Tissue- and Nuclear Receptor-Specific Function of the C-Terminal LXXLL Motif of Coactivator NCoA6/AIB3 in Mice

Abstract

Although the LXXLL motif of nuclear receptor (NR) coactivators is essential for interaction with NRs, its role has not been assessed in unbiased animal models. The nuclear receptor coactivator 6 (NCoA6; also AIB3, PRIP, ASC-2, TRBP, RAP250, or NRC) is a coactivator containing an N-terminal LXXLL-1 (L1) and a C-terminal L2. L1 interacts with many NRs, while L2 interacts with the liver X receptor α (LXRα) and the estrogen receptor α (ERα). We generated mice in which L2 was mutated into AXXAL (L2m) to disrupt its interaction with LXRα and ERα. NCoA6^L2m/L2m mice exhibited normal reproduction, mammary gland morphogenesis, and ERα target gene expression. In contrast, when treated with an LXRα agonist, lipogenesis and the LXRα target gene expression were significantly reduced in NCoA6^L2m/L2m mice. The induction of Cyp7A1 expression by a high-cholesterol diet was impaired in NCoA6^L2m/L2m mice, which reduced bile acid synthesis in the liver and excretion in the feces and resulted in cholesterol accumulation in the liver and blood. These results demonstrate that L2 plays a tissue- and NR-specific role: it is required for NCoA6 to mediate LXRα-regulated lipogenesis and cholesterol/bile acid homeostasis in the liver but not required for ERα function in the mammary gland.

The nuclear receptor (NR) superfamily consists of hormone-inducible transcription factors. Through regulation of gene expression, NRs control numerous biological events in development, growth, sexual maturation, reproduction, and metabolic homeostasis. Recent studies have identified a number of transcriptional coactivators and corepressors that modulate NR transcriptional activities and determine the expression levels of their target genes (28, 36, 41). Multiple coactivators usually form functional protein complexes which facilitate chromatin remodeling, general transcription factor assembly, RNA polymerase II recruitment, and transcriptional initiation. Mechanistically, most coactivators do not bind to DNA, and their recruitment to a gene promoter/enhancer is dependent on their interaction with specific agonist-bound NRs. Most coactivators contain one or more LXXLL (L, leucine; X, any amino acid) α-helix motifs required for interaction with the ligand-binding domain of NRs (10, 26, 27, 37). Although the molecular basis and requirement of the LXXLL motif for NR coactivator function have been investigated using biochemical and cell culture experiments (10, 26, 27, 37), the physiological function and the tissue and NR specificities of the LXXLL motif of coactivators have not been carefully assessed in an animal model using an unbiased molecular genetic approach.

The nuclear receptor coactivator 6 (NCoA-6; also AIB3, PRIP, ASC-2, TRBP, RAP250, or NRC) is an NR coactivator amplified and overexpressed in some breast, colon, and lung cancers (4, 9, 16, 20, 24, 25, 50). Biochemical and cell culture-based experiments have demonstrated that NCoA6 interacts with and coactivates many NRs, including estrogen receptor α (ERα) and liver X receptor α (LXRα) (15, 24, 25). NCoA6 may enhance NR-dependent transcription through its interaction and recruitment of multiple coactivator complexes, such as the ASC-2 (NCoA6) complex, the steroid receptor coactivator 1/cyclic AMP (cAMP) response element-binding protein binding protein (SRC-1/CBP) complex, thyroid receptor-associated protein, or the vitamin D receptor-interacting protein complex, and COAA (coactivator's coactivator) (4, 7, 11, 16, 20, 24, 25, 50). Disruption of both NCoA6 alleles in mice results in embryonic lethality due to defective development of the placenta, heart, and liver (3, 18, 23, 51). Disruption of one NCoA6 allele in mice accelerates polyomavirus middle-T-antigen-induced mammary tumorigenesis, reduces insulin secretion, and slightly affects postnatal growth (23, 43, 47). In addition, specific deletion of the NCoA6 gene in mouse mammary epithelial cells decreases mammary ductal growth regulated by estrogen and partially impairs milk synthesis (33). These findings indicate that NCoA6 is an essential coactivator in development, and it plays various physiological functions in different tissues. However, the molecular mechanisms for the tissue- and NR-specific activities of NCoA6 remain largely unknown.

NCoA6 contains two LXXLL motifs, the N-terminal LXXLL-1 (L1) and the C-terminal L2. The L1 motif interacts with many NRs, including ERα, but does not interact with LXRα and LXRβ (4, 9, 15, 16, 20, 24, 25, 50). The L2 motif mainly interacts with LXRs with high affinity and ERα with lower affinity (15, 24, 25). In the mammary gland and uterus, the ERα-mediated estrogen responses are essential for mammary gland morphogenesis and female reproduction (6, 17). In the liver, LXRα regulates the expression of the sterol response element-binding protein 1c (SREBP1c). SREBP1c is a master transcription factor that mediates the expression of key enzymes for lipogenesis in the liver, including fatty acid synthase (FAS), lipoprotein lipase (LPL), the acetyl coenzyme A (CoA) carboxylase (ACC), and the stearoyl-CoA desaturase 1 (SCD-1) (13, 34, 38, 45). Through direct and indirect induction of these lipogenic genes, LXRα enhances lipogenesis in the liver (38). More importantly, LXRα is a major sensor of dietary cholesterol. Oxysterols, the cholesterol metabolites, serve as endogenous ligands of LXRα. Upon activation, LXRα induces cholesterol 7α-hydroxylase (Cyp7A1) expression and controls the rate-limiting step of bile acid synthesis from cholesterol in the liver (5, 32, 46). Because bile acid synthesis and its fecal excretion play pivotal roles in the maintenance of cholesterol homeostasis upon cholesterol overfeeding, inactivation of LXRα function causes cholesterol accumulation in the liver and blood (5, 46). Therefore, we reasoned that if the L2 motif of NCoA6 mediates transcriptional function of ERα and LXRα, a loss-of-function mutation of the L2 motif would reduce ERα and LXRα function and affect mammary gland morphogenesis, liver lipogenesis, bile acid synthesis, and cholesterol homeostasis.

In this study, we have assessed the in vivo role of the NCoA6 L2 motif in LXRα and ERα function by generation and characterization of mice where L2 is mutated into AXXAL (L2m). We show that L2 is not required for ERα-regulated mammary gland development and gene expression but is absolutely required for LXRα-regulated lipogenesis, bile acid synthesis, and cholesterol homeostasis in the liver. Our results indicate that the L2 motif of NCoA6 plays a tissue- and NR-specific role.

MATERIALS AND METHODS

Construction of targeting vector.

The pLoxpI parent plasmid contained a neomycin-resistant expression cassette (neo) as a positive selection marker and a thymidine kinase expression cassette (tk) as a negative selection marker. In this vector, neo was flanked by a pair of LoxP sites. A 3-kb genomic DNA fragment in intron 8 of the mouse NCoA6 gene (GenBank NM_019825) was amplified by high-fidelity PCR from the 129SvEv mouse embryonic stem (ES) cell DNA template with the NCoA6-67F primer (5′-ATAGCGGCCGCACTGCATATACAC) and the NCoA6-67R primer (5′-TTCCTCGAGAAAACTTTGCTACCTTTAGAG). This PCR product was subcloned into the vector between the NotI and XhoI sites upstream of neo to serve as the 5′ targeting arm. Another 5-kb genomic DNA fragment from exon 9 to intron 9 was amplified by PCR using the NCoA6-63F primer (5′-TTTGTCGACTGAAAAATCTAACTTCCC) and the NCoA6-66R primer (5′-TTTATCGATCCCTAACTAAACTAGAAACAAAACC). This DNA fragment was first subcloned into the pBlueScript II SK plasmid (Stratagene, La Jolla, CA), and the DNA sequence encoding the LSQLL (L2) motif in exon 9 (5′-TTAAGCCAACTTCTT) was changed to the sequence 5′-GCAAGCCAAGCTCTT, encoding the ASQAL mutant motif (L2m) by PCR-based site-specific mutagenesis. This mutated DNA fragment was subsequently transferred into the targeting vector at the SalI and ClaI sites between the neo and the tk cassettes to serve as the 3′ targeting arm for knocking in the mutations. All exon sequences and the mutated sequences in the targeting vector were confirmed by DNA sequence analysis (Fig. 1A).

FIG. 1.

