Biochemical Journal

Research article

Interaction between the SH3 domain of Src family kinases and the proline-rich motif of HTLV-1 p13: a novel mechanism underlying delivery of Src family kinases to mitochondria

Elena Tibaldi , Andrea Venerando , Francesca Zonta , Carlo Bidoia , Elisa Magrin , Oriano Marin , Antonio Toninello , Luciana Bordin , Veronica Martini , Mario Angelo Pagano , Anna Maria Brunati

Abstract

The association of the SH3 (Src homology 3) domain of SFKs (Src family kinases) with protein partners bearing proline-rich motifs has been implicated in the regulation of SFK activity, and has been described as a possible mechanism of relocalization of SFKs to subcellular compartments. We demonstrate in the present study for the first time that p13, an accessory protein encoded by the HTLV-1 (human T-cell leukaemia virus type 1), binds the SH3 domain of SFKs via its C-terminal proline-rich motif, forming a stable heterodimer that translocates to mitochondria by virtue of its N-terminal mitochondrial localization signal. As a result, the activity of SFKs is dramatically enhanced, with a subsequent increase in mitochondrial tyrosine phosphorylation, and the recognized ability of p13 to insert itself into the inner mitochondrial membrane and to perturb the mitochondrial membrane potential is abolished. Overall, the present study, in addition to confirming that the catalytic activity of SFKs is modulated by interactors of their SH3 domain, leads us to hypothesize a general mechanism by which proteins bearing a proline-rich motif and a mitochondrial localization signal at the same time may act as carriers of SFKs into mitochondria, thus contributing to the regulation of mitochondrial functions under various pathophysiological conditions.

  • human T-cell leukaemia virus type 1 (HTLV-1) p13
  • mitochondrion
  • Src family kinase (SFK)
  • SH3 domain

INTRODUCTION

SFKs (Src family kinases) are non-receptor tyrosine kinases acting as molecular switches that are generally described as associated with the inner face of the plasma membrane, integrating diverse signals generated by the cell surface receptors in response to a large number of extracellular cues, and regulate a variety of cellular events, such as cell growth and proliferation, cell adhesion and migration, differentiation, survival and death [13]. In the last decade, SFKs have been reported to target other subcellular compartments, such as endosomes, secretory granules or phagosomes, the endoplasmic reticulum, the Golgi apparatus and mitochondria [49]. Although distinctive localizations of SFK members have been implicated in their specific functions, the mechanism of localization and their function exerted in these organelles still remain to be determined. SFKs comprise eight members classed into two groups on the basis of the phylogenetic tree inferred from the analysis of the complete sequence of each kinase [10], namely Src-related (Src, Yes, Fyn and Fgr) and Lyn-related (Lyn, Hck, Lck and Blk). SFKs possess a common modular structure consisting of: (i) a SH4 (Src homology 4) domain, a unique N-terminal sequence for myristoylation and/or palmitoylation; (ii) a SH3 domain, which associates with specific proline-rich sequences; (iii) a SH2 domain, which recognizes p-Tyr (phosphotyrosine) motifs; and (iv) a kinase domain, also referred to as a SH1 domain, followed by a C-terminal tail playing a negative regulatory role when phosphorylated by the inhibitory CSK (c-Src tyrosine kinase) [11].

SFKs are ordinarily kept in a closed inactive conformation through two major intramolecular inhibitory interactions, namely the binding of the C-terminal CSK-phosphorylated tyrosine (i.e. Tyr527 of c-Src) to the SH2 domain, and the interaction of a PPII (polyproline type II) helical motif (containing the core consensus sequence Pro-X-X-Pro; where X is any amino acid) within the SH2-kinase linker with the SH3 domain [12,13]. Conformational transition from the inactive to the active form of SFKs is induced by an array of factors, including physiological and pathological stimuli triggering dephosphorylation of the C-terminal tail by several tyrosine phosphatases, and displacement of the SH2 or SH3 domains by specific ligands that lead to the disruption of the inhibitory intramolecular interactions; these events can occur singly, sequentially or synergistically [1420]. Importantly, the open active form of SFKs is in turn susceptible to recognition by interacting proteins capable of engaging the non-catalytic domains, resulting in modulation of SFK activity [3,11].

The complex regulation of SFK activity is well exemplified by the specific binding of the SH3 domain to ligands displaying a proline-rich motif, which, on the basis of the activation state of SFKs and the type of interactors, can be a mechanism of activation or inhibition of SFKs themselves [3,11]. In fact, although the HIV accessory protein Nef was the first molecule shown to elicit the SFK activity through this mode of interaction [18], followed by other viral [2124] and cellular [2529] proteins, which all bear a proline-rich motif, p66shc (Src homology and collagen homology, 66 kDa isoform) has recently been described to interact with the active form of SFKs and initiate inhibition through a proline-rich motif/SH3 interaction [30]. Emerging evidence indicates that, in addition to the prevailing view of a regulatory role of the kinase activity, this type of interaction can influence the turnover of these kinases [31] and/or their localization in various subcellular compartments [27,32]. Recently, we have demonstrated that Lyn is not only resident in the intermembrane space of mitochondria [33], but can also be translocated from the plasma membrane to mitochondria in the early phases of liver regeneration, taking part in a multiprotein complex, ultimately preserving mitochondrial integrity.

The aim of the present study was to assess whether SFKs can traffic to mitochondria through assembly with proteins known to be imported into mitochondria and containing a proline-rich motif putatively capable of binding SH3 domains. In this regard, the HTLV-1 (human T-cell leukaemia virus type 1) accessory protein p13 fulfilled the above criteria, being targeted to mitochondria and possessing a well-defined C-terminal proline-rich motif [34]. In the present study we demonstrate that the interaction between p13 and distinct SFKs dramatically activates SFK activity and results in the import of SFKs into mitochondria. Intriguingly, the recognized capability of p13 to trigger an inward K+ current, leading to swelling, depolarization and increased respiratory chain activity in mitochondria [34], is strongly impaired by the SH3-dependent assembly with SFKs, supporting the notion of a scaffolding function of this family of protein kinases and disclosing novel insights into the action exerted by p13 in HTLV-1-infected cells.

