17

79

Research ArticleReceived: 14 July 2012 Revised: 8 September 2012 Accepted article published: 1 November 2012 Published online in Wiley Online Library: 29 November 2012

(wileyonlinelibrary.com) DOI 10.1002/jsfa.5977

Protein extraction and two-dimensional gelelectrophoresis of proteins in the marinemussel Mytilus galloprovincialis: an importanttool for protein expression studies, foodquality and safety assessmentAlexandre Campos,a∗ Maria Puerto,b Ana Prieto,b Ana Camean,b Andre MAlmeida,c,d Ana V Coelhod and Vitor Vasconcelosa,e

Abstract

BACKGROUND: Shellfish farming is an important economic activity that provides society with a valuable source of food.Analyses of the protein content and metabolism of shellfish are therefore of utmost importance to monitor the presence andeffects of environmental contaminants in these organisms and also to assess food quality and authenticity. The aim of thepresent study was to compare different protein extraction protocols commonly used in two-dimensional gel electrophoresis(2DE) research and select the most suitable for the analysis of gill and digestive gland proteomes from the marine musselMytilus galloprovincialis.

RESULTS: High-resolution protein separation was achieved by direct solubilisation of proteins from M. galloprovincialis tissueswith urea (7 mol L−1), thiourea (2 mol L−1), CHAPS (40 g L−1), DTT (65 mmol L−1) and ampholytes (pH 4–7, 8 mL L−1). Subsequentprotein identification from 2DE gels by MALDI-TOF/TOF mass spectrometry revealed a high number of proteins with functions incytoskeleton structure, dynamics and maintenance. Other proteins identified in the 2DE gels are involved in energy productionand carbohydrate metabolism, metal transport, chaperones and stress response, cell signalling and regulation, proteolysis andprotein transduction.

CONCLUSION: Important protein markers for contaminant and quality assessment of shellfish food products can be analysedusing 2DE.c© 2012 Society of Chemical Industry

Supporting information may be found in the online version of this article.

Keywords: Mytilus galloprovincialis; food safety; proteomics

INTRODUCTIONProteomics is a research field focused on the analysis of proteinfractions expressed by organisms, tissues and cells, contributingto a better knowledge of the biochemistry and physiologyof organisms challenged by a given environmental stimulus.Proteomics and mass spectrometry have gained importance inareas such as environmental and aquatic toxicology, allowingone to infer contaminants present in the environment byinterpretation of the response of organisms at the protein level.Such investigation relies on the identification of robust and specificpollution biomarkers.1 Two-dimensional gel electrophoresis(2DE) separates proteins according to their isoelectric pointand molecular mass.2 2DE combined with mass spectrometryconstitutes a powerful approach for high-throughput proteomicinvestigations.

Bivalves are filter-feeding organisms with a sessile natureand a widespread distribution.3 They can bioaccumulate water

∗ Correspondence to: Alexandre Campos. Centro Interdisciplinar de InvestigacaoMarinha e Ambiental, CIIMAR/CIMAR, Universidade do Porto, P-4050-123 Porto,Portugal. E-mail: acampos@ciimar.up.pt

a Centro Interdisciplinar de Investigacao Marinha e Ambiental, CIIMAR/CIMAR,Universidade do Porto, P-4050-123, Porto, Portugal

b Area of Toxicology, Faculty of Pharmacy, University of Seville, E-41012, Seville,Spain

c Instituto de Investigacao Cient ıfica Tropical, Lisboa, Portugal and CIISA-CentroInterdisciplinar de Investigacao em Sanidade Animal, P-1300, Lisboa, Portugal

d Instituto de Tecnologia Quımica e Biologica, Universidade Nova de Lisboa,P-2780-157, Oeiras, Portugal

e Departamento de Biologia, Faculdade de Ciencias, Universidade do Porto,P-4169-007, Porto, Portugal

J Sci Food Agric 2013; 93: 1779–1787 www.soci.org c© 2012 Society of Chemical Industry

17

80

www.soci.org A Campos et al.

contaminants and are extremely responsive to alterations in theenvironment. They are therefore considered ‘sentinel’ species ofaquatic pollution. Since some of these species, e.g. the marinemussels Mytilus edulis and Mytilus galloprovincialis, are exploitedas a food source, it is of utmost importance to monitor the qualityand safety of this food product. Proteomics may also play arole here, providing valuable protein markers for the presenceof several environmental contaminants and for food qualityassessment. For example, a group of cytosolic proteins, analysedby 2DE and mass spectrometry, has been related to the exposureof the clam Chamaelea gallina to several model contaminants(Aroclor 1254, copper, tributyltin, arsenic).3 Moreover protein post-translational modification profiles (ubiquitination, carbonylation,glutathionylation, disruption of disulfide bridges) were shown tobe linked with exposure to several pollutants and oxidative stress in

M. edulis.4–6 Notwithstanding, the proteome of peroxisomes mayprovide the most specific and robust markers of environmentalpollution, allowing the investigation of sets of proteins involvedin xenobiotic detoxification and biotransformation. Liquidchromatography and 2DE approaches have been proposed tostudy this subproteome and to relate its dynamics to pollution.7,8

