Mouse models of HPS
A major contribution of Dr. Richard T. Swank's laboratory was the collection, since the 1970s, of mouse mutants that mimic the human HPS phenotypes. Additional phenotypes such as susceptibility to anesthetics, protection from atherosclerosis, and otolith deficiency have been documented only in mouse HPS mutants as summarized in (Li et al., 2004; Swank et al., 1998). The characterization of more than a dozen mouse HPS mutants (Swank et al., 1998) led to a series of successful identifications of murine and human HPS genes (Li et al., 2004; Table 2). Currently, the 15 cloned mouse HPS genes are known to reside in several HPS Protein Associated Complexes (HPAC, shown in Table 2 and Figure 2) to mediate the biogenesis of LROs.
Figure 2. Illustration of the HPS protein associated complexes (HPAC). Seven HPACs are depicted based on the current understanding of the structural assembly and physical interactions described in the text. BLOC-1 is assembled linearly connected by two subcomplexes (BLOS1-pallidin-cappuccino and BLOS2-dysbindin-snapin), with muted and BLOS3 overhung. KXD1 is a BLOC-1 interactor bound to BLOS1. In BLOC-2, HPS3, HPS5, and HPS6 interact with each other. HPS1 and HPS4 are tightly bound in BLOC-3. In the HOPS complex, VPS18 connects the head (VPS16-VPS33A-VPS41) and tail (VPS11 and VPS39). The ubiquitous AP-3 complex contains the β3A, δ, μ3A and σ3 subunits. The core Rab GGTase II enzyme consists of α and β subunits. In the RabGGTase II holoenzyme, REP-1 (or component A) binds unprenylated Rab proteins and then presents them to the catalytic Rab GGTase II for the geranylgeranyl transfer reaction. The protein names with a star sign (*) indicate mutations in human and mouse HPS, those with a pound sign (#) indicate mutations in mouse HPS only. A question mark on the RAB38 indicates an unresolved HPS protein. HPS, Hermansky–Pudlak syndrome.
Download figure to PowerPoint
With the discovery of murine and human HPS genes, three biogenesis of lysosome-related organelles complexes (BLOC-1, BLOC-2, and BLOC-3) have been defined. That is, pallidin, muted, dysbindin, cappuccino, snapin, BLOS1, BLOS2, and BLOS3 in BLOC-1 (Ciciotte et al., 2003; Falcon-Perez et al., 2002; Gwynn et al., 2004; Li et al., 2003; Starcevic and Dell'Angelica, 2004); HPS3, HPS5, and HPS6 in BLOC-2 (Di Pietro et al., 2004; Gautam et al., 2004; Zhang et al., 2003); and HPS1 and HPS4 in BLOC-3 (Chiang et al., 2003; Martina et al., 2003; Nazarian et al., 2003; Suzuki et al., 2002). The biochemical features and assembling machineries of these BLOC complexes remain to be defined, although some pioneering studies have revealed binding domains in building these BLOCs (Dell'Angelica, 2004; Li et al., 2007) and the linear assembly of the BLOC-1 complex in vitro (Lee et al., 2012). Together with the well-known AP-3 (adaptor protein complex-3) and HOPS (homotypic fusion and protein sorting complex) complexes, the BLOC complexes function in endo-lysosomal trafficking. Emerging evidence has shown that these complexes direct cargoes from either de novo synthesis or endocytosis into lysosomes and LROs. In addition, whether there is a master regulator to coordinate the action of these complexes remains unknown. A key to understanding LRO biogenesis is to define the behaviors of these interacting complexes. An intriguing question is whether these complexes act synergistically or sequentially on cargo transport. Double or multiple mouse mutants played a pivotal role in dissecting these interactions as evidenced by epistatic or synergistic effects on coat color or LRO phenotypes (Gautam et al., 2006; Hoyle et al., 2011). At the cellular level, the mouse HPS mutants provide powerful tools for the dissection of cargo-specific endo-lysosomal-trafficking pathways in different tissues.
