Comparative Proteomics Provides Insights into ... - ACS Publications

Article pubs.acs.org/jpr

Comparative Proteomics Provides Insights into Metabolic Responses in Rat Liver to Isolated Soy and Meat Proteins Shangxin Song,† Guido J. Hooiveld,‡ Wei Zhang,∥ Mengjie Li,† Fan Zhao,† Jing Zhu,† Xinglian Xu,† Michael Muller,§ Chunbao Li,*,† and Guanghong Zhou*,† †

Key Laboratory of Meat Processing and Quality Control, MOE; Key Laboratory of Animal Products Processing, MOA; Jiang Synergetic Innovation Center of Meat Processing and Quality Control, Nanjing Agricultural University, Nanjing 210095, P.R. China ‡ Nutrition, Metabolism and Genomics Group, Division of Human Nutrition, Wageningen University, Wageningen 6703 HD, The Netherlands § Norwich Medical School, University of East Anglia, Norwich NR4 7TJ, United Kingdom ∥ Key Laboratory of Human Functional Genomics Jiangsu Province, Nanjing Medical University, Nanjing 210029, P. R. China S Supporting Information *

ABSTRACT: It has been reported that isolated dietary soy and meat proteins have distinct effects on physiology and liver gene expression, but the impact on protein expression responses are unknown. Because these may differ from gene expression responses, we investigated dietary protein-induced changes in liver proteome. Rats were fed for 1 week semisynthetic diets that differed only regarding protein source; casein (reference) was fully replaced by isolated soy, chicken, fish, or pork protein. Changes in liver proteome were measured by iTRAQ labeling and LC−ESI−MS/ MS. A robust set totaling 1437 unique proteins was identified and subjected to differential protein analysis and biological interpretation. Compared with casein, all other protein sources reduced the abundance of proteins involved in fatty acid metabolism and Pparα signaling pathway. All dietary proteins, except chicken, increased oxidoreductive transformation reactions but reduced energy and essential amino acid metabolic pathways. Only soy protein increased the metabolism of sulfur-containing and nonessential amino acids. Soy and fish proteins increased translation and mRNA processing, whereas only chicken protein increased TCA cycle but reduced immune responses. These findings were partially in line with previously reported transcriptome results. This study further shows the distinct effects of soy and meat proteins on liver metabolism in rats. KEYWORDS: metabolic syndrome, isolated protein, animal protein, chicken protein, fish protein, pork protein, molecular nutrition, nutrigenomics, proteomics

■

young rats;10,11 however, it is also recognized that the abundance of mRNA transcripts is not always representative of cellular protein levels. Thus, to further understand metabolic responses on the level of protein expression to different sources of dietary proteins, we analyzed and compared the liver proteome of rats fed nutritionally balanced semisynthetic diets that differed only in protein source. The liver proteome was investigated by using iTRAQ (isobaric tagging for relative and absolute quantitation) labeling and LC−ESI−MS/MS. Changes observed in the present iTRAQ data set were compared with our previous RNA-seq data using gene set enrichment analysis (GSEA).

INTRODUCTION Metabolic syndrome is becoming a global epidemic that increases risks of cardiovascular diseases and type 2 diabetes.1 An increasing number of studies show that metabolic syndrome parameters can be favorably modulated by altering total dietary protein2−6 or with whole food protein sources.7 Meat and soy are the major dietary protein sources for human nutrition. Compared with plant protein, meat protein distinguishes itself for its richness in all essential amino acids (AAs);8 however, in contrast with soy plant protein, which has been widely studied, the effects of meat proteins on physiological and metabolic health have been less well investigated, especially for meat proteins from different species. We previously demonstrated that the protein composition of meat proteins from pork, chicken, and fish and their digests were very different,9 and short-term feeding of dietary proteins from different sources resulted in distinct physiological and transcriptome changes in © 2016 American Chemical Society