Other putative strategies for contamination assessment rely on theselection of protein/peptide expression signatures (PESs). Kniggeet al.1 and Bjornstad et al.,9 for example, used surface-enhancedlaser desorption/ionisation time-of-flight mass spectrometry tocharacterise PESs with a high discriminatory potential betweenM. edulis groups exposed to contaminants and non-exposedindividuals. Azaspiracids are a class of alga-derived shellfish toxinswith a worldwide distribution and have been associated withseveral cases of food poisoning.10 The presence of this naturaltoxin in bivalve tissues has been studied by Nzoughet et al.11,12

using complementary proteomic techniques, chromatography, gelelectrophoresis and mass spectrometry. The authors reported thatthe toxin binds to specific M. edulis protein fractions. Biomarkersof exposure and accumulation of other natural toxins producedby cyanobacteria have been investigated in the freshwater clamCorbicula fluminea and M. galloprovincialis using 2DE and massspectrometry.13,14

Proteomics has been employed for pathogen detection, foodstorage optimisation and processing, playing a role in the qualityassessment of fish and crustacean food products.15 Accordingly,proteomics should also be implemented as a complementaryresearch tool in bivalve species of economic interest such as M.edulis and M. galloprovincialis.

The aim of this study was to compare the application of fourmethods of protein extraction commonly used in 2DE researchin the analysis of the proteome of gills and digestive glandsfrom M. galloprovincialis, a mussel species with a distributionalong the Mediterranean and Atlantic coasts of Southern Europeand North Africa16 and intensely cultured in regions of Galicia,Northwest Spain.17 Furthermore, a preliminary characterisation ofthe 2DE protein profiles from the two organs by matrix-assistedlaser desorption/ionisation time-of-flight/time-of-flight (MALDI-TOF/TOF) mass spectrometry (MS) is presented. This work providesnovel information on protein extraction method selection for 2DEstudies in marine mussels.

MATERIALS AND METHODSBiological materialMussels (M. galloprovincialis) with a dimension of 50–60 mm werecollected at Memoria Beach (Oporto, Northern Portugal) during

May 2009. One hundred and fifty animals were maintained in 30L aquaria with natural sea water and acclimatised to laboratoryconditions. The water temperature was 16 ± 1 ◦C. The animalswere fed twice a week with 1 × 108 cells L−1 of the green algaChlorella vulgaris. The water was renewed every 2 days. Gills anddigestive glands were dissected in ice, frozen in liquid nitrogenand stored at −80 ◦C until protein extraction.

Sample preparation for 2DEOrgans from four animals were collected, pooled and groundin liquid nitrogen with a mortar and pestle. Proteins weresubsequently extracted using four different methods. Threeindependent extractions were performed for each method.

Method 1Tissue powder of 0.1 g fresh weight (FW) was homogenisedfor 1 h in 250 µL of urea (7 mol L−1), thiourea (2mol L−1), 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) (40 g L−1), dithiothreitol (DTT) (65 mmol L−1)and ampholytes (pH 4–7, (8 mL L−1) (Bio-Rad, Hercules, CA, USA)(protein solubilisation buffer, SB).

Method 2Samples obtained by method 1 were subsequently precipitatedand resolubilised following instructions from the commercialkit ProteoExtract (Calbiochem, San Diego, CA, USA) to furthereliminate potential interfering substances.

Method 3Tissue powder (0.1 g FW) was homogenised in cold acetone withtrichloroacetic acid (TCA, 100 g L-1) and mercaptoethanol (0.7 mLL−1) for 1 h at −20 ◦C. The homogenate was centrifuged andthe pellet was washed with cold acetone with mercaptoethanol(0.7 mL L−1). The protein pellet was subsequently dried at roomtemperature and solubilised in SB (250 µL).18

Method 4Proteins were extracted with phenol by a protocol adaptedfrom Hurkman and Tanaka.19 Briefly, tissue powder (0.1 g FW)was homogenised in 400 µL of Tris–HCl (pH 8.8, 0.1 mol L−1),ethylenediaminetetraacetic acid (5 mmol L−1), DTT (20 mmol L−1)and sucrose (300 g L−1) and 400 µL of phenol for 30 min at 4 ◦C.After centrifugation at 14 000 × g for 10 min, the phenol (upper)phase was collected. The extraction from the aqueous phase wasrepeated by adding 400 µL of phenol. The phenol fractions werepooled and proteins were precipitated with an equal volume ofammonium acetate (0.1 mol L−1) in methanol at −20 ◦C for 30min. This last step was repeated using the same buffer, repeatedtwice with cold acetone (800 mL L−1) and finally with cold acetone(800 mL L−1) with DTT (10 mmol L−1). The protein pellet was driedat room temperature and proteins were solubilised in SB (250 µL).

Proteins were quantified using the Bradford method20 andsamples were stored at −20 ◦C until further analysis.

