Function and inhibition of intracellular calcium-independent phospholipase A2.

acyltransferases to form phosphatidic acid ( PA ). In mammalian cells, the phosphatidic acid can be converted to phosphatidylinositol ( PI ) or can be converted to diacyl-glycerol ( DG ), which is the precursor for phosphatidylcholine ( PC ) and phosphatidylethanolamine ( PE ), which in turn forms phosphatidyl-serine ( PS ). In contrast, in the remodeling pathway, preformed phos- phatidylinositol, phosphatidylserine, phosphatidylcholine, or phosphatidylethanolamine is acted on by iPLA 2 to produce lyso- phosphatidylinositol, lysophosphatidylserine, lysophosphatidylcholine, or lysophosphatidylethanolamine; these can be reacylated by acyltrans- ferases using fatty acyl-CoA. Adapted with permission from Ref. 22.


Function and Inhibition of Intracellular Calciumindependent Phospholipase
Our previous Minireview (1) considered the three main kinds of phospholipase A 2 (PLA 2 ) 1 : the well characterized Groups I, II, and III small Ca 2ϩ -dependent secretory phospholipase A 2 s (sPLA 2 ), the 85-kDa Group IV Ca 2ϩ -dependent cytosolic phospholipase A 2 (cPLA 2 ), and the 80-kDa Ca 2ϩ -independent cytosolic phospholipase A 2 (iPLA 2 ). In the ensuing years, it has become clear that PLA 2 represents a growing superfamily of enzymes with many additional sPLA 2 s (Groups IIC, V, and IX), further definition of the 80-kDa iPLA 2 (Group VI), and two Ca 2ϩ -independent PLA 2 s specific for platelet-activating factor (PAF) (Groups VII and VIII) (2).
All of the well studied sPLA 2 s appear to use a His-Asp catalytic mechanism and require Ca 2ϩ to be bound tightly in the active site of the enzyme. The well characterized iPLA 2 appears to require a central Ser for catalysis and of course, no Ca 2ϩ . Interestingly, the Group IV cPLA 2 does not use Ca 2ϩ for catalysis, but rather the Ca 2ϩ dependence seems to relate to a calcium lipid-binding domain (CaLB or C-2 domain) at the N-terminal end responsible for association of the enzyme with the membrane. Thus, the catalytic mechanism and active site Ser do not involve Ca 2ϩ (3)(4)(5); therefore a mechanistic distinction between the Group IV cPLA 2 and the iPLA 2 s may not be warranted at this time. This is relevant because most of the inhibitors that work on the Group IV cPLA 2 also act on the Group VI iPLA 2 (6,7). Inhibitor specificity will be discussed in the next section.
We (8) recently surveyed all of the reported Ca 2ϩ -independent PLA 2 activities. While there exists a group of lysosomal iPLA 2 s and a group of characterized ectoenzymes with broad specificity, which may actually be general lipases (8), sequenced and well characterized intracellular iPLA 2 s are limited to the 80-kDa Group VI iPLA 2 and the 29-kDa Group VIII enzyme, which is a PAF acetyl hydrolase (9). The latter hydrolyzes the acetyl chain present at the sn-2 position of PAF and perhaps acts on oxidized phospholipids as well but not on normal phospholipids carrying unoxidized long chain fatty acids at the sn-2 position (9). This enzyme and a secreted Group VII PAF acetyl hydrolase, both of which are really iPLA 2 s with a particular substrate specificity, have been considered elsewhere (2).
The Group VI 80-kDa iPLA 2 was first identified in P388D 1 macrophages (10), purified (11), further characterized (12), and then cloned and sequenced by Jones and co-workers (13) from CHO cells. The CHO iPLA 2 has been shown to represent a species variant of that present in P388D 1 macrophage-like cells, where the iPLA 2 has also been cloned and sequenced (14). The sequence of the Group VI iPLA 2 reveals the presence of eight ankyrin-like domains and the G-X-S-X-G motif commonly found in other lipases. Interestingly, no known consensus sequences for posttranslational modification, such as phosphorylation sites, are apparent in the Group VI iPLA 2 (13,14). This is compatible with the possibility that the Group VI iPLA 2 acts to remodel membrane phospholipids as a sort of housekeeping enzyme as will be discussed later.

