Rab27b regulates number and secretion of platelet dense granules

Abstract

The Rab27 GTPase subfamily consists of two closely related homologs, Rab27a and Rab27b. Rab27a has been shown previously to regulate organelle movement and regulated exocytosis in a wide variety of secretory cells. However, the role of the more restrictedly expressed Rab27b remains unclear. Here we describe the creation of Rab27b knockout (KO) strain that was subsequently crossed with the naturally occurring Rab27a KO line, ashen, to produce double KO (Rab27a^ash/ash Rab27b^−/−) mice. Rab27b KO (and double KO) exhibit significant hemorrhagic disease in contrast to ashen mice. In vitro assays demonstrated impaired aggregation with collagen and U46619 and reduced secretion of dense granules in both Rab27b and double KO strains. Additionally, we detected a 50% reduction in the number of dense granules per platelet and diminished platelet serotonin content, possibly due to a dense granule packaging defect into proplatelets during megakaryocyte maturation. The presence of Rab27a partially compensated for the secretory defect but not the reduced granule number. The morphology and function of platelet α-granules were unaffected. Our data suggest that Rab27b is a key regulator of dense granule secretion in platelets and thus a candidate gene for δ-storage pool deficiency in humans.

Rab proteins constitute a large group within the Ras superfamily of monomeric GTP-binding proteins (1, 2). Rab GTPases are critical regulators of vesicular transport steps in endocytic and exocytic pathways, acting as molecular switches oscillating between active GTP-bound and inactive GDP-bound states (3–5). Rab27a and Rab27b constitute the Rab27 subfamily and share 71% identity (2). Much recent interest has focused on Rab27a (6–8). Defects in RAB27A are responsible for Griscelli Syndrome in humans, a rare autosomal disorder characterized by partial albinism, variable cellular immunodeficiency, and an acute phase of uncontrolled T lymphocyte and macrophage activation (9). A spontaneous mouse knockout (KO) of Rab27a, designated ashen, possesses similar characteristics, with coat color dilution and defective cytotoxic T lymphocyte function (10, 11). Rab27a is expressed in a wide variety of secretory cell types, where it localizes to mature secretory vesicles and lysosome-related organelles (12). Proposed functions for Rab27a include regulation of organelle motility and promotion of the tethering/fusion of vesicles in regulated secretion (6–8). For example, in pigmented skin melanocytes, Rab27a mediates the interaction of melanosomes with the actin-based motor, myosin Va (8); when in cytotoxic T lymphocytes, it recruits Munc13-4, which is necessary for the release of lytic granules (13).

In contrast to Rab27a, Rab27b expression is much more restricted. Recent studies suggest that Rab27b is expressed mainly in platelets, stomach, large intestine, pancreas, pituitary, and bladder (14–17). The role of Rab27b remains unclear, although several facts suggest that Rab27a and Rab27b could be functionally redundant. First, Rab27a and Rab27b interact with the same family of effectors, including melanophilin, Myrip, the synaptotagmin-like proteins (Slp1–5), and Munc13-4 (6). Secondly, transgenic expression of Rab27b cDNA in melanocytes, where it is not normally expressed, rescues the coat color defect of ashen mice (14).

Rab27b (formerly named c25KG) was originally purified as an abundant GTP-binding protein in platelets (18), and our subsequent studies have confirmed the high level of Rab27b expression in platelets (14). We previously suggested a specific role for Rab27b in platelet function based on the finding that the expression of Rab27b, but not that of Rab27a or five other Rabs also expressed in platelets, appears to be regulated directly by the transcription factor nuclear factor-erythroid 2 (NF-E2) (19). NF-E2 is a critical controller of platelet biogenesis within megakaryocytes. NF-E2 KO mice lack circulating platelets (20), and these mice show reduced mRNA and undetectable protein levels of Rab27b (19). Here we generated Rab27b KO and double Rab27 KO (Rab27a^ash/ash Rab27b^−/−) mice, which exhibit defects in platelet function.

Results

Generation of Rab27b KO Mice.

