Comparative proteomics provides insights on

Journal of Proteome Research

This document is confidential and is proprietary to the American Chemical Society and its authors. Do not copy or disclose without written permission. If you have received this item in error, notify the sender and delete all copies.

Comparative proteomics provides insights on metabolic responses in rat liver to isolated soy and meat proteins

Journal: Manuscript ID Manuscript Type: Date Submitted by the Author: Complete List of Authors:

Journal of Proteome Research pr-2015-009225.R1 Article n/a Song, Shangxin; Nanjing Agricultural University Hooiveld, Guido; Wageningen University, Nutrition, Metabolism and Genomics Group, Division of Human Nutrition Zhang, Wei; Nanjing Medical University, Key Laboratory of Human Function Genomics Jiangsu Province Li, Mengjie; Nanjing Agricultural University Zhao, Fan; Nanjing Agricultural University Zhu, Jing; Nanjing Agricultural University Xu, Xinglian; Nanjing Agricultural University, College of Food Science and Technology Muller, Michael; University of East Anglia, Norwich Li, Chunbao; Nanjing Agricultural University, Zhou, Guanghong; Nanjing Agricultural University,

ACS Paragon Plus Environment

Page 1 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Comparative proteomics provides insights on metabolic responses in rat liver to isolated soy and meat proteins Shangxin Song1, Guido J. Hooiveld2, Wei Zhang4, Mengjie Li1, Fan Zhao1, Jing Zhu1, Xinglian Xu1, Michael Muller3, Chunbao Li1*, Guanghong Zhou1*

1

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

2

Nutrition, Metabolism and Genomics Group, Division of Human Nutrition, Wageningen University, Wageningen, the Netherlands

3

4

Norwich Medical School, University of East Anglia Norwich

Key Laboratory of Human Function Genomics Jiangsu Province, Nanjing Medical University, Nanjing, 210029, P.R. China

1



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 33

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. Since these may differ from gene expression responses, we investigated dietary protein-induced changes in liver proteome. Rats were fed for 1 week semi-synthetic 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 1,437 unique proteins was identified and subjected to differential protein analysis and biological interpretation. Compared to 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 metabolism of sulfur-containing and non-essential 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.

2


Page 3 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


KEYWORDS: metabolic syndrome, isolated protein, animal protein, chicken protein, fish protein, pork protein, molecular nutrition, nutrigenomics, proteomics

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 33

INTRODUCTION Metabolic syndrome is becoming a global epidemic which increases risks of cardiovascular diseases and type 2 diabetes1. 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 sources7. Meat and soy are the major dietary protein sources for human nutrition. Compared to plant protein, meat protein distinguishes itself for its richness in all the essential amino acids (AAs)8. However, in contrast to 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 different9, and short-term feeding of dietary proteins from different sources resulted in distinct physiological and transcriptome changes in young rats10,

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 semi-synthetic 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

4


Page 5 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


compared with our previous RNA-seq data using gene set enrichment analysis (GSEA).

EXPERIMENTAL SECTION 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 described10,

11

. Briefly, after a one-week

adaption period, 4 wk-old male Sprague Dawley rats were fed for 7 days the nutritionally balanced semi-synthetic 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 papers13, 14. Briefly, 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 5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 33

°C, 30,000 g 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 8.0-8.5). After sonication and centrifugation, the supernatant was transferred to new tube. To reduce the disulfide bonds in proteins of the supernatant, 10 mM DTT was added and incubated at 56 °C for 1 h. 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 0.5 M TEAB (Applied Biosystems, Milan, Italy), and sonicated at 200 W for 15 min. Finally, samples were centrifuged at 30,000 g (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, USA) 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). Briefly, one unit of iTRAQ reagent was thawed and reconstituted in 24 6


Page 7 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


µL isopropanol. Samples were labeled with the iTRAQ tags as follow: casein group (114 tag), soy protein group (116 tag), pork protein group (117 tag), chicken protein group (119 tag), 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 centrifugation. As there were three liver samples for each group, there was a total of three peptides mixtures. To reduce sample complexity, the Strong Cation Exchange (SCX) chromatography was used in the fractionation of iTRAQ labeled peptides. SCX chromatography was performed with a LC-20AB HPLC Pump system (Shimadzu, Kyoto, Japan). Firstly, the peptide mixtures were reconstituted in 4 mL 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, 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 20,000 g for 10 min, the average final concentration of peptide was about 0.5 µg/µL. Supernatant (10 7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 33