Two-dimensional gel electrophoresisProtein samples (150 µg of protein) were diluted to 125 µL in SB.The protein samples were loaded onto 7 cm, pH 4–7 immobilinedry strips (Bio-Rad) and proteins were separated by isoelectricfocusing (IEF) in a Protean IEF cell (Bio-Rad) with the following

wileyonlinelibrary.com/jsfa c© 2012 Society of Chemical Industry J Sci Food Agric 2013; 93: 1779–1787

17

81

Proteomics in marine mussels: food quality and safety assessment www.soci.org

programme: 16 h at 50 V (strip rehydration); step 1, 15 min at 250V; step 2, 2 h voltage gradient to 4000 V (linear ramp); step 3,4000 V until achieving 10 000 V h−1 (linear ramp). After the firstdimension, IEF gel strips were stored at −20 ◦C until performingthe second dimension, sodium dodecyl sulfate polyacrylamide gelelectrophoresis (SDS-PAGE). IEF gel strips were equilibrated using10 g L−1 DTT and 25 g L−1 iodoacetamide in urea (6 mol L−1),glycerol (300 mL L−1) and SDS (20 g L−1).18 Subsequently, IEF gelstrips were placed on top of acrylamide/bisacrylamide (37.5:1 w/w,120 g L−1) SDS-PAGE gels (size 8.3 cm × 7 cm) and proteins wereseparated in a miniProtean II cell (Bio-Rad) at 150 V.

The 2DE was subsequently adapted for large-format gels. Forthis purpose, proteins (400 µg) were loaded onto 17 cm, pH 4–7immobiline dry strips and IEF was performed with the followingprogramme: 16 h at 50 V (strip rehydration); step 1, 15 min at 250 V;step 2, 3 h voltage gradient to 10 000 V (linear ramp); step 3, 10 000V until achieving 60 000 V h−1 (linear ramp). Second-dimensionSDS-PAGE was performed in a Protean II XL cell (Bio-Rad) at 24 mAper gel using acrylamide/bisacrylamide (37.5:1 w/w, 120 g L−1)large-format gels (size 20 cm × 20 cm). Gel staining for proteinvisualisation was performed as described by Puerto et al.14

Gel image acquisition and analysis2DE gel images were acquired in a GS-800 calibrated densitometer(Bio-Rad) and protein spots were detected automatically withPDQuest 2D analysis software (Bio-Rad), reproducing thesensitivity parameters for every gel image. Spot detection and spotmatching were manually revised in the software. Reproducibleprotein patterns were determined by the presence of eachindividual protein in three replicate gels, corresponding to threeindependent pools of biological material. Protein spot intensitieswere normalised to the total spot density in the gel.

Protein identificationProtein identification was performed based on the methoddescribed by Puerto et al.14 Protein spots were excised fromgels and subjected to in-gel digestion using the protease trypsin,followed by desalting and concentration with reverse phasemicrocolumns.21 The peptides were eluted directly onto the MALDIplate using the matrix α-cyano-4-hydroxycinnamic acid (5 g L−1)prepared in acetonitrile (700 mL L−1) and trifluoroacetic acid (1 mLL−1). Protein identification was done by MALDI-TOF/TOF with anApplied Biosystem 4800 proteomics analyser (Applied Biosystems,Foster City, CA, USA) in MS and MS/MS mode. Each mass spectrumwas obtained in independent acquisition mode and externallycalibrated using des-Arg-Bradykinin (904.468 Da), angiotensin 1(1296.685 Da), Glu-Fibrinopeptide B (1570.677 Da), ACTH (1–17)(2093.087 Da) and ACTH (18–39) (2465.199) (calibration mix fromApplied Biosystems). The MS/MS analyses were performed withcollision-induced dissociation. The generated mass spectra wereused to search the NCBI predicted protein database with thealgorithms Paragon, from Protein Pilot 2.0 software (AppliedBiosystems/MDS Sciex), and Mowse, from Mascot-demon 2.1.0software (Matrix Science, London, UK). A protein score above 2.0(P < 0.01) for Paragon and a threshold of 95% (P < 0.05) for Mowsewere considered for confident protein identification and at leastone peptide with ion score corresponding to P < 0.05. In theanalysis using Protein Pilot, other parameters considered were:enzyme, trypsin; Cys alkylation, iodoacetamide; special factor,urea denaturation; species, none; ID focus, biological modification.Regarding Mascot, the analysis of results was performed in GPS

Table 1. Protein extraction from Mytilus galloprovincialis tissues:protein yield and concentration obtained with different extractionmethods

M. galloprovincialis

organ

Protein

extraction

method

Protein yield

(µg g−1 tissue FW)

Protein

concentration

(µg mL−1)

Gills 1 15.6 ± 1.2 6.3 ± 0.5

2 15.1 ± 1.1 6.1 ± 0.4

3 11.1 ± 4.3 4.4 ± 1.7

4 6.6 ± 2.5 2.7 ± 1.0

Digestive gland 1 27.8 ± 6.4 8.5 ± 2.0

2 29.6 ± 6.3 11.8 ± 2.6

3 11.2 ± 4.3 4.5 ± 1.7

4 9.3 ± 2.5 3.7 ± 1.0

Values are mean ± standard deviation of three replicates.

Explorer 3.5 software (Applied Biosystems) using the followingparameters: missed cleavage, one; peptide tolerance, 50 ppm;fragment mass.

RESULTSSample preparation and 2DE from gills and digestive gland ofMytilus galloprovincialisProteins from M. galloprovincialis gills and digestive glands wereextracted following the four methods described above. Threeindependent extractions were performed for each method todetermine the protein yield and concentration. As presented inTable 1, methods 1 and 2 allowed the highest protein yield,with increases of about two- to threefold in comparison with themethod with the lowest protein yield (method 4).

