MicroRNA-193a-5p Regulates the Synthesis of Polyunsaturated Fatty Acids by Targeting Fatty Acid Desaturase 1 (FADS1) in Bovine Mammary Epithelial Cells

Cardiovascular diseases (CVDs) are seriously threatening to human life and health. Polyunsaturated fatty acids (PUFAs) are known for their role in preventing CVDs. It is beneficial to population health to promote the content of PUFAs in bovine milk. In recent years, limited research based on molecular mechanisms has focused on this field. The biological roles of numerous microRNAs (miRNAs) remain unknown. In this study, a promising and negatively correlated pair of the miRNA (miRNA-193a-5p) and a fatty acid desaturase 1 (FADS1) gene are identified and screened to explore whether they are potential factors of PUFAs’ synthesis in bovine milk. The targeted relationship between miRNA-193a-5p and FADS1 in bovine mammary epithelial cells (BMECs) is demonstrated by dual luciferase reporter assays. qRT-PCR and western blot assays indicate that both the expression of mRNA and the protein FADS1 show a negative correlation with miRNA-193a-5p expression in BMECs. Also, miR-193a-5p expression is positively correlated with the expression of genes associated with milk fatty acid metabolism, including ELOVL fatty acid elongase 6 (ELOVL6) and diacylglycerol O-acyltransferase 2 (DGAT2). The expression of fatty acid desaturase 2 (FADS2) is negatively correlated with miR-193a-5p expression in BMECs. The contents of triglycerides (TAG), eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA) and docosahexaenoic acid (DHA) have a significant positive correlation with the expression of FADS1 and a significant negative correlation with the expression of miR-193a-5p in BMECs. For the first time, this study confirms that miRNA-193a-5p regulates PUFAs metabolism in BMECs by targeting FADS1, indicating that miRNA-193a-5p and FADS1 are underlying factors that improve PUFAs content in bovine milk.


Introduction
Fats are some of the major nutrients in bovine milk and are mainly triglycerides (TG) containing saturated fatty acids (SFAs), unsaturated fatty acids (UFAs) and several minor compounds [1,2]. Although milk fat rich in SFAs has been proposed as a potential risk factor for cardiovascular diseases (CVDs) [3,4], accumulating evidence shows the concern is unnecessary, whether it is for people who consume low-fat milk or high-fat milk [5,6]. A total of 17 fatty acids found in bovine milk, such as trans-palmitoleic acid and conjugated linoleic acid (LA), may prevent excessive obesity [7]. Duda et al. [8] found that dietary supplementation with n-3 Polyunsaturated fatty acids (PUFAs) attenuated pressure

Ethics Statement
Our experiment obtained ethics approval from the Institutional Animal Care and Use Committee (IACUC) of the Yangzhou University Animal Experiments Ethics Committee (Permit Number: SYXK (Su) IACUC 2016-0019).

Transcriptome Sequencing
Three Holstein dairy cows (A, B, C) in their second lactation period without a history of illness were selected from the farm of Yangzhou University. The three dairy cows were feed with total mixed ration (TMR), and its components are shown in Table S1 [24]. Mammary glands of dairy cows were collected by surgical methods at early lactation (postpartum day 30, 30 d) and mid lactation (postpartum day 180, 180 d) [25,26]. Total RNA was extracted using the mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA). RNA integrity was assessed on an Agilent 2100 Bioanalyzer (Agilent, Santa Clara, CA, USA), and samples with an RNA Integrity Number (RIN) ≥7 were used in transcriptome analysis.
The small RNA (sRNA) libraries were constructed according to Illumina ® TruSeq™ (Illumina, San Diego, CA, USA) Small RNA Sample Preparation protocol. Briefly, 3 adaptor and 5 adaptors were ligated to the total RNA. Then, the cDNA was obtained using reverse transcription with SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA). Finally, the cDNA was amplified by PCR with DNA polymerase to generate small RNA libraries. The libraries were sequenced on Illumina Hiseq 2500 at a sequence read length of 1 × 36 bp to obtain the original data. The quality of raw reads was elevated using the FastQC online tool (http://www.bioinformatics. babraham.ac.uk/projects/fastqc/) [27].
Datasets were obtained from the NCBI Sequence Read Archive (SRA) (https://www. ncbi.nlm.nih.gov/sra) with accessing numbers SRS6946983, SRS6946984 and SRR12149783, and the Genome Sequence Archive in the BIG Data Center, Beijing Institute of Genomics (BIG), Chinese Academy of Sciences, under accession number CRA002742, and are publicly accessible at http://bigd.big.ac.cn/gsa.

