Membrane Localization and Topology of Leukotriene C4 Synthase*

Leukotriene C4(LTC4) synthase conjugates LTA4 with GSH to form LTC4. Determining the site of LTC4synthesis and the topology of LTC4 synthase may uncover unappreciated intracellular roles for LTC4, as well as how LTC4 is transferred to its export carrier, the multidrug resistance protein-1. We have determined the membrane localization of LTC4 synthase by immunoelectron microscopy. In contrast to the closely related five-lipoxygenase-activating protein, LTC4 synthase is distributed in the outer nuclear membrane and peripheral endoplasmic reticulum but is excluded from the inner nuclear membrane. We have combined immunofluorescence with differential membrane permeabilization to determine the topology of LTC4 synthase. The active site of LTC4 synthase is localized in the lumen of the nuclear envelope and endoplasmic reticulum. These results indicate that the synthesis of LTB4 and LTC4 occurs in different subcellular locations and suggests that LTC4 must be returned to the cytoplasmic side of the membrane for export by multidrug resistance protein-1. The differential localization of two very similar integral membrane proteins suggests that mechanisms other than size-dependent exclusion regulate their passage to the inner nuclear membrane.


Leukotriene C 4 (LTC 4 ) synthase conjugates LTA 4 with GSH to form LTC
. Determining the site of LTC 4 synthesis and the topology of LTC 4 synthase may uncover unappreciated intracellular roles for LTC 4 , as well as how LTC 4 is transferred to its export carrier, the multidrug resistance protein-1. We have determined the membrane localization of LTC 4 synthase by immunoelectron microscopy. In contrast to the closely related five-lipoxygenase-activating protein, LTC 4 synthase is distributed in the outer nuclear membrane and peripheral endoplasmic reticulum but is excluded from the inner nuclear membrane. We have combined immunofluorescence with differential membrane permeabilization to determine the topology of LTC 4 synthase. The active site of LTC 4 synthase is localized in the lumen of the nuclear envelope and endoplasmic reticulum. These results indicate that the synthesis of LTB 4 4 must be returned to the cytoplasmic side of the membrane for export by multidrug resistance protein-1. The differential localization of two very similar integral membrane proteins suggests that mechanisms other than size-dependent exclusion regulate their passage to the inner nuclear membrane.
A major strategy used by cells to prevent the formation of these bioactive molecules under resting conditions is compartmentalization of the biosynthetic enzymes. In circulating, quiescent peripheral blood cells, both cytoplasmic phospholipase A 2 and 5-LO are localized to the cytoplasmic compartment, whereas FLAP is localized to the inner and outer nuclear membrane, preventing the release of AA and the formation of LTs. When leukocytes and mast cells are activated, cytoplasmic phospholipase A 2 translocates to the nuclear envelope. At the same time, 5-LO translocates to the inner and outer nuclear membranes. The association of 5-LO with these membranes is regulated, in part, by FLAP (7,16). LTC 4 synthase and FLAP are closely related 17-kDa transmembrane proteins, and both are predicted to have three hydrophobic domains. The region that extends from the first predicted hydrophilic loop to the third hydrophobic domain is highly homologous between the two proteins (Fig. 1). The localization of LTC 4 synthase has not been characterized in detail. FLAP is distributed between the inner and outer nuclear membranes and peripheral ER. The first hydrophilic loop of FLAP was localized to the lumen of the nuclear membrane (7), but the topology of the second hydrophilic loop and the C and N termini of FLAP are unknown, although a model with three transmembrane domains was proposed (8) (Fig. 1A). The first hydrophilic loop of LTC 4 synthase binds LTA 4 (8,17). GSH conjugation is determined by amino acid residue Tyr 93 in the second hydrophilic loop, suggesting that the hydrophilic loops are on the same, cytoplasmic side of the membrane (Fig. 1B). No experimental determination of the topology has been made to support this model, and other relationships between the hydrophilic loops and the C-and N termini may exist ( Fig. 1, C-G).
