Intestine-specific expression of MOGAT2 partially restores metabolic efficiency in Mogat2-deficient mice.

Acyl CoA:monoacylglycerol acyltransferase (MGAT) catalyzes the resynthesis of triacylglycerol, a crucial step in the absorption of dietary fat. Mice lacking the gene Mogat2, which codes for an MGAT highly expressed in the small intestine, are resistant to obesity and other metabolic disorders induced by high-fat feeding. Interestingly, these Mogat2−/− mice absorb normal amounts of dietary fat but exhibit a reduced rate of fat absorption, increased energy expenditure, decreased respiratory exchange ratio, and impaired metabolic efficiency. MGAT2 is expressed in tissues besides intestine. To test the hypothesis that intestinal MGAT2 enhances metabolic efficiency and promotes the storage of metabolic fuels, we introduced the human MOGAT2 gene driven by the intestine-specific villin promoter into Mogat2−/− mice. We found that the expression of MOGAT2 in the intestine increased intestinal MGAT activity, restored fat absorption rate, partially corrected energy expenditure, and promoted weight gain upon high-fat feeding. However, the changes in respiratory exchange ratio were not reverted, and the recoveries in metabolic efficiency and weight gain were incomplete. These data indicate that MGAT2 in the intestine plays an indispensable role in enhancing metabolic efficiency but also raise the possibility that MGAT2 in other tissues may contribute to the regulation of energy metabolism.

Diets, New Brunswick, NJ). In these diets, protein calories (from casein) were held constant at 20%, as were micronutrient and fi ber contents, while fat (lard) replaced carbohydrate (corn starch and sucrose) to increase fat calories from 10 to 45 or 60%. Each diet contained 3.8, 4.7, and 5.2 kcal/g metabolizable energy, respectively (formulas are available at www.researchdiets.com).

Quantitative PCR
Tissue samples from age-matched mice fed a 60 kcal% fat diet for two months were homogenized, and total RNA was extracted and passed through gDNA Eliminator spin columns to remove genomic DNA (RNeasy Plus Mini Kit, Qiagen, Valencia, CA). RNA (2 g) was reverse transcribed (Superscript III Supermix, Invitrogen, Grand Island, NJ), and quantitative PCR was performed with SYBR Green PCR mix (Applied Biosystems, Grand Island, NJ) and analyzed with the ABI PRISM 7000 Sequence Detection System (Applied Biosystems). Relative expression levels were calculated by the comparative C T (cycle of threshold detection) method as outlined in the manufacturer's technical bulletin. Cyclophilin B ( Cypb ) expression was used as an internal control. The 2 -⌬ ⌬ Ct method was used to calculate the fold change in gene expression ( 16 ). The primer pair used to detect both mouse and human MGAT2 mRNAs was designed to anneal to regions where the human and mouse Mogat2 cDNAs share 100% identity; the forward primer sequence is 5 ′ -CAGAA GA TCATG GGC ATCTC-3 ′ , and the reverse primer sequence is 5 ′ -CCAA AG CTGTAC TG G AAGAC-3 ′ . The primer sequences of Cypb gene are 5 ′ TGCCG GAGTCG-ACAATGAT-3 ′ (forward) and 5 ′ -TGG A AG A GCACCAA GACA G-ACA-3 ′ (reverse).

In vitro monoacylglycerol-O -acyltransferase assays
MGAT activity assays were performed with total tissue homogenates as described ( 7 ). Reactions were started by adding total tissue homogenates to the assay mix and were stopped after 2 min by adding 4 ml of chloroform:methanol (2:1, v:v). The extracted lipids were dried, separated by thin-layer chromatography (TLC) on silica gel G-60 TLC plates with hexane:ethyl ether:acetic acid (80:20:1, v:v:v), and visualized with iodine vapor. Products were identifi ed by comparison with the migration of lipid standards. The incorporation of radioactive substrates into lipid products was quantifi ed by an imaging scanner (Typhoon FLA 7000, GE Healthcare Life Sciences, Piscataway, NJ) followed by band scraping and counting in a scintillation counter.

