Abstract

Microsomes are derived mostly from endoplasmic reticulum and are an ideal target to investigate compound metabolism, membrane-bound enzyme functions, lipid-protein interactions, and drug-drug interactions. To better understand the molecular mechanisms of the liver and its diseases, mouse liver microsomes were isolated and enriched with differential centrifugation and sucrose gradient centrifugation, and microsome membrane proteins were further extracted from isolated microsomal fractions by the carbonate method. The enriched microsome proteins were arrayed with two-dimensional gel electrophoresis (2DE) and carbonate-extracted microsome membrane proteins with one-dimensional gel electrophoresis (1DE). A total of 183 2DE-arrayed proteins and 99 1DE-separated proteins were identified with tandem mass spectrometry. A total of 259 nonredundant microsomal proteins were obtained and represent the proteomic profile of mouse liver microsomes, including 62 definite microsome membrane proteins. The comprehensive bioinformatics analyses revealed the functional categories of those microsome proteins and provided clues into biological functions of the liver. The systematic analyses of the proteomic profile of mouse liver microsomes not only reveal essential, valuable information about the biological function of the liver, but they also provide important reference data to analyze liver disease-related microsome proteins for biomarker discovery and mechanism clarification of liver disease.

1. Introduction

The liver, a vital organ, has a wide range of physiological functions and plays a major role in metabolism, biosynthesis, and chemical neutralizing. Liver diseases, such as viral hepatitis and liver cancer, pose a worldwide public health challenge. The Human Liver Proteome Project (HLPP) was launched in 2002 to better understand molecular liver functions and diseases, and liver proteome expression profile is one of the major parts of HLPP [1]. Because of the complexity, no single proteomic analysis strategy can sufficiently address all components of a proteome. Analysis of the subcellular proteome would provide insight into the functions of a given tissue or cell line. Subcellular proteomics reduces the complexity of a proteome [2, 3], detects some low-abundance proteins, and offers more detailed information that would contribute to the understanding of the function of the entire proteome.

Microsomes are composed primarily of closed sacs of membrane called vesicles that are derived mostly from endoplasmic reticulum (ER). As for liver, in addition to components of the protein secretary pathway, microsomes contain a multitude of proteins that are involved in lipid/lipoprotein biosynthesis and drug metabolism. The liver microsome is an ideal way to study the metabolism of compounds, the functional properties of membrane-bound enzymes, lipid-protein interactions, and drug-drug interactions [4, 5]. The proteomic profiling of the microsomes combined with bioinformatics analysis can reveal more essential information about the biological function of the liver. The main goal of this study was to systematically identify the protein components of the liver microsomes, to conduct the functional annotation with bioinformatics analysis, and to provide insight into the biological functions of the liver.

Two-dimensional gel electrophoresis (2DE) is one of the most widespread techniques for the proteomic profiling of soluble proteins and visualizes isoforms and posttranslational modifications in a proteome [6, 7]. Membrane proteins, however, are less amenable to solubilization in protein extraction buffers and are also susceptible to precipitation during isoelectric focusing (IEF) because of their hydrophobicity and alkaline pH value. One study showed that the analytical performance of one-dimensional gel electrophoresis (1DE) that separates endoplasmic reticulum membrane proteins is incomparably greater than that of 2DE [8]. Other studies [7, 9] demonstrated that the proteomic analysis of subcellular organelles, such as microsomes that contain a considerable number of highly hydrophobic membrane proteins, should be performed by combining 1DE and 2DE.

Although many of microsome proteins have been studied, many more remain to be isolated and characterized. With the improvement of current methodologies and experimental techniques, more proteomic data will be obtained. Also, biological interpretation of proteomic data and extracting biological knowledge are essential to further understanding liver function.

