Bovine Milk Fat Globule Membrane Proteome - Semantic Scholar

Journal of Dairy Research (2006) 73 406–416. f Proprietors of Journal of Dairy Research 2006 doi:10.1017/S0022029906001889 Printed in the United Kingdom

406

Bovine Milk Fat Globule Membrane Proteome Timothy A Reinhardt* and John D Lippolis Periparturient Diseases of Cattle Research Unit, USDA-ARS, National Animal Disease Center, Ames, IA. 50010 USA Received 5 December 2005 and accepted for publication 6 March 2006

Milk fat globule membranes (MFGM) were isolated from the milk of mid-lactation Holstein cows. The purified MFGM were fractionated using 1-dimensional SDS gels. Tryptic peptides from gel slices were further fractionated on a micro-capillary high performance liquid chromatograph connected to a nanospray-tandem mass spectrometer. Analysis of the data resulted in 120 proteins being identified by two or more unique peptide sequences. Of these 120 proteins, 71 % are membrane associated proteins with the remainder being cytoplasmic or secreted proteins. Only 15 of the proteins identified in the cow MFGM were the same as proteins identified in previous mouse or human MFGM proteomic studies. Thus, the bulk of the proteins identified are new for bovine MFGM proteomics. The proteins identified were associated with membrane/protein trafficking (23 %), cell signalling (23 %), unknown functions (21 %), fat transport/metabolism (11 %), transport (9 %), protein synthesis/folding (7%), immune proteins (4 %) and milk proteins (2%). The proteins associated with cell signalling or membrane/ protein trafficking may provide insights into MFGM secretion mechanisms. The finding of CD14, toll like receptor (TLR2), and TLR4 on MFGM suggests a direct role for the mammary gland in detecting an infection. Keywords : Milk fat globule membrane, proteomics, lactation, mammary gland, mastitis.

Milk production and secretion is a complex process that warrants more basic research to further our understanding of these processes. Proteomics is a tool that will help identify proteins important to milk production and secretion. Identification of proteins associated with various aspects of milk production and secretion will provide a foundation for new research in lactation biology. The proteomic studies conducted thus far on mammary epithelial cells, organelles, membranes and secretion processes (Wu et al. 2000a, b ; Quaranta et al. 2001; Charlwood et al. 2002; Pucci-Minafra et al. 2002; Fortunato et al. 2003; Jacobs et al. 2004) are focused on breast cancer and/or rodent lactation. While these studies have advanced our understanding of mammary function and milk secretion, they may not address the unique aspects of milk secretion in dairy cattle. One area of milk secretion that is lacking in molecular details is that of milk fat secretion (Mather & Keenan, 1998a; Keenan, 2001). This scarcity of information is due in large part to the lack of cell lines that secrete milk and milk fat (Keenan, 2001). The MFGM is a rich source of membrane proteins and proteomic analysis of these

*For correspondence ; e-mail: [email protected]

membranes may highlight some of the signalling and secretory pathways used by the mammary gland. Furthermore, the proteome of the MFGM provides additional insight into this membrane’s cellular origin. The most widely accepted source of membrane for the MFGM is the apical membrane of the secretory cell (Mather & Keenan, 1998a ; Keenan, 2001). Specifically, the outer envelope of the MFGM is derived from the apical plasma membrane. Their conclusions are supported by biochemical, electron microscopy and immunocytochemical evidence. Many MFGM proteins are also derived from intracellular sources including the endoplasmic reticulum, and the surface on intracellular lipid droplets (Wu et al. 2000a).The other theory suggests that the MFGM originates from secretory vesicle membranes (Wooding, 1971, 1973). While this secretory vesicle mechanism may contribute to some membrane to the MFGM, the apical membrane of the secretory cell seems to be the primary source for MFGM (Mather & Keenan, 1998a; Keenan, 2001). The major proteins in the MFGM have been identified using traditional biochemical approaches (Mather, 2000). These methods are slow, laborious and address only one protein at a time. Proteomic approaches yield the identities of many proteins in one experiment. To date

Milk fat globule membrane proteome the proteomic approaches have been applied to human MFGM (Quaranta et al. 2001; Charlwood et al. 2002 ; Fortunato et al. 2003) and mouse MFGM (Wu et al. 2000a). In these 4 papers only 6–45 proteins were identified in the MFGM. These authors used 2 dimensionalelectrophoresis (2DE) protocols to do the initial fractionation of the MFGM proteins. However, 2DE is problematic with membrane proteins. Many intrinsic membrane proteins are lost in 2DE due to precipitation at the isoelectric point or the proteins are difficult to get into solution prior to iso-electric focusing. Omitting the iso-electric focusing step and fractionating the MFGM directly using 1 dimensional-electrophoresis SDS gel electrophoresis (1DE) overcomes the solubility problems (Gu et al. 2003; Peirce et al. 2004). The loss of the high resolution provided by 2DE can be overcome by high performance liquid chromatography tandem mass spectroscopy (MS/MS). In this report the MS/MS approach to analyse a complex mixture of tryptic peptides from 1DE slices. Using this approach, 120 proteins were identified in cow MFGM.

Materials and Methods Milk Fat Globule Membrane Preparation Milk was collected from Holstein cows in mid-lactation. The milk from 5 cows was cooled to 4 8C, pooled and centrifuged at 10 000 g for 15 min. at 4 8C. The floating milk fat pellet was removed, mixed with 5 volumes of ice cold phosphate buffered saline (pH 7)+complete protease inhibitor cocktail (complete mix of serine and cysteine protease inhibitors) from Boehringer Mannheim (Indianapolis, IN) and centrifuged at 10 000 g for 15 min. This washing step was repeated 3 times until the supernatant was clear. MFGM were prepared from the washed milk fat as previously described (Reinhardt et al. 2000). Washed milk fat was diluted in 10 volumes of Buffer A which contained: 10 mM-Tris-HCl, 2 mM-MgCl2, 0.1 mM-phenylmethylsulphonyl fluoride, 1 mM-EDTA, 4 mg aprotinin/ml and 4 mg leupeptin/ml at pH 7.5. The sample was homogenized using a Polytron PT-10 (Brinkman Instruments, Boston, MA) running at 12 000 rpm. Each homogenization step was for 12 sec with 30 sec of sample cooling between each homogenization run. A total of three 12 sec. homogenizations were performed on the sample. The homogenate was mixed with an equal volume of Buffer B (Buffer A plus 300 mM-KCl) and centrifuged at 100 000 g for 1 h. The supernatant was discarded and the membrane pellet was resuspended in Buffer C (Buffer A plus 150 mM-KCl). The resuspended membrane preparation was centrifuged at 100 000 g for 1 h. The supernatant was discarded and MFGM pellet was resuspended in buffer A. Protein concentration was determined using the BioRad Protein Assay Kit using a bovine serum albumin (BSA) standard. MFGM were stored at – 70 8C until needed.