To generate Rab27b KO mice, we used the conditional Cre–loxP system of site-specific recombination (Fig. 1A). Initially, we obtained two 125-kb mouse BACs containing the mouse Rab27b gene and cloned an 11.5-kb DNA fragment containing exons 2 and 3. In the 5′ end of intron 3 (700 bp downstream of exon 3), we inserted a cassette containing two loxP sites flanking a neomycin resistance gene (Neo^r) under control of the PGK promoter, followed by a polyA sequence (Fig. 1A). Another loxP site was inserted 400 bp upstream of the 5′ end of exon 2 in the same orientation as the other two loxP sites, enabling deletion of exon 2 and 3 after introduction of Cre recombinase. Deletion of exons 2 and 3 leads to the loss of the first 79 aa, with retention of only one methionine residue in exon 4. However, the resulting severely truncated protein product is predicted to be very unstable and quickly degraded.

Fig. 1.

Generation of mice carrying the conditional and KO Rab27b alleles. (A) Targeting vector pTT29 carrying three loxP sites, a neomycin-resistance gene (Neo^r), and two homology arms were used to generate the Rab27b^3lox allele in GSI-1 ES cells by homologous recombination. Diagnostic HindIII and XbaI restriction sites and corresponding 5′ and 3′ probes were used to identify correctly targeted ES clones. Cre-mediated recombination among the three loxP sites within the Rab27b^3lox allele results in three possible alleles: Rab27b^flox, Rab27b^null, and Rab27b^null+Neo, which were distinguished by Southern blot analysis by using EcoRI digestion and probe A. (B) Results of Southern blot analysis by using HindIII digestion and 5′ probe and XbaI digestion and 3′ probe are shown for five correctly targeted clones (clones 33, 99, 131, 181, and 200) and a wild-type (wt) ES clone. (C) Results of Southern blot analysis by using EcoRI digestion and probe A are shown for Rab27b^flox/WT, Rab27b^null/WT, and Rab27b^null/null (Rab27b KO) mice. (D) Western blotting by using platelet lysates from Rab27b KO and wild-type mice and anti-Rab27a antibody, 4B12, and anti-calnexin antibody as a control. GST-Rab27b protein was used as positive control.

Mouse GSI-1 ES cells isogenic with the Rab27b BAC clone were electroporated with linearized targeting vector, and clones were selected with G418. Initial screening of 336 G418-resistant clones by Southern blotting with XbaI digestion and a 3′-end probe resulted in seven positive clones, all of which had correct integration at the 5′-end confirmed by Southern blot with HindIII digestion and a 5′-end probe (Fig. 1B). Blots were stripped and reprobed with the neo probe to confirm single integration of the construct. To confirm the presence of the loxP site in intron 1, we used a PCR approach and identified five correctly targeted clones.

Rab27b^3lox/+ mice were crossed with a transgenic mouse line expressing Cre–recombinase under the control of the ubiquitous PGK promoter to produce Rab27b KO alleles. Both males and females were used as a PGK-Cre-positive parent. In crosses with PGK-Cre-positive males, deletion occurred strictly with inheritance of the transgene, whereas in crosses with PGK-Cre females deletion occurred even without inheritance of the Cre transgene, confirming previous reports that the cytoplasmic pool of Cre protein within the oocyte is sufficient to promote recombination (21). Animals carrying a Rab27b^null allele were identified by PCR, and Southern blotting and homozygous Rab27b^nul/nulll (Rab27b^−/−) mice were generated (Fig. 1C). We confirmed the absence of Rab27b protein expression in Rab27b^−/− mice by immunoblotting with the use of a specific anti-Rab27b antibody (Fig. 1D).

Rab27b^−/− mice were carefully inspected at regular times in search of any general phenotype, such as developmental abnormalities, feeding, and other behavioral changes (e.g., death rate and so on). No noticeable abnormalities were detected, including no evidence of coat color dilution. Rab27b^−/− mice were crossed with ashen mice to produce heterozygous animals (Rab27a^ash/+Rab27b^−/+), which were crossed to generate double KO animals (Rab27a^ash/ash Rab27b^−/−). Double KO animals were viable and similar to ashen mice in coat color.

Analysis of Platelet Function in Rab27 KO and Double KO Mice.

We have previously described that ashen mice do not exhibit platelet abnormalities (14). Our findings were contrary to results produced by Wilson et al. (11), but the same authors recently described that the bleeding phenotype observed in their ashen strain was in fact due to a defect in a second gene, Slc35d3 (22), thus supporting our conclusion that Rab27a defect does not affect platelet function.