We have analyzed the distribution and topology of LTC 4 synthase using a combination of immunofluorescence, confocal, and electron microscopy combined with differential membrane permeabilization by streptolysin O (SLO). Surprisingly, and in contrast to the closely related FLAP protein, LTC 4 synthase is distributed on the outer nuclear membrane and in the peripheral ER but is strictly excluded from the inner nuclear membrane. Furthermore, the active site of LTC 4 synthase is localized in the lumen of the nuclear envelope and ER. These results support the possibility that the intracellular synthesis of LTB 4 and LTC 4 may be differentially compartmentalized and that LTC 4 is synthesized on the luminal face of the ER membrane from where it must be returned to the cytoplasm for eventual export by MRP. In addition, the differential localization of two integral membrane proteins of essentially the same size, with a high degree of identity, indicates that mechanisms other than size-dependent exclusion regulate the passage of integral membrane proteins to the inner nuclear membrane.

EXPERIMENTAL PROCEDURES
Plasmid Constructs-LTC 4 synthase cDNA was cloned by reverse transcriptase-PCR. mRNA was prepared from KG-1 cells with Triazol (Invitrogen), and first strand cDNA synthesis was performed using poly(dT) primer and Superscript II reverse transcriptase (Invitrogen). The coding region of LTC 4 synthase was amplified using a forward primer corresponding to bp 67-94 with an EcoRI site and bases TCAG added to the 5Ј-end. The reverse primer corresponded to bp 527-510 with a BamHI site and GTCGAT added. The Expand High Fidelity PCR System (Roche Molecular Biochemicals) was used with a 250 M concentration of the primers and 1.5 mM MgCl 2 . Cycling conditions were as follows: 94°C, 1 min; 55°C, 1 min; and 72°C, 1 min (30 cycles) followed by 1 cycle with a 10-min extension time. The product was subcloned into pcDNA3.1(Ϫ)Myc-His (Invitrogen), which places a Myc epitope on the C terminus, or into pCMV5 (Sigma) to place a FLAG epitope on the N terminus. The full-length coding region of CYP4F3A cDNA (18) was subcloned into the EcoRI site of pcDNA3 (Invitrogen) or into pCMV5.
Cell Culture and Transfections-COS-7 cells were cultured to 50% confluence in Dulbecco's modified Eagle's medium with 10% fetal bovine serum on two-chambered slides. CHO cells were cultured in Ham's F-12 medium with 10% serum. COS cells were transfected using Superfect (Qiagen) and 2 g of plasmid/chamber. Transfection efficiencies were ϳ50%. Alternatively, CHO cell or COS cell transfections were performed with the ProFection system (Promega) using 5 g of plasmid; the efficiencies were ϳ10 -15%. RBL-2H3 cells were a gift of Dr. Howard Katz (Division of Rheumatology and Immunology, Brigham and Women's Hospital).
18 h after transfection, cells were washed three times in PBS, fixed in 4% paraformaldehyde in PBS for 30 min, incubated with 50 mM ammonium chloride for 10 min, and washed. To permeabilize all membranes, they were treated with 0.1% Triton X-100 in PBS for 4 min, washed three times in PBS, and blocked in 10% goat serum for 30 min prior to incubations with antibodies. For plasma membrane permeabilization, cells were washed three times in serum-free medium and then in HCMF ϩϩ (10 mM Hepes, pH 7.4, containing 137 mM NaCl, 5 mM KCl, 5 mM glucose, 0.3 mM Na 2 HPO 4 , supplemented with 1 mM Ca 2ϩ and Mg 2ϩ ). The cells were then incubated on ice for 10 min in HCMF ϩϩ containing 0.4 units/ml SLO (preactivated with dithiothreitol). The cells were washed on ice three times with HCMF ϪϪ and two times with 30 mM PIPES, pH 7.0, 10 mM NaCl, 30 mM potassium acetate, 3 mM magnesium acetate, 0.6 mM CaCl 2 , 1 mM EGTA containing 1 mM dithiothreitol (20). They were warmed to 37°C for 10 min, fixed with 4% paraformaldehyde for 30 min, and incubated in 50 mM ammonium chloride and 10% goat serum. Slides were treated with primary antibody for 1 h at room temperature, washed three times, and probed with secondary antibody for 1 h. They were washed, sealed under Gel/Mount (Biomeda), and analyzed. RBL cells were grown on slides, fixed, and stained directly.