One protein sample representative of each method wassubsequently tested for 2DE. The separation of proteins by 2DEwas qualitatively assessed considering protein resolution in thetotal area of the 2D gels between pI 4 and 7 and molecular mass(M) 20 and 100 kDa. All extraction methods allowed proteins tobe separated in the range of M and pI defined (Figs 1 and 2).Nevertheless, lower protein resolution and horizontal streakingwere verified for instance at the acidic end of 2DE gels loaded withgill samples obtained with methods 2 and 4 and digestive glandsamples obtained with method 4 (Figs 1 and 2, sections a). Verticalstreaking was observed in gel areas with a high abundance ofproteins with M varying between 40 and 60 kDa (Figs 1 and 2,sections b). The vertical streaking was more intense in the 2DEgels from digestive gland proteins extracted with method 4 andfrom gill proteins extracted with methods 2, 3 and 4 (Figs 1 and 2,sections b). In regard to these results, we suggest methods 1 and3 as most suitable for 2DE analysis of both M. galloprovincialistissues. Moreover they may be complementary in proteomicinvestigation, since we can verify differences in protein expressionin specific areas of the 2DE gels (Fig. 2, areas delineated by circles).Method 1 was further selected to extract and characterise gilland digestive gland proteins by mass spectrometry. An additionaladvantage of this method is the reduced sample handling, whichmay contribute to decrease postharvest modification of proteinsand sample loss during extraction. In this respect, method 1 can bemore adequate for proteomic analysis over acetone/TCA. In fact,method 3 extracted fewer proteins from gills and digestive glands,suggesting a loss of proteins during the precipitation step.22

J Sci Food Agric 2013; 93: 1779–1787 c© 2012 Society of Chemical Industry wileyonlinelibrary.com/jsfa

17

82

www.soci.org A Campos et al.

Figure 1. Two-dimensional gel electrophoresis of proteins from gills of Mytilus galloprovincialis obtained with extraction methods 1–4. Protein samples(150 µg) were loaded on 7 cm IEF gel strips. Sections show variations in protein resolution obtained with different extraction methods.

Figure 2. Two-dimensional gel electrophoresis of proteins from digestive gland of Mytilus galloprovincialis obtained with extraction methods 1–4. Proteinsamples (150 µg) were loaded on 7 cm IEF gel strips. Sections show variations in protein resolution obtained with different extraction methods. Circlesshow variations in protein abundance between methods 1 and 3.

Interestingly, this loss was not verified in method 2, probablyowing to precipitation being performed with proteins in solutionand not directly from fresh tissue as in method 3.

Protein identification in 2DE gels from Mytilusgalloprovincialis gills and digestive glandProteins from the gills and digestive gland of M. galloprovincialiswere extracted following method 1 and separated in large-format

(20 cm × 20 cm) 2DE gels with a pI range of 4–7 (Figs 3(a) and3(b)). The 2DE gels were loaded with 400 µg of protein and stainedwith colloidal Coomassie blue G-250. Computer-assisted analysisof three independent samples (n = 3) from M. galloprovincialis gillsand digestive gland by 2DE led to the detection of 146 and 97reproducible protein spots respectively. Mytilus galloprovincialisgill and digestive gland reference maps localising the identifiedproteins in a 2DE gel are presented in Figs 3(a) and 3(b) respectively.

wileyonlinelibrary.com/jsfa c© 2012 Society of Chemical Industry J Sci Food Agric 2013; 93: 1779–1787

17

83

Proteomics in marine mussels: food quality and safety assessment www.soci.org

(a)

(b)

Figure 3. Protein identification in Mytilus galloprovincialis (a) digestive gland and (b) gill 2DE gels. The first dimension was carried out on 17 cm, pH 4–7IEF gel strips and the second dimension on 120 g L−1 SDS-PAGE gels. Gels were stained with colloidal Coomassie blue G-250. Identified proteins areindicated with the most commonly used name followed by the respective spot number and the reference of the functional group assigned. Detailed dataconcerning protein identification by MALDI-TOF/TOF mass spectrometry are presented as supplementary tables A and B in ‘Supporting information’.

The 132 most abundant gill proteins from Coomassie blue-stained

2DE gels were excised manually and 41 were successfully identified

by MALDI-TOF/TOF mass spectrometry with 95% confidence,

representing a success rate of 31%. Proteins were grouped into six

functional classes: structural proteins (group 1), energy production

and carbohydrate metabolism (group 2), metal carriers (group 3),

chaperones (group 4), cell signalling and regulation (group 5)

and proteolysis (group 6) (Fig. 4). Functional group 1 was the

most representative, accounting for 61% of the identifications.

Regarding M. galloprovincialis digestive gland, all previously

found reproducible proteins (97) were excised from the 2DE

gels and analysed by MALDI-TOF/TOF mass spectrometry. Forty-

seven proteins were identified, representing a success rate of

48%. Most of the identified proteins were actin, tubulin and

J Sci Food Agric 2013; 93: 1779–1787 c© 2012 Society of Chemical Industry wileyonlinelibrary.com/jsfa

17

84

www.soci.org A Campos et al.