407

Extraction of Intrinsic Proteins MFGM prepared as described above were pelleted by centrifugation at 100 000 g for 1 h. All procedures were done at 4 8C. The membrane pellet was resuspended in a small amount of 300 mM-sucrose, 10 mM-Tris-HCL at pH 7.5. This suspension was diluted with ice-cold 0.1 M-sodium carbonate (pH 11.5) to a protein concentration 0.01 mg/ml (Fujiki et al. 1982). The sample was incubated on ice for 1 h and then centrifuged at 100 000 g for 1 h through a cushion of 300 mM-sucrose (10 % of the tube volume). The MFGM intrinsic protein pellet was resuspended in buffer A. Protein concentration was determined and the extracted MFGM were stored at – 70 8C until needed.

Gel Electrophoresis, Protein Alkylation and Ingel Trypsin Digestion One mg extracted MFGM was pelleted by centrifugation at 100 000 g for 1 h. The MFGM was resuspended in a modified Laemmli buffer containing 150 mg urea/ml and 65 mM-DTT and heated to 90 8C for 10 min. The sample was loaded and electrophoresed for 1.5 h at 125 volts on an 8–16 % Tris-glycine gel (Novex, San Diego, CA). At the end of electrophoresis the gel was stained with Pierce Gel Code Blue for 1 h and destained overnight in water. The gel was then cut into 40 slices from the bottom to the top of the gel. These slices represented molecular weights from y6–250 KDa. Each slice was cut into 1 mm cubes and placed into a siliconized microcentrifuge tube. All NH4HCO3 buffer used in this section is pH 8. The diced gel slices were destained with 3 washes of 50 : 50 (vol/vol) acetonitrile (ACN) : 25 mM-NH4HCO3 followed by one wash in 80 : 20 (vol/vol) ACN : 25 mM-NH4HCO3. The gel slices were then dried for 15 min in a vacuum centrifuge. Next, the samples were rehydrated in just enough freshly prepared 10 mM-Tris(2-carboxethyl)phosphine hydrochloride (Pierce Biotechnology Inc, Rockford, IL) in 25 mM-NH4HCO3, to cover the gel slices and incubated for 1 h at 56 8C. The samples were cooled to room temperature and the Tris(2-carboxethyl)phosphine hydrochloride solution removed and discarded. Freshly prepared 55 mM-Iodoacetamide in 25 mMNH4HCO3 was added to the gel slices and they were incubated for 1 h at room temperature in the dark. This solution was discarded and the slices were washedr2 with 50 : 50 (vol/vol) ACN : 25 mM-NH4HCO3 followed by one wash with 80 : 20 (vol/vol) ACN : 25 mM-NH4HCO3. The gels slices were dried for 15 min in a vacuum centrifuge. Next, sufficient 25 mM-NH4HCO3 was added to rehydrate the gel slices. After adding cap locks, the proteins were thermally denatured at 90 8C for 20 min as described (Park & Russell, 2000, 2001). The slices were then cooled on ice for 10 min, the supernatant discarded and the gel slices were washedr2 with 50 : 50 (vol/vol) ACN : 25 mM-NH4HCO3 followed by one wash with

408

TA Reinhardt and JD Lippolis

80 : 20 (vol/vol) ACN : 25 mM-NH4HCO3. The gel slices were then dried in a vacuum centrifuge. For trypsin digestion, just enough proteomic grade trypsin (20 ug/ml in 25 mM-NH4HCO3) was added to cover the gel slices. After the gel slices were rehydrated, ACN was added so that the solution was 30 % ACN (Russell et al. 2001). This ACN/trypsin solution was incubated at 37 8C overnight. The next day the samples were cooled to room temperature. The digest solution was transferred to clean siliconized tubes. The gel slices were extractedr2 with 50 : 50 (vol/vol) ACN : 5 % formic acid,r1 with 15 : 50 : 35 (vol/vol) isopropanol : ACN : 5 % formic acid, andr2 with 80 % ACN. All extracts were combined with the digest solution and dried in a vacuum centrifuge. The samples were stored dry at – 20 8C until used. High Performance Chromatography and Tandem Mass Spectroscopy of the Samples The tryptic peptides from each gel slice were analysed by CapLC-nanospray-MS/MS using a Q-TOF Ultima API mass spectrometer (Waters-Micromass, Milford Ma. ; Lippolis & Reinhardt, 2005). An Atlantis C18 NanoEase column (75 umr100 mm) was used for peptide separation. The system was configured to concentrate and wash the injected sample on a Symmetry 300 C18 precolumn. Seven min after the start of sample loading the precolumn was switched in line with the analytical column to allow the trapped peptides to be eluted onto the analytical column. Mobil phase A was 0.1 % formic acid in 5% ACN. Mobil phase B was 0.1 % formic acid in 95 % ACN. The gradient was 95 % A for 5 min and then was ramped linearly to 40 % B over 85 min. Over the next 2.5 min it was ramped to 90 % B and held for 10 additional min before re-equilibration of the column. The flow rate was 300 nl/min. The analytical column was connected to WatersMicromass Lockspray-nanospray interface on the front of the mass spectrometer. The lockspray used the peptides fibriopeptide b and leucine enkephalin as standards. The capillary voltage was 3500 V and was tuned for signal intensity. The 5 most intense ions, with charge states between 2 and 4, were selected in each survey scan if they met the switching criteria for a MS/MS scan. Three collision energies were used to fragment the peptides ions based on their m/z values. Protein Identification All peptide identifications were done using MS/MS data only. Post-run processing and mass correction of the MS/MS data was done using ProteinLynx Global Server 2.0 (Waters-Micromass, Milford Ma.) as was database searching. MS/MS scoring, statistics and protein identification algorithms are described in detail in appendix c of ProteinLynx Global Server 2.0 Users Guide (Waters, 2004). Proteins were identified using The Universal Protein

