The C2-like β-Barrel Domain Mediates the Ca2+-dependent Resistance of 5-Lipoxygenase Activity Against Inhibition by Glutathione Peroxidase-1*

Recently, we reported that in crude enzyme preparations, a monocyte-derived soluble protein (M-DSP) renders 5-lipoxygenase (5-LO) activity Ca2+-dependent. Here we provide evidence that this M-DSP is glutathione peroxidase (GPx)-1. Thus, the inhibitory effect of the M-DSP on 5-LO could be overcome by the GPx-1 inhibitor mercaptosuccinate and by the broad spectrum GPx inhibitor iodoacetate, as well as by addition of 13(S)-hydroperoxy-9Z,11E-octadecadienoic acid (13(S)-HPODE). Also, the chromatographic characteristics and the estimated molecular mass (80–100 kDa) of the M-DSP fit to GPx-1 (87 kDa), and GPx-1, isolated from bovine erythrocytes, mimicked the effects of the M-DSP. Intriguingly, only a trace amount of thiol (10 μm GSH) was required for reduction of 5-LO activity by GPx-1 or the M-DSP. Moreover, the requirement of Ca2+ allowing 5-LO product synthesis in various leukocytes correlated with the respective GPx-1 activities. Mutation of the Ca2+ binding sites within the C2-like domain of 5-LO resulted in strong reduction of 5-LO activity by M-DSP and GPx-1, also in the presence of Ca2+. In summary, our data suggest that interaction of Ca2+ at the C2-like domain of 5-LO protects the enzyme against the effect of GPx-1. Apparently, in the presence of Ca2+, a low lipid hydroperoxide level is sufficient for 5-LO activation.

In cell-free systems Ca 2ϩ , ATP, phospholipids (membranes), lipid hydroperoxides (LOOH), and leukocyte-derived proteins have been shown to enhance 5-LO catalysis (reviewed in Ref. 1). However, the degree of stimulation by each of these components depends on the assay conditions, i.e. the source of 5-LO (isolated, in crude homogenates or cellular fractions), presence of other cofactors, the concentration of AA, etc. LOOH are of importance for the initial conversion of the active site iron from the ferrous (resting) to the ferric (active) state (7,8). Accordingly, glutathione peroxidases (GPx) that reduce LOOH inhibit 5-LO product synthesis in vitro and in intact cells (5, 9 -15), and conditions that are associated with an increased peroxidetone promote 5-LO product formation (6,16,17). Two Ca 2ϩ ions bind to the N-terminal C2-like domain of 5-LO with a K d of 6 M (18,19). Half-maximal activation of purified 5-LO was determined at 1-2 M Ca 2ϩ , whereas 4 -10 M Ca 2ϩ causes maximal activation of the enzyme (20,21). In intact cells lower concentrations of Ca 2ϩ (200 -300 nM) seem to be sufficient for 5-LO activation (22,23). It was shown that Ca 2ϩ increases the hydrophobicity of 5-LO (18) and causes 5-LO binding to phosphatidylcholine (PC) vesicles or to cellular membranes (20, 24 -26), and that the C2-like domain is important also for membrane association (27,28). Nevertheless, the mechanisms of how Ca 2ϩ stimulates 5-LO activity may involve additional factors.
Several reports state that in cell-free systems 5-LO is catalytically active without Ca 2ϩ (see Ref. 29 and references therein), and in intact cells, 5-LO phosphorylation by members of the MAP kinase family stimulate 5-LO product synthesis in the absence of Ca 2ϩ (30 -32). Mg 2ϩ at concentrations that occur in intact cells, can substitute for Ca 2ϩ regarding binding and activation of 5-LO in vitro (19,33). We observed considerable 5-LO product synthesis in homogenates of human PMNL and rat basophilic leukemia (RBL)-1 cells in the absence of Ca 2ϩ , whereas Ca 2ϩ was required for 5-LO activity in homogenates of monocytic MM6 cells under the same assay conditions (29,34). Interestingly, we found that an 80 -100 kDa soluble protein from MM6 cells renders 5-LO activity Ca 2ϩ -dependent. In this study we provide evidence that this M-DSP is GPx-1 and we suggest that Ca 2ϩ , via interaction with the C2 domain of 5-LO, renders the enzyme resistant against GPx-1, possibly by increasing the affinity toward activating LOOH. * This study was supported by grants from the Fonds der Chemischen Industrie, the EU (LEUCHRON, QLRT-2000-01521), the Swedish Medical Research Council (03X-217), and the Deutsche Pharmazeutische Gesellschaft. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Human PMNL were promptly isolated from fresh leukocyte concentrates obtained from healthy donors at St Markus Hospital (Frankfurt, Germany) as described (31). Cells were finally resuspended in PBS plus 1 mg/ml glucose (PG buffer) or PBS plus 1 mg/ml glucose and 1 mM CaCl 2 (PGC buffer) as indicated.
