LPIN1 promotes triglycerides synthesis and is transcriptionally regulated by PPARG in buffalo mammary epithelial cells

Studies on 3T3-L1 cells and HepG2 hepatocytes have shown that phosphatidic acid phosphohydrolase1 (LPIN1) plays a key role in adipogenesis, acting as a co-activator of peroxisome proliferator-activated receptor gamma coactivator 1a (PGC-1a) to regulate fatty acid metabolism. However, the functional role and regulatory mechanism of LPIN1 gene in milk fat synthesis of buffalo are still unknown. In this study, overexpression of buffalo LPIN1 gene transfected with recombinant fusion expression vector significantly increased the expression of AGPAT6, DGAT1, DGAT2, GPAM and BTN1A1 genes involved in triglyceride (TAG) synthesis and secretion, as well as PPARG and SREBF1 genes regulating fatty acid metabolism in the buffalo mammary epithelial cells (BMECs), while the lentivirus-mediated knockdown of buffalo LPIN1 dramatically decreased the relative mRNA abundance of these genes. Correspondingly, total cellular TAG content in the BMECs increased significantly after LPIN1 overexpression, but decreased significantly after LPIN1 knockdown. In addition, the overexpression or knockdown of PPARG also enhanced or reduced the expression of LPIN1 and the transcriptional activity of its promoter. The core region of buffalo LPIN1 promoter spans from − 666 bp to + 42 bp, and two PPAR response elements (PPREs: PPRE1 and PPRE2) were identified in this region. Site mutagenesis analysis showed that PPARG directly regulated the transcription of buffalo LPIN1 by binding to the PPRE1 and PPRE2 on its core promoter. The results here reveal that the LPIN1 gene is involved in the milk fat synthesis of BMECs, and one of the important pathways is to participate in this process through direct transcriptional regulation of PPARG, which in turn significantly affects the content of TAG in BMECs.

In order to determine whether the expression of LPIN1 affects the accumulation of TAG in the BMECs, the intracellular TAG content was measured using the TAG kit, and the results showed that the content of TAG in the pEGFP-C1-LPIN1 group was significantly higher than that in the pEGFP-C1 group (P < 0.01; Fig. 2D).
The LPIN1-sh2 was selected to prepare lentivirus particles (Lv-LPIN1-sh2) and further transfected into BMECs. The results showed that the mRNA abundance of LPIN1 in the Lv-LPIN1-sh2 group was significantly lower than that in control (50%) (P < 0.01; Fig. 3C). The knockdown of LPIN1 significantly decreased the mRNA abundance of genes related to milk fat synthesis in the BMECs (Fig. 3C), and the relative expression of AGPAT6, DGAT1, DGAT2, GPAM and BTN1A1 genes involved in TAG synthesis decreased by 73%, 67%, 29%, 68% and 33%, respectively. Meanwhile, the knockdown of LPIN1 also decreased the transcriptional levels of genes involved Scientific Reports | (2022) 12:2390 | https://doi.org/10.1038/s41598-022-06114-w www.nature.com/scientificreports/ in regulating fatty acid metabolism. For example, the expression of PPARG and SREBF1 was reduced by 22% and 83%, respectively, but it had no significant effect on the expression of SREBF2 (Fig. 3C). Correspondingly, compared with the control group, the TAG content of the BMECs in the Lv-LPIN1-sh2 group was significantly reduced (P < 0.01) (Fig. 3D).
Characterization and activity analysis of buffalo LPIN1 promoter. We obtained a sequence of 1920 bp at the 5' end of the buffalo LPIN1 gene by PCR, which contains 1878 bp upstream of the transcription start site (TSS) (+ 1) (Fig. 4A). Bioinformatics analysis showed that the G + C content of the amplified sequence was as high as 56.67%, containing a GC box (− 11 bp to − 19 bp), but not a typical TATA box. The sequence also contained several important cis-acting elements: PPAR response elements (PPRE1: − 481 bp to − 493 bp; PPRE2: − 86 bp to − 100 bp), sterol regulatory element (SRE: − 691 bp to − 700 bp) and the sites of nuclear factor (NF-Y: − 709 bp to − 715 bp), C/EBPα (− 408 bp to − 425 bp) and cAMP-response element binding protein (CREB: − 827 bp to − 835 bp) (Fig. 4B). The analysis by Methprimer software showed that the amplified sequence also contains a CpG island (Fig. 4C).
