Mechanisms of αIIbβ3 Biogenesis in the Megakaryocyte: A Proteomics Approach

Platelets play a central role in hemostasis and thrombosis, initiating clot formation in response to vessel wall damage. Platelet aggregates are formed at sites of injury by the binding and crosslinking of the integrin IIb3 to fibrinogen, von Willebrand factor and other soluble ligands. Platelets also form pathological thrombi, and the resulting arterial occlusion can lead to myocardial infarction or stroke. Inhibition of IIb3 binding can decrease the formation of pathologic thrombi, and IIb3 has become an important pharmacological target. Anti-IIb3 therapies have been highly successful in preventing death following myocardial infarction and percutaneous arterial stent placement.(Topol, Lincoff et al. 2002; De Luca, Ucci et al. 2009) However, attempts to design novel, orally available anti-IIb3 agents have been hampered by what appears to be paradoxical activation of IIb3 by the drug, in some cases leading to an increased risk of mortality in patients who received the drug.(Quinn, Plow et al. 2002) A novel approach to manipulatingIIb3 would be to perturb its post-translational processing and trafficking within the megakaryocyte, prior to platelet formation. Like most membrane proteins, IIb3 is formed by concerted processes of protein sorting and trafficking. Some of the mechanisms underlying IIb3 biogenesis and expression in megakaryocytes have been described, such as the calnexin cycle of protein quality control.(King and Reed 2002; Tiwari, Italiano et al. 2003; El.Golli, Issertial et al. 2005; Lo, Li et al. 2005) Clues to the stringent protein quality control of IIb3 biogenesis come from the study of patients with defective or absent IIb3, who manifest the mucocutaneous bleeding disorder Glanzmann thrombasthenia. A subset of patients with mutations in the IIb gene produce full-length IIb that retains the ability to form a complex with 3 but is retained within the cell and degraded, resulting in disease. These patients demonstrate the existence of stringent quality control mechanisms acting post-translationally to control IIb3 biogenesis and expression. A clearer understanding of these mechanisms may lead to new possibilities of anti-integrin therapy. Toward the goal of identifying proteins involved in IIb3 biogenesis we performed a proteomic analysis of proteins interacting with IIb in megakaryocytes cultured from human umbilical cord blood (UCB) and in HEK293 cells expressing IIb and 3.


Introduction
Platelets play a central role in hemostasis and thrombosis, initiating clot formation in response to vessel wall damage.Platelet aggregates are formed at sites of injury by the binding and crosslinking of the integrin IIb3 to fibrinogen, von Willebrand factor and other soluble ligands.Platelets also form pathological thrombi, and the resulting arterial occlusion can lead to myocardial infarction or stroke.Inhibition of IIb3 binding can decrease the formation of pathologic thrombi, and IIb3 has become an important pharmacological target.Anti-IIb3 therapies have been highly successful in preventing death following myocardial infarction and percutaneous arterial stent placement.(Topol, Lincoff et al. 2002;De Luca, Ucci et al. 2009) However, attempts to design novel, orally available anti-IIb3 agents have been hampered by what appears to be paradoxical activation of IIb3 by the drug, in some cases leading to an increased risk of mortality in patients who received the drug.(Quinn, Plow et al. 2002) A novel approach to manipulatingIIb3 would be to perturb its post-translational processing and trafficking within the megakaryocyte, prior to platelet formation.Like most membrane proteins, IIb3 is formed by concerted processes of protein sorting and trafficking.Some of the mechanisms underlying IIb3 biogenesis and expression in megakaryocytes have been described, such as the calnexin cycle of protein quality control.(King and Reed 2002;Tiwari, Italiano et al. 2003;El.Golli, Issertial et al. 2005;Lo, Li et al. 2005) Clues to the stringent protein quality control of IIb3 biogenesis come from the study of patients with defective or absent IIb3, who manifest the mucocutaneous bleeding disorder Glanzmann thrombasthenia.A subset of patients with mutations in the IIb gene produce full-length IIb that retains the ability to form a complex with 3 but is retained within the cell and degraded, resulting in disease.These patients demonstrate the existence of stringent quality control mechanisms acting post-translationally to control IIb3 biogenesis and expression.A clearer understanding of these mechanisms may lead to new possibilities of anti-integrin therapy.
Toward the goal of identifying proteins involved in IIb3 biogenesis we performed a proteomic analysis of proteins interacting with IIb in megakaryocytes cultured from human umbilical cord blood (UCB) and in HEK293 cells expressing IIb and 3.
Megakaryocyte proteins were captured by poly-histidine tagged IIb, or by photoreactive crosslinking followed by immunoprecipitation with anti-IIb mAbs, and analyzed by mass spectrometry.
The IIb and b3 subunits are synthesized as a single-chain precursors in the chaperone-rich folding environment of the endoplasmic reticulum (ER) (Fig 1).The precursor IIb, pro-IIb, is glycosylated under control of the calnexin cycle of protein folding.The 3 precursor is also glycosylated but does not appear to interact with the calnexin cycle.The two precursors heterodimerize to briefly form pro-IIb3, and then pro-IIb is cleaved by one or more members of the furin family of proteases in the Golgi.Cleavage of pro-IIb to mature IIb marks the exit of IIb from the ER, and this cleavage occurs only when pro-IIb is in complex with 3.(Bray, Rosa et al. 1986;Rosa and McEver 1989) Both IIb and 3 are synthesized in excess of what will finally be processed to mature IIb3.The calnexin cycle exerts stringent quality control over IIb production and up to one half of all pro-IIb is targeted to the proteasome for degradation in megakaryocytes.Excess 3 is degraded by a non-proteasomal mechanism.Inhibition of the proteasome in megakaryocytes resulted in a build-up of pro-IIb that was not being degraded, but had no apparent effect on the level of mature IIb in the IIb3 complex.Thus, the expression level of mature IIb3 is not simply a result of the stoichiometry of production vs. degradation.Rather there appears to be a concerted mechanism that controls how much pro-IIb3 will be converted into mature IIb3, and this mechanism is not grossly responsive to excessive levels of pro-IIb or 3.

