Properties of the Mouse Intestinal Acyl-CoA:Monoacylglycerol Acyltransferase, MGAT2*

Acyl-CoA:monoacylglycerol acyltransferase (MGAT) plays an important role in dietary fat absorption by catalyzing a rate-limiting step in the re-synthesis of diacylglycerols in enterocytes. The present study reports further characterization of MGAT2, a newly identified intestinal MGAT (Cao, J., Lockwood, J., Burn, P., and Shi, Y. (2003) J. Biol. Chem. 278, 13860–13866) for its substrate specificity, requirement for lipid cofactors, optimum pH and Mg 2 (cid:1) , and other intrinsic properties. MGAT2 enzyme expressed in COS-7 cells displayed a broad range of substrate specificity toward fatty acyl-CoA derivatives and monoacylglycerols, among which the highest activities were observed with oleoyl-CoA and rac-1-monolauroylglycerol, respectively. MGAT2 appeared to acylate monoacylglycerols containing unsaturated fatty acyls in preference to saturated ones. Lipid cofactors that play roles in signal transduction were shown to modulate MGAT2 activities. In contrast to oleic acid and sphingosine that exhibited inhibitory effects, phosphatidylcholine, phosphatidylserine, and phosphatidic acid stimulated MGAT2 activities. Using recombinant murine MGAT2 expressed in Escherichia coli , we demonstrated conclusively that MGAT2 also possessed an intrinsic acyl-CoA:diacylglycerol acyltransferase (DGAT) activity, which could provide an alternative cells were harvested in ice-cold phosphate-buffered saline, pelleted by centrifugation, homogenized, and assayed immedi-ately or frozen in liquid N 2 for later use. In Vitro Assays for MGAT and DGAT Activity— Cell pellets were homogenized in 20 m M NaCl with three short 10-s pulses from a Brink- mann Polytron. The resultant homogenates were used to assess the activity of MGAT and DGAT in transfected mammalian cells. The protein concentration in homogenates was determined by a BCA Protein Assay Kit (Pierce) according to manufacturer’s instructions. MGAT and DGAT activity was determined at room temperature in a final volume of 100 or 200 (cid:1) l as previously described (28). MGAT activity was determined by measuring the incorporation of [ 14 C]oleoyl moiety into diacylglycerol with [ 14 C]oleoyl-CoA (acyl donor) and various monoacyl- glycerols (acyl acceptors) or by measuring the incorporation of acyl moiety into diacylglycerol with various acyl-CoAs and sn -2-[ 3 H]mono- oleoylglycerol. The incorporation of [ 14 C]oleoyl moiety into trioleoyl-glycerol with [ 14 C]oleoyl-CoA and sn -1,2-dioleoylglycerol was measured to obtain DGAT activity. The acyl acceptors were introduced into the reaction mixture by liposomes prepared with phosphatidylcholine/phos- phatidylserine (molar ratio (cid:3) 1:5). Unless indicated otherwise, the reaction mixture contained 100 m M Tris/HCl, pH 7.0, 5 m M MgCl 2 , 1 mg/ml bovine serum albumin free fatty acids (Sigma), 200 m M sucrose, 20 (cid:1) M of various cold acyl-CoAs or [ 14 C]oleoyl-CoA (50 mCi/mmol), 2 (cid:1) M sn -2-[ 3 H]monooleoylglycerol (60 Ci/mmol) or 200

monoacylglycerols are utilized to sequentially re-synthesize diacylglycerols and triacylglycerols by MGAT and acyl-CoA: diacylglycerol acyltransferase (DGAT), allowing for the lipid to be transported into the circulation system by chylomicron. Therefore, MGAT plays a critical role in intestinal dietary fat absorption. MGAT may also play a role in signaling since its enzymatic product diacylglycerol is an activator of protein kinase C as well as an intermediate in the synthesis of phospholipids (3)(4)(5)(6)(7). Hepatic MGAT activity was found at high levels in suckling rats, diabetic and hibernating animals (8 -10), suggesting that it could be regulated by the high influx of fatty acids. MGAT may also play an important role in systemic regulation of glycerolipid biosynthesis since its sn-2-monoacylglycerol is a competitive inhibitor of glycerol-3-phosphate acyltransferase (11,12). The contribution of MGAT to adipose glycerolipid synthesis is also demonstrated (13).