The pGL4-LPIN1 luciferase recombinant vector containing LPIN1 gene promoter (− 1878 bp to + 42 bp) was constructed, then transfected into the BMECs to determine the luciferase activity. Compared with the control (pGL4.11), the luciferase activity in the BMECs in the pGL4-LPIN1 group was significantly increased (~ 62 fold), indicating that the − 1878 to + 42 region of LPIN1 gene contains the promoter region of this gene ( Fig. 4D; P < 0.01).
Identification of the core promoter region of LPIN1 gene. A series of luciferase reporter recombinant vectors with 5' flanking deletion fragments of the LPIN1 promoter were constructed (− 1523/ + 42, − 1264/ + 42, − 953/ + 42, − 666/ + 42, − 416/ + 42, − 336/ + 42, − 172/ + 42 and − 88/ + 42). After transfecting these recombinant vectors into BMECs, luciferase activity was detected (Fig. 5). The deletion from − 1878 bp to − 1523 bp showed a decrease of luciferase activity that was trending toward significance (P = 0.07); when the deletion reached the − 1264 bp, the luciferase activity was significantly increased (P < 0.01); When the deletion reached − 953 bp, the luciferase activity was significantly reduced (P < 0.01), indicating that there are negative regulatory elements in the region from − 1264 bp to − 953 bp. In addition, the deletion from − 953 bp to − 666 bp has the highest luciferase activity (P < 0.01), and the deletion from − 666 bp to − 416 bp reduced the luciferase activity by about 56% (P < 0.01). Subsequently, with the deletion of the 5' flanking region, the activity continued to decline (Fig. 5). This indicates that the fragment between − 666 bp and + 42 bp contains cis-functional elements required for www.nature.com/scientificreports/ transcriptional activation of LPIN1 gene. The above results reveal that the core region of the LPIN1 promoter is located in the region from − 666 bp to + 42 bp, which has basic transcriptional activity.
Overexpression of PPARG increased the activity of LPIN1 promoter. In order to determine whether PPARG has a regulatory effect on the LPIN1 gene promoter, the pGL4-LPIN1 vector or 5' deletion mutagenesis constructs were transfected with pEGFP-C1-PPARG to detect their luciferase activity. The results showed PPARG gene was successfully overexpressed in the BMECs, and its expression level was 540 times that of the control (pEGFP-C1) (Fig. 6A). In addition, the overexpression of PPARG significantly increased the mRNA expression of LPIN1 gene (~ 6.8 fold; Fig. 6B). Compared with the control (pEGFP-C1), after overexpression of PPARG (pEGFP-C1-PPARG), the activity of the 5' deletion mutagenesis constructs was significantly increased except for the deletion fragment − 88/ + 42 (Fig. 6C). Therefore, it is speculated that there are PPRE elements in the LPIN1 promoter, and PPARG may regulate the expression of the LPIN1 gene by combining with the PPRE elements.
