Pathway for inositol 1,3,4-trisphosphate and 1,4-bisphosphate metabolism

Proc. Natl. Acad. Sci. USA Vol. 84, pp. 217-2174, April 1987 Biochemistry Pathway for inositol 1,3,4-trisphosphate and 1,4-bisphosphate metabolism ROGR C. INHORN, VINAY S. BANSAL, AND PHILIP W. MAJRUS* Division of Hematology-Oncology, Departments of Internal Medicine and Biological Chemistry, Washington University School of Medicine, St. Louis, MO 6311 Contributed by Philip W. Majerus, December 8, 1986 ABSTRACT We prepared [3H]inositol-, 3-[32P]phosphateand 4-[32P]phosphate-labeled inositol phosphate substrates to investigate the metabolism of inositol 1,3,4-trisphosphate and inositol 1,4-bisphosphate. In crude extracts of calf brain, inositol 1,3,4-trisphosphate is first converted to inositol 3,4- bisphosphate, then the inositol 3,4-bisphosphate intermediate is further converted to inositol 3-phosphate. Similarly, inositol 1,4-bisphosphate is converted to inositol 4-phosphate, and no inositol 1-phosphate is formed. We partially purified an enzyme that we tentatively name inositol polyphosphate 1- phosphatase. This cytosolic enzyme converts inositol 1,3,4- trisphosphate to inositol 3,4-bisphosphate and also converts inositol 1,4-bisphosphate to inositol 4-phosphate. The enzyme does not utilize inositol 1,3,4,5-tetrakisphosphate, inositol 1,4,5-trisphosphate, or inositol 1-phosphate as substrates. Thus we propose a new scheme for inositol phosphate metabolism. According to this pathway inositol 1,4,5-trisphosphate and inositol 1,4-bisphosphate are degraded to inositol 4- phosphate. Inositol 1-phosphate, which is the major inositol monophosphate formed in stimulated brain, is derived either from phospholipase C cleavage of phosphatidylinositol or from the degradation of inositol cyclic phosphates. Inositol phospholipids are membrane lipids that are rapidly metabolized in response to extracellular agonists to liberate several second messenger molecules and precursors of messenger molecules (for review, see refs. 1-3). A phospholipase C enzyme hydrolyzes all three inositol phospholipids (4) to yield mixtures of cyclic and noncyclic phosphate esters (5-7), namely, inositol 1-phosphate (Ins 1P), inositol 1,4-bisphosphate [Ins(1,4)P2], inositol 1,4,5-trisphosphate [Ins(1,4,5)P3], and the (1:2) cyclic counterparts of each. The metabolism of the water-soluble inositol phosphates is complex. An inositol phosphate 5-phosphomonoesterase converts Ins(1,4,5)P3 and its cyclic counterpart to Ins(1,4)P2 (8-1) and cyclic inositol 1:2,4-bisphosphate (11), respectively. Additionally, Ins(1,4,5)P3 is metabolized to inositol 1,3,4,5-tetrakisphosphate [Ins(1,3,4,5)P4] by a 3-kinase and subsequently converted to a different inositol trisphosphate isomer, inositol 1,3,4-trisphosphate [Ins(1,3,4)P3], by the same 5-phosphomonoesterase (12-17). In the current study, we have investigated the further metabolism of Ins(1,3,4)P3 in crude extracts from calf brain utilizing substrates with radiolabels in the inositol ring, the 3-phosphate, or the 4-phosphate. We find that inositol 3,4- bisphosphate [Ins(3,4)P2] is the first product formed from Ins(1,3,4)P3. In addition, we find that the enzyme that utilizes this substrate also converts Ins(1,4)P2 to inositol 4-phosphate (Ins 4P). Based on these results, a new pathway for inositol phosphate metabolism is proposed. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. 