Non-alcoholic steatohepatitis is a disease of emerging identity and importance and is now considered as one of the commonest liver diseases worldwide. It is frequently associated with severe obesity and is intimately related to various clinical and biological markers of hepatic lipid dysregulation(Reference Ooi, Meikle and Huynh1). Choline, as an essential dietary nutrient, plays a vital role in maintaining hepatic lipid homoeostasis(Reference Luo, Wei and Ye2); thus, its deficiency has been closely linked to non-alcoholic steatohepatitis(Reference Tahan, Sonsuz and Imeryuz3,Reference Corbin and Zeisel4) . Additionally, choline, functioning as a lipotropic agent, has been commonly used in animal feeds(Reference Liu, Pirozzi and Codabaccus5). Additionally, choline’s role in lipid metabolism is of particular interest for aquaculture finfish species given its recognition as a lipotropic agent(Reference Liu, Pirozzi and Codabaccus5). One of two major fates for choline are to be irreversibly oxidised by choline dehydrogenase (Chdh) and used as a donor of methyl groups(Reference Mahmoud and Ali6). Thus, emerging data support a role for dietary choline in modulating DNA methylation(Reference Park, Friso and Choi7,Reference Zeisel8) . Furthermore, the close association between DNA methylation and lipid metabolic disorder has been investigated(Reference Yoon, Tammen and Park9–Reference Jones, Irvin and Claas11). However, less research has been done on the contribution and mechanism of methyl donor function of choline in regulating hepatic lipid homoeostasis, although the crucial role of dietary choline in lipid metabolism and DNA methylation has been widely studied.
Endoplasmic reticulum (ER)-localised enzymes synthesise the vast majority of cellular lipids(Reference Jacquemyn, Cascalho and Goodchild12). Hepatic lipid homoeostasis therefore depends on the ER function(Reference Wilfling, Wang and Haas13). Naturally, as a conserved adaptive mechanism in response to ER dysfunction, ER unfolded protein response (UPRer) has been closely connected with hepatic lipid dysregulation(Reference Fu, Yang and Li14). Importantly, UPRer has an upstream and comprehensive role for regulating hepatic lipid metabolism. UPR is initially activated and transduced by three UPR branches: the perk-eif2α pathway, the ire1α-xbp1 pathway and the atf6-chop pathway, and then induces translational inhibition followed by up-regulation of ER-resident chaperones, such as grp78/bip, grp94 and crt (Reference Lebeaupin, Vallée and Hazari15). Given the potential role of methylation in regulating UPRer has been investigated(Reference Han, Hu and Lau16,Reference Ramos-Lopez, Riezu-Boj and Milagro17) , UPRer should be a good and novel viewpoint from which to understand the underlying mechanism governing choline-mediated DNA methylation regulating hepatic lipid metabolism; however, the relevant regulatory mechanism still remains to be explored. On the other hand, some studies, including our previous study, have indicated the differential effects and mechanisms of three UPRer branches on regulating cell function and metabolism(Reference Song, Hogstrand and Wei18,Reference Nasiri-Ansari, Nikolopoulou and Papoutsi19) . Accordingly, it is important to investigate the differential mechanism of three UPRer branches on choline regulating hepatic lipid metabolism.
Yellow catfish, Pelteobagrus fulvidraco, is an economically important freshwater teleost fish widely distributed in China and other countries(Reference Gong, Dan and Xiao20). Its genome has a high degree of similarity to that of humans(Reference Gong, Dan and Xiao20,Reference Sato, Hashiguchi and Nishida21) and is also a good model animal for studying lipid metabolism(Reference Wei, Luo and Hogstrand22). Our previous studies found UPRer was the major and most common inducements to hepatic lipid dysregulation in yellow catfish(Reference Song, Hogstrand and Wei18,Reference Song, Luo and Zhang23) . In addition, although choline’s role in lipid metabolism is of particular interest for aquaculture finfish species given its recognition as a lipotropic agent(Reference Liu, Pirozzi and Codabaccus5), the mechanism of choline acting as methyl donor alleviating hepatic lipid dysregulation is still not fully understood. Thus, our present study reveals dietary choline addition down-regulated ire1α expression by activating specific CpG methylation sites and also found choline prevented hepatic lipid dysregulation via controlling Ire-1α/fatty acid synthase (Fas) interaction. These results emphasise the critical contribution of methyl donor function for choline acting as a lipotropic agent.
Materials and methods
Experimental treatments
Huazhong Agricultural University’s (HZAU) institutional ethical guidelines for the care and use of laboratory animals were followed throughout all investigations and were approved by the Ethical Committee of HZAU (identification code: Fish-2020-07-21).
