Cell. Signal. Vol. 9, No. 5, pp. 329–336, 1997 Copyright 1997 Elsevier Science Inc.
ISSN 0898-6568/97 $17.00 PII S0898-6568(96)00175-1
TOPICAL REVIEW
Structural and Mechanistic Features of Phospholipases C: Effectors of Inositol Phospholipid-Mediated Signal Transduction Stephen R. James* and C. Peter Downes Department of Biochemistry, Medical Sciences Institute, University of Dundee, Dundee, DD1 4HN, Scotland, UK
ABSTRACT. The production of the intracellular second messengers inositol (1,4,5)-trisphosphate (InsP3) and sn 1,2-diacylglycerol (DG) in response to a wide variety of extracellular primary messengers is achieved by an extended family of inositol phospholipid phosphodiesterases termed phospholipases C (PLC, E.C. 3.1.4.11). This family has been the subject of extensive research and it is clear that the different isoenzymes exhibit some common characteristics (e.g., interactions with substrates) and other distinctive features (e.g., modes of regulation). The recent description of the X-ray crystal structure of a mammalian PLC has served to clarify much about the behaviour of the PLCs, emphasising the “modular” structure of these enzymes. The main focus of this review will concern the specific adaptations of PLC molecules which make them efficient lipid-metabolising enzymes. We also describe what is known about how these enzymes interact with their lipid substrates, which will serve as a basis for considering how PLCs may be activated. cell signal 9;5:329–336, 1997. 1997 Elsevier Science Inc. KEY WORDS. Phospholipases C; Signal transduction, effectors of inositol phospholipid-mediated
INTRODUCTION
Molecular Heterogeneity of PLC Isoforms
The production of the intracellular second messengers inositol (1,4,5)-trisphosphate (InsP 3) and sn 1,2-diacylglycerol (DG) in response to a wide variety of extracellular primary messengers is achieved by an extended family of inositol phospholipid phosphodiesterases termed phospholipases C (PLC, E.C. 3.1.4.11). This family has been the subject of extensive research and it is clear that the different isoenzymes exhibit some common characteristics (e.g. interactions with substrates) and other distinctive features (e.g. modes of regulation). The recent description of the X-ray crystal structure of a mammalian PLC [1] has served to clarify much about the behaviour of the PLCs, emphasising the ‘‘modular’’ structure of these enzymes. The main focus of this review will concern the specific adaptations of PLC molecules which make them efficient lipid-metabolising enzymes. We also describe what is known about how these enzymes interact with their lipid substrates, which will serve as a basis for considering how PLCs may be activated.
The PLC family comprises three different classes of enzyme termed b, g and d. At least ten different isoforms have been purified and cloned (PLCb1-4, PLCg1 and 2, PLCd1-4) [2– 8]. The reader is directed to other reviews for the relationships which exist between these different proteins [9–12]. Although it is not clear at present whether the activity of PLCd is regulated other than by intracellular Ca 11—only one paper has appeared which described the activation of PLCd by a rho-GAP-like protein [13]—significantly more is known about the activation of PLCb and PLCg. PLCb isoforms are activated in response to agonists which activate heptahelical receptors, via the intermediary activity of heterotrimeric GTP-binding proteins (G proteins). Thus, a subunits of the Gq family and bg subunits are able to stimulate PLCb1-3, with differing potencies and relative efficacies [14–20]. PLCb4, however, is refractory to bg subunits but can be activated by ribonucleotides [21, 22]. By contrast, PLCg isoforms are activated by receptors with intrinsic or associated tyrosine kinase activity [9–12].
* Author to whom all correspondence should be addressed. E-mail: ,
[email protected].. Abbreviations: PLC–phospholipase C; PLCbT—turkey erythrocyte PLC; PtdInsP—phosphatidylinositol 4-phosphate; PtdInsP2—phosphatidylinositol (4,5)-bisphosphate; PtdCho—phosphatidylcholine; PtdSer—phosphatidylserine; InsP3—inositol (1,4,5)-trisphosphate; DG—sn 1,2-diacylglycerol; PH domain—pleckstrin homology domain; PITP—phosphatidylinositol transfer protein; G protein—heterotrimeric GTP-binding protein. Received 12 August 1996; accepted 1 November 1996.
Modular Structure of PLCs and Relationship to Function Mammalian PLCs are modular proteins, each consisting of several distinct domains, some of which also occur in proteins unrelated to the PLCs. Examination of the secondary
330
S. R. James and C. P. Downes
FIGURE 1. Linear alignment of the three main members of the family of eukaryotic phosphoinositide-specific PLCs. PLCs comprise
different structural modules and from recent research, it is now possible to assign specific functions to them. During the normal catalytic cycle, PLCs appear to bind stably to membrane interfaces, as a pre-requisite for efficient substrate hydrolysis. At least three of the protein modules which make up PLCs (the PH domain, X/Y TIM barrel and the C2 domain) are involved in the regulated interaction of PLCs with membranes, whilst other modules contribute to receptor, G protein and tyrosine kinase-mediated regulation of the hydrolytic activity.