As described above, Rab27b is more likely to be functionally important in platelets, and the availability of the new strains allowed an examination of platelet morphology and function. We started with standard hematological analysis. First, we measured platelet numbers in wild-type, Rab27b KO, and double KO mice by using a ZM counter (Beckman Coulter, Fullerton, CA). We found no differences among wild-type (5.8 ± 1.0 × 10⁸ platelets per milliliter of blood), Rab27b KO (5.6 ± 0.9 × 10⁸), and double KO (5.7 ± 0.5 × 10⁸) blood samples. Cell parameters also were within the normal range according to FACS measurements that were performed with FITC-conjugated anti-CD41 antibody. The following values were obtained for wild type, Rab27b KO, and double KO, respectively: forward scatter (arbitrary units): 15.14 ± 1.00, 14.89 ± 0.52, and 16.28 ± 0.64; side scatter (arbitrary units): 204.11 ± 10.6, 206.86 ± 3.02, and 197.01 ± 4.81; platelets in relation to total number of cells (%): 8.29 ± 0.96, 9.54 ± 1.39, and 8.87 ± 1.63. These results suggest that neither Rab27b KO nor double KO exhibits reduced platelet number or size.

Next, we performed bleeding tests. Initially, we started with a common method of immersing the tail of a deeply anesthetized mouse in saline and measuring the time needed for the bleeding to stop. However, we found this method unreliable, because bleeding would frequently never stop completely, or could stop and restart. Hence, we developed a semiquantitative bleeding test based on measuring the weight loss after 10-min bleeding from the tail tip, which is a reflection of the blood loss during the test (see Materials and Methods). We observed a significant bleeding defect in all mouse strains homozygous for the Rab27b KO allele (Table 1). These mice lost significantly more weight (0.34–0.39 g) than mice heterozygous for Rab27b⁻ allele or wild-type mice (0.13–0.16 g). The presence or absence of Rab27a had no noticeable effect on weight loss (bleeding). As a positive control, we performed similar bleeding test for RabGGTase-deficient gunmetal mice, previously shown to exhibit a bleeding defect (23). We found that the bleeding defect caused by Rab27b KO is more severe than that of gunmetal mice, which lost, on average, 0.18 ± 0.1 g per mouse. These results indicate that Rab27b deletion causes a severe defect in platelet function in vivo.

Table 1.

Bleeding test in Rab27-deficient mice

Mouse genotype		No. of mice tested	Blood loss per mouse, g	P
Rab 27a	Rab27b
+/+	+/+	15	0.13 ± 0.08	n/a
+/+	−/+	26	0.13 ± 0.08	0.9920
ash/+	−/+	7	0.14 ± 0.11	0.7102
ash/ash	−/+	9	0.14 ± 0.1	0.6377
+/+	−/−	10	0.34 ± 0.2	0.0011
ash/+	−/−	9	0.37 ± 0.2	0.0004
ash/ash	−/−	7	0.39 ± 0.17	0.0001

Values for blood loss are averages ± SD. P values are according to Student's t test for two populations; each mouse strain was compared to Rab27a^+/+Rab27b^+/+. n/a, not applicable.

Next, we tested the aggregation response in vitro to various agonists. Aggregation in response to collagen was severely compromised in Rab27b KO and double KO. At a high dose of collagen (20 μg/ml), aggregation responses in Rab27b KO and double KO were reduced in comparison with wild-type mice (49 ± 8% and 45 ± 9% vs. 62 ± 5%) (Fig. 2A). At a lower dose of collagen (5 μg/ml), aggregation responses in Rab27b KO and double KO were severely reduced in comparison with wild-type mice (9 ± 6% and 7 ± 4% vs. 46 ± 7%). Consistently, aggregation responses to the thromboxane A₂ mimetic U46619 in Rab27b KO and double KO were noticeably reduced in comparison with control. At high dose (10 μM) of U46619, aggregation responses of Rab27b KO and double KO mice were 48 ± 9% and 47 ± 6% vs. 62 ± 7% for the wild type. At a low dose (1 μM) of U46619, aggregation responses were further reduced for Rab27b KO and double KO platelets (41 ± 7% and 34 ± 10%), whereas wild type remained unchanged (65 ± 8%). As low dose of collagen and U46619 require generation and release of secondary agonists for complete aggregation responses (e.g., secretion of ADP, serotonin from dense granules, and formation of thromboxane A₂), these results suggest that Rab27b may play a role in these amplification cascades. Conversely, the platelet aggregation responses to thrombin in mutant platelets were strong and similar to control, although the shape of the curve showed a reduced rate of primary aggregation (Fig. 2B).

