Rab geranylgeranyl transferase α mutation in the gunmetal mouse reduces Rab prenylation and platelet synthesis

Abstract

Few molecular events important to platelet biogenesis have been identified. Mice homozygous for the spontaneous, recessive mutation gunmetal (gm) have prolonged bleeding, thrombocytopenia, and reduced platelet α- and δ-granule contents. Here we show by positional cloning that gm results from a G→A substitution mutation in a splice acceptor site within the α-subunit of Rab geranylgeranyl transferase (Rabggta), an enzyme that attaches geranylgeranyl groups to Rab proteins. Most Rabggta mRNAs from gm tissues skipped exon 1 and lacked a start codon. Rabggta protein and Rab geranylgeranyl transferase (GGTase) activity were reduced 4-fold in gm platelets. Geranylgeranylation and membrane association of Rab27, a Rab GGTase substrate, were significantly decreased in gm platelets. These findings indicate that geranylgeranylation of Rab GTPases is critical for hemostasis. Rab GGTase inhibition may represent a new treatment for thrombocytosis and clotting disorders.

Platelets play a pivotal role in acute myocardial infarction, unstable angina, deep venous thrombosis, and stroke. About 4 million individuals are hospitalized each year in the United States with these disorders. Treatment with agents that prevent platelet activation reduces the risk of myocardial infarction and stroke by about 30% and death by about 15% (1). Novel platelet antagonists are needed, however, because platelet activation is not always inhibited by current drugs.

Mice homozygous for gunmetal (gm), a spontaneous, recessive mutation, have prolonged bleeding caused by defects in platelets and megakaryocytes (2, 3). These mice also have macrothrombocytopenia and reduced platelet α- and δ-granule contents (storage pool deficiency, SPD). Megakaryocytes, the progenitors of platelets, are more plentiful in gm mice, but have abnormal intracellular membranes, increased emperipolesis, and decreased platelet synthesis. In addition, gm homozygotes have partial cutaneous albinism (Fig. 1). These features closely resemble the rare human disorders gray platelet syndrome (GPS) and platelet α,δ-SPD. Positional cloning of gm was undertaken both to shed light on the mechanism of disease in platelet SPD disorders and to identify a possible target for antiplatelet drug development.

Figure 1.

C57BL/6J-gm/gm mice (Left) have a lighter coat color than coisogenic C57BL/6J-+/+ mice (Right).

Materials and Methods

Mice and Genetic Mapping.

Mice obtained from The Jackson Laboratory were bred at Roswell Park Cancer Institute. Backcross mice were phenotyped by coat color and genotyped for simple sequence length polymorphisms (SSLPs) and restriction fragment length polymorphism (RFLP) using standard techniques. Linkage relationships were determined by segregation analysis, and best locus order was decided by minimizing crossovers and eliminating double crossovers (4).

Physical Mapping.

Mouse yeast artificial chromosome (YAC) and bacterial artificial chromosome (BAC) clones were identified from libraries (Research Genetics, Huntsville, AL) by PCR or hybridization with oligonucleotides corresponding to chromosome 14 loci. Clones were oriented by sequence-tagged site content mapping and insert end-sequencing. Additional markers were isolated by inverse repetitive element PCR (5).

Inverse Repetitive Element-Direct cDNA Selection.

Expressed sequences were isolated from YACs by inverse repetitive element-direct cDNA selection (6). Briefly, outward-oriented, biotin-labeled, B1 repetitive element-specific primers were used for long-range PCR of YAC clones (5). Denatured PCR products were incubated with ribosomal DNA, Cot-1 DNA, and yeast DNA, and then with denatured mouse bone marrow cDNA (attached to amplification cassettes). PCR products (and annealed cDNAs) were captured on streptavidin-coated beads and washed stringently, and cDNAs were eluted. Eluted cDNAs were reamplified and used in a second selection cycle. Resultant cDNA fragments were subcloned and sequenced.

Full-Length Rabggta (α-Subunit of Rab Geranylgeranyl Transferase) cDNA Isolation.

