Identification of tyrosine-phosphorylation sites in the nuclear membrane protein emerin Andreas Schlosser 1, *, Ramars Amanchy 2, * and Henning Otto 3 1 Charite ´ , Institut fu ¨ r Medizinische Immunologie, Berlin, Germany 2 McKusick-Nathans Institute for Genetic Medicine and the Department of Biological Chemistry, Johns Hopkins University, Baltimore, MD, USA 3 Freie Universita ¨ t Berlin, Institut fu ¨ r Chemie und Biochemie, Germany The nuclear envelope encloses the genetic material of a eukaryotic cell and takes part in its structural and functional organization. It consists of interconnected membranes, an outer nuclear membrane (ONM) and an inner nuclear membrane (INM). The ONM is part of the rough endoplasmic reticulum and folds at the nuclear pores into the INM, which is firmly attached to the lamina by integral membrane proteins of the INM. The INM proteins form complexes, transiently or stably, with lamins, chromatin proteins and a vari- ety of regulatory proteins, including transcriptional regulators and splicing factors [1,2]. Attempts have been made to identify and catalogue the complete rep- ertoire of nuclear-envelope proteins by subcellular pro- teomics. These approaches resulted in several novel validated nuclear membrane proteins and also in long lists of putative protein constituents of the nuclear envelope awaiting their validation [3,4]. Such an inventory is just a first step that must be followed by the analysis of molecular interactions of the nuclear-envelope proteins. Well-characterized nuc- lear-envelope proteins like the lamin B receptor, the lamina-associated polypeptide 2 (LAP2) membrane iso- forms, emerin or the lamins, evidently participate in the formation of distinct complexes by the cell at the right place and the right time [5–12]. To regulate such complex interactions, cells use post-translational modi- fications; their regulatory repertoire relies mostly on the transient phosphorylation of either serine ⁄ threon- ine or tyrosine residues [13,14]. The identification of such post-translational modifications is efficiently addressed by specialized mass spectrometric techniques Keywords Emerin; Emery–Dreifuss muscular dystrophy; nuclear envelope; phosphorylation; proteomics Correspondence H. Otto, Freie Universita ¨ t Berlin, Institut fu ¨ r Chemie und Biochemie, Thielallee 63, D-14195 Berlin, Germany Fax: +49 30 83853753 Tel: +49 30 83856425 E-mail: hotto@chemie.fu-berlin.de *These authors contributed equally to this work (Received 17 January 2006, revised 27 April 2006, accepted 18 May 2006) doi:10.1111/j.1742-4658.2006.05329.x Although several proteins undergo tyrosine phosphorylation at the nuclear envelope, we achieved, for the first time, the identification of tyrosine-phos- phorylation sites of a nuclear-membrane protein, emerin, by applying two mass spectrometry-based techniques. With a multiprotease approach com- bined with highly specific phosphopeptide enrichment and nano liquid chromatography tandem mass spectrometry analysis, we identified three tyrosine-phosphorylation sites, Y-75, Y-95, and Y-106, in mouse emerin. Stable isotope labeling with amino acids in cell culture revealed phospho- tyrosines at Y-59, Y-74, Y-86, Y-161, and Y-167 of human emerin. The phosphorylation sites Y-74 ⁄ Y-75 (human ⁄ mouse emerin), Y-85 ⁄ Y-86, Y-94 ⁄ Y-95, and Y-105 ⁄ Y-106 are located in regions previously shown to be critical for interactions of emerin with lamin A, actin or the transcrip- tional regulators GCL and Btf, while the residues Y-161 and Y-167 are in a region linked to binding lamin-A or actin. Tyrosine Y-94 ⁄ Y-95 is located adjacent to a five-residue motif in human emerin, whose deletion has been associated with X-linked Emery–Dreifuss muscle dystrophy. Abbreviations EDMD, Emery–Dreifuss muscle dystrophy; INM, inner nuclear membrane; LC, liquid chromatography; MS ⁄ MS, tandem mass spectrometry; ONM, outer nuclear membrane; SILAC, stable isotope labeling with amino acids in cell culture. 3204 FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS [15–17], which allowed for example the identification of peptides from nuclear-envelope proteins phosphoryl- ated at serine or threonine residues [18]. However, tyrosine-phosphorylation sites have not been identified so far. While improving the search for phosphopeptides containing phosphotyrosine, we identified one of the well-defined integral INM proteins. Emerin is a type-II integral membrane protein of 34 kDa, which is inser- ted into the membrane with a single transmembrane sequence near its carboxy-terminus and is targeted to the inner nuclear membrane [19,20]. In humans, emerin is the gene product of the EMD gene that is associated with the X-chromosome-linked form of the inherited Emery–Dreifuss muscular dystrophy (EDMD), leading to slowly progressing muscle wasting and a cardiomy- opathy with conduction defects [19,21]. On the cellular level, EDMD is characterized by a mislocalization of emerin that is caused by the loss of emerin binding to lamin A (X- linked form) or by the loss of lamin A (autosomal dominant form) [22,23]. Several emerin binding partners have been detected and partial sequences required for their binding have been mapped [10,11,24–29], which will probably form different emerin complexes, whose formation may be regulated by transient phosphorylation. In this study, we describe the identification of emerin as the first tyrosine-phosphorylated nuclear-envelope protein. We have identified tyrosine phosphorylation sites on human and mouse emerin using independently two different strategies: (1) a multiprotease approach, where we combined subcellular fractionation of mouse N2a cells with in-gel digestion of emerin using a set of different proteases followed by phosphopeptide enrich- ment using the phosphopeptide affinity matrix titan- sphere [30]; and (2) stable isotope labeling with amino acids in cell culture (SILAC) in combination with antiphosphotyrosine immunoprecipitation and tryptic in-gel digestion to identify human emerin phosphory- lation sites in HeLa cells. This led to the identification of tyrosine phosphorylation sites of mouse and human emerin. Results To identify tyrosine-phosphorylated nuclear-envelope proteins and their phosphorylation sites, we used a mul- tiprotease approach on mouse cells (Fig. 1A) and the SILAC approach on human cells (Fig. 3A). The analy- sis of phosphorylated cellular proteins requires an effi- cient inhibition of endogenous protein phosphatases. This is particularly important for studying tyrosine phosphorylation, as it is highly transient due to very A B Fig. 1. Multi-protease approach. (A) Scheme of the approach. Nuc- lear envelopes were purified from BiPy-treated N2a cells (mouse neuroblastoma), and the protein mixtures separated by SDS ⁄ PAGE. An aliquot of the sample was used for western blot analysis. The pattern of tyrosine-phosphorylated nuclear-envelope proteins, visu- alized by using the phosphotyrosine-specific antibody PY99 (horse- radish peroxidase conjugate) and ECL, was used for sample selection on a Coomassie-stained reference gel. The protein bands cut from the gel were divided into four aliquots and digested with trypsin, elastase, proteinase K and thermolysin, respectively. The extracted peptides of all four digests were mixed, phosphopeptides were enriched on a titansphere column and analyzed by nanoLC- MS ⁄ MS. (B) Immunoblot of tyrosine-phosphorylated nuclear envelope proteins and the corresponding Coomassie-stained gel. According to the pattern of phosphotyrosine immunostaining (ECL), samples 1 and 2 were selected for further analysis. A. Schlosser et al. Tyrosine-phosphorylation of emerin FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS 3205 active phosphotyrosine-specific phosphatases. There- fore, dephosphorylation of phosphotyrosine was pre- vented by the addition of very potent, cell-permeable, highly specific tyrosine-phosphatase inhibitors to the cells in culture. In the case of the multiprotease approach (Fig. 1A), we added 100 lm of the tyrosine-phosphatase inhibitor BiPy [31] to the culture medium of mouse-neuroblast- oma N2a cells 10 min before harvesting the cells (BiPy-treated cells). This results in hyperphosphoryla- tion of proteins [Fig. 1B, left panel and Fig. 2, left panel (BiPy+)], which is required for a successful phosphopeptide analysis. Before starting with protein and phosphopeptide identification, we compared tyrosine-phosphorylation of nuclear-envelope proteins from control cells (no BiPy added prior to homogenization) and from BiPy- treated cells (Fig. 2). BiPy treatment should increase the amount of tyrosine phosphorylation but should not change the pattern of tyrosine-phosphorylated nuc- lear envelope proteins. To prevent, as far as possible, changes of tyrosine phosphorylation after breaking up the cells, we simultaneously added 500 nm of the broad-range protein kinase inhibitor staurosporine and 100 lm of phosphotyrosine phosphatase inhibitor BiPy (in addition to sodium vanadate and sodium molyb- date) to both control cells and BiPy-treated cells at the beginning of homogenization. Both inhibitors were then present throughout the preparation of nuclei and nuclear envelopes, although, in contrast to the phos- phatases, the kinases should not work efficiently anymore due to a lack of ATP. Then, the nuclear- envelope proteins were separated by SDS ⁄ PAGE, blot- ted onto nitrocellulose and sequentially immunostained for emerin and for phosphotyrosine. Control cells already show a weak pattern of tyro- sine-phosphorylated nuclear envelope proteins [Fig. 2, left panel (BiPy–)], with some of the phosphotyrosine immunostaining overlapping with emerin immuno- staining (Fig. 2, arrows). For BiPy-treated cells, this pattern of tyrosine-phosphorylated proteins increases in intensity but does not considerably change other- wise. This suggests that our approach of adding BiPy before harvesting the cells enhances physiologically relevant tyrosine phosphorylation of nuclear-envelope proteins, as the interaction of tyrosine kinases and sub- strate proteins is still restricted to their endogenous compartments at that point. An efficient hyperphosphorylation is achieved only, when BiPy is added before homogenizing the cells. Sta- urosporine, on the other hand, did not seem to have much influence on the tyrosine phosphorylation pat- tern after homogenization of the cells (data not shown). Therefore, we omitted staurosporine during preparation of hyperphosphorylated nuclear envelopes intended for phosphopeptide analysis. For the mass-spectrometric identification of phos- phopeptides, we purified nuclei from BiPy-treated cells, from which we obtained nuclear envelopes by digesting nucleic acids under hypo-osmotic conditions (Fig. 1B). Throughout the preparation, 100 lm BiPy and 1 mm each of sodium vanadate and sodium molybdate were present to preserve the phosphorylation obtained in the living cell immediately before homogenization. An aliquot of the protein mixture was then separ- ated by one-dimensional SDS ⁄ PAGE, blotted on nitrocellulose and immunostained with the phosphotyr- osine-specific antibody PY99. Nuclear membrane frac- tions were run in parallel on a second gel to separate the protein mixtures for mass spectrometric analysis of tyrosine-phosphorylated proteins. Figure 1B shows the pattern of tyrosine-phosphorylation (PY) for nuclear- envelope proteins (NE), and the corresponding Coo- massie-stained gel. A complex pattern of putatively tyrosine-phosphorylated proteins is visible. Regions of Fig. 2. Enhancement and preservation of tyrosine phosphorylation in nuclear envelopes from N2a cells. N2a cells in culture