When broken cell preparations or purified 5-LO were assayed, S100 or partially purified 5-LO enzyme (in the presence or absence of ATP-PTs) were diluted in ice-cold PBS containing 1 mM EDTA (final volume, 1 ml) and 1 mM ATP as well as other agents (as indicated) were added. The samples were preincubated for 30 s at 37°C, and the incubation was started by the addition of AA together with or without 2 mM CaCl 2 and other compounds as indicated. After 10 min at 37°C, the incubation was stopped with 1 ml methanol and the formed 5-LO products were extracted and analyzed by HPLC as described for intact cells.
Determination of Cellular Peroxide Formation-Measurement of peroxides was conducted using the peroxide-sensitive fluorescence dye 2Ј,7Ј-dichlorofluorescein diacetate (DCF-DA). Freshly isolated PMNL (5 ϫ 10 6 in 1 ml of PGC buffer), differentiated MM6 cells or RBL-1 cells (3 ϫ 10 6 cells in 1 ml PGC buffer) were preincubated with DCF-DA (1 g/ml) for 2 min at 37°C in a thermally controlled (37°C) fluorimeter cuvette with continuous stirring in a spectrofluorometer (Aminco-Bowman series 2). The fluorescence emission at 530 nm was measured after excitation at 480 nm. The mean fluorescence data measured 5 min after stimulus addition are expressed as fold increase over unstimulated cells.
Determination of Glutathione Peroxidase Activity-GPx activity in S100 of PMNL, MM6, and RBL-1 cells or fractions of the gel-permeation chromatography was measured according to the indirect GSH reductase-coupled method described by Wendel (39). One unit GPx activity was defined as the conversion of 0.5 mol of NADPH to NADP ϩ per min at 37°C and 1 mM GSH. GPx activity is expressed as milliunits/10 6 cells.
Measurement of Intracellular Ca 2ϩ Levels-Cells (1 ϫ 10 7 in 1 ml of PGC buffer) were incubated with 2 M Fura-2/AM for 30 min at 37°C, washed, resuspended in 1 ml of PGC buffer and transferred into a thermally controlled (37°C) fluorimeter cuvette in a spectrofluorometer (Aminco-Bowman series 2) with continuous stirring. The fluorescence emission at 510 nm was measured after excitation at 340 and 380 nm, respectively. Intracellular Ca 2ϩ levels were calculated according to the method of Grynkiewicz et al. (40). F max (maximal fluorescence) was obtained by lysing the cells with 1% Triton-X 100 and F min by chelating Ca 2ϩ with 10 mM EDTA.

Identification of Glutathione Peroxidase-1 as a M-DSP That
Renders 5-LO Activity Ca 2ϩ -dependent-We attempted to identify the recently described M-DSP, which suppresses the activity of 5-LO in the absence of Ca 2ϩ (29). Based on the apparent molecular mass (80 -100 kDa) determined by gel-permeation chromatography, it appeared possible that the M-DSP is GPx-1, which was recently identified as a regulator of cellular 5-LO activity in monocytic cells (12). As shown in Fig. 1A, GPx-1 protein and GPx activity indeed co-eluted with the 5-LO inhibitory activity of the M-DSP during gel-permeation chromatography.