Figure 4. Percentage of proteins identified in gills and digestive gland of Mytilus galloprovincialis according to their biological function. The functionalgroups were numbered for reference in Fig. 3.

tropomyosin isoforms belonging to functional group 1. Forty-fourproteins from group 1 were identified, accounting for 94% ofthe total identifications in this organ. We were able to identifyparamyosin, also belonging to functional group 1, a ribosomalprotein and calmodulin, belonging to the functional groups ofprotein transduction (group 7) and cell signalling and regulation(group 5) respectively (Fig. 4). A protein with unknown function(group 8) was also identified in the 2DE gel of M. galloprovincialisdigestive gland (Fig. 4).

DISCUSSIONThe gills and digestive gland are major respiratory and nutrientuptake organs in bivalves, being in direct contact with nutrients,particles in suspension and pollutants. The exposure of this groupof animals to water contaminants leads to a sensitive response,making them ideal candidates for the biomonitoring of pollution.7

Nevertheless, gill and digestive cells accumulate different typesof lipids (presence of neutral lipids and lipofucsin granules),polysaccharides and carbohydrates,23,24 compounds that caninterfere with 2DE and biomarker analysis. Proteomic investigationof these tissues therefore requires the utilisation of protocols thatallow efficient extraction of proteins and simultaneously theirseparation from non-protein material.

In this work we extracted proteins from M. galloprovincialisgills and digestive gland with different methods, aiming to assesstheir putative use in 2DE and analysis of protein expression inthis model species. The proteins were subsequently characterisedby mass spectrometry to assess the major functional groups ofproteins separated by 2DE. A pH separation gradient of 4–7 waschosen, instead of a wider pH interval, for the first-dimension IEFin an attempt to increase the resolution in a pI interval where theproteome is expected to increase in complexity.25

Protein extraction protocolsFour protein extraction procedures were tested regarding theefficiency to solubilise and separate proteins from the gills and

digestive gland by 2DE. The direct solubilisation of proteins fromthe disrupted tissues (method 1) proved to be more suitablefor 2DE analysis in line with the high protein resolution in thegels and the highest protein yield. Similar approaches have beenemployed for the analysis of foot and gill proteins, focusing on thegenetic diversity between mussel species and populations,26,27

and the investigation of M. galloprovincialis and Mytilus trossulusresponse to thermal stress.28 In comparison with other methodsemployed for soluble, cytoplasmic proteins,3,29 this method allowsthe extraction and investigation of more hydrophobic membraneproteins.

The TCA/acetone extraction (method 3) can be an alternativemethod if additional removal of contaminants is necessary toincrease 2DE resolution. Acetone precipitation has been employedin a study to identify protein markers of algal toxin contaminationin the digestive glands of shellfish.30 In our samples, this methodallowed high-quality 2DE separations, likely owing to the organicsolvent extraction of lipids and polyphenols from the tissues.31 Thetwo methods can be complementary, since the protein profilesobtained suggest different abilities to solubilise and detect specificproteins.

Method 4 (phenol extraction) has been used in recalcitrant planttissues.19,32 A phase partitioning of proteins and contaminants (e.g.nucleic acids) is made possible by the use of phenol in the initialsteps of the method. Further contaminant removal is achievedby protein precipitation. Nevertheless, application of this methodto the mussel tissues led to poor protein resolution in the acidicrange of the gels. This result suggests the presence of contaminantsaffecting protein migration. Owing to poor protein resolution, wemay also expect M. galloprovincialis gill samples obtained withmethod 2 to have high levels of contaminants.

Protein identificationProtein identification revealed a high number of proteins withfunctions in cytoskeleton structure, movement and maintenancein both organs. This group of proteins represented 61 and94% of the gill and digestive gland reproducible protein

wileyonlinelibrary.com/jsfa c© 2012 Society of Chemical Industry J Sci Food Agric 2013; 93: 1779–1787

17

85

Proteomics in marine mussels: food quality and safety assessment www.soci.org

profiles respectively. Proteins involved in energy production andcarbohydrate metabolism, metal transport, chaperones and stressresponse, cell signalling and regulation, proteolysis and proteintransduction complete the group of identifications (2–13% of totalidentifications in each proteome). The results show a low proteinidentification rate using MS/MS data and homology search (31 and48% for gill and digestive gland proteins analysed respectively).The lack of protein sequence information for this or closely relatedspecies may contribute to the reduced success of the analysis. Infact, presently, none of the three species from the M. edulis group(M. edulis, M. galloprovincialis and M. trossulus) has its genomesequenced, and the number of entries in generalist databasesis extremely low (3239, 1656 and 572 entries respectively in theNCBI database when accessed in February 2012). This is likely themost significant limitation on the use of proteomics in marinemussels, since the function of some of the important proteins inthe investigations may not be adequately assessed.

