- Top of page
- Results and Discussion
- Materials and Methods
Hermansky–Pudlak syndrome (HPS) defines a group of at least seven autosomal recessive disorders characterized by albinism and prolonged bleeding due to defects in the lysosome-related organelles, melanosomes and platelet-dense granules, respectively. Most HPS genes, including HPS3, HPS5 and HPS6, encode ubiquitously expressed novel proteins of unknown function. Here, we report the biochemical characterization of a stable protein complex named Biogenesis of Lysosome-related Organelles Complex-2 (BLOC-2), which contains the HPS3, HPS5 and HPS6 proteins as subunits. The endogenous HPS3, HPS5 and HPS6 proteins from human HeLa cells coimmunoprecipitated with each other from crude extracts as well as from fractions resulting from size-exclusion chromatography and density gradient centrifugation. The native molecular mass of BLOC-2 was estimated to be 340 ± 64 kDa. As inferred from the biochemical properties of the HPS6 subunit, BLOC-2 exists in a soluble pool and associates to membranes as a peripheral membrane protein. Fibroblasts deficient in the BLOC-2 subunits HPS3 or HPS6 displayed normal basal secretion of the lysosomal enzyme β-hexosaminidase. Our results suggest a common biological basis underlying the pathogenesis of HPS-3, -5 and -6 disease.
Hermansky–Pudlak syndrome (HPS) comprises at least seven different autosomal recessive disorders (referred to as HPS-1 through -7) that share the clinical manifestations of oculocutaneous albinism combined with platelet storage pool deficiency and do not present with the severe immunodeficiency and giant intracellular granules characteristic of Chediak–Higashi syndrome (1,2). The two manifestations common to all types of HPS are due to deficiencies in the biogenesis of two highly specialized, cell-type-specific organelles: albinism (of variable severity) is associated with abnormal melanosomes (the site of synthesis and storage of melanin pigments in mammals) and storage pool deficiency is due to defects in platelet-dense granules (the main site of storage of serotonin, calcium and ADP in platelets). Both melanosomes and platelet-dense granules are considered lysosome-related organelles because they share the characteristics of low intralumenal pH and the presence of lysosomal-associated membrane proteins (LAMPs) in their limiting membrane. Moreover, recessive mutations affecting both organelles can also cause abnormalities in lysosomes and other related organelles (3,4). Indeed, at least some forms of HPS are associated with additional manifestations that can be ascribed to defects in lysosomes and lysosome-related organelles. For example, autofluorescent undegraded materials (ceroid lipofuscin) have long been shown to accumulate in lysosomes of reticulo-endothelial cells of HPS patients (1). Also, patients suffering from HPS-1 or HPS-4 often develop progressive pulmonary fibrosis, which has been proposed to originate from defects in the lamellar bodies of type II lung epithelial cells (5). Finally, cytotoxic T lymphocytes from HPS-2 patients display impaired killing activity owing to abnormalities in their lytic granules (6), and the cyclic hematopoiesis observed in a dog model of HPS-2 has been proposed to result from defective biogenesis of elastase-containing granules in neutrophils (7). Therefore, the study of the molecular basis of HPS is contributing to our understanding of the mechanisms by which lysosomes and related organelles are formed.
Identification of the genes that are mutated in the different types of HPS, as well as of those that are defective in mouse models of the disease, has revealed a few general themes. First, all HPS genes are expressed ubiquitously, consistent with a function more general than that of regulating the biogenesis of melanosomes and platelet-dense granules. Second, most HPS genes encode novel proteins that have no homologs in unicellular organisms and no recognizable protein motifs; so far the only exceptions are the genes encoding subunits of the AP-3 complex involved in protein sorting and the mouse Vps33a gene encoding an ortholog of the yeast Vps33 protein involved in fusion of vesicles and organelles with the vacuole/lysosome (8,9). Third, most HPS gene products have been shown to assemble into distinct multisubunit protein complexes. For example, the gene defective in HPS-2 disease encodes the β3A subunit of the heterotetrameric AP-3 complex, the one associated with HPS-7 disease encodes a subunit of Biogenesis of Lysosome-related Organelles Complex-1 (BLOC-1), and the products of HPS1 and HPS4 are subunits of BLOC-3 (10–14).
