Proteomic Analysis of Cow, Yak, Buffalo, Goat and Camel Milk Whey

Mar 7, 2013 - This analysis is the first to produce proteomic data for the milk from the above-mentioned animal species: 211 proteins have been identi...
12 downloads 14 Views 388KB Size
Download PDF
Article pubs.acs.org/jpr

Proteomic Analysis of Cow, Yak, Buffalo, Goat and Camel Milk Whey Proteins: Quantitative Differential Expression Patterns Yongxin Yang,†,‡,§ Dengpan Bu,†,§ Xiaowei Zhao,†,‡ Peng Sun,† Jiaqi Wang,*,† and Lingyun Zhou† †

State Key Laboratory of Animal Nutrition, Institute of Animal Science, Chinese Academy of Agricultural Sciences, Beijing 100193, China ‡ Institute of Animal Science and Veterinary Medicine, Anhui Academy of Agricultural Sciences, Hefei 230031, China S Supporting Information *

ABSTRACT: To aid in unraveling diverse genetic and biological unknowns, a proteomic approach was used to analyze the whey proteome in cow, yak, buffalo, goat, and camel milk based on the isobaric tag for relative and absolute quantification (iTRAQ) techniques. This analysis is the first to produce proteomic data for the milk from the above-mentioned animal species: 211 proteins have been identified and 113 proteins have been categorized according to molecular function, cellular components, and biological processes based on gene ontology annotation. The results of principal component analysis showed significant differences in proteomic patterns among goat, camel, cow, buffalo, and yak milk. Furthermore, 177 differentially expressed proteins were submitted to advanced hierarchical clustering. The resulting clustering pattern included three major sample clusters: (1) cow, buffalo, and yak milk; (2) goat, cow, buffalo, and yak milk; and (3) camel milk. Certain proteins were chosen as characterization traits for a given species: whey acidic protein and quinone oxidoreductase for camel milk, biglycan for goat milk, uncharacterized protein (Accession Number: F1MK50) for yak milk, clusterin for buffalo milk, and primary amine oxidase for cow milk. These results help reveal the quantitative milk whey proteome pattern for analyzed species. This provides information for evaluating adulteration of specific specie milk and may provide potential directions for application of specific milk protein production based on physiological differences among animal species. KEYWORDS: iTRAQ, milk protein, animal proteomics, species, principal component analysis, clustering

1. INTRODUCTION Milk is a complex biological fluid produced by the mammary glands of mammals and is the essential source of nutrients that provides the primary nutrition for a newborn mammal. Cow milk is consumed by humans worldwide. Cow, goat, and sheep milk accounts for approximately 87% of global milk production. However, some minor dairy animal species such as buffalo, yak, horse, donkey, and camel are nutritionally and economically important in specific regions.1 The dissemination of information on milk nutritional properties from minor dairy animal species will help provide for conservation and sustainable use of milk from underutilized dairy breeds and species.1,2 Milk composition depends on several factors such as breed, season, animal health status, lactation phase, and dietary nutrition.3−5 Milk proteins are a major source of essential nutrients. To study the milk proteome, whole milk was separated to cream and skim milk. Skim milk was then centrifuged to obtain a casein pellet and a supernatant. Milk of most species contains three or four caseins and the supernatant contains a large number of low-abundance proteins. More importantly, proteomics approaches have the potential to facilitate simultaneous analysis of complex milk proteins.6 Proteomic © 2013 American Chemical Society

characterization of human milk proteins has been examined in numerous studies.7,8 Low-abundance proteins in the whey fraction of human milk were enriched by ProteoMiner beads and identified by liquid chromatography−tandem mass spectrometry (LC−MS/MS) during a 12-month lactation period.7 Comprehensive analysis of the human milk proteome was performed through anion exchange fractionation and protein chip array mass spectrometry.8 More recently, proteomic approaches based on mass spectrometry have been used for farm animal milk proteome research.2,9 Several researchers have aimed to study bovine milk components using two-dimensional electrophoresis (2DE),10,11 immuno-absorption techniques,12 and mass spectrometry (MS).5,12,13 Further, post-translational modifications including phosphorylation, glycosylation, and disulfide linkages were pursued and confirmed by employing tandem mass spectrometry (MS/MS) approaches.14−16 Using ion exchange chromatography and two-dimensional gel-based proteomic methods, protein isoform patterns in isolated bovine fractions Received: October 24, 2012 Published: March 7, 2013 1660

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667

Journal of Proteome Research

Article

filter (10 kDa cutoff, Sartorius, German), 200 μL UT buffer (8 M Urea and 150 mM Tris-HCl pH 8.0) were added to the filter and centrifuged at 14000× g for 30 min, and washed again with UT buffer. Subsequently, 100 μL of iodoacetamide solution (50 mM iodoacetamide in UT buffer) were added to the filter. The filter unit was mixed for 1 min, followed by incubation for 30 min at room temperature in the dark and centrifuged at 14000× g for 20 min. Two wash steps with 100 μL UT buffer were performed with centrifugation at 14000× g for 20 min after each wash step. Then, 100 μL dissolution buffer (Applied Biosystems, Foster City, CA) was added to the filter and centrifuged at 14000× g for 30 min, and this step was repeated twice. Finally, 40 μL of trypsin (Promega, Madison, WI) buffer (2 μg trypsin in 40 μL dissolution buffer) were added and digested at 37 °C for 16−18 h. The filter unit was transferred to a new tube and centrifuged at 14000× g for 20 min. Resulting peptides were collected as a filtrate and the peptide concentration was analyzed at OD280.30 Subsequently, a 100-μg peptide mixture was labeled with iTRAQ reagents according to the manufacturer’s instructions (Applied Biosystems, Foster City, CA). The 3.46%-protein (w/ v) cow milk from Beijing was labeled with reagent 114, 2.75%protein (w/v) cow milk from Shanxi was labeled with reagent 115, buffalo milk samples were labeled with reagent 116, goat milk samples were labeled with reagent 117, yak milk samples were labeled with reagent 118, and camel milk samples were labeled with reagent 121. Identical quality peptide mixture of all 6 samples from above-mentioned animals were labeled with reagent 113 and served as sample REF. Then, three independent biological experiments were performed. The labeling solution reaction was incubated at room temperature for 1 h prior to further analysis.

