MicroRNA analysis reveals the role of miR-214 in duck adipocyte differentiation

Objective Fat deposition in poultry is an important factor in production performance and meat quality research. miRNAs also play important roles in regulating adipocyte differentiation process. This study was to investigate the expression patterns of miRNAs in duck adipocytes after differentiation and explore the role of miR-214 in regulating carnitine palmitoyltransferases 2 (CPT2) gene expression during duck adipocyte differentiation. Methods Successful systems for the isolation, culture, and induction of duck primary fat cells was developed in the experiment. Using Illumina next-generation sequencing, the miRNAs libraries of duck adipocytes were established. miRanda was used to predict differentially expressed (DE) miRNAs and their target genes. The expression patterns of miR-214 and CPT2 during the differentiation were verified by quantitative real-time polymerase chain reaction and western blot. Luciferase reporter assays were used to explore the specific regions of CPT2 targeted by miR-214. We used a miR-214 over-expression strategy in vitro to further investigate its effect on differentiation process and CPT2 gene transcription. Results There were 481 miRNAs identified in duck adipocytes, included 57 DE miRNA candidates. And the 1,046 targets genes of DE miRNAs were mainly involved in p53 signaling, FoxO signaling, and fatty acid metabolism pathways. miR-214 and CPT2 showed contrasting expression patterns before and after differentiation, and they were selected for further research. The expression of miR-214 was decreased during the first 3 days of duck adipocytes differentiation, and then increased, while the expression of CPT2 increased both in the transcriptional and protein level. The luciferase assay suggested that miR-214 targets the 3′untranslated region of CPT2. Overexpression of miR-214 not only promoted the formation of lipid droplets but also decreased the protein abundance of CPT2. Conclusion Current study reports the expression profile of miRNAs in duck adipocytes differentiated for 4 days. And miR-214 has been proved to have the regulator potential for fat deposition in duck.


INTRODUCTION
Deposited fat in poultry mainly consists of abdominal fat, subcutaneous fat, and intra muscular fat (IMF).Excessive fat deposition can reduce feed efficiency and carcass yield [1].While the IMF is recognized as the predominant factor affecting meat quality due to its positive correlation with juiciness, tenderness, and flavor.Therefore, understanding fat deposition is critical in meat quality research and improving production [2].The molecular and cellular mechanisms of underlying lipid metabolism have been well studied in mam malian models; for instance, the increasing demands of muscle during exercise or heat production depend on fat mobilization [3], and fatty acid oxidation is essential for con straining adipocyte lipolysis and regulating systemic catabolism when glucose is limiting [4].In contrast, such information is scarce for birds.
Adipocytes, which are derived from mesenchymal stem cells, play an important role in energy balance in both poultry and mammals [5].Adipose tissue expansion is a result of adipocyte formation, involving several processes collectively referred to as adipogenesis, and cellular accumulation of tri glyceride (TG) inside lipid droplets [6].For the last 40 years, the cellular and molecular mechanisms of adipocyte differ entiation have been extensively studied using preadipocyte culture systems [7]; specifically, committed preadipocytes undergo growth arrest and subsequent terminal differentia tion into adipocytes.Similar vitro differentiation system has also been widely used in the adipocyte differentiation re searches in both chickens and ducks [8,9].
In this study, highthroughput sequencing technology was used to monitor miRNA levels in duck embryo primary preadipocytes and differentiated adipocytes.miR214 and CPT2 were selected for further analysis at different stages of differentiation.Overexpression of miR214 promoted adi pocytes differentiation and downregulated the protein level of CPT2.Collectively, these results not only expand the miRNA regulation network during adipogenesis but also provide potential molecular regulator for fat deposition in duck.

Ethics statement
All animal experimental procedures were approved and guided by the Institutional Animal Care and Use Committee of Yangzhou University (approval number 176-2020).

Duck pre-adipocyte isolation, culture, and differentiation
Primary preadipocytes were isolated from subcutaneous adipose tissue of twentydayold embryonic Cherry Valley ducks, using the method established previously [18].The preadipocytes were seeded at a density of 1×10 4 cells/cm 2 and maintained in complete medium (Dulbecco's modified eagle medium with 10% fetal bovine serum, and 1% penicillin and streptomycin) [20].Upon reaching 90% cell confluence, the medium was replaced with the differentiation medium, comprising the complete medium with an additional 300 μM oleic acid, 1 μm/L dexamethasone, 0.5 mmol/L 3isobutyl 1methylxanthine, 1 μM rosiglitazone, 10 μg/μL insulin, added.The medium was replaced every second day [8].Pre adipocyte morphology was observed using a microscope (Olympus, Tokyo, Japan).