Non-redundant Reference database UniRef 100 from Swiss Prot (http://www.pir.uniprot.org/database/nref_blast. shtml). Modifications allowed were one missed cleavage per peptide, methionine oxidation and carboxyaminomethylation in the first pass search. Mass accuracy of peptide fragment ions was set at 20 ppm for all searches. All protein identifications were P < 0.05 using the search and protein identification program ProteinLynx Global Server 2.0. In general, 3 or more peptides from a protein were required for a firm identification. However a very high quality MS/MS spectrum would provide sufficient confidence for positive protein identification using the PepSeq. For proteins just below P < 0.05 (those with only 2 peptides), the peptide sequences were manually checked using the PepSeq (Waters-Micromass, Milford Ma.) to confirm the sequence assignments at P < 0.05. Western Blotting MFGM were incubated for 15 min at room temperature in a modified Laemmli buffer containing 150 mg urea/ml, 65 mM-DTT and 40 g SDS/l. Samples were then electrophoresed for 1.5 h at 125 volts (20 mA at the start of a run) in a 6 % Tris-glycine gel for PMCA2 or a 12 % gel for Snap23 and Rab3A (Novex, San Diego, CA). Proteins were transferred to nitrocellulose membranes for 1 h at 25 volts (600 mA at the start of a run) in 0.7 M-glycine, 0.025 M-Tris at pH 7.4. Blots were developed using Pierce’s Supersignal (Pierce Products, Rockford IL) using the protocol provided by the manufacturer. Anti-PMCA2 antibody was described previously (Reinhardt et al. 2000, 2004). Anti-Rab3A and anti-SNAP23 were purchased from Affinity BioReagents (Deerfield, IL).

Results Figure 1 shows a 1DE gel image for the purified MFGM proteins. The large band just below the 64 kDa standard is butyrophilin. Forty gel slices from a larger gel were prepared as described in the Material and Methods section for tandem mass spectral analysis and protein identification. Table 1 presents a list of the 120 proteins, by general functional groups, found in MFGM prepared from Holstein cows in mid-lactation. Of these 120 proteins, 71 % are membrane associated proteins with the remainder being either cytoplasmic or secreted proteins (Fig. 2A). The proteins with an asterisk (Table 1, a, d, f and h) are proteins common to those found in other proteomic studies of the MFGM in human and mouse (Cavaletto et al. 1999; Wu et al. 2000a ; Fortunato et al. 2003; Jacobs et al. 2004). The identified proteins were divided into 8 broad functional categories based primarily on functions listed in the Swiss Prot database. Membrane/protein trafficking proteins (23 %) and cell signalling proteins (23 %) accounted for almost half of the proteins identified (Fig. 2B). The next most abundant group at 21 % was proteins with unknown

Milk fat globule membrane proteome 250 148

98 64 50 36

22 16

Fig. 1. Gel image of purified milk fat globule membrane proteins separated on a 1-dimensional SDS gel electrophoresis run using an 8–16 % gradient gel. The protein load was 30 mg. Left lane is molecular weight standards with weights of standards noted and the right lane is milk fat globule membrane proteins. A larger gel with 1 mg of membrane protein loaded was used to prepare the 40 gel slices used in this study.

or ‘poorly’ defined functions. The remainder of the proteins were associated with fat transport/metabolism (11 %), general transporters (9 %), protein synthesis/folding (7 %), immune proteins (4 %) and milk proteins (2 %). Figure 3 shows a MS/MS spectrum used to confirm the presence of Rab 7 in the MFGM proteome when only 2 peptides were identified by ProteinLynx Global Server 2.0 (P> 0.05). This MS/MS spectrum was de novo sequenced with PepSeq and the sequence was determined to be EAINVEQAFQTIAR with a 99 % probability (P < 0.01) thus confirming the presence of RAB7 in this proteome. Figure 4 shows Western blots confirming the presence of 3 proteins identified in the milk fat globule membrane proteome. Plasma membrane Ca2+ATPase isoform 2 (PMCA2), SNAP23 and Rab3A were identified in the milk fat globule proteome by 2, 6 and 7 peptides respectively and confirmed by Western blotting.

Discussion Previously, proteins in or associated with MFGM have been identified with traditional biochemical methods (Mather, 2000). To date, proteomics approaches have been applied to human and mouse MFGM with the identification of new proteins in the MFGM (Wu et al. 2000a ; Quaranta et al. 2001 ; Charlwood et al. 2002 ; Fortunato et al. 2003). The resulting data from these papers are somewhat limited due to their reliance on 2DE, which results in large losses of membrane proteins (Gu et al. 2003; Peirce et al. 2004). In an attempt to identify more proteins associated with the membrane of MFGM, the

