The Charcot–Leyden crystal protein revisited—A lysopalmitoylphospholipase and more

Abstract

The Charcot–Leyden crystal protein (CLC-P), a constituent of human and not mouse eosinophils, is one of the most abundant proteins within human eosinophils. It has a propensity to form crystalline structures, Charcot–Leyden crystals, which are hallmarks in their distinctive extracellular crystalline forms as markers of eosinophilic inflammation. The functions of CLC-P within eosinophils have been uncertain. Although the action of CLC-P as a lysophospholipase has been questioned, assays of chromatographically purified CLC-P and crystal-derived CLC-P as well as studies of transfected recombinant CLC-P have consistently documented that CLC-P endogenously expresses lysophospholipase activity, releasing free palmitate from substrate lysopalmitoylphosphatidylcholine. Rather than acting solely as a hydrolytic enzyme to release palmitate from a lysolipid substrate, some other lysophospholipases function more dominantly as acylprotein thioesterases (APTs), enzymes that catalyze the removal of thioester-linked, long chain fatty acids, such as palmitate, from cysteine residues of proteins. As such APTs participate in palmitoylation, a post-translational modification that can affect membrane localization, vesicular transport, and secretion. CLC-P has attributes of an APT. Thus, whereas CLC-P expresses inherent lysophospholipase activity, like some other lysophospholipase enzymes, it likely also functions in regulating the dynamic palmitoylation cycle, including, given its dominant subplasmalemmal location, at the human eosinophil’s plasma membrane.

1 |. INTRODUCTION

The Charcot–Leyden crystal protein (CLC-P) has long been enigmatic. The Charcot–Leyden crystal (CLC), a distinctive dipyramidally shaped hexagonal crystalline structure, has been recognized since the mid-19th century.^1,2

Although the CLC-P is also present in human basophils,^3–5 human T reg cells⁶ and in IL-22 producing T cells in atopic dermatitis,⁷ CLCs are recognized hallmarks in their distinctive extracellular crystalline forms as markers of eosinophils. The associations of CLCs with eosinophilrich inflammatory conditions have been recently tabulated and the histories of CLC studies have been reviewed.^8,9 Levels of CLC-P have been measured as biomarkers of eosinophil inflammation in esophageal and airway sites.^10,11

More recently, interest in CLCs has been engendered with demonstrations of their formation within the cytoplasm of human eosinophils undergoing eosinophil extracellular trap cell death (EETosis)¹² and by the demonstration that CLCs released extracellularly may be a druggable target to mitigate their extracellular drive of Th2 inflammation¹³ CLCs can activate the NLRP3 inflammasome and elicit IL1β-driven inflammation¹⁴ and can promote neutrophilic inflammation in nasal polyposis.¹⁵ Thus, the release of CLC-P from eosinophils in its extracellular crystalline form has the potential to mediate and enhance eosinophil-derived pro-inflammatory activities.

The proteinaceous nature of CLCs was clarified in 1976, when Gleich et al. identified that CLCs were composed of a single polypeptide that was distinct from human eosinophil granule-derived major basic protein (MBP), that in 1% SDS-PAGE yielded a single band with an apparent molecular weight of 13 kD, and had a remarkable propensity to aggregate.¹⁶ It was then noted that CLC-P from resolubilized crystals eluted abnormally, not as a discrete molecular weight-related peak, but smeared over a broad range on Sephadex G-50 gel filtration,¹⁷ likely related to its “stickiness” and hydrophobicity. Thereafter, CLC-P was shown to have sequence characteristics of galectins,^17–19 and CLC-P has been designated galectin-10 (Gal-10). By proteomic analyses of peripheral blood eosinophils, the CLC-P/Gal-10 was the fifth-most abundant human eosinophil protein (after actin, a nonsecretory ribonuclease and histones).²⁰ In an earlier proteomic study of blood eosinophils in which CLC-P was localized to a cytoplasmic subcellular fraction, CLC-P was the second-most prevalent human eosinophil protein.²¹

