A Label-Free Proteomic Strategy to Investigate the Intramuscular fat Proteomic Differences Between Biceps Femoris and Longissimus Dorsi in Inner Mongolian Cashmere Goats

Background: As a major raw-cashmere-producing province in China. Nearly 700,000 Aerbasi cashmere goats are fed per year, and the corresponding meat production is nearly 10,000 tons. However, there are no reports on the meat of this goat. To better understand the molecular variations underlying intramuscular fat (IMF) anabolism and catabolism in Inner Mongolian cashmere goats, the proteomic differences between the biceps femoris (BF) and longissimus dorsi (LD) were investigated by a label-free strategy. Then, the identied proteins were veried as being involved in IMF anabolism and catabolism by Western blot analysis. Results: The IMF content was signicantly higher in the BF than in the LD, suggesting that IMF accumulated more in the BF or was metabolized more in the LD. We performed proteomic analysis of IMF anabolism and catabolism at the proteomic level, and 1209 proteins were identied in the BF (high-IMF) and LD (low-IMF) groups. Among them, 110 were differentially expressed proteins (DEPs), 81 of which were upregulated in the high-IMF group, while 29 were upregulated in the low-IMF group. Gene ontology (GO) classication showed that the 110 DEPs were functionally classied into 100 annotation clusters. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed that the 110 DEPs covered 34 KEGG pathways. Three pathways were related to IMF metabolism and deposition—fatty acid metabolism, fatty acid degradation and fatty acid elongation—and included 7 proteins. Conclusion: GO and KEGG analyses showed that differentially expressed HADHA, HADHB, ACSL1, ACADS, ACAT1 and ACAA2 in the mitochondria act via fatty acid metabolism, fatty acid degradation and fatty acid elongation to inuence the metabolism and synthesis of long-, short- and medium-chain fatty acids and modulate IMF anabolism and catabolism. Protein-protein interaction (PPI) network analysis showed that IMF accumulation in different muscle tissues of Inner Mongolian cashmere goats was affected not only by 5 key enzymes and proteins involved in fatty acid synthesis and metabolism but also by ve DEPs (SUCLG1, SUCLG2, CS, DLST, and ACO2) in the TCA cycle. Our results provide new insights into IMF deposition in goats and improve our understanding of the molecular mechanisms underlying IMF anabolism and catabolism. long-chain acid, short-chain acid and medium-chain acid metabolism and synthesis inuence IMF anabolism and catabolism.

affected not only by 5 key enzymes and proteins involved in fatty acid synthesis and metabolism but also by ve DEPs (SUCLG1, SUCLG2, CS, DLST, and ACO2) in the TCA cycle. Our results provide new insights into IMF deposition in goats and improve our understanding of the molecular mechanisms underlying IMF anabolism and catabolism.

Background
As a major raw-cashmere-producing province in China, Inner Mongolia cashmere goats harbors three types -Erlangshan, Aerbasi and Alashan [1]. The Inner Mongolian cashmere goat is a local breed that provides both cashmere and meat, and the production of cashmere in Inner Mongolia accounts for approximately 40% of the total output of the whole country [2]. Nearly 700,000 Aerbasi cashmere goats are fed per year, and the corresponding meat production is nearly 10,000 tons. Aerbasi cashmere has been widely studied, but the meat value of this goat has not been reported. Therefore, the study of Aerbasi cashmere goat meat is of great signi cance.
Intramuscular fat (IMF) is mainly found in the sarcolemma, including the perimysium, epimysium and endomysium, and is mainly composed of triglycerides and phospholipids [3]. The IMF content and its fatty acid composition play an important role in meat quality, affecting the sensory properties ( avor, juiciness and tenderness) and nutritional value of meat [4]. The positive effect of IMF on the sensory quality of meat has been demonstrated in pork [5], mutton [6], and beef [7]. The IMF deposition capacity is regulated by 3 aspects: transport to fatty acids, fat anabolism and fat catabolism. Recent studies have demonstrated that FABP and CD 36 play important roles in the process of fat uptake [8,9], and DGAT 1 and SCD regulate the synthesis of IMF [10,11]. The HSL and LPL genes regulate fat breakdown [12,13]. However, it is not known which proteins affect the deposition and metabolism of IMF in cashmere goat meat, so research on the composition of protein in cashmere goat meat is urgently needed.
In recent years, the label-free approach to proteomics, which is known to be a reliable, versatile, and costeffective strategy, has received much attention [14]. Furthermore, it has been successfully applied to compare proteome changes in beef color stability [15] and gain knowledge of broiler self-regulation mechanisms under heat stress [16]. However, the application of a label-free proteomics strategy to investigate proteomic changes that occur during the process of IMF regulation has not yet been reported in cashmere goats. Moreover, the proteomic abundance in cashmere goats remains unclear. Therefore, the objective of the present study was to identify key proteins involved in IMF anabolism and catabolism and to reveal underlying biochemical events with the help of bioinformatics analyses.