Generation of NCoA6 L2m mutant ES cell lines and mice by homologous recombination. (A) The knock-in strategy for generation of the NCoA6-L2m allele. The WT NCoA6 allele (WA), targeting vector (TV), targeted mutant allele with the neo expression cassette (TA), and mutant allele postexcision of the neo cassette (L2m) are sketched. The locations of major restriction sites used for subcloning or Southern blotting and PCR primers P19, P20, and P21 used in genotyping analysis are indicated. Exons 8 to 10 are indicated by black rectangles. The LoxP sites are indicated by black triangles. The region used as a Southern blot probe is also indicated. L2, the LXXLL-2 motif in WT allele; L2m, the AXXAL-2 mutant in the targeting vector or in the mutant allele. (B and C) Identification of targeted ES clones. Southern blotting was performed first with HindIII-digested ES cell DNA (B) and then with BstEII-digested ES cell DNA (C). The 9-kb band in panel B and the 13.4-kb band in panel C are from the targeted NCoA6 allele with neo (nL2m), while the 7-kb and the 11.4-kb bands in panels B and C are from the WT NCoA6 allele (+). (D) Southern blot assay. Mouse tail DNA was digested with BstEII. The 13.4-kb band is from the NCoA6 nL2m allele. The 11.4-kb band is either the WT NCoA6 allele or the L2m allele without neo. (E) PCR-based genotype analysis. DNA samples were isolated from tail tips. The 550-bp band was amplified from the nL2m allele using primers P19 and P20. The 260-bp band was amplified from the L2m allele using primers P21 and P20. The 220-bp band was amplified from the WT allele (+) using primers P20 and P21.

Gene targeting in ES cells.

TC-1 ES cells with a 129Sv/Ev strain background were cultured on a feeder layer of G418-resistant mouse embryonic fibroblasts (MEFs) in a medium containing 750 U/ml of leukemia inhibitory factor. The targeting vector was linearized by NotI digestion and electroporated into ES cells (18). Growth selection was carried out in ES cell medium containing 300 μg/ml of G418 and 0.2 μM 1-(2-deoxy-2-fluoro-β-d-arabinofuranosyl)-5-iodouracil (FAIU). Surviving clones were isolated and expanded in 96-well plates. For Southern blot screening, purified DNA samples were digested with HindIII and separated by electrophoresis in a 0.7% agarose gel. The blots were hybridized with a ³²P-labeled DNA probe located upstream of the 5′ targeting arm. The 430-bp probe template was generated by PCR from the ES cell DNA with the NCoA6N-1F primer (5′-GAGTGACACTGCAGAGTACTGG) and the NCoA6N-1R primer (5′-GCCTGTGGTGTCCAAGTTGATTC). For the secondary screening, DNA of the candidate clones was digested with BstEII and hybridized to the same probe described above. The blots with the BstEII-digested DNA were rehybridized with a neo probe for identifying clones with homologous recombination but without any misinsertion of the targeting vector. The correctly targeted clones were further examined for the ASQAL mutant coding sequence by PCR using a pair of mutant-specific primers, which were NCoA6-L2MF (5′-CACCAACATCAGCAAGCCAAGC) and NCoA6-62R (5′-GGCGGATCCTGCAGTGTGG). PCR was performed with the following program: 95°C for 1 min; 94°C for 1 min, 65°C for 30 s, and 72°C for 40 s, 30 cycles; 72°C for 5 min.

Generation of mutant mice.

Chimeric mice were generated by microinjection of the targeted ES cells into 3.5-day-old blastocysts collected from C57BL pregnant mice (18, 42). The chimeric mice were bred with B6.FVB-TgN(EIIa-Cre)C5379Lmgd/J mice with a C57BL background from Jackson Laboratory to produce heterozygous mutant (NCoA6^L2m/+) mice. Subsequently, NCoA6^L2m/+ mice with the EIIa-Cre transgene were bred to obtain NCoA6^L2m founder mice without the EIIa-Cre transgene. These founder mice with a mixed strain background of 50% 129SvEv and 50% C57BL were used to expand the colony and to produce wild-type (WT), NCoA6^L2m/+, and NCoA6^L2m/L2m mice for experiments. Mice were maintained on a 12-hour light/12-hour dark cycle and fed normal rodent diet (TestDiet, Richmond, IN) ad libitum. Animal procedures were approved by the Animal Care and Use Committee at Baylor College of Medicine.

Genotype analysis.

For Southern blot analysis, genomic DNA was prepared from tail tips of mice, digested with BstEII, and hybridized with the ³²P-labeled probe as described for ES cell DNA. For PCR analysis, a small piece of ear tissue was collected and digested in a solution containing 1.5 mg/ml proteinase K, 20 mM Tris-HCl (pH 8.0), 50 mM KCl, 2.5 mM MgCl₂, and 0.5% Tween 20 at 55°C for at least 2 h. The solution was heated at 95°C for 5 min to inactivate proteinase K and then centrifuged for 10 min. The supernatant was collected, and 0.5 μl of the supernatant was used for each PCR. The PCR program was as follows: 95°C for 2 min; 94°C for 1 min, 57°C for 40 s, and 72°C for 1 min, 35 cycles; 72°C for 10 min. Primers P20 (5′-TGCGACATGCTGCATGAG) and P21 (5′-CTTGCTTGTGTCAAATAATGAGG) were used to detect the 220-bp band from the NCoA6⁺ allele and the 260-bp band from the NCoA6^L2m allele after the neo cassette was excised (Fig. 1A and E). Primers P19 (5′-GGTGAGAACAGAGTTACCTAC) and P20 were used to detect the 550-bp band from the NCoA6^L2m allele with the neo cassette (Fig. 1A and E). Genotype analysis of the EIIa-Cre mice was performed as instructed by Jackson Laboratory.

RNA extraction and RPA.

The mouse liver and the fourth pair of mammary glands were quickly isolated, frozen in liquid nitrogen, and granulated into a fine powder under frozen conditions. Total RNA was extracted from the tissue powder by using the TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. The purified RNA was resuspended in diethyl pyrocarbonate-treated water. RNA quality was estimated by 1% agarose gel electrophoresis. The 320-bp DNA template for the RNase protection assay (RPA) probe was amplified by reverse transcription-PCR (RT-PCR) using primers P16 (5′-CAGAGTGACATTTCTGCAGG) and P20 and subcloned into the SmaI site of the pBluescript II SK plasmid. The plasmid DNA was linearized by BamHI digestion, and the ³²P-labeled antisense riboprobe was transcribed using T7 RNA polymerase. The probe sequence was complementary to exons 7 to 9. Protected RNA was separated by 6% acrylamide gel electrophoresis and visualized by autoradiography. An RPA for the β-actin mRNA was performed to serve as a loading control.

qPCR.

RNA samples were treated with RNase-free DNase I to remove residual DNA. A 200-ng aliquot of RNA was reversely transcribed into cDNA by using the reverse transcriptase core kit (Eurogenec, San Diego, CA). Gene-specific primer pairs and TaqMan probes were designed by using the Primer Express software (Applied Biosystems, Foster City, CA). Real-time quantitative RT-PCR (qPCR) was performed with 1.5 μl of reverse transcription product on the ABI Prism PE7500 sequence analyzer (Applied Biosystems). The TaqMan probe-based qPCR was performed to measure the progesterone receptor (PR), cyclin D1, and 18S RNAs. The primers and probes used for these measurements were previously described (29). The SYBR Green qPCR was performed to measure the mRNA concentrations of other genes using the following pairs of primers: 5′-GAGTGTCGACTTCGCAAATGC and 5′-TCAAGCGGATCTGTTCTTCTGA for LXRα; 5′-AAGCAGGTGCCAGGGTTCT and 5′-TGCATTCTGTCTCGTGGTTGT for LXRβ; 5′-GGCCGAGATGTGCGAACT and 5′-TGGTTGTTGATGAGCTGGAGC for SREBP-1c; 5′-GTCCCAGAAATCGCCTATGG and 5′-TTCTTTTCCGGTACTTTCGATATATAAAT for FAS; 5′-TCTGACCTGAAAGCCGAGAAG and 5′-TGGGCAGGATGAAGCACAT for SCD-1; 5′-GTGGCCGAGAGCGAGAAC and 5′-CCACCTCCGTGTAAATCAAGAAG for LPL; 5′-TAAACAACCTGCCAGTACTAGATAGCA and 5′-GTCCGGATATTCAAGGATGCA for Cyp7A1; 5′-TCCTCATCCTCGTCATTCAAA and 5′-GGACTTGGTAGGACGGAACCT for ABCA1; 5′-TCAGGACCCCAAGGTCATGAT and 5′-AGGCTGGTGGATGGTGACAAT for ABCG5; 5′-GACAGCTTCACAGCCCACAA and 5′-GCCTGAAGATGTCAGAGCGA for ABCG8. Relative standard curves were generated by using serial dilutions of a mixture of both WT and NCoA6^L2m/L2m RNA samples. Relative mRNA expression levels were determined against the standard curves and normalized to the 18S RNA amount.

