THE CURRENT PROTEOMIC LANDSCAPE OF THE PORCINE LIVER

Gaining insight into the molecular changes in the proteome of different tissues and organs of farm animals allows to broaden our biological knowledge as well as the basis of disease pathology. Together with adipose tissue and skeletal muscles, the liver is the biggest and the most metabolically active organ. The liver has a very wide range of activities, including metabolic turnover of proteins, lipids and carbohydrates. Metabolic activity of the liver is especially important for protein metabolism. Kampf et al. (1) have demonstrated that 60% (288) of all genes identified by the authors as a liver-specific were responsible for metabolic processes both anabolic and catabolic. These include energy generation, DNA replication and repair, protein synthesis and degradation. The liver is the only organ that is able to synthesize urea and eliminate bilirubin. It is the main organ responsible for amino acids metabolism. The liver is the major site for blood plasma protein synthesis (i.e. albumins, a and b globulins, fibrinogen). Moreover, this organ also produces several enzymes, coagulation factors and bile. The liver also stores vitamins A, D and B12 as well as iron bound to apoferritin.

Uhlen et al. (2) have shown high expression of the liver genes involved in the production of blood plasma proteins, biliary proteins, detoxification proteins, as well as proteins associated with metabolic processes and glycogen storage. Comparative analysis of transcriptomes, protein profiles and specific metabolic pathways of different human tissues have demonstrated that the liver is the most active tissue, even surpassing adipose tissue and skeletal muscles. The authors highlight the large number of genes and metabolic reactions in the liver that do not occur in any other organ or tissue. Many of these reactions are linked to a specialized lipid metabolism, e.g., de novo synthesis and secretion of bile acids, including glycocholic, taurocholic, glycochenodeoxycholic and taurochenodeoxycholic acids. Uhlen et al. (2) also described other metabolic functions specific for the liver, such as ornithine degradation. Moreover, the authors found that the liver (together with the kidney and pancreas) is an organ, where some proteins show an increased levels of expression variability. These authors provided mitochondrial glycine amidinotransferase (GATM) as an example of a protein displaying a higher expression in the liver compared to other tissues.

Application of proteomic tools in studies on liver physiology is becoming increasingly important, especially for elucidating the mechanisms underlying pathophysiology of this organ. Creating a complete and reproducible map representing specific protein patterns is of particular importance in proteomic studies of tissues or body fluids. Reference protein maps are an innovative tool that provides a significant resource for studies aimed at characterizing changes in protein profiles induced by altered physiological conditions, pathological processes or certain experimental factors (3).

Farm animals are increasingly used as a model in proteomic research, which is related to the production of high quality food. The domestic pig is a valuable model for biomedical studies on physiological processes and mechanisms of pathological changes occurring during animal and human diseases. Anatomical, genetic and physiological similarities persuade to select this species for medical and biomedical research (4-6). The liver of Sus scrofa domestica is also an important tissue material for proteomic analyses due to its wide range of applications in pharmacological and toxicological research on drug and xenobiotic metabolism (7). Moreover, this organ is being studied as a potential solution for xenotransplantation in humans (8). This, however, requires a more detailed study on physiological and metabolic similarities between pig and human, also at the molecular level. Recent studies have shown that proteomics is extremely useful for investigating regeneration capacity by combining epigenome, transcriptome and proteome analyses (9).

One of the first studies related to the creation of protein maps of different tissues and body fluids of farm animals were conducted by D’Ambrosio et al. (10). The authors obtained protein maps of the kidney, liver, striated muscles, blood plasma, and erythrocytes in cattle using two-dimensional electrophoresis and mass spectrometry. Golovan et al. (11) performed the first proteomic analysis of porcine liver. These authors identified 880 proteins involved in energy metabolism, protein biosynthesis, electron transport and other oxidation-reduction reactions using isobaric tag for relative and absolute quantification (iTRAQ). The authors suggested that the higher expression of those proteins in the liver, compared with other tissues, was a confirmation of the key role of this organ as “a chemical and energy factory”. Moreover, by comparing the protein composition of human, murine and porcine livers, the authors found that these hepatic proteomes shared 80% of all proteins. Tsujita et al. (12), by means of electrospray-ionization quadrupole time-of-flight mass spectrometry (ESI Q-TOF MS)/MS identified 154 proteins of porcine liver, of which, 109 were specific for this organ. High expression of catalase, cytochrome P450 and ornithine carbamoyltransferase, i.e., proteins located in the cytoplasm with housekeeping functions, was demonstrated.