were either treated with 100 l M BiPy or left untreated as control cells. Then, nuclear envelopes were prepared in the constant presence of 500 n M staurosporine and 100 lM BiPy in order to preserve the phosphorylation status reached at the time of homogenization. The proteins were separated by SDS ⁄ PAGE, blotted and immuno- stained for emerin and for phosphotyrosine. In the absence of BiPy, a weak pattern of proteins phosphorylated at tyrosine residues appears, which is increased in intensity under hyperphosphorylating conditions. Arrows indicate the phosphotyrosine bands correspond- ing to the emerin bands on the left. Tyrosine-phosphorylation of emerin A. Schlosser et al. 3206 FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS the Coomassie-stained reference gel corresponding to the strongest PY99-reactivity (samples 1 and 2; Fig. 1B) were excised from the gel and further ana- lyzed. In-gel digestion of these bands was done in par- allel with four enzymes: trypsin, elastase, proteinase K, and thermolysin. The four digests were pooled; phos- phopeptides were enriched on a titansphere nano-col- umn, eluted, and analyzed by nanoLC-MS ⁄ MS. Four different proteins were detected in the two samples. In sample 1 (apparent molecular weight 35–45 kDa), LAP2, possibly the membrane isoform LAP 2 c (38.5 kDa calculated, accession number AAH64677), was identified. In sample 2 (apparent molecular weight 25–35 kDa), nucleophosmin-1 (nucleolar phosphoprotein B23, 28.4 kDa calculated, accession number NP_032748), cation- dependent mannose-6-phosphate receptor (31.1 kDa calculated, accession number NP_034879), and emerin (29.4 kDa calculated, accession number NP_031953) were present in addition to LAP2. First, surprisingly neither a phosphotyrosine-contain- ing peptide nor an emerin peptide could be detected in sample 1, although this sample should correspond to a region of phosphotyrosine-immunostaining stronger than that corresponding to sample 2. Secondly, sample 1 should also contain the upper emerin band, which most likely reflects a different, not yet characterized emerin phosphorylation state [20]. Both samples, cut from the gel, contain more than one protein. Also, it is not possible to exactly control the protein composition in such an excised gel piece. One explanation for this lack could therefore be that the phosphotyrosine immunostaining, despite overlap- ping with emerin immunostaining, may be caused by another protein. Another explanation could be a low content of tyrosine-phosphorylated peptides, for example, due to a high amount of other proteins in the sample. Also, the strength of the phosphotyrosine immunosignal may be misleading, since the affinity of phosphotyrosine-specific antibodies is always influ- enced by amino acid residues surrounding the phos- photyrosine. Finally, emerin in sample 1 could carry different phosphotyrosines that, despite using four different proteases, might be located in a sequence not suitable for mass spectrometric analysis. The method applied facilitates the identification of phosphopeptides in general. As most regulatory phosphorylation events occur at serine and threonine residues and persist longer in the cells than the highly transient tyrosine phosphorylation, their detec- tion is much more likely. It is therefore not surpri- sing that we detected in all identified proteins serine- and threonine phosphorylation sites. We found the following Ser ⁄ Thr-phosphorylation sites. Three new sites for LAP 2: (1) S-183, (2) T-316 or T-319, and (3) one in the region between T-153 and S-158) in addition to the previously identified sites [18]; for nucleophosmin, S-4, S-10, S-70, and S-125; and for A B Fig. 3. Stable isotope labeling with amino acids in cell culture (SILAC) approach. (A) Scheme for the identification of emerin phos- phorylation sites. HeLa cells were grown in two different popula- tions, one in normal medium and the other in medium containing arginine and lysine labeled with stable isotopes (described in meth- ods). The cells growing in heavy isotope medium were treated to 1m M sodium pervanadate. The cell lysates were mixed after deter- gent lysis of the cells, followed by the immunoprecipitation of tyrosine-phosphorylated proteins. The proteins were separated by SDS ⁄ PAGE. A protein band corresponding to 30 kDa was excised and digested with trypsin before analyzing the peptides by LC- MS ⁄ MS. (B) MS spectrum showing the doubly charged peptide pair (light and heavy isotope pair) with a mass shift of 6 Da, which corresponds to the unphosphorylated emerin peptide KIFEYETQR (aa residues 37–45, with and without one 13 C 6 -Arg and one 13 C 6 - Lys). The heavy peptide from the pervanadate-treated cells shows an increased intensity due to the increase of tyrosine-phosphorylat- ed emerin in these cells. A. Schlosser et al. Tyrosine-phosphorylation of emerin FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS 3207 mouse emerin, one Ser-phosphorylation site was detected on S-72. Tyrosine phosphorylation sites could only be detec- ted for emerin, which colocalizes in western blotting with a weak phosphotyrosine immunosignal in control cells and with a strong immunosignal under hyper- phosphorylating conditions (Fig. 2). The identified peptides from mouse emerin (including the S-72 phos- phorylation) are listed in Table 1. In total, 12 peptides were assigned to emerin, eight phosphopeptides and four acidic peptides. Although the phosphopeptide affinity material titansphere shows excellent selectivity for phosphopeptides, peptides with five or more acidic residues are sometimes coenriched under the applied conditions. Three tyrosine phosphorylation sites were identified: Y-75, Y-95, and Y-106. Figure 4A shows the MS ⁄ MS spectrum of the peptide DYNDD-pY-YE- ESYLTTK (aa residue 90–104, mouse emerin) as an