As revealed by spontaneous murine HPS mutants, one would expect the development of HPS when generating knockout mutants for the remaining subunits of the HPACs. While the mutants of known subunits of BLOC-2 and BLOC-3 are murine HPS models (Li et al., 2004), it is unknown whether mutation of the genes encoding snapin, BLOS1, and BLOS2 in the BLOC-1 complex would cause typical HPS. In contrast, the snapin knockout (KO) mice died at the perinatal stage, which is unlike other BLOC-1 mutants and showed developmental defects of the central nervous system (Tian et al., 2005; Zhou et al., 2011). This suggests that snapin may play extra roles independent of BLOC-1. Whether the BLOS1 or BLOS2 knockout mice exhibit typical HPS phenotypes with comparable survival remains to be investigated. Interestingly, multiple isoforms of dysbindin have been shown both in human and mouse tissues (Talbot et al., 2009). Studies have shown the different distribution pattern of these isoforms in developmental stages and in sub-brain regions and sub-synaptic regions (Ito et al., 2010; Talbot et al., 2011; Tang et al., 2009a). The involvement of dysbindin in the dystrobrevin complex (Benson et al., 2001) and WAVE2-Abi-1 complex (Ito et al., 2010) has raised the question whether all these isoforms function in a BLOC-1-dependent manner.
The AP-3 complex is a heterotetramer composed of two large adaptins (AP3D1/δ and AP3B1/β3A or AP3B2/β3B), a medium adaptin (AP3M1/μ3A or AP3M2/μ3B), and a small adaptin (AP3S1/σ3A or AP3S2/σ3B). There exist two types of mammalian AP-3 complexes: a ubiquitous AP-3 comprising AP3D1-AP3B1-AP3M1-AP3S1 (or AP3S2) subunits and a brain-specific AP-3 complex containing AP3D1-AP3B2-AP3M2-AP3S1 (or AP3S2) subunits as summarized in a recent review (Dell'Angelica, 2009). In mice, only mutation of the Ap3b1 (in pearl mice) or Ap3d1 (in mocha mice) have been shown to present HPS phenotypes (see Table 2). Knockout of the neuron-specific Ap3b2 or Ap3m2 results in neurological impairments but not HPS (Nakatsu et al., 2004; Newell-Litwa et al., 2009; Seong et al., 2005). This leads to the notion that tissue-specific complexes are formed to execute special physiological functions.
The class C Vps complexes (HOPS and CORVET) are involved in homotypic fusion and tethering during endosomal trafficking (Nickerson et al., 2009). Both HOPS and CORVET complexes share the four Vps proteins (class C core), VPS11-VPS16-VPS18-VPS33. The HOPS complex contains in addition two Rab-binding proteins, VPS41 and VPS39, whereas the CORVET complex has the VPS41 homolog VPS8 and the VPS39 homolog VPS3. The CORVET and HOPS complexes interconvert through two intermediate complexes consisting of the class C core bound to VPS39-VPS8 or VPS3-VPS41 (Peplowska et al., 2007). In the HOPS complex, the large head contains VPS41, VPS33, and VPS16, whereas Vps39 is found in the tip of its tail, with VPS11 and VPS18 connecting the head and tail (Brocker et al., 2012). VPS33 is a Sec1-/Munc18-like protein that interacts with SNAREs. In metazoans, two homologs, VPS33A and VPS33B, are present. It is uncertain whether VPS33B is a component of the COVERT complex (Zlatic et al., 2011), while VPS33A is a part of the HOPS complex (Sriram et al., 2003). Mutation of Vps33a in the bf mice leads to HPS with additional neurological lesions, although the underlying mechanism leading to Purkinje cell loss and neurological atrophy is unknown (Chintala et al., 2009; Suzuki et al., 2003b). A mutation in human VPS33B causes arthrogryposis-renal dysfunction-cholestasis syndrome but not HPS (Gissen et al., 2004). Again, this suggests that distinct mammalian Vps-C complexes function differently.
The HPS mouse model gunmetal (gm) is deficient in the α-subunit of Rab GGTase II (RABGGTA; Detter et al., 2000). Together with the β-subunit and an escort protein, REP-1, RABGGTA adds two geranyl 10-carbon isoprenoid groups to the C-termini of Rab proteins (Anant et al., 1998). In gm mutant, the prenylation of Rab proteins (such as Rab27a, 11a, and 4) is deficient in platelets and melanocytes (Zhang et al., 2002b), which may explain the defects in the biogenesis of platelet granules and melanosomes, thus mimicking the symptoms of HPS. Currently, no RABGGTA mutation has been reported in patients with HPS or storage pool deficiency (Li et al., 2000). However, mutations in the CHM gene that encodes REP-1 are a cause for choroideremia (also known as tapetochoroidal dystrophy). This X-linked disease is characterized by progressive dystrophy of the choroid, retinal pigment epithelium, and retina (Sankila et al., 1992). Defective Rab prenylation and depigmentation occur in Chm knockout mice (Tolmachova et al., 2006).
