Cloning and functional verification of a porcine adipose tissue-specific promoter

Fat deposition is an important economic trait in pigs. In the past decades, many genes regulating porcine fat deposition were identified by Omics technology and verified by cell biology studies. Using genetically modified pigs to investigate the function of these genes in vivo is necessary before applying in breeding. However, lack of tissue-specific promoters of pigs hinders the generation of adipose tissue-specific genetically modified pigs. In order to identify a porcine adipose tissue-specific promoter, we used the software Digital Differential Display (DDD) to screen 99 genes highly expressed in porcine adipose tissue. GO and KEGG enrichment analysis indicated that the 99 genes were mainly related to lipid metabolism. Q-PCR proved that LGALS12 was an adipose tissue-specific gene. Five truncated fragments of the LGALS12 promoter were cloned and the 4 kb fragment (L-4 kb) exhibited a high level of promoter activity in adipocytes and no promoter activity in non-adipocytes. Following co-transfection with adipogenic transcription factors, the promoter activity of L-4 kb was enhanced by PPARγ, C/EBPβ, and KLF15, whereas it was suppressed by KLF4. Finally, we demonstrated that L-4 kb can drive APOR gene expression to exert its function in adipocytes. This study demonstrates that porcine LGALS12 is an adipose tissue-specific gene, and identified the 4 kb fragment of LGALS12 promoter that exhibited adipocyte-specific promoter activity. These results provide new evidence for understanding porcine fat deposition and a promoter element for adipose tissue-specific genetic modification in pigs. Identified porcine LGALS12 as an adipose tissue-specific gene. Truncated LGALS12 promoter (L-4 kb) showed adipose tissue-specific promoter activity. Identified transcription factors involved in the regulation of L-4 kb promoter activity.


Background
Genetically modified (GM) pigs are becoming increasingly important for both breeding and biomedical research [1]. For breeding purposes, GM pig research has focused on economically important traits, such as the growth rate [2], meat quality [3], disease/stress resistance [4] and feed conversion efficiency [5]. Although genetic modifications have not yet made real contributions to pig breeding, these methods have surpassed traditional breeding methods with rapid development and impressive results in a short period. For biomedical purposes, disease models have been established in pigs by genetic manipulation of key genes in disease pathways, which might offer a better representation of human pathology than rodent models for certain disease contexts [6].
Open Access † Dawei Zhang, Liangcai Shen and Wenjing Wu contributed equally to this work.
*Correspondence: zhangjin7688@163.com Furthermore, pigs are also considered to be the best source of organs for xenotransplantation, due to their anatomical and physiological analogy to humans [7]. In the past decade, gene editing technology represented by CRISRP/Cas9 has developed rapidly, and has greatly facilitated the generation of GM pigs. However, wholebody gene-editing might cause unexpected side effects, which makes it different to obtain healthy offspring for study. In these cases, tissue-specific GM animal models are necessary. Hundreds of tissue-specific GM rodent models have been reported, but only a few have been reported in pigs. The main reason is the lack of tissuespecific promoters for pigs.
Digital differential display (DDD) is a free software for comparing expression profiles among the different library pools in UniGene (finished by 2012-4-26) [8], offering a quick method to identify genes with expression levels that differ between different tissues, stages, or conditions in organisms [9,10]. In this study, pig adipose tissue-specific genes were screened by DDD analysis and verified by Q-PCR analysis.
LGALS12 showed the best specificity and its promoter was cloned. Furthermore, the transcriptional factors involved in LGALS12 regulation were identified, which makes it possible to generate adipose tissue-specific GM pigs.