5′ and 3′ Rapid amplification of cDNA ends (RACE) PCR was performed with mouse fetal brain cDNA (CLONTECH). 5′ RACE used a Rabggta-specific reverse primer (5′-TTCACACGCAGACAGCTCTC-3′) and universal forward primer, followed by a second, nested PCR with another Rabggta primer (5′-CTCTAGCCTTTTGGCCTCTG-3′) and second universal primer. 3′ RACE was performed by using a Rabggta-specific forward primer (5′-CCAGCAGTCTGCTGCACTTC-3′) and universal reverse primer, and a second, nested PCR with the Rabggta primer (5′-GGAGCGTTTGGCTGAGATGC-3′) and another universal reverse primer. Products were subcloned and sequenced, and full-length Rabggta cDNA sequence was assembled.

Reverse Transcription–PCR (RT-PCR) Analysis.

Total RNA was prepared from mouse tissues with equilibrated phenol/guanidine isothiocyanate (GIBCO/BRL). Oligo(dT)-primed, first-strand cDNA was synthesized with SuperScript II (GIBCO/BRL). All Rabggta exons were amplified by PCR with gene-specific primers. Aberrant gm splicing was identified with Rabggta primers 5′-GTGCAAGGGTCCACGGGAC-3′ and 5′-TTCACACGCAGACAGCTCTC-3′.

Western Blot Analysis.

Platelet and bone marrow protein were isolated as described (7) except that platelets were washed twice with 0.38% sodium citrate in 0.85% NaCl and once with 1% ammonium oxalate to remove residual red blood cells. Twenty micrograms of protein in proteinase inhibitor mixture was incubated in SDS/mercaptoethanol and electrophoresed on denaturing 10% polyacrylamide gels, followed by Western blotting on nitrocellulose or poly(vinylidene difluoride) membranes (2). For subcellular fractionation studies, fresh platelets were sonicated and separated into membrane and soluble subcellular fractions by centrifugation as described (2). Samples were immediately separated on 12% polyacrylamide gels. Western blots were probed with specific antisera against Rabggta (8) or Rab27 (9), followed by peroxidase-labeled goat α-rabbit IgG secondary antibody (Kirkegaard & Perry Laboratories), and visualized by incubation with ECL-Plus reagent (Amersham Pharmacia) and autoradiography. Several exposures were taken and comparisons were made in the linear range. Blots were washed and reprobed with antibodies to mouse actin. Alternatively GTP-binding proteins were detected on Western blots by hybridization with [³²P]GTP, washing, and autoradiography (2).

Assay of Rab Geranylgeranyl Transferase (GGTase) Activity.

Frozen tissues (liver, kidney, or spleen) were thawed, immersed in 3 vol of buffer (50 mM sodium Hepes, pH 7.2/10 mM NaCl/1 mM DTT/0.5 mM PMSF/5 μg/ml aprotinin/5 μg/ml pepstatin/5 μg/ml leupeptin) and lysed by polytron homogenization. Lysates were spun at 100,000 × g for 1 h, and pellets were discarded. Frozen platelet samples (10⁷-10⁸ platelets) were lysed by repeated passages through a Hamilton syringe. Protein concentrations were determined by using Coomassie Plus reagent (Pierce).

Rab GGTase activity (9) was determined by measuring ³H transfer from [³H]geranylgeranyl pyrophosphate ([³H]GGPP) to Rab1a in 50-μl reactions containing 50 mM sodium Hepes (pH 7.2), 5 mM MgCl₂, 1 mM DTT, 4 μM [³H]GGPP, 2,365 dpm/pmol), 5 μM recombinant Rab1a, 2 μM recombinant Rab escort protein (REP) 1, and indicated amounts of cytosolic protein from gm/gm or +/+ tissues. After incubation for 15 min at 37°C, ethanol/HCl-precipitable radioactivity was measured on glass fiber filters. Assays with concentrations of cytosolic proteins <10 μg were carried out in 25-μl reactions containing 1 μM [³H]GGPP and incubated for 30 min at 37°C.

Assay of GGTase I Activity.

GGTase I activity was determined by measuring ³H transfer from [³H]GGPP to Rac1 in a 25-μl reaction containing 50 mM sodium Hepes (pH 7.2), 150 mM KCl, 5 mM MgCl₂, 1 mM DTT, 1.2 μM [³H]GGPP, 200 μM Zwittergent 3–14, 25 μM recombinant Rac1, and the indicated amounts of cytosolic proteins (9). After incubation for 30 min at 37°C, ethanol/HCl precipitable radioactivity was measured.

Results and Discussion

Genetic and Physical Mapping of gm.

