Src Inhibitors Modulate Frataxin Protein Levels

Human Molecular Genetics, 2015, Vol. 24, No. 15 4296–4305

doi: 10.1093/hmg/ddv162 Advance Access Publication Date: 6 May 2015 Original Article

ORIGINAL ARTICLE

Src inhibitors modulate frataxin protein levels Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021 Fabio Cherubini1, Dario Serio1, Ilaria Guccini1, Silvia Fortuni1,2, Gaetano Arcuri1, Ivano Condò1, Alessandra Ruﬁni1,2, Shadman Moiz1, Serena Camerini3, Marco Crescenzi3, Roberto Testi1,2 and Florence Malisan1,*

1Laboratory of Signal Transduction, Department of Biomedicine and Prevention, University of Rome ‘Tor Vergata’, Via Montpellier 1, 00133 Rome, Italy, 2Fratagene Therapeutics Ltd, 22 Northumberland Rd, Dublin, Ireland and 3Department of Cell Biology and Neurosciences, Italian National Institute of Health, Viale Regina Elena, 299, 00161 Rome, Italy

*To whom correspondence should be addressed at: Laboratory of Signal Transduction, Department of Biomedicine and Prevention, University of Rome ‘Tor Vergata’, Via Montpellier 1, 00133 Rome, Italy. Tel: +39 0672596501; Fax: +39 0672596505; Email: [email protected]

Abstract Defective expression of frataxin is responsible for the inherited, progressive degenerative disease Friedreich’s Ataxia (FRDA). There is currently no effective approved treatment for FRDA and patients die prematurely. Defective frataxin expression causes critical metabolic changes, including redox imbalance and ATP deﬁciency. As these alterations are known to regulate the tyrosine kinase Src, we investigated whether Src might in turn affect frataxin expression. We found that frataxin can be phosphorylated by Src. Phosphorylation occurs primarily on Y118 and promotes frataxin ubiquitination, a signal for degradation. Accordingly, Src inhibitors induce accumulation of frataxin but are ineffective on a non-phosphorylatable frataxin- Y118F mutant. Importantly, all the Src inhibitors tested, some of them already in the clinic, increase frataxin expression and rescue the aconitase defect in frataxin-deﬁcient cells derived from FRDA patients. Thus, Src inhibitors emerge as a new class of drugs able to promote frataxin accumulation, suggesting their possible use as therapeutics in FRDA.

of dysregulated mitochondrial metabolism, frataxin-defective Introduction cells have indeed reduced activity of iron sulfur cluster (ISC)-con- Friedreich’s Ataxia (FRDA) is an autosomal recessive disorder taining enzymes, a general imbalance in intracellular iron distri- characterized by progressive degeneration of the central and per- bution, reduced ATP content and increased sensitivity to ipheral nervous systems, cardiomyopathy and increased inci- oxidative stress with increased ROS generation. Low frataxin le- dence of diabetes mellitus. FRDA is caused by homozygous vels and disease severity have been correlated (5). Moreover, fra- hyperexpansion of GAA triplets in intron 1 of the frataxin gene taxin levels are not only crucial for cell survival but also for stress on chromosome 9q21 (1). Pathological GAA expansions severely handling responses (6–11). There is no current successful treat- reduce transcription of the FXN gene. Frataxin is an extremely ment, but considering that the frataxin coding sequence is intact conserved mitochondrial protein synthesized as a cytosolic 210 in most of the patients, therapies aiming at enhancing frataxin amino acid precursor, which is then imported into mitochondria levels are now being considered (11–15). Frataxin protein levels following a two-step proteolytic maturation by a mitochondrial are controlled by the proteasome upon ubiquitination of target processing peptidase (2,3). Low levels of expression of frataxin residue, K147, on frataxin (16). This lysine represents a crucial are responsible for all clinical and morphological manifestations site for frataxin stability because a frataxin mutant lacking of FRDA (4). In fact, frataxin deﬁciency in humans critically K147 cannot be ubiquitinated and is more stable. Therefore, pre- affects survival of large primary neurons of the dorsal root gan- venting ubiquitination on K147 is expected to grant frataxin an glia, cardiomyocytes and pancreatic β-cells. As a consequence increased stability and a prolonged half-life (16). Ubiquitination

4296 Human Molecular Genetics, 2015, Vol. 24, No. 15 | 4297

and phosphorylation are post-translational modiﬁcations (PTM) that often interact dictating the fate of proteins (17). In addition, PTM have emerged as powerful modulators of metabolic path- ways (18) and are important regulators of mitochondrial functions (19). Moreover, Src tyrosine kinase family members such as Lyn, Fgr, Fyn and c-Src have also been reported to associate with mitochondria (19). Considering that Src tyrosine kinases can be regulated by a variety of important mitochondrial signals such as ATP levels and redox state (20), which are indeed dysregulated in FRDA, we sought to investigate whether frataxin levels could in turn be modulated by phosphorylation. In this study, we report that frataxin is phosphorylated on Y118 by Src kinase. Interestingly, non-phosphorylatable frataxin-Y118F mutant is Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021 less ubiquitinated and blocking Src activity with speciﬁc inhibitors increases frataxin levels. Accordingly, Src inhibitors are ineffective in human cells in which a frataxin-Y118F mutant was expressed. Moreover, Src inhibitors induce frataxin expression and rescue the aconitase defect in cells derived from FRDA patients. Therefore, Src inhibitors can promote frataxin accumulation in living cells, strongly supporting their potential therapeutic application.