were elucidated.17 Furthermore, ProteoMiner combinatorial peptide ligand libraries provided enhancement of low-abundant proteins in the cow milk whey proteome, thereby promoting identification of 149 unique protein species using MS.18 D’Alessandro et al. (2011) merged data from previous investigations of bovine milk to compile an exhaustive list of 573 nonredundant annotated protein entries. Most of the milk proteins were grouped under pathways, networks, or ontologies referring to nutrient transport, lipid metabolism, and objectification of the immune system response, respectively.19 In addition, studies investigating minor dairy animals including sheep, buffalo, horse and donkey milk have also employed proteomics approaches.20−22 Some researchers have even compared the proteome profiles of some animal species milk for identifying sources of hypoallergenic alternatives to bovine milk and detection of milk adulteration.23−25 Changes in protein expression result in variations in milk composition. In this study, differential protein regulation will be characterized and compared among animal sources using proteomic approaches. Despite previous efforts, a complete understanding of the composition and functions of milk proteins has been hampered by incomplete knowledge of the milk proteome. There is increasing demand for natural milk from nonbovine mammals, particularly donkey, yak, and camel.26 However, little information about these milk proteomes is available. The isobaric tag for relative and absolute quantification (iTRAQ) is a quantitative proteomic approach with relatively high throughput that allows for simultaneous identification and peptide quantification by measuring the peak intensities of reporter ions with MS/MS.27−29 Here, we present an exploration of cow (2.75% (w/v) and 3.46% (w/v)−protein content, respectively), goat, buffalo, yak, and camel milk whey proteomes using the iTRAQ approach. We also provide the quantified protein profile for each species. The principal component analysis (PCA) and hierarchical clustering analysis of proteomic data highlighted significant differences in the milk of various animal species.

2.3. Strong Cationic-exchange Chromatography Separation

The combined sample was acidified with 1% trifluoroacetic acid and 400 μg subjected to strong cationic-exchange chromatography (SCX) fractionation using a Polysulfethyl (PolyLC Inc., Columbia, MD) column (4.6 × 100 mm, 5 μm, 200 Å). Solvent A was 10 mM KH2PO4 pH3.0 and 25% (v/v) acetonitrile; solvent B was solvent A with 500 mM KCl. The solvents were applied using a 20 min gradient from 10−45% (v/v) solvent B, followed by 20 min at 45% (v/v). These samples were combined into 10 fractions and then desalted on C18 cartridges (66872-U, Sigma).

2. MATERIALS AND METHODS 2.1. Sample Preparation

In the current study, 30 Chinese Holstein cow (Bos taurus) milk samples were collected from local dairy farms in Beijing and Shanxi, respectively, 27 goat (Capra hircus) milk samples were collected from a farm in Shanxi province, 21 Bactrian camel (Camelus bactrianus) milk samples were collected from a farm in Urumchi city in Xinjiang, 30 yak (Bos grunniens) milk samples were collected from a farm in Qinghai province, and 21 buffalo (Bubalus bubalis) milk samples were collected from a farm in Yunnan province. Raw milk samples from each farm were pooled into three fractions. Whole-milk samples were centrifuged at 3000× g for 15 min after collection at 4 °C to remove the fat. Skim-milk samples were centrifuged at 100 000× g for 1 h to obtain the supernatant containing whey proteins. Thereafter, the supernatant was collected and protein concentration was determined using the modified Bradford assay protocol (Bio-Rad, Berkeley, CA).

2.4. Reversed Phase Chromatography Separation

SCX fractions were thawed and dissolved in Buffer C (0.1% (v/ v) formic acid in Milli-Q water). The column was equilibrated for 20 min with 95% (v/v) solvent C. Peptide mixtures were then separated using reverse-phase chromatography and the Thermo EASY-nLC column with Zorbax 300SB-C18 peptide traps (Agilent Technologies, Santa Clara, CA) (100 mm × 75 μm, 5 μm) at 200 nL/min. Peptides were separated with solvent D (acetonitrile with 0.1% (v/v) formic acid) using a segmented gradient from 4−28% (v/v) in 110 min, from 28− 40% (v/v) in 20 min, from 40−90% (v/v) in 5 min and then at 90% (v/v) for 5 min.

2.2. Digestion and iTRAQ Labeling

2.5. MS/MS Analysis and Quantification

Total-protein samples (200 μg) diluted in 4% SDS, 100 mM Tris-HCl pH 8.0, and 100 mM dithiotreitol solution were heated at 95 °C for 5 min. After each sample was cooled to room temperature, the sample was loaded on an ultrafiltration

The Q-Exactive (Thermo Finnigan, San Jose, CA) mass spectrometer was set to perform data acquisition in the positive ion mode, with a selected mass range of 300−1800 mass/ charge (m/z). Resolving power for the Q-Exactive was set as 1661

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667

Journal of Proteome Research

Article

in the original data set. Ultimate objective of PCA is to find the smallest group of PCs that can explain the maximum variance. Cluster 3.0 software was used to investigate the hierarchical clustering of identified proteins. Java Treeview was used for visualization.

