The complementary roles of VAMP-2, -3, and -7 in platelet secretion and function

Abstract Platelet secretion requires Soluble N-ethylmaleimide Sensitive Attachment Protein Receptors (SNAREs). Vesicle SNAREs/Vesicle-Associated Membrane Proteins (v-SNAREs/VAMPs) on granules and t-SNAREs in plasma membranes mediate granule release. Platelet VAMP heterogeneity has complicated the assessment of how/if each is used and affects hemostasis. To address the importance of VAMP-7 (V7), we analyzed mice with global deletions of V3 and V7 together or platelet-specific deletions of V2, V3, and global deletion of V7. We measured the kinetics of cargo release, and its effects on three injury models to define the context-specific roles of these VAMPs. Loss of V7 minimally affected dense and α granule release but did affect lysosomal release. V3−/−7−/− and V2Δ3Δ7−/− platelets showed partial defects in α and lysosomal release; dense granule secretion was unaffected. In vivo assays showed that loss of V2, V3, and V7 caused no bleeding or occlusive thrombosis. These data indicate a role for V7 in lysosome release that is partially compensated by V3. V7 and V3, together, contribute to α granule release, however none of these deletions affected hemostasis/thrombosis. Our results confirm the dominance of V8. When it is present, deletion of V2, V3, or V7 alone or in combination minimally affects platelet secretion and hemostasis. Plain Language Summary What did we know? V8 is the primary VAMP isoform for platelet granule secretion, but V2 and V3 play compensatory roles. V3 is important for platelet endocytosis. V7 plays a minimal role in secretion and does not affect hemostasis. What did we discover? The loss of both V3 and V7 increases α and lysosomal secretion defects. Platelet-specific deletion of V2 and V3 with global V7-deletion causes defective α and lysosomal release. Secretion deficiencies in V3−/−7−/− and V2Δ3Δ7−/− have no effect on hemostasis or thrombosis. What is the impact? We show that endosomal v-SNAREs (V3 and V7) play minor roles in secretion. V3−/−7−/− and platelet-specific V2Δ3Δ7−/− mice are viable and will be valuable in in vivo studies of membrane trafficking.


