FACI Is a Novel CREB-H-Induced Protein That Inhibits Intestinal Lipid Absorption and Reverses Diet-Induced Obesity

Abstract

Background & aims: CREB-H is a key liver-enriched transcription factor governing lipid metabolism. Additional targets of CREB-H remain to be identified and characterized. Here, we identified a novel fasting- and CREB-H-induced (FACI) protein that inhibits intestinal lipid absorption and alleviates diet-induced obesity in mice.

Methods: FACI was identified by reanalysis of existing transcriptomic data. Faci^-/- mice were generated by clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated 9 (Cas9)-mediated genome engineering. RNA sequencing was performed to identify differentially expressed genes in Faci^-/- mice. Lipid accumulation in the villi was assessed by triglyceride measurement and Oil red O staining. In vitro fatty acid uptake assay was performed to verify in vivo findings.

Results: FACI expression was enriched in liver and intestine. FACI is a phospholipid-binding protein that localizes to plasma membrane and recycling endosomes. Hepatic transcription of Faci was regulated by not only CREB-H, but also nutrient-responsive transcription factors sterol regulatory element-binding protein 1 (SREBP1), hepatocyte nuclear factor 4α (HNF4α), peroxisome proliferator-activated receptor γ coactivator-1α (PGC1α), and CREB, as well as fasting-related cyclic adenosine monophosphate (cAMP) signaling. Genetic knockout of Faci in mice showed an increase in intestinal fat absorption. In accordance with this, Faci deficiency aggravated high-fat diet-induced obesity, hyperlipidemia, steatosis, and other obesity-related metabolic dysfunction in mice.

Conclusions: FACI is a novel CREB-H-induced protein. Genetic disruption of Faci in mice showed its inhibitory effect on fat absorption and obesity. Our findings shed light on a new target of CREB-H implicated in lipid homeostasis.

Keywords: Intestinal Fat Absorption; Lipid Homeostasis; Metabolic Syndrome; Phospholipid-Binding Protein; Recycling Endosome.

Graphical abstract

Figure 1

Faci is a CREB-H–regulated gene preferentially expressed in liver and intestine. (A) Identification of Faci as a target of CREB-H. (B and C) Faci expression in the liver of WT and Crebh^-/- mice. Data were retrieved from GEO data sets (B) GSE29643 and (C) GSE121096. Normalized mRNA expression intensities in arbitrary units (AU) for GSE29643 and log₂ mRNA expression intensities for GSE121096 were statistically analyzed by the 2-tailed Student t test. (D and E) Verification of Crebh and Faci mRNA expression by RT-qPCR. Six-week-old Crebh^-/- mice (male, n = 3 for each group) were injected intraperitoneally with AAV-green fluorescent protein (GFP) or AAV-CREB-H–ΔTC. AAV–CREB-H–ΔTC represents an AAV gene transfer vector expressing CREB-H–ΔTC driven by a liver-specific promoter and enhancers. Viral dose for injection was 1 × 10¹¹ genome copies/mouse. After 2 weeks, mice were killed and total RNA in the liver was extracted to detect (D) Crebh and (E) Faci mRNA expression by RT-qPCR. The Crebh mRNA detected encodes CREB-H–ΔTC protein. The result was normalized to β-tubulin expression level. Data are means ± SD. Statistical significance was evaluated by the 2-tailed Student t test. (F–K) Induction of Faci mRNA expression or promoter activity by CREB-H–ΔTC. Expression levels of (F, H, and J) Faci mRNA or (G, I and K) relative Faci-Luc activity in (F and G) HepG2, (H and I) Hep3B, and (J and K) AML12 cells that are mock-transfected or transfected with CREB-H–ΔTC or CREB-H–ΔTC–4A plasmid were analyzed by RT-qPCR or luciferase reporter assay, respectively. The mRNA level was normalized to that of β-tubulin. The readout of firefly luciferase activity was normalized to that of Renilla luciferase. Statistical significance was evaluated by 1-way ANOVA with the Tukey post hoc tests. (L) FACI protein expression in transfected cells. HEK293T cells were transfected with plasmids encoding FACI-V5His and V5-FACI. Although FACI-V5His carries C-terminal tags, the V5 tag appears at the N-terminal of V5-FACI. Proteasome inhibitor MG132 was added to prevent protein degradation. (M) Expression of endogenous FACI protein. Polyclonal antiserum αFACI was raised in rabbits against a synthetic peptide of FACI. HEK293T cells transfected with V5-FACI plasmid (lanes 1 and 2) and mouse liver tissue (lanes 3 and 4) were harvested and lysed for protein extraction. Total protein samples were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then probed with αFACI (lanes 1 and 3) and antibody-depleted αFACI (lanes 2 and 4). Antibody depletion was performed by pre-incubating αFACI for 2 hours with the synthetic peptide used as immunogen in antibody preparation. β-tubulin was detected as a loading control. (N) Faci mRNA expression profile in mouse tissues. Total RNA of indicated mouse tissues (n = 4) was extracted. The mRNA expression levels of Faci were analyzed by RT-qPCR and normalized to those of β-tubulin transcript. ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.