409

MFGM were enriched for intrinsic membrane proteins, separated the proteins by 1DE and identified them by tandem mass spectrometry. With this approach 120 proteins were identified in the bovine MFGM proteome (Table 1). Proteomic analysis is complicated by the extreme dynamic range seen in protein abundances. It is estimated that cells may contain proteins with as few as 10 copies per cell ranging up to y1 000 000 copies per cell (Moritz et al. 2004). For this reason it is important to simplify the proteome prior to analysis. This simplification was done by extracting the MFGM with carbonate to remove extrinsic proteins (Fujiki et al. 1982), in order to focus on a simpler MFGM proteome that was enriched for intrinsic proteins. The resulting MFGM preparation was judged highly enriched for intrinsic proteins by both the high percentage of membrane proteins identified and the almost total lack of skim milk peptides found in the sample. Tandem mass spectrometry instruments focus on the most abundant ions and need several seconds to obtain enough data to generate useful peptide spectra. During this time many peptides go undetected. However, for the bovine MFGM proteome, one of the greatest problems is the abundance butyrophilin. Butyrophilin constitutes 30–40% of the total protein in the Holstein MFGM (Mather, 2000). Due to this dominant protein there is constant noise from butyrophilin peptides that results in the mass spectrometer missing many co-eluting low abundance proteins. Several other abundant MFGM proteins such as adipophilin, xanthine oxidase etc also contribute to this problem in the analysis of the MFGM proteome. The consequence of these or any overly abundant proteins in a proteome is simply a reduction in the total number of proteins that can be identified with high confidence. The greatest roadblock to high throughput proteomic research with material from dairy cows is the incomplete nature of the genomic data. Proteomic analysis requires matching peptide spectra to potential peptides from known proteins. Where bovine sequence data is lacking, proteins identified were homologous to either human, mouse or rat. The peptide identification software allows for limited divergence of the peptide sequence from that reported in the protein database. Peptide spectra that represent peptides in polymorphic regions of unsequenced genome are not identified. The full power of proteomic research in dairy cows will be realized only upon completion of the bovine genome. There are ambiguities concerning the functions of many of the proteins in protein databases such as SwissProt. In addition, there are ambiguities specific to the functions of proteins present in the MFGM. For example, xanthine oxidase’s role in purine metabolism is likely not its function in the MFGM. Previous work strongly suggests it playing a role in milk fat secretion (Mather & Keenan, 1998a; McManaman et al. 2002; Vorbach et al. 2002). The transcription factor, nuclear coactivator protein p100,

410


Table 1a–h. Proteins Identified in Bovine Milk Fat Globule Membranes. * indicates the protein was previously identified in mouse or human MFGM proteomes MW

Peptides Matched

Protein name

SwissProt ID

a: Identified Proteins with Membrane/Protein Trafficking Function ADP ribosylation factor Calcium binding protein P22 Endobrevin MAL2 protein Rab 10 Rab 11A Rab 11B Rab 13 Rab 18 Rab 1a Rab 22A Rab 27b Rab 2A Rab 35 Rab 3A Rab 3B Rab 3C Rab 5B Rab 5C Rab 7 SAR1B protein SCAMP2 SNARE protein Ykt6 SNAP23 Syntaxin 3 Translocon associated protein alpha VAMP 2

P32889 Q99653 O60625 Q969L2 O88386 P24410 Q86YS6 P51153 Q9NP72 P11476 Q9UL26 Q8HZJ5 P08886 Q15286 P11023 P10948 P10949 Q8IXL2 P51148 P51149 Q9QVY2 O15127 O15498 O00161 Q13277 P43307 Q15836

20565 22325 11395 19125 22541 24393 23339 22774 22977 22678 21855 24612 23545 23025 24723 24766 25826 29007 23482 23490 22429 36649 22417 23354 33141 32235 11178

15 7 3 3 6 4 2 6 4 5 3 3 4 5 6 5 4 4 4 2 8 4 5 7 3 3 4

b: Identified Proteins with Cell Signalling Function 14-3-3 protein gamma 14-3-3 protein beta alpha A kinase anchor protein 4 Cell death activator CIDE-A GTP binding protein alpha 13 GTP binding protein alpha 14 GTP binding protein beta subunit-like protein 12.3 GTP binding protein G(i),alpha-2 subunit GTP binding protein GL2 alpha chain GTP binding protein Sara GTP binding proteinG I G S G T beta subunit 1 GTP binding regulatory protein G s alpha GTP binding regulatory protein beta 2 chain Ral B Rap 1A Rap 1b Rheb 1 Rheb 2 Rho A Rho C Rho GDP-dissociation inhibitor 1 Rit 1 Seven transmembrane helix receptor STAT5 MGF

P29359 P29358 Q9XS94 O60543 Q14344 P38408 P25388 P04899 P38409 Q8NG23 P04901 P04896 P11017 P11234 P10113 P09526 Q99444 Q15382 P06749 P08134 P19803 Q92963 Q8NHC8 Q9TUM3

28121 27950 93988 24686 44049 41498 35076 40319 42070 22337 37377 45708 35645 23408 20987 20825 20497 20497 21768 22006 23421 25145 34844 89991

5 7 4 5 4 3 3 6 4 5 4 7 9 5 7 5 6 4 4 3 14 3 3 7

c: Identified Proteins with Unknown Function 100 kDa coactivator AD158

Q863B3 Q8TDW0

101988 92391

4 8

Milk fat globule membrane proteome

411

Table 1a–h. (Cont.) MW

Peptides Matched

Protein name

SwissProt ID

BA304I5 2 CGI 49 protein DREV1 protein EH-domain containing protein 4 EF G protein XLalphas Galectin 7 Glial fibrillary acidic protein HCG-1 protein HSPC041 protein Hypothetical ADP Ribosylation factors family Hypothetical protein Hypothetical protein DKFZp451B1115 Lactophorin Leucine-rich repeat containing protein 8 MUC1 protein MUC15 protein Presenilin 1 Responsive to centrifugal force and shear Similar to SSA protein SS 56 Similar to transmembrane 4 superfamily member 1 Synaptic vesicle membrane protein VAT-1 homolog Uromodulin Hypothetical 23.0 kDa Protein

Q8WU06 Q9Y363 Q9H1A2 Q9EQP2 O95417 P47929 Q96P18 O95164 Q96EQ4 Q8BGR6 Q8IZ81 Q86T70 P80195 Q8IWT6 Q8WML4 Q8MI01 Q9XT97 AAS20596 Q8K243 Q96CE8 Q99536 P48733 Q8BGF6

21054 46921 36437 61480 28343 14944 49808 13157 15824 22905 34960 93736 17151 94198 58091 35715 53653 34623 56098 22277 32558 69898 34747