Received: October 1, 2015 Published: February 17, 2016 1135

DOI: 10.1021/acs.jproteome.5b00922 J. Proteome Res. 2016, 15, 1135−1142

Article

Journal of Proteome Research

■

EXPERIMENTAL SECTION

centrifugation. Because there were three liver samples for each group, there was a total of three peptides mixtures. To reduce sample complexity, we used the strong cation exchange (SCX) chromatography in the fractionation of iTRAQ-labeled peptides. SCX chromatography was performed with an LC-20AB HPLC pump system (Shimadzu, Kyoto, Japan). First, the peptide mixtures were reconstituted in 4 mL of buffer A (25 mM NaH2PO4 in 25% ACN, pH 2.7) and loaded onto a 4.6 × 250 mm Ultremex SCX column containing 5-μm particles (Phenomenex). The peptides were eluted at a flow rate of 1 mL/min using a gradient of buffer A for 10 min, 5−60% buffer B (25 mM NaH2PO4, 1 M KCl in 25% ACN, pH 2.7) for 27 min, and 60−100% buffer B for 1 min. The system was then maintained at 100% buffer B for 1 min before equilibrating with buffer A for 10 min prior to the next injection. Elution was monitored by measuring the absorbance at 214 nm, and fractions were collected every 1 min. The eluted peptides were pooled into 20 fractions, desalted with a Strata X C18 column (Phenomenex), and vacuum-dried. LC−ESI−MS/MS Analysis Based on Q EXACTIVE. Each fraction was resuspended in buffer A (2% ACN, 0.1% FA) and centrifuged at 20 000g for 10 min, and the average final concentration of peptide was ∼0.5 μg/μL. Supernatant (10 μL) was loaded on a LC-20AD nanoHPLC (Shimadzu, Kyoto, Japan) by the autosampler onto a 2 cm C18 trap column. Then, the peptides were eluted onto a 10 cm analytical C18 column (inner diameter 75 μm) packed in-house. The samples were loaded at 8 μL/min for 4 min; then, a 44 min gradient was run at 300 nL/min starting from 2 to 35% B (98% ACN, 0.1% FA), followed by a 2 min linear gradient to 80%, then maintenance at 80% B for 4 min, and finally returning to 5% in 1 min. The peptides were subjected to nanoelectrospray ionization, followed by tandem mass spectrometry (MS/MS) in an Q EXACTIVE (Thermo Fisher Scientific, San Jose, CA) coupled to the HPLC system. Intact peptides were detected in the Orbitrap at a resolution of 70 000. Peptides were selected for MS/MS using high-energy collision dissociation (HCD) operating mode with a normalized collision energy setting of 27.0; ion fragments were detected in the Orbitrap at a resolution of 17 500. A data-dependent procedure that alternated between one MS scan followed by 15 MS/MS scans was applied for the 15 most abundant precursor ions above a threshold ion count of 20 000 in the MS survey scan with a following dynamic exclusion duration of 15 s. The electrospray voltage applied was 1.6 kV. Automatic gain control (AGC) was used to optimize the spectra generated by the Orbitrap. The AGC target for full MS was 3e6 and 1e5 for MS2. For MS scans, the m/z scan range was 350−2000 Da. For MS2 scans, the m/z scan range was 100−1800. Mass Spectrometric Data Analysis. Raw mass spectrometric data acquired from the Orbitrap were analyzed in MaxQuant version 1.5.2.8.15 In total, 948 072 MS/MS spectra from raw files were searched against the UniProtKB Rattus norvegicus (2014.11 release, 34,165 entries) using the Andromeda search engine.16 The search type was set to Reporter Ion MS2. iTRAQ8plex-Nter 114/116/117/119/121 and iTRAQ8plex-Lys 114/116/117/119/121 were selected as the isobaric labels. The mass tolerance for the first search was 20 ppm, the results of which were used for mass recalibration.17 Mass tolerances for the main search and isotope match were 4.5 and 2 ppm, respectively. Enzyme specificity was set to trypsin. Variable modifications included acetyl (protein N-term), deamidation (NQ), Gln → pyro-Glu, and Oxidation (M),

Diets and Animals

All animals were handled in accordance with the guidelines of the Ethical Committee of Experimental Animal Center of Nanjing Agricultural University. The animal experiment has been previously described.10,11 In brief, after a 1 week adaptation period, 4 week old male Sprague−Dawley rats were fed for 7 days the nutritionally balanced semisynthetic AIN-93G diet12 or the same diet in which the protein source (casein; milk protein) was fully replaced by isolated proteins from soy, pork, chicken or fish (n = 10 rats in each group). Immediately following euthanasia, livers were obtained, snap frozen in liquid nitrogen, and stored at −80 °C until analysis. Quantitative Proteomic Analysis Protein Preparation

