Organelle Proteomics

Natural killer (NK) cells and cytotoxic T lymphocytes eliminate virally infected and transformed cells. Target cell killing is mediated by the regulated exocytosis of secretory lysosomes, which deliver perforin and proapoptotic granzymes to the infected or transformed cell. Yet despite the central role that secretory lysosome exocytosis plays in the immune response to viruses and tumors, little is known about the molecular machinery that regulates the docking and fusion of this organelle with the plasma membrane. To identify potential components of this exocytic machinery we used proteomics to define the protein composition of the NK cell secretory lysosome membrane. Secretory lysosomes were isolated from the NK cell line YTS by subcellular fractionation, integral membrane proteins and membrane-associated proteins were enriched using Triton X-114 and separated by SDS-PAGE, and tryptic peptides were identified by LC ESI-MS/MS. In total 221 proteins were identified unambiguously in the secretory lysosome membrane fraction of which 61% were predicted to be either integral membrane proteins or membrane-associated proteins. A significant proportion of the proteins identified play a role in vesicular trafficking, including members of both the Rab GTPase and soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) and protein families. These proteins include Rab27a and the SNARE vesicle-associated membrane protein-7, both of which were enriched in the secretory lysosome fraction and represent potential components of the machinery that regulates the exocytosis of this organelle in NK cells.


cells and cytotoxic T lymphocytes (CTLs) eliminate virally infected and transformed cells. Target cell
recognition by either an NK cell or a CTL promotes the formation of an immunological synapse at the point of contact between the two cells (1,2). Prestored perforin and granzymes are then secreted by the NK cell or CTL whereupon target cell death is induced. More specifically, perforin facilitates the entry of granzymes into the target cell where they induce apoptosis (3). Perforin and granzymes are stored within secretory lysosomes, specialized organelles found in both NK cells and CTLs (4). In these cells the secretory lysosome performs two distinct functions: acting as a degradative compartment and also functioning as a secretory organelle that delivers perforin and granzymes to the immunological synapse.
Exocytosis of secretory lysosomes by NK cells and CTLs involves a complex sequence of events that can be divided into three steps. The first step involves the polarization of secretory lysosomes toward the immunological synapse, next the secretory lysosomes are tethered to the plasma membrane, and in the final step they fuse with the plasma membrane. The identity of much of the molecular machinery that mediates these steps remains to be elucidated, although genetics studies in mice and humans have identified roles for a number of molecules in CTLs. The adaptor protein 3 complex is required for the movement of secretory lysosomes to the immunological synapse (5), the small GTPase Rab27a is required to tether the secretory lysosomes at the plasma membrane (6,7), whereas Munc13-4 is required immediately prior to fusion with the plasma membrane (8). Much less is known about the fusion of NK cell and CTL secretory lysosomes with the plasma membrane. Soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNAREs) are central components of the molecular machinery that catalyze membrane fusion reactions (9). Membrane fusion requires an R-SNARE on the donor membrane and either two or three Q-SNAREs on the recipient membrane; when these interact they form a trans-SNARE complex that catalyzes the fusion of the two membranes. However, the identity of the SNAREs associated with the secretory lysosome in NK cells and CTLs is unknown.
Proteomics can provide detailed information about the composition of intracellular organelles and hence reveal novel insights into their function (10). In this study we used proteom-45 min at 4°C) in a T1250 rotor (Sorvall) to separate subcellular organelles. Twenty-four 1-ml fractions were collected from the gradient. An aliquot of each fraction was added to an equal volume of 2ϫ Laemmli sample buffer and heated at 70°C for 10 min prior to SDS-PAGE and immunoblotting. Each fraction was assayed for N-acetyl-␤-D-glucosaminidase (NAGA) activity and alkaline phosphatase activity to determine the distribution of secretory lysosomes and plasma membrane, respectively, through the gradient using methods described previously (15). Fractions 21 and 22, which contained the highest NAGA activity, were pooled and centrifuged (166,000 ϫ g for 60 min at 4°C) in a TLS-55 rotor (Beckman) to pellet the membrane fraction, which was stored at Ϫ80°C. An aliquot of this secretory lysosome sample was removed for immunoblotting. In total, secretory lysosomes from 2.9 ϫ 10 9 cells were pooled for proteomics analysis.