Oil red O staining and triglyceride content analysis
Oil red O staining kit (Solarbio, Beijing, China) were used for the lipid droplets staining.Firstly, adipocytes were washed with phosphatebuffered saline (PBS) and fixed using 4% formaldehyde for 10 min.Then, stained the cells with 1% fil tered oil red O solution after two rinses in PBS.After 30 min, the cells were repeatedly washed with distilled water and ob served under a microscope.For oil red O quantitative analysis, the intracellular adsorbed oil red O was extracted in 100% isopropanol for 10 min, and absorbance was measured at 500 nm wavelength.
Triglyceride content was analyzed in three replicates using a TG detection kit (Applygen, Beijing, China) according to the manufacturer's protocols.Briefly, duck adipocytes were washed using preheated PBS and then centrifuged at 1,000 ×g for 5 min.The cells were ruptured using the ultrasonic cellbreak method.A bicinchoninic acid protein assay kit (Beyotime, Beijing, China) was used to measure the pro tein concentration of the ruptured cells.Optical density values were detected using a microplate spectrophotometer.

Small RNAs library construction and RNA-seq analysis
Three Cherry Valley duck preadipocytes samples (CVC) and those three samples that had undergone differentiation treatment for 3 d (CVT) were collected for RNA sequencing using the mirVana miRNA Isolation Kit (Ambion, Austin, TX, USA).After quality checking, total RNA for each sam ple was used to construct the small RNA library, according to the manufacturer' s protocol, using the Illumina Small RNA Sample Prep Kit (Illumina, San Diego, CA, USA).Libraries were sequenced using the Illumina HiSeq 2500 platform by pairedend sequencing, conducted by Novogene Biotech nology Co., Ltd.(Beijing, China).

Identification of miRNA and differentially expressed miRNAs
After conducting quality control, the clean reads for each sample were mapped using Bowtie [21] against the standard duck genome BGI_duck_1.0 (https://www.ncbi.nlm.nih.gov/assembly/GCA_000355885.1), then aligned with the known miRNAs in miRbase (https://www.mirbase.org/index.shtml) to identify known miRNAs, allowing less than one mismatch between sequences.The identified miRNAs were classified into families based on their sequence similarities.Unmatched reads were further processed to predict novel miRNAs using miREvo [22] and mirdeep2 [23].Differen tially expressed (DE) miRNA (read count threshold >10) analysis was performed using the DEseq 2 algorithm, with the cutoff of |Log 2 Fold Change|>1 and false discovery rate <0.05.Six miRNAs (miR10b3p, miR4555p, miR148a3p, miR1225p, miR15c5p, and miR10a5p) were randomly selected for validation using a replicated experiment.The quantitative realtime polymerase chain reaction (qRTPCR) primer sequences used are listed in Table 1.

Prediction of miRNA target genes and pathway enrichment analysis
Differentially expressed miRNAs and their target genes were predicted using miRanda (https://www.microrna.org/microrna/home.do)based on the previous transcriptome results [8].Gene ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analyses were performed us ing DAVID 6.8 (https://david.ncifcrf.gov/).Less than 0.05 p value was considered significantly enriched.Signal pathway diagram of target genes involved in fatty acid metabolism was performed by Microsoft PowerPoint.

Sample collection and quantitative real-time polymerase chain reaction
Three sevendayold male Cherry Valley ducks were eutha nized by cervical dislocation, and the heart, liver, spleen, lung, breast muscle, right leg muscle, and adipose tissues were collected to analyze gene expression.Total RNA was extracted from the tissue or cultured cells using Trizol re agent (Invitrogen, Carlsbad, CA, USA).Firststrand cDNA was synthesized from 1 μg of total RNA using the Prime Script RTPCR Kit (TaKaRa, Osaka, Japan), following the manufacturer's instructions.The mRNA expression levels of several marker genes were detected using a TB Green qRT PCR Premix (TaKaRa, Japan) and the QuantStudio 5 real time PCR instrument.MirX miRNA FirstStrand Synthesis (TaKaRa, Japan) and TB Green qRTPCR Premix (TaKaRa, Japan) were used for qRTPCR analysis of miR214.The qRT PCR primer sequences used are listed in Table 1 and Table 2, while the common downstream primers for miRNA were provided by Takara (undisclosed sequence).Betaactin or U6 were used as internal controls, and all assays were run in triplicate.Relative transcription was evaluated using the 2 -ΔΔCt method, and is presented as mean±standard deviation.

