Mass Spectrometric Identification of a Novel Phosphorylation Site

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

FEBS 29495 FEBS Letters 579 (2005) 2485–2490

Mass spectrometric identiﬁcation of a novel phosphorylation site in subunit NDUFA10 of bovine mitochondrial complex I

Birgit Schillinga, Robert Aggelerb, Birte Schulenbergc, James Murrayb, Richard H. Rowa, Roderick A. Capaldib, Bradford W. Gibsona,d,* a Buck Institute for Age Research, 8001 Redwood Building, Novato, CA 94945, USA b Department of Molecular Biology, University of Oregon, Eugene, OR 97403, USA c Invitrogen, 29851 Willow Creek Road, Eugene, OR 97402, USA d Department of Pharmaceutical Chemistry, The University of California, San Francisco, CA 94143-0446, USA

Received 15 February 2005; revised 18 March 2005; accepted 21 March 2005

Available online 7 April 2005

Edited by Peter Brzezinski

been identified, including acetylation, myristylation, and oxi- Abstract Mitochondrial Complex I (NADH:ubiquinone oxidoreductase) consists of at least 46 subunits. Phosphorylation of dative modifications [12–15]. One of the most important the 42-kDa subunit NDUFA10 was recently reported using a post-translational modifications affecting and/or regulating novel phosphoprotein stain [Schulenberg et al. (2003) Analysis enzymatic activity is protein phosphorylation. Several studies of steady-state protein phosphorylation in mitochondria using a have provided evidence of phosphorylation of subunits of novel fluorescent phosphosensor dye. J. Biol. Chem. 278, Complex I with effects on electron transfer and in superoxide 27251]. Two smaller Complex I phosphoproteins, ESSS and production by this complex. Scacco and Papa [16–18] found MWFE, and their sites of modification, have since been deter- phosphorylation of subunits of Complex I with apparent mined [Chen et al. (2004) The phosphorylation of subunits of molecular weights of 6 and 18 kDa after incubation with complex I from bovine heart mitochondria. J. Biol. Chem. 279, [32P]ATP and cAMP. These same researchers subsequently as- 26036]. Here we identify the site of phosphorylation in NDUFA10 signed the 18-kDa subunit as NDUFS4 (AQDQ) and demon- from bovine heart mitochondria by tandem mass spectrometry. A single phosphopeptide spanning residues 47–60 was identified and strated that when cAMP levels were increased by treatment of confirmed by synthesis to be (47)LITVDGNICSGKpSK(60), mouse fibroblasts with cholera toxin, phosphorylation of the establishing serine-59 as the site of phosphorylation. 18-kDa subunit increased along with Complex I activity [17]. 2005 Federation of European Biochemical Societies. Published Raha and Robinson [19] also found phosphorylation of sub- by Elsevier B.V. All rights reserved. units of Complex I with molecular weights of 18 and 6 kDa along with a 42-kDa polypeptide when radiolabeling was car- Keywords: Phosphorylation; Complex I; Mitochondria; Mass ried out in rat heart mitochondria in the presence of protein spectrometry; ParkinsonÕs disease kinase A. The 42-kDa protein was attributed to a contaminat- ing protein, the E1a subunit of pyruvate dehydrogenase, and the latter two bands to the Complex I subunits NDUFS4 (AQDQ) or NDUFB7 (B18), and NDUFA1 (MWFE), respec- 1. Introduction tively. Walker and colleagues [20] using Edman degradation and mass spectrometry sequence analysis, also obtained label- NADH:ubiquininone oxidoreductase or Complex I is the ing of two subunits of Complex I, which they identified as largest of the five mitochondrial electron transport chain com- ESSS (the 18-kDa protein) modified at serine-20 and MWFE plexes with an approximate molecular weight of 980 kDa. also called NDUFA1 (the 6-kDa protein) labeled at serine- Complex I is a membrane bound multimeric enzyme that is 55 when bovine Complex I was incubated with radiolabeled comprised of at least 46 subunits. Impairment of Complex I [32P]ATP and cAMP (and protein kinase A). Recently, a novel activity has been described in several neurodegenerative dis- phospho-specific, fluorescent dye has been introduced to iden- eases, including ParkinsonÕs disease (PD) [1,2], AlzheimerÕs dis- tify phosphoproteins. Using this method Schulenberg and col- ease [3], and Huntington disease [4]. There is strong evidence leagues [21,22] found staining of two subunits of Complex I that mitochondrial Complex I inhibition can be caused in part with molecular weights of 42 and 18 kDa, respectively. In these through accumulated damage caused by age-related oxidative experiments, the protein migrating at 42 kDa was assigned by stress mediated by reactive oxygen and nitrogen species mass spectrometry as NDUFA10. Unlike the 18-kDa subunit [5–7]. For example, oxidation of cysteine residues in Complex that required protein kinase treatment for significant labeling I [8], either through glutathionylation or S-nitrosylation by the phospho-specific stain (Pro-Q Diamond), the 42 kDa [9–11], is likely to play a role in PD. Bovine and human mito- protein was labeled by the dye in mitochondria isolated with- chondrial Complex I have been analyzed extensively and sev- out prior kinase treatment. The above results, when considered eral different types of post-translational modifications have together, raise obvious questions. First, is the staining of NDUFA10 really due to the presence of a covalent phosphate or is it an artifact of the staining procedure? Also, if NDU- *Corresponding author. Fax: +1 415 209 2231. FA10 is a phosphoprotein what residues are involved? In an E-mail address: [email protected] (B.W. Gibson). attempt to resolve these questions we have used on-line

