upon the amount of enzyme present. At several membrane onentrations the deline in reation rate was a funtion of time rather than of substrate onsumed. This suggested that the enzyme was being inativated during the reation period although preinubation of the isolated membranes at 37OC for several hours only slightly redued the ativity. In developing the standard assay, the best ativity had been obtained when the reations were started by adding the membrane suspensions to the mixture of buffer, Triton X-100, substrate, and CTAB. Triton X-100 was found to have several additional effets. Although enough detergent was present in the assay mixture to solubilize the added membranes, prior brief exposure of the membranes to Triton X-100 (2-10 min at 0 C in an ie bath) further inreased the PtdIns4P phosphatase ativity as measured in the standard assay (Table 4). The effet was maximal at Triton, the onentration at whih the membranes (3-6 mg protein ml-') were ompletely solubilized. Pretreatment alone was insuffiient for maximum ativity; it was still neessary to have detergent in the assay mixture. The inreased PtdIns4P phosphatase ativity was not EDTA (5 mm)-sensitive. The low level of PtdIns(4,5)P2 phosphatase was unhanged in these pretreated membranes. Similar pretreatment of the ytosol fration had no effet on the soluble PtdIns(4,5)P2 phosphatase ativity. Pretreatment stimulated membrane PtdIns4P phosphatase ativity from 2- to 4-fold. The greatest stimulation ourred with those membranes having the lowest unstimulated ativity. The mean stimulated ativity was 35.7 & 6.7(4) nmol - min-' mg protein-'. Unstimulated ativities with membranes isolated from fresh heparinized blood (1 3.8 & 5.3(7) nmol - min-' mg protein-') were lower than those with membranes from stored blood. However, Triton pretreatment inreased TABLE 4. Effet of Triton X-100 pretreatment of human ervthrovte membranes on the PtdIns4P DhosDhatase Triton X-100 Pretreatment Assay PtdInslP Phosphatase % nmol. min-'. mg prolein-' The protein onentration of the membranes was 5.25 mg. ml-' during the pretreatment period at ODC (16 hr) and 0.53 mg. ml-' in the 10 min assay at 37 C. Mak and Palmer Phosphatidylinositol 4phoephate phosphatme in erythroyte membranes 79

6 Duration of Exposure to Triton X-100 ( hr ) Fig. 2. Effet of temperature on the PtdIns4P phosphatase of Triton X-1 00 solubilized human erythroyte membranes. Membranes were solubilized in (w/v) Triton X-100 at OOC. The protein onentration was 4.2 mgvml-'. After 10 min, aliquots of the membrane solution were transferred to water baths at 20 C (A), 25 C (A), 3OoC (O), and 37 C (W). A ontrol aliquot was maintained at 0 C (0). PtdIns4P phosphatase ativity was determined at various times using the standard assay mixture with a IO-min inubation at 37 C. The onentration of membrane protein in the assays was 0.85 mg-ml-'. The relative ativity is expressed as a perent of the ativity after the initial treatment with detergent at 0 C for 10 min. this ativity to 36.8 f 6.6(5), the same as the stimulated value for membranes from blood whih had been olleted and stored in aid-itrate-dextrose. The ativities were the same using either the enzymially produed PtdIns4P or the naturally ourring PtdIns4P. No other detergent was as effetive for pretreatment as Triton X-100. The membranes were readily solubilized by most detergents at the onentrations used and there was no detergent-dependent release of inorgani phosphate from the membranes at 0 C. The solubilization detergent was diluted 5-fold when the membranes were added to the assay mixture ontaining Triton X Both Cutsum and otylgluoside were therefore not present in the assay mixtures at onentrations required to ause the release of inorgani phosphate from the membranes at 37 C. Cutsum stimulated PtdIns4P phosphatase ativity slightly at 0.5% (w/v) but higher onentrations inhibited. All the other detergents listed in Table 1 were ineffetive. Prior solubilization of the membranes rendered the PtdIns4P phosphatase very temperature-sensitive (Fig. 2). The ativity was not affeted at 0 C for up to 16 hr, and there was little loss of ativity over several hours at up to 20 C. At higher temperatures, the loss of ativity was inreasingly rapid. The data suggest that signifiant loss would our during the 10- or 15-min inubation at 37 C in the presene of detergents in the standard assay. Human erythroyte membranes ontain several protease ativities, some of whih are stimulated by detergents (30, 3 1). Pretreatment under onditions known to inhibit these protease ativities ( 10 mm phenylmethylsulfonylfluoride, 20 mm EDTA) neither redued the ontrol ativity (pretreatment at 0 C) nor prevented the loss of ativity that ourred after pretreatment at 37 C for 1 hr. The same result was obtained with several other protease inhibitors (10 mm aprotinin, 10 mm benzamidine, 0.5 mg/ml pepstatin). The PtdIns4P phosphatase ativity of pretreated membranes was abolished by exposure to 5 mm HgC12, another erythroyte membrane protease inhibitor. Although effetive inhibition of the proteases in suh a omplex mixture annot be assured, inativation of the PtdIns4P phosphatase in the presene of Triton X-1 00 seems more likely due to heat denaturation than to proteolysis. Detergent-indued temperature sensitivity was further studied using a 1-hr preinubation of membranes with one of several detergents. Fig. 3 relates enzymi ativity to preinubation temperature. Exept with deoxyholate, the ativity was stable up to a ritial temperature, above e - )r.- '2 80 I u a W m!z CL YI 60 0 a 40 W.- > : 20 E Temperature of Detergent Treatment ( OC ) Fig. 3. Effet of detergents on the temperature sensitivity of human erythroyte membrane PtdIns4P phosphatase. Membrane suspensions were mixed at O C for 10 min with detergent: (w/v) Triton X- 100 (O), 25 mm otylgluoside (A), (w/v) Ammonyx LO (W), or 5 mm deoxyholate (A). Detergent was omitted from the ontrol (0). Aliquots of eah treated suspension were inubated for 1 hr at several temperatures as indiated. The membrane protein onentration during detergent treatment was mg m1-l exept for samples treated with deoxyholate where it was 3.5 ml* ml-'. PtdIns4P phosphatase ativity was measured using the standard assay proedure with a IO-min inubation period at 37 C and mgem1-l of membrane protein. Relative ativity was alulated as desribed for Fig Journal of Lipid Researh Volume 25, 1984

7 whih it beame very unstable. For untreated membranes that temperature was about 35 C. Solubilization of the membranes in Triton X-100 lowered it to 20 C. Similar shifts in temperature sensitivity were observed with Ammonyx LO (25 C) and otylgluoside (ISOC), detergents that did not stimulate the PtdIns4P phosphatase ativity. The data suggest that some abrupt hange in the environment or onformation of the enzyme ours at the ritial temperature, most likely a fluidity or phase hange in the membranes or the mixed mielles of membrane protein, lipid, and detergent. For experiments desribed in Fig. 3, the membrane-detergent mixtures ontained approximately equal masses of detergent, membrane lipid, and membrane protein. With deoxyholate ( w/v) the membranes were only partially solubilized; nevertheless the PtdIns4P phosphatase was very unstable with some inativation ourring even at 0 C. The shape of the urve was different from those observed with either untreated membranes or membranes solubilized with the other detergents, and suggests that the deoxyholate-indued temperature sensitivity results from a progressive rather than an abrupt hange. An infletion in the Arrhenius plot, both in the presene and absene of Triton X-100, oinided with the temperature at whih the enzyme beame temperature labile (Fig. 4). Inativation of the enzyme during the assay ould ontribute to the lowering of the ativity observed above the ritial temperature. To minimize this effet the inubation times were redued. Assays in the Triton X solubilized system were shortened from 15 min at 0-20 C, to 5 min at 25" and 30"C, and to 2 min at 35" and 40 C. From the data in