EXPERIMENTAL

Materials

All analytical grade reagents, cell culture media and phosphatase inhibitor cocktails 2 and 3 were from Sigma–Aldrich. [γ-33P]ATP (3000 Ci/mmol) was purchased from PerkinElmer. Anti-Lyn (sc-15), anti-Src (sc-19), anti-Fyn (sc-434), anti-Fgr (sc-130) and anti-TOM20 rabbit polyclonal FL-145 (sc-11415) (anti-TOM20 R) and F-10 mouse monoclonal (sc-17764) (anti-TOM20 M) antibodies were from Santa Cruz Biotechnology. The monoclonal anti-p-Tyr antibody (clone PY-20) was from BioSource International. Anti-aconitase (ab102803), anti-AIF (apoptosis-inducing factor) (ab1998) and anti-LDH (lactate dehydrogenase) (ab52488) antibodies were from Abcam. Complete protease inhibitor cocktail tablets and anti-c-Myc antibody (clone 9E10) were from Roche Diagnostics. Anti-phospho-SFK (anti-Src pY416) antibody was from Cell Signaling Technology. PP2 {4-amino-5-(4-chlorophenyl)-7(t-butyl)pyrazolo[3,4-d]pyrimidine} as well as HRP (horseradish peroxidase)-conjugated antibodies were purchased from Calbiochem. The ECL (enhanced chemiluminescence) detection system was from GE Healthcare.

Peptide synthesis, production and purification of anti-p13 antibody

The full-length form of wild-type p13, the p13-derived peptide covering the N-terminal region (p139–41), and the p13-derived peptide covering the C-terminal proline-rich motif (p1361–87) were synthesized by solid-phase peptide synthesis as described previously [35]. The anti-p13 antibody was raised against p13 in New Zealand rabbits held in the animal house of the Department of Biological Chemistry, University of Padova, following the procedures approved by the Animal Ethical Committee of the University of Padova (Comitato Etico di Ateneo per la Sperimentazione Animale, CEMSA), in accordance with the guidelines issued by the Ministry of Health, Government of Italy. Consistently, rabbits were anaesthetized locally by applying EMLA cream (a mixture of lidocaine and 2.5% prilocaine) at the sizes of injection of the antigen (rabbit back) and collection of blood (rabbit ear). To purify the anti-p13 antibody, the serum obtained from blood samples were subjected to affinity chromatography after immobilization of p13 on SulfoLink Coupling Gel (Pierce), according to the manufacturer's instructions. The antigenicity of full-length wild-type p13 and the N- and C-terminal regions of p13 itself was tested by dot blot analysis.

Dot blot analysis

Increasing quantities (from 10 ng to 10 μg) of full-length synthetic p13, p139–41 or p1361–87 were spotted on to nitrocellulose membranes. The membranes were blocked with 3% (w/v) BSA in TBS (Tris-buffered saline) (50 mM Tris/HCl, pH 7.5, and 150 mM NaCl) for 1 h at room temperature (25°C), incubated with the anti-p13 antibody (1:500 or 1:2000 dilution) for 2 h at room temperature, washed three times with TBST (TBS supplemented with 0.1% Tween 20), and then incubated with HRP-conjugated anti-rabbit antibody (1:5000) in TBS at room temperature for 30 min. Membranes were washed three times with TBST and developed using the ECL detection system, captured using Kodak Image Station 2000R and visualized using Kodak 1D Image software (Eastman Kodak).

Far-Western blotting

Removal of brain and spleen from rats to obtain tyrosine kinases used throughout the present study was performed according to the guidelines approved by the Animal Ethical Committee of the University of Padova as described above. Src (0.1 μg), purified from rat brain as described previously [36], 0.1 μg of Fyn, Fgr and Lyn, purified from rat spleen as described previously [9], and 0.1 μg of BSA, as a negative control, were subjected to SDS/PAGE (10% gels) and blotted on to nitrocellulose membranes. The membranes were blocked by 3% (w/v) BSA in TBS and then incubated in a buffer containing 20 mM Tris/HCl, pH 7.5, 300 mM KCl and 0.1% Tween 20 for 30 min at 4°C. Synthetic full-length p13 (30 μg/ml) was then overlaid for 2 h at 20°C in the absence or presence of p1361–87 and probed with anti-p13 antibody. After washing with TBST, the membranes were incubated with HRP-conjugated polyclonal antibody for 1 h. Immunoblots were developed by the ECL detection system, captured using Kodak Image Station 2000R and visualized by Kodak 1D Image software.

Western blot analysis

Samples were rapidly solubilized in SDS buffer [62 mM Tris/HCl buffer, pH 6.8, containing 5% (v/v) glycerol, 0.5% 2-mercaptoethanol and 0.5% SDS] and subjected to SDS/PAGE before being transferred on to nitrocellulose membranes by electroblotting. After treatment with 3% (w/v) BSA in TBS at 4°C overnight, membranes were incubated for 2 h with primary antibodies (anti-Lyn, anti-Src, anti-Fyn and anti-Fgr antibodies diluted 1:400; anti-p-Tyr, anti-aconitase, anti-AIF and anti-LDH antibodies diluted 1:1000). After washing with TBST, the membranes were incubated with secondary HRP-conjugated polyclonal antibody for 1 h. Immunoblots were developed using the ECL detection system. Images were captured using Kodak Image Station 2000R and visualized by Kodak 1D Image software. Loading controls were performed by re-probing membranes with appropriate antibodies after stripping twice in 0.1 M glycine, pH 2.5, 0.5 M NaCl, 0.1% Tween 20, 1% (v/v) 2-mercaptoethanol and 0.1% NaN3, for 10 min each.

Cell culture and transfection

HeLa cells were maintained in DMEM (Dulbecco's modified Eagle's medium) supplemented with 10% FBS (fetal bovine serum), 100 units/ml penicillin and 20 units/ml streptomycin. The eukaryotic expression plasmids pcDNA3.1/myc-His C (empty vector) and pcDNA3.1/myc-His C/p13 (the latter containing the wild-type p13 coding sequence), provided by Dr Luc Willems (Cellular and Molecular Biology, Gembloux Agro-Bio Tech, Gembloux, Belgium), as well as pCMV6-XL4 (empty vector) and pCMV6-XL4/Lyn (the latter containing the Lyn coding sequence) (OriGene Technologies), were used to perform transfection experiments as described previously [39]. At 36 h after transfection, the medium was, when required, supplemented with 10 μM PP2, and the cells were further incubated for 12 h before being processed as described below.