Results Src kinase triggers frataxin phosphorylation

To assess whether frataxin could be a substrate for non-receptor tyrosine kinases, frataxin was transiently transfected in human embryo kidney (HEK) 293 cells, together with several constructs encoding different forms of Src and Abl kinases. The constitutively active Src, SrcY527F, but not its inactive kinase counterpart, SrcY527FKin−, caused retarded frataxin precursor electrophoret- ic migration as shown by immunoblotting (Fig. 1A). To address whether this shift migration was due to precursor phosphorylation, phosphatase assay on total lysates was performed. Follow- ing phosphatase treatment, the shifted form disappeared, indicating that the frataxin precursor is indeed phosphorylated in the cells (Fig. 1B). To further validate frataxin precursor phosphorylation, immunoprecipitation experiments were performed (Fig. 1C). Interestingly, the constitutively active mutant of c-Abl, Abl-PP, could not phosphorylate frataxin precursor suggesting that this form of frataxin is speciﬁcally phosphorylated by Src.

Src kinase phosphorylates frataxin on residue Y118 Figure 1. Src kinase phosphorylates frataxin. (A) HEK293 cells were transiently To identify Src-induced tyrosine phosphorylated site(s) on fra- transfected with frataxin (FXN), and either constitutively active Src (Y527F) or its − taxin, single non-phosphorylatable mutants of the eight tyro- kinase-inactive counterpart (Y527F-Kin ). Total protein extracts (TOT) were sines residues of frataxin (Y74, Y95, Y118, Y123, Y143, Y166, separated by SDS–PAGE and immunoblotted (WB) with specificantibodiesagainst Y175 and Y205) were generated converting each tyrosine into frataxin and tubulin (TUB) as loading control. Data are representative of ten phenylalanine residue. Mutants were analyzed in cotransfection independent experiments. (B) Total lysate of HEK293 transfected with frataxin and constitutively active Src (Y527F) was incubated for 50 min at 37°C with buffer alone, assays with constitutively active SrcY527F or its inactive kinase − CIP phosphatase (PPase) in the presence or absence of phosphatase inhibitors (Inh) counterpart SrcY527FKin as described in Figure 1. Interestingly, and analyzed after separation by SDS–PAGE by immunoblotting (WB) with specific mutation of residues Y95, Y118 and Y123 induced an electrophor- antibody against frataxin (FXN). Data are representative of three independent etic shift migration of all the frataxin forms (Fig. 2A). Though mu- experiments. (C) HEK293 cells were transiently transfected with frataxin, and either − tations are conservative, the shift migration could be due to constitutively active Src (Y527F), its kinase-inactive counterpart (Y527F-Kin )or charges modifications because the α-helix 1 (D91 to A114) is an constitutively active Abl (Abl-PP). The kinase activity of Abl-PP was indeed controlled (data not shown). Total protein extracts (TOT) or immunoprecipitated acidic residue-rich region and the loop 1 (D115 to Y123) is import- frataxin (IP) were separated by SDS–PAGE and immunoblotted (WB) with specific ant for proper protein folding and stability (21). antibodies against frataxin (FXN), phosphorylated tyrosine (pY) and tubulin (TUB) Figure 2A shows three representative mutants, out of the as loading control. Data are representative of three independent experiments. The eight analyzed, whereas the table summarizes the results for precursor (P), intermediate (I) and mature (M) frataxin forms are indicated. The all mutants. Only mutation of Y118 abrogated tyrosine phosphor- arrows show the phosphorylated shifted precursor form. ylation of frataxin precursor, indicating that Y118 is the main Src phosphorylation site. To further confirm that Y118 was the gel after visualization by Coomassie staining, digested with main phosphorylation site, mass spectrometry was performed. chymotrypsin and analyzed by LC–MS/MS. A phosphopeptide, – In vitro phosphorylation reaction using recombinant frataxin1 210 corresponding to amino acids 96–123, was isolated, identifying and recombinant Src was performed, recovered on SDS–PAGE the phosphorylated residue as Y118 (Fig. 2B). 4298 | Human Molecular Genetics, 2015, Vol. 24, No. 15 Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021

Figure 2. Src kinase phosphorylates frataxin on residue Y118. (A) HEK293 cells were transiently transfected with wild-type frataxin (FXN), or non-phosphorylatable − frataxin mutants Y175F, Y118F, Y123F and either constitutively active Src (Y527F) or its kinase-inactive counterpart (Y527F-Kin ). Total protein extracts (TOT) or immunoprecipitated frataxin (IP) were separated by SDS–PAGE and immunoblotted (WB) with speciﬁc antibodies against frataxin, phosphorylated tyrosine (pY) and tubulin (TUB) as loading control. Data are representative of three independent experiments. The precursor (P), intermediate (I) and mature (M) frataxin forms are indicated. (B) The picture shows the fragment ion spectrum of the phosphopeptide (96–123), with sequence ERLAEETLDSLAEFFEDLADKPpYTFEDY and triple-charged precursor ion at m/z 1146.7, obtained after in-gel digestion of frataxin by chymotrypsin. The fragment ion series are indicated on the sequence: b and y ions, which are detected in the spectrum, are in bold dark. The upper panel shows an enlarged region of the spectrum (m/z range from 1200 to 1440): here, the distance from the peaks corresponding to b22 to b23 ions coincides with a phosphotyrosine (pY) locating the phosphorylation on Y118.