Results
Screening of genes preferentially expressed in porcine adipose tissue DDD analysis was performed by comparing porcine adipose tissue libraries with porcine non-adipose tissue libraries including intestine, kidney, longissimus, lung, muscle, ovary, spleen, testis, thymus and uterus. A total of 5 adipose tissue libraries with 33,401 ESTs (pool A) versus 17 libraries of non-adipose tissues with 132,632 ESTs (pool B) were included in this analysis. The software identified 99 genes/transcripts with 10-fold higher expression levels in adipose tissue than in non-adipose tissues (Table S2).
These 99 genes/transcripts were subjected to Gene Ontology (GO) analysis to predict their potential functions. The top 20 enriched categories related to biological processes with P < 0.05 were listed (Fig. 1A). The results were mainly related to lipid metabolism (Lipid particle, Triglyceride catabolic process, Lipid catabolic process and Very-low-density lipoprotein particle) and ribosome function (Structure constituent of ribosome, Cytosolic large ribosomal unit, Large ribosomal subunit rRNA binding, Translation, Cytoplasmic translation and Cytosolic small ribosomal subunit). KEGG enrichment analysis was also conducted to identify the involved pathways (Fig. 1B). Four pathways, PPAR signaling pathway, Ribosome, AMPK signaling pathway and Fatty acid metabolism, were enriched, all of which are highly associated with lipid metabolism.
LGALS12 exhibited the highest specificity for porcine adipose tissue In order to confirm the results of bioinformatic analysis, 8 genes with more than 30-fold change were verified by Q-PCR analysis (Table 1 and Fig. 2). All 8 genes were preferentially expressed in adipose tissue, but their expression profiles varied significantly. The relative expression level of each gene in adipose tissue was compared, and was found to be in the order FABP4, ADIPOQ, CIDEC, LPL, LIPE, SDR16C5, LGALS12 and SMAF1. The expression of the most highly expressed gene FABP4 was 325 times higher than that of the least highly expressed SMAF1. The tissue expression patterns of these genes were analyzed and five genes, CIDEC, LPL, LIPE, SDR16C5 and SMAF1, were detected in more than three tissues, although the expression levels in non-adipose tissue were very low. FABP4 and ADIPOQ were detected in one tissue other than adipose tissue with very weak signals.
LGALS12 was specifically expressed in adipose tissue and was not detected in any other tissues. Therefore, LGALS12 exhibited the best specificity for porcine adipose tissue. However, few reports on porcine LGALS12 were found in the literature.

Cloning and activity analysis of the truncated LGALS12 promoter
To obtain an adipose tissue-specific promoter, a 5 kb genome sequence upstream of the LGALS12 start codon was analyzed. Five core promoter regions with scores over 0.85 were predicted ( Table 2). Binding sites for adipogenic transcriptional factors were also predicted ( Fig. 3A). Based on this analysis, the five truncated fragments indicated in Fig. 3 were cloned and named L-1kp, L-2 kb, L-3 kb, L-4 kb and L-5 kb. The fragments were used to construct the dual luciferase reporter vectors pGL3-L-1 kb/2 kb/3 kb/4 kb/5kB, respectively. Vectors with each truncated promoter were introduced

Identification of transcriptional factors affecting LGALS12 promoter activity
In order to screen the transcription factors affecting the activity of the LGALS12 promoter, 293 T cells were cotransfected with pGL3-L-4 kb and plasmids encoding five transcription factors, C/EBPα, C/EBPβ, PPARγ, KLF4 and KLF15. The effects of each transcription factor on the activity of the pGL3-L-4 kb promoter were analyzed using a dual luciferase reporter assay (Fig. 4A). The promoter activity of pGL3-L-4 kb was enhanced by PPARγ, C/EBPβ, and KLF15 while being suppressed by KLF4. In order to assess the combined effects of transcription factors, KLF15 was chosen as the highest promoting effector for co-transfection with C/EBPα, C/EBPβ and PPARγ, respectively (Fig. 4B). It was found that the transcriptional activity of pGL3-L-4 kb was improved by two combinations, KLF15 and C/EBPβ, as well as KLF15 and PPARγ. C/EBPα showed no effect on pGL3-L-4 kb in all assays.