Identifition of Target-Related miRNA and DEG
miRNA expression was calculated as transcripts per million (TPM = reads of each miRNA alignment/reads of samples' total alignment × 10 6 ) [30]. Principal component analysis (PCA) was performed to explore transcripts' distribution. The calculation of DE miRNAs was realized by the DEseq approach. DE miRNAs were defined as those with a fold change >2 and p-value < 0.05.
In addition, qRT-PCR was used for validation of the sequencing data based on the expression of DE miRNAs and DEGs. A PrimeScript™ RT Reagent Kit with a gDNA Eraser (Takara, Kyoto, Japan) was applied for the synthesis of cDNA. Expression of the DE miRNAs/DEGs was detected using the TB Green ® Premix Ex Taq™ II Kit (Takara, Kusatsu, Shiga, Japan)/Hairpin-it microRNA and the U6 snRNA Normalization RT-PCR Quantitation kit (GenePharma, Shanghai, China) [33]. All reactions were performed in triplicates using a Light Cycler ® 480 System (Roche Applied Science, Indianapolis, IN, USA). DE miRNAs/mRNAs were normalized for bovine U6/β-actin, and the primers used for qRT-PCR are shown in Tables S2 and S3. Relative expression was calculated using the 2 −∆∆Ct method for all samples [34,35].
A TargetScan program (http://www.targetscan.org) was used to identify the target gene of DE miRNAs. Only mRNAs showing a negative correlation with DE miRNAs were considered putative targets. A network map was built to investigate the interactions between DE miRNAs and their targeted DEGs. To find the target-related miRNAs and DEGs with the biological function of PUFAs synthesis, gene ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis were respectively achieved by DAVID 6.8 (https://david.ncifcrf.gov/) [36] and KOBAS 3.0 (http://kobas.cbi. pku.edu.cn/kobas3) [37]. A p-value < 0.05 was the threshold for identifying significantly GO terms and KEGG pathways. In this study, miR-193a-5p and FADS1 were considered as key DE miRNA and mRNA with the biological function of PUFAs synthesis.

Cell Culture and Transfection
BMECs were obtained from the Animal Genetics Laboratory of Yangzhou University. They were cultured in six-well plates with DMEM/F12 (Gibco, Waltham, MA, USA) complete media that contained 10% fetal bovine serum (FBS; Gibco), 50 U/mL penicillin/mL streptomycin, 5 mg/mL bovine insulin, 10 ng/mL EGF-1 (Gibco) and 0.25 mmol/L hydrocortisone. The plates were placed in an incubator with 5% CO 2 at 37 • C. When BMECs grew to approximately 80% per well, the cells were transfected with miR-193a-5p mimic (100 nM final concentration) and inhibitor (200 nM final concentration) using the Lipofectamine 2000 Transfection Kit (Sigma, St. Louis, MO, USA). Transfection of BMECs was carried out in Opti-MEM medium (Gibco, Waltham, MA, USA) for 6 h. Then, Opti-MEM medium was replaced with DMEM/F12 complete media for the next 48 h of culture. The sequence of miR-193a-5p mimic, inhibitor and negative control were shown in Table S4. Each transfection held three replicates.

Western Bolt
To obtain the total protein, the transfected BMECs were lysed in RIPA buffer containing 1% PMSF (Solarbio, Beijing, China). Protein concentrations were assessed using a BCA Protein Assay Kit (Solarbio, Beijing, China). Equal amounts of protein in each well were separated on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and then transferred to a PVDF membrane (Amersham Biosciences, Buckinghamshire, UK). Finally, the total protein was probed with the primary antibodies-monoclonal rabbit anti-FADS1 (Cell Signaling Technology, Beijing, China) and monoclonal mouse anti-βactin (Proteintech Gronup, Shanghai, China). The secondary antibody was a polyclonal goat antirabbit HRP-conjugated IgG (Tiangen, Sichuan, China). All antibodies were used in adherence to the instructions. Signals were measured on the chemiluminescent ECL western blot system (Pierce, Appleton, WI, USA).

Triglycerides Content Assay
Triglyceride was assayed using an enzymatic triglyceride assay kit (Applygen, Beijing, China). The BMECs were lysed after 48 h of transfection. The mixture was allowed to rest for 10 min, and the supernatant was transferred into a sterile 1.5 mL micro-centrifuge tube and then heated for 10 min in a 70 • C water bath. After centrifugation (5 min at 2000 rpm), the supernatant was collected for an enzymatic assay [38,39]. According to the manufacturer's instructions, absorbance was detected at 550 nm in a microplate reader (Tecan, Switzerland).