or non-phosphorylatable frataxin mutants Y118F, Y166F and Non-phosphorylatable Y118F frataxin mutant is less Y175F together with hemagglutinin (HA)-tagged ubiquitin (HA- ubiquitinated Ub) in the absence or presence of proteasome inhibitor MG132. To evaluate the impact of phosphorylation on frataxin ubiquiti- As previously reported (16), frataxin ubiquitination status was nation, HEK-293 cells were transfected with wild-type frataxin evaluated by immunoblotting with anti-frataxin monoclonal Human Molecular Genetics, 2015, Vol. 24, No. 15 | 4299 Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021

Figure 3. Phosphorylation on Y118 promotes ubiquitination. (A) HEK293 cells were transiently transfected with wild-type frataxin (FXN), or non-phosphorylatable mutants Y118F, Y166F, Y175F and hemagglutinin (HA)-tagged ubiquitin (HA-Ub). Total protein extracts (TOT) or immunoprecipitated ubiquitinated frataxin (IP α-HA) were separated by SDS–PAGE and immunoblotted (WB) with specific antibodies against frataxin (FXN) and tubulin (TUB) as loading control. Data are representative of five independent experiments. (B) The graph illustrates the relative ubiquitination level quantitated as the ratio between monoubiquitinated frataxin (Mono-Ub) level versus the frataxin precursor expression in the MG132-treated lanes. The precursor (P), intermediate (I) and mature (M) frataxin forms are indicated. antibodies on total lysates or after immunoprecipitation of ubi- the Src inhibitors tested could increase frataxin levels in cells quitinated forms with anti-HA antibody only in the presence of overexpressing wild-type frataxin, but not in cells overexpressing MG132. Frataxin monoubiquitinated forms can be detected as the non-phosphorylatable Y118F frataxin mutant (Fig. 4), sug- slower migrating bands above the frataxin precursor. Figure 3A il- gesting that they indeed act by inhibiting phosphorylation of lustrates that accumulation of ubiquitinated frataxin forms was Y118. In addition, these inhibitors are also effective on the levels reduced when non-phosphorylatable Y118F mutant, but not of endogenous frataxin as shown in HEK293 cells (Fig. 5). other tyrosine mutants such as Y166F and Y175F, was transfected. Relative ubiquitination level of frataxin was quantitated Src inhibitors promote frataxin accumulation in as the ratio between monoubiquitinated forms versus the non- frataxin-deficient cells and rescue the aconitase defect ubiquitinated precursor forms in the MG132-treated lanes (Fig. 3B). Non-phosphorylatable Y118F frataxin mutant is >60% As blocking Src activity with tyrosine kinase inhibitors increases less ubiquitinated compared with wild-type frataxin, thus suggest- frataxin levels in human cells, we therefore tested their efficacy ing that phosphorylation on Y118 is required for ubiquitination. on frataxin-deficient cells such as FRDA patient-derived B cells (GM16214). Among the different Src inhibitors, SU6656, PP2, and Dasatinib seemed to be best tolerated. FRDA lymphoblasts were Src inhibitors increase wild-type frataxin expression but exposed to these inhibitors for different time periods. As illu- not non-phosphorylatable Y118F frataxin mutant strated in Figure 6A, the upregulation of frataxin could be de- Considering that phosphorylation on Y118 promotes ubiquitina- tected as early as 24h of treatment and could be further tion and that preventing ubiquitin-dependent degradation in- accumulated within 72 h, particularly with SU6656 and Dasati- creases frataxin levels (16), we sought to investigate whether nib. Moreover, Dasatinib could upregulate frataxin in primary inhibiting frataxin phosphorylation by Src kinase would indeed fibroblasts (GM04078) derived from a FRDA patient in a dose- allow enhancement of frataxin expression. Different Src inhibi- dependent manner (Fig. 6B). Furthermore, aconitase activity tors such as PP2, SU6656, Saracatinib, Bosutinib and Dasatinib was also measured in FRDA lymphoblasts exposed to SU6656, were therefore used to treat HEK293FXN cells, stably expressing PP2 and Dasatinib inhibitors for different time periods. Interest- a single copy of wild-type frataxin or the non-phosphorylatable ingly, the aconitase defect caused by frataxin deficiency could Y118F frataxin mutant. As illustrated in Figure 4, all the Src inhi- be restored in 48 h and further rescued in 72 h by all the Src inhibitors can promote frataxin accumulation within 24 h in a dose- bitors tested (Fig. 7). dependent manner, although with different efficacy. Among the Src inhibitors tested, Dasatinib appeared to be the most efficient Discussion in promoting frataxin accumulation, being still active in the na- nomolar range of concentrations (data not shown). Interestingly, This study shows that frataxin precursor can be phosphorylated by frataxin accumulation was observed for all the different process- Src kinase on Y118, promoting its ubiquitination and degradation. ing forms such as precursor, intermediate and mature forms. All Moreover, Src inhibitors increase frataxin expression in human 4300 | Human Molecular Genetics, 2015, Vol. 24, No. 15 Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021

Figure 4. Src inhibitors upregulate wild-type frataxin but not the non-phosphorylatable Y118F frataxin mutant. HEK293FXN cells stably expressing single copy of wild-type frataxin (WT) or non-phosphorylatable Y118F frataxin mutant (Y118F) were treated for 24 h with 1, 3 and 10 μ of either Src inhibitor SU6656, PP2, Dasatinib, Bosutinib, Saracatinib or vehicle (−). Left panels: frataxin (FXN) and tubulin expression (TUB) was analyzed by western blot. Data are representative of three independent experiments. The precursor (P), intermediate (I) and mature (M) frataxin forms are indicated. Right panels: densitometric quantiﬁcation of frataxin accumulation. Frataxin expression was normalized with tubulin and frataxin expression in non-treated cells (−) set to one. Data represent the mean ± 1 S.E.M. from three different independent experiments performed for each inhibitor. P-values were calculated with Student’s t-test and were statistically signiﬁcant (*P <0.05;**P < 0.01) for each treatment compared with non-treated conditions. cells. Accordingly, Src inhibitors failed to upregulate frataxin in iron chelators with controversial results (23,24). Other strategies human cells in which a frataxin-Y118F mutant was expressed. aiming at improving mitochondrial function or restoring frataxin More importantly, Src inhibitors are effective in FRDA cells as levels are being developed (4,15). We recently provided evidence they promote frataxin accumulation and rescue aconitase defect. that preventing frataxin ubiquitination and degradation effect- Reduced frataxin expression, owing to abnormal GAA triplet ively results in frataxin accumulation in cells derived from expansion, gives rise to the degenerative disorder FRDA, a debili- FRDA patients (16). Modulating frataxin ubiquitination therefore tating disease that leads to a premature death of the patients. To represents an attractive therapeutic strategy. date, there is no approved treatment for FRDA (22). Main current Ubiquitination frequently interplays with phosphorylation to therapeutic approaches were based on the use of anti-oxidants or regulate many important signaling processes (17,25). Both serine/ Human Molecular Genetics, 2015, Vol. 24, No. 15 | 4301

to tightly tune crucial cellular processes. Here, we show that Src-mediated phosphorylation of frataxin promotes its ubiquitination. This prompted us to test whether Src inhibitors could modulate frataxin levels. We found that indeed all Src inhibitors tested promote frataxin accumulation and rescue aconitase defect in cells derived from FRDA patients. Src inhibitors might therefore have therapeutic potential in FRDA. Src inhibitors Dasatinib and Bosutinib, reported to cross the blood–brain barrier (28,29), have been approved for therapeutic use in humans (30,31) for the treatment of Philadelphia chromosome positive (Ph+) chronic myelogenous leukemia. Similarly to Dasatinib, Bosutinib is an ATP competitive Abl and Src inhibitor (32). However, chemical proteomics approaches have identified Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021 over 45 target kinases for Bosutinib (33) as well as Dasatinib (34). Interestingly, tyrosine kinases have emerged as powerful modulators of mitochondrial functions (18,19). Moreover, Src tyrosine kinases can actually be modulated by redox status and ATP levels that are indeed altered in FRDA patients (20). Considering the conserved domain structure of the Src family kinases (SFK) (35), it would be interesting to therefore study whether other SFK members could also specifically phosphorylate frataxin, especially Lyn, Fgr and Fyn described to associate with mitochondria (19). Further structural studies addressed to analyze frataxin and SFK binding would also help to develop more effective inhibitors blocking specific interaction between frataxin and SFK in order to improve efficacy and eventually circumvent possible toxicity.

Figure 5. Src inhibitors promote endogenous frataxin accumulation in HEK293 In summary, our results reveal a novel molecular mechanism cells. Human HEK293 cells were treated with 1, 3 and 10 μ of SU6656, PP2 and directly controlling frataxin protein levels in living cells. More- Dasatinib Src inhibitor or vehicle (−) for 24 h. Frataxin (FXN) and tubulin over, they provide the rationale for testing a whole class of expression (TUB) was analyzed by western blot. Data are representative of three drugs, some of them already available on the market, as thera- independent experiments. The precursor (P), intermediate (I) and mature (M) peutics for FRDA. frataxin forms are indicated. Right panels: densitometric quantification of frataxin accumulation. Frataxin expression was normalized with tubulin and frataxin expression in non-treated cells (−) set to one. Data represent the Material and Methods mean ± 1 S.E.M. from three different independent experiments performed for each inhibitor in the left panels. P-values were calculated with Student’s t-test Cell cultures and transfections and were statistically significant (*P < 0.05; **P < 0.01) for each treatment compared with non-treated conditions. Human embryonic kidney HEK-293 cells were cultured in Dulbec- co’smodified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS). HEK-293 cells were transfected with threonine and tyrosine phosphorylation of protein substrates the calcium/phosphate precipitation method, using 20 μgof can indeed promote or inhibit ubiquitination, in different man- total DNA (5 or 10 μg of pIRES2-Frataxin, 10 μgoftyrosine-mu- ners, leading to either proteasomal degradation or processing, tated non-phosphorylatable frataxin mutants, 10 μgofSrcor or the regulation of intracellular compartmentalization. Phos- Abl constructs and 10 μg of HA-Ub, or the corresponding empty phorylation can, for instance, promote ubiquitination by creating vectors) on 10-cm dishes. a phosphodegron, a recognition signal for some E3 ligases. Many HEK-293 Flp-In cells (Life Technologies) are HEK-293 variants of these short motifs are serine/threonine phosphorylation- allowing the stable and isogenic integration and expression of a dependent ubiquitination targets. Phosphorylated tyrosine in transfected gene. The HEK-293 clone stably expressing fratax- target proteins are also specifically recognized by SH2 domain- – in1 210 (2) and the clone stably expressing frataxinY118F were gen- containingE3ligasessuchasCblfamilymembers.Cblsare erated by mutagenesis as described later. HEK-293 Flp-In cells RING ubiquitin ligases that modulate receptor tyrosine kinases were maintained in DMEM supplemented with 10% FBS. (RTKs) by binding to phosphotyrosine residues on activated Immortalized GM16214 lymphoblasts, from a clinically af- RTK such as the epidermal growth factor receptor. This promotes fected FRDA patient, and immortalized GM16215 lymphoblasts ligand-dependent ubiquitination of these receptors, targeting from an unaffected carrier parent were obtained from NIGMS them for degradation. In addition, the phosphorylation of a crit- Human Genetic Cell Repository, Coriell Institute for Medical Re- ical tyrosine residue of Cbl stimulates its E3 ligase activity (26). search (Camden, NJ, USA) and cultured in RPMI 1640 supplemen- Furthermore, non-receptor tyrosine kinase activity can also be ted with 15% FBS. Primary fibroblasts GM04078 from a clinically regulated by ubiquitination. One example is given by Cbl- affected FRDA patient were obtained from NIGMS Human Genetic mediated ubiquitination of Src kinases that leads to their degrad- Cell Repository, Coriell Institute for Medical Research and cultured ation and results in the attenuation of antigen receptor signals, in DMEM supplemented with 15% FBS. thus controlling immune responses (27). Conversely, different kinases can also phosphorylate a target E3 ligase modulating its ac- DNA constructs tivity: the activity of the ubiquitin ligase ITCH is negatively regulated by Src-kinase family member Fyn but positively regu- The construct pIRES2-Frataxin1-210 contains human frataxin cDNA lated by JNK-mediated serine/threonine phosphorylation (25). cloned into the bicistronic expression vector pIRES2-EGFP (Clontech Therefore, phosphorylation and ubiquitination often crosstalk Laboratories) and was previously generated in this laboratory (6). 4302 | Human Molecular Genetics, 2015, Vol. 24, No. 15 Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021

