J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te / j p ro t

The sheep milk fat globule membrane proteome

Salvatore Pisanua,1, Stefania Ghisauraa,b,1, Daniela Pagnozzia, Grazia Biosaa,Alessandro Tancaa, Tonina Roggioa, Sergio Uzzaua,c, Maria Filippa Addisa,d,⁎aPorto Conte Ricerche Srl, Tramariglio, Alghero (SS), ItalybBiosistema Scrl, Sassari, ItalycDipartimento di Scienze Biomediche, Università di Sassari, Sassari, ItalydDipartimento di Patologia e Clinica Veterinaria, Università di Sassari, Sassari, Italy

A R T I C L E I N F O

⁎ Corresponding author. Porto Conte Ricerche079 998 526; fax: +39 079 998 567.

E-mail address: addis@portocontericerche1 These authors contributed equally to this

1874-3919/$ – see front matter © 2010 Elsevdoi:10.1016/j.jprot.2010.11.011

A B S T R A C T

Article history:Received 14 October 2010Accepted 29 November 2010Available online 13 December 2010

Milk fat globule membranes (MFGM) are three-layered structures that enclose fat droplets,and are composed by an internal monolayer of endoplasmic reticulum origin, surroundedby a bilayer derived from the apical membrane of the lactating cell. In this work, anoptimized protein extraction method was applied to sheep MFGM, and extracts weresubjected to SDS-PAGE separation followed by shotgun LC tandem mass spectrometry(GeLC-MS/MS) for identification and characterization. In total, 140 unique sheep MFGMproteins (MFGMPs) were identified. All protein identification data were subjected to GeneOntology (GO) classification for localization and function. Moreover, the relative abundanceof all identified MFGMPs was estimated by means of the normalized spectral abundancefactor (NSAF) approach, and GO abundance classes were obtained. The data gathered in thiswork provide a detailed picture of the proteome expressed in healthy sheep MFGs, and laythe foundations for future studies on sheep lactation physiology and on its alterations inpathological conditions.

© 2010 Elsevier B.V. All rights reserved.

Keywords:Sheep milkMilk fat globule membraneGeLC-MS/MSNormalized spectral abundanceOntology

1. Introduction

Milk is a biological fluid of unique quality and complexity,being composed of proteins, carbohydrates, lipids, water, andions in an optimal ratio [1]. Macro- andmicronutrients in milkprovide bioactive factors with beneficial effects for theimmune system, cognitive development, prevention of path-ogen colonization, andmodulation of the intestinalmicroflora[2–5]. Milk is dominated by presence of a few primary proteins,namely αs1-casein (αs1CN), β-casein (βCN), k-casein (kCN), αs2-casein (αs2CN), β-lactoglobulin (β-lg), α-lactalbumin (α-la), andserum albumin, having different relative abundances anddifferent functional activities [6]. In recent years, studies havefocused their attention on proteins of the milk fat globule

Srl; S.P. 55 Porto Conte/Ca

.it (M.F. Addis).work.

ier B.V. All rights reserve

membrane (MFGMPs). MFGMPs, which account for 1–2% of thetotal protein fraction in milk, are receiving a growingattention, but their functions have not yet been completelyelucidated [7].

The MFGM is the envelope surrounding each lipid dropletin milk. It is organized into a three-layered structure (an innerlipid monolayer and an outer lipid bi-layer), consisting of acomplex mixture of proteins, glycoproteins, enzymes, neutrallipids, and polar lipids, such as phospholipids and glyco-sphingolipids [8,9]. The MFGM is composed of proteins andlipids in a 1:1 weight ratio [10]. This coat material is partiallyoriginated from the apical plasma membrane, from theendoplasmic reticulum, and from other intracellular compart-ments of the mammary epithelial cell [11]. The intracellular

po Caccia Km 8.400, Tramariglio, 07041 Alghero (SS), Italy. Tel.: +39

d.

351J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

precursors of milk fat globules, milk secretory granules, ariseas milk lipids between the leaflets of the endoplasmicreticulum (ER) membrane, distending the leaflet facing thecytoplasm and causing primary droplets of accumulated lipidon the ER [12–14]. Milk secretory granules arise by the fusion oflipid droplets and are thus enveloped in a monolayer ofphospholipids from the cytoplasmic leaflet of the ER mem-brane [15–17]. However, in contrast to all other lipid droplets,milk secretory granules are transported to the cell surfacewhere they are pinched off into the alveolar space entirelysurrounded in a layer of plasma membrane [18,19]. Accord-ingly, the envelopes of mature milk fat globules consist ofthree phospholipid membrane monolayers. On the inside isthe cytoplasmic leaflet of the ER membrane arising from thephospholipid monolayer that surrounds the milk secretorygranule. On the outside is a classic membrane bilayercomprised of the extracellular and cytoplasmic leaflets ofthe plasma membrane. Variable amounts of cytoplasm areoften entrained between the inner monolayer and the outerbilayer.

To date, the proteomic approach has been applied tohuman MFGM [20–22], mouse MFGM [23], cow MFGM [24],water buffalo MFGM [25], and goat MFGM [26]. On the contrary,proteomic data on ovine MFGM are lacking, and theircharacterization is complicated by the higher amount of fatin sheep milk (almost double than cow milk), that hinders thecurrently available procedures reported in the literature forpurification of MFGMPs.

Dairy sheep farming is mostly concentrated in Europe andin the Mediterranean countries, where it often plays a primaryeconomical role. Most of the sheep milk produced worldwideis destined to cheese production, but it is also transformedinto other typical dairy products such as ricotta cheese andtraditional fermented milks. Characterization of MFGM pro-teins in sheep milk is therefore important, especially whenconsidering that traditional sheep dairy products produceddirectly from raw, unprocessed milk, are usually those withthe higher economical value.

