Diallyl disulfide alleviates hepatic steatosis by the conservative mechanism from fish to tetrapod: Augment Mfn2/Atgl-Mediated lipid droplet-mitochondria coupling

Abstract

Despite increasing evidences has highlighted the importance of mitochondria-lipid droplet (LD) coupling in maintaining lipid homeostasis, little progress in unraveling the role of mitochondria-LD coupling in hepatic lipid metabolism has been made. Additionally, diallyl disulfide (DADS), a garlic organosulfur compound, has been proposed to prevent hepatic steatosis; however, no studies have focused on the molecular mechanism to date. To address these gaps, this study investigated the systemic control mechanisms of mitochondria-LD coupling regulating hepatic lipid metabolism, and also explored their function in the process of DADS alleviating hepatic steatosis. To this end, an animal model of lipid metabolism, yellow catfish Pelteobagrus fulvidraco were fed four different diets (control, high-fat, DADS and high-fat + DADS diet) in vivo for 8 weeks; in vitro experiments were conducted to inhibit Mfn2/Atgl-mediated mitochondria-LD coupling in isolated hepatocytes. The key findings are: (1) the activations of hepatic LDs lipolysis and mitochondrial β-oxidation are likely the major drivers for DADS alleviating hepatic steatosis; (2) the underlying mechanism is that DADS enhances mitochondria-LD coupling by promoting the interaction between mitochondrion-localized Mfn2 with LD-localized Atgl, which facilitates the hepatic LDs lipolysis and the transfer of fatty acids (FAs) from LDs to mitochondria for subsequent β-oxidation; (3) Mfn2-mediated mitochondrial fusion facilitates mitochondria to form more PDM, which possess higher β-oxidation capacity in hepatocytes. Significantly, the present research unveils a previously undisclosed mechanism by which Mfn2/Atgl-mitochondria-LD coupling relieves hepatic LDs accumulation, which is a conserved strategy from fish to tetrapod. This study provides another dimension for mitochondria-LD coupling and opens up new avenues for the therapeutic interventions in hepatic steatosis.

Graphical abstract

Highlights

• •The activation of hepatic lipid droplets (LDs) lipolysis and mitochondrial β-oxidation are likely the major driver for diallyl disulfide (DADS) alleviating hepatic steatosis.
• •DADS enhances mitochondria-LD coupling by promoting the interaction between mitochondrion-localized Mfn2 with LD-localized Atgl, which facilitates the hepatic LDs lipolysis and the transfer of fatty acids (FAs) from LDs to mitochondria for subsequent β-oxidation.
• •Mfn2-mediated mitochondrial fusion facilitates mitochondria to form more peridroplet mitochondria (PDM), which possess higher β-oxidation capacity in hepatocytes.
• •Significantly, the present research unveils a previously undisclosed mechanism by which Mfn2/Atgl-mitochondria-LD coupling relieves hepatic LDs accumulation, which is a conserved strategy from fish to tetrapod.

1. Introduction

The cell is an organic entity, which means that organelles are both dynamic and not fully separated [1]. Lipid droplets (LDs) represent the primary energy storage organelle in most cell types (e.g., adipose cells, skeletal muscle cells, and hepatocytes) and act as the hub for lipid metabolism and signaling [2,3]. Interactions with other organelles, particularly the energy factories—mitochondria, are one way in which LDs influence cellular function [4]. Currently, The function of mitochondria-LD coupling mainly focuses on lipid metabolism. While it would intuitively seem that such close connections between the two organelles ensures an efficient transfer of lipolysis-derived fatty acids (FAs) from the LD surface to mitochondria for their subsequent oxidation [5,6]; paradoxically, mitochondria-LD contact is also responsible for enhancing LD biogenesis to shield the mitochondria from lipid induced toxicity [6,7]. This seemingly contradictory result suggests that currently the biological functions of mitochondria-LD interaction are multiplex and has not yet formed a unified notion in different species or tissues, especially in liver. In addition, recent research indicates that a disruption in effective mitochondria-LD communication is strongly associated with hepatic steatosis [8,9], which is the most common hepatic disease worldwide, affecting approximately one-third of adults [10]. Therefore, elucidation of the biological functions and mechanisms of mitochondria-LD contact could provide new insights into the hepatic lipid homeostasis in the subcellular (organelle) level, and enhance our understanding of the pathogenesis of hepatic steatosis.

Mitochondria-LD coupling, as dynamic organelle junctions, its integrity and function are precisely regulated by certain proteins localized to mitochondrion and/or LDs [11,12]. Mitofusin2 (Mfn2), a transmembrane (TM) protein localized to the outer mitochondrial membrane and involved in mitochondrial fusion, is highly conserved from Drosophila to humans [13]. It has been suggested that Mfn-2 has a crucial role in determining LD-mitochondria contacts, as its gene deletion blunts these interactions in brown adipocytes [14]. Theoretically, as a link between two originally independent organelles, mitochondria-localized Mfn2-mediated mitochondria-LD coupling requires the coordination with another LD-localized protein. A recent study has examined the Mfn2-mediated collaborative mechanisms with Hsc70 for coupling mitochondria-LD in the heart [15]. Nevertheless, this Mfn2-mediated coordination mechanism has yet to be defined in the liver. Adipose triacylglycerol lipase (Atgl) catalyzed the first step of lipolysis in cells (i.e. the mobilization of LDs) and also was conserved even from bacteria to humans [16,17]. Our previous study confirmed that Atgl could be recruited onto the surface of LDs as the LD-localized protein to participate in LDs/mitochondria interaction-mediated hepatic LDs lipolysis [18]. Thus, our hypothesis is that the existence of Mfn2/Atgl interaction in Mitochondria-LD coupling will result in a more effective shortcut link between LD lipolysis and mitochondrial β-oxidation by aiding FA trafficking directly toward mitochondria. Additionally, given the multifaceted roles of Mfn2 in both mitochondrial fusion and mitochondria-LD coupling, and the well-accepted mechanisms of mitochondrial fusion facilitating mitochondrial β-oxidation by enhancing mitochondrial matrix, we are interested in an open issue: is there another new mechanism of Mfn2-controlled mitochondrial fusion that promotes mitochondrial β-oxidation by regulating mitochondria-LD coupling?