Fig. 1. Biogenesis of IIb3
We hypothesized that whatever proteins underlie this mechanism must be interacting with pro-IIb and/or pro-IIb3 complex, but not with the mature IIb3 complex.We chose a proteomics approach to identify these interacting proteins.In order to enrich our assay for proteins that preferentially bound pro-IIb or pro-IIb3, we used an IIb subunit harboring R858G and R859G mutations that eliminates furin cleavage and traps IIb in its pro-IIb form.(Kolodziej, Vilaire et al. 1991) This IIbR858G/R859G construct can form a complex with 3, but only in small amounts compared to normal IIb3.In addition, while some of the mutant IIb3 complexes reach the cell surface and can mediate adherence to immobilized fibrinogen, the proportion of IIbR858G/R859G reaching the surface is very small compared to normal IIb.(Kolodziej, Vilaire et al. 1991) Thus, IIbR858G/R859G is a "nearly normal" mutant IIb subunit that is primarily retained within the cell, making it a useful bait to capture the proteins involved in that process.
The hsp40 type chaperone protein, DNAJC10, was captured by both the normal IIb and IIbR858G/R859G subunits, and was evaluated as a putative IIb interacting protein..We report that DNAJC10 interacted with pro-IIb and 3 in megakaryocytes, and appears to promote the degradation of pro-IIb.Notably, while IIb3-DNAJC10 interaction was evident in megakaryocytes, specific interaction was not detectable in HEK293 cells transfected with IIb cDNA, suggesting megakaryocyte specificity.Knockdown of DNAJC10 by siRNA increased IIb3 surface expression in UCB derived megakaryocytes, indicating that DNAJC10 negatively regulates IIb3 surface expression.Thus, DNAJC10-IIb interaction represents a novel post-translational mechanism regulating IIb3 surface expression.

Human Umbilical Cord Blood culture
This study used human UCB from the US National Cord Blood Bank that was deemed not suitable for clinical use due to low white blood cell number.Mothers who donate their UCB to the National Cord Blood Bank sign a consent giving permission to donate the UCB to research if it is inadequate for clinical use and will be discarded.Because the donated units are completely de-identified and are not collected prospectively specifically for research, the New York Blood Center Institutional Review Board (IRB), which oversees research ethics at the National Cord Blood Bank, considers their use exempt from IRB review.Human UCB was prepared as previously described.(Mitchell, Li et al. 2006) Briefly, leukocytes were separated from 1-3 units of human UCB judged to be inadequate for clinical purposes (generously provided by the New York Blood Center) by Dextran 70 sedimentation (Amersham Biosciences, Piscataway, NJ) for 1h, and then enriched for CD34+ progenitor cells by negative selection using a combination of antibodies against maturation/lineagespecific markers (RosetteSep, StemCell Technologies, Vancouver, BC) concomitant with density sedimentation using Ficoll-Paque Plus (Amersham Biosciences, Piscataway, NJ).These cells were cultured in serum-free medium (StemCell Technologies, Vancouver, BC) with 50 ng/ml thrombopoietin (TPO) plus 1 ng/ml SCF (both Millipore, Temecula,CA) for 3 days in 10 cm dishes in a 37°C incubator.At this point a portion of the cells will have died.The remaining living population of larger cells was gently washed and replated in fresh media with the same cytokines in 10 cm dishes and left until use (Day 8 or 9).Fresh media was added on day 6.We have previously reported that under these conditions the UCB differentiate into a population of large cells with > 90% expressing IIb3, > 80% expressing GPIb, about 50% expressing 21.(Mitchell, Li et al. 2007) For experimental use the cells were gently harvested, pelleted at 300 rpm for 5 min, or let settle by gravity, and gently resuspended in the appropriate buffer.

HEK293 cell culture
HEK293 cell lines (American Type Culture Collection (ATCC), Manassas, VA) that stably expressed human IIb3 receptors were established as previously described.(Mitchell, Li et al. 2006) Transfections were performed using Lipofectamine 2000 (Gibco-BRL, Carlsbad, CA) according to the manufacturer's instructions, followed by selection in media containing 500 g/ml G418 for 2-4 weeks.To obtain a population of cells uniformly expressing high levels of IIb3, cells were labeled with the mAb 10E5 (anti-IIb3) and sorted using a MoFlo cell sorter (Beckman Coulter, Fullerton, CA).