Triton X-114 Phase Partitioning-Integral membrane proteins and membrane-associated proteins were enriched by Triton X-114 phase partitioning. Solutions of Triton X-114 are homogenous at 0°C but separate into a detergent-enriched phase and a detergent-depleted phase at temperatures above 20°C (16). Proteins partition according to their hydrophobicity with integral membrane proteins and membrane-associated proteins partitioning in the lower detergent phase and hydrophilic proteins partitioning in the upper aqueous phase (16,17). Prior to use Triton X-114 was precondensed, to remove hydrophilic molecules to obtain a more homogeneous preparation, according to the method of Bordier (16). Extraction of integral membrane proteins and membrane-associated proteins was performed according to a modification to the method of Bordier (16). Briefly 1.2 ml of 5% (v/v) Triton X-114 containing Complete EDTA-free protease inhibitor mixture at 0°C was added to the secretory lysosomal sample (3.3 ml), and the mixture was kept on ice for 30 min before being centrifuged (10,000 ϫ g for 10 min at 4°C) to remove the insoluble residue. The supernatant was added to 10 ml of 0.2 mM EDTA, 5 mM MgCl 2 , 200 mM NaCl, and 40 mM Tris-HCl (pH 7.5) and incubated for 5 min at 37°C before being centrifuged at 4,470 ϫ g for 3 min at room temperature. The upper aqueous phase was discarded, and the lower detergent phase was partitioned a further four times. Proteins in the detergent phase were precipitated by adding 20% (w/v) TCA in acetone, incubating at Ϫ20°C for 4 h, and then centrifuging at 13,000 ϫ g for 20 min at 4°C. The pellet was washed twice in 90% (v/v) acetone and then resuspended in 95 l of 2ϫ Laemmli sample buffer (0.02% (w/v) bromphenol blue, 20% (v/v) glycerol, 5% (v/v) mercaptoethanol, 6% (w/v) SDS, and 125 mM Tris-HCl (pH 6.8)) and stored at Ϫ80°C until required.
SDS-PAGE and In-gel Trypsin Digestion of Secretory Lysosome Membrane Proteins-The Triton X-114-extracted membrane proteins (95 l) were loaded into a single well of a 1-mm-thick 4% (v/v) stacking gel above a 7-20% (v/v) gradient polyacrylamide gel. SDS-PAGE was performed in an SE 600 Ruby electrophoresis unit (Amersham Biosciences) at 20 mA/gel for 2 h followed by 30 mA/gel until the bromphenol blue dye front was 0.5 cm from the bottom of the gel. After electrophoresis the gel was fixed overnight in 50% (v/v) ethanol and 3% (v/v) phosphoric acid before the separated proteins were visualized by Coomassie Brilliant Blue staining (17% (w/v) ammonium sulfate, 34% (v/v) methanol, 3% (v/v) phosphoric acid, and 0.035% (w/v) Coomassie Brilliant Blue (G-250)) for 48 h. The stained gel was destained in deionized water for 16 -18 h and then scanned. The lane from the polyacrylamide gel was cut into eight segments (A-H) at points corresponding to molecular weight markers, and then each segment was further cut into 2-4-mm slices. Gel slices were subsequently cut into 1-mm 2 pieces with a scalpel before being destained and subjected to in-gel trypsin digestion as described previously (18). Peptides were extracted from gel pieces using vigorous shaking at room temperature using a modification to the method of Wilm et al. (19) as described previously (18). Extracted peptides were dissolved in 8 l of 2% (v/v) formic acid.

