BLOC-3 Mutated in Hermansky-Pudlak Syndrome Is a Rab32/38 Guanine Nucleotide Exchange Factor

1Mutations in components of the Rab prenylation machinery required to link Rabs to membranes result in pigmentation defects due to altered platelet and melanosome formation. Moreover, mutations in rat Rab38 cause pigmentation defects and other changes similar to those found in HPS, and Rab38 and the closely related Rab32 are important for trafficking of enzymes, such as Tyrp1, involved in pigmentation. However, it is not known how these Rabs become activated during melanosome biogenesis. This activation step requires the inactive Rab to release GDP and bind GTP, a process triggered by a specific guanine nucleotide exchange factor (GEF). Bioinformatic analysis of one of the genes mutated in Hermansky-Pudlak syndrome, Hps4, indicated the amino-terminal 120 amino acids were homologous to the equivalent region of the Ccz1 subunit of the yeast Ypt7 GEF (Figure 1A). In addition, similarities were noted between Hps1 and the Mon1 subunit of the Ypt7 GEF (Figure 1A). Intriguingly, given these similarities between Hps1-Hps4 and Mon1-Ccz1, Rab32 and Rab38 fall into the same subcategory of Rabs as Rab7. To test if BLOC-3 has Rab GEF activity, recombinant Hps1-Hps4 complexes and Rab7 GEF complexes were produced (see Figures S1A and S1B available online). Due to the low sequence similarity of human and the characterized yeast Mon1-Ccz1 complex, the human protein complex was first isolated and the subunit composition confirmed by mass spectrometry (Figure S1C). These complexes were then tested for GEF activity toward a panel of Rab7 subfamily Rabs, as well as a selection of other Rabs (Figure 1B). This analysis showed that the BLOC-3 complex but not the subunits alone had GEF activity toward Rab32 and the closely related Rab38, but little or no activity to the other Rabs tested (Figures 1B and S1D). Bacterially expressed BLOC-3 also showed activity to Rab32 and Rab38 but not Rab7 or Rab7-like (Figure S1E), eliminating the possibility that a eukaryotic cell contaminant was responsible for the Rab32/38 GEF activity. By contrast, human Mon1a-Ccz1 had activity toward Rab7 but little or no activity toward Rab32, Rab38, or other Rabs tested (Figure 1B). These findings supported the idea that BLOC-3 and Mon1a-Ccz1 are specific Rab GEF complexes for Rab32/38 and Rab7, respectively.

2If BLOC-3, the Hps1-Hps4 complex, acts as a specific GEF for Rab32/38 in vivo, then the recruitment of Rab32/38 to membranes, the localization of melanosomal proteins, and the formation of melanosomes should require the activity of this complex. All of these properties were therefore tested using MNT-1 cells, which produce melanosomes. Western blot analysis of HeLa and MNT-1 cells revealed that the BLOC-3 subunit Hps4 is enriched in MNT-1 cells, like its target GTPases Rab32 and Rab38, and the tyrosinase-related protein (Tyrp1), a marker for melanin biogenesis (Figure 2A). Rabs and regulators required for early endocytic trafficking and lysosomal sorting were present in both cell lines, although MNT-1 cells have lower levels of the Rab7 GEF subunit Mon1a (Figure 2A). Rab32 was present on small ring-like structures in control cells that overlapped with Tyrp1 (Figure 2B), characteristic of endosomal and melanosomal structures required for sorting of melanosome components. When either the Hps1 or Hps4 subunits of BLOC-3 were depleted using different siRNA duplexes, Rab32 was lost from these ring-like membranes and became redistributed to a diffuse or finely punctate cytoplasmic pattern closer to the nucleus (Figures 2B and S2A). This was in agreement with the measurements showing that both subunits of BLOC-3 are required for GEF activity toward Rab32 or 38. Furthermore, Rab32 localizes to structures adjacent to those defined by the Hps4 subunit of its GEF (Figure S2B).