Fig. 2.

Platelet aggregation studies. Washed platelets were prewarmed at 37°C for 3 min, and an agonist, either collagen (A) or thrombin (B), was added (marked by an arrow). The aggregation reaction was recorded at 37°C for 5 min by using a dual aggregometer (Chrono-log).

In vitro secretion of platelet dense granules was assessed by a 5-hydroxytryptamine (5-HT) release assay (Fig. 3A). The uptake of 2-(5-hydroxy-3-indolyl)[2-¹⁴C]ethylamine creatinine sulfate complex (¹⁴C-5-HT) by platelets in all strains during 1 h of incubation at 37°C was similar. Rab27b KO and double KO platelets exhibit a significant defect in ¹⁴C-5-HT release in response to stimulus. In contrast to nearly 100% release of the ¹⁴C-5-HT taken up by the wild-type platelets, maximal release for Rab27b KO was only 76.1 ± 7.5% and was even lower for the double KO (46.4 ± 6.5%). For the latter, the maximal release remained similar when the concentration of thrombin was increased further (up to 4 units/ml). Interestingly, we observed that the defect in ¹⁴C-5-HT secretion in Rab27a^ash/+Rab27b^−/− mice was as severe as in double KO mice (Fig. 3A). These data further suggest that the absence of Rab27b leads to a dense granule secretory defect. Furthermore, it suggests that Rab27a also participates in secretion of dense granules as double KO platelets exhibit a more severe defect than single Rab27b KO platelets. Consistently, a reduced level of Rab27a in platelets seems to limit the ability of Rab27a to functionally compensate for Rab27b.

Fig. 3.

In vitro secretion assays. (A) Secretion of dense granules. PRP was isolated from the blood of wild-type (black bars), Rab27b KO (gray bars), Rab27a^ash/+Rab27b^−/− (white bars), and double KO (hatched bars) mice and incubated with ¹⁴C-5-HT at 37°C for 1 h. Platelets were sedimented, washed, and challenged with different doses of thrombin (0.25, 0.5, 1, 2 units/ml) for 5 min at 37°C after prewarming for 3 min. Platelets were fixed, and release of 5-HT was measured by using a scintillation counter. The formula used to determine the percentage of 5-HT release was as follows: (release−background) × 100/(total − background) (%). The assay was repeated three times; each time, blood from three mice of the same genotype was pooled. Data are presented as averages ± SD. The initial uptake of ¹⁴C-5-HT (dpm per 10⁷ platelets) was as follows: wild type, 2,311 ± 79; Rab27b KO, 2,562 ± 301; double KO, 1,957 ± 167. (B) Secretion of α-granules. Whole blood from double KO mice (hatched bars) and double heterozygous control mice, Rab27a^ash/+Rab27b^−/+ (black bars), was stained with biotin-conjugated anti-CD62P antibody and phycoerythrin-conjugated streptavidin in the absence or presence of phorbol-12-myristate-13-acetate (0, 0.06, 0.125, 0.25, 0.5, and 1 μM) and measured by FACS. For each genotype, blood from three mice was used and counted individually. The data presented are geometric means. (P = 0.9498 as calculated by Student's t test for two populations.) The assay was repeated two times.

In the 5-HT release test, we noticed that background levels of radioactivity in the absence of agonist were consistently higher in Rab27b KO and double KO mice (12% of total loaded radioactivity compared with 5% for the wild type). These data suggested the possibility that Rab27b KO might lead to increase in constitutive secretion (unstimulated release) of dense granules. To address this question, we performed similar ¹⁴C-5-HT loading experiments and incubating platelets for up to 40 min at 37°C without agonist stimulation. The amount of released ¹⁴C-5-HT did not increase with time (wild type, 5.6 ± 0.8%; Rab27b KO, 12.7 ± 1.6%; and double KO, 12.5 ± 1.8%). These data suggest that Rab27b loss does not lead to constitutive secretion of dense granules.