Given LDs and mitochondria are common organelles in vertebrates, and Mfn2 and Atgl are conserved from lower organisms to humans, another unanswered question is whether the mechanisms of Mfn2/Atgl-regulated mitochondria-LD coupling and their functional role in maintaining hepatic lipid homeostasis are conservative in vertebrates? The most ancient bony vertebrates are teleost fish [16,19]. Therefore, understanding the evolutionary origins and ancient roles of mitochondria-LD coupling in early vertebrates is made possible by systematic studies in fish. Our previous study has indicated that yellow catfish, Pelteobagrus fulvidraco, a freshwater teleost widely distributed in East Asia, could be a good model for studying hepatic lipid metabolism in fish [[20], [21], [22]]. Thus, to obtain a comprehensive landscape of the evolutionary process and find clues to conservative mechanisms of mitochondria-LD coupling, this present study uses yellow catfish as a model to examine the underlying molecular mechanisms in fish and then conduct comparative studies among various vertebrates, including frogs, chickens, and mice. It is important to note that mitochondria-LD coupling is sensitive to nutrients and has different roles in different nutritional states [6,23]. Garlic (Allium sativum L.) as a widely consumed traditional food has various medicinal properties. A few pilot studies suggest that diallyl disulfide (DADS), a major component of the secondary organosulfur compounds derived from garlic, can activate Mfn2 and provide hepatoprotection against hepatic steatosis [24,25], although the underlying mechanisms remain largely unknown. Thus, this study also identifies the involvement and mechanism of Mfn2-mediated mitochondria-LD coupling in DADS-promoted hepatic lipid homeostasis, which may offer potential dietary avenues for intervention in hepatic steatosis.

2. Materials and methods

2.1. Ethical statement

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-2023-05-11).

2.2. Expt. 1: Animals feeding and sampling (in vivo experiment)

Dietary formula and yellow catfish feeding were determined according to our previous studies [26,27]. We formulated four experimental diets, as shown in Table S1. Dietary lipid concentrations were 10.8 % (control group), 16.1 % (high-fat diet group, HFD), 10.8 % (DADS diet group, DADS), and 16.1 % (high-fat + DADS diet group, HFD + DADS), respectively. Fish oil and soybean oil (1:1, w/w) were used as the lipid sources. The addition of DADS was 0 (Control), 0 (HFD), 18 (DADS), and 18 (HFD + DADS) mg of DADS per kg diet (≥99.0 % in purity, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China). At the start of the study, each tank was supplied with thirty fish of uniform size (4.71 ± 0.05 g, mean ± SEM). Each experimental diet was randomly assigned to the three water tanks, and 12 tanks were used in the experiment. The amount of feed consumed by the fish in each tank adopted quantitative feeding based on fish body weight and was recorded daily. Each diet was fed twice daily at 8:00 a.m. and 4:00 p.m. for 8 weeks. During the feeding experiment, the parameters in water quality were followed: dissolved oxygen 6.21 ± 0.11 mg/L; water temperature 28.0 ± 0.5 °C; pH 7.56 ± 0.23; NH4–N 0.051 ± 0.003 mg/L.

At the end of the feeding experiment, all fish were fasted for 24 h before sampling. MS-222 was used to euthanize all of the yellow catfish. Then all yellow catfish were counted and weighed to measure survival, weight gain (WG), specific growth rate (SGR) and feed conversion rate (FCR). The morphological parameters and growth performance are shown in Table S2. For one diet group, the liver tissues of nine fish from three tanks (3 fish from one tank) were sampled for the subsequent observations of histology and ultrastructure. Other fish liver samples were promptly frozen in liquid nitrogen for further examination, including the contents of triglyceride, PeriDroplet Mitochondria (PDM) extraction, gene and protein expression.

2.3. Expt. 2: Cultures and treatments of hepatocellular cell line or primary hepatocytes from varied vertebrates (in vitro experiment)

Primary hepatocyte culture and treatment: yellow catfish and frog (Rana chensinensis) were obtained from the laboratory animal center of HZAU. The primary hepatocytes were isolated and cultured from yellow catfish according to our recent publication [26,27]. The cells were incubated at 28 °C or 22 °C for 48 h. Each treatment was performed in triplicate plates and three biological replicates (3-well cells) in one plate. For each cell culture, a pool of cells from three fish or frog was used. In order to explore the role of Mfn2 and Atgl in PDM-mediated hepatic lipid metabolism, we designed and transfected siRNA against mfn2 and atgl, respectively, to hepatocellular cell line or primary hepatocytes from frog. The inhibition efficiency of siRNA on related protein expression in this analysis is given in Fig. S1 and Fig. S2. Meanwhile, in the present study, hepatocellular cell line (NCTC, mouse normal liver cell line; LMH, leghorn male hepatoma cell line; HepG2, human hepatocellular carcinoma cell line) were used to identify whether there are similar results in chickens, mice or human being cell line, with the transfection of siRNA against MFN2, and ATGL for chickens, mice or human being.

2.4. Sample Analysis

2.4.1. Oil red O, Hematoxylin and Eosin (H&E), Bodipy 493/503 staining and transmission electron microscopy (TEM) observation

For one diet group, the liver tissues of nine fish from three tanks (3 fish from one tank) were sampled for oil red O, H&E, Bodipy 493/503 staining and TEM observation. Oil red O and H&E staining tests were conducted followed the description in our publication [26]. The ten fields of each sample were quantified by the software Image J (NIH, Bethesda, MD, USA) to get the statistics of the relative areas of lipid droplets (LDs) in the Oil red O staining, the relative areas of vacuoles in the H&E staining. Bodipy 493/503 staining for the LDs and TEM observation of primary hepatocytes from yellow catfish were carried out according to our publication [26]. For TEM, ten pictures for one fish liver, and then ten fields of each sample were quantified by the software Image J (NIH, Bethesda, MD, USA) to get the statistics of length of mitochondrion.

2.4.2. Cell viability and determination of triglyceride, diacylglycerol and monoglyceride

Cell viability was measured with the use of 3-(4,5-dimethylthia -zol-2-yl)-2, 5-diphenyltetrazolium bromide (V13154; Thermo Fisher Scientific) and CellTiter-Meiluncell Luminescent cell viability assay kit (PWL111, ShangHai Meiluncell), following the manufacturer's instructions. The contents of triglyceride, diacylglycerol and monoglyceride were determined by commercial kits (A110-1-1, Nanjing Jiancheng Bioengineering Institute), according to the manufacturer's instructions. Soluble protein content was analyzed, based on protocols by Bradford [27].

2.4.3. The determination of nonesterified fatty acids (NEFAs), total bile acid, cholesterol and ATP contents

The contents of NEFAs, total bile acid and cholesterol in the tissue homogenate were evaluated using tissue NEFAs and cholesterol assay kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China; A042-2-1 for NEFAs, A111-1-1 for cholesterol and E003-2-1 for total bile acid), following the manufacturer's instructions. ATP was measured by using an ATP Colorimetric/Fluorometric Assay Kit (S0026; Beyotime, Nantong, China) according to the manufacturer's instructions. Briefly, cells were lysed and centrifuged at 12,000 RPM for 5 min. The supernatant (100 μL) was transferred to a 96-well plate and then mixed with ATP detection working solution (100 μL). Luminescence signals were measured by a microplate reader.

