Receptors for the 5-oxo class of eicosanoids in neutrophils.

5-Hydroxy- and 5-oxo-eicosatetraenoate (5-HETE and 5-oxoETE) activate polymorphonuclear neutrophils (PMNs) through a common, receptor-like recognition system. To define this system, we examined the interaction of these eicosanoids with human PMNs. PMNs esterified 5-[3H]HETE to glycerolipids at 37 and 4 degreesC. At 37 but not 4 degreesC, the cells also hydroxylated the label to 5, 20-[3H]diHETE. The acyl:CoA synthetase blocker, triacsin C, inhibited esterification but also led to an increase in the hydroxylation of the label. PMNs processed 5-[3H]oxoETE through the same pathways but only or principally after reducing it to 5-[3H]HETE (37 or 4 degreesC). In the presence of these varying metabolic reactions, PMNs (37 or 4 degreesC; +/- triacsin C) could not be shown to receptor bind either radiolabel. Plasma membranes isolated from PMNs esterified but unlike whole cells did not reduce or hydroxylate 5-[3H]oxoETE. Triacsin C blocked esterification, thereby rendering the membranes unable to metabolize this radiolabel. Indeed, triacsin C-treated membranes bound (Kd = 3.8 nM) 5-[3H]oxoETE specifically and reversibly to 86 pmol of sites per 25 micrograms of membrane protein. 5-OxoETE, 5-HETE, and 5,15-diHETE displaced this binding at concentrations correlating with their potency in eliciting PMN Ca2+ transients. GTP and GTPgammaS, but not ATP or ATPgammaS, also reduced 5-[3H]oxoETE binding, whereas 15-HETE, leukotriene B4, platelet-activating factor, IL-8, C5a, and N-formyl-Met-Leu-Phe lacked this effect. We conclude that PMNs and their plasma membranes use an acyl:CoA synthetase-dependent route to esterify 5-HETE and 5-oxoETE into lipids. Blockade of the synthetase uncovers cryptic plasmalemma sites that bind 5-oxoETE with exquisite specificity. These sites apparently mediate responses to the 5-oxo class of eicosanoids and are likely members of the serpentine superfamily of G protein-linked receptors.


EXPERIMENTAL PROCEDURES
Buffers and Other Reagents-Cells and membranes were suspended in a modified Hanks' balanced salt solution (9) containing 1.4 mM CaCl 2 and 0.7 mM MgCl 2 unless indicated otherwise. Stimuli, glycerolipids, and other reagents were obtained commercially (8,9). Triacsin C was purchased from Biomol (Plymouth Meeting, PA).

5-[ 3 H]HETE and [ 3 H]LTB
Ca 2ϩ Transient and Binding Assays-PMNs were isolated from normal donor blood (7). Ca 2ϩ transients were done on fura2-AM-loaded PMNs suspended in Hanks' buffer (ϩCa 2ϩ , no MgCl 2 ) (7). Binding was done on PMNs (10 7 ) suspended in 1 ml of Hanks' buffer (37°C) and treated with triacsin C or Me 2 SO for 30 min. Suspensions were placed on ice for 30 min, incubated (4°C) with label for 0 -120 min, layered on 0.4 ml of silicone oil, and centrifuged (12,000 ϫ g for 1 min at 4°C). Isolated supernatant fluids and pellets were counted for radioactivity (32). Results are given as the percentage of total recovered radioactivity in pellets. For membrane binding assays, plasma membranes (see below) were incubated (37°C) with radiolabel in 250 l of Hanks' buffer and passed through GF/C filters. Filters were washed with 5 ml of Hanks' buffer (no CaCl 2 or MgCl 2 ; 4°C), air-dried, and counted for radioactivity (32).