Systematic creation of proteomic catalogues is required to disseminate proteomic studies of different porcine organs and to improve the usefulness of this species as a model for clinical human studies. Knowledge of protein composition of porcine tissues enables complementing information concerning physiological similarities and differences at the molecular level, between human and pig.

An incomplete database related to peptide masses of porcine proteins is a significant limitation of proteomic studies in domestic pig, and it impedes searching for real sequences of porcine proteins. The homologous protein sequence of other species which is not sufficiently similar, does not provide appropriately high results for protein identification.

The work presented here may complement currently available protein profiles of growing pigs and serve as a reference for monitoring changes in protein expression induced by altered physiological and pathological factors.

MATERIAL AND METHODS

Animals and sample collection

The experiment was carried out on a total of 8 castrated male piglets (PIC × Penarlan P76) kept with sows until weaning at 28 day of life, and subsequently in pens (4 animals per pen) with free access to feed and water. The animals were fed a cereal-based diet from the 10th day of life that contained 20% of crude protein and 14.3 MJ/kg of metabolizable energy. The components of the diet were as follows: wheat (46.84%), barley (20%), corn starch (3%), full-fat soybean meal (5.9%), whey (9.7%), fish meal (4%), spray-dried blood plasma (4%), soybean oil (3.39%), calcium formate (0.3%), limestone (0.5%), dicalcium phosphate (0.6%), sodium chloride (0.07%), L-lysine (0.614%), DL-methionine (0.232%), L-threonine (0.264%), L-tryptophan (0.09%), mineral-vitamin premix (0.4%) and aroma (0.1%). Piglets were sacrificed at the age of 50 days at final body weight of about 20 kg. The livers were removed immediately after sacrifice and washed twice with 0.9% NaCl and thereafter twice with 20 mM Krebs-HEPES buffer (NaCl, KCl, CaCl2, MgSO4, K2HPO4, NaHCO3, pH 7.4). Then, dissected tissue fragments were placed in liquid nitrogen and stored at –80°C until further analysis.

Protein extract preparation

Frozen liver fragments in an amount of 0.1 g per sample were grounded to powder with steel beads in 2 ml tubes using mechanical homogenizer Tissue Lyser (QIAGEN) at a frequency of 22,000 Hz for 5 minutes. Next, samples were homogenized in 1500 µl of lysis buffer containing 7 M urea, 2 M thiourea, 4% w/v CHAPS, 1% w/v DTT, 2% v/v Biolyte, 1% v/v protease inhibitor cocktail (Sigma), 0.1% v/v nuclease (Sigma) at a frequency of 20,000 Hz for 60 minutes. Insoluble tissue debris were removed by centrifugation (4°C, 15,000 g, 15 minutes) and resulted supernatant was used for two-dimensional electrophoresis (2-DE).

Two-dimensional electrophoresis

To ensure reproducibility all 2-D gels were performed in duplicate and were run under the same electrophoresis conditions. Total protein concentration of each sample was estimated by the modified Bradford assay (Bio-Rad Protein Assay, Bio-Rad). Samples containing 800 µg of proteins, were mixed with lysis buffer (7 M urea, 2 M thiourea, 4% w/v CHAPS, 1% w/v DTT, 0.2% w/v 3 – 0 ampholytes) to the total volume of 650 µl and applied to 4 – 7, 24 cm NL (nonlinear) ReadyStrip™ IPG Strips (Bio-Rad). Isoelectrofocusing was run at 20°C by ramping the voltage to a maximum of 5000 V and finished at 90,000 Vh using a Protean® IEF Cell (Bio-Rad). Before the second dimension, the IPG strips were reduced with DTT in equilibration buffer (6 M urea, 0.5 M Tris/HCl, pH 6.8, 2% w/v SDS, 30% w/v glycerol and 1% w/v DTT) for 15 minutes and then alkylated with iodoacetamide (2.5% w/v) for 20 minutes. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed in 12% SDS polyacrylamide gels (20 × 25 cm) at 40 V for 2.5 hours and then at 100 V for 16 hours at 15°C using a Protean Plus™ Dodeca Cell™ electrophoretic chamber (Bio-Rad). After 2-DE separation, the gels were stained with colloidal Coomassie Brilliant Blue G-250 according to Pink et al. (13).