Immunoprecipitation

p13 and the different SFKs were immunoprecipitated either as purified proteins in interaction assays, or from lysates of purified RLM (rat liver mitochondria) or from mitochondrial lysates of transfected cultured cells. Immunoprecipitation was performed in the absence or presence of the recombinant GST (glutathione transferase) fusion form of the SH3 domain of Lyn (GST–Lyn SH3 domain), expressed and purified according to the protocol described previously [31] or, when required, with p1361–87 for 2 h at 4°C with the appropriate antibodies in competition assays. The resulting immunocomplexes were recovered by incubation for 1 h with Protein A/G–Sepharose previously saturated with BSA and washed three times with 50 mM Tris/HCl, pH 7.5, plus 1 mM orthovanadate, and phosphatase and protease inhibitor cocktails. Samples were then subjected to Western blot analysis with the appropriate antibodies.

Phosphorylation assays

Tyrosine kinase assays were performed in 40 μl of reaction mixture containing 50 mM Tris/HCl, pH 7.5, 10 mM MnCl2, 30 μM ATP/[γ-33P]ATP (specific activity 1000 c.p.m./pmol), 100 μM sodium orthovanadate, 200 μM cdc26–20 peptide as substrate, and 20 ng of Src, Fyn, Fgr or Lyn. Following incubation for 5 min at 30°C, the reaction was blocked by adding 5×SDS buffer and the samples were subjected to SDS/PAGE. Peptide phosphorylation was evaluated using the Cyclone Plus Storage Phosphor System (PerkinElmer).

Gel filtration

A Superdex 75 HR column (GE Healthcare) mounted on a fast-performance liquid chromatography system was equilibrated with 20 mM Tris/HCl, pH 7.5, 150 mM NaCl, 10% (v/v) glycerol, 10 mM 2-mercaptoethanol and 50 μM PMSF. The first 25 ml were collected together, and then fractions of 0.2 ml were collected at a flow rate of 0.4 ml/min.

Preparation of mitochondria

RLM were prepared as described previously [33]. Briefly, rat liver was homogenized in the isolation medium (250 mM sucrose, 5 mM Hepes and 0.5 mM EGTA, pH 7.4) and centrifuged at 900 g for 5 min. The supernatant was centrifuged again at 12000 g for 10 min to precipitate a crude mitochondrial pellet. The pellet obtained was resuspended in an isolation medium plus 1 mM ATP and layered on top of a discontinuous gradient of Ficoll diluted in the isolation medium, composed of 2 ml layers of 16, 14, and 12% (v/v) Ficoll and a 3 ml layer of 7% Ficoll. After centrifugation for 30 min at 75000 g, the mitochondrial pellet was suspended in the isolation medium and centrifuged again for 10 min at 12000 g. The resulting pellet was suspended in the isolation medium without EGTA, and the protein content was measured by the biuret method, using BSA as a standard. The absence of other contaminating subcellular compartments in the mitochondrial preparation was demonstrated as described previously [33].

Treatment of isolated RLM with p13 and SFK

p13 (0.2 μg) was incubated with single SFKs (0.1 μg) in the absence or presence of p1361–87 or GST–Lyn SH3 domain for 2 min at 30°C and subsequently incubated with 50 μg of RLM resuspended in 200 mM sucrose, 10 mM Hepes, pH 7.4, 5 mM succinate, 1.25 μM rotenone, 1 mM sodium phosphate and protease inhibitors for 5 min at 30°C.

The mitochondria were pelleted by centrifugation at 10000 g at 4°C for 10 min and the reaction was stopped by washing twice with isolation medium.

Proteinase K treatment

After treatment with p13 and/or SFKs, as described above, purified mitochondria were treated with 50 ng/ml proteinase K in the isolation medium without EGTA (see above) in the absence or presence of 0.5% Triton X-100 at room temperature for 30 min. The reaction was stopped by the addition of the protease inhibitor cocktail, and then analysed by Western blotting with anti-p13, anti- Lyn, anti-aconitase and anti-AIF antibodies.

Mitochondrial subfractionation

To separate the mitochondrial membranes from the soluble fractions, 50 μg of mitochondria, suspended in isolation medium, were sonicated in an MSE Sonicator and subjected to eight freeze–thaw cycles. Mitochondrial suspensions were then ultracentrifuged at 16000 rev./min for 30 min at 4°C in a Beckman MLA-130 rotor.

Digitonin treatment

Purified mitochondria (1 mg/ml) were incubated with increasing concentrations of digitonin (from 0.1 to 0.6 mg/ml) for 30 min at 4°C. The samples were then centrifuged at 22800 g for 20 min. The supernatant (S) and pellet (P) were subjected to SDS/PAGE and Western blotting analysis with the appropriate antibody.

Determination of mitochondrial membrane potential (ΔΨm)

Membrane potential (ΔΨm) was measured by monitoring the distribution of the lipophilic cation TPP+ (tetraphenylphosphonium) across the mitochondrial membrane with a selective electrode prepared in our laboratory according to published procedures [37] and an Ag/AgCl reference electrode. TPP+ was added at a final concentration of 2 μM in order to achieve high sensitivity in measurements and to avoid toxic effects on the proton ATPase and on calcium movements. The membrane potential measured with the TPP+-selective electrode was calibrated using the equation ΔΨm=(ΔΨelectrode−66.16 mV)/0.92 as proposed previously [38].

Confocal microscopy

HeLa cells were grown on 13 mm diameter glass coverslips and transfected as described above. At 48 h after transfection, cells were washed twice with PBS (1 mM Na2HPO4, 1.8 mM KH2PO4, 140 mM NaCl and 2.7 mM KCl, pH 7.4), fixed with 2% (w/v) paraformaldehyde in PBS for 15 min at room temperature, and then permeabilized with 0.2% Triton X-100 in PBS for 10 min at room temperature. Non-specific binding was prevented by the addition of 3% BSA in PBS for 30 min. Then, the cells were incubated for 1 h at 37°C with the indicated primary antibodies supplemented with 1% (w/v) BSA. After washing with PBS, the cells were incubated for 1 h at 37°C with TRITC (tetramethylrhodamine β-isothiocyanate)-conjugated secondary antibody against rabbit IgG and FITC-conjugated secondary antibody against mouse IgG. After a final washing step with PBS, the coverslips were placed on to glass slides in FluorSave™ reagent mounting medium (Calbiochem). Fluorescence was detected using an UltraView Living Cell Imaging (LCI) confocal system (PerkinElmer). FITC and TRITC were excited at 488 and 561 nm, and emission was collected at 541 and 594 nm respectively. Final image composites were created using Adobe Photoshop 6.0.