Immunoprecipitation and biosynthetic labeling
Samples were prepared as previously described and all steps were performed on ice unless otherwise stated.(Mitchell, Li et al. 2006) Briefly, cells (either day 8 megakaryocytes derived from UCB cells or HEK293 cells) were lysed in 1% Brij 98 lysis buffer containing protease inhibitors and 20uM N-methylmaleimide (NEM).Lysates were precleared with protein-G Sepharose beads (Amersham Biosciences, Piscataway, NJ), and then equivalent amounts of protein, usually 400 g, were incubated one h at 4C with one or more of the antibodies listed above (4 g/reaction).Samples were incubated with protein-G Sepharose beads for one h at 4C, washed twice, and incubated with SDS sample buffer for 10 min at 100°C.Reduced samples contained 10% beta mercaptoethanol (Sigma, Thermo Scientific, Rockford, IL).Samples were subjected to SDS-PAGE on 7% gels, and the gels were either stained for mass spectrometry identification (described below) or transferred to PVDF membranes for immunoblotting.Non-specific binding was determined by performing parallel immunoprecipitation with mouse or rabbit IgG in each experiment.In preliminary experiments, the production of IIbR858G/R859G and 3 was confirmed in the transfected HEK293 cell line by immunoprecipitation with both anti-IIb and anti-3 mAbs followed by immunoblot.For biosynthetic labeling, cells were incubated for 30 min at 37°C in methionine/cysteine-free medium, followed by pulse-labeling for 15 min at 37ºC in medium containing 35 S-methionine/cysteine (300 Ci/10 cm plate).The pulse was terminated by incubation in medium containing unlabeled methionine/cysteine (1 mg/ml each) and the cells were incubated at 37 °C until lysis in 1% Triton-X 100 lysis buffer.Following cell lysis, supernatants were prepared as above.Gels were dried and exposed to film.For inhibition of the proteasome cells were incubated in the proteasome inhibitor MG132 (10 M)(Sigma_Aldrich, St Louis, MO) in normal growth medium at 37C and then immediately lysed.

Histidine-tag/Nickel bead pulldown assay
HEK293 cells expressing poly-histidine tagged IIb cDNA (in vector pEF1/V5-His) and 3 cDNA (in vector pcDNA3.1)(Mitchell, Li et al. 2007) were lysed in 1% Triton, 150 mM NaCl, 10 mM imidazole buffer (lysis buffer) on ice for 30 m, centrifuged for 30m at 4C, and then the supernatant was reacted with 50 l of a 6:4 slurry of Ni beads:imidazole buffer (Qiagen, Inc.Valencia, CA) to bind the histidine-tagged subunits to the nickel beads.After incubating for 30 m the beads were washed four times with 10 mM imidazole lysis buffer by using a magnetic chamber to isolate the beads.Next, fresh whole cell lysates (1 mL) of UCB-derived megakaryocytes in 1% Triton, 150 mM NaCl, 10 mM imidazole buffer were incubated with the washed, Ni-bound IIb for 1 h, and then the beads were washed twice with 1 mL lysis buffer containing 20 mM imidazole.Ni-bound proteins were eluted with 250 mM imidazole and the entire eluate was subjected to SDS-PAGE on a 7% gel followed by staining with Imperial Stain (Pierce, Thermo Scientific, Rockford, IL).Control experiments were run in parallel in which no megakaryocyte lysate was added to the beads, but all other steps remained the same.These controls aimed to identify the remaining HEK293 cell proteins still bound to either the beads or to IIb after the washing steps, before incubation with megakaryocyte lysate.Experimental and control gels were run simultaneously, and the lanes were excised and analyzed by mass spectrometry.

Photocrosslinking amino acids
UCB cells were cultured as described above for eight days, then were washed and incubated for 24 hours in leucine-and methionine-free medium containing photoreactive methionine and leucine, dialyzed FBS (both Pierce, Thermo Scientific, Rockford, IL), and 50 ng/ml TPO (Millipore, Temecula,CA).The cells were exposed to 345 nm UV light for 15 minutes to crosslink the photoreactive amino acids, and harvested immediately, according to manufacturer's instructions.Whole cell lysates were immunoprecipitated with antibodies specific for IIb or 3, and the proteins were separated by SDS-PAGE using the same protocol as for the histidine-tag affinity capture.Controls were simultaneously immunoprecipitated with non-immune IgG.Experimental and control lanes were excised and analyzed by mass spectrometry.