Western blotting.

About 0.25 g of liver or mammary gland tissue from each mouse was homogenized in a lysis buffer consisting of 50 mM Tris-HCl (pH 7.4), 0.1% sodium dodecyl sulfate (SDS), 2 μM phenylmethylsulfonyl fluoride, and 10 μg/ml leupeptin. The tissue lysate was centrifuged, and the supernatant was mixed with the reducing sample buffer for SDS-polyacrylamide gel electrophoresis (SDS-PAGE). About 80 μg protein was separated in a 4.5% acrylamide gel and blotted onto a nitrocellulose membrane. The blot was probed with a specific primary antibody and then with goat anti-mouse or goat anti-rabbit immunoglobulin G (IgG) conjugated with horseradish peroxidase. The blot was developed using the chemiluminescent horseradish peroxidase substrate solution (Millipore Co., Billerica, MA) for 5 min and then exposed to X-ray films. The primary antibodies included monoclonal antibodies against NCoA6 (ASC-2) (20) and β-actin (Sigma, St. Louis, MO) and polyclonal antibodies against poly(ADP-ribose) polymerase (PARP; Upstate Biotechnology, Lake Placid, NY) and LXRα, LXRβ, and PR (Santa Cruz Biotechnology, Santa Cruz, CA).

MEFs, transient transfection, and co-IP.

The primary WT and NCoA6^L2m/L2m MEFs were isolated from embryos (two for each genotype) of a female mouse at 12.5 days postcoitum and cultured as we described previously (40). Frozen stocks of MEFs were prepared when they reached confluence in the initial culture. MEFs were thawed out and cultured for one passage and then set in six-well plates for transient transfection. For the ERα transactivation assay, MEFs were transiently transfected with the RSV-ERα, ERE-tk-Luc (reporter), and pCMV-RL (Renilla luciferase control; Promega, Madison, WI) plasmids using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA). These transfected MEFs were treated with 10⁻⁸ M 17β-estradiol for 24 h. For the LXRα transactivation assay, MEFs were transfected with the pCMX-mLXRα, pSG5-RXR (47), LXRE-tk-Luc (reporter), and pCMV-RL plasmids and treated with 10⁻⁶ M T0901317 for 24 h. Cells were harvested, and luciferase activities were measured with the dual-luciferase assay kit (Promega, Madison, WI). For coimmunoprecipitation (co-IP) experiments, transfected and hormone-treated MEFs were lysed for 30 min in the buffer containing 0.5% NP-40, 20 mM Tris-Cl (pH 7.5), 150 mM NaCl, 5 mM MgCl₂, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and proteinase inhibitor cocktail. After centrifugation at 16,000 × g for 18 min, the supernatant was incubated with 2 μg of mouse IgG and 10 μl of protein A-agarose beads for 1 h at 4°C for precleaning. The precleaned supernatant was mixed with 2 μl of mouse ascites fluid containing NCoA6 monoclonal antibody or a similar amount of normal mouse IgG as control for 15 h at 4°C and then with 10 μl of protein A-agarose beads for 1 h. Beads were collected by low-speed centrifugation and washed three times in the lysis buffer. Antibody-associated proteins were eluted by boiling the beads in 2× reducing sample buffer for SDS-PAGE and subjected to immunoblotting analysis with antibodies against NCoA6, ERα, or LXRα.

In vivo transfection.

The hepatocytes of 2-month-old male WT and NCoA6^L2m/L2m mice were transfected by tail vein injection of the LXRα (125 ng/g), RXRα (125 ng/g), LXRE-tk-luc (700 ng/g), and pCMV-RL (50 ng/g) plasmids using the TransIT in vivo gene delivery system (Mirus Bio Corp., Madison, WI). Two hours later, mice were fed T0901317 (50 mg/kg body weight) in sesame oil or only sesame oil. The treated mice were sacrificed at 24 h. The liver was homogenized in passive lysis buffer (0.2 g/ml) for a luciferase assay (Promega). After centrifugation, clear tissue lysate was collected for a luciferase assay using the dual-luciferase assay kit.

Chronic treatment of mice with the T0901317 LXR agonist.

Three-month-old male WT and NCoA6^L2m/L2m mice were fed T0901317 (50 mg/kg body weight/day) in sesame oil or only sesame oil for 6 days. Blood and liver samples were collected from the treated mice for triglyceride measurement, RNA preparation, histology examination, and Oil Red O staining.

Morphological and histological analyses.

For staining mouse mammary glands at different developmental stages, one of the inguinal mammary glands was dissected, mounted on a glass slide, fixed in Carnoy's fixative, and stained with carmine alum (Sigma, St. Louis, MO) (42, 47). The mammary gland morphology was imaged under a stereomicroscope mounted with a Carl Zeiss charge-coupled-device camera and analyzed by using the Axio Vision AC software (Carl Zeiss, Jena, Germany). For histological analysis, tissues were fixed overnight in neutralized formalin (10%, pH 7.2) at 4°C, washed in phosphate-buffered saline (PBS), dehydrated through serial ethanol solutions, and embedded in paraffin. Tissue sections were cut at 5 μm in thickness, and the rehydrated sections were stained with hematoxylin and eosin (H&E).

Oil Red O staining.

Fresh mouse liver tissue was sliced into small pieces (∼2 by 2 mm), immersed into Tissue Tech OCT compound (Sakura Finetek U.S.A. Inc., Torrance, CA), and immediately frozen on dry ice. Cryosections were cut at 8 μm in thickness and mounted on superfrosted slides. The slides were stained in the Oil Red O staining solution for 10 min, rinsed with water, and counterstained with hematoxylin for 10 seconds. After washed with water for 3 to 5 min, the slide was mounted with coverslips in PBS containing 50% glycerol.

Cholesterol and triglyceride measurements.

Five mice in each group were fasted for 5 h and anesthetized with avertin (20 μl of 1.25% solution per g body weight, intraperitoneal). About 100 μl of blood sample was collected from the retro-orbital vein of each mouse, and the serum sample was prepared by centrifugation at 5,000 × g for 25 min after the blood was clotted at 4°C. The concentrations of cholesterol and triglycerides in the serum were measured by the enzymatic method using the cholesterol quantitation kit (BioVision, Mountain View, CA) and the triglyceride determination kit (Sigma, St. Louis, MO). Hepatic lipids were extracted and analyzed as described elsewhere (44). Briefly, 10 mg of liver tissue was homogenized in 200 μl of chloroform-methanol (2:1) and then mixed with 40 μl of 50 mM NaCl. The mixture was centrifuged at 16,000 × g for 10 min. The organic phase was collected, vacuum dried, and redissolved in 20 μl of isopropanol containing 10% Triton X-100. One μl of the prepared sample was used for hepatic cholesterol and triglyceride measurements.

Bile acid assay.

The amount of bile acid excreted from the feces of WT and NCoA6^L2m/L2m mice was measured as described elsewhere (32). Briefly, mice were housed individually in the Nalgene diuresis metabolic cage (Fisher Scientific, Pittsburgh, PA) and fed either 2% cholesterol chow diet or normal diet (TestDiet, Richmond, IN) for 6 days. The feces were collected once a day from day 4 to 6. The feces from each mouse were dried overnight at 55°C and pooled. A total of 0.5 g dried feces was soaked in 2.5 ml water for 10 min, homogenized, and then extracted with 7.5 ml of ethanol at 50°C for 2 h. After centrifugation at 13,000 × g for 10 min, 1 ml of supernatant was collected and diluted to 3 ml with 0.4× PBS (pH 8.6). The fecal bile acid in the dilution was measured by using the bile acid assay kit (Sigma Diagnostics, St. Louis, MO).

RESULTS

Generation of the NCoA6 L2m mutant mice by a knock-in strategy.