Synthesis of TAG in the small intestine
To examine the uptake and processing of monoacylglycerol, micelles containing radiolabeled tracers were injected into ligated intestine pouches created in anesthetized mice ( 17 ). Taurocholate micelles were prepared by resuspending dried 2-monooleoyl-rac-glycerol [ 14 C (U); ‫ف‬ 15 Ci/mol] and unlabeled oleate with ethanol to fi nal concentrations of 1 and 2 mM, respectively. The mixture was diluted with 20 vol of 10 mM taurocholate in PBS and rotated at 80 rpm for 1 h. Mice were fasted for 4 h. After anesthesia, a longitudinal incision was made just over the midline of the abdomen in the skin and the abdominal wall. To create an isolated pouch in the proximal intestine, a 5-0 silk suture was passed under the small intestine approximately 2 cm inferior to the pyloric sphincter of the stomach and tied off, and another suture was passed and tied off 12 cm inferior to the pyloric sphincter of the stomach. In a similar fashion, an isolated pouch in the distal intestine was created in the terminal 10 cm of the ileum adjacent to the cecum. Of the micelle preparation, 150 µl was injected into each of the pouches. Two minutes later, the proximal and the distal small intestinal pouches were excised. The luminal content was washed out and collected. Lipids from assimilation of dietary fat and to determine the extent to which intestinal MGAT2 contributes to metabolic effi ciency, in this study we examined whether expressing MGAT2 specifi cally in the intestine of Mogat2 Ϫ / Ϫ mice is suffi cient to normalize intestinal lipid processing and systemic energy balance upon high-fat feeding.

Mice
To generate a transgene for overexpressing MGAT2 in the intestine specifi cally, human MGAT2 cDNA was cloned into the p12.4kbVil plasmid containing the 12.4 kilobase regulatory elements of the villin gene (a kind gift of Dr. D. Gumucio, University of Michigan, Ann Arbor, MI), which drives high level expression within the entire intestinal epithelium of transgenic mice ( 15 ). The DNA fragments containing the villin regulatory sequences linked to MGAT2 cDNA ( vil -hMOGAT2 transgene) were released, gel purifi ed, and injected into fertilized eggs of C57Bl/6J mice. Four male founders were crossed with C57Bl/6J mice to generate offspring, and two lines with relatively low or high expression of human MOGAT2 mRNA were selected and named M2 Int-L and M2 Int-H , respectively, for experiments reported in this study.

Mogat2
Ϫ / Ϫ mice in the C57Bl/6J genetic background were generated as reported previously ( 12 ). Male Mogat2 +/ Ϫ mice carrying the hMOGAT2 transgene were crossed with female Mogat2 +/ Ϫ mice without the transgene to generate littermates for experimental groups of each transgenic line of mice (Wild-type, mice with endogenous mMogat2 gene but not the hMOGAT2 transgene; Mogat2 Ϫ / Ϫ , mice without either mMogat2 or hMOGAT2 ; M2 Int-L and M2 Int-H , mice without endogenous mMogat2 gene but with a relative low or high level, respectively, of hMOGAT2 expression). Wild-type and Mogat2 Ϫ / Ϫ mice from both lines were included in the experiments. As differences in their metabolic phenotypes were more pronounced ( 12 ), adult male mice were used for experiments. Mice were housed at 22°C on a 12 h light, 12 h dark cycle. Weighing of mice and changes of diets and cages were performed between 3 and 6 PM. All animal procedures were approved by the University of Wisconsin-Madison Animal Care and Use Committee and were conducted in conformity with the Public Health Service Policy on Humane Care and Use of Laboratory Animals.

Diets
Mice were fed a complete, fi xed-formula diet (#8604, Teklad, Madison, WI). A series of semipurifi ed (defi ned) diets containing 10, 45, or 60% calories from fat were used to examine the effect of replacing carbohydrate with dietary fat on food intake and energy expenditure (D12450B, D12451, and D12492, Research and the nonhydrolyzed groups represented the glycogen measurement. Blood glucose was measured with a glucometer (One-Touch Ultra; Lifescan, Milpitas, CA). Insulin was measured by enzyme-linked immunoassay (Crystal Chem Inc., Downers Grove, IL).