Three liver samples were randomly selected from each group for quantitative proteomic analysis. The procedures for protein preparations were according to previous papers.13,14 In brief, liver samples were ground into powder in liquid nitrogen and extracted with lysis buffer I (7 M urea, 2 M thiourea, 4% CHAPS, 40 mM Tris-HCl, pH 8.5) containing 1 mM PMSF and 2 mM EDTA (final concentration). After 5 min, 10 mM DTT (final concentration) was added to the samples. The suspension was sonicated at 200 W for 15 min and then centrifuged at 4 °C, 30 000g for 15 min. The supernatant was mixed with 5× volume of chilled acetone containing 10% (v/v) TCA and kept at −20 °C overnight. The samples were centrifuged at 30 000 g (4 °C) and the supernatant was discarded. The precipitate was washed with chilled acetone three times. The pellet was air-dried and dissolved in lysis buffer II (7 M urea, 2 M thiourea, 4% NP40, 20 mM Tris-HCl, pH 8.0 to 8.5). After sonication and centrifugation, the supernatant was transferred to new tube. 10 mM DTT was added and incubated at 56 °C for 1 h to reduce the disulfide bonds in proteins of the supernatant. Subsequently, 55 mM iodoacetamide (IAM) was added to block the cysteines and then incubated for 1 h in a darkroom. The supernatant was mixed with 5× volume of chilled acetone for 2 h at −20 °C to precipitate proteins. After centrifugation at 30 000 g (4 °C) the supernatant was discarded and the pellet was air-dried for 5 min, dissolved in 500 μL of 0.5 M TEAB (Applied Biosystems, Milan, Italy), and sonicated at 200 W for 15 min. Finally, samples were centrifuged at 30 000g (4 °C) for 15 min. The supernatant was transferred to a new tube and quantified. The proteins in the supernatant were kept at −80 °C for further analysis. iTRAQ Labeling and Strong Cation Exchange Fractionation. Total protein (100 μg) was digested with Trypsin Gold (Promega, Madison, WI) using the ratio of protein/ trypsin 30:1 at 37 °C for 16 h. After digestion, peptides were dried by vacuum centrifugation. Peptides were reconstituted in 0.5 M TEAB and processed according to the manufacture’s protocol for 8-plex iTRAQ reagent (Applied Biosystems). In brief, one unit of iTRAQ reagent was thawed and reconstituted in 24 μL of isopropanol. Samples were labeled with the iTRAQ tags as follows: casein group (114 tag), soy protein group (116 tag), pork protein group (117 tag), chicken protein group (119 tag), and fish protein group (121 tag). The labeled peptides were incubated at room temperature for 2 h. Finally, the five labeled samples were pooled and dried by vacuum 1136


Article


Figure 1. (A) Venn plots of proteins (P < 0.05 by moderated t test) and (B) protein sets (P < 0.05 by GSEA analysis) that were significantly changed by isolated soy, chicken, fish, and pork proteins. C, casein group; down, down-regulated; up, up-regulated.

of “gene set” in the following statements or the other parts of the paper. The predefined protein sets were derived from the KEGG, Reactome, Biocarta, and WikiPathways databases. Only protein sets consisting of more than 15 and fewer than 500 proteins were taken into account. In this study, proteins that passed various strict filters were initially ranked based on their statistical T value obtained from the moderated t test. During the GSEA analysis, the enrichment score was calculated, and the statistical significance (P value) of enrichment score was determined using 1000 permutations. To improve the interpretation of the GSEA results, we visualized significant protein sets (P < 0.05) in the Enrichment Map plugin v2.1.026 in Cytoscape v3.2.1. For visual comparison, a merged enrichment map was generated that consisted of protein sets that were regulated by at least one of the dietary proteins. This merged network served as canvas on which the significantly regulated protein sets per dietary protein were overlaid. The gene sets from the GSEA results of the previous RNA-seq data10,11 were also shown in the network as node border area to compare the changes observed in the present iTRAQ data set with the previous RNA-seq data set.

and fixed modification was carbamidomethyl (C). Minimal peptide length was set to 7 AAs, and a maximum of two miscleavages were allowed. The second peptide option was activated to enable identification of coeluting peptides with very similar mass.16 For protein identification, a minimum of one peptide (razor+unique) was required. Proteins sharing the same peptides were combined and reported as one protein group. The expression amount of proteins was represented by the intensities of iTRAQ reporter ions in MS/MS spectra. Before statistical analysis, protein groups obtained from MaxQuant search were filtered in the Perseus software package v1.5.1.6. Potential contaminants were removed. Proteins recognized by the reverse database, or only identified by site, or mapped by fewer than two peptides were also removed. Each protein or protein group was required to have valid reporter intensity values in at least two of the three biological replicates in each experimental group. Proteins that could not be annotated to gene IDs were removed. Finally, only those proteins that passed these filters were subjected to subsequent statistical analyses. To simplify the interpretation of the rather complicated iTRAQ data set, we compared each protein group to only the casein group. A moderated t test implemented in the Bioconductor library18−20 was performed on the VSN (variance stabilization and normalization)-transformed21,22 intensities of proteins. Only those proteins with P < 0.05 were considered as being significantly different.