Protein Identification by Nanoflow LC ESI-MS/MS-LC ESI-MS/MS
analyses were performed using a Famos/Ultimate HPLC system (LC Packings) coupled to a Q-STAR PULSARi mass spectrometer (Applied Biosystems). Tryptic peptides were loaded onto a monolithic nanocolumn (5-cm length, 100-m diameter; LC Packings) and separated with a linear gradient of 2% (v/v) acetonitrile, 0.1% (v/v) formic acid to 100% (v/v) acetonitrile, 0.1% (v/v) formic acid over a period of 20 min. Peak lists from the MS/MS spectra were generated using Mascot script 1.6b21 (Matrix Science, London, UK) for Analyst QS 1.1 (Applied Biosystems/MDS Sciex). All MS/MS spectra were centroided and deisotoped. Charge states were determined from the MS survey scan. Analysis of MS/MS data was performed by searching against the National Center for Biotechnology Information (NCBI) non-redundant database; the Mascot search program was used to identify proteins of interest (Matrix Science). Database search criteria were: taxonomy, Homo sapiens; fixed modification, carbamidomethylation; peptide tolerance, 100 ppm; maximum allowed missed cleavage, 1; variable modification, methionine oxidation; mass tolerance, 0.1 Da. Proteins with at least one unique "identity" score peptide were considered as being unambiguously identified. TMHMM-2.0 was used to identify transmembrane proteins (20,21).
Immunoblotting-YTS cells or primary NK cells (ϳ4 ϫ 10 6 cells) were washed once in PBS and then lysed on ice for 30 min in 150 mM NaCl, 10 mM Tris-HCl (pH 7.4), and 1% (v/v) Triton X-100. Cell debris were removed by centrifugation (10,000 ϫ g for 10 min at 4°C). Protein concentrations in YTS cell lysates, secretory lysosome samples, and primary NK cell lysates were determined using the Pierce BCA TM protein assay kit. Aliquots of either gradient fractions or cell lysates were added to an equal volume of 2ϫ Laemmli sample buffer and separated by SDS-PAGE. Proteins were blotted onto PVDF membranes (Amersham Biosciences) and blocked for either 1 h in 2% (w/v) nonfat dry milk in TBST (150 mM NaCl, 10 mM Tris-HCl (pH 7.5), and 0.1% Tween 20) or 3% (w/v) bovine serum albumin in TBST. Membranes were probed with primary antibodies in blocking buffer for 1 h. After washing, blots were incubated for 1 h with goat anti-mouse, goat anti-rabbit, or donkey anti-sheep secondary antibodies coupled to horseradish peroxidase (Jackson ImmunoResearch Laboratories) in blocking buffers. Blots were washed, and proteins were visualized by enhanced chemiluminescence (Pierce).
Immunofluorescence Microscopy-Primary NK cells were washed three times in PBS and adhered onto 13-mm glass coverslips coated with (3-aminopropyl)triethoxysilane. Cells were fixed with 4% paraformaldehyde in PBS for 30 min, washed four times with PBS, and permeabilized/blocked for 1 h with 0.5% saponin, 1% bovine serum albumin, and 2% fetal bovine serum in PBS (PBS-SBF). Cells were incubated with mouse monoclonal anti-CD63 in combination with either rabbit anti-syntaxin 7, rabbit anti-syntaxin 11, or rabbit anti-VAMP7 antisera in PBS-SBF for 1 h. After washing three times with PBS-SBF, the cells were incubated with Texas Red-conjugated goat anti-mouse immunoglobulin (Jackson ImmunoResearch Laboratories) and FITC-conjugated goat anti-rabbit secondary antibodies (BD Biosciences) for 1 h. The coverslips were then washed three times with PBS and mounted onto glass slides with Vectorshield mounting medium (Vector Laboratories). Antibody fluorescence was visualized with a Zeiss Axioplan 2 microscope.

Isolation of Secretory Lysosomes from the NK Cell Line
YTS-Secretory lysosomes were separated from other organelles on the basis of their size and density by subcellular fractionation (Fig. 1A). To reduce disruption of intracellular organelles, YTS cells were homogenized using a ball bearing homogenizer. The homogenate was subsequently centrifuged at 3,000 ϫ g for 10 min to generate a PNS. This speed of centrifugation was used because it was also sufficient to sediment mitochondria as evidenced by immunoblotting with an antibody specific for mitochondrial NADH-ubiquinol oxi-FIG. 1. Isolation of secretory lysosomes by subcellular fraction. A, the fractionation procedure involved four consecutive steps. 1, YTS cells were homogenized. 2, the homogenate was centrifuged to remove nuclei (N) and mitochondria (M). 3, the PNS was layered onto a Percoll gradient, and the secretory lysosomes (SL) were separated from other less dense organelles by centrifugation. 4, fractions were collected, and the secretory lysosomes were pelleted by high speed centrifugation. B, the YTS cell homogenate and PNS were immunoblotted with an antibody specific for mitochondrial NADH-ubiquinol oxidoreductase. No immunoreactivity of mitochondrial NADH-ubiquinol oxidoreductase was observed in the PNS. doreductase (Fig. 1B). The PNS was then layered on top of a self-forming Percoll gradient, and after centrifugation fractions were collected and assayed for activity of the plasma membrane marker alkaline phosphatase and the lysosomal enzyme NAGA ( Fig. 2A). Membrane-associated alkaline phosphatase activity peaked in fractions 8 -10, whereas non-membrane-associated soluble alkaline phosphatase activity was present in fractions 1-5. In contrast the peak NAGA activity was much deeper in the gradient in fractions 21 and 22, which is consistent with the high density of secretory lysosomes when compared with that of many other organelles. Therefore fractions 21 and 22 were pooled and retained for proteomics analysis.