In our study, 2DE was first used to array the isolated microsome proteins of the liver. Because of the low performance of 2DE in separating membrane proteins [10] and the high efficiency of the carbonate procedure in separating membrane proteins [11, 12], the membrane proteins from Na2CO3-treated microsomes were separated by 1DE. Moreover, bioinformatics analysis of microsome proteomic data was performed to discover biological roles of the proteins. The results showed that the combination of 1DE and 2DE was more efficient for analyzing microsomes. Bioinformatics analysis can provide a valuable molecular basis to interpret the mechanisms underlying microsome biological functions and give insight into the biological function of the liver at the level of microsomes.

2. Material and Methods

2.1. Animals

Male C57 mice (9 weeks old) were purchased from the Experimental Animal Center of Central South University (Changsha, China). The mice were starved overnight for liver subcellular fractionation. All experiments were performed with the approval of the institutional ethics committee on animal research.

2.2. Preparation, Validation, and 2DE Analysis of Microsomes
2.2.1. Preparation of Microsomes

Microsome apparatus-rich fractions were prepared from mice livers with differential centrifugation and sucrose gradient centrifugation as described [13]. Mice livers (approximately 10 g each) were drained of blood, minced thoroughly with scalpels, and transferred to 50 mL of chilled homogenization medium (0.25 M sucrose, pH 7.4) for 5–10 min with occasional stirring. The liquid was decanted and replaced with 50 mL of fresh homogenization medium followed by homogenization (30–60 sec.) on a TAMATO homogenizer (1,000 rpm × 3 and 1,500 rpm × 3). The homogenate was squeezed through a single layer of microcloth and centrifuged (10 min, 1,000 g; HITACHI centrifuge). The supernatant was centrifuged (30 min, 3,000 g), and sequentially centrifuged (30 min, 8,000 g) after discarding the sediment. The remainder supernatant was centrifuged (30 min, 34,000 g), carefully decanted, and centrifuged again (130,000 g, 1 h; Beckman Instruments, Palo Alto, CA) to get the “light” microsomes. The pink sediment was gently resuspended with a glass homogenizer in ~7 mL of 52% sucrose-0.1 M H3PO4 buffer (pH 7.1), and the density of sucrose was adjusted to 43.7%. The fraction was placed in one type-70i rotor centrifuge tube; overlayered sequentially with 7 mL, 5 mL, 5 mL, and 6 mL of 38.7%, 36.0%, 33.0%, and 29.0% sucrose, respectively, and centrifuged (80,000 g, 1 h). The upper four layers of the sucrose gradient were discarded by aspiration, and the bottom layer (43.7%) was diluted with two volumes of cold distilled water and centrifuged (130,000 g, 1 h) in a type-70i rotor to get the “heavy” microsomes. The pellets, light and heavy microsomes, were suspended in 3 mL of 0.25 M sucrose (pH 7.0) and combined. The mixture was diluted to 14 mL with 0.25 M sucrose containing CsCl with its final concentration of 0.015 M. The suspension was layered into an equal volume of 1.3 M sucrose/0.015 M CsCl and then centrifuged (240,000 g, 1 h) in an SW 55Ti rotor. The rough microsomes were in the pink sediment, and the smooth microsomes were at the interface. The smooth microsomes were diluted with an equal volume of 0.25 M sucrose (pH 7.0) and centrifuged (140,000 g, 1 h) in an SW 55i rotor.

2.2.2. Detection and Validation of the Purity of Microsomes

Electron microscopy and Western blotting were used to detect and validate the purity of prepared microsomes. For electron microscope analysis, the prepared microsomes were fixed with 2.5% glutaraldehyde for 24 h and 2% OsO4 for 2 h, dehydrated with alcohol (50%, 70%, 90%, and 100% in turn), and processed into epoxy resin. Thin sections (500 Å) were prepared and stained with uranyl acetate and lead citrate then examined with a transmission electron microscope (H-600-1, Hitachi, Japan). For Western blotting analysis, the microsome fractions were lysed (4°C; 30 min) in lysis buffer (50 mM Tris-Hcl, 150 mM NaCl, 1 mM EDTA, 1% Triton-X100, and 0.1% SDS). The protein samples (50 μg) were subjected to electrophoresis on SDS-PAGE with 12% gel and transferred to PVDF membrane (Millipore). The PVDF membranes with proteins were immunoblotted with antibodies to endoplasmin (ER marker), OxPhos complex IV subunit I (mitochondrial marker), catalase (peroxisomal marker), and cadherin (cytoplasmic marker), respectively.