RAB38 forms a complex with VARP in mediating the melanosomal transport of TYRP1 (Tamura et al., 2009; Wang et al., 2008). Mutation of the melanosomal protein RAB38 in chocolate (cht) mice causes dilution in coat color (Loftus et al., 2002) and ocular hypopigmentation (Brooks et al., 2007), suggesting that cht is a mouse model of OCA. In addition, defects in melanosomes (Brooks et al., 2007) and lamellar bodies (LBs; Osanai et al., 2008) were observed in cht mice, suggesting that RAB38 is a candidate HPS gene affecting multiple LROs (Brooks et al., 2007; Osanai and Voelker, 2008). Consistent with this possibility, the rat RAB38 is null in Fawn-Hooded and Tester-Moriyama rats which mimic the human HPS phenotype (Oiso et al., 2004). Other evidence to show the involvement of RAB38 in the development of HPS is the interactions between RAB38 and BLOC-1, BLOC-2, and AP-3, which likely transport the cargoes of BLOC-2, AP-3, and AP-1 into melanosomes (Bultema et al., 2012). In addition, BLOC-3 has been shown to function as a guanine nucleotide exchange factor for RAB38 and RAB32 (Gerondopoulos et al., 2012). Taken together, these data suggest that RAB38 and RAB32 function with other HPS protein complexes, implicating the loss-of-function of RAB38 or RAB32 may develop similar phenotypes as HPS. However, no prolonged bleeding times or blood defects occur in cht mice (Brooks et al., 2007; Loftus et al., 2002). One possible explanation is a hypermorphic effect in cht mice. The existing mutant hydrophilic RAB38 (Osanai et al., 2008) may function in platelets without affecting the biogenesis of DG. Another explanation is redundant function of a RAB38 homolog, RAB32, in platelets (Wasmeier et al., 2006). In a recent study, RAB38 and RAB32 are involved in the cargo vesicle fusion with mature organelles during the biogenesis of platelet-DG (Ambrosio et al., 2012). Currently, no mutation of the RAB38 gene has been reported in patients with HPS or OCA (Brooks et al., 2007; Suzuki et al., 2003a).
Misty (m) mice show generalized hypopigmentation and prolonged bleeding times similar to HPS mouse mutants (Sviderskaya et al., 1998; Swank et al., 1998), but no evidence of a generalized LRO anomaly (Blasius et al., 2009). Thus, the mutation of Dock7 (a Rho family guanine nucleotide exchange factor) in misty mice (Blasius et al., 2009) is unlikely to cause murine HPS. The subtle gray (sut) mouse was regarded as a model for a mild form of HPS (Swank et al., 1996, 1998). However, the hypopigmentation is mainly caused by the reduction in pheomelanin production due to the mutation of the Slc7a11 gene which encodes the functional subunit of the cystine/glutamate antiporter, xCT (Chintala et al., 2005). Although the function of xCT in the biogenesis of platelet-dense granule has not been intensively studied, no generalized LRO anomaly in sut mice has been described, excluding it as a typical mouse HPS model. The ashen (ash) line maintained at Roswell Park Cancer Institute (referred as ash-Roswell) that was regarded as a mouse model of HPS (Li et al., 2004; Wilson et al., 2000) has been identified to be a double mutant with mutations in both the Rab27a gene and the Slc35d3 gene (Chintala et al., 2007; Wilson et al., 2000). The deficiency in Rab27a causes a defect in melanosomal transport which leads to hypopigmentation (Wilson et al., 2000), while the deficiency of Slc35d3 leads to defects in the biogenesis of platelet-DG (Chintala et al., 2007; Meng et al., 2012). The Rab27a mutation does not cause defects in platelet-DG or prolonged bleeding (Barral et al., 2002). In contrast, hypopigmentation is not seen in the Roswell (ros)-mutant mice, which carry only the Slc35d3 mutation. SLC35D3 may not be expressed in melanocytes (our unpublished data). Thus, the identified HPS mouse models have been updated to 15 lines as listed in Table 2.