The importance that MFG play in determining the qualityof dairy products is known, and several MFG components havebeen recently identified as being beneficial for human health[27]. This has led to an increased awareness that MFG can beconsidered as a food ingredient with functional properties andhealth benefits [28]. In recent years, different factors withbeneficial properties (e.g. cholesterolemia-lowering factor,inhibition of cancer cell growth, and antibacterial effect)have been detected in bovine MFGM components [8,28]. Onthe other hand, some researchers reported a link betweenMFGM proteins and diseases (multiple sclerosis, autism andcoronary heart disease) [7]. The identification of proteinsassociated with various aspects of milk production and MFGsecretion can also provide fundamental information forresearch in lactation biology and for monitoring the animaludder health. Proteomic analysis of MFGM has the potential touncover many of the possible signaling and secretory path-ways used by the mammary gland [29].

Aim of this study was to characterize ovine MFGMPs, toestablish a reference protein map in physiological conditions,and to lay the foundations for monitoring the animal udderhealth on the basis of their proteomic assortment. To this

purpose, a method for purification of sheep MFGMPs wasoptimized, extracted proteins were subjected to SDS-PAGEseparation followed by shotgun LC tandem mass spectrome-try identification (GeLC-MS/MS), and characterization forlocalization, function, and relative abundance was performedby means of the spectral counting approach.

2. Materials and methods

2.1. Preparation of the ovine milk fat globule membraneprotein fraction

Fresh milk was collected from different healthy Sarda dairysheep in mid-lactation (n=6). The milk from each collectionwas brought immediately to the laboratory and centrifuged at5000×g for 15 min at 4 °C to obtain the cream fractioncontaining MFGM. After centrifugation, the tubes with milksamples were turned upside down on ice for 20 min. The skimmilk was recovered while the cream layer was washed twotimes in phosphate saline buffer at 37 °C for 10 min in slowagitation. Eachwashing stepwas followed by centrifugation at5000×g for 15 min at 4 °C. Fat globules were washed in tri-distilled water at 37 °C for 10 min in slow agitation andcentrifuged at 5000×g for 15 min at 4 °C. After removing theaqueous phase, the fat globules were crystallized at 4 °Covernight. Fat globule membranes were homogenized in tri-distilled water with the TissueLyser mechanical homogenizer(Qiagen, Hilden, Germany) for 8 min at a frequency of 20 Hz.The shaken cream fraction was warmed to 45 °C for 30 min inorder to melt the fat. The fat fraction was centrifuged at10,000×g for 15 min at 4 °C to separate the fat from theproteins. The top fat layer was removed from the Eppendorfsafe-lock tubes (Eppendorf, Hamburg, Germany) using aspatula. The resulting pellet was solubilized in tri-distilledwater and protein concentration was determined using the 2DQuant Kit (GE Healthcare, Uppsala, Sweden). Samples werestored at −20 °C prior to analysis.

2.2. Protein precipitation

The residual fat layer remaining after MFGMP extraction wasresuspended with 4% SDS, 10% β-mercaptoethanol and0.125 M Tris–HCl at 56 °C for 1 h. Three alternative precipita-tion methods were then applied. Method 1: The sample wassubjected to protein precipitation in a cold mixture of acetoneandmethanol (v/v ratio 8:1) and stored at −20 °C overnight [30].The precipitate was pelleted by centrifugation at 5000×g for15 min at 4 °C, washed sequentially with 1 mL of acetonefollowed by methanol, and then air-dried. Method 2: TCA wasadded to the sample at a final concentration of 10% in acetonewith 20 mM DTT [31]. Following an overnight incubation at−20 °C, the protein pellet was collected by centrifugation at5000×g for 10 min at 4 °C. The supernatant was discarded andthe pellet was washed three times with acetone, subjected tocentrifugation, and then air-dried. Method 3: A mixture of ice-cold tri-n-butylphosphate:acetone:methanol (1:12:1) wasadded to the sample and incubated overnight at 4 °C [32].The precipitate was pelleted by centrifugation (5000×g for15 min at 4 °C), washed sequentially with 1 ml of tri-n-

Fig. 1 – SheepMFGMPs separated on a 10%SDS-PAGE gel. Theprotein loadwas 10 μg per lane. On the left, molecular weightstandards (lane M) are reported. Lanes 1 to 6 are milk fatglobule membrane proteins from six different sheep.

352 J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

butylphosphate, acetone, and methanol, and then air-dried.After precipitation, proteins were quantified using the 2DQuant Kit (GE Healthcare, Uppsala, Sweden) and resuspendedwith Laemmli buffer [33] for SDS-PAGE.

2.3. Sodium dodecyl sulphate-polyacrylamide gelelectrophoresis

Prior to electrophoresis, samples were resuspended inLaemmli buffer [33], heated at 100 °C for 5 min, and loadedon a 10% acrylamide separating gel. Electrophoresis wasperformed at a constant voltage of 200 Vwith a Mini PROTEANTetra Cell slab gel apparatus (Bio-rad Laboratories, Hercules,California, USA) using running buffer containing 250 mM Tris,1.92 M glycine, 1% SDS, pH 8.3. The gel was then stained withcolloidal Coomassie as described by Candiano and coworkers[34].