Figure 6. Src inhibitors promote frataxin accumulation in frataxin-deficient cells. (A). FRDA patient-derived B cells were treated with 10 μ of either Src inhibitor SU6656, PP2, Dasatinib or vehicle (−) for the time indicated. Left panels: mature frataxin (FXN) and tubulin expression (TUB) were analyzed by western blot. Data are representative of three independent experiments. Right panels: densitometric quantification of frataxin accumulation. Frataxin expression was normalized with tubulin and frataxin expression in non-treated cells (NT) set to one. Data represent the mean ± 1 S.E.M. from three different independent experiments performed for each inhibitor in the left panels. P-values were calculated with Student’s t-test and were statistically significant (*P < 0.05; **P < 0.01) for each treatment compared with non-treated conditions. (B). FRDA patient-derived primary fibroblast cells were treated with 0.1, 1 and 10 μ of Src inhibitor Dasatinib or vehicle (−) for 24 h. Left panels: mature frataxin (FXN) and tubulin expression (TUB) was analyzed by western blot. Data are representative of four independent experiments. Right panels: densitometric quantification of frataxin accumulation. Frataxin expression was normalized with tubulin and frataxin expression in non-treated cells (NT) set to one. Data represent the mean ± 1 S.E.M. from four different independent experiments performed. P-values (*P < 0.05; **P < 0.01) were calculated using Student’s t-test.

All the tyrosine-mutant constructs were generated using the Quick- Fifty units of CIP (Alkaline Phosphatase, Calf Intestinal) were Change site-directed mutagenesis kit (Agilent Technologies) with added to total cell extracts resuspended in NE3 buffer pH 7.9 fi 1–210   –   speci c primers using pIRES2-Frataxin as template. The HA- (100 m NaCl, 50 m Tris HCl, 10 m MgCl2 and 1 m dithio- Ub construct was generated by M. Treier in Dirk Bohmann’s lab threitol) and incubated for 60 min at 37°C. Sodium orthovanadate (36). Constitutively active Src (pSGTSrcY527F), its inactive kinase 4m and EDTA 50 m were used to inhibit CIP activity. counterpart (pSGTSrcY527F-kin−) and constitutively active Abl (pSGTAbl-PP) have been previously described (37,38). All the constructs generated were verified by DNA sequencing. Immunoprecipitation and western blot Total cell extracts were prepared in ice-cold modified RIPA buf- Dephosphorylation assay fer (10 m sodium phosphate, pH 7.2, 150 m NaCl, 1% Na deox- CIP dephosphorylation assay kit (New England BioLabs®,Inc.) ycholate,0.1%SDS,1%IgepalCA-630and2m EDTA) or IP was used to release phosphate groups from residues of tyrosine. buffer (50 m Tris–HCl, pH 7.5, 150 m NaCl, 1% Igepal CA- Human Molecular Genetics, 2015, Vol. 24, No. 15 | 4303