IMF content of the LD and BF muscles
To assess the IMF content, the longissimus dorsi (LD) and biceps femoris (BF) tissues were collected from 20 goats. Independent sample t-tests for fatty acids in different muscles showed that the IMF content was signi cantly higher in the BF than in the LD (P < 0.01) (Fig. 1). This suggests that IMF is deposited at higher levels in the BF or metabolized to a greater degree in the LD. To explore IMF anabolism and catabolism at the proteomic level, we selected the BF and LD from 6 goats for the next experiment.

Protein Identi cation And Comparative Analysis
To understand the protein composition of these six goats, the UniProt/Swiss-Prot/Capra hircus database was used as a reference to investigate the proteome during IMF anabolism and catabolism using a labelfree mass spectrometry strategy. A total of 1209 proteins were identi ed in the BF (high-IMF group) and LD (low-IMF group), and 993 and 896 proteins were identi ed in each group with a false discovery rate (FDR) ≤ 0.01 (supplementary material 1). Therefore, there are many proteins that are yet to be studied in Inner Mongolian cashmere goat.
To visualize and differentiate the observed sample clusters, Principal component analysis (PCA) was performed to compare the high-IMF group and low-IMF group based on the proteins present in both groups. PCA is an unsupervised method that can condense a large number of variables (proteins) into a set of representative and uncorrelated principal components by means of their variance-covariance structure. The score plot of the PCA of the high-IMF group and low-IMF group (Fig. 2a) showed that 45.9% of the variability was explained by the rst two principal components, which accounted for 31.8% and 14.5% of the total variance. Samples from the two groups could be separated completely and were located in different quadrants, which indicated the existence of differentially abundant proteins between the high-IMF group and low-IMF group.
The volcano plots present samples with fold changes > 2 or < 0.5 and P-value < 0.05 (Fig. 2b). In the comparison of the high-IMF group and low-IMF group, there were a total of 110 signi cant differentially expressed proteins (DEPs) that were associated with the accumulation or metabolism of IMF; 81 proteins were upregulated in the high-IMF group, whereas 29 proteins were upregulated in the low-IMF group. The detailed results for the DEPs are presented in supplementary material 2.