2 6 6 4 5 5 5 2 2 4 3 4 6 8 10 5 5 3 4 3 3 3 4

d: Identified Proteins with Fat Transport/metabolism Function E-NPP 3 5 nucleotidase 6 phosphogluconolactonase Acetyl-CoA carboxylase 1 Adipophilin Apolipoprotein A I Apolipoprotein E Butyrophilin CD36 Cytochrome b5 reductase Fatty acid synthase Fatty acid-binding protein PAS6/7 Long-chain-fatty-acid CoA ligase 1 Retinal short-chain Dehydrogenase/reductase Serum amyloid A protein Transport secretion protein 2 2 VAT-1 homolog Xanthine dehydrogenase/oxidase

O14638 Q05927 O95336 Q9TTS3 Q9TUM6 P15497 Q03247 P18892 P26201 P07514 P49327 Q09139 Q95114 P33121 O77769 Q8SQ28 Q8BJ56 Q99536 P80457

100096 63085 27546 265301 49368 30276 35980 59276 52737 33990 273100 14590 47411 77943 33496 14723 53656 32558 146681

6 2 2 5 31 8 8 14 12 9 6 6 28 11 7 3 4 3 29

e : Identified Proteins with General Transport Functions Breast cancer resistance protein Excitatory amino acid transporter 3 Hypothetical protein FLJ37958 Plasma membrane calcium transporting ATPase 2 Sodium and chloride dependent creatine transporter 1 Sodium dependent glucose transporter SGLT I Sodium dependent phosphate transporter Stomatin protein 1 9330174J19Rik protein Probable cation transporting ATPase 3

Q9UNQ0 Q95135 Q8N1Q9 Q01814 O18875 Q8MKB7 Q27960 P27105 Q8BUP1 Q9H7F0

72343 57297 93955 136876 70675 73076 75826 31599 136730 127138

12 3 4 3 7 4 12 5 4 3

f : Protein Synthesis, Binding and/or Folding Elongation factor 1 alpha 90 kDa heat shock protein beta

O18787 BAC82487

34379 84542

4 7


412

Table 1a–h. (Cont.) Protein name

SwissProt ID

MW

Peptides Matched

Calnexin Cofilin non muscle isoform Cyclophilin I Profilin I Rho F Tight junction protein 1

P27824 P10668 Q864S5 P02584 Q9HBH0 Q95M47

67568 18518 17738 14926 23625 194721

3 2 12 3 7 4

g: Identified Proteins with Immune Functions Immunoglobulin M heavy chain Toll like receptor 4 Toll-like receptor 2 CD14 CD9 antigen

P01871 Q8SQ55 Q95LA9 Q95122 P30932

49556 96054 90204 39666 25126

6 3 12 6 6

h: Identified Milk Proteins Beta casein Kappa casein

P02666 P02668

25107 21269

4 4

Secreted 5%

Cytoplasmic 24%

A

Membrane 71%

Milk proteins 2% Immune proteins 4%

Protein synthesis/binding/folding 7%

B

Membrane/protein trafficking 23%

Transporters 9% Fat transport/metabolism 11%

Unknown 21%

Cell signaling 23%

Fig. 2. The functional category assignments for the 120 proteins identified in this study. Assignments were made based primarily on information in the SwissProt database. Panel A shows the general cell location of the proteins. Panel B shows the distribution of proteins identified by general function.

presence in endoplasmic reticulum, lipid droplets and cytosol of lactating mammary gland led the authors of these finding to suggest that it may have additional functions beyond a transcription factor (Keenan et al. 2000). The

finding of nuclear coactivator protein p100 in the MFGM further supports their speculation. These ambiguities cannot be ignored and must be considered in any discussion of function in the absence of true functional data. But, for

Milk fat globule membrane proteome

413

1589.83(M+H) +

100 R

A

L

T

Q

F

A

Q

E

V

N

L

A

E

% 795.37

201.08 246.15 y2 b2 175.11 y1 86.09 L

314.17 460.28 y4 b3

527.28 b5

1063.55 y9

735.41 806.44 y7 934.49 y6 y8 588.35 y5 723.36

837.42

1017.45

1276.66 1389.73 1590.95 y11 y12 1571.82 1162.60 1600.79 y10 1277.55 1445.19 M/z

0

100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600

Fig. 3. A tandem MS/MS spectrum of the peptide ion 795.37 (M+2H) + selected from the survey scan for MS/MS analysis. This peptide ion was identified along with one other peptide ion as Rab 7 using the ProteinLynx Global Server 2.0 software. However, with only two peptide ions identified the probability for a protein match was just below the level of significance needed. Therefore this peptide’s MS/MS spectrum was sequenced using the de novo sequencing program PepSeq. The peptide sequence was determined to be EAINVEQAFQTIAR with a 99 % probability using this program thus confirming the presence of RAB7 in this proteome. The sequence is read right to left using the y ions in the spectrum.

PMCA2

SNAP23

Rab3A

Fig. 4. Western blots confirming the presence of 3 proteins identified in the milk fat globule membrane proteome. Plasma membrane Ca2+ATPase isoform 2 (PMCA2), SNAP23 and Rab3A were identified in the milk fat globule proteome by 2, 6 and 7 peptides, respectively, and confirmed by Western blotting.

the purpose of discussion, proteins were grouped into functional groups. Proteins from the bovine MFGM proteome were grouped by both cell location and general function (Fig. 2A&B and Table 1, a–h) based primarily on functions listed in the SwissProt database. Because of our purification process, the bulk of the proteins found were membraneassociated proteins (71 %). The 24 % cytoplasmic proteins found in the MFGM proteome are likely a combination of cytoplasmic proteins tightly associated with the membrane proteins or due to some cytoplasmic contamination of the MFGM preparations resulting from cytoplasmic crescents in the milk fat droplets (Wu et al. 2000a). Despite the potential for contamination of the MFGM by casein and other milk proteins, we were able to purify our membranes to a point that skim milk proteins were a very minor component of this proteome (Table 1h).