The CLC-P is genetically encoded only in humans and higher non-human primates and not other mammals, including mice;^22–24 hence precluding its functional roles from being studied in murine models of eosinophil-associated tissue- and disease-related responses. In humans, the CLC-P is the rare protein that forms crystals in vivo.^22,25

Thus, we have in human eosinophils a quantitatively very abundant, predominantly cytoplasmic protein, a unique protein with a proclivity to crystallize as CLCs, and a protein whose functional roles within human eosinophils remain to be delineated. Moreover, evolutionarily, given CLC-P’s genetically limited expression in only human and higher primates, the question arises as to what are its specific capabilities that this highly abundant cytoplasmic protein in human eosinophils brings to human eosinophils in contrast to other nonprimate, mammalian eosinophils? Fundamentally, the functions of the CLC-P within human eosinophils and other cells that contain CLC-P remain to be ascertained.

In this review, we first consider the body of experimental evidence that bears on the recognition that the CLC-P of human eosinophils exhibits lysopalmitoylphospholipase activity. The lysophospholipase activity of the CLC-P had been reported in varied studies in the 1980s and 1990s.^18,26–34 In 1999, however, Wang and Dennis noted that in comparison with other mammalian lysophospholipases known at the time, CLC-P often lacked the much greater specific activities of other lysophospholipases and showed no homologous structural analogs with these other lysophospholipases leading to the suggestion that the low lysophospholipase activity may be due to trace contaminants of other highly active lysophospholipases.³⁵ In 2002, Ackerman et al.’s study entitled “Charcot-Leyden crystal protein (galectin-10) is not a dual function galectin with lysophospholipase activity but binds a lysophospholipase inhibitor in a novel structural fashion” concluded: “Our results definitively show that CLC protein is not one of the eosinophil’s lysophospholipases but that it does interact with eosinophil lysophospholipases and known inhibitors of this lipolytic activity.”³⁶ This study has been cited to refute earlier studies that demonstrated that CLC-P exhibited endogenous lysophospholipase activity. For this review, we critically reevaluate prior studies and provide original unpublished figures that support our published findings that bear on issues of concern raised in the 2002 report.³⁶ With passage of time and newer insights into potential roles of lysophospholipases as more than hydrolytic, palmitate releasing enzymes, we integrate prior studies into a consideration of the potential functions of the CLC-P so highly abundant in human eosinophils.

2 |. INITIAL STUDIES OF THE LYSOPHOSPHOLIPASE ACTIVITY OF CLC-P

As was the standard at the time,³⁵ the assay of lysophospholipase activity evaluated free ¹⁴C-palmitate released from ¹⁴C-lysopalmitoylphosphatidylcholine, and it was assumed that the only function of a lysophospholipase was to mediate the hydrolytic cleavage of palmitate releasing a free fatty acid reaction product.³⁵ Based on several investigators’ studies, the CLC-P’s lysophospholipase enzymatic activities have at times exhibited a low turnover number in terms of measured released free palmitate relative to substrate; and CLC-P thus was considered a weak lysophospholipase. As likewise measured by the release of free ¹⁴C-palmitate, other lysophospholipases exhibited specific activities in the micromoles/hour (h)/mg range.³⁵ In varied studies cited below, CLC-P has exhibited lysophospholipase specific activities both in the weaker nanomoles/h/mg protein magnitudes as well as greater specific activities in the micromoles/h/mg range.