Expt. 1: Animals feeding, management and sampling in vivo study
Dietary formula and yellow catfish feeding were determined according to our previous studies(Reference Luo, Wei and Ye2). We formulated two experimental diets, which are shown in online Supplementary Table S1. Fish oil and soyabean oil (1:1, w/w) were used as the lipid sources. The addition of choline chloride was 1658·4 (adequate choline) and 264·7 (choline deficiency) mg of choline per kg diet (≥ 99·0 % in purity, Sinopharm Chemical Reagent Co. Ltd.). One hundred and fifty uniform size juvenile yellow catfish, obtained from a local fish pond (Wuhan, China, mean initial weight: 3·83 (sem 0·01) g, mixed sex), were randomly stocked in six tanks (300-l water volume), with triplicates for two treatment, twenty-five fish/tank. All fish were fed to the satiation twice daily at 08.00 and 16.30 hours. The experiment continued for 10 weeks. During the experiment, the parameters in water quality were as follows: water temperature 28 (sem 0·6)°C; dissolved oxygen 6·53 (sem 0·11) mg/l; pH 7·46 (sem 0·17) and NH4-N 0·09 (sem 0·02) mg/l.
At the end of the experiment, only before sampling, yellow catfish were fasted for 24 h to avoid the prandial effects. Yellow catfish were euthanised with MS-222 (tricaine methanesulfonate, 100 mg/l water). All yellow catfish were counted and weighed in bulk. The liver tissues from three fish of each tank were sampled for histological and ultrastructural observation, respectively. The liver samples from other fish were frozen immediately in liquid N2 for other analysis, including the contents of TAG, methylation analysis, gene and protein expression.
Expt. 2: Cell culture and treatments in vitro study
Primary hepatocyte culture and treatment: yellow catfish (mean weight: 6·83 (sem 0·24) g) were obtained from the in vivo experiment after 2 weeks of acclimation. Primary hepatocytes were isolated and cultured as previously described(Reference Song, Hogstrand and Ling24). Briefly, hepatocytes were seeded in cell culture plates at a density of 1·4 × 106 cells/ml, and primary hepatocytes were cultured in RPMI 1640 media (Thermo Fisher Scientific) (supplemented with 10 % FBS, 1 mmol/l glutamine, 100 μg/ml streptomycin and 100 μg/ml penicillin) and adhered 4 h before treatment. To investigate the mechanisms of choline influencing hepatic lipid metabolism, we used choline to incubate yellow catfish primary hepatocytes. The total concentration of choline chloride was 0·6 mg/l for the adequate choline group and no extra addition for choline deficiency group based on our previous study and the cell viability in our pilot experiment. The cells were incubated at 28°C for 48 h. Each treatment was performed in triplicate. For each cell culture, a pool of cells from three fish was used.
If choline serves as a donor of methyl groups, it needed to be irreversibly oxidised by choline dehydrogenase (CHDH)(Reference Corbin and Zeisel4). Thus, in order to intercept the methyl donor function of choline, we designed and transfected siRNA against chdh to hepatocytes of yellow catfish. On the other hand, human embryonic kidney cells (HEK293T cells) have high transfection efficiency and have been widely used to explore genetic function in fish(Reference Zhao, Yang and Chen25). In the present study, HEK293T cells were used to explore the effect of special methylation sites on ire1α promoter activity.
Sample analysis
Oil red O, haematoxylin and eosin, Bodipy 493/503 staining and transmission electron microscopy observation
Oil red O and haematoxylin and eosin staining tests were conducted according to the manufacture’s instruction and followed the description in our publication(Reference Wei, Luo and Hogstrand22,Reference Song, Hogstrand and Ling24) . The ten fields of each sample were quantified by the software Image J (NIH) to get the statistics of the relative areas of lipid droplets (LD) in the Oil red O staining, the relative areas of vacuoles in the haematoxylin and eosin staining.
Bodipy 493/503 staining for the LD and transmission electron microscopy (TEM) observation of hepatocytes were carried out according to the protocol, which was described in our previous publications(Reference Wei, Luo and Hogstrand22,Reference Song, Hogstrand and Ling24) . For Bodipy 493/503 staining, briefly, hepatocytes were cultured in twelve well plates, treated with the corresponding treatments for 48 h, washed twice with PBS and incubated with 5 μg/ml Bodipy 493/503 (D3922; Thermos Fisher Scientific) for 30 min, next to thrice PBS washes. The hepatocytes were observed with a laser scanning confocal microscope (Leica DMI8) to visualise the intensity of fluorescence. The green dots were defined as lipophagic vacuoles, which were quantified with a CytoFlex flow cytometer (Beckman Coulter). The data analysis was conducted using FlowJo v.10 software. For transmission electron microscopy observation of hepatocytes, in brief, samples were fixed in 2·5% glutaraldehyde, followed by post-fixation in osmium tetroxide. The ultrathin sections were dehydrated in ethanol, embedded in resin, stained with uranyl acetate followed by lead citrate and then prepared for EM.