In this paper we report the biochemical characterization of BLOC-2, a complex previously proposed to contain the HPS5 and HPS6 proteins on the basis of:
We have identified the endogenous HPS5 and HPS6 proteins from cultured HeLa cells and corroborated that they are components of a stable complex (BLOC-2). Interestingly, we have found that the complex also contains the product of the HPS3 gene.
Results and Discussion
- Top of page
- Results and Discussion
- Materials and Methods
To biochemically characterize the products of the HPS5 and HPS6 genes, we raised rabbit polyclonal antibodies using as immunogens selected regions of each gene product that were expressed and purified as glutathione S-transferase (GST)-fusion proteins. We obtained two different antibodies against HPS5, named HP5c and HP5d, and two against HPS6, named HP6b and HP6d. By immunoblotting analysis, the four antibodies recognized recombinant forms of their corresponding antigens, although only HP6d was able to detect the endogenous protein from a crude cell extract. As shown in Figure 1A, immunoblotting analysis of HeLa cytosol using the HP6d antibody revealed a major protein band of apparent molecular mass ∼ 90 kDa, a value that was similar to the predicted molecular mass of HPS6 (83 kDa). To verify that the ∼ 90-kDa band represents HPS6, and to identify endogenous HPS5, two different immunoprecipitation-based approaches were undertaken. One approach (‘immunoprecipitation-recapture’) consisted of two tandem immunoprecipitations, the first one under nondenaturing conditions and the second after denaturation of the first immunoprecipitate, using extracts from HeLa cells metabolically labeled with [35S]methionine and [35S]cysteine. Following this approach, a single radiolabeled protein of apparent molecular mass ∼ 90 kDa was isolated by using the HP6d antibody in both immunoprecipitation steps (Figure 1B, left panel). As expected, the 90-kDa protein was not detected in control samples in which the specific antibody was replaced by irrelevant rabbit IgG at either immunoprecipitation step (Figure 1B, left panel). In the case of HPS5, immunoprecipitation-recapture using the HP5c antibody at both steps resulted in the isolation of a major radiolabeled protein of apparent molecular mass ∼ 170 kDa, which was not detected in control immunoprecipitations using irrelevant IgG at either step (Figure 1B, right panel). The second approach (‘immunoprecipitation-immunoblotting’) consisted of immunoprecipitation of an unlabeled HeLa cell extract followed by immunoblotting analysis. Here, the same ∼ 90-kDa protein band was detected in samples that had been immunoprecipitated using two antibodies against nonoverlapping regions of HPS6 (HP6b and HP6d) but not in control immunoprecipitations using irrelevant IgG (Figure 1C). Likewise, the same ∼ 170-kDa protein band was detected in samples immunoprecipitated using two antibodies against nonoverlapping regions of HPS5 (HP5c and HP5d), whereas the band was not present in samples immunoprecipitated using preimmune serum or purified irrelevant IgG (Figure 1D). Together, these experiments suggest that the ∼ 90-kDa and ∼ 170-kDa proteins represent endogenous HPS6 and HPS5, respectively. Neither the HP5c nor HP6d antibodies were able to recognize the mouse counterparts of their antigens by immunoblotting or immunoprecipitation-immunoblotting analyses of cultured skin fibroblasts (data not shown). The electrophoretic mobility of HPS5 on SDS-PAGE was somewhat slower than that expected from its calculated molecular mass (∼ 127 kDa for the longest predicted isoform). Although the molecular basis for this abnormal electrophoretic mobility remains unclear, a similar phenomenon was observed for other HPS proteins, including HPS3 and HPS4 (e.g. ).