The biochemical properties of MGAT has been extensively investigated in intestine of various animal species (14 -17) as well as in the intestine and liver of suckling and adult rats (8, 18 -22), kidney (23)(24)(25), and rat adipocytes (13). Because of its association with the microsomal membranes and its alleged involvement in an enzyme complex, MGAT has been difficult to purify to homogeneity (26). Several partial purifications of MGAT from rat intestinal membranes and neonatal liver have been reported previously (17,27). However, properties on a pure MGAT have never been extensively studied because of the lack of a cloned gene encoding the enzyme.
The recent cloning and identification of an intestinal MGAT enzyme, MGAT2 (28), allows us to evaluate the intrinsic characteristics of the enzyme. The mouse MGAT2 is most abundantly expressed in the small intestine (28) where the highest MGAT activity was detected. MGAT2 can catalyze the acylation of each of sn-1-monoacylglycerol, sn-2-monoacylglycerol, and sn-3-monoacylglycerol (28). MGAT2-transfected cells also displayed DGAT activity. However, many biochemical characteristics such as acyl donor and acceptor preference and specificity, pH and magnesium optimum, potential activators and inhibitors, and other intrinsic properties await further investigation. By expression of MGAT2 in mammalian cells as well as in bacterial cells, this study examined these properties of the enzyme, mMGAT2.
In Vitro Assays for MGAT and DGAT Activity-Cell pellets were homogenized in 20 mM NaCl with three short 10-s pulses from a Brinkmann Polytron. The resultant homogenates were used to assess the activity of MGAT and DGAT in transfected mammalian cells. The protein concentration in homogenates was determined by a BCA Protein Assay Kit (Pierce) according to manufacturer's instructions. MGAT and DGAT activity was determined at room temperature in a final volume of 100 or 200 l as previously described (28). MGAT activity was determined by measuring the incorporation of [ 14 C]oleoyl moiety into diacylglycerol with [ 14 C]oleoyl-CoA (acyl donor) and various monoacylglycerols (acyl acceptors) or by measuring the incorporation of acyl moiety into diacylglycerol with various acyl-CoAs and sn-2-[ 3 H]monooleoylglycerol. The incorporation of [ 14 C]oleoyl moiety into trioleoylglycerol with [ 14 C]oleoyl-CoA and sn-1,2-dioleoylglycerol was measured to obtain DGAT activity. The acyl acceptors were introduced into the reaction mixture by liposomes prepared with phosphatidylcholine/phosphatidylserine (molar ratio Ϸ1:5). Unless indicated otherwise, the reaction mixture contained 100 mM Tris/HCl, pH 7.0, 5 mM MgCl 2 , 1 mg/ml bovine serum albumin free fatty acids (Sigma), 200 mM sucrose, 20 M of various cold acyl-CoAs or [ 14 C]oleoyl-CoA (50 mCi/mmol), 2 M sn-2-[ 3 H]monooleoylglycerol (60 Ci/mmol) or 200 M of various cold acyl acceptors, and 50 -100 g of cell homogenate protein or 0.5 g of partially purified protein from Escherichia coli. The indicated concentration of specific phospholipids, detergents, or other inhibitors and activators was delivered into reactions together with substrates. After a 10-20-min incubation at room temperature, lipids were extracted with chloroform/methanol (2:1, v/v). After centrifugation to remove debris, aliquots of the organic phase-containing lipids were dried under a speed vacuum and separated by the Linear-K Preadsorbent TLC plate (Waterman Inc., Clifton, NJ) with hexane:ethyl ether:acetic acid (80: 20:1, v/v/v). The separation was always performed under the conditions where sn-1,2-(2,3)-diacylglycerol and sn-1,3-diacylglycerols were clearly resolved. Individual lipid moieties were identified by standards with exposure to I 2 vapor. The TLC plates were exposed to a PhosphorScreen to assess the formation of 14 C-or 3 H-labeled lipid products. Phosphorimaging signals were visualized using a Storm 860 (Amersham Biosciences) and quantitated using ImageQuant software.