2.4. GeLC-MS/MS

Proteins (30 μg) were resolved on 10% acrylamide separatinggels. After staining by colloidal Coomassie [34], the gel lanewas cut into 30 slices and subjected to in-gel tryptic digestionas described [35,36]. Briefly, the gel slices were de-stained andwashed, and, after dithiothreitol reduction and iodoacetamidealkylation, the proteins were digested with porcine trypsin(Sigma) overnight at 37 °C. The resulting tryptic peptides wereextracted from the gel pieces with 100% acetonitrile. Theextracts were evaporated in a vacuum centrifuge to removeorganic solvent and resuspended in 0.02% formic acid. LC-MS/MS analyses were performed on a Q-TOF hybrid massspectrometer equipped with a nano-lock Z-spray source andcoupled on-line with a capillary chromatography systemCapLC (Waters, Manchester, UK). After loading, the peptidemixture was first concentrated and washed at 20 μL/min on aRP pre-column (Symmetry 300, C18, 5 μm, NanoEase, Waters,Manchester, UK) using 0.2% formic acid as eluent. The samplewas then separated on a C18 RP column (Symmetry,75 μm×20 cm, Waters, Manchester, UK) at a flow rate of250 nL/min, using a linear gradient of eluent B (0.2% formicacid in 95% ACN) in A (0.2% formic acid in 5% ACN) from 7 to50% in 40 min. The mass spectrometer was set up in a data-dependent MS/MS mode where a full-scan spectrum (m/zacquisition range from 400 to 1600 Da/e) was followed by atandem mass spectrum (m/z acquisition range from 100 to2000 Da/e). Peptide ions were selected as the three mostintense peaks of the previous scan. A suitable collision energywas applied depending on the mass and charge of theprecursor ion. Argon was used as the collision gas. Protein-Lynx software (Version 2.2.5), provided by the manufacturers,was used to analyze raw MS and MS/MS spectra and togenerate a peak list which was introduced in MASCOTDaemon software (version 2.3.0, Matrix Science, Boston, MA)for protein identification after having merged all the dataresulting from the same lane into a single large MGF file. NCBIwas used as sequence database. Search parameters were asfollows: peptide tolerance 30 ppm, MS/MS tolerance 0.4 Da,charge state +2 and +3, enzyme trypsin, allowing up to twomissed cleavages. For false positive analysis, a decoy searchwas performed automatically by choosing the Decoy checkbox

on the search form. Mascot results were parsed with the IRMatoolbox (version 1.26.1) [37], using the following inclusioncriteria: protein significance threshold p<0.01, peptide rank 1,peptide ion score cut-off 20, in order to keep FDR values always≤2% per each analysis. Unique peptides (UP), spectral counts(SpC), and exponentially modified protein abundance index(emPAI) values were reported as calculated by IRMa [37]. Skinkeratins were excluded from the final protein list. Proteinsthat contained similar peptides and could not be differentiat-ed based on MS/MS analysis alone were grouped [38].

Spectral counts (SpC) of each protein or group of proteinswere used to estimate protein abundance.

Protein abundance was expressed by means of thenormalized spectral abundance factor (NSAF). NSAF wascalculated as follows:

NSAF =SAFi

∑Ni = 1SAFi

where subscript i denotes a protein identity and N is the totalnumber of proteins, while SAF is a protein spectral abundancefactor that is defined as a protein spectral counts divided by itslength, as described by Zhang and coworkers [39].

3. Results and discussion

3.1. Extraction and isolation of sheep MFGMPs

In order to enable analysis of sheep MFGMPs, it was necessaryto remove fat and the major milk proteins. In fact, the veryhigh abundance proteins present in milk, such as caseins andwhey proteins, would interfere with the study of this lowerabundance fraction. Moreover, extraction of MFGMPs iscomplicated by their strong interaction with lipids. Conse-quently, efficient wash, lysis, lipid removal, and proteinsolubilization steps are key for extraction of these transmem-brane or membrane-bound proteins, especially when dealingwith a high fat content material such as sheep milk. Several

Fig. 2 – Colloidal Coomassie stained gel showing proteinsrecovered from the residual fat fraction by precipitation withthree different methods. Lane M, molecular weightstandards; Lane A, MFGMPs extracted with the optimizedprotocol (protein load: 20 μg); Lane B, precipitation withacetone and methanol (v/v ratio 8:1); Lane C, precipitationwith 10% TCA and 20 mM DTT; Lane D, precipitation withtri-n-butylphosphate:acetone:methanol (1:12:1).

353J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

methods developed to extract and isolate MFGMPs from milkof other species were initially applied to sheep milk [11,40].However, application of these techniques led to unsatisfactory

Fig. 3 – Qualitative GO classification of sheep MFGMPs accordingpercentage of protein identities falling in the indicated class.

results. We therefore optimized a method through theintroduction of washing, lysis, heating, and centrifugationsteps. The yield obtained with the optimized procedure was2.27±0.54 g (%CV 23.89) of MFGs per 100 g of milk, from whichthe average yield of protein extracted was found to be 74.70±18.28 mg (%CV 24.47) per 100 g of MFGs.

Extracted proteins were separated by SDS-PAGE andvisualized by Coomassie staining. As shown in Fig. 1, thecomplexity of isolated MFGMPs ranges from 15 kDa toover 250 kDa in molecular mass. Moreover, Fig. 1 shows thatthe electrophoretic profiles of MFGMPs obtained frommilk samples (n=6) of different Sarda sheep are highlycomparable.

3.2. Precipitation of proteins in the residual fat fraction

In order to assess the extraction efficiency of MFGMPs, the fatfraction remaining after protein extraction was solubilizedwith SDS and subjected to three different precipitationprocedures. The amount of protein recovered was quantifiedand comparedwith the total amount of protein extracted fromthe MFGM, revealing the presence of only 0.5% of total MFGMPin residual fat with the best extraction method, 10% TCA inacetone. SDS-PAGE analysis revealed that proteins remainingin the residual fat fraction were represented mainly by themost abundant proteins, showing that there was no selectiveloss of proteins during extraction. As shown in Fig. 2, theamount of proteins lost in the residual fat was minimal evenwith the best method used for recovering the proteins

to cellular localization (A) and function (B). Bars represent the

Fig. 4 – Quantitative GO classification of sheepMFGMPs according to cellular localization (A) and function (B). The average NSAFvalue of two biological replicates is reported. Error bars represent one SD from the mean.