1734 solely to indicate this fact. MATRIALS AND MTHODS Materials. [3H]Ins(1,4)P2 and [3H]Ins(1,4,5)P3 were from New ngland Nuclear. [-32P]ATP was from Amersham. Matrix gel PBA-6 (phenyl boronate agarose) was from Amicon. DA-Sepharose CL-6B was from Sigma. The Bio-Gel TSK DA-5 PW HPLC column, AG 1-X8 Dowex (formate form, 2-4 mesh) and Chelex 1 (sodium form, 2-4 mesh) were from Bio-Rad. The Whatman Partisil 1 SAX column was from Cobert Associates (St. Louis, MO). All other materials were from Sigma or Fisher. Preparation of Radiolabeled Ins 1P, Ins(1,4)P2, and Ins(1,4,5)P3. [3H]Ins 1P was prepared by incubation of lipid vesicles containing phosphatidyl[3h]inositol with ram seminal vesicle phospholipase C (18, 19). Ins(1,4-[4-32P])P2 and Ins(1,4,5-[4,5-32P])P3 were prepared from erythrocytes as described by Downes et al. (9). Preparation of Unlabeled Ins(1,4,5)P3. Unlabeled Ins(1,4,5)- P3 was prepared from Folch fraction I (Sigma) by the method of Irvine et al. (2) as detailed (21). Dowex Chromatography of Inositol Phosphates. Reaction mixtures were diluted to 1 ml with H2 containing 5 nmol Pi and poured onto a 1.5 x 11 cm Dowex-formate column equilibrated in 5 mm NH4COOH (ph 3.45). Columns were eluted with linear gradients (2 ml) of.5-1.5 M NH4COOH. Fractions (2 ml) were collected, and radioactivity was estimated by liquid scintillation counting with Scintiverse I. Rat Brain Homogenate. Crude rat brain cytosol was prepared by the method of Irvine et al. (14) as described (17). Briefly, rat brain cortex in 16 mm sucrose/1 mm Hepes, ph 7.5, was homogenized with a Teflon-glass homogenizer, and the homogenate was centrifuged at 5, x g for 3 min. The supernatant was used as a crude source ofboth Ins(1,4,5)- P3 3-kinase and inositol phosphate 5-phosphomonoesterase. In some cases the 3-kinase was partially purified as described (17). Preparation of Radiolabeled Ins(1,3,4,5)P4 and Ins(1,3,4)P3. Reaction conditions were found in which incubation of Ins(1,4,5)P3 and ATP with crude rat brain cytosol yielded primarily Ins(1,3,4,5)P4 at early time points. Ins(1,3,4,5)P4 was further converted to Ins(1,3,4)P3 with longer incubations. 3H-labeled substrates. Tritium-labeled Ins(1,3,4,5)P4 and Ins(1,3,4)P3 were prepared by incubating [3H]Ins(1,4,5)P3 (3 nmol,.2,aci/nmol; 1 Ci = 37 GBq) in 5 Al containing 5 mm glycine-naoh (ph 1.), 2 mm MgCl2, 1mM ATP, 4 mm Na2HPO4, and crude rat brain cytosol (5,ug of protein) Abbreviations: Ins(1,4,5)P3, inositol 1,4,5-trisphosphate; Ins 1P, inositol 1-phosphate; Ins(1,4)P2, inositol 1,4-bisphosphate; Ins(1,3,4,5)P4, inositol 1,3,4,5-tetrakisphosphate; Ins(1,3,4)P3, inositol 1,3,4-trisphosphate; Ins 3P, inositol 3-phosphate; Ins 4P, inositol 4-phosphate. *To whom reprint requests should be addressed at: Division of Hematology-Oncology, Washington University School of Medicine, 66 South uclid, St. Louis, MO 6311. 217

Biochemistry: Inhorn et al. for 45 min at 37C. The products of this reaction were separated by Dowex-formate chromatography (17) using Ins(1,4,5-[4,5-32P])P3 as an internal standard. Approximately 2/3 of the radioactivity was recovered in peaks migrating in the positions of Ins(1,3,4)P3 and Ins(1,3,4,5)P4 with equal amounts of both products. These products result from the 3-kinase and 5-phosphatase activities in this preparation. The [3H]Ins(1,3,4,5)P4 product was further characterized by converting a sample of it to [3H]Ins(1,3,4)P3 with the 5-phosphatase enzyme from platelets (1). The [3H]Ins(1,3,4)P3 was characterized by showing that it was resistant to 5- phosphomonoesterase and that it yielded altritol after periodate oxidation, borohydride reduction, and dephosphorylation (13, 17, 2). 4-Labeled Ins(1,3,4)P3. Ins(1,3,4-[4-32P])P3 was prepared in the same way except that Ins(1,4,5-[4,5-32P])P3 prepared from erythrocyte ghosts was used as the initial substrate. The erythrocyte substrate is labeled with 32p in both the 4- and 5-phosphates, with -2/3 of the radioactivity in the 5-position (9). The 5-phosphate is removed by the 5-phosphomonoesterase activity in the rat brain cytosol, thereby yielding Ins(1,3,4)P3 specifically labeled in the 4-position. Starting with 9 nmol of Ins(1,4,5-[4,5-32P])P3 (25, cpm, Cerenkov) we obtained 35, cpm in Ins(1,3,4-[4-32P])P3 product. This reflects =4% yield with a specific activity of about 9 cpm/nmol. 3-Labeled Ins(1,3,4)P3. [3H]Ins(1,3,4-[3-32P])P3 was prepared using purified 3-kinase and 5-phosphomonoesterase enzymes. [y-32p]atp (.33 uci/nmol) was