Exploring the Possible Link between the Gut Microbiome and Fat Deposition in Pigs

Excessive lipid accumulation and high oxidative stress have become a serious health and economic problem in the pig industry. Fatness characteristics are crucial in pig production since they are closely related to meat quality. The gut microbiome is well acknowledged as a key element in fat deposition. But the link between gut microbiota and fat accumulation in pigs remains elusive. To examine whether there is a link between pigs' gut microbiome, lipogenic properties, and oxidative stress, we selected 5 high-fat pigs and 5 low-fat pigs from 60 250-day-old Jinhua pigs in the present study and collected the colon content, serum sample, and liver and abdominal fat segments from each pig for metagenomic analysis, the oxidative stress assay, and RT-qPCR analysis, respectively. The backfat thickness and fat content of the longissimus dorsi muscle were considerably higher in the high-fat pigs than in the low-fat pigs (P < 0.05). An obvious difference in GSH-Px and MDA in the serum between the high- and low-fat pigs was observed. After RT-qPCR analysis, we found the gene expression of ACC1 and SREBP1 in the liver and FAS, PPARγ, and LPL in the abdominal fat were significantly higher in high-fat pigs than in low-fat pigs (P < 0.05). Additionally, metagenomic sequencing revealed that high-fat pigs had a higher abundance of Archaeal species with methanogenesis functions, leading to more-efficient fat deposition, while low-fat pigs had higher abundances of butyrate-producing bacteria species that improved the formation of SCFAs, especially butyrate, thus alleviating fat deposition in pigs. Furthermore, a total of 17 CAZyme families were identified to give significant enrichments in different fat phenotypes of pigs. This study would provide a detailed understanding of how the gut microbiome influences fat deposition in pigs, as well as a hint for improving growth performance and fatness traits by manipulating the gut microbiome.


Introduction
The gut microbiota is a complex and dynamic ecosystem composed of trillions of microorganisms living in the digestive tract and acting as a regulator and barrier for other metabolic organs [1]. It has been linked to the onset of metabolic disorders such as obesity and oxidative stress. A chronic inflammatory process such as oxidative stress and metabolic disorders may result from the alteration in the structure of gut microbiota [2].
Fatness traits are not only a characteristic of obesity and oxidation injury but also very important in pig production since they are linked to fattening features and meat quality. The accumulation of excessive fat in animals and humans has become an increasing threat to the animal production industry as well as human health, which would lead to obesity and could further set off many other health problems such as cardiovascular disease, arthritis, and dementia. Furthermore, lipid accumulation and impaired lipid metabolism are associated with pathophysiological phenotypes in pigs, which has turned into a severe economic and health problem in pig production [3].
Chronic inflammation in adipose tissue due to excessive fat accumulation would promote proinflammatory status and oxidative stress [4]. The substantial rise in the generation of free radicals in humans and animals that disrupts the antioxidation-oxidation equilibrium is referred to as oxidative stress [5]. A variety of illnesses have been reported to be pathologically caused by oxidative stress, which has been shown to be correlated with fat accumulation [4]. In pigs, a significant number of free radicals are produced because of oxidative damage, which leads to several diseases [6]. More importantly, the development of the pig industry would be hampered by oxidative stress due to oxidative stress directly limiting the growth performance and health of pigs [5]. Therefore, studying the oxidative stress in pigs is of great significance to pig production, as well as other animals and human health.
Similar to humans, the pig gut microbiota is also a huge, sophisticated, and dynamic microbial population with a variety of vital biological functions [7], including energy harvest, methane synthesis, and the synthesis of short-chain fatty acids (SCFAs) [8]. Pig fatness phenotypes are closely associated with the composition and diversity of the microbial community in the gastrointestinal gut tract. The composition of fatty acids in the adipose tissues and liver could be altered by the activities of the gut microbiota [9]. The Jinhua pig in Zhejiang Province, China, is characterized by its slow growth speed and high intramuscular fat content, which is considered an excellent model for studying fat deposition [10]. In the previous study, we compared the bacterial community structure of obese Jinhua pigs and lean Landrace pigs and illustrated a clear causal connection between gut microbiota and fat deposition by fecal microbiota transplantation [11].
Most of the related studies focused on the differences in the structure of the intestinal microflora and fat deposition among different pig species. However, the relationship between fat-related phenotypes and the gut microbiome in pigs is still unclear. Therefore, in this study, fatness characteristics, fat-related gene expression in abdominal fat and liver, and oxidative stress in serum were measured in highand low-fat pigs, respectively. Furthermore, the composition, diversity, and potential functions of the gut microbiota between the two groups were studied using a highthroughput metagenomic sequencing technique, further analyzing the association between the gut microbiome and fat deposition. This study will provide basic data for improving fatness traits in pig production by manipulating the gut microbiome.