To investigate the specific physiological role of the L2 motif of NCoA6, we used a knock-in strategy to generate the L2m mutant mice. In the L2m mutant mice, the LSQLL-2 motif of NCoA6 was mutated to ASQAL. This mutation was designed to disrupt the interaction of the NCoA6 L2 motif with LXR and ERα according to previous in vitro studies (15, 21, 24). The targeting vector consisted of a 3-kb 5′ arm from intron 8 and a 5-kb 3′ arm from exon 9 and intron 9 of the mouse NCoA6 gene. In the 3′ arm, the coding sequence for the LSQLL motif located in exon 9 was mutated to the sequence encoding the ASQAL amino acid residues. Between the two targeting arms, we inserted a neo positive selection marker that was flanked by two LoxP sites. We also added a tk negative selection marker downstream of the 3′ targeting arm (Fig. 1A). After a total of 6.6 × 10⁷ ES cells were electroporated with the targeting vector DNA and cultured in a selection medium containing G418 and FAIU, about 96 colonies formed. These colonies were screened by Southern blotting for 5′ homologous recombination by digesting the genomic DNA samples with HindIII and using a probe upstream of the 5′ targeting arm. Subsequently, we screened these colonies for both 5′ and 3′ homologous recombination by digesting the DNA samples with BstEII. Among 96 clones examined, 3 of them were identified as correctly targeted clones that exhibited a 7-kb band for the WT NCoA6 allele and a 9-kb band for the targeted allele after HindIII digestion and an 11.4-kb band for the WT allele and a 13.4-kb band for the targeted allele after BstEII digestion (Fig. 1B and C and data not shown). All three clones were further examined by Southern blotting using the neo probe after the DNA was digested with BstEII. The neo probe detected only the expected 13.4-kb band for the targeted NCoA6 allele (data not shown), indicating that the genome of these clones does not contain any random insertion of the targeting vector DNA.

Two of the targeted ES clones were injected into the blastocysts of C57BL/6J mice to produce chimeric mice. The chimeric mice were bred with the EIIa-Cre mice with a C57BL/6J background. This breeding strategy had two purposes: first, it served to test the mutant germ line transmission from the chimeric mice by coat colors. F₁ pups with a dominant agouti coat color from the 129SvEv ES cells indicated a germ line transmission, but F₁ pups with a recessive C57BL/6J black coat color represented a non-germ line transmission. Second, the EIIa-Cre transgene was expressed in early embryonic stages, and it served to excise the neo cassette bracketed by the LoxP sites from the targeted NCoA6 allele. From this breeding, F₁ offspring with WT, WT/EIIa-Cre, NCoA6^+/neo-L2m, and NCoA6^+/L2m/EIIa-Cre genotypes were generated. NCoA6^+/L2m/EIIa-Cre mice were further interbred to produce WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice with or without the EIIa-Cre transgene (Fig. 1E and data not shown). Finally, NCoA6^+/L2m mice without the EIIa-Cre transgene were used to establish the NCoA6 L2m mutant colony to produce WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice for experiments. The genotypes of these mice were verified by Southern blot analysis. As expected, after BstEII digestion the L2m DNA fragment showed a similar length to the WT DNA fragment after the floxed neo cassette was excised (Fig. 1D). The established mouse colony had a mixed 129SvEv and C57BL/6J strain background.

Next, we confirmed the L2m (ASQAL) mutations in mice by RT-PCR using RNA samples prepared from the liver and by sequencing the PCR products. WT mice had the expected nucleotide sequence encoding the L2 (LSQLL) motif. NCoA6^+/L2m mice showed both nucleotide sequences encoding the L2 motif from the WT allele and the L2m mutant from the targeted allele. NCoA6^L2m/L2m mice only expressed the nucleotide sequence encoding the ASQAL mutant motif (Fig. 2A). These results demonstrate that the L2 motif was successfully mutated to L2m in the NCoA6 L2m mutant mice. In order to examine if the single LoxP site left in intron 8 would affect splicing efficiency between exons 8 and 9 (Fig. 1A) of the mutant allele and to compare the mRNA expression levels of the L2m mutant and WT NCoA6, we performed RPA analyses with liver RNA samples from WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice. The 320-nucleotide antisense riboprobe complementary to a partial sequence of exons 8 and 9 located upstream of the mutant motif detected similar expression levels of the WT and mutant NCoA6 mRNAs in these mice (Fig. 2B). This indicates that the genetic manipulation and the mutation of the L2 motif had no effects on NCoA6 mRNA splicing and expression. Furthermore, Western blot analysis using a monoclonal antibody against the NCoA6 C terminus demonstrated that in the NCoA6^L2m/L2m mouse liver the NCoA6 L2m mutant protein was similarly expressed as the WT NCoA6 protein in the WT mouse liver (Fig. 2C). Similar data were also obtained from the mammary glands of WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice (data not shown).

FIG. 2.

Analysis of L2m mutant expression in knock-in mutant mice. (A) Sequence analysis of NCoA6 L2m mRNA. The expected cDNA sequences and peptide sequences of WT and L2m are listed on the top left. Total RNA samples were prepared from the livers of WT (+/+), NCoA6^+/L2m, and NCoA6^L2m/L2m mice and were reverse transcribed into cDNA as PCR templates. The DNA fragment spanning the L2m mutation region was amplified by PCR, and the PCR product was sequenced. The sequence data of WT (+/+), NCoA6^+/L2m, and NCoA6^L2m/L2m samples are presented. Note that NCoA6^+/L2m mice show both WT and L2m sequences. (B) RPA. Liver RNA samples (20 μg) were prepared from WT (+/+), NCoA6^+/L2m, and NCoA6^L2m/L2m mice. The RPA was performed with a ³²P-labeled antisense riboprobe complementary to a common region of NCoA6 WT and L2m mRNAs. Analysis of β-actin mRNA levels served as a loading control. C. Western blot assay. Cell nuclear lysates were prepared from the livers of WT (+/+), NCoA6^+/L2m, and NCoA6^L2m/L2m mice. Fifty μg of nuclear protein from each sample was separated by SDS-PAGE, and the blot was assayed using antibodies against NCoA6 and PARP. The nuclear protein PARP served as a loading control.

The NCoA6 LXXLL-2 motif is not required for normal survival, somatic growth, and reproduction.

We have previously demonstrated that NCoA6 null embryos die at a middle pregnant stage (18). To examine whether mutation of the NCoA6 L2 motif would affect animal survival and development, we recorded the number and genotypes of pups derived from the NCoA6^+/L2m breeding pairs at weaning. Among a total of 244 pups analyzed, there were 59 (24%) WT mice, 132 (54%) NCoA6^+/L2m mice, and 53 (22%) NCoA6^L2m/L2m mice. This distribution was consistent with the Mendelian ratio (1:2:1). In addition, NCoA6^L2m/L2m mice showed normal development. Together, these data suggest that the L2 motif is not required for NCoA6 to support animal survival and development.

One study has shown that disruption of one of the NCoA6 alleles (NCoA6^+/−) in mice causes growth retardation around the weaning stage (23). To examine whether mutation of the L2 motif would affect the molecular function of NCoA6 in growth regulation, we compared the body weights of mice with different genotypes. The average body weights of 26-day-old male WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice were 13.9 ± 1.2 g (n = 7), 14.0 ± 1.5 g (n = 22), and 13.9 ± 2.0 g (n = 12), respectively. The average body weights of 26-day-old female WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice were 13.4 ± 0.8 g (n = 6), 13.2 ± 1.8 g (n = 23), and 12.4 ± 2.1 g (n = 14), respectively. These body weights were not statistically different. Furthermore, the average body weights of 2-month-old WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice also showed no significant differences (data not shown). These results indicate that the L2 motif is not required for NCoA6 function in somatic growth.

We also compared the reproductive capability of NCoA6^+/L2m and NCoA6^L2m/L2m female mice with WT female mice after pairing with fertile WT males for 6 months. The female WT, NCoA6^+/L2m, and NCoA6^L2m/L2m mice produced 5.5, 5.7, and 5.3 litters on average per mother, with 8.2 ± 2.1, 7.9 ± 1.5, and 8.1 ± 1.9 pups per litter, respectively. Neither the litter numbers per mother nor the pup numbers per litter showed any statistical differences. The NCoA6^+/L2m and NCoA6^L2m/L2m male mice also exhibited similar reproductive functions. These data indicate that the NCoA6 L2 motif is not required for mouse reproductive function.

The NCoA6 LXXLL-2 motif is not essential for steroid-regulated mammary gland morphogenesis.