Next, we investigated secretion of α-granules by measuring exposure of P-selectin on the platelet surface by FACS after activation with phorbol-12-myristate-13-acetate (Fig. 3B). Platelets were identified by staining with FITC-conjugated anti-CD41 antibody, whereas P-selectin was detected by using biotinylated anti-P-selectin antibody and phycoerythrin-conjugated streptavidin. We observed no significant differences in P-selectin levels on the surface of activated platelets from Rab27b KO (data not shown) and double KO mice, suggesting that Rab27b does not play a major role in secretion of platelet α-granules.

Morphologic Analysis of Dense Granules in Rab27 KO Mice.

Platelet morphology was studied by conventional EM, and no gross abnormalities were detected (Fig. 4). However, quantitative analysis of granule numbers revealed that Rab27b KO and double KO platelets exhibited 50% reduction in the number of dense granules per platelet (Table 2). Consistently, the total amount of platelet endogenous 5-HT in Rab27b KO lines was reduced ≈2-fold (Table 2). Both observations are consistent with the compromised platelet responses in vivo (tail bleeding) and in vitro (aggregation) in these mice. Moreover, the total platelet serotonin content was lower in Rab27b^−/− mice irrespective of Rab27a (Table 2). These data suggest that Rab27b, not Rab27a, is implicated in controlling platelet dense granule number.

Fig. 4.

Transmission electron microscopy of wild-type (A), Rab27b KO (B), double KO (C), and Rab27a^ash/ashRab27b^−/+ (D). Insets show dense granules in higher-magnification views of the boxed areas. (Scale bars, 500 nm.)

Table 2.

Number and content of dense granules

Mouse genotype		Total serotonin content, μg per 109 platelets	Dense granules per cell
Rab27a	Rab27b
+/+	+/+	5.76 ± 0.73	0.33 ± 0.07
+/+	−/+	5.58 ± 1.31	ND
ash/+	−/+	5.37 ± 1.00	0.33 ± 0.02
ash/ash	−/+	4.88 ± 0.78	0.34 ± 0.04
+/+	−/−	2.64 ± 0.56	0.16 ± 0.02
ash/+	−/−	2.36 ± 0.59	ND
ash/ash	−/−	2.21 ± 0.31	0.19 ± 0.01

Data presented are averages ± SD. To calculate the dense granules per cell, the total number of granules visible was divided by the total number of cells in the same field. For each genotype, three of four fields were quantitated corresponding to 200–300 cells examined. ND, not determined.

Discussion

The aim of our study was to unravel a biological role for Rab27b. For this purpose, we created a Rab27b gene deletion in mice and subsequently a double KO by crossing with the natural Rab27a KO strain, ashen (Rab27a^ash/ash). Rab27b KO and double KO mice exhibit a serious bleeding defect, and aggregation studies in vitro showed reduced responses to several agonists, such as collagen and U46619. In vitro secretion of 5-HT but not P-selectin was compromised in both Rab27b KO and double KO mice. Finally, absence of Rab27b leads to a 50% reduction in the number of dense granules per platelet. Our data suggest that Rab27b is critical for the packaging and secretion of platelet dense granules, resulting in the bleeding diathesis observed in vivo.

The function of Rab27b in platelet dense granule secretion is consistent with a flurry of recent evidence linking Rab27 proteins with regulated exocytosis in nonneuronal cells (6, 7, 12). Rab27a has been more thoroughly investigated to date, presumably because it is widely expressed in secretory cells (12). However, defects in RAB27A in humans result in Griscelli syndrome affecting primarily only two cell types, melanocytes and cytotoxic T lymphocytes (9). To explain this intriguing disparity, we have previously suggested that Rab27a activity can be compensated in other cell types by Rab27b and possibly other related proteins, such as the Rab3 isoforms (14). The study of the newly generated Rab27b KO mice and double Rab27 KO mice allowed us to further address this issue, and the results presented here suggest several novel ideas. First, the double KO mice are viable and have no gross organ/tissue abnormality, suggesting that potential widespread defects in regulated secretion due to loss of Rab27a and Rab27b could be compensated for by other Rab proteins, at least partially. Second, our studies suggest that specific cell types rely differentially on a single Rab27 isoform, such as melanocytes and cytotoxic T lymphocytes for Rab27a, and platelets for Rab27b. Third, the present study suggests that Rab27 isoforms may perform similar (redundant) but also nonredundant functions in the same cell type. Our present results suggest that both Rab27a and Rab27b participate in secretion of platelet dense granules. Rab27a can indeed partially compensate for the absence of Rab27b in dense granule secretion as the in vitro secretion tests demonstrated a more severe defect in the double KO vs. the single Rab27b KO. This partial compensation could be because of lower Rab27a expression in platelets compared with Rab27b (14), a hypothesis supported by the fact that we observed a sharp decline in functional compensation in Rab27b KO mice that were heterozygous for the Rab27a^ash allele (Fig. 3). Nevertheless other Rab27b functions appeared nonredundant, such as a possible role in dense granule formation or “packaging” (see below). Finally, we detected no structural or functional impairment in α-granules, despite the localization of (a minor pool of) Rab27 proteins to these granules observed previously (14). These findings suggest that the Rab27 proteins perform nonessential functions on these organelles.