Phospholipid Analysis-PMNs (10 7 in 1 ml of Hanks' buffer) were incubated at 37°C with triacsin C or Me 2 SO and placed on ice for 30 min (for 4°C experiments) or used directly (for 37°C experiments). Suspensions were incubated (4 or 37°C, respectively) with a radiolabel for 2.5-120 min and centrifuged (12,000 ϫ g for 5 s at 4°C). Cell pellets were twice washed with 1 ml of Hanks' buffer (4°C; no CaCl 2 or MgCl 2 ) and then suspended in 0.5 ml of the same buffer. For plasmalemmal studies, plasma membranes were isolated from PMNs (see below), suspended in Hanks' buffer (37°C), treated with triacsin C or Me 2 SO for 30 min, and incubated (37°C) with label. Final washed PMN suspensions, pools of the supernatant, and washes of the original cell suspensions and membrane suspensions were twice extracted with an equal volume of chloroform:methanol, 2:1, v/v, containing enough HCl to make the final mixture pH 3-4. Pooled extracts were evaporated to dryness, taken up in 40 l of chloroform:methanol (2:1, v/v), and applied to heat-activated (180°C, 3 h) Silica Gel G TLC plates (Analtech, Newark, DE). Plates were developed to 15 cm in TLC system I (ethyl ether:hexane:acetic acid, 60:40:1, v/v) or II (ethyl ether:hexane: NH 3 (OH), 60:40:1, v/v) and scraped in 0.5-cm zones. Scrapings were suspended in 200 l of methanol for 10 min and counted for radioactivity (18). In selected experiments, extracts were evaporated to dryness and suspended in 500 l of Tris (1 M) buffer (pH 8.0) containing 500 ng of 5-oxoETE, 500 ng of rac 5-HETE, and 5 g of triglyceride. Reactions were incubated (20°C) with BSA or triglyceride lipase (Calbiochem, 50,000 units/mg, or Sigma, 60,000 units/mg) for 20 min and then extracted and examined by TLC. In other analyses, extracts of supernatant fluids were dried, taken up in 30 l of methanol or isopropanol, and analyzed by reversed-phase or normal-phase HPLC, respectively.

Effect of Triacsin C on 5-[ 3 H]HETE
Metabolism-PMN suspensions were treated with Me 2 SO or the acyl-CoA synthetase inhibitor triacsin C (33) at 37°C for 30 min, equilibrated to 4°C or kept at 37°C, incubated with 100 fmol/ml of 5-[ 3 H]HETE, separated into cells and media, and analyzed by TLC. PMNs progressively incorporated the label to equilibrium levels over 40 min at 37°C or 60 min at 4°C. After 80 min at 37°C, Me 2 SO-treated PMNs had taken up Ͼ90% of the label. On TLC system I, cellular radioactivity migrated with triglyceride and to a small extent phospholipid ( Fig. 1, top left panel), whereas medium radioactivity moved with 5,20-diHETE (Fig. 1, bottom left panel). In studies at 4°C, in contrast, PMNs took up only ϳ25% of label, and most of this migrated with triglyceride. Media label migrated with 5-HETE (Fig. 2, left panels). The pattern of these results was not limited to 80-min incubations. Analyses of suspensions incubated (37 or 4°C) with label for 5, 10, 20, 40, or 60 min, as well as 80 min, by TLC, HPLC, and other systems revealed that PMNs contained glycerolipid-associated but virtually no intact 5-[ 3 H]HETE (21). We next examined the effect of triacsin C with the expectation that this drug would block the charging of 5-HETE with CoA and thereby reduce its esterification while increasing its accumulation as intact, receptor-bound ligand. Triacsin C (Ն 1 M) did in fact reduce esterification. PMNs treated with a 20 M concentration of the drug had only a small amount of triglycerideassociated radioactivity during an 80-min incubation with 5- radioactivity migrating with 5,20-diHETE and to a lesser extent 5-HETE in TLC system I ( Fig. 1, bottom right panel), TLC system II (not shown), and HPLC ( Fig. 3, top panel). In 4°C experiments, cells had small amounts (Ͻ3%), and media had far larger amounts (Ͼ70%), of radioactivity that moved with 5-HETE in TLC system I (Fig. 2, bottom right panel), TLC system II (data not shown), and reversed-phase HPLC (data not shown). Analysis of triacsin C-treated PMNs exposed to 5-[ 3 H]HETE (4 or 37°C) for 20 min gave results paralleling those shown in Figs. 1 and 2. In particular, these PMNs never amassed more than 3% of total radioactivity as intact 5-[ 3 H]HETE. We conclude that triacsin C blocks the esterification of 5-HETE into glycerolipids. It also leads to an increase in the oxidation of the fatty acid at 37°C and promotes accumulation of intact 5-[ 3 H]HETE with cells at 4°C.