2.2.3. Separation of Microsome Proteins by 2DE

2DE was performed as described by the manufacturer (Amersham Biosciences). Protein samples (400 μg) were diluted to 450 μL with rehydration solution (7 mol/L urea, 2 mol/L thiourea, 0.2% DTT, 0.5% (v/v) pH3–10 NL IPG buffer, and trace bromophenol blue) and applied to IPG strips (pH 3–10 NL; 24 cm) for rehydration (14 h; 30 V). Proteins were focused successively (1 h at 500 V, 1 h at 1,000 V, and 8.5 h at 8,000 V) to give a total of 68 kVh on an IPGphor. After equilibration, SDS-PAGE was performed with 12% gel on Ettan DALT II system. Then, the blue silver staining method was used to visualize the protein spots on the 2DE gels [14].

2.3. Na2CO3 Extraction and 1DE Analysis of Microsome Membrane Proteins

Microsome membrane proteins were further extracted by the carbonate method from isolated microsomal fractions [12]. Microsomal fractions were diluted 50- to 1,000-fold with 100 mM sodium carbonate (pH 11.5; final protein concentration to 0.02 to 1 mg/mL), and incubated (0°C; 30 min) with slow stirring and accompanying sonication for 15 sec at 3-4 W at 0 min, 15 min, and 30 min. The suspensions were centrifuged and decanted, and the membrane pellets were gently rinsed three times with ice-cold distilled water. These pellets were diluted with denaturing sample buffer (5% mercaptoethanol, 2% SDS, 0.06 M Tris-HCl, pH 6.8, and 10% glycerol), heated (95°C; 5 min), and then subjected to 1D SDS-PAGE with a 12% gel. Electrophoresis was performed at 80 V for 20 min, followed by 100 V for 2 h. Gels were visualized with Coomassie Brilliant Blue G [14].

2.4. Tandem Mass Spectrometry (MS/MS) Identification of Proteins
2.4.1. In-Gel Digestion

The proteins contained in the 2D gel spots and 1D gel bands were subjected to in-gel digestion with trypsin. Gel spots or bands were excised and destained with 100 mM NH4HCO3 in 50% acetonitrile (ACN) at room temperature. The proteins were reduced with 10 mM dithiothreitol (DDT) (56°C; 30 min) and alkylated with 50 mM iodoacetamide in 100 mM NH4HCO3 (dark, room temperature, 30 min). The gel pieces that contained proteins were dried and then incubated in the digestion solution (40 mM NH4HCO3, 9% ACN, and 20 μg/mL trypsin; 18 h, 37°C). The tryptic peptides were extracted with 50% ACN/2.5% TFA and then dried using a Speed-Vac.

2.4.2. Nanoliquid Chromatography (LC) MS/MS and Protein Identification

The tryptic peptide mixture was fractionated with reverse-phase (RP) high-performance liquid chromatography (HPLC) by using an Ultimate nano-HPLC system (Dionex). Peptide samples were purified and concentrated with a C18-PepMap precolumn and then separated on an analytical C18-PepMap column (75 μm ID × 150 mm, 100 Å pore size, 3 mm particle size) at a column flow rate of 300 nL/min. The ACN gradient (solution A: 0.1% formic acid, 2% ACN; solution B: 0.1% formic acid, 80% ACN) started at 5% B and ended at 70% B in 45 min. Mass spectrometry (MS) and MS/MS data were acquired using a Micromass quadrupole time of flight Micromass spectrometer (Waters). Database searches were carried out with the MASCOT server by using a decoy database (concatenated forward-reverse mouse IPI database, version 3.07; release date June 20, 2005). A mass tolerance of 0.3 Da for both parent (MS) and fragmented (MS/MS) ions, allowance for up to one trypsin miscleavage, variable amino acid modifications consisting of methionine oxidation and cysteine carbamidomethylation were used. MS/MS ion score threshold was determined to produce a false-positive rate less than 5% for a significant hit . The false-positive rate was calculated with 2* reverse/(reverse + forward)/100. In the current study, the MS/MS ion score threshold was 23 and a false-positive rate was approximately 3.1%. For all the proteins that were identified with only one peptide, each MS/MS spectrum was checked manually.

