A novel lncRNA BADLNCR1 inhibits bovine adipogenesis by repressing GLRX5 expression

Abstract Adipogenesis is a complex cellular process, which needs a series of molecular events, including long non‐coding RNA (lncRNA). In the present study, a novel lncRNA named BADLNCR1 was identified as a regulator during bovine adipocyte differentiation, which plays an inhibitory role in lipid droplet formation and adipogenic marker gene expression. CHIPR‐seq data demonstrated a potential competitive binding motif between BADLNCR1 and sterol regulatory element‐binding proteins 1 and 2 (SREBP1/2). Dual‐luciferase reporter assay indicated target relationship between KLF2 and BADLNCR1. Moreover, after the induction of KLF2, the expression of adipogenic gene reduced, while the expression of BADLNCR1 increased. Real‐time quantitative PCR (qPCR) showed that BADLNCR1 negatively regulated mRNA expression of GLRX5 gene, a stimulator of genes that promoted formation of lipid droplets and expression of adipogenic genes. GLRX5 could partially reverse the effect of BADLNCR1 in bovine adipocyte differentiation. Dual‐luciferase reporter assay stated that BADLNCR1 significantly reduced the enhancement of C/EBPα on promoter activity of GLRX5 gene. Furthermore, CHIP‐PCR and CHIRP‐PCR confirmed the suppressing effect of BADLNCR1 on binding of C/EBPα to GLRX5 promoter. Collectively, this study revealed the molecular mechanisms underlying the negative regulation of BADLNCR1 in bovine adipogenic differentiation.

adipogenic genes, such as adipocyte fatty acid-binding protein/adipocyte P2 (FABP4/AP2) and fatty acid translocase (FAT/CD36) that give rise to adipocyte phenotype. 5,6 Except for coding genes, lncRNA is a set of non-coding RNA considered to be longer than 200 nt with little or no coding potential. 6 Recently, a plenty of lncRNA has been identified to be involved in diverse biological processes, including genomic imprinting, 7 chromatin modification, 8 cancer metastasis, 9,10 neurogenesis, 11 myogenic development 12 and adipogenesis. 13,14 Li et al 15 discovered hundreds of lncRNA that differently expressed in mature adipocytes compared to pre-adipocytes of cattle. However, the detailed function of only one lncRNA was revealed. The roles of most lncRNA in bovine adipogenesis are unknown.
Glutaredoxin 5 (GLRX5) is a 156 amino acid mitochondrial protein, which is evolutionarily conserved. 16 It is necessary for iron-sulphur clusters transfer, which is required for normal iron homoeostasis. 17 Also, GLRX5 is involved in protein lipoylation, since mutation in GLRX5 gene could impair transfer of [Fe-S] to lipoate synthase enzyme. 18 As a family member of GLRX5, GLRX1 was elevated 70% in the adipose tissue of the obese and 45% fat calorie-fed rats. 19 Additionally, there is a great difference in mitochondrial proteome between the different adipose tissues. 20 But it is still unclear how GLRX5 participates in adipogenesis.
Given this, in this study, the negative role of a novel lncRNA, bovine adipocyte differentiation-related long non-coding RNA 1 (BADLNCR1) in bovine adipogenic differentiation, is revealed. Its genome-wide binding is analysed to reveal its regulation mechanisms.
As its neighbour, GLRX5 is found to be its target, which positively regulates bovine adipocyte differentiation. Finally, why BADLNCR1 represses GLRX5 transcription is investigated. These data provide a novel insight into lncRNA and GLRX5 in molecular regulation of bovine adipogenesis.

| MATERIAL S AND ME THODS
All experiments were approved by the Review Committee for the Use of Animal Subjects of Northwest A&F University. All experiments were performed in accordance with relevant guidelines and regulations.

| Rapid amplification of cDNA ends (RACE)
In order to gain the full-length sequence of BADLNCR1, SMARTer RACE cDNA Amplification Kit (Clotech, 634859) was used to perform 5'RACE according to the instructions. And 3'RACE was performed referring to Frohman Scotto-Lavino. 23 The template RNA was extracted from adult adipose tissue of cattle. The gene-specific primer used for 5' RACE was 5'-GATTACGCCAAGCTTCTGCCAGTTTTCTCTTCCTGTCGG-3', and two gene-specific primers used for 3'RACE were inner: 5'-TGCCATGTGCAATTTTCC-3' and outer: 5'-AGAGGAAGCTGAGGCATG-3'.

