The mPlrp2 and mClps genes are involved in the hydrolysis of retinyl esters in the mouse liver.

Retinyl esters are the major chemical forms of vitamin A stored in the liver, and can be delivered to peripheral tissues for conversion into biologically active forms. The function and regulation of the hepatic genes that are potentially involved in catalyzing the hydrolysis of retinyl esters remain unclear. Here we show that two lipid hydrolytic genes, pancreatic-related protein 2 (mPlrp2) and procolipase (mClps), expressed specifically in the mouse pancreas, are associated with the ratio of S-adenosylmethionine (AdoMet) to S-adenosylhomocysteine (AdoHcy). Light illumination deficiency or administration of 5′-AMP elevated the ratio of AdoMet to AdoHcy and induced the expression in the liver of mPlrp2 and mClps, which was blocked by all-trans retinoic acid. Mice fed a vitamin A-free diet exhibited increased activation of hepatic mPlrp2 and mClps expression, which was associated with increased methylation of histone H3K4 residues located near the mPlrp2 and mClps promoters. Inhibition of hepatic mPlrp2 and mClps expression by a methylase inhibitor, methylthioadenosine, markedly decreased plasma retinol levels in these mice. The activated hepatic stellate cell (HSC)-T6 cell line specifically expressed mClps and mPlrp2. Inhibition of mClps gene expressions by short hairpin RNA (shRNA) decreased hydrolysis of retinyl esters in the HSC-T6 cell line. These data suggest that the conditional expression of mPlrp2 and mClps is involved in the hydrolysis of retinyl esters in the mouse liver.

ule as the white cycle for 72 h. For light intensity experiments, animals were transferred from standard light (about 250 lux) to weak light (about 20 lux at the cage cover) from a white light lamp in 12 h/12 h L/D cycles for 72 h. Samples were collected at the time points indicated in the fi gure legends. Light intensity was measured by Laboratory Digital Light Meter CT-1330B (CETO, Nanjing, China).

Western blotting
The protein samples were separated by 10% homogeneous SDS-PAGE before the proteins were transferred to nitrocellulose membranes. After blocking with milk in PBS-Tween (0.1%) buffer overnight at 4°C, the transferred membranes were incubated with PLRP2 antibody (Santa Cruz Biotechnology) for 2 h at 26°C, followed by HRP-conjugated secondary antibody (Santa Cruz Biotechnology) for 1 h at 26°C. The blotting was developed by using ECL detection reagent (Keygen, China) and visualized by autoradiography. HRP-conjugated monoclonal mouse ␤ -actin was introduced as a reference.

Determination of hepatic levels of AdoMet and AdoHcy
Extracts were separated and quantifi ed by using reverse-phase HPLC (Waters 1525 System; Millipore Corp., Bedford, MA) analysis on a Partisphere bonded phase C18 (reverse phase) cartridge column, according to procedures previously described by She et al. ( 22 ). The mobile phase contained 0.1 M sodium acetate, 5 mM heptanesulfonic acid adjusted to pH 4.5 with acetic acid, and 4.2% acetonitrile. The samples were eluted at room temperature with an invariable gradient at a fl ow rate of 0.8 ml/min. Characteristic peak spectra and retention times compared with those of the standard were used to identify AdoMet and AdoHcy. Quantitation was based on peak areas. AdoMet and AdoHcy standards were purchased from Sigma (St. Louis, MO).