applied to a 2-ml phenyl boronate agarose column to remove radiocontaminants from the ATP. The column was washed with 12 ml of.2 M triethylammonium bicarbonate, ph 8.8, containing 15 mm MgCl2. ATP was eluted with H2 and lyophilized. [3H]Ins(1,4,5)P3 (3 nmol,.33,ci/nmol) and [y- 2P]ATP (.25,umol) were incubated at 37 C for 1 min in 5,ul of 5 mm Tris maleate (ph 7.5), 1 mm MgCl2, 1 mm sodium phosphate (ph 7.5), and partially purified inositol trisphosphate 3-kinase (17). The reaction was stopped by adding 1 ml of 5 mm NH4COOH (ph 3.45), and the products were separated by Dowex-formate chromatography (17). Fractions containing Ins(1,3,4,5)P4 were pooled, diluted to 2 ml with H2, then rechromatographed on Dowex-formate to further resolve the [3H]Ins(1,3,4,5-[3-32P])P4 from contaminating labeled ATP. The [3H]Ins(1,3,4,5-[3-32P])P4 (21, cpmof3h]) was combined with 7 nmol unlabeled Ins(l,3,4,5)- P4 prepared in a parallel reaction, then desalted (22). The [3H]Ins(1, 3,4,5-[3-_2p])p4 was incubated 1 hr at 37 C with inositol phosphate 5-phosphomonoesterase (7 nmol/min), 5 mm Mes (ph 6.5), and 3 mm MgCl2 in 4,d. The product was chromatographed on Dowex-formate as described above. [3H]Ins(1,3,4-[3-32P])P3, which was separated from unreacted Ins(1,3,4,5)P4, was desalted as before. The final yield of product was 52, cpm of tritium and 55, cpm of 32p with a specific activity of 3 cpm of 3H per nmol. Inositol Monophosphate Phosphatase Assay. Inositol monophosphate phosphatase activity was assayed by measuring the formation of [3H]inositol from [3H]Ins 1P as described (1). Inositol Polyphosphate 1-Phosphatase Assay. Inositol polyphosphate 1-phosphatase activity was assayed by one of two methods. Assay 1: 4-Phosphate release. Inositol polyphosphate 1-phosphatase converts Ins(1,4)P2 to Ins 4P, then inositol monophosphate phosphatase hydrolyzes the 4-phosphate as described below. Since crude homogenates and fractions from early steps of the inositol polyphosphate 1-phosphatase purification contain contaminating inositol monophosphate phosphatase, 1-phosphatase activity in these fractions can be estimated by measuring 4-phosphate release. Samples were incubated in 3 mm MgCl2, 5 mm Hepes (ph 7.5), and Proc. Natl. Acad. Sci. USA 84 (1987) 2171 Ins(1,4-[4-32P])P2 (1-2,uM) in 5 ud at 37C. Reactions were initiated by the addition of enzyme and stopped by addition of 75 /.l of 1.67 M perchloric acid. Inorganic phosphate was complexed with 1% ammonium molybdate, extracted with 2 volumes of isobutanol/toluene, 1:1 (vol/vol) (9, 23), and measured as Cerenkov radioactivity. This assay was linear until 35% of the substrate was hydrolyzed. Most assays were for 15 min. Assay 2: Coupled 4-phosphate release. Samples were assayed as in assay 1 with the addition of partially purified inositol monophosphate phosphatase (1 nmol of Ins 1P hydrolyzed per min) (see below). The reaction was initiated by the addition of inositol polyphosphate 1-phosphatase and carried out for 15 min. [32P]Pi release was measured as described for assay 1. Inositol monophosphate phosphatase alone hydrolyzed no [32P]Pj from Ins(1,4-[4-32P])P2. Purification of Phosphatases. Inositol polyphosphate 1- phosphatase and inositol monophosphate phosphatase were partially purified from calf brain (unpublished results). Briefly, 1 calf brains from a local abattoir were rapidly frozen on dry ice. Brains were shredded in a commercial vegetable shredder and homogenized in buffer (1:2, wt/vol) with a Polytron homogenizer in 2 mm Hepes, ph 7.5/1 mm GTA/2 mm DTA/.25 M sucrose/1,um phenylmethylsulfonyl fluoride. The homogenate was centrifuged 1 hr at 3, X g, then the supernatant was recentrifuged 1 hr at 48,4 x g. The supernatant was chromatographed sequentially on a DA-Sepharose CL-6B column and a Bio-Gel TSK DA-5 PW HPLC column. Inositol polyphosphate 1-phosphatase and inositol monophosphate phosphatase activities were partially resolved on the DA-resin