Rab27b function in dense granule secretion may be mediated by Munc13-4, a member of a protein family implicated in regulated exocytosis (24). Munc13-4 binds directly to both Rab27a and Rab27b and has been implicated in regulated secretion of platelet dense granules (25). A physiological role for Munc13-4 in hematopoietic cell secretion has been confirmed recently in a study of patients with familial hemophagocytic lymphohistiocytosis caused by mutations in this gene (26). Munc13-4-deficient familial hemophagocytic lymphohistiocytosis patients exhibit low platelet count, together with other characteristics of the disease, such as fever, hepatosplenomegaly, and hemophagocytosis (27). In addition to Munc13-4, Rab27 proteins bind numerous other effectors (19), which also could be important for platelet function. Future work should help elucidate this issue.

It is now accepted that complete aggregation response to many agonists requires an amplification cascade that involves the release of ADP and other mediators from dense granules followed by stimulation of P2Y₁₂ receptors enhancing G_i-mediated effects (28). The reduced ability of Rab27b KO and double KO mouse platelets to secrete dense granules explains the markedly reduced aggregation responses to U46619 and especially collagen. Our data showing that aggregation is unaffected for thrombin also are consistent with the above model, because thrombin itself can activate the guanine nucleotide inhibitor factor-mediated pathway. Still, much remains to be established regarding the signaling mechanism associated with granule secretion post-surface-receptor activation. PKC activity has long been established as being required, including pharmacological evidence for PKCδ-dependent PAR1-mediated dense granule secretion (29). However, little is known linking post-PKC events to granule movement and fusion to the plasma membrane events that involve Rab27b, its effectors, and SNARE members.

In addition to defective secretion, we observed a significant reduction in the number of dense granules by using EM quantitative analysis and measurement of total platelet 5-HT. Interestingly, this defect appears to be independent of Rab27a because the results were similar whether none, one, or two copies of Rab27a were present (Table 2 and Fig. 4). The reduction in dense granule number could reflect a defect in organelle biogenesis. Rab27b may be controlling a post-Golgi trafficking pathway necessary for the maturation of dense granules in megakaryocytes, as recently shown for Rab38 and Rab32 in melanosome biogenesis, a related lysosome-related organelle (30). However, it is more likely that the reduction in dense granule number in mature platelets is due to a dense granule packaging defect. Platelets derive from megakaryocytes during a complex differentiation process that involves the packaging of cytoplasmic components onto discrete areas of megakaryocyte dendritic projections, called proplatelets (31). Rab27b may control the cytoskeletal-mediated transport and distribution of dense granules during proplatelet formation in megakaryocytes, as described for Rab27a in melanosomes peripheral distribution (8, 19). The absence of Rab27b may lead to reduced or absent transition from microtubule-mediated to actin-mediated transport and hence to reduced numbers of dense granules retained in proplatelets.

The morphology and function of α-granules was not affected in any of the KO strains, in agreement with only minor association of Rab27a and Rab27b with α-granule membranes (14). This suggests that Rab27 proteins are not important for secretion of α-granules or that their loss could be compensated by other Rabs. One possible candidate would be Rab4, which was shown to be an essential regulator of Ca²⁺-induced exocytosis of α-granules (32). Specific dense granules defects have been observed in a rare genetic condition called δ-storage pool disease (33). The similarities between the patients reported and the mice in this study led us to speculate that RAB27B is a candidate gene for δ-storage pool disease.

Materials and Methods

Generation of Rab27b KO Mice.