MS/MS identiﬁcation of nitrated or phosphorylated residues in frataxin Human Recombinant Frataxin protein (GenScript Corp., New Jer- sey, USA) was previously treated with the Src Kinase Assay (Milli- pore) and then separated on a 1D-gel NuPAGE 4–12% (Novex, Invitrogen) run in morpholinepropanesulfoninic acid (MOPS) buffer and stained with the Colloidal Blue Staining kit (Invitrogen). The stained bands were cut from the gel and destained with a solution containing 50 m ammonium bicarbonate/acetonitrile (1:1v/v) (ACN, Merck Darmstadt, Germany). Protein bands were subsequently subjected to cysteine reduction by 10 m DTT for 1 h at 56°C and alkylation by 50 m iodoacetamide for 45 min

at RT in the dark and then dried by acetonitrile treatment and Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021 evaporation in a SpeedVac concentrator. In-gel digestion was performed by incubating gel slices with a solution containing 12.5 ng/ml chymotrypsin (Promega, Madi- son, WI, USA) in 25 m ammonium bicarbonate at 37°C overnight under stirring. To recognize phosphorylated residues, the peptide mixture was analyzed by nanoflow-reversed-phase liquid chromatography tandem mass spectrometry (RP-LC–MS/MS) using an HPLC Ultimate 3000 (Dionex, Sunnyvale, CA, USA) connected on line with a linear Ion Trap (LTQ, ThermoElectron, San Jose, CA, USA). Peptide mix- tures were desalted in a trap-column (AcclaimPepMap100 C18, LC Packings, Dionex) and then separated in a reverse phase column, a 10-cm-long fused silica capillary (SilicaTipsFS 360-75-8, New Ob- jective, Woburn, MA, USA), slurry-packed in-house with 5 μmand 200 Å pore size C18 resin (Michrom BioResources, CA, USA). Pep- tides were eluted using a 30-min linear gradient from 96% aqueous fi Figure 7. Src inhibitors rescue the aconitase defect in frataxin-de cient cells. phase (H2O with 5% ACN, 0.1% formic acid) to 60% organic buffer FRDA patient-derived B cells were treated with 10 μ of either Src inhibitor fl (ACN with 5% H2O with 0.1% formic acid) at 300 nl/min ow rate. SU6656, PP2, Dasatinib or vehicle (FRDA) for the time indicated. B cells derived Analyses were performed in positive ion mode, and the HV poten- from unaffected carrier parent (Healthy) were treated with the vehicle for the tial was set up around 1.5–1.8 kV. The LTQ mass spectrometer op- time indicated. Total aconitase activity was measured and normalized as described in the ‘Material and Methods’ section. Data represent the mean ± 1 S. erated in a data-dependent mode: each full MS scan was followed E.M. from three different independent experiments performed for each by collision-induced dissociation (CID) fragmentation of the five inhibitor. P-values (*P < 0.05; **P < 0.01) were calculated using Student’s t-test. most abundant molecular ions, using a normalized collision en- ergy of 35%. Tandem mass spectra were matched against the Uni- ProtKB/Swiss-Prot protein database with the SEQUEST algorithm (39) incorporated in the Bioworks software (version3.3, Thermo 630, 5 m EDTA and 5 m EGTA) supplemented with complete Electron), using no enzyme constraint, static cysteine alkylation protease inhibitor cocktail (Roche Diagnostics, Milan, Italy), by iodoacetamide, dynamic modification by oxidation on sodium orthovanadate 1 m and NaF 25 m to inhibit methionine and phosphorylation on tyrosine residues (Δm: phosphatases. +80 Da). A peptide was considered reliably identified when it For in vivo detection of ubiquitin conjugates, 100 μ MG132, achieved cross-correlation scores of 1.8, 2.5 and 3 for single-, dou- 50 ng/ml Ub-aldehyde and N-ethylmaleimide 2 m (NEM; ble- and triple-charged ions, respectively, and a probability cutoff Sigma–Aldrich) were added to the lysis buffer. Cell lysates for randomized identification of P < 0.001. (100 μg) were resolved by SDS–PAGE and analyzed by immunoblot with specific mAb anti-frataxin clone 1G2 and STR-23 (Immuno- logical Sciences, Rome, Italy), mAb anti-tubulin (Sigma), mAb Reagents and treatments anti-actin (Sigma), mAb anti-phosphotyrosine (Millipore), pAb The proteasome inhibitor MG132 (Sigma–Aldrich) and the deubi- anti-phospho-Src (Life Technologies), mAb anti-GFP (Takara), quitinase inhibitor Ub-aldehyde (Biomol) were added to the cell secondary antibody horseradish peroxidase-conjugated goat culture at the final concentration of 10 μ and 50 ng/ml respect- anti-mouse (Pierce), secondary antibody horseradish peroxid- ively. Src Inhibitors: PP2, SU6656 (EMD Millipore), Bosutinib, Da- ase-conjugated mouse anti-rabbit (Pierce), secondary antibody satinib and Saracatinib (Selleckchem) were added to cells at the horseradish peroxidase-conjugated goat anti-Fc anti-mouse minimum concentration of 10 n to a maximum concentration fi (Thermo Scienti c) using ECL system detection (GE Healthcare of 10 μ. For treatment longer than 24 h, inhibitors were added Europe GmbH, Milan, Italy). every 2 days and each time lymphoblasts were centrifuged and For immunoprecipitation, 5 mg of total protein extract pre- resuspended at 400.000 cells/ml with the indicated concentration – pared as mentioned above were incubated for 1 2 h at 4°C with of inhibitor. specific antibodies, previously conjugated to protein G-Sephar- ose (GE Healthcare). Immunocomplexes were then resolved and Evaluation of aconitase activity analyzed by SDS–PAGE. Densitometric analyses were performed using ImageLab Whole-cell extracts from immortalized lymphoblasts were pre- software (Biorad). pared in ice-cold CelLytic M buffer (Sigma–Aldrich) supplemented 4304 | Human Molecular Genetics, 2015, Vol. 24, No. 15