Previously, gm had been localized to a 1.5-centimorgan (cM) interval on mouse chromosome 14 (10). Genotyping of 29 loci in 2,462 gm backcross mice narrowed the gm critical region to 0.64 cM between D14Mit63 and D14Mit122 (Fig. 2). Three backcrosses were used to minimize the effect of a local recombination cold spot observed in a Mus spretus backcross. Genetic markers flanking and within the gm critical region were used to isolate contiguous YAC and BAC clones encompassing gm. To complete the cloning of the interval, novel, polymorphic simple sequence repeats (D14Sfk1, D14Sfk3, and D14Sfk5) and sequence-tagged sites were isolated from critical region YACs by inverse repetitive element-PCR (5). They were mapped and used to isolate additional YACs and BACs to cover the interval (Fig. 2). A family of serine protease genes (Mcpt1, Mcpt2, Mcpt4, Mcpt5, Mcpt9, McptL, Ctla1, Ctla5, Ctla6, Ctla7, and Ctsg) were mapped to the gm critical region but excluded from gm candidacy by virtue of incompatible distribution of expression or absence of mutations (data not shown).

Figure 2.

Genetic and physical maps of mouse chromosome 14 in the vicinity of gm. (A) Consensus genetic map of the gm region on chromosome 14 derived from three backcrosses (MGD J:52722). The circle represents the centromere. Distances between markers are in centimorgans (cM). Numbered loci represent microsatellite markers. Relative positions of loci were ascertained from 769 [(C57BL/6J-gm × SPRET)F₁ × C57BL/6J-gm], 727 [(C57BL/6J-gm × PWK)F₁ × C57BL/6J-gm], and 966 [(C57BL/6J-gm × DBA)F₁ × C57BL/6J-gm] backcross mice. No recombination was found between gm and Rabggta. Dotted lines link the consensus gm genetic map to the physical map. (B) Physical map of the gm nonrecombinant interval (MGD J:52722). The gm critical region is delineated by chromosome crossovers (denoted by red X). gm was flanked by D14Mit63 (proximal), and D14Mit153 and D14Mit122 (distal). YAC (blue lines) and BAC (red lines) clones are shown. Loci physically mapped within YAC and BAC clones are indicated with colored circles. The location of Rabggta is indicated by red circles. Green circles indicate genes near gm that (C) also map to the homologous human chromosomal segment (14q11.2).

Candidate Gene Isolation.

Additional candidate genes were identified from pooled gm contig YACs and BACs by using a modified direct cDNA selection method (6). Nineteen of 160 selected cDNA fragments were derived from Rabggta. Rabggta-derived sequence-tagged sites were physically mapped on six gm critical region YAC and BAC clones (Fig. 2). A 2.2-kb M. spretus BamHI restriction fragment length polymorphism enabled genetic mapping of Rabggta in the M. spretus backcross. Rabggta cosegregated with gm (Fig. 2), confirming derivation from the gm critical region.

Because mouse Rabggta was novel, cDNAs corresponding to the full-length coding domain were isolated and sequenced. The mouse cDNA shared 94% and 86% nucleotide identity to rat and human Rabggta, respectively.

Rabggta Mutation Identification.

Analysis of the genomic sequence of Rabggta revealed a single difference between C57BL/6J-+/+ and C57BL/6J-gm/gm DNA: The normal 3′ terminal nucleotide of intron α (guanine) was replaced with an adenine in gm DNA (Fig. 3). This substitution created a Tsp509I restriction site that was used to show perfect cosegregation of Rabggta and gm in 2,462 meioses.

Figure 3.

A splice acceptor mutation in Rabggta causes abnormal splicing. (A) Genomic structure of the 5′ end of mouse Rabggta. Exon α encodes 5′ untranslated region. The initiation methionine (bold and underlined ATG) occurs at the 3′ end of exon 1. Rabggta nucleotide sequence above is C57BL/6J-+/+ and below is C57BL/6J-gm/gm. The splice acceptor mutation is indicated by *. Rabggta exons connected by solid green lines represent normal splicing. Dashed red lines indicate exon 1 skipping in gm/gm RNA, and dotted blue lines indicate cryptic splice sites used by gm/gm. (B) RT-PCR analysis of RNA from wild-type (+/+) and mutant (gm/gm) bone marrow. The Rabggta oligonucleotides used amplified the region between exon α and exon 2. Each band was excised and sequenced: Normal bone marrow (+/+) shows the expected product of 486 bp (band a) (GenBank accession no. AF127656), and a product (638 bp) retaining intron α (band b) (GenBank accession no. AF127658). gm/gm bone marrow shows three bands: Band c (429 bp) (GenBank accession no. AF127657) represents exon 1 skipping; band b′ (638 bp) (GenBank accession no. AF127659) retains intron α; band d (557 bp) (GenBank accession no. AF127662) represents utilization of cryptic splice sites in intron α and exon 1. (C) Sequence analysis of Rabggta genomic DNA. Arrows indicate the splice acceptor mutation: +/+ = G, gm/gm = A, gm/+ exhibits both nucleotides (G and A).