5-[ 3 H]OxoETE Metabolism and
Effects of Triacsin C-PMNs incubated with 100 pmol of 5-[ 3 H]oxoETE, Ϯ 20 M triacsin C, for 80 min at 4 or 37°C incorporated and processed the label in a fashion similar to their processing of 5-[ 3 H]HETE. More particularly, cell label migrated with glycerolipids, media label migrated with 5,20-diHETE, and triacsin C reduced glycerolipid-co-migrating and increased 5,20-diHETE-co-migrating radioactivity (Figs. 4 and 5). However, the media from these incubations also had radioactivity that migrated with 5-HETE in TLC system I (Figs. 4 and 5, bottom panels), TLC system II (data not shown), reversed-phase HPLC (Fig. 3, bottom panel), and normal-phase HPLC (data not shown). Hence, PMNs readily reduce 5-[ 3 H]oxoETE and might process the latter metabolite further. We accordingly examined the radioactivity migrating with lipids and 5,20-diHETE in greater detail. Cell extracts were digested with triglyceride lipase. The label recovered from these digests migrated with 5-HETE rather than 5-oxoETE on TLC (Fig. 6) and reversed-phase HPLC (not shown). Similarly, label in the media from these suspensions eluted on reversed-phase HPLC not only with 5-HETE but also with 5,20-diHETE (Fig. 3, bottom panel). We did not characterize the latter species because studies (12) indicate that it is composed of -oxidized metabolites of both 5-HETE and 5-ox-oETE. In a final set of experiments, PMNs (Ϯ triacsin C; 37 or 4°C) were exposed to 5-[ 3 H]oxoETE for 20 rather than 80 min. These cells had radioactivity patterns similar to that seen in Figs. 4 and 5, e.g. cells from these shorter term incubations did not contain appreciable levels of intact 5-[ 3 H]oxoETE. We con- also metabolized 5-[ 3 H]oxoETE, converting it to slower migrating species in TLC system I (Fig. 7, top panel). Triacsin C (10 M) inhibited the latter reaction (Fig. 8, top panel). Furthermore, triglyceride lipase digestion of extracts from these membranes recovered intact 5-[ 3 H]oxoETE (Fig. 7, bottom panel).
The data indicate that plasma membranes readily esterify 5-oxoETE into glycerolipids, a result differing from whole cells, which reduce 5-oxoETE before esterification. The membranes also differed from PMNs in that after triacsin C treatment they did not metabolize 5-[ 3 H]oxoETE over 60 min (Fig. 8). The last finding points to a condition that is appropriate for testing receptor binding.
Membrane Binding of 5-[ 3 H]oxoETE-Plasma membranes were treated with 10 M triacsin C, incubated with 25 fmol of 5-[ 3 H]oxoETE in a volume of 250 l at 37°C, and freed of soluble label by passage through GF/C filters. The membranes took up 5-[ 3 H]oxoETE to apparent equilibrium levels of ϳ1 fmol/25 g of membrane protein over 20 min (Fig. 9, top panel, solid line). Binding fell by 80% when 25 pmol of 5-oxoETE was added simultaneously with, 20 min after, or 60 min after (Fig.  9, top panel, interrupted lines) 5-[ 3 H]oxoETE. 5-OxoETE binding thus was readily saturated, specific, and reversible. The parameter of specific binding rose linearly with membrane mass over a range of 0 -40 g/ml (Fig. 9, middle panel). Scatchard analysis revealed a single class of binding sites with an average K d of 3.8 nM and a B max of 86 pmol/25 g of membrane protein (Fig. 9, bottom panel). To examine the specificity of these results, membranes were incubated with 5-[ 3 H]oxoETE in the presence of selected agents. 5-OxoETE, 5-HETE, and 5,15-diHETE displaced 5-[ 3 H]oxoETE (Fig. 10, top panel) with potencies that corresponded to their strength in eliciting PMN Ca 2ϩ transients as represented in Fig. 10, bottom panel, and elsewhere in this (5-10, 12) as well as other PMN assays (9, 11-13, 16, 31). 15-HETE had little radiolabel-displacing, Ca 2ϩ transient-stimulating (Fig. 10), or other actions (9,11,16,31 (Table I). FMLP, PAF, IL-8, and C5a did not displace binding (Table I). Finally, GTP and GTP␥S, but not ATP or ATP␥S, reduced 5-[ 3 H]oxoETE binding.  (Table I). DISCUSSION PMNs presented with pM levels of 5-[ 3 H]HETE at 37°C rapidly took up the label and, in pace with this, deposited it in glycerolipids, principally triglyceride (Fig. 1, left panels). Arai et al. (22) first observed that triacsin C blocks this reaction. Our data agree with theirs but also indicate that PMNs -hydroxylate 5-[ 3 H]HETE and rapidly released the product. Triacsin C increased 5,20-[ 3 H]diHETE formation (Figs. 1 and 3, right panels). Evidently, then, the deposition of 5-HETE in lipid by PMNs involves a standard fatty acid processing pathway: triacsin C-sensitive synthetase forms 5-HETE-CoA, and a transacylase esterifies the 5-HETE adduct to membrane lipid. Concurrently, an -oxidizing activity, possibly NADP ϩdependent P-450 LTB 4 C-20-hydroxylase (21,34,35), handles a small percentage of 5-[ 3 H]HETE in control, but a far larger percentage of radiolabel in triacsin C-treated, PMNs. -Oxidation, we note, also predominates over esterification in cells presented with M levels of 5-HETE (21). The acylation pathway thus appears to have a high affinity but low capacity for substrate, whereas the oxidation pathway has a low affinity but high capacity. In the cases studied here, the two pathways in cooperation or the -oxidation acting alone can scavenge virtually all 5-[ 3 H]HETE as it enters PMNs at 37°C (Fig. 1).