Statistical analysis

Data are presented as means±S.D. and compared using 1-way ANOVA followed by a Bonferroni post-hoc test. A P value of less than 0.05 was considered statistically significant. All statistics were performed using GraphPad Prism (GraphPad Software) statistical software.

RESULTS

p13 interacts with SFKs and stimulates their activity via its C-terminal proline-rich region

p13 is an 87-amino acid accessory protein of HTLV-1, the organization of which is shown in Figure 1(A). The primary structure of p13 contains a series of regions, including a proline-rich motif putatively capable of interacting with SH3 domains [34], albeit not fulfilling a canonical consensus, because of the absence of a basic amino acid residue upstream or downstream of the proline residues [3,13].

Figure 1 p13 binds the SH3 domain of Src, Fyn, Fgr and Lyn via its proline-rich region

(A) Schematic representation of the sequence and relevant functional motifs of p13. MTS, mitochondrial targeting signal (residues 21–35); TM, transmembrane region (residues 30–40); H, region with a high flexibility score (residues 42–48); B, predicted β-sheet hairpin structure (residues 65–75); PRM, proline-rich motif (residues 75–84). (B) Src, Fyn, Fgr, Lyn and BSA (0.1 μg), the latter used as an irrelevant control protein, were subjected to SDS/PAGE and transferred on to nitrocellulose membranes. p13, in the absence or presence of p1361–87, was then overlaid on the membranes and probed with anti-p13 antibody. The molecular mass of protein standards is indicated in the middle. (C) p13 (0.2 μg) was incubated with 0.1 μg of each single SFK in the absence or presence of p1361–87 or GST–Lyn SH3 domain, for 10 min at 30°C. The subsequent immunoprecipitation (Ip) was performed with anti-p13 antibody or with specific antibodies against each single SFK. Immunocomplexes were then probed (Wb, Western blot) with antibodies to each single SFK or p13.

To assess whether the C-terminal proline-rich motif of p13 was capable of interacting with the SH3 domain of SFKs, some of the SFK members, Src, Fyn, Fgr and Lyn, were run on SDS/PAGE and transferred on to nitrocellulose to undergo far-Western blot analysis. The immobilized enzymes were then incubated with p13 in the absence or presence of p1361–87, a p13-derived peptide covering its C-terminal proline-rich motif. As shown in Figure 1(B), p13 was detected with of all the SFKs tested (left-hand panel), whereas this interaction was abolished by the presence of p1361–87, indicating a competition with full-length p13 and confirming the inability of the anti-p13 antibody to recognize the C-terminal region of p13 (see Supplementary Figure S1 at http://www.BiochemJ.org/bj/439/bj4390505add.htm). To demonstrate that the SH3 domain of SFKs was implicated in the interaction with p13, and in particular with its C-terminal proline-rich motif, p13 was incubated with single SFKs (molar ratio 3:1) in the absence or presence of p1361–87 or, alternatively, of the GST–Lyn SH3 domain (Figure 1C). By subsequent immunoprecipitation of p13 or the SFKs and Western blot analysis with the respective antibodies, the presence of p1361–87 or the GST–Lyn SH3 domain proved to prevent the formation of the SFK–p13 complex, confirming that this interaction was highly specific and mediated by the C-terminus of the viral protein and the SH3 domain of the SFKs tested. Under these experimental conditions, the residual quantity of SFKs in the supernatant after immunoprecipitating p13 alone was negligible, whereas p13 in the supernatant after immunoprecipitating SFKs was approximately two-thirds (results not shown).

As a number of viral proteins containing a proline-rich motif in their sequence have been shown to enhance SFK activity after engaging the SH3 domain of SFKs [11], we evaluated the ability of p13 to act as a positive regulator of SFKs. For this purpose, SFK activity was tested in vitro using the Src-specific peptide cdc26–20 as a substrate in the presence of increasing concentrations of full-length p13 or p1361–87, in the absence or presence of the GST–Lyn SH3 domain. Figure 2(A) shows that both p13 and p1361–87 exerted an activatory effect on all of the SFKs assayed in a dose-dependent manner, although the full-length form of p13 was more potent in stimulating SFK activity, even at much lower concentrations. On the other hand, the presence of the GST–Lyn SH3 domain blocked the increase in SFK activity due to the full-length viral protein or its C-terminal-derived peptide, highlighting that the proline-rich motif–SH3 domain interaction had a role in modulating SFK activity.

Figure 2 p13 and p1361–87 act as positive modulators of Src, Fyn, Fgr and Lyn

(A) Tyrosine kinase activity of Src, Fyn, Fgr and Lyn was tested on the Src-specific peptide substrate cdc26–20 in the absence (column 1) or presence of increasing concentrations of p13 (columns 2–5) or p1361–87 (columns 6–9) alternatively supplemented with the GST–Lyn SH3 domain (columns 4–5 and 8–10) as described in the Experimental section. Results are expressed as means±S.D. from three separate experiments.**P<0.001 and *P<0.05. (B) Elution profile of activity on Src-specific peptide substrate cdc26–20 of Lyn (top panel), Lyn plus p13 (middle panel) and Lyn plus p13 supplemented with GST–Lyn SH3 domain (bottom panel), obtained from a Superdex 75 HR column as described in the Experimental section. Downward arrows indicate the position of molecular mass standards on the Superdex 75 HR column: BSA (66 kDa), ovalbumin (43 kDa), chymotrypsinogen (26 kDa) and lysozyme (14.5 kDa).