and tertiary structures of enzymes of the PLC family reveals significant similarities between different isoforms, consistent with the similar catalytic activities they perform. There are also some notable differences, reflecting different modes of regulation. Figure 1 shows a linear representation of the domains or modules which comprise these enzymes. At the amino terminus, all PLC isozymes are thought to contain a pleckstrin homology (PH) domain [23], which may facilitate binding of the enzymes to PtdInsP2 molecules. This is based on the observations that purified recombinant PH domains from a selection of proteins are able to bind PtdInsP2 and inositol phosphate head groups [24–32]. Binding is largely an electrostatic phenomenon, in which conserved basic residues subtend phosphate groups of the inositol phosphates, with further binding energy to stabilise the interaction supplied by tyrosine, tryptophan and serine residues [29, 32]. PH domains from different proteins are able to bind inositol phosphates and/or PtdInsP2 with different affinities, and the PH domain of PLCd1 has a notably high affinity for both PtdInsP2 and its head group, Ins(1,4,5)P3 [30–32]. The distinctive specificity of the PLCdl PH domain appears to be due to several influences, including the more buried binding of the inositol phosphate and the fact that the loop between the third and fourth b-sheets interacts with the ligand, contributing to the 12 hydrogen bonds formed, compared to only seven in the spectrin PH domain [29, 32]. The PH domain of PLCd1 appears to function physiologically by targeting the enzyme to membranes and also in a putative feedback inhibitory loop since the competitive binding of Ins (1,4,5)P3 inhibits PLCd1 activity [33–35]. Whether the PH domains of other PLC isoforms have similar binding specificities or physiological functions have not yet been determined. Signal transducing PLCs are Ca11 -dependent enzymes and they appear to bind Ca 11 ions at distinct regions of the
molecule. The Ca11 -dependency is accounted for by a Ca11 ion subtended within the active site, which is possibly structural in nature, and which is involved in stabilising the transition state of the enzyme-substrate complex during catalysis (a pentavalent phosphoryl PtdInsP2-enzyme complex) [1]. However, PLCs also contain four repeated EF hand domains which are commonly thought to be Ca11 binding sites. The role of this module in PLC-mediated catalysis is not clear although deletion studies demonstrated that it is essential for activity [36]. The two most highly-conserved domains which occur in all the mammalian PLC isozymes are the X and Y boxes which are known to be essential for catalytic activity. The crystal structure of PLCd reveals that these form a distorted TIM barrel at the level of tertiary structure. This module is so-called, based upon the triose phosphate isomerase enzyme [37] and it is now apparent that many unrelated proteins possess similar structures [e.g. 38]. Of greatest interest to the PLCs is the fact that B. cereus phosphatidylinositolspecific PLC also forms a TIM barrel [39], whereas bacterial PtdCho-specific PLC does not [40], raising the intriguing possibility that this tertiary structural motif is particularly well-suited to the substrates which it can accommodate and the reactions which it supports. (Further details of the catalytic mechanism are discussed below). The final domain common to all eukaryotic PLCs is a C2 domain which forms interfaces with the catalytic domain and the EF hand module. C2 domains were originally described as Ca11-dependent phospholipid binding domains in protein kinase C, although different C2 domains have been described in synaptotagmins, termed A and B, which bind lipid in Ca11 -dependent and -independent manners, respectively [41, 42]. It is possible that this represents a third lipid-binding region of PLCs, although its Ca11 dependency is not clear.
Phospholipase C
PLCb and PLCg also contain distinct protein domains which are peculiar to themselves within this family. C-terminal to the C2 domain, PLCb contains a region which is essential for interaction with a subunits of Gq-related proteins [43, 44] and a further region apparently required for association with the particulate fraction of cell lysates [44]. The X and Y boxes of PLCg are separated by a complex region which contains a split PH domain, two SH2 domains and one SH3 domain [10, 45, 46]. The SH2 and SH3 domains reflect the importance of this enzyme in growth factor signalling. Presumably extensive re-folding of this enzyme brings the split TIM barrel and split PH domain together for complete function. The roles of these domains in receptor-mediated regulation of PLCs is discussed below.
Receptor-Catalysed Activation of PLCb and PLCg The protein-protein interactions which allow activated receptors to stimulate PLC activity have been extensively studied and reviewed elsewhere [10, 47–49]. b-isoforms of PLC are all capable of being activated by a-subunits of the Gq family of heterotrimeric G proteins. In addition, PLCb2 and PLCb3 are prominently activated by G protein bg subunits and are subject to inhibition via protein kinase A-mediated phosphorylation [50]. Initially it was thought that PLCs would be activated by G proteins in a manner exactly analagous to the adenylyl cyclases. However, it has been established that the b1-isoform of PLC acts as a GTPase-activating protein (GAP) for the cognitive Ga subunits of the Gq family which activates it [51]. PLCb1 (and presumably other PLCb isoforms) therefore acts to terminate the signals which stimulate it and overall enzyme activity is derived from a balance between receptor-catalysed GTP/GDP exchange on Gqa and PLCb-induced GTP hydrolysis by Gqa. Indeed, the rapid inositol phosphate and hence Ca11 responses to G proteincoupled agonists is apparently due to these antagonising influences [52]. The rate of activation of G protein-coupled effectors has traditionally been thought to be limited by the rate of association of agonist-occupied receptor with the G protein. In the case of PLCbs, it now appears that a second G protein cycle can be proposed, specific for effectors which act as GAPs, based on in vitro reconstitution experiments of PLCb, Gq and m1 muscarinic receptor [52]. The rate of this hypothetical G protein cycle is limited by receptor-catalysed GTP/GDP exchange on Gqa. Rather than the receptor and G protein dissociating every cycle, as viewed for Gsa-dependent regulation of adenylyl cyclase, the rate of GTP hydrolysis by Gqa stimulated by PLCb is so fast that the receptor-Gqa complex may not dissociate before the G protein decays to the GDP-liganded state. Thus, the complex of receptor-Gqa-PLCb may be characterised by an apparent stability. Rapid responsiveness of PLCbs to receptor agonists is intrinsic to this cycle, accounting for the observed kinetics of inositol phosphate and Ca11 signals. By contrast, although it is clear that PLCb isoforms are activated by G protein bg subunits, less is known about such
331
regulatory influences on PtdInsP2 hydrolysis. For example, the factors which determine the efficacy of bg subunits and the specificity with respect to the makeup of the bg complex have not been clearly determined [53]. PLCg isoforms, by contrast, are activated by receptors which have integral or associated tyrosine kinase activity. This process involves SH2 domain-mediated binding of PLCg1 to tyrosine phosphorylated target molecules (such as autophosphorylated tyrosine kinase receptors) and subsequent tyrosine phosphorylation of PLCg1 itself (tyrosines 771, 783 and 1254). Although it is not yet clear how this mechanism leads to enhanced PLC activity, phosphorylation of PLCg1 has been reported to facilitate hydrolysis of profilin-bound PtdInsP2 [54–57]. The SH3 domain of PLCg1 has also been proposed to participate in the regulatory mechanism by targeting the activated enzyme to the cytoskeleton [58]. Although it is clear that PLCg isoenzymes are key elements in growth factor signalling and subsequent cellular responses [59, 60], progress towards developing a more detailed understanding of their mode of regulation has been surprisingly slow.