Next, we investigated if agents that suppress or counteract GPx-1 activity could overcome the effect of the M-DSP. Control experiments confirmed that partially purified 5-LO from MM6 cells alone is catalytically active, regardless of Ca 2ϩ . When aliquots of the MM6 ATP-PT were added back to 5-LO, 5-LO activity was enhanced about 4-fold in the presence of Ca 2ϩ . However, without Ca 2ϩ , 5-LO activity was completely suppressed (Fig. 1B). The specific GPx-1 inhibitor mercaptosuccinate (30 M) (41) as well as the broad spectrum GPx inhibitor iodoacetate (2 mM, not shown) (42), were capable of reconstituting 5-LO activity. Also, 13(S)-HPODE (3 M), that counteracts GPx-1 activity (16) was able to prevent the 5-LO inhibitory action of the MM6-ATP-PT, whereas H 2 O 2 (1-100 M) and 13-oxo-octadecadienoic acid (3 up to 30 M) had no effect. Similarly, 5-LO suppression induced by aliquots of the 80 -100 kDa fractions of the gel-permeation chromatography was inhibited by mercaptosuccinate, iodoacetate or 13(S)-HPODE (not shown). By contrast, these agents did not alter 5-LO activity in the presence of Ca 2ϩ (Fig. 1B). Notably, in all these experiments, no exogenous thiols such as GSH or dithiothreitol were added to the incubation mixtures, but it should be noted that the MM6-ATP-PT by itself contains traces of thiols, originally derived from the intracellular environment.
In S100 of MM6 cells, 5-LO is catalytically active only when Ca 2ϩ is present, but not when Ca 2ϩ is omitted (29). Also under such experimental settings mercaptosuccinate, iodoacetate, or 13(S)-HPODE were able to overcome 5-LO suppression in the absence of Ca 2ϩ (not shown). In contrast to MM6 cells, Ca 2ϩ is not needed for 5-LO activity in S100 from PMNL (34), and these cells exert only low GPx activity (Table I). However, when aliquots of the MM6-ATP-PT were added to S100 of PMNL in the absence of Ca 2ϩ (but not in its presence), 5-LO activity was strongly suppressed ( Fig. 2A). At a fixed concentration of GPx-1, variation of the amount of 5-LO enzyme in the incubation mixture did not alter the degree of 5-LO inhibition. Thus, the amounts of 5-LO products formed in 1 ml incubations containing S100 of PMNL derived from 2.5, 5, 10 or 15 ϫ 10 6 cells was 215 Ϯ 52, 462 Ϯ 147, 745 Ϯ 170, and 988 Ϯ 282 ng/ml, respectively. Inclusion of MM6-ATP-PT corresponding to 2 ϫ 10 6 cells suppressed the 5-LO activities in all of these incubations by 93-97%. Again, mercaptosuccinate, iodoacetate or 13(S)-HPODE, but not H 2 O 2 , counteracted the 5-LO inhibitory effects of the MM6-ATP-PT ( Fig. 2A). Interestingly, 5-LO activity in PMNL-S100 was considerably suppressed when 5 mM GSH, the main co-substrate of GPx-1, was included, leading to strong catalysis of GPx (Fig. 2B). Replenishment of Ca 2ϩ restored 5-LO product synthesis under these conditions. Bovine GPx-1 Mimics the 5-LO Inhibitory Effect of M-DSP-Next, we investigated if GPx-1 could mimic the inhibitory effects of the MM6-ATP-PT toward 5-LO. GPx-1, isolated from bovine erythrocytes, was added to S100 of PMNL and 5-LO activity, without supplementation of thiols, was determined. Addition of isolated GPx-1 to PMNL-S100 inhibited 5-LO activity in a dose-dependent manner with an EC 50 of ϳ70 mU/ml TABLE I GPx-1 activity, 5-LO product synthesis and peroxide formation in PMNL, RBL-1 and MM6 cells RBL-1 cells were grown in the absence (RBL-1) or presence (RBL-1 ϩ Se) of exogenously added Se 4ϩ (100 ng/ml). GPx activity was determined using S100 of PMNL (corresponding to 5-15 ϫ 10 6 ), MM6 (0.5-2 ϫ 10 6 ), RBL-1 cells (0.5-2 ϫ 10 6 ) as described. Results are given as mean ϩ SE, n ϭ 3. Peroxide formation: PMNL (5 ϫ 10 6 ), MM6 cells (3 ϫ 10 6 ), and RBL-1 cells (3 ϫ 10 6 ) were resuspended in 1 ml PGC buffer, DCF-DA (1 g/ml) was added and after 2 min at 37°C, cells were stimulated with 60 M AA. Data determined 5 min after addition of AA are expressed as -fold increase of the mean fluorescence ϩ/Ϫ S.E. over unstimulated cells, n ϭ 5. 