RNAi
HEK293 cells stably expressing high levels of IIb and 3 were transfected with 100 nM siRNA duplexes (Dharmacon, Thermo Scientific, Rockford, IL and Qiagen, Valencia, CA) using Dharmafect-1 reagent (Dharmacon, Thermo Scientific, Rockford, IL), then analyzed by flow cytometry using a FacsCanto (Becton Dickenson, ) at 48-96 h after transfection.Cultured UCB cells were transfected twice, on culture days 3 and 5, with 100 nM of siRNA duplexes or controls also using Dharmafect-1 reagent, and then analyzed 48 -72 h later by flow cytometry.In order to identify individual transfected cells by FACS, cells were cotransfected with 10 nM of fluorescent-labeled non-coding siRNA (Qiagen, Valencia, CA).In some experiments, Cy3-labeled siRNA duplexes were used (Dharmacon, Thermo Scientific, Rockford, IL).Controls were: no treatment, transfection reagent but no siRNA duplex, Cy3labeled negative control siRNA duplex only (Qiagen), siControl non targeting siRNA duplex, and siRNA duplex against cyclophillin B (positive control) (both Dharmacon).

Quantitative RT-PCR
For analysis of RNA content, cells were collected in RNAlater (Applied Biosystems, Foster City, CA) and RNA was purified using the RNEasy mini kit (Qiagen, Valencia, CA).For some analyses, siRNA transfected cells were sorted for expression of the fluorescent marker using a MoFlo sorter, and the sorted cells were subjected to QRT-PCR.Analysis was performed on an ABI 7900 or an ABI 7700 thermocycler/ fluorescence analyzer (Applied Biosystems, Foster City, CA) using the SYBR green probe (Qiagen or Invitrogen) and Quantitect primer assays (Qiagen).Relative mRNA levels were calculated using the Ct method that corrects for GAPDH expression in all samples.Fold-reduction was corrected for percent transfection in each experiment.

Immunofluorescence analysis
Mks were cytospun onto poly L-lysine-coated glass coverslips and fixed in methanol for 5 minutes at RT. Cells were then washed with saponin buffer (PBS containing 0.02% BSA, 0.005% saponin), then blocked with saponin buffer for 1 hour at RT. Cells were then incubated with the primary antibody diluted in saponin buffer for 1 hour at RT.After washing with PBS, cells were incubated with the appropriate conjugated secondary antibody in PBS for 1 hour at RT. Specimens were mounted with Prolong gold (Invitrogen).Images were acquired using a Zeiss LSM 510 META (Axiovert 200M Inverted Microscope Stand) confocal laser scanning microsope through a 100x objective.Each set of staining conditions was repeated 3 times.Background fluorescence was measured by incubating fixed cells with the secondary antibody only and then acquiring images with the exact settings used to obtain the experimental images.

Mass spectrometry data analysis
The protein gel bands were excised from the SDS-PAGE.The gel bands were reduced with 10 mM of DTT and alkylated with 55 mM iodoacetamide, and then digested with Sequence Grade Modified Trypsin (Promega, Madison, Wisconsin) in ammonium bicarbonate buffer at 37 overnight.(Kumarathasan, Mohottalage et al. 2005) The digestion products were extracted twice with 0.1% trifluoroacetic acid and 50% Acetonitrile and 1.0% trifluoroacetic acid respectively.The extracted mixture was dried by Speed-Vac and redisolved in 10 L 0.1% trifluoroacetic acid.Half of the extracts were injected by LC-MS/MS analysis.For LC-MS/MS analysis, each digestion product was separated by a 60 min gradient elution with the Dionex U3000 capillary/nano-HPLC system (Dionex, Sunnyvale, California) at a flow rate of 0.250 L/min that is directly interfaced with the Thermo-Fisher LTQ-Obritrap mass spectrometer (Thermo Fisher, San Jose, California) operated in data-dependent scan mode.The analytical column was a home-made fused silica capillary column (75 m ID, 100 mm length; Upchurch, Oak Harbor, Washington) packed with C-18 resin (300 A, 5 m, Varian, Palo Alto, California).Mobile phase A consisted of 0.1% formic acid, and mobile phase B consisted of 100% acetonitrile and 0.1% formic acid.The 60 min gradients with 250 nL/min flow rate for B solvent went from 0 to 55% in 34 minutes and then in 4 min to 80%.The B solvent stayed at 80% for another 8 min and then decreased to 5% in 8 min.Another 6 min was used for equilibration, loading and washing.The mass acquisition method was one FT-MS scan followed by 6 subsequent MS/MS scan in the ion trap.The FT-MS scan was acquired at resolution 30,000 in the Orbi-trap.The six most intense peaks from the FT full scan were selected in the ion trap for MS/MS.The selected ions were excluded for further selection for 180 seconds.The following search parameters were used in all MASCOT searches: maximum of 1 missed trypsin cleavages, cysteine carbamidomethylation, methionine oxidation.The maximum error tolerance for MS scans was 10 ppm for MS and 1.0 Da for MS/MS respectively.Proteins were designated as "hits" if they matched at least 2 distinctive peptides with a MASCOT score of at least 40.For proteins matching the same sets of peptides, only the protein with the greater percentage of coverage was selected.In the one case where 2 isoforms could not be distinguished (HSP70A/B), both proteins are reported.Proteins identified in the control gels were considered "negative," and these proteins were removed from further analysis.The interaction data were further analyzed using the Cytoscape (Cline, Smoot et al. 2007) software and publically available protein interaction databases (e.g.INTACT, NCBI, UniProt) to generate a network of first-degree interactions with IIb.The network was further expanded to include proteins that were previously reported to interact with IIb and known interactions between any of the proteins identified in our primary assays.