Upstream Regulator Analysis

The Ingenuity Pathway Analysis (content version 26127183 released November 30, 2015, Ingenuity Systems) was conducted to identify upstream regulators for protein expression changes under different protein diet interventions. For each potential upstream regulator, two statistical measures including an overlap P value and an activation Z score were computed. The overlap P value measures whether there is a statistically significant overlap between the data set proteins and the proteins that are regulated by an upstream regulator. It is calculated by using the Fisher’s Exact Test, and significance is attributed to P < 0.01. The activation Z score is used to determine whether an upstream regulator is significantly activated (Z score >2.0) or inhibited (Z score < −2.0). To clearly see the relation between these predicted upstream regulators and the significant protein sets changed by different dietary proteins, we overlapped the targets of upstream regulators to the proteins in protein sets in the network using Post Analysis tool in the Enrichment Map v2.1.0. Only

Biological Interpretation of Protein Expression Data

KEGG Pathway Overrepresentation Analysis Using Enrichr. To obtain a better biological interpretation for the significantly changed proteins, we conducted a KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway overrepresentation analysis by using Enrichr.23 Only KEGG terms with P < 0.01 getting from Fisher Exact Test were considered as being significantly different. Gene Set Enrichment Analysis (GSEA). It is well accepted that GSEA has multiple advantages over analysis performed on the level of individual genes or proteins.24,25 The GSEA method was initially developed for the gene expression profiling data, but it can also be applied to the proteomics data.24 To avoid causing confusion, we use “protein set” instead 1137


Article

Journal of Proteome Research those upstream regulators having at least five common downstream target genes with protein sets are shown in the network.

Table 1. Enriched KEGG Pathways for the Chicken and Soy Protein Groups KEGG term

Quantitative PCR Verifying Upstream Regulators

P value

Chicken vs Casein_157 Downregulated Proteins fatty acid metabolism 6/45 0.008 butanoate metabolism 7/45 0.002 chicken vs casein_151 upregulated proteins glycolysis and gluconeogenesis 7/63 0.005 valine leucine and isoleucine degradation 6/44 0.005 Soy vs Casein_27 Downregulated Proteins fatty acid metabolism 4/45 0.00023 propanoate metabolism 5/33 9.03 × 10−06 Ppar signaling pathway 5/67 0.00012 pyruvate metabolism 4/41 0.00021 histidine metabolism 4/40 0.00021 soy vs casein_26 upregulated proteins alanine and aspartate metabolism 4/33 0.00031

Four upstream regulators predicted for chicken and soy protein groups, including Pten (phosphatase and tensin homologue), Ppara (peroxisome proliferator-activated receptor alpha), Pparg (peroxisome proliferator-activated receptor gamma), and Ppargc1a (peroxisome proliferator-activated receptor gamma coactivator 1-alpha), were analyzed by Q-PCR array (330231 PARN-14 9ZA, Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Gene amplification was performed on an Applied Biosystems 7500 real-time PCR system (Foster City, CA). Gene expression changes of soy or meat protein groups relative to casein were determined by 2−ΔΔCT method and tested using Student’s t test in SPSS version 16.0 (Chicago, IL). Statistical significance was set at P < 0.05.

■

overlapa

a

Overlap between the proteins in our database and protein members of KEGG terms (e.g., 6/45 means that 6 proteins in our database belong to the KEGG term of fatty acid metabolism which contains totally 45 protein members).