It has been reported that specific NCoA6 inactivation in mammary epithelial cells attenuates mammary gland ductal growth at puberty, decreases lobular alveolar development during pregnancy, and reduces milk production during lactation (33). It also has been shown that the NCoA6 L2 motif could interact with ERα, although the affinity of the interaction was not high (24). To examine if the L2 motif was required for NCoA6 activity in steroid-regulated mammary gland development, we compared the mammary gland morphology and its response to estrogen treatment in WT and NCoA6^L2m/L2m mice. The mammary ductal growth in 40-day-old WT and NCoA6^L2m/L2m virgin females was comparable (Fig. 3A). There were 61.8 ± 8.7 (n = 5) and 60 ± 9.9 (n = 5) ductal branches and 80.2 ± 5.9 (n = 5) and 78.2 ± 7.1 (n = 5) terminal end buds per mammary gland in the virgin WT and NCoA6^L2m/L2m mice, respectively. Analysis of whole-mounted mammary glands also revealed that the overall morphology and alveolar formation were similar between mammary glands of WT and NCoA6^L2m/L2m mice gestation day 15 and lactation day 3 (Fig. 3A). To precisely evaluate the possible role of the NCoA6 L2 motif in the estrogen/ERα-mediated mammary ductal growth, 17-day-old prepubertal WT (n = 5) and NCoA6^L2m/L2m (n = 5) mice were ovariectomized and treated with placebo or 17β-estradiol for 15 days. As expected, the placebo did not induce any mammary ductal growth. In contrast, the estrogen treatment stimulated a significant mammary ductal growth in both WT and NCoA6^L2m/L2m mice, and the extents of growth in these two groups of mice were similar (Fig. 3B). Furthermore, the expression levels of estrogen/ERα-responsive genes, including PR, cyclin D1, and c-Myc, were comparably decreased in placebo-treated WT and NCoA6^L2m/L2m mice and also comparably induced after estrogen treatment of these ovariectomized mice (Fig. 3C and data not shown). The protein levels of NCoA6, ERα, PRA, and PRB were also similar in the mammary glands of WT and NCoA6^L2m/L2m mice at gestation day 14 when examined by Western blotting (Fig. 3D). Mammary glands at the pregnancy stage were used because they are not influenced by the estrus cycle and had many more epithelial cells from which ERα and PR were expressed. Taken together, these results demonstrate that the L2 motif of NCoA6 is not essential for estrogen/ERα-regulated gene expression and mammary gland development.

FIG. 3.

NCoA6^L2m/L2m mice exhibit normal steroid hormone-regulated mammary gland development. (A) Comparison of mammary gland morphogenesis between WT and NCoA6^L2m/L2m mice. The whole-mounted mouse mammary glands were prepared from WT and NCoA6^L2m/L2m female littermates at different developmental stages, including the 40-day-old virgin stage, gestation day 15, and lactation day 3. Five mice were examined for each stage per genotype group, and the morphologies of representative pairs are presented. LN, mammary gland lymph node. (B) Comparison of estrogen-stimulated mammary ductal growth in WT and NCoA6^L2m/L2m mice. WT and NCoA6^L2m/L2m female littermates (five in each group) were ovariectomized at the age of 17 days and then treated with either placebo or slow-releasing 17β-estradiol (E2) pellets for 15 days. The whole-mounted mammary glands were stained with carmine alum. Mammary glands with representative morphologies are presented. (C) Real-time RT-PCR analysis of PR mRNA expression. Intact WT and NCoA6^L2m/L2m female adult virgin mice without any treatment or WT and NCoA6^L2m/L2m female mice that were ovariectomized and treated with placebo or E2 were used. Each group had five mice. RNA samples were prepared from the third pairs of mammary glands for real-time RT-PCR analysis. The relative expression levels of PR mRNA were normalized to the 18S RNA levels. No statistical differences (P > 0.05) were observed between WT and NCoA6^L2m/L2m groups. (D) Western blot analysis. Fifty μg of protein isolated from the mammary glands of three pairs of WT and L2m littermates on gestation day 14 were used for Western blotting, and the blots were reacted with antibodies against NCoA6, ERα, a common region of PRA and PRB, and β-actin. β-Actin served as a loading control. Band intensity was quantified by densitometry. The average band intensities showed no statistical differences between the WT and NCoA6^L2m/L2m groups.

Mutation of the LXXLL-2 motif does not affect ERα but attenuates LXRα transcriptional function.

To directly evaluate the effects of L2m mutant on ERα- and LXRα-mediated transactivation, we prepared primary WT and NCoA6^L2m/L2m MEFs at early passages from mouse embryos and performed coimmunoprecipitation and reporter-based transactivation assays. In the WT and NCoA6^L2m/L2m MEFs, the endogenous levels of NCoA6 and NCoA6-L2m proteins were similar and were not affected by estrogen or T0901317 treatment (Fig. 4A). The levels of eptopically expressed ERα and LXRα were also similar in WT and NCoA6^L2m/L2m MEFs (Fig. 4A and data not shown). In MEFs transfected with ERα and treated with 17β-estradiol, co-IP assays revealed that ERα was associated with both NCoA6 in WT MEFs and NCoA6-L2m in NCoA6^L2m MEFs (Fig. 4B). Accordingly, the estrogen-induced increase in ERE-Luc reporter expression mediated by ERα was similar in both WT and NCoA6^L2m/L2m MEFs (Fig. 4C). These results suggest that the L2 motif is not required for NCoA6 to interact and coactivate ERα in MEFs. In contrast, in MEFs transfected with LXRα and treated with T0901317, coimmunoprecipitation assays revealed that LXRα was associated with NCoA6 in WT MEFs but not NCoA6-L2m in NCoA6^L2m/L2m MEFs (Fig. 4D). The transcriptional function of LXRα was severely impaired in NCoA6^L2m/L2m MEFs compared with WT MEFs. The LXRα reporter expression was increased 2.8-fold in WT MEFs treated with T0901317, but it was not significantly increased in NCoA6^L2m/L2m MEFs treated with T0901317 (Fig. 4E). These results suggest that the L2 motif of NCoA6 is required for LXRα-mediated gene transcription in MEFs.

FIG. 4.

Differential effects of the L2m mutation on ERα and LXRα. (A) Western blot analysis. WT and NCoA6^L2m/L2m primary MEFs were transfected with LXRα expression plasmids and LXRE-tk-Luc reporter plasmids and treated with either vehicle or 10⁻⁶ M T0901317. Cell lysates with 50 μg protein were used in a Western blot analysis using antibodies against NCoA6, LXRα, and β-actin. The relative levels of endogenous NCoA6 in WT MEFs and NCoA6-L2m in NCoA6^L2m/L2m MEFs were comparable when normalized to β-actin. The LXRα levels were also similar regardless of T0901317 treatment. (B) Coimmunoprecipitation of NCoA6 and ERα. WT and L2m primary MEFs were transfected with ERα and treated with 17β-estradiol (E2; 10⁻⁸ M). Cell lysates were subjected to IP using NCoA6 antibody (α-NCoA6) or control IgG. Immunoprecipitates were separated by SDS-PAGE. Immunoblotting (IB) analysis was performed using NCoA6 and ERα antibodies. (C) ERα transfection assay. Primary WT and NCoA6^L2m/L2m MEFs were transfected with ERα, ERE-Luc, and pCMV-RL (Renilla luciferase) plasmids and treated with vehicle (V) or E2 for 24 h. The luciferase activity was assayed and normalized to RL activity. Data were obtained from three independent experiments using different batches of MEFs derived from littermate embryos. Data from three assays are presented as means ± standard deviations (SD). (D) Coimmunoprecipitation of NCoA6 and LXRα. WT and L2m primary MEFs were transfected with LXRα and treated with T0901317 (T; 10⁻⁶ M). Cell lysates were subjected to IP with NCoA6 antibody (α-NCoA6) or control IgG. Immunoprecipitates were subjected to IB analysis using NCoA6 and LXRα antibodies. (E) LXRα transfection assay. WT and NCoA6^L2m/L2m primary MEFs were transfected with LXRα, LXRE-tk-Luc reporter, and pCMV-RL plasmids and treated with vehicle (V) or T0901317 (T) for 24 h. The luciferase activity was assayed and normalized to RL activity. Data from three independent experiments are presented as means ± SD. ***, P < 0.001 by unpaired t test. (F) In vivo hepatocyte transfection assay for LXRα transactivation. LXRα, RXRα, LXRE-tk-luc, and pCMV-RL plasmids were mixed with the in vivo transfection reagent and injected into the tail vein of WT and NCoA6^L2m/L2m mice (n = 5 in each group). After 2 h, mice were orally administrated vehicle (V) or T0901317 (T). Mice were sacrificed 24 h post-plasmid injection. Luciferase activity in liver lysates was assayed and normalized to RL activity. Data are presented as means ± SD. ***, P < 0.001 by unpaired t test.

The impaired LXRα transcriptional function in the hepatocytes of NCoA6^L2m/L2m mice was further confirmed by in vivo transfection assay using the TransIT gene delivery system. In the hepatocytes of WT mice, the transcriptional activity of LXRα was induced 17-fold by T0901317 treatment. However, it was induced only 7-fold by T0901317 in the hepatocytes of NCoA6^L2m/L2m mice (Fig. 4F).