2.5. Bioinformatics Analysis of Identified Proteins

Protein annotations were obtained primarily from UniProt 7.0 including accession, entry name, comments such as function, catalytic activity, subcellular location, and similarity. The Cytoscape plugin, Biological Networks Gene Ontology (BinGO), was used to find statistically overrepresented GO categories of the protein dataset. An online tool, WebGestalt (http://bioinfo.vanderbilt.edu/webgestalt/), was used to map target proteins to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. The pathway visualization was based on the pathway mapping service provided in KEGG.

3. Results

3.1. Characterization and Detection of Liver Microsomes

It was essential to obtain a highly pure fraction to conduct proteomic characterization of microsomes. The purity of prepared microsomes was monitored with electron microscope and Western blotting analysis. A large number of nearly spherical membrane vesicles were visualized with electron microscope without other contaminated organelle compositions (see Supplemental Figure  1(a) in Supplementary Material available online at http://dx.doi.org/10.1155/2012/832569). Western blotting analyses showed that, with the standard immunoblotting protocol, the ER marker endoplasmin was enriched in the isolated microsome fractions without the contamination marker (mitochondrial marker OxPhos Complex IV subunit I, peroxisomal marker catalase, and cytoplasmic marker cadherin) being detected (Supplemental Figure  1(b)). The results demonstrated an optimized preparation of microsomes.

3.2. Fractionation and Identification of Microsome Proteins Identified by 2DE and MS/MS

The 2DE reference maps display protein spots ( gels). A representative 2DE map of microsome proteins was shown (Figure 1). A total of 183 proteins were identified with ESI-Q-TOF MS/MS from 204 excised gel spots. Those proteins are summarized (Table 1 and Supplemental Table  1), including 2D gel-spot number, IPI number, protein name, predicted TMD, and subcellular location. The microsomal marker proteins such as endoplasmin (Spot 2) and UDP glucuronosyltransferase (Spots 6 and 7) were identified. Those proteins were located in different subcellular locations (Table 1) including ER, mitochondrial membrane, cytoplasmic, ribosome, microbody, microsome membrane, nuclear, vesicular membrane, sarcolemma, extracellular space, cilium, ER-Golgi intermediate compartment, and secreted proteins. Supplemental Figure  2 shows the percentage of each group of proteins, according to their subcellular locations, derived from the annotations in the Swiss-Prot database and Gene Ontology: 22% of proteins ( ) from ER and Golgi, 11% of proteins ( ) from mitochondria and other membranes, 50% of proteins ( ) from cytosolic and other soluble proteins, 8% of secreted proteins ( ), and 9% of proteins without unambiguous location ( ).