Introduction
Platelet granule cargo release is involved in various aspects of platelet activation and function. 1Platelets contain three types of granules: dense granules that contain small molecules, e.g., ATP, serotonin, and calcium, which are important for platelet recruitment.Alpha granules contain a multitude of proteins, e.g., cytokines, chemokines, and growth factors.Lysosomes that contain cathepsins and β hexosaminidase, important for clot remodeling. 2 The release process is mediated by membrane proteins collectively known as Soluble N-ethylmaleimide sensitive factor Attachment protein Receptors (SNAREs).Based on their membrane localization and the amino acids at the center of their conserved SNARE domains, SNAREs are classified as target (t/Q, Glu) SNAREs and vesicle (v/ R, Arg) SNAREs (also known as vesicle-associated membrane proteins [VAMPs]).These v-and t-SNAREs form a trans-bilayer SNARE complex that promotes granule fusion to the plasma membrane/PM and subsequent release of granule cargo and externalization of luminal granule membrane proteins.While the dominant t/Q SNAREs in platelets are Syntaxin-11 (Q a -SNARE) and SNAP-23 (Synaptosomal Associated Protein 23, Q bc SNARE), [3][4][5] the reported roles of the v/R SNAREs have been more heterogeneous.Platelets contain multiple v-SNAREs: VAMP-2, -3, -4, -5, -7, -8, and Ykt6.Previous work identified V8 as the major VAMP for platelet granule secretion with V2 and V3 as secondary VAMPs, which compensate in V8's absence. 6,7Deletion of V3 alone or in combination with V2 does not affect granule secretion. 6,8,9Two points from these previous studies should be noted: 1) deletion of V2, V3, and V8 together also reduced V7; and 2) despite the deletion of three platelet VAMPs (and reduction of a fourth), secretion was not completely abolished. 67 differs from the other platelet "brevin" v/R SNAREs because of its additional N-terminal "Longin" domain that interacts with unique regulators.One characterized binding partner, the retrograde trafficking adaptor Leucine Rich Repeat Kinase 1 (LRRK1), regulates V7-mediated lysosomal exocytosis. 10Overexpression of the Longin domain in PC12 cells inhibits neurite outgrowth, while deleting this domain stimulates neurite outgrowth. 11In MDCK cells, the expression of Longin domain inhibits lysosomal secretion. 12V7 is also reported to play a role in transporting Golgiderived vesicles to late endosomal compartments and their fusion to PM. 13 V7 mediates ATP release from autophagic vesicles in HeLa cells 14 and IL12 release from dendritic cells. 15Thus V7, through its Longin domain, appears to be involved in several different endosomal trafficking and secretion pathways in nucleated cells.
V7's role(s) in platelets has been enigmatic.Its deletion mildly affects platelet granule secretion but has no significant effect on thrombosis and hemostasis ( 16 and reported here).Interestingly, V7 is localized to the periphery of human platelets upon spreading, while V3 and V8 localize to the central granulomere. 17Based on these reports, we wondered if V7 had a role in secretion, potentially masked by the compensatory activity of another platelet VAMP (i.e., V2 or V3).To address this question, we generated mouse strains lacking V7 in combination with V3 as a global, double-deletion, or with V2 and V3 as a conditional, triple-deletion.We confirmed that single deletion of V7 had only a modest effect on granule secretion, specifically lysosome release.However, the additional loss of either V3 or both V2 and V3 did exacerbate the secretion defects, notably for α granule and lysosome release but not for dense granule release.These results were confirmed using cytometry of α granule and lysosome membrane proteins.Consistent with the mild secretion defect and lack of an effect on dense granule release, mice from these strains had minimal defects in thrombosis or hemostasis in three different injury models.Our data show that V7 deletion, either alone or in combination with V2 and V3, had no effect on hemostasis and only significantly affected α and lysosome release.These data confirm the importance of V8, which was unaffected in these strains, and suggest that V7 May have only a minor subtle role in platelet exocytosis.

Reagents
Apyrase, hirudin, and proteinase K were from Sigma (St Louis, Mo).Thrombin was from Chronolog (Havertown, PA).Prostaglandin I 2 (PGI 2 ) was from Cayman (Ann Arbor, MI).Other reagents used were at least in laboratory grade.

Flow cytometry analysis
Washed mouse platelets (20 μL of 5 × 10 7 /mL) were either in resting state (no agonist) or were stimulated with thrombin (0.1 U/mL) for 2 min at RT (Room Temperature).The reactions were stopped with the addition of a twofold excess of hirudin.Platelets were incubated with 2.5 μL FITC-conjugated or PE-conjugated antibodies for 20 min at RT.The platelets were 10-fold diluted by adding HEPES-Tyrode's buffer (pH 6.5) and transferred to a polystyrene FalconTM tube (BD Biosciences, San Jose, CA).Fluorescent intensity was measured using FACScanTM flow cytometer (BD Biosciences) or CellQuestTM (BD Biosciences) with FlowJo TM (Version 10.8.0) analysis package.It should be noted that in our hands, the signals from FACS scan are higher than with CellQuest.A total of 20 000 platelets were analyzed, and fluorescent intensities were plotted as a histogram with statistical values.

Quantitative western blotting
Mouse platelets were prepared from the indicated strains, and 5 × 10 7 platelet equivalents were loaded per well.The proteins were separated by SDS-PAGE and transferred to Immobilon-P polyvinylidene fluoride (PVDF) membranes (Millipore Corp., Bedford, MA).The PVDF membrane was blocked and incubated with indicated primary antibodies.The immuno-decorated proteins were detected with alkaline phosphatase-conjugated secondary antibodies using Vista-ECF substrate (Amersham Biosciences).The proteins were visualized with a Typhoon FLA 9500 scanner and quantified using ImageQuantTL software (v 7.0, GE Healthcare).

