Abstract
Sirtuin 2 (Sirt2) belongs to the NAD+-dependent deacetylase family, is more highly expressed than other family members in adipocytes, and plays crucial roles in a wide range of biological processes. However, the mechanisms underlying Sirt2 expression during adipogenesis are poorly studied. In this study, the transcriptional start site (TSS) of Sirt2 was identified and two alternative transcript variants were spliced from Sirt2. The 5′-regulatory region of Sirt2 was also characterized; no TATA-box or CCAAT-box was presented in the 5′-flanking region. Two cytosine-phosphate diester-guanine (CpG) islands were also identified between nucleotides −563 and +4. A dual-luciferase reporter assay revealed that a 178 base pair sequence upstream from the TSS (+1) was the core promoter of Sirt2. Results from a site-directed mutagenesis experiment, electrophoretic mobility shift assay, and chromatin immunoprecipitation assay indicated Yin Yang 1 (YY1) to be a positive regulator of bovine Sirt2 in preadipocytes. YY1 is likely to suppress adipogenesis in two different ways by regulating peroxisome proliferator-activated receptor gamma expression. Our results expand the information on the regulatory network of adipogenesis, which is an important basis for improving beef quality, treating obesity, and other related diseases.
Introduction
Sirtuins consist of seven members (sirtuin 1–7) and are NAD-dependent histone deacetylases (Ma et al., 2019). Sirtuins play roles in various physiological processes through modulating epigenetic changes (Bosch-Presegue and Vaquero, 2015) and are involved in diabetes (Aditya et al., 2017), cardiac diseases (Matsushima and Sadoshima, 2015), neurodegenerative diseases (Herskovits and Guarente, 2013), pulmonary fibrosis (Shaikh et al., 2018), and cancer (O'Callaghan and Vassilopoulos, 2017; Geng et al., 2018).
Sirtuin 2 (Sirt2) differs from other sirtuins, as it can shuttle between the cytoplasm and nucleus (Dryden et al., 2003; North and Verdin, 2007). Sirt2 is involved in a wide range of biological processes, including fatty acid oxidation, adipogenesis, and insulin sensitivity (Gomes et al., 2015). In mice, Sirt2 knockouts displayed reduced skeletal muscle insulin-induced glucose uptake (Lantier et al., 2018). Overexpression of Sirt2 increased hepatic glucose uptake in high-fat diet mice and mitigated impaired glucose tolerance by deacetylating glucokinase regulatory protein (Watanabe et al., 2018). In humans, overexpression of Sirt2 in the visceral adipose stem cells of obese subjects inhibited differentiation into adipocytes, while knockdown of Sirt2 promoted adipogenesis (Perrini et al., 2020). In Qinchuan cattle, the expression level of Sirt2 was associated with the body size traits (Gui et al., 2015).
Intramuscular fat content is associated with juiciness, tenderness, and flavor of meat (Platter et al., 2005), and is largely controlled by adipogenesis. Adipogenesis is highly regulated by a series of transcription factors, including peroxisome proliferator-activated receptor gamma (PPARγ), and members of the CCAAT/enhancer-binding protein family (C/EBPα, C/EBPβ, and C/EBPδ) (Mota de Sa et al., 2017). However, the full regulatory network of adipogenesis is far from defined. A previous study showed Sirt2 was more highly expressed than other members in adipocytes, and was downregulated during differentiation of 3T3-L1 cells (Jing et al., 2007). Overexpression of Sirt2 suppressed 3T3-L1 cells differentiation, while Sirt2 knockdown promoted differentiation, implying that Sirt2 negatively regulates adipogenesis (Jing et al., 2007). Furthermore, Sirt2 was reported to promote the binding of FOXO1 to PPARγ by deacetylating FOXO1 to inhibit the expression of PPARγ, resulting in the suppression of adipogenesis (Wang and Tong,
Materials and Methods
Ethical approval
All animal procedures were performed according to the guidelines of the China Council on Animal Care and the protocols were approved by the Experimental Animal Management Committee (EAMC) of Northwest A&F University (Approval ID: 2015-ZX120101).
Sample collection and preparation
Qinchuan cattle were fed in the National Beef Cattle Improvement Center's farm belonged to Northwest A&F University. Three bulls at 3 days of age were slaughtered after stunned. The blood were obtained for DNA extraction using a blood DNA kit (Omega Bio-Tek, Doravile, USA). Inguinal adipose tissues were dissected and collected within 20 min after slaugher, then snap-frozen in liquid nitrogen for chromatin immunoprecipitation (ChIP) assay.