The Catalytic Mechanisms of Mammalian PLCs SUBSTRATES OF PLCS MEMBRANES OR MONOMERIC
Inositol phospholipid-specific PLCs can hydrolyse PtdIns, PtdIns 4P and PtdIns (4,5) P2 in vitro when these substrates are presented in vesicular form, as mixed micelles with anionic and non-ionic detergents or in monolayers, but PtdInsP2 is the preferred substrate, particularly under physiologically-relevant conditions. Interestingly, inositol phospholipids in which the 3-OH of the inositol ring is phosphorylated, the products of PI 3-kinase activity, appear not to be substrates for these enzymes [61]. In vivo, PtdInsP2 resides mainly in biological membranes and physiological concentrations of monomeric lipid dissolved in the cytosol are probably negligible. Recent experiments have demonstrated that G proteinstimulated PLC activity in permeabilised cells is greatly enhanced in the presence of an additional cytosolic protein, the PtdIns transfer protein (PITP) [62, 63]. Since this enhancement occurred apparently in the absence of any detectable effect on the level of cellular PtdInsP2, it was proposed that the PITP functions as a cofactor, perhaps presenting PtdIns to lipid kinases and PLC in a coupled system. An alternative interpretation of these results is that PITP transfers PtdIns to replenish a small pool of this lipid at the plasma membrane. In this model, the phosphorylation of transferred PtdIns to PtdInsP2 and the latter’s subsequent hydrolysis would be independent of PITP. The failure to observe enhanced PtdInsP2 levels in the presence of PITP could be explained if the relevant pool was a small proportion of total cellular PtdInsP2. In support of the latter model, reconstitution of homogeneous PLCb1 with aq and the muscarinic m1 receptor in mixed phospholipid vesicles achieved up to 100-fold activation of PLC in response to agonist and GTP, compatible with estimates of the extent of LIPIDS?
332
S. R. James and C. P. Downes
as distinct steps in the case of PLCs. Evidence supporting each aspect of the model is presented below. FIGURE 2. A potential reaction scheme for signal-transducing
PLCs. In a similar fashion to other lipid-metabolising enzymes, PLCs form multiple contacts with membrane substrates, which may occur in an ordered-sequential manner, which contributes to the efficiency of second messenger formation. PLC binds to lipid interfaces forming a stable anchorage at the membrane. This may occur at a non-catalytic site of the enzyme, presumably involving the PH domain, which accounts for the PtdInsP2-directed binding of PLCs to lipid vesicles. PLC which is anchored at the membrane (PLCb) then penetrates the membrane (becoming PLCb*) as a prelude to binding a second PtdInsP2 molecule, in the active site, (PLCb*-PtdInsP2). These two steps are analagous to the ‘‘tether and fix’’ mechanism proposed by Essen et al [1]. The second molecule of PtdInsP2 is then hydrolysed to the second messenger products InsP3 and DG. Remaining stably attached to the membrane, PLC can subsequently bind other PtdInsP2 molecules at its active site, thus catalysing production of multiple second messenger molecules before it detaches from the membrane and re-dissolves in the cytosol. The rate of binding of PLC to the interface is dictated by the association and dissociation rate constants (ka and kd), the rate of penetration by the penetration and release rate constants kpen and krel, formation of active site-occupied PLCb* is determined by the binding rates k1 and k21, and the rate of hydrolysis of PtdInsP2 is shown as kcat.