Microscopy-Immunofluorescence microscopy was performed using a Nikon FXA photomicroscope and IP Spectrum (Scanalytics, Vienna, VA) acquisition analysis software. Data were imported to PowerPoint and Adobe Photoshop version 5.0 software. Confocal images were obtained using a Radiance 2000 microscope controlled by an Optix Pentium 5133 Dell computer. For electron microscopy, RBL cells were fixed for 1 h at room temperature in 4% paraformaldehyde (Electron Microscopy Sciences, Ft. Washington, PA) in 0.05 M sodium phosphate buffer, pH 7.4. Cells were rinsed in PBS and then scraped and pelleted (10 min at 2,000 rpm). The pellet was resuspended in a small amount of warm 2% agarose (Sigma) and allowed to harden; it was then cut into very small blocks, and these were sucrose-protected in 2.3 M sucrose in PBS. Cryosections (70 nm) were cut on a Leica EM FCS at Ϫ80°C. They were collected on Formvar-coated nickel grids and floated on PBS.
For EM and immunolabeling, grids were blocked on drops of PBS plus 5% normal goat serum and 1% bovine serum albumin for 10 min. They were incubated on drops of primary antibody (1:50 in DAKO diluent, DAKO Corp., Carpinteria, CA) or diluent alone for 2 h at room temperature. After rinsing on drops of PBS, the grids were incubated on drops of goat anti-rabbit gold (10 nm, 1:5 in DAKO diluent; Electron Microscopy Sciences, Fort Washington, PA) for 1 h at room temperature. The grids were floated on drops of tylose plus uranyl acetate for 10 min and then allowed to dry. The grids were examined in a Philips CM 10 TEM at 80 kV. 4 Synthase-We initially analyzed the subcellular localization of endogenous LTC 4 synthase in RBL-2H3 cells. These cells were chosen because they express high levels of endogenous LTC 4 synthase and generate LTC 4 after stimulation with IgE or ionophore. The cells were permeabilized with 0.1% Triton X-100 and analyzed by conventional immunofluorescence. As shown in Fig. 2A, the enzyme was detected with a perinuclear staining pattern that extended out toward the peripheral ER, a distribution typical of intrinsic ER proteins. We next compared the distribution of LTC 4 synthase-Myc-His expressed in CHO cells by conventional (Fig. 2B) and confocal microscopy (Fig. 2, C and E) with that of the inner nuclear membrane protein lamin A/C (Fig. 2D). Transfected cells were fixed, permeabilized with Triton X-100, and then probed with both anti-Myc monoclonal antibody and anti-lamin A/C polyclonal antibody (Fig. 2, B-E), followed by rhodamineconjugated donkey anti-mouse IgG and fluorescein-conjugated goat anti-rabbit secondary antibody. When the localization of LTC 4 synthase was determined by immunofluorescent microscopy, the highest intensity of rhodamine staining was detected at the nuclear rim (Fig. 2B); however, the Myc epitope was detected with a distribution consistent with that of the peripheral ER and with a pattern matching that of the endogenous protein. To further analyze the nuclear envelope, we performed confocal microscopy. In this case, a clear nuclear rim was observed (Fig. 2C), but additional staining was observed extending outwards with a pattern consistent with an ER distribution. Lamin A/C was also detected with a clear green ring around the nucleus (Fig. 2D). Digital overlay indicated a yellow rim around the nucleus and indicates co-localization of LTC 4 synthase and lamin A/C at this level (Fig. 2E). However, as discussed below, these findings cannot discern whether LTC 4 synthase is present in both the inner and outer nuclear membrane.