2.4.4. Analysis of mitochondrial FAs β-oxidation rate and Cpt1 activity

Briefly, mitochondrial FAs-oxidation rate was determined using labeled [1⁻¹⁴ C] PA (NEC075H050UC; PerkinElmer, Pittsburgh, PA, United States) as a substrate. As for Cpt1 activity assay, mitochondria were isolated from the liver or hepatocytes according to Morash et al. [28,29]. Cpt1 activity was determined using the method of Bieber and Fiol [28], based on measurement of the initial CoA-SH formation by the 5,5-dithio- bis-(2-nitrobenzoic acid) (DTNB) reaction from palmitoyl-CoA by mitochondrial samples with l-carnitine at 412 nm. One unit (IU) of Cpt1 activity was defined as 1 μmol of product formed per min per mg of mitochondrial protein at 25 °C.

2.4.5. Isolated hepatocellular mitochondria subpopulation

Cytoplasmic mitochondria and PDM were isolated by differential centrifugation as described in other studies [30,31]. Briefly, approximately ∼1.5 g of liver was washed with PBS, and treated with 7 mL of mitochondrial isolation buffer and BSA. After 10 up-down strokes at 200 rpm in a homogenizer, the homogenate was centrifuged at 900×g for 11 min at 4 °C. The lipid and aqueous phases were recovered from each tube and centrifuged at 9000×g for 18 min at 4 °C. After recovering the lipid phase from one tube and the aqueous phase from the other tube, the resulting pellets from both tubes (PDM and CM) were resuspended in mitochondrial isolation buffer without BSA.

2.4.6. RNA isolation and quantitative real-time PCR analysis (qPCR)

Total RNA was isolated by using the Trizol reagent and then transcribed into the cDNA by using the Reverse Transcription Kit. Analyses of gene transcript levels were conducted through the real-time quantitative PCR (qPCR) method. The primer sequences used in this analysis are given in Table S3. A set of nine housekeeping genes (gapdh, b2m, ef1α, 18s rRNA, tuba, β-actin, hprt1, ubce, rpl7) were selected from our transcriptome database in order to test their transcription stability. Our pilot experiment indicated that gapdh and β-actin (M = 0.28) showed the most stable levels of expression across the experimental conditions, as suggested by geNorm. Thus, the relative expression levels were normalized to the geometric mean of the combination of gapdh and β-actin and calculated using the 2^−ΔΔCt method. For siRNA experiment, the primer sequences of siRNA are given in Table S4.

2.4.7. DNA Isolation and rt-PCR for mtDNA

Based on our previously published study [21], the mtDNA was isolated by using a commercial kit (K280- 50; MT DNA Isolation Kit; Biovision, Mountain View, CA, USA). The mitochondrial atp8 gene was used to measure mtDNA copy number and the 18s rRNA gene as the standardized reference. The sequences of the primers are listed in Table S3. Concentrations of cytoplasm mtDNA were converted to copy number via a DNA copy number calculator [32]. All samples were measured with standards at the same time.

2.4.8. Oxygen consumption rate (OCR) measured by Seahorse

We determined the cellular OCR by Seahorse according to previously published studies [33,34]. Briefly, cells were inoculated on XF24 cell culture plates in 100 μL growth media per well on day 1. Check cells to ensure that they have adhered to plate reaching almost 90–100 % confluence, and carefully add 150 μL growth media to each well. Add 1 mL of Seahorse XF24 calibrant solution (pH 7.4) to each well of the Seahorse 24-well plate for preparing Sensor Cartridge. Incubate entire cartridge into a non-CO₂ incubator at 37 °C overnight or for up to 72 h (Day 2) All reagents should be at pH 7.4 and 37 °C at all times for the XF24 assay. Prepare XF media. Wash once with 1 mL XF assay media. Aspirate it off, and add 500 μL XF media to each well. Place plate in non-CO₂ incubator at 37 °C until ready to assay. The sensor cartridge is equipped with four injection ports for each well. Loading testing compounds in the Sensor Cartridge. Calibrating the Sensors and running the Seahorse XF24 Assay. Seahorse XF24 data can be analyzed directly by the Seahorse software.

2.4.9. RNAi and gene transfection

To generate mfn2 and atgl knockdown cells, hepatocellular cell line or primary hepatocytes from yellow catfish or frog were transfected with 103 nM of siRNA against mfn2 and atgl. Transfection was performed with EntransterTM-R4000 Transfection Reagent (4000-3; Engreen Corp, Beijing, China). Target sequences for preparing the siRNAs of different species mfn2 and atgl are shown in Table S4. The transfection of siRNA was performed using the EntransterTM-R4000 transfection reagent according to the supplier's protocol. Western blot was used to measure the knockdown efficiency of mfn2 and atgl, and the siRNA with the highest knockdown efficiency was selected for subsequent research (Supplementary Fig. S2A and Fig. S2B).

2.4.10. Immunoprecipitation and western blot

Immunoprecipitation was performed to identify the interaction between Mfn2 and Atgl. Briefly, in order to conduct the immunoprecipitation analysis, we lysed the cells in NP-40 buffer with the addition of a protease inhibitor cocktail. The cell lysate was then mixed with anti-Mfn2 (Cat No. 12186-1-AP; Proteintech group, America) or anti-Atgl (Cat No. 55190-1-AP; Proteintech group, America), respectively, as the bait protein, at 4 °C overnight. The incubation was sustained for 3 h, followed by the addition of protein A/G agarose. The immunocomplexes were washed 5 times by PBS supplemented with PMSF protease inhibitor. Finally, the western blot analysis was performed with anti-Mfn2 (Cat No. 12186-1-AP; Proteintech group, America) or anti-Atgl (Cat No. 55190-1-AP; Proteintech group, America), as the prey protein. In addition, for the mutation analysis of Mfn2 and Atgl, the full open reading frames of Mfn2/or lacks interaction sequence (Mfn2 Δ93-342, Mfn2 Δ343-629, Mfn2 Δ648-754) and the full open reading frame of Atgl or the Atgl reading frame lacking the interaction (Atgl Δ4-252, Atgl Δ253-393) were subcloned into the pcDNA3.1 (+) vector with the His-tag and HA-tag sequences, respectively. To identify the protein levels of Srebp1, Atgl, Cpt1α, Tom20, Opa1, Mfn2, Mfn1 and Gapdh western blot analysis was performed according to our previous study [21]. Image J was used to measure the densitometry of these bands (Fig.S3E, Fig. 3F and Fig. S3L). For the western blot trials, all the biological replicates for western blots are 3 (n = 3).

2.4.11. Protein docking for Mfn2 and Atgl

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 (Mfn2 PDB ID: 6jfm, Atgl PDB ID: 5fya) and the appropriate templates were selected. The ZDOCK module of Discovery Studio 4.0 was used to perform Mfn2-Atgl docking. The docking pose was picked up by using pymol from the given top models.