RESULTS

Differentially Expressed Proteins

To understand the effects of isolated soy and meat proteins on liver metabolism regarding protein expression levels, we applied iTRAQ technology in combination with LC−ESI−MS/MS to investigate differentially expressed proteins in liver. In total, 3150 unique proteins (Supporting Information Table S-1) were identified and quantified from 16 842 unique peptides (Supporting Information Table S-2) that were deduced from 147 705 unique MS/MS spectra. After further strict filtering (see “Mass Spectrometric Data Analysis” section for details), 1437 high-quality proteins were retained for the further analysis of differential regulation (Supporting Information Table S-3). Compared with the casein group (reference), the abundance of 308, 53, 10, and 9 proteins was significantly changed by isolated chicken, soy, fish, and pork proteins, respectively (P < 0.05, Supporting Information Table S-4). When presented in a Venn plot (Figure 1A), it became clear that very few proteins were located in the overlapping areas. Most of the significantly regulated proteins in the soy and chicken protein groups were specific for each group.

proteins, and therefore GSEA was performed. The advantage of GSEA is that instead of only focusing on individual proteins, it takes all 1437 proteins into account when evaluating at the level of protein sets, which tends to be more interpretable.24 The number of significantly changed protein sets (pathways) did not reflect the number of individual significant proteins. In total, 41, 36, 28, and 22 protein sets were significantly changed (P < 0.05, Supporting Information Table S-5) by soy, fish, pork, and chicken proteins, respectively. Venn plot showed that soy, fish, and pork protein groups had considerable overlapped protein sets with each other (Figure 1B); however, the chicken protein group almost had no overlapped protein sets with other meat and soy protein groups. Two protein sets were commonly downregulated by all dietary proteins. To improve the interpretation, all significant protein sets were summarized in enrichment maps (Figure 2). Functionally related protein sets were semiautomatically annotated and manually labeled to highlight their prevalent biologic functions. The regulated gene sets identified by GSEA analysis of RNA-seq data were also included in the network and shown as node border to compare the iTRAQ results with our previous RNA-seq results.10,11 A high-resolution map that includes the names of all protein sets is shown in the Supporting Information Figure S-1. In general, the proteomics data revealed that various metabolic processes as well as protein biosynthesis and immune system were regulated. In detail, the cluster of protein sets related to fatty acid oxidation and Pparα signaling pathway were significantly inhibited by all dietary proteins. Protein sets related to glucose and energy metabolism were inhibited by soy (inhibited glycolysis), fish (inhibited electron transport chain), and pork (inhibited TCA cycle, oxidative phosphorylation, and electron transport chain) proteins but were increased by the chicken protein (increased TCA cycle); however, protein sets related to oxidoreductive transformation including biological oxidations and Nrf2 target genes were increased by soy, pork, and fish proteins but were not affected by chicken protein. In addition, protein sets related to AA metabolism were not affected by the chicken protein either but were significantly changed by soy, fish, and pork proteins. Specifically, the metabolism of sulfurcontaining AA and nonessential AA was increased by soy

KEGG Pathway Overrepresentation Analysis Results

To gain better insight into the underlying biologic phenomena that were affected by the differentially regulated proteins, KEGG pathway overrepresentation analysis was conducted using the 308 differentially expressed proteins in the chicken protein group and 53 differentially expressed proteins in the soy protein group. Because of the limited number of significantly regulated proteins this was not done for the pork and fish protein groups. For the chicken protein group, the significantly overrepresented KEGG terms (P < 0.01, Table 1) that were derived from 157 downregulated proteins were related to fatty acid metabolism, while the terms derived from 151 upregulated proteins were related to glucose and branched chain AA metabolism. For the soy protein group, the over-represented terms derived from 27 downregulated proteins were related to fatty acid metabolism, Ppar signaling pathway, and pyruvate metabolism, while the term derived from 26 upregulated proteins was alanine and aspartate metabolism. GSEA Analysis Results

Because the KEGG pathway overrepresentation analysis is only based on individual, significantly regulated proteins, this method was not suitable for the analyses of the pork and fish 1138


Article


Figure 2. Network of protein sets enriched by GSEA analysis. This network was produced by using Cytoscape v3.2.1 and Enrichment Map v2.1.0. Nodes represent the enriched protein set in GSEA analysis of liver proteome (iTRAQ, inner area) in the present study and gene sets in GSEA analysis of liver transcriptome (RNA-seq, border area) in the previous study (Song et al, submitted). The colors of nodes indicate the directions of changes of gene sets with red for up-regulation and blue for down-regulation. The node size is proportional to the total number of proteins within each set (from 15 to 500). The lines between nodes represent the “overlap” score (Jaccard and overlap coefficients >0.375) depending on the number of proteins two protein sets share. Nodes of high similarity were automatically arranged close together, and circles were semiautomatically annotated and manually labeled. A high-resolution map that includes the names of all protein sets is shown in Supporting Information Figure S-1.