Although our initial evaluation of human eosinophil lysophospholipase activities was engendered by reports of heightened lysophospholipase (then known as phospholipase B or lysolecithinase) activities in rat eosinophils,³⁷ we focused on a single protein with lysophospholipase activities that was prominent in eosinophil cytoplasm-rich fractions. Unanticipated by us at the time, CLCs had been identified previously as arising from the cytoplasm and not granules of human eosinophils.³ Based on our initial findings that lysophospholipase activity was prominent in the cytoplasm-rich fractions of eosinophil sonicates (Fig. 1), and that these cytoplasm-rich sonicates yielded on size-exclusion gel filtration chromatography a dominant peak of lysolecithinase/lysophospholipase activity (Fig. 2) when assayed with a cytoplasmic-biased neutral 7.5 pH, we focused on purifying the protein responsible within this gel-filtered, ∼28 kD peak of lysolecithinase/lysophospholipase activity. We never sought to detect all human eosinophil-derived proteins with lysophospholipase activities, such as potential granule-based lysosomal enzymes that might require enzymatic assays at different pHs or conditions. Whereas there are other lysophospholipases in human eosinophils, including a 74 kD pancreatic lipase-like lysophospholipase that was immunolocalized to eosinophil granules,³⁸ our focus was solely on the lysophospholipase expressing protein that we would thereafter identify as the CLC-P.

FIGURE 1. Subcellular fractionation of eosinophil “lysolecithinase” lysophosphatidylcholine lysophospholipase activity.

Human eosinophils were disrupted by sonication and subjected to subcellular fractionation on continuous sucrose gradients.³⁹ Enzyme activity, using lysolecithinase terminology in use at the time, was assessed based on the release of ¹⁴C-palmitate from 1-lyso¹⁴C-palmitoylphosphatidylcholine, where one unit is 1 nanomole of released fatty acid/hour

FIGURE 2. Lysophospholipase (lysolecithinase) activity of human eosinophil sonicates resolved on Sephadex G-100 gel filtration chromatography.

The predominant peak of lysophospholipase activity, eluting with an approximate 28 kD, was collected and subjected to further chromatographic purification steps on organomercurial affinity chromatography and heparin-Sepharose chromatography. Enzyme activity, using lysolecithinase terminology in use at the time, was assessed based on the release of ¹⁴C-palmitate from 1-lyso¹⁴C-palmitoylphosphatidylcholine, where one unit is 1 nanomole of released fatty acid/hour. Any higher molecular weight lysophospholipases were excluded at the outset from further analyses of what proved to be the Charcot–Leyden crystal protein (CLC-P)

We initially subjected eosinophils to disruption by sonication and subcellular fractionation on sucrose gradients. As shown in Fig. 1, under the conditions we used, the then termed lysolecithinase activity was not detectable in granules but was prominent in cytosolic membrane-containing fractions. Does this finding exclude the potential presence of lysophospholipase enzymes in eosinophil granules? No, our cell disruption may have liberated granules, but when assayed for lysophospholipase activity the intact granules might not have been fractured open and/or any lysophospholipase within granules might have been enzymatically latent. We had reported that CLC/lysophospholipase was 10% of the eosinophil’s protein,²⁹ but now recognizing the bias based on eosinophil disruption, more specifically it should have been 10% of the granule-depleted eosinophil sonicate, because the substantial quantities of eosinophil granule cationic proteins and other granule-included proteins were not measured. We focused on the cytosolic/membrane bearing eosinophil sonicates for our further lysophospholipase purifications and studies.^26–29,40

In the premolecular era of protein chemistry and enzymology, we developed a scheme to purify the eosinophil lysophospholipase of interest to homogeneity, including in this purification scheme the initial limiting step in preferentially utilizing our eosinophil sonification conditions to release a cytosolic-rich starting material. The propensity of this “sticky” protein to adhere to various ion exchanges, hydrophobic, and other chromatographic media confounded the development of our chromatographic purifications. Sephadex G-100 size-exclusion gel filtration chromatography became our initial step (Fig. 2), although as we noted that as the protein was further purified, it aberrantly absorbed to gel filtration media,³⁰ as also noted earlier by Gleich et al.¹⁶ The chromatography steps we employed utilized an opening Sephadex G-100 gel filtration step based on size exclusion (Fig. 2), followed by an organomercurial agarose affinity chromatographic column that bound cysteinyl residues and was eluted with a dithiothreitol gradient (see Fig. 1 in Weller, 1988²⁶) and thereafter a heparin-Sepharose column in which the enzymatic activity appeared in the fall-through (see Fig. 2 in Weller, 1988²⁶). The chromatographically purified protein on 0.1% SDS-PAGE was homogeneous by protein staining and had an apparent molecular weight of 17.4 kD.