Functional verification of LGALS12 promoter activity
To verify the function of LGALS12 promoter activity, the L-4 kb fragment was cloned into the multiple cloning sites of the vector pDsRed-Express-N1, resulting in pDsRed-L-4 kb. The pDsRed-Express-N1 vector contains a red fluorescent protein (RFP) reporter gene without a promoter, which is used to detect promoter activity of cloned segments. Apolipoprotein R (APOR) can promote lipolysis in adipocytes and its coding sequence without a stop codon was cloned in to the vector downstream of the L-4 kb sequence, resulting in pDsRed-L-4 kb-APOR. The red signal could be seen when 3 T3-L1 cells were transfected with pDsRed-L-4 kb-APOR, while no signal was detected in the control group transfected with pDsRed-APOR. After 8 days of differentiation, oil red O staining and triglyceride analysis were conducted. Cells transfected with pDsRed-L-4 kb-APOR showed lower Table 2 The Prediction of the Core Promoter of the LGALS12 Gene levels of lipid accumulation than control group (Fig. 5A  and B). Furthermore, the group transfected with pDsRed-L-4 kb-APOR exhibited higher levels of free fatty acids in the culture supernatant (Fig. 5C). All these data proved that L-4 kb possesses promoter activity and can drive a gene to play its function in adipocytes.

Discussion
Adipose tissue performs various physiological functions, including storing excess energy as fat, protecting inner organs from physical impact, keeping warm and secreting adipokines. Due to their highly developed adipose tissue, pigs are regarded as an ideal model for studying adipogenesis. Many important regulatory genes (including non-coding RNAs) for porcine fat accumulation have been identified by Omics technology and verified by cell biology studies in the past decades. Genetically modified (GM) pigs will be a powerful tool to unveil the functions of these genes and investigate possible side effects in vivo, which is necessary before application. However, the lack of porcine adipose tissue-specific genetic elements hinders the adipogenesis research based on tissue-specific GM pigs. There have been hundreds of research papers on adipose tissue-specific deficient/transgenic mice, but none on pigs so far. In this study, 99 genes/transcripts were identified highly expressed genes in porcine adipose tissue. GO analysis and KEEG enrichment indicated that the 99 genes were mainly related to lipid metabolism. By Q-PCR analysis, LGALS12 were proved to be adipose tissue specific. According to bioinformatic analysis, five truncated fragments of the LGALS12 promoter were cloned and the 4 kb fragment (L-4 kb) exhibited adipose tissue-specific promoter activity. Mechanistically, the promoter activity of L-4 kb was enhanced by PPARγ, C/ EBPβ, and KLF15, while it was suppressed by KLF4. Finally, we proved that L-4 kb can drive APOR gene expression to play its function in adipocytes.
With the development of Omics technology, a large amount of data has been accumulated in the field of biology, and making efficient use of these data is a focus of research. Digital Differential Display (DDD) is a free online tool that can be used to compare EST-based expression profiles among different libraries, or pools of libraries, represented in UniGene [8], which is easy to learn and use. It can quickly screen the target gene set in the database, and is mainly used to screen genes with time-or space-specific expression [11,12]. We used DDD to screen 99 genes out of 33,401 ESTs and focused on the top eight genes. All these eight genes were preferentially expressed in porcine adipose tissue, which indicated that DDD analysis could efficiently provide useful information.
According to Q-PCR analysis, we identified LGALS12 as a porcine adipose tissue-specific gene.
LGALS12 was firstly reported in 2001 by two research groups independently [14,15]. In 2011, one of the two groups proved that LGALS12 deficiency promoted lipolysis in knockout mice [16]. We found that LGALS12 was a key molecule in porcine fat deposition and proved that LGALS12 affected porcine intramuscular and subcutaneous adipogenesis via different signaling pathways [17]. All these data support the idea that LGALS12 is a candidate gene for genetic improvement of fat-related traits in pigs.
Although it was demonstrated to be an adipocyte-specific gene, the transcriptional mechanism of LGALS12 remains unclear. Here, we found that several transcription factors related to lipid metabolism play a pivotal role in LGALS12 transcription. PPARγ, C/EBPβ and KLF15 could promote LGALS12 transcription, while KLF4 had the opposite effect. The expression of C/EBPβ is an early event during the differentiation of adipocytes, and it can induce the expression of C/EBPα and PPARγ, which are the master transcription factors for adipogenesis. Since LGALS12 was expressed after adipogenic induction, it is expected that the transcription of LGALS12 is regulated by C/EBPβ and PPARγ. KLF15 and KLF4 belong to a family of zinc finger transcription factors, which play diverse roles during mammalian cell differentiation and development. KLF15 expression is markedly increased during the differentiation of 3 T3-L1 pre-adipocytes. Knockdown of KLF15 reduced the expression of PPARγ and suspended adipogenesis in 3 T3-L1 cells. Ectopic expression of KLF15 in NIH 3 T3 or C2C12 cells induced PPARγ expression and promoted triglyceride accumulation when the cells were exposed to an adipogenic medium [18].. Adipose tissue-specific KLF15 knockout (AK15KO) mice exhibited decreased adiposity and increased lipolysis. Moreover, AK15KO mice showed resistance to high-fat diet induced obesity and insulin resistance [19]. Mechanistic studies revealed that KLF15 regulates genes related to triglyceride synthesis and suppresses lipolysis. In this study, KLF15 led to the largest increase of LGALS12 promoter activity, contributing new evidence for the role of KLF15 in adipocytes. However, KLF4 suppressed LGALS12 promoter activity in this study. KLF4 was discovered in 1996 and was proved to be one of four factors required for the induction of pluripotent stem cells (iPSCs) in 2006 [20]. The function of KLF4 in adipocytes was not elucidated so far, but a study showed that KLF4 inhibits the differentiation of intramuscular preadipocytes by targeting C/EBPβ [21], which supports our finding that KLF4 can suppress adipogenesis.
We cloned a 4 kb fragment of the LGALS12 promoter (L-4 kb) and proved its tissue-specific promoter activity in adipocytes using a dual-luciferase reporter assay. In order to test its function, L-4 kb was used to drive apolipoprotein R (APOR) gene expression in adipocytes. In a previous study, we found that overexpression of APOR could increase lipolysis in adipocytes [22], and adipose tissue-specific expression of pig APOR protected mice from diet-induced obesity. In this study, decreased triglyceride accumulation in cells and an increased level of non-ester fatty acids in the culture supernatant were observed, which showed that APOR expression driven by L-4 kb recapitulated its physiological role in 3 T3-L1 cells.