3.3. Fractionation and Identification of Microsomal Membrane Proteins Identified by 1DE and MS/MS

The Na2CO3-treated microsome membrane proteins were separated on SDS-PAGE gels and visualized with Coomassie brilliant blue staining (Figure 2(a)). A total of 99 proteins (Table 2 and Supplemental Table  2) was identified with electrospray ionization- (ESI-) Q-TOF MS/MS from 17 gel bands (Figure 2(a)). Those proteins were derived from the ER, type I/II membrane proteins, integral membrane proteins, major histocompatibility complex class I protein, ER-Golgi intermediate compartment, mitochondrial membrane, nuclear, cytoplasm, microbody, sarcolemma, and secreted and unknown proteins (Table 2). Those membrane proteins were classified into three categories (Figure 2(b)): (a) proteins with known membrane associations (55%; ), (b) putative membrane proteins (5%; ), and (c) other proteins (40%; ). Those identified proteins were categorized according to the reported annotation in the UniProt database (http://www.uniprot.org/) and predictions for transmembrane regions (http://www.cbs.dtu.dk/services/TMHMM/). Of the 99 proteins, 59 (60%) were described as “membrane-associated” proteins (category (a) and (b)), including ER-characteristic proteins (cytochromes P-450 and b5, calnexin, integral membrane enzymes such as NADPH-cytochrome c reductase, and microsomal glutathione S-transferase 1).

Hydrophobicity is an important characteristic of a membrane protein. The grand average of hydropathy (GRAVY) scores (>−0.4) (http://us.expasy.org/tools/protparam.html) is an index to evaluate the hydrophobic status of a protein, indicates a hydrophobic protein, and suggests a membrane association. In the current study, 69 (70%) of the 99 proteins identified from 1DE had a (Supplemental Figure  3), a score indicating the probability for membrane association. Moreover, some alkaline proteins with PI values close to or greater than 10 were separated by 1DE (Supplemental Figure  4), but they could not be detected in a conventional 2DE map.

3.4. Comparison of 2DE and 1DE Datasets

Among the 2DE dataset ( proteins; Table 1) and 1DE dataset ( proteins; Table 2), only 23 proteins (Table 3) were consistent between 2DE and 1DE datasets (23% of 1DE dataset, and 13% of 2DE dataset). A total of 259 nonredundant proteins ( were identified in the microsome fraction through the strategy of combining 2DE with 1DE protein-separation technologies followed by ESI-Q-TOF MS/MS. The microsome consisted of a complex network of continuous membranes including ER, ER-Golgi intermediate complex—also referred to as the vesiculotubular clusters or pre-Golgi intermediates—and the Golgi apparatus [5]. Among those identified proteins, 62 located in ER and Golgi were definitely classified as microsome proteins by annotation in the Swiss-Prot database and the Gene Ontology (GO).

3.5. Significantly Enriched GO Terms for Mouse Liver Microsome Proteins

Biological Networks Gene Ontology [15] and Cytoscape [16] plugins to find statistically overrepresented GO categories were used for the enrichment analysis of our protein dataset. The microsome protein dataset ( , from 1DE and 2DE datasets) was compared to a reference set of complete mouse proteome (IPI mouse) that was provided by Biological Networks Gene Ontology. The analysis was done with a hypergeometric test, and all significant ( ) GO terms were selected after correcting for a multiple term testing with a Benjamini and Hochberg false discovery rate. The analysis was performed separately for molecular function, cellular component, and biological process categories, and x-fold enrichment for every overrepresented term in three GO categories was calculated (Supplemental Figure  5). The results showed that the terms were related to mostly catalytic activity in terms of molecular function, including metabolism-related oxidoreductase, hydrolase, and dehydrogenase. Similarly, terms belonging to the cellular component namespace include mitochondrion, ER, and ribosome. Finally, terms from the biological process namespace included metabolic process, localization, transport, and translation. All of the information suggested the main functions and compositions of microsome.

3.6. Significant Enrichment of KEGG Pathway for Mouse Liver Microsome Proteins

Biological pathways analysis based on KEGG pathway database was performed with an analysis toolkit—WebGestalt (http://bioinfo.vanderbilt.edu/webgestalt/) [17]. This toolkit allowed the functional annotation of gene/protein sets into well-characterized functional signaling pathways (KEGG: http://www.genome.jp/kegg/). In addition, an enrichment score was obtained of the frequency of occurrence of a specific protein (or gene) within any given experimental subset with respect to a species-specific background set. Thus, an enrichment factor (the observed frequency in input set/the expected frequency in background set) was created with a statistical value that indicated that the protein (or gene) was specifically overrepresented in the input dataset. In this current study, all the proteins except 81 were linked to a total of 99 biological pathways in the KEGG, including metabolic pathway, glycolysis/gluconeogenesis, metabolism of xenobiotics by cytochrome P450, and PPAR signaling pathway. Among those pathways, 34 significantly ( ) enriched biological processes analyzed by WebGestalt were obtained (Figure 3). Those biological processes were involved in cell metabolism, benzoate degradation, metabolism of xenobiotics, ribosome, biosynthesis, signaling pathway, and oxidative stress. Those results are known to be related to microsome.