5-LO product synthesis: PMNL (7.5 ϫ 10 6 ), MM6 cells (2 ϫ 10 6 ), and RBL-1 cells (2 ϫ 10 6 ) were diluted to 1 ml in PG. Cells were either stimulated with 60 M AA in presence of 1 mM EDTA, or alternatively with ionophore plus 60 M AA in presence of 1 mM CaCl 2 . 5-LO product formation was determined 10 min after addition of the stimuli at 37°C as described. Results are given as mean ϩ S.E., n ϭ 5.  GPx-1 in the absence but not in the presence of Ca 2ϩ (Fig. 3A). As found for MM6-ATP-PT, the degree of 5-LO inhibition at a fixed amount of GPx-1 (200 mU) was about the same for S100 of 2.5, 5, 10 or 15 ϫ 10 6 PMNL as source of 5-LO (not shown). As shown in Fig. 3B, the strong inhibition of 5-LO in PMNL-S100 by 300 mU bovine GPx-1 was almost completely reversed by mercaptosuccinate or 13(S)-HPODE (but not by H 2 O 2 ), resembling the counteracting effects observed with the MM6-ATP-PT. Again, the agents had no such up-regulatory effects on 5-LO activity when Ca 2ϩ was present.
In contrast to the PMNL-S100, for purified 5-LO enzyme (from PMNL, RBL-1, or MM6 cells), addition of isolated bovine GPx-1 failed to suppress 5-LO activity in the absence of Ca 2ϩ (Fig. 4), implying that an additional component, present in the PMNL-S100, appears to be operative. Thus, the PMNL-S100 was subjected to ATP affinity chromatography, and the pass through fraction, was separated into a high (Ͼ10 kDa) and a low molecular mass (Ͻ10 kDa) fraction by gel-permeation chromatography using a PD-10 column (Amersham Biosciences). When the low molecular mass fraction was included in the incubation mixture (containing purified 5-LO and isolated GPx-1), 5-LO was suppressed by GPx-1 in the absence of Ca 2ϩ (but not in its presence) (Fig. 4). The high molecular mass fraction rather increased 5-LO activity regardless of Ca 2ϩ . Notably, addition of 10 M GSH (0.307 kDa), could replace the low molecular mass fraction, rendering GPx-1 a potent 5-LO inhibitory enzyme (not shown). were added as indicated together with 40 M AA. After another 10 min at 37°C, 5-LO product formation was determined as described. The control represents 5-LO activity in the absence of MM6 ATP-PT. B, S100 of PMNL was diluted in ice-cold PBS containing 1 mM EDTA and 1 mM ATP, and 5 mM GSH were added as indicated. After 5 min on ice, samples were prewarmed for 30 s at 37°C, and CaCl 2 (2 mM) was added as indicated together with 40 M AA. After another 10 min at 37°C, 5-LO product formation was determined as described. Results are given as mean Ϯ S.E., n ϭ 4.

FIG. 3. GPx-1 inhibits 5-LO in S100 of PMNL in the absence of Ca 2؉ ; reversal by mercaptosuccinate and 13(S)-HPODE. A, 5-LO inhibition by
GPx-1. The S100 of PMNL, corresponding to 10 7 cells, was diluted in 1 ml of ice-cold PBS containing 1 mM EDTA and 1 mM ATP. Then, GPx-1, isolated from bovine erythrocytes, was added at the indicated amounts. After 5 min on ice, samples were prewarmed for 30 s at 37°C, and CaCl 2 (2 mM) was added as indicated together with 40 M AA. After another 10 min at 37°C, 5-LO product formation was determined. B, effects of mercaptosuccinate (merc.suc.), 13(S)-HPODE, and hydrogen peroxide. The S100 of PMNL, corresponding to 10 7 cells, were diluted in 1 ml of ice-cold PBS containing 1 mM EDTA and 1 mM ATP. Then, GPx-1 (300 mU, isolated from bovine erythrocytes) and mercaptosuccinate (30 M) were added as indicated. After 5 min on ice, samples were prewarmed for 30 s at 37°C and CaCl 2 (2 mM), 13(S)-HPODE (3 M), and hydrogen peroxide (3 M) were added as indicated together with 40 M AA. After another 10 min at 37°C, 5-LO product formation was determined as described. Results are given as mean Ϯ S.E., n ϭ 5.