Cell viability and determination of lipid and TAG contents, and Fas activity
Cell viability was measured with the use of 3-(4,5-dimethylthia-zol-2-yl)-2, 5-diphenyltetrazolium bromide (V13154; Thermo Fisher Scientific) according to our previous publication(Reference Wei, Luo and Hogstrand22,Reference Song, Hogstrand and Ling24) . Hepatic lipid content was determined by the ether extraction according to our previous publication(Reference Luo, Wei and Ye2). The contents of TAG were determined by commercial kits (A110-1-1, Nanjing Jiancheng Bioengineering Institute), according to the manufacturer’s instructions. Soluble protein content was analysed, based on protocols by Bradford(Reference Wei, Luo and Hogstrand22,Reference Song, Hogstrand and Ling24) .
For Fas activity assays, the liver and cells samples were homogenised in three volumes of ice-cold buffer (0·02 M Tris–HCl, 0·25 M sucrose, 2 M MEDTA, 0·1 M sodium fluoride, 0·5 mM phenylmethylsulphonyl fluoride, 0·01 M β-mercapto-ethanol, pH 7·4) and centrifuged at 20 000 × g at 4°C for 30 min. The assay was carried out at 25°C on several dilutions of the supernatant fluid, in the final volume of 1 ml, and started by the addition of malonyl-CoA. The enzyme activity was calculated after subtraction of the non-specific oxidation of NADPH in the absence of malonyl-CoA. 2 mol NADPH was taken to correspond to 1 mol malonyl-CoA utilised. The reaction was started by addition of the tissue extract. The changes in absorbance at 340 nm were monitored at intervals of 15 s for 3 min.
RNA isolation and real-time quantitative PCR analysis
Total RNA was isolated by using the Trizol reagent and then transcribed into the cDNA by using the Reverse Transcription Kit. Analyses on gene transcript levels were conducted through the real-time quantitative PCR method described before(Reference Wei, Luo and Hogstrand22,Reference Song, Hogstrand and Ling24) . The primer sequences used in this analysis are given in online Supplementary Table S3. A set of nine housekeeping genes (gapdh, b2m, ef1α, 18s rRNA, tuba, β-actin, hprt1, ubc9, tbp) were selected from our transcriptome database in order to test their transcription stability. Our pilot experiment indicated that gapdh and β-actin (M = 0·27) showed the most stable levels of expression across the experimental conditions, as suggested by geNorm(Reference Vandesompele, Preter and Pattyn26). Thus, the relative expression levels were normalised to the geometric mean of the combination of gapdh and β-actin and calculated using the 2−ΔΔCt method.
Immunoprecipitation and western blot
Immunoprecipitation was performed to identify Fas–Ire1α interaction based on the protocols by Su et al (Reference Su, Yang and Wang27).
Briefly, in order to conduct the immunoprecipitation analysis, we lysed the cells in NP-40 buffer (Beyotime, p0013F) with the addition of a protease inhibitor cocktail (Beyotime, P1010). The cell lysate was then mixed with anti-Ire1α (ab37073; Abcam), as the bait protein, at 4°C overnight. The incubation was sustained for 3 h, followed by the addition of protein A/G agarose (P2012; Beyotime). The immunocomplexes were washed five times by PBS supplemented with Phenylmethanesulfonyl fluoride (PMSF) protease inhibitor. Finally, the western blot analysis was performed with anti-Fas (ab133619; Abcam) or anti-Ire1α (ab37073; Abcam), as the prey protein. In addition, for the mutation analysis of Ire1α, the open reading frames of Fas and Ire1α sequences/or lacks interaction sequence (Ire1α Δ836–963) were subcloned into the pcDNA3·1 (+) vector with the Flag-tag and HA-tag sequences, respectively. Then, the cell lysate was then mixed with anti-HA-tag (ab236632; Abcam), as the bait protein, at 4°C overnight. The incubation was sustained for 3 h, followed by the addition of protein A/G agarose. Lastly, western blot analysis was performed with anti-HA-tag (ab236632; Abcam) or anti-Flag-tag (ab1162; Abcam), as the prey protein.
To identify the protein levels of Fas, Ire1α, Chdh, Dnmt1, Grp78 and Gapdh, western blot analysis was performed according to our previous study(Reference Wei, Luo and Hogstrand22). In brief, the protein was loaded onto the SDS-PAGE gel and then transferred to the Polyvinylidene fluoride (PVDF) membrane. Membranes were blocked with 5 % skimmed milk and then incubated overnight at 4°C with one of the following primary antibodies: anti-Fas, anti-Ire1α, anti-Gapdh (ab198233 1:10 000, 12 118, Cell Signaling Technology), anti-Chdh (1:1000, A16545, ABclonal), anti-Dnmt1 (1:1000, A5495, ABclonal) and anti-Grp78 (1:1000, A0241, ABclonal), respectively. The secondary antibodies, including an HRP-conjugated anti-rabbit IgG antibody (1:10 000, 7074, Cell Signaling Technology), were then incubated with the membranes. The membranes were seen by ECL (1 705 060, Bio-Rad) after additional washing. The membranes were seen using enhanced chemiluminescence, and Image J was used to measure the densitometry of these bands.