Tail bleeding time assay
Briefly, 6-to 8-week-old sex and age-matched mice were anesthetized using ketamine 75 mg/kg i.p.A transverse incision was made ~3 mm from the tail tip, the tail was immersed in 37°C normal saline, and the time to bleeding cessation was recorded.After initial bleeding cessation, mice were observed for an additional minute to exclude re-bleeding.Bleeding in animals with bleeding times >10 min was stopped manually. 6

FeCl 3 -Induced carotid injury model
The FeCl 3 -induced carotid injury model was performed as described in Joshi et al. 2023.Briefly, 8-to 12-week-old sex and age matched mice were anesthetized with Avertin (0.2 g/kg, i.p.).The left carotid artery was exposed and a miniature Doppler flow probe (0.5VB, Transonic system Inc., Ithaca, NY, USA) was placed to monitor arterial blood flow.Thrombus formation was induced by placing a filter paper saturated with 6% FeCl 3 solution on the top of the vessel for 3 min.Time from the removal of filter paper to the cessation of blood flow was recorded.The recording was stopped after 30 min.

Jugular puncture injury model
Briefly, 8-to 12-week-old sex and age-matched mice were anesthetized with Avertin (0.2 g/kg, i.p.).The right jugular vein was exposed, and the surrounding tissue was removed.Using a syringe pump, the saline flow started at 1 mL/min rate to keep the area irrigated.The vein was punctured using a 30 G needle, and the time to bleeding cessation was recorded.

Statistical analysis
All statistical analysis was performed using SigmaPlot v15 or GraphPad Prism9.The data from bleeding time and occlusion assays were analyzed using the Log-rank test.The data from secretion assays and FACS-based experiments were analyzed by two-way ANOVA (Analysis of Variance).In all cases, the p values are as indicated and values less than 0.05 were considered significant.

Study approval
All animal work was approved by the Institutional Animal Care and Use Committee at the University of Kentucky (protocol # 2019-3384).