To assess the formation of the Lyn–p13 complex, we also performed size-exclusion chromatography, showing that Lyn, detected at the molecular mass of approximately 60 kDa when alone (Figure 2B, top panel), was eluted at approximately 75 kDa when pre-incubated with p13, indicating that these two proteins form a stable heterodimer with a 5-fold increase in the enzyme activity (Figure 2B, middle panel). In contrast, the presence of the GST–Lyn SH3 domain reversed the activatory effect by competing with Lyn's SH3 domain and hence disrupting the complex (Figure 2B, bottom panel). Similar data were obtained with Src, Fyn and Fgr (results not shown).

SFKs traffic to mitochondria when complexed with p13

p13 is known to mainly target mitochondria through the atypical N-terminal MLS (mitochondrial localization sequence) L21RVWRLCTRRLVPHL35. This sequence is predicted to be part of an α-helix, the amphipathic properties of which are conferred by the four arginine residues, finally localizing to the mitochondrial inner membrane [34]. In contrast with other proteins bearing the canonical MLS, the targeting signal of p13 is not cleaved upon import into mitochondria [34]. After anchoring to the mitochondrial inner membrane, p13 has been recognized to trigger an influx of K+ into the mitochondrial matrix accompanied by a concentration-dependent decrease in the mitochondrial inner membrane potential (ΔΨm) [35]. Since we demonstrated that p13 forms a stable heterodimer with the SFKs examined, we evaluated the ability of p13 to act as a carrier for mitochondrial import of SFKs by virtue of its MLS. For this purpose, highly purified RLM, in which we have previously demonstrated that the protein level of SFKs is undetectable [33], were incubated with Src, Fyn, Fgr and Lyn, alternately, in the absence or presence of p13. After a 5 min incubation, mitochondria were spun down and the resulting fractions, mitochondria (M) and supernatant (S), underwent Western blot analysis with antibodies against p13 and the SFK being studied. As shown in Figure 3, Src and Lyn were found in the mitochondrial fraction only after pre-incubation with p13 (lane 4 compared with lane 6, M), whereas they were totally recovered in the supernatant fraction when the sample was either devoid of p13 (lane 4 compared with lane 6, S) or pre-incubated with p13 in the presence of the GST–Lyn SH3 domain (lane 6 compared with lane 7, M and S). Similar results were also achieved with Fyn and Fgr (results not shown).

Figure 3 p13 acts as a carrier of Src and Lyn for mitochondrial import

(A) RLM (50 μg) was incubated with p13 and/or Src in the absence or presence of the GST–Lyn SH3 domain. Afterwards, mitochondria were spun down and the resulting fractions, mitochondria (M) and supernatant (S), were subjected to Western blot (Wb) analysis with anti-Src antibody and subsequently with anti-p13 antibody. The membranes were re-probed with anti-aconitase antibody as loading control. (B) RLM (50 μg) was incubated with p13 and/or Lyn in the absence or presence of the GST–Lyn SH3 domain, and treated as described in (A). The Figure is representative of four experiments performed in triplicate.

The p13–SFK complex localizes to the intermembrane space of mitochondria

To establish in which mitochondrial compartment p13 and SFKs were localized, RLM were incubated with p13 alone or in the presence of a slight molar excess of each tyrosine kinase available, and further processed as described below. First, treatment with proteinase K showed susceptibility of p13 and Lyn to the protease only when mitochondria were solubilized with Triton X-100 (Figure 4A), indicating that p13 alone or in association with Lyn was localized inside mitochondria. AIF, a structural component of the mitochondrial inner membrane, and aconitase, a mitochondrial matrix protein, were used as intramitochondrial markers. Secondly, after separation of mitochondrial membranes from the soluble fractions, as described in the Experimental section, p13 proved to be bound to the membrane fraction when incubated alone with mitochondria (Figure 4B, left-hand panel), whereas it was found in the soluble fraction when complexed with Lyn (Figure 4B, right-hand panel). Thirdly, treatment with increasing concentrations of digitonin resulted in a selective release of p13 depending on the absence or presence of Lyn. In particular, p13 was released from mitochondria, if previously incubated in the absence of Lyn, only at concentrations of digitonin higher than 0.4 mg/ml, similarly to AIF, which is suggestive of either a tight binding to the inner mitochondrial membrane or localization in the mitochondrial matrix; on the other hand, concentrations of digitonin as low as 0.1 mg/ml caused release of p13, if the latter was pre-incubated with Lyn, suggesting that this interaction prevented p13 from reaching the mitochondrial inner membrane, thus segregating it in the intermembrane space. Similar results were obtained using Src, the prototypical member of the Src-related group (see Supplementary Figure S2 at http://www.BiochemJ.org/bj/439/bj4390505add.htm).

Figure 4 p13 and Lyn are detected as a component of a soluble complex in the intermembrane space of isolated mitochondria

RLM were incubated with p13 only (left-hand panels) or in the presence of Lyn (right-hand panels) and treated as described below. (A) RLM were incubated in the absence or presence of proteinase K (PK) and Triton X-100 (TX-100) according to the Experimental section. Aliquots of the mitochondrial lysate were subjected to Western blot (Wb) analysis with anti-Lyn, anti-p13, anti-AIF and anti-aconitase antibodies as indicated. (B) RLM were subjected to treatment to obtain mitochondrial membranes and soluble fractions as described in the Experimental section. Intact mitochondria (Mit.), mitochondrial membranes (Membr.) and soluble fractions were subjected to Western blot analysis with anti-Lyn, anti-p13, anti-AIF and anti-aconitase antibodies as indicated. (C) RLM were incubated in the presence of increasing concentrations of digitonin. After each treatment, the samples were centrifuged to separate the pellet (P) from the soluble fraction (S). Aliquots of such fractions after differential digitonin treatment were subjected to Western blot analysis with anti-Lyn, anti-p13, anti-AIF and anti-aconitase antibodies as indicated. The Figure is representative of four independent experiments performed in triplicate.

Tyrosine phosphorylation in isolated mitochondria is induced by p13-mediated import of SFKs

To assess whether the association between SFKs and p13 could change Lyn's activity within mitochondria, as occurred in vitro, we incubated purified RLM in the presence of Lyn and p13, as described above, and evaluated the level of mitochondrial tyrosine phosphorylation by Western blot analysis with the anti-pTyr antibody.