PLC activation in vivo [52]. Therefore, although PITP may be instrumental in supplying PtdIns to the plasma membrane, an essential role for it as a cofactor in PLC activity is unlikely and PLC isoforms probably act on substrate molecules located in the membrane structures to which they bind. The above arguments lead us to believe that PLCs must be able to interact with complex cellular membranes in a stable fashion, bind substrate molecules and catalyse the production of second messengers very rapidly in response to cell stimulation. Different in vitro assays have been utilised to investigate the mechanisms by which PLCs interact with aggregated substrates and it is becoming clear that these are complex processes, likely to involve combinations of the protein modules comprising PLCs which were described above. Mechanistic Aspects of Catalysis by PLCs The following discussion centers on the kinetic model depicted in Figure 2 which is based on the results of classical studies of another class of phospholipase, the phospholipase A2s, which exhibit distinct interfacial binding and catalytic steps, since it is apparent from several lines of evidence that PLCs behave in a similar fashion. In this model, the first step involves membrane association of the enzyme, represented by the equilibrium PLC* )PLGb, which is followed by membrane penetration of the enzyme active site in order to engage the scissile bond of the substrate (PLCb* )PLCb*). In the case of enzymes such as PLA2, the interfacial binding and penetration steps are usually regarded as being synonymous, but there are compelling reasons for regarding these
INTERFACIAL BINDING. As noted above, many interfacial enzymes, particularly those involved in lipolysis, are thought to anchor themselves to the surface of lipid substrates and act processively, scooting over the surface of the aggregate performing multiple catalytic cycles [64]. This has been shown for secretory [65, 66] and cytosolic PLA2 [67– 69], yeast PtdIns 4-kinase [70], PtdCho-specific PLC [71] and PI 3-kinase [72]. Carefully-designed assays which allow Michaelis-Menten formalised analysis of data have shown that PLCb1, PLCb2, PLCb from turkey erythrocytes (PLCbT) [73], PLCg [74] and PLCd [34] form multiple contacts with lipid interfaces, leading to processive PtdInsP2 hydrolysis. This is based partly on the fact that PLC activity is dependent both on the bulk concentration of PtdInsP2 (the number of moles per unit volume) and the mole fraction of PtdInsP2 (the percentage of the surface area comprising PtdInsP2), which are thought to regulate enzyme binding to the interface (ka in Figure 2) and within the interface (kpen 1 k1 in Figure 2), respectively. A variety of vesicle-binding studies, together with the kinetic analyses noted above, indicate that PLCb1, b2, bT and PLCdl exhibit interfacial binding which is PtdInsP2dependent and probably distinct from interactions at the active site [73, 75, 76]. This is a significant issue because lipolytic enzymes have been described adhering to one of two interfacial binding models: the surface dilution model and the dual substrate model [64–66]. In the former, the enzyme recognises and binds to a lipid interface, regardless of its composition, whereas in the latter, enzyme binding is directed towards specific molecules, often the substrate. Thus it appears that PLCs conform with the dual substrate model. A consequence of this would be that the interaction of PLC with the plasma membrane in a stimulated cell directs the enzyme to a PtdInsP2-rich domain, permitting a rapid increase in InsP 3 and DG production. THE ROLE OF THE AMINO TERMINAL PH DOMAIN OF
When the amino terminus of PLCd was removed by proteolysis, the dependence of PtdInsP2 hydrolysis on the mole fraction of PtdInsP2 was lost, suggesting that this region of the molecule (containing the PH domain) is instrumental in anchoring the enzyme to the interface [34]. It is possible that this region of all PLCs is important for stable interfacial binding of the enzymes, requisite for efficient PtdInsP2 hydrolysis, a function which is compatible with the ligand-binding specificities of PH domains. The significance of the PH domain of PLCd to its function in situ was recently demonstrated by micro injection of recombinant full length and truncated PLCd, into MDCK cells [35]. Truncated enzyme lacking the amino terminal PH domain showed altered subcellular distribution with greatly-diminished association with the plasma membrane. It is evident therefore that the PH domain is crucial to the normal intracellular function of PLCd, possibly targeting PLCS IN MEMBRANE ANCHORING.
Phospholipase C
the enzyme to membrane domains which are enriched in its substrate. Such a role of the PH domain in PLC function may also contribute to the observed substrate specificities of these enzymes (PtdInsP2 $ PtdInsP .. PtdIns [77, 78]. Although some PH domains have been shown to bind PtdInsP [26, 31], the affinity of the interaction is between 5–20 fold lower than that for PtdInsP2. These interactions may therefore be of questionable physiological relevance on their own, unless localised concentrations of PtdInsP in the plasma membrane are high. Alternatively, it is possible that they might act additively with other ligands to bind proteins at particular subcellular locations. It seems likely that PLCs bind only to PtdInsP2 by virtue of their PH domains with sufficient affinity for the enzyme to remain localised at the plasma membrane long enough to hydrolyse multiple molecules of substrate lipid. PtdInsP2 is able to bind to other proteins, some of which contain PH domains, and is thought to play a key role in regulating diverse biochemical events such as cytoskeletal activity, ARF-dependent phospholipase D activity and secretory mechanisms (see ref. 10 and references therein). Typical cellular PtdInsP2 concentrations are of the order of 100 mM whilst concentrations in the plasma membrane are likely to be greater. Therefore, although a variety of other proteins bind to PtdInsP2 they are unlikely to modulate the concentration of this lipid sufficiently to affect PLC binding and subsequent activity. EFFECTS OF MEMBRANE COMPOSITION ON PLC ACTIVITY. Factors which modulate PLC binding to lipid inter-
faces are likely to be important regulatory influences on InsP3 and DG production in stimulated cells. Binding of PLCs to membranes is in large part an electrostatic event, with PLC binding directed to acidic phospholipids (see above). The binding of inositol phosphates to PH domains and possibly also to X boxes (see below) of PLCs is stabilised by the electrostatic interplay between strategicallypositioned lysine and arginine residues and phosphate groups. Disruption of these interactions can be envisaged to have marked effects on PtdInsP2 hydrolysis. One way such disruption could occur would be by reductions in the surface potential of lipid interfaces by inclusion of zwitterionic lipids such as PtdCho and exclusion of acidic lipids. The binding and activity of PLCd and PLCb in lipid vesicles and monolayers has been shown to be markedly reduced by reductions in the PtdSer content of the surfaces [76, 79, S. R. James, unpublished data]. The affinity of binding of PLCbT to lipid vesicles was reduced by an order of magnitude when PtdSer was omitted [73]. Thus localised regions of the plasma membrane of reduced potential are unlikely to support PLC-catalysed PtdInsP2 hydrolysis. Other factors may also affect the ability of PLC to hydrolyse PtdInsP2. In mixed micellar and vesicular PLC assays, PtdInsP2 hydrolysis ceases when less than 30% PtdInsP2 has been hydrolysed at physiological pH, which we concluded was possibly due to DG-induced changes in the lipid aggregates [80]. It may be that DG causes sequestration of PtdInsP2 from PLC or alters surface properties of the lipid interface such that InsP 3 generation is necessarily restricted.