As we purified this major eosinophil protein based on its lysophospholipase activity, we hypothesized that this protein might be the CLC-P. We showed that this chromatographically purified eosinophil lysophospholipase could crystallize as CLCs.²⁶ Although this was notable, definitive evidence was warranted to understand nominal physicochemical differences between the CLC molecular weights of 13 kD reported by Gleich et al.¹⁶ vs. our 17.4 kD and importantly to preclude a protein contaminant in mediating the lysophospholipase activity we ascribed to the CLC-P.

We reconciled the 13 kD vs. 17.4 kD differences by identifying that the CLC-P, both derived from solubilized crystals and from chromatographic purifications, migrated aberrantly faster in 1% SDS-PAGE, with the lower apparent 13 kD molecular weight than in 0.1% SDS with a 17.4 kD apparent molecular weight.²⁹ Such aberrant migration in 1% SDS-PAGE was likely due to enhanced binding of SDS; and indeed we documented that CLC-P bound levels higher than most proteins of the anionic detergents, SDS (e.g., 3.820 g SDS/g protein vs. ∼1.4 g SDS/g for soluble proteins) and sodium deoxycholate.²⁹ The markedly enhanced binding of anionic detergents can be found with some membrane proteins⁴¹; however, neither membrane-inserting regions nor regions of sufficient hydrophobicity were identified in the CLC-P.^17,29 As noted below, the subsequent recognition that CLC-P contained a thioester-linked fatty acid would explain this enhanced detergent binding.

With regard to later concerns that the lysophospholipase activity of the chromatographically purified protein and the crystal-derived CLC-P might be attributable to an alternative contaminating eosinophil-derived lysophospholipase, there are many lines of evidence that preclude that possibility. Of note, specifically whether 74 kD or 85 kD proteins^36,38 were the true lysophospholipases, on the initial G-100 gel filtration step (Fig. 2) lysophospholipase activity gel-filtered with an approximate molecular weight of 28 kD, which is consistent with the same gel-filtered 28 kD more recently reported for dimeric recombinant Gal-10.⁴² On our initial G-100 gel-filtration step, higher molecular weight lysophospholipases would have been excluded in the pooled fractions utilized for our subsequent chromatographic purification steps. Moreover, in all of our electrophoretic analyses, there was only a single protein band that migrated identically for both chromatographically purified and crystal-derived CLC-P (including the aberrant migration in 1% SDS noted above).^26–29,40

On nondenaturing alkaline PAGE gels, lysophospholipase activity was associated solely with the single protein staining band (see Figure 2 in Weller et al., 1980²⁶). Both chromatographically purified and crystal-derived lysophospholipase preparations exhibited identical Michaelis constants (Km, 22 μM) for the substrate lysopalmitoylphosphatidylcholine.²⁹ Monospecific rabbit antisera, prepared against homogeneous, chromatographically purified eosinophil lysophospholipase, and separately against CLC formed in vitro, yielded only precipitin lines fusing in a pattern of immuno-chemical identity on Ouchterlony analyses with disrupted eosinophils, purified lysophospholipase, and solubilized CLC-P (see Figure 1 in Weller et al., 1982²⁸). Moreover, both chromatographically purified and crystal-derived CLC-Ps had similar amino acid compositions and on Edman degradation exhibited blocked amino termini.²⁹ Thus, by multiple analytic approaches, both chromatographically purified and CLC crystal-derived proteins were the sole, identical eosinophil protein expressing inherent lysophospholipase activities.

