Proteome analysis of differentiating humanmyoblasts by dialysis-assisted two-dimensional gelelectrophoresis (DAGE)
Florence Gonnet1, Belaid Bouazza2, Gal Armel Millot3, 4, Simin Ziaei5, Luis Garcia5,Gillian S. Butler-Browne2, Vincent Mouly2, Jeanine Tortajada1, Olivier Danos6 andFdor Svinartchouk5
1 Laboratoire Analyse et Modlisation pour la Biologie et lEnvironnement (LAMBE),Universit dEvry Val dEssonne, CNRS UMR 8587, Evry, France
2 UMR-S 787 Inserm, Institut de Myologie, Universit Pierre et Marie Curie Paris VI, Paris, France3 UMR 7147, CNRS, Institut Curie, Paris, France4 Universit Pierre et Marie Curie, 4 place Jussieu, France5 GENETHON, CNRS UMR 8115, Evry, France6 Hpital Necker Enfants Malades, INSERM U781, Paris, France
In the present study, modifications in cytosolic expressed proteins during human myoblast dif-ferentiation were studied by dialysis-assisted 2-DE (DAGE, ). About 1000 spots were analysedon the 5th and 13th day of differentiation with a dynamic range of protein expression exceeding1000-fold. During myogenic differentiation, the number of nonmatching spots as well as theextent of quantitative differences between matched spots significantly increased. Over one hun-dred differentially expressed spots were excised and identified by MALDI-TOF MS. The differ-entiation-associated expression pattern of eight proteins was validated by Western blot analysis.Differential expression of several proteins was demonstrated for the first time in human myo-tubes. Interestingly, Ingenuity pathway analysis grouped 30 of these proteins into two over-lapping networks containing as principal nodes IGF-1 and tumour necrosis factor, two proteinsknown to play a crucial role in cytogenesis. Our results illustrate the large rearrangement of theproteome during the differentiation of human myoblasts and provide evidence for new partnersinvolved in this complex process.
Received: March 20, 2007Revised: July 23, 2007
Accepted: October 12, 2007
Keywords:2-D PAGE / Differential expression / Human / Mass spectrometry / Protein identification
264 Proteomics 2008, 8, 264278
In mature muscle most satellite cells are in a quiescent state,but will become activated and proliferate in response to
extrinsic signals. The process of myogenesis from com-mitted muscle precursor cells to differentiated myotubes hasbeen studied in a number of cell and tissue culture models.As a result of these studies, key molecular aspects of thismultistep process have been elucidated [2, 3]. Previous stud-ies have mainly concentrated on the contractile proteinswhich are the major proteins expressed by the differentiatedmyotubes. However, no proteomic study of differentiation ofhuman myoblasts has been published to date and this will beessential for studies on muscle disease and ageing.
In the present study, we have applied a recently pub-lished modification of the classical 2-DE technique (dialysis-assisted 2-DE (DAGE), ) in order to investigate the mod-
Correspondence: Dr. Fdor Svinartchouk, GENETHON, 1 bis ruede lInternationale 91002 Evry, FranceE-mail: firstname.lastname@example.orgFax:133-(0)-1-69-47-28-38
Abbreviations: DAGE, dialysis-assisted 2-DE; Mb, myoblasts;MudPIT, multidimensional protein identification; M5, myotubesat the 5th day of differentiation; M13, myotubes at the13th dayof differentiation
2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com
Proteomics 2008, 8, 264278 Cell Biology 265
ification in the cytosolic proteins which occurs during dif-ferentiation of human myoblasts isolated from skeletalmuscle biopsies and cultured in vitro. In the absence ofserum, proliferating myoblasts withdraw from the cell cycleand differentiate into polynucleated myotubes, mimickingthe first steps of muscular development and regeneration. Comparative proteomic studies have been performed fordifferentiated mouse myoblasts using 2-DE  and multi-dimensional protein identification (MudPIT) shotgun pro-filing . About 2000 proteins were identified in thesestudies. However, a surprisingly small number of proteinswas found to be regulated in the study of Tannu et al. using a 2-DE-based proteomic method. Using the MudPITanalysis and the total cumulative spectral count recordedfrom each protein as a semiquantitative measure of relativeabundance, Kislinger and colleagues  detected changes inthe abundance of hundreds of proteins linked to cell adhe-sion, intracellular signalling, metabolism and muscle con-traction.
DAGE substantially improves spot matching and dimin-ishes quantitative variations when comparing complex pro-tein mixtures . In the present study using this methodabout 1000 spots from human myoblasts and myotubes wereanalysed on the 5th and 13th days of differentiation with thedynamic range of protein expression exceeding 1000-fold.During the differentiation process the number of non-matching spots as well as the extent of quantitative differ-ences between matched spots significantly increased. Overone hundred differentially expressed spots were excised andidentified by MALDI-TOF MS and the expression of eight ofthese proteins was confirmed by Western blot analysis.