example. Although the peptide contains four tyrosine residues, the phosphorylated tyrosine can be clearly located on Y-95 (mass difference of a phosphotyrosine residue (243.03 Da) between carboxy-terminal frag- ment ions y 9 and y 10 , which comprise the 9 and 10 carboxy-terminal amino-acid residues, respectively). The SILAC approach (Fig. 3A) is based on in vivo labeling of all the cellular proteins by isotope-coded amino acids. In addition, we used the determination of relative ratios of peptide abundance obtained from proteins isolated from hyperphosphorylated and refer- ence cells to distinguish between nonspecifically cap- tured vs. true IP-captured tyrosine-phosphorylated proteins. To achieve labeling, we added a mixture of argin- ine and lysine, each containing six 13 C atoms ( 13 C 6 - Arg and 13 C 6 -Lys), to HeLa cells in culture. The two amino acids were chosen because the tryptic protein digestion applied later in the procedure would gener- ate peptides ending with arginine and lysine residues. This ensures the generation of labeled peptides and nonlabeled but otherwise identical reference peptides. As reference, HeLa cells were grown with unlabeled arginine and lysine (Fig. 3A). The cells first were serum-starved for 12 h, prior to 1 mm sodium per- vanadate treatment for 30 min. Sodium pervanadate treatment of HeLa cells (20 large dishes) grown in the presence of the heavy amino acids 13 C 6 -arginine and 13 C 6 -lysine created a state of hyperphosphoryla- tion of all cellular proteins. For comparison, HeLa cells (20 large dishes) were grown with unlabeled arginine and lysine. In cells growing in five (i.e. one quarter) of these reference dishes, tyrosine phosphory- lation was also stimulated to generate a certain amount of tyrosine phosphorylation necessary for the final comparison of tyrosine-phosphorylated proteins Table 1. Phosphorylation sites of mouse and of human emerin. Phosphorylation site Number of phosphorylated residues Peptide sequence Residues Additional modifications Species S-72 or Y-75 1 AVDSDMYDLPKKEDAL 69–85 Oxidation (M) Mouse Y-75 1 AVDSDMYDLPKKEDA 69–84 Mouse S-72 and Y-75 2 AVDSDMYDLPKKE 69–82 Mouse S-72 and Y-75 2 AVDSDMYDLPKKEDAL 69–85 Oxidation (M) Mouse Y-75 1 MYDLPKKE 74–81 Mouse 0 DYNDDYYEE 90–98 Mouse 0 DYNDDYYEESY 90–100 Mouse 0 DYNDDYYEESYLTTK 90–104 Mouse Y-95 1 DYNDDYYEESYLTTK 90–104 Mouse Y-106 1 LTTKTYGEPES 101–111 Mouse Y-106 1 LTTKTYGEPESVGMSKS 101–117 Oxidation (M) Mouse 0 DDIFSSLEEEGKDR 138–150 Mouse 0 RYNIPHGPVVGSTR 17–31 2 13 C 6 -Arg Human 0 YNIPHGPVVGSTR 18–31 13 C 6 -Arg Human 0 KIFEYETQR 37–45 13 C 6 -Arg and 13 C 6 -Lys Human 0 IFEYETQR 38–45 Human S-49 and Y-59 2 RLSPPSSSAASSYSFSDLNSTR 47–68 2 13 C 6 -Arg Human Y-74 1 GDADMYDLPKKEDALLYQSK 69–88 Oxidation (M) 3 13 C 6 -Lys Human 0 KEDALLYQSK 79–88 Human 0 KEDALLYQSK 79–88 2 13 C 6 -Lys Human Y-85 1 KEDALLYQSK 79–88 2 13 C 6 -Lys Human Y-161 and Y-167 2 DSAYQSITHYRPVSASR 158–174 2 13 C 6 -Arg Human Tyrosine-phosphorylation of emerin A. Schlosser et al. 3208 FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS in both cell populations. The cells were lyzed, the lysates from the two states were mixed, and tyrosine- phosphorylated proteins were extracted by immuno- precipitation applying the phosphotyrosine-specific antibodies 4G10 and RC20. Proteins were eluted from the precipitated immune complexes with phenyl- phosphate, separated by SDS ⁄ PAGE and stained with colloidal Coomassie blue. Bands of stained proteins were then excised from the gel. The proteins were reduced, alkylated, and digested with trypsin within the gel matrix. The peptides extracted from the gel matrix were finally analyzed by reversed-phase liquid chromatography tandem mass spectrometry (LC- MS ⁄ MS). The sequences obtained from MS ⁄ MS spec- tra were analyzed and potential phosphopeptides con- taining tyrosine were scanned by plotting the relevant extracted ion chromatograms for the corresponding unphosphorylated peptide (80 Da mass difference for single-charged peptides). Unphosphorylated peptides could be detected for all identified tyrosine-phosphor- ylated peptides giving an additional confirmation for the correct assignment. In comparison to the reference cells, tryptic peptides from the tyrosine-phosphorylated cells labeled with 13 C 6 -Arg and 13 C 6 -Lys show up with a mass difference of 6 or multiples of 6. As a quarter of the reference Fig. 4. Identification of emerin and of tyro- sine-phosphorylation sites. (A) MS ⁄ MS spectrum of the mouse-emerin peptide DYNDD-pY-YEESYLTTK (aa residues 90– 104), phosphorylated at Y-95, which was obtained by the multiprotease approach. (B) MS ⁄ MS spectrum of the human-emerin peptide RL-pS-PPSSSAASS-pY-SFSDLNSTR (aa residues 47–68), which is phosphorylat- ed at S-49 and Y-59, obtained by using the SILAC approach. In both figure parts, the phosphotyrosine-specific mass difference between the appropriate carboxy-terminal y-ions is indicated by a bar labeled ‘pY’. A. Schlosser et al. Tyrosine-phosphorylation of emerin FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS 3209 cells were also stimulated to undergo tyrosine-phos- phorylation, peptide pairs that are tyrosine-phosphor- ylated are expected to show such a mass difference. Peptides derived from proteins that undergo tyrosine- phosphorylation due to pervanadate treatment but are not tyrosine-phosphorylated in control cells appear as pairs, which show an ion ratio (peptides from hyper- phosphorylated cells (pervanadate treatment) ⁄ peptides from control cells) close to 4. In contrast, all proteo- lytic peptides from nonspecifically captured proteins show a ratio close to 1. For the identification of tyrosine phosphorylation sites, only such peptide pairs were used, where the quantification showed a significantly increased amount of the tyrosine-phosphorylated peptides obtained from