Localization of LTC
Membrane Topology of CYP4F3A-To determine the membrane topology of LTC 4 synthase, we used a strategy based on the ability of SLO to permeabilize the plasma membrane as opposed to intracellular membranes. Triton X-100 will permeabilize all cellular membranes including the ER and both membranes of the nuclear envelope. Thus, antigenic epitopes that are in the cytosol are revealed by treatment with SLO, but not those oriented to the ER lumen or contained within the nucleus; these require Triton X-100 to make them accessible to immunological detection. To establish this method, we employed a protein with a known topology that is a member of a class of proteins in which this approach has been successfully used. The protein chosen was CYP4F3A. CYPs are localized to the ER and possess a single transmembrane domain with the N terminus oriented to the lumen (21). We generated a functional FIG. 2. Immunofluorescent localization of LTC 4 synthase and the topology of the C and N termini. Endogenous LTC 4 synthase was localized in RBL cells (A) using primary antibody to hydrophilic loop 1. Alternatively, CHO cells expressing LTC 4 Myc-His were permeabilized with Triton X-100 (B-E) or SLO (F and G). Transfected CHO cells were probed with mouse monoclonal anti-Myc antibody and rabbit anti-lamin A/C antibody followed by Texas Red-conjugated donkey anti-mouse IgG and fluoresceinated goat anti-rabbit IgG (B-G). Alternatively, the cells were transfected with pCMV5-LTC 4 synthase, permeabilized with streptolysin O, and probed with anti-FLAG antibody followed by rhodamineconjugated anti-mouse secondary antibody (H). The cells were analyzed for rhodamine (A, B, C, F, and H), or fluorescein (D and G) fluorescence or by digital overlay (E). Cells in A, B, and F-H were analyzed by conventional immunofluorescence; cells in C-E were analyzed by confocal microscopy. variant of CYP4F3A that contained an N-terminal FLAG epitope. This allowed detection of the luminal N terminus using anti-FLAG antibody and the cytoplasmic C terminus using affinity-purified anti-CYP4F3A-(410 -520) antibody generated in our laboratory. COS cells were transfected with the pCMV5-CYP4F3A plasmid, fixed, and permeabilized with Triton X-100. They were probed simultaneously with fluoresceinated anti-CYP4F3A-(410 -520) to detect the cytoplasmic portion of the molecule and anti-FLAG antibody to detect the N terminus. The FLAG epitope was detected with Texas Red-conjugated donkey anti-mouse IgG. When analyzed for fluorescein and rhodamine (Fig. 3, A and B), both the C and N termini showed an identical pattern consistent with ER localization. COS cells were then transfected and selectively permeabilized with SLO and probed with both antibodies. The cells demonstrated a pattern consistent with ER localization when analyzed for flu-

FIG. 4. Membrane topology of hydrophilic loops 1 and 2 of LTC 4 synthase. CHO cells expressing LTC 4 synthase-Myc-
His were fixed and permeabilized with 0.1% Triton X-100 (A-C) or streptolysin (D-F). Cells were stained with anti-loop 1 antibody (A and D) and antibody to loop 2 (B and E) followed by fluoresceinated secondary antibody. Alternatively, cells were probed with antibody to nucleoredoxin followed by Texas Red-conjugated secondary antibody (C and F).

FIG. 5. Differential membrane topologies of the C-terminal and hydrophilic loop 1 of LTC 4 synthase. CHO cells expressing LTC 4 synthase Myc-His
were fixed, permeabilized with Triton X-100, and then simultaneously probed with anti-Myc antibody and anti-loop 1 antibody. Anti-loop 1 antibody was detected using fluorescein-conjugated secondary antibody (A), and the Myc epitope was detected using rhodamine-conjugated secondary antibody (B). In parallel, cells were probed simultaneously with antibody to loop 2 and to the Myc epitope (C and D). The cells were then examined under rhodamine (D) or fluorescein (C). Alternatively, the cells were permeabilized with SLO, fixed, and then simultaneously probed for loop 1 and the Myc epitope (E and F). orescein (Fig. 3C). However, no fluorescence was detected when they were analyzed for rhodamine (Fig. 3D), indicating that the FLAG epitope was not accessible to antibody. The detection of the cytoplasmic C terminus, but not the N terminus, after treatment with SLO confirms that this approach is effective in determining the topology of protein domains relative to the peripheral ER and nuclear envelope.
Topology of the C and N Termini of LTC 4 Synthase-We employed differential permeabilization to determine the topology of the C terminus of LTC 4 synthase. CHO cells that expressed LTC 4 synthase Myc-His were treated with SLO using the conditions established for CYP4F3A. The cells were then simultaneously probed for the Myc epitope and lamin A/C as described above. The same staining pattern was observed for the Myc epitope when cells were permeabilized with SLO (Fig.  2F). However, lamin C was not detected (Fig. 2G). These results indicated that the C terminus of LTC 4 synthase and the C terminus of CYP4F3A shared the same relationship (cytosolic) to the ER and outer nuclear envelope. To determine the topology of the N terminus, a FLAG-LTC 4 synthase vector was created that places a FLAG epitope on the N terminus of LTC 4 synthase. This vector was expressed in CHO cells, and differential permeabilization with SLO was performed. In these experiments (Fig. 2H), the FLAG epitope was detected, indicating its cytoplasmic orientation.