Conclusions
In summary, the present study demonstrates that LGALS12 is an adipose tissue-specific gene in pigs. Transcription factors that regulate the LGALS12 promoter were identified and the 4 kb fragment of porcine LGALS12 promoter exhibited adipocyte-specific promoter activity. Our finding provides new evidence for understanding porcine fat deposition and a promoter element for adipose tissue-specific genetic modification in pigs.

Animals
Three 4-month-old large white pig were provided by Zhejiang Huateng Agricultural Technology Co., Ltd. Pigs were dissected for sampling after euthanatizing in CO 2 euthanasia box. The longissimus dorsi muscle, subcutaneous adipose tissues, cardiac apex, the right lobe of liver, spleen, right kidney, the upper lobe of the right lung and intestine tenue (jejunum) were collected in liquid nitrogen and stored at − 80 °C until RNA extraction. Jiaxing University Animal Care Committee approved and verified all the experimental procedures and followed ARRIVE guidelines to perform the experiments [23].

Digital differential display (DDD) analysis
Digital Differential Display (DDD) is an algorithmic system for the identification of differentially expressed genes based on the relative abundance of expressed sequence tags (ESTs) from two or more contrasting cDNA libraries, which are deposited in the NCBI UniGene database (http:// www. ncbi. nlm. nih. gov/ UniGe ne/). DDD compares the number of assignments of ESTs from several different libraries, or pools of libraries, to a specific UniGene cluster. To account for the unequal number of ESTs in each library, DDD utilizes Fisher's exact test to restrict the output to statistically significant differences (P ≤ 0.05) [9]. Gene expression levels of adipose tissue derived cDNA libraries (pool A) against the other organ-specific cDNA libraries (pool B including intestine, kidney, longissimus, lung, muscle, ovary, spleen, testis, thymus, uterus) were compared. Genes with statistically significant differential expression between adipose and non-adipose tissues were recorded. For these recorded genes, fold changes of expression levels were calculated by dividing the frequency of that gene in pool A (adipose tissue) to the frequency in pool B (non-adipose tissue). More than 10-fold changed genes were subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Gene and Genome (KEGG) [24] pathway analysis using the online databases DAVID 6.8.