Structural proteins (1)Proteins involved in cytoskeleton structure and dynamicsidentified in this work were actin, tubulin, myosin, paramyosin,tropomyosin, gelsolin, centrin and tektin. The identification ofactin, tubulin and tropomyosin has been frequently reported inother proteomic studies, emphasising their relative abundancein tissue samples of bivalves and their prevalence in publicdatabases.3,4 Moreover, some of these proteins have beenrecognised as valuable biomarkers to assess quality traits andstorage conditions in fish and shrimp food products.33,34

Energy production and carbohydrate metabolism (2)Triosephosphate isomerase (TPI), beta subunit ATPase, NAD-dependent epimerase/dehydratase, malate dehydrogenase(MDH) and citrate synthase identified in this work are key enzymesof cellular energy metabolism, participating in glycolysis, the citricacid cycle and ATP synthesis. Alterations in the expression of theseproteins may reflect physiological stress and exposure to envi-ronmental stressors and pollutants. Moreover, TPI was considereda good candidate marker to monitor alterations in fish productsduring frozen storage.34

Metal transport (3)Another protein identified that can provide additional informationon the physiological condition of organisms, in particular relatingto the levels of free/stored iron and oxidative stress, is the irontransporter ferritin. This protein has been shown to increase inresponse to stresses such as anoxia, being considered an acutephase protein.35 A homologue of extrapallial (EP) protein wasidentified in the gill proteome. The EP protein is a calcium-binding glycoprotein identified in the extrapallial fluid of marinebivalves and considered to play a role as a building block ofthe shell-soluble organic matrix.36,37 The protein has similaritiesto the histidine-rich glycoproteins (HRGs), with which it sharesover 95% sequence identity, expressed in the blood plasma ofmarine bivalves and involved in metal transport.36,37 Thus themonitoring of EP expression may provide additional informationon metal transport and accumulation in bivalves, growth and shelldevelopment.

Chaperones (4)The heat shock proteins HSP 60 and HSP 70 as well as theP4hb protein are molecular chaperones, being involved in

protein folding and stabilisation. These proteins are recognised asimportant molecular stress markers.38 HSP 70 isoforms seem tobe particularly affected by the polycyclic aromatic hydrocarbonbenzo[a]pyrene in the bivalve species C. gallina.39 P4hb is alsoa regulator of the protein redox state and is a subunit of themicrosomal triglyceride transfer protein complex.40 Its expressionis affected also by chronic hypoxia stress.41 Of note also is thepotential of specific HSPs as animal food sensory quality markers.42

Cell signalling and regulation (5)The 14-3-3 protein epsilon belongs to a family of conservedregulatory molecules expressed in eukaryotic cells that bind to amultitude of functionally diverse signalling proteins, includingkinases and phosphatases, being thereby involved in theregulation of cellular processes such as cell cycle, cell growth,differentiation, survival, apoptosis, migration and spreading.43

The visual G protein identified in M. galloprovincialis gills belongsto the guanine nucleotide-binding family of proteins that playan important role in signal transduction in cells, mediated byhormones, neurotransmitters, chemokines and autocrine andparacrine factors.44 Calmodulin, on the other hand, regulates anumber of different protein targets, mediated by the binding ofcalcium to the helix-loop-helix structural domains of the protein.This protein is important for diverse cellular processes, includinginflammation, apoptosis, smooth muscle contraction, intracellularmovement, short-term and long-term memory, nerve growth andthe immune response.45 Variations in the expression of theseregulatory proteins may relate to significant alterations in themetabolism of cells exposed to pollutants.

Proteolysis (6) and protein transduction (7)The proteasome subunit alpha identified in M. galloprovincialisgills is a constituent of the proteasome complex, whose mainfunction is to degrade inactive or damaged proteins by proteolysis.Moreover, a homologue of ribosome’s structural constituent wasidentified in the digestive gland (protein spot 46). Together, theseproteins may reflect the effects of pollutants in the activity ofprotein degradation or synthesis pathways, providing additionalinformation on cellular stress.

CONCLUSIONThe 2DE proteomic approach described in this work is animportant tool to investigate the expression in M. galloprovincialisof proteins with functions in energy metabolism, cell signallingand regulation and stress response. With particular relevance tometal contamination are the metal-binding proteins (ferritin andEP protein). Chaperones instead may provide information on otherenvironmental stressors. Notwithstanding, the methods describedin this paper can be relevant to elucidate other biochemicalprocesses of interest in shellfish farming and control of foodquality and also to further develop/improve gel-based proteomicapproaches in this and other affiliated species. Particularlyimportant for method development would be to reduce thepresence of structural proteins that are highly abundant in2DE gels, opening up an opportunity to study other sets ofproteins biologically more relevant for environmental monitoringand biomarker discovery. Sample fractionation by differentialcentrifugation followed by high-pH extraction was proposed toremove several structural proteins from skeletal muscle samples.46

The 2DE reference map described here is also a valuable tool for

J Sci Food Agric 2013; 93: 1779–1787 c© 2012 Society of Chemical Industry wileyonlinelibrary.com/jsfa

17

86

www.soci.org A Campos et al.

laboratories to standardise protocols and establish direct datacomparison and interpretation.

ACKNOWLEDGEMENTSAlexandre Campos’s and Andre de Almeida’s contract work issupported by the Ciencia 2007 programme of the Fundacao paraa Ciencia e a Tecnologia (FCT, Lisbon, Portugal). Anabel Prietoand Maria Puerto acknowledge Jose Castillejo and Ministeriode Ciencia e Innovacion (MICINN, Madrid, Spain) scholarshipsrespectively. This work was funded by the research projectPTDC/MAR/115347/2009 from FCT.

SUPPORTING INFORMATIONSupporting information may be found in the online version of thisarticle.

REFERENCES1 Knigge T, Monsinjon T and Andersen O-K, Surface-enhanced laser

desorption/ionization-time of flight-mass spectrometry approachto biomarker discovery in blue mussels (Mytilus edulis) exposedto polyaromatic hydrocarbons and heavy metals under fieldconditions. Proteomics 4:2722–2727 (2004).