70000 for the MS scan and 17500 for the MS/MS scan at m/z 200. MS/MS data were acquired using the top 10 most abundant precursor ions with charge ≥2 as determined from the MS scan. These were selected with an isolation window of 1.6 m/z and fragmented by higher energy collisional dissociation with normalized collision energies of 30 eV. The maximum ion injection times for the survey scan and the MS/ MS scans were 10 and 60 ms, respectively, and the automatic gain control target values for both scan modes were set to 1E6. Dynamic exclusion for selected precursor ions was set at 40 s. Underfill ratio was defined as 0.1% on the Q-Exactive. Raw files were processed using ProteomicsTools (Version 3.0.5) (http://www.proteomics.ac.cn/).31 Mascot generic format was used for data files, which were then reverse-imported to the database using ProteomicsTools and Mascot. Fragmentation spectra were searched using the MASCOT search engine (version 2.2; Matrix Science) against the Cetartiodactyla in the in-house uniprot database (02−2012, 79442 entries). The following search parameters were set: monoisotopic mass, MS/ MS tolerance at 0.1 Da and peptide mass tolerance at ±20 ppm, trypsin as the enzyme and allowing up to two missed cleavages, peptide charges of 2+, 3+, and 4+. Fixed modifications were defined as iTRAQ labeling and carbamidomethylation of cysteine; oxidation of methionine was specified as a variable modification. The decoy database pattern was set as reversed version of the target database. All reported data were based on 99% confidence for proteins and peptides identification as determined by false discovery rate (FDR) of no more than 1%, as well as 2*N(decoy)/((N(decoy) + N(target)) as the formula for computing FDR, in which the decoy is reversed database and the target is target database.31 Protein identification was supported by at least one unique peptide identification. The iTRAQ analysis of relative protein quantification levels across multiple samples follows. ProteomicsTools (Version: 3.0.5) was used to calculate relative ratios of identified peptides among labeled samples using relative peak intensities of released iTRAQ reporter ions in each of the MS/MS spectra, while relative protein quantification among samples was based on weighted ratios of uniquely identified peptides which belonged to the specific individual protein in which sample REF was used as reference. Final ratios of protein quantification were then normalized by the median average protein quantification ratio for unequally mixed different labeled samples. This correction is based on the assumption that the expression of most proteins does not change. Thus, if samples from each experimental condition are not combined in exactly equal amounts, this normalization fixes the systematic error. Only protein identification that was inferred from the unique peptide identification in all three independent experiments was considered. Statistical analysis was conducted using a one-way ANOVA. P-values of less than 0.05 were considered significant, and Tukey test was used to evaluate differences among group samples.

3. RESULTS 3.1. Identified Proteins Analysis

In the current study, 211 milk proteins were identified in cow, goat, buffalo, yak, and camel milk whey using the iTRAQlabeled proteomic approach (Table S1, Supporting Information). Of these, five proteins identified were inferred from the single, unique peptide identification assignment with annotated tandem mass spectra (Figure S1−S5, Supporting Information). The 108 kinds of proteins identified were uniquely inferred from peptide identification by consulting databases for cow and a small number of proteins by consulting databases for buffalo, yak, and camel. These results provide more information and enable us to understand milk whey protein composition in these minor mammalian species. In addition, proteins such as serotransferrin, albumin, apolipoprotein, and complement C3 that could have originated from blood were also identified in milk whey of the above-mentioned animals. All 211 identified proteins were classified using the gene ontology (GO) annotation (http://david.abcc.ncifcrf.gov/ home.jsp) and further categorized into three functional groups: those related to molecular function, cellular components, and biological processes. The GO annotations were available for 113 identified proteins including 86 identified proteins in biological processes, 98 proteins in cellular components, and 97 proteins in molecular function (Table S2, Supporting Information). Proteins assigned to each category are presented in Figure 1. The most common molecular function was binding activity, which included ion binding, protein binding, carbohydrate binding, pattern binding, cell surface binding, and lipid binding. Another major functional category was enzyme regulator activity. The most prevalent cellular components were located in the intracellular part and extracellular region. The most frequently encountered biological processes were biological regulation, response to stress, localization, and establishment of localization. 3.2. Statistical Analysis of Identified Proteins

Results of ANOVA analysis indicated existence of 177 differential proteins (P < 0.05) in animal milk whey proteome and whey proteins listed in Table S3, Supporting Information. These results demonstrated that primary amine oxidase and epididymal secretory protein E1 were highest in cow milk whey than in the milk whey of buffalo, goat, yak, and camel. Biglycan, C-X-C motif chemokine 6, serum amyloid A protein, polymeric immunoglobulin receptor (Fragment), fatty acid synthase, and disulfide-isomerase protein were highest in goat milk whey. Alpha-lactalbumin (AC: P00712), CD14 (Fragment), thrombospondin-1, haptoglobin (Fragment), mannan-binding lectin serine peptidase 1, clusterin, and uncharacterized protein (AC: Q3T0Z0) were highest in buffalo milk whey. Folate receptor alpha, uncharacterized protein (AC: F1MK50), and adipose differentiation-related protein were highest in yak milk whey. Alpha-lactalbumin (AC: P00710), glycosylation-dependent cell adhesion molecule 1, quinone oxidoreductase, whey acidic protein, and peptidoglycan recognition protein 1 were highest in camel milk whey. These results suggest that there is a significant difference in composition of milk whey protein