Figure 2

Faci is preferentially expressed in hepatocytes and enterocytes. (A) Faci mRNA expression in different cell clusters of liver. The single-cell RNA-seq data were retrieved from GEO: GSE109774. (B) Faci mRNA expression in mesenchymal and epithelial compartments of the developing intestine. The data were retrieved from GEO: GSE6383. Values represent log₂ probe signal intensities. Statistical analysis was performed with the 2-tailed Student t test. (C) RT-qPCR analysis. Epithelial and mesenchymal compartments of the jejunum were isolated from mice (n = 4). The total RNA was extracted and Faci mRNA levels were analyzed by RT-qPCR. Relative mRNA expression levels were derived by normalizing to the levels of β-tubulin transcript. The 2-tailed Student t test was performed to judge statistical significance. Faci expression in different cell clusters of the (D) small intestine epithelium and (E) large intestine epithelium. The data on the small intestine epithelium were retrieved from GEO: GSE92332. The data on the large intestine epithelium were retrieved from GEO: GSE109774. ∗∗∗P < .001. CPM, counts per million; NK, natural killer.

Figure 3

FACI sequence alignment and phylogenetic analysis. (A) Secondary structure prediction. Secondary structure of FACI was predicted using JPred (www.compbio.dundee.ac.uk/jpred) and IUPred2A (iupred2a.elte.hu). Two α-helical regions (red and orange) and 1 intrinsic disordered region (blue) are highlighted. (B) Multiple alignments of FACI homologs among 65 mammalian species. Five conserved regions are boxed and labeled as A to E. (C) A phylogenetic tree of FACI. FACI homologs exist in lizards, turtles, and mammals (blue), but are not found in cartilaginous fishes, snakes, crocodiles, or birds (green). Uncharacterized ancestral versions of FACI were identified in amphibians and bony fishes (red). (D) Sequence alignment of FACI homologs and ancestral forms. FACI homologs from Chelonia mydas (mammalia) and Phascolarctos cinereus (reptilia), as well as FACI ancestral proteins from Rhinatrema bivittatum (amphibia), Erpetoichthys calabaricus (osteichthyes), and Danio rerio (osteichthyes) were selected for analysis. Conserved motifs A, B, D, and E are boxed. N, species without FACI homologs; P, species with ancestral versions of FACI; Y, species with FACI homologs.