Generation and characterization of V3 −/−-/-and V2 ∆∆-/- mouse models
To determine if the loss of other VAMPs would unmask a role for V7 in platelets, we created a series of mouse lines (Table I).A global V3 −/ − strain was crossed to the V7 −/− strain to generate a global V3 −/− 7 −/− strain.Viable mice, which grew into adulthood, recovered and displayed normal fertility and gross anatomy.The recovered allele distributions, after matings, were at the expected Mendelian ratios.The hematological parameters (RBC, WBC, and platelet counts) were indistinguishable from wild type (Table S1).Thus, globally deleting both V3 and V7 in the same strain was not lethal.This is in stark contrast to our attempts to generate a V3 −/− 8 −/− strain, which resulted in embryonic lethality at a very early stage (~E10; Whiteheart, unpublished).Since global deletion of V2 is perinatal lethal, 23 we chose a Cre/Lox system that drives tissue-specific expression of the catalytic subunit of tetanus toxin, 21,22 an endopeptidase that specifically cleaves V2 and V3. 24,25This enzymatic strategy for post-translationally removing the protein differs from the classical deletion of exons which affects mRNAs and thus may not be as efficient.The threeway cross PF4Cre x TetTox x V7 −/− generated a mouse strain globally lacking V7 and lacking V2 and V3 in megakaryocytes and platelets.Viable and fertile mice were recovered from these crosses at the expected Mendelian ratios.There were no gross anatomical alterations in these animals.There was a trending decrease in RBCs in the females of this strain that did not reach statistical significance (Table S1).
When platelet number and mean platelet volume (MPV) were analyzed, there were some differences among the strains.Platelet number was elevated in the male V2 ∆ 3 ∆ 7 −/− mice, and an increase was trending in the females.Such differences were not seen in the V3 −/− 7 −/− mice.Males from both strains showed increased mean platelet volume (MPV), with only female V2 ∆ 3 ∆ 7 −/− mice showing a similar increase (Table S1).These differences suggest some effect of these genetic manipulations on platelet biogenesis, but this was not pursued here.The V7 gene is on the X-chromosome; however, it is unclear if that is related to the sexual dimorphism observed in these strains.
Platelet extracts from V3 −/− 7 −/− and V2 ∆ 3 ∆ 7 −/− mice were prepared and probed by immunoblotting to confirm the losses of the VAMPs and to determine if other SNARE proteins were altered in these strains.There were no significant differences in the levels of the t-SNAREs, SNAP23, and syntaxin 11 nor were two α granule proteins, Platelet Factor 4 (PF4) and P-selectin, affected (Figure 1).Levels of granule marker cargos, e.g., serotonin-uptake for dense granules, PF4 for α granules, and β-hexosaminidase for lysosomes, were not altered in platelets from any of the strains when compared to wild-type (Figure S2).Most notably, V8 was unchanged in both strains.As expected, fibrinogen levels were reduced in both V3 −/ − 7 −/− and V2 ∆ 3 ∆ 7 −/− platelets.This is consistent with our earlier report that V3 plays a role in endocytosis. 9 Granule cargo secretion from V3 −/−-/-and V2 ∆∆-/-mouse platelets A single deletion of V7 had minimal effects on stimulation-or timedependent release of serotonin from dense granules and PF4 from α granules (Figure S1). 16Notably, the secretion of β-hexosaminidase, from lysosomes, was decreased in V7's absence (Figure S1).When the secretion of the soluble granule markers was measured from the global V3 −/− 7 −/− platelets, serotonin release from dense granules was unaffected (Figure 2a-d).As with the V7 −/− platelets, the release of βhexosaminidase was reduced (Figure 2c-f).PF4 release was also VAMP-2, -3, and -7 and platelet granule secretion 3 reduced, mostly at low agonist levels and early time points, and reached wild-type levels at longer time points and higher agonist concentrations (Figure 2b-e).The release of β-hexosaminidase was reduced under all conditions tested.Analysis of V2 ∆ 3 ∆ 7 −/− platelets gave comparable results (Figure 3).Dense granule release was not affected (Figure 3a-d), and lysosomal release showed a significant decrease under most of the conditions tested (Figure 3c-f).Release of α granules was also affected (Figure 3b-e) but was further reduced at low agonist concentrations and early time points.The amplitude of the secretion defects was higher in the global knockout strain, V3 −/− 7 −/− , which could be indicative of a more complete deletion of the VAMPs compared to enzymatically deleted strains (see Methods for The actin loading control in a was run on the same gel as SNAP23 and that in B was run on the same gel as fibrinogen.(c) Ratio of fibrinogen/actin in each of the strains studied-WT, V3 −/− 7 −/− and V2 ∆ 3 ∆ 7 −/− .Each data point represents an individual experiment.Data are from n > 3. Statistical significance was calculated using a two-tailed, paired t-test.ns not significant, *p ≤ .05. explanation).These data suggest that V7 is more critical for lysosome release and that the losses of V2, V3, and V7 combined do not affect dense granule release.Single deletions of V7 (Figure S1), V3, 8 or double deletions of V2 and V3 6 do not cause a defect in α granule release, but in combination they do.Thus, the loss of V3 and/or V2 and V3 unmasked a role for V7, or potentially all three proteins, in α granule release.Release still occurs because V8, the primary platelet v-SNARE is still present.
Upon fusion, the granule membranes become one with the plasma membrane and thus granule membrane proteins become detectable on the platelet surface.We used cytometry to probe the agonist-induced exposure of two granule membrane proteins, P-selectin (α granules), and LAMP1 (lysosomes; Figure 4).Activation stimulated exposure of both markers was significantly reduced in V2 ∆ 3 ∆ 7 −/− and even more drastically in V3 −/− 7 −/− platelets (Figure 4a,b).Consistent with Figure S1, no defect was detected in P-selectin exposure in the single V7 deleted platelets, although LAMP1 exposure was reduced significantly.While this experiment did not specifically examine dense granule release, the data confirm the defects in α granule and lysosome release noted in Figure 3, using soluble granule markers.
Activation of surface integrin α IIb β 3 occurs upon platelet activation and can be measured by cytometry with JonA antibody, as can the total levels of surface α IIb β 3 .Loss of V7 either alone or in combination with V2 or V3 or both failed to robustly affect JonA binding (Figure 4c-e).In mice, there is a pool of internal α IIb β 3 that appears on the surface upon activation, much like the α granule proteins measured above.This can be detected as an activation-dependent increase of surface α IIb β 3 (Figure 4d-f).Interestingly, there is a slight decrease in α IIb β 3 exposure in the V3 −/− 7 −/− platelets (Figure 4e).Perhaps V3 and V7, together, play a role in mobilizing α IIb β 3 from this internal compartment.However, despite being statistically significant, it is hard to determine if this small difference is physiologically relevant.