Bisulphite sequencing of ire-1α
To determine the methylation level of ire-1α, ire-1α gene was sequenced by bisulphite based on our published protocol(Reference Xu, Xu and Hogstrand28). Briefly, tissue DNA kit (Omega Biotek) was used to isolate genomic DNA from the intestine samples of yellow catfish and then the DNA Methylation-Gold Kit was used to modify them, based on the manufacturer’s instructions. Then, PCR was used to amplify the bisulphite-modified DNA with specific primer pairs. The operation of PCR included 95°C denaturation for 3 min, then 35 cycles of 95°C for 30 s, 53°C for 30 s and 72°C for 40 s and end of 72°C extension for 5 min. The PCR products were purified and cloned into pMD19-T Vectors (TaKaRa). After cloning, nine clones were chosen from each sample and used for DNA sequencing. Sequencing analysis was performed by TsingKe Biological Technology.
RNAi and gene transfection
To generate a chdh knockdown cell, hepatocytes of yellow catfish were transfected with 103 nM of siRNA against chdh from Sigma-Aldrich based on our previous protocols(Reference Su, Yang and Wang27). Transfection was performed with Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific). Target sequences for preparing the siRNA of yellow catfish chdh are shown in online Supplementary Table S3. The transfection of siRNA was performed using the Lipo2000 transfection reagent according to the supplier’s protocol with transfection efficiencies of 42·9 (sem 3·4) % with low levels of cell death (4·2 (sem 1·0) %). Hepatocytes were transfected with indicated siRNA at 30 % confluence. The siRNA was added to one tube and mixed gently; 5 μl of Lipo2000 transfection reagent was added to the other tube and gently mixed. Then, the culture medium containing the siRNA was gently added to the culture solution containing the Lipo2000 transfection reagent, and the tube had a gently inverted mixture. A mixture of the Lipo2000 transfection reagent and siRNA was added to each well, and the culture medium was changed after 5 h.
Plasmid and cell transfection
To identify the role of Ire1α/Fas interaction on Fas activity, we constructed the Ire1α and Fas expression vectors. The open reading frames of Ire1α and Fas sequences were subcloned into the pcDNA3·1 (+) vector with the HA-tag and Flag-tag sequences inserted at the N-terminus of Ire1α and Fas sequences, respectively. Mutations of amino acid from 836 to 963, which are the interaction sequence for Ire1α, were produced in the HA-Ire1α plasmid by the Mut Express II Fast Mutagenesis Kit (Vazyme). The transient transfection of the plasmids into hepatocytes from yellow catfish was conducted using Lipofectamine 2000 (Invitrogen) based on the manufacturer’s instructions.
Plasmid construction, transfections and luciferase assays
We constructed perk promoters into the pGl3 basic vector by using the ClonExpress™ II One Step Cloning Kit (C112, Vazyme) based on our published protocol(Reference Song, Xu and Zhuo29). CpG islands of perk were predicted by the MethPrimer (http://www.urogene.org/methprimer/). Mutations of CG sites in ire-1α promoter regions were performed according to the manufacturer’ instructions of Mut Express II Fast Mutagenesis Kit (C214-01, Vazyme) and verified by sequencing. All these plasmids were transiently transfected into the HEK293T cells using the Lipofectamine 2000 (12566014, Invitrogen) at about 80 % confluence to measure the luciferase activity. Primers used for site-mutation analysis are presented in online Supplementary Table S3.
Protein docking for Fas and Ire-1α
Swiss-Model server (https://swissmodel.expasy.org/) was used to construct the protein structure of yellow catfish, from which the crystal structures were downloaded from the PDB protein data (Fas PDB template ID: 2vz9, Ire-1α PDB template ID: 6urc) and the appropriate templates were selected. The ZDOCK module of Discovery Studio 4.0 was used to perform Fas-Ire1α docking. Docking pose was picked up by using pymol from the given top models. The information for structural protein prediction model and Fas/Ire-1α protein interaction is presented in online Supplementary Figs S2, S3 and S4.
Statistical analysis
All data were expressed as mean values with their standard error of means. The normality of data distribution and the homogeneity of variances were analysed using the Kolmogorov–Smirnov test and Bartlett’s test, respectively. Differences between adequate choline and choline deficiency groups or between si-NC and si-chdh groups were analysed using a Student’s t test for independent samples using SPSS 19.0 software, and the minimum significance level was set at P < 0·05.