Fig. 2 it an be determined that there would be less than 10% inativation ourring under these onditions. Sine this is not suffiient to aount for the observed infletion, the data suggest that a temperature-dependent transition in the detergent-membrane mixed mielles at 20 C alters the ativation energy as well as inreases the sensitivity of the enzyme to inativation. A similar transition ours in the Triton X-1 00-free membranes, although at a higher temperature, produing a similar hange in the ativation energy and inreased suseptibility to inativation above the ritial temperature. Effet of dithioerythritol Inhibition by sulfhydryl-reating reagents suggests that one or more sulfhydryl group(s) are neessary for enzymi ativity. Dithioerythritol had no effet when added to the reation mixture in the assay of unsolubilized membranes (Table 5). It did retard the slow loss of ativity during prolonged preinubation of membranes at 0 C. It also stimulated the ativity of Triton-solubilized membranes when added either during the initial pretreatment period at 0 C or later in the assay. This result suggests Temperature ( OC ) I I I I ,0001T ( OK ) Fig. 4. Arrhenius plot of the temperature dependene of the human erythroyte Ptdlns4P phosphatase. Triton X-1 00-pretreated membranes were assayed in the standard assay mixture ontaining Triton X-100 (0). Untreated membranes were assayed in the absene (0) of Triton X-100 in the assay mixture. Inubation times were shorter at the higher temperatures (see text). that the initial solubilization in pure detergent permits dithioerythritol aess to the enzyme that annot be ahieved when membranes are solubilized by the detergent-ctab-substrate mixture in the ourse of the assay. None None TABLE 5. Effet of dithioerythritol and Triton X-100 pretreatment on human erythroyte membrane PtdIns4P phosphatase Dithioerythritol (5 mm) Pre- Pretreatment Conditions treatment Assay Ativity 30 min, 37OC 30 min, 37 C 10 min, O'C, Triton X min, O'C, Triton X min, O'C, Triton X min, 37OC. Triton X min, 37"C, Triton X min, 37"C, Triton X t t 7.9 PtdIns4P phosphatase ativity was measured in the standard assay mixture ontaining Triton X-100 (IO-min inubation at 37 C). Mak and Palmer Phosphatidylinositol 4-phosphate phosphatase in erythroyte membranes 81

8 Dithioerythritol did not, however, prevent the inativation of the Triton-solubilized enzyme at 37 C. This further suggests that inativation is the result of thermal denaturation rather than, for instane, oxidation of an essential sulfhydryl group. Phosphoinositide phosphatases in other mammalian erythroytes The results of a survey of PtdIns4P and PtdIns(4,5)P2 phosphatase ativities in the ytosol and membrane frations of erythroytes from several mammalian speies are given in Table 6. Sheep erythroytes were atypial, having both PtdIns4P and PtdIns(4,5)P2 phosphatase ativities loated in the ytosol fration. However, two different enzymes were present sine the PtdIns(4,5)P2 phosphatase but not the PtdIns4P phosphatase was inhibited by EDTA. Preferential loation of an EDTA-insensitive PtdIns4P phosphatase in the membranes and of a Mg2+-dependent PtdIns(4,5)P2 phosphatase in the ytosol was observed for all other speies. Separation of ativities was learest in human erythroytes. Membranes from rabbit, rat, and dog erythroytes all had more PtdIns(4,5)P2 phosphatase ativity than human erythroyte membranes (1 0-25% of the ativity with PtdIns4P), but sine this ativity was inhibited by EDTA it ould not be attributed to the PtdIns4P phosphatase. The differene may be due to the proedure for isolating and washing the erythroyte membranes, whih had been optimized for human erythroytes. Membranes prepared from all animal erythroytes exept sheep were pink, indiating inomplete removal of hemoglobin. PtdIns(4,5)P2 phosphatase may also have been less effetively removed from these membranes. There was no similarity in the relative speifi ativities TABLE 6. of the two enzymes among the different speies. Human erythroytes had the highest speifi