Mutation of the LXXLL-2 motif reduces LXRα target gene expression and lipogenesis in the liver.

LXRα plays an important role in enhancing hepatic lipogenesis through up-regulation of SREBP-1c and FAS expression (13, 38, 45). The increase in SREBP-1c expression subsequently induces multiple lipogenic genes in the liver, including FAS, LPL, and SCD-1 (13). In addition, LXRα up-regulates Cyp7A1 expression in the liver, which controls bile acid synthesis from cholesterol (32). To define the contribution of the NCoA6 L2 motif to LXRα-mediated endogenous gene expression, we treated WT and NCoA6^L2m/L2m mice with T0901317 for six consecutive days to activate the LXRα-mediated lipogenic pathways and measured the expression levels of the hepatic lipogenic genes mentioned above. The average expression levels of the LXRα and LXRβ mRNAs were similar in the livers of WT and NCoA6^L2m/L2m mice regardless of T0901317 treatment (Fig. 5A). In the placebo-treated WT and NCoA6^L2m/L2m mice, the average expression levels of SREBP-1c, FAS, LPL, SCD-1, and Cyp7A1 were also comparable. After LXRα was activated by T0901317 treatment, the expression levels of these LXRα-regulated genes were robustly induced in WT mice (Fig. 5A). Importantly, the induction of SREBP-1c, LPL, SCD-1, and Cyp7A1 expression by T0901317 treatment was largely abolished and the induction of FAS expression by T0901317 was significantly reduced in NCoA6^L2m/L2m mice compared with WT mice (Fig. 5A). These results demonstrate that the L2 motif is required for NCoA6 to mediate the transcriptional function of LXRα in the liver.

FIG. 5.

LXRα-mediated lipogenesis in the liver of NCoA6^L2m/L2m mice is impaired. (A) Real-time RT-PCR analysis of T0901317-induced gene expression in the liver. RNA samples were prepared from the livers of vehicle-treated (V) or T0901317-treated (T) WT and NCoA6^L2m/L2m male mice (five in each group). Relative concentrations of the LXRα, LXRβ, SREBP-1c, FAS, LPL, SCD-1, and Cyp7A1 mRNAs were measured by real-time RT-PCR, and their relative expression levels were normalized to the 18S RNA concentrations in the respective samples. Data are presented as means ± standard deviations (SD). *, P < 0.05; **, P < 0.01; ***, P < 0.001. (B) Hematoxylin and eosin-stained liver sections prepared from vehicle- or T0901317-treated WT and NCoA6^L2m/L2m mice. A higher number of unstained intracellular vacuoles (lipid droplets, indicated by green asterisks) were visible in WT liver sections compared with NCoA6^L2m/L2m liver sections after T0901317 treatment (right upper panel). Bar, 50 μm. (C) Oil Red O-stained liver cryosections prepared from vehicle- or T0901317-treated WT and NCoA6^L2m/L2m mice. The Oil Red O-stained lipid droplets were much larger in WT liver sections compared with that in NCoA6^L2m/L2m liver sections after T0901317 treatment (right upper panel). Bar, 25 μm. D. NCoA6^L2m/L2m mice have lower serum and hepatic triglycerides after T0901317 treatment. WT and NCoA6^L2m/L2m mice (five each) were treated with vehicle (V) or T0901317 (T). Triglycerides were measured by the enzymatic method. Data are presented as means ± SD. *, P < 0.05; ***, P < 0.001.

Next, we examined the effect of L2m mutation on LXRα-regulated lipogenesis in the liver. In the H&E-stained liver sections prepared through a paraffin-embedding procedure, the lipid droplets were displayed as empty vacuoles because their lipid components were extracted in organic solvents. As expected, the lipid droplets were barely observed in the H&E-stained liver sections of placebo-treated WT and NCoA6^L2m/L2m mice. After WT mice were treated with T0901317, numerous intracellular lipid droplets were observed in the H&E-stained liver sections. Importantly, both the average number per hepatocyte and the average size of the intracellular droplets were remarkably reduced in the H&E-stained liver sections of NCoA6^L2m/L2m mice treated with T0901317 (Fig. 5B). These results suggest that disruption of the NCoA6 L2 motif attenuates the LXRα-mediated liver lipogenesis. This was confirmed by further examination of the lipid droplets by Oil Red O staining of the liver cryosections. The Oil Red O staining detected only a few lipid droplets in the liver of placebo-treated WT and NCoA6^L2/m/L2m mice. In contrast, many lipid droplets with various sizes were detected in the liver of T0901317-treated WT mice. However, the lipid droplets detected in the liver of T0901317-treated NCoA6^L2m/L2m mice were reduced in both number and size (Fig. 5C). Quantitatively, the Oil Red O-stained areas in the liver of vehicle-treated WT and NCoA6^L2m/L2m mice showed no significant difference, with values of 0.12% ± 0.06% (n = 5) and 0.05% ± 0.05% (n = 5) of total areas measured. The Oil Red O-stained areas in the liver of T0901317-treated WT and NCoA6^L2m/L2m mice were increased to 12.9% ± 1.3% (n = 5) and 8.8% ± 1.1% (n = 5), respectively, and the latter value was significantly lower than the former one (P < 0.01, unpaired t test). In conclusion, these results demonstrate that the L2 motif of NCoA6 is required for LXRα-mediated liver lipogenesis.

Furthermore, since the activation of LXRα and elevation of its direct and indirect target gene expression levels are known to enhance lipogenesis and increase triglyceride (TG) levels in the circulation and liver (8), we measured the serum and hepatic TG levels. The TG concentrations in the serum of placebo-treated WT and NCoA6^L2m/L2m mice were similar, 88.1 ± 11.3 and 83.6 ± 7.3 mg/dl, respectively. The T0901317 treatment increased the serum TG concentration in WT mice to 194.2 ± 10.3 mg/dl, which was a 2.2-fold induction. However, the serum TG concentration in the T0901317-treated NCoA6^L2m/L2m mice was only increased to 141.2 ± 32.5 mg/dl, which was a 1.7-fold induction and was significantly lower than that in the serum of T0901317-treated WT mice (Fig. 5D). Consistent results were also obtained from the liver. The average hepatic TG levels in placebo-treated WT and NCoA6^L2m/L2m mice were 10.0 ± 3.2 and 11.2 ± 3.3 mg/g and showed no statistical difference between them. The average hepatic TG level in the T0901317-treated WT mice was 35.9 ± 6.0 mg/g, a 3.6-fold increase compared with the placebo-treated WT mice. However, the average hepatic TG level in the T0901317-treated NCoA6^L2m/L2m mice was 22.1 ± 6.1 mg/g, which was only a twofold increase compared with the placebo-treated NCoA6^L2m/L2m mice and was significantly lower than the hepatic TG levels observed in the T0901317-treated WT mice (Fig. 5D). These results demonstrate that LXRα-mediated TG synthesis is partially impaired when the L2 motif of NCoA6 is mutated.

Disruption of the NCoA6 LXXLL-2 motif inhibits high-cholesterol-diet-induced hepatic Cyp7A1 expression and fecal bile acid excretion.

Oxysterols, the cholesterol metabolites, serve as physiological agonists for LXRα to induce the expression of Cyp7A1, the rate-limiting enzyme in bile acid synthesis in the liver (5, 32, 46). We showed that T0901317-induced Cyp7A1 expression was attenuated in NCoA6^L2m/L2m liver compared with WT liver (Fig. 5A), suggesting that the NCoA6 L2 motif is required for LXRα-mediated Cyp7A1 expression. To further assess the physiological contribution of the L2 motif in LXRα-regulated Cyp7A1 expression and bile acid synthesis and excretion, we fed 3-month-old WT and NCoA6^L2m/L2m mice (matched littermates) with either a normal or a high-cholesterol (2%) diet for 3 months and compared high-cholesterol-diet-induced Cy7A1 expression and fecal bile acid excretion in these mice. Cyp7A1 expression in the liver of WT mice was induced more than fivefold after feeding with the high-cholesterol diet. In contrast, Cyp7A1 expression was not significantly induced in the liver of NCoA6^L2m/L2m mice fed the high-cholesterol diet (Fig. 6A). The expression levels of LXRα and LXRβ were comparable in the livers of WT and NCoA6^L2m/L2m mice regardless of the cholesterol amounts in their diets (Fig. 6B and C). In agreement with the levels of Cyp7A1 expression in the liver, the total fecal bile acid excretion was similar in WT and NCoA6^L2m/L2m mice fed a normal diet, but the fecal bile acid excretion in NCoA6^L2m/L2m mice was much lower than that in WT mice after a high dietary cholesterol challenge. When WT mice were fed the high-cholesterol diet, their total fecal bile acid excretion was increased 88%. However, when NCoA6^L2m/L2m mice were fed the high-cholesterol diet, their total fecal bile acid excretion was only increased 33%, which was significantly lower than that induced by the high-cholesterol diet for WT mice (Fig. 6D). These results demonstrate that disruption of the L2 motif in NCoA6^L2m/L2m mice inhibits dietary cholesterol-induced and LXRα-mediated Cyp7A1 expression in the liver and thereby significantly attenuates the increase in bile acid synthesis and total fecal bile acid excretion upon an excess dietary cholesterol intake.