Statistical analysis
Flow cytometry data was summarized as the geographic mean fluorescence intensity (MFI) of antibody binding, and normalized to the MFI of control siRNA treated cells, so that the experimental MFI is expressed as a percentage of the control MFI.The overall percent change of the replicates was expressed as the average percent change +/-the confidence interval.The two-sided, paired t-test was used to determine whether there were differences in antibody binding between the different experimental groups.In the siRNA experiments, relative mRNA levels were calculated using the Ct method that corrects for GAPDH expression.

The IIb interactome
Proteins putatively interacting with IIb were isolated from UCB derived megakaryocytes and from HEK293 cells expressing IIb3, and these proteins were analyzed by mass spectrometry.Samples were processed by the two methods depicted in Figure 2. In the first method, recombinant IIb or IIbR858G/R859G subunits bearing a polyhistidine tag were expressed in HEK293 cells along with normal 3.The cells were lysed with 1% Triton buffer, and the lysates were reacted with Ni beads that bind polyhistidine.The Ni beads were washed using a magnetic separator to reduce non-specific binding to the Ni-bound IIb, and then incubated with fresh whole cell lysate of umbilical cord blood derived megakaryocyte.Since we are interested in the early process of IIb3 formation and intracellular trafficking we used megakaryocytes at day 8 or 9 of culture; at that point in our system IIb3 is highly expressed but there is not yet any proplatelet formation.The Nibound histidine-tagged IIb and IIb-bound proteins were washed twice and then were eluted from the Ni beads with 250 mM imidazole in a 1% Triton buffer.Simultaneous control experiments were performed without megakaryocyte lysate in order to identify proteins non-specifically binding to either the Ni beads or the polyhistidine tag.The proteins thus collected were separated by one dimensional SDS-PAGE on a 7% gel.After staining, the experimental and control lanes were cut out and analyzed by mass spectrometry.This entire process was repeated in three separate experiments with normal IIb, and in four separate experiments with IIbR858G/R859G.Fig. 2. Methods used to isolate and identify IIb-interacting proteins.A): Polyhistidinetagged IIb subunits were captured on Ni beads and then incubated with whole cell lysate of day 8 UCB-derived megakaryocytes.Protein complexes thus captured were processed as described in the text, and then the proteins were identified by FT-MS and MS/MS.B) Crosslinking experiments were performed on day 8 UCB-derived megakaryocytes and transfected HEK293 cells.Cells were fed photoactivatable methionine and leucine, exposed to UV light and lysed immediately.Samples were processed for mass spectrometry as described.
The second strategy used photo-activated amino acids to crosslink IIb to its binding partners (Figure 2B).UCB-derived megakaryocytes were starved for methionine and leucine, then fed photo-activatable methionine and leucine overnight, and then exposed to UV light to crosslink the amino acids of neighboring proteins.The megakaryocytes were immediately lysed and the IIb-bound proteins were immunoprecipitated by anti-IIb mAb.These proteins were separated by one dimensional SDS-PAGE and then the appropriate bands were cut out and analyzed by mass spectrometry.Because we were looking only for proteins crosslinked to IIb, only the portions of the bands with M r > 120 (M r of the mature IIb subunit) were analyzed.Five separate crosslinking experiments were performed with UCB-derived megakaryocytes and two on HEK293 cells expressing IIbR858G/R859G and normal 3.
Minimum criteria for protein inclusion in data analysis were at least 2 distinctive peptides with a MASCOT score of at least 40, and absence of the protein in the control lanes.I m p o r t a n t l y , s i n c e w e w e r e i n t e r e s t e d i n p r o t e i n s t h a t a r e m o r e a b u n d a n t i n t h e megakaryocyte than in HEK293 cells, even proteins known to interact with IIb3, such as talin, were removed from the final results because they were present in the control lanes.Proteins identified in the Ni bead extraction experiments that are known to harbor natural polyhistidine sequences (such as the DEAH boxes) that could independently bind to the Ni beads were also excluded.Combining the results of both methods, 98 proteins were identified as potentially interacting with the normal IIb subunit, and 79 proteins were identified as putatively interacting with the IIbR858G/R859G subunit (Table 1).These 163 proteins putatively constitute a portion of the IIb interactome, a network of protein-protein interactions relevant to the trafficking and function of IIb in megakaryocytes (Figure 3).