Animal Experiments and Sample Collection.
A total of 60 newly born Jinhua pigs were fed in 6 pens, 10 pigs per pen, in a commercial pig facility with a standard corn-soybeanbased diet and water ad libitum as described previously [12]. Within two months before slaughter, each pig was healthy and had not been treated with antibiotics. At 250 days of age, pigs were killed to acquire blood samples, liver segments, and abdominal adipose tissue. Carcass traits, namely, body weight, loin muscle area, backfat thickness, and fat content in longissimus dorsi muscle, were analyzed by a principal component analysis (PCA). These statistics were utilized to determine which pigs were the most extreme for the selection of the most extreme pigs [13,14]. Five highfat (H) and five low-fat (L) pigs were chosen to collect luminal samples from the same colon location. Briefly, the gastrointestinal tract was peeled from the enterocoelia. The luminal samples were collected from the middle section of each pig's colon. Within 30 min after slaughter, all of the luminal samples were taken and divided into two parts: one was for the measurement of SCFAs, and the other one was placed in a -80°C refrigerator until DNA extraction after being dipped in liquid nitrogen.
All the procedures for animal experiments were approved by the Zhejiang Academy of Agricultural Sciences Institutional Animal Care strictly according to the relevant rules and regulations (Ethic code: ZAAS-2017-009).

Luminal DNA Extraction, Metagenomic Sequencing.
According to the manufacturer's recommendations, the QIAamp Fast DNA Stool Mini Kit (Qiagen, Germany) was used to extract the luminal DNA from each colon content sample. A NanoDrop 1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) was used to measure the amount and quality of DNA, and sterile water was used to dilute the DNA concentration to a final concentration of 1 ng/L. The Illumina TruSeq™ DNA Sample Prep Kit was used to generate libraries, which were sequenced on an Illumina HiSeq 2500 platform by a commercial sequencing company, Novogene (Beijing, China).

De Novo
Assembly of Short Reads. We used trimmomatic [15] for the quality control of raw datasets to remove the 3′and 5′-end of reads, eliminate poor quality bases (<20), and trim containing 10% N of reads, and small segments (<75 bp). The BWA [16] was performed to align the reads with the pig genome to filter out the host DNA from the reads. Megahit [17] (https://github.com/voutcn/megahit) was used to de novo assemble the filtered reads for each sample. Contigs were continuous sequences that had clear linkages between each other.