Image analysis

Gel image acquisition was performed using a GS-800™ Calibrated Densitometer (Bio-Rad). 2-D images were analysed with the aid of PDQuest analysis software version 8.0.1 advanced (Bio-Rad). To measure the variability within the group, the coefficient of variation (CV) was calculated. Additionally to compare the CV of technical replication sets CV was calculated for replication set 1 and replication set 2 (duplicates). Next, distribution of spot expression and distribution data were plotted to analyse the coefficient of correlation for both replication sets. Experiment normalization was performed using a local regression model (LOESS). On the basis of the standard 2-D markers the experimental pI and molecular weight (kDa) values were computed for each identified protein spot.

Calculated and theoretical pIs and MWs for each identified protein spot were compared. Linear regression was calculated for pIs and power regression for protein MWs, calibration for both was obtained using coordinates of proteins that showed minimal deviation between their theoretical and experimental pI (0.2 pH) and Mw (5 kDa) as described by Rogowska-Wrzesinska et al. (14). Results were plotted and visually typed outlier spots were marked.

In-gel digestion of proteins

Protein spots were manually excised from two 2-D gels and decolorized by washing in buffer containing 25 mM NH4HCO3 in 5% v/v acetonitrile (ACN), followed by two washes with a solution of 25 mM NH4HCO3 in 50% v/v ACN. The gel pieces were dehydrated (100% ACN), vacuum dried and incubated overnight with trypsin (20 µl/spot of 12.5 µg trypsin/ml in 25 mM NH4HCO3; Promega, Madison, USA) at 37°C.

MALDI-TOF mass spectrometric analysis

A Microflex™ MALDI-TOF (matrix-assisted laser desorption/ionization time of flight) mass spectrometer (Bruker Daltonics, Germany) was operated in positive ion reflector mode. The resulting peptides were extracted with 100% ACN, combined with the matrix solution (5 mg/ml CHCA, 0.1% v/v TFA, 50% v/v ACN) and were loaded onto a MALDI-MSP AnchorChip™ 600/96 plate (Bruker Daltonics, Germany) in a final volume of 1 µl. Droplets were allowed to dry at room temperature. External calibration was performed using Peptide Mass Standard II (Bruker Daltonics) based on monoisotopic values of seven peptides: angiotensin II (1046.5420 Da), angiotensin I (1296.6853 Da), substance P (1347.7361 Da), bombesin (1619.8230 Da), ACTH clip 1-17 (2093.0868 Da), ACTH clip 18-39 (2465.1990 Da) and somatostatin 28 (3147.4714 Da). The mass spectra were acquired with 150 shots of a nitrogen laser operating at 20 Hz and were internally calibrated using porcine tryptic autolytic products (842.51 and 2211.10 m/z) as previously described by Ozgo et al. (15).

Database search

Peptide mass fingerprinting (PMF) data were compared to mammalian databases (SWISS-PROT; http://us.expasy.org/uniprot/ and NCBI; http://www.ncbi.nlm.nih.gov/) with the aid of MASCOT search engine (http://www.matrixscience.com/). Search criteria included: trypsin as an enzyme, carbamidomethylation of cysteine as a fixed modification, methionine oxidation as a variable modification, mass tolerance to 150 ppm and a maximum of one missed cleavage site. Contaminating peaks of keratin and trypsin were removed from the peptide mass list prior database search. The results of PMF-based identification were accepted when the protein score was significant (P < 0.05) with at least five matching peptides and 15% peptide coverage.

Bioinformatic data analysis

In order to obtain a global overview of the molecular functions, gen ontology (GO) biological processes and KEGG biological pathways, the identified proteins were uploaded to the STRING v10 (http://string-db.org/) protein-protein interaction database (16). Adjusted P-value (Benjamini correction) was used to gain a score and to rank our data set.

Euk-mPLoc 2.0. (http://www.csbio.sjtu.edu.cn/bioinf/euk-multi-2/) was used to define the subcellular localisation of the identified liver proteins (17).

InteractiVenn (18) software was used to compare the set of identified gene products with the results of other authors and to visualize the effect of this comparison on Venn diagram. For the gene products specified as non-previously described in the pig liver proteome the analysis of their relative abundance in pig and in human liver were performed using PaxDb: Protein Abundance Database 4.1 (19). The human liver was selected as a reference due to lack of available data for that organ in pigs and according to the anatomical and physiological similarities existing between human and pig.