Plasma retinol analysis by HPLC
Analysis of plasma retinol levels was performed by reversephase HPLC according to the method described by Teerlink et al. ( 23 ). Plasma was collected in EDTA-coated tubes, isolated, and frozen in the dark at Ϫ 70°C until the determination of retinol concentrations. Retinoid extraction and analytical procedures were carried out in a darkened room under dim yellow light (15 w) to protect retinoids from exposure to light. The extraction procedure involved in single-phase fl uid extraction consisted of adding 350 l of plasma to 50 l of 1.0 M sodium acetate (PTL) and its partner, colipase, have been demonstrated to be major pancreatic REHs in mice. Like classical PTL, the pancreatic lipase-related protein 2 (PLRP2), with strong nucleotide and amino acid sequence homology to PTL, also cleaves triglycerides and retinyl esters, but PLRP2 has broader substrate specifi city ( 15,16 ). Although PLRP2 does not absolutely require colipase, colipase may increase PLRP2 protein activity ( 17 ). In suckling mice, PLRP2 protein is the major colipase-dependent pancreatic lipase ( 18 ).
The murine genes encoding pancreatic-related protein 2 ( mPlrp2 ) and procolipase ( mClps ) are expressed specifically in the pancreas to hydrolyze fat and retinyl esters in the diet ( 19 ). Our recent studies demonstrate that a reduction in light illumination results in upregulated expression of mPlrp2 and mClps in the mouse liver ( 20 ), and that the light-modulated all-trans retinoic acid (ATRA) signal plays a key role in the regulation of the expression of these genes ( 21 ). In this study, we characterized the molecular mechanism by which mPlrp2 and mClps are expressed specifi cally in the pancreas by assessing the difference in the ratio of S -adenosylmethionine (AdoMet) to S -adenosylhomocysteine (AdoHcy) between the pancreas and the liver. A temporary, light-modulated change in the ratio of AdoMet to AdoHcy regulates hepatic mPlrp2 and mClps gene expression in mice. Using mice fed a vitamin A-free diet, hepatic mPlrp2 and mClps gene expression was found to be directly related to the level of retinol in the plasma. The results of the studies strongly suggest that hepatic mPlrp2 and mClps gene expression is associated with the hydrolysis of retinyl esters in the mouse liver.

Animals
Female C57BL/6 mice, 6 to 8 weeks old, were used in this work. Mice were maintained in 12 h light/12 h dark (LD) cycles, with lights on at 7:00 AM and off at 7:00 PM, and were allowed free access to food and water. For the vitamin A-free diet experiment, mice were placed on a vitamin A-free diet for 1 week, and the control animals were given normal chow. The vitamin A-free diet (Laboratory, Nanjing, China) for 100 g food by dry weight was composed of vitamin-free casein (18%), sucrose plus maize starch (68.4%), cellulose (2%), hydrogenated coconut oil (4.6%), salt (4.8%), yeast (2%), and a vitamin integration lacking vitamin A (0.2%). A control group of mice was fed using standard laboratory chow. For consistent observations, all samples were collected at 7:00 PM, unless otherwise indicated. All animal care and use procedures were in accordance with the guidelines of the Institutional Animal Care and Use Committee at Nanjing University of Science & Technology.

Light illumination analysis
For light quality experiments, animals were transferred to closed chambers illuminated by light-emitting diode (LED) illuminators (PARA; Nanjing, China). The intensity of colored light was matched to the intensity of the incident white light source by adjusting the distance between the light and the cage covers (about 250 lux). Colored light was generated by the following LEDs (LED-Z, Burbank, CA): blue (470 nm), green (528 nm), yellow (593 nm), and red (625 nm). Mice were maintained in colored L/D (12 h/12 h) or constant darkness at the same sched-Cell culture HSC-T6 cells were cultured in DMEM (Invitrogen) containing 25 mM glucose, 10% (v/v) FBS, and 100 U/ml penicillin/streptomycin at 37°C/5% CO 2 . HSC-T6 cells were incubated for 12 h with palmitic acid (100 M), after which the cells were treated with 5 M retinol for an additional 12 h, and then shRNAs or the indicated dosages of MTA (4 M and 12 M) were introduced into the cells. Retinyl ester concentrations were determined 12 h later.

PLRP2 immunohistochemistry
To detect PLRP2, slides of paraffi n-embedded liver sections were deparaffi nized and rehydrated. Immunohistochemical staining for PLRP2 with a mouse antibody against PLRP2 (Santa Cruz Biotechnology) was performed according to the manufacturer's protocol.

Statistical analysis
Data are presented as mean ± SEM. Statistical analysis was performed with a Student's t -test. Signifi cance was defi ned as P < 0.05.