2.4.12. Immunofluorescence and 2D/or 3D confocal analysis

Immunofluorescence and confocal analysis were utilized to analyze the colocalization of Mfn2, Atgl, mitochondria and lipid droplets in the hepatic cell line or primary hepatocytes from yellow catfish or frog, according to previous publications [35,36]. For immunofluorescence analysis, in brief, cells were incubated with particular primary antibodies, anti-Mfn2, anti-Atgl (1:200). Then the cells were incubated for 60 min at room temperature and in the dark with a secondary anti-body made of goat anti-rabbit IgG H&L. The pictures were captured by a laser scanning confocal microscope (Leica Wetzlar, Germany). For the 3D imaging, the images were recorded by employing the Plan-Apochromat × 40 objective with a triple zoom. For mitochondria and lipid droplets confocal analysis, hepatic cell line or primary hepatocytes were stained by corresponding fluorescent probes, respectively, and then were captured by the laser scanning confocal microscope. The CytoFlex flow cytometer (Beckman Coulter, Miami, FL, USA) was used to measure the mitochondria, which were identified as the red dots. The software FlowJo v.10 was used to analyze the data.

2.5. Statistical Analysis

All data were expressed as the mean ± SEM, and statistical analysis was performed using GraphPad Prism software 10.0. For western blot, n = 3 three biological replicates for one treatment. For other data, n = 9, in vivo at least three fish form one tank and then three tanks for one treatment; in vitro each treatment was performed in triplicate plates and then at least three biological replicates (3-well cells) for one plate. The normality of data distribution and the homogeneity of variances were analyzed using the Kolmogorov–Smirnov test and Bartlett's test, respectively. Differences between the two groups (Si-NC and Si-RNA group) were analyzed using an unpaired Student's t-test, and multiple comparisons were assessed using a one-way analysis of variance (ANOVA) with Tukey's multiple comparison test. A significance level of P < 0.05 was applied for all analyses. Significance levels were denoted as ∗ or #P < 0.05, ∗∗ or ##P < 0.01, ∗∗∗ or ###P < 0.001, and ∗∗∗∗ or ####P < 0.0001. In addition, no significant differences were labeled as “ns”.

3. Results

3.1. Boosting hepatic lipolysis and β-oxidation are the major drivers for DADS in alleviating HFD-induced hepatic steatosis

Growth performance, morphological parameters and feed utilization are shown in Table S2. The survival rate was 100 % among the four treatments. When compared to other treatments, DADS diets have a significant increase in WG and SGR, but a significant decrease in FCR. Meanwhile, when compared to control treatments, HFD has a significant decrease in WG and SGR.

This finding suggests that HFD caused hepatic lipid accumulation and metabolism dysregulation. Fig. 1A–E shows that the HFD-induced hepatic steatosis in yellow catfish can be modelled by the significant area for triglyceride staining in the liver of the HFD group, which exceeded 25 %. Meanwhile, DADS diets apparently alleviate HFD-induced hepatic steatosis (Fig. 1A–G). In general, hepatic lipid accumulation results from the balance between dietary absorbed lipid, de novo lipogenesis and fat catabolism via lipolysis and β-oxidation. Interestingly, further studies indicates that hepatic lipolysis and its subsequent β-oxidation are likely the major driver for DADS in alleviating hepatic steatosis (Fig. 1H–N), evidenced by the up-regulated expression of genes involved in lipolysis and β-oxidation (g6pd, 6pgd, acca, fas, dgat1, srebp1, atgl, hsl, mgl, acsl1, cpt1, cpt2, lcad, mcad, scad and echs) in HFD + DADS group (Fig. 1H). The expression of Atgl and Cpt1 proteins support this further (Fig. 1I).

Fig. 1.

Boosting hepatic lipolysis and β-oxidation are the major drivers for DADS in alleviating HFD-induced hepatic steatosis. (A) Representative images of liver tissues stained by H&E and oil red O. 200 × magnification. scale bars, 50 μm; HE, hepatocytes. VA, vacuoles. (B) Representative images of hepatocyte stained by oil red O and bodipy, respectively. (C–D) Relative areas for hepatic vacuoles after H&E staining and LDs after oil red O staining, respectively. (E) Relative green fluorescence intensity in panel B. (F) Triglyceride content. (G) NEFA content. (H) The mRNA levels of genes involved in hepatic lipid homeostasis. (I) Western blot analysis of Srebp1, Atgl and Cpt1α. (J) Atgl activity. (K) DG content. (L) MG content. (M) Cpt1 activity. (N) FAs β-oxidation rate. (O) Cholesterol content. (P) HDL-C content. (Q) Total bile acid in liver. (R) The mRNA levels of genes involved in hepatic glycolysis, Cholesterol metabolism and bile acid metabolism. Lipid overload: high-fat diet (HFD) for in vivo in A, C, D, F, G, H, I, J, K, L, M, N pannels; FA (OA:PA/1:1) incubate for in vitro in B, E pannels. Values are means ± SEM. For western blot, n = 3 three biological replicates for one treatment. For other data, n = 9, in vivo at least three fish form one tank and then three tanks for one treatment; in vitro each treatment was performed in triplicate plates and then at least three biological replicates (3-well cells) for one plate. Significance levels were denoted as ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001. In addition, no significant differences were labeled as “ns”. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

On the other hand, except hepatic lipid metabolism, glycolysis and the biosynthesis of both bile acids and cholesterol, are also closely associated with hepatic steatosis. In this present study, comparing to HFD group, HFD + DADS diet up-regulate the content of HDL-C content and total bile acid, but down-regulate cholesterol content (Fig. 1O–Q). Further, the expression of genes involved in cholesterol efflux (sr-b1, ldlr, abcg1) and also bile acid cycle (fxr, shp) in liver are significantly up-regulated in HFD + DADS group (Fig. 1R), when comparing to HFD group. These results indicate DADS diet-activated bile acid cycle and cholesterol efflux also contribute to its role in alleviating hepatic steatosis.