333
The binding of a peptide corresponding to the basic region of MARCKS to phospholipid vesicles was recently reported to induce formation of lateral domains enriched in acidic phospholipids [81]. Considering the greater affinity PLCs exhibit for lipid interfaces containing significant amounts of acidic lipids, such laterally-enriched lipid domains are likely to be important in PLC activity. Indeed, PLCs may be associated with such domains specifically, and may even facilitate their formation in binding to their substrates. Relocation of PLC molecules from regions of the plasma membrane low in PtdInsP2 to regions which are relatively enriched in PtdInsP2 may markedly increase PLCcatalysed InsP3 and DG production and this may form part of the activation process in a hormone or growth factor-induced response. MEMBRANE PENETRATION BY PLCs. Although it was perhaps not necessary to consider PLCs actually penetrating cell membranes to perform catalysis in a manner similar to PLA2s (whose scissile ester bond is arguably deeper in the hydrophobic regions of the membrane), work with phospholipid monolayers suggests that PLCs penetrate the lipid surface when hydrolysing PtdInsP2 [77, 79, 82, 83 S. R. James unpublished observations]. This step corresponds to the equilibrium PLCb* )PLCb* in Figure 2. Increasing monolayer surface pressure is accompanied by reduced InsP3 and DG production, by PLCb1, PLCb2, PLCbT, PLCg1 and PLCd1, suggesting that PLCs must do more work to penetrate the monolayer and bind PtdInsP2 under conditions of close lateral packing of phospholipid molecules. Rebecchi, McLaughlin and colleagues calculated that approximately 1 nm2/surface area of PLCd penetrates the monolayer during catalysis [79]. This estimate appears to have been borne out rather nicely in the recent description of the crystal structure of PLCd [1]. The hydrophobic rim of the active site of PLCd was proposed to be that part of the molecule involved in penetrating the membrane to interact with the more hydrophobic regions of PtdInsP2. Such behaviour presumably stabilises and retains PtdInsP2 in the active site suitably orientated for hydrolysis, and dictates specificity of binding for intact lipid rather than simply the InsP3 head group. Considering the conservation of the ridge between each PLC isoform, it is possible that all enzymes will penetrate lipid surfaces with at least this region of the molecule (and maybe other regions in PLCb and PLCg) and perform similar amounts of work. It is possible that this process is also a target for the influences of G proteins and tyrosine phosphorylation in activation of PLCb and PLCg, manifested as an increase in the active site binding of PtdInsP2 (reduced interfacial Km, or increased k1 in Figure 2). The active site of PLCs forms a depression in the TIM barrel into which the polar head group of the substrate lipid fits, whilst the hydrophobic regions may interact with a ridge of amino acids which form part of the rim of the active site. Detailed comparison of the X boxes of many eukaryotic PLCs has revealed a consensus of basic amino acids (KxxxKxKK) which would be well-suited to binding acidic
334
S. R. James and C. P. Downes
inositol phosphates. Experiments with peptides based on this sequence showed that they stimulated basal PLCb2 activity [84–86] leading to the hypothesis that this region of the active site presents the PtdInsP2 molecule to the hydrolysis reaction centre, facilitating InsP 3 production. The crystal structure of PLCd has shown that two of these lysine residues (K438 and K440) form interactions with the 4- and 5-phosphates of the lipid head group, respectively [1], contributing to the stable binding of the substrate. PTDINS P2 HYDROLYSIS. Once PLC has successfully penetrated the membrane and a PtdInsP2 molecule is located at the active site, cleavage of the phosphodiester bond appears to be achieved by a pseudo charge relay catalytic triad akin to that of triglyceride lipases [87]. Thus, in a general acidgeneral base mechanism, a particular residue (possibly His 311 in PLCd (His 331 in rat PLCb1, His 327 in human PLCb2 and His 335 in human PLCg1)) held in place by a serine residue, deprotonates the 2-OH of the inositol ring, which than acts as a nucleophile, attacking the phosphodiester bond. His 356 (in PLCd) acts as an acid, protonating the DG leaving group. In a reverse direction, wherein His 356 is the base and His 311 the acid, the 1,2 cyclic inositol phosphate intermediate is linearised to form InsP 3. A similar mechanism has been proposed for B. cereus PLC [39], and for each enzyme, the third member of the catalytic triad is the substrate itself.