GPx-1 Activity, Peroxide Formation, and 5-LO Product Synthesis in Various Cell
Types-The activities of GPx-1 and cellular 5-LO as well as the capacity to elevate the cellular peroxide tone (in response to 60 M AA) were determined in intact isolated PMNL, MM6 and RBL-1 cells. Cellular 5-LO product synthesis was induced either by stimulation with 60 M AA after chelation of extracellular Ca 2ϩ by EDTA in order to detect Ca 2ϩ -independent 5-LO activity, or alternatively with ionophore and 10 M AA in the presence of Ca 2ϩ (compare Ref. 30). As shown in Table I, PMNL and RBL-1 cells exert only low GPx-1 activity and are clearly capable of elevating the cellular redox tone by producing peroxides in response to AA. Interestingly, when the culture medium of RBL-1 cells has been supplemented with 100 ng/ml Se 4ϩ , a determinant for GPx-1 expression (43), GPx-1 activity was increased about 6.7-fold, and at the same time the peroxide level in AA-stimulated cells was considerably reduced. Compared with PMNL and RBL-1 cells, the GPx-1 activity was about 17-and 3.8-fold higher in MM6 cells, respectively, and stimulation with AA caused an only marginal increase of the cellular peroxide level. Intriguingly, cells possessing high GPx-1 activity and thus low capacity to elevate the cellular peroxide levels (MM6 or RBL-1 with Se 4ϩ supplementation), gave low AA-induced 5-LO product synthesis, whereas 5-LO product formation was high in cells (PMNL and RBL-1 cells) exhibiting low GPx-1 activity and enhanced peroxide levels. Notably, 5-LO product synthesis was substantial in all cell types, when Ca 2ϩ levels were elevated by stimulation with ionophore.
Effects of Ebselen on 5-LO Product Synthesis in PMNL-Ebselen, a cell-permeable organo-selenium compound that mimics GPx activity in the presence of thiols (44), was described as an inhibitor of 5-LO (45,46). PMNL were preincubated with ebselen and 5-LO product synthesis was induced either by stimulation with ionophore in the presence of 10 M AA and 1 mM Ca 2ϩ , or with 60 M AA in the presence of 1 mM EDTA and 30 M BAPTA/AM (in order to remove Ca 2ϩ ). Under conditions where Ca 2ϩ is elevated in the cell by ionophore stimulation, the IC 50 value of ebselen was 12.6 M (Fig. 6). However, removal of Ca 2ϩ leads to a considerable shift of the IC 50 value to 1.2 M ebselen.
Interaction of Ca 2ϩ with the C2 Domain Protects 5-LO Against Inhibition by GPx-1-It appeared possible that Ca 2ϩ could suppress the activity of GPx-1, thereby rendering 5-LO activity resistant against GPx-1. However, we found that Ca 2ϩ does not significantly suppress GPx-1 activity in MM6-ATP-PT or the activity of GPx-1 isolated from bovine erythrocytes (not shown), implying that Ca 2ϩ rather protects 5-LO against GPx-1, instead of abolishing GPx-1 activity. In order to investigate if a functional C2 domain is important for protection of 5-LO against GPx-1 activity by Ca 2ϩ , a mutated 5-LO (N43A, D44A, and E46A, loop2 mut-5LO, Ref. 19), which requires about 10 -100 fold higher Ca 2ϩ concentrations for binding and stimulation of 5-LO reactions, was tested. wt 5-LO and loop2 mut-5LO were purified and incubated in the absence and in the presence of 10 M Ca 2ϩ with or without soluble fractions of undifferentiated MM6 cells (which are devoid of 5-LO (36) but possess high GPx-1 activity, Ref. 12) and 5-LO activity (at 20 M AA) was determined. In the absence of Ca 2ϩ , the capacity for 5-LO product synthesis was substantial for wt-but also for loop2 mut-5LO, although the