To assess the purity of the fractions chosen for proteomics analysis, immunoblots were performed using antibodies specific to proteins localized to a range of different intracellular organelles (Fig. 2B). The presence of secretory lysosomes in fractions 21 and 22 was corroborated by immunoblotting with an antibody specific for CD63, an integral membrane protein associated with secretory lysosomes (22). CD63 was enriched in the fractions with the highest NAGA activity. In contrast, the fractions chosen for proteomics analysis showed no detectable contamination with markers of other organelles, namely calnexin (endoplasmic reticulum), TGN46 (trans-Golgi network), EEA1 (early endosome), MPR (late endosome/trans-Golgi network), and GLUT1 (plasma membrane). Instead each of these proteins was detected in less dense fractions higher in the gradient. For example the majority of GLUT1 was present in fractions 7-13, paralleling the distribution of membrane-associated alkaline phosphatase activity. These immunoblots demonstrate that the fractions chosen for proteomics analysis were highly enriched for secretory lysosomes and had little or no contamination with other organelles.
Protein Separation and LC ESI-MS/MS Identification-Because the primary aim of this study was to identify the exocytic machinery associated with the secretory lysosome membrane, integral membrane proteins and membrane-associated proteins were enriched prior to LC ESI-MS/MS by Triton X-114 phase partitioning. This enrichment step was included because peptides from highly abundant luminal proteins may have suppressed the ionization of peptides from less abundant integral membrane proteins and membraneassociated proteins. Triton X-114 solutions are homogenous at 0°C but separate into a detergent phase and an aqueous phase at 30°C (16,(23)(24)(25)(26). Integral membrane proteins and membrane-associated proteins partition predominantly into the detergent phase, whereas hydrophilic proteins partition into the aqueous phase, and proteins that do not partition into either phase form an insoluble residue (Supplemental Fig. 1) (16). Furthermore previous studies have shown that integral and membrane-associated proteins involved in vesicular trafficking partition into the detergent phase (23)(24)(25)(26). Therefore, the pooled secretory lysosome fractions were solubilized in Triton X-114, and the detergent phase was retained for proteomics analysis. The detergent phase was then separated by SDS-PAGE, and protein bands were visualized with Coomassie Brilliant Blue (Fig. 3). The lane was cut into eight segments (A-H) at points determined by the location of molecular weight

FIG. 2. Isolation of secretory lysosomes from YTS NK cells by fractionation on Percoll gradients. A, YTS cells
were subjected to ball bearing homogenization, and the resulting homogenate was centrifuged to yield a PNS that was subsequently centrifuged on a 27% Percoll gradient. A total of 24 fractions were collected from the gradient of which fractions 1 and 24 represent the top and bottom of the gradient, respectively. Fractions were assayed for NAGA activity (secretory lysosome marker) and alkaline phosphatase activity (plasma membrane marker). Secretory lysosome-enriched fractions (fractions 21 and 22) were pooled and retained for proteomics analysis. Results are expressed as mean Ϯ S.E. (n ϭ 24). B, gradient fractions were immunoblotted for calnexin (endoplasmic reticulum marker), TGN46 (trans-Golgi network marker), EEA1 (early endosome marker), MPR (late endosome and trans-Golgi network marker), GLUT1 (plasma membrane marker), and CD63 (secretory lysosome marker). Immunoblots are representative of four different experiments. markers, and then each segment was further cut into 2-4-mm slices. Following in-gel digestion of each gel slice, peptides were separated and identified by LC ESI-MS/MS.