Results
Effects of dietary choline on growth performance and hepatic lipid metabolism
Weight Gain (WG) in the group fed the choline-deficient diet were higher than in the adequate choline diet group. FCR, HSI, CF and survival showed no significant differences between two treatments (online Supplementary Table S2). In the present study, when compared with adequate choline diet group, choline-deficient diet caused hepatic lipid dysregulation supported by increasing the vacuoles in haematoxylin and eosin and LD in Oil Red O (Fig. 1(a)–(c)). Meanwhile, at the transcription level, choline-deficient diet significantly up-regulated the expression of genes involved in lipogenesis (srebp-1, pparγ, 6pgd, accα and fas) but down-regulated lipolysis and FA β-oxidation (hsl, cpt-1α, echs1) (Fig. 1(d)). Further study indicated, a key enzyme in lipogenesis, Fas was significantly up-regulated by choline-deficient diet at both protein expression and enzyme activity levels (Fig. 1(e)–(g)). All these results suggested choline-deficient diet caused hepatic lipid dysregulation in the livers of yellow catfish, which was further proved by the significant increased trend of TAG and lipid content (Fig. 1(h) and (i)).
Choline-deficient diet caused lipid dysregulation in the liver of yellow catfish. (a) Representative images of hepatic H&E and Oil red O stained. Scale bar, 30 μm; hepatocytes (he); vacuoles (va); lipid droplets (ld). (b)–(c) Relative areas for hepatic vacuoles in H&E staining and LD in Oil Red O staining. (d) mRNA levels of the genes related to hepatic lipid metabolism. (e)–(f) Western blot analysis and quantification analysis for Fas. (g) Fas activity. (h) TAG content. (i) Hepatic lipid contents. Data are mean values with their standard error of the means (n 3 replicate tanks). * indicates significant differences between adequate choline diet and choline-deficient diet group. H&E, haematoxylin and eosin; LD, lipid droplet.
Choline-deficient diet-induced hepatic unfolded protein response main via ire-1α signalling
We examined whether choline-deficient diet could cause UPRer in the liver of yellow catfish. As shown in Fig. 2(a), compared with adequate choline diet, transmission electron microscopy observation found severe swelling of ER in choline-deficient diet groups. Also, choline-deficient diet apparently up-regulated mRNA abundance of UPRer markers, including grp78, grp94, crt, perk, eif2α, ire-1α, xbp-1, aft-6α and chop. Furthermore, among these three UPRer signalling pathways, ire-1α signalling shown the most sensitive to choline-deficient diet (Fig. 2(b)). Meanwhile, this ire-1α signalling-mediated UPRer was further confirmed by the protein level of Grp78 and Ire-1α (Fig. 2(c)–(e)). All these indicated that choline-deficient diet caused severe hepatic UPRer main via ire-1α signalling in yellow catfish.
Choline-deficient diet induced hepatic UPRer main via ire-1α signalling. (a) TEM structures of the liver, ER and LD; scale bars, 2 μm. (b) mRNA levels of the genes related to UPRer. (c)–(e) Western blot analysis and quantification analysis for Grp78 and Ire-1α. Data are mean values with their standard error of the means (n 3 replicate tanks). * indicates significant differences between adequate choline diet and choline-deficient diet group. UPRer, unfolded protein response; TEM, transmission electron microscopy; ER, endoplasmic reticulum; LD, lipid droplet.
Dietary choline increased CpG methylation in the promoter of ire-1α
Choline can be irreversibly oxidised to yield betaine in a two-step process catalysed by CHDH and betaine aldehyde dehydrogenase mainly in the liver [4]. Betaine is an important methyl group donor. Thus, next we examined the dietary choline-induced alteration in DNA methylation reaction in the liver of yellow catfish. First, choline-deficient diet significantly reduced the concentration of choline and betaine and also the ratio of S-adenosylmethionine (SAM)/S-adenosylhomocysteine (SAH) (Fig. 3(a) and (b)). Second, the down-regulation of genes and/or protein involved in methionine cycle and methyltransferase (Fig. 3(c) and (d)). All these results suggested choline-deficient diet caused the inhibition of DNA methylation reaction. Since the main role of ire-1α signalling in choline-deficiency-induced up-regulation of UPRer markers was determined, and also given three CpG islands in the promoter of ire-1α from yellow catfish were predicted (Fig. 3(e)), we studied the effect of dietary choline on CpG methylation in the promoter of ire-1α. Compared with choline-deficient diet, adequate choline diet significantly up-regulated the levels of ire-1α promoter methylation (Fig. 3(f)). Importantly, we identified the significant choline-induced methylation sites, including the −994, −1229, −1287 and −1497 sites in the promoter of ire-1α (Fig. 3(g)), implying the potential role of these methylation sites in regulating CpG methylation of ire-1α.
Dietary choline increased CpG methylation in the promoter of ire-1α. (a) The concentration of choline, betaine, SAM and SAH. (b) The ratio of SAM/SAH. (c) mRNA levels of the genes related to methionine cycle and methyltransferase. (d) Western blot analysis for Chdh and Dnmt1. (e) Prediction analysis of CpG islands in the sequence range of 2000 bp upstream from the transcriptional start site in the ire-1α promoter region. (f) The methylation level of ire-1α promoter region. (g) The methylation level of specific sites in ire-1α promoter region. Data are mean values with their standard error of the means (n 3 replicate tanks). * indicates significant differences between adequate choline diet and choline-deficient diet group. SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine.