3 |. FURTHER MOLECULAR CHARACTERIZATION OF CLC-P

Subsequent studies after our biochemical and immunochemical characterizations of CLC-P and its lysophospholipase activity led to the molecular cloning and characterization of recombinant CLC-P, which began to identify its similarities with S-type animal lectins and its lack of sequence similarities with then recognized mammalian lysophospholipases.¹⁷ Although CLC lacked similarities with then recognized lysophospholipases, subsequent studies with recombinant CLC continued to document the inherent lysophospholipase activities of the CLC-P. In 1991, Zhou et al. reported the initial successful stable transfection of Chinese hamster ovary cells with a cloned full length CLC construct. Transfected cells expressed the CLC-P that formed CLCs and expressed “marked lysophospholipase activity (2.7 μmol/h/mg of expressed rCLC-P at 37°C, pH 7.5) comparable to that of eosinophils.”³¹ In 1992, Zhou et al. reported that the gene for human eosinophil CLC-P directs expression of lysophospholipase activity and spontaneous crystallization in transiently transfected COS cells.³² Transfected COS cells expressed lysophospholipase activity of 6.3 μmol/h/mg CLC-P in accord with 8 μmol/h/mg of purified native CLC-P. Notably, COS cell expressed recombinant CLC-P and native “crystal-derived” eosinophil CLC-P both demonstrated their substantial, high level endogenous lysopalmitoylphospholipase activities and released ¹⁴C-palmitate from the ¹⁴C-palmitoyl-lysophosphatidylcholine substrate. In 1993, Zhou et al. found that site-directed mutagenesis at Cys²⁹ and Cys⁵⁷ enhanced the enzyme activity of human eosinophil lysophospholipase.³³ Recombinant CLC-P transfected into COS cells again exhibited substantial lysophospholipase activity (∼2.2 μmol/h/mg); and moreover site-directed mutagenesis of either the Cys²⁹ to a glycine or the Cys⁵⁷ to glycine in transfected CLC recombinants yielded even 2–3 fold greater lysophospholipase activities.³³ Thus, CLC-P whether native or transfected clearly could also exhibit endogenous lysophospholipase activities, with specific activities substantially greater than the “weak” activities that we had found with native eosinophil CLC-P, either obtained by sequential chromatographic purification or from solubilized crystalline CLCs.^26,29

In 1993, Garsetti et al. presented a comparison of six mammalian lysophospholipases, two from HL-60 cells, two from WEHI cells, one from the pancreas, and one from human blood eosinophils.³⁴ Using a different protein purification scheme, they identified a single human eosinophil protein with lysophospholipase activity that had a molecular weight estimate by their SDS-PAGE of ∼15 kD and had an amino acid composition identical to our CLC/lysophospholipase protein.³⁴ They concluded that a single lysophospholipase, ∼15 kDa, is a major eosinophil protein that was different from the other cell’s lysophospholipases.³⁴ This finding from a group that would later also find 74 kD pancreatic-like lysophospholipase in eosinophils³⁸ was notable for independently confirming that the CLC-P is a distinct human eosinophil-derived lysophospholipase.

In 1995, Leonidas et al. reported an analysis of the crystal structure of human CLC-P and identified it as a new member of the carbohydrate-binding family of galectins.¹⁸ In concert, they assayed the lysophospholipase activity not only of recombinant CLC-P transfected into COS cells but also three recombinant mutated variants of the CLC-P in which Cys²⁹ and Cys⁵⁷ were singly or doubly mutated to glycine residues. Whereas wild type CLC expressed 1.84 μmol/h/mg of lysophospholipase activity, each of the mutants exhibited even greater lysophospholipase activities (ranging from 2.45 to 6.46 μmol/h/mg). Again, this study documented that transfected CLC-P and its cysteinyl mutants endogenously expressed high level (μmol/h/mg) lysophospholipase activities.

