Dystrobrevin family members (α and β) are cytoplasmic components of the dystrophin-associated glycoprotein complex, a multimeric protein complex first isolated from skeletal muscle, which links the extracellular matrix to the actin cytoskeleton. Dystrobrevin shares high homology with the cysteine-rich and C-terminal domains of dystrophin and a common domain organization. The β-dystrobrevin isoform is restricted to nonmuscle tissues, serves as a scaffold for signaling complexes, and may participate in intracellular transport through its interaction with kinesin heavy chain. We have previously characterized the molecular determinants affecting the β-dystrobrevin–kinesin heavy chain interaction, among which is cAMP-dependent protein kinase [protein kinase A (PKA)] phosphorylation of β-dystrobrevin. In this study, we have identified β-dystrobrevin residues phosphorylated in vitro by PKA with pull-down assays, surface plasmon resonance measurements, and MS analysis. Among the identified phosphorylated residues, we demonstrated, by site-directed mutagenesis, that Thr11 is the regulatory site for the β-dystrobrevin–kinesin interaction. As dystrobrevin may function as a signaling scaffold for kinases/phosphatases, we also investigated whether β-dystrobrevin is phosphorylated in vitro by kinases other than PKA. Thr11 was phosphorylated by protein kinase C, suggesting that this represents a key residue modified by the activation of different signaling pathways.
A kinase anchoring protein
dystrophin-associated glycoprotein complex
immobilized metal affinity chromatography
linear trap quadrupole
protein kinase A (cAMP-dependent protein kinase)
protein kinase C
surface plasmon resonance
Dystrobrevins are a family of proteins related to and associated with dystrophin. Together with the transmembrane dystroglycan and sarcoglycans, and the cytoplasmic proteins dystrophin and syntrophin, dystrobrevin is a component of the macromolecular membrane complex known as the dystrophin-associated glycoprotein complex (DGC). The DGC was first isolated and characterized from skeletal muscle, where, by linking the extracellular matrix with the actin cytoskeleton, it stabilizes the sarcolemma during muscle contraction and relaxation cycles [1, 2]. Mutations of the coding genes for some of the DGC components result in several forms of muscular dystrophies .
DGCs have also been widely studied in nonmuscle tissues, such as brain, lung, and kidney [4, 5]. Various DGCs can originate from the assembly of dystrophin, dystrobrevin, and syntrophin-specific isoforms. Different cellular compartmentalization and/or cell type expression derive from this diverse assembly, and add functional complexity to the DGC. Besides its structural role, the DGC is now thought to function as a scaffold that concentrates and organizes interrelated signaling molecules at the plasma membrane .
Dystrobrevin shares sequence homology with the C-terminus of dystrophin, but does not bind dystroglycan. It contains two EF-hands, a zinc-binding domain (ZZ), syntrophin-binding sites, and two coiled-coil domains that mediate the association between dystrobrevin and dystrophin . Two separate genes, both of which yield multiple transcripts, code for two dystrobrevin isoforms, α and β [7, 8]. Whereas α-dystrobrevin is a ubiquitous isoform, β-dystrobrevin (β-DB) is expressed in nonmuscle tissues, including brain [4, 5].
Besides dystrophin and syntrophin, a number of dystrobrevin-interacting partners have so far been identified. These include: α-sarcoglycan ; intermediate filament proteins such as syncoilin  and desmuslin ; dysbindin , a subunit of biogenesis of lysosome-related organelles complex-1 and the product of a potential susceptibility gene for schizophrenia [13, 14]; DAMAGE, an α1-dystrobrevin-associated protein that is highly expressed in brain ; pancortin, an extracellular matrix protein that has been proposed to play a role in neuronal differentiation ; kinesin heavy chain Kif5 , a motor protein involved in axonal and dendritic transport of vesicles, organelles, and protein complexes ; the regulatory subunits RIα and RIIβ of cAMP-dependent kinase [protein kinase A (PKA)] , and α-catulin, a protein recruited to the α(1D)-adrenergic receptor signalosome by the C-terminal domain of α1-dystrobrevin . Moreover, we have recently demonstrated and characterized the interaction between β-DB and the high mobility group-domain proteins, iBRAF and BRAF35, suggesting an involvement of β-DB in the regulation of chromatin dynamics . Through exploration of the activity of these proteins, a multifaceted function of dystrobrevin in distinct cell compartments emerges. By these multiple interactions, dystrobrevin can organize molecular platforms and link them to the cytoskeleton and intracellular organelles.