0014-5793/$30.00 2005 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2005.03.061 2486 B. Schilling et al. / FEBS Letters 579 (2005) 2485–2490

HPLC-MS/MS to analyze NDUFA10 in Complex I obtained ProXeon nanospray ion source (ProXeon Biosystems). ESI-MS/MS by sucrose gradient fractionation or immunoprecipitation of mass spectra were acquired as previously described [27]. Mass spectro- untreated bovine heart mitochondria. metric data were analyzed using the search engine Mascot (Matrix Sci- ence) [28]. A custom-designed ‘‘bovine Complex I’’ protein database was incorporated into our in-house licensed Mascot server as previously described [26]. 2. Materials and methods 2.6. Synthetic phosphopeptides 2.1. Materials The two custom synthesized phosphopeptide isomers, LITVDG- NuPAGE 12.5% Bis–Tris gels, MES running buffer, SYPRO Ruby NICpSGKSK, and LITVDGNICSGKpSK were diluted with 50 mM and Pro-Q Diamond fluorescent protein gel stains were obtained from ammonium bicarbonate to a concentration of 400 lM. The peptides Molecular Probes/Invitrogen. For proteolysis, sequencing grade, mod- were reduced with a 500-fold molar excess of DTT (60 C, 60 min), ified trypsin (porcine) was purchased from Promega. N-dodecyl-b-DD- and then alkylated with a 1100-fold molar excess of iodoacetamide maltoside was purchased from Calbiochem. Additional reagents for (37 C, 45 min). protein chemistry were obtained from Sigma. HPLC solvents were obtained from Burdick & Jackson. For MALDI-MS experiments, a matrix solution of a-cyano-4-hydroxycinnamic acid in acetonitrile/ 3. Results methanol was purchased from Agilent Technologies. The two synthetic phosphopeptides, LITVDGNICpSGKSK and LITVDGNICSGKpSK (pS is phosphoserine), were custom synthesized and purchased from In this current study, we first investigated the phosphoryla- Princeton Biomolecules Corporation. tion status of bovine Complex I under steady-state conditions isolated from heart tissue by either sucrose gradient fraction- 2.2. Isolation of bovine heart Complex I ation [24,25] or immunoprecipitation [13]. This material was Mitochondria were prepared from bovine heart by homogenizing, separated by 1D SDS–PAGE and stained first with a phos- filtration, and differential centrifugation as described previously [23]. Complex I was then isolated and purified from bovine heart mitochon- pho-specific fluorescent dye (Pro-Q Diamond) and then with dria by either a sucrose gradient centrifugation protocol [24,25] or an a total-protein stain (SYPRO Ruby) so that all subunits could immunoisolation protocol where Complex I was immunocapturing in- be visualized. The methods used for Complex I purification tact as reported by Murray et al. [13]. and subunit separation and visualization had been previously described in several recent reports, including a one-step purifi- 2.3. Dephosphorylation of Complex I cation method for human Complex I from mitochondria [13] Complex I obtained from the sucrose gradient was treated with bo- and detailed mass spectrometry analysis of both bovine and vine alkaline phosphatase (AP) after adjusting pH to 8.8 and supple- mouse Complex I subunits [26]. menting the buffer with 2 mM MgCl2 and 1% SDS. The dephosphorylation of 50 lg Complex I was carried out in 100 llat The subunit profile of the bovine Complex I proteins puri- 35 C for 2 h using 5 U of bovine AP. After neutralizing the pH with fied by sucrose gradient fractionation and then separated by 3 ll 10% TFA samples were pelleted with chloroform/methanol. Pellets 1D SDS–PAGE is shown in Fig. 1. The major band showing were redissolved in 50 ll dissociation buffer containing 50 mM DTT. Samples were then heated to 95 C for 5 min, spun down and 15 lg significant phospho-specific staining (Fig. 