Rabggta RNA Splicing Defects.

Typically, substitution of the obligatory 3′ terminal guanine of the consensus splice acceptor site prevents normal mRNA splicing and leads to intron retention, exon skipping, or cryptic splice site utilization (11). RT-PCR of the 5′ end of Rabggta RNA (corresponding to exons α, 1, and 2) disclosed aberrantly spliced Rabggta transcripts in gm bone marrow (Fig. 3). This tissue was used because the hematologic abnormalities of gm homozygotes are corrected by bone marrow transplantation (3). In normal mice, RT-PCR of the 5′ end of Rabggta gave two products. The most abundant product was correctly spliced (exon α - exon 1 - exon 2; Fig. 3, +/+ product a). A second, minor product in normal mice retained intron α (exon α - intron α - exon 1 - exon 2; Fig. 3, +/+ product b) and may be immature, because it could only be marginally detected by RT-PCR with cytosolic or poly(A)⁺ RNA (data not shown).

In gm homozygotes, the major Rabggta RT-PCR product of +/+ bone marrow was not observed (product a, exon α - exon 1 - exon 2). Instead, the most abundant gm product skipped Rabggta exon 1 (Fig. 3, gm/gm product c; exon α - exon 2). Translation of this transcript was not anticipated to yield functional Rabggta protein because the initiation codon was absent, and no alternative, in-frame start codon was present (Fig. 3).

A second novel, but much less abundant, gm Rabggta RT-PCR product represented activation of a cryptic splice donor (within intron α, 84 bp downstream of the normal splice donor site) and acceptor (within exon 1, 13 bp downstream of the G→A substitution) sites (Fig. 3, gm/gm product d). Utilization of these cryptic splice sites did not result in removal of the initiation codon, and the existence of this transcript may explain the presence of residual Rabggta in gm homozygotes.

Like +/+ bone marrow, a rare product that retained intron α occurred in gm/gm RNA (exon α - intron α - exon 1 - exon 2; Fig. 3, product b′). This gm/gm transcript also was anticipated to be functional if polyadenylated, because the initiation codon of Rabggta is in exon 1, and therefore retention of gm intron α (with G→A substitution) should not disrupt translation. However, RT-PCR of cytosolic or poly(A)⁺ RNA suggested that it was an immature, nonpolyadenylated RNA, like the corresponding transcript in +/+ bone marrow (data not shown).

These aberrant Rabggta splice variants were found in all tissues tested (gm bone marrow, melanocyte, and kidney; data not shown). As expected, both normal and mutant products were observed in gm/+ heterozygote RNA (data not shown). No other defect in splicing was observed in gm/gm; other regions of Rabggta mRNA produced RT-PCR products of expected size and sequence, with the exception of a Rabggta isoform that retained intron 9 (in both +/+ and gm/gm mice).

The enzyme Rab GGTase is composed of Rabggta and a β-subunit (Rabggtb) and absolutely requires Rab escort protein (Rep) 1 or 2 for activity (12, 13). The gm reduction in Rabggta was specific. Quantitative RT-PCR revealed no changes in the abundance of transcripts of Rabggtb, Rep1 or Rep2 in gm/gm bone marrow or kidney (data not shown). Furthermore, Northern blots disclosed no difference between gm/gm and +/+ tissues in the overall abundance of Rabggta mRNA (data not shown).

Rabggta Protein Abundance.

Western blots of gm platelets showed one consequence of aberrant Rabggta mRNA splicing to be ≈70% reduction in 60-kDa Rabggta protein in gm platelets compared with +/+ (Fig. 4). It is likely that residual Rabggta protein in gm/gm platelets resulted from translation of the novel gm Rabggta transcript that used cryptic splice sites (Fig. 3, gm/gm product d).

Figure 4.