How the interactions of dystrobrevin with its binding partners are spatially and temporally regulated is still not understood. Post-translational modifications such as phosphorylation may be among the regulatory factors. We found, indeed, that dystrobrevin is a substrate of PKA, and that PKA phosphorylation regulates its binding affinity for the cargo-binding domain of kinesin heavy chain (Kif5A) .
In this study, we used biochemical, molecular and proteomic analyses to investigate which serine/threonine residues of β-DB were phosphorylated by PKA. Among the phosphorylated residues identified, we found that Thr11 is the regulatory residue for the β-DB–kinesin interaction. As our previous observations indicated that β-DB behaves, similarly to A kinase anchoring proteins (AKAPs), as a platform for phosphatases and kinases , we investigated whether β-DB is a substrate for protein kinases other than PKA. We found that Thr11 is also phosphorylated by protein kinase C (PKC), suggesting that it may be a key residue for docking and undocking of the vesicular or multiprotein complex cargoes from the motor protein.
β-DB is phosphorylated on multiple sites by PKA
Our previous studies indicated that β-DB can be phosphorylated in vitro by PKA, and that its phosphorylation regulates the interaction with the neuronal kinesin heavy chain . Here, to map the phosphorylation site(s) of β-DB and identify the serine/threonine residue(s) involved in the regulation of the binding to kinesin, glutathione-S-transferase (GST)–β-DB full-length and GST–β-DB-deleted proteins (Fig. 1A) were phosphorylated in vitro by the PKA catalytic subunit in the presence of [32P]ATP[γP]. Autoradiography after SDS/PAGE showed that β-DB was efficiently phosphorylated on multiple sites, as indicated by the fact that most of the β-DB deletion mutants, with the mutations covering different regions of the molecule, were labeled (Fig. 1B). Dystrobrevin shares a common domain organization with the C-terminal domain of dystrophin containing two EF-hands, a ZZ, and two coiled-coil domains. Our data indicate that at least one phosphorylation site is present in the N-terminal region (amino acids 1–352), comprising EF-hands and ZZ domains (Fig. 1B, left panels), as well as in the C-terminal region containing coiled-coil regions (amino acids 399–615) (Fig. 1B, right panels). Further analysis of the β-DB C-terminal region indicated that these phosphorylated residue(s) were located in the amino acid sequence 399–468, which contains the first coiled-coil motif (H1) (Fig. 1A), whereas no phosphorylation sites were present in the β-DB region that contains the second coiled-coil motif (H2) and the C-terminus, as indicated by the absence of any phosphorylation signal in β-DB469–615 (Fig. 1B, right panels).
β-DB is phosphorylated on Thr424 by PKA
In order to identify the phosphorylated residue(s) present in β-DB399–468, the band corresponding to this mutant was subjected to MS analysis (Fig. 2). After enzymatic digestion and immobilized metal affinity chromatography (IMAC) enrichment, MALDI-TOF MS analysis revealed only one phosphorylated peptide corresponding to phospho-LAAEAGNMTRPPT(412–424) (m/z 1408.22) (Fig. 2A), and Thr424 was identified as the phosphorylated residue by MS/MS analysis (Fig. 2B). To verify whether Thr424 could have a regulatory role in modulating the interaction between β-DB and kinesin, we performed surface plasmon resonance (SPR) experiments with the kinesin heavy chain cargo-binding domain (GST–Kif5A804–1027) as analyte and β-DB deletion mutants as ligands. We previously demonstrated that the binding between β-DB and kinesin is a high-affinity interaction (Kd ~ 30 nm), and that the β-DBΔ411–530 deletion mutant, which lacks the whole coiled-coil region (H1 and H2), binds to kinesin with slightly reduced affinity . The deletion of this region results in the removal of Thr424 (Fig. 3A). We phosphorylated in vitro full-length β-DB and β-DBΔ411–530 (Fig. 3B) and used them in SPR experiments, together with their nonphosphorylated counterparts. We found that phosphorylated β-DBΔ411–530 showed a reduced sensorgram as compared with the nonphosphorylated protein (Fig. 3C), indicating reduced binding to kinesin. This result is similar to that obtained with full-length β-DB1–615 (Fig. 3C), implying that the lack of Thr424 does not affect the phosphorylation-dependent regulation of β-DB binding to kinesin. Other phosphorylated residue(s) are therefore responsible for the reduced sensorgram.