Function Analysis Of Differentially Expressed Proteins
According to the Gene ontology (GO) classi cation statistics (Fig. 3), DEPs from IMF metabolism and deposition can be divided into three main categories: biological processes, cell components, and molecular functions. In this study, the 110 DEPs were functionally classi ed into 100 annotation clusters (supplementary material 3). The top 10 annotated GO terms in the biological process category showed that the DEPs participated in the oxidation-reduction process, small molecule metabolic process, oxoacid metabolic process, organic acid metabolic process, generation of precursor metabolites and energy, and carboxylic acid metabolic process (Fig. 3a). The top 10 GO terms in the cellular component category indicated that the DEPs were mainly enriched in mitochondria, the intracellular region, organelles, intracellular organelles, the cytoplasm, and the mitochondrial envelope. The top 10 GO terms in the molecular function category indicated that the DEPs were mainly involved in catalytic activity, ion binding, oxidoreductase activity, cytoskeletal protein binding, and actin binding.
It is speculated that during IMF metabolism and deposition, the increase or decrease in proteins or enzymes leads to changes in the metabolic rate or accumulation, which further leads to differences in the IMF content.
Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was performed to evaluate the potential functions of exclusively detected proteins and DEPs. The top 20 enriched KEGG terms of these proteins are described in Fig. 3b. The results show that the 110 investigated proteins cover 34 KEGG pathways. In total, three of the pathways were related to IMF metabolism and deposition, including fatty acid metabolism, fatty acid degradation and fatty acid elongation. Seven proteins were included in these three pathways, such as acetyl-CoA acyltransferase 2 (ACAA 2 ), cholesterol acyltransferase 1 (ACAT 1 ), and Enoyl-CoA hydratase (HADHA) (Supplementary material 3). The KEGG analysis results provide further insight into IMF anabolism and catabolism.
Protein-protein interaction (PPI) network analyses were performed to construct the speci c molecular network involving the key DEPs related to IMF anabolism and catabolism between the high-IMF group and the low-IMF group.
To search for hub proteins among the DEPs, we used the software CytoHubba. One PPI network was determined, covering 33 DEPs and 147 edges (Fig. 4) targeted or were targeted by another 9 proteins; SUCLG 1 and SUCLG 2 targeted or were targeted by another 8 proteins (except HADHA); HADHB targeted or was targeted by another 8 proteins (except ACO 2 ); CS and DLST targeted or were targeted by another 7 proteins (except HADHA and ECHS1); ECHS 1 targeted or was targeted by another 7 proteins (except CS and DLST); ACO2 targeted or was targeted by another 7 proteins (except HADHA and HADHB); and HADHA targeted or was targeted by another 5 proteins (except DLST, SUCLG 1 , SUCLG 2 and ACO 2 ). These 10 hub genes play a vital role in regulating IMF anabolism and catabolism.
Both functional analysis and PPI network analysis showed that HADHA, ACAA 2 and ACAT 1 might play an important role in regulating IMF anabolism and catabolism in Inner Mongolian cashmere goat.

Protein Determination By Western Blot Analysis
Protein determination by Western blot analysis Western blot quanti cation is a critical method that provides accurate and reproducible results (Fig. 5). Therefore, this method was used to detect the expression levels of three DEPs (HADHA, fatty acid-binding protein (FABP 3 ) and AMP-binding domain-containing protein (ACSL 1 )) from the high-IMF group and low-IMF group. All three proteins were upregulated in the high-IMF groups. The average band intensities of HADHA, FABP 3 and ACSL 1 (normalized to β-tubulin) had similar variation tendencies according to the results of MS analyses. The band intensities of these three proteins were signi cantly higher in the BF than in the LD (p < 0.05).