Figure 5(A) shows a striking elevation in tyrosine phosphorylation of a number of mitochondrial proteins distributed over a wide range of molecular masses once Lyn translocated into mitochondria. The involvement of Lyn in this event was further confirmed by the effect of PP2, a specific SFK inhibitor (Figure 5A, lane 6 compared with lane 7), and its activation was confirmed by Western blot analysis with the anti-Src pY416 antibody. To establish whether the level of tyrosine phosphorylation observed inside mitochondria was determined by the stability of the p13–Lyn complex, mitochondria pre-incubated with p13 and Lyn in the above conditions were treated with 0.1 mg/ml digitonin. After centrifugation, the resulting supernatant was incubated in the presence of increasing concentrations of the GST–Lyn SH3 domain, subsequently subjected to immunoprecipitation with anti-p13 antibody and assayed for SFK activity. Figure 5(C) shows that the p13–Lyn interaction was still stable after translocation across the mitochondrial outer membrane, with a disruption of the complex occurring in a manner proportional to the concentration of the GST–Lyn SH3 domain (Figure 5B). Since the presence of the GST–Lyn SH3 domain blocked the stimulation of SFK activity due to p13 (Figure 5C), we presumed that the increase in SFK activity depended upon the stability of the p13–Lyn interaction. Similar findings were achieved for the other SFKs by examining Fyn, Fgr (results not shown) and Src (Supplementary Figure S3 at http://www.BiochemJ.org/bj/439/bj4390505add.htm).

Figure 5 p13-mediated import of Lyn increased tyrosine phosphorylation in isolated mitochondria

(A) RLM (50 μg) were incubated with p13 and/or Lyn in the absence or presence of 1 μM PP2. RLM were subsequently centrifuged and subjected to Western blot (Wb) analysis with anti-p-Tyr (top panel), anti-Src pY416 (middle panel) and anti-Lyn (bottom panel) antibodies. The molecular mass markers are indicated on the right-hand side in kDa. (B) RLM (250 μg) were incubated with p13 and Lyn. RLM were then spun down and aliquots of 50 μg of mitochondria were lysed and immunoprecipitated (Ip) with anti-p13 antibody in the absence or presence of increasing concentrations of GST–Lyn SH3. The immunocomplexes were probed with anti-Lyn and subsequently with anti-p13 antibodies. (C) RLM (250 μg) were incubated with p13 and Lyn. After a 5 min incubation, mitochondria were spun down and aliquots of 50 μg of mitochondria were analysed for in vitro Lyn activity in the absence or presence of increasing concentrations of GST–Lyn SH3. The Figure is representative of four independent experiments; results are means±S.D.

The interaction between p13 and SFKs impairs the ability of p13 to induce mitochondrial K+ influx

p13 is known to target the inner mitochondrial membrane, bringing about substantial structural and functional changes of these organelles, such as fragmentation in a cellular context as well as an increase in K+ permeability. These effects result in depolarization of the inner mitochondrial membrane and enhancement of respiratory chain activity with concurrent elevation in production of ROS (reactive oxygen species) in isolated mitochondria [34]. Interestingly, we observed that p13 together with SFKs via a SH3/proline-rich motif association in vitro formed a stable complex still capable of crossing the outer mitochondrial membrane of highly purified RLM, but that it was unable to insert itself into the inner mitochondrial membrane where it usually exerts its perturbing action (Figure 4). Therefore, we sought to verify whether the change in localization of p13 also affected the functional effects that it is known to provoke. For this purpose, RLM were tested for membrane potential (ΔΨm) before and after incubation with p13, the latter in turn mixed with either GST–Lyn SH3 or Lyn. Figure 6(A) shows that free p13 caused a sharp drop in ΔΨm, as expected in accordance with the previously documented increase in the permeability of the inner mitochondrial membrane to small cations [35]. Conversely, when p13 was complexed with Lyn, its effect on ΔΨm was dampened in a manner proportional to the molar ratios between the two proteins. In fact, decreasing p13/Lyn ratios as far as 1:1, as shown in Figure 6(A), progressively abolished the capability of p13 to alter ΔΨm. Similar effects were obtained by using the GST–Lyn SH3 domain as a ligand for p13 (Figure 6B), indicating a role for the SH3 domain in affecting both function and localization of p13 itself. These findings also suggest a possible non-catalytic function of SFKs mediated by their SH3 domain.

Figure 6 Interaction between the SH3 domain of Lyn and p13 proline-rich motif impairs p13 ability to induce a collapse in the inner mitochondrial potential (ΔΨm)

RLM were incubated in 200 mM sucrose, 10 mM Hepes, pH 7.4, 5 mM succinate, 1.25 μM rotenone, 1 mM sodium phosphate and protease inhibitors in the presence of p13 only or with increasing concentrations of Lyn (A) or GST–Lyn SH3 domain (B). The downward arrows indicate the time point at which p13 or p13 complex was added. The Figure is representative of four independent experiments.

The p13–Lyn complex enters mitochondria in p13-transfected HeLa cells

Once we verified that p13 interacted with SFKs to form a complex capable of translocating to isolated mitochondria, we investigated the presence and the possible localization of such a complex in living cells, using Lyn- and p13-overexpressing cells. Among SFKs, we monitored Lyn expression because it has already been shown to translocate from the plasma membrane to mitochondria [33]. Therefore HeLa cells were transfected with expression vectors for Lyn or p13 singly or in combination, or both empty vectors as a control, in the absence or in the presence of PP2, as described in the Experimental section. We examined the protein level of p13 and Lyn by Western blot analysis on the total cell lysate and on highly purified mitochondria obtained by protocols described previously [33]. Figure 7(A) shows an increase in tyrosine phosphorylation in the whole cell lysate (left-hand panel) upon transfection of Lyn (lane 3), as expected, and of p13 (lane 5), indicating that the latter was capable of eliciting an endogenous SFK-dependent kinase activity, as elucidated by the inhibitory effect of PP2 (lanes 4 and 6). This effect was dramatically enhanced by the co-transfection of both proteins (lanes 3 and 5 compared with lane 7), further supporting the model of interaction inferred from the in vitro data. Notably, Lyn, both endogenous and transfected, was detected inside purified mitochondria only when p13 was present (Figure 7A, right-hand panel, lanes 5–8 compared with lanes 1–4), clearly paralleling the level of mitochondrial tyrosine phosphorylation (lane 5 compared with lane 1 and lane 7 compared with lane 3). The relative purity of the mitochondrial fraction was assessed by using antibodies against specific cellular markers.