Figure 1. Immunochemical detection of endogenous HPS5 and HPS6 proteins from HeLa cells.A) Immunoblotting analysis of a recombinant histidine-tagged fusion protein containing residues 551–775 of human HPS6 (∼ 0.1 ng, apparent molecular mass ∼ 34 kDa), and a cytosolic extract prepared from HeLa cells (∼ 10 μg total protein), using affinity-purified antibody HP6d against HPS6. The positions of molecular mass standards are indicated on the left. B) HeLa cells metabolically labeled with [35S]methionine and [35S]cysteine were lyzed under nondenaturing conditions, and the cleared lysate was subjected to a first immunoprecipitation step (1st IP) using the purified antibodies HP6d and HP5c (against HPS6 and HPS5, respectively) or irrelevant rabbit IgG. Following extensive washing, each immunoprecipitate was denatured by heating in the presence of SDS and dithiothreitol, diluted, and subjected to a recapture immunoprecipitation (2nd IP) using the indicated antibodies. The final immunoprecipitates were analyzed by 4–20% SDS-PAGE followed by fluorography. The positions of molecular mass standards are indicated on the left. C) A whole-cell extract from unlabeled HeLa cells was divided into aliquots (∼ 2.4 mg total protein each) and immunoprecipitated using irrelevant rabbit IgG or the purified HP6b and HP6d antibodies against HPS6. Bound proteins were analyzed by immunoblotting using the HP6d antibody. The arrow indicates the position of the HPS6 protein (∼ 90 kDa). D) A whole-cell extract from unlabeled HeLa cells was divided into aliquots (∼ 2.4 mg total protein each) and immunoprecipitated using control preimmune rabbit serum, purified rabbit IgG, HP5d antiserum or purified HP5c antibody (the last two against HPS5). Bound proteins were analyzed by immunoblotting using the HP5c antibody. The arrow indicates the position of the HPS5 protein (∼ 170 kDa).
Next, we used the HP6d antibody to estimate the relative abundance of the HPS6 protein in HeLa cells by means of quantitative immunoblotting. To this end, the signal obtained from known total protein amounts of Triton X-100 extracts from HeLa cells was referred to that of purified His-HPS6 (551–775) protein analyzed in the same membrane. The obtained value was ∼ 90 μg of HPS6 per gram of total protein.
All HPS gene products characterized so far were found to exist in soluble and membrane-associated forms. To test whether this was also the case for HPS6, HeLa cells were lyzed in the absence of detergents and the soluble and microsomal membrane fractions were isolated by differential centrifugation. The membranes were washed once with buffer to reduce contamination by cytosolic proteins. As shown in Figure 2A, HPS6 was consistently recovered from both fractions. Furthermore, the membrane-bound form of HPS6 was partially extracted using a detergent-free, low-salt buffer, and quantitatively extracted using the same buffer supplemented with 1 m NaCl or 0.2 m Na2CO3 (Figure 2B). These results suggest that HPS6 (and, by extension, BLOC-2; see below) exists as both a soluble protein and a peripheral membrane protein.
Figure 2. Biochemical characterization of the human HPS6 protein.A) HeLa cells were lyzed in detergent-free buffers A or B by 15–20 passages through a 25-G needle (buffer compositions are described under Materials and Methods). The crude lysates were centrifuged at 15 000 × g for 10 min, and the resulting supernatants were further centrifuged at 120 000 × g for 90 min to prepare cytosolic and microsomal membrane fractions, which were analyzed in comparable amounts for the presence of HPS6 protein, using the HP6d antibody. B) Washed microsomal membranes prepared as in (A) were resuspended in buffer containing no additive (–), 1 m NaCl or 0.2 m Na2CO3 (pH 11). The resuspended membranes were incubated for 1 h at room temperature and then centrifuged at 120 000 × g for 90 min. Comparable amounts of the resulting supernatants (S) and membrane pellets (P) were analyzed by immunoblotting using the HP6d antibody to HPS6. C) HeLa cytosol was fractionated by ultracentrifugation on a linear 5–20% (w/v) sucrose gradient, as described under Materials and Methods. The resulting fractions were analyzed by immunoblotting using the HP6d antibody. Fractions 1 and 23 correspond to the top and bottom ends of the gradient, respectively. The positions of standard proteins of known sedimentation coefficient (in Svedberg units) are indicated on the top.