ativity for the membrane PtdIns4P phosphatase and a low ativity for the ytosol PtdIns(4,5)P2 phosphatase, while the onverse was true for erythroytes from the other speies. Hemoglobin ontamination may ontribute to the lower speifi ativity of PtdIns4P phosphatase in nonhuman erythroyte membranes. DISCUSSION We have previously reported on the PtdIns(4,5)P2 phosphatases of human erythroytes and protozoa (1 2, 13). Both are highly speifi and do not hydrolyze PtdIns4P under a variety of assay onditions. Whether or not these enzymes an degrade PtdIns4P in vivo, the rapid metaboli turnover of both monoesterified phosphates of the polyphosphoinositides requires suh an ativity. Previously reported PtdIns4P phosphatase ativity was assoiated either with rude preparations (6, 7, 9, 10) or purified preparations isolated for their PtdIns(4,5)P2 phosphatase ativity (5, 8, 1 l), whih ompletely degraded PtdIns(4,5)P2 to PtdIns with the transient appearane of PtdIns4P. Few studies have used purified PtdIns4P and PtdIns(4,5)P2 to assess both ativities. Phosphomonoesterase ativity of rabbit iris smooth musle homogenates hydrolyzes PtdIns4P at SO% of the rate for PtdIns(4,5)P2 (9). The highly purified polyphosphoinositide phosphatase from rat brain ytosol is equally ative with PtdIns4P and PtdIns(4,5)P2 (1 l), but its speifiity with respet to other possible substrates remains to be established. PtdIns4P in hromaffin vesile membranes Polyphosphoinositide phosphatases of other mammalian erythroytes PtdIns(4,5)P~ PtdIns4P Phosphatase Phosphatase EDTA Speies (n) (I 0 mm) Cytosol Membrane Cytosol Membrane Monkey" (1) o Rabbit (3) f Rat (3) f 0.1 ~ Dog (3) 1.7 f Sheep (3) f f f f f f f f f f f f f f f f f f k f f f 0.1 Eah phosphatase was assayed under optimum onditions as determined with human erythroyte preparations. All preparations were from blood freshly olleted in heparin. Ativity is expressed as nmol * min-' mg protein-' f standard deviation for n determinations when n is 3 or more. I' Mnriri forrul(iri$. 82 Journal of Lipid Researh Volume 25, 1984

9 is hydrolyzed by both phosphatase and phosphodiesterase ativities in frations of rat adrenal medulla homogenates (32). The vesile membranes do not ontain PtdIns(4,5)P2 and the ability of these preparations to hydrolyze PtdIns(4,5)P2 was not determined. The apparent similarity in the harateristis and subellular distribution of PtdIns4P and PtdIns(4,5)P2 phosphatase ativities of several tissues (6,lO) led to the idea that one phosphatase removes both monoester phosphates from PtdIns(4,5)P2. However, the EDTA-insensitive PtdIns4P phosphatase ativity of human erythroyte membranes reported here is learly distint from the PtdIns(4,5)P2 phosphatase desribed earlier (1 3). That enzyme is loalized in the ytosol and exhibits an absolute requirement for Mg2+ ions. The differential loation and ation requirement of the two phosphatase ativities was onfirmed in erythroytes from several other mammalian speies. Human erythroyte membranes have phosphatidi aid phosphatase ativity (33). Lak of phosphatidi aid phosphatase ativity in our preparations is onsistent with the reported low speifi ativity (<1 nmol min-' * mg protein-') whih is at the limit of detetion in our assay system. Furthermore, the Mg2+-dependene of this enzyme exludes it as a soure of PtdIns4P phosphatase ativity. The absene of ativity with 2,3-bisphosphoglyerate or phosphoglyolate (34, 35) argues against the possible ontamination with ytosol phosphatases, whih might otherwise be implied by the low but onsistent PtdIns(4,5)P2 phosphatase ativity present in the membranes. Lak of PtdIns4P phosphatase ativity in the ytosol also exludes this enzyme as the soure of the ativity. Membrane-bound nuleotidases and both speifi and nonspeifi phosphatase were also exluded by negative results with a variety of possible substrates. The erythroyte membrane PtdIns4P phosphatase is quite speifi. Only PtdIns4P and lysoptdins4p, but not PtdIns(4,5)P2 or phosphate monoesters derived from it, were hydrolyzed in the