2 Materials and methods
2.1 Equipment, chemicals and antibodies
The Protean IEF cell for IEFand criterion cell for SDS-PAGE,IGP strips (11 cm 310 nonlinear), BioRad Bradford proteinassay, Coomassie R250, mineral oil, DTT and thiourea werepurchased from BioRad. Carrier ampholytes (3/5 and 6/11)were purchased from Amersham Pharmacia Biotech.CHAPS and Triton X-100 were purchased from Sigma.Amidosulphobetaine (ASB-14) was purchased from Calbio-chem. Complete protease inhibitor cocktail tablets were pur-chased from Roche. Benzonase was purchased from Merck.
Mouse anti-annexin I mAb, rabbit polyclonal antibodiesagainst alpha-actinin, TrxR (H-200), creatine-kinase M(N-13), factor XIII (C-20), DJ-1 (N-20), DDAH II (N-20) werepurchased from Santa Cruz (CA, USA). Mouse antidesminantibodies D33, was purchased from DAKO (Denmark).Rabbit anti-STAT1 antibodies were provided by New EnglandBiolabs (Beverly, MA, USA).
2.2 Sample preparations
2.2.1 Cell culture and harvesting
The isolation and culture of human muscle cells were carriedout in agreement with the French legislation on ethical rules.The human primary myoblast strain MCQ31 was derivedfrom a quadriceps biopsy of a stillborn boy that was born at31 wks of gestation. The myoblasts were isolated from biop-sies as described previously . Briefly, the muscle was finelyminced and 2550 explants were plated onto noncoated Petridishes (diameter 60 mm) first overnight in serum drops andthen in 2 mL of growth medium consisting of DMEM con-taining 0.1 g/L gentamicin and 20% FCS (Gibco). Serumwas selected to optimise the growth capacity of human myo-blasts and maintain myogenicity. Human myoblasts werecultivated at 377C in a humid atmosphere containing 7%CO2. Once the mononucleated cells had migrated out fromthe explants (between 2 and 3 days), they were removed bytrypsinisation (1.5% trypsin, 0.04% EDTA) and replanted at1.86103 cells/cm2 in 5 mL of the same growth medium in60 mm dishes. These human myoblasts were routinely pro-pagated in DMEM (Gibco BRL), supplemented with 20% v/vFCS (Hyclone), 100 U/mL penicillin and 100 mg/mL strep-tomycin. To induce the myoblasts to differentiate into myo-tubes, the cells were placed in the same medium withoutserum, supplemented with 10 mg/mL of insulin and 100 mg/mL of apotransferrin. All cultures were incubated at 377Cunder a humidified atmosphere of air with 7% CO2. Initialplating density was between 26103 and 56103 cells/cm2.The myogenic purity of the human myoblast cultures wasmonitored by immunocytochemistry using an antibody spe-cific for desmin.
2.2.2 Protein extractions
Cells were rinsed five times with PBS and then 56106 cellswere lysed in 1 mL of homogenisation buffer (20 mM Tris-HCl, pH 8.0, 0.32 M sucrose, 3.0 mM CaCl2, 2.0 mM Mgacetate, 0.1 mM EDTA, 1% Triton X-100, 1 mM DTT and1 mM sodium vanadate in the presence of a protease in-hibitor cocktail). The homogenate was centrifuged for10 min at 20006g to remove nuclei and then centrifugedfor 40 min at 25 0006g. The supernatant was then incu-bated with benzonase (final concentration 100 U/mL) for20 min at room temperature and precipitated by adding 1.2volume of precooled acetone for at least 4 h at 07C. Theprecipitate was collected by centrifugation at 20 0006g for10 min; the pellet was washed once with methanol andstored under methanol at 207C. Protein estimation wasperformed using the commercial Bradford reagent usingBSA as a protein standard. On the day of utilisation, thesample was collected by centrifugation and the pellet wasdried and resuspended in IEF buffer (final protein con-centration 2 mg/mL).