Choline down-regulated ire-1α expression by controlling site-specific DNA methylations
We further examined the mechanism of choline-mediated CpG methylation regulating ire-1α expression. By choline dehydrogenase (chdh), choline could be irreversibly oxidised and then serves as a donor of methyl groups. Thus, si-chdh could intercept the methyl donor function of choline. First, for adequate choline groups, si-chdh apparently up-regulated the concentration of choline but reduced the level of betaine, SAM and SAH; meanwhile, the ratio of SAM/SAH has also been decreased (Fig. 4(a) and (b)), suggesting the interception of choline oxidisation metabolism. In addition, in adequate choline groups, the si-chdh markedly down-regulated the mRNA abundance of genes involved in methionine cycle and methyltransferase (Fig. 4(c)). All these results suggested that si-chdh caused inhibition of choline-mediated methionine cycle and its follow-up DNA methylation reaction, which was further confirmed by the lower protein levels of Chdh and Dnmt1 (Fig. 4(d)).
Choline down-regulated ire-1α expression by controlling site-specific DNA methylations. (a) The concentration of choline, betaine, SAM and SAH. (b) The ratio of SAM/SAH. (c) mRNA levels of the genes related to methionine cycle and methyltransferase. (d) Western blot analysis for Chdh and Dnmt1. (e) mRNA levels of ire-1α. (f) Western blot analysis for Ire-1α. (g) The methylation level of ire-1α promoter region. (h) The methylation level of specific sites in ire-1α promoter region. (i) Site mutation analysis of −994, −1229 and −1497 methylation site on pGl3-perk −1680/+116 vectors. Data are mean values with their standard error of the means (n 3 independent biological experiments). * indicates significant differences among same si-NC or si-chdh groups. SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine.
Then si-chdh significantly up-regulated the expression of ire-1α at both protein and mRNA levels (Fig. 4(e) and (f)), suggesting the potential regulatory role of DNA methylation on ire-1α expression. Thus, next we focus on the relationship between choline-mediated DNA methylation and ire-1α expression. As shown in Fig. 4(g), the levels of DNA methylation in the promoter of ire-1α were significantly down-regulated by si-chdh. In addition, using the site-specific CpG methylations analysis, we identified remarkable methylation sites, including the −994, −1229 and −1497 sites in the promoter of ire-1α (Fig. 4(h)). Further, site mutation analysis shown only double mutation of −994 and −1229 methylation sites or triple mutation of −994, −1229 and −1497 methylation sites significantly up-regulated the promoter methylation activity of ire-1α, suggesting the collaborative relationships of these specific methylation sites for regulating ire-1α expression. All these results indicated the key regulatory role of CpG methylation on ire-1α expression by controlling specific CpG methylation sites.
Choline prevents hepatic lipid dysregulation via reducing Ire1α–Fas interaction
The contribution and mechanism of choline acting as methyl donor in preventing hepatic lipid dysregulation were tested. First, transmission electron microscopy observation found si-chdh significantly aggravated swelling of ER in hepatocytes (Fig. 5(a)), suggesting the potential regulatory role of methyl donor function of choline mitigating hepatic UPRer. Further analysis with respect to protein–protein interaction prediction and immunoprecipitation of Fas and Ire-1α clearly demonstrated an interaction between these two proteins. This interaction was significantly down-regulated in the presence of adequate choline (Fig. 5(b) and (d)). In addition, si-chdh significantly inhibited the down-regulation for Ire1α–Fas interaction induced by choline addition (Fig. 5(c) and (d)). To determine whether Ire1α–Fas interaction is associated with Fas activity, the hepatocytes were co-transfected with Ire-1α that lacks the Ire1α–Fas interaction sequence (Ire1α Δ836–963) (Fig. 5(f) and (g)). The loss of Ire1α–Fas interaction sequence in ire-1α impaired Fas activity indicating a potential role of Ire-1α in regulating Fas activity. On the other hand, the content of LD in hepatocytes was significantly improved by si-chdh (Fig. 5(i)). These results indicated the contribution of choline acting as methyl donor in preventing hepatic lipid dysregulation, which was also further confirmed by the content of TAG in hepatocytes of yellow catfish (Fig. 5(j)).
Choline prevents hepatic lipid dysregulation via reducing Ire1α–Fas interaction. (a) TEM structures of the hepatocytes, ER, LD, SER and UN; scale bars, 2 μm. (b) The structural protein prediction model between Fas and Ire-1α. (c)–(d) IP analysis and quantification analysis for Fas-Ire1α complex. Flag-tag Fas and HA-tag Ire1 α were transfected into hepatocytes from yellow catfish. The interaction between Fas and Ire1 α was determined with IP and western blot. (e) Fas activity. (f)–(g) IP analysis and quantification analysis for Fas-Ire1α complex in hepatocytes transducing with the full-length Ire1α or Ire1α that lacks the Fas interaction sequence (Ire1α Δ836–963). (h) Fas activity in hepatocytes transducing with the full-length Ire1α or Ire1α that lacks the Fas interaction sequence (Ire1α Δ836–963). (i) Representative confocal microscopic image of lipid droplets (green) hepatocytes of yellow catfish (bar = 15 μm). (j) TAG content. Data are mean values with their standard error of the means (n 3 independent biological experiments). * indicates significant differences among same si-NC or si-chdh groups and also among same full-length Ire1α or Ire1α Δ836–963 groups. TEM, transmission electron microscopy; ER, endoplasmic reticulum; SER, swelling endoplasmic reticulum; UN, nucleus; IP, immunoprecipitation.
