IntroductionO-linked N-acetylglucosamine transferase (OGT) catalyzes the addition of a single monosaccharide O-linked N-acetylglucosamine to serine and threonine residues of numerous cytosolic and nuclear proteins1,2,3. This enzymatic activity of OGT — called O-GlcNAcylation — is achieved through a complex of the enzyme O-GlcNAc transferase and the sugar nucleotide donor co-substrate UDP-N-acetylglucosamine (UDP-GlcNAc). O-GlcNAcylation is a dynamic protein post-translational modification (PTM), which is reversed through the action of a glycosidase enzyme called O-GlcNAcase (OGA). As UDP-GlcNAc is synthesized through the hexosamine biosynthetic pathway (HBP), which directly uses glucose and glutamine, two extracellular nutrient inputs, O-GlcNAcylation is considered a nutrient-sensing PTM.Host cell factor-1 (HCF-1) is a transcriptional co-factor that serves as an “adaptor” protein to connect DNA-sequence-specific transcription factors with numerous chromatin-modifying enzymes4,5. HCF-1 is critical in cell cycle regulation during the G1/S phase transition and for proper progression through the mitosis6. Several O-GlcNAc-modified serine/threonine sites have been mapped on HCF-1, suggesting HCF-1’s activity can be regulated, making HCF-1 a potential target of OGT regulatory activities5,7,8,9. In addition to glycosyltransferase activity, OGT also mediates HCF-1 proteolytic maturation. It specifically cleaves HCF-1 at six highly conserved 26 amino acid repeat sequences known as HCF-1PRO repeats10,11. Therefore, OGT represents a unique intersection of glycosyltransferase and protease activities, both regulated by HCF-1.Both OGT and HCF-1 have essential roles in metabolic processes. As a nutrient sensor, OGT has been shown to interact with proteins involved with the regulation of cellular metabolism, such as carbohydrate-responsive element-binding protein (ChREBP)12, PGC1α13 as well as glycolytic enzymes such as PFK114. Alterations in O-GlcNAc signaling are linked to diseases like diabetes. Moreover, OGT has been shown to link the hexosamine biosynthesis pathway to oncogenic signaling, impacting glucose and lipid metabolism. This connection aligns with the phenomenon of metabolic reprogramming in cancer, where aberrant glucose metabolism (e.g., the Warburg effect) and epigenetic changes support rapid growth and proliferation15,16. As a prominent substrate and interacting partner of OGT and as a critical component of active transcriptional hubs17, HCF-1 has been implicated in regulating lipogenesis12, gluconeogenesis13, non-alcoholic fatty liver disease (NAFLD) progression18 as well as cobalamin metabolism disorders19.The liver is the second largest organ in humans and is vital for maintaining nutrient homeostasis20. Hepatocytes, which comprise 80% of the liver mass, majorly regulate metabolism, detoxification, and bile synthesis21. Owing to OGT’s role in metabolic processes, we here investigated O-GlcNAcylation perturbations in liver hepatocytes after an induced loss of HCF-1 in these cells. We show that hepatocyte-specific loss of HCF-1 leads to a parallel decline in levels of enzyme OGT and its O-GlcNAcylation activity. Under fasted conditions, we show a rapid nuclear-to-cytoplasmic redistribution of OGT and O-GlcNAcylated proteins, which mimics the observed changes upon loss of HCF-1. These findings suggest that HCF-1 plays a crucial role in maintaining the stability and localization of OGT, impacting cellular metabolic responses. Given the indispensable role of OGT in regulating critical metabolic pathways, these alterations may have profound effects on liver function, offering new insights into how cellular stress conditions, such as fasting, influence metabolic regulation.ResultsWe utilized a previously