Figure 7 The p13–Lyn complex localizes in mitochondria of p13-transfected HeLa cells

(A) HeLa cells were transfected with expression vectors for Lyn (pCMV6-XL4/Lyn, lanes 3–4 and 7–8), p13 (pcDNA3.1/myc-His C/p13, lanes 5–8), or with both empty vectors (pCMV6-XL4 and pcDNA3.1/myc-His C, lanes 1 and 2 respectively). At 36 h after transfection, cells were treated with 10 μM PP2 (lanes 2, 4, 6 and 8) for 12 h. Whole cell lysates (left-hand panels) and whole mitochondrial lysates (right-hand panels) were assayed by Western blot (Wb) analysis with anti-Lyn, anti-p13, anti-p-Tyr, anti-LDH (cytosolic marker), anti-calnexin (microsomal marker), anti-lamin (nuclear marker) and anti-aconitase (mitochondrial marker) antibodies as indicated. (B) Mitochondrial lysates from HeLa cells transfected with expression vectors for Lyn and p13 (lanes 3 and 4), or with both empty vectors (lanes 1 and 2) were analysed for in vitro Lyn activity on SFK-specific peptide substrate cdc26–20 (right-hand panel) or were immunoprecipitated with anti-p13 antibody (left-hand panel) in the absence (lanes 1 and 3) or presence (lanes 2 and 4) of GST–Lyn SH3. The immunocomplexes were probed with anti-Lyn antibody and subsequently with anti-p13 antibody. The Figure is representative of three independent experiments.

To examine whether the interaction between Lyn and p13 occurred according to the mechanisms previously identified in vitro and whether a stable complex persisted within mitochondria in cultured cells, p13 was immunoprecipitated from the mitochondrial lysates of HeLa cells transfected with expression vectors for Lyn and p13, or with both empty vectors. As shown in Figure 7(B), Lyn co-immunoprecipitated with p13 and the GST–Lyn SH3 domain was able to disrupt the p13–Lyn complex (left-hand panel, lane 4 compared with lane 3), highlighting the role of Lyn's SH3 domain in the interaction with p13. Moreover, the same mitochondrial lysates were assayed for SFK activity in the absence or presence of the GST–Lyn SH3 domain, demonstrating that Lyn in mitochondria displayed a higher activity when associated with p13, further confirming the activatory role of p13 on SFKs (right-hand panel, column 3 compared with column 1).

To corroborate the biochemical data obtained in HeLa cells with regard to the p13–Lyn interaction and import of Lyn into mitochondria, the subcellular localization of Lyn was analysed in relation with p13 expression by confocal microscopy. Since the anti-p13 antibody was useless in the immunofluorescence assay, we used the anti-c-Myc antibody to follow p13 cellular localization. The pcDNA3.1 expression vector for p13 also contains a Myc tag. As shown in Figure 8, we observed that p13 mainly localizes to mitochondria, as visualized by double immunofluorescence staining with the anti-TOM20 antibody as mitochondrial marker and the anti-c-Myc antibody, and that its subcellular distribution was unaltered, even upon Lyn co-transfection (Figure 8A, merge). In contrast, overexpressed Lyn was shown to co-localize with TOM20 only when p13 was expressed, indicating that p13 was the driving force for Lyn to be recruited to mitochondria (Figure 8B). Finally, as shown in Figure 8(C), p13 and Lyn widely overlap when co-expressed in HeLa cells, supporting the association of the two proteins in vivo.

Figure 8 Lyn localizes to the mitochondria of HeLa cells only upon co-expression of p13

HeLa cells were transfected with expression vectors for Lyn (pCMV6-XL4/Lyn), p13 (pcDNA3.1/myc-His C/p13), or with both empty vectors. At 48 h post-transfection, cells were fixed, permeabilized and incubated with the indicated antibodies. (A) Cells transfected with p13 only or Lyn and p13 were incubated with anti-c-Myc and anti-TOM20 R (mitochondrial marker) antibodies followed by FITC-conjugated anti-mouse secondary antibody (green) and with TRITC-conjugated anti-rabbit secondary antibody (red) respectively. (B) Cells transfected with Lyn only or Lyn and p13 were incubated with anti-Lyn and anti-TOM20 M antibodies, followed by incubation with TRITC-conjugated anti-rabbit secondary antibody (red) and with FITC-conjugated anti-mouse secondary antibody (green) respectively. (C) Cells transfected with expression vectors for Lyn, p13 or Lyn/p13 were probed with anti-Lyn and anti-c-Myc antibodies, followed by the appropriate secondary antibodies, as described above. Co-localization is visualized by the yellow fluorescence appearing after merging of both signals. The Figure is representative of three independent experiments. Scale bar, 10 μm.

DISCUSSION

In the present study, we demonstrate that p13, a HTLV-1 accessory protein containing an N-terminal MLS, associates with the SH3 domain of SFKs through its C-terminal proline-rich motif, forming a stable heterodimer that migrates into mitochondria and confers novel functional properties to the single components of the complex. It is noted that mitochondria have only been described as a location for SFKs within the last ten years [7]. In fact, although SFKs have commonly been thought to be located at the inner leaflet of the plasma membrane to relay extracellular cues from different types of transmembrane receptors [13], accumulating evidence has been reported for a wider intracellular distribution of this family of protein kinases as well as for peculiar functions of SFKs associated with their distinct intracellular localizations [49]. In this regard, post-translational modifications such as palmitoylation, and lately the redox state, have been recognized as crucial factors in driving trafficking and localization of SFKs. In particular, three modes of SFK trafficking have been proposed according to the palmitoylation state: (i) the cycling pathway between plasma membrane and endosomes/lysosomes for non-palmitoylated SFKs [8,40,41]; (ii) the secretory pathway from the Golgi apparatus to the plasma membrane for mono-palmitoylated SFK [8,42]; and (iii) the direct plasma membrane-targeting pathway for doubly palmitoylated SFKs [8]. The redox state has been recently demonstrated to profoundly regulate the trafficking of non-palmitoylated Src as well, with a remarkable shift of this SFK from the plasma membrane to endosomes/lysosomes accompanied by a decrease in enzyme activity under reducing conditions [36]. In addition to post-translational modifications, protein–protein interactions due to the multimodular structure of SFKs have also been implicated in their targeting to specific subcellular compartments. Accumulating data show that the SH3 domain, besides modulating the SFK activity by binding the PPII motif of SFKs themselves, is directly implicated in the localization of SFKs to cellular compartments other than the plasma membrane by recognizing specific ligands. An example of this mode of trafficking is given by the interaction between Hck, a SFK highly expressed in macrophages, and Nef, a HIV-1 multifunctional protein, resulting in the activation of Hck and its accumulation to the Golgi apparatus, thereby causing the arrest of the maturation of the nascent cytokine receptor Fms [31].