using HPLC. The separated pooled fractions of each enzyme were individually rechromatographed on the DA-resin by HPLC, and the resulting further-resolved peaks of activity were used for the experiments described below. The inositol monophosphate phosphatase pool was heated to 7 C for 3 min to inactivate small amounts of contaminating inositol polyphosphate 1-phosphatase activity (see Results). RSULTS Metabolism of [3H]Ins(1,3,4-[3-32P])P3 in Crude Calf Brain Supernatant. We prepared [3H]Ins(1,3,4-[3-32P])P3 to investigate the order in which phosphates are removed from Ins(1,3,4)P3. [3H]Ins(1,3,4-[4-32P])P3 degradation was measured at 2, 5, 1, 2, and 6 min after addition of.29 mg of protein per ml from a calf brain supernatant fraction. The products were analyzed by Dowex-formate chromatography. The earliest product detected was an inositol bisphosphate eluting before Ins(1,4)P2. In six experiments, the ratio of 32P/3H in the unknown inositol bisphosphate was.99 ±.8 (SD), while that in the unreacted substrate Ins(1,3,4)P3 was 1.6 ±.5. The inositol bisphosphate isomer, therefore, retained the 32P-labeled 3-phosphate, indicating that either the 1- or the 4-phosphate was removed first, yielding Ins(3,4)P2 or Ins(1,3)P2. Purification of Inositol Polyphosphate 1-Phosphatase. We set out to purify the enzyme catalyzing the first step in the degradation of Ins(1,3,4)P3 to determine whether the 1- or the 4-phosphate is removed first. As a substrate, we used Ins(1,4-[4-32P])P2 prepared from erythrocyte ghosts due to the relative ease of its preparation. Our original assumption was that the 4-phosphate was removed first from Ins(1,3,4)P3 and that the putative 4-phosphatase enzyme would also utilize Ins(1,4)P2. Calf brain was used as the starting material for enzyme purification. Both the crude brain supernatant and the fractions from the initial chromatography step (DA-Sepharose CL-6B) showed activity in the 4-phosphate release assay (see assay 1). Subsequent HPLC using a DA-resin column, however, led to an apparent >9%o loss

2172 Biochemistry: Inhorn et al. of activity using this assay. A mixture of [3H]Ins(1,4)P2 and Ins(1,4-[4-32P])P2 was incubated with a sample of the enzyme pooled from this column, and the products were separated using Dowex-formate chromatography. The product of this reaction was an inositol monophosphate with the same 32P/3H ratio as the unreacted Ins(1,4)P2 substrate. This indicates that the inositol bisphosphatase is in fact a 1- phosphatase that yields Ins 4P as a product. This enzyme does not utilize Ins1P as a substrate (see below) and will thus be termed inositol polyphosphate 1-phosphatase. An enzyme that hydrolyzes the phosphate from Ins1P has been purified from rat (24) and bovine brain (25). This enzyme, termed myo-inositol-1-phosphatase (C 3.1.3.25), has been shown to utilize Ins 4P and inositol 5-phosphate as substrates (26) in addition to both d- and i-ins1p. Therefore, this enzyme will be referred to in this paper as inositol monophosphate phosphatase. Inositol monophosphate phosphatase was partially purified from the same calf brain preparation. While the inositol monophosphate phosphatase alone showed no activity with Ins(1,4-[4-32P])P2, mixture of the enzyme with HPLC fractions enriched in inositol polyphosphate 1-phosphatase restored 4-phosphate release activity to the fractions. This result indicates that the inositol monophosphate phosphatase hydrolyzes the phosphate from the Ins 4P intermediate. Heat Lability of Inositol Polyphosphate 1-Phosphatase. Inositol monophosphate phosphatase is stable at 7'C (24, 25). Table 1 shows an experiment in which a fraction from the DA-Sepharose CL-6B column containing both inositol monophosphate phosphatase and inositol polyphosphate 1- phosphatase activities was incubated at 7 C for various lengths of time. While the rate of hydrolysis of Ins1P was stable over the time course examined, greater than 9% of the inositol