3.2. Dietary DADS promotes mitochondrial fusion by activating the pparα binding sites of the Mfn2 promoter, but shows no impact on mitochondrial quantity

Given that the mitochondrial matrix is the main site where the β-oxidation occurs and the mitochondrial quantity and dynamics are the cornerstone for mitochondrial function [21], it would be intriguing to investigate whether dietary DADS promote β-oxidation mainly through mitochondrial quantity or dynamics. First, mtDNA fold analysis clearly indicates the dietary DADS or HFD show no significant impact on mitochondrial quantity, confirmed by the flow cytometry detection of mitochondrial mass, mtDNA copy number and the Tom20 protein expression (Fig. 2A–C). Second, TEM analysis and confocal images plainly reveal, as depicted in Fig. 2D–F, dietary DADS significantly increase the number of elongated morphology of mitochondria, suggesting the dietary DADS activating mitochondrial fusion. Further, the expression of genes and/or proteins involved in mitochondrial dynamics (mfn1, mfn2, opa1, drp1, fis1, mff, mid49, mid51) (Fig. 2G–I) confirms the dietary DADS-activated mitochondrial fusion. Among these key genes for mitochondrial fusion, the Mfn2 expression seems to be the most noticeable phenotype, implying that Mfn2 plays a central role as the primary contributor to the dietary DADS-activated mitochondrial fusion. Finally, thus, we further explore the mechanisms by which dietary DADS activates Mfn2-mediated mitochondrial fusion. Here, dual-luciferase reporter assay and EMSA show DADS up-regulated Mfn2 expression main by activating the cis-acting elements of pparα in Mfn2 promoter, which are confirmed by the result from site-mutation assay of pparα binding sites and pparα antagonist treatment (Fig. 2J and K). Further, pparα antagonist treatment also eliminate DADS-activated Mfn2 protein expression and the function on mitochondrial fusion and β-oxidation (Fig. 2M and Q), further indicating that DADS activates Mfn2 by the cis-acting elements of pparα in Mfn2 promoter. These results provide the first molecular evidence for DADS-activated mitochondrial fusion.

Fig. 2.

Dietary DADS promotes mitochondrial fusion by activating the pparα binding sites of the Mfn2 promoter, but shows no impact on mitochondrial quantity. (A) The mitochondrial content analyzed by flow cytometry, staining by Cy5 probe (mitochondrion) in primary hepatocytes from yellow catfish, and data used FlowJo for analysis. (B) mtDNA fold analysis. (C) Western blot analysis of Tom20. (D) Length of mitochondrion in panel E (TEM). (E) Representative images of hepatic ultrastructure (TEM) and confocal images of mitochondria. scale bars, 1 μm or 4 μm. (F) Quantification of the primary hepatocytes from yellow catfish with different states for mitochondria (fragmented, intermediate or elongated, n = 300 cells counted per condition). (G) The ratio of mfn1+2/drp1+fis1 in mRNA levels. (K) Relative protein levels in pannel H. (H) The mRNA levels of genes involved in mitochondrial dynamics. (I) Western blot analysis of Opa1, Mfn2 and Mfn1. (J) Site-mutation assay of RUNX1 and PPARα binding sites on mfn2-405 vector in 293T cells. (K) Site-mutation assay with pparα antagonists (GW 6471) in 293T cells. (L) EMSA of PPARα binding sequence. The 5′-biotin labeled double-stranded oligomers were incubated with nuclear protein (10 μg). A 100-fold excess of the competitor or mutative competitor oligomers was added to compete with biotin-labeled oligomers. (M) Western blot analysis of Mfn2. (N) Quantification of the primary hepatocytes from yellow catfish with different states for mitochondria (fragmented, intermediate or elongated, n = 300 cells counted per condition) in Pannel O. (O) Confocal images of mitochondria. scale bars, 4 μm. (P) Cpt1 activity. (Q) FAs β-oxidation rate. Lipid overload: high-fat diet (HFD) for in vivo in E, F, H, I, J, K pannels; FA (OA:PA/1:1) incubate for in vitro in A, B, C, D, E, G pannels. Values are means ± SEM. For western blot, n = 3 three biological replicates for one treatment. For other data, n = 9, in vivo at least three fish form one tank and then three tanks for one treatment; in vitro each treatment was performed in triplicate plates and then at least three biological replicates (3-well cells) for one plate. Significance levels were denoted as ∗ or #P < 0.05, ∗∗ or ##P < 0.01, ∗∗∗ or ###P < 0.001, and ∗∗∗∗ or ####P < 0.0001. In addition, no significant differences were labeled as “ns”. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

3.3. Mfn2-mediated mitochondrial fusion enables the generation of more PDM, which possess a higher capacity for β-oxidation

Since Mfn2 has a dual regulatory function in both mitochondrial fusion and mitochondria-LD coupling [2], we will next examine the relationship between mitochondrial fusion and mitochondria-LD coupling. As shown in Fig. 3A–G, TEM and confocal images observations show that fused-mitochondria are more likely to bond to LDs, which suggests the potential association between mitochondrial fusion and mitochondria-LD coupling. And for this reason, also given that DADS activates Mfn2-mediated mitochondrial fusion, we clearly found the higher proportion of PDM in primary hepatocytes from yellow catfish after DADS treatment.

Fig. 3.

Mfn2-mediated mitochondrial fusion enables the generation of more PDM, which possess a higher capacity for β-oxidation. (A) Representative images of hepatic ultrastructure (TEM). scale bars, 2 μm or 1 μm. (B–C) Quantification of mitochondrial recruitment to LDs assessed as the area of mitochondria within 0.5 mm of LD border. (D) Representative confocal microscopy image after MitoTracker Deep Red and BODIPY493/503 staining. (E) The contact site between mitochondria and LDs. (F) Mitochondria in contact with lipid droplets were quantified by % lipid droplet perimeter. (G) Quantification of the primary hepatocytes from yellow catfish with PDM and CM. (H) Number of LDs. (I) Diameter of LDs. (J) OCR was measured by consecutive injections of oligomycin (1.5 μM), FCCP (0.2 μM), and antimycin A (0.5 μM)/rotenone (0.5 μM) (n = 3). (K) PDM are stripped from LDs by high-speed centrifugation. (L) Western blot analysis of Tgn46, Plin2, Gapdh and Tom20. (M) Cpt1 activity. (N) FAs β-oxidation rate. Lipid overload: high-fat diet (HFD) for in vivo in A, B, C pannels; FA (OA:PA/1:1) incubate for in vitro in D, E, F, G pannels. Values are means ± SEM. For western blot, n = 3 three biological replicates for one treatment. For other data, n = 9, in vivo at least three fish form one tank and then three tanks for one treatment; in vitro each treatment was performed in triplicate plates and then at least three biological replicates (3-well cells) for one plate. Significance levels were denoted as ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001. In addition, no significant differences were labeled as “ns”. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Having confirmed that Mfn2-mediated mitochondrial fusion forms more PDM, next we are interested in the alteration in hepatic lipid metabolism and mitochondrial function. As shown in Fig. 3H–J, along with the reduction of LDs' size and number, DADS treatment significantly promoted mitochondrial function, evidenced by the oxygen consumption rate (OCR) analysis. Given the higher proportion of PDM in DADS treatment group, thus this study proposes that PDM may also possess a higher β-oxidation capacity than cytoplasmic mitochondria (CM). To verify this hypothesis, we developed an approach to isolate intact respiring PDM and determine their bioenergetic function for the first time in fish. As shown in Fig. 3K and L, confocal images of PDM and the detection of marker proteins from different organelles confirm the isolation of PDM and CM from hepatocytes of yellow catfish. Importantly, the PDM's β-oxidation capacity is higher, as evidenced by its Cpt-1 activity and FAs β-oxidation rate (Fig. 3M and N).