Expression and Purification of mMGAT2 from Escherichia coli-mMGAT2 cDNA was subcloned into an E. coli expression vector for bacterial expression and purification of mouse MGAT2 enzyme. E. coli cell line DH5␣ was transformed with either empty vector (mock-transformed) or vector containing mMGAT2. Two liters of cultures in LB medium were incubated at 37°C until A 600 reached 0.7. Expression of recombinant mMGAT2 was then induced by adding 1 mM isopropyl-␤-D-thiogalactopyranoside for 16 h at 18°C. Cells were homogenized by sonication in an extraction buffer containing 50 mM Tris/HCl, pH 7.5, 150 mM NaCl, 5 mM ␤-mercaptoethanol, and protease inhibitors. The clarified extract was loaded onto a heparin column (Amersham Biosciences) followed by washing the column sequentially with 40 ml of Tris buffer A (50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 10% glycerol, 5 mM mercaptoethanol, and protease inhibitors) and 50 ml of Tris buffer A containing 2 M NaCl. Bound proteins were eluted with a linear gradient of 150 mM to 1.5 M NaCl. The protein was identified with 12.5% SDS-PAGE gel by Coomassie Blue staining and Western blot analysis.
pH and Mg 2ϩ Profiles of MGAT2-The pH profile of the MGAT2 was determined using 100 mM Tris buffers between pH 6.0 and 9.0 with an interval of 0.5. As shown in Fig. 2, A and B (quantitative analysis), the optimal activity of MGAT2 was found at pH 7.0 with a broad profile between pH 7 and 9 with more 1,3-diacylglycerol formed at higher pH values. The effect of MgCl 2 on the activity of MGAT was also examined in this study because it was reported that the addition of magnesium into the reaction affected MGAT and DGAT activities from various tissues or isolated microsomes (13,29,31). The presence of lower concentrations of MgCl 2 (Ͻ30 mM) showed no significant effects on MGAT2 activity, whereas higher concentrations (Ͼ100 mM) of magnesium produced marked inhibition of the reaction (Fig. 2, C and D).
Effects of Detergents on MGAT and DGAT Activities-It has been suggested that some of the detergents may affect the MGAT activity by serving as competitive substrate because of their structural similarities to fatty acids (26). The effects of various detergents on MGAT and DGAT activity from primary tissues such as small intestine (26,29), neonatal liver (27), and adipocytes (29) have been extensively investigated because the presence of detergents is usually necessary to solubilize the enzymes from membranes. However, these experiments were conducted using tissues that may contain several isoforms of each enzyme, making the results difficult to interpret. In this study, we examined the effects of nonionic (Triton X-100), ionic (SDS), or zwitterionic (CHAPS) detergents on MGAT and DGAT activity from MGAT2-and DGAT1-transfected mammalian cells. At a low concentration, none of the detergents posed any measurable effect on DGAT activities. However, Triton X-100, SDS, and CHAPS exhibited severe inhibitory effects on MGAT activity of MGAT2 when the concentration of detergents reached 1.0, 0.1, and 1.0%, respectively (Fig. 3, A and B). The observed DGAT activity from MGAT2-transfected cell homogenate was more liable and sensitive to inactivation by detergents as determined by the formation of triacylglycerol, which started diminishing in the presence of 0.01% Triton X-100, 0.01% SDS, and 0.1% CHAPS, respectively (Fig. 3, A and B). At 0.1% detergent concentration, SDS abolished both MGAT2 and DGAT1 activities and Triton X-100 only diminished the innate DGAT activity of MGAT2 as shown in Fig. 3, A and C, with quantitative analysis shown in Fig. 3, B and D. Surprisingly, at 1.0% concentration, both Triton X-100 and CHAPS significantly enhanced the DGAT activity from DGAT1-transfected cell homogenate. Thus, both Triton X-100 and CHAPS can be used at relatively high concentration (1.0%) to distinguish the MGAT2 activity from that of DGAT1.
