Fructose-Induced Progression of Steatohepatitis Involves Disrupting Aldolase B-AMPK Signaling in Methionine Adenosyltransferase 1A Deficient Mice

Abstract

Objective:

Aldolases (ALDO) are sensors that regulate AMPK via binding to fructose 1,6-biphosphate (FBP), an intermediate of glucose and fructose metabolism. Fructose consumption is linked to metabolic dysfunction-associated steatotic liver disease (MASLD) progression but whether ALDO-AMPK signaling is involved is unknown. Methionine adenosyltransferase alpha 1 (Mat1a) knockout (KO) mice have low hepatic S-adenosylmethionine (SAMe) level and spontaneously develop steatohepatitis. ALDOB methylation has not been reported and here we investigated whether SAMe level regulates ALDOB and ALDOB-AMPK signaling and whether fructose feeding accelerates MASLD progression by disrupting ALDOB-AMPK signaling.

Methods:

Mass spectrometry identified ALDOB methylation sites and recombinant in vitro approaches assessed how methylation at those sites affects ALDOB oligomerization and activity. Primary hepatocytes cultured with high/low glucose and/or fructose and wild type (WT) and Mat1a KO mice fed with a high-fructose diet examined AMPK-ALDOB signaling and MASLD progression.

Results:

In Mat1a KO livers ALDOB R173 is hypomethylated while ALDOB activity is enhanced. Recombinant ALDOB is methylated at R173 and R304 by protein arginine methyltransferase 1. Low hepatic SAMe level results in hypomethylated ALDOB, which favors the tetrameric form that has higher enzymatic activity, and higher capacity to signal to activate AMPK. Fructose, independently of glucose levels, inhibited AMPK activity and induced lipid accumulation in hepatocytes. Mat1a KO mice have hyperactivated AMPK and fructose feeding inhibits it, enhancing the accumulation of fat in the liver and the progression of MASLD.

Conclusion:

Hepatic SAMe levels regulate ALDOB oligomeric state and enzymatic activity impacting on AMPK signaling and fructose-induced MASLD progression.

1. INTRODUCTION

Fructose consumption is a major contributor of the metabolic syndrome and accelerates the progression of metabolic dysfunction-associated steatotic liver disease (MASLD, previously known as non-alcoholic fatty liver disease) [1–3]. Fructose is found naturally in fruits and certain vegetables, but industrial fructose like the high-fructose corn syrup is widely used in almost all kinds of processed foods and beverages [4–5]. In the United States fructose makes up between 5 and 15% of the daily caloric intake [6] and its consumption positively correlates with obesity, diabetes, cardiovascular disease, hypertension, and also MASLD [7], whose incidence keeps increasing annually at an alarming level [8]. Fructose is rapidly taken up by the liver and preferentially converted to fructose 1-phosphate (F1P). Thereafter, F1P is channeled into the glycolytic pathway by aldolase B (ALDOB, the main form of aldolase (ALDO) expressed in adult liver), a reversible enzyme that supports both glycolysis and gluconeogenesis depending on the metabolic state of the cell and the direction of the metabolic flux. ALDOB catalyzes the reversible cleavage of fructose 1,6-bisphosphate (FBP) into dihydroxyacetone phosphate (DHAP) and glyceraldehyde 3-phosphate (G3P) [9]. Recently, ALDOs have been shown to function as sensors of FBP that regulate AMP-activated protein kinase (AMPK) activity independent of the canonical regulation by the adenosine monophosphate (AMP)/adenosine triphosphate (ATP) ratio [10–11]. When unoccupied by FBP, ALDO promotes the formation of a lysosomal complex containing the v-ATPase, Ragulator and Rag guanosine triphosphatases (GTPases) that activates AMPK [10–11]. This way, high glucose or fructose would signal to v-ATPase via binding of ALDOB to FBP, blocking the activation of AMPK. Since higher AMPK activity enhances mitochondrial fatty acid (FA) oxidation [12], fructose consumption could potentially worsen MASLD by this mechanism. Consistently, a recent work showed that blocking this pathway with the aldolase-FBP inhibitor aldometanib provides beneficial metabolic effects and protects against fatty liver in rodents by selectively activating the lysosomal pool of AMPK [13].

In this communication we provide evidence of a novel regulation of ALDOB in the liver. This was discovered while investigating the methyl-proteome of mice lacking methionine adenosyltransferase 1A (MAT1A), which exhibit chronically low hepatic S-adenosylmethionine (SAMe) level, and spontaneously development metabolic dysfunction-associated steatohepatitis (MASH) [14] and hepatocellular carcinoma (HCC) [15]. While analyzing the methyl-proteome of Mat1a knockout (KO) mice, we found they have lower levels of monomethyl-ALDOB at R173 and this correlated with higher ALDOB enzymatic activity and AMPK activation. Using an in vitro system, we confirmed that ALDOB is regulated by protein arginine N-methyltransferase 1 (PRMT1) methylation at R173 and R304, and that this affects its structure and enzymatic activity. ALDOB methylated at R173 or R304 has higher dimer:tetramer ratios and lower activity. We observed that SAMe regulates ALDOB activity, ALDOB-AMPK interaction and ultimately AMPK activation, suggesting that ALDOB methylation prevents FBP sensing. We speculate that the higher ALDOB-AMPK activation in Mat1a KO mice is a compensatory response to enhance FA β-oxidation, which is disrupted by fructose feeding leading to marked acceleration of MASLD. Altogether, our data unravels a new mechanism of regulation of ALDOB that contributes to accelerate MASLD in livers where SAMe level is low.