polyphosphate 1-phosphatase activity was inactivated within 5 min. In further purification of inositol polyphosphate 1-phosphatase, we used assay 2 to resolve the enzyme from inositol monophosphate phosphatase. The ratio of the inositol polyphosphate 1-phosphatase activity to inositol monophosphate phosphatase activity in the most highly purified enzyme pool is 11-fold greater than that in the starting supernatant. The residual activity using Ins1P as substrate is heat stable, suggesting that it results from inositol monophosphate phosphatase contamination and not slight activity of inositol polyphosphate 1-phosphatase on Ins 1P. Action of Inositol Polyphosphate 1-Phosphatase on Ins- (1,3,4)P3. Ins(1,3,4-[4-32P])P3 substrate was prepared. This substrate was incubated with inositol polyphosphate 1- phosphatase, and the products were resolved as shown in Fig. LA. The product is a 4-labeled inositol bisphosphate that migrates faster than Ins(1,4)P2 in a position identical to the 3-labeled bisphosphate that resulted from metabolism of Ins(1,3,4-[3-32P])P3. This product must, therefore, be Ins- Table 1. Heat inactivation of inositol polyphosphate 1-phosphatase Incubation, Inositol, 4-Phosphate release, min nmol min-1 ml-' nmol min-1ml-1 318 64 5 285 5.3 15 269 6.2 3 262 5.9 Hydrolysis of Ins 1P was used to determine inositol monophosphate phosphatase activity, while release of [32P]P, from Ins(1,4-[4-32P])P2 was used to estimate inositol polyphosphate 1-phosphatase activity. A pool from the DA-Sepharose CL-6B column enriched in inositol monophosphate phosphatase activity was incubated at 7TC. Samples were removed at the indicated times, denatured protein was pelleted by centrifugation for 15 min at 1, x g, and assays were performed. - L1.1. a- cm in, 2- A 1-8- 6-4- 2- o 2- u 16 u- 12 8 -! 4 - Proc. Natl. Acad. Sci. USA 84 (1987) 1 A ~~~wu-wm '--.-- 2 3 4 5 6 7T -- 8U in nmyrrrrern-e 431%M 2 3 4 I 1 2F 3 4 5 6 P pi Ylns(1,4)P2 Iris(1,3,4)P3 I I I i Ins(1,4)PIns(1,3,4)P3 I I FIG. 1. Hydrolysis of Ins(1,3,4)P3 by partially purified inositol polyphosphate 1-phosphatase. (A)Ins(1,3,4-[4-32P])P3 (1 nmol) was incubated for 3 min at 37C with partially purified inositol polyphosphate 1-phosphatase (6,Ag) in 2 A.l containing 5mM Hepes (ph 7.5) and 3 mm MgCl2. The reaction mixture was chromatographed on Dowex-formate; after collecting 6 fractions, the gradient was stopped, and the column was step-eluted with 1.5 M NH4COOH, ph 3.45. The elution position of a [3H]Ins(1,4)P2 standard is indicated. (B) [3H]Ins(1,3,4)P3 (.1 nmol) was incubated for 15 min at 37C with 1.5,tg of inositol polyphosphate 1- phosphatase, 5mM Hepes (ph 7.5), and 3mM MgCl2. The reaction mixture was chromatographed on an HPLC SAX column as described (21). The elution positions of 32P-labeled Pi and Ins(1,4)P2 standards are indicated; note that the bisphosphate reaction product elutes after Ins(1,4)P2 in this chromatography system. The position of unreacted Ins(1,3,4)P3 is also indicated. (3,4)P2 resulting from removal of the 1-phosphate from Ins(1,3,4)P3, since both the specifically labeled 4-phosphate (Fig. 1) and 3-phosphate groups are retained. Fig. 1B shows the HPLC elution position of [3H]Ins(3,4)P2 on a strong anion exchange column. Ins(3,4)P2 elutes after Ins(1,4)P2 on HPLC in contrast to its behavior in Dowexformate chromatography. Ins(3,4)P2 is further metabolized to inositol 3-phosphate (Ins 3P) as shown in Fig. 2. In separate incubations containing 2 mm inositol 2-phosphate to inhibit the inositol monophosphate phosphatase, only Ins 3P was formed. Action of InositolPolyphosphate 1-PhosphataseonIns(1,4,5)- P3 and Ins(1,3,4,5)P4. [3H]Ins(1,4,5)P3 (1 nmol, 15 cpm/ nmol) was incubated in 5,ul with an amount of inositol polyphosphate 1-phosphatase capable of hydrolyzing 6 nmol of Ins(1,4)P2. No hydrolysis of the trisphosphate was observed under these conditions. In a separate experiment, Ins(1,4-[4-32P])P2 (1 nmol) included