To generate targeting vector pTT29, a 11.5-kb ClaI–KpnI DNA fragment containing exons 2 and 3 of mouse Rab27b gene was subcloned into pBlueScript vector. Thirty micrograms of targeting vector pTT29 were linearized with NotI endonuclease, purified by ethanol precipitation, and used for electroporation of 3 × 10⁷ GSI-1 cells, which were then cultured in the presence of 0.4 mg/ml G418. To test for correct integration, the following probes were generated by PCR with a 5′-end probe (forward primer, 5′-agcctttactagcagggcaaaccaagatgc; reverse primer, 5′-ttaaggaagggatttggtagagacccaccg; size, 1 kb) and a 3′-end probe (forward primer, 5′-accactgccttcttcagagatgccatgggc; reverse primer, 5′-agctacagacatgaagtccaggcagagggg; size, 1.1 kb). To confirm single integration of the targeting construct, a 500-bp neo probe was generated by using forward primer 5′-ccgcgctgttctcctcttcc and reverse primer 5′-gtacgtgctcgctcgatgcg. To confirm the presence of the most 5′ loxP site, PCR was performed with forward primer 5′-ttggatccaggaaacataggtactgaaatgg and reverse primer 5′-ttcagctggcagaatgaagaagttggagat. This PCR amplifies the wild-type allele (product size, 520 bp) and modified allele (570 bp). Two male chimeras were generated by blastocyst injections of targeted ES clone 181 into C57BL/6 embryos. Both chimeras were 80–90% chimeric and were transmitting the targeted allele.

Mouse Strains.

Mice were bred and maintained in the Imperial College animal facility in accordance with the rules and regulations of the Home Office on project license 70/6176. The transgenic line of PGK-Cre mice was obtained from Ian Roswell (Cancer Research UK, London, U.K.) and out-crossed for three generations with C57BL/6J mice. Probe A is a 1.2-kb PCR fragment generated with forward primer 5′-tccacataaatcgtgtgtgtcctcttctcc and reverse primer 5′-gtgcagaatgaagaagttggagagttttgc. Rab27b^null allele was identified with forward primer B19 (5′-ctgctgcaggatctcacatcagtg) and reverse primer B21 (5′-gaaatgggacattgggacaggagg); the size of the PCR product was 250 bp. Rab27b^WT and Rab27b^3lox alleles were identified with forward primer B19 and reverse primer B20 (5′-agcatctgtaacctagacattggc); the sizes of the PCR products were 320 and 370 bp, respectively. Routine genotyping was performed by using three primers (B19, B20, and B21) in the same PCR. The diagnostic PCR assay for Rab27b^flox allele was preformed with forward primer B5 (5′-aacactgttgcatgagctgatcgc) and reverse primer B21; the size of a product was 350 bp, whereas, for the wild type, the allele size of the product was 440 bp. The PCR cycling conditions were as follows: 94°C for 3 min, followed by 32 cycles of 94°C for 40 sec, 62°C for 40 sec, 72°C for 1 min, and, finally, 72°C for 10 min.

Immunoblotting.

Blood was drawn by cardiac puncture under terminal anesthesia in 0.4 M sodium citrate (ratio 1:9; Sigma, St. Louis, MO) or ACD (20 mM citric acid/110 mM sodium citrate/5 mM dextrose; ratio 2:8). Blood was diluted with a half volume of EHS (150 mM NaCl/1 mM EDTA/10 mM Hepes, pH 7.6) and spun at 125 × g for 15 min. Platelet-rich plasma (PRP) was collected and centrifuged at 1,000 × g for 15 min. The pellet was resuspended in lysis buffer (50 mM Hepes/10 mM NaCl/1 mM DTT, pH 7.2) supplemented with Complete protease inhibitor mixture (Roche Diagnostics, Uppsala, Sweden). Protein extracts were subjected to SDS/PAGE and transferred to PVDF membranes. The antibodies used were polyclonal anti-Rab27b antibody S086 at a 1:400 dilution (13) and polyclonal anti-calnexin antibody at a 1:5,000 dilution (Stressgen Bioreagents, Ann Arbor, MI).

Determination of Platelet Counts and Size.

To obtain platelet counts, 50 μl of a 1:2,000 dilution of the PRP was counted in a ZM counter. To assess platelet size, 50 μl of whole blood was stained with FITC-conjugated rat anti-mouse CD41 (integrin α_IIb chain) at a final concentration or 10 μg/ml (BD Pharmingen, Franklin Lakes, NJ) for 30 min. FITC-conjugated rat IgG1, κ monoclonal Ig isotype control was used at similar concentration. The analysis was performed by using FACS Calibur (BD Immunocytochemistry Systems) by using CellQuest version 3.1f software (BD Biosciences). Measurements were made on pools of blood from three mice of the same genotype and performed four to five times for each genotype. Results are presented as averages ± SD.