Inhibition of iPLA 2
The functional significance of intracellular iPLA 2 can most easily be investigated using selective inhibitors. Unfortunately, no specific iPLA 2 inhibitors currently exist. As indicated above, the apparent presence of an active site Ser residue in Group VI iPLA 2 is similar to that of the Group IV cPLA 2 . Thus, the cPLA 2 inhibitors currently available, which were designed as site-directed inhibitors, all inhibit the Group VI iPLA 2 as well. These include arachidonyl trifluoromethyl ketone (6), arachidonyl tricarbonyl (6), and methyl arachidonyl fluorophosphonate (7). These three compounds contain an arachidonyl tail and function as transition-state analogues in a reversible or irreversible manner. The arachidonyl tail was intended to confer selectivity to the inhibitors and to facilitate their access to the cPLA 2 active site (15,16), as this enzyme selectively hydrolyzes arachidonate-containing phospholipids (17,18). Remarkably, even though the iPLA 2 is not AA-specific (11), these inhibitors work even better on the iPLA 2 than on the cPLA 2 (6,7). Furthermore, palmityl trifluoromethyl ketone and palmityl tricarbonyl are as good inhibitors of both the Group IV cPLA 2 and Group VI iPLA 2 as their arachidonyl analogs (6,12).
Due to the lack of selectivity of the aforementioned compounds, it is unlikely that they will find much use in defining the role of the iPLA 2 in cell function, unless the process under study is truly Ca 2ϩ -independent. Inhibition studies employing the fatty acyl trifluoromethyl ketones, tricarbonyls, or fluorophosphonates in the absence of Ca 2ϩ might selectively target the iPLA 2 , as this is the only one of the well studied cellular PLA 2 s that remains active under Ca 2ϩ -depleted conditions.
One common feature of the two best characterized intracellular iPLA 2 s, namely the Group VI enzyme present in P388D 1 macrophages (12) and CHO cells (14) and a 40-kDa iPLA 2 present in myocardial tissue and pancreatic islets (19), is their complete and irreversible inhibition by the mechanism-based inhibitor BEL. This inhibitor was first introduced as a serine protease inhibitor (20) but has been shown to be specific for iPLA 2 over Ca 2ϩ -dependent sPLA 2 s (19,21) and Group IV Ca 2ϩ -dependent cPLA 2 (21). In addition, BEL does not affect a number of enzyme activities directly involved in cellular AA metabolism (22). Thus BEL has received great attention because of its possible use as a selective iPLA 2 inhibitor in whole cell studies. As a matter of fact, much of what is currently believed to be mediated by iPLA 2 enzymes has been derived * This minireview will be reprinted in the 1997 Minireview Compendium, which will be available in December, 1997. This is the third article of six in "A Thematic Series on Phospholipases." Work in the authors' laboratory was supported by National Institutes of Health Grants HD 26171, GM 20501, and GM 51606.
from studies using BEL. Unfortunately, BEL has recently been found to inhibit another key enzyme in cellular phospholipid metabolism, the Mg 2ϩ -dependent phosphatidic acid phosphohydrolase (PAP-1) (23). The latter enzyme catalyzes the dephosphorylation of phosphatidic acid to yield 1,2-diacylglycerol (DAG), a central intermediate in glycerolipid synthesis as well as an important intracellular messenger molecule. Therefore, BEL cannot be used as a selective iPLA 2 inhibitor in whole cell studies unless it is demonstrated that the process under study is independent of variations in cellular DAG levels as well as Ca 2ϩ ; even then BEL is known to inhibit proteases and may affect other enzymes as well. The implications will be discussed in the next section.

Function of Group VI iPLA 2 in Membrane Phospholipid Remodeling
There exists in cells an ongoing deacylation/reacylation cycle of membrane phospholipids, the so-called Lands cycle, whereby a pre-existing phospholipid is cleaved by an intracellular PLA 2 to generate a 2-lysophospholipid, which in turn may be reacylated with a different fatty acid to generate a new phospholipid (22,24). This remodeling cycle constitutes the major route for incorporation of free AA into the phospholipids of cells at nanomolar levels of the free fatty acid (Fig. 1) (22, 24). The de novo route or Kennedy pathway constitutes, in addition, a second route for incorporation of AA into cellular phospholipids (Fig. 1). However, the Kennedy pathway appears to be relevant in terms of AA incorporation only when high, micromolar levels of free AA are available (25).