Gene Prediction and Functional Annotation.
The ORFs (open reading frames) from each sample's contigs were predicted using METAProdigal (http://prodigal.ornl.gov/). We used cd-hit [18] software (http://www.bioinformatics.org/ cd-hit/) to exclude the redundant genes (parameters for 95% identity; 90% coverage) from all the predicted ORFs. And then, Salmon [19] (https://github.com/COMBINE-lab/ salmon) was used to determine gene abundances by mapping the original sequences to anticipated genes. Finally, with BLASTP [20] (BLAST Version 2.2.28+, http://blast .ncbi.nlm.nih.gov/Blast.cgi), the taxonomy of the colon microbiota was evaluated against the NR database. The relative abundances of taxa at the domain, phylum, genus, and species levels were determined using taxonomic profiles. The NMDS was conducted at the gene level based on the 2.7. Oxidative Stress Determination. Serum antioxidant levels were determined after homogenization with saline (1 : 9 w/v) followed by centrifugation at 11,000 × g for 15 minutes at 4°C. As recommended by the manufacturer's instructions, we used Nanjing Jiancheng Bio (Nanjing, China) diagnostic kits to detect indicators including malondialdehyde (MDA), superoxide dismutase (SOD), and glutathione peroxidase (GSH-Px).
2.8. SCFA Measurement. As mentioned in our previous study [24], the gas chromatographic (GC) was used to determine the levels of SCFAs in each luminal content sample. Shortly, the 100 mg luminal content sample was weighed into a 1.5 mL centrifuge tube and suspended in 9 volumes of Milli-Q water. Following a 10 min centrifugation at 10,000 rpm, 1,000 μL of the supernatant was supplemented with 0.2 mL of crotonic acid (internal standard). Finally, following the membrane filtering (0.22 μm), the mixture was put into the GC-2010 plus (Shimadzu, Kyoto, Japan) with an FID detector operating at 180°C. The chromatographic conditions were as follows: column 110°C, vaporization chamber 180°C. The carrier gas was nitrogen at 0.06 MPa while the auxiliary gas was hydrogen and air with the pressure of 0.05 MPa and 0.05 MPa, respectively.
2.9. Statistical Analysis. All statistical analyses were conducted using the unpaired two-tailed Students' t-test with a P value < 0.05 as the level of statistical significance. Data were presented as mean ± standard deviation (SD).

Fatness Phenotypes between the High-and Low-Fat Pigs.
To determine whether there is a significant difference in the fatness phenotypes between the two groups, pigs were raised under standard management and sacrificed at 250 days old to determine body weight, loin muscle area, backfat thickness, and fat content of longissimus dorsi muscle. The high-fat pigs showed significantly higher backfat thickness and fat content of longissimus dorsi muscle than the lowfat pigs (P < 0:05, see Figures 1(b) and 1(d)). However, the two groups had no significant differences in body weight and loin muscle area (see Figures 1(a) and 1(c)).

Expression of Lipid Metabolism Genes.
To further test the gene expression of lipid metabolism in the liver and abdominal fat tissue, we collected liver and abdominal fat segments from each of the high-and low-fat pigs, followed by RNA isolation and RT-qPCR analysis. In the liver, the two key lipogenic genes, ACC1 (acetyl-CoA carboxylase-1) and FAS (fatty acid synthase), showed higher levels in the high-fat pigs than in the low-fat pigs, with ACC1 being significantly higher (P < 0:05, see Figure 2(a)). Additionally, SREBP1 (Sterol Regulatory Element Binding Protein-1) was also significantly higher in the liver of high-fat pigs (P < 0:05), while the gene expression of MLXIPL (Mlx-interacting protein-like) showed no significant difference (see Figure 2(a)).
In the abdominal fat tissue, a significant increase in the gene expression of lipogenesis genes, FAS and LPL (lipoprotein lipase), and the adipogenesis gene PPARγ (peroxisome proliferator-activated receptor-γ) in the high-fat pigs was observed (P < 0:05). The ACC1 and FABP4 (fat acidbinding proteins 4) showed higher levels in the high-fat pigs than the low-fat pigs with no significant difference (P > 0:05, see Figure 2(b)).

Oxidative Stress Levels in Pigs with Different Fat
Deposition. To evaluate whether oxidative stress functions in the lipid metabolism, the key markers of oxidative stress, including MDA, GSH-Px, and SOD, were further analyzed. The low-fat pigs had decreased MDA content and increased GSH-Px activity compared to the high-fat pigs (see Figure 3, P < 0:05).

3.4.
Profiling of the Colon Metagenome in Pigs. Next, the composition and diversity of the colonic microbiota between the two groups were investigated by metagenomic sequencing. For the construction of libraries, we pooled 10 samples 3 Oxidative Medicine and Cellular Longevity of luminal DNA from Jinhua pigs. Each DNA pool included an average of 12.84 Gb of raw data (9.94-19.00 Gb). Clean data ranged from 9.33 Gb to 17.90 Gb after removing lowquality reads and host contamination. All short sequence data was assembled by Megahit [25] (https://github.com/ voutcn/megahit). Ten DNA pool samples had a total number of contigs ranging from 243,029 to 710,018 with a length greater than 500 bp. METAProdigal [26] (http://prodigal  Oxidative Medicine and Cellular Longevity .ornl.gov/) predicted the open reading frame (ORF) of each contig. A total of 5,617,408 ORFs that were longer than 60 bp were obtained in the pool of ten samples (see Table 1).