We next determined the orientation of the two hydrophilic loops of the active site. When CHO cells expressing LTC 4 synthase-Myc-His were permeabilized with Triton X-100 and analyzed by immunofluorescent microscopy (Fig. 4, A-C), both loop 1 (Fig. 4A) and loop 2 (Fig. 4B) were detected. In addition, the nuclear protein nucleoredoxin (Fig. 4C) was detected (as was lamin; data not shown). In these experiments, the nonspecific background was secondary to antisera, and the dark red staining nucleus was clearly visible, with nucleoli visible as dark holes. In contrast, neither loop 1 nor loop 2 could be detected following permeabilization with SLO (Fig. 4, D-F). When cells permeabilized with SLO were probed with antinucleoredoxin antibody, the background remained unchanged, whereas the nuclei appeared black secondary to a complete lack of staining. These results localize the active site of LTC 4 synthase to the same luminal side of the membrane as the N terminus of CYP4F3A, and the opposite side of the membrane as the N-and C termini of LTC 4 synthase.
We next used a second, intramolecular approach to confirm these observations. Cells were transfected with LTC 4 synthase Myc-His, fixed, permeabilized with Triton X-100, and then simultaneously probed with anti-Myc and antibodies to either loop 1 or loop 2. As shown in Fig. 5, these antibodies gave the same intracellular distribution when analyzed under fluorescein (loop 1 or 2) or rhodamine (Myc epitope) (Fig. 5, A-D).  F) were expressed in CHO cells and analyzed after permeabilization with 0.1% triton X-100 (TX-100) or after extraction with 1% Triton X-100 and NaCl (Extraction). Alternatively, they were analyzed before and after extraction using antibody to lamin A/C (B and E). In the latter experiments, nonextracted cells were stained with fluoresceinated secondary antibody, and extracted cells were stained with rhodamine-conjugated secondary antibody. When cells were permeabilized with SLO and analyzed simultaneously for loop 1 and for the C-terminal Myc epitope, only the C-terminal Myc epitope was detected (Fig. 5, E and F).
We next determined whether LTC 4 synthase possessed physical properties similar to an ER protein. ER or nuclear envelope proteins that are not linked to cytoskeletal or nucleoskeletal structures are completely extracted from membranes with a combination of 1% Triton X-100 and 0.35 M NaCl (22). As shown in Fig. 6, both LTC 4 synthase-EGFP and CYP4F3A-EGFP were both visualized with a characteristic ER pattern (Fig. 6, A and C) and were completely extracted with 1% Triton X-100 and 350 mM NaCl (Fig. 6, D and F). When cells were fixed and probed with antisera to lamin A/C before (Fig. 6B) and after (Fig. 6E) extraction with Triton X-100 and 350 mM NaCl, the protein was detectable. These results indicate that LTC 4 synthase does not interact with any structural protein in any cellular compartment in a manner similar to lamin and lamin-associated protein. These experiments characterize a molecule localized to the peripheral ER and the outer nuclear envelope that, based on its size of 17 kDa and relationship to FLAP, would be expected to freely distribute to all contiguous domains of the ER. To determine whether LTC 4 synthase could also distribute to the inner nuclear membrane, we identified the ultrastructural localization of endogenous LTC 4 synthase in RBL cells. Surprisingly, endogenous LTC 4 synthase was distributed throughout the ER and was also localized to the outer nuclear envelope but was excluded from the inner nuclear membrane in any section or cell (Fig. 7). Only very rare gold particles in occasional cells were detected in the inner membrane. These findings are strikingly different from the localization of FLAP, which was shown to be preferentially localized to the inner nuclear membrane.