3.4. Mfn2-mediated mitochondrial fusion control mitochondria-LD coupling may via interaction with Atgl

After confirming that the potential association between mitochondrial fusion and mitochondria-LD coupling, the role and related mechanism of Mfn2 controlling mitochondria-LD coupling were further tested. First, as depicted in Fig. 4A–C, the inhibition of Mfn2 inevitably causes a large number of rod-shaped mitochondria and LDs to accumulate in primary hepatocytes from yellow catfish. Second, si-mfn2 has no discernible impact on mitochondrial quantity, which is confirmed by the flow cytometry detection, mtDNA copy number and the Tom20 protein expression (Fig. 4D–G). However, si-mfn2 transfection causes substantial damage to structural integrity of mitochondria-LD coupling, clearly evidenced by the 2D and 3D confocal images (Fig. 4H and I), indicating the indispensable role of mfn2 in maintaining the mitochondria-LD coupling. Further, along with the reduction of mitochondria-LD coupling and PDM, the inhibition of Mfn2 also results in the impaired capability of FA trafficking toward mitochondria (Fig. 4J–L). This also from one perspective supports the notion that PDM have higher β-oxidation capacity may via increasing the capability of FA trafficking toward mitochondria.

Fig. 4.

Mfn2-mediated mitochondrial fusion control mitochondria-LD coupling may via interaction with Atgl. (A) Representative confocal microscopy image after Cy5 (mitochondrion) and BODIPY493/503 staining, respectively. (B) Relative mitochondrion fluorescence intensity in panel A. (C) Relative green fluorescence intensity in panel A. (D–E) mtDNA fold analysis. (F) Western blot analysis of Tom20. (G) The mitochondrial content analyzed by flow cytometry, staining by Cy5 probe (mitochondrion) in primary hepatocytes from yellow catfish, and data used FlowJo for analysis. (H) Representative confocal microscopy image after MitoTracker Deep Red and BODIPY493/503 staining. (I) The fusion rate between MitoTracker and LDs in pannel H. (J) Atgl activity. (K) Representative confocal images of mitochondria and Red C12 480/508, scale bars, 5 μm. (L) The fusion rate between mitochondria and Red C12 in pannel K. (M) Representative confocal microscopy image after MitoTracker Deep Red and BODIPY493/503 staining. (N) The fusion rate between MitoTracker and LDs in pannel M. (O) Cpt1 activity. (P) FAs β-oxidation rate. Values are means ± SEM. For western blot, n = 3 three biological replicates for one treatment. For other data, n = 9, in vitro each treatment was performed in triplicate plates and then at least three biological replicates (3-well cells) for one plate. Significance levels were denoted as ∗ or #P < 0.05, ∗∗ or ##P < 0.01, ∗∗∗ or ###P < 0.001, and ∗∗∗∗ or ####P < 0.0001. In addition, no significant differences were labeled as “ns”. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Our previous study has confirmed that Atgl could be recruited into the LDs surface as a LDs-localized protein to mediate the interaction between LDs and mitochondria [18]; however, the related mechanism has yet to be defined. Given that the DADS diet also activates LDs lipolysis and Atgl catalyzes the first step of lipolysis, we hypothesized that Mfn2-mediated mitochondrial fusion control mitochondria-LD coupling via the interaction with Atgl. 2D and 3D confocal images clearly indicate that si-atgl cause the loss of DADS diet-induced decrease of mitochondria-LD coupling (Fig. 4M and N), suggesting the potential involvement of Atgl in DADS diet-activated mitochondrial fusion in controlling mitochondria-LD coupling. Further, along with the reduction of mitochondria-LD coupling, the DADS diet-activated β-oxidation is also significantly suppressed, which is proven by the Cpt1 activity and FAs β-oxidation rate (Fig. 4O and P).

3.5. Mfn2 maintains the formation of mitochondria-LD coupling by coordinating with Atgl's C-terminal domain through its GTPase domain

Following our above-mentioned findings that Mfn2/Atgl can regulate the formation of mitochondria-LD coupling, the in-depth molecular mechanisms by which Mfn2 coordinated with Atgl to regulate the formation of mitochondria-LD coupling and their roles in hepatic lipid homeostasis are further examined. First, co-staining Mfn2 with Atgl clearly shows colocalization between Mfn2 and Atgl. However, the DADS treatment significantly further promoted this colocalization (Fig. 5A). The protein docking further validate the protein-protein interaction between Mfn2 and Atgl. It is noteworthy that subcellular Co-IP analysis indicates that this Mfn2/Atgl co-localization mainly exists in the PDM region, not the whole-cell intracytoplasmic (Fig. 5B). Second, our fragment analysis indicates that this Mfn2/Atgl relies on Mfn2's GTPase-domain (93–342 aa) coordinating with Atgl's C-terminal domain (253–393 aa), evidenced by the result of Co-IP analysis and Mfn2/Atgl co-staining after the transfection of mfn2 or atgl deletion mutant, respectively (Fig. 5C–G). Last and most important, 3D confocal images clearly show that the fragments deletion of both Mfn2's GTPase-domain and Atgl's C-terminal domain caused the massive dismiss of mitochondria-LD coupling and PDM (Fig. 5H–J). Altogether, these results indicate that Mfn2 maintains the formation of mitochondria-LD coupling by coordinating its GTPase domain with Atgl's C-terminal domain. These regulated mechanisms mainly exist in the PDM region.

Fig. 5.

Mfn2 maintains the formation of mitochondria-LD coupling by coordinating with Atgl's C-terminal domain through its GTPase domain. (A) Representative confocal microscopy image of Mfn2 (green) and Atgl (red) in yellow catfish hepatocytes. (B) CoIP of Mfn2 with Atgl from yellow catfish hepatic fat laying and except PDM of whole cell lysate. (C) Protein docking prediction for Mfn2/Atgl interaction and their core binding domain. (D–G) 293T cell were transfected with wild-type Mfn2/Atgl and its truncated mutant and applied to immunoprecipitation assay, respectively. (H) Representative confocal microscopy image after MitoTracker Deep Red and BODIPY493/503 staining. (I) The fusion rate between MitoTracker and LDs in pannel H. (J) Mitochondria in contact with lipid droplets were quantified by % lipid droplet perimeter in pannel H. Values are means ± SEM. For western blot, n = 3 three biological replicates for one treatment. For other data, n = 9, in vitro each treatment was performed in triplicate plates and then at least three biological replicates (3-well cells) for one plate. Significance levels were denoted as ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001. In addition, no significant differences were labeled as “ns”. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

3.6. Mfn2/Atgl-mediated mitochondria-LD coupling has been a conserved strategy for relieving hepatic LD accumulation throughout vertebrate evolution