Weak light illumination activated the expression of mPlrp2 and mClps in the liver
The lipase gene mPlrp2 and its partner, mClps , are expressed specifi cally in the mouse pancreas under general conditions ( 26 ). Our previous work demonstrated that light exposure inhibits both mPlrp2 and mClps gene expression and that darkness induces their expression in the mouse liver ( 20,21 ). We hypothesized that a specifi c light wavelength or light intensity in illumination was a modulator for mPlrp2 and mClps gene expression. Therefore, we carried out specifi c wavelength illumination studies to determine whether there was a difference in the responses of these genes to light. Confi rming previous observations, no hepatic mPlrp2 and mClps gene expression was observed under LD cycles ( Fig. 1A ). Mice maintained in similar intensities of various wavelengths, including red light, yellow light, blue light, and green light, failed to exhibit mPlrp2 buffer (pH 4.0). After mixing, 600 l acetonitrile was added to the mixture, and the samples were immediately vortex-mixed and centrifuged for 5 min at 12,000 g and 4°C, and 450 l clear supernatant was decanted and diluted with 150 l water. For HPLC analysis, 400 l of this solution was injected. The analysis was performed on a Waters HPLC system (Waters 1525) equipped with a vacuum degasser, quaternary pump, autosampler, column heater, and diode array detector. HPLC analysis was conducted on a 5 m octadecyl silane C 18 column (25 cm × 4.6 mm internal diameter; CEL Associates, Pearland, TX). The mobile phase consisted of acetonitrile-methanol-2% ammonium acetate solutionglacial acetic acid (79:2:16:3) with a fl ow rate of 1.0 ml/min. The temperature of the column was maintained at 30°C. The diode array detector was set at a wavelength of 325 nm to monitor all compounds simultaneously. Characteristic peak spectra and retention times were compared with a retinol standard (Sigma) to identify retinol. Quantitation was based on peak areas.

Treating animals with 5 ′ -AMP, MTA, and ATRA
The indicated dosages of 5 ′ -AMP, methylthioadenosine (MTA), and ATRA (Sigma) were administered by intraperitoneal injections in various groups of mice. After injection, mice were maintained for the desired period of time and then euthanized. Plasma or liver tissue was collected and extracted for analysis of AdoMet and AdoHcy ratios or retinol levels. Total RNA was isolated from liver tissues for RT-PCR analysis.

Chromatin immunoprecipitation assay
The chromatin immunoprecipitation (ChIP) assay was performed as previously described ( 24,25 ), with some modifi cations for the liver. Briefl y, 100 mg of liver tissue was cut into small pieces with a razor blade and cross-linked with 1% formaldehyde for 15 min. Cross-linking was stopped with 0.125 M glycine, and samples were rinsed with 1× PBS containing 1 mM PMSF and disaggregated with a Dounce homogenizer. The nuclei were pelleted and resuspended in lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl at pH 8.1, 0.8 g/ml pepstatin A, 0.6 g/ml leupeptin, and 1 mM PMSF). The suspension was sonicated using a sonicator (KS-130; Ninbo Kesheng, China) to generate DNA fragments averaging 400-500 bp in length and clarifi ed by centrifugation. For immunoprecipitation, the supernatant was diluted 10-fold with dilution buffer (1% Triton X-100, 2 mM EDTA, 150 mM NaCl, and 20 mM Tris-HCl at pH 8.1) and divided into fractions for control IgG and anti-trimethyl-H3K4 (Millipore). Protein A-Sepharose beads, preblocked with 300 g/ml of sheared salmon sperm DNA and 0.1% BSA, were used to precipitate antibody-chromatin complexes. Beads were washed sequentially for 5 min each with TSE buffer (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl, pH8.1) with 150 and 500 mM NaCl, buffer III, and TE (10 mM Tris-HCl, 1 mM EDTA PH = 8.0). Immunocomplexes were eluted, the elution was heated to reverse the formaldehyde cross-linking, and DNA was phenol-chloroform extracted. For PCR analysis of the ChIP samples, purifi ed immunoprecipitates were dissolved in 20 l water.

Preparation of shRNA
The shRNAs targeting rat mClps mRNA were designed and synthesized by GenScript. Sequences used for the shRNAs in the experiments were: shRNA-1, 5 ′ -TGCCAACATGACACCATCCTG-3 ′ ; and shRNA-2, 5 ′ -TGCAGTGTAAGAGCAGATGCT-3 ′ . All of these sequences were shown by BLAST search not to share sequence homology with any rat mRNA. Negative control shRNA was used to assess nonspecifi c gene-silencing effects. The mammalian expression vector pRNA-U6.1 was used to express shRNAs in hepatic stellate cell (HSC)-T6. ShRNAs were introduced into HSC-T6 cells by GenEscort™II, according to the manufacturer's protocol. . This analysis revealed that the ratio of AdoMet to AdoHcy was markedly higher in the pancreas, compared with that found in the liver.