specific activity of the wt-5-LO (112.3 g 5-H(P)ETE/mg of protein) was about 3-fold greater than that of loop2 mut-5LO (36.5 g 5-H(P)ETE/mg of protein). In agreement with previous studies (19), in the presence of 10 M Ca 2ϩ , the activity of loop2 mut-5LO was unchanged, whereas the activity of wt-5-LO was increased about 2-to 3-fold. To ensure comparable capacities of 5-LO product synthesis in incubations containing either wt-or loop2 mut-5LO, the amounts of wt-5-LO and loop2 mut-5LO were adjusted to obtain comparable 5-HPETE formation. Addition of MM6 soluble fractions in the absence of Ca 2ϩ suppressed the activity of both enzymes. Of interest, in the presence of 10 M Ca 2ϩ , MM6 soluble fractions (or GPx-1 plus 10 M GSH, not shown) suppressed the activity of loop2-5LO, whereas the activity of wt-5-LO (that binds Ca 2ϩ ) was rather increased (Fig. 7A). Thus, Ca 2ϩ binding at the C2 domain protects 5-LO activity against the M-DSP or GPx-1. Importantly, the GPx inhibitors iodoacetate or mercaptosuccinate and also 13(S)-HPODE were able to restore the activity of the loop2 mut-5LO, suppressed by the MM6 soluble fraction in the presence of 10 M Ca 2ϩ (Fig. 7B). DISCUSSION We have recently published that in the absence of Ca 2ϩ , 5-LO in broken cell preparations of PMNL and RBL-1 cells has considerable catalytic activity, whereas in broken cell preparations of the monocytic cell line MM6, a soluble protein confers FIG. 4. Cellular low molecular mass components cooperate with GPx-1 to inhibit 5-LO. The S100 of PMNL was subjected to ATP-affinity chromatography to obtain partially purified 5-LO and the PMNL ATP-PT, respectively. The PMNL ATP-PT was further separated into a high (Ͼ10 kDa) and a low molecular mass (Ͻ10 kDa) fraction using a PD-10 column (Amersham Biosciences). Partially purified 5-LO, corresponding to 2 ϫ 10 6 cells, was diluted in ice-cold PBS containing 1 mM EDTA and 1 mM ATP. Then, GPx-1 (500 mU, isolated from bovine erythrocytes) and aliquots of high and low molecular mass fractions, corresponding to 10 7 cells, were added; final volume was 1 ml. After 5 min on ice, samples were prewarmed for 30 s at 37°C and CaCl 2 (2 mM) was added as indicated together with 40 M AA. After another 10 min at 37°C, 5-LO product formation was determined. The control represents 5-LO activity in the absence of GPx-1 and cellular components. Results are given as mean Ϯ S.E., n ϭ 3. (29,34). Although we were unable to completely purify this protein in an active state, we provide strong evidence that this M-DSP is GPx-1. First, the apparent size of the M-DSP (80 -100 kDa) fits well with the molecular mass (87 kDa) of GPx-1, and the M-DSP shares chromatographic properties with GPx-1 (Fig. 1A and Refs. 12 and 29). Second, mercaptosuccinate, a specific inhibitor of GPx-1 (41) as well as the broad spectrum GPx inhibitor iodoacetate (42) reversed the M-DSP-induced 5-LO inhibition in the absence of Ca 2ϩ , without significantly affecting 5-LO catalysis in the presence of Ca 2ϩ (Figs. 1-3). Similar effects were obtained with 13(S)-HPODE, which counteracts GPx activities. Third, GPx-1 activity in MM6, RBL-1 cells and PMNL is negatively correlated to the cellЈs capacity to form 5-LO products at low intracellular Ca 2ϩ -levels (Table I). Finally, GPx-1, isolated from bovine erythrocytes mimicked the inhibitory effects of the M-DSP with respect to reduction of 5-LO activity, and this GPx-1-mediated 5-LO suppression could be overcome by mercaptosuccinate, iodoacetate, or 13(S)-HPODE.

5-LO activity Ca 2ϩ -dependent
Members of the GPx family are well recognized endogenous inhibitors of LOs (5,9,10,12,15,47), which act by reducing the level of LOOH. A certain level of LOOH is required for the conversion of the active site iron from the ferrous to the ferric state thereby initializing the LO reaction (6 -8). Thus, 15-HPETE, 12-HPETE, 5-HPETE, and 13-HPODE, but not H 2 O 2 , could stimulate 5-LO in vitro (8). In accordance, 13(S)-HPODE but not H 2 O 2 counteracted M-DSP-or GPx-1-induced 5-LO suppression. Although GPx-4 has been reported to inhibit 5-LO activity in differentiated HL-60 and BL41-E95-A cells (11), RBL-2H3 cells (14) and A431 cells (13), it is unlikely that this peroxidase is the M-DSP, since GPx-4 is a monomeric 19 kDa protein (48) that is not sensitive to mercaptosuccinate (5) and requires millimolar concentrations of thiols for efficient inhibition of 5-LO activity (11,12). Interestingly, Coffey et al. (49) reported about a cytosolic protein of mononuclear phagocytes that reduced 5-LO activity in broken cell preparations and prolonged the lag phase of soybean LO, which was reversed by addition of 13(S)-HPODE. Recently we showed that in MM6 cells, GPx-1 but not GPx-4 is involved in the regulation of cellular 5-LO activity (12). For the classical, high turnover rate reduction of peroxides, GPx-1 utilizes GSH and to a lesser extent dithiothreitol in the millimolar range (50,51). Surprisingly, in this as well as in our previous study (12), reduction of 5-LO activity by GPx-1 occurred also in the absence of millimolar concentrations of GSH or dithiothreitol, and was supported also by 0.5 mM ␤-mercaptoethanol (12) (which functions as a less efficient cofactor for classical peroxidase activity). Also, protection of 5-LO against inactivation during storage by GPx-1 was supported best by ␤-mercaptoethanol (46). The reduction of 5-LO activity by M-DSP did not depend on millimolar concentrations of thiols. Thus, sole addition of the 80 -100 kDa fraction from gel-permeation chromatography, which should not contain endogenous thiols of low molecular mass, suppressed 5-LO activity in the absence of Ca 2ϩ . Presumably, after cell lysis, GPx-1 exists in the reduced selenol form, the species that is responsible for rapid reduction of hydroperoxides. Such hydroperoxide reduction leads to an oxidation of the active-site selenol to selenenic acid, which can oxidize GSH to recover the active selenol form of GPx-1 (52). Possibly, this selenol species could become oxidized during the extended time-consuming purification procedures, explaining why the activity of the M-DSP was lost. Also, GPx-1, purified from erythrocytes, may exist in the selenenic form, explaining why GSH (at least in the micromolar range), supernatants containing traces of thiols, or cellular low molecular mass components were necessary for suppression of 5-LO in the absence of Ca 2ϩ . However, for PMNL, which possess only low GPx levels (Table  I), 5-LO in S100 was only suppressed when GSH in the millimolar range was present, but also under these experimental settings, Ca 2ϩ could protect 5-LO against GPx activity (Fig.  2B). The finding that Ca 2ϩ is required to allow 5-LO activity in the presence of GPx-1 in vitro, may also be relevant for 5-LO catalysis in intact cells. In fact, the capacities of MM6 cells to form 5-LO products correlate with the intracellular Ca 2ϩ levels (Fig. 5, Table I, and Ref. 29). Thus, when the Ca 2ϩ supply was substantial, 5-LO product synthesis was prominent, and elimination of GPx activity by iodoacetate or 13(S)-HPODE caused no further enhancement. However, when cells were depleted of extracellular Ca 2ϩ by chelation, counteraction of GPx-1 potently enhanced the low 5-LO product synthesis. Similarly, only Se 4ϩ -supplemented RBL-1 cells (possessing high GPx-1 activity) required elevation of intracellular Ca 2ϩ for prominent 5-LO product synthesis, whereas the Ca 2ϩ levels played a minor role in cells exerting low GPx-1 activity (Table I). Finally, ebselen, which mimics GPx-1 activity in the presence of thiols (44), suppressed 5-LO activity much more efficiently in PMNL when Ca 2ϩ levels were reduced, as compared with conditions where Ca 2ϩ was elevated (Fig. 6).
Enzymatic active 5-LO forms the LOOH 5-HPETE and it appeared possible that the effectiveness of GPx-1 to suppress 5-LO activity depends on the ratio of GPx-1 capacity to 5-LO capacity. However, increasing the amounts of 5-LO in the incubation mixtures could not impair the degree of 5-LO inhibition, indicating that the effectiveness of GPx-1 to inhibit 5-LO was seemingly independent of the amounts of 5-LO products (5-HPETE) formed during catalysis. Also, in the incubations of wild-type and loop2 mut-5LO the absolute amounts of 5-LO products formed were about the same, but the enzyme activities were differentially suppressed by GPx-1. Presumably, instead the affinity of 5-LO for LOOH prior to the initiation of the 5-LO reaction determines enzyme activation.