the pervanadate-treated population of cells [32]. As an example, the peptide pair corresponding to the non- phosphorylated peptide KIFEYETQR (aa residues 37– 45) of human emerin is shown in Fig. 3B. Since tyro- sine-phosphorylated emerin has been enriched by the phosphotyrosine affinity-purification step, this partic- ular peptide shows a 3.5-fold increase in signal inten- sity between the peptide from the control cells and from the pervanadate-treated cells. The heavy peptide contains one 13 C 6 -lysine and one 13 C 6 -arginine, which result in a m ⁄ z-difference of +6 for the doubly charged peptide. Analyzing all Coomassie-stained bands of the SDS ⁄ PAGE, several tyrosine-phosphorylated proteins have been identified. However, emerin (human emerin, accession number NP_000108) was the only identified protein with known localization at the inner nuclear membrane. For emerin, we identified the tyrosine residues Y-59, Y-74, Y-85, Y-161 and Y-167 as phosphorylation sites of human emerin. The observed ratios between heavy and light emerin peptides range from 2.5 to 6.5. Sim- ilar values are obtained for other tyrosine-phosphoryl- ated proteins. This fluctuation is greater than typically observed for classic SILAC experiments. However, this is not particularly relevant for our approach, since we only have to be able to distinguish between ratios close to 1 (nonspecific impurities) and ratios close to 4 (pro- teins that are tyrosine-phosphorylated upon pervana- date treatment). Ratios smaller than 4 are expected, if a protein is already partially tyrosine-phosphorylated before treatment with pervanadate. All of these tyrosines are conserved between human and mouse emerin (equivalent positions of mouse emerin are the aa residues Y-60, Y-75, Y-86, Y-161 and Y-167). Figure 4B, for example, shows the emerin peptide RL-pS-PPSSSAASS-pY-SFSDLNSTR (aa res- idues 47–68), which is phosphorylated at S-49 and Y-59. The serine phosphorylation site, unique for human emerin, reflects probably a basal phosphoryla- tion not attributed to the pervanadate treatment. All identified phosphopeptides are summarized in Table 1 and the phosphorylation sites are indicated in the emerin alignment (Fig. 5A). Fig. 5. Scheme of emerin interactions and EDMD mutations. (A) Alignment of human and mouse emerin. ‘P’ indicates identified tyrosine-phosphorylation sites. (B) Schematic representation of the binding interactions mapped onto the emerin structure. Black lines indicate the different phosphorylation sites, grey bars and a star the EDMD mutations S-54 F, Del95–99, and P-183 H ⁄ T. The numbers shown are based on the human emerin sequence. Equivalent sequence positions (human ⁄ mouse emerin) are Y-59 ⁄ Y-60, Y-74 ⁄ Y-75, Y-86 ⁄ Y-87, Y-94 ⁄ Y-95, Y-105 ⁄ Y-106, Y-161, and Y-167. LEM, LEM domain; TM, membrane-spanning sequence. Tyrosine-phosphorylation of emerin A. Schlosser et al. 3210 FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS Discussion By applying two different approaches to either lysates from human cells or to isolated mouse nuclear enve- lopes, we identified emerin as a tyrosine-phosphoryla- ted protein of the inner-nuclear membrane, which seems to be a key protein in building different com- plexes with other proteins at the nuclear envelope. For both human and mouse emerin, we were able to deter- mine a few sites that are targets of tyrosine kinase activity. In this study, the multiprotease approach and the SILAC approach complement each other. While the phosphorylation site Y-74 ⁄ 75 (human ⁄ mouse emerin) has been identified with both methods, the phosphorylation sites Y-94 ⁄ 95 and Y-105⁄ 106 have been detected only in mouse emerin and the sites Y-59 ⁄ Y-60, Y-161 and Y-167 only in human emerin. Although a different phosphorylation may result from species-specific differences of the phosphorylation machinery, it seems more likely that the different cell types with their specific kinase and phosphatase equip- ment and their differing regulatory properties account for the differences in the usage of the emerin phos- phorylation sites in mouse N2a and human HeLa cells. In addition, the different methods applied may also contribute to the differences in identified phosphoryla- tion sites. The multiprotease approach was used in a subcellular-proteomics background. As it does not rely on the comparison of emerin from differently treated cells but on the analysis of isolated emerin from a quasi homogeneous source, this approach should enable the identification of all tyrosine-phosphorylated emerin peptides provided that their quantity and their affinity to the titansphere column are sufficient to pass a detectable amount to the mass spectrometer. The SILAC approach, on the other hand, uses two differ- ent filters for identifying phosphorylation sites. First, binding to phosphotyrosine-specific antibodies is used to enrich tyrosine-phosphorylated proteins. As residues surrounding the phosphotyrosine also influence bind- ing to such antibodies, this step might favor sub- populations of differentially phosphorylated emerin. Secondly, this method filters for such peptides, which show an increase in quantity of tyrosine-phosphorylat- ed peptides from the unlabeled reference sample to the labeled tyrosine-phosphorylation sample. Although this might discriminate against peptides that may be con- siderably phosphorylated in the unlabeled reference cells, this filter was applied to prevent false positives due to nonspecific binders that would appear with the same intensity in both samples. As all identified tyro- sine-phosphorylation sites seem to be conserved in mammalian emerin, they could as well be used simi- larly in all species for differentially regulating the diverse interactions demonstrated for emerin. Emerin is the product of a gene linked to EDMD [19,21]. An integral membrane