Darkness or weak light illumination elevated the ratio of AdoMet to AdoHcy in livers
To demonstrate that the ratio of Adomet to AdoHcy was associated with hepatic mPlrp2 and mClps expression, we measured AdoMet and AdoHcy in the mouse livers under darkness and weak light illumination. HPLC profi les showed that peak AdoHcy declined and peak AdoMet rose in mouse livers in the DD cycle compared with the normal LD cycle ( Fig. 3A ). A similar change was observed in mice maintained under weak light illumination. Moreover, in DD animals, light exposure elevated the AdoHcy peak to a level comparable to that of LD animals ( Fig. 3A ). Quantitative analysis of peak levels revealed that the ratio of AdoMet to AdoHcy was markedly elevated in mouse livers under darkness compared with light illumination; the elevated level of the ratio in the liver was close to that found in the pancreas, and light exposure in DD mice for 4 h reversibly returned this ratio to the normal levels found in LD animals ( Fig. 3B ). Previous observations demonstrated that light exposure in DD mice inhibited mPlrp2 and mClps expression in the liver ( 20,21 ).

Administration of ATRA lowered the ratio of AdoMet to AdoHcy in DD livers, and administration of 5 ′ -AMP increased the ratio of AdoMet to AdoHcy in LD livers
Our recent studies demonstrate that ATRA has a feedback inhibition to hepatic mPlrp2 and mClps . The lightmodulated ATRA signal regulated ecto-5 ′ -nucleotidase (CD73) gene expression by the RAR ␣ receptor, affecting the circulatory levels of 5 ′ -AMP, resulting in mPlrp2 and mClps expression ( 21 ). Therefore, we carried out experiments to analyze the alteration of hepatic AdoMet and AdoHcy in ATRA-and 5 ′ -AMP-treated animals. RT-PCR analysis showed that administration of 5 ′ -AMP in LD animals activated mPlrp2 and mClps gene expression in the liver ( Fig. 4A ), confi rming a previous Northern analysis; expectedly, LD livers indicated that the ratio of AdoMet to AdoHcy increased notably after 5 ′ -AMP injection ( Fig.  4B ). Next, ATRA (2 nmol/g and 10 nmol/g)-treated mice exhibited reduced mPlrp2 and mClps expression under DD conditions ( Fig. 4C ), and the ratio of AdoMet to AdoHcy and mClps gene expression in the liver, indicating that the inhibition of hepatic mPlrp2 and mClps expression by white light was not due to a specifi c wavelength ( Fig. 1B ). Next, we investigated whether absolute light defi ciency [dark/ dark (DD)] was necessary to induce hepatic mPlrp2 and mClps expression. RT-PCR analysis showed that light intensity below 20 lux activated mPlrp2 and mClps gene expression in the liver ( Fig. 1B ), and no expression of these genes could be determined under high light intensity. These observations suggest that light intensity is a regulatory signal in modulating hepatic mPlrp2 and mClps gene expression. Moreover, Western blotting showed that DD induced PLRP2 protein expression, whereas light exposure for 2 h decreased the protein level, and at 6 h light illumination, PLRP2 almost cannot be detected (see supplementary Fig. I ).