described hepatocyte-specific HCF-1 knockout model18 in which tamoxifen-dependent Cre-mediated recombination deletes Hcfc1 exons 2 and 3 in hepatocytes of mice carrying the Albumin-Cre-ERT2tg transgene. This generates complete hepatocyte-specific knockout (Hcfc1hepKO allele) in Hcfc1hepKO/Y male mice, while heterozygous Hcfc1hepKO/+ females display mosaic patches of HCF-1-positive and HCF-1-negative hepatocytes due to X-inactivation22. Control mice lacking the Albumin-Cre-ERT2tg transgene were treated identically. While this HCF-1 hepatocyte-specific knockout model was previously shown to develop NAFLD-like features18, the current study focuses specifically on elucidating the molecular mechanisms underlying HCF-1's regulation of OGT function and O-GlcNAcylation in hepatocytes.To confirm hepatocyte-specific HCF-1 depletion, we performed co-immunostaining for HCF-1 and the hepatocyte marker HNF4α23 on liver sections from control Hcfc1lox/Y and Hcfc1hepKO/Y males seven days post-tamoxifen injection. As previously reported18, targeted HCF-1 loss was observed specifically in HNF4α-positive hepatocytes (white arrowheads) in Hcfc1hepKO/Y mice (compare Fig. 1A to B), while HNF4α-negative non-hepatocytes (open arrowheads) retained normal HCF-1 levels.Fig. 1Loss of HCF-1 affects OGT and O-GlcNAcylation levels. (A and B). Immunofluorescence analysis of paraffin-embedded sections from control (0d) liver (A) and Alb-Cre-ERT2tg; Hcfc1hepKO/Y (B) livers seven days post-tamoxifen treatment. Sections were stained with 4’,6-diamidino-2-phenylindole (DAPI) (blue), together with antibodies against HCF-1 (green) and HNF4α (red). Scale bars, 50 μm. (C). Immunoblot analysis from liver lysates isolated from two (technical replicates #1 and #2) Alb-Cre-ERT2tg; Hcfc1hepKO/Y male mice at 0 day (lanes 1 and 3) and 5 d post-tamoxifen injection (lanes 2 and 4). OGT panel: OGT protein levels; HCF-1 panel: HCF-1 protein levels showing full-length HCF-1 (*) and mature C-terminal HCF-1 cleavage fragments (.); O-GlcNAc panel: O-GlcNAcylated proteins (RL2 antibody). Anti-U2AF65 was used as a loading control. Results confirm the decrease in HCF-1, O-GlcNAcylation, and OGT levels upon hepatocyte-specific loss of HCF-1.Full size imageLoss of HCF-1 affects OGT and O-GlcNAcylation levels in the murine liverA previous study by Daou et al. reported a decrease in OGT protein levels after siRNA-mediated knockdown of HCF-1 in HeLa cells8. We investigated whether a similar effect can be observed upon hepatocyte-specific loss of HCF-1 in Hcfc1hepKO/Y mice through a post tamoxifen injection time course. Identically treated Hcfc1lox/Y males lacking the Albumin-Cre-ERT2tg transgene served as controls. HCF-1 is initially formed as a precursor protein of 2035 amino acids. OGT-mediated proteolytic activity results in the formation of two subunits – one from the N-terminus and one from the C-terminus. Consistent with the long half-life of HCF-1 protein24, tamoxifen treatment induced a progressive loss of HCF-1 protein, which can be seen in whole-liver lysates (probed with a c-terminal specific anti-HCF-1 antibody) from day 2 onwards, with a significant loss by day 5 (Fig. 1C, HCF-1 panel, Supplementary Fig. 1). Probing the same lysates with an anti-OGT antibody showed a reduction in OGT protein levels upon loss of HCF-1 by day 5 (Fig. 1C, OGT panel, Supplementary Fig. 1). These lysates were also probed with anti-O-GlcNAc antibody (RL2) to detect any changes in global protein O-GlcNAcylation levels post-tamoxifen treatment. Correlated with the observed decrease in enzyme OGT levels, we observe a reduction in global O-GlcNAcylation in Hcfc1hepKO/Y lysates (Fig. 1C, O-GlcNAc panel, Supplementary Fig. 1). For quantification refer to Supplementary Fig. 2.Loss of HCF-1 