Discussion
Choline’s role in hepatic lipid metabolism has been broadly investigated. Given choline’s recognition as a lipotropic agent and an important component of several phospholipids, its role in lipid metabolism is of particular interest for aquaculture finfish species(Reference Liu, Pirozzi and Codabaccus5). Although choline is primarily known for its function as a donor of methyl groups(Reference Corbin and Zeisel4), the exact mechanisms by which it contributes to preventing excessive hepatic lipid deposition are not well understood. Furthermore, since UPRer serves as the key upstream pathway for lipid metabolism, it may have a regulatory role in choline-mediated hepatic lipid metabolism. However, the specific mechanism underlying this relationship has yet to be elucidated.
The present study indicated that choline-deficient diet caused lipid dysregulation in the liver of yellow catfish as expected, which has been confirmed in our previous studies(Reference Song, Luo and Zhang23). The dietary choline deficiency markedly up-regulated the mRNA expression of genes involved in de novo lipogenesis, especially fas. This suggested the close association between lipogenesis disorders and choline-deficient model of hepatic lipid dysregulation, which also been proved by other study(Reference Sun, ShangGuan and Zhu30). On the other hand, given ER is the main organelle for de novo lipogenesis and the crucial role of UPRer on hepatic lipid metabolism, we next evaluated the choline diet-induced alteration of UPRer. Choline plays a crucial role in the formation of several phospholipids and is essential for preserving the stability of the ER membrane structure. The choline deficiency has been linked to the activation of the UPRer in the ER(Reference Yao, Jamil and Vance31). Not surprisingly, our present study clearly indicated that choline-deficient diet generated severe swelling ER, indicating the inducement of UPRer in the liver of yellow catfish, in agreement with other studies(Reference Muraki, Makita and Yamasaki32,Reference Minami, Miura and Ishioka33) . Interestingly, mRNA expression analysis indicated that the choline-deficiency-induced UPRer is mainly originated from ire-1α signalling pathway compared with perk and atf-6α signalling pathways. This suggests a major role of ire-1α in choline-deficiency-induced UPRer. Similarly, the differential effects and mechanisms of three UPRer branches on regulating cell function and metabolism have been confirmed by other studies, including our previous study(Reference Song, Hogstrand and Wei18). On the other hand, our in vivo studies indicated that WG was higher in choline-deficient diet. This may be attributed to lipid accumulation in whole body in response to dietary deficiency of choline.
DNA methylation, an important epigenetic modification, has been closely associated with hepatic lipogenesis and fatty liver(Reference Zeisel8,Reference Tan, Liu and Zhang34) . In addition, diet intake of choline can modulate methylation because, via choline dehydrogenase (Chdh), this nutrient (and its metabolite, betaine) regulates the concentrations of SAM and SAH(Reference Zeisel8). The present study found dietary choline addition increased the concentrations of choline and betaine and the ratio of SAM/SAH. Meanwhile, the process of DNA methylation is catalysed by a group of enzymes called DNA methyltransferases (Dnmts) mainly including Dnmt1, Dnmt3a and Dnmt3b(Reference Mahmoud and Ali6,Reference Zeisel8) . Our current study found that dietary choline addition up-regulated the expression of genes and/or protein involved in methionine cycle and Dnmts. All these findings indicated that the dietary choline could activate DNA methylation in the liver of yellow catfish.
Given the major role of ire-1α for choline-deficient diet-caused UPRer and multiple CpG islands in ire-1α promoter region has been confirmed, we next explored the mechanism of choline-activated DNA methylation regulating ire-1α expression and their roles on UPRer and lipid metabolism. First, in vivo, dietary choline up-regulated the methylation level of ire-1α promoter region by activating specific CpG methylation sites. Then, in vitro, by using si-chdh transfection and mutation analysis of specific CpG methylation sites, we found ire-1α promoter methylation was inhibited by the interception of the methyl donor function of choline. Further, choline activated ire-1α promoter methylation, by controlling specific CpG methylation sites the in ire-1α promoter region. The activation of promoter methylation was closely correlated with down-regulation of gene expression(Reference Matsuzaki, Miyajima and Fukamizu35). Thus, we concluded that the decrease in ire-1α expression is attributed to choline-mediated CpG methylation. Importantly, in vitro experiments demonstrated the involvement of choline-mediated methylation of ire-1α promoter down-regulating UPRer pathway. Thus, our present study clearly showed that choline serving as methyl donor activated specific CpG methylation sites of ire-1α, which caused the down-regulation of ire-1α expression. Furthermore, in vitro experiments demonstrated that si-chdh counteracted the beneficial effect of choline on ER membrane structure. Collectively, the interception of the methyl donor function of choline could down-regulate CpG methylation of ire-1α promoter and then induce UPRer in liver. Similarly, studies have pointed out that alcohol-induced UPRer is manifested through altered DNA methylation, which subsequently regulates the expression of UPRer-associated factors(Reference Han, Hu and Lau36). Additionally, in rodents, dietary restriction of the methyl donors (methionine and/or choline) rapidly and reliably induces a spectrum of liver injury histologically(Reference Cordero, Gomez-Uriz and Campion37).