Upon fractionation of HeLa cytosol by size exclusion chromatography (data not shown) and density gradient centrifugation (Figure 2C), HPS6 was recovered in fractions that were consistent with it being part of a large protein complex, in agreement with the idea that HPS6 is a component of BLOC-2 (15). To test for stable association between the endogenous HPS5 and HPS6 proteins, we performed immunoprecipitation-immunoblotting experiments using our antibodies first to immunoprecipitate HPS5 or HPS6 from HeLa cell extracts (under nondenaturing conditions) and then to assay by immunoblotting for the presence of each protein in the immunoprecipitates. As shown in Figure 3 (top and middle panels), both HPS5 and HPS6 proteins were detected in immunoprecipitates obtained using antibodies to either protein, and neither was found in control immunoprecipitations using irrelevant IgG. Association between endogenous HPS5 and HPS6 proteins was also demonstrated by immunoprecipitation-recapture (data not shown), and by immunoprecipitation-immunoblotting analyses that were performed subsequent to fractionation of HeLa cytosol by size-exclusion chromatography (Figure 4A) or density gradient centrifugation (Figure 4B). Taken together, these results demonstrate that, as previously suggested (15), HPS5 and HPS6 are components of a stable protein complex (BLOC-2).
Figure 3. Coimmunoprecipitation of endogenous HPS3, HPS5 and HPS6 proteins. HeLa cells were lyzed in the presence of Triton X-100, and the clear lysate was divided into aliquots that were subjected to immunoprecipitation (IP) using irrelevant rabbit IgG or the purified antibodies HP5c (to HPS5) and HP6d (to HPS6). Following extensive washing, immunoprecipitated proteins were analyzed by immunoblotting (IB) using specific antibodies to HPS3, HPS5 and HPS6, as indicated. The HPS4 protein, which had been found to be a subunit of another complex and not to coimmunoprecipitate with HPS3 (13), was not detected in any of these immunoprecipitates (not shown).
Figure 4. Cofractionation of HPS6 with the associated HPS3 and HPS5 proteins upon size-exclusion chromatography(A) and density gradient centrifugation (B).A) Cytosol from HeLa cells was fractionated on a Superose 6 gel filtration column as described under Materials and Methods, and the resulting fractions 21, 23–40, 42, 44, 46 and 48 were immunoprecipitated using the HP6d antibody against HPS6. Following washing, bound proteins were analyzed by immunoblotting using antibodies to HPS3 (HP3c), HPS5 (HP5c) and HPS6 (HP6d). The exclusion volume (Vo) and the elution position of standard proteins (Stokes radii given in Ångstroms) are indicated on the top. B) Cytosol from HeLa cells was fractionated by centrifugation on a 5–20% (w/v) sucrose gradient as described under Materials and Methods. The resulting fractions were analyzed by immunoprecipitation using the HP6d antibody to HPS6, followed by immunoblotting using the same antibody or antibodies to HPS3 (HP3c) or HPS5 (HP5c). Fractions 1 and 23 correspond to the top and bottom ends of the gradient, respectively. The positions of standard proteins of known sedimentation coefficient (in Svedberg units) are indicated on the top.
From the sedimentation velocity analyses shown in Figures 2(C) and 4(B), the sedimentation coefficient of BLOC-2 was estimated to be 8.2 ± 0.5 S. Interestingly, this value was in close agreement with that previously calculated for the HPS3 protein (8.8 ± 1.0 S ), suggesting that HPS3 might also be a component of BLOC-2. This idea was tested directly by immunoprecipitation-immunoblotting analyses using our antibody HP3c against human HPS3 (13). As shown in Figure 3, lower panel, significant amounts of HPS3 were recovered from immunoprecipitates that had been obtained using antibodies to HPS5 or HPS6, but not in those using control IgG. Conversely, both HPS5 and HPS6 were detected in samples immunoprecipitated using HP3c (data not shown). Stable association of HPS3 with HPS5 (data not shown) and HPS6 was also demonstrated by immunoprecipitation-immunoblotting following fractionation of HeLa cytosol by size-exclusion chromatography (Figure 4A) and density gradient centrifugation (Figure 4B). Taken together, the above results indicate that BLOC-2 contains not only HPS5 and HPS6 but also HPS3.