2.6. Multivariate Analysis

Quantification of proteins from different species was analyzed using a PCA program by Unscrambler software (Camo, version 9.8, Norway). PCA is a multivariate statistical method that transforms the original data set into noncorrelated variables called principal components (PC). Each PC is a linear combination of original variables and has the property of explaining the maximum possible amount of variance contained 1662

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667

Journal of Proteome Research

Article

cell adhesion molecule 1, α-lactalbumin (AC: P00710), quinone oxidoreductase, whey acidic protein, peptidoglycan recognition protein 1, and lactalbumin (Fragment) (AC: O97723); and PC2 principal contained biglycan, serum amyloid A protein, uncharacterized protein (AC: F1MK50), C-X-C motif chemokine 6, and polymeric immunoglobulin receptor that explained 72% of the total variance based on quantitative iTRAQ expression proteins among species milk whey. Score plot of PCA is shown in Figure 2A. Camel milk is separated from others’ milk in the direction of PC1, goat milk is separated from others’ milk in the direction of PC2, which indicates changes in the milk whey proteins pattern among species milk. The PCA results demonstrate a very effective discrimination between camel and goat milk, and others’ milk. This discrimination is less effective but also clear among yak, cow, and buffalo milk. Corresponding correlation loading plots shown in Figure 2B, in which each dot represents a protein identified by quantitative iTRAQ techniques, are used to identify which variables of quantitative proteins contribute to separation of samples on scores plot. The first four PCs explained more than 87% of total variance in the original data set, and were able to differentiate clearly among species. According to PCA results, animal species milk could be identified on the basis of quantitative whey proteins listed in Table 2. 3.4. Cluster Analysis

Finally, 177 differential proteins were submitted to further hierarchical clustering using Cluster 3.0 software. This analysis yielded a clustering pattern with three major sample clusters. Cow milk also presented subclusters for both protein concentrations. Cow, buffalo, and yak milk had similar proteomic patterns and comprised a major sample cluster. Goat milk added to this group yielded another cluster; and camel milk clustered alone. Interestingly, proteomic patterns in cow, buffalo, and yak milk are similar, but significant differences were observed for uncharacterized protein (AC: F1MK50) and folate receptor alpha in yak milk; haptoglobin, cysteine-rich secretory protein 3, and clusterin in buffalo milk; serum amyloid A protein, biglycan, and C-X-C motif chemokine 6 in goat milk; and glycosylation-dependent cell adhesion molecule 1, alphalactalbumin (AC: P00710), whey acidic protein, and quinone oxidoreductase in camel milk. In addition, although proteomic patterns in 3.46% (w/v)- and 2.75% (w/v)-protein content cow milk are very similar, protein expression patterns revealed differences for secretoglobin family 1D member (Figure 3).

Figure 1. Classification of 113 identified proteins in milk whey according to gene ontology annotation; (A) 97 identified proteins in molecular function, (B) 98 identified proteins in cellular components, and (C) 86 identified proteins in biological processes.

profile among species. In addition, the osteopontin, secretoglobin family 1D member, lactoferrin, and lactoferrin (fragment) in the whey from 3.46%-protein (w/v) cow milk samples were significantly greater compared with whey from 2.75%protein (w/v) cow milk samples. In other words, effect of cow milk protein concentration on the protein composition profile was very limited. 3.3. PCA Analysis

4. DISCUSSION There are two major findings presented in this study. These included enhancing knowledge of protein composition in milk from minor animal species, and establishment of some of the quantitative differential expression protein profile among animal species’ milk. Previously, low-abundance proteins in cow milk were enhanced by peptide ligand libraries and identified by nanoLC−MS/MS. A total of 149 unique protein species were submitted to functional characterization and demonstrated that binding activity is most common in molecular function.18 Subsequently, human and bovine milk were analyzed using a shotgun proteomics approach and total of 192 proteins were identified in bovine milk serum.13 Le et al. (2011) reported that 293 unique gene products were identified in bovine milk whey using the ion-exchange protein

In the present study, the matching procedure yielded 211 variables and 18 samples (data matrix, 211 × 18). Animal milk whey samples were analyzed in triplicate. All variables were autoscaled before performing PCA (Table 1). The PC1 principal contained lactotransferrin, glycosylation-dependent Table 1. PCA Results (for the First Four PCs) Performed on the Overall Data Set

PC1 PC2 PC3 PC4

explained variance %

cumulative explained variance %

60.14 12.63 8.86 5.93

60.14 72.77 81.63 87.56 1663

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667

Journal of Proteome Research

Article

Figure 2. (A) PCA scores plot and (B) corresponding correlation loadings plot of the first two principal components based on quantitative iTRAQ expression proteins. Dots ID show identified proteins by quantitative iTRAQ techniques.