RESULTS

In the present study, we employed 2-DE and MALDI TOF-based analysis to obtain a global view of the hepatic proteome of 50-day-old piglets. The number of protein spots obtained from all 16 2-D gels (8 gels performed in duplicate) was 470 ± 44. The most prominent spots (420), i.e., those displaying similar locations and staining intensities on each gel were selected for the MS analysis. The analysis showed an average coefficient of variation (CV%) of 49.58. It was estimated for all gels after the LOESS normalization, the average CV for first replicate set was estimated at the level of 49.26 and for the second replicate set (duplicates) at the level of 51.11. Scatter plot comparing two replicate sets is shown in Fig. 1. Correlation between the gel sets was at the level of 0.980677. A total of 265 protein spots (63%), representing 142 unique gene products were positively identified and their position on the 2-D porcine liver map is shown in Figs. 2-6. The majority of these proteins (93.58%) were identified based on Sus scrofa identities. Protein identification was validated when matches had a mascot score greater than 61. Proteins were identified with peptide matches ranging from 6 to 46. More than 90% of proteins were identified by 8-37 peptides. The list of identified proteins, including detailed information on 2-DE and MALDI-TOF coordinates of each individual protein spot is given in Table 1.

Fig. 1. The plot presenting the comparison of expression pattern between the technical replication sets.

Table 1. A summary of all protein spots identified by MALDI-TOF MS analysis in the porcine liver. Values used to create a scatter plot of the molecular weight and pI values distributions are marked with bold.

Abbreviations: CM, cell membrane; CP, cytoplasm; CS, cytoskeleton; EC, extracellular; ER, endoplasmic reticulum; GA, Golgi apparatus; M, mitochondrion; N, nucleus; P, peroxisome; SL, subcellular localization.

Fig. 2. A representative 2-DE gel image compiling all identified porcine hepatic protein spots. 800 µg of proteins were applied to the IPG strip (24 cm, pH 4 – 7) for the first dimension, the second dimension was performed on 12% SDS-PAGE gels and the gels were stained with Coomassie brilliant blue G-250. The 2-DE gel image was divided into 4 panels (A, B, C and D) to highlight the location of identified protein spots.

Fig. 3. Panel presenting an enlarged fragment of the gel image shown in Fig. 1A. Spot numbers correspond to those in Table 1.

2-D gel analysis showed that 87 proteins were expressed as a multiple spots, indicating that they were isoforms, whereas the remaining 58 proteins were resolved as a single spots. Albumin (spots no. 86 – 94) and catalase (spots no. 112 – 120) are examples of a multiple spots, as both were represented by 9 spots, while selenium-binding protein 1 was represented by 4 spots (spots No. 145 – 148). These spots showed slight changes in molecular weight and pI probably resulting from different posttranslational modifications or other potential proteolytic products (precursor versus mature protein form).

The experimental molecular mass (Mr) and pI value were calculated for each identified spot using a PDQuest 8.01 advanced bioinformatic tool and compared to the predicted Mw/pI coordinates accessible in the Uniprot or NCBI databases (Table 1). The distribution of the experimental Mr and pI value are shown in Fig. 7. All coordinates were plotted after calibration of the Mw and pI scales using protein spots that showed minimal deviation between the computed and predicted pI/Mw values to correlate experimental and theoretical spots distribution patterns. The majority of the computed and theoretical pI/Mw coordinates were consistent with a minimal variation. Nevertheless, among all 265 identified liver protein spots, 29 were classified as significant pI outliers (Fig. 7B). The calculated molecular mass (Mr) and pI value of the identified proteins ranged from 19.30 to 139.50 kDa and from 4.30 to 6.90, respectively. It should be noted that the majority of the identified proteins were high-molecular weight proteins (> 30 kDa).

Fig. 4. Panel presenting an enlarged fragment of the gel image shown in Fig. 1B. Spot numbers correspond to those in Table 1.

Fig. 5. Panel presenting an enlarged fragment of the gel image shown in Fig. 1C. Spot numbers correspond to those in Table 1.

Fig. 6. Panel presenting an enlarged fragment of the gel image shown in Fig. 1D. Spot numbers correspond to those in Table 1.