in the incubation was quantitatively converted to Ins 4P while no Ins(1,4,5)P3 was hydrolyzed. Similarly, [3H]Ins(1,3,4,5)P4 (.1 nmol, 1, cpm/nmol) was incubated in 1,ul with inositol polyphosphate 1-phosphatase, both with and without Ins(1,4-[4-32P])P2 (1 nmol) in the reaction. No hydrolysis of Ins(1,3,4,5)P4 was observed in either condition, though the Ins(1,4)P2 was converted to Ins 4P. Metabolism of Ins(1,4)P2 in Crude Brain Homogenate. We investigated the pathway of Ins(1,4)P2 metabolism in the crude calf brain supernatant to determine if all Ins(1,4)P2 is converted to Ins 4P or whether an alternative pathway involving Ins 1P exists. We incubated [3H]Ins(1,4-[4-32P])P2

. I'4 5-4 3-2- 1 Biochemistry: Inhorn et al. Ins-3-P Ins-4-P 1 2 3 4 5 6 7 8 9 FIG. 2. Inositol monophosphate product of Ins(3,4)P2 metabolism by crude calf brain supernatant. [3H]Ins(3,4)P2 was prepared by incubation of [3H]Ins(1,3,4)P3 with inositol polyphosphate 1-phosphatase and isolated by Dowex-formate chromatography using a linear gradient (2 ml) of.5-1 M NH4COOH (ph 3.45). [3H]Ins- (3,4)P2 (.3 nmol) was incubated 2 min at 37C with 25,g of crude calf brain supernatant, 5 mm Mes (ph 6.5), and 3 mm MgCl2 in.1 ml, and the products were separated by Dowex-formate chromatography. The elution positions of standards are indicated. with crude calf brain supernatant using conditions in which -5% of the substrate was hydrolyzed as shown in Fig. 3. dl-ins 1P was included in this reaction to saturate the inositol monophosphate phosphatase and thus trap any inositol monophosphate intermediate of Ins(1,4)P2 metabolism. The 32P/3H ratio in the inositol monophosphate peak is identical to that of the unreacted Ins(1,4)P2, indicating that all Ins(1,4)P2 breakdown occurred via 1-phosphate removal. Moreover, no peak of 32p inorganic phosphate was observed, 2 18 _ 16 4-2 1 T~ 8 ~6-4- 2- Ins-4-P P. Ins( 1,4) P2 1 2 3 4 5 6 Pi Ins(3,4)R Ins(1,4)P2 i I2f- FIG. 3. Inositol monophosphate product of Ins(1,4)P2 metabolism by crude calf brain supernatant. A mixture of [3H]Ins(1,4)P2 and Ins(1,4-[4-32P])P2 (2.5 nmol total) was incubated 1 min at 37 C with 25,g of crude calf brain supernatant, 5 mm Hepes (ph 7.5), and 3 mm MgCl2 in 5,ul. dl-ins 1P (1.2 mm) was included to saturate the inositol monophosphate phosphatase and thus trap the inositol monophosphate product of Ins(1,4)P2 metabolism. The reaction mixture was chromatographed on Dowex-formate. Proc. Natl. Acad. Sci. USA 84 (1987) 2173 which demonstrates that no phosphate was removed from the 4-position of Ins(1,4)P2. This experiment was repeated with a sample of unwashed membrane pellet from the 48,4 x g centrifugation (21 Ag ofprotein, resuspended by sonication in 5 mm Hepes, ph 7.5) to investigate whether a membraneassociated 4-phosphatase exists. Again, the 32P/3H ratio in the inositol monophosphate peak was identical to that of the unreacted Ins(1,4)P2 and no peak of [32p]p, was observed. We conclude that under these experimental conditions all Ins(1,4)P2 is hydrolyzed to Ins 4P. DISCUSSION This study describes a pathway for the metabolism of Ins(1,3,4)P3 and presents information on the breakdown of Ins(1,4)P2. Our current understanding of inositol phosphate metabolism is illustrated in Fig. 4. We demonstrated that the first product of Ins(1,3,4)P3 hydrolysis in calf brain is Ins(3,4)P2, an inositol bisphosphate isomer. Ins(3,4)P2 elutes after Ins(1,4)P2 on HPLC using a SAX column (Fig. 2B). This elution position is consistent with that of the inositol bisphosphates resulting from Ins(1,3,4)P3 metabolism that were described in rat liver (27) and rat parotid glands (16). Hansen et al. (27) suggested that the new inositol bisphosphate was Ins(1,3)P2, which is inconsistent with our results.4: Preliminary evidence from our laboratory suggests that human platelets also hydrolyze Ins(1,3,4)P3 to yield Ins(3,4)- P2. Removal of the 1-phosphate from Ins(1,3,4)P3 may thus be a general