Bleeding Test.

Eight-week-old mice (not previously genotyped) were anesthetized by i.p. injections of 30 μl of a 2:1 ketamine/xylazine mixture. Mouse body weight was measured, a 5-mm tail tip was cut off, and tail was immersed into saline prewarmed to 37°C. Bleeding was allowed for 10 min, and then the tail was cauterized and mouse weight was determined.

Aggregation.

PRP was spun at 1,000 × g for 15 min, resuspended in Tyrode's buffer (138 mM NaCl/2.9 mM KCl/12 mM NaHCO₃/0.36 mM NaH₂PO₄/5.5 mM glucose/10 mM Hepes/0.4 mM MgCl₂, pH 7.4), spun at 1,000 × g for 15 min, and resuspended in Tyrode's buffer at a final concentration of 2.5 × 10⁸ platelets per milliliter. The aggregation traces were registered by using a dual-channel aggregometer (Chrono-log, Havertown, PA) at 37°C, with stirring at 1,000 rpm. Platelets were prewarmed at 37°C for 3 min, and then agonist was added and traces were recorded for 5 min. The agonists used were thrombin (Sigma), U46619 (Sigma), and Horm collagen (Nycomed, Roskilde, Denmark).

5-HT Release Assay.

PRP was incubated with 0.1 μCi/ml of ¹⁴C-5-HT for 1 h at 37°C. Platelets were spun at 1,000 × g for 15 min and washed and resuspended in Tyrode's buffer at 2.5 × 10⁸ platelets per milliliter; 200 μl of washed platelets was used per reaction. 5-HT release assay was performed in a shaking water bath at 37°C with a 3-min warming-up period before the addition of thrombin (Sigma). Reaction was stopped after 5 min by addition of 1/3 volume of ice-cold 4% paraformaldehyde in 50 mM EDTA, spun at 9,300 × g for 5 min. Supernatant was collected and radioactivity was counted by using a liquid scintillation analyzer (model 1900TR; Packard Instrument Co., Meridan, CT). “Total” radioactivity values were obtained by counting 200 μl of labeled, washed, unstimulated platelets. To obtain a “background” value, 200 μl of labeled and washed platelets was kept at 37°C for 8 min, fixed, and spun, and radioactivity of the supernatant was measured. To measure platelet uptake, platelets were labeled with ¹⁴C-5-HT for 1 h at 37°C and washed, and incorporated radioactivity was counted. The uptake per 10⁷ platelets is calculated as (uptake × 10⁷)/(total number of platelets).

Determination of Endogenous 5-HT Levels.

The method used was a modified procedure of Drummond and Gordon (34) that has been described previously in detail (14).

Detection of P-Selectin at the Platelet Surface.

Blood was drawn with 4% sodium citrate as above and diluted 1:6 with Tyrode's buffer. Biotinylated anti-P-selectin antibody (5 μg/ml), FITC-conjugated anti-CD41 antibody (10 μg/ml), and phycoerythrin-conjugated streptavidin (10 μg/ml) were added to 50 μl of diluted blood, together with different concentrations of phorbol-12-myristate-13-acetate, and incubated for 30 min at room temperature. Samples were fixed with 1 volume of 1% paraformaldehyde in Tyrode's buffer and analyzed by FACS.

Electron Microscopy.

PRP was collected and spun at 1,000 × g for 10 min. Pellet was washed twice with 0.4% sodium citrate in 1× PBS containing 8 μM prostaglandin E1. Pellet was resuspended in 2% paraformaldehyde and 1.5% glutaraldehyde and fixed for 1–2 h. Platelets were then osmicated, stained with tannic acid, and embedded in Epon. Ultrathin sections stained with lead citrate were viewed on a Joel 1010 transmission electron microscope (Joel, Tokyo, Japan).

Abbreviations

5-HT

5-hydroxytryptamine

¹⁴C-5-HT

2-(5-hydroxy-3-indolyl)[2-¹⁴C]ethylamine creatinine sulfate complex

KO

knockout

PRP

platelet-rich plasma.