At 4°C, PMNs did not -oxidize yet still acylated 5-[ 3 H]HETE (Fig. 2, left panels). Triacsin C blocked this reac-tion and resulted in the accumulation of 5-[ 3 H]HETE in PMNs (Fig. 2, right panels). Quantitatively, this accumulation was minor. Furthermore, the accumulation of intact radiolabel was neither reduced nor reversed with excess 5-HETE and therefore could not be ascribed unambiguously to a receptor binding event. The data indicate that either PMNs lack discrete 5-HETE receptors or 5-[ 3 H]HETE is unsuited to detect these receptors in whole PMNs even at 4°C and in the presence of triacsin C.
Powell et al. (6,34) found that the endoplasmic reticulum of PMNs has both NADPH-dependent reductase and NADP ϩ -dependent dehydrogenase activities that, in the presence of the appropriate cofactor, convert 5-HETE to 5-oxoETE and 5-ox-oETE to 5-HETE, respectively, in vitro. Whole PMNs, however, reduce 5-oxETE but do not dehydrogenate 5-HETE, perhaps because the cells maintain a high ratio of NADPH to NADP ϩ (6). Our findings agree with this notion in that PMNs converted 5-oxoETE to 5-HETE ( Fig. 4) but had no propensity to conduct the reverse reaction (Fig. 1). Moreover, 5-oxoETE reduction occurred at both 4 and 37°C and appeared to dominate esterification. That is, PMNs incubated with 5-[ 3 H]oxoETE deposited the label in triglyceride-co-migrating material. Triglyceride lipase digestion of the extracts from these cells recovered  (Fig. 6). Because plasmalemma can esterify 5-[ 3 H]oxoETE to glycerolipids (Fig. 7, top  panel), whole PMNs likely have a latent capacity to esterify 5-oxoETE. Failure to observe this esterification may be due to the enzymatic reduction of substrate as it enters the cell, an esterification pathway that prefers 5-HETE to 5-oxoETE, or an NADPH-dependent dehydrogenase that rapidly attacks esterified 5-oxoETE.
Plasmalemma from PMNs readily takes up 5-[ 3 H]HETE but then deposits virtually all of the incorporated label in membrane lipid (8). Fig. 7 shows that membranes also metabolize 5-[ 3 H]oxoETE, converting it to a triglyceride lipase-sensitive species. This finding indicates that the membranes, unlike whole PMNs, readily esterify 5-oxoETE. Triacsin C inhibited this reaction (Fig. 8, top panel). Because our membrane isolates have almost no endoplasmic reticulum marker enzymes or LTB 4 -hydroxylase activity (36), it is unlikely they would have endoplasmic reticulum-associated (6) 5-oxoETE reductase or -hydroxylase activities. This notion was confirmed here: triacsin C-treated membranes failed to metabolize 5-[ 3 H]oxoETE to any measurable extent. The last finding also defined an experimental condition for detecting 5-oxoETE receptors.