In total, 221 proteins were unambiguously identified from one or more peptide sequences (Table I and the supplemental  table). However, there was one example when an unambiguous identification was not possible. In this instance we were unable to discriminate between the closely related proteins Rab11a and Rab11b, which share 90.8% sequence identity. Of the 221 proteins unambiguously identified, 103 were predicted to have at least one hydrophobic membrane-spanning domain, 31 had a predicted lipoprotein motif, and 87 were soluble proteins (Table I). Thus 61% of the proteins identified were likely to be integral membrane proteins or membraneassociated proteins. The remaining proteins represent, at least in part, soluble proteins that were not completely removed by Triton X-114 phase partitioning. These include granzyme B and granzyme H, which are abundant within the NK cell secretory lysosome lumen (27). Alternatively some luminal proteins may have physiochemical properties that cause them to partition into the Triton X-114 detergent phase. One such example is perforin, which although found within the secretory lysosome lumen can insert itself into the membrane of a target cell to form pores (3).
Experimental molecular mass values generally compared well with predicted values based upon cDNA sequences. However, there were some notable exceptions. Thirty-seven proteins were found to have higher than expected experimental molecular mass values. The vast majority of these proteins were either known glycoproteins or had predicted glycosylation sites in their sequences. These include LAMP1 and LAMP2, which based upon their cDNA sequences have predicted molecular mass values of 38.3 and 44.8 kDa, respectively, but due to extensive glycosylation both proteins have experimental molecular mass values of ϳ120 kDa (28). Fortyfive proteins displayed lower than expected molecular mass values; this may be because some of these proteins were in the process of being proteolytically degraded within secretory lysosomes.
Functional Classification of Identified Proteins-Functional classification of proteins was based upon information obtained from published literature and from the Swiss-Prot Protein Knowledgebase (www.expasy.org/sprot/) (Table I and (6, 7, 29 -32), whereas the three SNARE proteins identified by LC ESI-MS/MS, namely syntaxin 7, syntaxin 11, and VAMP7, are either associated with conventional lysosomes or have been implicated in NK cell function (33)(34)(35)(36)(37). Nonetheless it was important to validate the data obtained by LC ESI-MS/MS with an alternative method to demonstrate that these proteins were indeed expressed by YTS NK cells and present within the secretory lysosome fraction. To this end immunoblots of YTS cell lysates and secre-     tory lysosome fractions were probed with antibodies that recognize Rab8, Rab27a, RalA, syntaxin 7, syntaxin 11, and VAMP7 (Fig. 5). Immunoblots probed with each antibody revealed the presence of a protein of the corresponding size in both the YTS lysate and secretory lysosome fraction, hence providing independent corroboration of the data obtained by LC ESI-MS/MS. Immunoblots probed with the Rab8 antibody (which recognizes both Rab8a and Rab8b) revealed the presence of a significant proportion of one or both isoforms in the secretory lysosome fraction. The small GTPase Rab27a was enriched in the secretory lysosome fraction, which is consistent with observations in the closely related CTL cell type where this protein is associated with secretory lysosomes (7). RalA, although present, was not enriched in the secretory lysosome fraction implying that the majority of this protein was localized elsewhere in the cell. Of the three SNAREs identified, both syntaxin 7 and VAMP7 were enriched within the secretory lysosome fraction, consistent with a significant proportion of these proteins being localized to this organelle. In contrast syntaxin 11 was not enriched in the secretory lysosome fraction, suggesting that this organelle is not the principal intracellular location of this protein.
Although the YTS cell line is a well established model for NK cells, it was necessary to confirm that the proteins identified in the YTS cells are also expressed in primary human NK cells. Therefore, primary NK cell lysates were immunoblotted with Rab8, Rab27a, RalA, syntaxin 7, syntaxin 11, and VAMP7 antibodies; in each instance the corresponding protein was detected (Fig. 5). Because rabbit antisera were available to detect the SNARE proteins it was also possible to use immunofluorescence microscopy to test for co-localization of syntaxin 7, syntaxin 11, and VAMP7 with secretory lysosomes in primary NK cells by co-staining with a mouse monoclonal antibody specific for CD63 (Fig. 6). CD63 staining was concentrated into large punctate structures in the cytoplasm of the primary NK cells. The intracellular distribution of both syntaxin 7 and VAMP7 paralleled that of CD63 with a high  degree of overlap in large punctate structures. This is consistent with the localization of syntaxin 7 and VAMP7 to the secretory lysosome and corroborates the data obtained from the immunoblots of the YTS cell secretory lysosome fractions.