In 2002, without recognizing their antecedent studies demonstrating that transfected recombinant CLC-Ps endogenously exhibited consistently high levels of lysophospholipase activities (i.e., μmoles/h/mg), Ackerman et al.’s study, “Charcot–Leyden crystal protein (Gal-10) is not a dual function galectin with lysophospholipase activity but binds a lysophospholipase inhibitor in a novel structural fashion³⁶” reported that: “Our results definitively show that CLC-P is not one of the eosinophil’s lysophospholipases but that it does interact with eosinophil lysophospholipases and known inhibitors of this lipolytic activity.³⁶” This study that nominally refuted that the CLC-P can express lysophospholipase activity has been subsequently cited in “A brief history of Charcot-Leyden crystal protein/galectin-10 research”⁸ “before the CLC protein gene was sequenced, it was inappropriately considered to be a lysophospholipase. While a 74 kDa enzyme was found to be the actual lysophospholipase, a pull-down assay showed that CLC protein could interact with this lysophospholipase. This indicates that the initial enzyme assay was actually contaminated by lysophospholipase.⁸” Another recent report noted that “Initially, galectin-10 was falsely considered to have weak lipase activity but was later shown to bind pancreatic-like lysophospholipase in human eosinophils and to inhibit lipolytic activity.”⁴³ In response to the 2002 publication³⁶ and these two more recent interpretations,^8,43 there are several issues that merit attention.

4 |. REEVALUATING PREVIOUS STUDIES

First, in contrast to the reported conclusion that CLC-P is not a lysophospholipase, the study by Ackerman et al. indeed documents the inherent lysophospholipase activity of the CLC-P.³⁶ As reported in their discussion: “In the original work by Weller and colleagues in which eosinophil LPLase was first identified as CLC protein and characterized enzymatically, both the chromatographically purified protein and the protein resolubilized from Charcot-Leyden crystals were reported to express LPLase activities with Km and Vmax values of 23.7 μM and 31.8 nmol/h/mg and 21.9 μM and 46.8 nmol/h/mg of CLC protein, respectively. In our present work, affinity-purified CLC protein from both AML14.3D10 eosinophils and blood eosinophils expressed some residual LPLase activity (130 and 3.6 nmol/h/mg, respectively), comparable with the published Km values. These figures represent a minimally active LPLase compared with the activities of other known eukaryotic (mammalian) and prokaryotic LPLases. These enzymes typically have 1,000-fold greater specific activities, falling in the μmol/h/mg range, suggesting that CLC protein would at best be a very poor LPLase.” Thus, this study is fully consonant with our prior findings that the CLC-P from eosinophils indeed expresses inherent lysophospholipase activity, albeit with low specific activities. Moreover, as cited earlier, recombinant CLC-P has uniformly and consistently expressed higher, μmol/h/mg, specific activities, so the inherent lysophospholipase activity of CLC-P has been repetitively documented.

Second, there is no need to seek out a higher molecular weight enzyme with lysophospholipase activities that would have been excluded in our starting Sephadex G-100 size exclusion gel filtration pooling of an apparent 28 kD peak of activity (Fig. 2), to account for the documented inherent lysophospholipase activity of the CLC-P. Given the hydrophobicity and “stickiness” of the CLC-P, it is not surprising that rabbit polyclonal antibody-bound CLC-P would interact with varied proteins, even from nonnative murine eosinophils. That some of these could include an 85 kD putative lysophospholipase or other lysophospholipase proteins does not detract from the inherent lysophospholipase activity of the CLC-P. Moreover, none of these putative binding lysophospholipases were documented to contribute to the levels of lysophospholipase activity specifically attributable to the CLC-P.