Specifi c expression of mPlrp2 and mClps was associated with the ratio of AdoMet-AdoHcy in the pancreas
Although light intensity plays a dominant role in inducing mPlrp2 and mClps gene expression in the liver, light exposure or darkness failed to induce a signifi cant difference in the expression pattern of either gene in the pancreas. Under normal conditions for maintaining animals, the expression of the mPlrp2 and mClps genes was tissue specifi c and restricted to the pancreas; both gene products functioned in hydrolyzing lipid from diet fat and retinyl esters ( 19 ). The conditional expression of mPlrp2 and mClps in the liver by weak light illumination or darkness was reversible. To identify whether various states of energy metabolism in tissues could result in the specifi c expression of mPlrp2 and mClps , we examined the difference between the pancreas and liver tissues. First, nucleotides were analyzed from extracts of both tissues. There was no signifi cant difference in the levels of ATP, ADP, and AMP between the pancreas and the liver, excluding the possibility that nucleotide metabolism results in the specifi c expression of mPlrp2 and mClps . Then, we compared the ratio of AdoMet to AdoHcy, which represents the potential of methylation in tissue ( 27 ). HPLC profi les showed that the peak of AdoMet was much higher than that of AdoHcy in the pancreas ( Fig. 2A ), whereas liver tissues had no obvious differences in the peaks of AdoMet and AdoHcy ( Fig. 2B ). Next, we quantifi ed the levels of the peaks obtained from the liver and pancreas ( Fig. 2C ,   Fig. 2. Increased ratios of AdoMet and AdoHcy present in the pancreas compared with the liver under LD conditions. A: Representative HPLC profi les of AdoMet and AdoHcy extracted from the pancreas. B: HPLC profi les from the liver. Peak 1 is AdoHcy and peak 2 is AdoMet. The detection wavelength was 260 nm. C: Quantifi cation of AdoMet and AdoHcy in the pancreas and liver with respect to their internal protein concentration. Data are means ± SEM; * P < 0.05; n = 6. correlated with mPlrp2 and mClps expression by the regulation of methylase, we examined the effects of a methylase inhibitor on mPlrp2 and mClps gene expression in the liver. MTA, an effective methylation blocker, was used to treat mice fed the vitamin A-free diet. Figure 6A, B showed that declined in ATRA-treated livers compared with control livers ( Fig. 4D ). These observations implied that the temporary changes of the ratio of AdoMet to AdoHcy in the livers were associated with hepatic mPlrp2 and mClps expression.

Mice fed a vitamin A-free diet activated hepatic mPlrp2 and mClps expression, along with related increases in the methylation of histone H3K4 associated with the mPlrp2 and mClps promoters
The gene products of mPlrp2 and mClps have been demonstrated to function in the hydrolysis of dietary retinyl esters in the gastrointestinal organs ( 19 ), implying the possibility of similar hydrolysis activities for hepatic retinyl esters. To determine whether hepatic mPlrp2 and mClps were related with the hepatic utilization of retinoids, mice were placed on a vitamin A-free diet for 7 days. Surprisingly, hepatic mPlrp2 and mClps mRNA were upregulated strongly in the livers of mice fed a vitamin A-free diet ( Fig. 5A ). Next, we detected the expression levels of other lipases in the liver, such as hepatic lipase, lipoprotein lipase, and carboxyl ester lipase, and found no obvious reproducible changes in these lipases in mice fed the vitamin A-free diet compared with that observed in mice fed a normal diet (data not shown). A subsequent analysis of AdoMet and AdoHcy showed that the ratio of AdoMet to AdoHcy in the liver is elevated in these animals compared with mice fed a normal diet. ( Fig. 5B ). These changes in AdoMet and AdoHcy levels resulted in an increase of total trimethylation of histone H3K4-related liver DNA in mice fed a vitamin A-free diet compared with control mice, and ChIP analysis indicated an increase in the binding of trimethylated H3K4 to mPlrp2 and mClps promoters in the liver DNA of mice fed vitamin A-free diets ( Fig. 5C, D ).

A methylase inhibitor blocked hepatic mPlrp2 and mClps expression and markedly lowered the levels of free retinol in the plasma and impaired hydrolysis of retinyl ester in the HSC-T6 cell line
AdoMet serves as a methyl donor in a series of biomethylation reactions, and the by-product, AdoHcy, is a potent feedback inhibitor of AdoMet-dependent methyltransferase. Therefore, the ratio of AdoMet to AdoHcy refl ects the methylating capacity of tissues ( 27,28 ). To clarify whether the changes of the ratio of AdoMet to AdoHcy are  Results are expressed as double samples, representative of six animals in every group. Gapdh mRNA was monitored as an internal control. B: Quantifi cation of AdoMet and AdoHcy in livers of LD mice treated by 5 ′ -AMP compared with mice treated with saline. An elevated ratio of AdoMet to AdoHcy was observed in mice treated by 5 ′ -AMP. Data are means ± SEM; n = 6; * P < 0.05. C: Injection of ATRA (2 nmol/g and 10 nmol/g) for 4 h inhibited hepatic mPlrp2 and mClps gene expression in DD mice. Data are means ± SEM; n = 6; * P < 0.05, ** P < 0.01, compared with DD. D: Quantifi cation of AdoMet and AdoHcy in the livers of DD mice treated with ATRA, compared with mice treated with oil. Lowered ratios of AdoMet to AdoHcy were observed in mice treated by ATRA. Data are means ± SEM; n = 6; * P < 0.05. PLRP2 immunohistochemistry for establishing the intracellular site of mPlrp2 expression in the liver. As shown in supplementary Fig. II , mPlrp2 is especially expressed in portal tracts (rich HSCs) in the liver.