protein specifically loca- ted at the inner nuclear membrane, emerin, like other INM proteins, binds to lamins. It is linked to EDMD by its interaction with lamin A [20]. In EDMD this interaction is weakened or lost either by mutations in emerin itself, which leads to the X-linked form of EDMD [20,25,33,34], or by a loss of lamin A, which causes the autosomal-dominant form of EDMD [22,23,35]. In both forms, emerin is mislocalized [36] and cannot efficiently accumulate at the inner nuclear membrane [37]. For several proteins shown to interact with emerin, binding regions were mapped onto the emerin sequence (Fig. 5B). The best characterized of these interactions occurs between the DNA ⁄ protein com- plexes of the heterochromatin protein BAF and emer- in’s amino-terminal LEM domain (aa residues 2–44) [38], which similarly exists in the INM proteins MAN1 and LAP2 [24,39]. This interaction is also necessary for a correct localization after mitosis [37]. Recently, Hirano and coworkers identified with a similar approach four serine and one threonine residues of human emerin that in vitro are strongly phosphoryla- ted in M-phase extracts prepared from Xenopus eggs. They showed that phosphorylation of one site, S-175, causes BAF dissociation from emerin. Remarkably, this site is at the primary structure level far away from the N-terminal LEM-domain that is thought to medi- ate this interaction [40]. Lamin-A binding has been mapped to the central region of emerin (aa residues 70–178) [38], which overlaps with a region capping actin filaments [10,26]. Binding of transcriptional regu- lators like GCL and the death-promoting transcrip- tional repressor Btf or the splicing factor YT521-B involves emerin sequences on both sides of the central region that interacts with cytoskeletal elements, parti- ally overlapping with the LEM domain as well as with the lamin-A binding region [11,27,41]. Not as well characterized but important in the context of a muscle dystrophy is a muscle-specific interaction with the actin-binding spectrin-repeat proteins nesprin 1a and nesprin 2 [28,29]. Since the different binding regions overlap at least partially, a simultaneous binding of some binding part- ners may be excluded, giving rise to distinct emerin complexes [41]. Cells may be able to control the forma- tion of such complexes by phosphorylating critical resi- dues of emerin. In fact, the disruption of BAF binding A. Schlosser et al. Tyrosine-phosphorylation of emerin FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS 3211 to emerin in mitosis seems to be mediated by phos- phorylation of S-175 of emerin [40]. Interestingly, the identified tyrosine phosphorylation sites all are located near regions that have been proven as disease-related. Several mutations in human emerin have been linked to EDMD. Among these are the mis- sense mutations S54F, P183H, P183T, and a deletion of five amino acids (Del 95–99, YEESY), which hint at regions critical for regulated complex formation [34,37]. S54F and Del 95–99 disrupt indeed the Btf binding to emerin. Del 95–99 also disrupts the interac- tion with lamin A and GCL [27]. The tyrosine-phosphorylation sites Y-59 ⁄ Y-60 (human ⁄ mouse emerin) and Y-74 ⁄ Y-75 are exactly positioned in a region of overlapping binding sites and could help to control interactions with lamin A, actin and the transcriptional regulators GCL and Btf. For comparing mouse and human emerin sequences, note that the sequence alignment shown for mouse emerin has an additional amino acid residue after residue 57 and a gap after residue 138. Thus, equivalent amino acid residues of human and mouse emerin differ in numbering by one between these sequence positions (Fig. 5A). The phosphorylation sites Y-94 ⁄ Y-95 (human ⁄ mouse emerin) and Y-105 ⁄ Y-106 are even more directly linked to an EDMD mutation. Y-94 ⁄ 95 is almost directly affected by the EDMD-linked dele- tion mutation Del 95–99 of human emerin, while Y105 ⁄ 106 seems near enough to this region to influ- ence emerin binding to other proteins. Phosphorylation of both tyrosine residues may differentially influence emerin binding to actin and lamin A. The phosphorylated tyrosine residues Y-161 and Y-167, identified in human emerin, are located within a region of emerin that has been linked to lamin A ⁄ ac- tin or to lamin-A binding, respectively, which is likely to be regulated by such phosphorylation. Ellis et al. [20] have shown that different phosphorylation states of emerin are linked to EDMD [20]. Four different cell cycle-dependent phosphorylation states have been iden- tified. In patients, mutated emerin forms, which show a changed solubility or extractability compared with wild-type emerin, undergo also aberrant phosphoryla- tion. The corresponding phosphorylation sites, how- ever, remained elusive. If and how the phosphorylation sites determined here may be correlated to those aberrant phosphoryla- tion states described for EDMD patients and how these sites take part in intermolecular interactions, remains to be investigated. The analysis of phosphory- lation sites that we present here requires hyperphos- phorylation conditions. Phosphorylation, however, was initiated in both approaches by already adding a cell-permeable tyrosine-phosphatase inhibitor based on pervanadate before opening up the cells. The compar- ison of the tyrosine-phosphorylation status of nuclear envelopes from BiPy-treated and control cells (Fig. 2) indicates that the regular pattern of tyrosine phos- phorylation of the control cells is just enhanced in the presence of the pervanadate compounds. Also, the ATP concentration drops after homogenization to lev- els that do not allow any further phosphorylation. Therefore, the phosphorylation at the determined sites occurred already in the living cells, while emerin was still in its native