Figure 4

Subcellular localization of FACI. (A and C) AML12 and Caco2 cell lines stably expressing (A) V5-FACI or (C) mEmerald-FACI were fixed and then subjected to either immunostaining followed by confocal microscopic analysis or confocal microscopic analysis directly. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). FACI-mEmerald or V5-FACI is shown in green. (B) Immunoblotting. Expression of V5-FACI in AML2-V5-FACI and Caco2-V5-FACI stable cell lines was induced successfully by Dox. (D) AML12 cells were co-transfected with plasmids encoding V5-FACI or mEmerald-FACI plus 1 of the endosomal markers mCherry-Rab11a, mCherry-FYVE, mCherry-Rab7a, and Tag-RFP-Rab4a. Cells were fixed and subjected to either immunostaining followed by confocal microscopic analysis or confocal microscopic analysis directly. Scale bars: 20 μm. (E) Immunoblotting. Total membrane and cytosolic protein fractions of AML12-V5-FACI cells were isolated. V5-FACI was enriched in the total membrane fraction but not the cytosolic fraction. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

Figure 5

Further localization analysis of FACI. (A) Co-localization of FACI with various organelle markers was examined. AML12 cells transiently expressing mEmerald-FACI (left) or mCherry-FACI (right) were probed or stained with various organellar markers including mCherry-ER3 for ER, GM130 for Golgi apparatus, MitoTracker for mitochondria, mCherry-lysosome for lysosome, EGFP peroxisome for peroxisome, EGFP-LC3 for autophagosome, and EGFP-PLIN2 for lipid droplets. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). (B) Distribution of mEmerald-FACI in polarized Caco2 (left) and Calu3 (right) cells was examined. Caco2-mEmerald-FACI and Calu3-mEmerald-FACI stable cells were first differentiated into polarized states. Expression of mEmerald-FACI protein (green) then was induced with Dox. Representative Z-stack layers as indicated, including the apical layer, the layers with tight junction and recycling endosome, respectively, as well as the basal–lateral layer with nucleus, were imaged by confocal microscopy. Tight junction marker zonula occludens-1 (ZO-1) was stained with antibody and nucleus was stained with DAPI. Scale bars: 20 μm.

Figure 6

Determination of the membrane binding region of FACI. (A) Subcellular localization of FACI-mEmerald with C-terminal mEmerald. (B) A list of truncated FACI proteins. The region in red represents the conserved motif E. The α-helical regions within motif E were impaired in FACI-E and FACI-F mutants. The adjacent regions within motif E were disrupted in FACI-G and FACI-H mutants. The underlined sequence (ETLLDTNN) is present in mice but absent in human beings. (C) Co-localization analysis of FACI mutants with Rab11a. AML12 cells were transfected with plasmids expressing mCherry-Rab11a and the indicated mEmerald-FACI mutants. Cells were fixed and imaged by confocal microscopy. (D) Structured illumination imaging showing co-localization of mEmerald-FACI-mutant C, which contains motif E, with mCherry-FYVE in AML12 cells. Scale bars: 20 μm. DAPI, 4′,6-diamidino-2-phenylindole.

Figure 7

Phospholipid-binding property of FACI. (A) Helical region in FACI. Amphipathic helices were identified in the C-terminus of FACI using Heliquest (heliquest.ipmc.cnrs.fr). The hydrophobic face is highlighted in yellow. The values of hydrophobic moment (<μH>), hydrophobicity (<H>), and net charge (Z) are indicated. (B–D) Fat blot assays. Synthetic peptide of 46 residues corresponding to the C-terminus of FACI (FACI-C) was biotin-labeled and incubated with the (B) Membrane Lipid Strip, (C) Sphingo Strip, and (D) MultiPIP Grip. HRP-conjugated streptavidin was used to visualize the biotin-labeled FACI-C peptide by enhanced chemiluminescence (ECL) reagents. Both ECL blot and bright field (BF) are shown. (E) Co-localization of FACI protein with phosphoinositides to plasma membrane. mEmerald-FACI plus PI4, 5P₂ marker PLCD-PH or PI3, 4P₂ marker TAPP-PH were expressed in AML12 cells. PLCD-PH or TAPP-PH is indicated in red, while FACI is indicated in green. For cells transfected with PLCD-PH, the plasma membrane layer was selected for confocal imaging. (F) AML12 cells were co-transfected with plasmids expressing mCherry-FACI and either AKT-PH or AKT-PH-dominant negative (DN). AKT-PH-DN is the dominant-negative form of AKT-PH. AKT-PH or AKT-PH-DN is shown in green. The red signals indicate FACI proteins. Nuclei were stained by 4′,6-diamidino-2-phenylindole (DAPI). Co-localization in the merged panels is shown in yellow. Scale bars: 20 μm.