Mather and Keenan (1998b), point out in their review that the general progress in understanding the nature and regulation of milk secretion at the cellular and molecular level has lagged behind the progress in other secretory systems. Furthermore, much of what is proposed is by analogy with other quite different secretory model systems. Twenty three % of the proteins identified from purified MFGM have functions associated with membrane/protein trafficking (Table 1a). The membrane/protein trafficking proteins identified included 16 Rab proteins, SNAP23, syntaxin 3, SNARE protein Ykt6 and VAMP2 (Table 1a). The Rab proteins are low-molecular-weight GTP-binding proteins that form the largest branch of the Ras superfamily of GTPases. Rab proteins and their effectors coordinate stages of transport in the secretory pathway (Zerial & McBride, 2001). The Rab3 isoforms and Rab27A are known to function in regulated exocytosis (Burgoyne & Morgan, 2003). The Rab5 isoforms regulate vesiclemediated transport from the plasma membrane to the early endosomes. Rab11 regulates recycling endosomes between the plasma membrane and Golgi (Zerial & McBride, 2001). Three membrane proteins, SNAP-25, synaptobrevin, and syntaxin, form the core of a ubiquitous membrane fusion machine that interacts with the soluble proteins N-ethylmaleimide-sensitive factor (NSF) and a-SNAP. Rab proteins, in coordination with the core fusion machinery and Munc-18, help to mediate vesicle docking and fusion. SNAPs are cytosolic proteins that play a key role in the process of membrane fusion in intracellular vesicle trafficking. In eukaryotic cells, the SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein

414


receptor) complex is critical to membrane docking and fusion and is believed to impart some degree of specificity between vesicle SNARE and target organelle SNARE. In neurons and neuroendocrine cells, the SNARE complex consists of the integral membrane proteins VAMP (vesicleassociated membrane protein), syntaxin and SNAP25. In non-neuronal tissues, such as mammary tissue, SNAP23 (a SNAP25 homolog) functionally replaces SNAP25 in the SNARE complex. Studies show that VAMP, syntaxin and SNAP23 are required for SNARE function. This complex exists as a heterotrimer of the three proteins (Zerial & McBride, 2001; Bonifacino & Glick, 2004). The high expression of SNAP23 and Rab3A in MFGM was confirmed by Western blotting (Fig. 4). Thus, the firm identification SNAP23, VAMP 2, syntaxin 3 along with several of the Rabs in purified MFGM provides a foundation of proteins likely key for the study and understanding of secretory processes in the cow. The SNAREs in MFGM could have originated from the fusion of secretory vesicles with the apical plasma membrane with subsequent carry over of the SNARE complexes into the lipid droplets. The proteins in Table 1b represent an additional 23 % of the proteome in this study. Their functions are primarily associated with cell signaling. In this group, 2 isoforms of the 14-3-3 proteins were identified. These 14-3-3 proteins have been shown to be associated with the stimulation of calcium dependant exocytosis (Roth et al. 1993, 1994). The secretion of milk proteins occurs by both constitutive and calcium regulated exocytosis (Turner et al. 1992a, b). The presence of the 14-3-3 proteins suggests they may participate in the regulation of the calcium dependant exocytosis of milk proteins. The identification of a large group of proteins linked to cell signalling, in a membrane preparation, is not surprising. However, linking them to any specific function in milk secretion is beyond the scope of this study. The next largest group of proteins identified is those with unknown or ‘ poorly’ defined functions (Table 1c). The presence of the transcription factor nuclear coactivator protein p100 in MFGM was unexpected. Its presence in the MFGM proteome raises questions regarding its function which are not easily addressed. The finding of MUC1 in MFGM is not surprising, as it has been shown to be a major MFGM protein in prior studies and has been used as marker for some elegant studies on membrane movement in lactating mammary tissue (Mather et al. 2001). However, its function in the MFGM remains elusive as MUC1 knockout mice lactate normally (Spicer et al. 1995) but MUC1 may have protective functions in the neonatal intestine (Peterson et al. 1998). Most of the other proteins in Table 1c are new identifications for the MFGM proteome. They are not extensively characterized as to their functions in general and not at all for the MFGM. The next grouping was for proteins believed to be involved in fat transport or general metabolism (Table 1d). This group constituted 11 % of this proteome. This group contains most of the major proteins of the MFGM

(butyrophilin, xanthine oxidase, PAS 6/7, adipophilin, and CD36) (Mather & Keenan, 1998a). Specific interactions have been proposed between the butyrophilin, xanthine oxidase, and adipophilin (Mather & Keenan, 1998a) in the MFGM and supporting data for this interaction has been presented (McManaman et al. 2002). The mechanisms controlling milk fat secretion are not well understood, but these proteins are thought to play a critical role (Mather & Keenan, 1998a). Using xanthine oxidase knockout mice, it was shown that females with 1 allele knocked out could not maintain lactation (Vorbach et al. 2002). These authors went on to show that xanthine oxidase may have a primary role in milk fat droplet secretion. The recent examination of mice with the butyrophilin gene knocked out demonstrates the importance of butyrophilin expression on milk fat secretion (Ogg et al. 2004). Several proteins were identified (9 % of the proteome) whose functions are transport related (Table 1e). The major calcium pump (plasma membrane calcium transporting ATPase 2) found in mammary tissue and the MFGM (Reinhardt et al. 2000) was identified in this group. This calcium pump has recently been shown to be required for secretion of much of the calcium in milk (Reinhardt et al. 2004). No phosphate transport is thought to occur across the apical membrane of the mammary secretory cell (Shennan, 1998; Shennan & Peaker, 2000). It is therefore somewhat surprising to find a sodium dependent phosphate transporter on the MFGM. Previous work has shown the presence of a sodium dependent glucose transporter (SGLT I) in mammary tissue but its location was not determined (Shennan & Peaker, 2000). The data presented here suggest an apical location for this transporter. One of the more interesting findings in this study was the identification of the Toll like receptors (TLR) TLR2 and TLR4 along with CD14 in the MFGM (Table 1g). CD14 has been shown previously on general secretory vesicles (Landmann et al. 2000) and has most recently been shown to be expressed in the mammary gland and to play a role in mammary involution (Stein et al. 2004). Our data show that mammary expression of CD14 is on the apical membrane of the secretory cell and the MFGM. The TLRs are predominantly expressed on antigen presenting cells and signal the presence of pathogens to the host. A recent study showed that mastitis increases mammary tissue mRNA expression of TLR2&4 (Goldammer et al. 2004). These authors were unable to determine the cell types expressing the TLRs in the mammary gland. They speculated that the bulk of TLR mRNA measured likely arose from infiltrating neutrophils and that some may be on the secretory epithelial cell. The data demonstrates that TLR2&4 are on the apical membrane of the secretory cell. The finding of CD14, TLR2, and TLR4 on MFGM suggests a direct role for the mammary gland in detecting an infection. The MFGM also may act as a sink for lipopolysaccaride (LPS) in an infection. The possibility that the CD14, TLR2, and TLR4 found in this study are the result of