Statistical analyses
All data are presented as mean ± SEM, except for indirect calorimetry data presented in Fig. 4 , in which error bars were omitted to enhance clarity. Each experiment was repeated at least once to confi rm reproducibility of the results. For oxygen consumption per day, differences between genotypes were assessed by the mixed-effects model with repeated measures and adjusted for body weight within the four diet strata using the PROC MIXED procedure in SAS (Cary, NC) ( 24 ). For weight patterns and glucose tolerance, differences between genotype were assessed by repeated measures ANOVA, followed by Tukey's multiple comparison test (Prism 5.01, GraphPad Inc., La Jolla, CA). For all other parameters, differences between genotypes were determined by one-way ANOVA followed by the protected least significant differences technique (Prism 5.01). P < 0.05 was considered statistically signifi cant.

Generation of Mogat2
Ϫ / Ϫ mice expressing hMOGAT2 in the intestine To generate mice expressing MGAT2 in an intestinespecifi c manner, we injected into fertilized eggs from C57Bl/6 mice a transgene linking the cDNA of human MGAT2 ( MOGAT2 ) to the cis -regulatory element of the villin gene ( Fig. 1A ), which drives expression in all intestinal epithelia, beginning by day 12.5 postcoital and throughout life ( 15 ). Using PCR with forward primers specifi c to human or mouse MGAT2 genes and a shared reverse primer ( Fig. 1A ), we identifi ed four founder transgenic mice ( Vil-MOGAT2 ) ( Fig. 1B ). The presence of the transgene in the genome was also confi rmed with Southern blotting (data not shown). These founders were crossed with MGAT2-targeted C57Bl/6 mice ( Mogat2 Ϫ / Ϫ ) to generate four lines of mice expressing MGAT2 only in the intestine ( Vil-MOGAT2 , Mogat2 Ϫ / Ϫ ).
To compare the relative expression levels of human MGAT2 in the transgenic mice to those of endogenous mouse MGAT2 in the wild-type mice, we used a pair of primers annealing to the regions identical between mouse and human cDNA in the quantitative PCR analysis. Two lines of mice, referred to as M2 Int-L and M2 Int-H , were selected for further characterization in this study, as they expressed MGAT2 mRNA at 45% and 160%, respectively, of the wild-type levels in the proximal intestine but not in other tissues, including the liver and reproductive fat pad Int-H lines, respectively; however, the differences between these three genotypes did not reach statistical signifi cance. Low levels of MGAT2 expression were detectable in the kidney of wild-type mice ( ‫ف‬ 3% of levels found in the proximal intestine), while villin the proximal small intestine or the distal small intestine were extracted with chloroform:methanol (2:1, v:v) and separated by TLC using a two-solvent system [fi rst by chloroform:acetone:met hanol:acetic acid:water (50:20:10:10:5) and then by hexane:ethyl ether:acetic acid (80:20:1)]. The bands corresponding to triacylglycerol (TAG), FFA, diacylglycerol (DAG), monoacylglycerol (MAG), and phospholipids (PL) were scrapped after iodine staining. Scraped silica was transferred into scintillation vials for counting of radioactivity. To verify that few radiolabeled tracers were released from the small intestine during the 2 min incubation, blood, liver, and adipose tissues were also collected and their radioactivity was measured.

Absorption of dietary fat
Fat absorption was assessed after an acute challenge of an oil bolus. Mice acclimatized to high-fat feeding for one week were fasted for 6 h and, 45 min before intragastric gavage of a bolus of oil, were injected with 500 mg/kg of the surfactant Pluronic F127 NF Prill poloxamer 407 (a gift from BASF, Florham Park, NJ) to inhibit the clearance of plasma TAG ( 18 ). To assess the rate at which dietary fat entered circulation, blood was collected before (time 0) and at indicated times after mice were challenged with 200 µl of olive oil containing 2 µCi trioleoylglycerol [carboxyl- 14 C] for TAG measurement (Wako Diagnostics, Richmond, VA) and scintillation counting.