The underlying molecular mechanism of how Ca 2ϩ renders 5-LO resistant against GPx-1 activity is not clear. Although conceivable, our data show that the 5-LO protective effect of Ca 2ϩ is not related to a direct suppression of GPx-1 activity by Ca 2ϩ . Another possibility could be that Ca 2ϩ directly facilitates the conversion of the active site iron from the ferrous to the active ferric state at suboptimal LOOH concentrations, for example by alteration of the iron coordination, in particular by the flexible ligands His-367, Asn-554, or a putative water molecule (7). However, no significant effects of Ca 2ϩ regarding the redox state of the active site iron were observed when 5-LO was investigated by electron paramagnetic resonance spectroscopy (7), and based on a proposed model of 5-LO (19), the Ca 2ϩ binding sites are rather distant from the active site iron.
Ca 2ϩ binds to the N-terminal C2 domain of the 5-LO enzyme (19) and such Ca 2ϩ binding increases the hydrophobicity of the protein (19) allowing an association with phospholipids or cellular membranes (27,28). Mutation of Asn-43, Asp-44, and Glu-46 to alanine within the C2 domain of 5-LO (loop2 mut-5LO) caused decreased Ca 2ϩ binding and a requirement for higher Ca 2ϩ concentrations to stimulate enzyme activity (19). Interestingly, in contrast to wt-5-LO, Ca 2ϩ was not able to protect the loop2 mut-5LO against the effect of M-DSP or GPx-1 (Fig. 7A). It should be noted that wild type and loop2 mut-5LO exert no significant variations in the uninhibited (no GPx) enzyme activities, neither in the absence nor in the presence of Ca 2ϩ . Again, mercaptosuccinate and 13(S)-HPODE, but not Ca 2ϩ , were capable of counteracting the suppressed activity of mutated loop2 mut-5LO (Fig. 7B). Thus, the Ca 2ϩ binding sites within the C2 domain seem to mediate the protective effects of Ca 2ϩ against GPx-1 and the M-DSP. It was shown that Ca 2ϩ lowers the K m value of 5-LO for enzymatic conversion of AA (33,53) and preliminary results from our group indicate that Ca 2ϩ strongly increases the binding of AA to 5-LO. 2 Along these lines it is conceivable, that in analogy to AA, the Ca 2ϩ -mediated increase in hydrophobicity of 5-LO could augment also the affinity toward LOOH, thus allowing 5-LO activation at lower LOOH levels. Experimental data (7,37,53,54) suggest that 5-LO (as other LOs, Ref. 55) may have two fatty acid binding sites, one catalytic and one regulatory, where the latter may be the primary site for LOOH binding. Interestingly, nonredox-type 5-LO inhibitors, which presumably act at such a regulatory fatty acid (LOOH) binding site, require low LOOH levels or GPx activity, respectively, for efficient 5-LO inhibition (37) in relation to Ca 2ϩ is in progress in our laboratory.
Taken together our data suggest that Ca 2ϩ binding at the C2 domain facilitates 5-LO activation at low LOOH levels, which are controlled by GPx-1. In such a scenario, Ca 2ϩ may lead to an increased affinity of 5-LO for LOOH at a putative regulatory fatty acid binding site. 5-LO can be phosphorylated at serine residues by MAPKAPK-2 and ERKs (30,32), and stimulation of these kinases activated 5-LO in intact cells in the absence of Ca 2ϩ (30,31,34). Possibly, also phosphorylation could reduce the requirement of LOOH, it is tempting to speculate that Ca 2ϩ and phosphorylation may act together to activate 5-LO, by reducing the LOOH requirement for conversion of the active site iron from the ferrous to the ferric form. Alternatively, as discussed before (30), phosphorylation might lead to a small pool of active cellular 5-LO, which could form LOOH, and thus activate the bulk of 5-LO in the cell. In this context it is of interest that phosphorylation-induced 5-LO activity was rather resistant against non-redox-type 5-LO inhibitors (56), which require low LOOH level for efficient 5-LO inhibition (37).