Activators and Inhibitors-Bhat et al. (18) reported that anionic phospholipids and anionic lysophospholipids stimulated MGAT activity, whereas fatty acids and sphingosine inhibited enzyme activity derived from neonatal liver microsomes. The addition of phospholipids greatly increased DGAT activity in the lipid body fraction of an oleaginous fungus (32). In contrast, phosphatidylcholine was shown to inhibit triacyl- glycerol synthetase activity derived from intestinal mucosa (26). Such a discrepancy may be attributed to the presence of different isoforms of MGAT in these studies as reported in small intestine and neonatal liver (20). In order to clarify the issue with the MGAT2 enzyme, we characterize the properties of the intestinal MGAT2 using a variety of known inhibitors and activators of MGAT including phospholipids, oleic acid, and sphingosine. Acyl acceptor, sn-2-monooleoylglycerol, in ethanol was delivered into reaction mixture together with the indicated amounts of activators or inhibitors. The amount of ethanol used in the assay was Ͻ1%, which did not affect the enzyme activity. phosphatidylcholine, phosphatidylserine, and phosphatidic acid activated MGAT activity in a dose-dependent manner as shown in Fig. 4A. The three phospholipids showed similar potency of activation. Lysophosphatidic acid displayed a biphasic effect. Whereas lysophosphatidic acid activated the enzyme activity at relatively lower concentrations, a high concentration of lysophosphatidic acid exhibited a marked inhibitory effect (Fig. 4A). In contrast, oleic acid and sphingosine were potent inhibitors for MGAT2 activity (Fig. 4B).
Characteristics of MGAT and DGAT Activity of Partially Purified MGAT2 from Escherichia coli-To further investigate the intrinsic properties of MGAT2 enzyme, we expressed and purified recombinant MGAT2 in E. coli. Bacterial expression offers a unique advantage over the mammalian and insect expression systems since E. coli does not express endogenous MGAT, DGAT, and other synthetic enzymes of triacylglycerols. The cell extract was loaded on a peptide-tagged affinity column and eluted with a linear gradient of 150 mM to 1.5 M NaCl. The eluted fractions were subjected to SDS-PAGE or enzyme assays. The Coomassie Blue-stained SDS-PAGE profile from MGAT2-transformed cells revealed a major polypeptide band migrating at an apparent molecular mass of 38 kDa (Fig. 5A, fraction [27][28][29] that was absent in mock-transformed cells. The 38-kDa peptide is consistent with the molecular mass of 38.6 kDa predicted from the open reading frame of the mouse MGAT2 gene and human MGAT2 expressed in mammalian cells with an apparent molecular mass of 39 kDa on SDS-PAGE (data not shown). We next tested the fractions for both MGAT and DGAT activities. As shown in Fig. 5B, MGAT activity was only detected from fractions that contains the partially purified MGAT2 enzyme but not from similar fractions purified from the mock-transformed E. coli cells. Although the detected MGAT activity is much weaker than that from the MGAT2 enzyme expressed in the mammalian cells, the results suggest that the bacterial expressed enzyme is active. To address the issue of whether the MGAT2 enzyme possesses intrinsic DGAT activity as detected in previous experiments using MGAT expressed in mammalian cells (28), we also tested the purified fractions for DGAT activity. As demonstrated in Fig. 5C, fractions that showed MGAT activity also catalyzed the synthesis of triacylglycerol, thus confirming that MGAT2 enzyme possesses intrinsic DGAT activity. Furthermore, treatment with Triton X-100 abolished the DGAT activity of recombinant enzyme from the bacterial source (Fig. 5D), which is consistent with the property of the MGAT2 enzyme expressed in the mammalian cells shown in Fig. 3A. DISCUSSION Although the biochemical aspects of MGAT have been extensively studied in the last three decades, those efforts have been hindered by the lack of a cloned gene encoding MGAT. Because of the importance of these enzymes, considerable attempts have been made to purify them (17,26,27). However, the purification has proved to be difficult because of their association with the membranes and sensitivity to inactivation of detergents in the purification process. Thus, many experiments were carried out using isolated microsomes that contained both MGAT and DGAT together with other fatty acid biosynthetic enzymes, making the outcome difficult to interpret. Recently, the successful cloning and identification of these enzymes, particularly an intestinal MGAT (MGAT2), has facilitated the further characterization of this important enzyme in dietary fat absorption (28,30,31,(33)(34)(35). Studies employing recombi-nant MGAT enabled us to clarify the reported differences in activities, substrate specificities, stereospecificities, and association with other enzymes and to elucidate its catalytic mechanisms and regulations.