Comparison of Microbial Domains between High-and
Low-Fat Pigs. Nonmetric multidimensional scaling (NMDS) revealed a robust separation between high-and low-fat pigs (see Figure 4(a)). The microbial structure in the colons of the two groups was compared. The relative abundance of Archaea in the high-fat group was obviously higher than that of low-fat pigs (P < 0:05) while eukaryota and viruses were not significantly different (P > 0:05, see Figure 4(b)).
We then compared the relative abundance of Archaea bacterial genera between the colon microbiomes of the two groups with distinct fatness. Two genera (see Figure 4(c)) and seven species (see Figure 4(d)) were identified to give a significant difference in the relative abundances between the two groups (P < 0:05), which all belong to the methanogen genus.

Significant Difference in Bacterial Community
Compositions of Pigs with Distinct Fatness Phenotypes. As shown in Supplementary Figure S1, the low-fat pigs exhibited a higher α diversity of gut microbiota than the high-fat pigs. Further studies were performed on the microbial community structures between the two groups with distinct fatness phenotypes in the colon. Bacteroidetes were found in significantly higher abundance in the colons of low-fat pigs than in high-fat pigs (P < 0:05, Supplementary Figure 2). At the genus level, 13 genera were significantly different between the two groups. Prevotella and Bacteroides were significantly more abundant in the colons of low-fat pigs (P < 0:05). Prevotella relative abundance was 2.979% and 5.659% between the two groups, respectively, while Bacteroides relative abundance was 1.300% and 1.935%, respectively. Additionally, the relative abundance of Prevotella and Bacteroides was above 1.0%, whereas most other genera were below 1.0% in the colon. It is worth noting that Acidaminococcus was significantly enriched in the low-fat pigs (P < 0:05, see Figure 5(a)). At the species level, we discovered 33 bacterial species giving different enrichments between the two groups. Among these species, 12 species were enriched in the highfat pigs, while the other 21 species were more abundant in the low-fat pigs (see Figure 5(b)). Furthermore, we observed that ten of the 12 species with higher abundances in the high-fat pigs are Firmicutes and Tenericutes. The bacteria abundant in low-fat pigs have been correlated with fiber fermentation and butyrate production, namely, Ruminococcus sp. AF12-5 [27], Faecalibacterium sp. OF04-11AC [28], and Oscillibacter sp. CAG:155 [29].  The functional capacity of the gut microbiome between the two groups was further analyzed using KEGG annotation. At Level 2 of KEGG function analysis, "Glycan biosynthesis and metabolism" was the most significantly different metabolic function between the two groups (see Figure 6(a)). At Level 3, "Methane metabolism," "Other glycan degradation," and "RNA polymerase" were found to be the significant markers in the pigs with a distinct fatness phenotype. More importantly, the relative abundance in the function pathway profile of "Other glycan degradation" was higher in the low-fat pigs than in the highfat pigs (see Figure 6(b)).
To further study the colonic microbiota's functional capacity, we analyzed the enzymes that break down glycans (CAZymes: Carbohydrate-Active Enzymes) [30]. There were significant differences in the enrichments of 17 CAZyme families between the two groups of pigs (P < 0:05, see Figure 6(c)). We found 2 enriched CAZyme genes, which function to degrade carbohydrates such as cellulose,    3.8. Changes in Colonic SCFA Levels in the Pigs with Different Fatness Phenotypes. To examine whether pigs with different fatness phenotypes would alter the SCFA content in the colon, we quantified the absolute concentrations of the total SCFAs, namely, propionate, acetate, and butyrate. In comparison to the high-fat pigs, the low-fat pigs' colons had much higher levels of acetate, propionate, and butyrate (see Figure 7(a)). Furthermore, the colon of low-fat pigs had a relatively higher abundance of butyryl-CoA acetate-CoA transferase (see Figure 7(b)), which functions to modulate the butyrate production in the colonic microbiome.
To reveal the potential linkage between gut microbiota and SCFAs, we correlated between differential gut microbiota and SCFA levels.       Oxidative Medicine and Cellular Longevity were negatively correlated with acetate, propionate, and butyrate levels. These results revealed the changed composition and function of the gut microbiota might well have contributed to the SCFA production in the colon (see Figure 7(c)).