354 J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

remaining in the residual fat fraction (Fig. 2, lane C). Thisanalysis further confirmed that the extraction procedure hadan extremely high efficiency and enabled the almost completerecovery of MFGMPs from ovine milk.

3.3. GeLC-MS/MS of sheep MFGMPs

AGeLC-MS/MS analysiswas performed on theMFGMP fractionof two different sheep in order to gather systematic proteomeinformation and to enable application of the label-freequantification approach. A total of 30 slices were cut fromthe SDS-PAGE gel lane containing the separated MFGMPs, andsubjected to MS/MS identification. Upon application of thismethod, 121 unique MFGM proteins were identified in bothsamples, plus 19 proteins identified only in replicate 1. In total,140 unique MFGM proteins were identified (Additional file 1reports the detailed list of all protein identifications. Addi-tional files 2 and 3 report annotated MS/MS spectra for singlepeptide identifications of replicates 1 and 2, respectively).

3.4. Data analysis and classification

A gene ontology (GO) classification was carried out on allproteins identified by GeLC-MS/MS. As shown in Fig. 3, uniqueproteins (140) were found to bemainly divided into: cytoplasm(38.37%), membrane (26.16%), endoplasmic reticulum (16.86%),

secreted (13.95%), nucleus (2.32%), mitochondrion (1.74%), andlysosome (0.58%). As we observed also in previous studies [41],the elevated number of membrane and endoplasmic reticu-lum classes highlights the high efficiency of SDS-PAGE inanalyzing liposoluble proteins, such as those extracted frommembrane compartments. All protein identifications werethen classified according to function (Fig. 3) into eight groups:membrane and vesicular trafficking (34%), protein synthesis,binding and folding (15%), enzymatic activity (15%), cellsignalling (14%), immune function (7%), fat transport/metab-olism (6%), milk proteins (4%), redox regulation (4%), andapoptosis (2%).

Comparison of these qualitative data with those availablefor the cow MFGM proteome showed a higher presence ofcytoplasmic and secreted proteins in sheep milk. Indeed, thebovine proteome appears composed for 71% of membrane,24% of cytoplasmic, and 5% of secreted proteins [29]. Thesedifferences could be due to numerous factors, such asdifferences in the efficiency of the mechanism governingformation and secretion of milk fat globules in sheep. Duringthis process, in fact, cytoplasmic material can be entrainedbetween the inner coat and the outer double membrane layerresulting in cytoplasmic crescents. These crescents can varyfrom thin slivers of cellular material to cases where thecrescent volume exceeds the fat globule core volume [18].Crescents have been observed in themilk of numerous species

Tab

le1–Listingof

them

ostab

unda

ntMFG

Mpr

oteinsob

tained

intw

obiolog

ical

replicates

byGeL

C-M

S/MSan

alys

is.

Rep

licate1

Rep

licate2

Proteinnam

eAcc

no.

Loca

lization

NSA

F1

Proteinnam

eAcc

no.

Loca

lization

NSA

F2

Butyro

philinsu

bfam

ily1mem

berA1

gi|125

6621

32Mem

bran

e0.15

8211

Butyro

philinsu

bfam

ily1mem

berA1

gi|125

6621

32Mem

bran

e0.21

0415

Xan

thinede

hyd

roge

nas

e/ox

idas

egi|119

7121

45Cytop

lasm

0.08

6803

Xan

thinede

hyd

roge

nas

e/ox

idas

egi|119

7121

45Cytop

lasm

0.11

9932

Milk

fatglob

ule-EGFfactor

8(Lac

tadh

erin)

gi|259

5096

35Mem

bran

e0.08

1086

Milk

fatglob

ule-EGFfactor

8(Lac

tadh

erin)