FIG. 6.

Reduction in LXR target gene expression and fecal bile acid excretion in NCoA6^L2m/L2m mice fed a high-cholesterol diet. (A to C) Relative expression levels of Cyp7A1, LXRα, and LXRβ mRNAs in the liver. RNA was prepared from the livers of WT and NCoA6^L2m/L2m mice (five in each group) fed either a normal diet (ND) or a high-cholesterol diet (HCD). Relative Cyp7A1, LXRα, and LXRβ mRNA levels were measured by real-time RT-PCR and normalized to the 18S RNA. Data are presented as means ± standard deviations (SD). ***, P < 0.001. (D) NCoA6^L2m/L2m mice have reduced fecal bile acid excretion. Feces were collected from WT and NCoA6^L2m/L2m mice (five in each group) fed ND or HCD. Total fecal bile acid was extracted, measured, and normalized to body weight. Data are presented as means ± SD. *, P < 0.05; **, P < 0.01. (E to G) Relative expression levels of the ABCA1, ABCG5, and ABCG8 mRNAs in the intestine. RNA was prepared from the small intestine of WT and NCoA6^L2m/L2m mice (five in each group) fed ND or HCD. Relative ABCA1, ABCG5, and ABCG8 mRNA levels were measured by real-time RT-PCR and normalized to the 18S RNA. Data are presented as means ± SD. *, P < 0.05; **, P < 0.01.

Mutation of the NCoA6 LXXLL-2 motif results in high-cholesterol-diet-induced hepatic cholesterosis and hypercholesterolemia.

Activation of LXRs also decreases intestinal absorption of cholesterol and plant sterol through induction of ABCA1, ABCG5, and ABCG8 expression (46). These transporters mediate cholesterol efflux from intestinal absorptive cells to the lumen or basolateral compartment (22, 46). The high-cholesterol diet significantly induced ABCA1, ABCG5, and ABCG8 expression in the intestine of WT mice as assayed by qPCR. However, the high-cholesterol diet only partially induced ABCA1, ABCG5, and ABCG8 expression in the intestine of NCoA6^L2m/L2m mice (Fig. 6E to G). These results suggest that the NCoA6 L2 motif is also required for LXR-mediated normal gene expression in the intestine.

Since cholesterol catabolism through bile acid synthesis and fecal bile acid excretion and regulation of cholesterol absorption play inevitable roles in maintenance of cholesterol homeostasis in mice fed a high-cholesterol diet (5, 22, 46), we addressed whether inhibited induction of Cyp7A1, ABCA1, ABCG5, and ABCG8 expression and bile acid synthesis and excretion caused by the L2m mutation of NCoA6 would result in hepatic cholesterosis and hypercholesterolemia. We compared the general liver morphology and histology in WT and NCoA6^L2m/L2m mice fed either a normal or high-cholesterol diet for 3 months. On the normal diet, the relative liver weights of WT and NCoA6^L2m/L2m mice were similar, 3.88% ± 0.11% (n = 4) and 3.74% ± 0.12% (n = 5) of body weight, respectively. However, on the high-cholesterol diet the relative liver weight of NCoA6^L2m/L2m mice (4.41% ± 0.08%, n = 6) was heavier than that of WT mice (4.15% ± 0.08%, n = 5), and the difference was statistically significant (P < 0.001, unpaired t test). The liver color of WT and NCoA6^L2m/L2m mice fed the normal diet was similar. However, after being fed the high-cholesterol diet, the liver color of NCoA6^L2m/L2m mice became much yellower than the liver color of WT mice (Fig. 7A), suggesting a possible accumulation of cholesterol in the liver of NCoA6^L2m/L2m mice. There were no obvious changes in hepatocyte morphology between WT and NCoA6^L2m/L2m mice on the normal diet when H&E-stained sections were examined (Fig. 7B). Nevertheless, after the high dietary cholesterol challenge, both the number and size of the lipid droplets (unstained vacuoles) in the liver sections of NCoA6^L2m/L2m mice were more significantly increased compared with WT mice (Fig. 7B). To validate this observation, cryosections were prepared and Oil Red O staining was performed to estimate the cholesteryl ester contents in the liver of these mice. Indeed, after high dietary cholesterol challenge, the numbers and sizes of the lipid droplets were significantly increased in the liver of NCoA6^L2m/L2m mice compared with the liver of WT mice, while no differences were observed in the liver of these mice fed a normal diet (Fig. 7C). Quantitatively, the ratios of Oil Red O-stained areas to total measured areas in the liver sections of WT and NCoA6^L2m/L2m mice fed the normal diet were 0.9% ± 0.9% (n = 4) and 2.1% ± 0.4% (n = 4), showing no significant difference (P > 0.05, unpaired t test). The Oil Red O-stained areas in WT and NCoA6^L2m/L2m mice fed the high-cholesterol diet increased to 9.5% ± 0.5% (n = 5) and 19.2% ±1.5% (n = 4) of total liver area, respectively. The Oil Red O-stained areas in NCoA6^L2m/L2m mice were significantly larger than that in WT mice under high dietary cholesterol challenge (P < 0.0001, unpaired t test).

FIG. 7.

NCoA6^L2m/L2m mice fed a high-cholesterol diet exhibit impaired cholesterol homeostasis. (A) Photograph of livers of WT and NCoA6^L2m/L2m mice fed either a normal diet (ND) or a high-cholesterol diet (HCD). The livers of NCoA6^L2m/L2m mice fed HCD were bigger and yellower than other livers. (B) H&E-stained liver sections prepared from WT and NCoA6^L2m/L2m mice following the paraffin-embedding procedure. Many more unstained intracellular vacuoles (lipid droplets, indicated by green asterisks) were visible in NCoA6^L2m/L2m liver sections compared with other liver sections (right lower panel). CV, central vein. Bar, 25 μm. (C) Oil Red O-stained liver sections prepared from ND- or HCD-fed WT and NCoA6^L2m/L2m mice by using a cryostat. The Oil Red O-stained lipid droplets are larger in livers of NCoA6^L2m/L2m mice fed HCD (right lower panel). Bar, 50 μm. (D and E) Levels of hepatic (D) and serum (E) cholesterol in WT and NCoA6^L2m/L2m mice fed ND or HCD. Data are presented as means ± standard deviations (n = 5). *, P < 0.05; **, P < 0.01.

We further compared cholesterol levels in the liver and serum of NCoA6^L2m/L2m and WT mice by quantitative measurement. On the normal diet, the average cholesterol levels in the liver and serum of WT and NCoA6^L2m/L2m mice were comparable (Fig. 7D). After high dietary cholesterol challenge, the hepatic cholesterol was increased 82% in WT mice and 160% in NCoA6^L2m/L2m mice (Fig. 7D). Under the challenge of high dietary cholesterol, the serum cholesterol concentration was maintained at normal levels in WT mice, but it was increased about 25% in NCoA6^L2m/L2m mice. The average level of serum cholesterol in NCoA6^L2m/L2m mice was statistically higher than that in WT mice after feeding them the high-cholesterol diet (Fig. 7E). Taken together, these results demonstrate that mutation of the L2 motif of NCoA6 results in hepatic cholesterosis and hypercholesterolemia when the mutant mice are challenged with a high-cholesterol diet, which most likely is a consequence of impaired Cyp7A1 and other gene induction levels mediated by LXRs and NCoA6.