Discussion
Our previous studies suggested heterogeneity in the usage of VAMPs for platelet granule exocytosis. 6,7Deletion of V8 leads to a clear secretion defect that can be further exacerbated by the additional deletion of V2 and V3.Thus, these v-SNAREs may have compensatory activity.Other reports suggested V7's contribution to platelet secretion, but the effects were modest ( 16 and Figure S1).In this study, we sought to determine if V7's roles were masked by the compensatory activity of another VAMP in platelets.Ex vivo granule secretion assays show that the deletion   of V3 alone or in concert with V2 in platelets does not influence granule secretion. 6This double deletion of V2 and V3, however, exacerbated the secretory defect of V7-deficient platelets, but only for α granules and lysosomes, not for dense granules.This was unlike what was seen when V8 was absent, and both V2 and V3 were deleted (i.e., V2 ∆ 3 ∆ 8 −/− platelets).In those platelets, dense granule release was also drastically affected. 6The dense granules harbor secondary agonists important for amplification of platelet activation and recruitment, including calcium, ADP, and polyphosphates.Defective dense granule biogenesis (as observed in Hermansky Pudlak Syndrome and Chediak Higashi Syndrome) and/or secretion, therefore leads to hemostatic impairment.Based on our previous report and ongoing studies, we believe that as long as V8 is present, there is no dense granule secretion defect.Our data suggest that V7 plays a role in lysosome release and that both V7 and V3 contribute, in some way, to α granule release.Given that there are only minimal effects when also deleting V2, one must conclude that V7 and V3 have some overlapping compensatory role in release from these two compartments for which V2 cannot substitute.Using three in vivo models, we show that the defects in release caused by deleting these v-SNAREs are insufficient to cause a thrombosis or hemostasis defect in arterial or venous models.
Although the platelet number was increased in V2 ∆ 3 ∆ 7 −/− animals and mean platelet volume was increased in V2 ∆ 3 ∆ 7 −/− and V3 −/− 7 −/− platelets, there were no significant changes in serotonin uptake, PF4, and β-hexosaminidase levels (Figure S2).Preliminary RNAseq and proteomic analysis do not show any major differences in cargo levels (Joshi et al., in preparation).Based on these observations, we find that the granule secretion deficiency in these strains is mainly, if not exclusively, influenced by the loss of VAMP isoforms.Future morphology studies of these platelets to evaluate granule numbers, size, and distribution will be needed to fully understand if granule biogenesis is affected in these strains.
Compensating roles for VAMPs in granule secretion have been studied in other systems.Zhao et al. showed that V2, V3, and V8 have a major role in GLUT4 trafficking in adipocytes, but V7 does not. 26In chromaffin cells, V2 is the dominant isoform for catecholamine secretion and in its absence, V3 can efficiently compensate. 27In mast cells, the cargoes are packed into distinct subsets of secretory granules, and the release of at least some of these granules is mediated by V8, while V2 and V3 appear not to be important. 28The release of granule cargo from neutrophils is mediated by the SNARE complexes consisting of different VAMPs. 29While V2 mediated release of cargo from tertiary granules, V7 played a role in azurophilic granule release.V1 was found to be crucial for both of these release events.These studies underline the heterogeneous nature of VAMP usage in the secretory granule release from different cell types.Whether this is due to the promiscuous pairing of SNAREs to mediate membrane fusion or represents a pathway to fine-tune release kinetics in a context-specific manner remains to be determined.
V7 has been assigned various roles (e.g., neurite outgrowth 11 ).It colocalizes with the lysosomal marker CD63 in mouse fibroblast cells 30 and autophagosomal marker LC3 in HeLa cells. 31In rat liver cells, V7 is responsible for heterotypic fusion of lysosomes with late endosomes. 32Fader et al., showed how V3 and V7 are required for different trafficking steps in autophagic pathway; however, they did not examine V8's role. 31Multiple reports suggest that V7 plays a role in lysosomal secretion from fibroblasts 12,33,34 and ATP secretion from astrocytes. 35Our data further support this connection between V7 and lysosomal release 31 and are consistent with α granules being endolysosomal in nature.