Gene Ontology analysis
Gene Ontology (GO) analysis using the DAVID Bioinformatics Resources (Dennis, Sherman et al. 2003;Huang, Sherman et al. 2008) revealed enrichment for protein functions related to processing and trafficking, and ER, Golgi or vesicle components (Figure 3).Comparison of the proteins isolated with the normal IIb subunits vs with the IIbR858G/R859G subunits showed similar distributions of their localization and molecular functions (Table 2).
Notably, the IIbR858G/R859G subunits captured a greater percentage of these types of proteins, suggesting that a greater percentage of the IIbR858G/R859G subunits were associated with the ER, Golgi and vesicles.Comparison of our data set to a recently published analysis of gene expression in megakaryocytes (Watkins, Gusnanto et al. 2009) showed that approximately 70% of the proteins identified herein as potentially interacting with IIb were identified as expressed in the megakaryocyte transcriptome.
GO function/location % of total proteins IIb IIbR858G/R859G Despite the similarity in distributions into GO categories pertaining to protein processing and localization, there were only 14 proteins in common between the normal IIb and IIbR858G/R859G captured proteins (Table 1).This may indicate a difference in the relative amount of time spent by the normal and mutant subunits in the protein processing environment.One protein, DNAJC10, was captured in all experiments by both IIb and IIbR858G/R859G.Like other DNAJ proteins, DNAJC10 contains a binding domain for the chaperone BiP, but is unique in that its second domain has four protein-disulfide-isomerase (PDI) consensus sequences.(Cunnea, Miranda-Vizuete et al. 2003;Dong, Bridges et al. 2008) Because disulfide bond rearrangement is important for both biogenesis and function of IIb3, we investigated the putative interaction between IIb and DNAJC10.To our knowledge, DNAJC10 has not previously been reported to interact with IIb or 3.

DNAJC10 interacts with IIb3 during biogenesis
DNAJC10 coimmunoprecipitated with IIb and 3 from UCB derived megakaryocytes using anti-IIb mAbs 10E5, B1B5, M148 and CA3, and the anti 3 mAb 7H2.DNAJC10 either interacted with IIb and 3 directly or was part of a complex with them (Figure 3).Because DNAJC10 was captured by the IIbR858G/R859G subunit, we explored the possibility that DNAJC10 interacted with IIb prior to IIb3 complex formation and IIb cleavage to its mature form.We have previously shown that pro-IIb is degraded by the proteasome and that proteasome inhibition leads to an increase of pro-IIb within the cells.(Mitchell, Li et al. 2006) The undegraded pro-IIb is trapped in a "pre-degradation" state in which normally transient protein interactions, which usually lead to degradation, may become more long-lived and may be captured by co-immunoprecipitation.Incubation of megakaryocytes with MG132 resulted in increased DNAJC10 co-immunoprecipitation with IIb by all mAbs, but particularly M148 and CA3.Thus, DNAJC10 may associate with IIb early in biogenesis, before excess pro-IIb is targeted to the proteasome.To test this possibility we used a panel of conformation-specific mAbs that we have previously used to track the conformational changes of IIb as it proceeds through biogenesis (Figure 3).(Mitchell, Li et al. 2007) Specifically: mAb10E5 binds to the IIb head region and recognizes both the pro-IIb3 and mature IIb3 complex; mAb 7E3 binds to the 3 head region and also recognizes both the pro-and mature complex; mAb B1B5 binds to the IIb tail and preferentially recognizes pro-IIb and pro-IIb3 complex, and mAb M148 preferentially recognizes mature IIb3 but its epitope is not known.(Mitchell, Li et al. 2007) To determine whether DNAJC10 interacts with IIb during a specific stage of biogenesis, IIb was immunoprecipitated from megakaryocytes in the presence of MG132 using this panel of conformation-specific mAbs.DNAJC10 was most strongly coimmunoprecipitated by mAb B1B5, suggesting that DNAJC10 preferentially interacts with pro-IIb or the pro-IIb3 complex.In contrast, DNAJC10 coimmunoprecipitated less well with mAb M148, suggesting less interaction with mature IIb.DNAJC10 also interacted with the precursor and/or mature IIb3 complex, since it coimmunoprecipitated with the complex-specific mAb10E5.Since the ratio of pro-IIb3 to mature IIb3 is small in megakaryocytes, (Mitchell, Li et al. 2007) this binding pattern is consistent with DNAJC10 having bound preferentially to the small amount of pro-IIb3 present in the cells.Together these findings suggest that DNAJC10 preferentially interacted with pro-IIb and the pro-IIb3 complex, rather than the mature complex.

DNAJC10 depletion increases surface expression of IIb3
To determine whether the IIb-DNAJC10 interaction had physiological relevance, we assessed its impact on the surface expression of IIb3 in megakaryocytes.siRNA mediated knockdown of DNAJC10 was performed on both human megakaryocytes derived from UCB and on HEK293 cells expressing normal IIb and 3.At least an 80% decrease in RNA level was achieved (data not shown).Knockdown of DNAJC10 increased IIb3 surface expression on megakaryocytes by 25% +/-11% (n = 4, p = 0.02), and on HEK293 cells expressing IIb3 by 35% +/-12% (n=4, p = 0.01) (Figure 3D).

Intracellular localization of DNAJC10 in megakaryocytes
DNAJC10 was localized within megakaryocytes by immunofluorescence microscopy (Figure 5).The distribution of DNAJC10 was compared to that of IIb, 3, and the ER and Golgi compartments.Both IIb and 3 were distributed throughout the ER as well as on the cell surface.DNAJC10 had a diffuse punctate distribution in the periphery of the cell, away from the nucleus, and colocalized with only a portion of aIIb and 3, consistent with their presumably transient interaction.There was partial overlap between the ER marker calnexin and DNAJC10, suggesting that part of the DNAJC10 distribution is outside the ER or at least separate from the distribution of calnexin in the ER.This finding was surprising since DNAJC10 has the ER-retention signal KDEL.DNAJC10 did not localize with markers for the cis-or trans-Golgi compartments.