All identified liver proteins were categorized according to their subcellular location and biological and molecular function; in addition, enrichment analysis was performed using annotations from KEGG pathways. Regarding the subcellular distribution, slightly less than a half of the identified proteins were classified as cytoplasmic, while the remaining were assigned to the mitochondrion, endoplasmic reticulum, cytoskeleton, nucleus, extracellular, peroxisome, cell membrane and Golgi apparatus (Fig. 8, Table 1). Biological function analysis revealed that proteins involved in metabolic processes (P = 9.1 e-20), organic substance metabolic process (P = 4.86 e-18) and those engaged in the primary metabolic process (P = 8.63 e-16) were predominant among all the identified porcine hepatic gene products (Fig. 9). We showed using the KEGG pathway analysis that 43 of the identified proteins were significantly associated with metabolism (P = 8.61 e-19), while 16 were found to be involved in protein processing in the endoplasmic reticulum (P = 2.62 e-09) (Fig. 9).

Fig. 7. Comparison of experimentally determined and theoretical pI (A) and Mw (B) values of the identified protein spots from the pig liver 2-D protein maps. Spots used for the curve preparation and other plotted spots are marked with black rhombs and grey squares, respectively. Spots found as an outliers are circled with dashed line. Number of protein spots correspond to those in Table 1, Figs. 3, 4, 5 and 6.

Fig. 8. Diagram presenting the percentage of the identified on 2-D maps of porcine liver gene products distribution according to their subcellular localization based on the Euk-mPLoc 2.0 bioinformatic tool.

Fig. 9. Diagram representing GO analysis of biological processes and KEGG pathway enrichment. The top fifteen significant terms with the highest protein counts are shown.

DISCUSSION

The implementation of 2DE and MALDI TOF MS allowed us to establish Sus scrofa liver proteome map. Until now, two proteome maps of porcine liver have been published by Caperna et al. (20) and Golovan et al. (11), where the authors reported the identification of 282 and 880 unique proteins, respectively. The present study was conducted to establish a 2-DE map of porcine liver proteins. We identified 265 proteins using 2-DE and MALDI-TOF that corresponded to 142 distinct gene products. Proteins were positively identified at the identification rate of 63%. Similar results (53.4%) were obtained by Caperna et al. (20), who used MALDI-TOF MS as an initial screening tool to identify porcine hepatocyte proteins. Of all proteins identified by Caperna et al. (20), only 167 (47.9%) were annotated based on Sus scrofa identities. In our study, the majority of proteins (93.58%) were identified based on Sus scrofa identities. This indicates significant progress in the development of porcine protein database in the last decade.

Among successfully identified protein spots, 29 were classified as spots with significant pI shifts. Those protein spots represented 17 distinct gene products. Most gene products with a significantly shifted pI coordinates were of mitochondrial origin, namely: glycine amidinotransferase, mitochondrial (GATM), medium-chain specific acyl-CoA dehydrogenase, mitochondrial (ACADM), short-chain specific acyl CoA dehydrogenase, mitochondrial (ACADS), short/branched chain specific acyl-CoA dehydrogenase, mitochondrial (ACADB), isocitrate dehydrogenase (NADP), mitochondrial (IDH2), ornithine carbamoyltransferase, mitochondrial (OTC) and glutamate dehydrogenase 1, mitochondrial (GLUD1). The analysis of the GATM, ACADM, ACADS and IDH2 amino acid sequences using information available in the databases, such as Uniprot and NCBI showed that the predicted pI values corresponded to the sequence that contained mitochondrion transit peptide (mtP). The analysis of protein sequence of haptoglobin (HP), showed that it also contains the 18 aa signal peptide. Most probably, the observed pI shifts of other mitochondrial proteins were of the same nature, since their sequences had a status of unreviewed/predicted protein in the Uniprot/NCBI databases and their presence and sequences were predicted based on cDNA sequences and did not include any PTM information of those proteins. Additionally, the TargetP mitochondrial transition peptide predictor (21) confirmed the existence of mtP in the sequence of GLUD1, OTC and ACADB. Virtual cleavage of the predicted 53-aa, 32-aa, 18-aa and 12-aa peptides for GLUD1, OTC, HP and ACADB, respectively, shifted the predicted pI values towards the computed ones. The KHK (accession No. NP_001230537) and LHPP (XP_003125348) proteins were identified as specific isoforms, but the analysis of other known isoforms of those proteins with high sequence similarity to the identified proteins showed that their pIs and molecular masses closely correspond to the values calculated for those proteins in our study, respectively: 6.66 and 26 kDa for KHK (XP_020943371.1) and 6.2 and 23 kDa for LHPP (XP_020927622.1). The observed pI shifts for both cytoplasmic proteins, GST-class alpha and PBLD (phenazine biosynthesis-like domain-containing protein), were most likely the results of known PTMs, such modifications changes protein pI to more acidic. The moderate changes in the pI/Mw values observed for other protein spots may reflect the occurrence of other posttranslational modifications, such as glycosylation, deamination or proteolytic cleavage, which are mainly responsible for charge and protein size modifications, and in consequence, may influence their electrophoretic mobility (14, 17, 22, 23). In case of protein spots identified as Jaculus Jaculus HSPD1, we have found that its homologue for Sus scrofa has isoelectric point and molecular mass 5.71 and 60.1 kDa which is close to the values calculated for both spots identified in our study.