Figure 8

Transcriptional regulation of FACI. (A) Schematic diagrams of the promoter regions of mouse Faci and human FACI. Conserved regions CR1 (green), CR2 (red), CR3 (blue), AG-repeat region (yellow), and the putative genomic repeat sequences (black) are highlighted. (B) Dual-luciferase reporter assay in HepG2 cells. Stimulatory effects on the FACI promoter were tested. HepG2 cells were co-transfected with pFACI-Luc (pW) and plasmids encoding the indicated transcription factors or transcriptional regulators. The readouts of firefly luciferase were measured and normalized to those of Renilla luciferase. Statistical significance for all groups was evaluated by 1-way ANOVA with the Dunnett post hoc tests. (C) Mutational analysis of FACI promoter. pGL3-basic reporter plasmids driven by 3 truncated versions of FACI promoter, designated pA, pB, and pC, were constructed (left). HepG2 cells were transfected with the indicated transcriptional regulators. The activities of pFACI-Luc (pW) and the 3 truncated versions (pA, pB, and pC) were measured. The readouts of firefly luciferase were measured and normalized to those of Renilla luciferase (right). Statistical significance was evaluated by 2-way ANOVA with the Tukey post hoc tests. ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001. The indicated groups were compared. The asterisks were placed directly on top of the error bars when the comparison was made with the mock. (D) Faci (left) and Crebh (right) mRNA expression in livers of fed and fasted mice. Mice with or without overnight fasting were killed and livers were collected for total RNA extraction. The mRNA expression level of Faci and Crebh were measured by RT-qPCR. Statistical analysis was performed with 2-tailed Student t test. ∗∗∗P < .001. (E) Induction of Faci mRNA expression in gonadal white adipose tissue (gWAT) by forskolin (FSK). Adipose tissue explants from 3 male mice were mock-treated with vehicle or treated with forskolin for 2 hours. Total RNA was extracted and Faci mRNA expression was analyzed by RT-qPCR. ∗P < .05 by 2-tailed paired Student t test. (F) Induction of FACI mRNA expression in human adipocytes by norepinephrine (norepi) treatment. Biopsy specimens from abdominal subcutaneous adipose tissue of nondiabetic human subjects were obtained. Differentiated adipocytes were stimulated for 4 hours with norepinephrine. Expression values represent transcripts per kilobase million. The data were retrieved from GEO: GSE150119. Statistical analysis was performed with 2-tailed Student t test. ∗∗P < .01.

Figure 9

Generation and genotyping of Faci^-/-mice. (A) Schematic diagram illustrating the strategy for generation of Faci^-/- mice. Two gRNAs (gRNA1 and gRNA2) were designed for FACI knockout. Primers indicated were used for genotyping. A 3200-bp genotyping product was predicted for WT mice, while a 200-bp genotyping product was expected for Faci^-/- mice. A DNA sequence chromatogram of Faci^-/- mice also is shown. (B) Genotyping of Faci^-/- mice. Genotyping PCR produced distinct products of 3200 bp for WT and 200 bp for Faci^-/- mice. Both 3200-bp and 200-bp products were present in the heterozygous Faci^+/- mice. (C) RT-qPCR analysis. Total mRNA of either intestinal epithelium from jejunum or liver of WT and Faci^-/- mice (n = 4) was extracted. Faci mRNA expression was analyzed by RT-qPCR. Faci mRNA was undetectable in Faci^-/- mice. ∗∗∗P < .001 by unpaired 2-tailed Student t test. HET, heterozygous; KO, Faci knockout; M, DNA markers.