Implications for Activation of PLCs The reaction which PLCs catalyse, namely production of InsP3 and DG from PtdInsP2 in response to a stimulus, is clearly not a simple one. Although the reaction occurs in the two dimensions of the cell membrane rather than the three dimensions of bulk solution, this has served to complicate matters, because the enzyme must do work (penetration) to engage the scissile bond of its substrate and this can be hampered by the characteristics of the lipid interface. Measurements of the specific activities of purified PLCs give high values, even in the absence of physiological activators such as G proteins, which led to the hypothesis that PLC activity must be restrained in situ to avoid uncontrolled hydrolysis of all cellular PtdInsP2. The existence of an unidentified inhibitor protein of PLCs was hypothesised to account for the low basal rates of PLC activity in unstimulated tissues [9]. However, it is now clear that biological membranes and complex lipid substrates used in in vitro assays are not readily accessible to PLCs, but rather represent a significant barrier by virtue of their innate characteristics (e.g., surface pressure, surface potential, degree of hydration and lipid composition) which PLC isoforms must overcome to produce an inositol phosphate response. However, PLCs are multi-modular proteins, well-suited to the reactions they perform and the problems imposed by cellular membranes. Although the modes of regulation of PLCs are quite distinct (GTP-binding proteins and tyrosine phosphorylation), they are nevertheless likely to enhance catalysis by inducing similar structural and functional effects at the active site. Within the models of membrane-binding alluded
FIGURE 3. Proteins which activate PLCs may act at several dif-
ferent parts of the enzyme’s catalytic cycle. Regulatory influences on PLC may serve to enhance binding and penetration of the enzyme to the membrane (increasing ka/kpen or decreasing kd/ krel), binding of PtdInsP2 at the active site (increasing k1 or reducing k2 1) or may increase the rate of hydrolysis of PtdInsP2. G protein and tyrosine kinase-mediated effects at any or all of these parts of the reaction would be predicted to increase the catalytic rate of PLC. In the case of PLCd, InsP3 has been demonstrated to bind to the enzyme and inhibit PtdInsP2 hydrolysis [33], presumably by competing with PtdInsP2 at the amino terminus of the enzyme, leading to reduced ka .
to above, it is easy to see potential sites for activation by G proteins and tyrosine phosphorylation (Figure 3). Such activation influences may operate at the level of ka/kpen (or kd/ krel) enhancing the anchorage of PLCs to the membrane, at the level of k1 increasing the rate of binding of PtdInsP2 at the active site, or at the level of kcat, increasing the rate of hydrolysis of the phosphodiester bond. Indeed, tyrosine phosphorylation has been reported to reduce the equilibrium dissociation constant of PLCg binding to PtdInsP2-detergent mixed micelles by 8 fold (Ks, which equals kd/ka) [88], thereby enhancing enzyme association with the lipid interface. Such effects may be achieved by conformational changes in PLCs induced by the activation event, such as relocation of the amino terminus or the TIM barrel appropriately for binding PtdInsP2. Such thoughts are purely speculative without knowledge of the structures of these enzymes themselves, but it seems reasonable to assume that marked conformational alterations of these proteins will occur as they become activated and interact with the membrane. CONCLUSION Members of the family of phosphoinositide-specific PLCs have been characterised in great detail in recent years, and much is known about the molecular events which lead to the activation of two groups of isoenzymes within the extended family. Although the regulation of the three subfamilies is distinct, which reflects the wide range of signalling events in which phosphoinositide metabolism is involved, a core of common characteristics is being established between all isoforms. The interaction of proteins with cellular
Phospholipase C
membranes is a significant biochemical obstacle to be overcome in mounting a response to a stimulus, and it appears that phosphoinositide-specific PLCs utilise similar mechanisms with distinct embellishments to bind to their substrates. The result is that enzymes in this family are welladapted to the catalytic mechanism they perform with great scope for regulation, allowing a finely-controlled intracellular signal transduction cascade to be activated. Many questions remain, however, including which intracellular influences regulate subcellular localisation of PLCs; why PLCs which demonstrate relatively high affinities for PtdInsP2 are not constantly bound at the plasma membrane; which parts of the catalytic mechanism are affected by G protein subunits and tyrosine phosphorylation; and the structural basis for activation of PLCb and PLCg by receptor-regulated G proteins and tyrosine kinases. Work in the authors’ laboratory was supported by the B. B.S.R.C.. We are grateful to Drs. Andrew Paterson, T. Kendall Harden and Rudy A. Demel for their collaborations, and Ian Batty and Richard Currie for their help with the manuscript.
References 1. Essen L. O., Perisic O., Cheung R., Katan M. and Williams R. L. (1996) Nature 380, 595–602. 2. Meldrum E., Parker P. J. and Carozzi A. (1991) Biochim. Biophys. Acta 1092, 49–71. 3. Cockroft S. and Thomas G. M. H. (1992) Biochem. J. 288, 1–14. 4. Lee C. W., Park D. J., Lee K.-H., Kim C. G. and Rhee S. G. (1993) J. Biol. Chem. 268, 21318–21327. 5. Ferreira P. A., Shortridge R. D. and Park W. L. (1993). Proc. Natl. Acad. Sci. USA 90, 6042–6046. 6. Bahk Y. Y., Lee Y. H., Lee T. G., Seo J., Ryu S. H. and Suh P.-G. (1994) J. Biol. Chem. 269, 8240–8245. 7. Lee S. B. and Rhee S. G. (1996) J. Biol. Chem. 271, 25–31. 8. Liu N., Fukami K., Yu H. and Takenawa T. (1996) J. Biol. Chem. 271, 355–360. 9. Rhee S. G., Suh P.-G., Ryu S.-H. and Lee S. Y. (1989) Science 244, 546–550. 10. Lee S. B. and Rhee S. G. (1995) Curr. Op. Cell Biol. 7, 183– 189. 11. Rhee S. G., and Choi K. D. (1992) J. Biol. Chem. 267, 12393– 12396. 12. Rhee S. G. (1991) Trends Biochem. Sci. 16, 297–301. 13. Homma, Y. and Emori, Y. (1995) EMBO J. 14, 286–291. 14. Blank J. L., Ross A. H. and Exton J. H. (1991) J. Biol. Chem. 266, 18206–18216. 15. Taylor S. J., Chae H. Z., Rhee S. G. and Exton J. H. (1991) Nature 350, 516–518. 16. Smrcka A. V., Hepler J. R., Brown K. O. and Sternweis P. C. (1991) Science 251, 804–807. 17. Wu D., Lee C. H., Rhee S. G. and Simon M. I. (1992) J. Biol. Chem. 267, 1811–1817. 18. Camps M., Hou C., Sidiropoulos D., Stock J. B., Jakobs K. H. and Gierschik P. (1992) Eur. J. Biochem. 206, 821–831. 19. Park D., Jhon D. Y., Lee C.-W., Lee K-H. and Rhee S. G. (1993) J. Biol. Chem. 268, 4573–4576. 20. Camps M., Carozzi A., Schnable P., Scheer A., Parker P. J. and Gierschik P. (1992) Nature 360, 684–686. 21. Jiang H., Wu D. and Simon M. I. (1994) J. Biol. Chem. 269, 7593–7596. 22. Lee C-W., Lee K-H., Lee S. B., Park D. and Rhee S. G. (1994) J. Biol. Chem. 269, 25335–25338.