protein only; valine, leucine, and isoleucine degradation were inhibited by soy, fish, and pork proteins; arginine and proline metabolism and alanine, aspartate, and glutamate metabolism were inhibited by fish and pork proteins; and lysine degradation was inhibited by the fish protein only. Protein biosynthesis (translation, mRNA processing, and tRNA aminoacylation) was increased by all dietray proteins except by chicken protein. Only soy protein inhibited protein folding. A large cluster of protein sets related to the immune system was reduced by dietary chicken protein only. When comparing the protein expression changes (Figure 2, node inner area) with our previous mRNA expression results (node border), feeding soy or fish protein changed both liver protein and mRNA expression of proteins involved in AA and fatty acid metabolism, oxidoreductive transformation, and mRNA translation. Pork protein increased oxidoreductive transformation and chicken protein reduced Pparα signaling pathway in both protein and mRNA expression levels.

Table 2. Upstream Regulators Predicted for the Chicken and Soy Protein Groups upstream regulator Adipoq Pten Mknk1

Pparg Ppara Nr1i2 Adipoq Srebf2

Upstream Regulators

Ppargc1a

The underlying mechanisms by which the dietary proteins modulated protein expression changes are not well understood. We therefore aimed to identify potential upstream regulators that could explain the observed shifts in protein expression profiles (Table 2). Three regulators were predicted for the chicken protein group, namely, Adipoq (adiponectin, C1Q, and collagen-domain containing), Pten (phosphatase and tensin homologue), and Mknk1 (MAP kinase-interacting serine/ threonine-protein kinase 1). Six regulators were predicted to be inhibited by the soy protein, that is, Pparg (peroxisome proliferator-activated receptor gamma), Ppargc1a (peroxisome proliferator-activated receptor gamma coactivator 1-alpha), Ppara (peroxisome proliferator-activated receptor alpha), Srebf2 (sterol regulatory element-binding protein 2), Nr1i2

molecule type

Z scorea

name

Chicken vs Casein other −2.219 adiponectin, C1Q, and collagen domain containing phosphatase 2.314 phosphatase and tensin homologue kinase 2.449 MAP kinase-interacting serine/ threonine-protein kinase 1 Soy vs Casein ligand-dependent −3.038 peroxisome proliferator-activated nuclear receptor receptor gamma ligand-dependent −2.377 peroxisome proliferator-activated nuclear receptor receptor alpha ligand-dependent −2.213 nuclear receptor subfamily 1, nuclear receptor group I, member 2 other −2.205 adiponectin, C1Q, and collagen domain containing transcription −2.200 sterol regulatory element-binding regulator protein 2 transcription −2.169 peroxisome proliferator-activated regulator receptor gamma coactivator 1-alpha

a

Upstream regulator analysis were done in the Ingenuity Systems (Ingenuity Pathway Analysis, content version 26127183 released November 30, 2015). All upstream regulators were significant at P < 0.01 by the Fisher’s Exact Test. 1: Z score is used to determine whether an upstream regulator is significantly activated (>2.0) or inhibited (5, Figure 2). This showed that Pten, predicted to be an upstream regulator for chicken protein, overlapped the proteins sets related to TCA cycle and Ppara target genes. The regulator Ppara, predicted to be inhibited by the soy protein, overlapped with protein sets related to fatty acid oxidation and Ppar signaling pathway. Q-PCR Verifying Upstream Regulators

Q-PCR analyses revealed that the mRNA expression of Ppargc1, that was predicted to be an inhibited upstream regulator of the soy protein group, was reduced by soy protein (fold change = −1.40) (Figure 3); however, the mRNA

Figure 3. mRNA expressions of four predicted upstream regulators detected by Q-PCR. Fold changes were tested using Student’s t test in SPSS version 16.0 (Chicago, IL). *P < 0.05 vs casein. Values were shown as mean ± SD (n = 4). Ppargc1a: peroxisome proliferatoractivated receptor gamma coactivator 1-alpha; Pparg: peroxisome proliferator-activated receptor gamma; Ppara: peroxisome proliferatoractivated receptor alpha; Pten: phosphatase and tensin homologue.

expression of Pparg was significantly increased by soy protein (fold change = 2.16, P < 0.05), which was opposite to its predicted activity. Expression of two other upstream regulators, Ppara predicted for soy protein group and Pten predicted for the chicken protein group, was not changed (fold change