MGAT2 expressed in COS-7 cells displayed a broad range of substrates toward either fatty acyl-CoA derivatives or monoacylglycerols containing various fatty acyl chains. Consistent with the substrate specificity of MGAT1 expressed in insect cells (34) and MGAT activity in human intestinal mucosa (29), MGAT2 preferred fatty acyl-CoAs with longer carbon chains in the range below C18 while the activity began to decline when the length of carbon chains exceeded C20. With regards to monoacylglycerols, MGAT2 appeared to prefer acyls with short carbon chain, contrasting with a reported preference of longer ones (29). MGAT2 acylated monoacylglycerols containing unsaturated fatty acyls, such as essential linoleic acid, in preference to saturated ones. These findings coincide with the concept that the essential fatty acids play important roles in normal development of the retina and brain and thus could be selectively preserved by specific tissues (36,37). Consistent with our findings, human MGAT2 and human MGAT3 were reported to possess the similar substrate specificity pattern (30,35). Variations of substrate specificity among different investigators may be attributed to different resources of enzymes, substrate concentrations, and assay systems employed (29,34,37,38).
The properties of MGAT2 and MGAT1 were similar in several respects but differed from the MGAT activity observed in neonatal rat liver. For example, both MGAT1 and MGAT2 expressed in mammalian cells utilized sn-1-monoacylglycerol and sn-2-monoacylglycerol (28,34) while partially purified MGAT from suckling rat liver acylated 1-monoacylglycerol at only 4% of activity observed with sn-2-monoacylglycerol (27). In addition, MGAT2 expressed in mammalian cells acylated 1-Ohexadecyl-rac-glycerol efficiently (Fig. 1B), whereas the ether analog of monoacylglycerol was catalyzed very poorly by MGAT from rat liver (27). These findings together with previously reported different responses to various treatments including temperature, proteolysis, detergents, protein modification reagents, and divalent cations between liver MGAT and intestine MGAT (20) strongly indicate that liver might express isoforms different from MGAT1 and MGAT2. However, this is not conclusive as the MGAT enzyme may be differently anchored in liver, intestine, and mammalian cell lines or tissue-specific cofactors may exist to modulate the function of MGAT. It was reported that the hepatic isoform could be developmentally up-regulated to meet the need to use fatty acids for energy or to preserve essential fatty acids (20). Further in vivo studies detecting hepatic expression of MGAT1, MGAT2, and MGAT3 during these specific physiological conditions will be very informative in clarifying this issue.
Current reports also investigated possible roles of phospholipids and lipid cofactors in regulating MGAT2 activities. MGAT activities in microsomes has previously been shown to be affected by lipid cofactors that modulate the activity of most membrane-bound enzymes (18,22). Some of the lipid cofactors may serve as coordinators between the monoacylglycerol and glycerol 3-phosphate pathways of triacylglycerol biosynthesis. It remains to be elucidated on the degree of interaction between the two pathways since both routes presumably compete for the same pool of acyl-CoA and both produce diacylglycerol intermediates. Indirectly, the monoacylglycerol route may also affect the rate of fatty acid oxidation by modulating the level of acyl-CoA. Our data show that several metabolic intermediates and their derivatives from the glycerol 3-phosphate pathway impose mildly positive effects on the MGAT activities. In contrast, sphingosine and oleic acid demonstrated a potent inhibitory effect on MGAT2 enzyme activity, which is supported by previous reports (18,22).
It has been proposed that the biosynthesis of triacylglycerols in the enterocyte was catalyzed by a synthetase complex consisting of MGAT, DGAT, and fatty acyl-CoA synthetase (39). This study reports for the first time the properties of an individual recombinant MGAT expressed in E. coli. This E. coli expression system offers a unique advantage to the mammalian expression system since the bacterial system does not express endogenous MGAT and DGAT. Our finding that E. coli expressed MGAT2 is enzymatically active suggests that MGAT2 can function without forming a complex with other biosynthetic enzymes. It should be noted that the activity of MGAT2 expressed in E. coli appears much weaker than in mammalian cells, suggesting the existence of other cofactors for full activation of MGAT2. In addition to the formation of a synthase complex, the requirement of a small cytoplasmic protein for the activation of rat hepatic and intestinal synthesis of triacylglycerol, which has been purified from plasma, was reported previously (40 -43).