Discussion
IMF is formed by the deposition of fat in muscle, which is composed of IMF and myo brils. A previous study showed that the IMF content ultimately depends on fatty acid transport, IMF anabolism and IMF catabolism [17,18]. The process of fatty acid transport involves fatty acids entering intramuscular cells to provide the necessary substrates for IMF synthesis; IMF anabolism includes the synthesis, elongation or desaturation of fatty acid chains and the synthesis of triglycerides; IMF catabolism includes mobilization of fat in intramuscular cells and hydrolysis of triglycerides in lipoproteins. Interestingly, in our study, we found that the IMF content was signi cantly higher in the BF than in the LD. This suggests that IMF is deposited at greater levels in the BF or metabolized to a greater degree in the LD. This provides a good model for studying the regulatory mechanisms of IMF anabolism and catabolism.
IMF anabolism and catabolism are closely associated with many critical cellular functions and biological processes. For example, IMF or fatty acid triglycerides are metabolized in the mitochondrial matrix, and this process is called fatty acid beta-oxidation [19]. Fatty acid beta-oxidation generally includes four steps: oxidation, hydration, oxidation and cleavage [20]. In our study, the GO classi cation statistics of 110 DEPs showed that the top biological process was the oxidation-reduction process, which included 36 DEPs. In the cellular component analysis, the most enriched cellular component was the mitochondrion, which included 39 DEPs. In the molecular function analysis, 65 DEPs were mainly enriched in catalytic activity. Overall, the DEPs were mostly located in mitochondria, had catalytic activity, and participated in oxidation processes. In summary, the GO analysis results were consistent with the function of fatty acid beta-oxidation steps, and DEPs involved in these processes were the key proteins contributing to the differences in fatty acid beta-oxidation, further in uencing IMF metabolism.
The KEGG enrichment analyses indicated that there were 3 terms (fatty acid metabolism, fatty acid degradation and fatty acid elongation) related to IMF, and 7 of the DEPs were associated with these three pathways. The seven proteins included in these three pathways were ACAA 2 , short-chain acyl-coenzyme A dehydrogenase (ACADS), ACAT 1 , ACSL 1 , short-chain enoyl-CoA hydratase (ECHS 1 ), HADHA, and HADHB.
Fatty acids can be classi ed as long-chain (containing more than 12 carbon atoms), medium--chain (containing 6-12 carbon atoms) and short-chain (containing less than 6 carbon atoms) fatty acids [21]. Short-chain fatty acids can directly cross the outer mitochondrial membrane and enter the mitochondrial matrix for fatty acid oxidation, but long-chain and medium-chain fatty acids need to be transported through the inner mitochondrial membrane under the catalysis of carnitine acyltransferase 1, which is located on the outer mitochondrial membrane. Among the 7 DEPs enriched by KEGG analysis, ACSL1, HADHA and HADHB are mitochondrial membrane proteins that act on long-chain fatty acids. ACSL 1 is known to catalyze the rst step (oxidation) of the activation of long-chain fatty acids by converting them into long-chain acyl-CoA thioesters for channeling toward chain elongation, triacylglyceride synthesis or fatty acid oxidation [22]. ACSL 1 is necessary for the synthesis of long-chain acyl-CoA esters, fatty acid degradation and phospholipid remodeling [23]. HADHA and HADHB, which break down fatty acids to acetyl-CoA, are speci c for long-chain fatty acids. HADHA is involved in fatty acid beta-oxidation, which is a part of lipid metabolism. Research has shown that HADHA overexpression signi cantly inhibits cell growth, induces cell apoptosis, and decreases the formation of cytoplasmic lipid droplets [24]. ACADS is one of the acyl-CoA dehydrogenases that catalyze the rst step (oxidation) of mitochondrial fatty acid beta-oxidation, an aerobic process that breaks down fatty acids to acetyl-CoA. It acts speci cally on acyl-CoAs with saturated 4-to 6-carbon-long primary chains. Studies have shown that ACADS not only plays a vital role in free fatty acid β-oxidation but also regulates energy homeostasis [25]. ECHS 1 , similar to ACADS in function, also participates