Fatty Acid Profiles Assay
Intracellular fatty acids were assayed by referring to the method of Zhou et al. [40]. In detail, BMECs were transfected with miR-193a-5p mimic, inhibitor and negative control. A total of 100 mg BMECs mixed with 2 mg of methanol containing 0.25% sulfuric acid were sonicated by ultrasound and then incubated at 80 • C for 1 h to methylate the fatty acids. The solution, returned to room temperature, was mixed with 2 mL hydrochloric acid solution (0.1 M) and then 800 µL of n-hexane was added, obtaining a mixed solution. The mixed solution was centrifuged at 900× g for 5 min, and the supernatant was transferred to a siliconized glass tube. The supernatant was added to 0.5 g anhydrous sodium sulfate and vortexed. After centrifugation at 13,800× g for 5 min, the supernatant and anhydrous sodium sulfate were separated from the solution. The supernatant was applied to the detected fatty acid composition and content with gas chromatography−mass spectrometry (Thermo Fisher, Waltham, MA, USA) analysis of fatty acid composition and content. The relative content of fatty acid was evaluated using the percentage of peak area for individual fatty acids.

Statistical Analysis
Statistical analysis Software (SAS) 9.4 (SAS Institute, Cary, NC, USA) was used for statistical analyses. Data were presented as mean ± standard error (SE) from triplicate experiments. p < 0.05 indicated significant differences between the groups.

Identification of Differentially Expressed miRNAs and mRNAs
A total of six sRNA/cDNA libraries were established and sequenced to identify DE miRNAs/mRNAs based on the mammary glands of dairy cows during early and midlactation. The six sRNA/cDNA libraries produced 916.29/3913.31 million (M) raw reads (Tables S8 and S9). After removing the unqualified reads, a total of 847.55/3859.12 M clean reads were obtained from the sRNA/cDNA libraries and sequenced on Illumina Hiseq 2500. In each sRNA library, the Q20 percentage was higher than 99.88%. In each cDNA library, the Q30 percentage was greater than 96.19%, and the GC content ranges from 47.50% to 50.00%. Consequently, the sequencing results are reliable and could be used for subsequent analysis.
PCA based on the sRNA/cDNA libraries was performed, respectively generating two clusters: early and mid-lactation ( Figure 1). A cluster included miRNAs/mRNAs from different dairy cows who were in the same lactation stage, indicating that the main differences of the miRNA/mRNA expression profiles appeared at different stages of lactation. DE miRNAs/mRNAs were identified by comparing the miRNA/mRNA expression profiles at two time points. Subsequently, a total of 13 DE miRNAs and 632 DEGs were identified by fold change >2 and p-value < 0.05. The 13 DE miRNAs were 6 upregulation miRNAs and 7 downregulation miRNAs in mid-lactation, compared with early lactation (Figure 2a). The 632 DEGs include 332 upregulation genes and 300 downregulation genes in mid-lactation compared with early lactation (Figure 2b).
Additionally, the validation of sequencing data used qRT-PCR based on the expression of DE miRNAs and DEGs. The qRT-PCR results demonstrate that the expression of miR-223, miR-2346, miR-338, FADS1, PHLDA2 and SLC13A5 are upregulation in mid-lactation verse early lactation, while miR-193a-5p, miR-124a, ACSM1, ACPP, CYP4B1, ACACB and CCDC3 are downregulation in mid-lactation verse early lactation (Tables S10 and S11). The results of qRT-PCR were in accordance with transcriptome sequencing data.