Our results from the present study corroborate the notion that the SH3 domain of SFKs is one of the players involved in directing SFKs themselves to final destinations upon the basis of localization signals present on interacting partners. In fact, p13, known to enter mitochondria through its MLS, behaves analogously to Nef, by virtue of its C-terminal proline-rich motif theoretically suitable for binding SH3 domains. In the present study, we demonstrate for the first time that p13 associates indifferently with distinct members of the Src family (Src, Fyn, Fgr and Lyn) via SH3 domain–proline-rich motif interaction in vitro, hence forming a heterodimer in which the activity of all the SFKs tested are dramatically enhanced (Figures 1 and 2). Importantly, p13 conserves its ability to translocate into isolated mitochondria, concomitantly participating in a stable complex with SFKs, ultimately carrying them into the mitochondrial intermembrane space. Likewise, both biochemical analysis and confocal microscopy provide compelling evidence that Lyn is also detected inside mitochondria when p13 is co-expressed in HeLa cells, and is associated with p13 (Figures 7 and 8). This data supports the hypothesis that p13 may act as a carrier for SFKs into mitochondria, even in living cells. Moreover, the striking elevation in tyrosine phosphorylation observed in either freshly isolated mitochondria in the presence of synthetic p13 or in p13-overexpressing HeLa cells highlights the role of SFKs, as shown by the use of the specific SFK inhibitor PP2 (Figures 5A and 7A), confirming the ability of p13 to function as positive modulator of SFKs.

Notably, we observed a similar scenario while investigating the role of SFKs, and in particular of Lyn, in the context of early liver regeneration, assessing that Lyn translocates from the plasma membrane to mitochondria, more precisely into the mitochondrial intermembrane space, where it is part of a 230 kDa multiprotein complex [33]. It has been demonstrated only recently that this complex, similar to the results shown in the present study, is disrupted by the GST–Lyn SH3 domain, in agreement with the model proposed. Furthermore, we also showed that Lyn, as part of the multiprotein complex, protected the structural and functional integrity as well as the bioenergetic competence of mitochondria by contrasting the potentially harmful effects resulting from ROS elevation and Ca2+ overload, which would fatally lead to cell death (E. Tibaldi, M. A. Pagano and A. M. Brunati, unpublished work).

An experimental setting that exploits p13 as a Trojan horse for SFKs might provide a useful model to elucidate the role of SFKs themselves within mitochondria and how tyrosine phosphorylation influences mitochondrial physiology. In this regard, it is worthwhile to remember that functional studies on isolated mitochondria with full-length synthetic p13 showed that this protein inserts itself into the inner mitochondrial membrane, thereby inducing an inward K+ current with a drop in ΔΨm and activation of the electron transport chain accompanied by increased mitochondrial ROS production [34,35]. In contrast, as highlighted by our results, the SFK–p13 complex is still taken up by these organelles, but p13 is unable to target the inner mitochondrial membrane and to perturb ΔΨm. These observations are in agreement with studies that recognize a scaffolding role for SFKs mediated by the SH3 domain, in our hands resulting in impairment of the activity of p13, indicating a non-catalytic function of SFKs [4246]. Importantly, we note that this property of SFKs does not prove detrimental to their catalytic action, which instead was remarkably increased upon interaction with p13, suggesting that the catalytic and non-catalytic function might not be mutually exclusive, as is sometimes reported in the literature [46], but might even be independent or synergistic, and potentially takes part in pathophysiological conditions.

In summary, the SFK–p13 interaction and the resulting effects described above lead us to hypothesize a mechanism by which the transfer of SFKs to different cellular districts, mediated by the SH3 domain and directed by the localization signals of interacting proteins, may exert a more general action on cell fate by modifying the phosphorylation state and hence the function of key substrates resident in the targeted cellular compartments.

AUTHOR CONTRIBUTION

Elena Tibaldi performed the majority of the research, analysed the data and wrote some of the manuscript. Andrea Venerando synthesized and purified the peptides used in this work and reviewed the manuscript prior to submission. Francesca Zonta performed in vitro research. Carlo Bidoia actively contributed to the setting up of genetic assays. Elisa Magrin performed some of the in vitro research. Oriano Marin purified the antibody against p13. Antonio Toninello provided cultural background and a location for assays on mitochondria. Luciana Bordin performed some of the in vitro research. Veronica Martini performed the immunofluorescence assays. Mario Angelo Pagano supported the work intellectually, reviewed all of the data and wrote the manuscript. Anna Maria Brunati designed the research, reviewed all of the data, participated in analysis of data and wrote the manuscript.

FUNDING

This work was funded by Fondazione CARIPARO (Progetto di Eccellenza 2008) (to A.M.B.).

Acknowledgments

We thank Luc Willems (Gembloux, Belgium) for kindly providing pcDNA3.1/myc-His C/p13.

Abbreviations: AIF, apoptosis-inducing factor; CSK, c-Src tyrosine kinase; GST, glutathione transferase; HRP, horseradish peroxidase; HTLV-1, human T-cell leukaemia virus type 1; LDH, lactate dehydrogenase; MLS, mitochondrial localization sequence; PP2, 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine; PPII, motif, polyproline type II motif; p-Tyr, phosphotyrosine; RLM, rat liver mitochondria; ROS, reactive oxygen species; SFK, Src family kinase; SH, domain, Src homology domain; TBS, Tris-buffered saline; TPP+, tetraphenylphosphonium; TRITC, tetramethylrhodamine β-isothiocyanate

References

View Abstract