Discussion
The gut microbiota is considered essential in the utilization of nutrients and energy and the maintenance of healthy status in farm animals. Thus, the gut microbiota is regarded as an important component affecting the growth performance and development of pigs. Nowadays, numerous studies investigating the structure of the gut microbial community in pigs by metagenomics have been proposed to reveal the association between gut bacterial species and porcine fatness in different breeds [31]. We explored the interaction among the gut microbiota, lipogenic features, and oxidative stress in pigs with two different fat phenotypes. Furthermore, we tried to implicate the possible relationship between the gut microbiome and fat deposition in pigs using metagenomic analysis.
The liver and adipose tissue are the main parts of the body that deposit, metabolize, and transport fat. ACC1, FAS, SREBP-1, and ChREBP are the key genes in de novo fat synthesis [32]. The gene expression of ACC1 and SREBP-1 was significantly higher in the liver of the highfat pigs than in the low-fat pigs, indicating that high-fat pigs might have a stronger ability to synthesize fat than low-fat pigs. The abdomen and intestines are the main sites of fat accumulation. The "master regulator" of adipogenesis is PPARγ, required for fat cell production [33]. The upregulation of PPARγ, LPL, and FAS in the high-fat pigs indicated enhanced lipid synthesis in the abdomen. Collectively, the imbalance in lipid metabolism might contribute to the abnormal fat accumulation in pigs.
On the other hand, the development of the pig industry would be hampered by oxidative stress due to oxidative stress directly limiting the growth performance and health of pigs. MDA is a biomarker for free radical species-related damage [34]. In the present study, MDA levels in serum were obviously higher in high-fat pigs than in low-fat pigs. Increased oxidative stress associated with fat consumption may alter the bacterial composition and the expression of lipogenic genes. GSH-Px is a critical enzyme that catalyzes hydrogen peroxide decomposition, which would protect the structure and function of the cell membrane [35]. GSH-Px levels were significantly lower in high-fat pigs. These findings show that high-fat pigs are more susceptible to oxidative stress than low-fat pigs, and hence more susceptible to obesity.
Consistent with our earlier investigation of 16S rRNA gene sequencing in pigs [23], the bacterial community of the two groups showed taxonomic discrepancies. It implies that bacteria contribute more to host fatness than other microbial kingdoms. Bacteria, not other microbial kingdoms, are responsible for the majority of the breakdown and fermentation of feed biopolymers [36]. Interestingly, the majority of the species with significantly increased abun-dances in the low-fat pigs were from the Bacteroides genus, which is one of the most frequent core genera in the pig colony. Bacteroides was able to utilize the fermentation of dietary fibers to produce acetate, propionate, and butyrate [24]. Prevotella and Bacteroides were found to be significantly less abundant in the colons of high-fat pigs than in low-fat pigs. The more Bacteroides-Prevotella-Porphyromonas there are in developing pigs, the better their ability to ferment polysaccharides to SCFAs [37]. Furthermore, Bacteroides-Prevotella was negatively correlated with inflammation and fat mass development in diet-induced obese mice [38]. Taken together, these results suggest that enhanced SCFA production in the low-fat pigs might be linked with the higher abundance of Prevotella and Bacteroidetes, which further alleviate host lipid accumulation.
Accordingly, our findings indicated that colonic SCFA levels were significantly higher in pigs with low-fat in comparison to pigs with high-fat traits. Butyrate could directly activate AMP kinase to prevent mice from having excess fat deposition in the liver [39]. The mice transplanted from lean cotwin's fecal microbiota showed lower fat storage in adipose tissue compared to the mice transplanted from obese cotwin's fecal microbiota [40]. Importantly, we found correlations between SCFAs and differential gut microbiota in pigs. The phylum Firmicutes and its subordinate species were related to reduced SCFAs, whereas the phylum Bacteroidetes and its subordinate species were related to increased SCFAs. All SCFAs showed synchronous correlations with the differential gut microbiota, suggesting that the differential gut microbiota and the SCFAs they produced might work synchronously in fat deposition.