In contrast, although there may have been some overlap with CD63, syntaxin 11 staining was more diffuse with some immunoreactivity observed in small punctate structures (Fig. 6). This implies that the majority of syntaxin 11 is not localized to the secretory lysosome in these cells and is in agreement with the results obtained by immunoblotting.

DISCUSSION
Subcellular fractionation and proteomics represent a powerful combination when used to analyze the composition and function of intracellular organelles (10). In this study we applied these techniques to NK cell secretory lysosomes with the specific aim of identifying potential components of the exocytic machinery associated with this organelle. Subcellular fractionation of YTS NK cells was performed using self-forming Percoll gradients. Due to their high density it was possible to use this single step fractionation procedure to separate secretory lysosomes from other organelles. Immunoblotting demonstrated that the secretory lysosome fractions were highly enriched with little or no contamination of other organelles. However, LC ESI-MS/MS did identify a number of proteins not normally associated with lysosomes and lysosome-like organelles. One possible interpretation of this finding is that the secretory lysosome fraction may have been contaminated to a small degree with other organelles; this is something that we cannot exclude. However, it is important to note that the secretory lysosome is a degradative organelle, and as such many proteins not classically associated with the secretory lysosome may have been sorted to this organelle to be degraded. This process could occur either via autophagy, in which cytoplasmic components including organelles are delivered to the secretory lysosome (38), or via the endocytic pathway. Consistent with the latter, ubiquitin was identified throughout the molecular mass range of the SDS-PAGE gel suggesting that it was covalently conjugated to a number of different proteins. This is significant because ubiquitin conjugation acts as a signal for the trafficking to the lysosome for membrane proteins that are down-regulated from the plasma membrane (39).
Because the secretory lysosome is a specialized organelle that functions both as a degradative compartment and as a secretory organelle, it is not surprising that many of the proteins identified are also present in conventional lysosomes. Indeed a significant proportion of the proteins identified in this study are orthologues of proteins identified in the proteomics analysis of conventional lysosomes isolated from rat liver (40). These include LAMP1 and LAMP2, both of which are associated with the membrane of conventional lysosomes. However, a substantial number of the proteins identified in this study were not found in rat liver lysosomes, and these proteins reflect the specialized function of the NK cell secretory lysosome. One important function of NK cells is cytotoxicity; correspondingly we identified perforin, granzyme B, and granzyme H in the NK cell secretory lysosome. Another more recently described function of NK cells is antigen presentation to CD4 ϩ T cells (41); this is also clearly reflected in the composition of the secretory lysosome. Antigen-presenting cells activate CD4 ϩ T cells by presenting antigens on MHC class II molecules, and these antigens are loaded onto the MHC class II molecules within endocytic/lysosomal compartments (42). Indeed not only were 10 different MHC class II molecule subunits identified in the secretory lysosome fraction, but both invariant chain (also known as CD74) and a subunit of human leukocyte antigen-DM (HLA-DM) were identified. Invariant chain is required to target newly synthesized MHC class II molecules to lysosomes, whereas HLA-DM acts as a chaperone to facilitate the removal of the invariant chain and promote the subsequent peptide loading of MHC class II molecules (42). Consistent with the NK cell secretory lysosome serving as a compartment for antigenic peptide loading of MHC class II molecules, HLA-DM is enriched in YTS cell secretory lysosome fractions (data not shown).