335 23. Parker P. J., Hemmings B. A. and Gierschik P. (1994) Trends Biochem. Sci. 19, 54–55. 24. Touhara K., Koch W. J., Hawes B. E. and Lefkowitz R. J. (1995) J. Biol. Chem. 270, 17000–17005. 25. Harlan J. E., Hajduk P. J., Yoon H. S. and Fesik S. W. (1994) Nature 371, 168–170. 26. Harlan J. E., Yoon H. S., Hajduk P. J. and Fesik S. W. (1995) Biochemistry 34, 9859–9564. 27. Saraste M. and Hyvonen M. (1995) Curr. Op. Struct. Biol. 5, 403–408. 28. Lemmon M. A., Ferguson K. M. and Schlessinger J. (1996) Cell 85, 621–624. 29. Hyvonen M., Macias M. J., Nilges M., Oschkinat H., Saraste M. and Wilmanns M. (1995) EMBO J. 14, 4676–4685. 30. Lemmon M. A., Ferguson K. M., O’Brien R., Sigler P. B. and Schlessinger J. (1995) Proc. Natl. Acad. Sci. USA 92, 10472– 10476. 31. Garcia P., Gupta R., Shah S., Morris A. J., Rudge S. A., Scarlata S., Petrova V., McLaughlin S. and Rebecchi M. J. (1995) Biochemistry 34, 16228–16234. 32. Ferguson K. M., Lemmon M. A., Schlessinger J. and Sigler P. B. (1995) Cell 83, 1037–1046. 33. Cifuentes M. E., Delaney T. and Rebecchi M. J. (1994) J. Biol. Chem. 269, 1945–1948. 34. Cifuentes M. E., Honkanen L. and Rebecchi M. J. (1993) J. Biol. Chem. 268, 11586–11592. 35. Paterson H. F., Savopoulos J. W., Perisic O., Cheung R., Ellis M. V., Williams R. L. and Katan M. (1995) Biochem. J. 312, 661–666. 36. Ellis M. V., Carne A. and Katan M. (1993) Eur. J. Biochem. 213, 339–347. 37. Banner D. W., Bloomer A. C., Petsko G. A., Phillips D. C., Pogson C. I., Wilson I. A., Corran P. H., Furth A. J., Milman J. D., Offord R. E. Priddle J. D. and Warley S. C. (1975) Nature 255, 609–614. 38. Baldwin T. O., Christopher J. A., Raushel F. M., Sinclair J. F., Ziegler M. M., Fisher A. J. and Rayment I. (1995) Curr. Op. Struct. Biol. 5, 798–809. 39. Heinz D. W., Ryan M., Bullock T. L. and Griffith O. H. (1995) EMBO J. 14, 3855–3863. 40. Hansen S., Hough E., Svensson L. A., Wong Y. L. and Martin S. F. (1993) J. Mol. Biol. 234, 179–187. 41. Fukuda M., Aruga J., Niinobe M., Aimoto S. and Mikoshiba K. (1994) J. Bio. Chem. 269, 29206–29211. 42. Fukuda M., Kojima T., Aruga J., Niinobe M. and Mikoshiba K. (1995) J. Biol. Chem. 270, 26523–26527. 43. Park D., Jhon D-Y., Lee C-W., Ryu S. H. and Rhee S. G. (1993) J. Biol. Chem. 268, 3710–3714. 44. Wu D., Jiang H., Katz A. and Simon M. I. (1993) J. Biol. Chem 268, 3704–3709. 45. Suh P. G., Ryu S. H., Moon K. H., Suh H. W. and Rhee S. G. (1988) Proc. Natl. Acad. Sci. USA 85, 5419–5423. 46. Stahl M. L., Ferenz C. R., Kelleher K. L., Kriz R. W. and Knopf J. L. (1988) Nature 332, 269–272. 47. Noh D. Y., Shin S. H. and Rhee S. G. (1995) Biochim. Biophys. Acta. 1242, 99–113. 48. Sternweis P. C. and Smrcka A. V. (1993) CIBA Found. Symp. 176, 96–106. 49. Exton J. H. (1993) Adv. Second Mess. Phosphoprotein Res. (eds. Brown B. L. and Dobson P. R. M.) 28, 65–72, Raven Press Ltd., New York. 50. Liu M. and Simon M. I. (1996) Nature 382, 83–87. 51. Berstein G., Blank J. L., Jhon D-Y., Exton J. H., Rhee S. G. and Ross E. M. (1992) Cell 70, 411–418. 52. Biddlecome G. H., Bernstein G. and Ross E. M. (1996) J. Biol. Chem 271 7999–8007. 53. Sternweis P. C. (1994) Curr. Op. Cell Biol. 6, 198–203.