The native molecular mass of BLOC-2 was calculated (16) from its sedimentation coefficient (see above) and Stokes radius, the latter estimated to be 96 ± 7 Å from the size-exclusion chromatography experiments. The calculated mass is 340 ± 64 kDa, which fits, within the experimental error, the theoretical mass of a complex consisting of one molecule each of HPS3, HPS5 and HPS6 (∼ 324 kDa). However, the possible existence of one or more additional subunits cannot be ruled out at this point. Moreover, the molecular composition of BLOC-2 may not be homogeneous, as seems to be suggested by a subtle shift in the position of the HPS5 intensity peak, relative to that of HPS6 and HPS3, in the gel filtration fractions analyzed by immunoprecipitation-immunoblotting (Figure 4A), and by the observation that immunodepletion of HPS6 from a HeLa cell extract resulted in incomplete depletion of the HPS3 and HPS5 proteins (data not shown). Nevertheless, the significant extent of cofractionation observed for the HPS3, HPS5 and HPS6 proteins using two distinct biochemical methods (i.e. size exclusion chromatography and density gradient centrifugation; Figures 2C and 4 and data not shown) and genetic evidence recently obtained by another group (see below) suggest that the bulk of these proteins may exist in vivo as subunits of BLOC-2. Future experiments, such as biochemical purification of BLOC-2 followed by mass spectrometry analysis, will be required to clarify this point.
While this manuscript was under revision, a paper reporting the association of the mouse Hps3 protein with Hps5 and Hps6 was published online (17). The paper included the following genetic evidence:
the melanosome and platelet-dense granule phenotypes characterized for cocoa (Hps3coa) single mutant and cocoa/ruby eye double mutant mice were virtually identical to those of ruby eye (Hps6ru) and ruby eye-2 (Hps5ru2) mutant mice;
the steady-state protein levels of Hps5 and Hps6 were drastically reduced in tissues from ruby eye, ruby eye-2 and cocoa mutants.
The results of the two studies are mostly complementary and strongly support the idea that mammalian BLOC-2 contains the HPS3/Hps3, HPS5/Hps5 and HPS6/Hps6 proteins.
Although the molecular function of BLOC-2 remains largely unknown, published data on the phenotypic characterization of mouse cells deficient in either Hps6 (ruby eye) or HPS5 (ruby eye-2) hints at a regulatory role in the secretion of lysosomes and other organelles. For instance, a ∼ 3-fold increase in the frequency and duration of transient fusion events (i.e. kiss and run) was observed in mast cells from ruby eye mutant mice during the initial phase of stimulation of degranulation (18). In addition, reduced secretion of lysosomal enzymes by kidney and platelets from both ruby eye and ruby eye-2 mice has been documented (reviewed in ), although normal secretion of lysosomal enzymes by kidney and platelets has been reported for Hps3 null mutant (cocoa) mice (20). Here, we decided to examine whether the constitutive secretion of lysosomal enzymes by fibroblasts in culture would be affected by a deficiency in BLOC-2. To this end, immortalized skin fibroblast lines derived from wild-type, cocoa and ruby eye mice were grown on monolayers, and the activity of β-hexosaminidase released into the culture medium during a 15-h period was measured and normalized to the intracellular protein content of each culture. Each independent cell line was analyzed in two separate cultures, and each enzymatic determination was performed in duplicate. The secreted enzymatic activities (expressed as means ± SD of three independent cell lines) were 10.1 ± 3.7, 12.1 ± 1.2 and 9.6 ± 1.1 units per gram of cellular protein of wild-type, cocoa and ruby eye fibroblasts, respectively. These results suggest that BLOC-2 is not critical for basal levels of secretion of lysosomal enzymes by fibroblasts.
In summary, our results demonstrate that BLOC-2 is a stable protein complex containing the human HPS3, HPS5 and HPS6 proteins. This finding has implications for both basic and clinical aspects of HPS. In particular, it predicts that HPS-5 and HPS-6 diseases should be clinically similar, if not identical, to HPS-3 disease, which has been characterized more extensively and found to represent a mild form of HPS (21).