with 2 m sodium citrate, 1 m sodium orthovanadate, 25 m 11. Hayashi, G., Shen, Y., Pedersen, T.L., Newman, J.W., Pook, M. NaF and Complete EDTA-free protease inhibitor cocktail (Roche and Cortopassi, G. (2014) Frataxin deficiency increases cy- Diagnostic). Spectrophotometric aconitase assays were performed clooxygenase 2 and prostaglandins in cell and animal models at 25°C with 150 μg of cell extracts using the BIOXYTECH Aconi- of Friedreich’s ataxia. Hum. Mol. Genet., 23, 6838–6847. tase-340TM Assay (OxisResearchTM 21041). Spectrophotometric cit- 12. Herman, D., Jenssen, K., Burnett, R., Soragni, E., Perlman, S.L. rate synthase activities were assessed at 25°C with 15 μgofcell and Gottesfeld, J.M. (2006) Histone deacetylase inhibitors re- extracts using the Citrate Synthase Assay Kit (Sigma–Aldrich verse gene silencing in Friedreich’s ataxia. Nat. Chem. Biol., 2, CS0720). Aconitase activities were referred to the specific activity 551–558. of citrate synthase to correct for mitochondrial content. For the 13. Marmolino, D., Acquaviva, F., Pinelli, M., Monticelli, A., Cas- calculation of the activities, one unit of enzyme was expressed taldo, I., Filla, A. and Cocozza, S. (2009) PPAR-gamma agonist as the amount of protein that converted 1 μmol of substrate per Azelaoyl PAF increases frataxin protein and mRNA expres- minute at 25°C. sion: new implications for the Friedreich’s ataxia therapy. Cerebellum, 8,98–103. Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021 Acknowledgements 14. Tomassini, B., Arcuri, G., Fortuni, S., Sandi, C., Ezzatizadeh, V., Casali, C., Condo, I., Malisan, F., Al-Mahdawi, S., Pook, M. et al. We thank all colleagues in our laboratory, especially Dr. Simona (2012) Interferon gamma upregulates frataxin and corrects Bagedda. We also thank Allegra Via for bioinformatics support the functional deficits in a Friedreich ataxia model. Hum. and Dr. Daniela Barilà for helpful discussion and careful reading Mol. Genet., 21, 2855–2861. fl of the manuscript. Con ict of Interest statement. None declared. 15. Puccio, H., Anheim, M. and Tranchant, C. (2014) Pathophysio- gical and therapeutic progress in Friedreich ataxia. Rev. Neu- Funding rol. (Paris), 170, 355–365. 16. Rufini, A., Fortuni, S., Arcuri, G., Condo, I., Serio, D., Incani, O., This work was supported by European Research Council (Ad- Malisan, F., Ventura, N. and Testi, R. (2011) Preventing the ubi- vanced Grant #293699, FAST), Friedreich Ataxia Research Alliance quitin/proteasome-dependent degradation of frataxin, the USA and Telethon Italy (GGP11102). protein defective in Friedreich’s Ataxia. Hum. Mol. Genet., 20, 1253–1261. References 17. Hunter, T. (2007) The age of crosstalk: phosphorylation, 1. Pastore, A. and Puccio, H. (2013) Frataxin: a protein in search ubiquitination, and beyond. Mol. Cell, 28, 730–738. for a function. J. Neurochem., 126 Suppl 1, 43–52. 18. Hofer, A. and Wenz, T. (2014) Post-translational modification 2.Condo,I.,Ventura,N.,Malisan,F.,Rufini,A.,Tomassini,B. of mitochondria as a novel mode of regulation. Exp. Gerontol., and Testi, R. (2007) In vivo maturation of human frataxin. 56, 202–220. Hum. Mol. Genet., 16, 1534–1540. 19. Hebert-Chatelain, E. (2013) Src kinases are important regula- 3. Schmucker, S., Argentini, M., Carelle-Calmels, N., Martelli, A. tors of mitochondrial functions. Int. J. Biochem. Cell Biol., 45, and Puccio, H. (2008) The in vivo mitochondrial two-step mat- 90–98. uration of human frataxin. Hum. Mol. Genet., 17, 3521–3531. 20. Corcoran, A. and Cotter, T.G. (2013) Redox regulation of 4. Evans-Galea, M.V., Lockhart, P.J., Galea, C.A., Hannan, A.J. and protein kinases. FEBS J., 280, 1944–1965. Delatycki, M.B. (2014) Beyond loss of frataxin: the complex 21. Roman, E.A., Faraj, S.E., Gallo, M., Salvay, A.G., Ferreiro, D.U. molecular pathology of Friedreich ataxia. Discov. Med., 17, and Santos, J. (2012) Protein stability and dynamics modula- 25–35. tion: the case of human frataxin. PLoS One, 7, e45743. 5. Parkinson, M.H., Boesch, S., Nachbauer, W., Mariotti, C. and 22. Strawser, C.J., Schadt, K.A. and Lynch, D.R. (2014) Therapeutic Giunti, P. (2013) Clinical features of Friedreich’s ataxia: clas- approaches for the treatment of Friedreich’sataxia.Expert sical and atypical phenotypes. J. Neurochem., 126 Suppl 1, Rev. Neurother., 14, 949–957. 103–117. 23. Perlman, S.L. (2012) A review of Friedreich ataxia clinical trial 6. Condo, I., Ventura, N., Malisan, F., Tomassini, B. and Testi, R. results. J. Child Neurol., 27, 1217–1222. (2006) A pool of extramitochondrial frataxin that promotes 24. Richardson, T.E., Kelly, H.N., Yu, A.E. and Simpkins, J.W. cell survival. J. Biol. Chem., 281, 16750–16756. (2013) Therapeutic strategies in Friedreich’sataxia.Brain 7. Condo, I., Malisan, F., Guccini, I., Serio, D., Rufini, A. and Testi, Res., 1514,91–97. R. (2010) Molecular control of the cytosolic aconitase/IRP1 25. Nguyen, L.K., Kolch, W. and Kholodenko, B.N. (2013) When switch by extramitochondrial frataxin. Hum. Mol. Genet., 19, ubiquitination meets phosphorylation: a systems biology 1221–1229. perspective of EGFR/MAPK signalling. Cell Commun. Signal, 8.Paupe,V.,Dassa,E.P.,Goncalves,S.,Auchere,F.,Lonn,M., 11, 52. Holmgren, A. and Rustin, P. (2009) Impaired nuclear Nrf2 26. Dou, H., Buetow, L., Hock, A., Sibbet, G.J., Vousden, K.H. and translocation undermines the oxidative stress response in Huang, D.T. (2012) Structural basis for autoinhibition and Friedreich ataxia. PLoS One, 4, e4253. phosphorylation-dependent activation of c-Cbl. Nat. Struct. 9. Guccini, I., Serio, D., Condo, I., Rufini, A., Tomassini, B., Man- Mol. Biol., 19, 184–192. giola, A., Maira, G., Anile, C., Fina, D., Pallone, F. et al. (2011) 27. Rao, N., Dodge, I. and Band, H. (2002) The Cbl family of ubiqui- Frataxin participates to the hypoxia-induced response in tin ligases: critical negative regulators of tyrosine kinase sig- tumors. Cell Death Dis., 2, e123. naling in the immune system. J. Leukoc. Biol., 71, 753–763. 10. Schiavi, A., Torgovnick, A., Kell, A., Megalou, E., Castelein, N., 28. Porkka, K., Koskenvesa, P., Lundan, T., Rimpilainen, J., Must- Guccini, I., Marzocchella, L., Gelino, S., Hansen, M., Malisan, F. joki,S.,Smykla,R.,Wild,R.,Luo,R.,Arnan,M.,Brethon,B. et al. (2013) Autophagy induction extends lifespan and et al. (2008) Dasatinib crosses the blood–brain barrier and is reduces lipid content in response to frataxin silencing in C. an efficient therapy for central nervous system Philadelphia elegans. Exp. Gerontol., 48, 191–201. chromosome-positive leukemia. Blood, 112, 1005–1012. Human Molecular Genetics, 2015, Vol. 24, No. 15 | 4305