Interference of PPARG decreased the activity of LPIN1 promoter. After the shRNA mediatedknockdown of PPARG was conducted in BMECs, the mRNA expression of LPIN1 and PPARG and the luciferase activity of each promoter fragment were measured. The results showed that after interference with the PPARG gene (Lv-PPARG-sh3), the mRNA level of this gene was significantly reduced (P < 0.01; Fig. 7A), and the mRNA level of the LPIN1 gene also decreased significantly (P < 0.01; 43%; Fig. 7B). In addition, compared with the control (Lv-NC), after interfering with PPARG, except for that of the − 88 bp/ + 42 bp fragment vector, the luciferase activity of all the 5' deletion fragment vectors decreased significantly (Fig. 7C). This result further confirms that there are PPARG binding target sites (PPREs) in the LPIN1 promoter, and PPARG may directly regulate its expression by binding to the PPREs on the promoter of LPIN1 gene. To determine whether PPARG regulates the expression of the gene by binding to the two predicted PPREs (PPRE1 and PPRE2) of the LPIN1 promoter, recombinant vectors with the site-directed mutations on the PPREs located in the core promoter region of LPIN1 (− 666/ + 42 bp) were constructed, respectively, including the recombinants for site-directed mutagenesis of PPRE1 and PPRE2, and vector for simultaneous mutagenesis of both. Luciferase detection showed that both PPRE1 and PPRE2 site-directed mutagenesis could significantly decrease the promoter activity of LPIN1 gene. The decrease of LPIN1 promoter activity after site-directed mutation of PPRE1 (62%) was greater than that after site-directed mutation of PPRE2 (44%) (Fig. 8). In addition, after the simultaneous site-directed mutation of PPRE1 and PPRE2, the LPIN1 promoter activity decreased to a greater extent compared to PPRE1 or PPRE2 mutation alone (82%) (Fig. 8). It is suggested that PPARG can directly regulate the transcription of LPIN1 gene by binding to the PPRE1 and PPRE2, and the PPRE1 plays a more important role in the binding of PPARG.

Discussion
The mammary gland of lactating cows is a formidable TAG synthesis machine 7 . The synthesis of milk fat in the mammary gland includes biological processes such as uptake and transport of long-chain fatty acids, de novo synthesis of fatty acids, desaturation, synthesis of triglycerides, formation and desaturation of lipid droplets. Milk fat synthesis is regulated by PPARG, SREBF and other transcription factors 17 . The lipid droplets formed in the endoplasmic reticulum of BMECs are released into the acinar cavity of the mammary gland in the form of milk fat globules 18 . In the TAG synthesis pathway, glycerol-3-phosphate acyltransferase (GPAM) first converts glycerol-3-phosphate (G-3-P) and fatty acyl-CoA into a lysophosphatidic acid (LPA) in the endoplasmic reticulum. Then, LPA is converted into a phosphatidic acid (PA) by the AGPAT enzyme. PA is subsequently dephosphorylated by the enzyme LPIN1 to form a diacylglycerol (DAG). Then, DGAT makes DAG plus fatty acid acyl-CoA to form a TAG, which is the rate-limiting enzyme for TAG synthesis 19 . As a member of the immunoglobulin superfamily, butyrophilin A1 (BTN1A1) is essential for regulating the secretion of milk fat droplets 20 . In this study, the functional role of LPIN1 in buffalo milk fat synthesis was investigated by lentivirus-mediated interference and overexpression experiments in vitro. The overexpression of LPIN1 gene led to a significant increase in mRNA expression of GPAM, DGAT1, DGAT2 and BTN1A1 genes involved in TAG synthesis and secretion in the BMECs,   www.nature.com/scientificreports/ as well as a significant increase in TAG content (Fig. 2). On the contrary, after the knockdown of LPIN1 gene, the expression of these genes in the BMECs was decreased, and the TAG content was also significantly decreased ( Fig. 3). This suggests that buffalo LPIN1 gene plays an important role in the triglyceride pathway of buffalo milk fat synthesis, and it is speculated that this gene also participates in the synthesis of TAG in BMECs by catalyzing the dephosphorylation of PA to form DAG. Previous studies in goat mammary epithelial cells showed that knockdown of PPARG gene led to a significant decrease in LPIN1 gene, suggesting that LPIN1 gene may be a downstream target gene of PPARG during milk fat synthesis in ruminants 14 . Our previous research revealed that PPARG is a central regulator of buffalo milk fat synthesis 21 . Another study found that LPIN1, as a transcription co-activator, its C-terminal amino acid residues at 217-399 can enhance the transcriptional activation activity of PPARG 12 . In this study, the expression of PPARG gene was significantly increased after the overexpression of LPIN1 gene (Fig. 2). Moreover, the overexpression of PPARG gene also significantly increased the expression of LPIN1 gene (Fig. 6). Therefore, it was speculated that in the BMECs, the LPIN1 gene involved in milk fat synthesis may be regulated by the PPARG. Meanwhile, LPIN1 may also activate PPARG through transcriptional co-activation, and then PPARG regulates the milk fat synthesis of BMECs, which needs further confirmation in subsequent experiments. SREBF1 is considered to be an important central regulator in the regulatory network of milk fat synthesis 22,23 . Previous studies have shown that SREBF1 gene is a target gene of PPARG 24 . Based on the fact that the expression of SREBF1 gene was significantly down-regulated after LPIN1 interference in this study (Fig. 3), it is speculated that the interference of LPIN1 gene initially reduced the expression of PPARG , which further led to the downregulation of SREBF1 expression. The main function of SREBF2 is to specifically regulate cholesterol synthesis 22 . In this study, the overexpression or interference of LPIN1 gene had no significant effect on the expression of SREBF2, suggesting that LPIN1 gene may not be involved in the synthesis of cholesterol in BMECs.