As the most complex organ in the body, the liver has a number of physiological functions. The liver shows extreme metabolic activity and governs body energy metabolism by connecting carbohydrate, lipid and protein metabolic pathways (24). Moreover, the liver is involved in production of bile, storage of glycogen, as well as detoxification of harmful substances such as ammonia, xenobiotics and drugs (1, 11). Here, we demonstrate hepatic expression of proteins involved in metabolic processes specific for this organ. We observed the expression of proteins involved in carbohydrate metabolism, including processes of glycolysis and gluconeogenesis (ENO1, ALDH2, ALDH9A1, PFKFB3) and metabolism of sugars other than glucose (GALK1, KHK). Several of proteins identified in this study are engaged in lipid metabolism including fatty acid synthesis and degradation (ACADM, ACADS, ACADSB, ALDH2, PCCA), cholesterol (APOA1, MVK) and bile acid metabolism (AKR1D1, BAAT). Moreover, we demonstrated the expression of several proteins that participate in amino-acid metabolism, including metabolism of arginine and proline (AGMAT, ALDH9A1, ARG1, GATM, LAP3), glycine, serine and threonine (BHMT, GATM, SDSL), alanine, aspartate and glutamate (GLUD1, GLUL) and sulfur amino acid metabolism (SUOX). We also identified proteins associated with drug/xenobiotics biodegradation and metabolism (UPB1, TPMT, INMT, HPRT1, GSTO1, CES1, AS3MT, APLE), as well as CPS1 involved in removing excess ammonia from the cell (25).

Protein spots identified in the present study, representing 142 different gene products, were compared with the gene products identified in all available proteomic studies using 2-DE and MS approaches performed on porcine liver (8, 12, 20, 26-33). The results of this comparison are presented as a Venn diagram (Fig. 10). It was found that 26 out of 142 gene products identified in the present study have never been detected on the proteomic maps representing sus scrofa liver proteome. The unique gene products were categorized according to their contribution to the known biological processes, such as: cytoskeleton organization; stress response and detoxification, secondary messenger, energetic metabolism, proteasome structure, vesicular transport as described in Table 2. The abundance analysis indicated that most of the proteins not detected previously on the 2-DE hepatic proteome maps are high abundant liver proteins, but we also noticed proteins showing very low or low expression in the human liver, including: indolethylamine N-methyltransferase - 0.24 ppm; olfactomedin-like protein 3 - 7.14 ppm; dipeptyl peptidase 9 - 9.32 ppm; arsenite methyltransferase - 10 ppm. That may be the effect of different expression of those proteins in liver between human and pigs. We are sure that a set of data of previously not described on 2-DE maps porcine proteins may be of considerable value in future studies aimed at understanding the molecular basis of this organ biology and pathophysiology.

Fig. 10. Venn diagram presenting comparison between the gene products identified in the present study and these that have been previously published using 2-DE and MS or LC-MS approaches performed in pig liver. The particular number of the identified gene products are given before the following references: Caperna et al., 2008, Golovan et al., 2008, Lee et al., 2011, Wang et al., 2011, Takeda et al., 2012, Bondzio et al., 2013, Sejersen et al., 2013, Yi et al., 2013, Grubbs et al., 2014, Li et al., 2017, Bovo et al., 2018, the reference numbers are given in square brackets. * marks the number of gene products that gives a set of individual gene products after the substraction of 3 proteins that have been present in more than one subsets analyzed as one group.

Table 2. Identified gene products previously not reported in other proteomic studies of liver proteome in horses with uniport/ncbi database description and relative abundance with rank per number of gene products available for the analysed subset 2959 for pigs body and 14774 for human liver in PaxDb: Protein Abundance Database 4.1 (Wang et al., 2015). With * the gene product that belong to the two different groups of proteins (Energetic metabolism; Stress response and detoxification) have been marked.

ND, no data available.