To ascertain the coverage of our dataset with the enriched pathways or biological processes, the KEGG search service was used to map our dataset on KEGG pathways. Two of the aforementioned enriched KEGG pathways (metabolism of xenobiotics and ribosome) were related to the well-known function and composition of the microsome (Figure 4). Enzyme Commission numbers (EC no., e.g, 1.14.14.1) are used to represent enzymes in metabolism. Highlighted in green background are known mouse enzymes annotated in the KEGG database and the red boxed are enzymes in our dataset (Figure 4(a)). All enzymes ( ) that played a key role in every pathway of metabolism of xenobiotics were included in our dataset (Table 4). Thirteen proteins from large and small subunits of ribosome were also found in our dataset (Table 4) and are indicated with a red box (Figure 4(b)). These proteins interact physically with each other and form a large protein complex—the ribosome. All the identified proteins that are involved in those two pathways are summarized in Table 4, including their KEGG pathway, protein ID, and protein name.

4. Discussion

Proteome analysis of the cell membrane-bound organelles is a daunting task mainly because of (a) isolation of membrane that is free from nonconstituents and (b) solubilization of membrane proteins in a manner amenable to isoelectric focusing [10]. 2DE is an effective tool to survey biological complexity at the molecular level and provides a systematic and comprehensive study of the proteins. However, because of the PI value range limited by the IPG strip and the high dependence on sample preparation, some problems exist for the available 2DE protocols to resolve membrane-associated proteins [10, 22]. Therefore, in the current study, the whole microsome lysate was arrayed with 2DE, and the membrane fraction of microsomes purified by the carbonate procedure was separated with 1DE. The complementary 2DE and 1DE approaches provided a much wider coverage of microsome proteome.

Hydrophobicity and relatively low abundance causes a challenge for proteomic technology to separate and identify membrane proteins. The hydrophobicity of proteins is frequently expressed as GRAVY scores (http://us.expasy.org/tools/protparam.html). A calculated GRAVY score of up to –0.4 indicates a hydrophobic protein, suggesting a membrane association [21]. In the current study, 69 (70%) of the 99 proteins identified from 1DE had a (Supplemental Figure  3), indicating the probability for membrane association [21]. As shown in Supplemental Figure  4, some alkaline proteins with PI values close to or greater than 10 were separated by 1DE; they could not be detected in conventional 2DE map. Only 23 proteins were found to be consistent between 2DE and 1DE datasets with 6 proteins classified as membrane proteins (Table 3). All these results indicate that 1DE is a potent supplement to 2DE, and the combination of the two approaches is necessary in protein profiling of microsomes.