gi|259

5096

35Mem

bran

e0.07

1699

Adipo

sedifferen

tiation-related

protein

gi|157

2786

06Mem

bran

e0.06

661

Adipo

sedifferen

tiation-related

protein

gi|157

2786

06Mem

bran

e0.03

9217

Stom

atin

gi|157

4280

70Mem

bran

e0.03

5548

Stom

atin

gi|157

4280

70Mem

bran

e0.03

5995

Actin

gi|809

561

Cytop

lasm

0.02

9916

Ubiqu

itin

gi|229

532

Cytop

lasm

0.03

2251

Cas

einka

ppa

gi|571

6438

1Se

creted

0.02

8619

Annex

inI

gi|74

Cytop

lasm

0.02

8308

ATP-bindingca

ssette,s

ub-family

Ggi|118

4033

04Mem

bran

e0.01

9622

Alpha-S1

-cas

ein

gi|575

2646

9Se

creted

0.02

7748

Beta-lac

toglob

ulin

gi|540

3771

2Se

creted

0.01

8893

Clusterin

gi|278

0690

7Cytop

lasm

0.02

7176

Immunog

lobu

linlambd

alig

htch

ain

gi|523

6698

6Se

creted

0.01

854

Immunog

lobu

linlambd

alig

htch

ain

gi|523

6698

6Se

creted

0.02

5925

Clusterin

gi|278

0690

7Cytop

lasm

0.01

7566

Actin

gi|283

36Cytop

lasm

0.02

2458

Annex

inA3

gi|740

0200

9Cytop

lasm

0.01

7529

Lactoferrin

gi|254

6561

13Se

creted

0.02

1787

Ras

-related

proteinRab

-1A

isofor

m1

gi|475

8988

Mem

bran

e0.01

6855

Beta-lac

toglob

ulin

gi|540

3771

2Se

creted

0.01

7613

Alpha-S1

-cas

ein

gi|575

2646

9Se

creted

0.01

6537

ATP-bindingca

ssette,s

ub-family

Ggi|118

4033

04Mem

bran

e0.01

6097

Proteindisu

lfide-isom

eras

eA3

gi|729

433

Endo

plas

mic

reticu

lum

0.01

2135

Pero

xiredo

xin1

gi|278

0608

1Cytop

lasm

0.01

4489

Pero

xiredo

xin1

gi|278

0608

1Cytop

lasm

0.01

2089

PAS-4

gi|132

2373

Mem

bran

e0.01

418

Acc

.n,a

cces

sion

numbe

r;Lo

caliz

ation,C

ellularloca

lization;N

SAF,

nor

malized

spec

tral

abunda

nce

factor

.

355J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

and, most importantly, the proportion of globules withcrescents appears to vary both among species and amongindividuals within a species [18]. This would explain thequalitative enrichment in cytoplasmic proteins found in thesheep MFGM proteome. However, it needs to be pointed outthat the higher amount of cytoplasmic proteins found insheep MFGs compared to cow MFGs may be also due toclassification of some highly abundant proteins as cytoplas-mic protein in some studies, and as membrane-associated inothers; the subtle boundaries among classes, themanifold andintertwined protein functions, and the different roles that thesame protein plays in different cellular or tissue contextscreate difficulties in generation of GO classifications, andconsequently hinder their comparability among studies. Thisapplies especially since the results generated by most GOclassification algorithms are corrected by researchers basedon scientific knowledge, or left unchanged by “purists” of GOclassifications. For example, xanthine dehydrogenase/oxidase(XDH/XO) is a cytoplasmic enzyme, but it has been correctlyclassified as membrane-associated in studies on MFGMPs [24].

Quantitative estimations can undoubtedly provide betterinformation on MFGM protein composition. In fact, someprotein classesmay appear as beingmore represented under aqualitative perspective (i.e., number of protein identities), butother classes can be those showing the higher expressionlevels (i.e. protein abundance). In order to gain a quantitativepicture of the sheep MFGM proteome, abundance estimateswere obtained by means of the spectral counting approach[39,42–47]. The normalized spectral abundance factor (NSAF)generated from shotgun proteomics data sets was applied inthis study. In spectral counting, in fact, larger proteins areexpected to generate more peptides and therefore morespectral counts than smaller proteins. Consequently, thenumber of spectral counts for each protein must be dividedby the mass [42] or length, defining SAF. In this approach, thespectral counts of each protein were divided by its length andnormalized to the total sum of spectral counts/length in agiven analysis [39]. The use of NSAF showed to be a powerfultool for obtaining information about the abundance ofproteins present in the MFGM. This analysis revealed thatmembrane proteins are the highest expressed protein cate-gory according to cellular localization (Fig. 4A). The use ofNSAF for proteins categorized according to function showedthe considerably higher expression levels of membrane andvesicular trafficking proteins (Fig. 4B), as expected. In fact,membrane proteins are those that play a major role in theformation and budding of milk fat globules.

The most abundant proteins in sheep MFGM were butyr-ophilin subfamily 1 member A1 (BTN) (mean NSAF 0.185±0.026), xanthine dehydrogenase/oxidase (XDH/XO) (meanNSAF 0.103±0.017), lactadherin (milk fat globule-EGF factor8) (mean NSAF 0.076±0.004), and adipophilin (ADPH) (adiposedifferentiation-related protein, perilipin-2) (mean NSAF0.053±0.014), as shown in Table 1 and Fig. 5 (for a detailedlist, see Additional file 1).

BTN, a type I transmembrane glycoprotein with a cyto-plasmic C-terminal tail, is concentrated in the apical plasmamembranes of mammary epithelial cells [18,48]. BTN isexpressed only during lactation, and it appears to be essentialfor milk fat globule production [18,49,50]. Two potential

Fig. 5 – Most abundant proteins in sheep MFGM. The average NSAF value of two biological replicates is reported. Error barsrepresent one SD from the mean.

356 J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

interactive partners of BTN are microtubules [51] and XDH/XO[52], which has been shown to bind to the cytoplasmic tail ofBTN in vitro assays [53]. Furthermore, formation of a supra-molecular complex among XDH/XO, BTN, and proteins on thesurface of cytoplasmic lipid droplets, such as ADPH [54,55],may be an essential step in the assembly of the MFGM andsubsequent expulsion of lipid droplets from the cell. XDH/XOmay be viewed as a linker protein between BTN in themembrane bilayer and proteins on the surface of the lipiddroplet. Apart from being a major protein of MFGM, XDH/XOis a well-known and extensively characterized redox/purinecatabolizing enzyme [56]. XDH/XO is a soluble cytosolicenzyme, but in milk-secreting cells this protein is concentrat-ed along the inner face of the apical plasma membrane andthe MFGM [52]. XDH/XO may play a structural and functionalrole in the formation of the MFGM by binding to thecytoplasmic domain of BTN probably by means of disulfidebonds [53], thus forming a protein complex that interacts withlipid droplets at the apical cell surface. The presence of highamounts of disulfide isomerase in the MFG, detected as one ofthe highest abundance proteins also in this study (Table 1),does also support this hypothesis. In sheep, we observed aBTN to XDH/XO ratio of about 1.8:1 (Table 1). The ratio betweenthese two proteins, whose role is crucial formilk secretion andshows apparent relationships with fat content and MFG size,varies according to the species. Using molecular weights of149.107 kDa for XDH/XO and 59.912 kDa for BTN, the xanthineoxidase to butyrophilin molar ratio resulted approximately ofabout 1:4.44. In cow milk, XDH/XO and BTN are present inconstant molar proportions of about 1:4 [57,58], suggestingthat the same constant molar ratio exists between these twoproteins in theMFGMof sheep and cow. In humanmilk, on thecontrary, the levels of XO have been reported to be muchhigher than those of BTN [59]. In goat, Zamora and coworkers[60] calculated a 1:1 molar ratio between the two proteins. Itwill be interesting to investigate the reasons for such species-to species variation, and for its stability within the species, in

the relative amounts of these two proteins with a crucial rolein MFG secretion.