Figure 10

Faci deficiency aggravates diet-induced obesity, dyslipidemia, and hepatic steatosis in mice. (A) Body weight and (B) weekly weight gain of male WT and Faci^-/- mice on NCD or HFD diet (4 mice per NCD group, 6 mice per HFD-WT group, and 7 mice per HFD-Faci^-/- group). Results are means ± SD. Statistical significance of body weights or weekly weight gain between HFD-WT and HFD-Faci^-/- were assessed by 2-way ANOVA with repeated measures followed by the Tukey test. Body weight of (C) female and (D) male Faci^-/- mice at 4 months. For this measurement, 18–20 female mice or 13–16 male mice were used in each group. ∗P < .05 by 2-tailed Student t test. (E) Fat percentages of male WT and Faci^-/- mice on NCD or HFD diet for 11 weeks. Statistical analysis was performed with 1-way ANOVA with the Tukey post hoc test. (F and G) Cumulative food intake of male WT and Faci^-/- mice on NCD or HFD diet. (H and I) Plasma (H) total cholesterol (TC) and (I) total TG levels of fasted male WT and Faci^-/- mice on NCD or HFD diet for 12 weeks. Results were assessed statistically with 1-way ANOVA with the Tukey post hoc test. (J) Representative images of livers and hearts isolated from WT and Faci^-/- mice on HFD diet for 12 weeks. (K and L) Liver weight and liver-weight percentages (liver weight/body weight) of WT and Faci^-/- male mice on NCD or HFD diet for 12 weeks. Statistical analysis was performed with 1-way ANOVA with Welch correction followed by Games Howell post hoc tests. (M) Hepatic alanine aminotransferase (ALT) levels of male WT and Faci^-/- mice on NCD or HFD diet for 12 weeks. Difference between HFD-WT and HFD-Faci^-/- was assessed statistically with the Kruskal–Wallis test with the Dunn post hoc tests. (N) Hepatic TG and (O) cholesterol levels of male WT and Faci^-/- mice on NCD or HFD diet for 12 weeks. Results were assessed statistically by 1-way ANOVA with Tukey post hoc tests. (P) Representative images of H&E-stained (top row) and Oil red O–stained (bottom row) liver sections from WT and Faci^-/- mice on NCD or HFD diet for 12 weeks. Scale bars: 100 μm. ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.

Figure 11

Faci deficiency exacerbates insulin resistance in mice. Overnight fasting (A) blood glucose, (B) blood insulin, and (C) homeostasis model assessment of the insulin resistance index (HOMA-IR) of male WT and Faci^-/- mice on NCD or HFD diet for 11 weeks. Data were statistically analyzed with 1-way ANOVA with Tukey post hoc tests. The HOMA-IR values were calculated using iHOMA2 software. (D and E) Intraperitoneal glucose tolerance test (IPGTT) of male WT and Faci^-/- mice on NCD or HFD diet for 11 weeks (4 mice per NCD group, 6 mice per HFD-WT group, and 7 mice per HFD-Faci^-/- group). Mice were overnight-fasted and injected with 1 g/kg glucose intraperitoneally. The blood glucose levels were examined at regular intervals. Results were statistically analyzed by 2-way ANOVA with repeated measures followed by the Tukey test. (E) Areas under the curve (AUC) for IPGTT were calculated using Y = 0 as the baseline. Results were statistically analyzed by 1-way ANOVA with Tukey post hoc tests. (F) Blood insulin levels during IPGTT were examined and statistically assessed by 2-way ANOVA with repeated measures followed by the Tukey test. (G) The AUC was calculated using Y = 0 as the baseline and judged statistically by 1-way ANOVA with Tukey post hoc tests. ∗P < .05, and ∗∗P < .01.