2. MATERIALS AND METHODS

2.1. Animals

Eight-month-old Mat1a KO (C57BL/6) male mice with increased alanine transaminase (ALT) and aspartate transferase (AST) levels and hepatic lipid accumulation (determined by ultrasound) were given orally by gavage with vehicle (water) (n = 12) or SAMe (30 mg/kg/day, n = 12; Abbott, Chicago, IL) for 8 weeks before death. Age-matched WT male littermates showing normal liver serum enzyme levels and ultrasound also were treated with vehicle for the same duration (n = 11). Animals were bred and housed in the CIC bioGUNE animal unit, accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. The mice were housed in groups using high-quality wood pellet hygienic litter bedding (Lignocel HBK 1500–3000; Rettenmaier & Sönne, Rosenberg, Germany) and in the presence of enrichment materials. Animals were fed with standard commercial chow animal diet (ref. 2914; Envigo, Barcelona, Spain). All procedures were performed during the light cycle and were approved by the Diputación de Bizkaia upon a favorable assessment by the Institutional Animal Care and Use Committee at CIC bioGUNE (Spain) [16].

Six-month-old male WT and Mat1a KO (C57BL/6) littermates were used for the fructose feeding study. Mice were randomly assigned to a low-fructose (Cat.: D08111507, Research Diets, Inc., NJ, USA) or high-fructose diet (Cat.: D08111508, Research Diets, Inc., NJ, USA) for 8 weeks. Mice were monitored weekly. All procedure protocols, use, and the care of the animals were reviewed and approved by the Institutional Animal Care and Use Committee at Cedars-Sinai Medical Center and all experiments involving research animals were conducted in accordance with all relevant ethical regulations. All mice are housed under 12-h light/12-h dark cycle at an average temperature of 74 F and 40% humidity.

2.2. Cells

Primary mouse hepatocytes were isolated following the standard two-step collagenase perfusion technique. Briefly, mice received a terminal dose of anesthetic and underwent laparotomy. The portal vein was cannulated and then perfused with 50 mL Liver Perfusion Medium (Thermo Scientific) followed by 50 mL Liver Digest Medium (Thermo Scientific) pre-warmed to 37 °C. The liver was removed and then hepatocytes were mechanically disassociated and released from the Glisson capsule, filtered through 70 μM filter and purified with a density gradient centrifugation. The cells were seeded at a concentration of 5×10⁴ cells/6-well plate on collagen-coated dishes. After 3 hours, the medium was changed, and cultures were collected (t₀) or maintained in DMEM (Thermo Scientific) containing 10% fetal bovine serum (FBS) and treated with SAMe 1mM for 24 hours. The medium was changed to low (5mM, Cat.: 11885084, Thermo Scientific) or high glucose (25mM, Cat.: 11965092, Thermo Scientific) DMEM and hepatocytes were treated with water of fructose 12mM for 18 hours. SAMe-D [17] and OKER [18] cells were cultured in DMEM (Thermo Scientific) containing 10% FBS and 1% penicillin/streptomycin. All cells were cultured at 37°C in a humidified 5% CO₂ incubator.

2.3. Aldolase activity

Aldolase activity in cells and livers was determined using the Aldolase Activity Colorimetric Assay Kit (Cat.: K665, Biovision) according to the manufacture’s guide. Measurement was normalized to total protein.

2.4. Western blotting

For western blotting, 10–30 μg of whole-cell lysates were separated by 10% SDS-PAGE. Blots were probed with antibodies diluted 1:1000 in 5% skim milk against ALDOB (18065-1-AP, Proteintech), PRMT1 (11279-1-AP, Proteintech), MATα1 (ab129176, Abcam), AMPK (2532S Cell signaling), p-AMPK (2535S, Cell signaling), β-actin (A3854 Sigma), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (5174 Cell signaling). Horseradish peroxidase (HRP)-linked secondary antibodies anti-rabbit (7074S Cell Signaling) and anti-mouse (7076S Cell Signaling) were diluted 1:5000 in 5% skim milk. Densitometry analyses were carried out using ImageJ Software.

2.5. Immunoprecipitation (IP) analyses

50–100mg of liver tissues and primary hepatocytes seeded in 10 cm dishes were homogenized using RIPA buffer (ab156034) with a tissue homogenizer and by scraping, respectively. 700 μg of whole liver or cells lysates diluted in 700 μl of IP buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 25 mM NaFl, 1% (vol/vol) Nonidet P-40) with protease inhibitors were pre-cleared by adding 0.7 μg of the appropriate normal IgG together with 20 μl of protein A/G agarose beads (Santa Cruz Biotechnology, Dallas, TX) for 1 hour at 4 °C. The IPs were performed with 3 μg of anti-mono/dimethyl-Arg antibody or anti-ALDOB antibody overnight at 4 °C. The protein complexes bound to the protein A/G agarose beads were washed five times with IP buffer with protease inhibitors. The beads were boiled in sample buffer at 95 °C for 5 minutes and the samples were separated by 10% SDS-PAGE for immunoblot. The gel was transferred to nitrocellulose membrane and probed with antibodies as described above. Trueblot IP Detector Reagent (18-8816-31, Rockland) was used to reduce background. Normal IgG (Cell Signaling) was used as control.