Vector construction and luciferase assays
A luciferase reporter, with two wild type target sites of the CPT2 3′ UTR (WT), was constructed by subcloning the fragment into the region directly downstream of the cyto megalovirus promoterdriven firefly luciferase cassette in the

Primers
Sequence (5′-3′)  To validate the miRNA targets, we used 293T cells pro vided by Professor Song (Yangzhou University, China).Cells were seeded into 24well plates at a density of 1×10 5 cells/well, and incubated at 37°C overnight.Cotransfection with 100 ng of the WT or MUT vectors, and with 2 μL of the miR214 mimics or negative control, were performed using Lipo fectamine 2000 (Invitrogen, USA).Luciferase activity was measured using the Dual Luciferase Reporter Assay System (Promega, USA) at 48 h posttransfection.Finally, the fold change in expression was calculated by comparing the expression of miR214 with that of the miRNANC.

Statistical analysis
Statistical analysis was performed using SPSS 22.0 and Graph Pad Prism 6. Meaningful differences between the groups were calculated using Student's ttests.Statistical significance is defined when p values are less than 0.05.

Summary of small RNA-seq data
In total, 12,244,755, 10,749,053, 11,688,816, 11,178,383, 11,706,221, and 11,456,503 raw reads were obtained from the CVC1, CVC2, CVC3, CVT1, CVT2, and CVT3 libraries, respectively (Supplementary Table S1).After removing con taminant reads, we obtained 11,614,612, 10,227,841, and 11,095,066 clean reads for CVC, and 10,650,292, 11,006,776, and 10,615,983 clean reads for CVT, of 18 to 35 nt long, which were used for subsequent analyses.These clean reads were uniquely mapped to the standard duck genome BGI_duck_1.0 (https://www.ncbi.nlm.nih.gov/assembly/GCA_000355885.1); on average, 10,979,173 and 10,757,684 clean reads were mapped for the CVC and CVT libraries, respectively.Further, >90% of the bases had base quality >Q20.Most of the clean reads were 21 to 24 nt; 22 nt reads were the most abundant, followed by 23 nt reads (Supplementary Figure S2A).The sequenced sRNAs mapped to >60% of miRNAs in the miRBase database (Supplementary Table S2).However, the other types of RNAs (including rRNAs, tRNAs, mRNAs, and snoRNA) were mapped in relatively low proportions.Reproducibility between samples was good (Supplementary Figure S2B).This indicates that the deep sequencing data were highly enriched in mature miRNA sequences and were suitable for subsequent profiling analysis.

miR-214 and CPT2 expression in preadipocytes and adipocytes
The seed region of the mature miR214 sequence was highly conserved in duck, chickens, mice, humans, and other spe cies (Figure 3C).Two potential binding sites for miR214 were identified in the 3′ UTR of the CPT2 gene; these sites were not highly conserved among poultry (Figure 3D).There was a negative association between miR214 expression and CPT2 gene expression by the transcriptome results.We veri fied miR214 expression levels in CVC and CVT cells (Figure 3E), and found that it was expressed in the heart, liver, breast muscle, thigh muscle, and fat tissue of 7 dayold Cherry Valley ducks (Figure 3F), but mostly in the breast muscle and adi pose tissue.
To fully elucidate miR214 and CPT2 expression during adipocyte differentiation, we assessed their expression at 0, 2, 4, 6, and 8 d postdifferentiation.Lipids gradually accumu lated to form larger lipid droplets as the differentiation time increased (Figure 4A).The expression level of miR214 was decreased in the first 3d differentiation time and then in creased rapidly (Figure 4B).The adipocytes differentiation induced the expression of CPT2 at the transcriptional level and accumulation of the protein level (Figure 4C).

miR-214 overexpression promoted adipocyte differentiation and targeting CPT2
In order to verify the miR214 target sites on CPT2, a plas mid containing the WT or mutant target sites of the 3′ UTR region (Figure 4D) of the CPT2 gene was constructed.Be sides, miR214 expression level in 293T was lower than that in duck preadipocyte, and the effect of miR214 mimics overexpression could meet the experimental requirements (Figure 4E).The plasmid were transfected into 293T cells, together with miR214 mimics or miRNC.miR214 mimics significantly reduced the luciferase activity of the WT CPT2 3′ UTR, but did not affect the activity of the three CPT23′ UTR mutants (Figure 4F).This indicates that miR214 binds directly to the two target sites in the CPT2 3′ UTR, and that there might be a cascade effect between the two target sites.
To investigate the regulatory role of miR214 in adipocyte differentiation, an overexpression experiment was conducted, by transfecting miR214 mimics.miR214 expression was significantly higher in the overexpression groups (those transfected with the miR214 mimics), for both preadipocytes and differentiated adipocytes, than in the groups transfected with the miRNC (Figure 5A).FABP4, C/EBPα, and CPT1A expression was higher, and that of DLK1 was lower, in the miR214 overexpression group after differentiation, than in the negative control group (Figure 5B).miR214 over expres sion inhibited the expression level of CPT273kD protein but not at the level of mRNA (Figure 5C).Besides, triglyceride content was higher in the overexpression group than in the negative control group (Figure 5D).The number of lipid droplets was increased and the lipid droplets become bigger after miR214 overexpression (Figures 5E, F; Supplementary Figure S2).Together, these results suggested that miR214 overexpression promoted adipocyte differentiation process, reduced fatty acid transfer to mitochondria.