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 %, and then maintenance at 80 % B for 4 min, and finally returning to 5 % in 1 min. The peptides were subjected to nano electrospray 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 to 2,000 Da. For MS2 scans, the m/z scan range was 100-1,800. Mass Spectrometric Data Analysis. Raw mass spectrometric data acquired from the Orbitrap were analyzed in MaxQuant version 1.5.2.815. In total 948,072 MS/MS 8


Page 9 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


spectra from raw files were searched against the UniProtKB Rattus Norvegicus (2014.11 release, 34,165 entries) using the Andromeda search engine16. 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 recalibration17. Mass tolerances for the main search and isotope match were 4.5 ppm 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), and fixed modification was Carbamidomethyl (C). Minimal peptide length was set to 7 AAs and a maximum of two mis-cleavages were allowed. The second peptide option was activated to enable identification of co-eluting peptides with very similar mass16. For protein identification, minimal 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 less than two peptides were also removed. Each protein or protein group was required to have valid reporter intensity values in at least 2 of the 3 biological replicates in each experimental group. Proteins that could not be annotated to gene 9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 33

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 dataset, each protein group was only compared to the casein group. A moderated t-test implemented in the Bioconductor library18-20 was performed on the VSN (variance stabilization and normalization)21, 22 transformed intensities of proteins. Only those proteins with P < 0.05 were considered as being significantly different. Biological Interpretation of Protein Expression Data KEGG Pathway Overrepresentation Analysis Using Enrichr. In order to obtain a better biological interpretation for the significantly changed proteins, a KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway overrepresentation analysis was conducted by using Enrichr23. 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 proteins24, 25. The GSEA method was initially developed for the gene expression profiling data, but it can also be applied to the proteomics data24. To avoid causing confusion, we use “protein set” instead 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, 10


Page 11 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


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 1,000 permutations. In order to improve the interpretation of the GSEA results, significant protein sets (P < 0.05) were visualized 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. To compare the changes observed in the present iTRAQ data set with the previous RNA-seq data set, the gene sets from the GSEA results of the previous RNA-seq data (submitted, Song et al.) were also shown in the network as node border area. Upstream Regulator Analysis The Ingenuity Pathway Analysis (content version 26127183 released 30 November 2015, Ingenuity Systems) was conducted to identify upstream regulators for protein expression changes under different protein diet interventions. For each potential upstream regulators, 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 dataset proteins and the proteins that are regulated by an upstream regulator. It is calculated by using the Fisher’s Exact Test, 11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 33

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, the targets of upstream regulators were overlapped to the proteins in protein sets in the network using Post Analysis tool in the Enrichment Map v2.1.0. Only those upstream regulators having at least 5 common downstream target genes with protein sets are shown in the network. Quantitative PCR Verifying Upstream Regulators Four upstream regulators predicted for chicken and soy protein groups, including Pten (phosphatase and tensin homolog), 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-∆∆CT method and tested using Student's t-test in SPSS version 16.0 (Chicago, IL). Statistical significance was set at P < 0.05.

RESULTS 12


Page 13 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Differentially Expressed Proteins In order to understand the effects of isolated soy and meat proteins on liver metabolism regarding protein expression levels, iTRAQ technology in combination with LC-ESI-MS/MS was applied to investigate differentially expressed proteins in liver. In total, 3,150 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), 1,437 high quality proteins were retained for the further analysis of differential regulation (Supporting Information Table S-3). Compared to the casein group (reference), the abundance of 308, 53, 10 and 9 proteins were 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. 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. Due to the limited number of significantly regulated proteins this was not done for the pork and fish 13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 33

protein groups. For the chicken protein group, the significantly overrepresented KEGG terms (P < 0.01, Table 1) which 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 Since 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 proteins, and therefore GSEA was performed. The advantage of GSEA is that instead of only focusing on individual proteins, it takes all 1,437 proteins into account when evaluating at the level of protein sets, which tends to be more interpretable24. 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 14


Page 15 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


by all dietary proteins. To improve the interpretation, all significant protein sets were summarized in Enrichment maps (Figure 2). Functionally related protein sets were semi-automatically annotated and manually labeled to highlight their prevalent biologic functions. To compare the iTRAQ results with our previous RNA-seq results10, 11, the regulated gene sets identified by GSEA analysis of RNA-seq data were also included in the network and shown as node border. A high-resolution map that includes 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, metabolism of sulfur-containing AA and non-essential AA were increased by soy protein only; valine, leucine and isoleucine degradation were inhibited by soy, fish and pork 15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 33

proteins; arginine and proline metabolism and alanine, aspartate and glutamate metabolism were inhibited by fish and pork proteins; lysine degradation was inhibited by the fish protein only. Protein biosynthesis (translation, mRNA processing and tRNA aminoacylation) were 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 were 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. Upstream Regulators 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

homolog)

and

Mknk1

(MAP

kinase-interacting

serine/threonine-protein kinase 1). Six regulators were predicted to be inhibited by the soy protein, i.e. Pparg (peroxisome proliferator-activated receptor gamma), Ppargc1a 16