Identification of Differentially Expressed miRNAs and mRNAs
A total of six sRNA/cDNA libraries were established and sequenced to identify DE miRNAs/mRNAs based on the mammary glands of dairy cows during early and midlactation. The six sRNA/cDNA libraries produced 916.29/3913.31 million (M) raw reads (Tables S8 and S9). After removing the unqualified reads, a total of 847.55/3859.12 M clean reads were obtained from the sRNA/cDNA libraries and sequenced on Illumina Hiseq 2500. In each sRNA library, the Q20 percentage was higher than 99.88%. In each cDNA library, the Q30 percentage was greater than 96.19%, and the GC content ranges from 47.50% to 50.00%. Consequently, the sequencing results are reliable and could be used for subsequent analysis.
PCA based on the sRNA/cDNA libraries was performed, respectively generating two clusters: early and mid-lactation ( Figure 1). A cluster included miRNAs/mRNAs from different dairy cows who were in the same lactation stage, indicating that the main differences of the miRNA/mRNA expression profiles appeared at different stages of lactation. DE miRNAs/mRNAs were identified by comparing the miRNA/mRNA expression profiles at two time points. Subsequently, a total of 13 DE miRNAs and 632 DEGs were identified by fold change >2 and p-value < 0.05. The 13 DE miRNAs were 6 upregulation miR-NAs and 7 downregulation miRNAs in mid-lactation, compared with early lactation (Figure 2a). The 632 DEGs include 332 upregulation genes and 300 downregulation genes in mid-lactation compared with early lactation (Figure 2b).
Additionally, the validation of sequencing data used qRT-PCR based on the expression of DE miRNAs and DEGs. The qRT-PCR results demonstrate that the expression of miR-223, miR-2346, miR-338, FADS1, PHLDA2 and SLC13A5 are upregulation in mid-lactation verse early lactation, while miR-193a-5p, miR-124a, ACSM1, ACPP, CYP4B1, ACACB and CCDC3 are downregulation in mid-lactation verse early lactation (Tables S10 and S11). The results of qRT-PCR were in accordance with transcriptome sequencing data.

MiR-193a-5p Specific Targeting of FADS1
Given that the function of the fatty acid metabolism of FADS1 remains to be fully clarified in BMECs, we selected the miR-193a-5p-FADS1 candidate pair and attempted to verify their biological function and regulatory relationships. The BMECs were transfected with miR-193a-5p mimic, inhibitor and miRNA negative control. After 48 h, the expression of miR-193a-5p was 55 times greater in the mimic group compared with the negative control (Figure 5a). In contrast, compared with the negative control, the expression of miR-193a-5p was more than 70% lower in BMECs whose miRNA-193a-5p was silenced (Figure 5b). The results were successful for the transfection of miR-193a-5p mimic and inhibitor. Figure 5c-e shows the expression of FADS1 in BMECs transfected with mimic and an inhibitor of miR-193a-5p. Both mRNA and protein results indicate that FADS1 expression is upregulated in BMECs transfected by the inhibitor of miR-193a-5p, while the overexpression of miR-193a-5p leads to the downregulation of FADS1 expression in BMECs.
Dual luciferase reporter gene assay evaluated whether miR-193a-5p targets FADS1 in BMECs. We predicted the 3′ UTR region of FADS1 in dairy cows using TargetScan and constructed wild-type (pmirGLO-FADS1-WT) and mutant (pmirGLO-FADS1-MUT) vector plasmids. The two types of vectors were co-transfected with miR-193a-5p mimic into BMECs. In the group of co-transfections with pmirGLO-FADS1-WT and miR-193a-5p mimic, luciferase activity significantly decreased to 40% of the negative control group (Figure 5f). In the group of co-transfections with pmirGLO-FADS1-MUT and miR-193a-5p mimic, luciferase activity had no statistical significance compared with the negative control group (Figure 5f). These results confirm that miR-193a-5p regulates FADS1 expression by binding to the predicted target site in the 3′ UTR region of FADS1 in dairy cows.

MiR-193a-5p Specific Targeting of FADS1
Given that the function of the fatty acid metabolism of FADS1 remains to be fully clarified in BMECs, we selected the miR-193a-5p-FADS1 candidate pair and attempted to verify their biological function and regulatory relationships. The BMECs were transfected with miR-193a-5p mimic, inhibitor and miRNA negative control. After 48 h, the expression of miR-193a-5p was 55 times greater in the mimic group compared with the negative control ( Figure 5a). In contrast, compared with the negative control, the expression of miR-193a-5p was more than 70% lower in BMECs whose miRNA-193a-5p was silenced (Figure 5b). The results were successful for the transfection of miR-193a-5p mimic and inhibitor. Figure 5c-e shows the expression of FADS1 in BMECs transfected with mimic and an inhibitor of miR-193a-5p. Both mRNA and protein results indicate that FADS1 expression is upregulated in BMECs transfected by the inhibitor of miR-193a-5p, while the overexpression of miR-193a-5p leads to the downregulation of FADS1 expression in BMECs.
Dual luciferase reporter gene assay evaluated whether miR-193a-5p targets FADS1 in BMECs. We predicted the 3 UTR region of FADS1 in dairy cows using TargetScan and constructed wild-type (pmirGLO-FADS1-WT) and mutant (pmirGLO-FADS1-MUT) vector plasmids. The two types of vectors were co-transfected with miR-193a-5p mimic into BMECs. In the group of co-transfections with pmirGLO-FADS1-WT and miR-193a-5p mimic, luciferase activity significantly decreased to 40% of the negative control group (Figure 5f). In the group of co-transfections with pmirGLO-FADS1-MUT and miR-193a-5p mimic, luciferase activity had no statistical significance compared with the negative control group (Figure 5f). These results confirm that miR-193a-5p regulates FADS1 expression by binding to the predicted target site in the 3 UTR region of FADS1 in dairy cows. . Data are presented as mean ± standard error (SE). * represents p < 0.05, indicating significant differences between the groups.