2.6. RT-PCR

Total RNA was isolated with Trizol (Invitrogen). Total RNA (1–2 μg) was reverse transcribed into cDNA using M-MLV Reverse Transcriptase (Invitrogen). Two microliters of RT product were subjected to quantitative real-time PCR analysis. TaqMan probes for Aldolase a (Mm00833172_g1), Aldolase b (Mm00523293_m1), Aldolase c (Mm01298116_g1), Mat1a (Mm00522563_m1), Col1a2 (Mm00483888_m1), and Acta2 (Mm00725412_s1) were purchased from Applied Biosystems. Universal PCR Master Mix was purchased from Bio-Rad. Gapdh was used as housekeeping gene (Mm99999915_g1). The cycle Ct value of the target genes was normalized to that of the housekeeping gene to obtain the delta Ct (ΔCt). The ΔCt obtained was used to find the relative expression of target genes according to the formula: relative expression = 2 −ΔΔCt, where ΔΔCt = ΔCt of target genes in experimental groups −ΔCt of target genes in control group.

2.7. Site-directed mutagenesis and DNA sequencing

For the generation of ALDOB R173L and R173F mutants, the Aldob vector was purchased from Origene (Cat.: MR205585) and the primers from Eurofins Genomics. Primers were as follows: R173L Fwd: 5’-CAATGCTCTGGCTCTCTATGCCAGCATC-3’, R173L Rev: 5’-GATGCTGGCATAGAGAGCCAGAGCATTG-3’, R173F Fwd: 5’-CCAATGCTCTGGCTTTCTATGCCAGCATCTG-3’ and R173F Rev: 5’-CAGATGCTGGCATAGAAAGCCAGAGCATTGG-3’. The quick-change lightning site-directed mutagenesis kit (Agilent Technologies) was used to generate mutants. Mutations were confirmed through sequencing by GENEWIZ.

2.8. In vitro translation

The PURExpress^® In Vitro Protein Synthesis Kit (Cat.: E6800S, New England Biolabs) was used to generate ALDOB WT, R173L and R173F recombinant proteins following the manufacture’s guidelines.

2.9. In vitro methylation assay

Recombinant ALDOB (WT, R173L or R173F) (100nM to 1μM) and recombinant PRTM1 (1 μg, Cat.: enz-365) were mixed in 40ul of SAMe 1mM solution (20mM Tris pH 7.6, 4mM EDTA, 1mM PMSF, 0.5mM DTT). The mix was incubated at 25°C for 3 hours.

2.10. PRMT1 silencing

For PRMT1 knockdown 20 nM siRNA against mouse Prmt1 (SC-41070, Santa Cruz) or equivalent scramble control (SC) were delivered into primary mouse hepatocytes using Lipofectamine RNAiMAX (Life Technologies) for 24 h following the manufacturer’s protocol.

2.11. Bodipy staining

Primary hepatocytes were incubated with BODIPY 493/503 (Molecular Probes) at 1 mg/ml during 30 min prior to fixation (4% paraformaldehyde). Imaging was performed using a fluorescence microscope (BZ-X800, Keyence, Osaka, Japan) and quantification of lipid bodies was performed using ImageJ Software.

2.12. Immunohistochemistry

Mouse liver sections were fixed with 10% neutral formalin and paraffin-embedded, sectioned and stained with hematoxylin and eosin (H&E) and Sirius red by the liver histology core of the University of Southern California research center for liver diseases. To perform Oil Red O staining, frozen tissues embedded in OCT were cryosectioned and stained with 0.2% Oil Red O and counterstained with hematoxylin. For F4/80 staining, paraffin-embedded liver sections were incubated with citrate buffer (ab93678), permeabilized with PBS-Triton 0.5% for 30 min, blocked with hydrogen peroxide for 10 min, then sections were blocked with 5% goat serum in TBS-T (Tween 0.1%) for 10 min and incubated with F4/80 (ab300421, 1:100, 1 h at 37 °C) overnight at 4 °C followed by 30 min incubation with anti-rabbit HRP-conjugate. Colorimetric detection was completed with DAB Chromogen. Slides were counterstained with Hematoxylin (MHS32-1L; Sigma), and finally samples were dehydrated, cleared, and mounted in mounting media. Images (5/sample) were captured using a microscope (BZ-X800, Keyence, Osaka, Japan) and quantified with ImageJ Software.

2.13. p65 immunofluorescence

The tissue sections were deparaffinized, rehydrated and rinsed under running water. Hydrated sections were then boiled with citrate buffer for antigen retrieval. The sections were then blocked and incubated overnight with NF-κB p65 primary antibody (ab16502, Abcam; Dilution 1:100). Next day, sections were rinsed with PBS and incubated with Alexa fluor 488 secondary antibody (ab150081, Abcam) for 1hour at 37°C in dark. The stained sections were then mounted with mounting medium containing DAPI (ab104139, Abcam) and imaged under fluorescent microscope. Positive nuclei were quantified manually using ImageJ.

2.14. Hepatic and serum triglycerides (TGs) measurement

TGs were measured using the Triglyceride Colorimetric Assay Kit from Cayman (Cat.: 10010303) according to the manufacture’s guide. Measurement in liver was normalized to total protein.

2.15. Serum ALT/AST levels

ALT and AST levels in serum were measured using the ALT kit (Cat.: TR71121, Thermo Scientific) and AST kit (Cat.: TR70121, Thermo Scientific).

2.16. Serum glucose levels

Glucose levels in serum were measured using the Glucose Colorimetric Assay Kit from RayBiotech (Cat.: MA-GLU-2) according to the manufacture’s guide.

2.17. Interleukin-6 (IL-6), Interleukin-1β (IL-1β) and tumor necrosis α (TNFα) serum levels

IL-6, IL-1β and TNFα serum levels were measured using the kits E-UNEL-M0070 (Elabscience), E-UNEL-M0064 (Elabscience), and 88-7324-88 (Thermo Fisher Scientific) respectively, following the manufacturer’s instructions.

2.18. Hydroxyproline measurement

The hydroxyproline content in liver was measured using the General Hydroxyproline Assay Kit from MyBioSource (Cat.: MBS8243208) according to the manufacture’s guide.