Milk fat globule membrane proteome neutrophil membrane contamination of our MFGM fraction is unlikely. This is because none of the major neutrophil membrane proteins (Lippolis, Reinhardt unpublished observations) were found in our MFGM preparation. The significant of finding CD14, TLR2, and TLR4 on MFGM in mastitis will await direct studies. In conclusion, a proteomic analysis of the bovine MFGM resulted in 120 proteins being identified. Of these, the majority were membrane associated proteins, mainly associated with membrane/protein trafficking or cell signalling, with the remainder being either cytoplasmic or secreted proteins. These data may provide insights into MFGM secretion mechanisms. The finding of CD14, TLR2, and TLR4 on MFGM suggests that pathogens entering the mammary gland are signaled to the host directly through the mammary gland. This proteomic study provides a foundation for research into secretory mechanisms of the mammary gland. It should be pointed out that many of the identified proteins are likely minor contaminants that have been carried over into the MFGM during its birth from the cell and consequently do not have any functional significance in the formation and secretion of the milk-fat globule. Future proteomic studies will address the differential expression of MFGM proteins, as a function of stage of lactation, production level and mastitis. We thank Tera Nyholm, Katie Bradshaw, Derrel Hoy and Duane Zimmerman for their technical assistance and animal care.

References Bonifacino JS & Glick BS 2004 The mechanisms of vesicle budding and fusion. Cell 116 (2) 153–166 Burgoyne RD & Morgan A 2003 Secretory granule exocytosis. Physiological Reviews 83 (2) 581–632 Cavaletto M, Giuffrida MG, Giunta C, Vellano C, Fabris C, Bertino E, Godovac-Zimmermann J & Conti A 1999 Multiple forms of lactadherin (breast antigen BA46) and butyrophilin are secreted into human milk as major components of milk fat globule membrane. Journal of Dairy Research 66 (2) 295–301 Charlwood J, Hanrahan S, Tyldesley R, Langridge J, Dwek M & Camilleri P 2002 Use of proteomic methodology for the characterization of human milk fat globular membrane proteins. Analytical Biochemistry 301 (2) 314–324 Fortunato D, Giuffrida MG, Cavaletto M, Garoffo LP, Dellavalle G, Napolitano L, Giunta C, Fabris C, Bertino E, Coscia A & Conti A 2003 Structural proteome of human colostral fat globule membrane proteins. Proteomics 3 (6) 897–905 Fujiki Y, Hubbard AL, Fowler S & Lazarow PB 1982 Isolation of intracellular membranes by means of sodium carbonate treatment : application to endoplasmic reticulum. Journal of Cell Biology 93 (7) 97–102 Goldammer T, Zerbe H, Molenaar A, Schuberth HJ, Brunner RM, Kata SR & Seyfert HM 2004 Mastitis increases mammary mRNA abundance of beta-defensin 5, toll-like-receptor 2 (TLR2), and TLR4 but not TLR9 in cattle. Clinical and Diagnostic Laboratory Immunology 11 (7) 174–185 Gu S, Chen J, Dobos KM, Bradbury EM, Belisle JT & Chen X 2003 Comprehensive Proteomic Profiling of the Membrane Constituents of a Mycobacterium tuberculosis Strain. Molecular and Cellular Proteomics 2 (12) 1284–1296