Molecular cloning and bioinformatics analysis
Based on the bovine Sirt2 reference sequence (UMD3.1.1, GenBank accession no. AC_000175.1), a pair of primers (Sirt2-PF/PR; Table 1) was designed to amplify the 5′-regulatory region of the Sirt2 gene. Using DNA extracted from the whole blood of Qinchuan cattle as a template, polymerase chain reaction (PCR) was performed according to the instructions of KOD-Plus high-fidelity DNA polymerase (TOYOBO; Biotechnology Co. Ltd., Shanghai, P.R. China). PCR products were purified and inserted into a pMD 19-T vector (TaKaRa, Biotechnology Co. Ltd, Dalian, P.R. China) before sequencing. To uncover potential transcription factor binding sites, the 5′-regulatory region was analyzed using Patch 1.0 (
Sequence Information of Primers
Note: The bold sequence “GGTACC” is the KpnI restriction endonuclease site and “AAGCTT” is the HindIII restriction endonuclease site. The italicized sequence “ACATGG” is the predicted YY1-binding site. The sequence “ATG” wthin YY1-binding site “ACATGG” is mutated to the sequence “GAA.”
ChIP, chromatin immunoprecipitation.
5′-Rapid amplification of cDNA ends
The PCR primers (Sirt2-GSP and Sirt2-NGSP; Table 1) were used to identify TSSs of the Sirt2 gene. All other steps were performed as previously described (Li et al., 2010, 2015, 2016a).
Cell culture, transfection, and dual-luciferase reporter assay
3T3-L1 preadipocytes were incubated in 24-well plates containing 10% newborn calf serum (Gibco-Invitrogen, Carlsbad, CA, USA) at 37°C and 5% CO2. Eight serial deletion products of the 5′-regulatory region were amplified with 5′-nested PCR primers (Sirt2-KF1∼KF8; Table 1) and a common 3′-primer (Sirt2-HR; Table 1). All remaining steps were performed as previously described (Li et al., 2010, 2015).
Conservation analysis of YY1
Sequence similarities of YY1 amino acid sequences among vertebrates were analyzed using protein BLAST. To further analyze the conserved protein domains within mammals, YY1 amino acid sequences were aligned using Clustal × 2.0 software.
Site-directed mutagenesis
Corresponding primers (YY1-MT-F/R; Table 1) were used in the site-directed mutagenesis experiment. All other steps were performed as previously described (Li et al., 2010, 2015).
Electrophoretic mobility shift analysis
Nuclear proteins from 3T3-L1 cells were obtained according to the guidelines provided by the manufacturer of the Nuclear Extract Kit (Active Motif, Carlsbad, NM, USA). 100 fmol of 5′ biotin-labeled YY1, 1 μL of poly(dI.dC), 2 μL of 10 × binding buffer, and 10 μg of nuclear protein extracts were incubated on ice. Meanwhile, in the competition experiments, unlabeled cold competitor YY1-WT and mutated probe YY1-MT were separately added in 50-fold excess. To test the interaction of anti-YY1 antibody (sc-281; Santa Cruz Biotechnology, CA, USA) with the mobility complex, 5 μg of YY1 antibody was added to nuclear protein extracts before the addition of 5′-biotin-labeled YY1. After incubation on the ice, all mixtures were analyzed through electrophoresis on 6% polyacrylamide gels. Detail of the steps were supplied in the instructions for the LightShift® Chemiluminescent EMSA Kit (Pierce Corp., Rockford, IL, USA). Images were obtained using an Amersham Imager 600.
Chromatin immunoprecipitation assay
Inguinal adipose tissue was obtained from three bulls of Qinchuan cattle at 3 days of age. The crosslinking of DNA-protein complexes was achieved by incubation at 37°C with 37% formaldehyde, then terminated by adding glycine. After sonication, chromatin was sheared into 200–500 bp DNA fragments. Fifty microliters of sonicated sample was immunoprecipitated overnight at 4°C with anti-YY1 antibody or normal mouse anti-IgG antibody. The primers ChIP-YY1-F/R (Table 1) were used for reverse transcription (RT)-PCR and ChIP-qPCR, and the immunoprecipitated product from the normal mouse IgG group was used as a negative control. RT-PCR products were verified by sequencing. The remaining steps were performed as described in our previous publication (Li et al., 2016b).
Statistical analyses
All statistical analyses were performed as described in our previous study (Li et al., 2016b; Mi et al., 2019).
Results and Discussion
Identification of the bovine Sirt2 gene TSS
To identify the TSS of bovine Sirt2, a 5′-rapid amplification of cDNA ends (5′-RACE) experiment was performed. Two contiguous products with different lengths were obtained (Fig. 1a). The sequencing results from 50-positive clones revealed 6 different 5′ ends, which were 85, 66, 52, 42, 37, and 33 bp upstream of the translational start site (ATG) (Fig. 1b). The guanine (G) residue, 85 bp upstream of “ATG,” was designated as +1. Furthermore, sequence alignment results showed that the shorter of the two contiguous products was missing a 47-bp fragment deleted from +102 to +148 (covering the full exon 2 of the bovine Sirt2 gene) compared with the longer fragment (Fig. 1c). These results revealed that the Sirt2 gene had two alternatively spliced transcript variants, which is generally consistent with that in humans, mice, and pigs (North and Verdin, 2007; Liu et al., 2010).