1A) migrated at were applied on a 12.5% minigel, which was stained with Pro-Q Dia- 42-kDa and corresponds to the NDUFA10 subunit. This mond (D) phosphoprotein stain followed by SYPRO Ruby (S) protein observation is consistent with previous data by Schulenberg gel stain according to the manufacturerÕs instructions (Invitrogen). Ra- et al. [21,22] where a similar high staining band was identified tios of the staining intensities D/S were determined with ImageGauge as this subunit. In addition, Fig. 1A and B lanes 2–4 show (FUJI Film, Tokyo, Japan) using a 473 nm excitation for the SYPRO Ruby dye and a 532 nm excitation for Pro-Q Diamond dye (580 nm incubation experiments of mitochondria with AP under dif- emission filter). ferent conditions of incubation (see Fig. 1, legend). The fluo- rescence signal obtained after Pro-Q Diamond (Fig. 1A) and 2.4. In-gel tryptic digestion of proteins SYPRO Ruby (Fig. 1B) staining was quantified both before Protein bands of interest were manually excised out of the gel and and after AP treatment using Image Gauge (FUJI Film, To- processed with an automatic in-gel digester robot (ProGest, Genomic kyo) to determine the phosphorylation level for the NDU- Solutions). The gel bands were destained and dehydrated with acetonitrile, the proteins were reduced with 10 mM DTT (60 C, 30 min), and FA10 gel band. The phospho-specific fluorescent signal was the reduced cysteine residues were then alkylated with 100 mM iodoa- compared as a ratio with the same sample stained with SY- cetamide (37 C, 45 min). Prior to enzymatic digestion excess reagents PRO Ruby (Pro-Q Diamond/SYPRO Ruby, D/S) and was were removed and the gel pieces were washed twice with 25 mM expressed in percent of the control lane 1 where no AP was ammonium bicarbonate, dehydrated, and incubated with 250 ng added which was defined as 100% (lane 1, D/S: 0.47). AP sequencing grade trypsin (37 C for 4 h). The resulting tryptic peptides were extracted from the gel with 10% formic acid as previously treatment was then carried out with and without the addition described [26]. of detergent, the latter of which led to a slight decrease in fluorescent signal to 97% of control (lane 2, D/S: 0.46). When 2.5. Mass spectrometry and database searches treatment with AP was carried out in the presence of 1% lau- Digested protein gel bands and synthetic peptide samples were ini- rylmaltoside (lane 3) a further but incremental decrease was tially analyzed by matrix-assisted laser desorption ionization time- observed to 93% of control (D/S: 0.44). In the presence of of-flight (MALDI-TOF) mass spectrometry on a Voyager DE STR 1% SDS (lane 4), however, a significant decrease to 47% of plus instrument (Applied Biosystems). Mass spectra were acquired in positive-ionization mode with reflectron optics as described previously the control was observed (D/S: 0.22), suggesting the presence [27]. All samples were then analyzed by reverse-phase nano-HPLC- of a phosphoprotein. The small decrease in signal after AP MS/MS. Briefly, peptides were separated on an Ultimate nanocapillary treatment without detergent or with 1% laurylmaltoside is HPLC system equipped with a PepMape C18 nano-column (75 lm presumably due to low accessibility of this phosphorylation l I.D. · 15 cm) (Dionex) and CapTrap Micro guard column (0.5 l site under the conditions used in the isolation of Complex bed volume, Michrom) as recently reported [27]. The nano-HPLC column eluant was directly coupled to a ‘‘QSTAR Pulsar i’’ quadrupole I, conditions that would keep Complex I largely intact and orthogonal TOF mass spectrometer (MDS Sciex) equipped with a NDUFA10 subunit in its native configuration, although not B. Schilling et al. / FEBS Letters 579 (2005) 2485–2490 2487