DISCUSSION
We are endeavoring to understand the basic biochemistry of retinoid metabolism, not only because of its fundamental scientifi c interest, but also for a better understanding of the role that vitamin A nutrition and retinoids play in health and disease. Under normal nutritional conditions, the main site of vitamin A storage is the liver, where over 95% of total retinoids are found as retinyl esters ( 29,30 ). The hydrolysis of retinyl esters is especially crucial in the digestion and intestinal absorption of dietary vitamin A and in the hepatic uptake, storage, and mobilization of vitamin A. Although PTL, PLRP2, and their partner, CLPS, were found to be the main hydrolases of retinyl esters in intestinal juice ( 19 ), the enzymes physiologically involved in hepatic retinyl ester hydrolysis remain unclear. Much early work showed that retinyl palmitate hydrolyzing activities are present in rat liver homogenates and that the activity varied markedly, with over a 50-fold range among the livers of individual rats ( 31 ), raising the question of whether these REH activities detected in liver homogenates actually contributed to retinoids released in physiological livers. A few types of lipases, such as hepatic lipase, carboxylesterases, and lipoprotein lipase, are found in hepatocytes. All of these known enzymes hydrolyze a variety of substrates, including retinoids, and no enzyme has been shown to be absolutely specifi c for retinyl esters ( 11 ).
In our previous work, a reduction in light illumination activated the expression of the pancreas-specifi c lipases mPlrp2 and mClps in the liver ( 20 ), and light-modulated retinoic acid signals regulated circulatory 5 ′ -AMP levels by modulating CD73 activity, switching on mPlrp2 and mClps gene expression in the liver ( 21 ). This study showed that temporary changes in the ratio of AdoMet to AdoHcy in specifi c tissues were associated with mPlrp2 and mClps expression. That correlative relationship between this ratio and mPlrp2 and mClps gene expression was supported by several fi ndings: i ) a higher ratio of AdoMet to AdoHcy in the pancreas compared with the liver; ii ) an elevated hepatic ratio of AdoMet to AdoHcy in darkness or weak light illumination; and iii ) an increasing ratio of AdoMet to AdoHcy after 5 ′ -AMP administration. In these independent studies on the elevation of the ratio of AdoMet to AdoHcy, the mPlrp2 and mClps genes were found to be activated in liver tissues. This sensitive switch to both genes by temporary changes of the ratio of AdoMet to AdoHcy implied specifi c physiological functions. Homeostasis of free retinol in the plasma was controlled by vitamin A uptake from the diet and release of liver retinoid stores. A stable hydrolase activity in the degradation of liver retinoid stores was unreasonable for maintaining a relative constant circulatory level of retinol. The specifi c genes for hydrolyzing hepatic retinyl esters should be regulated by the status of the utilization of vitamin A under physiological conditions. Thus, severe attenuation of hepatic mPlrp2 and mClps gene expression was observed in MTA-treated animals. Subsequent HPLC analysis showed that the peak of retinol from plasma extraction markedly declined in MTA-treated mice compared with control animals treated with saline. Quantitative analysis of peak levels revealed that MTA-treated mice exhibited signifi cantly decreased levels of free retinol in the plasma compared with control animals ( Fig. 6C , n = 5). MTA inhibited mPlrp2 expression and hydrolysis of retinyl ester in a dose-dependent manner in the HSC-T6 cell line ( Fig. 6D, E) , and a linear correlation was found between the expression of mPlrp2 and the levels of retinyl ester. These data indicated that mPlrp2 and mClps are involved in the hydrolysis of retinyl esters in HSCs in the liver and that their expression is regulated by DNA methylation.