We previously reported that MGAT2 expressed in mammalian cell lines also possess weak but significant DGAT activity (28). The same phenomenon was observed in MGAT1 and MGAT3 (30,35). We postulated that this DGAT activity could be due to either the intrinsic properties of MGAT2 or enhanced endogenous DGAT activity by the presence of MGAT2 in mammalian cell lines. To clarify this issue, we analyzed MGAT2 for both MGAT and DGAT activities using purified MGAT2 enzyme expressed in E. coli where MGAT and DGAT background should be negligible. Consistent with what was observed in mammalian systems, the purified MGAT2 from the bacterial resource also displayed weak DGAT activity in addition to the predicted MGAT activity (Fig. 5, B and C). This finding provides convincing evidence that MGAT2 possesses intrinsic DGAT, which is further supported by the observation that treatment with a low dose of nonionic detergent, Triton X-100, abolished the DGAT activity of MGAT2 without affecting the MGAT activity itself. The intrinsic DGAT activity of mouse MGAT2 is not surprising since mouse MGAT2 shares 47.5% sequence homology with DGAT2. Likewise, DGAT2 also possesses a MGAT activity (34). Additionally, the human MGAT2 is closely adjacent to the DGAT2 gene located at 11q13.5, suggesting that the two genes originated from duplication. Hence, it will be interesting to investigate whether DGAT enzymes expressed in the E. coli system also possess intrinsic MGAT activities, and if so, whether DGAT2 enzymes have more MGAT activity than DGAT1, which shares much less homologies with MGAT genes. Such a dual function of intestinal MGAT2 and MGAT3 offers an alternative pathway for triacylglycerol synthesis when DGAT enzymes are deficient. It is possible that overexpression of MGAT2 could have a synergistic effect on the endogenous DGAT activity since they are reported to form a complex with other triacylglycerol biosynthetic enzymes. However, our preliminary experiments demonstrated that the addition of partially purified MGAT2 from E. coli cells did not increase the activity of DGAT overexpressed in COS-7 cell. 2 The interaction among MGAT, DGAT, and other triacylglycerol biosynthetic enzymes such as fatty acyl-CoA synthetase could be addressed in future studies by cotransfection experiments in mammalian cells or other systems.
Detergents have been widely used in solubilizing membraneassociated proteins and protein complexes including MGAT, DGAT, and acyl-CoA acyltransferase from various tissues, although they mostly led to a substantial loss of enzyme activities (17,26,27,29). In view of our results, MGAT activity of MGAT2 was more stable and resistant against treatment with detergents than its intrinsic DGAT activity since 0.1% Triton X-100 or CHAPS did not result in the loss of MGAT activity, whereas they did so with DGAT activity from the same preparation. However, high concentration of detergents (1.0%) led to a substantial loss of activities of MGAT2. It is possible that the nonionic detergents served as competitive inhibitors for the substrate-binding site of MGAT because of their structural similarities with fatty acids or that the lipid-protein interaction is needed for optimal activity of MGAT2. One of the most striking effects of detergent treatment is that both Triton X-100 and CHAPS at high concentration (1%) significantly enhanced the DGAT activity of DGAT1 expressed in COS-7 cells. Although the underlining mechanism remains elusive, it is possible that these detergents could serve as dispersing agents for the hydrophobic diacylglycerols or lead to the exposure of more active sites of DGAT1 to its substrates, therefore increasing the opportunity of interaction between the enzyme and substrates. The finding that MGAT2 also has intrinsic DGAT activity and that nonionic and zwitterionic detergents demonstrated an activating effect on DGAT1 activities may have repercussions on the interpretations of previously published work on MGAT using microsomes that contain both MGAT and DGAT enzyme solubilized with detergents.