Triglycerides Content Analysis
After confirming the regulation relationships between miR-193a-5p and FADS1, we sought to explore the special biological function of the miR-193a-5p-FADS1 candidate pair in BMECs. As triglycerides (TAG) is the main composition of milk fat and the main carrier of milk fatty acids, we examined the TAG levels in BMECs transfected with miR-193a-5p mimic/inhibitor. Compared with the controls, the miR-193a-5p mimics reduced the TAG level to 20% (p < 0.05) (Figure 6a). In contrast, TAG levels increased to 150% when miR-193a-5p was inhibited in BMECs (Figure 6b).

Triglycerides Content Analysis
After confirming the regulation relationships between miR-193a-5p and FADS1, we sought to explore the special biological function of the miR-193a-5p-FADS1 candidate pair in BMECs. As triglycerides (TAG) is the main composition of milk fat and the main carrier of milk fatty acids, we examined the TAG levels in BMECs transfected with miR-193a-5p mimic/inhibitor. Compared with the controls, the miR-193a-5p mimics reduced the TAG level to 20% (p < 0.05) (Figure 6a). In contrast, TAG levels increased to 150% when miR-193a-5p was inhibited in BMECs (Figure 6b).

Partial Rescue of the Triglycerides Reduction
The experiments have indicated that miR-193a-5p regulates FADS1 expression by binding to the predicted target site in the 3′ UTR region of FADS1, and miR-193a-5p inhibits the synthesis of cellular TAG. A rescue experiment was performed to verify whether miR-193a-5p induces the increase of TAG levels by targeting FADS1. Rescues of siRNA-FADS1 increase the TAG levels in BMECs. siRNA-FADS1 partially alleviated the increase in TAG content in BMECs treated with miR-193a-5p inhibitor (Figure 6d). The results demonstrate that miR-193a-5p induces an increase in TAG levels by targeting FADS1.

Fatty Acid Profiles Analysis
To further understand the role of miRNA-193a-5p in the fatty acid metabolism of BMECs, we detected changes in the fatty acid content of BMECs transfected with miR-193a-5p mimic/inhibitor. BMECs transfected with miR-193a-5p inhibitor show an increase in concentration of PUFAs such as eicosapentaenoic acid (EPA, C20:5n-3), docosapentaenoic acid (DPA, C22:5n-3) and docosahexaenoic acid (DHA, C22:6n-3) compared with the control group (Table 1). BMECs transfected with miR-193a-5p mimic show a decrease in concentration of PUFAs such as EPA, DPA and DHA compared with the control group (Figure 7).

Partial Rescue of the Triglycerides Reduction
The experiments have indicated that miR-193a-5p regulates FADS1 expression by binding to the predicted target site in the 3 UTR region of FADS1, and miR-193a-5p inhibits the synthesis of cellular TAG. A rescue experiment was performed to verify whether miR-193a-5p induces the increase of TAG levels by targeting FADS1. Rescues of siRNA-FADS1 increase the TAG levels in BMECs. siRNA-FADS1 partially alleviated the increase in TAG content in BMECs treated with miR-193a-5p inhibitor (Figure 6d). The results demonstrate that miR-193a-5p induces an increase in TAG levels by targeting FADS1.

Fatty Acid Profiles Analysis
To further understand the role of miRNA-193a-5p in the fatty acid metabolism of BMECs, we detected changes in the fatty acid content of BMECs transfected with miR-193a-5p mimic/inhibitor. BMECs transfected with miR-193a-5p inhibitor show an increase in concentration of PUFAs such as eicosapentaenoic acid (EPA, C20:5n-3), docosapentaenoic acid (DPA, C22:5n-3) and docosahexaenoic acid (DHA, C22:6n-3) compared with the control group (Table 1). BMECs transfected with miR-193a-5p mimic show a decrease in concentration of PUFAs such as EPA, DPA and DHA compared with the control group (Figure 7).