absene of divalent ations. Lak of ativity with either glyerophosphoinositol4-phosphate or myoinositol 1,4-bisphosphate indiates that this phosphatase requires a lipid substrate. It is also onsistent with the reported inability of the membrane inositol trisphosphate phosphatase to hydrolyze these ompounds (1 6). Enzymes ating on polyphosphoinositides are frequently assayed in omplex reation mixtures and there is onsiderable variety in the reported requirements of PtdIns(4,5)P2 phosphatases. In general, maximum ativity is obtained when requirements for Mg2+, harge redution of the substrate, and detergent solubilization are all met. Mg2+-dependene has been shown for partially purified PtdIns(4,5)Pp phosphatase from bovine brain, human erythroyte ytosol, and protozoa (5, 12, 13) but not for any other system studied, inluding the highly purified enzyme from rat brain (1 1). Charge redution is most readily ahieved by forming mixed mielles of the substrate and a ationi detergent suh as CTAB, although inreasing the divalent ation onentration an ahieve similar results (5). Prior studies of PtdIns4P hydrolysis have simply substituted PtdIns4P for PtdIns(4,5)P2 in assay systems optimized for PtdIns(4,5)P2. There is some evidene that PtdIns4P hydrolysis requires Mg2+ in kidney homogenates (10). Ativity was stimulated by Mg2+ and inhibited by EDTA but there was a large residual EDTA-resistant ativity. Comparison of PtdIns4P and PtdIns(4,5)P2 hydrolysis by the purified rat brain PtdIns(4,5)P2 phosphatase was limited to determining saturation urves in the absene of detergent but at the high Mg2+ onentration that is optimal for PtdIns(4,5)P2 (1 1). Mg2+ was not required for the erythroyte PtdIns4P phosphatase, the ativity of whih was not affeted by either Mg2+ or EDTA. Our reation onditions were optimized for PtdIns4P phosphatase ativity and were different from the optimum PtdIns(4,5)P2 phosphatase onditions. PtdIns4P is negatively harged and might be expeted to require some harge moderation for maximum enzymi ativity, although less than the more highly harged PtdIns(4,5)P2. However, there is no evidene of suh a requirement. CTAB had no effet on this ativity in kidney homogenates when premixed with PtdIns4P and presented as mixed mielles (1 0). Our experiene with erythroyte membranes was similar but high CTAB onentrations, as are used in PtdIns(4,5)Pz phosphatase assays, were inhibitory. The effet of non-ioni detergents is omplex. With soluble enzymes suh as the PtdIns(4,5)P2 phosphatases of C. fasiulata and human erythroyte ytosol, the major advantage is redued dependene on ationi ompounds with little or no effet on maximum ativity or the optimum CTAB/substrate ratio (1 2, 13). This suggests that dispersal of the substrate over the surfae of large mixed mielles redues the harge density enountered by the enzyme at the mielle surfae. The major effet on the membrane-bound erythroyte PtdIns4P phosphatase was to inrease the aessibility of the enzyme by solubilizing the membranes. This view is supported by inreased sensitivity to thermal inativation and dithioerythritol stimulation following solubilization of the membranes with Triton X-100. Furthermore, the higher ativity of untreated membranes with lysoptdins4p was omparable to that of solubilized membranes with PtdIns4P. This is probably due to the greater ability of lysolipids to penetrate and solubilize membranes. However, this may be an oversimplifiation, sine other detergents, whih dissolved the membranes, inreased the temperature sensitivity but not the enzymi ativity. This enzyme is learly Mak and Palmer Phosphatidylinositol Pphosphate phosphatase in erythroyte membranes 83