affects OGT and O-GlcNAcylation levels in pure hepatocyte populationsHepatocytes contribute to 60% by cell number and 70–85% of liver cell mass20; other cells, such as endothelial cells, Kupffer cells, and stellate cells, make up the rest. The results described above were obtained with whole-liver lysates, where the protein levels detected through immunoblotting have a significant contribution from non-hepatocytes where the Hcfc1 allele has not been targeted. To determine the precise impact of HCF-1-protein loss in hepatocytes, we isolated primary murine hepatocytes from tamoxifen-injected control Hcfc1lox/Y and Hcfc1hepKO/Y male mice (hereby referred to as HCF-1 WT hepatocytes and HCF-1 KO hepatocytes, respectively), using a two-step collagenase digestion process as previously described25. As shown in Fig. 2A, compared to HCF-1 WT hepatocytes, we observed a significant loss of HCF-1 proteins in HCF-1 KO hepatocytes on day 5 and an almost 75% reduction in HCF-1 intensity by day 7 (Fig. 2A, HCF-1 panel, lanes 1–3). Next, we probed these lysates for changes in levels of OGT and overall O-GlcNAcylated proteins. We observed a substantial decrease in the abundance of OGT and O-GlcNAcylated proteins after the loss of HCF-1 protein (Fig. 2A, OGT and O-GlcNAc panel).Fig. 2Loss of OGT and O-GlcNAcylation levels in HCF-1 negative hepatocytes (A). Immunoblot analysis of purified hepatocytes isolated from control (5d post-tamoxifen; lane 1) and Alb-Cre-ERT2tg; Hcfc1hepKO/Y male mice (5 and 7 d post-tamoxifen injection; lanes 2 and 3) for HCF-1, O-GlcNAcylation, and OGT levels. Anti-LSD2 antibody was used to detect levels of E3 ligase LSD2. Anti-actin was used as a loading control. Full-length HCF-1 (*); mature C-terminal HCF-1 cleavage fragments (.) and non-specific band in OGT (#). (B). The total OGT, LSD2 (KDM1B), and XIAP transcript expression levels in control Hcfc1lox/Y (blue), Alb-Cre-ERT2tg; Hcfc1hepKO/+ (orange), and Alb-Cre-ERT2tg; Hcfc1hepKO/Y (red) livers. Graphs display RPKM reads per kilobase of transcript per million mapped reads. (C). Phase contrast micrographs of isolated hepatocytes kept in culture conditions for 4–8 h (HCF-1 WT hepatocytes) and 8 h (HCF-1 KO hepatocytes). The scale bar represents 200 μM.Full size imageTo investigate the cause for the decrease in OGT-protein levels, we first checked for OGT encoding mRNA levels in our previously described RNA-Seq dataset18. We found no reduction in OGT mRNA levels upon loss of HCF-1 (Fig. 2B, upper panel) — suggesting a post-translational control of OGT protein stability through HCF-1. As a control, we checked for the mRNA levels of Mmachc and Mdh1 that exhibited altered trends (Supplementary Fig. 3 A and 3B). Recently, two E3 ligases — LSD2 (also known as KDM1B)26 and XIAP27, were shown to regulate OGT’s stability via proteasomal pathway-induced degradation. We investigated whether the loss of HCF-1 leads to any changes in LSD2 and XIAP mRNA levels. We observed no significant change in the levels of either transcript in our RNA-Seq data (Fig. 2B, LSD2 (Kdm1b) and Xiap panels). We also probed for LSD2 protein levels in HCF-1 KO hepatocyte lysates and found no change upon loss of HCF-1 (Fig. 2A).Overall, the results obtained through pure hepatocyte isolation showed a more significant decrease in the levels of HCF-1, OGT, and global O-GlcNAcylated proteins compared to whole liver lysate preparations described in Fig. 1, underscoring the striking impact of HCF-1 loss on OGT protein and its enzymatic activity in murine hepatocytes. Interestingly, when plated in DMEM plating media, the purified HCF-1 KO hepatocytes failed to adhere to plates properly after 8 h (Fig. 2C). This was in contrast to the adherence exhibited by HCF-1 WT hepatocytes isolated from