Microsome-sealed vesicles could be converted into flat membrane sheets with cisternal contents that were released effectively with the treatment solution (100 mM Na2CO3; 0°C). It appears to be as effective as the low detergent procedure in selectively releasing microsomal content. In the current study, some proteins that were identified from Na2CO3-extracted fraction were classified as membrane associated mainly based on published reports, even though their predicted transmembrane domains (TMDs) did not suggest a membrane origin. The observations point out the fact that structure alone may not be the deciding factor, as far as the association of proteins with cell membrane is concerned. First, the proteins may be bound to the membrane simply to perform their functional obligations. Consequently, they could become part of complexes involving membrane proteins and may not depart from them easily under the conditions of sample preparation. For example, many enzymes were identified in the extracted membrane fraction, such as Cis-retinol androgen dehydrogenase 1 (short-chain dehydrogenase family). It is anchored to the ER membrane facing the cytoplasm by an N-terminal signaling sequence of 22 residues and takes part in the membrane-associated retinoid metabolism [23], so is fatty acid-binding protein, which participates in the palmitic acid or retinylester metabolism that is incorporated in microsomal membranes [24] and the free fatty acid transferation to the membrane. Second, some truly cytosolic proteins may simply integrate with membrane vesicles during the sonication process and become difficult to remove by the extraction methods [25]. Studies [5] have demonstrated that hepatic microsomes are derived from the ER and other cell organelles. The ER represents a membrane tubular network that crosses the cytoplasm from the nucleus membrane to the plasma membrane. Moreover, some proteins perform their functions between cytoplasm and ER, such as fatty-acid-binding proteins [26]. From this point of view, taking all of the portions into account, 60%–70% of the proteins identified can be regarded as microsome proteins in this research. A part (~15%) of identified proteins did not have unambiguous locations in published reports or annotations in the genome database. This current study provides information relevant to subcellular locations of these proteins for subsequent studies.

Two datasets from 1DE and 2DE are part of the complete protein composition of microsomes. A bioinformatics analysis of the two datasets combined offers more information. For an overview of the proteomic data and comprehending their biological importance, biological networks GO (BinGO) (http://www.psb.ugent.be/cbd/papers/BiNGO/index.html) was used to identify GO-category significant enrichment with all the identified proteins. BiNGO is a plugin for Cytoscape, which is an open source bioinformatics software platform to visualize and integrate molecular interaction networks. BinGO maps the predominant functional themes of a given gene set on the GO hierarchy. Of the 259 target proteins and direct partners analyzed, 182 target proteins linked to one or more GO terms. GO-term enrichment analysis revealed that the most highly represented GO terms in the cellular GO category component were organelles such as ER, mitochondrial, and organelle membrane. An analysis of the proteins that were identified according to their potential roles in biological processes indicated that the proteins were mainly involved in metabolic process, localization, transport, and translation. All the results were highly statistically significant.

The KEGG pathway database integrates current knowledge on molecular interaction networks in biological processes. To gain a broad understanding of our dataset, WebGestalt (a web-based gene set analysis toolkit) was used to map the identified proteins to KEGG pathways. The results showed that 112 of the total proteins were associated with one or more KEGG pathways. Meanwhile, 97 of 112 target proteins (87%) fell into 34 KEGG pathways; they were specifically enriched ( ) compared to statistical expectations. Pathways that are involved in benzoate degradation, metabolism of xenobiotic, glutamate metabolism, and cysteine metabolism were among the most enriched biologically. This finding was consistent with the fact that microsomes were used to investigate the metabolism of compounds and to examine drug-drug interaction by in vitro studies.

Collectively, the bioinformatics analysis via enrichment analysis of GO annotation and KEGG pathways derived meaning from the proteomic data and assisted in the understanding of the function of liver at the subcellular level.