Characteristics of the bovine Sirt2 gene promoter.
Characterization of the bovine Sirt2 gene 5′-regulatory region
To characterize the 5′-regulatory region of the Sirt2 gene, a 2017 bp fragment spanning from −1956 to +61 was cloned from bovine blood genomic DNA. As shown in Figure 1b, many transcription factor binding sites were identified, including PPARγ, C/EBPα, C/EBPβ, stimulating protein 1 (SP1), lymphoid enhancer-binding factor 1 (LEF1), YY1, activator protein 2 alpha (AP2), cAMP response element-binding protein (CREB), and E2F-myc activator/cell cycle regulator 1 (E2F1). Interestingly, neither a TATA-box nor CCAAT-box was found near the TSS. However, two CpG islands spanning the regions −563 to −322 (island 1) and −281 to +4 (island 2) were predicted using MethPrimer (Fig. 1d), which is similar to that in humans (Voelter-Mahlknecht et al., 2005). In our previous study, Sirt2 was widely expressed in tissues (Gui et al., 2015), which is consistent with the hypothesis that multiple TSSs, no TATA box, and CpG islands in the promoter region are associated with widespread expression of a gene (Lenhard et al., 2012; Danino et al., 2015).
Identification of the bovine Sirt2 gene core promoter
To identify the core promoter of Sirt2, eight serial deletion constructs in pGL3-basic were generated containing −1956/+61, −1333/+61, −775/+61, −513/+61, −234/+61, −178/+61, −71/+61, and +4/+61, respectively. These constructs were transfected into preadipocytes: the promoter activity of construct −1956 to +61 was ∼100-fold higher than the empty pGL3-basic vector (p < 0.01) (Fig. 2), indicating that a functional bovine Sirt2 promoter was present within this region. Deletion of the region from −1956 to −775 revealed no significant change (p > 0.05) in luciferase activity. Indeed, upon deleting the sequences from −775 to −234, promoter activity significantly increased (p < 0.01) (Fig. 2), suggesting that the region from −775 to −234 could contain crucial inhibitory elements. Strikingly, upon deleting the region from −178 to −71, the promoter activity decreased sharply (p < 0.01) (Fig. 2), suggesting that this region might contain crucial activated elements. Deletion of the region from −71 to +4 led to almost complete abrogation of transcriptional activity (p < 0.01) (Fig. 2). These results indicate that the sequence of 178bp upstream from the TSS (+1) is the core promoter of Sirt2.

Promoter activity analysis of the bovine Sirt2 gene. Eight serial deletion constructs in pGL3-basic were transfected into 3T3-L1 cells. Firefly luciferase activities were normalized to renilla luciferase activity; relative luciferase activities were normalized to that of pGL3-basic. Relative luciferase activities are averages of three independent transfections. **p < 0.01. Error bars represent the SD (n = 3). Numbering is relative to the TSS (+1). SD, standard deviation.
To identify conserved regulatory elements among cattle, humans, mice, and rats, the region of nucleotides from −178 to +4 was analyzed using Patch 1.0 and MatInspector software to predict transcription factor binding sites. Multiple sequence alignments showed that the binding sites for SP1, YY1, and LUN-1 were conserved among the four species, while the E2F1-binding site was conserved between cattle and humans (Fig. 3a).

Multiple sequence alignment analysis of the core functional promoter of bovine Sirt2 and YY1.
Han et al. (2015) showed that YY1 overexpression suppressed preadipocyte differentiation, while knockdown of YY1 promoted preadipocyte differentiation. To better understand the transcription factor YY1, the amino acid sequence similarities between bovine YY1 and other vertebrate YY1 proteins were analyzed. Bovine YY1 shared high sequence similarity with YYI from other vertebrates, including dog (99%), pig (99%), human (98%), rat (96%), mouse (94%), frog (78%), and fish (76%). The DNA-binding domain (C-terminal tail, residues 295–414) and the transcriptional activation domain (N-terminal domain, residues 1–69) of YY1 were completely conserved among cattle, humans, mice, and rats (Fig. 3b).
YY1 was required for the transcription of Sirt2 gene
YY1 participates as an initiator, activator, or repressor in the transcriptional control of ∼10% of known mammalian genes (Khachigian, 2018). Therefore, we hypothesized that YY1 regulated the transcription of bovine Sirt2 in preadipocytes.
Mutation of the YY1-binding site from position −116 to −114 (“ATG” mutated to “GAA”) resulted in a significant decrease in luciferase activity, which was ∼60% compared to that of wild-type samples (Fig. 4a). This suggests that the YY1-binding site promotes Sirt2 transcription.