control Hcfc1lox/Y mice (Fig. 2C). These results suggest that loss of HCF-1 impacts the cell adherence functions of hepatocytes.HCF-1-negative hepatocytes selectively display loss of nuclear OGT levelsTo visually detect the changes in the levels and sub-cellular localization of OGT and O-GlcNAcylated protein after the loss of HCF-1, we performed immunofluorescence with paraffin-embedded liver sections from the control Hcfc1lox/Y and HCF-1 deficient Hcfc1hepKO/Y mice. We immunostained these sections using anti-OGT, anti-O-GlcNAc (RL2), and anti-HCF-1 antibodies. In control Hcfc1lox/Y mice, OGT was found to be nucleocytoplasmic in hepatocytes (Fig. 3A). Co-staining with RL2 antibody for O-GlcNAcylated protein showed a predominantly dense nuclear localization, with some diffuse cytoplasm presence (Fig. 3A). Isolated pure hepatocytes from control Hcfc1lox/Y mice also showed a similar staining pattern for HCF-1 (nuclear), OGT (nucleocytoplasmic), and O-GlcNAcylated protein (prominently nuclear) (Supplementary Fig. 4 A and B).Fig. 3Changes in sub-cellular localization of OGT and O-GlcNAcylated proteins after the loss of HCF-1 (A and B). Immunofluorescence analysis of paraffin-embedded sections from liver (A) control Alb-Cre-ERT2tg; Hcfc1lox/Y mouse without tamoxifen treatment (A) and (B) Alb-Cre-ERT2tg; Hcfc1hepKO/Y mouse seven days after tamoxifen treatment stained with DAPI (blue), anti-O-GlcNAc RL2 (red), and anti-OGT (green) antibodies. Scale bars, 50 μm. (C). Immunofluorescence analysis of paraffin-embedded liver sections from Alb-Cre-ERT2tg; Hcfc1hepKO/+ mouse seven days after tamoxifen treatment stained with DAPI (blue) as well as anti-O-GlcNAc RL2 (red) and anti-HCF-1 (green) antibodies. HCF-1-positive hepatocytes (marked zone) display nuclear O-GlcNAc staining. Scale bars, 50 μm. (D). Magnified image of HCF1 positive cells (pink square) and HCF1 negative cells (blue square) stained with anti-O-GlcNAc RL2 (red) and anti-HCF-1 (green) antibodies. Scale bar, 20 μm.Full size imageIn contrast, Hcfc1hepKO/Y liver sections showed a marked reduction in the presence of nuclear OGT, while cytoplasmic OGT levels appeared unchanged (Fig. 3B). This result agrees with previously described observations from siRNA-mediated depletion of HCF-1 in HeLa cells, where nuclear OGT levels had decreased in cells lacking HCF-1 protein8. In contrast to control Hcfc1lox/Y liver sections, Hcfc1hepKO/Y liver sections showed a marked disappearance of nuclear O-GlcNAcylation, while only a diffused cytosolic staining could be seen (Fig. 3B). As isolated pure hepatocytes from Hcfc1hepKO mice had adhesion problems; they could not be grown over coverslips to stain them for HCF-1, OGT, and glycosylated protein to make a direct comparison vis-a-vis control hepatocytes grown on coverslips.In Hcfc1hepKO/+ female mice, owing to the random inactivation of the X-linked Hcfc1 gene22, approximately half the hepatocytes in the liver lose HCF-1 after Cre-recombinase induction post-tamoxifen injection18. Using tissue sections from Hcfc1hepKO/+ female mouse livers, we investigated further the impact of such a cell-selective loss of HCF-1 on OGT’s enzymatic O-GlcNAcylation activity. HCF-1 positive hepatocytes also stained positive for dense nuclear O-GlcNAcylation (Fig. 3C and Fig. 3D panel a; pink square), while HCF-1 negative hepatocytes stained for cytoplasmic O-GlcNAcylation (Fig. 3C and Fig. 3D panel b; blue square).OGT nuclear localization is sensitive to the nutrient state of the cellOGT is considered a nutrient-sensing enzyme (see introduction), and O-GlcNAcylation levels are highly responsive to changes in nutrient conditions15,28,29. Since the liver has an essential function in metabolic process regulation20, we investigated the impact of