Fig. 1. Alkaline phosphatase (AP) treatment of Complex I from sucrose gradient. A 1D SDS–PAGE gel was stained for phosphoprotein using Pro-Q Diamond phosphoprotein gel stain (A) before post-staining with SYPRO Ruby protein gel stain (B) for a total protein proﬁle. Lane 1 (both panels) shows a control sample of 15 lg Complex I without addition of detergent or AP. Alkaline phosphatase treatment was carried out without addition of detergent (lane 2), and in the presence of 1% laurylmaltoside (lane 3) or 1% SDS (lane 4). Lane 5 (named Std) shows a broad range molecular weight marker containing ovalbumin as a positive control (Bio-Rad Laboratories). NDUFA10 is indicated (as ‘‘A10’’) in both panels at a molecular weight of around 42 kDa.

necessarily identical to its configuration or conformation 648.3 (y6/y6-98), 1030.6/932.5 (y9/y9-98), 1145.5/1047.5 (y10/ within the mitochondria. y10-98), and 1345.7/1247.7 (y12/y12-98), indicating the presence To determine the site-specific features of Complex I phos- of a phosphate in the C-terminal part of this peptide at either phorylation, we analyzed the NDUFA10 band by mass spec- Ser-56 or Ser-59. The presence of a few smaller fragment ions trometry. Previously, we performed a thorough mass at m/z 488.3, 401.3, 344.2, and 216.1, were tentatively assigned spectrometric analysis of immunocaptured bovine Complex I as yn-P ions (n = 2–5), suggesting Ser-59 as the site of phosphor- that was separated by 1D SDS–PAGE as shown in Fig. 1A ylation and not Ser-56, i.e., LITVDGNICSGKpS59K. To eliminate and B, identifying 43 out of 46 bovine Complex I protein sub- the possible phosphorylation of serine-56, the two mono- units [26]. In this current work, we focused our attention to- phosphorylated peptide isomers LITVDGNICSGKpS59K and wards an in depth analysis and sequence assignment of the LITVDGNICpS56GKSK were synthesized de novo, alkylated, ‘‘phosphorylated’’ NDUFA10 subunit. After excision of the and analyzed by nano-HPLC-ESI-MS/MS mass spectrometry protein from the gel and digestion with trypsin, the sample for comparison. was subjected to initial MALDI MS analysis. In this experi- Fig. 3 shows the ESI-MS/MS spectra of carbamidomethy- ment, 10 peptide masses (or fingerprints) were detected by lated synthetic peptides LITVDGNICSGKpS59K(Fig. 3A) MALDI MS for subunit NDUFA10, yielding a sequence cov- and LITVDGNICpS56GKSK (Fig. 3B). Fragmentation pat- erage of 35% (data not shown). tern of the two mono-phosphorylated peptide isomers appear To obtain more sequence coverage and to identify the site(s) of very similar and correspond well to the overall fragmentation phosphorylation, the tryptic digestion mixture of the NDU- of the native phosphopeptide isolated from bovine Complex I FA10 gel band was then investigated by nano-HPLC-ESI-MS/ (compare Fig. 2). Upon detailed inspection of the tandem mass MS. The resulting tandem mass spectra were analyzed using a spectra for the synthetic peptides, subtle but significant differ- Mascot database search that allowed for post-translational ences were delineated that allowed for the determination of the modifications, including phosphorylation. From this data set, specific phosphorylation site as Ser-59 (see Fig. 4). In these a precursor ion selected for MS/MS, [M + 2H]2+ at m/z spectra, the mass ranges from m/z 140–500 are displayed for 786.32+ (M = 1570.62), was matched to a mono-phosphorylated the native phosphopeptide isolated from bovine bovine NDU- peptide with the sequence corresponding to an expected tryptic FA10 (Fig. 4A), and the two synthetic peptides, LIT- peptide spanning Leu-47 to Lys-60 of the NDUFA10 subunit, VDGNICSGKpS59K(Fig. 4B) and LITVDGNICpS56GKSK i.e., L47ITVDGNICSGKSK60 (NDUFA10 sequence number- (Fig. 4C); these peptides are further referred to as native pep- ing is based on the protein precursor and not the mature protein tide LITV-A10 and synthetic peptides LITV-pS-59 and LITV- which contains a N-terminal 23 amino acid import sequence pS-56, respectively. Significant differences in fragmentation which is removed post-translationally). Fragmentation analysis can be seen among these peptides in the low mass region for clearly showed the presence of one phosphate group on either the relevant y-ions. In particular, peptide LITV-pS-56 shows serine residue Ser-56 or Ser-59 (shown in bold type above). Spe- an abundant y2-ion at m/z 234.1 that indicates an unmodifed cifically, a series of y-type fragment ions and several b-type frag- Ser-59 (Fig. 4C), whereas peptide LITV-pS-59 (Fig. 4B) and ment ions were observed. For example, the b3 and b3-H2O ion native peptide LITV-A10 (Fig. 4A) do not show a peak at this pair at m/z 328.2 and 310.2, clearly indicate a non-phosphory- m/z value (m/z 234). Instead LITV-pS-59 and LITV-A10 show lated threonine at position 49. Several y-type ions were also a relatively weak y2-ion at m/z 314.1 and a much stronger char- observed along with their yn-98 ion counterparts, the latter of acteristic y2-98 (y2-P) ion at m/z 216.1 corresponding to a phos- which results from neutral loss of phosphoric acid (–H3PO4) phate group on Ser-59 (the abundant m/z 216.1 ion is not [29]. These fragment ion pairs were observed at m/z 746.4/ present in LITV-pS-56 as expected). Several other distinguishing 2488 B. Schilling et al. / FEBS Letters 579 (2005) 2485–2490