Thus, AA incorporation into phospholipids under normal conditions is strikingly dependent on a PLA 2 that generates the 2-lysophospholipid used in the acylation reaction. Macrophages and macrophage cell lines possess a high capacity to incorporate AA into their membrane phospholipids (26,27). This process occurs in a Ca 2ϩ -independent manner (22,26), suggesting the involvement of an iPLA 2 . Consistent with this observation, BEL inhibits AA esterification in a dose-dependent and saturatable manner, and the decrease in AA incorporation directly correlates with inhibition of both cellular iPLA 2 activity and steady-state lysophospholipid levels (22). Although the lack of BEL specificity now raises some doubt about the firmness of this conclusion, it is important to stress here that AA esterification via phospholipid remodeling is independent of variations in DAG levels. Moreover, BEL does not reduce the cellular steady-state levels of DAG. 2 Thus, the possible parallel inhibition of PAP-1 by BEL should not affect the basal rate of AA incorporation into phospholipids.
The nucleotide sequence for the Group VI murine iPLA 2 is now available (14). This has enabled us to utilize more convincing molecular biological approaches, such as antisense inhibition studies, to ascertain the role of the Group VI iPLA 2 in cellular phospholipid metabolism. Antisense inhibition of the Group VI iPLA 2 has confirmed that this enzyme does play a role in phospholipid remodeling as iPLA 2 -depleted cells show a significant reduction of their capacity to incorporate AA into membrane phospholipids. 2 Moreover, the decreased incorporation of AA into phospholipids that iPLA 2 -depleted cells manifest is not further decreased by BEL, 2 demonstrating that this compound is indeed targeting the iPLA 2 in the previous experiments (22).
Collectively, these findings suggest that the Group VI iPLA 2 is responsible for phospholipid fatty acid remodeling under resting conditions. Hence this enzyme appears to regulate the main pathway through which the cells incorporate AA and other unsaturated fatty acids into their membrane phospholipids. In addition to its obvious importance in cellular metabolism, the rate of AA incorporation into phospholipids also determines the amount of free fatty acid available under resting conditions. This is relevant because free AA availability is a limiting factor for eicosanoid biosynthesis. By regulating basal AA esterification reactions, the Group VI iPLA 2 may also play a key role in regulating the amount of prostaglandins synthesized by resting cells.
On the other hand, there is now strong evidence for the existence of different AA pools within the cells (21, 28) that can be utilized by distinct Ca 2ϩ -dependent PLA 2 s during cell activation (21). The role of CoA-independent transacylase in regulating the asymmetric distribution of AA among different phospholipid subclasses has recently become evident (24). However, as the iPLA 2 -mediated reaction precedes the action of the CoAindependent transacylase, it is possible that the iPLA 2 determines both the subcellular distribution of this fatty acid among the different cellular compartments and the relative amount of fatty acid present in each compartment before further remodeling reactions catalyzed by the CoA-independent transacylase take place (Fig. 2).

Other Proposed Functions for Intracellular iPLA 2 s
Based solely on BEL effects, the iPLA 2 has been suggested to mediate AA release in different cells stimulated with various agonists (29 -31). It is known, however, that during cell stimulation DAG levels rise appreciably, and this may influence FIG. 1. De novo and remodeling pathways for incorporation of free fatty acid into phospholipids. In the de novo pathway free fatty acid, provided to the cell or liberated by endogenous phospholipases, is incorporated via fatty acyl-CoA into glycerol phosphate (GP) or dihydroxyacetone phosphate (DHAP) and into the resulting lysophosphatidic acid (LysoPA) by fatty acyl-CoA acyltransferases to form phosphatidic acid (PA). In mammalian cells, the phosphatidic acid can be converted to phosphatidylinositol (PI) or can be converted to diacylglycerol (DG), which is the precursor for phosphatidylcholine (PC) and phosphatidylethanolamine (PE), which in turn forms phosphatidylserine (PS). In contrast, in the remodeling pathway, preformed phosphatidylinositol, phosphatidylserine, phosphatidylcholine, or phosphatidylethanolamine is acted on by iPLA 2 to produce lysophosphatidylinositol, lysophosphatidylserine, lysophosphatidylcholine, or lysophosphatidylethanolamine; these can be reacylated by acyltransferases using fatty acyl-CoA. Adapted with permission from Ref. 22. Minireview: Intracellular Calcium-independent Phospholipase A 2 16070 agonist-induced AA release, either directly (by providing DAG substrate for the release) (32) or indirectly (by activating protein kinase C) (33). As discussed above, BEL has recently been found to inhibit cellular PAP-1 activity (23). It is likely that PAP-1 plays a role in raising intracellular DAG levels during signal transduction, particularly in those settings where the de novo phospholipid biosynthetic pathway is involved (34). Thus, the reported effects of BEL on agonist-induced AA release might also be due, at least in part, to inhibition of PAP-1 in addition to the iPLA 2 .