initial step in the pathway of Ins(1,3,4)P3 metabolism. Our experiments with [3H]Ins(3,4)P2 showing conversion to inositol 3-phosphate (see Fig. 2) suggest that Ins(3,4)P2 is further hydrolyzed to Ins 3P by a 4-phosphatase. We find that Ins(1,4)P2 metabolism in calf brain yields Ins 4P as the sole monophosphate product (see Fig. 3). This finding contradicts the conclusions of several other investigators. Storey et al. (28) claim that Ins(1,4)P2 is acted upon by both a 1- and a 4-phosphatase in rat liver. They support this hypothesis with the observation that Ins(1,4-[4-32P])P2 hydrolysis yielded both a labeled inositol monophosphate (Ins 4P) and labeled Pi. We included Ins IP in our incubations to suppress inositol monophosphate phosphatase activity and detected no release of labeled Pi from Ins(1,4-[4-32P]P2. Since no such inhibitors were included in the former study, it is possible that the [32P]P, measured by Storey et al. (28) resulted from inositol monophosphate phosphatase action upon the Ins 4[32p]p intermediate rather than direct action of a 4-phosphatase upon Ins(1,4-[4-32P])P2. Berridge et al. (29) detected a 4-phosphate specific bisphosphatase in blowfly salivary glands, though these data were not specifically reported. It is unclear whether this represents a difference among species or whether the investigators were unable to resolve Ins 4P and Pi on the Dowex chromatography system used. Finally, Ackermann et al. (26) incubated 7,uM Ins(1,4)P2 with crude rat brain cytosol and detected formation of both Ins 4P and Ins 1P, though Ins 4P was the major product. The difference between this result and the data presented here might result from action of a 4-phosphatase with low affinity for Ins(1,4)P2 that we were unable to detect at low substrate concentrations. An interesting feature of the pathway described here is that Ins(1,4,5)P3 recycles to inositol via Ins 4P rather than Ins 1P. Sherman et al. (3) directly measured the levels of Ins 1P and Ins 4P in rat cerebral cortex using gas chromatography. In both control and lithium-treated animals, Ins 1P was the predominant inositol monophosphate. In lithium-treated tunder some conditions we find evidence for a 4-phosphatase that utilizes Ins(1,3,4)P3 to form Ins(1,3)P2 (unpublished results).

2174 Biochemistry: Inhorn et al. Proc. Natl. Acad. Sci. USA 84 (1987) /mpkc -activation PA= 1,2 dioclglycerol -rochidonte icosanoid c1:2p 11:2,4P2--- -CII2 4.5P3 14oPdI 3$4-7m1ACO 3P --- I3,P2 - [ 1,3.4P3 mobilization FIG. 4. Scheme for inositol phospholipid breakdown and resynthesis. I, inositol; P, phosphate; numbers, phosphate position on the inositol ring. Note that the noncyclic inositol phosphates are metabolized to inositol via inositol 4-phosphate and Ins 3P, while Ins 1P is derived from cyclic inositol phosphate metabolism and direct phosphatidylinositol breakdown. rats, 92% of the inositol monophosphate was Ins 1P, while 8% was Ins 4P. Similarly, Siess (31) found that Ins 1P was the most abundant inositol monophosphate in thrombin-stimulated platelets using an HPLC system that separates Ins 1P from Ins 4P. The current results suggest that the presence of Ins 1P in cells results from either direct action of phospholipase C upon phosphatidylinositol or from hydrolysis of cyclic inositol 1:2 phosphate to Ins 1P (see Fig. 4). We have partially purified an enzyme that removes the 1-phosphate from both Ins(1,3,4)P3 and Ins(1,4)P2. Our preliminary work suggests that this enzyme does not utilize Ins 1P as a substrate, so we have tentatively named the enzyme inositol polyphosphate 1-phosphatase. We find no activity of this enzyme upon Ins(1,4,5)P3 or Ins(1,3,4,5)P4. Our preliminary data show that 5 mm Li' causes >5% inhibition of the partially purified inositol polyphosphate 1-phosphatase activity. This result is consistent with the findings of several investigators who noted increases in Ins(1,3,4)P3 and Ins(1,4)P2 after Li' treatment (27, 32, 33). In addition, Storey et al. (28) found that Li' inhibited bisphosphatase activity in rat liver cytosol. An enzyme activity similar to the inositol polyphosphate 1-phosphatase described in this paper may thus account for the inositol phosphate metabolism reported by those investigators. We thank Dan Lips, Hans Deckmyn, Teresa Bross, and Brian Whiteley for advice and assistance; Susan Brodack and Cecil Buchanan for substrate preparation; Lois Isenberg for expert typing; and Tom Connolly for invaluable discussions throughout the course of this work. This research was supported by Grants HLBI 14147 (Specialized Center for Research in Thrombosis), HL 16634, and Training Grant T32 HLBI 788 from the National Institutes of Health, and by National Institutes of Health Research Service Award GM-72, Medical Scientist, from the National Institute of General Medical Sciences. 1. Berridge, M. J. & Irvine, R. F. (1984) Nature (London) 312, 315-321. 2. Nishizuka, Y. (1986) Science 233, 35-312. 3. Majerus, P. W., Connolly, T. M., Deckmyn, H., Ross, T. S., Bross, T.., Ishii, H., Bansal, V. S. & Wilson, D. B. (1986) Science 234, 1519-1526. 4. Wilson, D. B., Bross, T.., Hofmann, S. L. & Majerus, P. W. (1984) J. Biol. Chem. 259, 11718-11724. 5. Dawson, R. M. C., Freinkel, N., Jungalwala, F. B. & Clarke, N. (1971) Biochem. J. 122, 65-67. 6. Wilson, D. B., Bross, T.., Sherman, W. R., Berger, R. A. & Majerus, P. W. (1985) Proc. Nati. Acad. Sci. USA 82, 413-417. 7. Wilson, D. B., Connolly, T. M., Bross, T.., Majerus, P. W., Sherman, W. R., Tyler, A. N., Rubin, L. J. & Brown, J.. (1985) J. Biol. Chem. 26, 13496-1351. 8. Roach, P. D. & Palmer, F. B. St. C. (1981) Biochim. Biophys. Acta 661, 323-333. 9. Downes, C. P., Mussat, M. C. & Michell, R. H. (1982) Biochem. J. 23, 169-177. 1. Connolly, T. M., Bross, T.. & Majerus, P. W. (1985) J. Biol. Chem. 26, 7868-7874. 11. Connolly, T. M., Wilson, D. B., Bross, T.. & Majerus, P. W. (1986) J. Biol. Chem. 261, 122-126. 12. Irvine, R. F., Letcher, A. J., Lander, D. J. & Downes, C. P. (1984) Biochem. J. 223, 237-243. 13. Batty, I. R., Nahorski, S. R. & Irvine, R. F. (1985) Biochem. J. 232, 211-215. 14. Irvine, R. F., Letcher, A. J., Heslop, J. P. & Berridge, M. J. (1986) Nature (London) 32, 631-634. 15. Downes, C. P., Hawkins, P. T. & Irvine, R. F. (1986) Biochem. J. 238, 51-56. 16. Hawkins, P. T., Stephens, L. & Downes, C. P. (1986) Biochem. J. 238, 57-516. 17. Connolly, T. M., Bansal, V. S., Bross, T.., Irvine, R. F. & Majerus, P. W. (1987) J. Biol. Chem., in press. 18. Hofmann, S. L. & Majerus, P. W. (1982) J. Biol. Chem. 257, 6461-6469. 19. Ross, T. S. & Majerus, P. W. (1986) J. Biol. Chem. 261, 11119-11123. 2. Irvine, R. F., Letcher, A. J., Lander, D. J. & Berridge, M. J. (1986) Biochem. J. 24, 31-34. 21. Auchus, R. J., Kaiser, S. L. & Majerus, P. W. (1987) Proc. Natl. Acad. Sci. USA 84, 126-129. 22. Burgess, G. M., Godfrey, P. P., McKinney, J. S., Berridge, M. J., Irvine, R. F. & Putney, J. W. (1984) Nature (London) 39, 63-66. 23. Martin, J. B. & Doty, D. M. (1949) Anal. Chem. 21, 965-967. 24. Takimoto, K., Okada, M., Matsuda, Y. & Nakagawa, H. (1985) J. Biochem. (Tokyo) 98, 363-37. 25. Hallcher, L. M. & Sherman, W. R. (198) J. Biol. Chem. 255, 1896-191. 26. Ackermann, K.., Gish, B. G., Honchar, M. P. & Sherman, W. R. (1987) Biochem. J., in press. 27. Hansen, C. A., Mah, S. & Williamson, J. R. (1986) J. Biol. Chem. 261, 81-813. 28. Storey, D. J., Shears, S. B., Kirk, C. J. & Michell, R. H. (1984) Nature (London) 312, 374-376. 29. Berridge, M. J., Dawson, R. M. C., Downes, C. P., Heslop, J. P. & Irvine, R; F. (1983) Biochem. J. 212, 473-482. 3. Sherman, W. R., Munsell, L. Y., Gish, B. G. & Honchar, M. P. (1985) J. Neurochem. 44, 798-87. 31. Siess, W. (1985) FBS Lett. 185, 151-156. 32. Burgess, G. M., McKinney, J. S., Irvine, R. F. & Putney, J. W. (1985) Biochem. J. 232, 237-243. 33. Turk, J., Wolf, B. A. & McDaniel, M. L. (1986) Biochem. J. 237, 259-263.