Phosphorylation of the N-terminal domain regulates β-DB binding to kinesin
To verify whether the phosphorylation of serine/threonine residue(s) present in the N-terminal region of β-DB affects its capacity to bind to GST–Kif5A804–1027, we tested the deletion mutant β-DB1–52 (Fig. 4A) in SPR experiments after in vitro phosphorylation by PKA. We observed a reduction of the sensorgram when phosphorylated GST–β-DB1–352 was passed over GST–Kif5A804–1027 immobilized on the sensor chip, as compared with the nonphosphorylated protein (Fig. 4B). This indicates that phosphorylation site(s) localized in the N-terminal region of the protein are involved in regulating the interaction of β-DB and kinesin. To identify these modified residue(s), phosphorylated and nonphosphorylated β-DB1–352 proteins were analyzed by MS. By combining different digestion strategies (trypsin, LysC, AspN, and chymotrypsin) and CNBr treatment, we obtained 83% sequence coverage of β-DB1–352 (Fig. 4C). Nonetheless, we were not able to detect any phosphorylated peptide by fingerprint (MALDI-TOF MS) or by MS/MS [linear trap quadrupole (LTQ)] analysis, even after IMAC enrichment (data not shown). There could be different reasons for this, the main one being the low abundance of the phosphorylated protein moiety.
β-DB is phosphorylated on Thr11 by PKA
To overcome this difficulty, we took advantage of the observation that in vitro expression of N-terminal β-DB mutants shorter than β-DB1–352 resulted in products with molecular masses lower than expected. In fact, recombinant GST–β-DB1–236 (Fig. 5A), when subjected to SDS/PAGE, showed a band of ~ 30 kDa, which was dominant over the expected one of 54 kDa (Fig. 5B). As the 30-kDa fragment was shown to be phosphorylated in vitro by PKA (Fig. 5B, 32P), we decided to use it for MS analysis. After PreScission Protease digestion to remove GST, the digested sample was subjected to Tricine/SDS gel electrophoresis. Autoradiography of the gel revealed two 32P-labeled bands with apparent molecular masses of 2–4 kDa (Fig. 5C). MALDI-TOF MS analysis of the phosphorylated peptides digested with CNBr identified the N-terminal sequence (amino acids 1–22) of β-DB (Fig. 5D). However, the spectra did not reveal peptide 2–12 of m/z 1214.6493, perhaps because of ion intensity suppression by the flanking peptide 13–22 of m/z 1216.6690. Within the undetected peptide β-DB2–12, there is only one phosphorylatable residue, Thr11, which is localized within a consensus sequence Lys-Arg-Lys-Thr (amino acids 8–11), for PKA phosphorylation, suggesting that Thr11 could indeed be the phosphorylated residue.
Therefore, to verify this hypothesis, we performed in vitro phosphorylation of the synthetic peptide biotin-Met-Ile-Glu-Glu-Gly-Gly-Asn-Lys-Arg-Lys-Thr-Met-Ala-Glu-Leu-Arg-Gln-Lys-Phe-Ile-Glu-Met-Arg (corresponding to amino acids 1–23 of β-DB) by incubation with PKA and [32P]ATP[γP]. This resulted in effective phosphorylation of the peptide (data not shown). This synthetic peptide, either phosphorylated or not, was further analyzed by MS/MS, finally allowing the detection of phosphorylated Thr11 (Fig. 6A,B).
Phosphorylation of Thr11 reduces binding of β-DB to kinesin
To investigate whether phosphorylation of Thr11 regulates the binding of β-DB to kinesin, we used site-directed mutagenesis to mutate Thr11 either into Asp, which mimics a persistent phosphorylated residue, or into Ala. Wild-type β-DB and β-DB mutants (β-DBT11D and β-DBT11A, respectively) were obtained as GST-fused proteins, and used to pull down kinesin from rat brain extracts. Pulled-down proteins were analyzed by SDS/PAGE and western blot. The amount of kinesin bound to β-DBT11D was ~ 80% lower than that bound to wild-type β-DB and β-DBT11A (Fig. 7), indicating that phosphorylation on Thr11 reduces the ability of β-DB ability to bind to kinesin. As a control, we tested the blot with an antibody against another β-DB-binding protein, syntrophin. The amount of pulled-down syntrophin was similar for the three β-DB proteins (Fig. 7), indicating that phosphorylation on Thr11 does not affect the binding of β-DB to other partners.