In the presence of triacsin C, PMN plasma membranes bound 5-[ 3 H]oxoETE. Binding was time-dependent, proportional to membrane mass, and displaced as well as reversed by excess 5-oxoETE (Fig. 9). The ligand bound to 86 fmol per 25 g of membrane protein of sites with an average dissociation affinity of 3.8 nM. PMN plasma membranes treated with triacsin C thus express a finite number of sites that reversibly bind 5-oxoETE. To establish the relationship of these sites to 5-HETE receptors, we performed three sets of studies. First, 5-oxoETE, 5-HETE, and 5,15-diHETE use the same recognition system to stimulate PMNs (31,37). The eicosanoids elicited Ca 2ϩ transients with ED 50 potencies of 10, 200, and 7000 nM, respectively (Fig. 10, bottom panel). Their respective ID 50 (37) values for reducing 5-[ 3 H]oxoETE binding were 7, 50, and 5000 nM. In addition, 15-HETE failed to stimulate PMNs and only marginally (ID 50 Ͼ10,000 nM) altered 5-[ 3 H]oxoETE binding (Fig. 10). The correlation between these two actions of the eicosanoids was 0.99 (p Ͻ 0.05), a correspondence implying that the binding sites transmit the Ca 2ϩ -mobilizing action of the analogs. There is a similarly high correlation between the binding displacement of the analogs and PMN degranulating (9), chemotactic (16), and other effects (9,12,13,16). Second, binding sites for 5-oxoETE and LTB 4 were essentially nonin- were treated with of 10 M triacsin C for 30 min, incubated with 100 pM 5-[ 3 H]oxoETE plus the indicated concentration of eicosanoid for 60 min at 37°C, and passed through GF/C filters. Test eicosanoids were 5-ox-oETE (1), 5-HETE (2), 5,15-diHETE (3), and 15-HETE (4). Data are corrected for the amount of radioactivity adherent to filters from experiments incubating label without membranes. Results are the mean percentages of radioactivity trapped by filters Ϯ S.E. for 3-6 experiments. In the bottom panel, fura2-treated PMNs were challenged with an eicosanoid for 1 min and assayed for the concentration of cytosolic Ca 2ϩ (7). Results are the mean concentrations (nM) of cytosolic Ca 2ϩ for four experiments (Ca 2ϩ was ϳ45 nM in resting and BSA-stimulated PMNs). 4 to membranes PMN membranes were treated with a stimulus for 0 min or nucleotide for 30 min, incubated with a 100 pM concentration of the indicated labels for 60 min, and passed through GF/C filters. Filters were washed and assayed for radioactivity. teracting (Table I). This agrees with studies (7)(8)(9)(10)(11) showing that these two agents share 5-hydroxy, 6 -7 double bond motifs yet still act on different recognition systems. FMLP, PAF, IL-8, and C5a likewise did not interfere with 5-[ 3 H]oxoETE binding (Table I). The latter chemotactic factors act through respective receptors that do not recognize 5-oxoETE (9). The 5-oxoETE binding site defined here, then, is unique and does not involve receptors for other chemotactic factors or metabolic enzymes, such as LTB 4 -hydoxylase. Third, GTP and GTP␥S, but not their adenosine analogs, reduced the membrane binding of 5-[ 3 H]oxoETE. The effect paralleled that found for the binding of [ 3 H]LTB 4 (Table I), an agent that operates on G proteinlinked receptors (4). Such receptors induce G proteins to exchange GTP for GDP and thence to dissociate into ␣ and ␤␥ subunits. The receptors also must associate with G proteins in order to maintain a high affinity for their ligand. GTP analogs, by disrupting G proteins, convert serpentine receptors to a low affinity or nonbinding configuration (38). The nucleotide effects of Table I thus implicate receptor-G protein linkages in 5-[ 3 H]oxoETE binding. Because the stimulating actions of 5-HETE and 5-oxoETE require pertussis toxin-sensitive G proteins (8,18), the structurally specific effects of GTP analogs reinforce a notion that 5-oxoETE binding sites are in fact receptors.

TABLE I Effects of selected agents on the binding of 5-[ 3 H]oxoETE and [ 3 H]LTB
In conclusion, PMNs and their isolated plasma membranes utilize a triacsin C-sensitive acyl-CoA synthetase pathway to esterify 5-HETE and 5-oxoETE to glycerolipids. The pathway operates at 37 or 4°C. A second, and also temperature-insensitive, pathway reduces 5-oxoETE to 5-HETE and, in whole cells, dominates esterification: PMNs reduce 5-oxoETE before depositing it in glycerolipid. In a third pathway, 5-HETE and possibly 5-oxoETE are -oxidized. This route requires physiological temperatures and processes large amounts of substrate when acyl-CoA synthetase is blocked or saturated. Operation of these varying pathways obscures the whole-cell interactions of 5-HETE and 5-oxoETE with receptors. Plasmalemma, however, fails to oxidize, reduce, or, in the presence of triacsin C, esterify 5-oxoETE. Rather, it binds 5-oxoETE specifically and reversibly. Stringent criteria implicate these binding sites in mediating PMN responses to the 5-oxo class of eicosanoids. Finally, the effect of GTP analogs on 5-[ 3 H]oxoETE binding suggests that the latter sites belong to the serpentine superfamily of G protein-linked receptors.