Novelty and Limitation
Mammalian liver microsome proteomes have been studied by several groups [1820]. Comparison of the current study with the literature data [1820] was shown in Tables 5 and 6. Zgoda et al. [18] studied differential ultracentrifugation-separated mouse liver microsome proteome; 2DE and silver stain yielded 1,100 protein spots, and 138 proteins contained in 2D gel spots were identified with peptide mass fingerprint (PMF). Zgoda et al. [19] also studied differential ultracentrifugation-separated mouse liver microsome proteome with 1DE and MS/MS; 519 proteins were identified including 138 (138/519 = 27%) predicted membrane proteins. Gilchrist et al. [20] used 1DE and MS/MS to analyze rat ER and Golgi that were separated with differential ultracentrifugation and density gradient centrifugation; 832 ER proteins were identified including 183 (183/832 = 22%) membrane proteins. This current study combined differential ultracentrifugation and sucrose gradient centrifugation to prepare mouse liver microsomes; 2DE and Coomassie brilliant blue stain yielded 514 protein spots, and 183 proteins were identified with MS/MS from 204 excised gel spots, including 41 (41/183 = 22%) membrane proteins. Na2CO3 was used to further extract membrane proteins from isolated microsomes; 1DE and Coomassie brilliant blue stain yield 17 protein bands, and 99 proteins were identified with MS/MS from those 17 protein bands, including 54 (54/99 = 55%) membrane proteins. A total of 259 nonredundant proteins were identified including 62 (62/259 = 24%) membrane proteins. Compared to the documented data [1820], the novelty of this current study is that the carbonate method significantly increased the identification rate of microsomal membrane proteins, that some proteins and functional annotations from this current study have not been identified in other literature, which expanded and enriched the documented data, and that the established analysis system and data will benefit the discovery of liver disease-related microsomal membrane proteins. Meanwhile, we also noted that the current study had a relatively low coverage ( proteins) of mouse liver microsome proteome relative to the documented data ( proteins [19] and 832 proteins [20]), which might be derived from several factors: (i) inconsistent protein-extracted procedures and protein-stained methods were used, (ii) only part of 2D gel spots were excised to identify proteins, (iii) only visualized 1D gel bands (not the entire 1D gel lane) were used for protein identification, (iv) MS/MS (not PMF) was used to identify 2D gel proteins, (v) different sensitivity mass spectrometers were used, (vi) different parameters were used to search protein database. The use of 2D/3D LC-MS/MS [19] and carbonate extraction of isolated microsomes would significantly improve the coverage of microsomal membrane proteome.

5. Conclusions

The preparation of liver microsomes was optimized. The data presented here demonstrated that 1DE and 2DE are complementary approaches to analyze the intracellular microsomes that contain considerable numbers of highly hydrophobic membrane proteins. An integrated bioinformatics analysis of all of the microsome proteins identified with 1DE and 2DE can provide a relatively complete understanding of the protein composition and cellular function of the target microsome organelles. The information presented here will be useful for successful analysis of other membranous organelles. Our data will assist in understanding the function of liver and are an important reference for subsequent analysis of liver disease-related microsome proteins for biomarker discovery and mechanism clarification of a liver disease.

Abbreviations

ACN:Acetonitrile
BinGO:Biological Networks Gene Ontology
DTT:Dithiothreitol
1DE:One-dimensional gel electrophoresis
2DE:Two-dimensional gel electrophoresis
ER:Endoplasmic reticulum
GO:Gene ontology
GRAVY:Grand average of hydropathy
HLPP:Human Liver Proteome Project
IEF:Isoelectric focusing
KEGG:Kyoto Encyclopedia of Genes and Genomes
LC:Liquid chromatography
MS:Mass spectrometry
MS/MS:Tandem mass spectrometry
Q-TOF:Quadrupole-time of flight
RP:Reverse phase
TMD:Transmembrane domains.

Acknowledgments

This work was supported by China National Haman Liver Proteome Project (Grant no. 2004 BA711A18) and The National Basic Research Program of China (Grant No. 2011CB910704).

Supplementary Materials

Supplemental figure 1: is detection and validation of the purity of isolated microsomes.

Supplemental figure 2: is distribution of subcellular locations of 2DE-derived proteins.

Supplemental figure 3: is distribution of 1DE-derived proteins over the GRAVY scores.

Supplemental figure 4: is distribution of 1DE-derived proteins over the pI values.

Supplemental figure 5: is significant enrichment of GO terms for mouse liver microsome proteins (n = 259) that were derived from 1DE and 2DE strategies.

Supplemental Table 1: MS/MS identification of 2DE-arrayed proteins from mouse liver microsomal preparations.

Supplemental Table 2: MS/MS identification of 1DE-separated proteins from Na2CO3-extracted mouse liver microsomal membrane preparations

  1. Supplementary Figures
  2. Supplementary Table 1
  3. Supplementary Table 2