β-DB is a substrate for PKC
We have previously reported that β-DB behaves as an AKAP-like protein, tethering PKA and other signaling proteins to defined intracellular sites . Besides PKA, AKAPs can directly bind various signaling proteins such as PKC, protein phosphatases, PKA substrates, and adaptor proteins [19, 23]. The finding that 4β-phorbol 12-myristate 13-acetate treatment of MDCK cells alters dystrobrevin association with ZO-1 at the cell junctions by activating PKC  prompted us to investigate whether dystrobrevin could also be a substrate for PKC phosphorylation. To verify this hypothesis, GST–β-DB deletion mutants (Fig. 8A) were phosphorylated in vitro by the PKC catalytic subunit in the presence of [32P]ATP[γP], and analyzed with SDS/PAGE (Fig. 8B, CB) followed by autoradiography (Fig. 7B, 32P). We found that β-DB1–615, β-DB1–352 and β-DB399–468 were phosphorylated by PKC, whereas β-DB499–615 was not (Fig. 8A,B). SPR experiments showed that PKC phosphorylation of β-DB1–352 reduced the binding to kinesin, as compared with nonphosphorylated β-DB1–352 (Fig. 8C).
We performed MS analysis on β-DB1–352 phosphorylated by PKC. By combining different digestion strategies (trypsin, LysC, AspN, and chymotrypsin) and CNBr treatment, we obtained a sequence coverage of 73% (Fig. 9). We were able to identify, by LC-MS/MS, the following phosphorylated residues: Thr69, Thr179, and Thr212 (Fig. 10). When we tested the synthetic peptide (amino acids 1–23), we found that it was phosphorylated by PKC (data not shown), suggesting that Thr11 could be an additional target of this kinase.
In this study, we have demonstrated that β-DB is a phosphorylation substrate for both PKA and PKC, and have shown that Thr11 is a key residue for regulating the binding of β-DB to kinesin heavy chain. Thr11 is a target for both kinases, suggesting that phosphorylation at this site may occur following the activation of different signaling pathways.
Besides being a component of the DGCs in muscle and nonmuscle tissues, dystrobrevin is now thought to participate, through its multiple interactions, to protein networks connecting different cell compartments, also playing a role in signaling. Numerous dystrobrevin-binding proteins, which may specify dystrobrevin function, have been identified so far [9-12, 15-17, 19, 20]. How these multiple associations take place in the cell may depend on several factors, such as the dystrobrevin isoform (α, β, and their spliced variants) [25-28], the cell type , the stage of development of the tissue , and the binding affinity for the different partners . For instance, as we reported previously, both α-dystrobrevin and β-DB can bind to the cargo-binding domain of kinesin heavy chain, but with different binding affinities, as measured by SPR .
The function of dystrobrevins and the interaction with their binding partners can also be regulated by post-translational modifications such as phosphorylation. It is known that α-dystrobrevin, which was first isolated and characterized in the torpedo electric organ neuromuscular junctions, is a tyrosine-phosphorylated protein . The phosphorylated tyrosines are located in the unique α1-dystrobrevin C-terminal domain , a region that is absent in other α-dystrobrevin isoforms as well as in β-DB [7, 8]. The function of α-dystrobrevin tyrosine phosphorylation has been investigated in the skeletal muscle of transgenic mice . The findings revealed that the effectiveness of α1-dystrobrevin in stabilizing the neuromuscular junction depends, in part, on its ability to serve as a tyrosine kinase substrate, and that α-dystrobrevin tyrosine phosphorylation may be a key regulatory point for synaptic remodeling . Moreover, both α-dystrobrevin and β-DB are in vitro substrates of PKA . β-DB phosphorylation specifically reduces the binding to kinesin heavy chain but not to other dystrobrevin-binding partners, such as dysbindin .