ADPH, which has a very high affinity for triglycerides, islocated in the inner polar lipid monolayer [22]. The onlyknown proteins similar to ADPH are members of perilipinfamily [18] and TIP47, a protein that functions in membranetraffic between endosomes and the trans-Golgi network [61].The function of ADPH remains unknown. Possible roles in thecellular uptake of long-chain fatty acids and in the accretionand transport of intracellular lipid have been suggested[54,55,62,63]. Lactadherin (milk fat globule-epidermal growthfactor 8) is one of the major glycoproteins of the MFGM ofmany species [13]. As suggested by Oshimal and coworkers,lactadherin may play a possible role in the membrane vesiclesecretion, such as budding or shedding of plasma membrane(micro-vesicles) and exocytosis of endocytic multivesicularbodies (exosomes) [64]. Other membrane and vesiculartrafficking proteins identified in sheep MFGMs include theRab proteins, lowmolecular-weight GTP-binding proteins thatcoordinate stages of transport in the secretory pathway, andannexins, which include a group of calcium-dependentmembrane aggregating proteins that can initiate contactsbetween secretory vesicle membranes, which subsequentlyfuse [65]. The role of cytoskeletal proteins in the excretionmechanism emerges also from the detection of actin, myosin,and gelsolin, which play a role in cytoskeletal processesmediating secretion of MFGs. Our analysis detected also anumber of proteins involved in transport of fatty acids, in thecontrol of lipid metabolism and in the accretion of lipiddroplets in the cytoplasm. These proteins were fatty acid-binding protein, fatty acid synthase, apolipoprotein A and E.Typical milk proteins, such as caseins, lactoferrin and beta-lactoglobulin, accounted for 4% of the identified spots; thesewere probably present as contaminants, since they are usuallydetected in the skimmed milk fraction, although theirpresence might also be derived by inclusion of secretionvescicles within the cytoplasmic crescents of MFGs.

357J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

4. Conclusion

In conclusion, this study reports an optimized purificationmethod and the qualitative and quantitative characterization ofthe sheep MFGM proteome. The data gathered in this studyprovide a comprehensive view of this important milk fraction,and lay the foundations for future studies on physiology ofsheep lactationandformonitoring its alterations inpathologicalconditions.

Acknowledgements

This work was supported by funding from Regione Autonomadella Sardegna – “Progetto Cluster Proteomica”.

Appendix A. Supplementary Data

Supplementary data to this article can be found online atdoi:10.1016/j.jprot.2010.11.011.

R E F E R E N C E S

[1] Bauman DE, Mather IH, Wall RJ, Lock AL. Major advancesassociated with the biosynthesis ofmilk. J Dairy Sci 2006;89(4):1235–43.

[2] Affolter M, Grass L, Vanrobaeys F, Casado B, Kussmann M.Qualitative and quantitative profiling of the bovine milk fatglobule membrane proteome. J Proteomics 2009;73(6):1079–88.

[3] Daniels MC, Adair LS. Breast-feeding influences cognitivedevelopment in Filipino children. J Nutr 2005;135(11):2589–95.

[4] German JB, Dillard CJ, Ward RE. Bioactive components inmilk.Curr Opin Clin Nutr Metab Care 2002;5(6):653–8.

[5] Harmsen HJM, Wildeboer-Veloo ACM, Raangs GC, WagendorpAA, Klijn N, Bindels JG, et al. Analysis of intestinal floradevelopment in beast-fed and formula-fed infants by usingmolecular identification and detection methods. J PediatrGastroenterol Nutr 2000;30(1):61–7.

[6] O'Donnell R, Holland JW, Deeth HC, Alewood P. Milkproteomics. Int Dairy J 2004;14(12):1013–23.

[7] Riccio P. The proteins of the milk fat globule membrane in thebalance. Trends Food Sci Technol 2004;15(9):458–61.

[8] Dewettinck K, Rombaut R, Thienpont N, Le TT, Messens K,Van Camp J. Nutritional and technological aspects of milk fatglobule membrane material. Int Dairy J 2008;18(5):436–57.

[9] Fong BY, Norris CS. Quantification of milk fat globulemembrane proteins using selected reaction monitoring massspectrometry. J Agric Food Chem 2009;57(14):6021–8.

[10] Kanno C. Secretory membranes of the lactating mammarygland. Protoplasma 1990;159(2):184–208.

[11] Vanderghem C, Blecker C, Danthine S, Deroanne C, HaubrugeE, Guillonneau F, et al. Proteome analysis of the bovine milkfat globule: enhancement ofmembrane purification. Int DairyJ 2008;18(9):885–93.

[12] KeenanTW.Historical perspective:milk lipid globules and theirsurrounding membrane: a brief history and perspectives forfuture research. J Mammary Gland Biol Neoplasia 2001;6(3):365–71.

[13] Mather IH. A review and proposed nomenclature for majorproteins of the milk-fat globule membrane. J Dairy Sci 2000;83(2):203–47.

[14] Mather IH, Keenan TW. Origin and secretion of milk lipids. JMammary Gland Biol Neoplasia 1998;3(3):259–73.