DISCUSSION
In this study, we found that, after 4d of culture in differentia tion medium, TG content increased, and lipid droplets gradually accumulated inside cells.The expression level of fatty acid synthesis genes increased, and that of fatty acid transport genes increased slightly.And, several of the miRNAs identified in our miRNAseq analysis, such as miR103, miR 21, miR30, miR27, miR130 and let7, were also detected by other researches [11,24].Further, these miRNAs are known to be associated with adipocyte differentiation; in porcine preadipocytes, for instance, miR130a overexpres sion can inhibit adipocyte differentiation [25]; in Zebrafish, miR27b depletion promotes lipid accumulation and weight gain [26], and in chicken, inhibiting miR30a promotes adi pocyte differentiation [27].Our study revealed differential miRNA expression related to adipocyte differentiation.
We conducted KEGG pathway analysis to reveal the bio logical significance of the target genes of the DE miRNAs.The "fatty acid metabolism" pathway was enriched in KEGG analysis and 8 genes (HADHB, ACAA2, CPT2, ACSL1, FASN, HADH, HACD3, and ACSBG2) were involved in this pathway.They play important roles in fatty acid synthesis, transport, and oxidative utilization.FASN and HACD3 are two key en zymes required for fatty acid synthesis [28,29].ACSL1 and ACSBG2 enable fatty acids to enter into lipid pools, and acti vate fatty acids [30,31].CPT1 and CPT2 mediate fatty acid transfer into the mitochondrial matrix for βoxidation [32].And the study of Qiu proved that the expression level of CPT1 regulated fatty acid oxidation and thus affects intra muscular adipogenesis [33].CPT2 was proved to be a common upregulated protein during differentiation of preadipocytes from bovine omental, subcutaneous and in tramuscular adipose depots [34].HADH, HADHB, and ACAA2 are key enzymes in fatty acid oxidation [3537].Some of these enzymes have been shown to be regulated by miRNAs; for instance, miR33 regulates fatty acid me tabolism by targeting HADHB, in chickens [38], geese [39], and African green monkeys [40]; and miR214 is predicted to regulate ACSBG2 expression [27].We suggest that the transport and utilization of fat is closely related to the ac cumulation of fat droplets.Considering that two target sites of miR214 in the 3'UTR of CPT2, miR214 and CPT2 are selected to explore their regulatory mechanism during adi pocyte differentiation in duck.
miR214 is an important miRNA that has been well stud ied in humans.It is also identified in chicken [27], duck [18], and pigeon [41].Although, miR214 is a highly conserved miRNA between humans and the poultry above, its function in adipocyte differentiation and fatty acid metabolism in ducks has not been fully reported.It has multiple functions in both normal and cancer cells, helps to coordinate essen tial signaling pathways [42].It is also a potential biomarker or target for therapeutic intervention [43].Besides, miR214 is implicated in promoting fibroblast differentiation of adipose derived mesenchymal stem cells by targeting mitofusin2 (MFN2) [44], and suppressing the osteogenic differentiation of bone marrowderived mesenchymal stem cells and human hair follicle stem cells [45,46].We found that the expression level of miR214 was decreased in the early adipocyte differ entiation, then increased in the later adipocyte differentiation.The research on 3T3L1 preadipocyte differentiation re vealed that the level of miR214 downregulated on 36h of differentiation and upregulated on 6d of differentiation [17].Further, miR214 overexpression promoted duck adi pocyte lipid droplets accumulation, this is consistent with findings for 3T3L1 preadipocytes [17].Taken together, our findings provided that miR214 may be an important posi tive regulator in duck adipogenic development.
The carnitine palmitoyltransferase (CPT) enzyme system, which operates throughout the body, enables longchain fatty acids to enter the mitochondrial matrix [47].The CPT sys tem comprises two separate proteins located in the outer and inner mitochondrial membranes (CPT1 and CPT2, re spectively) [32].Carnitine is important for mitochondrial metabolism; however, two carnitine acyltransferases, carni tine octanoyltransferase and carnitine acetyltransferase, are peroxisomal enzymes, indicating that it is also important for peroxisomal metabolism.There is some evidence that per oxisomes can degrade fatty acids that are typically degraded by mitochondria, possibly after transport as acylcarnitines [47].During early adipocyte differentiation, the lipid droplets accumulated and CPT2 expression increased with differenti ation time which is coincide with the result of bovine during differentiation of adipocytes [34].Here, we investigated that, during the differentiation time, miR214 and CPT2 did not show obvious opposite expression patterns, these might ex plained by the complex regulatory systems during adipocyte differentiation.Furthermore, CPT2 regulated fatty acid transportation and differentiation in duck adipocytes: lipid droplets accumulated and CPT2 expression increased with differentiation time, enhancing fatty acid transport; while the later upregulated miR214 promoted the differentiation process, inhibited the increasing expression of CPT2, and these further increased the accumulation of lipid droplets.CPT2 is also found to be under the expression regulation of peroxisome proliferatoractivated receptor γ (PPARγ) [34].Besides, miR214 overexpression enhanced duck adipocyte lipid droplets accumulation and reduced the protein level of CPT2, thereby reducing fatty acid transport and oxidation.Altogether, we suggest that miR214 and CPT2 are involved in the complex regulatory network during the duck adipo cytes differentiation.