in the metabolism of fatty acyl coenzyme A ester and is an important mitochondrial fatty acid beta oxidase but acts on straight-chain enoyl-CoA thioesters that have 4 to at least 16 carbons [26]. ACAT 1 and ACAA 2 are involved in lipid metabolism. They are two key enzymes of the fatty acid oxidation pathway, catalyzing the last step (cleavage) of mitochondrial betaoxidation. They use free coenzyme A/CoA and catalyze the thiolytic cleavage of medium-to long-chain unbranched 3-oxoacyl-CoAs to acetyl-CoA and fatty acyl-CoA, which are shorter by two carbon atoms, thus playing an important role in fatty acid metabolism [27]. GO and KEGG analyses showed that in the mitochondrion differentially expressed HADHA, HADHB and ACSL 1 participate in the fatty acid metabolism, fatty acid degradation and fatty acid elongation pathways to in uence long-chain fatty acid metabolism and synthesis, further in uencing IMF anabolism and catabolism. Differentially expressed ACADS affects fatty acid metabolism, fatty acid degradation and fatty acid elongation, in uencing shortchain fatty acid metabolism and synthesis and further in uencing IMF anabolism and catabolism; Differentially expressed ACAT 1 and ACAA 2 act on fatty acid metabolism, fatty acid degradation and fatty acid elongation, in uencing long-chain and medium-long-chain fatty acid metabolism and synthesis and further in uencing IMF anabolism and catabolism.
In the construction of the PPI network, a total of 10 hub proteins were found. Among them, ACAT 1 , ACAA 2 , HADHB, ECHS 1 , and HADHA were also identi ed by GO and KEGG analyses, and the other 5 interacting proteins were SUCLG 1 , SUCLG 2 , CS, DLST, and ACO 2 . All ve of these proteins are involved in the tricarboxylic acid cycle (TCA). Succinate-CoA ligase (SUCL) is a heterodimer consisting of an alpha subunit encoded by SUCLG 1 and a beta subunit encoded by SUCLG 2 , catalyzing an ATP-and a GTPforming reaction, respectively [28,29]. SUCL is at the intersection of several metabolic pathways [30]. For example, SUCLA 2 rebound increases, pleiotropically affecting metabolic pathways associated with SUCL [31]. CS synthesizes isocitrate from oxaloacetate, and Tereza Škorpilová, et al showed that there could be speci c limits designed for CS activity in chilled and frozen/thawed meats [32]. DLST catalyzes the overall conversion of 2-oxoglutarate to succinyl-CoA and CO 2 [33]. ACO 2 catalyzes the isomerization of citrate to isocitrate, and Aco2 is needed for mitochondrial translation [34]. Analysis of the PPI network showed that in addition to the 5 key enzymes and proteins involved in the process of fatty acid synthesis and metabolism affecting the nal accumulation of IMF in different muscle tissues of Inner Mongolian cashmere goat, there are also ve proteins in the TCA cycle that exhibit expression, thereby affecting the accumulation of IMF.
In our study, the proteomic pro les of the high-IMF and low-IMF muscles of Inner Mongolia cashmere goats were evaluated. Our results provide new insights into IMF deposition in goats and improve our understanding of the molecular mechanisms associated with IMF anabolism and catabolism.
Conclusion: GO and KEGG analysis results show that differentially expressed HADHA, HADHB, ACSL 1 , ACADS, ACAT 1 and ACAA 2 in mitochondria act on fatty acid metabolism, fatty acid degradation and fatty acid elongation pathways to in uence long-chain fatty acid, short-chain fatty acid and medium-chain fatty acid metabolism and synthesis and to in uence IMF anabolism and catabolism. PPI network analysis results showed that IMF accumulation in different muscle tissues of Inner Mongolian cashmere goats was affected not only by 5 key enzymes and proteins involved in the process of fatty acid synthesis and metabolism but also by ve DEPs (SUCLG 1 , SUCLG 2 , CS, DLST, ACO 2 ) involved in the TCA cycle. Our results provide new insights into IMF deposition in goats and improve our understanding of the molecular mechanisms associated with IMF anabolism and catabolism. And permission were obtained from the farm owner. Animals were slaughtered under controlled conditions after being electrically stunned, and then, the muscles were collected aseptically into enzyme-free tubes, immediately immersed in liquid nitrogen and subsequently stored at -80°C until analysis.