Given LDs and mitochondria are common organelles in vertebrates, and Mfn2 and Atgl are conserved from lower organisms to humans [13], more research was conducted to examine whether the mechanisms of Mfn2/Atgl-regulated mitochondria-LD coupling and their functional role in maintaining hepatic lipid homeostasis are conservative in vertebrates. By using hepatocellular cell line or primary hepatocytes (frog) from varied vertebrates, including frogs, chickens, mice and human being, both si-MFN2 and si-ATGL transfection significantly caused a substantial dismissal of mitochondria-LD coupling and PDM in hepatocellular cell line or primary hepatocytes (frog) among these diverse vertebrates (Fig. 6A–C, G-I, M − O and S–U). This is consistent with the findings from yellow catfish, suggesting that the mechanisms of Mfn2/Atgl complex-mediated mitochondria-LD coupling are highly conserved across diverse vertebrate lineages. Meanwhile, along with the suppression of mitochondria-LD coupling and PDM caused by Mfn2/Atgl inhibition, hepatic LD lipolysis and its subsequent mitochondrial -oxidation are also significantly suppressed (Fig. 6D–F, J-L, P–R and V-X) in hepatocellular cell line or primary hepatocytes (frog) from varied vertebrates. These suggest the regulated role of Mfn2/Atgl -mediated mitochondria-LD coupling in promoting hepatic lipolysis and β-oxidation is also conservative in vertebrates.

Fig. 6.

Mfn2/Atgl-mediated mitochondria-LD coupling has been a conserved strategy for relieving hepatic LD accumulation throughout vertebrate evolution. (A, G, M, S) Representative confocal microscopy image after MitoTracker Deep Red and BODIPY493/503 staining in frog (hepatocyte), chicken (LMH) and mouse (NCTC), human being (HepG2), respectively. (B, C, H, I, N, O, T, U) The fusion rate between MitoTracker and LDs. (D, J, P, V) Atgl activity. (E, K, Q, W) FAs β-oxidation rate. (F, L, R, X) Cpt1 activity. Values are means ± SEM. n = 9, in vitro each treatment was performed in triplicate plates and then at least three biological replicates (3-well cells) for one plate. Significance levels were denoted as ∗ or #P < 0.05, ∗∗ or ##P < 0.01, ∗∗∗ or ###P < 0.001, and ∗∗∗∗ or ####P < 0.0001. In addition, no significant differences were labeled as “ns”. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

4. Discussion

The importance of mitochondria-LD coupling for maintaining lipid metabolism is being highlighted by increasing evidence [7,37]. However, the role of mitochondria-LD coupling in hepatic lipid homeostasis is less clear, with even contradictory results have been mixed. DADS, one of the organosulfur compounds found in garlic, has demonstrated protective effects against hepatic steatosis [38]; however, the molecular mechanism is largely unknown. In this study, we unveil a previously undisclosed mechanism wherein DADS enhances mitochondria-LD coupling by mediating the mitochondrial-localized Mfn2 binding to LD-localized Atgl, thereby facilitating the lipolysis of hepatic LDs and the subsequent transfer of FAs from LDs to mitochondria for β-oxidation, ultimately alleviating HFD-induced hepatic steatosis. Significantly, for the first time, a conservative mechanism underlying Mfn2-Atgl-mediated mitochondria-LD coupling promoting hepatic lipolysis from fish to tetrapod has been discovered. This may be a strategy conserved throughout vertebrate evolution in response to lipid overload, which offers novel insights into the function of mitochondria-LD coupling and also inform the development of future therapeutic strategies for treating hepatic steatosis.

Consistent with our previous findings [32,39], our current study established a model of hepatic steatosis caused by HFD in yellow catfish. In accordance with previous studies [40], the DADS diet significantly alleviated the phenotype of hepatic LDs accumulation, indicating its protective effects on hepatic steatosis. Lipogenesis, fatty acids (FAs) uptake, LDs lipolysis, FAs β-oxidation, and lipid transport/secretion are all involved in the regulation of hepatic lipid accumulation through a complex network [32]. It is our finding that hepatic lipolysis and its subsequent β-oxidation are likely the primary cause of DADS in alleviating hepatic steatosis. In addition to hepatic lipid metabolism, other metabolites, including glycolysis and the biosynthesis of both bile acids and cholesterol, are also closely associated with hepatic steatosis [41,42]. Similarly with findings from previous studies partially [43], the present findings also indicate that DADS alleviated HFD-induced metabolic disorder involved in biosynthesis of bile acids and cholesterol. These results show the alteration biosynthesis of both bile acids and cholesterol also contribute to the DADS-alleviated hepatic steatosis. Given that the mitochondrial matrix is the main site where FAs β-oxidation occurs, and the quantity and dynamics of mitochondria are the foundation for mitochondrial function [44], it would be intriguing to investigate the role of mitochondrial quantity and dynamics in this process. Unexpectedly, the present results show that the DADS diet has no a significant impact on the mitochondrial quantity, but it clearly changed the mitochondrial dynamics, increasing the elongated morphology of mitochondria. Our findings suggest that, compared to Mfn1, Mfn2 could be the main contributor to mitochondrial fusion in DADS diet, partially consistent with previous studies [21,25]. Moreover, our present study also further elucidates the underlying molecular mechanism of the DADS diet activating the binding sites of the pparα transcription factor on the Mfn2 gene promoter, evidenced by both site-mutation assay and pparα antagonists. This provides the first molecular evidence of DADS diet boosting mitochondrial fusion, although the promoted role of garlic diet on mitochondrial function has been well-identified.

Currently, the best-known way to facilitate mitochondrial fusion and facilitate FAs -oxidation is by increasing mitochondrial matrix, which then provides a larger space for β-oxidation [21]. Given Mfn2 exerts dual regulatory functions in both mitochondrial fusion and mitochondria-LD coupling [2], one unanswered question is whether Mfn2-mediated mitochondrial fusion can be directly mediated through mitochondria-LD coupling to promote FAs β-oxidation, apart from enhancing mitochondrial matrix? First, in accordance with previous studies [22], our current study confirms the crucial role of Mfn2 in maintaining the structure of mitochondria-LD coupling. Next, interestingly, further research has revealed that mitochondria in fused states are more susceptible to binding to LDs, which can be clearly seen in TEM and confocal images in hepatocytes. Given that DADS diet could promote mitochondrial fusion [25], we found a higher proportion of peridroplet mitochondria (PDM) in hepatocytes from the DADS diet group. In totality, this observation validates the correlation between Mfn2-controlled mitochondrial fusion and the proportion of PDM. Importantly, our analyses of isolated mitochondria first report that PDM has a higher capacity for fatty acid oxidation than cytoplasmic mitochondria (CM) in hepatocytes. From another angle, in DADS diet group, the higher oxygen consumption rate (‌OCR) and the superior trafficking capability of FA into MT can also support this conclusion. However, on the contrary, a recent study revealed that PDM, which was isolated from brown adipose tissue, has low fatty acid oxidation [6]. These completely opposite results indicate that the role of PDM in fatty acid oxidation is tissue-dependent, which required additional in-depth mechanistic investigations. Nevertheless, in the present study, after confirming the positive correlation between mitochondrial fusion and the proportion of PDM in hepatocytes, and PDM isolated from liver tissue have higher FAs β-oxidation capacity, we are convinced that PDM enhancement is also the major route for mitochondrial fusion facilitating FAs β-oxidation. Collectively, these data demonstrate fused-mitochondria have a higher chance of binding to LDs, resulting in greater PDM formation, which leads to FA trafficking toward fused-mitochondria for a higher β-oxidation efficiency.