[15] Brown DA. Lipid droplets: proteins floating on a pool of fat.Curr Biol 2001;11(11):R446–9.

[16] Deurs BV, Roepstorff K, Hommelgaard AM, Sandvig K.Caveolae: anchored, multifunctional platforms in the lipidocean. Trends Cell Biol 2003;13(2):92–100.

[17] Ostermeyer AG, Paci JM, Zeng Y, Lublin DM, Munro S, BrownDA. Accumulation of caveolin in the endoplasmic reticulumredirects the protein to lipid storage droplets. J Cell Biol2001;152(5):1071–8.

[18] Heid HW, Keenan TW. Intracellular origin and secretion ofmilk fat globules. Eur J Cell Biol 2005;84(2–3):245–58.

[19] Murphy DJ. The biogenesis and functions of lipid bodies inanimals, plants and microorganisms. Prog Lipid Res 2001;40(5):325–438.

[20] Charlwood J, Hanrahan S, Tyldesley R, Langridge J, Dwek M,Camilleri P. Use of proteomic methodology for thecharacterization of human milk fat globular membraneproteins. Anal Biochem 2002;301(2):314–24.

[21] Fortunato D, Giuffrida MG, Cavaletto M, Garoffo LP, DellavalleG, Napolitano L, et al. Structural proteome of human colostralfat globule membrane proteins. Proteomics 2003;3(6):897–905.

[22] Quaranta S, Giuffrida MG, Cavaletto M, Giunta C, Godovac-Zimmermann J, CañasB, et al. Humanproteomeenhancement:high-recovery method and improved two-dimensional map ofcolostral fat globule membrane proteins. Electrophoresis2001;22(9):1810–8.

[23] Wu CC, Howell KE, Neville MC, Yates 3rd JR, McManaman JL.Proteomics reveals a link between the endoplasmic reticulumand lipid secretory mechanisms in mammary epithelial cells.Electrophoresis 2000;21(16):3470–82.

[24] Bianchi L, Puglia M, Landi C, Matteoni S, Perini D, Armini A,et al. Solubilization methods and reference 2-DE map of cowmilk fat globules. J Proteomics 2009;72(5):853–64.

[25] D'Ambrosio C, Arena S, Salzano AM, Renzone G, Ledda L,Scaloni A. A proteomic characterization of water buffalo milkfractionsdescribing PTMofmajor species and the identificationofminor components involved in nutrient delivery anddefenseagainst pathogens. Proteomics 2008;8(17):3657–66.

[26] Cebo C, Caillat H, Bouvier F, Martin P. Major proteins of thegoat milk fat globule membrane. J Dairy Sci 2010;93(3):868–76.

[27] Martini M, Salari F, Pesi R, Tozzi MG. Relationship betweenactivity of some fat globule membrane enzymes and thelipidic fraction in ewes' milk: Preliminary studies. Int Dairy J2010;20(1):61–4.

[28] Spitsberg VL. Bovinemilk fat globulemembrane as a potentialnutraceutical. J Dairy Sci 2005;88(7):2289–94.

[29] Reinhardt TA, Lippolis JD. Bovine milk fat globule membraneproteome. J Dairy Res 2006;73(04):406–16.

[30] Carboni L, Piubelli C, Righetti PG, Jansson B, Domenici E.Proteomic analysis of rat brain tissue: comparison of protocolsfor two-dimensional gel electrophoresis analysis based ondifferent solubilizing agents. Electrophoresis 2002;23(24):4132–41.

[31] Jiang L, He L, Fountoulakis M. Comparison of proteinprecipitation methods for sample preparation prior toproteomic analysis. J Chromatogr A 2004;1023(2):317–20.

[32] Mastro R, Hall M. Protein delipidation and precipitation bytri-n-butylphosphate, acetone, and methanol treatment forisoelectric focusing and two-dimensional gel electrophoresis.Anal Biochem 1999;273(2):313–5.

[33] Laemmli UK. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature 1970;227(5259):680–5.

[34] Candiano G, Bruschi M, Musante L, Santucci L, Ghiggeri GM,Carnemolla B, et al. Blue silver: a very sensitive colloidalCoomassie G-250 staining for proteome analysis.Electrophoresis 2004;25(9):1327–33.

358 J O U R N A L O F P R O T E O M I C S 7 4 ( 2 0 1 1 ) 3 5 0 – 3 5 8

[35] Addis MF, Tanca A, Pagnozzi D, Rocca S, Uzzau S. 2-D PAGEand MS analysis of proteins from formalin-fixed,paraffin-embedded tissues. Proteomics 2009;9(18):4329–39.

[36] Addis MF, Tanca A, Pagnozzi D, Crobu S, Fanciulli G,Cossu-Rocca P, et al. Generation of high-quality proteinextracts from formalin-fixed, paraffin-embedded tissues.Proteomics 2009;9(15):3815–23.

[37] Dupierris V, Masselon C, Court M, Kieffer-Jaquinod S, Bruley C.A toolbox for validation of mass spectrometry peptidesidentification and generation of database: IRMa.Bioinformatics 2009;25(15):1980–1.

[38] Saydam O, Senol O, Schaaij-Visser TBM, Pham TV, PiersmaSR, Stemmer-Rachamimov AO, et al. Comparative proteinprofiling reveals minichromosome maintenance (MCM)proteins as novel potential tumormarkers for meningiomas. JProteome Res 2009;9(1):485–94.

[39] Zhang Y, Wen Z, Washburn MP, Florens L. Refinements tolabel free proteome quantitation: how to deal with peptidesshared by multiple proteins. Anal Chem 2010;82(6):2272–81.

[40] Fong BY, Norris CS, MacGibbon AKH. Protein and lipidcomposition of bovinemilk-fat-globule membrane. Int Dairy J2007;17(4):275–88.