Decreased Rabggta expression and Rab GGTase activity in gm/gm mice. (A) Western blots of wild-type (+/+), heterozygous (gm/+), and mutant (gm/gm) platelets probed with Rabggta and β-actin antibodies. Rabggta (60 kDa) was reduced ≈70% in gm/gm. This experiment was repeated four times with the same results. (B) Rab GGTase activity of gm/gm and +/+ platelets. Duplicate cytosolic protein extracts from two samples of pooled platelets from gm (open symbols) and C57BL/6J mice (solid symbols) were assayed for Rab GGTase activity by measuring transfer of [³H]geranylgeranyl ([³H]GG) to Rab1a (15). This experiment was repeated four times with platelets, and twice with liver, spleen, and kidney protein extracts with similar results. (C) GGTase-I activity of gm/gm and +/+ pooled platelets. GGTase-I activity was determined by measuring transfer of [³H]GG to Rac1. Each point represents an average of duplicate reactions. For Rab GGTase reactions, a control reaction incubated after addition of 50 mM EDTA was subtracted from all samples. For GGTase-I reactions, a control reaction incubated with buffer alone was subtracted from all samples. These experiments were repeated at least twice in each of the liver, kidney, and spleen extracts with similar results to platelets.

Rab GGTase Enzymatic Activity.

Rab GGTase activity in gm/gm and +/+ platelets was examined with an in vitro assay of Rab1a prenylation (Fig. 4) (9). No other enzyme (including GGTase I and farnesyl-transferase) can prenylate Rab1a (14, 15). gm/gm Rab GGTase activity was reduced ≈70% compared with +/+ platelets (Fig. 4), matching the decrease in Rabggta protein. Similar results were reproducibly obtained with liver, kidney, and spleen extracts (Fig. 4 and data not shown). The gm reduction in Rab GGTase was specific: no difference was observed between gm/gm and +/+ in activity of GGTase-I, a related but different enzyme (16) (Fig. 4). Thus the Rabggta splicing defect in gm results in decreased Rabggta and Rab GGTase activity.

Rab27 Hypoprenylation.

A previous study showed that gel migration of some small GTP-binding proteins was abnormal in gm platelets (2). Novel bands (≈25 and ≈28.5 kDa) of unknown identity were observed on GTP overlays of Western blots of gm platelets (Fig. 5) (2). These were not observed in gm liver, kidney, brain, spleen, macrophages or neutrophils (2). Rab proteins represented candidates for these abnormal GTP binding proteins. Specifically, Rab27a or Rab27b were excellent candidates because they are abundant in platelets (17). Blots stained with an antibody specific for both Rab27 isoforms disclosed a single band of 27.5 kDa in normal platelets and bone marrow, but an additional, mobility-shifted (28.5 kDa) isoform in gm/gm (Fig. 5). This isoform may represent unprenylated Rab27 because the electrophoretic mobility of unmodified Rab proteins is slower than prenylated forms in SDS-containing gels (14, 18). Not all small GTPases showed abnormal mobility; no differences between normal and gm platelets were observed on immunoblots of Rab4 or Ras (2).

Figure 5.

Deficient prenylation of Rab27 in gm platelets. (A) Western blots of wild-type (+/+), heterozygous (gm/+), and mutant (gm/gm) platelets and bone marrow incubated with Rab27 and β-actin antibodies. This experiment was repeated five times with the same results. (B) Western blots of subcellular fractions of wild-type (+/+) and mutant (gm/gm) platelets incubated with radiolabeled GTP or Rab27 antibodies. The subcellular fractions were total protein (T), membrane-associated protein (M), and soluble-fraction protein (S). This experiment was repeated five times with similar results. (C) Photomicrographs (×200) of +/+ and gm/gm platelets stained with Rab27 antibodies and detected with a FITC-conjugated secondary antibody. This experiment was repeated three times with similar results.

Additional evidence for Rab27 hypoprenylation was sought by examining gm platelets for membrane-associated Rab27 (which depends on prenylation) (9). Rab27 was found mainly within the membrane fraction of normal platelets, but predominantly in the soluble fraction of gm platelets (Fig. 5). GTP overlay studies showed the distribution and mobility of the abnormal Rab27 band to match that of a previously described “novel” gm GTP-binding protein (Fig. 5) (2).