2 O’Farrell PH, High resolution two-dimensional electrophoresis ofproteins. J Biol Chem 250:4007–4021 (1975).

3 Rodrıguez-Ortega MJ, Grosvik BE, Rodriguez-Ariza A, Goksoyr A andLopez-Barea J, Changes in protein expression profiles in bivalvemolluscs (Chamaelea gallina) exposed to four model environmentalpollutants. Proteomics 3:1535–1543 (2003).

4 McDonagh B, Tyther R and Sheehan D, Redox proteomics in themussel, Mytilus edulis. Mar Environ Res 62:S101–S104 (2006).

5 McDonagh B and Sheehan D, Effects of oxidative stress on proteinthiols and disulphides in Mytilus edulis revealed by proteomics: actinand protein disulphide isomerase are redox targets. Mar Environ Res66:193–195 (2008).

6 Tedesco S, Doyle H, Redmond G and Sheehan D, Gold nanoparticlesand oxidative stress in Mytilus edulis. Mar Environ Res 66:131–133(2008).

7 Apraiz I, Mi J and Cristobal S, Identification of proteomic signatures ofexposure to marine pollutants in mussels (Mytilus edulis). Mol CellProteomics 5:1274–1285 (2006).

8 Mi J, Apraiz I and Cristobal S, Peroxisomal proteomic approach forprotein profiling in blue mussels (Mytilus edulis) exposed to crudeoil. Biomarkers 12:47–60 (2007).

9 Bjornstad A, Larsen BK, Skadsheim A, Jones MB and Andersen OK, Thepotential of ecotoxicoproteomics in environmental monitoring:biomarker profiling in mussel plasma using proteinchip arraytechnology. J Toxicol Environ Health A 69:77–96 (2006).

10 Campos A, Tedesco S, Vasconcelos V and Cristobal S, Proteomicresearch in bivalves: towards the identification of molecular markersof aquatic pollution. J Proteomics 75:4346–4359 (2012).

11 Nzoughet KJ, Hamilton JTG, Floyd SD, Douglas A, Nelson J, DevineL, et al., Azaspiracid: first evidence of protein binding in shellfish.Toxicon 51:1255–1263 (2008).

12 Nzoughet JK, Hamilton JTG, Botting CH, Douglas A, Devine L, NelsonJ, et al., Proteomics identification of azaspiracid toxin biomarkersin blue mussels, Mytilus edulis. Mol Cell Proteomics 8:1811–1822(2009).

13 Martins JC, Leao PN and Vasconcelos V, Differential protein expressionin Corbicula fluminea upon exposure to a Microcystis aeruginosatoxic strain. Toxicon 53:409–416 (2009).

14 Puerto M, Campos A, Prieto A, Camean A, Almeida AMD, CoelhoAV, et al., Differential protein expression in two bivalve species;Mytilus galloprovincialis and Corbicula fluminea; exposed toCylindrospermopsis raciborskii cells. Aquat Toxicol 101:109–116(2011).

15 Gaso-Sokac D, Kovac S and Josic D, Use of proteomic methodologyin optimization of processing and quality control of food of animalorigin. Food Technol Biotechnol 49:397–412 (2011).

16 Hilbish TJ, Mullinax A, Dolven SI, Meyer A, Koehn RK and Rawson PD,Origin of the antitropical distribution pattern in marine mussels(Mytilus spp.): routes and timing of transequatorial migration. MarBiol 136:69–77 (2000).

17 Villalba A, Mourelle SG, Carballal MJ and Lopez C, Symbionts anddiseases of farmed mussels Mytilus galloprovincialis throughout theculture process in the Rias of Galicia (NW Spain). Dis Aquat Org31:127–139 (1997).

18 Campos A, da Costa G, Coelho A and Fevereiro P, Identification ofbacterial protein markers and enolase as a plant response proteinin the infection of Olea europaea subsp. europaea by Pseudomonassavastanoi pv. savastanoi. Eur J Plant Pathol 125:603–616 (2009).

19 Hurkman WJ and Tanaka CK, Solubilization of plant membrane proteinsfor analysis by two-dimensional gel electrophoresis. Plant Physiol81:802–806 (1986).

20 Bradford MM, A rapid and sensitive method for the quantitationof microgram quantities of protein utilizing the principle ofprotein–dye binding. Anal Biochem 72:248–254 (1976).

21 Almeida AM, Campos A, Francisco R, Harten SV, Cardoso LA andCoelho AV, Proteomic investigation of the effects of weight lossin the gastrocnemius muscle of wild and NZW rabbits via 2D-electrophoresis and MALDI-TOF MS. Anim Genet 41:260–272 (2010).

22 Fic E, Kedracka-Krok S, Jankowska U, Pirog A and Dziedzicka-Wasylewska M, Comparison of protein precipitation methodsfor various rat brain structures prior to proteomic analysis.Electrophoresis 31:3573–3579 (2010).

23 Dimitriadis VK, Domouhtsidou GP and Cajaraville MP, Cytochemicaland histochemical aspects of the digestive gland cells of the musselMytilus galloprovincialis (L.) in relation to function. J Mol Histol35:501–509 (2004).