Figure 7.
Effect of miR-193a-5p mimic and inhibitor on the composition of n-3 fatty acids in BMECs (n = 6). * represents p < 0.05, indicating significant differences between the groups.

Discussion
Research showed that FADS1 encodes rate-limiting enzymes in the endogenous formation of PUFAs [41,42]. EPA and DHA are endogenous produced in a pathway where essential n-3 fatty acids and α-linolenic acid are desaturated and elongated. In this pathway, FADS1 plays a critical role [43,44]. Approximately 17.9 million people die from CVD every year [45,46]. Increasing evidence shows that nutrition plays a pivotal role in the development of chronic diseases, especially CVD [47]. Diet and lifestyle are the key risk factors for CVD prevention and have been the focus of intense research [46]. Dairy products, as an integral part of human nutrition, contain n-3 fatty acids, usually at concentrations below biological requirements. Thus, in recent years, some research has been devoted to increasing the content of PUFAs in bovine milk based on genome, transcription and post-transcriptional levels [48][49][50]. In this study, we confirmed that miR-193a-5p directly binds to the FADS1 3′-UTR in dairy cows. qRT-PCR and western bolt assays indicated that overexpression of miR-193a-5p downregulated expressions of FADS1 regardless of mRNA or protein levels, and the inhibition of miR-193a-5p had the opposite effect. In addition, the content of TAG, EPA, DPA and DHA has a significant positive correlation with the expression of FADS1 and a significant negative correlation with the expression of miR-193a-5p in BMECs. These results confirm that miR-193a-5p regulates the synthesis of PUFAs by targeting FADS1 in BMECs. Increasing PUFAs intake is beneficial for global health and can prevent human beings from metabolic, cardiovascular, cognitive and developmental health conditions [51][52][53]. From a global health perspective, the downregulation of miRNA-193a-5p and the upregulation of FADS1 are beneficial.
Polymorphisms in FADS1-and FADS2-altered fatty acid profiles or the activity of FADS1-and FADS2-. He et al. [54] identified strong associations between SNPs in the fatty acid desaturase (FADS) region and multiple PUFAs, suggesting that genetic variants in the FADS region were major genetic modifiers that could regulate fatty acid metabolism through epigenetic gene regulation. Koletzko et al. [55] found that genotype distribution differs markedly among ethnicities, apparently reflecting an evolutionary advantage of genotypes in activating LC-PUFAs synthesis when the introduction of agriculture provided diets rich in LA but with little arachidonic acid (AA) and EPA. Vegetarians with a TT genotype of the rs174547 in FADS1 gene had higher odds of developing metabolic syndrome (MetS), larger waist circumference (WC), higher blood pressure (BP), and lower level of HDL-c, which suggested that the dietary intake of LA might interact with rs174547 in the FADS1 gene and affect HDL-c levels [56]. Joshi et al. [57] demonstrated that an increase the FADS1 expression of vegetarian pregnant women during pregnancy might maintain AA, EPA and DHA levels to ensure the overall LC-PUFA levels of the neonate. Populations had an increase in EPA in plasma when intake of flaxseed oil rich in α-linolenic acids (ALA) increased beyond that of major allele homozygotes consuming a typical "western" diet [58]. Individuals homozygous for the minor allele of FADS1/FADS2 had a