2.19. Hepatic ATP levels

The levels of ATP were determined using the ATPlite luminescence assay (PerkinElmer) according to the manufacture’s guide. Measurement was normalized to total protein.

2.20. Proximity ligation assay (PLA)

For detection of ALDOB-AMPK interaction using PLA, WT and Mat1a KO mice fed low- and high-fructose diets liver paraffin-embedded tissue sections (n=4) were deparaffinized followed by antigen retrieval using the Universal HIER antigen retrieval reagent (ab208572, Abcam, Cambridge, UK). After permeabilization, tissue sections were processed using duolink PLA probe anti-rabbit PLUS (DUO92002, Millipore Sigma, St. Louis, MO, USA) and duolink PLA probe anti-mouse MINUS (DUO92004, Millipore Sigma, St. Louis, MO, USA) as per manufacturer’s instructions with some modifications. Briefly, tissues were blocked with blocking solution for 1 hour in humidity chamber at 37°C followed by incubation with primary ALDOB (Rabbit, 18065-1-AP, Proteintech) and AMPK (Mouse, 66536-1-Ig, Proteintech) antibodies (1:150) overnight at 4°C in humidity chamber. Then, tissues were washed with wash buffer A and incubated with PLUS as well as MINUS PLA probes (1:250) overnight at 4°C in humidity chamber. Tissues were washed, and protein interaction was detected with duolink in-situ detection reagent kit (red, DUO92008, Millipore Sigma, St. Louis, MO, USA) according to manufacturer’s instructions with slight modifications. Tissues were viewed with confocal microscope (BZ-X800, Keyence, Osaka, Japan) after counterstaining with Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA, USA) and PLA signal was recognized as red fluorescent spots. ImageJ was used to quantify and analyze the PLA fluorescence dots.

2.21. Mass spectrometry (MS)

ALDOB protein intensity and R[173] methylation was assessed using animals and technique described [19–20]. Briefly, spectra were obtained a hyper-methylated data-independent acquisition (DIA) Assay Library through a series of data-dependent acquisition (DDA)-MS acquisitions of SCX fractionated Mat1a KO and Gnmt KO liver lysate. Individual samples were analyzed via OpenSWATH. The total ion current (TIC) associated with the MS2 signal across the chromatogram was calculated for normalization using in-house software. This ‘MS2 Signal’ of each file was used to adjust the transition intensity of each peptide in a corresponding file. To obtain a measure of total protein quantity, all methylated peptides were removed and the remaining unmodified peptides analyzed separately. Normalized transition-level data of the unmodified peptides were subsequently processed using mapDIA (https://pubmed.ncbi.nlm.nih.gov/26381204) to obtain protein and methylated peptide level quantitation. For ALDOB R[173] methylation, DIA-MS data was imported into Skyline (Skyline Daily, version 19.1.1.309), which was then used to visualize and manually validate the co-eluting precursor and fragment trace of a peptide containing R[173] methylation. The raw and processed data for the hyper-methylated DIA Assay Library and ALDOB R[173] methylation is available at Panorama (https://panoramaweb.org/methylation_methods_1.url, Proteome Exchange ID: PXD012621) and the total protein quantitation for the animals used in this study is also available at Panorama (ProteomeXchange ID: PXD022122, https://panoramaweb.org/Larp1.url).

2.22. Statistical analysis

Data are given as mean ± standard error (SEM). Sample data statistical analysis was done using Student t-test for comparison between two groups and one-way ANOVA, followed by Tukey’s test for multiple comparisons. All in vitro experiments were repeated at least three times, and p<0.05 was considered statistical significance. For in vivo comparison of WT and KO mice on high fructose diet, two-way ANOVA, followed by Tukey’s test was used. For mRNA and protein levels, the ratios of genes and proteins to respective housekeeping densitometric values were compared. Densitometry analyses were carried out using Image J.

3. RESULTS

3.1. Low SAMe levels correlate with hypomethylated and more active ALDOB

One of the most striking differences found in the methyl-proteome of Mat1a KO mice livers was a significant drop in ALDOB R173 mono-methylation (Fig. 1A) in the livers of 10-month-old KO mice that had MASH and fibrosis [16] as compared to WT littermates despite a slight increase in the total ALDOB protein level (Fig. 1B). These changes in ALDOB were associated with an increase in hepatic ALDO activity (Fig. 1C), and SAMe treatment for two months, which nearly normalized the liver histology [16], also corrected ALDOB hypomethylation, total protein expression, and ALDO activity (Fig. 1A–C). To further examine if SAMe level regulates ALDOB activity, we measured ALDO activity in SAMe-D (HCC cell line isolated from Mat1a KO that has low SAMe level) [17] and OKER (HCC cell line isolated from Glycine N-methyltransferase KO that has high SAMe level) cells [18]. SAMe-D cells showed much higher ALDO activity as compared to OKER cells (Fig. 1D), supporting our previous data that ALDO activity is inversely correlated with SAMe level.

Figure 1. Low SAMe levels correlate with hypomethylated and more active ALDOB.

(A) Quantification of the MS2 TIC normalized intensity of peptide containing ALDOB R[173] monomethylation in livers from WT littermates, Mat1a KO mice, and Mat1a KO mice treated with SAMe for each DIA-MS acquisition using MAPDIA software. (B) Quantification of the MS2 TIC normalized intensity of ALDOB total protein was determined in the same samples for each DIA-MS acquisition for all unmodified peptides using MAPDIA software. (C) Aldolase activity in livers from WT littermates, Mat1a KO, and Mat1a KO treated with SAMe as above. (D) Aldolase activity in OKER (high SAMe) and SAMe-D (low SAMe) cells. (E) In vitro methylation assay using ALDOB and PRMT1 recombinant proteins. (F) Mouse hepatocytes were treated with siRNAs against PRMT1, subjected to co-IP with anti-mono/dimethyl-Arg antibody and blotted for ALDOB. ALDOB and PRMT1 levels in the inputs are shown. (G-H) Aldolase activity in a recombinant system of ALDOB alone, ALDOB+PRMT1, and ALDOB+PRMT1+SAMe. *p<0.05 Mat1a KO vs. WT; Mat1a KO+SAMe vs. Mat1a KO; siPrmt1 vs. SC; ALDOB+PRMT1+SAMe vs. ALDOB.

ALDOB A174D mutation has been found in 16% of patients with hereditary fructose intolerance (HFI) [21], a disease associated with MASH and death of liver cells [22]. In HFI, ALDOB mutants such as A174D and L256P cannot form homotetramers because these residues are embedded in the subunits interface and have greatly impaired enzymatic activity [23]. R173 is also buried in the subunits interface of ALDOB [24] and, therefore, would not be accessible by any of the PRMTs unless the methylation reaction occurs on the monomeric form of the enzyme. We next examined which PRMT can methylate ALDOB and what the consequences are in terms of ALDOB structure and function. There are nine different PRMTs in mammals (PRMT1-9) [25]; PRMT1 can catalyze monomethylation and asymmetric dimethylation and is responsible for ~ 85% of protein arginine methylation activity in mouse liver [26, 27]. We found that recombinant PRMT1 was able to methylate recombinant ALDOB (Fig. 1E). To see if PRMT1 is responsible for ALDOB methylation in vitro we next treated primary mouse hepatocytes with a siRNA against Prmt1. We found that Prmt1 silencing in normal mouse hepatocytes resulted in almost 50% drop in the methyl-ALDOB:total ALDOB ratio (Fig. 1F). Moreover, we observed using recombinant ALDOB and PRMT1 proteins, that the addition of SAMe to the incubation buffer significantly inhibited ALDOB activity (Fig. 1G–H), supporting the hypothesis that ALDOB methylation alters its enzymatic activity and PRMT1 is the methyltransferase responsible for it.

3.2. ALDOB methylation at R173 and R304 influences its oligomerization and enzymatic activity

We next evaluated whether PRMT1 methylates ALDOB at R173. We performed site-directed mutagenesis and generated methylation (R173L) and methyl-mimetic (R173F) ALDOB mutants (Fig. S1). Using an in vitro translation system that allowed generation of recombinant proteins ALDOB R173L and R173F, we performed methylation in vitro assays with recombinant PRMT1 and SAMe, including WT ALDOB as positive control. As shown in Figure 2A, we found that PRMT1 methylated all three, which suggests that there are other arginine residues in ALDOB besides R173 that are methylated by PRMT1. Interestingly, we observed that the ALDOB R173F methyl-mimetic had about 20–30% lower activity as compared to ALDOB WT or the R173L methyl-mutant (Fig. 2B). In addition, supporting that methylation at R173 has a negative effect on ALDOB activity, the incubation of ALDOB R173F methyl-mimetic with PRMT1 and SAMe had no additional significant effect on its activity (Fig. 2C).

Figure 2. ALDOB methylation at R173 and R304 regulates its oligomerization and enzymatic activity.

(A) In vitro methylation assay using ALDOB WT, R173L methylmutant and R173F methylmimetic and PRMT1 recombinant proteins. (B) Aldolase activity in recombinant ALDOB WT, R173L and R173F and (C) in a recombinant system of R173F alone, R173F+PRMT1, and R173F+PRMT1+SAMe. (D) Aldolase activity in recombinant ALDOB WT and R304F and (E) in a recombinant system of R304F alone, R304F+PRMT1, and R304F+PRMT1+SAMe. (F) Tetramer-dimer equilibrium in recombinant ALDOB WT, R173L and R173F and (G) R304F. (H) ALDOB exists as a dimer and tetramer in equilibrium and methylation at R173 and R304 by PRMT1 favors a dimeric and less active form. *p<0.05 vs. WT.

To confirm ALDOB R173 methylation by PRMT1, we employed in vitro methylation of recombinant ALDOB in the presence of PRMT1 and SAMe, followed by methyl-proteomics. Although we were not able to identify R173 using this method, this led us to identify R304, a second ALDOB methylation site. The explanation why R173 was not picked up in the methyl-proteomics from the in vitro methylation experiment may be due to the oligomeric structure of the recombinant ALDOB, making R173 inaccessible to PRMT1. To study how methylation at R304 site affects ALDOB activity, we generated a methyl-mimetic for R304 (R304F) and like R173F mutant, we observed that R304F mutant had lower enzymatic activity (Fig. 2D). Supporting this, similar to R173F, the incubation of ALDOB R304F with PRMT1 and SAMe had no effect on its activity (Fig. 2E).

Using a dimer-tetramer equilibrium assay we also confirmed that ALDOB exists as a dimer and tetramer in equilibrium and that both the R173F and R304F mutants favors the dimeric form (Fig. 2F–G). Taken together, we found that methylation of R173 and R304 results in less tetramer form and as such, lower enzymatic activity of ALDOB (Fig. 2H).

3.3. Methylation of ALDOB influences its sensing towards AMPK

To further examine if SAMe level regulates ALDOB activity and ALDOB-regulated AMPK activity, we measured ALDO activity in cultured mouse hepatocytes for up to 24 hours, an in vitro model where SAMe level drops because of a fall in MAT1A expression and ALDOB levels remain unchanged (Fig. 3A). Correlating with lower SAMe levels, we found a lower methyl-ALDOB:total ALDOB ratio in hepatocytes cultured for 24 hours that was normalized when SAMe was added to the culture media (Fig. 3A). We measured ALDO activity at different culture time points up to 24 hours and found a significant increase with time, which was reduced when the hepatocytes were treated with SAMe (Fig. 3B), condition where me-ALDOB was normalized (Fig. 3A). Interestingly, ALDOB-AMPK signaling, as indicated by the interaction of ALDOB with AMPK, and AMPK activity increased as SAMe level fell but these were blunted by SAMe treatment (Fig. 3C). Taken together, these results show that SAMe level regulates ALDOB methylation, enzymatic activity, and its function as sensor to activate AMPK.

Figure 3. SAMe level regulates ALDOB-AMPK signaling and fructose inhibits AMPK activation and exacerbates lipid accumulation under low glucose.

(A) ALDOB methylation was evaluated in lysates of hepatocytes cultured for 0 hour (h), 24h and 24h+SAMe by co-IP using an anti-mono/dimethyl-Arg antibody. Densitometric values are shown on the right. (B) ALDO activity in hepatocytes cultured for 0h, 3h, 12h, 24h and 24h+SAMe. (C) ALDOB-AMPK interaction was evaluated in lysates of hepatocytes cultured for 0h, 24h and 24h+SAMe by co-IP using an ALDOB antibody. Densitometric values are shown on the right. Primary mouse hepatocytes were cultured for 18h under low or high glucose media ± fructose. (D) Western blot and densitometry analysis of AMPK activation, E) BODIPY staining, and (F) quantification of lipid bodies. *p<0.05 3h, 12h, and 24h vs. 0h; 24h+SAMe vs 24h; high glucose vs. low glucose; and fructose vs. no fructose.

3.4. Fructose inhibits AMPK activation and exacerbates lipid accumulation

Our data suggest that in the liver, SAMe level and ALDOB regulate liver AMPK activity, with AMPK activated when SAMe level is reduced. We hypothesize that in Mat1a KO mice, AMPK activation is an adaptive response to prevent further FA accumulation, generation of lipotoxic species and MASH progression. High fructose can lead to an increase in FBP in the liver because fructose metabolism generates excess DHAP and G3P, which can be converted to FBP by ALDOB, blocking AMPK activation and accelerating the progression to MASH. ALDOA was reported to be induced in a rat fatty liver model [28], but ALDOB remains the dominant form expressed (AldoA and AldoC mRNA levels are 5% and 1% of AldoB, respectively, and are unaffected by Mat1a deletion or SAMe treatment (Fig. S2)). Therefore, our hypothesized mechanism may be unique for the liver, the main organ where de novo lipogenesis, FA oxidation and gluconeogenesis take place. In support of this hypothesis, hepatocytes cultured under low glucose condition show AMPK activation, but this was blocked by fructose addition (Fig. 3D), which also increased lipid accumulation, similar to high glucose (Fig. 3E–F). This result supports the hypothesis that FBP, independent of glucose availability and cellular energy status, regulates AMPK activity in the liver.

Lastly, we tested the response of 6-month-old Mat1a KO mice (prior to development of overt fatty liver) to high-fructose diet for two months as compared to WT mice. Mat1a KO mice developed more liver injury as compared to WT or low fructose diet (Fig. 4A–F). The food intake was comparable between the two groups and neither WT nor Mat1a KO mice gained weight with the high fructose diet (Fig. S3A). However, the livers of Mat1a KO mice increased significantly after high fructose feeding (Fig. S3B–C). Mat1a KO mice accumulated more fat in the liver, as measured by oil red o staining (Fig. 4AB) and triglyceride levels (TGs) (Fig. 4C). The levels of TGs in serum were slightly higher in Mat1a KO mice and following the high fructose diet, but the difference was not statistically significant (Fig. S3D). On the other hand, the levels of glucose in serum were unaltered (Fig. S3E). In correlation with higher levels of fat in the liver, Mat1a KO mice had higher ALT (Fig. 4D) and AST levels (Fig. 4E). Although F4/80 staining, p65 nuclear translocation (Fig. 4A–B) and serum TNFα and IL-6 levels (Fig. S3F–G) suggested the same degree of inflammation, IL-1β in serum was higher in Mat1a KO mice (Fig. S3H). On top of that, Mat1a KO mice showed signs of fibrogenesis not present in WT mice, as seen by Sirius red staining (Fig. 4A–B), hydroxyproline hepatic levels (Fig. 4F), and the induction of Col1a2 and Acta2 mRNA levels (Fig. S3I).

Figure 4. Mat1a KO mice are more sensitive to high-fructose feeding.

(A) Representative images of hematoxylin and eosin (H&E), Oil red o, Sirius red, and F4/80 immunostaining and p65 immunofluorescence of liver tissues of WT and Mat1a KO mice fed a low or a high-fructose diet for 8 weeks. (B) Quantitative analyses of the staining. (C) Liver TGs levels, (D) ALT, (E) AST, and (F) hydroxyproline content in liver. (G) Western blot and densitometric analyses of p-AMPK/AMPK, ALDOB and MATα1 in liver lysates of WT and Mat1a KO mice fed a low or a high-fructose diet. (H) Liver sections were stained with ALDOB and AMPK followed by secondary antibody-proximity ligation assay (PLA) probe conjugates and detected using red duolink detection reagent kit. A representative area is shown. Each tissue was quantified using ImageJ and represented as the PLA fluorescence (% area). *p<0.05 high fructose vs. low fructose; Mat1a KO vs. WT. In G: *p<0.05 high fructose vs. low fructose and ^#p<0.05 Mat1a KO vs. WT.

As previously observed and correlating with lower SAMe level and me-ALDOB, Mat1a KO mice had higher basal AMPK activity in the liver, which was strongly inhibited by high fructose (Fig. 4G). ALDOB expression was equally induced in both WT and Mat1a KO mice on the high-fructose diet (Fig. 4G and Fig. S3I). However, using the proximity ligation assay (PLA), we found higher interaction between ALDOB and AMPK (Fig. 4H), similar to what we observed in vitro in a condition of lower SAMe level (Fig. 3C), suggesting that ALDOB-AMPK signaling is hyperactivated in Mat1a KO mice. This higher interaction and pathway were inhibited when these mice were put on a high-fructose diet (Fig. 4H), via higher FBP, which worsened the progression of the disease. Supporting this pathway as the main mechanism by which fructose worsens the progression of MASLD, we observed a significant reduction in the levels of ATP in the livers of WT mice fed a high-fructose diet. On the other hand, Mat1a KO mice showed lower hepatic ATP levels, which correlate with higher p-AMPK at baseline compared to WT mice, and these levels remained unchanged upon high-fructose feeding (Fig. S3J).

In summary, our work unveiled a novel regulatory mechanism of ALDOB function that impacts on AMPK signaling. Low hepatic SAMe level results in hypomethylated ALDOB at R173 and R304, which leads to a more tetrameric form, with higher enzymatic activity and higher capacity to signal to activate AMPK. Mat1a KO mice utilize this mechanism as an adaptive response to offset lipid accumulation. Nearly half of the patients with MASLD have a metabolomics profile similar to that of Mat1a KO mice [16], and based on this mechanism, fructose consumption may be particularly harmful to them.

4. DISCUSSION

Fructose is primarily metabolized in the liver and when consumed in excess, it can contribute to liver steatosis and inflammation, key features of MASLD [1–3]. High fructose intake has been linked to insulin resistance, dyslipidemia, and increased visceral adiposity [29], all of which are risk factors for MASLD development and progression. Although recent publications have elucidated several molecular mechanisms by which fructose accelerates the disease, the exact process is still incompletely defined.

One key mechanism involves the rapid metabolism of fructose in the liver, which leads to increased de novo lipogenesis (DNL), resulting in the accumulation of TGs in hepatocytes [30–31]. Fructose can also disrupt hepatic ATP production and promote the production of advanced glycation end products and reactive oxygen species [32–33], exacerbating liver damage and fibrosis [34]. Fructose metabolism generates intermediate metabolites such as diacylglycerol and acylcarnitines, which can activate protein kinase C and induce mitochondrial dysfunction, contributing to insulin resistance and hepatic lipid accumulation [35–36]. Moreover, fructose generates uric acid, which has been linked to oxidative stress, inflammation, and mitochondrial dysfunction in the liver [32]. Emerging evidence suggests that fructose metabolism may also disrupt gut barrier function and alter the composition of the gut microbiota, leading to increased gut permeability and bacterial translocation [37], which can further exacerbate liver inflammation and injury. Overall, these molecular mechanisms underscore the multifaceted role of fructose in the pathogenesis of MASLD. Novel therapeutic approaches targeting fructose metabolism, such as inhibitors of fructose-metabolizing enzymes or modulation of gut microbiota composition have shown promise in preclinical studies for the treatment of MASLD [37–38]. However, reducing fructose intake seems to be crucial in the prevention and management of MASLD, emphasizing the importance of dietary strategies in mitigating this increasingly prevalent liver condition.

FBP, a key intermediate in the metabolism of fructose, represents a potential molecular target for MASLD. It plays a central role in both glycolysis and gluconeogenesis [39] and can contribute to DNL, inflammation, and oxidative stress in the liver. ALDOB catalyzes the cleavage of FBP [9] and a recent publication has described a new role for ALDOB, acting as a sensor of AMPK by a mechanism that involves FBP. When FBP levels are low, ALDOB is not occupied and can form a lysosomal complex with the v-ATPase that leads to AMPK activation. On the contrary, when FBP levels are elevated –under high glucose or fructose conditions– ALDOB is occupied by FBP and not able to form that complex to activate AMPK [11]. Supporting this finding, the same group developed an ALDOB-FBP specific inhibitor that activates AMPK and protected the liver against MASLD [13]. Although DNL has been suggested as the main mechanism by which fructose intake accelerates MASLD, other recent publications suggest that a decrease in β-oxidation is another major mechanism. It is well known that higher AMPK activity enhances mitochondrial FA oxidation [40] but fructose itself has been found to suppress the expression of genes involved in FA oxidation and modify mitochondrial proteins related to lipid metabolism [41]. Overall, suppression of AMPK activity seems to be a key mechanism by which fructose may accelerate the progression of MASLD.

ALDOB deficiency and mutations, such as A149P, A174D or L256P, that negatively impact its activity have been identified as the underlying cause of HFI, a rare genetic disorder characterized by the inability to properly metabolize fructose [9]. Although these findings underscored ALDOB’s critical role in fructose metabolism decades ago, how ALDOB protein is regulated is still poorly understood and to the best of our knowledge, ALDOB has not been shown to be regulated by methylation.

While studying the methyl-proteome of Mat1a KO mice, which have chronic hepatic SAMe deficiency and spontaneously develop MASH [14], we found that ALDOB is hypomethylated at R173 in KO mouse livers, which was normalized when SAMe was administered for 2 months, treatment that improved the phenotype [16]. Further studies using recombinant ALDOB and mass spectrometry (MS) allowed us to identify R304 as another methylation site in ALDOB. Mutation at A174 is one of the most commonly observed genetic variants found in HFI patients and alters the structure and function of ALDOB, resulting in impaired enzymatic activity [42]. Although less common, R304W is considered a severe mutation that results in a complete loss of ALDOB activity [43]. Since R173 is buried in the subunits interface of ALDOB and R304 sits within the active site pocket, we hypothesized that their methylation could affect ALDOB oligomerization and as a result, its activity. Using WT and Mat1a KO mice, cell lines that express low or high SAMe, and recombinant systems, we found that ALDOB activity is negatively regulated by SAMe and that its methylation regulates its oligomeric state. While ALDO activity was increased in the livers of Mat1a KO mice and the cell line with low SAMe levels, its activity was reduced in the livers Mat1a KO mice after SAMe treatment and the cell line with high SAMe levels. Consistently, the methyl-mimetic mutants R173F and R304F showed lower activity as compared to WT ALDOB. Interestingly, using a dimer-tetramer equilibrium assay we confirmed that ALDOB exists as a dimer and tetramer in equilibrium and that methylation at R173 and R304 favored the dimeric form. Our data supports that the oligomeric state of ALDOB, which is modulated by methylation at two different residues, correlates with its activity.

Because ALDOB has been recently identified as a new AMPK sensor, we wondered whether its methylation –oligomeric state and activity– could have an impact on its sensing towards AMPK. Primary hepatocytes have lower MAT1A expression and SAMe level, and consequently lower me-ALDOB, after 24 hours in culture. We found higher ALDO activity, ALDOB-AMPK complex, and AMPK activity under this condition, which were all normalized by SAMe administration. Similarly, we found that AMPK was activated and ALDOB-AMPK interaction enhanced in Mat1a KO livers. Although further experiments would be needed to prove that ALDOB methylation regulates its ability to sense FBP, our data support this hypothesis.

AMPK is inhibited in MASLD and multiple studies have shown the beneficial effect of its activation in the progression of the disease [44]. How fructose metabolism influences AMPK activity is not well understood and although no study has shown that high-fructose downregulates AMPK, its activation in the liver prevented steatosis in mice on a high-fructose diet [45]. Here, we observed that high fructose inhibited AMPK in the liver both in vitro and in vivo. Although the inhibition of AMPK may occur at least in part due to alterations in cellular energy status, we hypothesized that FBP levels and ALDOB-AMPK signaling could be also involved. In fact, we found that fructose inhibited AMPK in hepatocytes regardless of the glucose level, supporting that FBP, independent of glucose availability and cellular energy status, regulates AMPK activity in the liver.

Lastly, aiming to evaluate the link between ALDOB methylation status and ALDOB-AMPK signaling in vivo, we fed WT and Mat1a KO mice with a low or a high-fructose diet for 8 weeks. As compared to WT mice, Mat1a KO mice showed more liver injury and developed more fatty liver and fibrogenesis. As discussed earlier, Mat1a KO mice showed higher AMPK activation under a low-fructose diet, which we hypothesize is a compensatory mechanism to overcome their spontaneous phenotype and lower hepatic ATP levels. Interestingly, both ALDOB-AMPK interaction and AMPK activity were blunted in these mice under high-fructose diet, correlating with a more severe phenotype. This result is of high importance as the metabolic phenotype of these mice resembles to 50% of MASLD human patients [16] and could explain why the impact of fructose on the acceleration of MASLD can vary among individuals.

In summary, we have shown that ALDOB methylation at R173 and R304 affects its oligomerization and activity. Mat1a KO mice livers have hypomethylated ALDOB, which leads to a more active form of the enzyme that senses FBP and is more susceptible to AMPK inhibition. These findings represent a mechanism through which fructose may contribute to the development and progression of MASLD under conditions of low SAMe level, paving the way for the development of new strategic approaches. A summary scheme is shown in Fig. S4.

Highlights:

• PRMT1 methylates ALDOB at R173 and R304 favoring a more dimeric and inactive form

• Hepatic SAMe levels inversely correlate with ALDOB activity and sensing towards AMPK

• Fructose, independently of glucose, inhibits AMPK and induces hepatic steatosis

• AMPK activity and ALDOB-AMPK interaction are enhanced in Mat1a KO livers

• Fructose inhibits ALDOB-AMPK interaction and enhances MASLD progression in Mat1a KO mice

ABBREVIATIONS

ALDO

Aldolases

ALDOB

Aldolase B

ALT

alanine transaminase

AST

aspartate transferase

AMPK

AMP-activated protein kinase

AMP

Adenosine monophosphate

ATP

adenosine triphosphate

DDA

data-dependent acquisition

DHAP

dihydroxyacetone phosphate

DIA

data-independent acquisition

DNL

de novo lipogenesis

FA

fatty acid

FBP

fructose 1,6-biphosphate

FBS

fetal bovine serum

F1P

fructose 1-phosphate

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

G3P

glyceraldehyde 3-phosphate

GTP

guanosine triphosphate

H&E

hematoxylin and eosin

HCC

hepatocellular carcinoma

HFI

hereditary fructose intolerance

HRP

horseradish peroxidase

IP

immunoprecipitation

KO

knockout

MAT1A

Methionine adenosyltransferase alpha 1

MASH

metabolic dysfunction-associated steatohepatitis

MASLD

metabolic dysfunction-associated steatotic liver disease

MS

mass spectrometry

PLA

proximity ligation assay

PRMT1

protein arginine methyltransferase 1

SAMe

S-adenosylmethionine

SEM

standard error

SC

scramble control

TGs

triglycerides

TIC

total ion current

WT

wild type