336 54. Nishibe S., Wahl M. I., Hernandez-Sotomayor S. M. T., Tonks N. K., Rhee S. G. and Carpenter G. (1990) Science 250, 1253–1256. 55. Goldschmidt-Clermont P., Kim J. W., Machecky L. M., Rhee S. G. and Pollard T. D. (1991) Science 251, 1231–1233. 56. Goldschmidt-Clermont P. J., Machesky L. M., Baldassare J. J. and Pollard T. D. (1990) Science 247, 1575–1578. 57. Kim H. K., Kim J. W., Zilberstein A., Margolis B., Kim J. G., Schlessinger J. and Rhee S. G. (1991) Cell 65, 435–441. 58. Rhee S. G. (1991) Trends Biochem. Sci. 16, 297–301. 59. Valius M. and Kazlauskas A. (1993) Cell 73, 321–334. 60. Kundra V., Escobedo J. A., Kazlauskas A., Kim H. K., Rhee S. G., Williams L. T. and Zetter B. R. (1994) Nature 367, 474–476. 61. Serunian L., Haber M. T., Fukui T., Kim J. W., Rhee S. G., Lowenstein J. M. and Cantley L. C. (1989) J. Biol. Chem. 264, 17809–17815. 62. Cunningham E., Thomas G. M. H., Ball A., Hilse I. and Cockroft S. (1995) Curr. Biol. 5, 775–783. 63. Thomas G. M. H., Cunningham E., Fensome A., Ball A., Totty N. F., Truong O. Hsuan J. J. and Cockroft S. (1993) Cell 74, 919–928. 64. Carman G. M., Deems R. A. and Dennis E. A. (1995) J. Biol. Chem. 270 18711–18714. 65. Deems R. A., Eaton B. R. and Dennis E. A. (1975) J. Biol. Chem. 250, 9013–9020. 66. Hendrickson H. S. and Dennis E. A. (1984). J. Biol. Chem. 259, 5734–5739. 67. Hanel A. M., Schutel S. and Gelb M. (1993) Biochemistry 32, 5949–5958. 68. Creaney A., Masters D. J., Needham M. B., Gordon R. D., Mott R. and Wilton D. G. (1995) Biochem. J. 306, 857–864. 69. Kucera G. L., Miller C., Sisson P. J., Wilcox R. W., Wiemer Z. and Waite M. (1988) J. Biol. Chem. 263, 12964–12969. 70. Buxeda R. J., Nickels J. T., Belunis C. J. and Carman G. M. (1991) J. Biol. Chem. 266, 13859–13865. 71. Eaton B. R. and Dennis E. A. (1976) Arch. Biochem. Biophys. 176, 604–609.
S. R. James and C. P. Downes 72. Barnett S. F., Ledder M., Stirdivant S. M., Ahern J., Conroy R. R. and Heimbrook D. C. (1995) Biochemistry 34, 14254– 14262. 73. James S. R., Paterson A., Harden T. K. and Downes C. P. (1995) J. Biol. Chem. 270, 11872–11881. 74. Wahl M. I., Jones G. A., Nishibe S., Rhee S. G. and Carpenter G. (1992) J. biol. Chem. 267, 10447–10456. 75. Pawelczyk T. and Lowenstein J. M. (1993) Biochem. J. 291, 693–696. 76. Rebecchi M., Peterson A. and McLaughlin S. (1992) Biochemistry 31, 12742–12747. 77. James S. R., Demel R. A. and Downes C. P. (1994) Biochem. J. 298, 499–506. 78. Park D., Jhon D-Y., Kriz R., Knopf J. and Rhee S. G. (1992) J. Biol. Chem. 267, 16048–16055. 79. Rebecchi M., Boguslavsky V., Boguslavsky L. and McLaughlin S. (1992) Biochemistry 31, 12748–12753. 80. James S. R., Smith S., Paterson A., Harden T. K. and Downes C. P. (1996) Biochem. J. 314, 917–921. 81. Yang L. and Glaser M. (1995) Biochemistry 34, 1500–1506. 82. Hirasawa K., Irvine R. F. and Dawson R. M. C. (1981) Biochem. J. 193, 607–614. 83. Boguslavsky V., Rebecchi M., Morris A. J., Jhon D-Y., Rhee S. G. and McLaughlin S. (1994) Biochemistry 33, 3032–3037. 84. Simoes A. P., Schnabel P., Pipkorn R., Camps M. and Gierschik P. (1993) FEBS Lett. 331, 248–251. 85. Simoes A. P., Camps M., Schnabel P. and Gierschik P. (1995) FEBS Lett. 365, 155–158. 86. Simoes A. P., Reed J., Schnabel P., Camps M. and Gierschik P. (1995) Biochemistry 34, 5113–5119. 87. Brady L., Brzozowski A. M., Derewends Z. S., Dodson E., Dodson G., Tolky S., Turkenburg J. P., Christiansen L., Huge-Jensen B., Norskov L., Thim L. and Menge U. (1990) Nature 343, 767–770. 88. Carpenter G., Hernandez-Sotomayor T. and Jones G. (1993) Adv. Second Mess. Phosphoprotein Res. (ed. Brown B. L. and Dobson P. R. M.) 28, 179–185, Raven Press Ltd., New York.