environment, involved in its normal molecular interactions. Thus, the determined sites most likely reflect sites that are used under physiological conditions. Experimental procedures Multi-protease approach: sample preparation Mouse-neuroblastoma N2a cells were cultured in Dulbec- co’s modified Eagle’s medium containing 10% fetal bovine serum, 100 mgÆmL )1 streptomycin, and 100 mgÆmL )1 peni- cillin at 37 °C in a humidified atmosphere with 5% CO 2 . Nuclei and nuclear envelopes were prepared from N2a cells [42]. Ten minutes before the cells were harvested for nuclear preparation, the selective tyrosine-phosphatase inhibitor potassium-(2,2¢-bipyridine)-oxobisperoxovanadate (BiPy) [31] was added to the culture medium at a concen- tration of 100 lm. Throughout the purification, BiPy (100 lm), as well as the phosphatase inhibitors sodium vanadate (1 mm) and sodium molybdate (1 mm), were present to prevent dephosphorylation of the proteins. Tyrosine-phosphorylated proteins were separated by SDS ⁄ PAGE, blotted onto a nitrocellulose blot membrane [43,44] and visualized on the membrane by enhanced chemi- luminescence (ECL), using the phosphotyrosine-specific antibody PY99 conjugated to horseradish peroxidase (BD Biosciences-Pharmingen, Heidelberg, Germany) [42]. Like- wise, blotted protein mixtures were probed for emerin by applying the antibody Emerin (FL-254) (Santa Cruz Bio- technology, Inc., Heidelberg, Germany). When staurosporine was used, 500 nm were added at the time of homogenization together with the 100 lm BiPy. Staurosporine and BiPy were then kept present throughout the preparation. SILAC: sample preparation Human cervical carcinoma (HeLa) cells were grown in Dul- becco’s modified Eagle’s medium containing ‘light’ arginine and lysine or 13 C 6 -arginine and 13 C 6 -lysine supplemented with 10% dialyzed fetal bovine serum plus antibiotics [32]. The cells were grown for five passages in the above medium Tyrosine-phosphorylation of emerin A. Schlosser et al. 3212 FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS prior to initiating these experiments. HeLa cells were serum starved for 12 h before treatment with 1 mm pervanadate for 30 min and were subsequently lysed in a modified RIPA buffer (50 mm Tris ⁄ HCl, pH 7.4, 150 mm NaCl, 1 mm EDTA, 1% Nonidet P-40, 0.25% sodium deoxycholate, and 1 mm sodium orthovanadate in the presence of prote- ase inhibitors). Light and heavy cell lysates were precleared with protein A-agarose, mixed and incubated with 400 lg of 4G10 monoclonal antibodies coupled to agarose beads and 75 lg of RC20 antibodies, overnight at 4 °C. Precipita- ted immune complexes were washed three times with lysis buffer and then eluted three times with 100 mm phenyl- phosphate in lysis buffer at 37 °C. The eluted phosphopro- teins were dialyzed and resolved by 10% SDS ⁄ PAGE. The gels were stained using colloidal Coomassie stain. Proteolytic digestion for MS analysis Proteins were excised from the gel. In the multiprotease approach, the excision was guided by the pattern of the immunoblot of an identical reference gel. The gel pieces were destained with 30% acetonitrile. After reduction and alkylation of the proteins, the gel pieces were dehydrated with 100% acetonitrile and dried in a vacuum centrifuge. In the multiprotease approach, the proteins were digested in parallel with trypsin, elastase, proteinase K, and thermo- lysin (about 0.1 lg of each protease) in 0.1 m NH 4 HCO 3 (pH 8) at 30 °C overnight. Peptides were then extracted from the gel slices with 5% formic acid. All supernatants and extracts were combined, dried in a vacuum centrifuge, and redissolved in 10 lL of 30% acetonitrile and 2% for- mic acid. Phosphopeptides were enriched on nano-columns (inner diameter: 50 lm; length: 1.5 cm) packed with titan- sphere (5 lm particles, GL Sciences Inc., Tokyo, Japan), which were washed with 30% acetonitrile, 2% formic acid, and were eluted with 10 lL of 0.1 m NH 4 HCO 3 (pH 9). The eluate was acidified by mixing with formic acid and was analyzed with nanoLC-MS ⁄ MS using a reversed-phase column with an inner diameter of 25 lm [30]. In the SILAC experiment, the 30 kDa band correspond- ing to emerin was digested by trypsin using an in-gel diges- tion protocol. The peptides were extracted as described above and the peptide mixture was analyzed by reversed- phase LC-MS ⁄ MS [32]. Mass spectrometry All mass spectra were recorded on a quadrupole time-of- flight tandem mass spectrometer, type Q-TOF (Waters Micromass, Manchester, UK), equipped with a nanoESI source. The parameters for data-dependent MS ⁄ MS were set to fragment up to three precursors at a time (charge states +1 to +4). The intensity threshold for precursor selection was set to 20 countsÆs )1 , which indicates sufficient ion intensity for recording MS ⁄ MS spectra. The scan time for MS ⁄ MS spectra was set to 2 s using two different colli- sion offsets. A Mascot Server (Matrix Science, London, UK) was used for database searching, as follows. (1) Multi- protease approach: the mass tolerance was set to ± 0.1 Da for both precursor mass and fragment ion mass. Searches were performed in SwissProt without protease specificity and without any taxonomic restrictions [30]. (2) SILAC: searches with tryptic peptides were done in RefSeq (http:// www.ncbi.nlm.nih.gov/RefSeq/) with a mass tolerance of 0.3 Da and up to two missed tryptic cleavages [32]. Phosphopeptides identified by the search engine Mascot have been verified by manual inspection of the MS ⁄ MS spectra. Acknowledgements RA would like to thank Dr Akhilesh Pandey and Dr Dario Kalume, Institute of Genetic Medicine, Johns Hopkins University, Baltimore, MD, USA for finan- cial support and help on ESI-qTOF mass spectro- meter, respectively, and for fruitful scientific discussions. HO is grateful for all the support provided by Dr Ferdinand Hucho and his laboratory at the Freie Universita ¨ t Berlin, Germany. References 1 Holaska JM, Wilson KL & Mansharamani M (2002) The nuclear envelope, lamins and nuclear assembly. Curr Opin Cell Biol 14, 357–364. 