10 quite sensitive to its environment, a onlusion that is further supported by the distint transition in the Arrhenius plots of the ativity whih oinided with the onset of temperature sensitivity in both whole and solubilized membranes. Under the most favorable assay onditions, the PtdIns4P phosphatase ativity of human erythroyte membranes was nmol min-' mg protein-'. This is an order of magnitude greater than the inositol trisphosphate phosphatase ativity measured in vitro with exogenous substrate (16) and approximately two orders of magnitude greater than the polyphosphoinositide phosphodiesterase ativity measured with endogenous membrane-bound substrate (15). In view of the large effets of non-ioni detergents and of temperature on the PtdIns4P phosphatase in vitro, it is very likely that the environment of the enzyme in the membrane will greatly influene the in vivo ativity and ativities measured in the optimized assay may be muh greater than an our in vivo. Although no speifially identified PtdIns4P phosphatase ativity has been deteted with endogenous membrane-bound substrate either in vitro or in vivo, there is some indiret evidene of suh ativity. Inubation of human erythroytes with inorgani 32P labels only the monoesterified phosphate groups of phosphatidi aid, PtdIns4P and PtdIns(4,5)P2. Following lysis of the ells there is a ation-independent, EDTA-insensitive loss of both phosphatidi aid and PtdIns4P from the membrane (36). Ca2+-ativation of the polyphosphoinositide phosphodiesterase inreases PtdIns4P loss and initiates PtdIns(4,5)P2 loss. The ation-independent PtdIns4P phosphatase seems the most likely ause of the EDTAinsensitive omponent of PtdIns4P breakdown in these membranes. In more reent experiments (15) a small ation-independent release of inorgani phosphate from similarly labeled membranes was observed. Ativation of the polyphosphoinositide phosphodiesterase aused release of inositol 1,4-bisphosphate and inositol 1,4,5-trisphosphate from PtdIns4P and PtdIns(4,5)P2, but inorgani phosphate release was unhanged. The amount of radioativity released as inorgani phosphate was approximately the same as that in the inositol 1,4-bisphosphate. If PtdIns4P phosphatase ativity is truly the soure of the inorgani phosphate, the in vivo ativity of this enzyme is omparable to that of the polyphosphoinositide phosphodiesterase and muh lower than that observed with the optimized in vitro assay. The possibility that the PtdIns4P phosphatase is more widely distributed remains to be explored. Complete dephosphorylation of PtdIns(4,5)P2 by a single phosphatase would require either an enzyme having dual ativities or some other mehanism to explain the highly speifi, sequential hydrolysis of the phosphate from the 5 and then the 4 position. Sine neither phosphate is sterially hindered, it has been suggested that PtdIns(4,5)P2 must be oriented in a membrane lipoprotein omplex in suh a way that only the 5-phosphate is initially available to a single phosphatase (2). Sequential ation of two highly speifi phosphatases, as now appears to be the ase in erythroytes, provides a simpler explanation for the speifiity of the system.l This work was supported by a grant from the Medial Researh Counil of Canada (MT-2949). Some of the experiments were done by Ms. Rita Brekon. Analysis of erythroyte preparations for ellular ontamination was done by Dr. B. L. Sheridan, Hematology Dept., Vitoria General Hospital, Halifax. The editorial assistane of Dr. D. W. Russell is greatly appreiated. Manusrript reeived 25 July REFERENCES 1. Abdel-Latif, A. A Metabolism of phosphoinositides. In Handbook of Neurohemistry. Vol. 3.2nd ed. A. Lajtha, editor. Plenum Publishing Corp., New York Chang, M., and C. E. Ballou Speifiity of ox brain triphosphoinositide phosphomonoesterase. Biohem. Biophys. Res. Commun. 26: Prottey, C., J. G. Salway, and J. N. Hawthorne The strutures of enzymially produed diphosphoinositide and triphosphoinositide. Biohim. Biophys. Ata. 164: Palmer, F. B. St. C The enzymi preparation of diphosphoinositides. Prep. Bzohem. 7: Dawson, R. M. C., and W. Thompson The triphosphoinositide phosphomonoesterase of brain tissue. Biohem. J. 91: Salway, J. G., M. Kai, and J. N. Hawthorne Triphosphoinositide phosphomonoesterase ativity in nerve ell bodies, neuroglia and subellular frations from whole rat brain. J. Neurohem. 