| Plasmid, nucleotide and cell transfection
To construct the expression vectors, the full length of BADLNCR1, GLRX5, KLF2 and CEBPα was amplified and constructed into pcDNA3.1(+) vector. To translate BADLNCR1 in vitro, the fragment of BADLNCR1 was cloned into pET-28a vector. Also, the genomic fragments containing upstream 2,000 bp of transcriptional start site of BADLNCR1 and GLRX5 gene were constructed into pGL3basic vector to construct luciferase reporter plasmids, respectively.
Primers used in plasmid construction are listed in Table S3.

| Cell fractionation and fluorescent in situ hybridization (FISH)
Cytoplasmic and nuclear RNA of pre-adipocytes and mature adipocytes were extracted using PARIS Kit (Invitrogen, AM1921) based on its instruction. Cy3-labelled BADLNCR1 probes for FISH assay were purchased from RiboBio. And RNA FISH was carried out using FISH Kit (RiboBio, lnc1100285) following its instruction.

| CHIRP-seq, RNA pull-down and mass spectrum, and CHIP
In order to obtain genome-wide binding of BADLNCR1, CHIRP assays with both odd and even probes (

| Statistical analysis
Data were analysed by Student's t test using SPSS software (version 20). The results were presented as mean ± SE (Standard Error), and P value < .05 was considered statistically significant.

| Identification of BADLNCR1
Previous RNA-seq data have discovered differently expressed lncR-NAs when compared bovine pre-adipocytes to mature adipocytes. 15 qPCR was applied to identify their expression. A novel lncRNA, NONBAT013210, had higher expression level than other lncR-NAs and showed significantly decreasing expression in adipocytes ( Figure 1A). Also, tissue expression pattern indicated that it highly expressed in adipose tissue of cattle ( Figure 1B). Other lncRNAs showed low expression level in adipose tissue (data was not shown).
So NONBAT013210 was chosen as candidate lncRNA and named as BADLNCR1.
RACE revealed that BADLNCR1 transcript was polyadenylated and reversely transcribed from chromosome 21 with two exons of 1029 bp in length ( Figure S1A). A BANLNCR1 expression vector failed to produce a protein by translation assay in vitro ( Figure 1C).
Moreover, CPC analysis indicated that compared to protein-coding genes, GLRX5 and GAPDH, and a famous lncRNAs, Xist, BADLNCR1 had a lower probability than all of them ( Figure 1D). Secondary structure of BADLNCR1 was shown in Figure S1B.
qPCR analysis of fractionated nuclear and cytoplasmic RNA stated that BADLNCR1 was primarily expressed in nuclear; as a control, GAPDH gene was mainly detected in cytoplasmic ( Figure 1E).
This result was confirmed by FISH ( Figure 1F), which also indicated that BADLNCR1 was mostly localized in nuclear.

| Genome-wide binding of BADLNCR1
As previous reports described, lncRNAs modulate transcription of its target genes that locates closed to the lncRNA transcription site were located in intergenic region. In terms of the distance to genes,  Figure 3B), which is the gene expression regulation region. 37 This indicated the potential role of BADLNCR1 in regulation of gene expression. Using DMEM motif analysis, top five significantly enriched DNA-binding motifs of BADLNCR1, ACGTGATH, AGTGCRTG, RCTGATCA, CWCGWGA and AGATGAGS were found ( Figure 3C).
Interestingly, after comparing them to known motif using JASPAR analysis, motif 1 and motif 2 highly resemble binding sequence for sterol regulatory element-binding protein 1 (SREBP1) and SREBR2 (Table S1)

| GLRX5 is selected as target of BADLNCR1
To understand the molecular basis of how BADLNCR1 regulates adipocyte differentiation, using the University of California Santa Cruz GLRX5 gene and BADLNCR1 in 11 adult adipose tissues were evaluated ( Figure 4D). Consistently, correlation analysis indicated that there was a significant negative correlation (P < .05, r < 0) between the expression levels of them ( Figure 4E). Therefore, GLRX5 gene was selected as the cis-regulation of molecular target by BADLNCR1.