Here, we have identified, by MS, two β-DB residues that are phosphorylated in vitro by PKA, Thr11 and Thr424. We first analyzed the functional role of phosphorylation on Thr424. We had previously demonstrated that deletion of the coiled-coil or the C-terminal region decreased the kinetics of the binding of β-DB to kinesin, suggesting that factors affecting β-DB tertiary structure may play a role in regulating its associations . As Thr424 is located next to the first coiled-coil domain, we wondered whether Thr424 phosphorylation was involved in the regulation of β-DB–kinesin association, and found that it was not. Nonetheless, Thr424 is adjacent to motifs involved in the interaction with other partners (syntrophin and dystrophin), and it is therefore conceivable that phosphorylation on this residue may influence the functionality of this region. In this context, our preliminary data (not shown) suggest that Thr424 phosphorylation does not directly affect the binding to syntrophin; further studies are needed to reveal the function of this phosphorylation site.
We previously reported that the binding site for kinesin heavy chain is located in the N-terminal region of β-DB (amino acids 1–236) and that phosphorylation influences the interaction between the two proteins. We subjected the N-terminal deletion mutant β-DB1–352 to MS analysis to detect β-DB phosphorylated residues. Despite a sequence coverage of 83%, we failed to identify any phosphorylated peptide. We overcame this problem by analyzing the phosphorylated short products derived from β-DB1–236, which allowed us to restrict our search to the first 22 amino acids of the molecule. In this peptide, a putative PKA consensus sequence (Lys-Arg-X-Thr/Ser) precedes the threonine at position 11. By the use of a synthetic peptide (amino acids 1–23), we finally identified Thr11 as the PKA target in the N-terminus of β-DB, and by site-directed mutagenesis we demonstrated that phosphorylation of Thr11 regulates the interaction of β-DB with kinesin. Indeed, when Thr11 was mutated into aspartic acid, which mimics a phosphorylated residue, the ability of β-DB to bind kinesin from rat brain extract was reduced by ~ 80% in comparison with wild-type or T11A mutated β-DB. Phosphorylation on Thr11 may function as a mechanism to undock β-DB-bound cargoes transported by the kinesin molecular motor to specific cell compartments. So, it appears that the high-affinity interaction of β-DB with kinesin is a highly dynamic one, as it can be regulated by very fast events such as phosphorylation. The modification of the molecule induced by phosphorylation on Thr11 does not affect binding to molecules that involves regions other than the N-terminus, such as syntrophin (Fig. 7) or dysbindin , suggesting that the release of the motor protein does not affect the binding of β-DB to other partners.
Besides being phosphorylated by PKA, we found that β-DB is also an in vitro substrate for PKC phosphorylation. This finding supports our suggestion that β-DB behaves as an AKAP-like protein. AKAPs assemble and organize kinases and phosphatases on the same molecular platform, and they are themselves enzymatic substrates . β-DB phosphorylation by PKC resulted in a reduction of the binding to kinesin similar to that following phosphorylation by PKA. We found that the synthetic peptide corresponding to amino acids 1–23 of β-DB, which contains Thr11, was phosphorylated by PKC, suggesting that Thr11 may be a key residue modified by the activation of different signaling pathways. Moreover, we identified other threonines that are phosphorylated in vitro by PKC (Thr69, Thr179, and Thr212), all of which are in the EF-hand-like domains of β-DB. From the crystallography structure of the homologous dystrophin region , we can infer information about the tertiary structure of the N-terminal region of dystrobrevin. Thr69 is located between the first and the second α-helical structures of the EF1-hand-like domain, whereas Thr179 and Thr212 are located at the end of the second and fifth α-helical structures of the EF2-hand-like domain, respectively . Phosphorylation at these sites could alter the conformation of β-DB and, as a consequence, its association with partners that bind to the N-terminal region. So, even if β-DB is the target of different kinases, phosphorylation on diverse residues could result in different signaling pathways/cascades, depending on the β-DB-specific binding partners in different cell types.
Phosphorylation plays an essential role in the regulation of a large number of processes, from cell metabolism to higher-level functions of learning and memory. Phosphorylation events are controlled through the compartmentalization of protein kinases and phosphatases provided by scaffold and anchor proteins. When extracellular or intracellular stimuli reach the subcellular compartment and activate signaling components, scaffold proteins themselves may be the targets of phosphorylation/dephosphorylation processes. The fact that dystrobrevin is phosphorylated by both PKA and PKC suggests that, like other scaffold proteins, dystrobrevin can play a role not only as a stationary anchor but also as a dynamic signaling component. We have not yet investigated the possible implications of Thr11 and Thr424 phosphorylation on β-DB functions in terms of, for instance, affecting cell localization or protein lifetime. During our experiments, we observed that, in solution, the phosphorylated dystrobrevin is slightly unstable as compared with the unphosphorylated one (not shown). If this is true, phosphorylation could be also regarded as a mechanism to regulate protein lifetime. Further studies are necessary to test this hypothesis.