Page 17 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(peroxisome proliferator-activated receptor gamma coactivator 1-alpha), Ppara (peroxisome proliferator-activated receptor alpha), Srebf2 (sterol regulatory element-binding protein 2), Nr1i2 (nuclear receptor subfamily 1, group I, member 2) and Adipoq. No potential upstream regulators were predicted for the pork and fish protein groups due to their limited number of regulated proteins. To visualize the relation between upstream regulators and the significant changed protein sets, regulators were overlapped to the protein sets in the networks (number of common gene > 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 expression of Pparg was significantly increased by soy protein (fold change = 2.16, P 2.0) or inhibited (< -2.0). 27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 33

FIGURE CAPTIONS

Figure 1. Venn plots of proteins (A, P < 0.05 by moderated t-test) and protein sets (B, 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.

Figure 2. The 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 enriched protein set in GSEA analysis of liver proteome (iTRAQ, inner area) in present study and gene sets in GSEA analysis of liver transcriptome (RNA-seq, border area) in 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 semi-automatically annotated and manually labelled. A high-resolution map that includes names of all protein sets is shown in the Supporting Information Figure S-1.

28


Page 29 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


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 proliferator-activated receptor gamma coactivator 1-alpha; Pparg: peroxisome proliferator-activated receptor gamma; Ppara: peroxisome proliferator-activated receptor alpha; Pten: phosphatase and tensin homolog.

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42

Page 30 of 33

Figure 1 B. Protein Set (GSEA)

A. Protein

27 26 Soy vs. C

3 6 Pork vs. C 0 0 0 3

157 151 4 Chicken vs. C 6 141 Fish vs. C 142 1 0 2

0

12 14

17 24 Soy vs. C

2 0

4 4 1 2

1 1

0 0

15 13 Pork vs. C

6 9

Down Up

3 0

0 0

0 0

16 2

2 6

2 0 0 2

0 0

1 0 Down Up ACS Paragon Plus Environment

0 0

3 8

15 21 Fish vs. C

0 0 0 1

0 0

0 1

10 6

19 3 Chicken vs. C

4 5

4 1

Page 31 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42


Figure 2 Soy vs Casein

Pork vs Casein

Oxidoreductive Transformation Glycolysis

Oxidoreductive Transformation

Fatty Acid Oxidation PPARA

Amino Acid Metabolism

Energy Metabolsm

Chicken vs Casein Fatty Acid Oxidation

Fish vs Casein

Fatty Acid Oxidation PTEN




Energy Metabolsm

Fatty Acid Oxidation


Immune System mRNA Processing

mRNA Processing Translation

tRNA Aminoacylation

Translation tRNA Aminoacylation

Protein Folding Other RNA-seq iTRAQ

Other Upstream Regulator

Up

Down


tRNA Aminoacylation Cell Junction

Other

Protein Cell Folding Junction

Other


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

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 proliferator-activated receptor gamma coactivator 1-alpha; Pparg: peroxisome proliferator-activated receptor gamma; Ppara: peroxisome proliferator-activated receptor alpha; Pten: phosphatase and tensin homolog. 106x66mm (300 x 300 DPI)


Page 32 of 33

Page 33 of 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58


Table of Contents (TOC)/Abstract (ABS) Graphic

MILK

casein

VS isolated protein sources Proteome: iTRAQ & LC-ESI-MS/MS

Soy vs Casein

Pork vs Casein

Oxidoreductive Transformation Glycolysis


Fatty Acid Oxidation PPARA

AA Metabolism

Energy Metabolsm

Chicken vs Casein Fatty Acid Oxidation

Fish vs Casein

Fatty Acid Oxidation PTEN

AA Metabolism

AA Metabolism


Energy Metabolsm

Fatty Acid Oxidation

AA Metabolism

Immune System mRNA Processing

mRNA Processing Translation

tRNA Aminoacylation

Translation tRNA Aminoacylation

tRNA Aminoacylation

Protein Folding Other RNA-seq iTRAQ

Other

Cell Junction

Other

Protein Cell Folding Junction

Other

Upstream Regulator Up

Down ACS Paragon Plus Environment

Photograph courtesy of ‘Shangxin Song’. Copyright 2016.