DISCUSSION
Endogenous or overexpressed LTC 4 synthase show the same characteristic dense staining in the nuclear envelope extending to the peripheral ER (Fig. 2), which was clearly distinct from that of lamin A/C. The digital overlap with lamin showed a bright yellow rim surrounding the nucleus. Since lamin A/C was restricted to a tight nuclear rim, this suggested that the distribution of LTC 4 synthase extended well into the peripheral ER. This distribution was seen for FLAP in peripheral blood human monocytes (7). When examined by EM, FLAP was distributed almost equally to the inner and outer nuclear membrane (7). Furthermore, 5-LO, which was localized in the cytosol of resting monocytes and neutrophils, was associated with the outer and inner nuclear membrane after cell activation. This movement of 5-LO through the nuclear pore has been well characterized (23,24) and is mediated by a nuclear localization sequence at the C-terminal of the molecule (24,25). Recently, LTA 4 hydrolase has been shown to have the potential to move through nuclear pores and target to the nucleus in RBL cells (26), indicating that LTB 4 would be synthesized in the nuclear compartment. It was not determined whether LTA 4 hydrolase was translocated to the cell membrane after activation. LTC 4 synthase and FLAP share 52% identity between amino acids 41 and 97 of FLAP and amino acids 45-101 of LTC 4 synthase, which includes the two loops of the active site. Because of this high identity and the small 17-kDa size of these two proteins, it was highly surprising that the distribution of LTC 4 synthase between the inner and outer nuclear membrane was distinct from that of FLAP (Fig. 7). This restriction of LTC 4 synthase to the outer nuclear membrane combined with the observation that LTB 4 may be made within the nucleus suggests that the synthesis of LTB 4 and LTC 4 can be differentially compartmentalized within cells. This points to potential differences in their role in intracellular function and intracellular trafficking. Various studies have suggested a role for LTB 4 in gene transcription as a ligand for peroxisome proliferator-activated receptor ␥ (27). The exclusion of LTC 4 synthase from the inner nuclear membrane suggests that LTC 4 is less likely than LTB 4 to play an intracellular role in modulating nuclear function and in transcriptional regulation.
It is generally accepted that after their synthesis and insertion in the ER, integral membrane proteins become localized to the inner nuclear membrane by lateral diffusion through the proteolipid bilayer of the outer nuclear membrane followed by diffusion around the nuclear pore (28,29). In this model, the proteins are subsequently immobilized in the inner membrane by binding to immobile nucleoskeletal or nucleoplasmic ligands or by multimerization. The main mechanism excluding proteins from entry into the inner membrane is based on size, so that proteins with cytosolic/nucleoplasmic domains greater than 70 kDa fail to localize to the inner nuclear membrane (28). Proteins that do not fall into either category would be potentially free to diffuse between all contiguous membrane domains and be equally represented in the inner and outer nuclear membrane. FLAP fulfills these postulates and fits in the latter group, being equally represented in the inner and outer membrane (7). LTC 4 synthase contradicts them, being a small protein essentially the same size as FLAP but being excluded from the inner nuclear membrane. Thus, a different mechanism must exist that allows FLAP free entry into the inner nuclear membrane and excludes LTC 4 synthase. This includes the possibility that the primary sequence of FLAP contains a signal that allows it entry to the inner membrane or that LTC 4 synthase contains a sequence that signals its exclusion.
The orientation of the N and C termini to the cytoplasm (Fig.  2) and the loops of the active site to the ER lumen (Figs. 4 and 5) support the model of membrane topology shown in Fig 1F. As described above, the release of LTC 4 is dependent on its export from cells by the MRP-1 protein, which translocates its substrates from the cytosol to the extracellular space (12)(13)(14)(15). LTC 4 does not diffuse across membranes and must reenter the cytoplasmic compartment to be accessible to the MRP-1 protein. The mechanism by which this occurs is not known. LTC 4 is formed at the luminal face of the ER, where GSH concentrations are 2-3 mM (30). GSH within the ER has been suggested to play a role mostly as a redox buffer controlling the state of sulfhydryl bonds (31). Our data suggest that GSH in the ER can serve as an enzymatic substrate in additional reactions. Both prostaglandin H synthases have also been shown to have their active site oriented toward the ER lumen. As for prostaglandin endoperoxides generated by prostaglandin H synthase-1 and -2 (32), how LTC 4 moves from the luminal surface to be accessible to intracellular transport and export by the MRP protein remains an open question.