328227 A y2-P x3 a2 a2 L I T V D G N I C* S G K pS K 216.1 y -P C*SGK 199.2 y2 4 b 1345 1145 1030 746 314 147 b4-18 2 b2 b3 227.2 -P DG 488 344 y5-P 1247 1047 932 648 401 216 y2-P % b3-18 DG y12 y3-P 216.1 y -P y y1 173.1 12 9 y10 1345.7 624.32+ 1030.6 b3 1145.5 % y3-P y5-P y -P y9-P y 6 y10-P 1 -18 344.2 488.3 648.3 932.5 1047.5 y12-P y -P B y2-P K 4 1247.7 401.3 y6 216.1 L I T V D G N I C* S G K pS K 746.4 y2 L,I a2 b2 b 314 147 3 b -18 -P b -18 4 3 488 200 400 600 800 1000 1200 DG 401 y4-P C*SGK m/z % 344 y -P 216 3 y5-P Fig. 2. ESI-MS/MS spectrum of mono-phosphorylated peptide LIT- y1 VDGNIC\S56GKS59K (residues 47–60) after tryptic digestion of Complex I subunit NDUFA10 (42-kDa subunit) obtained from puriﬁed bovine heart Complex I. The molecular ion, [M + 2H]2+ at m/z 786.32+ (M = 1570.62), was selected for collision induced dissoci- C VD, IT ation. Several y-fragment ions featured loss of À98 Da due to loss of 215.1 y L I T V D G N I C* pS G K S K phosphoric acid as indicated by vertical arrows, suggesting phosphor- a2 b 2 2 b3 y ylation at Ser-59 (C\ is carbamidomethylated cysteine). 234.1 4 419 234 147 -P 362 TVD y DG 3 488 b3-18 % b4-18 A y5-P b2 y a2 1 227.2 L I T V D G N I C* S G K pS K y -P 2 b 3 y5-P 140 200 260 300 360 420 480 216.1 328.2 488.3 DG Fig. 4. Staggered plot of partial ESI-MS/MS spectra (m/z range 140– b3 C*SGK y12-P 173.1 500) of phosphorylated peptides: (A) native LITV-A10 obtained form -18 433.2 624.32+ y12-P puriﬁed bovine heart Complex I after tryptic digestion of Complex I % y -P y -P \ 59 y4-P 6 10 subunit NDUFA10, (B) synthetic peptide LITVDGNIC SGKpS K y1 y9-P 1247.5 \ 648.3 1047.5 (LITV-pS-59), and (C) synthetic peptide LITVDGNIC pS56GKSK 932.4 (LITV-pS-56). In all cases, the molecular ion, [M + 2H]2+ at m/z K y10-P y12 2+ y3-P 2+ 786.3 (M = 1570.62), was selected for collision induced dissociation. 2+ MH2 -P y10 524.2 y9 1345.6 2+ Several y-fragment ions featured the À98 Da loss of phosphoric acid as 737.4 1145.5 \ L,I indicated by vertical arrows (C is carbamidomethylated cysteine).

B a2 b2 of LITV-pS-59 and LITV-A10 (Fig. 4B and A). Peptides 227.2 LITV-pS-59 and LITV-A10, both show ions at m/z 401.1 y L I T V D G N I C* pS G K S K 2 (y -98), 344.2 (y -98), and 433.2 (CS56GK, internal ion, 234.2 y -P 4 3 DG 5 unmodiﬁed Ser-56), that on the other hand were not observed b 488.3 173.1 3 in LITV-pS-56. Thus, we can conclude that the phosphate 328.2 y -P 12 y12-P 2+ group of the native peptide LITV-A10 isolated from bovine y b3 y 624.3 y9-P y -P % 1 4 10 1247.5 Complex I (NDUFA10) is located on serine-59 alone. -18 419.3 y -P 932.4 1047.5 6 Phosphorylation of Ser-59 is also consistent with the ob- K y 648.3 3 540.3 y12 served proteolytic properties of bovine protein NDUFA10. 2+ y10 MH2 -P 1345.6 While the phosphorylated peptide was observed as LIT- 2+ y 737.4 9 1145.5 VDGNICSGKpS59K, the same non-phoshorylated version of L,I this peptide was not observed. Rather, a shorter peptide LIT- VDGNICSGK was observed, suggesting that the phosphory- 600 200 400 800 1000 1200 lation of Ser-59 blocks the proteolytic cleavage at Lys-58; in m/z the absence of the Ser-59 phosphate group tryptic cleavage Fig. 3. ESI-MS/MS spectrum of mono-phosphorylated, custom syn- immediately occurs again at Lys-58 and thus the longer thesized peptide isomers LITVDGNIC\SGKpS59K (A), and LIT- non-phosphorylated peptide LITVDGNICSGKS59K is not \ 56 2+ VDGNIC pS GKSK (B). In each case, the molecular ion, [M + 2H] observed. at m/z 786.32+ (M = 1570.62), was selected for collision induced dissociation (C\ is carbamidomethylated cysteine).