Mfn2, as the mitochondrion-localized protein, requires coordination with another LDs-localized protein to mediate the mitochondria-LDs coupling [15,22]. Our previous sub-proteome analysis of LDs proteins, which was isolated from liver of yellow catfish, confirmed that Atgl could be recruited into the LDs surface as a LDs-localized protein to mediate LDs/mitochondrial interaction in response to HFD [18]. The first step of LDs lipolysis was triggered by Atgl [45,46], which is also a prerequisite for the subsequent mitochondrial FAs β-oxidation. Given DADS diet also activates the hepatic lipolysis, thus, in exploring the underlying molecular mechanism of Mfn2 coordinating with another LDs-localized proteins, our focus was on whether Mfn2 and Atgl interact and their roles in mitochondria-LDs coupling. First, Atgl inhibition does not cause a significant change in mitochondrial quantity and dynamics, but results in a decrease in the number of PDM, suggesting that Atgl is involved in PDM formation. Furthermore, our subcellular Co-IP analysis clearly indicates that MT-localized Mfn2 has an interaction with LD-localized Atgl in the PDM region. Next, the interaction is based on the GTPase domains of Mfn2 and the C-terminal catalytic domain of Atgl, as revealed by fragment analysis. Importantly, both the ablation of GTPase domains of Mfn2 and the C-terminal catalytic domain of Atgl results in the disassociation between mitochondria and LDs, leading to a significant reduction in PDM. This indicates the critical role of Mfn2/Atgl complex on maintaining the PDM formation. The Mfn2-coordinated molecular machinery that regulates the structural formation of PDM with LD-localized Atgl has been identified for the first time in these findings. On the other hand, surprisingly, ablation of the Mfn2 GTPase domain cause a decrease in the lipolysis capacity of LDs, while disruption of the Atgl C-terminal catalytic domain leads to a decrease in the efficiency of β-oxidation. These results indicate that the interaction between Mfn2/Atgl affects each other's functions in lipid metabolism. In summary, the relationship between mitochondrial fusion and PDM has been made clearer by these data: Mfn2-activated mitochondrial fusion is more prone to form more PDM, wherein Mfn2 coordinate with LD-localized Atgl, which further promote LDs lipolysis and facilitate FAs trafficking toward fused-mitochondria via PDM, ultimately leading to higher FAs β-oxidation efficiency. Thus, our data also supports the concept that this efficient way of FAs β-oxidation is the basis for DADS diet to prevent hepatic steatosis. However, on the contrary, in the brown adipose tissue, PDM have reduced β-oxidation capacity, although PDM have increased pyruvate oxidation, electron transport, and ATP synthesis capacities [6]. These opposite results indicate that the β-oxidation capacity of PDM shown tissue-depend, which need further verified.

The most numerous vertebrates on the planet are teleost fish, which represent the most ancient bony vertebrates [47]. Due to the completely different living environment and the existence of the fish-specific genome duplication event (FSGD), there may be some differences between fish, including yellow catfish, and other vertebrates in mechanism of hepatic lipid metabolism. Thus, having identified the mechanisms underpinning the Mfn2/Atgl complex, which mediates mitochondria-LD coupling and their roles in hepatic lipid metabolism in fish, next we explored whether this mechanism and its function in response to lipid overload possess conservatism among different species. First, both Mfn2 and Atgl inhibition causes a decoupling between mitochondria and LDs, leading to a significant reduction of PDM number in hepatocytes from frogs to human being. This indicates the Mfn2/Atgl complex-mediated mitochondria-LD coupling is highly conserved across diverse vertebrate lineages such as yellow catfish, frog, chicken, mice and human being. Ablation of mitochondria-localized Mfn2 or LD-localized Atgl could have a similar effect on each other's functions in hepatic lipid metabolism, as previously concluded in fish; for instance, Mfn2 disruption caused the lower LDs lipolysis capacity in hepatocytes. Meanwhile, lipid overload reduces the proportion of PDM in hepatocytes, resulting in the accumulation of hepatic LDs. This suggests that it could play a crucial role in supporting the functional similarities of Mfn2/Atgl complex-mediated mitochondria-LD coupling in vertebrates. Collectively, these findings support a novel notion that Mfn2/Atgl-mitochondria-LD coupling relieving hepatic LDs accumulation is an ancient and conserved strategy throughout vertebrate evolution, which has already emerged in some teleost species.

To summarize, our study reveals an important role of Mfn2-Atgl-mediated mitochondria-LD coupling in hepatic lipid homeostasis. On this basis, we proposed a new concept for DADS diet alleviating hepatic steatosis via the efficient way of FAs β-oxidation: Mfn2-activated mitochondrial fusion contribute to formation of PDM, wherein Mfn2 coordinate with LD-localized Atgl, which further promotes LDs lipolysis and facilitates FAs trafficking toward fused-mitochondria via PDM, ultimately leading to higher FAs β-oxidation efficiency. Importantly, our findings offer novel perspectives from an evolutionary perspective to comprehend the evolution of mitochondria-LD coupling and its function in hepatic lipid homeostasis. A working model of aforementioned mechanism is shown in Fig. 7.

Fig. 7.

Graphical conclusions for the mechanism of diallyl disulfide alleviate hepatic steatosis by the conservative mechanism from fish to tetrapod.

CRediT authorship contribution statement

Ling-Jiao Wang: Formal analysis, Methodology, Project administration. Xiao-Hong Lai: Conceptualization, Methodology. Zhi Luo: Writing – review & editing. Guang-Li Feng: Data curation. Yu-Feng Song: Funding acquisition, Resources, Writing – original draft, Writing – review & editing.

Data availability statement

The datasets in the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

The study protocol was reviewed and approved by the Ethical Committee of Huazhong Agricultural University's (HZAU) (identification code: Fish-2023-05-11) and conformed to the ethical standards for the care and use of laboratory animals (no human subjects), as laid out in the 1964 Declaration of Helsinki and its later amendments.

Funding

National Natural Science Foundation of China (grant No. 32273156), Biological Breeding-Major Projects (2023ZD04054) and Fundamental Research Funds for the Central Universities, China (grant No. 20062023SCYJ003).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: