Phosphoinositide-Specific Phospholipase C δ1 Activity toward Micellar Substrates, Inositol 1,2-Cyclic Phosphate, and Other Water-Soluble Substrates: A Sequential Mechanism and Allosteric Activation†

The kinetics of full-length and PH domain truncated cloned PI-PLC δ1 from rat toward soluble substrates [inositol 1,2-(cyclic)-phosphate (cIP) and glycerophosphoinositol phosphates (GPIPx)] as well as PI in detergent micelles provide the following insights into the mechanism of this enzyme. (i) That cIP is a substrate for the enzyme implies a two-step mechanism for PI hydrolysis [intramolecular phosphotransferase reaction to form cIP followed by cyclic phosphodiesterase activity to form inositol1-phosphate (I-1-P)]. The dependence of enzyme activity on cIP is sigmoidal, suggesting a transition between less active and more active forms of the enzyme that is affected by substrate. (ii) Interfaces increase the kcat for cIP (but do not affect the cooperativity), and this allosteric activation requires an intact PH domain. (iii) Phosphorylation of the soluble inositol phosphodiesters GPI, GPIP, and GPIP2 enhances PI-PLC δ1 activity by dramatically increasing kcat and decreasing Km. For these phosphodiesters, the substrate saturation curve is no longer sigmoidal but hyperbolic, indicating the phosphorylated substrate can shift the enzyme to the activated form. (iv) Given the kinetic parameters for cIP hydrolysis and the constant ratio of cIP/I-1-P generated during PI hydrolysis, the cIP produced in situ is either released (and not readily rebound since its concentration is well below Km) or attacked by a water molecule for the generation of the acyclic product. Phosphoinositide-specific phospholipase C (EC 3.1.4.11) enzymes play a central role in many signal transduction cascades (Lee & Rhee, 1995; Rhee et al., 1989; Rhee & Choi, 1992). Mammalian PI-PLC1 hydrolysis of phosphatidylinositol 4,5-bisphosphate (PIP2) generates two second messengers: water-soluble D-myo-inositol 1,4,5-trisphosphate (IP3), which elevates the intracellular calcium level, and membrane-associated diacylglycerol (DAG), which activates many protein kinase C isozymes. Three classes of mammalian PI-PLCs with 10 different isozymes have been characterized (â1-â4, γ1-γ2, δ1-δ4). They are all calcium dependent and prefer phosphorylated inositols (Ryu et al., 1987), though to different extents. Bacterial PI-PLC, with little homology to the catalytic portion of the mammalian sequences, exhibits a phosphotransferase activity that is responsible for the formation of 1,2-cyclic inositol phosphate derivatives and a cyclic phosphodiesterase activity that leads to acyclic inositol phosphates (Volwerk et al., 1990; Bruzik & Tsai, 1994). A comparison of the structure of the bacterial enzyme (Heinz et al., 1995) to PI-PLC δ1 from rat (Essen et al., 1996) shows them both to have imperfect â-barrels with similar placement of catalytic residues. Interestingly, whereas cIP is the initial (and under most circumstances the only) product detected for PI hydrolysis by the bacterial enzymes (Volwerk et al., 1990; Bruzik et al., 1992), mammalian PI-PLCs generate both cyclic and acyclic inositol phosphates simultaneously. The ratio of cyclic to acyclic products, which appears constant during the reaction time course, depends on the isozyme class (â > δ > γ), substrate (PI > PIP > PIP2), pH, and calcium concentration (Kim et al., 1989). Early interpretations explained this kinetic behavior as a competitive attack of an activated water molecule instead of the inositol 2-hydroxyl group on the bound phosphodiester linkage (Dawson et al., 1971; Kim et al., 1989) to produce both products in parallel. However, this parallel mechanism where cIP is not an intermediate but a product contradicts the observed retention of the configuration at the 1-phosphorus on formation of acyclic inositol phosphates by PI-PLC â1 (Bruzik et al., 1992). Furthermore, substrate analogs lacking the 2-hydroxyl group were shown not to be hydrolyzable by several PIPLC enzymes (Seitz et al., 1992). Clearly, a critical experiment in understanding the detailed mechanism of mammalian PI-PLC enzymes is to determine the kinetic parameters for cIP hydrolysis. † This work has been supported by NIH Grant GM 26762 (M.F.R.), by the British Heart Foundation (R.L.W.), and by the MRC/DTI/ ZENECA/LINK Programme (R.L.W.). * To whom correspondence should be addressed. ‡ Boston College. § MRC Laboratory of Molecular Biology. | Chester Beatty Laboratories. X Abstract published in AdVance ACS Abstracts, September 1, 1997. 1 Abbreviations: PI-PLC, phosphoinositide-specific phospholipase C; PI, phosphatidylinositol; PIP2, phosphatidylinositol 4,5-bisphosphate; cIP, D-myo-inositol 1,2-(cyclic)-phosphate; I-1-P, D-myo-inositol 1-phosphate; DAG, diacylglycerol; GPI, glycerophosphoinositol; GPIP, glycerophosphoinositol 4-phosphate; GPIP2, glycerophosphoinositol 4,5bisphosphate; diC7PC, diheptanoylphosphatidylcholine; IP3, D-myoinositol 1,4,5-trisphosphate; DMSO, dimethyl sulfoxide; DMF, dimethylformamide; iPrOH, isopropyl alcohol; PH, Pleckstrin homology; ∆H, apparent activation enthalpy; ∆G, apparent activation free energy; ∆S, apparent activation entropy. 11223 Biochemistry 1997, 36, 11223-11233 S0006-2960(97)01039-8 CCC: $14.00 © 1997 American Chemical Society The activity of bacterial PI-PLC toward cIP is modulated by several factors (Zhou et al., 1997; Wu & Roberts, 1997) that could also explain the difference in mammalian versus bacterial enzyme hydrolysis of PI and cyclic versus acyclic product generation. Bacterial PI-PLC hydrolysis of cIP can be dramatically activated allosterically by the presence of an interface of PC or PE (Zhou et al., 1997). Water-miscible organic solvents such as isopropyl alcohol, dimethylformamide, or dimethyl sulfoxide also activate PI-PLC hydrolysis of cIP by mimicking the polarity of the interface and thereby stabilizing the active form of the enzyme (Wu & Roberts, 1997). Here we report a kinetic analysis of PI-PLC δ1 hydrolysis of cIP and other water-soluble substrates (glycerophosphoinositol phosphates). Determining the kinetic parameters for cIP hydrolysis, and exploring how soluble substrate phosphorylation affects these parameters, allows one to better define the mechanism for PI hydrolysis as well. Modulations of water activity were also examined for their effects on hydrolysis of cIP. All of our kinetic observations strongly support a sequential mechanism for PI hydrolysis with the rate of release of cIP from the enzyme comparable to the rate of an activated water attacking the enzyme-bound cIP to form I-1-P. Comparison of activities of the full-length enzyme versus a mutant, ∆(1-132)PI-PLC δ1, with the PH domain removed also implies that the PH domain participates not only in binding to interfaces but also in allosterically enhancing enzyme activity. MATERIALS AND METHODS Chemicals, Enzyme. DiC7PC was obtained from Avanti and used without further purification. iPrOH, DMSO, and other organic solvents were purchased from Aldrich. Triton X-100, crude PI, PIP, GPI, GPIP, GPIP2, and IP3 were all purchased from Sigma. Crude soybean PI (50% PI, purchased from Sigma) was used for the enzymatic generation of D-myo-inositol 1,2-cyclic phosphate (cIP) as described previously (Zhou et al., 1997). About 110 mg of pure cIP was obtained from 1.0 g of crude PI. Purification of PIPLC δ1 and ∆(1-132) PI-PLC-δ1, a catalytically active deletion variant lacking the N-terminal PH domain, was as described previously (Essen et al., 1997). Preparation of cIP Assay Solutions. The buffer used in all cIP assays was 50 mM Hepes, pH 7.5. A stock solution of cIP (600 mM) was prepared by dissolving cIP in D2O and adjusting the pH to 7.5 using NaOD. The range of cIP concentrations examined was 5-80 mM. A 100 mM diC7PC stock solution was prepared in D2O; the pH was adjusted to 7.5. In the cIP assays, 8 mM diC7PC (predominantly micelles given the CMC of 1.5 mM) and different amounts of DMSO were used to study the interfacial activation and organic solvent activation of the cyclic phosphodiesterase reaction. Stock solutions of GPIP2, GPIP, GPI, and IP3 were also prepared in D2O with the pH adjusted to 7.5. The calcium concentration was fixed to 0.5 mM unless otherwise indicated. The total volume of each assay solution was fixed to 350 μL. Preparation of PI Assay Solutions. PI and PIP solubilized in mixed micelles with Triton X-100 were also examined as substrates for PI-PLC δ1. Triton X-100 was used as the matrix to solubilize PI since it has been shown that (i) the extremely fast micelle exchange kinetics of Triton X-100 ensure that substrate depletion is not a problem (Soltys & Roberts, 1994), (ii) Triton X-100 is relatively noninteractive with Ca2+, and (iii) Triton X-100 micelles did not affect PIPLC δ1 hydrolysis of cIP (vide infra); hence, this interfacial matrix was relatively inert in its interactions with PI-PLC δ1. The buffer used in the PI assays was 50 mM Hepes, pH 7.5. A stock solution of PI (40 mM) was prepared by dissolving PI in D2O containing 80 mM Triton X-100 and incubating the sample in a bath sonicator for a few minutes. The ratio of PIPx/Triton X-100 was maintained at 1:2 (the minimum ratio of detergent to PI needed to solubilize all the phospholipid). The pH of the stock solution was adjusted to 7.5 using NaOD; the stock solution was optically clear. The PI concentration used in the assays was 8 mM unless otherwise indicated; sample volumes were 350 μL. After optimization, the calcium concentration was fixed at 0.5 mM. 31P NMR Assays of PI-PLC δ1 ActiVity. 31P NMR parameters were based on those used previously (Volwerk et al., 1990; Zhou et al., 1997). 31P NMR (202.3 MHz) spectra were acquired using a Varian Unity 500 spectrometer with samples in 5 mm tubes, and 5% phosphoric acid was used as an external reference. For all kinetic runs, a control spectrum (t ) 0 min) was performed prior to the addition of enzyme. The amount of enzyme added to initiate hydrolysis of water-soluble substrates varied between 18 and 160 μg, depending on the substrates used and assay conditions. For PI and PIP hydrolysis, 6 μg and 50 ng of enzyme were used. After the addition of the appropriate amount of PI-PLC, an arrayed experiment was initiated, and the hydrolysis rates were measured from the integrated intensity of the resonance corresponding to the phosphorylated product as a function of incubation time, typically 2-3 h, at 30 °C unless otherwise indicated. At all temperatures except 55 °C, the reaction time course was linear throughout most of the NMR experiment. Determination of the Energetic Parameters of PI-PLC Kinetics. The temperature dependence of kcat was analyzed according to transition state theory (Eyring, 1935; Fersht, 1985) which relates the rate constant of a reaction to an equilibrium constant between the reactants and the transition state, a transient high-energy species that decays to form product. The activation free energy, ∆G, was calculated by ∆G ) -RT ln (kcath/kBT), where h, kB, and R are Planck’s, Boltzmann’s, and the gas constant, respectively, and kcat (s-1) is the experimentally determined turnover number. The transmission coefficient (Eyring, 1935) is assumed to be unity (Fersht, 1985) and can be ignored. The activation enthalpy, ∆H, was calculated from the slope of plotting ln (kcat/T) versus 1/T, based on the Eyring equation: ln (kcat/T) ) ln (kB/h) + ∆S/R ∆H/RT. The activation entropy, ∆S, was estimated from ∆S ) (∆H ∆G)/T.