415

Jacobs JM, Mottaz HM, Yu LR, Anderson DJ, Moore RJ, Chen WN, Auberry KJ, Strittmatter EF, Monroe ME, Thrall BD, Camp DG 2nd & Smith RD 2004 Multidimensional proteome analysis of human mammary epithelial cells. Journal of Proteome Research 3 (7) 68–75 Keenan TW 2001 Milk lipid globules and their surrounding membrane: a brief history and perspectives for future research. Journal of Mammary Gland Biology and Neoplasia 6 (3) 365–371 Keenan TW, Winter S, Rackwitz HR & Heid HW 2000 Nuclear coactivator protein p100 is present in endoplasmic reticulum and lipid droplets of milk secreting cells. Biochimica et Biophysica Acta 1523 (7) 84–90 Landmann R, Muller B & Zimmerli W 2000 CD14, new aspects of ligand and signal diversity. Microbes and Infection 2 (3) 295–304 Lippolis JD & Reinhardt TA 2005 Proteomic survey of bovine neutrophils. Veterinary Immunology and Immunopathology 103 (1–2) 53–65 Mather IH 2000 A review and proposed nomenclature for major proteins of the milk-fat globule membrane. Journal of Dairy Science 83 (2) 203–247 Mather IH, Jack LJ, Madara PJ & Johnson VG 2001 The distribution of MUC1, an apical membrane glycoprotein, in mammary epithelial cells at the resolution of the electron microscope : implications for the mechanism of milk secretion. Cell and Tissue Research 304 (7) 91–101 Mather IH & Keenan TW 1998a Origin and secretion of milk lipids. Journal of Mammary Gland Biology and Neoplasia 3 (3) 259–273 Mather IH & Keenan TW 1998b The cell biology of milk secretion : historical notes. Introduction. Journal of Mammary Gland Biology and Neoplasia 3 (3) 227–232 McManaman JL, Palmer CA, Wright RM & Neville MC 2002 Functional regulation of xanthine oxidoreductase expression and localization in the mouse mammary gland : evidence of a role in lipid secretion. Journal of Physiology 545 (Pt 2) 567–579 Moritz RL, Ji H, Schutz F, Connolly LM, Kapp EA, Speed TP & Simpson RJ 2004 A Proteome Strategy for Fractionating Proteins and Peptides Using Continuous Free-Flow Electrophoresis Coupled Off-Line to Reversed-Phase High-Performance Liquid Chromatography. Analytical Chemistry 76 (16) 4811–4824 Ogg SL, Weldon AK, Dobbie L, Smith AJ & Mather IH 2004 Expression of butyrophilin (Btn1a1) in lactating mammary gland is essential for the regulated secretion of milk-lipid droplets. Proceedings of the National Academy of Sciences of the United States of America 101 (27) 10084–10089 Park ZY & Russell DH 2000 Thermal denaturation : a useful technique in peptide mass mapping. Analytical Chemistry 72 (11) 2667–2670 Park ZY & Russell DH 2001 Identification of individual proteins in complex protein mixtures by high-resolution, high-mass-accuracy MALDI TOF-mass spectrometry analysis of in-solution thermal denaturation/ enzymatic digestion. Analytical Chemistry 73 (11) 2558–2564 Peirce MJ, Wait R, Begum S, Saklatvala J & Cope AP 2004 Expression profiling of lymphocyte plasma membrane proteins. Molecular and Cellular Proteomics 3 (7) 56–65 Peterson JA, Patton S & Hamosh M 1998 Glycoproteins of the human milk fat globule in the protection of the breast-fed infant against infections. Biology of the Neonate 74 (2) 143–162 Pucci-Minafra I, Fontana S, Cancemi P, Alaimo G & Minafra S 2002 Proteomic patterns of cultured breast cancer cells and epithelial mammary cells. Annals of the New York Academy of Sciences 963, 122–139 Quaranta S, Giuffrida MG, Cavaletto M, Giunta C, GodovacZimmermann J, Canas B, Fabris C, Bertino E, Mombro M & Conti A 2001 Human proteome enhancement : high-recovery method and improved two-dimensional map of colostral fat globule membrane proteins. Electrophoresis 22 (9) 1810–1818 Reinhardt TA, Filoteo AG, Penniston JT & Horst RL 2000 Ca(2+)-ATPase protein expression in mammary tissue. American Journal of Physiology Cell Physiology 279 (5) C1595–1602 Reinhardt TA, Lippolis JD, Shull GE & Horst RL 2004 Null mutation in the gene encoding plasma membrane Ca2+-ATPase isoform 2 impairs

416


calcium transport into milk. Journal of Biololgical Chemistry 279 (41) 42369–42373 Roth D, Morgan A & Burgoyne RD 1993 Identification of a key domain in annexin and 14-3-3 proteins that stimulate calcium-dependent exocytosis in permeabilized adrenal chromaffin cells. FEBS Letters 320 (3) 207–210 Roth D, Morgan A, Martin H, Jones D, Martens GJ, Aitken A & Burgoyne RD 1994 Characterization of 14-3-3 proteins in adrenal chromaffin cells and demonstration of isoform-specific phospholipid binding. Biochemical Journal 301 (Pt 1) 305–310 Russell WK, Park ZY & Russell DH 2001 Proteolysis in mixed organicaqueous solvent systems : applications for peptide mass mapping using mass spectrometry. Analytical Chemistry 73 (11) 2682–2685 Shennan DB 1998 Mammary gland membrane transport systems. Journal of Mammary Gland Biology and Neoplasia 3 (3) 247–258 Shennan DB & Peaker M 2000 Transport of milk constituents by the mammary gland. Physiological Reviews 80 (3) 925–951 Spicer AP, Rowse GJ, Lidner TK & Gendler SJ 1995 Delayed mammary tumor progression in Muc-1 null mice. Journal of Biological Chemistry 270 (50) 30093–30101 Stein T, Morris JS, Davies CR, Weber-Hall SJ, Duffy MA, Heath VJ, Bell AK, Ferrier RK, Sandilands GP & Gusterson BA 2004 Involution of the mouse mammary gland is associated with an immune cascade and an acute-phase response, involving LBP, CD14 and STAT3. Breast Cancer Research 6 (2) R75–R91 Turner MD, Rennison ME, Handel SE, Wilde CJ & Burgoyne RD 1992a Proteins are secreted by both constitutive and regulated secretory

pathways in lactating mouse mammary epithelial cells. Journal of Cell Biology 117 (2) 269–278 Turner MD, Wilde CJ & Burgoyne RD 1992b Exocytosis from permeabilized lactating mouse mammary epithelial cells. Stimulation by Ca2+ and phorbol ester, but inhibition of regulated exocytosis by guanosine 5k-[gamma-thio]triphosphate. Biochemical Journal 286 (Pt 1) 13–15 Vorbach C, Scriven A & Capecchi MR 2002 The housekeeping gene xanthine oxidoreductase is necessary for milk fat droplet enveloping and secretion : gene sharing in the lactating mammary gland. Genes and Development 16 (24) 3223–3235 Waters 2004 ProteinLynx Global Server User Guide. Milford, MA : Waters Corporation Wooding FB 1971 The mechanism of secretion of the milk fat globule. Journal of Cell Science 9 (3) 805–821 Wooding FB 1973 Formation of the milk fat globule membrane without participation of the plasmalemma. Journal of Cell Science 13 (7) 221–235 Wu CC, Howell KE, Neville MC, Yates JR 3rd & McManaman JL 2000a Proteomics reveal a link between the endoplasmic reticulum and lipid secretory mechanisms in mammary epithelial cells. Electrophoresis 21 (16) 3470–3482 Wu CC, Yates JR 3rd, Neville MC & Howell KE 2000b Proteomic analysis of two functional states of the Golgi complex in mammary epithelial cells. Traffic 1 (10) 769–782 Zerial M & McBride H 2001 Rab proteins as membrane organizers. Nature Reviews Molecular Cell Biology 2 (2) 107–117