In conclusion, we have established a reproducible 2-D map of porcine liver proteome that represents 142 unique gene products, 26 of which have been confirmed for the first time in pig liver proteomes. Its applicability in the present form was confirmed by our group in nutriproteomic studies, in which we proved the effects of dietary factors: inulin (34) and rich in prebiotics chicory root (35) on the liver protein expression patterns, manifested by alternations in abundance of proteins involved in energetic metabolism, stress prevention and cell architecture. This provides a solid basis to believe that the established tool represents a useful and necessary prerequisite for further research on the liver response at the protein level to various stimuli affecting the organ, e.g. toxic factors, dietary factors, medical agents and various physiological and pathophysiological conditions. Besides the above-mentioned benefits, systematic cataloguing of the protein constituents of porcine liver might also be of great importance to farm animal production science as well as animal health and welfare.

Acknowledgements: This work was supported by grant from National Centre of Science, Poland (Project No. 2012/05/D/NZ9/01604).

Conflict of interest: None declared.

REFERENCES

1. Kampf C, Mardinoglu A, Fagerberg L, et al. The human liver-specific proteome defined by transcriptomics and antibody-based profiling. FASEB J 2014; 28: 2901-2914.
2. Uhlen M, Fagerberg L, Hallstrom BM, et al. Proteomics. tissue-based map of the human proteome. Science 2015; 347: 1260419. doi: 10.1126/science.1260419
3. Gromotowicz-Poplawska A, Kasprzyk J, Marcinczyk N, Stepien E, Piekoszewski W, Chabielska E. The proteome analysis of rat platelet with nano-liquid chromatography-matrix assisted laser desorption/ionization time-of-flight mass spectrometry technique. J Physiol Pharmacol 2018; 69: 963-968.
4. Almeida AM, Bendixen E. Pig proteomics: a review of a species in the crossroad between biomedical and food sciences. J Proteom 2012; 75: 4296-4314.
5. Zuwala-Jagiello J, Simon K, Kukla M, et al. Increased circulating endocan in patients with cirrhosis: relation to bacterial infection and severity of disease. J Physiol Pharmacol 2017; 68: 273-282.
6. Ferenc K, Pietrzak P, Wierzbicka M, et al. Alternations in the liver of intrauterine growth retarded piglets may predispose to development of insulin resistance and obesity in later life. J Physiol Pharmacol 2018; 69: 211-218.
7. Amiri-Dashatan N, Koushki M, Abbaszadeh HA, Rostami-Nejad M, Rezaei-Tavirani M. Proteomics applications in health: biomarker and drug discovery and food industry. Iran J Pharm Res 2018; 17: 1523-1536.
8. Yi SS, Oh SJ, Kim IY, et al. Proteomic analysis of liver in miniature pigs according to developmental stages using two-dimensional electrophoresis and matrix-assisted laser desorption/ionization-time of flight mass spectrometry. Lab Anim Res 2013; 29: 162-167.
9. Sato Y, Katoh Y, Matsumoto M, et al. Regulatory signatures of liver regeneration distilled by integrative analysis of mRNA, histone methylation, and proteomics. J Biol Chem 2017; 292: 8019-8037.
10. D’Ambrosio C, Arena S, Talamo F, et al. Comparative proteomic analysis of mammalian animal tissues and body fluids: bovine proteome database. J Chromatogr B 2005; 815: 157-168.
11. Golovan SP, Hakimov HA, Verschoor CP, et al. Analysis of Sus scrofa liver proteome and identification of proteins differentially expressed between genders, and conventional and genetically enhanced lines. Comp Biochem Physiol Part D Genomics Proteomics 2008; 3: 234-242.
12. Tsujita T, Kang D, Moon MH, et al. Large-scale identification by shotgun proteomics of proteins expressed in porcine liver and salivary gland. Zool Sci 2008; 25: 129-138.
13. Pink M, Verma N, Rettenmeier AW, Schmitz-Spanke S. CBB staining protocol with higher sensitivity and mass spectrometric compatibility. Electrophoresis 2010; 31: 593-598.
14. Rogowska-Wrzesinska A, Bihan MC, Thaysen-Andersen M, Roepstorffm P. 2D gels still have a niche in proteomics. J Proteom 2013; 88: 4-13.