Discussion
We have used a proteomics approach to identify novel proteins interacting with IIb in megakaryocytes.Two different constructs of IIb, representing normal IIb and pro-IIb subunits, were used to capture proteins interacting with both nascent and mature subunits.
Megakaryocytes were derived from human UCB and used on day 8, which in our system yields high expression IIb3 on the surface but no proplatelet formation.(Mitchell, Li et al. 2006) In all, 163 proteins were identified as potentially interacting with IIb subunits; 98 were captured with normal IIb and 79 with aIIbR858/G859G, with 14 overlapping (Table 1).Day 8 megakaryocytes express very high amounts of IIb, most of which is mature IIb3, resulting in a relatively small proportion of nascent IIb.However, a large proportion of the mutant aIIbR858/G859G subunits are retained within the cell and degraded.Thus the difference between the two protein lists could partly be due to their differences in localization within the cell.This may be reflected in the larger proportion of ER and Golgi related proteins that were captured with the mutant subunit (Table 2).Only a few intracellular proteins have been reported to interact with IIb and IIb3, most notably talin, calnexin, and calreticulin (Intact and NCBI).Our two-step protein capture method was designed to isolate proteins with low affinity binding to IIb, such as calnexin, while screening out higher affinity binding proteins, such as talin and 3.In accord with this expectation, both talin and 3 were identified in both the control and experimental lanes, and so were excluded from the final interaction list, while calnexin and calreticulin were identified only in the experimental lanes.Surprisingly, while DNAJC10 readily coimmunoprecipitates with IIb3 from megakaryocytes, we have been unable to replicate this finding in HEK293 cells transfected with IIb and 3, despite an abundance of DNAJC10 in HEK293 cells (data not shown).While this is in no way conclusive, it is suggestive of cell-specific interaction of IIb3 and DNAJC10 in megakaryocytes.
Comparison of our experimental results with previously reported platelet proteomic data and IIb3 interaction data showed good correlation.About 70% of the proteins identified in our screens were reported as "present" in platelets in the Haem Atlas, a proteomic analysis of platelet protein content (Watkins, Gusnanto et al. 2009).
Two protein capturing strategies were used, each with strengths and weaknesses.The twocell pull-down assay using Ni beads to capture poly-histidine-tagged IIb, allowed the use of mutant cDNA constructs, such as IIbR858G/R859G, as bait.However, by introducing a protein synthesized in HEK293 cells as bait, there was the potential for false positive identification of proteins that associated with IIb in the HEK293 cells but not in megakaryocytes.Since the interactions of chaperone proteins are typically of low affinity, these proteins were most likely cleared by the washing steps and did not appear in the control lanes.Another source of false positives was non-specific binding to the Ni beads.Proteins with poly-His sequences (such as DEAH boxes) or naturally occurring Ni binding activity (such as keratin) could have bound to the beads, constituting false positives.However, virtually all of these potential false positives appeared in the control lanes as well and were excluded from analysis.The photoreactive crosslinking assay was intended to capture proteins in situ with IIb in megakaryocytes.While crosslinking experiments typically produce high numbers of false positives, in our experiments we identified low numbers of proteins from both the experimental and control lanes.The low yield may be due in part to the short crosslinking time used.We found that more than 15 minutes of UV exposure caused excessive protein degradation, while shorter exposure resulted in low crosslinking activity.
Of the proteins captured using both IIb and IIbR858G/R859G the Hsp40-type chaperone protein, DNAJC10 (ERdj5), was notable due to its disulfide isomerase activity, since both IIb and 3 require disulfide bond rearrangement for both biogenesis and function.(Shen, Meunier et al. 2002;Cunnea, Miranda-Vizuete et al. 2003;Dong, Bridges et al. 2008) Among its several functions, the ER chaperone protein BiP protects nearly-folded proteins against aggregation by binding to exposed hydrophobic patches.(Hendershot 2004) The Hsp40 chaperones bind to BiP and increase its efficiency of ATP hydrolysis, which allows BiP to release its substrate.DNAJC10 has been shown to be induced during ER stress, and may assist in delivering misfolded ER proteins to the proteasome for degradation.(Shen, Meunier et al. 2002;Cunnea, Miranda-Vizuete et al. 2003;Dong, Bridges et al. 2008) DNAJC10 coimmunoprecipitated with both IIb and 3 subunits in megakaryocytes, suggesting that it may bind the IIb3 complex.The immunoprecipitation pattern obtained using a panel of conformation-specific mAbs (Mitchell, Li et al. 2007) indicated that DNAJC10 preferentially interacted with pro-IIb or pro-IIb3 rather than mature IIb3.Together these findings suggest that DNAJC10 interacted with pro-IIb up to the point of complex formation, but not after pro-IIb cleavage (Figure 2).Thus DNAJC10 appears to be present and interacting with IIb at a critical decision point during IIb3 biogenesis, i.e. when pro-IIb will either form the mature IIb3 complex or be targeted to degradation.
Surprisingly, the distribution of DNAJC10, which has a KDEL ER-localization signal, was not confined to the ER, as judged by the distribution of calnexin.To determine if DNAJC10 was cycling to the Golgi and back, as many ER packaging proteins do, we looked for colocalization of DNAJC10 with cis and trans Golgi markers, and found none.The identity of the organelle(s) where DNAJC10 resides remains to be determined.
Depletion of DNAJC10 by siRNA resulted in an increase in surface expression of IIb3 on both human megakaryocytes and transfected HEK293 cells (Figure 2D) Since DNAJC10 depletion led to an increase in IIb3 surface expression, it appears to be a negative regulator of IIb3 surface expression.These findings make DNAJC10 an interesting and potentially targetable protein for perturbing IIb3 biogenesis.