24 Gomez-Mendikute A, Elizondo M, Venier P and Cajaraville MP,Characterization of mussel gill cells in vivo and in vitro. Cell TissueRes 321:131–140 (2005).

25 Wu S, Wan P, Li J, Li D, Zhu Y and He F, Multi-modality of pI distributionin whole proteome. Proteomics 6:449–455 (2006).

26 Lopez JL, Marina A, Vazquez J and Alvarez G, A proteomic approachto the study of the marine mussels Mytilus edulis and M.galloprovincialis. Mar Biol 141:217–223 (2002).

27 Mosquera E, Lopez JL and Alvarez G, Genetic variability of the marinemussel Mytilus galloprovincialis assessed using two-dimensionalelectrophoresis. Heredity 90:432–442 (2003).

28 Tomanek L and Zuzow MJ, The proteomic response of the musselcongeners Mytilus galloprovincialis and M. trossulus to acute heatstress: implications for thermal tolerance limits and metabolic costsof thermal stress. J Exp Biol 213:3559–3574 (2010).

29 Riva C, Binelli A, Rusconi F, Colombo G, Pedriali A, Zippel R, et al., Aproteomic study using zebra mussels (D. polymorpha) exposed tobenzo(α)pyrene: the role of gender and exposure concentrations.Aquat Toxicol 104:14–22 (2011).

30 Ronzitti G, Milandri A, Scortichini G, Poletti R and Rossini GP,Protein markers of algal toxin contamination in shellfish. Toxicon52:705–713 (2008).

31 Carpentier SC, Witters E, Laukens K, Deckers P, Swennen R andPanis B, Preparation of protein extracts from recalcitrant planttissues: an evaluation of different methods for two-dimensional gelelectrophoresis analysis. Proteomics 5:2497–2507 (2005).

32 Faurobert M, Pelpoir E and Chaib J, Phenol extraction of proteinsfor proteomic studies of recalcitrant plant tissues. Meth Mol Biol355:9–14 (2007).

33 Martinez I, Friis TJ and Careche M, Post mortem muscle proteindegradation during ice-storage of Arctic (Pandalus borealis) andtropical (Penaeus japonicus and Penaeus monodon) shrimps: acomparative electrophoretic and immunological study. J Sci FoodAgric 81:1199–1208 (2001).

34 Kjærsgard IVH, Nørrelykke MR and Jessen F, Changes in cod muscleproteins during frozen storage revealed by proteome analysis andmultivariate data analysis. Proteomics 6:1606–1618 (2006).

35 Larade K and Storey KB, Accumulation and translation of ferritin heavychain transcripts following anoxia exposure in a marine invertebrate.J Exp Biol 207:1353–1360 (2004).

36 Yin Y, Huang J, Paine ML, Reinhold VN and Chasteen ND, Structuralcharacterization of the major extrapallial fluid protein of themollusc Mytilus edulis: implications for function. Biochemistry44:10720–10731 (2005).

37 Abebe AT, Devoid SJ, Sugumaran M, Etter R and Robinson WE,Identification and quantification of histidine-rich glycoprotein

wileyonlinelibrary.com/jsfa c© 2012 Society of Chemical Industry J Sci Food Agric 2013; 93: 1779–1787

17

87

Proteomics in marine mussels: food quality and safety assessment www.soci.org

(HRG) in the blood plasma of six marine bivalves. Comp BiochemPhysiol B 147:74–81 (2007).

38 Henderson B, Integrating the cell stress response: a new viewof molecular chaperones as immunological and physiologicalhomeostatic regulators. Cell Biochem Funct 28:1–14 (2010).

39 Jurgen F, Valerio M, Roberto R, Paolo SG and Marta M, 2-DE proteomicanalysis of HSP70 in mollusc Chamelea gallina. Fish Shellfish Immunol30:739–743 (2011).

40 Schultz-Norton JR, McDonald WH, Yates JR and Nardulli AM, Proteindisulfide isomerase serves as a molecular chaperone to maintainestrogen receptor alpha structure and function. Mol Endocrinol20:1982–1995 (2006).

41 Tian F, Zhou X, Wikstrom J, Karlsson H, Sjoland H, Gan LM, et al., Proteindisulfide isomerase increases in myocardial endothelial cells in miceexposed to chronic hypoxia: a stimulatory role in angiogenesis. AmJ Physiol Heart Circ Physiol 297:1078–1086 (2009).

42 Guillemin N, Jurie C, Cassar-Malek I, Hocquette JF, Renand G and PicardB, Variations in the abundance of 24 protein biomarkers of beeftenderness according to muscle and animal type. Animal 5:885–894(2011).

43 Mhawech P, 14-3-3 proteins – an update. Cell Res 15:228–236 (2005).44 Neves SR, Ram PT and Iyengar R, G protein pathways. Science

296:1636–1639 (2002).45 Chin D and Means AR, Calmodulin: a prototypical calcium sensor.

Trends Cell Biol 10:322–328 (2000).46 Jarrold B, DeMuth J, Greis K, Burt T and Wang F, An effective skeletal

muscle prefractionation method to remove abundant structuralproteins for optimized two-dimensional gel electrophoresis.Electrophoresis 26:2269–2278 (2005).

J Sci Food Agric 2013; 93: 1779–1787 c© 2012 Society of Chemical Industry wileyonlinelibrary.com/jsfa