identification among labeled samples; and discriminating peptides from cow, goat, buffalo, yak, and camel milk whey. Principal proteins in bovine, caprine, buffalo, equine, and camel milk were separated and identified using the 2-DE-based proteomics approach.24 Thus far, few studies have examined milk whey proteome of buffalo, yak, and camel milk. This limited knowledge store implies that few proteins could be identified by consulting databases for buffalo, yak, and camel, particularly in our analysis of camel milk. In this paper, we characterized proteomic patterns of milk from various animal species using the iTRAQ quantitative proteomic method. These results may provide valuable information in elucidating the differential expression pattern of milk proteins as well as the composition of milk protein in minor mammalian species. In addition, some proteins originating from blood and associated with diseases or stress were also identified in the milk whey. Turk et al. (2012) reported that serpin A3-1, haptoglobin, and apolipoprotein A-I in serum were upregulated in cows with subclinical and clinical mastitis.32 Other research indicated that serotransferrin, apolipoprotein A1, and complement C3 were present in serum and increased in milk of cows infected with mastitis by 2-DE coupled with MALDI-TOF MS.33 However, these proteins originating from blood were also identified in milk proteome after milk whey protein was identified by the higher sensitivity method, including fractions in samples determined with LC−MS/ MS.13,18 Multivariate statistical methods, such as cluster and PCA, comprised a powerful data-mining approach for exploration of proteomic data.34 These methods have also been used to

Table 2. Differential Proteins among Animal Species According to PCA Results species Camel milk

Goat milk

Yak milk Buffalo milk

Cow milk

accession no. P15522 P00710 P09837 Q28452 Q9GK12 O97723 Q9TUM0 A5JST2 P21809 Q8SQ82 P80221 F1MK50 D2U6Q1 P17697 Q08DW4 Q28178 Q3T0Z0 B0JYP6 P79345 Q29437

protein name glycosylation-dependent cell adhesion molecule 1 alpha-lactalbumin whey acidic protein quinone oxidoreductase peptidoglycan recognition protein 1 lactalbumin (Fragment) lactotransferrin serum amyloid A protein biglycan polymeric immunoglobulin receptor (Fragment) C-X-C motif chemokine 6 uncharacterized protein haptoglobin clusterin mannan-binding lectin serine peptidase 1 thrombospondin-1 uncharacterized protein IGK protein epididymal secretory protein E1 primary amine oxidase

fractionation method. In our study, 211 milk proteins were identified in milk whey from above-mentioned animal species. This is the result of protein identification inferred from peptide 1664

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667

Journal of Proteome Research

Article

residues and a molecular mass of 16.2 kDa.24,37 Based on these results, milk mixture samples were separated by conventional SDS-PAGE or 2-DE for detection of α-lactalbumin molecular mass for milk adulteration analyses. Whey acidic protein was one of the most abundant proteins in camel milk, but it is absent in bovine and goat milk due to a frameshift mutation in the whey acidic protein gene of these animals.38 To investigate milk adulteration, whey acidic protein may serve as a speciesspecific marker as detected by Western blot or ELISA method. Thus, α-lactalbumin (AC: P00710) and whey acidic protein were found to be characteristic molecular traits of camel milk, and could be used to distinguish camel milk from milk of other analyzed species. Biglycan is a member of the family of small leucine-rich proteoglycans. Relative abundance of biglycan in goat milk is 4.54, which is significantly (P = 3.5 × 10−08) higher than that of cow, buffalo, yak and camel milk. Biglycan is a newer identified protein in milk whey in our study, and may act as a speciesspecific marker in goat milk. Moreover, overexpression of biglycan has been strongly associated with tumor and disease.39,40 Clusterin, a highly glycosylated protein, was one of the most abundant proteins in buffalo milk. Previously, it was identified in animal and human milk fat globular membrane and milk whey by proteomic methodology.41−43 In addition, single nucleotide polymorphisms of clusterin gene were associated with milk production traits including milk yield, fat rate, and protein rate in Chinese Holstein cows.44 However, little information suggested that clusterin served as the speciesspecific markers in the buffalo milk in previous study. As reported by Zimin et al. (2009) uncharacterized protein (AC: F1MK50) is one of the most abundant proteins (relative abundance is 6.30) in yak milk, and no significant similarity was found in the blast search of uniprot and NCBI database.45 Bioinformatic software TMHMM 2.0 and SignalP 4.0 were applied to characterize the uncharacterized protein that was predicted to be soluble and harbored a signal peptide.46,47 These proteins could serve as the species-specific markers based on our current study, but lack of this differentiation limited investigation in previous studies. Furthermore, primary amine oxidase was found to be one of the most abundant proteins in cow milk, is involved in the amine metabolic process,48 and was observed in milk by Affolter et al.49 A previous study investigated joining and constant IgK in Holstein Friesian, German Black Pied, German Simmental and Aubrac cattle. That presented three alleles coding for two putative allotypic variants in kappa-light chain regions.50 The IgK is another abundant protein in cow milk and serves as the species-specific marker according to PCA analysis, but was not reported in previous studies. However, previous studies demonstrated that another immunoglobulin determined by an ELISA assay, IgG, was a target molecule for detection of cow’s milk versus milk of goat, sheep, and buffalo.51 Osteopontin is another abundant protein in cow milk. Leonard et al. (2005) reported that osteopontin variants were associated with milk protein percentage and milk fat percentage.52 Although some identified proteins served as molecular markers in specific milk according to PCA analysis, there is little evidence to support these results and this needs further investigation with Western blot or ELISA assays. Nonetheless, the current study is an important step in elucidating differences of milk whey proteome in analyzed animal species. The proteome identified here should be explored further in future research for better understanding

Figure 3. Hierarchical clustering of differentially expressed proteins. Clustering was based on protein expression levels in goat, camel, cow, buffalo and yak milk. Bar color represents a logarithmic scale from −3.0 to 3.0.