4. Discussion diﬀerences were also observed as follows: LITV-pS-56 (Fig. 4C) shows an abundant ion at m/z 419.3 (y4), with a less abun- In bovine NDUFA10, a single amino acid, serine-59, was dant ion at m/z 362.2 (y3) which are both absent in the spectra identiﬁed by tandem mass spectrometry to be phosphorylated. B. Schilling et al. / FEBS Letters 579 (2005) 2485–2490 2489

While there may be other sites of phosphorylation within ine (or threonine) at this position. However, examination of NDUFA10 still to be identified, careful analysis of all MS the immediate flanking regions of serine-59 among bovine, hu- and MS/MS data revealed no evidence for additional sites. man, rodent, C. elegans and Drosophila, showed that there are This observation represents the first time that an endogenous several other sites that are conserved in two or more of the spe- phosphorylation site has been identified unequivocally at the cies that have consensus motifs for S/T kinases. Experiments amino acid level in mammalian Complex I that was not in- are now underway to investigate these potential phosphoryla- duced by prior kinase treatment. In a recent report by Walker tion sites in more detail. and colleagues [20], for example, neither of the two subunits The mitochondrial kinase that is responsible for the phos- that were found to be phosphorylated after kinase treatment, phorylation of NDUFA10 has not yet been determined. How- ESSS and MWFE, appear to be phosphorylated when isolated ever, the consensus motif surrounding Ser-59 suggests a casein directly from mitochondria. kinase I-like activity. Recently, a mutation in the gene encod- NDUFA10 is located in the hydrophobic membrane arm ing a novel mitochondrial kinase, PTEN-induced kinase 1 (or (subcomplex Ic) that is embedded in the mitochondrial inner PINK1), was shown to result in an hereditary form of early on- membrane [30]. However, unlike other protein components set PD [33]. While it is too early to speculate whether a corre- of the membrane arm, the NDUFA10 subunit is a relatively lation exists between PINK1 kinase and Complex I hydrophilic protein and is thought to be only loosely associ- phosphorylation, including the novel phosphorylation site ated with the Ic subcomplex [30]. Presumably, this would make identified here, it is worth noting that Complex I dysfunction the NDUA10 subunit more accessible to external interactions is one of the earliest detectable biochemical features of PD. with other non-Complex I proteins, such as kinases and phos- It is also worth noting that NDUA10 has been referred to as phatases, although details of the structure of NDUA10 that ‘‘mammalian-specific’’ subunit of Complex I [34], even though might support such interactions are currently unavailable. Re- it is present in some higher non-mammalian eukaryotes. None- cently, published data from 2D gel separations of bovine Com- theless, its apparent absence in lower eukaryotes such as N. plex I showed multiple spots for the NDUFA10 subunit crassa (fungus), Arabidopsis thaliana and Chlamydomonas rein- reflecting a shift in its pI values [21,31]. These multiply resolved hardtii (plant/green algae) [34], suggests that NDUA10 has ap- ‘‘trains’’ of spots seen on a 2D SDS–PAGE gel may indicate peared late in mitochondrial evolution, consistent with a more differential phosphorylation, although other possible explana- regulatory role for this phosphoprotein. Indeed, future studies tions, such as deamidation [32], acetylation or methylation, are planned to examine the consequence of NDUFA10 phos- could also produce such a pattern. However, the only other phorylation and identify the kinase involved, information that modification that we found for bovine protein subunit NDU- will hopefully provide some insight on its function in Complex FA10 besides phosphorylation was a tryptophan oxidation to I function or assembly. formylkynurenin, which would not cause a change in pI values. Several post-translational and non-enzymatic modifica- Acknowledgments: We thank Dr. Wayne Patton for helpful discus- tions have been reported for protein subunits of bovine and sions. This work was supported by NIH R21 NS043620-01 grant to BWG, NIH 2 P50 NS039764-06 grant to RAC, and a National Cancer mouse Complex I [26,31], although no phosphorylation sites Institute Grant R33 CA093293-01 awarded to Molecular Probes, Inc. have been identified without prior kinase treatment [16,17,19,20]. In NDUFA10 isolated from bovine Complex I, however, Schulenberg et al. [21] observed several NDUFA10 References isoforms after 2D gel separation, some of which stained pref- erentially with the phospho-specific stain Pro-Q Diamond. [1] Schapira, A.H., Cooper, J.M., Dexter, D., Clark, J.B., Jenner, P. This latter observation indicated that protein phosphorylation and Marsden, C.D. (1990) Mitochondrial complex I deficiency in was the basis for this heterogeneity, and this has now been con- ParkinsonÕs disease. J. Neurochem. 54, 823–827. [2] Betarbet, R., Sherer, T.B., MacKenzie, G., Garcia-Osuna, M., firmed in this current mass spectrometric study. Similarly, we Panov, A.V. and Greenamyre, J.T. (2000) Chronic systemic have also observed a multiple spot pattern in the 2D gel anal- pesticide exposure reproduces features of ParkinsonÕs disease. ysis of mouse NDUFA10 (unpublished data, Gibson and Nat. Neurosci. 3, 1301–1306. Andersen) suggesting that there is likely other site(s) of phos- [3] Kim, S.H., Vlkolinsky, R., Cairns, N., Fountoulakis, M. and phorylation in this subunit distinct from the site identified in Lubec, G. (2001) The reduction of NADH ubiquinone oxidoreductase 24- and 75-kDa subunits in brains of patients with Down this study, as there is no corresponding hydroxyl-containing syndrome and AlzheimerÕs disease. Life Sci. 68, 2741–2750. amino acid at position Ser-59 in the mouse sequence (see [4] Arenas, J., Campos, Y., Ribacoba, R., Martin, M.A., Rubio, J.C., Fig. 5). Indeed, if one considers other animal models com- Ablanedo, P. and Cabello, A. (1998) Complex I defect in muscle monly used in aging and neurodegenerative disease research, from patients with HuntingtonÕs disease. Ann. Neurol. 43, 397– 400. only Caenorhabditis elegans and Drosophila (Fig. 5) have a ser- [5] Lenaz, G., Bovina, C., Castelluccio, C., Fato, R., Formiggini, G., Genova, M.L., Marchetti, M., Pich, M.M., Pallotti, F., Parenti Castelli, G. and Biagini, G. (1997) Mitochondrial complex I defects in aging. Mol. Cell. Biochem. 174, 329–333. [6] Lenaz, G., Bovina, C., DÕAurelio, M., Fato, R., Formiggini, G., Genova, M.L., Giuliano, G., Merlo Pich, M., Paolucci, U., Parenti Castelli, G. and Ventura, B. (2002) Role of mitochondria in oxidative stress and aging. Ann. N. Y. Acad. Sci. 959, 199–213. [7] Tretter, L., Sipos, I. and Adam-Vizi, V. (2004) Initiation of neuronal damage by complex I deficiency and oxidative stress in ParkinsonÕs disease. Neurochem. Res. 29, 569–577. Fig. 5. Sequence homology of bovine a10 subunits among other [8] Jha, N., Jurma, O., Lalli, G., Liu, Y., Pettus, E.H., Greenamyre, species surrounding phosphorylation site serine-59 (boxed), other J.T., Liu, R.M., Forman, H.J. and Andersen, J.K. (2000) potential sites are shown in bold. Glutathione depletion in PC12 results in selective inhibition of 2490 B. Schilling et al. / FEBS Letters 579 (2005) 2485–2490