15. Ozgo M, Lepczynski A, Herosimczyk A. Two-dimensional gel-based serum protein profile of growing piglets. Turk J Biol 2015; 39: 320-327.
16. Szklarczyk D, Franceschini A, Wyder S, et al. STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res 2015; 43: D447-D452.
17. Rabilloud T, Lelong C. Two-dimensional gel electrophoresis in proteomics: a tutorial. J Proteom 2011; 74: 1829-1841.
18. Heberle H, Meirelles GV, da Silva FR, Telles GP, Minghim R. InteractiVenn: a web-based tool for the analysis of sets through Venn diagrams. BMC Bioinformatics 2015; 16: 169. doi: 10.1186/s12859-015-0611-3
19. Wang M, Herrmann CJ, Simonovic M, Szklarczyk D, von Mering C. Version 4.0 of PaxDb: Protein abundance data, integrated across model organisms, tissues, and cell-lines. Proteomics 2015; 15: 3163-3168.
20. Caperna TJ, Shannon AE, Garrett WM. A gel-based reference map of the porcine hepatocyte proteome. Domest Anim Endocrinol 2008; 35: 142-156.
21. Nielsen H, Engelbrecht J, Brunak S, von Heijne G. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng 1997; 10: 1-6.
22. Bouley J, Chambon C, Picard B. Mapping of bovine skeletal muscle proteins using two-dimensional gel electrophoresis and mass spectrometry. Proteomics 2004; 4: 1811-1824.
23. Lepczynski A, Ozgo M, Dratwa-Chalupnik A, et al. An update on medium- and low-abundant blood plasma proteome of horse. Animal 2018; 12: 76-87.
24. Rui L. Energy metabolism in the liver. Compr Physiol 2014; 4: 177-197.
25. Chen Z, Tang N, Wang X, Chen Y. The activity of the carbamoyl phosphate synthase 1 promoter in human liver-derived cells is dependent on hepatocyte nuclear factor 3-beta. J Cell Mol Med. 2017; 21: 2036-2045.
26. Lee G, Han D, Song JY, Kim JH, Yoon S. Proteomic analysis of swine hepatitis E virus (sHEV)-infected livers reveals upregulation of apolipoprotein and downregulation of ferritin heavy chain. FEMS Immunol Med Microbiol 2011; 61: 359-363.
27. Wang J, Jiang J, Zhang H, et al. Integrated transcriptional and proteomic analysis with in vitro biochemical assay reveal the important role of CYP3A46 in T-2 toxin hydroxylation in porcine primary hepatocytes. Mol Cell Proteomics 2011; 10: M111.008748. doi: 10.1074/mcp.M111.008748
28. Takeda K, Tasai M, Iwamoto M, et al. Comparative proteomic analysis of liver mitochondrial proteins derived from cloned adult pigs reconstructed with Meishan pig fibroblast cells and European pig enucleated oocytes. J Reprod Dev 2012; 58: 248-253.
29. Bondzio A, Pieper R, Gabler C, et al. Feeding low or pharmacological concentrations of zinc oxide changes the hepatic proteome profiles in weaned piglets. PLoS One 2013; 8: e81202. doi: 10.1371/journal.pone.0081202
30. Sejersen H, Sorensen MT, Larsen T, Bendixen E, Ingvartsen KL. Liver protein expression in young pigs in response to a high-fat diet and diet restriction. J Anim Sci 2013; 91: 147-158.
31. Grubbs JK, Huff-Lonergan E, Gabler NK, M Dekkers JC, Lonergan SM. Liver and skeletal muscle mitochondria proteomes are altered in pigs divergently selected for residual feed intake. J Anim Sci 2014; 92: 1995-2007.
32. Li Y, Mou Y, Thunders M, et al. Effects of enrofloxacin on antioxidant system, microsomal enzymatic activity and proteomics in porcine liver. J Vet Pharmacol Ther 2018; 41: 562-571.
33. Bovo S, Di Luca A, Galimberti G, Dall’Olio S, Fontanesi L. A comparative analysis of label-free liquid chromatography-mass spectrometry liver proteomic profiles highlights metabolic differences between pig breeds. PLoS One 2018; 13: e0199649. doi: 10.1371/journal.pone.0199649
34. Herosimczyk A, Lepczynski A, Ozgo M, et al. Hepatic proteome changes induced by dietary supplementation with two levels of native chicory inulin in young pigs. Livest Sci 2017; 203: 54-62.
35. Lepczynski A, Herosimczyk A, Ozgo M, et al. Dietary chicory root and chicory inulin trigger changes in energetic metabolism, stress prevention and cytoskeletal proteins in the liver of growing pigs – a proteomic study. J Anim Physiol Anim Nutr (Berl) 2017; 101: e225-e236.