Third, with regard to CLC binding a lysophospholipase inhibitor in a novel structural fashion, the title has proven misleading. A recent commentary noted “Initially thought to be a lysophospholipase on the basis of gel chromatography studies, the observed increase in lysophospholipase activity was subsequently shown to be due to binding of galectin10 to a lysophospholipase inhibitor.”⁴⁴ This interpretation would suggest that CLC-P (#1) binds an endogenous inhibitor of lysophospholipases (#2) and thereby enabling an intracellular lysophospholipase (#3) to act, a curious ménage à trois scenario involving the CLC-P, a lysophospholipase inhibitor, and finally a lysophospholipase. Rather, what is reported in their elegant crystallographic studies, they found that N-ethylmaleimide, a known inhibitor of CLC lysophospholipase activity,²⁹ bound near the carbohydrate recognition domain of Gal-10.³⁶

Fourth, galectins can be lysophospholipases, although the 2002 study noted for Gal-10 that no members of the galectin superfamily had been reported to express any lysophospholipase or other lipolytic activities.³⁶ In contrast, galectin-13 (Gal-13, human placental tissue protein 13), which is highly homologous with CLC-P, was reported in 1999 to exhibit lysophospholipase activity as assessed by ³¹P and ¹H NMR analyses for generation of glycerophosphorylcholine (GPC) and fatty acids, respectively.⁴⁵ The endogenous lysophospholipase of activity of Gal-13 was confirmed by ³¹P-NMR using both purified and recombinant Gal-13.⁴⁶ Notably, this focused not on assaying liberated fatty acid but rather on the formation of the alternative cleavage product, GPC, released from the substrate lysophospholipid. We have confirmed by similar NMR analyses that CLC-P likewise as a lysophospholipase cleaves lysopalmitoylphosphatidylcholine to release GPC (Fig. 3).

FIGURE 3. Palmitic acid-related activities of the Charcot–Leyden crystal protein (CLC-P).

In (A) CLC-P exhibits lysophospholipase activity by catalytically releasing both reaction products, palmitate (Palm) and glycerophosphorylcholine (GPC), from substrate lysopalmitoylphosphatidylcholine. In (B) CLC-P isolated from human eosinophils contains hydroxylamine-cleavable, cysteinyl-bound palmitate (Palm) (B). In (C) CLC-P intracellularly within human eosinophils undergoes autopalmitoylation to incorporate free palmitate (Palm) attached to cysteine residues in CLC-P. CLC-P is richly localized immediately beneath the plasma membrane where it is a candidate to participate in regulating palmitoylation, including of many secretory vesicle proteins, such as vesicle associated membrane protein-2 (VAMP-2), whose functions maybe mediated by their palmitoylation status

5 |. ON THE FUNCTIONS OF CLC-P

CLC, like other galectins, lacks a N-terminal secretory sequence and is not a classically secreted protein. Although an early report seemed to localize the CLC-P to “primary” coreless granules in eosinophils from a hypereosinophilic syndrome subject,⁴⁷ our contemporary understandings are that coreless granules are early developmental stages of eosinophil specific granules.⁴⁸ It is now clearer that CLC-P is not granule associated but is a major, highly abundant protein within the cytoplasm of human eosinophils.⁴⁹ By immunofluorescence microscopy and at higher resolution by immunoelectron microscopy, we have localized CLC-P richly just under the plasma membrane (concentrated in a palmitoylation status narrow [∼250 nm wide], circumferential cytoplasmic band) and within the plasma membrane.⁴⁹ What may be the function of this protein with this dominant subplasmalemmal location?