The following example may better illustrate the uncertainty of BEL effects on agonist-induced AA mobilization. In studies with glucose-and carbachol-stimulated pancreatic islets, Ramanadham et al. (29) proposed the iPLA 2 as the major mediator of AA release on the basis of its inhibition by BEL. However, Konrad et al. (35) have suggested that the DAG lipase pathway constitutes the major route for AA release in the same system. The results by these two groups could be reconciled if BEL was inhibiting DAG generation in the work by Ramanadham et al. (29). On the other hand, it is interesting to note that in systems where agonist-stimulated AA release appears not to depend on DAG-induced protein kinase C activation, such as PAF-stimulated P388D 1 macrophages (36), BEL is ineffective in inhibiting this release (21).
Involvement of an iPLA 2 in stimulus-induced AA release has also been suggested by Lennartz and colleagues (37,38) in studies with human monocytes. During immunoglobulin G-mediated phagocytosis, human monocytes release AA in a Ca 2ϩ -independent manner (37). Consistent with the possible involvement of an iPLA 2 , AA release in this system is blocked by BEL (37). The process was later found to be dependent on protein kinase C activation (38). Thus, the BEL effects could be partly due to protein kinase C dependence on DAG, although the finding that the process takes place in the absence of Ca 2ϩ lends credence to the possible involvement of an iPLA 2 .
It is generally difficult to rationalize the involvement of a Ca 2ϩ -independent enzyme in processes such as AA release, which in most cases is strongly Ca 2ϩ -dependent. Wolf and Gross (39) have recently reported that a 40-kDa myocardial iPLA 2 associates with calmodulin in a Ca 2ϩ -dependent manner, providing a mechanism through which Ca 2ϩ may regulate a Ca 2ϩ -independent enzyme. According to these authors, myocardial iPLA 2 is inactive when associated with calmodulin. Dissociation of the complex due to decreases in the Ca 2ϩ concentration or addition of calmodulin antagonists renders the iPLA 2 ready to attack phospholipids and release AA (39). This mechanism has been proposed to mediate the cardiac cycle-dependent alterations in PLA 2 -catalyzed release of AA (39).
Before the reported association with calmodulin, the 40-kDa iPLA 2 activity from myocardial tissue and pancreatic islets was reported to associate to phosphofructokinase or an antigenically related protein (40,41). Due to their very distinct biochemical properties and molecular sizes, it appears clear that the 40-kDa iPLA 2 activity identified in myocardium and pan- Minireview: Intracellular Calcium-independent Phospholipase A 2 16071 creatic islets is different from the Group VI enzyme present in P388D 1 macrophages and CHO cells (41). However, as a common feature, the two proteins exist as catalytic complexes of about 340 kDa (11). The Group VI iPLA 2 has been shown to possess eight ankyrin repeats, which may mediate self-aggregation or interaction with other proteins (13). It is possible that the 40-kDa myocardial iPLA 2 activity possesses similar motifs that mediate its interaction with other proteins such as phosphofructokinase and calmodulin; however, the sequence of the 40-kDa protein has not yet been elucidated. Thus, its relationship with the Group VI iPLA 2 , if any, cannot be ascertained at this time. However, analogous to the Group VI iPLA 2 (11), the 40-kDa activity is stimulated by ATP in vitro (40). Recent data by our laboratory have shown that, rather than stimulating enzyme activity, ATP stabilizes and protects the iPLA 2 from inactivation during the assay; hence, higher activity is found in the presence than in the absence of ATP and other nucleotides (14). Thus there is no evidence that ATP plays a regulatory role for the Group VI iPLA 2 in vivo.

Concluding Remarks
The importance of the intracellular iPLA 2 in control of cell function has not been clearly established at present, despite the fact that iPLA 2 s have been found to exist in all cells and tissues examined. Currently, many new iPLA 2 s are being purified and characterized (8,(42)(43)(44). The widespread occurrence of iPLA 2 s suggests that this class of enzymes may play important roles in cell physiology. Fatty acid remodeling of membrane phospholipids in macrophages appears to be an event most likely mediated by intracellular iPLA 2 s. Currently, much of the data available on cellular iPLA 2 function relies on the use of inhibitors that have been shown not to be selective for this class of enzymes. However, these inhibitors may offer leads for the development of more selective agents that may help to uncover new roles for intracellular iPLA 2 s in cellular functioning.