2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com
266 F. Gonnet et al. Proteomics 2008, 8, 264278
2-DE was performed as described previously . IEF wasperformed using a Protean IEF cell and its 17 cm tray. pH 310 nonlinear strips (11 cm) were rehydrated for at least 6 h ina rehydration buffer consisting of 7 M urea, 2 M thiourea,1% v/v Triton X-100, 1% w/v ASB, 2% w/v CHAPS and20 mM DTT (IEF buffer). IEF for each pair of samples wasconducted on two half-strips obtained from one strip cutlongitudinally. This was done in order to avoid differences inthe IPG. The sample (100 mg of protein solubilised in 50 mLof the rehydration buffer containing 2% v/v IPG buffer 35and 2% v/v IPG buffer 611 (Amersham)) was loaded on awick cut from GF/Fglass microfibre filter (Whatman) 2.5 cmlong and 3 mm wide. Long wicks required for this adaptationwere cut 2.0 cm long by 0.5 cm wide from Whatman 3 mmpaper. Anode wicks were saturated in IEF buffer supple-mented with 5 mM acetic acid and cathode wicks were satu-rated with the same buffer supplemented with 10 mMlysine/arginine. The wicks were patted dry and placed in theIEF cell in the following order: (i) anode wick with one endoverlapping the platinum wire electrode and the other endoverlapping the sample loaded wick; (ii) sample wick over-lapping with the anode end of the strip; (iii) cathode wickwith one end overlapping the platinum wire electrode andthe other end overlapping with the cathode end of the IGPstrip. In order to run the two samples under identical condi-tions this cup-loading construct was covered with a 4 mmwidth dialysis membrane (Spectra/Por, MWCO 3500, Spec-trum Laboratories) and the second sample was loaded onexactly the same cup-loading construct on the surface of themembrane. In this construct two strips are placed face-to-face with a dialysis membrane between them. IEF was per-formed using the following voltage program: 50 V constantfor 5 h, linear ramp to 4000 Vover 6 h, then 8000 V for a totalof 30 000 V?h using the Protean IEF cell (BioRad). The cur-rent was limited to 50 mA per strip and the temperaturemaintained at 187C.
For SDS-PAGE, the IPG strips were incubated in equili-bration buffer containing 37.5 mM Tris-HCl (pH 8.8), 6 Murea, 4% w/v SDS, 20% glycerol and 20 mM DTT for 20 min.The equilibrated IPG strips were transferred for the seconddimension onto Criterium 1020% linear gradient gel(BioRad). Electrophoresis was carried out at room tempera-ture using a Criterium gel system (BioRad) with runningbuffer (25 mM Tris, 192 mM glycine), containing 0.1% w/vSDS, at 150 V until the front of fast migrating ions reachedthe bottom of the gel. Upon completion of SDS-PAGE, gelswere stained for 34 h with gentle agitation in 0.02% CBBsolution containing 5% acetic acid and 10% ethanol and thendestained by several changes of 5% acetic acid.
2.2.4 2-D gel image analysis
A region on the gels was selected corresponding to pI 58and MW from 15 to about 250 kDa. We have selected this
region to avoid known problem areas in 2-DE gels becausethe goal is to generate the most reproducible data as abenchmark for further studies and because the majority ofthe mouse myoblast and myotube proteins were found on2-DE gels in these pI and MW intervals . 2-D gels werecompared using Image Master Platinum software (Amers-ham). Spot detection settings were adjusted so that a max-imum of visible spots was detected. Spot volumes and per-centage of spot volumes were calculated using Image MasterPlatinum build-in options. Spot detection parameters weresettled as follows: smooth 1, minimal area 40 and saliency 70. These parameters give about 900 spots in the selectedregion of 767 cm. An automated matching algorithm withfive manually settled landmarks was used to match spots onthe gels. Landmarks were selected from all regions of the geland were only assigned when a spot was clearly present in allanalysed gels. After automatic spot matching, additional spotediting and matching were performed manually. The sameperson performed all assessments across the experiment.After spot matching the following parameters were takeninto account to estimate proteomics variations during myo-blasts differentiation: percent of spots matched which isindicative of qualitative differences among gels and the nor-malised distribution of the log2 ratio of the percentage ofspot volumes for each matching pair of spots in the two gels,which reflects the quantitative differences (see below). Foreach condition, several pairs (three to four) of DAGE wereperformed until two of them showing a good spot focalisa-tion were obtained. These two pairs were then retained forfurther analyses.
2.2.5 Statistical analyses
The R environment was used for all the analyses . Statis-tical significance was set at p0.05. Pairs of 2-D gels wereperformed in duplicated for each technical (DAGE or tradi-tional) and biological (myoblasts (Mb), myotubes at the 5thday of differentiation (M5) or myotubes at the 13th day ofdifferentiation (M13)) conditions. The proportions of mat-ched spots between conditions were compared with a w2 testwithout continuity correction (Figs. 3 and 5B). The mostunfavourable pair was chosen in each test (for example, inthe comparison between Mb/M5 and Mb/M13 (Fig. 3B), thetest was performed with the two grey bars which correspondto the minimal distance between the two conditions, mean-ing that in the three other pairs, the statistical result will bemore significant). Because significant differences wereexpected, the three w2 tests were penalised using the Holmprocedure.
Variations of spot intensities between pairs of gels werestudied by a method similar to chip analysis. The log2 ratioof the percentage of spot volumes for each matching pair ofspots in the two gels was established respecting the followingrule: myoblasts derived percentage of spot volumes were inthe numerator and differentiating cell derived percentage ofspot volumes in the denominator (Mb/M5 for 5 days of dif-
2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com
Proteomics 2008, 8, 264278 Cell Biology 267