Conclusion
While the details of DNAJC10-IIb interaction remain to be investigated, the current findings provide proof of principle that manipulation of early events in IIb biogenesis can result in altered expression levels of the mature IIb3 receptor, thereby setting a precedent for a novel approach to integrin-related therapy.These studies also support the validity of the data set, although other putative interactions must be explored for greater validation.We hope that the data set created will be a u s e f u l t o o l f o r s t u d y i n g i n t e g r i n a n d megakaryocyte biology.

www.intechopen.com
Mechanisms of IIb3 Biogenesis in the Megakaryocyte: A Proteomics Approach 193 By deciphering the IIb3 biogenesis pathway we hope to gain an inroad into controlling the level of IIb3 expression on platelets with the long-term goal of developing novel antithrombotic therapies.These types of therapy would not just inactivate the circulating platelets, but would modulate the megakaryocytes to make less adhesive platelets.One can imagine a scenario where patients at high risk of heart attack or stroke could be maintained on a drug that decreases their platelet IIb3 expression.Below a certain level of expression, platelet activation and aggregation would be diminished but not completely eliminated, resulting in an overall decrease in platelet thrombus formation but not complete loss of platelet function.This type of therapy could potentially have a greater safety profile than current therapies that summarily inactivate circulating platelets.
This study also assembles some of the wide range of research methods available to hematology research.No single technique could have discovered, validated and explored the function of DNAJC10 in megakaryocytes: rather, a broad range of methods was required.This wide variety is part of what makes research exciting and underscores the benefits of collaboration.

Fig. 3 .
Fig. 3. Network derived from Table 1 data.Proteins are represented as nodes (circles) labeled with their gene symbol, and putative protein interactions with IIb are indicated by thin green connecting lines.Thick blue lines and blues circles represent interactions or proteins, respectively, retrieved from online databases (Intact and NCBI) using the Cytoscape software.Arrow indicates DNAJC10. Figure prepared using Cytoscape.

Fig. 4 .
Fig. 4. Interaction of DNAJC10 with IIb3 during biogenesis.A) Schematic of changing conformations of IIb3 as it progresses through biogenesis.B) Changing specificity of 3 mAbs against IIb as it progresses through distinct conformations during biogenesis.B1B5 preferentially recognizes pro-IIb, while M148 shows preference for mature IIb, and 10E5 recognizes the heterodimer complex.C) Immunoprecipitation and Western blot of proteins isolated from megakaryocytes, as described in text.DNAJC10 coimmunoprecipitated preferentially with IIb that was pulled down by 10E5 and B1B5, but not M148, suggesting that DNAJC10 preferentially interacts with pro-IIb and pro-IIb3.The last lane is an IgG control.D) siRNA mediated knockdown of DNAJC10 increased surface expression of IIb3 as measured by flow cytometry using an Alexa647-labeled anti-IIb3 mAb (10E5).Expression was increased by 25 +\-11% (p=0.02,n=4) in UCB-derived megakaryocytes (UCB) and by 35 +\-12% (p=0.01,n=3) in HEK293 cells (HEK293) compared to control siRNA transfection (CTL).

Fig. 4 .
Fig. 4. Intracellular localization of DNAJC10.Cultured megakaryocytes were fixed and immunostained as described in the text.DNAJC10 (Blue) exhibited a punctate staining in the periphery of the cells, away from the nucleus.There was partial overlap of DNAJC10 with calnexin (Red) staining.IIb (Green) was distributed throughout the ER and cell surface.While IIb partially colocalized with the cis-and trans-Golgi, DNAJC10 did not.

Table 1 .
Proteins captured with IIb in megakaryocytes and HEK293 cells.Results are separated according to the IIb subunit they were captured with.Column labels: Gene www.intechopen.comMechanisms of IIb3 Biogenesis in the Megakaryocyte: A Proteomics Approach 187 Symbol, EntrezGene ID, Protein Accession Number and Description are from the NCBI database.SwissProt Accession Number is from the UniProt database.Number of Unique Peptides is cumulative from all experiments.Coverage percent is cumulative for all experiments.Tally is shown of the total number of experiments in which a protein was identified (No Expts), and whether captured by histidine tag/Ni affinity (Ni-His) or by crosslinking (X-link).Presence or absence in a recently published expression study (HaemAtlas) is indicated.* Proteins captured by both normal and mutant IIb subunits.

Table 2 .
Distribution of captured proteins into Gene Ontology categories.Some proteins appear in more than one category.