The results reported in this article contribute to our understanding of how dystrobrevin molecular interactions may be regulated, and add new evidence to the previously suggested hypothesis [17, 19, 22] of dystrobrevin involvement in trafficking and signaling mechanisms.
Preparation of recombinant proteins
Full-length β-dystrobrevin (β-DB1–615) and β-DB deletion mutants were subcloned into the pGEX-6P-1 vector (GE Healthcare, Princeton, NJ, USA), expressed in Escherichia coli BL21(DE3) cells as GST-tagged proteins, and purified by affinity chromatography on GST-Bind Resin (Novagen, Merk Chemical, Darmstadt, Germany) [16, 17, 19, 22]. β-DB399–468 was obtained with basically the same strategy; β-DBΔ411–430 has been previously described as β-DBΔH1–H2 in Ceccarini et al. . In some experiments, GST was removed from GST–β-DB polypeptides by cleavage with PreScission protease (GE Healthcare). Point mutations were introduced into full-length β-DB with the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA). All of the constructs were sequenced. The C-terminal cargo-binding domain (amino acids 804–1027) of the neuronal kinesin heavy chain (Kif5A) was expressed as a GST-fusion protein, as previously described .
Electrophoresis and western blot
Protein samples were separated by SDS/PAGE and stained with Coomassie blue. Small peptides were analyzed by Tricine/SDS/PAGE. For western blot analysis, protein samples were transferred to nitrocellulose membrane (Whatman, Dassel, Germany) and incubated overnight at 4 °C with kinesin heavy chain polyclonal antibody or with syntrophin mAb (MA-1-745, 1 : 2000; Affinity BioReagents, Golden, CO, USA), and then incubated for 1 h at room temperature with horseradish peroxidase-conjugated anti-mouse or anti-rabbit IgG (1 : 500; Pierce, Rockford, IL, USA). Immunoreactive bands were visualized with an enhanced chemiluminescence reagent (Pierce), and revealed on X-ray films.
In vitro phosphorylation
Agarose-bound full-length GST–β-DB1–615 (β-DB1–615), GST–β-DB deletion mutants (β-DB1–236, β-DB1–352, β-DB214–539, β-DB399–615, β-DB399–468, and β-DB469–615) and GST were phosphorylated by the PKA catalytic subunit as previously described . For PKC phosphorylation, 0.5 mL of agarose-bound GST–β-DB polypeptide suspension was equilibrated in 20 mm Hepes (pH 7.4), 1.67 mm CaCl2, 1 mm dithiothreitol, 10 mm MgCl2, and 200 μm ATP, and incubated with 0.5 μg·mL−1 PKC (Promega, Madison, WI, USA) in the presence of 0.6 mg·mL−1 phosphatidylserine (Sigma, St Louis, MO, USA) (kinase mixture) for 30 min at 30 °C. As a control for the phosphorylation reaction, 10 μCi of [32P]ATP[γP] (3.000 Ci·mmol−1) was added to 50 μL of kinase mixture. Radioactive samples were directly solubilized in SDS/PAGE sample buffer, and analyzed by gel electrophoresis. The 32P-labeled bands were revealed with InstantImager (Packard Instrument Company, Meriden, CT, USA) or by autoradiography.
SPR analysis was performed with a BIAcoreX instrument equipped with two flow cell sensor chips (BIAcore Intl. AB, Uppsala, Sweden). Immobilization of the proteins on the two flow cell sensor chips was achieved by covalently coupling them to CM5 sensor chips, as previously described . Recombinant kinesin heavy chain cargo-binding domain, GST–Kif5A804–1027, was immobilized on one flow cell and GST on the reference cell of the sensor chip. Experiments were performed in HBS-EP buffer [10 mm Hepes, pH 7.4, 0.15 m NaCl, 3 mm EDTA, 0.005% (v/v) surfactant P20] with a flow rate of 30 μL·min−1 at 25 °C. After each injection of β-DB polypeptides, the amount of protein bound to the sensor chip was monitored by the change in the refractive index [given in arbitrary response units (RU)]. The response was monitored as a function of time (sensorgram), as the difference (Δ) between signals arising from the cell with the immobilized GST–Kif5A804–1027 and signals from the reference cell with immobilized GST. At the end of the sample plug, HBS-EP buffer was passed over the sensor surface to allow dissociation. Regeneration of the sensor surface was obtained by injecting a 2-μL pulse of regenerating buffer (50 mm NaOH) at a flow rate of 20 μL·min−1. SPR data were analyzed with biaevaluation software Version 4.1 and the Langmuir model for 1 : 1 binding.