On the other hand, we further found the interception of methyl donor function of choline caused the loss of choline-mediated prevention of hepatic LD deposition. This suggested choline serving as a lipotropic agent may depend on the methyl donor function, preventing the hepatic lipid dysregulation. Analogously, choline improved lipid homoeostasis in obesity through oxidative demethylation in murine model(Reference Sivanesan, Taylor and Zhang38). More interestingly, immunoprecipitation analysis showed that there is an interaction between Ire-1α and Fas, and this interaction was obviously down-regulated by choline, suggesting the involvement of Ire-1α/Fas interaction in choline mediating hepatic lipid metabolism. This also provided the first experimental evidence for direct crosstalk between Ire-1α signalling-mediated UPRer and hepatic lipid dysregulation. Importantly, the choline-induced down-regulation of Ire-1α/Fas interaction was lost along with si-chdh treatment, which further confirmed the contribution of choline acting as methyl donor. Our findings further revealed that, when Ire1α–Fas interaction was impaired, Fas activity was dwindled indicating novel regulatory role of Ire-1α protein in regulating Fas activity in choline-mediated lipid homoeostasis in hepatocytes. In a nut shell, by providing methyl group, choline activated CpG methylation of ire-1α promoter and then down-regulated ire-1α expression, which sequentially down-regulated the interaction with Fas and finally relieved hepatic lipid accumulation. On the other hand, along with the down-regulation of ire-1α, choline also improved ER membrane structure and alleviated hepatic UPRer, which also provided favourable assistance for hepatic lipid metabolism. Similarly, in fish, a potential causal relationship between UPRer-associated methylation and excessive lipid deposition was suggested in our previous study(Reference Xu, Xu and Hogstrand28). However, further investigation is necessary to elucidate the mechanisms underlying the improvement of the ER membrane integrity in response to adequate dietary choline levels, particularly considering the dual role of choline as both a methyl donor and a component of phospholipids.
It needs to be mentioned that the current study still has certain limitations. It is well established that dietary choline has direct role in promoting hepatic lipid homoeostasis by regulating the expression of many key enzymes and transcription factors(Reference Hansen, Kortner and Denstadli39), which also have been confirmed in the present study and our previous studies(Reference Luo, Wei and Ye2). Thus, it is hard for us to remove this direct role of dietary choline regulating other critical enzymes and transcription factors, although the significance of choline as a methyl donor for hepatic lipid homoeostasis has been indicated in our current study.
In conclusion, the present study clearly demonstrated that dietary choline serving as methyl donor, by controlling specific CpG methylation sites of ire-1α, activated ire-1α promoter methylation, which led to the down-regulation of ire-1α expression and alleviated hepatic UPRer, and then consequently prevented hepatic lipid dysregulation via Ire-1α/Fas interaction. These results emphasise the critical contribution of methyl donor function of choline for its lipotropic agent role. Meanwhile, choline-prevented hepatic lipid dysregulation might also directly via reducing Fas expression, which need further investigation to confirm. The detail mechanism has been shown in Fig. 6.
Graphical conclusions for the mechanism of dietary choline prevent hepatic lipid dysregulation via specific CpG methylation of ire-1α. Choline (CH); dehydrogenase (CHDH); fatty acid synthase (Fas); TAG; unfolded protein response (UPR). Dietary choline serving as methyl donor, by controlling specific CpG methylation sites of ire-1α (–1497 and −1287), activated ire-1α promoter methylation, which down-regulated the ire-1α expression and alleviated UPRer, and then consequently prevented hepatic lipid dysregulation via Ire-1α/Fas interaction. Meanwhile, choline-prevented hepatic lipid dysregulation might also directly via reducing Fas expression.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (grant No. 32273156 and grant No. 31902380).
Z. Y. B. and Y. F. S. designed the experiment. Y. F. S. conducted the experiment and data analysis with the help of L. J. W. and H. Z. Y. F. S. drafted the manuscript. Z. L. revised the manuscript. All the authors reviewed and approved the manuscript.
The authors declared that they had no conflicts of interest with the contents of this article.
Supplementary material
For supplementary material referred to in this article, please visit https://doi.org/10.1017/S000711452300185X