Metabolic phenotyping studies
A metabolic phenotyping system (LabMaster modular animal monitoring system, TSE Systems, Chesterfi eld, MO) with housing and wood chip bedding similar to the home-cage environment was used to continuously monitor phenotypes related to acute energy balance, including production of carbon dioxide and consumption of food, drink, and oxygen, as previously described ( 13 ). Male mice (2 Ϫ 3 months old) were acclimated to individual housing and metabolic cages for one week before experiments and were fed indicated diets sequentially for three days each. To examine whether there was a genotype effect on oxygen consumption (VO 2 ), total oxygen consumption per day, calculated by area under the curve, was plotted against baseline body weight for each mouse on each diet. In addition, VO 2 was calculated and analyzed both with and without adjusting for body weight (19)(20)(21).

Biochemical assays
Plasma triacylglycerol was measured by enzymatic assays (L-Type Triglyceride M; Wako Diagnostics) on samples collected in the afternoon after a 6 h fast. Hepatic triacylglycerol of samples collected under the same fasting condition was measured as described ( 22 ). Briefl y, liver samples previously snap-frozen in liquid nitrogen were homogenized in 50 mM Tris-HCl (pH 7.4) and 250 mM sucrose. Lipids were extracted in chloroform:methanol (2:1) and separated by TLC with the solvent system hexane:ethyl ether:acetic acid (80:20:1). TAG was visualized by exposure to iodine, and TAG bands were scraped and analyzed as described using triolein as a standard. Protein concentration was measured by Pierce BCA Protein Assay Kit (Thermo, Rockford, IL). Hepatic glucose and glycogen were measured using an adapted acid hydrolysis method ( 23 ). Briefl y, 40 mg of liver was homogenized in 1 ml of 1 M HCl. Half of the homogenate was neutralized with 0.5 ml of 1 M NaOH (nonhydrolyzed group). The other half was incubated at 95°C for 90 min before being neutralized with NaOH (hydrolyzed group). Glucosyl units released in both groups were measured using a colorimetric glucose kit (Wako Diagnostics). The nonhydrolyzed group represented the free glucose measurement, while the difference between the hydrolyzed activity, such as acyl CoA:diacylglycerol acyltransferase (DGAT)1 ( 26 ), which are highly expressed in the intestine. In contrast, M2 Int-L mice showed 75% of the MGAT activity seen in wild-type mice, whereas there was a 31% increase in MGAT activity in M2 Int-H mice compared with wild-type mice in the proximal intestine ( Fig. 1D ). A similar pattern of differences in MGAT activity between genotypes was also seen in the distal intestine. These data indicate that the hMOGAT2 transgene was expressed in M2 Int-L and M2 Int-H mice and that the expression reinstated MGAT2 activity in the intestine of Mogat2 Ϫ / Ϫ mice.

Intestinal hMOGAT2 restores uptake and esterifi cation of monoacylglycerol
We next examined whether expression of MGAT2 increases the ability of the intestine to take up and esterify monoacylglycerol. To avoid potential effects of gastric emptying and digestion, we directly injected micelles containing 14 C-monooleoylglycerol and fatty acids into ligated pouches that were created in the proximal and distal intestine. We traced the radioactivity 2 min later before any signifi cant amount of dietary fat could be secreted from the intestine.
The levels of 14 C-monoacylglycerol uptake, as indicated by the sum of tracers in four lipid species recovered from the intestine segments, correlated with the levels of MGAT2 expression. For example, the uptake levels were highest in M2 Int-H , followed by wild-type, M2 Int-L , and the lowest in Mogat2 Ϫ / Ϫ mice in the proximal intestine ( Fig. 2 ). These uptake levels were generally higher than their counterparts in the distal intestine, with the exception of the Mogat2 Ϫ / Ϫ mice where uptakes were low in both segments ( Fig. 2 ). In each line of mice, the majority of 14 C-monoacylglycerol promoter moderately increased kidney expression in M2 Int-L and M2 Int-H mice (supplementary Fig. I), which has been reported in other studies ( 15,25 ).
In in vitro acyltransferase assay, Mogat2 Ϫ / Ϫ mice showed MGAT activity at around 30% of wild-type levels in both proximal and distal small intestine. This residual activity presumably came from other enzymes exhibiting MGAT  levels of oxygen as wild-type mice when fed the 10% fat diet. When fed 45% or 60% high-fat diet, they both exhibited oxygen consumption levels that were intermediate between Mogat2 Ϫ / Ϫ and wild-type mice, although the differences compared with either control group, decreases relative to Mogat2 Ϫ / Ϫ mice and increases relative to wild-type mice, did not reach statistical signifi cance ( Fig. 4A ).
Likewise, when total oxygen consumption per day was plotted against the baseline body weight of each mouse, regression lines of Mogat2 Ϫ / Ϫ mice were signifi cantly elevated relative to wild-type controls (data not shown). Based on the mixed-effects model within the three diet strata, Mogat2 Ϫ / Ϫ mice consumed approximately 7-9% more oxygen than did controls across diets at any given body weight. In contrast, both M2 Int-L and M2 Int-H mice exhibited an intermediate phenotype, such that regression lines of each were not signifi cantly different from that of Mogat2 Ϫ / Ϫ mice or wild-type controls.
As more fat replaced carbohydrate in their diets, all four groups of mice showed the expected decreases in overall respiratory exchange ratio (RER), indicating an increased proportion of fatty acid oxidation ( Fig. 4B ). On the low-fat diet, their RERs rose during the night, indicating that dietary carbohydrate was a preferred energy substrate when mice were active and eating ( Fig. 4B, C ). Across diets, RER did not differ among groups during the dark phase of the diurnal cycle when mice were mostly in the postprandial state. In contrast, Mogat2 Ϫ / Ϫ mice showed lower RER compared with wild-type mice during the day, when mice were mostly dormant and fasting ( Fig. 4C ). The differences in RER reached statistical signifi cance when mice were fed taken up was incorporated into triacylglycerol. In the proximal intestine of wild-type mice, we found 93.3% of recovered tracer as triacylglycerol, compared with 1.5% as diacylglycerol, 4.7% as phospholipids, and 0.5% as the tracer monoacylglycerol ( Fig. 2 ). In Mogat2 Ϫ / Ϫ mice, the production of diacylglycerol and triacylglycerol was greatly reduced compared with wild-type levels, illustrating the predominate role of intestinal MGAT2 in catalyzing the process. Expression of hMOGAT2 increased the esterifi cation of monoacylglycerol into di-and triacylglycerol; in M2 Int-L mice, the esterifi cation was partially restored to 91.4% and 51.4% of wild-type levels, respectively, in the proximal intestine, whereas in M2 Int-H mice, the levels exceeded those of wild-type mice in both segments ( Fig. 2 ). In contrast, the amounts of tracer incorporated into phospholipids or remaining as monoacylglycerol did not correlate with the expression levels of intestinal MGAT2. In the distal intestine, corresponding to lower levels of monoacylglycerol uptake, the percentage of tracers recovered as triacylglycerol was lower as compared with their counterpart in the proximal intestine in all genotypes. Nonetheless, the levels of radiolabeled triacylglycerol remained correlative to the levels of MGAT2 expression among genotypes.

Expression of hMOGAT2 in the intestine restores the rate of fat absorption
Mogat2 Ϫ / Ϫ mice absorb normal quantities of dietary fat; however, the rate of fat absorption in these mice is reduced compared with wild-type controls, as the entry of dietary fat into the circulation is delayed ( 12 ). Thus, we examined whether expressing hMOGAT2 in the intestine increases the rate of fat absorption in Mogat2 Ϫ / Ϫ mice. An intragastric bolus of olive oil containing 14 C-trioleoylglycerol was administered after mice were injected with a lipoprotein lipase inhibitor. The labeled dietary fat entered the circulation at a reduced rate in Mogat2 Ϫ / Ϫ mice compared with wild-type mice, as expected ( Fig. 3A ). In contrast, the radio activity increased rapidly over time in both M2 Int-L and M2 Int-H mice to the same extent as in wild-type mice ( Fig. 3A ), indicating that restoration of intestinal MGAT activity at as low as 75% of the wild-type level is suffi cient to reestablish the rate of fat absorption. In all four genotypes, the majority of the radioactivity in blood was found as triacylglycerol ( Fig. 3B ).

Intestinal hMOGAT2 normalizes energy expenditure but not substrate partitioning
Mogat2 Ϫ / Ϫ mice exhibit increased energy expenditure, as evidenced by greater oxygen consumption than in wildtype mice, even when fed a low-fat diet ( 13 ). To determine whether intestinal MGAT2 is suffi cient to prevent the increases in energy expenditure, wild-type, Mogat2 Ϫ / Ϫ , M2 Int-L , and M2 Int-H mice were housed in metabolic chambers and fed sequentially for three days each of the three semipurifi ed diets containing 10, 45, or 60% of calories from fat. Mogat2 Ϫ / Ϫ mice indeed consumed more oxygen than wild-type mice did across the diets, with statistically signifi cant 10, 10, and 12% increases, respectively ( Fig. 4A ). In contrast, both M2 Int-L and M2 Int-H mice consumed similar there were no statistically signifi cant differences in fasting blood glucose levels between them ( Fig. 6C, D ), consistent with the idea that wild-type mice became less sensitive to insulin. When challenged intraperitoneally with a dose of glucose, the rises in blood glucose were greater and lasted longer in wild-type mice than in the other three groups ( Fig. 6D ). The area under the curve of wild-type mice was 14% greater than that of Mogat2 Ϫ / Ϫ mice, although the difference did not reach statistical signifi cance. There were no differences in the response to glucose challenge between Mogat2 Ϫ / Ϫ , M2 Int-L , and M2 Int-H mice ( Fig. 6D ).

DISCUSSION
We previously reported that Mogat2 Ϫ / Ϫ mice exhibit increased energy expenditure and are protected from obesity induced by diets or the agouti mutation, demonstrating a role of the MGAT2 enzyme in the effi cient use and storage of metabolic energy. In this study, we report that MGAT2 the high-fat diets (45 and 60 kcal%). Interestingly, despite partial normalization of oxygen consumption, both M2 Int-L and M2 Int-H mice showed RER profi les that tracked with Mogat2 Ϫ / Ϫ mice, indicating that alterations in the use of carbohydrate versus fat are not corrected . The amounts of total food intake decreased in all mice when they were fed more calorie-dense, high-fat diets ( Fig. 4C ). However, regardless of the fat content, food intake did not differ between genotypes.

Expression of hMOGAT2 partially restores metabolic effi ciency in diet-induced weight gain
We next determined whether expressing MGAT2 only in the intestine is suffi cient in promoting weight gain in response to high-fat feeding. All mice gained weight upon high-fat feeding but not to the same degree ( Fig. 5 ). Whereas wild-type mice gained on average 17.3 g of body weight, Mogat2 Ϫ / Ϫ mice gained only 6.5 g after 10 weeks of high-fat feeding ( Fig. 5 ). As a result, they weighed only 75.2% as much as wild-type mice. In contrast, M2 Int-L and M2 Int-H mice gained 10.8 and 11.9 g, respectively, which was signifi cantly more than Mogat2 Ϫ / Ϫ mice but signifi cantly less than wild-type mice ( Fig. 5 ). Both lines of mice weighed around 88% of wild-type mice after the same duration of high-fat feeding.
After high-fat feeding, wild-type mice, but not Mogat2 Ϫ / Ϫ mice, developed hepatic steatosis as indicated by increases in liver mass and triacylglycerol content ( Fig. 6A , B ). Both M2 Int-L and M2 Int-H mice showed an intermediate phenotype; their livers did not enlarge like those of wild-type mice, but their hepatic triacylglycerol concentrations were higher than that of Mogat2 Ϫ / Ϫ mice. In contrast, there was no signifi cant difference in hepatic glycogen levels among all four groups (data not shown). After high-fat feeding, wildtype mice also showed a higher level of fasting plasma insulin levels than the other three groups of mice, whereas  Int-H mice were measured weekly after switching from chow to 60% diet for 10 weeks. n = 13 Ϫ 26 per group.

of Mogat2
Ϫ / Ϫ mice. They were restored proportionally to the levels of MGAT2 expression when MGAT2 was reintroduced in the intestine of M2 Int-L and M2 Int-H mice ( Figs. 1  and 2 ). Earlier biochemical studies of the intestine suggested that MGAT activity produces diacylglycerol designated for triacylglycerol synthesis ( 27 ). Diacylglycerol produced in the GPAT pathway, on the other hand, is a precursor for both triacylglycerol and phospholipid synthesis ( 28 ). Consistent with these earlier reports, we found that inactivating or reintroducing MGAT2 modulated the amounts of monoacylglycerol incorporated into triacylglycerol ( Fig. 2 ). In contrast, MGAT2 expression did not signifi cantly affect the levels of radioactivity detected in phospholipids, whose production was likely mediated by the GPAT pathway. That the diacylglycerols generated from these two pathways did not enter the same pool is in keeping with the idea that lipid metabolizing enzymes interact with specifi c partners to channel lipid substrates toward different fates.
Our data suggest that MGAT2 controls the rate of monoacylglycerol uptake by the intestine, as the levels of MGAT2 expression determine the cumulative levels of all lipids derived from the 14 C-monoacylglycerol tracer in the intestine. By injecting substrates directly into the intestinal pouches, these results demonstrated that the decreased lipid uptake is a cell-autonomous event due to the absence of MGAT2 in the enterocytes rather than a consequence of systemic effects, such as delayed gastric emptying, which has been reported in Mogat2 Ϫ / Ϫ mice ( 12 ). In wild-type mice, we found more tracer was taken up in the proximal intestine, where MGAT2 expression is higher, than in the distal intestine ( Fig. 2 ). The rest of the tracer remained in the corresponding lumen (data not shown). In Mogat2 Ϫ / Ϫ mice, the amounts of tracer taken up were low in both sections, with more pronounced difference from the wildtype mice in the proximal intestine. Expression of MOGAT2 increased the monoacylglycerol uptake, and the levels of expression correlated with the levels of the uptake in both M2 Int-L and M2 Int-H mice. In contrast, the uptake and distribution of radiolabeled fatty acids were not different between wild-type and Mogat2 Ϫ / Ϫ mice (data not shown), consistent with an intact GPAT pathway in these mice. The mechanism for monoacylglycerol uptake in the intestine remains unclear. A transporter for monoacylglycerol, likely shared with fatty acids, has been proposed, as the uptake of monoacylglycerol is saturable and is hindered by excess fatty acids ( 29 ). Our data suggest that MGAT2 may facilitate uptake by esterifying monoacylglycerol, analogous to how long-chain fatty acyl CoA synthetases facilitate fatty acid uptake.
Consistent with reduced monoacylglycerol uptake and esterifi cation in the intestine, defi ciency of MGAT2 delays the entry of dietary fat into circulation. Reintroducing MGAT2 in the intestine corrected it ( Fig. 3 ). Both M2 Int-L and M2 Int-H mice show rates of fat absorption similar to wild-type mice, despite their differences in the levels of intestinal MGAT activity. This fi nding suggests that MGAT activity at 75% of wild-type levels as exhibited in M2 Int-L mice was suffi cient to facilitate fat absorption. Additional in the intestine regulates the uptake and reesterifi cation of monoacylglycerol and promotes assimilation of dietary fat into body fat in mice. MGAT2 likely serves the same biochemical and physiological functions in humans, as expressing human MOGAT2 , specifi cally in the intestine of Mogat2 Ϫ / Ϫ mice, increased intestinal MGAT activity, restored fat absorption, and partially corrected energy expenditure. After high-fat feeding, both lines of transgenic mice that express MGAT2 only in the intestine gained more weight than their Mogat2 Ϫ / Ϫ littermates without the transgene. This fi nding indicates that intestinal MGAT2 is suffi cient, to a signifi cant extent, to promote effi cient assimilation of dietary fat. However, these mice did not gain as much weight as their wild-type littermates, raising the possibility that the low levels of MGAT2 expressed in other tissues may also contribute to enhancing metabolic effi ciency.
The best known function of MGAT activity is in the absorption of dietary fat, where MGAT catalyzes the reassembly of digested triacylglycerol ( 2 ). MGAT2 is the major contributor of MGAT activity in the intestine ( 8,12 ). MGAT activity in vitro and the ability to incorporate monoacylglycerol into triacylglycerol were signifi cantly reduced in the intestine gain induced by high-fat feeding have been reported in rodents (35)(36)(37). Whether some of the inhibitors are efficacious therapeutic agents for metabolic diseases in humans remains to be determined. In this study, we showed that expressing human MGAT2 in the intestine of Mogat2 Ϫ / Ϫ mice is suffi cient to enhance metabolic effi ciency, providing further evidence that MGAT2-mediated triacylglycerol metabolism in the intestine controls fat absorption and systemic energy balance and that the human enzyme may serve the same functions. These fi ndings raise the prospect of inhibiting MGAT2 in the intestine as an approach to prevent or treat obesity. In addition, we found that reinstituting intestinal MGAT activity only partially restores metabolic effi ciency and does not correct substrate partitioning in Mogat2 Ϫ / Ϫ mice, raising the possibility that MGAT2 in other tissues may play a functional role in the use and storage of energy substrates.