mitochondrial complex I activity. Implications for ParkinsonÕs ylation in mitochondria using a novel fluorescent phosphosensor disease. J. Biol. Chem. 275, 26096–26101. dye. J. Biol. Chem. 278, 27251–27255. [9] Carreras, M.C., Franco, M.C., Peralta, J.G. and Poderoso, J.J. [22] Schulenberg, B., Goodman, T.N., Aggeler, R., Capaldi, R.A. and (2004) Nitric oxide, complex I, and the modulation of mitochon- Patton, W.F. (2004) Characterization of dynamic and steady-state drial reactive species in biology and disease. Mol. Aspects Med. protein phosphorylation using a fluorophore phosphoprotein 25, 125–139. stain and mass spectrometry. Electrophoresis 25, 2526–2532. [10] Taylor, E.R., Hurrell, F., Shannon, R.J., Lin, T.K., Hirst, J. and [23] Smith, A. (1967) Preparation, properties, and conditions for assay Murphy, M.P. (2003) Reversible glutathionylation of complex I of mitochondria: slaughterhouse material. Methods Enzymol. 10, increases mitochondrial superoxide formation. J. Biol. Chem. 278, 81–86. 19603–19610. [24] Hanson, B.J., Schulenberg, B., Patton, W.F. and Capaldi, R.A. [11] Hsu, M., Srinivas, B., Kumar, J., Subramanian, R. and Andersen, (2001) A novel subfractionation approach for mitochondrial J. (2005) Glutathione depletion resulting in selective mitochon- proteins: a three-dimensional mitochondrial proteome map. drial complex I inhibition in dopaminergic cells is via an NO- Electrophoresis 22, 950–959. mediated pathway not involving peroxynitrite: implications for [25] Taylor, S.W., Warnock, D.E., Glenn, G.M., Zhang, B., Fahy, E., ParkinsonÕs disease. J. Neurochem. 92, 1091–1103. Gaucher, S.P., Capaldi, R.A., Gibson, B.W. and Ghosh, S.S. [12] Carroll, J., Shannon, R.J., Fearnley, I.M., Walker, J.E. and Hirst, (2002) An alternative strategy to determine the mitochondrial J. (2002) Definition of the nuclear encoded protein composition of proteome using sucrose gradient fractionation and 1D PAGE on bovine heart mitochondrial complex I. Identification of two new highly purified human heart mitochondria. J. Proteome Res. 1, subunits. J. Biol. Chem. 277, 50311–50317. 451–458. [13] Murray, J., Zhang, B., Taylor, S.W., Oglesbee, D., Fahy, E., [26] Schilling, B., Bharath, M.M.S., Row, R.H., Murray, J., Cusack, Marusich, M.F., Ghosh, S.S. and Capaldi, R.A. (2003) The M.P., Capaldi, R.A., Freed, C.R., Prasad, K.N., Andersen, J.K. subunit composition of the human NADH dehydrogenase and Gibson, B.W. (2005) Rapid purification and mass spectro- obtained by rapid one-step immunopurification. J. Biol. Chem. metric characterization of mitochondrial NADH dehydrogenase 278, 13619–13622. (complex I) from rodent brain and a dopaminergic neuronal cell [14] Murray, J., Taylor, S.W., Zhang, B., Ghosh, S.S. and Capaldi, line. Mol. Cell. Proteomics 4, 84–96. R.A. (2003) Oxidative damage to mitochondrial complex I due to [27] Schilling, B., Yoo, C.B., Collins, C.J. and Gibson, B.W. (2004) peroxynitrite: identification of reactive tyrosines by mass spec- Determining cysteine oxidation status using differential alkyl- trometry. J. Biol. Chem. 278, 37223–37230. ation. Int. J. Mass Spectrom. 