Study has shown that the expression of LPIN1 gene is regulated by cholesterol, and SREBF1 can directly bind to the sterol regulatory element (SRE) on the LPIN1 promoter to promote the transcription of this gene 25 . In humans and mice, LPIN1 is regulated by a variety of transcription factors, including SREBF1 26 , estrogen-related receptor γ (ERRγ) 27 , hepatic nuclear factor 4a (HNF4a) 28 and C/EBPα 6 . They directly bind to the corresponding sites in the LPIN1 promoter to promote LPIN1 transcription and lipid synthesis. In this study, not only the binding sites of SREBF1 and C/EBPα were found in the promoter of the buffalo LPIN1, but also the PPREs sites. In the fatty acid synthesis pathway, PPARG binds to RXR to form a heterodimer, which then binds to the ligand www.nature.com/scientificreports/ and enters the nucleus, and then recognizes the PPREs on the target gene promoter, thus activating the transcription of downstream genes 13,29 . We found two PPREs (PPRE1: TGG CCT TGT GAC AG; PPRE2: GGG TCG AAA GAT CT) in the promoter of buffalo LPIN1, which were highly consistent with the recognized PPRE 13 . So far, PPRE has been found in the promoter of many genes related to lipid synthesis, such as thrombospondin receptor (CD36) 30 , adipose differentiation-related protein (ADRP) and stearoyl-CoA desaturase 1 (SCD) 15 . In this study, after the overexpression of PPARG , the luciferase activity of all recombinants with 5' deletion fragments of the LPIN1 promoter increased significantly except for fragment − 88/ + 42 (Fig. 6). In contrast, after interference with PPARG , the luciferase activity of all recombinants with 5' deletion fragments of the LPIN1 promoter was significantly decreased except for fragment − 88/ + 42 (Fig. 7). Furthermore, the site-directed mutagenesis of PPRE1  www.nature.com/scientificreports/ and PPRE2 in the core promoter of LPIN1 greatly reduced the transcriptional activity of the promoter fragment, and after the simultaneous site-directed mutagenesis of PPRE1 and PPRE2, the activity of the core promoter was significantly lower than that of PPRE1 and PPRE2 alone (Fig. 8). The results here indicate that PPARG can directly regulate the transcription of buffalo LPIN1 by binding to the PPRE1 and PPRE2 sites on its promoter.

Materials and methods
Ethics declarations. Sample collection in the described experiments was approved by the Animal Care and Use Committee of Yunnan Agricultural University (No. YNAU2019llwyh019). Every effort has been made to minimize suffering. All methods were performed in accordance with the relevant guidelines and regulations and the study is reported in accordance with ARRIVE guidelines.