IMF content of meat
The IMF level for the BF and LD was determined using the Soxhlet extraction protocol following the method of Hopkins et al. [35]. Three grams of freeze-dried meat was weighed into a thimble and extracted in 85 mL of hexane for 60 minutes within individual extraction tins. The solvent was then evaporated off for an additional 20 minutes. The tin was then dried for 30 minutes at 105°C to remove any residual solvent. The variation in meat weight before and after extraction was used to calculate IMF content. The nal value is expressed as a percentage of meat weight.

Protein extraction and digestion
Each frozen sample was ground to powder in liquid nitrogen and dissolved in lysis buffer (proteinase inhibitors (Roche, Basel, Switzerland), 1% SDS (Coolaber, China)). The samples were then incubated at room temperature for 20 minutes, vortexing for 30 seconds every minute. After 20 minutes of ultrasonication, the samples were centrifuged at 12,000 rpm and 4°C for 30 minutes. The supernatant was collected for further study, and the protein concentration was measured with a bicinchoninic acid (BCA) kit (Tiangen, China).
One hundred micrograms of protein was added to 200 µL of 8 M urea (Sigma, Germany) and 10 mM DLdithiothreitol (Sigma-Aldrich, Germany) and incubated at 37°C for 1 hour. After centrifugation at 12,000 rpm for 40 minutes, 200 µL of urea was added to each ltrate tube, which was then agitated. Next, the ltrate tubes were centrifuged twice at 12,000 rpm for 30 minutes each. Then, 200 µL of 50 mM iodoacetamide (Sigma-Aldrich, Germany) was added to each ltrate tube; the reaction was allowed to proceed in the dark for 30 minutes, and then, the liquid was removed. Next, 100 µL of ammonium bicarbonate (Fluka, Germany) was added to each ltrate tube, and the samples were centrifuged at 12,000 rpm for 20 minutes. This step was performed 3 times, and then, the liquid was removed. The samples were incubated overnight with trypsin at 37°C and centrifuged at 12,000 rpm for 30 minutes. Then, 50 µL of ammonium bicarbonate (Fluka, Germany) was added to each ltrate tube. The samples were centrifuged at 12,000 rpm for 30 minutes, and this step was repeated. The ltrate was collected, freeze-dried, and stored at -20°C [36].

HPLC-MS/MS analysis
Two methods, namely, information-dependent acquisition (IDA) and sequential window acquisition of all theoretical fragment ion spectra (SWATH), were used to acquire data from separated peptides on the LC-MS/MS system (Sciex, Framingham, MA, USA). Approximately 2 μg of peptides was injected and separated on a C18 HPLC column (75 μm×15 cm). A linear gradient (120 minutes, going from 5 to 80% B at 500 nL/minute) of 0.1% formic acid in water and 0.1% formic acid in acetonitrile was used to separate peptides. The conditions for IDA were as follows: nominal resolving power of 30,000, time-of-ight (TOF)-MS collection from 350 to 1800 m/z and automated collision energy for MS/MS with IDA scanned from 400 to 1800 m/z. The conditions for SWATH-MS were as follows: 150-1200 m/z MS1 mass range, 100-1500 m/z MS2 spectra, and nominal resolving power of 30,000 and 15,000 for MS1 and MS2, respectively.

Data processing
Protein Pilot 4.5 software (Sciex, Framingham, MA, USA) was used with the UniProt/SWISS-PROT (https://www.UniProt.org/#) database (downloaded from https://www.UniProt.org; 556,388 proteins) to identify peptides. The results were ltered at a 1% FDR. The selected search parameters included the use of trypsin as the enzyme, allowing up to two missed cleavage sites. The peptide mass tolerance was ± 15 ppm, and the fragment mass tolerance was 20 mmu. The data were loaded into PeakView (Sciex, Framingham, MA, USA) software to search the SWATH databank using the ion library generated in Protein Pilot. PeakView generated extracted ion chromatograms (XICs) after processing targeted and nontargeted data. Then, the results were interpreted and quantitatively analyzed using MarkerView software (Sciex, Framingham, MA, USA). MarkerView allows a rapid review of data to determine the DEPs. PCA and volcano plot analysis, which combined fold change analysis and t-tests, were performed. A fold change >2 or fold change < 0.5 and statistical signi cance (p value < 0.05) were used to identify DEPs [37].

Bioinformatics analysis
The DEPs were subjected to bioinformatics analyses. g:Pro ler (https://biit.cs.ut.ee/gpro ler/gost) online software was used to perform the GO and KEGG pathway analyses. Cytoscape_v3.6.1 was used to perform the PPI analysis of DEPs, and cytoHubba was used for screening hub genes [38].

Western blotting
The homogenized denatured protein was separated by SDS-PAGE and transferred in a semidry state to a PVDF membrane (Bio-Rad, USA). The PVDF membrane was blocked in blocking buffer (Li-CoR, USA) for 2 hours and incubated overnight with murine monoclonal primary antibody (Abcam, Germany). The membrane was washed 3 times for 10 minutes each time and then incubated with uorophoreconjugated goat anti-mouse antibody (Li-CoR, USA) for 1 hour. Finally, the membrane was rinsed with water, and the immunoreactive bands were examined using a LI-COR Odyssey (CLX-0496, USA) nearinfrared imager.