HSCs specifi cally expressed mClps and mPlrp2
HSCs are primarily involved in the storage of retinol esters. Interestingly, cultured HSC-T6 cells in the activated state specifi cally expressed the mClps and mPlrp2 genes ( Fig. 7A ), whereas other cell lines, including some hepatic cell lines, failed to exhibit detectable expression of mClps and mPlrp2 . To examine the role of mPlrp2 and mClps in the retinoid metabolism of HSCs, we performed gene-silencing studies in the rat HSC-T6 cell line. As shown in Fig. 7B, C , the knockdown of mClps by 80% in HSC-T6 cells led to the retention of more retinyl esters compared with control and scrambled shRNA-treated cells, indicating a role for this gene in the hydrolysis of retinyl esters associated with the metabolism of retinoids in HSCs. Then we performed Quantifi cation of AdoMet and AdoHcy in the livers of mice fed a vitamin A-free diet compared with mice fed a normal diet. Increased ratios of AdoMet and AdoHcy were observed in vitamin A-free-diet mice. C: Representative PCR profi les, and D, quantitative real-time RT-PCR analysis of the binding of histone H3K4-3Me to the mPlrp2 and mClps promoter sequences in vitamin A-free and control mice, respectively. Data are means ± SEM; n = 6; * P < 0.05. creasing the ratio of AdoMet to AdoHcy. Moreover, mice fed a vitamin A-free diet also exhibited an elevated ratio of AdoMet to AdoHcy, strongly activating hepatic mPlrp2 and mClps expression. We observed an increase in the binding of trimethylated H3K4 to the promoter sequences of mPlrp2 and mClps in the liver of vitamin A-free mice. H3K4 trimethylation is a primary signal that is suffi cient for initiating a gene repression pathway in vivo ( 33 ). In vitamin both lipase genes mPlrp2 and mClps could be optimal candidate genes involved in hepatic hydrolysis of retinol esters.
ATRA has been shown to play a negative role in retinol metabolism in the ferret liver ( 32 ). Our current study demonstrates that in mice exposed to darkness, the ratio of AdoMet to AdoHcy decreased, inhibiting both mPlrp2 and mClps gene expression in the liver, and that ATRA feedback inhibited hepatic mPlrp2 and mClps gene expression by de-  and mClps mRNA in multiple cell lines. B: Total RNA from mClps knockdown HSCs was subjected to realtime RT-PCR analysis. C: Quantifi cation of retinyl esters from knockdown HSCs compared with control and scrambled shRNA. HSC-T6 cells were incubated for 12 h with palmitic acid (100 M), after which the cells were treated with 5 M retinol for an additional 12 h, and then shRNAs were introduced into the cells. Retinyl ester concentrations were determined 12 h later. Results represent means ± SEM from fi ve HSC preparations. * P < 0.05; ** P < 0.01. A-free mice, increased H3K4 trimethylation relaxed this repression of related gene transcription. Moreover, a methylase inhibitor impaired both hepatic mPlrp2 and mClps gene expression, resulting in the decrease of retinol levels in the plasma. Importantly, both of these genes were expressed specifi cally in an activated HSC-T6 cell line, and were only detected in portal tracts in the liver. Knockdown of the expression of these genes decreased retinol loss in vitro. Together, these observations strongly support the conclusion that hepatic mPlrp2 and mClps were the primary genes involved in the hydrolysis of hepatic retinyl esters.
With the regulation of the light-modulated retinoic acid signal, liver tissues displayed a reverse alteration in the ratio of AdoMet to AdoHcy, which was an important metabolic indicator for cellular methylation status, also known as methylation potential ( 27 ). An altered ratio of AdoMet to AdoHcy resulted in protein carboxymethylation ( 34,35 ), inhibiting or activating the expression of specifi c genes. In conclusion, the results of this study demonstrate that temporary changes in the ratio of AdoMet to AdoHcy switch hepatic mPlrp2 and mClps gene expression and are associated with the hydrolysis of hepatic retinyl esters for the utilization of vitamin A in the mouse liver.