Q-PCR
Total RNA was isolated from pig tissues using Trizol reagent (Invitrogen, USA). The HiFiScript gDNA Removal cDNA Synthesis Kit (CWBiotech, Beijing, China) was sued for cDNA synthesis. Q-PCR was conducted using the 2 × plus SYBR real-time PCR mixture (BioTek, Beijing, China) on a QuantStudio 3 system (Thermo fisher, USA). The primer sequences are listed in Table S1. Relative gene expression levels were calculated using the 2 -ΔΔCt method with GAPDH as the internal control.

Promoter analysis
The BDGP [25] online tool was used to predict core promoter sequence, and the type of organism selects eukaryote and the threshold score was set by default to 0.8. The JASPAR [26] website was used to identify the transcription factor binding sites, and the taxonomic group was set to vertebrata.

Vector construction
The pDsRed-Express-N1 vector was amplified by PCR using primers 5′-ATA GAG CTC CGC GG.
AAC TCC ATA TAT GGG-3′ and 5′-CTC GAG CTC AAG CTT CGA ATT CTG C-3′, then digested by SacI and ligated to construct a new tool vector without CMV promoter. The LGALS12 promoter (− 4061 to + 121) was cloned into the new vector digested by SacI and SalI, and the primers were 5′-ATA GAG CTC AGC ATA CAT GGA ATA CAT GCT G-3′ and 5′-ATA GTC GAC GCC CAA CTG AGCC.

Cell culture and transfection
Preadipocyte cell line 3 T3-L1 was purchased from ATCC, and cultured in DMEM containing 10% fetal bovine serum (FBS, Gibco, LOT2206993CP) to confluence (day 0), and then shifted to adipocyte differentiation medium, which was DMEM with 10% FBS, 0.5 mM dexamethasone, 20 nM insulin, and 0.5 mM isobutyl methylxanthine (IBMX) for 2 days (Day 3). From day 4 to day 8, cells were maintained in DMEM with 10% FBS and 20 nM insulin, and the medium was replaced every other day. Lipofectamine ™ 2000 (11668027, Invitrogen, USA) was used for plasmid transfection at the indicated time points. Then, the medium was replaced with fresh medium in 24 hours.

Luciferase reporter assays
HEK293T were seeded in 24-well plates and co-transfected with LGALS12 promoter and transcription factor (PPARα、C/EBPα、CEBPβ、KLF4 and KLF15) vector by VigoFect (Vigorous Biotech, Beijing). The pGL3-Basic and pCR3.1 vector were used to insert promoter and transcription factor sequence, respectively. The pTK-Renila luciferase reporter (Promega) was included in all transfections for normalization. Luciferase activities were measured after transfection for 24 h using the dual-luciferase reporter assay system (Promega), and each combined set of vectors had three independent replicates. The measure was performed as described [27] with some modification. Tecan's Spark multimode microplate reader was used to measure fluorescence intensity, and the comparison was conducted based on the ratio of firefly luciferase activity to Renilla luciferase activity.

Oil red O staining
The matured 3 T3-L1 cells were washed with PBS and fixed with 4% paraformaldehyde for 30 min at room temperature, then washed three times with PBS. The fixed cells were then covered with a mixture of Oil Red O solution (0.6% Oil Red O dye in isopropanol) and water at a 6:4 ratio for 30 min, followed by washing four times with PBS, and images were captured under an optical microscope (Leica Microsystems, Germany).

Statistical analysis
The data were obtained from at least three independent experiments, and presented as the means ± standard error (SE). GraphPad Prism 8.0 (GraphPad, San Diego, CA, USA) were used to conduct the statistical analysis, and the assessment of normality are Kolmogorov-Smirnov (K-S) test. Individual comparisons were assessed using Student's t-test. P-values of less than 0.05 were considered to indicate significant differences, are displayed as * p < 0.05, ** p < 0.01, and *** p < 0.001.