Based on the metagenomics, KEGG analysis revealed the gut microbiome functional changes in the two groups. The metagenomes of the high-fat pigs had a significantly higher enrichment of methane metabolism and RNA polymerase. Methane metabolism, in particular, has been reported to produce methane from hydrogen with carbon dioxide, acetate, and various methyl metabolites [41]. Bacterial NADH dehydrogenases could be suppressed by the improvement of H 2 , resulting in a decrease in ATP output [42]. Methane is able to slow down the gut transport rate, which allows nutrients to stay in contact with the gut for longer, giving more time for nutrients and energy to be absorbed. This might explain why the high-fat pigs gained more weight than the low-fat pigs.
Since glycan degradation was found to be significantly enriched in the functional microbiome of low-fat pigs, we further determined that CAZymes shifted between the two groups in the present study. The results showed that GH, CE, PL, AA, and CBM, which exert a function of carbohydrate degradation, were enriched in low-fat pigs, implying that low-fat pigs might have a better ability to deconstruct complex substrates. For example, the colon of the low-fat pigs showed a higher relative abundance of GH28, GH76, GH81, and GH106, which are polygalacturonosidase [43], α-1,6-mannanase [44], β-1,3-glucanase [45], and α-L-rhamnosidase [46], respectively.
In contrast to the low-fat pigs, the higher abundance of GTs in the high-fat pigs showed that the colon microbiomes 10 Oxidative Medicine and Cellular Longevity of high-fat pigs might have a better capability to utilize hydrolytic products to produce SCFAs, thus providing more energy to high-fat pigs. Generally speaking, feed-efficient animals generate more SCFAs and less methane [36]. The higher level of SCFAs and lower abundance of methanogenic functions in the colon of low-fat pigs suggest that low-fat pigs may be more feed-efficient than high-fat pigs. To confirm our assumptions, further research on feed efficiency and methane emissions is required in the future. Notably, the higher relative abundance in genus Methanobrevibacter and various species, namely, M. millerae, M. gottschalkii, M. sp. YE315, M. thaueri, M. smithii, M. sp. A27, and M. oralis, was observed in the colons of high-fat animals, which suggests that more methane might be produced, resulting in more efficient fat deposition in the high-fat pigs [47]. The metagenome gave us a chance to study the gut microbiome at different kingdom levels, such as eukaryote and virus, except for bacteria and Archaea. The interactions between bacteria and eukaryotes or viruses might also change the status of host fat deposition, because our main concern was not with eukaryotes or viruses in the present study. Additional research in the future is needed.
In summary, we found that there were abundant methanogenic Archaea and relatively lower SCFAproducing bacteria in the colons of high-fat pigs compared to low-fat pigs. The increased methane produced by the Archaea and the decrease of SCFAs would promote fat deposition in the liver and abdomen, which is manifested by the high expression of adipogenic genes in the liver (ACC1 and SREBP1) and abdominal fat (FAS, PPARγ, and LPL), furthering the oxidative stress injury (see Figure 8).

Conclusions
In summary, the high oxidative stress and lipid metabolism dysbiosis in the high-fat pigs could contribute to the fat deposition in Jinhua pigs. Host fat deposition is influenced by both the methanogenesis functions of Archaea and the short-chain fatty acids produced by bacteria. In addition, we discovered 17 CAZyme families and 3 KEGG pathways 11 Oxidative Medicine and Cellular Longevity (Level 3) with distinct enrichments in the high-fat and lowfat pigs. The present study would give a deep insight into how gut microbiomes influence fat deposition in pigs and provide a hint for improving growth performance and fatness traits by manipulating gut microbiomes.