3To provide further support for the idea that BLOC-3 is important for determining the localization of Rab32, the Hps1-Hps4 complex was relocated to the surface of a heterologous membrane, the mitochondria (Figure 3A). This resulted in the recruitment of Rab32 to the surface of the mitochondria only when both Hps1 and Hps4 were expressed (Figures 3B, 3C, and S3A). Importantly, in this system Hps1 and Hps4 formed a complex at the surface of mitochondria (Figures S3A and S3B), consistent with the idea that the BLOC-3 complex and not the subunits alone have GEF activity. Other Rabs were tested in this assay but did not show mitochondrial targeting (Figures 3B and 3C), supporting the conclusion that this reflected a specific activity of BLOC-3 on Rab32 and Rab38. Together, these data showed that BLOC-3 activity is required for Rab32 and Rab38 activation and is an important determinant of its localization to premelanosomal membrane structures.

4Finally, the role of BLOC-3 in trafficking of melanosomal cargo and production of melanosomes was tested. Previous studies have shown that cells from mice lacking Rab32/38 display deficiencies in the formation of mature melanosomes and accumulate early stage melanosomes. If BLOC-3 is required for Rab32/38 activation, then cells depleted of Hps1, Hps4, and Rab32/38 should show a similar accumulation of premelanosomal structures and reduced numbers of mature heavily pigmented melanosomes. For this purpose the premelanosome marker PMEL was used. PMEL is a transmembrane glycoprotein proteolytically cleaved following exit from the TGN as part of the normal melanosome biogenesis pathway. Precursor forms of PMEL are incorporated into insoluble fibrils and cleaved to 35–45 kDa fragments that become buried by melanin within melanosomes. In control cells, PMEL localizes to a punctate pattern clearly discrete from the mature melanosomes seen in the inverted bright-field image (Figure 4). Depletion of Rab32/38, Hps1, or Hps4 resulted in a strongly increased PMEL signal (Figure 4A). Western blotting showed this was due to increased levels of an 80 kDa nonprocessed form of PMEL in the cells (Figure 4B), suggesting that the proteolytic maturation of PMEL and its burial by melanin was blocked. Furthermore, there was a decrease in densely pigmented melanosomes in nearly 50% of cells (Figures 4A and 4C) and a 20%–30% reduction in pigment production (Figure 4D). This was especially obvious in bright-field images where the dark, round melanosomes are clearly reduced or show altered distribution, sometimes clustering together in large aggregates (Figure S4). BLOC-3 and its targets, Rab32 and Rab38, therefore play an important role in the control of pigment production and melanosome biogenesis in cultured MNT-1 cells.

5These findings provide insights into membrane trafficking pathways of great relevance for both normal cellular function and human diseases. Together the data demonstrate that Hps1-Hps4 (BLOC-3) and Mon1a-Ccz1 form specific Rab GEF complexes for Rab7 and Rab32/38, with discrete functions in the biogenesis of lysosome-related organelles and lysosomes, respectively. The characterization of BLOC-3 and Mon1-Ccz1 therefore defines a new family of Rab GEFs and supports the view that the longin domain is a signature for Rab GEF activity. Intriguingly, some SNAREs possess longin domains, suggesting that these might directly link Rab regulation to vesicle tethering and fusion events in a manner not previously suspected.

6Other unanswered questions relate to the mechanism by which BLOC-3 is recruited and triggers local activation of Rab32 and Rab38. One possibility is that this takes the form of a Rab cascade, where an upstream Rab promotes the recruitment of the exchange factor for downstream Rabs in the pathway. Fitting with this idea, BLOC-3 has been shown to bind to the activated GTP form of Rab9. Intriguingly, the Rab32 and Rab38 GTPase-activating protein RUTBC1 is also a Rab9 effector. These observations suggest that Rab9 may play a previously unappreciated role in the formation of lysosome-related organelles by coordinating the activation and inactivation of Rab32 and Rab38. In turn, Rab32 and Rab38 may then feed back to regulate yet further Rabs, since the VARP exchange factor for Rab21 is a Rab32/38-binding protein. Further work will be required to shed light on the currently murky picture of how such a complex cascade might function.