14: Sheltawy, A., M. Brammer, and D. Borrill The subellular distribution of triphosphoinositide phosphomonoesterase in guinea pig brain. Biohem. J. 128: Lee, T-C., and C. G. Huggins Triphosphoinositide phosphomonoesterase in rat kidney ortex. Arh. Biohem. Biophys. 126: Akhtar, R. A., and A. A. Abdel-Latif Studies on the properties of triphosphoinositide phosphomonoesterase and phosphodiesterase of rabbit iris smooth musle. Biohim. Biophys. Ata. 527: Cooper, P. H., and J. N. Hawthorne Phosphomonoesterase hydrolysis of polyphosphoinositides in rat kidney. Biohem. J. 150: Niijar, M. S., and J. N. Hawthorne Purifiation and properties of polyphosphoinositide phosphomonoesterase from rat brain. Biohim. Biophys. Ata Palmer, F. B. St. C The phosphatidyl-myo-inositol- 4,5-bisphosphate phosphatase from Crithidia fasiulata. Can. J. Biohem. 59: Roah, P. D., and F. B. St. C. Palmer Human erythroyte ytosol phosphatidyl-inositol-bisphosphate phosphatase (EC ). Biohim. Biophys. Ata. 661: Garrett, R. J. B., and C. M. Redman Loalization of enzymes involved in polyphosphoinositide metabolism 84 Journal of Lipid Researh Volume 25, 1984

11 on the ytoplasmi surfae of the human erythroyte membrane. Biohim. Biophys. Ata. 382: Downes, C. P., and R. H. Mihe!l The polyphosphoinositide phosphodiesterase of erythroyte membranes. Biohem. J Downes, C. P., M. C. Mussat, and R. H. Mihell The inositol trisphosphate phosphomonoesterase of the human erythroyte membrane. Biohem. J. 203: Downes, C. P., and R. H. Mihell The ontrol by Ca2* of the polyphosphoinositide phosphodiesterase and the Ca2+ pump ATPase in human erythroytes. Biohem. J. 202: Lang, V., G. Pryhitka, and J. T. Bukley Effet of neomyin and ionophore A23187 on ATP levels and turnover of polyphosphoinositides in human erythroytes. Can. J. Biohem. 55: Shneider, R. P., and L. B. Kirshner Di- and triphosphoinositide metabolism in swine erythroyte membranes. Biohim. Biophys. Ata. 202: Dittmer, J. C., and R. M. C. Dawson The isolation of a new lipid, triphosphoinositide, and monophosphoinositide from ox brain. Biohem. J. 81: Palmer, F. B. St. C Chromatography of aidi phospholipids on immobilized ne0myin.j. Lipid Res. 22: Ikezawa, H., and R. Taguhi Phosphatidylinositolspeifi phospholipase C from Baillus ereus and Baillus thuringiensis. Adv. Enrymol. 71: Palmer, F. B. St. C Hydrolysis of triphosphoinositides by a soluble fration of Crithidiafasiulata. Biohim. Biophys. Ata. 441: Wells, M. A., and J. C. Dittmer The quantitative extration and analysis of brain polyphosphoinositides. Biohemistry. 4: Bartlett, G. R Phosphorus assay in olumn hromatography. J. Biol. Chem Peterson, G. L A simplifiation of the protein assay method of Lowry et al. whih is more appliable. Anal. Biohem. 83: Coakley, W. T., and C. T. James A simple linear transform for the Folin-Lowry protein alibration urve to 1.O mg/ml. Anal. Biohem. 85: Hegyvary, C., K. Kang, and Z. Bandi Automated assay of phosphohydrolases by measuring the released phosphate without deproteinization. Anal. Biohem. 94: Bauer, J. D., P. C. Akerman, and G. Toro Clinial Laboratory Methods, 8th ed. The C. V. Mosby Co., St. Louis, MO Tokes, Z. A., and S. M. Chambers Proteolyti ativity assoiated with human erythroyte membranes. Self-digestion of isolated human erythroyte membranes. Biohim. Biophys. Ata. 389: Sott, G. K., and T. B. Kee Neutral proteases from human and ovine erythroyte membranes. Int. J. Biohem Lefebvre, Y. A., D. A. White, and J. N. Hawthorne Diphosphoinositide metabolism in bovine adrenal meduka. Can. J. Biohem. 54: Hokin, L. E., and M. R. Hokin Phosphatidi aid phosphatase in the erythroyte membrane. Biohim. Biophys. Ata. 67: Rose, Z. A The enzymology of 2,3-bisphosphoglyerate. Adu. Enzymol. 51: Badwey, J. A Phosphoglyolate phosphatase in human erythroytes. J. Biol. Chem. 252: Garrett, N. E., R. J. B. Garrett, R. T. Talwalkar, and R. L. Lester Rapid breakdown of diphosphoinositide and triphosphoinositide in erythroyte membranes. J. Cell. Physiol. 87: Mak and Palmer Phosphatidylinositol 4-phosphate phosphatase in erythroyte membranes 85