Discussion
Research showed that FADS1 encodes rate-limiting enzymes in the endogenous formation of PUFAs [41,42]. EPA and DHA are endogenous produced in a pathway where essential n-3 fatty acids and α-linolenic acid are desaturated and elongated. In this pathway, FADS1 plays a critical role [43,44]. Approximately 17.9 million people die from CVD every year [45,46]. Increasing evidence shows that nutrition plays a pivotal role in the development of chronic diseases, especially CVD [47]. Diet and lifestyle are the key risk factors for CVD prevention and have been the focus of intense research [46]. Dairy products, as an integral part of human nutrition, contain n-3 fatty acids, usually at concentrations below biological requirements. Thus, in recent years, some research has been devoted to increasing the content of PUFAs in bovine milk based on genome, transcription and post-transcriptional levels [48][49][50]. In this study, we confirmed that miR-193a-5p directly binds to the FADS1 3 -UTR in dairy cows. qRT-PCR and western bolt assays indicated that overexpression of miR-193a-5p downregulated expressions of FADS1 regardless of mRNA or protein levels, and the inhibition of miR-193a-5p had the opposite effect. In addition, the content of TAG, EPA, DPA and DHA has a significant positive correlation with the expression of FADS1 and a significant negative correlation with the expression of miR-193a-5p in BMECs. These results confirm that miR-193a-5p regulates the synthesis of PUFAs by targeting FADS1 in BMECs. Increasing PUFAs intake is beneficial for global health and can prevent human beings from metabolic, cardiovascular, cognitive and developmental health conditions [51][52][53]. From a global health perspective, the downregulation of miRNA-193a-5p and the upregulation of FADS1 are beneficial.
Polymorphisms in FADS1and FADS2-altered fatty acid profiles or the activity of FADS1and FADS2-. He et al. [54] identified strong associations between SNPs in the fatty acid desaturase (FADS) region and multiple PUFAs, suggesting that genetic variants in the FADS region were major genetic modifiers that could regulate fatty acid metabolism through epigenetic gene regulation. Koletzko et al. [55] found that genotype distribution differs markedly among ethnicities, apparently reflecting an evolutionary advantage of genotypes in activating LC-PUFAs synthesis when the introduction of agriculture provided diets rich in LA but with little arachidonic acid (AA) and EPA. Vegetarians with a TT genotype of the rs174547 in FADS1 gene had higher odds of developing metabolic syndrome (MetS), larger waist circumference (WC), higher blood pressure (BP), and lower level of HDL-c, which suggested that the dietary intake of LA might interact with rs174547 in the FADS1 gene and affect HDL-c levels [56]. Joshi et al. [57] demonstrated that an increase the FADS1 expression of vegetarian pregnant women during pregnancy might maintain AA, EPA and DHA levels to ensure the overall LC-PUFA levels of the neonate. Populations had an increase in EPA in plasma when intake of flaxseed oil rich in α-linolenic acids (ALA) increased beyond that of major allele homozygotes consuming a typical "western" diet [58]. Individuals homozygous for the minor allele of FADS1/FADS2 had a lower plasma AA and AA/LA ratio when compared with the major allele carriers after each diet, while ELOVL fatty acid elongase 2 (ELOVL2) had no effect on PUFAs [58]. Their findings provided strong supporting evidence for the role of FADS1 in regulating PUFAs metabolism. In our study, no SNPs were found in FADS1 3 UTR. The inhibitor of miR-193a-5p significantly increased FADS1 expression, either mRNA or protein, which upregulated the n-3 fatty acid levels. Our experimental results further clarify that miR-193a-5p plays an important role in the regulation of milk fatty acid profiles at a post-transcriptional level.
According to the regulatory relationship network (Figure 9), miRNA-193a-5p regulates EPA, DPA and DHA synthesis by targeting FADS1 in BMECs. In PPI and regulatory relationship networks, FADS1 respectively interacts with FADS2, ELOVL6 and DGAT1, and the expression of FADS1 could influence the expression of DGAT2 through FADS2, ELOVL6 and DGAT1. As is well known, ELOVL6 is a key enzyme in intracellular lipid metabolism [59]. Du et al. [60] found that ELOVL6 was identified as a target gene of miR-125a-5p during porcine intramuscular adipogenesis, and that the overexpression of miR-125a-5p decreased the content ratios of monounsaturated fatty acids (MUFAs) and total SFAs, while having no significant impact on PUFAs/SFAs and n-6/n-3 ratios. Takamura et al. [61] proved that the knockout of ELOVL6 in mice could inhibit lipid accumulation and increase cholesterol consumption. Shi et al. [48,62] demonstrated that ELOVL5 and ELOVL7 play roles in the synthesis of long-chain unsaturated fatty acids (LC-UFAs) in goat mammary epithelial cells. In the current study, ELOVL6 expression shows a positive correlation with FADS1 expression and a negative correlation with miR-193a-5p expression. Also, DGAT2 expression shows a positively correlation with FADS1 expression and a negative correlation with miR-193a-5p expression. DGAT2, belonging to the DGAT