Aliquots of PreScission-cleaved deletion mutants of β-DB were separated by SDS/PAGE and stained with the Colloidal Blue Staining kit (Invitrogen, Carlsbad, CA, USA). Slices were excised and in-gel digested with sequencing-grade trypsin (Promega) . The same protocol was used with the endoproteinase LysC, AspN and chymotrypsin (Roche, Penzberg, Germany). CNBr (Fluka, Milwaukee, WI, USA) was dissolved in 70% formic acid and 1% β-mercaptoethanol, and hydrolysis was performed for 20 h in the dark. IMAC phosphopeptide enrichment was performed following the manufacturer's guidelines (Phosphopeptide Enrichment Kit; Pierce). Peptide mass fingerprinting was performed with a MALDI-TOF Voyager DE-STR (Applied Biosystems, Foster City, CA, USA) in positive reflectron mode, with phospho-DHB as matrix . MS spectra were processed with data explorer (Applied Biosystems) and gpmaw software . LC-MS/MS analysis was performed with a nano-HPLC 3000 Ultimate (Dionex, Sunnyvale, CA, USA) connected in-line to an LTQ XL Linear Ion Trap (ThermoFisher, Waltham, MA, USA). Peptide mixtures were eluted along a 60-min linear gradient of buffer B (95% acetonitrile, 0.1% formic acid). Spectra were acquired in data-dependent mode with five precursor scans and active exclusion list. Optional wideband activation, neutral loss MS3 for phosphate groups (m/z 32.6, 49.0, and 98.0) and, eventually, parent mass data-dependent methods were enabled. Raw spectra were analyzed with bioworks 3.3.1 (ThermoFisher). Carboamidomethylation of cysteines was specified as a fixed modification, and oxidation of methionine and phosphorylation of serine and threonine were set as variable modifications; mass tolerance was set to 1.0 Da for the precursor ion and 0.8 Da for fragment ions, and a maximum of two missed cleavages was allowed. The minima criteria for protein identification were as follows: peptide probability was < 0.001; and peptide XCorr versus charge was 1.5 for charge +1, 2 for charge +2, 2.5 for charge +3, and 3 for charge +4.
Agarose-bound GST, full-length GST–β-DB and GST–β-DB mutants were used in in vitro protein-binding assays. Rat brain (1.4 g) was homogenized in 15 mL of ice-cold 0.32 m sucrose, 10 mm Hepes (pH 7.4), and 0.5 mm EDTA, in the presence of protease inhibitors (1 μg·mL−1 leupeptin, 2 μg·mL−1 benzamidine, 0.15 mm phenylmethanesulfonyl fluoride), and centrifuged at 650 g for 20 min at 4 °C. The supernatant was centrifuged at 35 000 g for 20 min at 4 °C. The pellet was solubilized (1 : 1, v/v) in extraction buffer (20 mm Hepes, pH 7.4, 150 mm NaCl, 5 mm MgSO4, 1 mm dithiothreitol, 0.2% Triton X-100) in the presence of protease inhibitors, and, after 30 min of incubation on ice, the sample was centrifuged at 35 000 g for 30 min at 4 °C. The supernatant was precleared by incubation with agarose-bound GST, and 0.5 mL (1 mg·mL−1) was then incubated with 40 μL of a 50% (v/v) suspension of agarose-bound GST–β-DB, GST–β-DBT11A, GST–β-DBT11D and GST overnight at 4 °C. Following exhaustive washes with extraction buffer, pulled-down proteins were separated by gel electrophoresis and analyzed by western blot.
We are indebted to C. Bucalossi for her collaboration and L. Dimiziani for his contribution during the preliminary phases of the project. We thank S. Cecchetti for helpful discussions. We are also grateful to M. Delle Femmine for helping with the figures. The project was partially supported by the Italian Ministero della Salute (Grant PS-NEURO ex 56/05/20) and by the Italy–USA Collaborative Programme (Grant Rare Diseases, 7DR1).