Isolation, culture and identification of buffalo mammary epithelial cells. BMECs were isolated,
purified and cultured from buffalo mammary gland tissue at the peak of lactation using the previously described method 21 . The female buffalo used for sampling (Binglangjiang buffalo, an indigenous river type buffalo breed distributed in western Yunnan Province, China) was healthy, five years old and at the 3rd parity (about 60 d postpartum). In short, the fresh buffalo mammary tissue was collected and washed with the phosphate buffer saline (PBS, Gibco, USA) for three times. Part of the acini was washed by highly resistant PBS (containing 400 IU mL-1 penicillin and 400 IU mL-1 streptomycin). The tissue blocks were then cut into 1 to 2 mm 3 pieces and spread at the petri dishes, then cultured in an incubator with the condition of 37 °C, 5% CO 2 and 95% air humidity. According to the mRNA sequence of buffalo LPIN1 gene (XM_006073743), three pairs of short hairpin RNAs (shRNAs) targeting the LPIN1 gene were designed using the online software BLOCK-iT RNAi Designer (http:// rnaid esign er. invit rogen. com/ rnaie xpress/). After the shRNA sequences were annealed, it was ligated into the pLKO.1 vector to generate final recombinant plasmids (pLKO.1-shRNAs). All the recombinant vectors were sequenced for verification and then purified using EndoFree Maxi Plasmid Kit (QIAGEN, Germany). The sequences of shRNAs are presented in Table 1.
Overexpression of buffalo LPIN1 gene. When the confluence of BMECs in a 6-well plate is 70-80%, pEGFP-C1-LPIN1 was transfected into the BMECs using Lipo6000™ transfection reagent (Beyotime Biotechnology, China) in accordance with the manufacturer's protocol, and pEGFP-C1 was transfected as negative control. Forty-eight hours after transfection, the green fluorescent protein (GFP) was monitored using a Leica fluores- Table 1. Information of shRNAs used for the knockdown of LPIN1.

Knockdown of buffalo LPIN1 gene. HEK-293 T cells (purchased from Kunming Institute of Zoology,
Chinese Academy of Sciences) were cultured in DMEM/F12 medium containing 10% fetal bovine serum and 1% penicillin/streptomycin (10 kU/L, Gibco, USA). HEK-293 T cells with a confluence rate of 70-80% were cotransfected with pLKO.1-shRNA and pEGFP-C1-LPIN1 at a ratio of 3:1 to screen for the most effective shRNA targeting the LPIN1 gene. The pEGFP-C1-LPIN1 vector was transfected alone as a control. Forty-eight hours after transfection, the expression of GFP was monitored using a Leica fluorescent microscope (DMI4000B, Germany). The HEK-293 T cells were then collected for RT-qPCR analysis. The PLKO.1 system used for the RNAi was a gift from David Root Lab 32 . When HEK-293 T cells were cultured to 70-80% confluence in a 10 cm plate, Lipo6000 (Beyotime Biotechnology, China) was used to co-transfection pLKO.1-shRNA, pMD2G (enveloped plasmid) and pPAX2 (packaged plasmid) into HEK-293 T cells at a ratio of 5:3:2 to produce lentivirus (Lv-LPIN1-shRNA). In addition, pLKO.1, pMD2G and psPAX2 were co-transfected into HEK-293 T cells as a negative control group (Lv-NC). The culture medium contained the lentivirus was centrifuged by 400 g for 5 min at 4 °C by a high-speed freezing centrifuge (Sorvall ST8; Thermo Fisher Scientific, USA) and then filtered with a 0.45 μm filter. The lentivirus particle was stored at − 80 °C for long-term storage. When the BMECs were cultured in a 60 mm plate to 70-80% confluence, the lentivirus (Lv-LPIN1-shRNA or Lv-NC) was added to culture medium. Meanwhile, 2 mL of polybrene (2 μg/mL, Sigma, USA) was added to the culture medium to improve the efficiency of virus infection. The medium was replaced with fresh medium after 24 h. The BMECs were harvested after 48 h for RNA extraction or TAG content analysis. Since PLKO.1 vector does not express fluorescent protein, multiplicity of infection (MOI) was not estimated in this study, and only RT-qPCR was used to ensure sufficient interference efficiency.