29. Liang, S., Pong, K., Gonzales, C., Chen, Y., Ling, H.P., Mark, R.J., 34. Li, J., Rix, U., Fang, B., Bai, Y., Edwards, A., Colinge, J., Bennett, Boschelli,F.,Boschelli,D.H.,Ye,F.,BarriosSosa,A.C.et al. K.L., Gao, J., Song, L., Eschrich, S. et al. (2010) A chemical and (2009) Neuroprotective profile of novel SRC kinase inhibitors phosphoproteomic characterization of dasatinib action in in rodent models of cerebral ischemia. J. Pharmacol. Exp. lung cancer. Nat. Chem. Biol., 6, 291–299. Ther., 331, 827–835. 35. Bradshaw, J.M. (2010) The Src, Syk, and Tec family kinases: 30. Hochhaus, A. and Kantarjian, H. (2013) The development of distinct types of molecular switches. Cell Signal., 22, 1175– dasatinib as a treatment for chronic myeloid leukemia 1184. (CML): from initial studies to application in newly diagnosed 36. Treier, M., Staszewski, L.M. and Bohmann, D. (1994) Ubiqui- patients. J. Cancer Res. Clin. Oncol., 139, 1971–1984. tin-dependent c-Jun degradation in vivo is mediated by the 31. Hill, B.G., Kota, V.K. and Khoury, H.J. (2014) Bosutinib: a third delta domain. Cell, 78, 787–798. generation tyrosine kinase inhibitor for the treatment of 37. Barila, D. and Superti-Furga, G. (1998) An intramolecular SH3- chronic myeloid leukemia. Expert Rev. Anticancer Ther., 14, domain interaction regulates c-Abl activity. Nat. Genet., 18, 765–770. 280–282. Downloaded from https://academic.oup.com/hmg/article/24/15/4296/2453025 by guest on 28 September 2021 32. Aleshin, A. and Finn, R.S. (2010) SRC: a century of science 38. Barila, D., Rufini, A., Condo, I., Ventura, N., Dorey, K., Superti- brought to the clinic. Neoplasia, 12, 599–607. Furga, G. and Testi, R. (2003) Caspase-dependent cleavage of 33. Remsing Rix, L.L., Rix, U., Colinge, J., Hantschel, O., Bennett, K. c-Abl contributes to apoptosis. Mol. Cell. Biol., 23, 2790–2799. L., Stranzl, T., Muller, A., Baumgartner, C., Valent, P., Augus- 39. Yates, J.R. 3rd, Eng, J.K., McCormack, A.L. and Schieltz, D. tin, M. et al. (2009) Global target profile of the kinase inhibitor (1995) Method to correlate tandem mass spectra of modified bosutinib in primary chronic myeloid leukemia cells. Leuke- peptides to amino acid sequences in the protein database. mia, 23, 477–485. Anal. Chem., 67, 1426–1436.