| GLRX5 promotes bovine adipogenic differentiation
Next, the involvement of GLRX5 in bovine adipogenic differentiation was tested. GLRX5 gene shows a rising expression trend after adipogenic induction ( Figure 5A), which is opposite to the decreasing expression trend of BADLNCR1. After GLRX5 was knocked down by siRNA and siGLRX5, expression levels of PPARγ, CEBPα and FABP4 were significantly reduced (P < .05) ( Figure 5B).
Accordingly, significantly reducing amount of lipid droplets was observed by Red Oil Staining ( Figure 5C,D). On the contrary, after the overexpression of GLRX5, the expression levels of CEBPα and FABP4 significantly increased (P < .05) and almost a twice rising of PPARγ was observed (P = .056) ( Figure 5E). Also, more lipid droplets were formed ( Figure 5F,G). Collectively, GLRX5 gene is a positive regulator of bovine adipocyte differentiation. The effects of GLRX5 and BADLNCR1 are opposite.

| BADLNCR1 represses GLRX5 transcription activity
Since BADLNCR1 regulates the mRNA expression of GLRX5, and it is transcribed from the upstream promoter region of GLRX5, it was supposed that BADLNCR1 was related to the transcription of GLRX5.
Based on previous reports, there are several ways in which lncRNAs could regulate transcription. For example, lncRNAs function through binding to histone-modifying complexes, to DNA-binding proteins (including transcription factors) and even to RNA polymerase II. 41 In In order to test the former hypothesis, RNA pull-down assay and mass spectrum were performed. Sixty-two proteins were enriched.
However, there were no proteins related to histone-modifying complexes, transcriptional factors or RNA polymerase II found. As a result, BADLNCR1 cannot recruit a regulatory protein to regulate the transcription of GLRX5.
To test the second hypothesis, CHIRP-PCR was applied to verify the binding sites of BADLNCR1 on GLRX5. According to the CHIRP-seq data, BADLNCR1 was found to bind on the upstream region (from −1271 to −476 bp) of GLRX5 gene ( Figure 6A). Based on the sequence of binding region, 5 pairs of primers (Table S2) were designed to perform PCR with DNA products from CHIRP assay as template to verify the sequencing data ( Figure 6A). The lengths of amplified fragments were the same as expect. And the sequencing results of these fragments are the same as reference sequence. So BADLNCR1 binds on −1271 to −476 bp region of GLRX5 gene. Bioinformatics analysis of this binding region revealed 5 binding sites for adipogenic transcriptional factors including CEBPα, activator protein 2, PPARα, KLF5, E2F transcription factor 1 and nuclear transcription factor Y ( Figure 6B). As CEBPα is the most significantly predicted and its number of binding sites was the most, it was taken as an example for following assay. −1851 to −96 bp upstream region of GLRX5 gene was constructed into pGL3 vector to be the reporter. As shown in Figure 6C, dual-luciferase activity assay results indicated that overexpression of BADLNCR1 significantly reduces the relative luciferase activity of GLRX5 promoter (P < .05), while overexpression of CEBPα significantly rises the relative luciferase activity of GLRX5 promoter (P < .01).
However, BADLNCR1 also significantly eliminates the enhancement of CEBPα on GLRX5 promoter (P < .05). On the other hand, CHIP-PCR and CHIRP-PCR were performed using chromatins from  Figure 6D). Taken together, these results indicated that BADLNCR1 may prevent the binding and activation of CEBPα to GLRX5 promoter, ultimately resulting in obstruction to bovine adipogenic differentiation.