Final evidence for Rab27 hypoprenylation in gm platelets was found by immunohistochemisty. Rab27 antibodies labeled punctate structures in normal mouse platelets, consistent with Rab27 localization on intracellular granules (Fig. 5). In contrast, gm platelets displayed more diffuse cytoplasmic Rab27 staining (Fig. 5). Collectively, the data suggest that Rab GGTase deficiency in gm platelets results in deficient Rab27 prenylation. The existence of other “novel” bands in GTP overlays of Western blots suggests that additional GTP-binding proteins also may be hypoprenylated in gm platelets (Fig. 5). Preliminary analyses suggest, as expected, deficiencies in prenylation of other Rabs in gm platelets (data not shown).

Correlation of gm Genotype and Phenotype.

Evidence has been presented that a mutation in Rabggta, the gm genotype, results in impaired prenylation of Rab GGTase substrates, such as Rab27. Unprenylated Rab proteins lack the ability to attach to membranes and may account for the hematologic phenotype of gm (2, 3, 9, 12, 23). Because Rab proteins regulate intracellular protein trafficking and vesicular transport by cycling between membrane-bound and cytosolic forms, failure to attach to intracellular membranes ablates this regulatory activity. Previous studies have documented disordered vesicular transport in gm platelets and megakaryocytes: Proteins destined for delivery from the trans-Golgi network to α-granules are misrouted in gm platelets (2, 3, 12). As a result, gm platelet α-granules lack fibrinogen, von Willebrand factor (vWF) and platelet factor 4. vWF is secreted from gm megakaryocytes instead of being transported to α-granules (2). In addition to these qualitative trafficking defects in α- and δ-granules, platelet synthesis is decreased in gm mice. Blood platelets are synthesized at the ends of proplatelet processes, long tube-like extensions produced by differentiated megakaryoctyes (24). During platelet assembly the megakaryocyte demarcation membrane system is extensively and dynamically remodeled within the proplatelet processes. Further, platelet production within proplatelet processes critically depends on both microtubule and actin cytoskeletal components (24). Potential roles of Rab proteins in platelet biogenesis are thus suggested. First, Rab proteins are well known to play key roles in membrane remodeling and trafficking (25). Second, several lines of evidence link Rab-mediated processes of vesicle trafficking with the actin and microtubule-based cytoskeletons (25). Finally, the distribution of membrane complexes is abnormal in gm megakaryocytes, and gm platelets are larger and more heterogeneous in size than normal (2). Thus, phenotypic effects of gm in bone marrow are explicable on the basis of Rabggta mutation.

Association of Rab GGTase with maintenance of normal platelet count and bleeding time implies that Rab GGTase represents a novel therapeutic target for thrombocytosis and clotting disorders (such as myocardial infarction, stroke, or deep venous thrombosis). Compounds that selectively inhibit Rab GGTase may be effective in these disorders when administered acutely to mimic gm by reducing platelet Rab GGTase activity.

Comparison of gm and Choroideremia (CHM).

The molecular pathology, but not the phenotype, of gm resembles the X-linked human disorder CHM. In both, mutated genes encoding proteins involved in Rab geranylgeranylation result in Rab dysfunction: Rabggta is mutated in gm and REP1 is mutated in CHM (19, 20). Some Rab GGTase activity is retained in gm, probably by aberrant splicing that rescues certain Rabggta transcripts. In CHM, partial activity remains because two redundant loci encode REPs (20, 21). In both, a consequence of Rab GGTase deficiency is decreased Rab27 prenylation, as evidenced by change in mobility and membrane association (9). The phenotype of gm, however, is different from CHM. Although pigmented tissues are affected in both, gm features partial cutaneous albinism whereas CHM has progressive atrophy of the choroid and retina (9). Furthermore, CHM patients do not have prolonged bleeding time or platelet SPD. These differences may reflect genetic complementation by REP2 in CHM, but not gm, or species differences as suggested by phenotypic differences between CHM patients and Chm knockout mice (22).

Human Platelet SPD Diseases.

The human genetic disorders most similar to gm are GPS and α/δ-SPD (26). Like gm, GPS platelets are reduced in number, are increased in size, and lack α-granule proteins (26–28). Mis-sorting and inappropriate secretion of α-granule proteins occurs in GPS, as in gm (28). GPS patients, however, do not exhibit δ-granule storage pool deficiency or partial albinism. Like gm, α/δ-SPD is characterized by reduction in the number and contents of both α- and δ-granules (29). RABGGTA and RABGGTB are excellent candidate genes for GPS and α/δ-SPD.