DISCUSSION

To date, more than 250 nuclear receptor coregulators have been identified and characterized, mostly by biochemical and cell culture-based experiments (31). Although these studies have provided basic conceptual models concerning the functional mechanisms of NRs and coregulators, these results may not be easily extrapolated to physiological conditions due to inherent limitations of in vitro experiments performed under nonphysiological conditions. Therefore, mutant mice are valuable models to explore the biological functions of these coregulators. Many gene-targeted mouse models for nuclear receptor coregulators have been reported, and the characterizations of these models have demonstrated the pleiotropic physiological functions of NR coregulators. In the case of the p160 SRC gene family, mice lacking any one of the three family members are viable and exhibit different phenotypes from one to another mutant line, but for mice lacking two of the three family members the effects are partially or completely lethal, suggesting that these family members have both specific and redundant physiological functions (41). In contrast, for mice with disruption of the single NCoA6 coactivator the effects are lethal, indicating that NCoA6 is an essential coactivator for embryonic development (3, 18, 23, 51). In adult mice, NCoA6 is widely expressed in different tissues, including liver and intestine (48). Unlike the p160 SRC family members containing six or seven LXXLL motifs (41), NCoA6 contains only two of this kind of motif. Its N-terminal L1 motif interacts with a number of NRs except for LXRs, while its C-terminal L2 motif mainly interacts with ERα and LXRs (4, 9, 15, 16, 20, 21, 24, 25, 50). Based on these unique structural and functional features, we selected NCoA6 as a coactivator model to investigate the tissue- and NR-specific roles of the LXXLL motifs using gene-targeted knock-in mice.

We have previously shown that NCoA6 null mice die in the uterus (18). The present study showed that NCoA6^L2m/L2m mice developed and grew normally, indicating that the L2 motif is not required for the essential function of NCoA6 in embryonic development and growth. The viable phenotype of NCoA6^L2m/L2m mice is consistent with the biochemical data showing that the L2 motif mainly interacts with ERα and LXRα (15, 24, 25), because both ERα and LXRα null mice exhibit normal embryonic development and viability (6, 32). Our analysis also revealed that NCoA6^L2m/L2m mice exhibit normal female reproductive capability and mammary gland development, which are events regulated by ovarian steroids and their receptors. This is consistent with the normal expression of ERα and its target PR in the mammary gland, and it is also consistent with the normal interaction between NCoA6-L2m mutant and ERα in MEFs. In addition, previous studies have demonstrated that the affinity between the L2 motif and ERα was much lower than the affinity between the L1 motif and ERα (24). These results indicate that the L2 motif is not required for NCoA6 to mediate ERα function in mammary gland development. In the future, it will be interesting to generate and characterize L1 mutant mice to discern whether the L1 motif is essential for NCoA6-mediated development in the embryo and ERα function in the adult mammary gland.

The LXR family contains LXRα and LXRβ. LXRβ is ubiquitously expressed, while LXRα is highly expressed in the liver and at lower levels in the intestine and other lipid metabolism tissues (35). It has been shown that LXRα null mice exhibit massive cholesteryl ester accumulation in their livers when challenged with high dietary cholesterol but LXRβ null mice do not show any obvious hepatic phenotype under the same challenge (2, 32). Therefore, LXRα plays a dominant role over LXRβ in cholesterol metabolism and homeostasis. The major role of the activated LXRα in the liver is to enhance lipogenesis and bile acid synthesis. Its role in lipogenesis is through direct induction of SREBP1c, FAS, and LPL expression, followed by the SREBP1c-dependent expression of FAS, LPL, ACC, and SCD-1 (13, 34, 38, 45, 49). Its role in bile acid synthesis is through stimulation of the rate-limiting enzyme Cyp7A1 (5, 32, 46).

The present study clearly demonstrates an essential role of the L2 motif in NCoA6 and LXRα interactions and their function. We showed that mutation of the L2 motif in MEFs diminished the interaction between NCoA6 and LXRα and the coactivation of LXRα by NCoA6, suggesting that the L2 motif is required for NCoA6 to interact with and coactivate LXRα. These results are consistent with those obtained previously from a different in vitro experimental system (15). Furthermore, the ligand-induced transcriptional activity of LXRs is compromised in NCoA6^L2m/L2m mice, as evidenced by the in vivo hepatocyte transfection assay and by measurement of LXR target gene expression in the liver and intestine.

Because the L2m mutation of NCoA6 impaired LXR transcriptional activity, the T0901317-induced expression of the direct and indirect LXRα target genes in the liver, including SREBP1c, FAS, LPL, ACC, and SCD-1, was either abolished or partially reduced in NCoA6^L2m/L2m mice. Since these LXR target genes are key enzymes for liver lipogenesis, their insufficient induction by the ligand-activated LXRα should be responsible for the reduction of hepatic lipogenesis and serum TG levels in NCoA6^L2m/L2m mice treated with T0901317.

In addition to the role of LXRα in hepatic lipogenesis, LXRα serves as an intracellular nuclear sensor of dietary cholesterol load to maintain cholesterol homeostasis through regulation of bile acid synthesis and excretion (46). The rate-limiting enzyme Cyp7A1 for bile acid synthesis is a direct target gene of LXRα, and inactivation of LXRα in mice blocks the increase in oxysterol-induced bile acid synthesis and excretion and leads to severe cholesterol accumulation in the liver and blood after mice are challenged with high dietary cholesterol (32, 46). Similarly, because of the reduction of LXR transcriptional activity caused by the NCoA6 L2m mutation, the high-cholesterol diet was unable to induce sufficient Cyp7A1 expression in the liver of NCoA6^L2m/L2m mice, resulting in a decrease in fecal bile acid excretion and a massive accumulation of cholesteryl ester in the liver and serum of NCoA6^L2m/L2m mice fed the high-cholesterol diet. In addition, the lower induction by the high-cholesterol diet of ABCA1, ABCG5, and ABCG8 expression in the intestine of NCoA6^L2m/L2m mice could reduce cholesterol efflux to the lumen and contribute to cholesterol accumulation in these mutant mice. These findings indicate that NCoA6 is required for LXRs to regulate bile acid synthesis and excretion and cholesterol absorption in response to dietary cholesterol load and that the role of NCoA6 in LXR function is mediated by its L2 motif.

Transgenic mouse lines that ubiquitously overexpress NCoA6 fragments spanning amino acid residues 849 to 929 (DN1) or 1431 to 1511 (DN2) have been previously reported (14, 15). DN1 and DN2 fragments contain the L1 and L2 motifs, respectively. These fragments are believed to play dominant negative roles to inhibit the interaction between NRs and NCoA6 (14, 15). DN1 transgenic mice exhibit various defects different from those observed in NCoA6 null mice (14, 18). The DN2 transgenic mice display disorders in lipogenesis and cholesterol homeostasis (15). While DN1 and DN2 mouse models have provided valuable information for understanding the in vivo role of NCoA6, these types of models suffer from their inherent limitations. As pointed out by the same laboratory, the DN1 fragment could block not only NCoA6 function but also the function of other essential NR box-containing coactivators (19). In addition, the ubiquitously overexpressed DN fragments might cause other effects unrelevant to the in vivo function of NCoA6. Therefore, it is necessary to reexamine and confirm the biological function of these LXXLL motifs in unbiased molecular genetic mouse models with specific knock-in mutations.

In comparison with previously reported LXRα null mice and DN2 transgenic mice (15, 32), NCoA6^L2m/L2m mice showed a less severe phenotype of T0901317-induced liver lipogenesis and high-cholesterol-diet-induced cholesterol accumulation in the liver and serum. In LXRα null mice, the LXRα function was completely lost (32). In DN2 transgenic mice, the overexpressed DN2 fragment might interfere with the interaction of LXRα with NCoA6 and other coactivators, since other coactivators, including TRRAP, PGC-1α, SRC-1, and SRC-3, have been shown to interact with LXRα in the presence of T0901317 (1, 30, 39). In NCoA6^L2m/L2m mice, only the L2 motif of NCoA6 was mutated, and this mutation should not affect the interaction of other coactivators with LXRα. Therefore, the partial phenotype of NCoA6^L2m/L2m mice may be explained by partial compensation of LXR transcriptional activity from other coactivators when NCoA6-L2m failed to coactivate LXR in these mice. In addition, the different strain backgrounds of these mouse lines might also make some contribution to the phenotype severity of these mutant mice, since mouse genetic background has been reported to affect cholesterol absorption (12).

In summary, we have generated the NCoA6-L2m mutant mouse line using a targeted knock-in strategy. Through characterization of the NCoA6^L2m/L2m mouse model, we have demonstrated that the L2 motif of NCoA6 is not required for estrogen/ERα- and progesterone/PR-regulated mammary gland morphogenesis and function, but it is required for LXR-mediated gene expression, lipogenesis, bile acid synthesis, and excretion and cholesterol homeostasis in mice. Since the L2 motif of NCoA6 is required for NCoA6 to help LXR to sense dietary cholesterol intake and to get rid of excess cholesterol through bile acid synthesis and excretion, loss-of-function mutations of the L2 motif of NCoA6 or the entire NCoA6 molecule may be implicated in human diseases, including hepatocirrhosis and atherosclerosis.