TAG content assay. After the overexpression or knockdown of LPIN1 for 48 h, the BMECs were rinsed twice with PBS. The TAG content in the BMECs was determined using the TAG kit (GPO-POD; Applygen Technologies Inc., Beijing, China) following the manufacturer's instructions. Meanwhile, the intracellular total protein content was measured using the BCA protein assay kit (Thermo Fisher Scientific, USA). The TAG content was normalized by per milligram of protein.
Cloning and bioinformatics analysis of buffalo LPIN1 promoter. Genomic DNA was extracted from the blood samples of Binglangjiang buffalo (adult, healthy) using the method of phenol/chloroform purification-based protocol. A pair of primers named LPIN1_F/R were designed to conduct the PCR for isolating the promoter of the LPIN1 (Table 3). Amplified products were detected by 1.5% agarose gel electrophoresis. The target bands were excised from the agarose gel and then purified by Gel Extraction Kit (OMEGA, China). The purified PCR products were cloned into pMD18-T vector (TaKaRa, China) and further bidirectionally sequenced. Prediction of transcription factor binding sites in the promoter sequence was executed using online software JASPAR database (http:// jaspar. gener eg. net/) and Peroxisome Proliferator Response Elements Search (http:// www. class icrus. com/ PPRE/).  (Table 3). The recombinant vectors of these 5' flanking deleted promoter fragments were constructed, and all the fragments inserted into the vectors were confirmed by bidirectional sequencing. Then, the overlapping PCR method was used to generate site-directed mutagenesis with TIANSeq HIFI Amplification Mix (QIAGEN, Germany). The specific primers of overlapping PCR for site-directed mutagenesis promoters are shown in Table 3. All the PCR products were subsequently cloned into pGL4.11 vector, and all recombinant vectors were confirmed by DNA sequencing.
Luciferase-based assays. When the confluence of BMECs in 12-well plates was 70-80%, the pGL4-LPIN1 vector or all recombinant vectors with 5' deletion mutagenesis were co-transfected with pRL-TK vector in a ratio of 10:1 using Lipo6000™ transfection reagent (Beyotime Biotechnology). The pGL4.11 vector was co-transfected with pRL-TK as negative control. After 48 h of culture, the BMECs were collected for the determination of luciferase activity. The lentivirus (Lv-PPARG-sh3) used for PPARG gene knockdown and pEGFP-C1-PPARG used for overexpression of PPARG gene were constructed by our laboratory in the previous research 21 . In order to investigate the effect of PPARG overexpression and knockdown on the activity of the LPIN1 promoter, after the BMECs were transfected with Lv-PPARG-sh3 or pEGFP-C1-PPARG for 12 h, the pGL4-LPIN1 plasmid and 5' deletion recombinant constructs of LPIN1 promoter were transiently co-transfected the BMECs from the previous step with pRL-TK, and the BMECs were collected for RT-qPCR analysis and luciferase activity analysis after 48 h of culture. The BMECs were washed twice with cold PBS and luciferase activity was measured by Dual-Luciferase ® Reporter Assay System (Promega, Wisconsin, USA). The relative luciferase activity of each sample was normalized with pRL-TK.
Statistical analysis. Three repetitions were set for all treatments, and the results were expressed as mean ± standard error of means (means ± SEM). Statistical comparisons (t-test) between the treatment and the control group were performed using software SPSS 19.0 (SPSS Inc., Chicago, IL) with a significance level of 0.05 and an extremely significant level of 0.01.

Conclusion
In this study, in vitro overexpression and interference experiments showed that the interference with LPIN1 significantly reduced the expression of lipid synthesis related genes and TAG content in the BMECs, while the overexpression of LPIN1 significantly increased the expression of lipid synthesis related genes and TAG content in the BMECs. Further transcriptional regulation experiments showed that the overexpression or interference of PPARG could enhance or decrease the expression and promoter activity of buffalo LPIN1, and the site-directed mutagenesis of PPRE1 and PPRE2 significantly reduced the promoter activity of buffalo LPIN1. The results here suggest that buffalo LPIN1 is involved in the synthesis of milk fat, and the PPARG can directly regulate the transcription of buffalo LPIN1 by binding to the PPRE1 and PPRE2 in its core promoter.