investigate differential proteome patterns of different classes of tissues on the basis of either quantitative or qualitative changes in the level of identified proteins.35,36 According to multivariate statistical analysis of quantitatively identified proteins, some proteins can be useful for discrimination of milk from analyzed species. The primary whey protein was α-lactalbumin (AC: P00710) whose relative abundance was 9.08 and was found in significantly (P = 1.73 × 10−08) greatest abundance in camel milk as compared to all other investigated animal milk. This is because α-lactalbumin comprises 123 amino acid residues and has a molecular mass of 14.6 kDa that exhibits comparatively large differences from proteins of other investigated species, such as bovine and goat which consisted of 142 amino acid 1665

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667

Journal of Proteome Research

Article

comparison of whey from healthy animals and from those with clinical mastitis. Proteomics 2004, 4 (7), 2094−100. (5) Le, A.; Barton, L. D.; Sanders, J. T.; Zhang, Q. Exploration of bovine milk proteome in colostral and mature whey using an ionexchange approach. J. Proteome Res. 2011, 10 (2), 692−704. (6) O’Donnell, R.; Holland, J.; Deeth, H.; Alewood, P. Milk proteomics. Int. Dairy J. 2004, 14 (12), 1013−23. (7) Liao, Y.; Alvarado, R.; Phinney, B.; Lonnerdal, B. Proteomic characterization of human milk whey proteins during a twelve-month lactation period. J. Proteome Res. 2011, 10 (4), 1746−54. (8) Mange, A.; Bellet, V.; Tuaillon, E.; Van de Perre, P.; Solassol, J. Comprehensive proteomic analysis of the human milk proteome: contribution of protein fractionation. J. Chromatogr., B: Analyt. Technol. Biomed. Life Sci. 2008, 876 (2), 252−6. (9) Cunsolo, V.; Muccilli, V.; Saletti, R.; Foti, S. Review: applications of mass spectrometry techniques in the investigation of milk proteome. Eur. J. Mass Spectrom. (Chichester, Eng) 2011, 17 (4), 305−20. (10) Lindmark-Månsson, H.; Timgren, A.; Aldén, G.; Paulsson, M. Two-dimensional gel electrophoresis of proteins and peptides in bovine milk. Int. Dairy J. 2005, 15 (2), 111−21. (11) Galvani, M.; Hamdan, M.; Righetti, P. G. Two-dimensional gel electrophoresis/matrix-assisted laser desorption/ionisation mass spectrometry of commercial bovine milk. Rapid Commun. Mass Spectrom. 2001, 15 (4), 258−64. (12) Yamada, M.; Murakami, K.; Wallingford, J. C.; Yuki, Y. Identification of low-abundance proteins of bovine colostral and mature milk using two-dimensional electrophoresis followed by microsequencing and mass spectrometry. Electrophoresis 2002, 23 (7−8), 1153−60. (13) Hettinga, K.; van Valenberg, H.; de Vries, S.; Boeren, S.; van Hooijdonk, T.; van Arendonk, J.; Vervoort, J. The host defense proteome of human and bovine milk. PLoS One 2011, 6 (4), e19433. (14) Holland, J. W.; Deeth, H. C.; Alewood, P. F. Analysis of disulphide linkages in bovine kappa-casein oligomers using twodimensional electrophoresis. Electrophoresis 2008, 29 (11), 2402−10. (15) Holland, J. W.; Deeth, H. C.; Alewood, P. F. Analysis of Oglycosylation site occupancy in bovine kappa-casein glycoforms separated by two-dimensional gel electrophoresis. Proteomics 2005, 5 (4), 990−1002. (16) Holland, J. W.; Deeth, H. C.; Alewood, P. F. Proteomic analysis of kappa-casein micro-heterogeneity. Proteomics 2004, 4 (3), 743−52. (17) Fong, B. Y.; Norris, C. S.; Palmano, K. P. Fractionation of bovine whey proteins and characterisation by proteomic techniques. Int. Dairy J. 2008, 18 (1), 23−46. (18) D’Amato, A.; Bachi, A.; Fasoli, E.; Boschetti, E.; Peltre, G.; Senechal, H.; Righetti, P. G. In-depth exploration of cow’s whey proteome via combinatorial peptide ligand libraries. J. Proteome Res. 2009, 8 (8), 3925−36. (19) D’Alessandro, A.; Zolla, L.; Scaloni, A. The bovine milk proteome: cherishing, nourishing and fostering molecular complexity. An interactomics and functional overview. Mol. Biosyst. 2011, 7 (3), 579−97. (20) Chianese, L.; Calabrese, M. G.; Ferranti, P.; Mauriello, R.; Garro, G.; De Simone, C.; Quarto, M.; Addeo, F.; Cosenza, G.; Ramunno, L. Proteomic characterization of donkey milk “caseome”. J. Chromatogr., A 2010, 1217 (29), 4834−40. (21) Roncada, P.; Gaviraghi, A.; Liberatori, S.; Canas, B.; Bini, L.; Greppi, G. F. Identification of caseins in goat milk. Proteomics 2002, 2 (6), 723−6. (22) Pisanu, S.; Ghisaura, S.; Pagnozzi, D.; Biosa, G.; Tanca, A.; Roggio, T.; Uzzau, S.; Addis, M. F. The sheep milk fat globule membrane proteome. J. Proteomics 2011, 74 (3), 350−8. (23) Spertino, S.; Cipriani, V.; De Angelis, C.; Giuffrida, M. G.; Marsano, F.; Cavaletto, M. Proteome profile and biological activity of caprine, bovine and human milk fat globules. Mol. Biosyst. 2012, 8 (4), 967−74. (24) Hinz, K.; O’Connor, P. M.; Huppertz, T.; Ross, R. P.; Kelly, A. L. Comparison of the principal proteins in bovine, caprine, buffalo, equine and camel milk. J. Dairy Res. 2012, 79 (2), 185−91.

the protein composition profile in animal species milk and identification of specific milk.