The capacity of the NK cell for regulated secretion is reflected by the identification of a number of small GTPases that are known to be associated with this process in other cell types. Rab27a is a key component of the exocytic machinery in CTLs in which it is required to tether secretory lysosomes to the plasma membrane prior to fusion (6,7). Given that Rab27a was enriched in the YTS NK cell secretory lysosome fraction, it seems highly likely that Rab27a will perform the same function in NK cells. Another potential component of the NK cell exocytic machinery is Rab8b, which also was identified in the secretory lysosome fraction by LC ESI-MS/MS. In the neuroendocrine cell line AtT20, Rab8b co-localizes with adrenocorticotropic hormone, and overexpression of this GTPase stimulates adrenocorticotropic hormone release (30). By analogy Rab8b may also participate in secretory lysosome exocytosis in NK cells, perhaps acting in concert with Rab27a. Indeed a precedent for this notion is granule exocytosis by PC12 cells in which both Rab27a and Rab3a are required for docking of the granules to the plasma membrane (43). Additionally the GTPase RalA was identified by LC ESI-MS/MS, but unlike Rab27a it was not obviously enriched in the secretory lysosome fraction. Nonetheless RalA represents a candidate component for the NK cell exocytic machinery because it is also required for granule exocytosis in PC12 cells, although it is localized primarily to the plasma membrane in these cells (31,32). Clearly future studies are required to dissect what role RalA and the other small GTPases associated with the secretory lysosome play in the exocytosis of this organelle in NK cells.
LC ESI-MS/MS identified three SNAREs in the secretory lysosome fraction of the YTS NK cell line, and this finding was corroborated by immunoblotting with antibodies specific for each SNARE. Additionally syntaxin 7 and VAMP7 co-localized to a significant degree with the secretory lysosome membrane protein CD63 in primary NK cells when analyzed by immunofluorescence microscopy. Taken together, these results demonstrate the presence of syntaxin 7 and VAMP7 in the NK cell secretory lysosome. This is consistent with previous observations in which syntaxin 7 and VAMP7 were localized to conventional lysosomes (33)(34)(35). VAMP7 is of particular interest because based upon literature precedents it is likely to play a central role in the exocytosis of the secretory lysosome. Conventional lysosomes can also undergo exocytosis in nonspecialized cells to seal holes in the plasma membrane (44). In this wound repair response, VAMP7 forms a complex with the plasma membrane Q-SNAREs syntaxin 4 and SNAP23 (33); as such VAMP7 may form a complex with the same plasma membrane Q-SNAREs to facilitate secretory lysosome exocytosis.
Results obtained by immunoblotting demonstrate that syntaxin 11 was present but not enriched in secretory lysosomes. In addition, immunofluorescence microscopy revealed that although there was a small degree of co-localization with CD63 in primary NK cells, the majority of syntaxin 11 was located on other intracellular organelles. Indeed previous studies have localized syntaxin 11 to a variety of different intracellular locations but not to lysosomes (13,45,46). However, our data demonstrate for the first time that syntaxin 11 is expressed in both YTS cells and primary NK cells. This is an important observation because mutations in syntaxin 11 are FIG. 6. Localization of SNARE proteins in primary NK cells. Primary NK cells were stained with an antibody specific for CD63 in combination with rabbit polyclonal antisera specific for VAMP7, syntaxin 7, and syntaxin 11. CD63 staining was detected with an anti-mouse Texas Red-conjugated antibody (middle panels), and SNARE staining was detected with anti-rabbit FITC-conjugated antibody (left panels). Antibody fluorescence was detected with a Zeiss Axioplan 2 microscope. The right panels are merges of the red and green channels where a yellow color indicates co-localization. The scale bar is 10 m.
linked to familial hemophagocytic lymphohistiocytosis type 4 (36,37). Familial hemophagocytic lymphohistiocytosis is an autosomal recessive disorder characterized by defective NK cell cytotoxicity, hence implying a role for syntaxin 11 in target cell killing. Given that SNAREs mediate membrane fusion reactions, syntaxin 11 may either be required for the exocytosis of the secretory lysosome or be involved in the trafficking of cytotoxic proteins to this organelle, although further work will be required to determine the precise role of this SNARE.
In summary we describe for the first time the protein composition of the NK cell secretory lysosome membrane. Although many proteins localized to this organelle are also associated with conventional lysosomes, a significant proportion of the proteins identified reflect both the capacity of this organelle for regulated secretion and the immunological role that it plays. Crucially we have used proteomics as a tool to identify the small GTPases and SNARE proteins that represent potential components of the exocytic machinery of the secretory lysosome.