| GLRX5 rescues inhibitory role of BALDNCR1 in bovine adipogenic differentiation
As GLRX5 was selected as the target gene of BADLNCR1, and it plays opposite part to BADLNCR1 in bovine adipogenic differentiation, whether GLRX5 could rescue the inhibitory role of BADLNCR1 was tested. As mentioned above, in bovine adipocyte, overexpression of BADLNCR1 inhibits the expression of PPARγ, CEBPα and FABP4. However, if GLRX5 was overexpressed at the same time, their repressed expression levels were increased significantly (P < .01) ( Figure 6E). Additionally, when BADLNCR1 was knocked down, these adipogenic genes displayed decreasing expression. But these decreasing levels were significantly increased after GLRX5 was knocked down in the meantime (P < .01) ( Figure 6F). Taken together, GLRX5 may help to rescue the prevention of BADLNCR1 in bovine adipocyte differentiation.

| D ISCUSS I ON
Recently, more and more studies highlighted the importance of lncRNA in diverse biological processes, including regulating cell showed that BADLNCR1 inhibits terminal adipocyte differentiation, which is consistent with its decreasing expression in terminal differentiation stage. No change in expression was found on 2nd day may indicate no effect of BADLNCR1 on pre-adipocyte proliferation.
However, further gain-and loss-of-function experiments are needed to explore whether BADLNCR1 regulates pre-adipocyte proliferation or not.
As a member of GLRX family, a group of oxidoreductases catalyse the reversible reduction in protein disulphides 48 and were involved in the biogenesis of iron-sulphur clusters, GLRX5 was identified as the downstream target of BADLNCR1. Using gain-and loss-of-function experiments, GLRX5 was found to perform as a simulative gene that regulates bovine adipocyte differentiation. Iron-sulphur cluster is necessary for the activity of a great number of mitochondrial proteins that participate in oxidative pathways, including complexes I and II of the respiratory chain. 49,50 Previous reports explained that oxidative pathways are associated with adipogenesis. 51 And several iron-sulphur cluster-related proteins were identified to be involved in adipose development, such as CISD1 52 and MitoNEET. 53  promoter, which indicated that BADLNCR1 impairs the binding of CEBPα on GLRX5 promoter.
Conservation is not a general feature for lncRNAs, especially in their nucleotide sequences. However, recent studies have indicated that lncRNAs show cross-species conservation of their genomic position. 61 For example, both mouse linc-YY1 and human linc-YY1 are transcribed from upstream genomic region of YY1 gene. 12 , 26 Both mouse lncMyoD and human lncMyoD are generated from upstream of MyoD gene in both mouse and human genomes. 62 Bioinformatics analysis indicated that mouse lncRNA Snhg10 and human lncRNA SNHG10 are reversely transcribed from upstream genomic region of GLRX5 gene ( Figure S3A-B). Consistently, BADLNCR1 is reversely transcribed from upstream genomic region of bovine GLRX5 gene.
AnnoLnc web server shows that human lncRNA SNHG10 is detected in all tissues and highly expressed in fat ( Figure S3C), and SNHG10 is mainly expressed in nuclear in most of cell lines as determined using the lncATLAS annotation database ( Figure S3D), these patterns similar to those of bovine BADLNCR1. However, expression levels of mouse Snhg10 in pre-adipocytes and mature adipocytes have no significant difference in 3T3-L1 ( Figure S3E). Considering the important role of BADLNCR1 in bovine adipogenesis, further researches of the function and mechanisms of SHNG10 in human adipogenesis and mouse adipogenesis should be performed.
In a conclusion, KLF2 regulated lncRNA, and BADLNCR1 appears to act as a suppressor during bovine adipogenic differentiation. This process might perform by its negative regulation on transcriptional activity and mRNA expression of GLRX5, which is benefit for bovine adipogenic differentiation (Figure 7). These findings support a mechanism underlying how lncRNA regulates bovine adipose development.

ACK N OWLED G EM ENTS
This study was supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflicts of interest with the contents of this article.

AUTH O R CO NTR I B UTI O N S
HC designed and performed research and wrote paper; ML and JW performed experiments; CS analysed data; YH and XL collected samples; CL revised manuscript; HC conducted this research.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets generated or analysed during the current study are available from the first authors and corresponding author on reasonable request.

F I G U R E 7
Model depicting how BADLNCR1 inhibits bovine adipocyte differentiation. TF, transcriptional factor