YY1 is a positive transcription factor of the Sirt2 gene.
To verify the interaction between YY1 and the promoter region of Sirt2, an electrophoretic mobility shift assay (EMSA) experiment was performed. The results showed that 5′ biotin-labeled YY1 probes associated with nuclear proteins of preadipocytes, which resulted in upshifts of bands (Fig. 4b, lane 2). When unlabeled wild-type YY1 was mixed, an upshifted band resembling the complex was barely visible (Fig. 4b, lane 3). When unlabeled mutant YY1 was mixed, an upshifted band did not disappear (Fig. 4b, lane 4). Furthermore, the addition of anti-YY1 antibody led to a strong decrease in the upshifted band (Fig. 4b, lane 5). Taken together, these results demonstrate that YY1 interacted with YY1-binding sites of the bovine Sirt2 promoter in vitro.
To confirm an interaction between the transcription factor YY1 and the 5′-regulatory region of Sirt2 in vivo, a ChIP assay was carried out. A 193 bp of DNA fragment was amplified from the precipitate obtained using anti-YY1 antibody, but not from mouse anti-IgG antibody (Fig. 4c, left). Furthermore, the enrichment of anti-YY1 antibody was significantly higher than that of anti-IgG antibody (Fig. 4c, right). These results indicate that YY1 can specifically bind to the YY1-binding site of the Sirt2 promoter in vivo.
YY1 is involved in a range of biological processes with key roles in cellular differentiation and tissue development (Chen et al., 2017). In 3T3-L1 cells, YY1 inhibited adipogenesis by competing with C/EBPβ for the PPARγ promoter, which decreased the expression of PPARγ (Han et al., 2015). A previous study showed that Sirt2 suppressed adipogenesis by deacetylating FOXO1 to promote FOXO1's binding to PPARγ (Wang and Tong, 2009). In the present study, results suggest that YY1 promoted transcription of the bovine Sirt2 gene. Both the YY1 amino acid sequence and the YY1-binding site in the Sirt2 promoter were conserved among cattle, humans, mice, and rats. Therefore, we speculate that YY1 can suppress adipogenesis via two ways by regulating the expression of PPARγ. One passway is YY1 suppresses PPARγ expression by binding to the PPARγ promoter region. Another passway is YY1 positively regulates the transcription of SIRT2, and SIRT2 deacetylates FOXO1 to inhibit the transcriptional activity of PPARγ (Fig. 5). Our study expands the information on the regulatory network of adipogenesis, which provides an important basis for improving beef quality, treating obesity, and related diseases.

Sketch map of cross-talk among YY1, SIRT2, and PPARγ during adipogenesis. This model reflects both prior findings and current findings. Dotted box with “Ac” represents deacetylation of protein. “A” represents the inhibitory effects of YY1 on PPARγ by competing with C/EBPβ binding to the PPARγ promoter region, thereby suppressing PPARγ expression, or directly repressing transcriptional activity of PPARγ by interaction with YY1. “B1” indicates that YY1 positively regulates the transcription of SIRT2. “B2” indicates that SIRT2 deacetylates FOXO1 in adipocytes. “B3” indicates that deacetylated FOXO1 represses the transcriptional activity of PPARγ by interaction with FOXO1. Taken together, YY1 could inhibit PPARγ through the “A” pathway and/or “B” (“B1” + “B2” + “B3”) pathway, thereby suppressing adipogenesis. PPARγ, peroxisome proliferator-activated receptor gamma.
Conclusion
In summary, this study identified the transcription start site and two alternatively spliced transcript variants of Sirt2; a 5′ regulatory region sequence of the Sirt2 gene was also characterized. A sequence of 178 bp upstream from the TSS (+1) was identified as the core promoter of Sirt2. YY1 was also an important positive regulator of the bovine Sirt2 gene. These results provide an important basis for understanding the transcriptional regulation of Sirt2, which could provide insight into its role in modulating beef marbling, obesity, and related diseases.
Footnotes
Acknowledgment
We are very grateful to Stephen S. Moore (University of Queensland) for revising this article.
Authors' Contributions
A.L., Y.Z., and L.Z. conceived and designed the experiments. H.W., C.M., and Y.L. participated in the sample collection. Y.Z. performed all the experiments. A.L. wrote the article and Y.Z. participated in the methods writing. H.W., C.M., and Y.L. revised the article. L.Z. and A.L. contributed reagents and materials. All authors read and approved the final article.
Disclosure Statement
No competing financial interests exist.
Funding Information
This research was supported by the National 863 Program of China (#2013 AA102505), National Science and Technology Support Projects (#2015BAD03B04), National Modern Agricultural Industry Special Program (#CARS-37), and Technical Innovation Engineering Project of Shaanxi Province (#2016 KTCL02-15).