2 Gant TM & Wilson KL (1997) Nuclear assembly. Annu Rev Cell Dev Biol 13, 669–695. 3 Dreger M, Bengtsson L, Schoneberg T, Otto H & Hucho F (2001) Nuclear envelope proteomics: novel integral membrane proteins of the inner nuclear mem- brane. Proc Natl Acad Sci USA 98, 11943–11948. 4 Schirmer EC, Florens L, Guan T, Yates JR 3rd & Ger- ace L (2003) Nuclear membrane proteins with potential disease links found by subtractive proteomics. Science 301, 1380–1382. 5 Nikolakaki E, Simos G, Georgatos SD & Giannakouros T (1996) A nuclear envelope-associated kinase phos- phorylates arginine-serine motifs and modulates interac- tions between the lamin B receptor and other nuclear proteins. J Biol Chem 271, 8365–8372. 6 Polioudaki H, Kourmouli N, Drosou V, Bakou A, Theodoropoulos PA, Singh PB, Giannakouros T & Georgatos SD (2001) Histones H3 ⁄ H4 form a tight complex with the inner nuclear membrane protein LBR and heterochromatin protein 1. EMBO Rep 2, 920–925. 7 Nili E, Cojocaru GS, Kalma Y, Ginsberg D, Copeland NG, Gilbert DJ, Jenkins NA, Berger R, Shaklai S, Amariglio N et al. (2001) Nuclear membrane protein A. Schlosser et al. Tyrosine-phosphorylation of emerin FEBS Journal 273 (2006) 3204–3215 ª 2006 The Authors Journal compilation ª 2006 FEBS 3213 [...]... Morris GE (2003) Emerin interacts in vitro with the splicingassociated factor, YT521-B Eur J Biochem 270, 2459– 2466 Lee KK, Starr D, Cohen M, Liu J, Han M, Wilson KL & Gruenbaum Y (2002) Lamin-dependent localization of UNC-84, a protein required for nuclear migration in Caenorhabditis elegans Mol Biol Cell 13, 892–901 Hunter T (1995) Protein kinases and phosphatases: the yin and yang of protein phosphorylation... (1996) Emerin deficiency at the nuclear membrane in patients with Emery–Dreifuss muscular dystrophy Nat Genet 12, 254–259 Ellis JA, Yates JR, Kendrick-Jones J & Brown CA (1999) Changes at P183 of emerin weaken its protein protein interactions resulting in X-linked Emery–Dreifuss muscular dystrophy Hum Genet 104, 262–268 Clements L, Manilal S, Love DR & Morris GE (2000) Direct interaction between emerin. .. to the Emery–Dreifuss muscular dystrophy phenotype J Cell Sci 115, 341–354 Lee KK, Haraguchi T, Lee RS, Koujin T, Hiraoka Y & Wilson KL (2001) Distinct functional domains in emerin bind lamin A and DNA-bridging protein BAF J Cell Sci 114, 4567–4573 Tyrosine-phosphorylation of emerin 39 Lin F, Blake DL, Callebaut I, Skerjanc IS, Holmer L, McBurney MW, Paulin-Levasseur M & Worman HJ (2000) MAN1, an inner... 82, 143–153 Martins S, Eikvar S, Furukawa K & Collas P (2003) HA95 and LAP2beta mediate a novel chromatin nuclear envelope interaction implicated in initiation of DNA replication J Cell Biol 160, 177–188 Holaska JM, Kowalski AK & Wilson KL (2004) Emerin caps the pointed end of actin filaments: evidence for an actin cortical network at the nuclear inner membrane Plos Biol 2, E231 Wilkinson FL, Holaska... Kendrick-Jones J & Ellis JA (1999) The Emery–Dreifuss muscular dystrophy phenotype arises from aberrant targeting and binding of emerin at the inner nuclear membrane J Cell Sci 112, 2571–2582 Lattanzi G, Cenni V, Marmiroli S, Capanni C, Mattioli E, Merlini L, Squarzoni S & Maraldi NM (2003) Association of emerin with nuclear and cytoplasmic actin is regulated in differentiating myoblasts Biochem Biophys... cell-less (GCL) and barrier to autointegration factor (BAF) compete for binding to emerin in vitro J Biol Chem 278, 6969– 6975 42 Otto H, Dreger M, Bengtsson L & Hucho F (2001) Identification of tyrosine-phosphorylated proteins associated with the nuclear envelope Eur J Biochem 268, 420–428 43 Towbin H, Staehelin T & Gordon J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose... & Worman HJ (2000) MAN1, an inner nuclear membrane protein that shares the LEM domain with lamina-associated polypeptide 2 and emerin J Biol Chem 275, 4840–4847 40 Hirano Y, Segawa M, Ouchi FS, Yamakawa Y, Furukawa K, Takeyasu K & Horigome T (2005) Dissociation of emerin from barrier-to-autointegration factor is regulated through mitotic phosphorylation of emerin in a xenopus egg cell-free system J... K, Stewart CL & Burke B (1999) Loss of A-type lamin expression compromises nuclear envelope integrity leading to muscular dystrophy J Cell Biol 147, 913–920 Cai M, Huang Y, Ghirlando R, Wilson KL, Craigie R & Clore GM (2001) Solution structure of the constant region of nuclear envelope protein LAP2 reveals two LEM-domain structures: one binds BAF and the other binds DNA EMBO J 20, 4399–4407 Fairley.. .Tyrosine-phosphorylation of emerin 8 9 10 11 12 13 14 15 16 17 18 19 20 21 A Schlosser et al LAP2beta mediates transcriptional repression alone and together with its binding partner GCL (germ-cell-less) J Cell Sci 114, 3297–3307 Lang C & Krohne G (2003) Lamina-associated polypeptide 2beta (LAP2beta) is contained in a protein complex together with A- and B-type lamins Eur J Cell Biol... localization of emerin: new insights into Emery–Dreifuss muscular dystrophy Hum Mol Genet 6, 2257–2264 Ellis JA, Craxton M, Yates JR & Kendrick-Jones J (1998) Aberrant intracellular targeting and cell cycledependent phosphorylation of emerin contribute to the Emery–Dreifuss muscular dystrophy phenotype J Cell Sci 111, 781–792 Bione S, Maestrini E, Rivella S, Mancini M, Regis S, Romeo G & Toniolo D (1994) Identification . attached to the lamina by integral membrane proteins of the INM. The INM proteins form complexes, transiently or stably, with lamins, chromatin proteins and. the inner nuclear membrane, emerin, like other INM proteins, binds to lamins. It is linked to EDMD by its interaction with lamin A [20]. In EDMD this interaction