Figure 12

Inhibition of intestinal lipid absorption by FACI. (A–D) Transcriptomic profiles of intestinal epithelium from jejunum between WT and Faci^-/- mice (male, n = 3) by RNA-seq analysis. (A) Heatmap depicting the correlation of 6 RNA-seq samples, that is, intestinal epithelium from 3 WT (WT1, WT2, and WT3) and 3 Faci^-/- mice (KO1, KO2, and KO3) by Pearson correlation coefficient analyses. (B) The volcano plot illustrates DEGs. DEGs were selected with the criteria of FDR (false discovery rate) less than 0.01 and log₂ (fold change) (FC) of 2 or greater. Up-regulated and down-regulated DEGs are shown in red and green, respectively. (C) The bubble plot depicts gene ontology of up-regulated DEGs. The Y-axis represents GO terms. The X-axis indicates the gene ratio. Bubble colors represent log₁₀ (FDR) and bubble sizes indicate gene counts. (D) Heatmap illustrating the fold changes of lipid absorption–related DEGs in WT (WT1, WT2, and WT3) and Faci^-/- mice (KO1, KO2, and KO3). Scaled FPKM values were used for heatmap generation (Supplementary Table 2). Up-regulation and down-regulation are highlighted in yellow and blue, respectively. (E) RT-qPCR analysis. Total mRNA of intestinal epithelium from jejunum of WT and Faci^-/- mice (male, n = 4) was extracted. The mRNA levels of the indicated genes were analyzed by RT-qPCR. Results were statistically assessed by unpaired 2-tailed Student t test. (F) Immunoblotting. Total proteins of intestinal epithelium from jejunum of WT and Faci^-/- mice (male, n = 3) were extracted. Expression of the indicated proteins was analyzed. β-actin was detected as the internal control. (G) Small intestine length of male WT and Faci^-/- mice at 3 months (n = 18–20 mice per group). (H) Duodenum villus length of male WT and Faci^-/- mice at 3 months (n = 4 per group). (I) Plasma TG measurement. Male mice (n = 4 per group) were HFD-fed for 5 days, fasted (6 hours), and injected intraperitoneally with 30% wt/wt lipoprotein lipase inhibitor poloxamer 407 (1.5 g/kg body weight). After 30 minutes, mice were orally gavaged with olive oil (10 mL/g body weight). Plasma TGs at the indicated time points were measured. Statistical analysis was based on 2-way ANOVA with repeated measures, followed by the Sidak test. (J) Very-low-density lipoprotein secretion assay. Female mice (n = 3 per group) were fed with HFD for 5 days, fasted overnight, and injected intraperitoneally with Poloxamer 407 (1.5 g/kg). Plasma TG was measured at the indicated time points. Data are shown as means ± SD. (K) Measurement of TG in small intestines. Male mice (n = 3 per group) were HFD-fed for 5 days, fasted overnight, followed by high-fat refeeding. Small intestines of mice were removed, washed, and homogenized. Lipids were extracted following the Bligh and Dyer method. Statistical analysis was performed with an unpaired 2-tailed Student t test. (L) Oil red O staining of proximal jejunum. Male mice were HFD-fed for 5 days, fasted overnight, followed by high-fat refeeding. Intestinal neutral lipids were visualized with Oil red O staining. (M and N) Fatty acid uptake assay. BODIPY-C12 fatty acid uptake was measured using the QBT fatty acid uptake kit (Molecular Devices). (M) Intracellular fluorescence signals were detected every 20 seconds for up to 80 minutes. Fatty acid uptake was compared between FACI-expressing (Dox group, n = 6 wells) and mock-treated (no Dox group, n = 3 wells) Caco2 cells. (N) Areas under the curve (AUC) for kinetic FA uptake were calculated. An unpaired 2-tailed Student t test was performed to assess statistical significance. ∗P < .05, ∗∗P < .01. adj, adjusted; BP, biological process; CC, cellular component; MF, molecular function; RFU, Relative fluorescence unit.