nutrient availability on OGT and its enzymatic activity (O-GlcNAcylation) in liver cells. We prepared liver sections of C57BL/6 mice (12–14 weeks old), which were either fed ad-libitum or subjected to a fasted state for 13-h and immune-stained them for O-GlcNAcylated proteins and OGT. As shown in Fig. 4A, in ad-libitum-fed mice, O-GlcNAcylated proteins were predominantly enriched in the nucleus of hepatocytes, whereas OGT was observed to be nucleocytoplasmic. However, upon 13-h nutrient starvation, we observed the disappearance of dense nuclear staining for O-GlcNAcylated proteins — which now appear thoroughly diffused in the cytoplasm (Fig. 4B). Additionally, there was a reduction in the nuclear OGT immunostaining in fasting conditions compared to the ad-libitum feeding state.Fig. 4Changes in sub-cellular localization of OGT and O-GlcNAcylated proteins after nutrient fasting (A and B). Immunofluorescence analysis of paraffin-embedded liver sections from liver of ad-libitum fed mice (A) and 13-h nutrient-fasted mice (B). Sections are stained with DAPI (blue), anti-O-GlcNAc RL2 (red), and anti-OGT-1 (green) antibodies to show cellular localization. Scale bars, 50 μm. (C and bD). Paraffin-embedded liver sections of ad-libitum fed mice (C) and 13-h nutrient-fasted mice (D) shown in (A and B) are stained with DAPI (blue) and anti-HCF-1 (green) antibody.Full size imageThe changes in cellular localization of O-GlcNAcylated proteins and OGT in hepatocytes of 13-h fasted mice appear identical to the changes observed in the HCF-1 lacking hepatocytes. Thus, we investigated whether the fasting condition also impacted HCF-1 protein nuclear distribution. As shown in Fig. 4C and Fig. 4D, HCF-1 protein levels (as noted by immunofluorescence) were not perturbed by the 13-h fasting period.DiscussionLoss of OGT in murine liver exhibits severe defects in liver regeneration with increased pro-fibrotic and pro-inflammatory genes, resulting in hepatic dysplasia, NAFLD, and NASH30,31,32. We previously described a similar model for rapid recapitulation of NAFLD in the livers of adult male mice upon inactivation of the Hcfc1 allele. In this study, we focused on the role of HCF-1 on the nutrient-sensing enzyme OGT in this model. OGT features an N-terminal TPR domain conducive to protein–protein interaction and a C-terminal catalytic domain containing binding sites for UDP-GlcNAc and substrate peptides11,33,34,35.Using tamoxifen-induced Cre-mediated loxP site recombination, we generated a conditional hepatocyte-specific Hcfc1 KO in Hcfc1lox/Y mice carrying the Albumin-Cre-ERT2tg transgene. As Hcfc1 is an X-linked allele, complete KO was created in hemizygous Hcfc1lox/Y male mice, whereas female Hcfc1lox/+ mice produced a heterozygous HCF-1 KO pattern22.A large proportion of OGT forms a complex with HCF-1, and this HCF-1-OGT interaction is necessary for O-GlcNAcylation and cleavage of the 2035 amino acid precursor HCF-1 protein of a glutamate residue at position 10 (E10) of HCF-1PRO repeat8,10. Our findings demonstrate that HCF-1 loss significantly impacts OGT protein levels and enzymatic activity in hepatocytes. The progressive loss of OGT protein following HCF-1 depletion, observed in both whole liver lysates and isolated hepatocytes, indicates a post-translational regulatory mechanism. This reduction occurred without changes in Ogt mRNA levels, suggesting that HCF-1 controls OGT protein stability rather than transcriptional regulation.We suggest that HCF-1 regulates OGT activity, stability, and localization via multiple post-translational regulatory mechanisms. HCF-1 has been shown to regulate OGT protein stability through recruitment of the BAP1 deubiquitinase. BAP1 forms a tri-complex with HCF-1 and OGT36, and BAP1 can directly deubiquitylate and stabilize OGT protein37. BAP1-deficient cells show decreased OGT levels and increased OGT protein turnover37, indicating that HCF-1-mediated BAP1 recruitment protects OGT from proteasomal degradation. This mechanism is consistent with our observation that OGT protein levels decrease following HCF-1 loss without changes in OGT mRNA. This represents a mutual regulatory circuit where HCF-1 recruits BAP1 to stabilize OGT, while OGT reciprocally O-GlcNAcylates HCF-1 at over 30 sites and performs HCF-1 proteolytic maturation7.While BAP1-mediated stabilization explains OGT protection, other degradation pathways may also be involved. Previous studies have shown that OGT’s stability is regulated via degradation through the ubiquitin–proteasome pathway32. Two E3-ligases, XIAP and LSD2, were shown to induce OGT ubiquitination. Although we observed no changes in mRNA levels for these two E3-ligases or LSD2 protein levels upon HCF-1 KO, the possibility of OGT ubiquitination and degradation through activation of these E3-ligases or a novel E3-ligase remains possible. Future studies employing proximity labeling, chromatin immunoprecipitation, and biochemical reconstitution approaches will be essential to fully define the molecular details of these HCF-1-OGT regulatory interactions.Beyond affecting overall OGT protein levels, HCF-1 loss also altered the subcellular distribution of OGT. The KO of HCF-1 results in a significant reduction of the nuclear localization of OGT compared to the corresponding cytoplasmic pool, as revealed by immunostaining Hcfc1hepKO/Y mice. These findings are consistent with prior research conducted in HeLa cells8 , which showed that depletion of HCF-1 via shRNA also led to a perturbation of OGT’s nuclear localization as observed by subcellular fractionation.Interestingly, nutrient starvation also showed a similar reduction in nuclear OGT and O-GlcNAcylation staining in normal C57BL/6 mice, suggesting that OGT’s nuclear localization may be sensitive to nutrient conditions. The observation that fasting reduces nuclear O-GlcNAcylation and OGT localization without altering HCF-1 protein levels (Fig. 4) initially appears contradictory to our findings that HCF-1 regulates OGT function. However, this phenomenon can be explained by several nutrient-sensitive mechanisms that may disrupt OGT activity and its interactions without requiring changes in HCF-1 abundance. First, fasting has been shown to deplete cellular UDP-GlcNAc levels by 65% 38, directly impairing OGT function regardless of HCF-1 protein levels. Second, OGT nuclear localization requires auto-O-GlcNAcylation at Ser38939 creating a substrate-dependent mechanism where reduced UDP-GlcNAc availability during fasting would prevent OGT’s nuclear import, thereby limiting access to nuclear substrate proteins. Third, fasting activates AMPK, which directly phosphorylates OGT at Thr44440, altering OGT’s substrate selectivity. While the direct impact of this phosphorylation on HCF-1-OGT binding remains to be determined, it could potentially affect their interaction. Finally, reduced UDP-GlcNAc availability during fasting would lead to HCF-1 hypoglycosylation, which may impair its ability to interact with OGT. As HCF-1 contains numerous O-GlcNAcylation sites and serves as a major nuclear platform for OGT recruitment, disruption of HCF-1-OGT interactions would simultaneously reduce both HCF-1's own glycosylation and that of HCF-1-associated nuclear proteins,. These converging mechanisms allow nutrient starvation to phenocopy HCF-1 loss effects, explaining the dramatic loss of nuclear O-GlcNAcylation observed during fasting despite unchanged HCF-1 protein levels.Collectively, our findings reveal that HCF-1 serves as a master regulator of hepatic OGT function through multiple post-translational mechanisms, establishing a nutrient-sensitive regulatory circuit essential for proper liver metabolic function.Materials and methodsMiceAll experimental protocols were approved by the European Union (EU) and national legislation. These regulations were advised by the Lemanique Animal Facility Network (RESAL) to ensure ethical considerations were met regarding the transportation, housing, strain maintenance, breeding, and experimental use of animals.Mice bearing the Hcfc1 conditional (lox) allele were generated by Ozgene Pty Ltd. (Described in detail in Minocha et al.,2016). Homozygous mice carrying the Hcfc1 conditional (lox) allele are referred to as Hcfc1lox/lox22. When Cre recombinase is present, the Hcfc1lox allele is transformed into the Hcfc1cKO allele, resulting in a highly truncated HCF-1 protein comprising 66 amino acids22. Additionally, the study involved the use of wild-type C57BL/6 mice and Alb-Cre-ERT2tg transgenic mice (provided by Daniel Metzger, IGBMC Strasbourg). For the control group, control female Hcfc1lox/+ and male Hcfc1lox/Y mice were utilized. For the experimental group, female Alb-Cre-ERT2tg; Hcfc1lox/+ and male Alb-Cre-ERT2tg; Hcfc1lox/Y mice were employed. All mice used in this study were between 10-to-14 weeks old.Littermate female and male mice were housed in groups of four or five per cage. The housing conditions included a temperature of 23°C, a 12-h light/dark cycle, and unrestricted access to food and water unless otherwise mentioned. Mice were anesthetized with isoflurane and euthanized by cervical dislocation.DNA isolation and genotypingTo perform genotyping, genomic DNA was extracted from postnatal mouse ear tags, following previously established protocols41. These DNA samples were then subjected to PCR amplification using the following primer sets, utilizing the KAPA2G Fast HotStart genotyping PCR mix (no. KK5621).Primers and PCR conditions for genotypingFor HCF-1: p1 (5’-GGAGGAACATGAGCTTTAGG-3’), p2 (5’-CAATAGGCGAGTACCATCACAC-3’), and p3 (5’-GGGAAAGTAGACCCACTCTG-3’). The annealing was done at 62°C for 15 s with an extension at 72°C for 10 s.For AlbCre: p1 (5’-ATCATTTCTTTGTTTTCAGG-3’), p2 (5’-GGAACCCAAACTGATGACCA-3’), and p3 (5’-TTAAACAAGCAAAACCAAAT-3’). The annealing was done at 53°C for 1 min with an extension at 72°C for 1 min. Combination of p1 and p2 was used to detect the wildtype allele (229 bp). Combination of p2 and p3 was used to detect the Cre allele (444 bp).Tamoxifen introductionIntraperitoneal injections of tamoxifen were administered to both control and experimental groups of mice. Each mouse received 1 mg/mouse tamoxifen (100 ul of 10 mg/ml [1:10 ethanol-corn oil]; 10,540–29-1; Sigma-Aldrich) three times at 24-h intervals from day 0 to day 2.Reagents and antibodiesCollagenase type II was purchased from Worthington. Anti-OGT (sc-32921) anti-LSD2 (sc- 517,222) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX). Anti-O-GlcNAc (RL2) antibody was purchased from Abcam (Cambridge, UK). Anti-HCF-1 antibody was purchased from Bethyl Laboratories (Montgomery, TX). Anti-actin antibody and anti-U2AF65 were purchased from Sigma Aldrich (St. Louis, MS). Anti-HNF4α was purchased from R&D Systems (cat. # PP-H1415-00).The secondary antibodies used were: goat anti-rabbit Alexa 488 (1:400, Molecular Probes cat. # A11034), goat anti-mouse Alexa 568 (1:500, Molecular Probes cat. # A11019), goat anti-rabbit Alexa 568 (1:1000, Molecular Probes cat. # A21069), goat anti-mouse Alexa 488 (1:400, Molecular Probes cat. # A11029), and donkey anti-mouse Alexa 594 (1:500, Molecular Probes cat. # A11005).Protein isolation & Western blottingApproximately 100 mg of liver tissue was homogenized in RIPA buffer (50 mM Tris–HCl ph7.4, 150 mM NaCl, 1 mM EDTA, 0.2% 133 sodium deoxycholate, 1 mM DTT, 1 mM PMSF, and 1% Triton X) containing protease inhibitor as described previously18. Proteins were resolved using SDS–polyacrylamide gels and blotted onto nitrocellulose membranes. Membranes were blocked for 60 min with 5 ml of LI-COR blocking buffer, incubated with primary antibody in 50% LI-COR blocking buffer and 50% PBST (PBS containing 0.1% Tween 20) overnight at 4 °C, washed three times, and incubated with secondary antibody (dilution 1:10,000) for 30 min at RT. The membranes were washed three times and scanned with an Odyssey infrared imager (LI-COR). The following primary antibodies were used at the following dilutions: anti-O-GlcNAc RL2 (1:1000), anti-OGT (1:5000), anti-HCF-1 (1:1000), anti-U2AF65 (1:1000), anti-LSD2 (1:1000). IRDye 680 donkey anti-rabbit and IRDye 800 donkey anti-mouse were used at dilutions of 1:10,000. Blots were imaged using the LI-COR Odyssey IR imaging system (LI-COR, Lincoln, NE).ImmunofluorescenceFor fluorescence immunostaining, the liver tissues were paraffin-embedded and sectioned into 4 μm thick sections using a MICROM HM325 microtome. The paraffin-embedded sections were first (i) deparaffinized in xylene, (ii) rehydrated through graded alcohol washes, and (iii) rinsed twice with PBS. Antigen retrieval was done by heating in a 750 W microwave oven until boiling for approximately 10 min in citrate buffer (10 mM, pH 6.0), allowed to cool to 4°C slowly, washed twice with PBS, and then blocked for 30 min with 2% normal goat serum (NGS) (Sigma-Aldrich, cat. # G9023) in PBS at room temperature (RT). After blocking, primary immunostaining was performed by incubating the slices with a specific primary antibody diluted in 2% NGS overnight at 4°C, followed by three washes with PBS. For secondary fluorescence immunostaining, incubation with the appropriate secondary antibody was for 30 min in the dark at RT, followed by (i) three PBS washes, (ii) counterstaining with 4’,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich, CAS # 28,718–90-3), (iii) two PBS washes, and (iv) embedding with Mowiol mounting medium (Sigma-Aldrich, CAS # 9002–89-5). The sections were subsequently analyzed using an AxioImager M1 microscope with AxioCam MRm monochrome and AxioCam MRc color cameras (Carl Zeiss AG, Oberkochen, Germany) or a Zeiss CLSM 710 spectral confocal laser scanning microscope. Images were processed using AxioVision 4.8.2 (Carl Zeiss AG, Oberkochen, Germany) or Imaris 8.2 (Bitplane Inc.) software.Two-step hepatocyte isolation and culturePrimary murine hepatocytes were isolated using a rapid two-step isolation protocol as described previously42. Briefly, following anesthesia using Isoflurane, the vena cava is cannulated, and the liver is perfused to chelate calcium and wash out blood. Then, pre-warmed collagenase solution is perfused to the liver in order to dissociate the extracellular matrix. Finally, the liver is dissected, and hepatocytes are purified by density-based separation. Isolated hepatocytes were allowed to adhere to a plastic plate in DMEM and grow in the incubator (37°C and 5% CO2).RNA sequencingRNA sequencing of the control and Hcfc1lox/Y murine liver was performed as described previously18.Quantification and statistical analysisIn our study, we began by assessing the normality of the data obtained from at least three independent experiments. We also checked for equality of variance among the different groups. Following these checks, we performed a Student’s t-test to analyze the data. The intensity of western blot bands and the lengths of neurites were measured using ImageJ software. To indicate the significance levels in our quantification, we used asterisks according to standard P-value thresholds: ***P