■

CONCLUSION In the current study, whey protein from cow, yak, buffalo, goat and camel milk were analyzed quantitatively using iTRAQ techniques. Proteins identified were classified according to associated gene ontology annotations. Differential proteomic patterns were detected using multivariate methods for evaluation of quantitatively identified proteins and investigated by hierarchical clustering of the differentially expressed proteins. Hierarchical clustering was also used to select a reduced set of potential trait markers for the characterization of proteomic patterns in milk from different animal species. The following proteins were selected in this manner: α-lactalbumin (AC: P00710), whey acidic protein, and quinone oxidoreductase in camel milk; fatty acid synthase and biglycan in goat milk; uncharacterized protein (AC: F1MK50) and folate receptor alpha in yak milk; thrombospondin-1 and clusterin in buffalo milk; and primary amine oxidase in cow milk. Quantifying the levels of these proteins in milk is a molecular tool that can be used for the accurate identification of specific milk from a particular mixture of milk. These findings elucidate the proteome composition of milk whey and quantitative protein profile in analyzed animal milk. This may be used to evaluate milk adulteration of specific milk and expand the potential direction for production of specific milk proteins.

■

ASSOCIATED CONTENT

S Supporting Information *

Supplementary tables and figures. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: 86-10-62890458. Fax: 86-10-62897587. Author Contributions §

Y.Y. and D.B. contributed equally to this work

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The project was supported by the National Key Basic Research Program of China (Project No. 2011CB100805) and grants (2012BAD12B02-5; 2004DA125184G1103) by Ministry of Science and Technology.

■

REFERENCES

(1) Medhammar, E.; Wijesinha-Bettoni, R.; Stadlmayr, B.; Nilsson, E.; Charrondiere, U. R.; Burlingame, B. Composition of milk from minor dairy animals and buffalo breeds: a biodiversity perspective. J. Sci. Food Agric. 2012, 92 (3), 445−74. (2) Roncada, P.; Piras, C.; Soggiu, A.; Turk, R.; Urbani, A.; Bonizzi, L. Farm animal milk proteomics. J. Proteomics 2012, 75 (14), 4259− 74. (3) Dorji, T.; Namikawa, T.; Mannen, H.; Kawamoto, Y. Milk protein polymorphisms in cattle (Bos indicus), mithun (Bos frontalis) and yak (Bos grunniens) breeds and their hybrids indigenous to Bhutan. Anim. Sci. J. 2010, 81 (5), 523−9. (4) Hogarth, C. J.; Fitzpatrick, J. L.; Nolan, A. M.; Young, F. J.; Pitt, A.; Eckersall, P. D. Differential protein composition of bovine whey: a 1666

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667

Journal of Proteome Research

Article

ization of human milk fat globular membrane proteins. Anal. Biochem. 2002, 301 (2), 314−24. (44) Wang, Z.; Huang, J.; Zhong, J.; Wang, G. Molecular cloning, promoter analysis, SNP detection of Clusterin gene and their associations with mastitis in Chinese Holstein cows. Mol. Biol. Rep. 2012, 39 (3), 2439−45. (45) Zimin, A. V.; Delcher, A. L.; Florea, L.; Kelley, D. R.; Schatz, M. C.; Puiu, D.; Hanrahan, F.; Pertea, G.; Van Tassell, C. P.; Sonstegard, T. S.; Marcais, G.; Roberts, M.; Subramanian, P.; Yorke, J. A.; Salzberg, S. L. A whole-genome assembly of the domestic cow, Bos taurus. Genome Biol. 2009, 10 (4), R42. (46) Petersen, T. N.; Brunak, S.; von Heijne, G.; Nielsen, H. SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat. Methods 2011, 8 (10), 785−6. (47) Krogh, A.; Larsson, B.; von Heijne, G.; Sonnhammer, E. L. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J. Mol. Biol. 2001, 305 (3), 567−80. (48) Zhang, X.; McIntire, W. S. Cloning and sequencing of a coppercontaining, topa quinone-containing monoamine oxidase from human placenta. Gene 1996, 179 (2), 279−86. (49) Affolter, M.; Grass, L.; Vanrobaeys, F.; Casado, B.; Kussmann, M. Qualitative and quantitative profiling of the bovine milk fat globule membrane proteome. J. Proteomics 2010, 73 (6), 1079−88. (50) Stein, S. K.; Diesterbeck, U. S.; Aboelhassan, D. M.; Czerny, C. P. Comparison of joining and constant kappa-light chain regions in different cattle breeds. Anim. Genet. 2012, 43 (6), 776−80. (51) Hurley, I. P.; Coleman, R. C.; Ireland, H. E.; Williams, J. H. Measurement of bovine IgG by indirect competitive ELISA as a means of detecting milk adulteration. J. Dairy Sci. 2004, 87 (3), 543−9. (52) Leonard, S.; Khatib, H.; Schutzkus, V.; Chang, Y. M.; Maltecca, C. Effects of the osteopontin gene variants on milk production traits in dairy cattle. J. Dairy Sci. 2005, 88 (11), 4083−6.