The initial focus on lysophospholipase expressing proteins was on their capacities to catabolize lysophospholipids, and for eosinophils there was early speculation that such an enzyme might be pertinent in responses to helminth parasites. When Wang and Dennis categorized lysophospholipases in 1989,³⁵ the principal focus was on hydrolytic enzymes that released fatty acid, but they did note a then contemporary report of an atypical lysophospholipase that functioned as a cytoplasmic acyl-protein thioesterase (APT) and removed palmitate from G proteins and p21 RAS.⁵⁰ This began to lead to a broader recognition that some lysophospholipases functioned more dominantly as APTs, enzymes that catalyze the removal of thioester-linked long chain fatty acids, such as palmitate, from cysteine residues of proteins. Indeed, what had been termed lysophospholipase, LYPLA1, became APT1 and another lysophospholipase, LYPLA2, became APT2.⁵¹

Palmitoylation, the covalent attachment of palmitate via thioester bonds to cysteines, is the most common post-translational modification of proteins, affecting over 12% of the human proteome.⁵² Palmitoylation cycles regulate, even with rapid events, the functions of many proteins in other cells, including membrane localization, subcellular trafficking, and vesicular transport and secretion.^53–55 The addition of palmitate to proteins is mediated by palmitoyl-acyl transferases and the removal of palmitate is mediated by APTs.^51,56,57 Some APTs are themselves palmitoylated through fatty acid-thioester links to cysteines and may also undergo autopalmitoylation, mediating the self-incorporation of thioester-linked palmitate.⁵⁸ We are finding that crystal-derived CLC-P contains a hydroxylamine labile, thioester-linked fatty acid, likely palmitate, and that within human eosinophils CLC-P undergoes autopalmitoylation to incorporate palmitic acid (Fig. 3). The covalent attachment of a C16 fatty acid, palmitate, to CLC-P may help explain the marked hydrophobicity and anionic detergent binding previously documented to be exhibited by CLC-P.²⁹ Thus, CLC-P, as it has evolved in human eosinophils, is involved in palmitoylation likely in the subplasmalemmal space where it is richly localized.⁴⁹

Although the “palmitome” of human eosinophils has not yet been studied, the human eosinophil contains many proteins that in other cell types are regulated by palmitoylation, including the integrin β4, the tetraspanins, CD9 and CD63, and secretory vesicle associated proteins.^59–61 Secretory vesicles in human eosinophils are critically important because human eosinophils secrete their preformed proteins, including MBP and even cytokines derived from granule sites, not by wholesale granule exocytosis but by a well-developed and regulated process of piecemeal degranulation (PMD).⁶² In PMD vesicles traffic from intracellular granules to transport proteins to subplasmalemmal space where docking and fusion of vesicles leads to extracellular secretion.^62,63 Proteins well implicated in the secretory responses and PMD of human eosinophils include CD63, VAMP-2 (vesicle associated membrane protein-2), and SNAP-23 (soluble N-ethylmaleimide-sensitive factor attachment protein).^64,65 These and other associated secretory vesicle proteins have been localized by confocal microscopy^64,65 and by pre-embedding immunonanogold EM for VAMP-2 to the subplasmalemmal region of human eosinophils.^64,66 Although not studied in human eosinophils, in other cells VAMP-2 and SNAP-23 are membrane anchored and regulated by their state of cysteinyl palmitoylation.^60,61 Just as palmitoylation of vesicle-associated proteins has been identified in mediating neuronal synaptic vesicle secretion,⁶¹ it is likely that palmitoylation plays a regulatory role in the subplasmalemmal space of eosinophils where the terminal vesicular secretory events of PMD govern the secretory responses of human eosinophils. Thus, although CLC-P expresses endogenous lysophospholipase activity, like some other lysophospholipase enzymes, it likely also functions in regulating dynamic palmitoylation cycles, including, given its subplasmalemmal location, at the human eosinophil’s plasma membrane.

Abbreviations:

APT

acyl-protein thioesterase

CLC

Charcot–Leyden crystal

CLC-P

Charcot–Leyden crystal protein

EETosis

eosinophil extracellular trap cell death

Gal-10

galectin-10

Gal-13

galectin-13

GPC

glycerophosphorylcholine

MBP

major basic protein

Palm

palmitate

PMD

piecemeal degranulation

SNAP-23

soluble N-ethylmaleimide-sensitive factor attachment protein-23

VAMP-2

vesicle associated membrane protein-2