[41] Cacciotto C, Addis M, Pagnozzi D, Chessa B, Coradduzza E,Carcangiu L, et al. The liposoluble proteome of Mycoplasmaagalactiae: an insight into theminimal protein complement ofa bacterial membrane. BMC Microbiol 2010;10(1):225.

[42] Blondeau F, Ritter B, Allaire PD, Wasiak S, Girard M, HussainNK, et al. Tandem MS analysis of brain clathrin-coatedvesicles reveals their critical involvement in synaptic vesiclerecycling. Proc Natl Acad Sci USA 2004;101(11):3833–8.

[43] Girard M, Allaire PD, McPherson PS, Blondeau F. Non-stoichiometric relationship between clathrin heavy and lightchains revealed by quantitative comparative proteomics ofclathrin-coated vesicles from brain and liver. Mol CellProteomics 2005;4(8):1145–54.

[44] Liu H, Sadygov RG, Yates 3rd JR. Amodel for random samplingand estimation of relative protein abundance in shotgunproteomics. Anal Chem 2004;76(14):4193–201.

[45] OldWM,Meyer-Arendt K, Aveline-Wolf L, Pierce KG, MendozaA, Sevinsky JR, et al. Comparison of label-free methods forquantifying human proteins by shotgun proteomics. Mol CellProteomics 2005;4(10):1487–502.

[46] Powell DW,Weaver CM, Jennings JL, McAfee KJ, He Y,Weil PA,et al. Cluster analysis of mass spectrometry data reveals anovel component of SAGA. Mol Cell Biol 2004;24(16):7249–59.

[47] Zybailov B, ColemanMK, Florens L,WashburnMP. Correlationof relative abundance ratios derived from peptide ionchromatograms and spectrum counting for quantitativeproteomic analysis using stable isotope labeling. Anal Chem2005;77(19):6218–24.

[48] Singh H. The milk fat globule membrane – a biophysicalsystem for food applications. Curr Opin Colloid Interface Sci2006;11(2–3):154–63.

[49] McManaman JL, Russell TD, Schaack J, Orlicky DJ, Robenek H.Molecular determinants of milk lipid secretion. J MammaryGland Biol Neoplasia 2007;12(4):259–68.

[50] Robenek H, Hofnagel O, Buers I, Lorkowski S, Schnoor M,Robenek MJ, et al. Butyrophilin controls milk fat globulesecretion. Proc Natl Acad Sci USA 2006;103(27):10385–90.

[51] Schweiger S, Foerster J, Lehmann T, Suckow V, Muller YA,Walter G, et al. The Opitz syndrome gene product, MID1,associates with microtubules. Proc Natl Acad Sci USA 1999;96(6):2794–9.

[52] Aoki N. Regulation and functional relevance of milk fatglobules and their components in themammary gland. BiosciBiotechnol Biochem 2006;70(9):2019–27.

[53] Ishii T, Aoki N, Noda A, Adachi T, Nakamura R, Matsuda T.Carboxy-terminal cytoplasmic domain of mouse butyrophilinspecifically associates with a 150-kDa protein of mammaryepithelial cells and milk fat globule membrane. BiochimBiophys Acta 1995;1245(3):285–92.

[54] Heid HW, Moll R, Schwetlick I, Rackwitz HR, Keenan TW.Adipophilin is a specific marker of lipid accumulation indiverse cell types and diseases. Cell Tissue Res 1998;294(2):309–21.

[55] Heid HW, Schnolzer M, Keenan TW. Adipocytedifferentiation-related protein is secreted into milk as aconstituent of milk lipid globule membrane. Biochem J1996;320(Pt 3):1025–30.

[56] Harrison R. Milk xanthine oxidase: properties andphysiological roles. Int Dairy J 2006;16(6):546–54.

[57] Mondy BL, Keenan TW. Butyrophilin and xanthine oxidaseoccur in constant molar proportions in milk lipid globulemembrane but vary in amount with breed and stage oflactation. Protoplasma 1993;177(1):32–6.

[58] Ye A, Singh H, Taylor MW, Anema S. Characterization ofprotein components of natural and heat-treated milk fatglobule membranes. Int Dairy J 2002;12(4):393–402.

[59] Freudenstein C, Keenan TW, Eigel WN, Sasaki M, Stadler J,Franke WW. Preparation and characterization of the innercoat material associated with fat globule membranes frombovine and human milk. Exp Cell Res 1979;118(2):277–94.

[60] Zamora A, Guamis B, Trujillo AJ. Protein composition ofcaprine milk fat globule membrane. Small Rumin Res 2009;82(2):122–9.

[61] Wolins NE, Rubin B, Brasaemle DL. TIP47 associates with lipiddroplets. J Biol Chem 2001;276(7):5101–8.

[62] Brasaemle DL, Barber T, Wolins NE, Serrero G,Blanchette-Mackie EJ, Londos C. Adiposedifferentiation-related protein is an ubiquitously expressedlipid storage droplet-associated protein. J Lipid Res 1997;38(11):2249–63.

[63] Gao J, Serrero G. Adipose differentiation related protein(ADRP) expressed in transfected COS-7 cells selectivelystimulates long chain fatty acid uptake. J Biol Chem 1999;274(24):16825–30.

[64] Oshima K, Aoki N, Kato T, Kitajima K, Matsuda T. Secretion ofa peripheral membrane protein, MFG-E8, as a complex withmembrane vesicles. Eur J Biochem 2002;269(4):1209–18.

[65] Wu CC, Yates JR, Neville MC, Howell KE. Proteomic analysis oftwo functional states of the Golgi complex in mammaryepithelial cells. Traffic 2000;1(10):769–82.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具