236, 117–127. [15] Taylor, S.W., Fahy, E., Murray, J., Capaldi, R.A. and Ghosh, [28] Perkins, D.N., Pappin, D.J., Creasy, D.M. and Cottrell, J.S. S.S. (2003) Oxidative post-translational modification of trypto- (1999) Probability-based protein identification by searching phan residues in cardiac mitochondrial proteins. J. Biol. Chem. sequence databases using mass spectrometry data. Electrophoresis 278, 19587–19590. 20, 3551–3567. [16] Papa, S., Sardanelli, A.M., Cocco, T., Speranza, F., Scacco, S.C. [29] DeGnore, J.P. and Qin, J. (1998) Fragmentation of phosphopep- and Technikova-Dobrova, Z. (1996) The nuclear-encoded 18 kDa tides in an ion trap mass spectrometer. J. Am. Soc. Mass (IP) AQDQ subunit of bovine heart complex I is phosphorylated Spectrom. 9, 1175–1188. by the mitochondrial cAMP-dependent protein kinase. FEBS [30] Sazanov, L.A., Peak-Chew, S.Y., Fearnley, I.M. and Walker, J.E. Lett. 379, 299–301. (2000) Resolution of the membrane domain of bovine complex I [17] Scacco, S., Vergari, R., Scarpulla, R.C., Technikova-Dobrova, Z., into subcomplexes: implications for the structural organization of Sardanelli, A., Lambo, R., Lorusso, V. and Papa, S. (2000) the enzyme. Biochemistry 39, 7229–7235. cAMP-dependent phosphorylation of the nuclear encoded 18- [31] Carroll, J., Fearnley, I.M., Shannon, R.J., Hirst, J. and Walker, kDa (IP) subunit of respiratory complex I and activation of the J.E. (2003) Analysis of the subunit composition of complex I complex in serum-starved mouse fibroblast cultures. J. Biol. from bovine heart mitochondria. Mol. Cell. Proteomics 2, 117– Chem. 275, 17578–17582. 126. [18] Papa, S., Sardanelli, A.M., Scacco, S., Petruzzella, V., Technik- [32] Robinson, N.E. and Robinson, A.B. (2001) Deamidation of ova-Dobrova, Z., Vergari, R. and Signorile, A. (2002) The human proteins. Proc. Natl. Acad. Sci. USA 98, 12409–12413. NADH: ubiquinone oxidoreductase (complex I) of the mamma- [33] Valente, E.M., Abou-Sleiman, P.M., Caputo, V., Muqit, M.M., lian respiratory chain and the cAMP cascade. J. Bioenerg. Harvey, K., Gispert, S., Ali, Z., Del Turco, D., Bentivoglio, A.R., Biomembr. 34, 1–10. Healy, D.G., Albanese, A., Nussbaum, R., Gonzalez-Maldonado, [19] Raha, S., Myint, A.T., Johnstone, L. and Robinson, B.H. (2002) R., Deller, T., Salvi, S., Cortelli, P., Gilks, W.P., Latchman, D.S., Control of oxygen free radical formation from mitochondrial Harvey, R.J., Dallapiccola, B., Auburger, G. and Wood, N.W. complex I: roles for protein kinase A and pyruvate dehydrogenase (2004) Hereditary early-onset ParkinsonÕs disease caused by kinase. Free Radic. Biol. Med. 32, 421–430. mutations in PINK1. Science 304, 1158–1160. [20] Chen, R., Fearnley, I.M., Peak-Chew, S.Y. and Walker, J.E. [34] Cardol, P., Vanrobaeys, F., Devreese, B., Van Beeumen, J., (2004) The phosphorylation of subunits of complex I from bovine Matagne, R.F. and Remacle, C. (2004) Higher plant-like subunit heart mitochondria. J. Biol. Chem. 279, 26036–26045. composition of mitochondrial complex I from Chlamydomonas [21] Schulenberg, B., Aggeler, R., Beechem, J.M., Capaldi, R.A. and reinhardtii: 31 conserved components among eukaryotes. Bio- Patton, W.F. (2003) Analysis of steady-state protein phosphor- chim. Biophys. Acta 1658, 212–224.