gene superfamily, catalyzes the final, committed step in acyl-CoA-dependent TAG biosynthesis [63]. Stone et al. [64] confirmed that DGAT2 plays a fundamental role in mammalian TAG synthesis, and DGAT1 plays an important part in regulating energy metabolism. In addition, multiple studies concluded that DGAT1 and DGAT2 share no homology [64][65][66]. In our study, the expression of miR-193a-5p and FADS1 have statistical significance with DGAT1. each diet, while ELOVL fatty acid elongase 2 (ELOVL2) had no effect on PUFAs [58]. Their findings provided strong supporting evidence for the role of FADS1 in regulating PUFAs metabolism. In our study, no SNPs were found in FADS1 3′ UTR. The inhibitor of miR-193a-5p significantly increased FADS1 expression, either mRNA or protein, which upregulated the n-3 fatty acid levels. Our experimental results further clarify that miR-193a-5p plays an important role in the regulation of milk fatty acid profiles at a post-transcriptional level.
According to the regulatory relationship network (Figure 9), miRNA-193a-5p regulates EPA, DPA and DHA synthesis by targeting FADS1 in BMECs. In PPI and regulatory relationship networks, FADS1 respectively interacts with FADS2, ELOVL6 and DGAT1, and the expression of FADS1 could influence the expression of DGAT2 through FADS2, ELOVL6 and DGAT1. As is well known, ELOVL6 is a key enzyme in intracellular lipid metabolism [59]. Du et al. [60] found that ELOVL6 was identified as a target gene of miR-125a-5p during porcine intramuscular adipogenesis, and that the overexpression of miR-125a-5p decreased the content ratios of monounsaturated fatty acids (MUFAs) and total SFAs, while having no significant impact on PUFAs/SFAs and n-6/n-3 ratios. Takamura et al. [61] proved that the knockout of ELOVL6 in mice could inhibit lipid accumulation and increase cholesterol consumption. Shi et al. [48,62] demonstrated that ELOVL5 and ELOVL7 play roles in the synthesis of long-chain unsaturated fatty acids (LC-UFAs) in goat mammary epithelial cells. In the current study, ELOVL6 expression shows a positive correlation with FADS1 expression and a negative correlation with miR-193a-5p expression. Also, DGAT2 expression shows a positively correlation with FADS1 expression and a negative correlation with miR-193a-5p expression. DGAT2, belonging to the DGAT gene superfamily, catalyzes the final, committed step in acyl-CoA-dependent TAG biosynthesis [63]. Stone et al. [64] confirmed that DGAT2 plays a fundamental role in mammalian TAG synthesis, and DGAT1 plays an important part in regulating energy metabolism. In addition, multiple studies concluded that DGAT1 and DGAT2 share no homology [64][65][66]. In our study, the expression of miR-193a-5p and FADS1 have statistical significance with DGAT1.
The downregulation of miRNA-193a-5p and upregulation of FADS1 significantly increased the content of dairy fat in BMECs. Some studies are skeptical about the safety of high-fat milk intake [67]. An early study reported that a strong correlation existed between dairy fat consumption and CHD [68]. However, in recent years, a growing number of studies have found no significant correlation between milk fat content and CVDs [5]. The effect of dairy fat on health is demonstrated by the fatty acid profile [16,69,70]. Thus, improving the fatty acid profile of dairy fat continues to be a worthy topic for future studies. The downregulation of miRNA-193a-5p and upregulation of FADS1 significantly increased the content of dairy fat in BMECs. Some studies are skeptical about the safety of high-fat milk intake [67]. An early study reported that a strong correlation existed between dairy fat consumption and CHD [68]. However, in recent years, a growing number of studies have found no significant correlation between milk fat content and CVDs [5].
The effect of dairy fat on health is demonstrated by the fatty acid profile [16,69,70]. Thus, improving the fatty acid profile of dairy fat continues to be a worthy topic for future studies.

Conclusions
For the first time, our data identify that miRNA-193a-5p regulates PUFAs metabolism in BMECs by targeting FADS1, suggesting that miRNA-193a-5p and FADS1 are underlying factors for improving PUFAs content in bovine milk (Figure 9).

Supplementary Materials:
The following are available online at https://www.mdpi.com/2218-273 X/11/2/157/s1. Table S1. Composition and nutrient levels of total mixed ration (TMR). Table S2. The primers used for qRT-PCR to validate the Small RNA sequencing. Table S3. The primers used for qRT-PCR to validate the RNA sequencing. Table S4. The sequences to alter miR-193a-5p and FADS1 expression in BMECs. Table S5. The primers used for qRT-PCR to detect miR-193a-5p expression in BMECs. Table S6. The primers used for qRT-PCR to detect the expression of FADS1 and other genes related to fatty acid metabolism in BMECs. Table S7. A primer containing the miR-193a-5p action site. Table S8. Statistics of miRNA sequencing reads. Table S9. Statistics of mRNA sequencing reads. Table S10. Verification results of Small RNA sequencing by qRT-PCR. Table S11. Verification results of RNA sequencing by qRT-PCR.  Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available from supplementary materials.