MGAT activity in M2
Int-H mice did not confer a higher rate of fat absorption, as other steps in the formation and secretion of chylomicrons may become limiting. It remains to be determined whether MGAT2 regulates the size and/ or composition of chylomicrons, which may in turn modulate lipid and energy metabolism.
The importance of intestinal MGAT2 in systemic energy balance was illustrated by our fi ndings that expressing MOGAT2 in the intestine is suffi cient to rescue metabolic effi ciency and promote positive energy balance in Mogat2 Ϫ / Ϫ mice. Human MGAT2 shares 81% identity in amino acid sequences with mouse enzyme and, in our studies, carried out similar biochemical and physiological functions to compensate for the mouse enzyme. Thus, MGAT2 likely serves similar physiological functions and enhances metabolic effi ciency in humans. Both M2 Int-L and M2 Int-H mice exhibited a similar reduction in energy expenditure and increases in weight gain in response to high-fat feeding compared with Mogat2 Ϫ / Ϫ mice. However, these recoveries did not reach the wild-type levels and thus appeared to be incomplete. It is possible that human MGAT2 cannot fully substitute for the functions of mouse MGAT2. Likewise, we cannot exclude the possibility that the expression pattern, both location and timing, driven by the villin cisregulatory element differs from endogenous expression of MGAT2 functionally in our study. Because of our fi ndings with indirect calorimetry, we postulate that the incomplete recovery implicates MGAT2 expressed in the extra-intestinal tissues in regulating metabolic effi ciency. We noted that energy expenditure differences were more pronounced during the postprandial state when nutrients are being processed through the intestine; in contrast, the differences in the RER (and thus the relative uses of carbohydrates versus fatty acids) were more pronounced during the postabsorptive state when absorption of nutrients is complete. Both M2 Int-L and M2 Int-H mice showed RER profi les that tracked with Mogat2 Ϫ / Ϫ mice. This fi nding suggests that they too have alterations in substrate partitioning compared with wild-type mice and that their extra-intestinal tissues may have altered energy metabolism like Mogat2 Ϫ / Ϫ mice, leading to partial metabolic ineffi ciency.
The energy balance phenotypes of Mogat2 Ϫ / Ϫ mice are in many aspects similar to mice defi cient in DGAT1, one of the two known DGAT enzymes catalyzing the fi nal step of triacylglycerol synthesis (30)(31)(32). Also associated with a reduction in rate but normal quantity of fat absorption ( 33 ), Dgat1 Ϫ / Ϫ mice exhibit increased energy expenditure, impaired metabolic effi ciency, and resistance to obesity. They are also protected from hepatic steatosis and other metabolic disorders linked to obesity. DGAT1 is highly expressed in the intestine and other tissues, including the adipose tissues. Reintroduction of intestinal DGAT1 using the same villin promoter as used in our study restores the rate of fat absorption and susceptibility to diet-induced hepatic steatosis and obesity, highlighting the role of intestinal lipid metabolism in systemic energy balance ( 34 ). Several inhibitors of DGAT1 have been shown to blunt postprandial increases in plasma triacylglycerol in both humans and rodents. Their effects on preventing weight