Figure 1 .
Figure 1.In vitro comparison of differentiated and non-differentiated Cherry Valley duck preadipocytes.(A) Comparison of (a) non-differentiated Cherry Valley duck preadipocytes (CVC) and (b) Cherry Valley duck adipocytes differentiated for 3 d (CVT); Oil Red O staining of (c) CVC (d) CVT cells.(B) Triglyceride (TG) content of CVC and CVT cells.Triglyceride content is expressed in mmol/g protein.(C) Relative mRNA expression of six genes related to differentiation and fatty acid metabolism in adipocytes.Beta-actin was used as an endogenous control for relative quantification.(B, C) Data are shown as the mean±standard deviation.n = 3, * p<0.05, ** p<0.01.

Figure 2 .
Figure 2. Analysis of differentially expressed miRNAs, comparing differentiated and non-differentiated Cherry Valley duck preadipocytes (CVT and CVC cells, respectively).(A) Venn diagrams showing the numbers of common and unique miRNAs, comparing CVC and CVT cells.(B) Volcano plot of differentially expressed miRNAs, comparing CVC and CVT cells.(C) quantitative real-time polymerase chain reaction (qRT-PCR) verification of sequencing results.

Figure 3 .
Figure 3. Functional enrichment analysis of differentially expressed miRNAs and their target genes, comparing differentiated and non-differentiated Cherry Valley duck preadipocytes (CVT and CVC cells, respectively).(A) KEGG analysis of the targets of differentially expressed miRNAs.(B) Signal pathway diagram of target genes involved in fatty acid metabolism.(C) Conservatism analysis of the miR-214 seed region.(D) Conservatism analysis of the target sites of miR-214 in the 3′ UTR of CPT2.(E) qRT-PCR verification of miR-214 expression.(F) Expression of miR-214 in various duck tissues, detected using qPCR.U6 small nuclear RNA (snRNA) was used as an endogenous control for miR-214 relative quantification.Data are shown as the mean±standard deviation from three individuals.KEGG, Kyoto encyclopedia of genes and genomes; UTR, untranslated region; CPT2, carnitine palmitoyltransferases 2; qRT-PCR, quantitative real-time polymerase chain reaction.** p<0.01.

Figure 5 .
Figure 5.The targeting effect of miR-214 and CPT2 in Cherry Valley duck adipocyte differentiation.(A) Efficacy of miR-214 mimics.(B) Gene expression and (C) CPT2 protein expression, in pre-adipocytes treated with the miR-214-mimic or the negative control (NC).(D) Triglyceride content of adipocytes transfected for 48 h with miR-214 mimics or with the miR-NC, and following differentiation for 4 d.(E) Comparison of preadipocytes transfected with miR-NC (a) or with miR-214 mimics (b) for 48 h, and following differentiation for 4 d.Oil red O staining of adipocytes transfected with miR-NC (c) or with miR-214 mimics (d) for 48 h, and following differentiation for 4 d.(F) Spectrophotometric analysis following oil red O staining.CPT2, carnitine palmitoyltransferases 2. * p<0.05, ** p<0.01.

Table 1 .
Primer sequences for miRNA quantification

Table 2 .
Primer sequences for polymerase chain reaction