The interactions between V3 and V7 must differ from what has been noted for V3 and V8.Specifically, the global deletion of both V3 and V8 together (V3 −/− 8 −/− ) is synthetically lethal, suggesting that these two VAMPs support redundant parallel pathways that are critical for viability (Whiteheart, unpublished).The V3 −/− 7 −/− and V7 −/− 8 −/− , strains are viable, suggesting that other VAMPs (e.g., V3 or V8) can compensate for their combined loss (this manuscript and Joshi et al. in preparation).V3 and V8 have been found in endosomal compartments and perhaps play crucial roles in endosomal trafficking. 25,36,37V3 and V8 have also been reported to be associated with Weibel-Palade bodies in endothelial cells.While V3 was important for the release of von Willebrand Factor, V8 was not. 38Interestingly, a report by the Lowenstein lab suggests V8's role in secretion from endothelial cells. 39In human platelets, Peters et al. showed that when spreading over fibrinogen, V3 and V8 colocalize into the central granulomere, while V7 localizes to the platelet periphery. 17Loss of V3 negatively affected fibrinogen uptake and endocytosis of virions but had no effect on fluid-phase endocytosis of dextran by platelets. 9,40Single deletion of V8 or V7 has no effect on endocytosis, but these deletions affect platelet secretion ( 7 and present manuscript).The three major platelet v-SNAREs, V3, V7, and V8 May 2001 have overlapping subcellular localizations/ functions, but their ability to compensate for each other may be contextually limited.Such limitations could be controlled by their expression levels or restricted by their ability to interact with specific t-SNARE isoforms and SNARE regulators.More analysis of the strains reported here will be needed to identify the mechanisms by which v-SNAREs are used and the ramifications of that usage in different cell types.
Our data show a role for V7 in lysosome release and suggest overlapping roles for V3 and V7 in lysosome and α granule cargo release.In none of the strains created was there a substantial defect in dense granule release and, consistently, there was no gross defect in thrombosis and hemostasis noted in any of the three injury models tested.These data support the dominant role for V8 in platelet function but do underline the compensatory nature of VAMP usage in platelets.Whether these overlapping VAMP functions are simply due to promiscuity of SNARE pairing, which has been noted in in vitro SNARE complex assembly assays, 41 or represent a pathway for fine-tuning platelet exocytosis for specific outcomes, remains to be addressed.The animals described in this report will be valuable in making these distinctions and better understanding the in vivo roles of platelet secretion.

Figure 2 .
Figure 2. Loss of V3, and V7 together affects the kinetics and extent of α granule and lysosomal secretion.[ 3 H]-5-[HT] (serotonin)labeled platelets from wild type (WT) and V3 −/− 7 −/− mice were prepared as described in methods.The release of [ 3 H]-5-[HT] from dense granules (a, d), PF4 from α granules (b, e), and β-hexosaminidase from lysosomes (c, f) was measured, and percentage secretion was calculated.(a-c) for the thrombin doseresponse experiment, platelets were stimulated for 2 min with the indicated concentrations of thrombin.(d-f) for the time-course experiments, platelets were stimulated with 0.05 U/ mL thrombin for the indicated times.Data are mean ± standard error of the mean of triplicate measurements and is representative of ≥ 3 independent experiments.Statistical significance was calculated using a two-way ANOVA.ns not significant, *p ≤ .05,**p ≤ .01,***p ≤ .001,**** p ≤ .0001.

Table I .
Summary of murine strains.The nature of murine strains (platelet-specific or global) and the references that characterized them.: https://doi.org/10.1080/09537104.2023.2237114 DOI