(25) D’Auria, E.; Agostoni, C.; Giovannini, M.; Riva, E.; Zetterstrom, R.; Fortin, R.; Greppi, G. F.; Bonizzi, L.; Roncada, P. Proteomic evaluation of milk from different mammalian species as a substitute for breast milk. Acta Paediatr. 2005, 94 (12), 1708−13. (26) Nikkhah, A. Equidae, camel, and yak milks as functional foods: a review. J. Nutr. Food Sci. 2011, 1 (116), 2. (27) Wu, W. W.; Wang, G.; Baek, S. J.; Shen, R. F. Comparative study of three proteomic quantitative methods, DIGE, cICAT, and iTRAQ, using 2D gel- or LC-MALDI TOF/TOF. J. Proteome Res. 2006, 5 (3), 651−8. (28) Ow, S. Y.; Salim, M.; Noirel, J.; Evans, C.; Rehman, I.; Wright, P. C. iTRAQ underestimation in simple and complex mixtures: “the good, the bad and the ugly”. J. Proteome Res. 2009, 8 (11), 5347−55. (29) Pottiez, G.; Wiederin, J.; Fox, H. S.; Ciborowski, P. Comparison of 4-plex to 8-plex iTRAQ quantitative measurements of proteins in human plasma samples. J. Proteome Res. 2012, 11 (7), 3774−81. (30) Wisniewski, J. R.; Zougman, A.; Nagaraj, N.; Mann, M. Universal sample preparation method for proteome analysis. Nat. Methods 2009, 6 (5), 359−62. (31) Sheng, Q.; Dai, J.; Wu, Y.; Tang, H.; Zeng, R. BuildSummary: using a group-based approach to improve the sensitivity of peptide/ protein identification in shotgun proteomics. J. Proteome Res. 2012, 11 (3), 1494−502. (32) Turk, R.; Piras, C.; Kovacic, M.; Samardzija, M.; Ahmed, H.; De Canio, M.; Urbani, A.; Mestric, Z. F.; Soggiu, A.; Bonizzi, L.; Roncada, P. Proteomics of inflammatory and oxidative stress response in cows with subclinical and clinical mastitis. J. Proteomics 2012, 75 (14), 4412−28. (33) Alonso-Fauste, I.; Andres, M.; Iturralde, M.; Lampreave, F.; Gallart, J.; Alava, M. A. Proteomic characterization by 2-DE in bovine serum and whey from healthy and mastitis affected farm animals. J. Proteomics 2012, 75 (10), 3015−30. (34) Meunier, B.; Dumas, E.; Piec, I.; Bechet, D.; Hebraud, M.; Hocquette, J. F. Assessment of hierarchical clustering methodologies for proteomic data mining. J. Proteome Res. 2007, 6 (1), 358−66. (35) Marengo, E.; Robotti, E.; Righetti, P. G.; Campostrini, N.; Pascali, J.; Ponzoni, M.; Hamdan, M.; Astner, H. Study of proteomic changes associated with healthy and tumoral murine samples in neuroblastoma by principal component analysis and classification methods. Clin. Chim. Acta 2004, 345 (1−2), 55−67. (36) Gomez, A.; Lopez, J. A.; Pintos, B.; Camafeita, E.; Bueno, M. A. Proteomic analysis from haploid and diploid embryos of Quercus suber L. identifies qualitative and quantitative differential expression patterns. Proteomics 2009, 9 (18), 4355−67. (37) Beg, O. U.; von Bahr-Lindstrom, H.; Zaidi, Z. H.; Jornvall, H. The primary structure of alpha-lactalbumin from camel milk. Eur. J. Biochem. 1985, 147 (2), 233−9. (38) Hajjoubi, S.; Rival-Gervier, S.; Hayes, H.; Floriot, S.; Eggen, A.; Piumi, F.; Chardon, P.; Houdebine, L. M.; Thepot, D. Ruminants genome no longer contains Whey Acidic Protein gene but only a pseudogene. Gene 2006, 370, 104−12. (39) Wang, B.; Li, G. X.; Zhang, S. G.; Wang, Q.; Wen, Y. G.; Tang, H. M.; Zhou, C. Z.; Xing, A. Y.; Fan, J. W.; Yan, D. W.; Qiu, G. Q.; Yu, Z. H.; Peng, Z. H. Biglycan expression correlates with aggressiveness and poor prognosis of gastric cancer. Exp. Biol. Med. (Maywood) 2011, 236 (11), 1247−53. (40) Gu, X.; Ma, Y.; Xiao, J.; Zheng, H.; Song, C.; Gong, Y.; Xing, X. Up-regulated biglycan expression correlates with the malignancy in human colorectal cancers. Clin. Exp. Med. 2012, 12 (3), 195−9. (41) Murakami, K.; Lagarde, M.; Yuki, Y. Identification of minor proteins of human colostrum and mature milk by two-dimensional electrophoresis. Electrophoresis 1998, 19 (14), 2521−7. (42) Kuy, S.; Kelly, V. C.; Smit, A. M.; Palmer, D. J.; Cooper, G. J. Proteomic analysis of whey and casein proteins in early milk from the marsupial Trichosurus vulpecula, the common brushtail possum. Comp. Biochem. Physiol., Part D: Genomics Proteomics 2007, 2 (2), 112−20. (43) Charlwood, J.; Hanrahan, S.; Tyldesley, R.; Langridge, J.; Dwek, M.; Camilleri, P. Use of proteomic methodology for the character1667

dx.doi.org/10.1021/pr301001m | J. Proteome Res. 2013, 12, 1660−1667