Dgat1 and Dgat2 regulate enterocyte triacylglycerol distribution and alter proteins associated with cytoplasmic lipid droplets in response to dietary fat

Abstract

Enterocytes, the absorptive cells of the small intestine, mediate efficient absorption of dietary fat (triacylglycerol, TAG). The digestive products of dietary fat are taken up by enterocytes, re-esterified into TAG, and packaged on chylomicrons (CMs) for secretion into blood or temporarily stored within cytoplasmic lipid droplets (CLDs). Altered enterocyte TAG distribution impacts susceptibility to high fat diet associated diseases, but molecular mechanisms directing TAG toward these fates are unclear. Two enzymes, acyl CoA: diacylglycerol acyltransferase 1 (Dgat1) and Dgat2, catalyze the final, committed step of TAG synthesis within enterocytes. Mice with intestine-specific overexpression of Dgat1 (Dgat1^Int) or Dgat2 (Dgat2^Int), or lack of Dgat1 (Dgat1^−/−), were previously found to have altered intestinal TAG secretion and storage. We hypothesized that varying intestinal Dgat1 and Dgat2 levels alters TAG distribution in subcellular pools for CM synthesis as well as the morphology and proteome of CLDs. To test this we used ultrastructural and proteomic methods to investigate intracellular TAG distribution and CLD-associated proteins in enterocytes from Dgat1^Int, Dgat2^Int, and Dgat1^−/− mice 2 hours after a 200 μl oral olive oil gavage. We found that varying levels of intestinal Dgat1 and Dgat2 altered TAG pools involved in CM assembly and secretion, the number or size of CLDs present in enterocytes, and the enterocyte CLD proteome. Overall, these results support a model where Dgat1 and Dgat2 function coordinately to regulate the process of dietary fat absorption by preferentially synthesizing TAG for incorporation into distinct subcellular TAG pools in enterocytes.

1. Introduction

The highly efficient absorption of dietary fat (triacylglycerol, TAG) is mediated by enterocytes, the absorptive cells of the small intestine. This process is important in the promotion of health as well as in contributing to disease when dysregulated [1]. Within this process, TAG is first hydrolyzed to fatty acids and monoacylglycerol in the intestinal lumen. These digestive products are then taken up by enterocytes and rapidly re-synthesized into TAG at the ER membrane. The resulting TAG is packaged in chylomicrons (CMs) for secretion into blood or in cytoplasmic lipid droplets (CLDs) for temporary storage within enterocytes [2]. The number and size of CLDs increase and then decrease over time in response to dietary fat consumption [3], suggesting that the stored TAG is hydrolyzed at later time points and either re-esterified into TAG at the ER membrane and used for CM synthesis or used for other cellular functions. The mechanism(s) through which TAG is partitioned for CMs and CLDs in enterocytes is unclear.

The partitioning of TAG for CM and CLD synthesis may be mediated by specific TAG synthesis enzymes. Acyl CoA: diacylglycerol acyltransferase 1 (DGAT1) and DGAT2 both catalyze the final, committed step of TAG synthesis (the acylation of diacylglycerol with a fatty acyl-CoA); however, they share no sequence homology, are members of different gene families and exhibit different biochemical, cellular, and physiological functions [4]. Although both DGAT1 and DGAT2 are expressed ubiquitously, they have different tissue expression patterns [5, 6]. Within the cell, both DGAT1 and DGAT2 localize to the ER, but DGAT2 has also been identified on CLDs and mitochondria-associated membranes [7, 8]. DGAT activity occurs on the lumenal and cytosolic sides of the ER membrane[9]. Topological studies demonstrate that the active site of DGAT1 may be present on either side of the ER membrane as the protein has dual topology [10], while the active site of DGAT2 is oriented toward the cytoplasm [11]. However, additional studies investigating the roles of DGAT1 and DGAT2 in hepatocytes suggest that this partitioning may be more complex [12–15]. Furthermore, the distinct physiological roles of Dgat1 and Dgat2 have been demonstrated in knockout mouse models. Dgat1^−/− mice are viable and resistant to diet-induced obesity [16]. Dgat2^−/− mice die within a few hours, likely due to extremely low whole body TAG content and an impaired skin barrier [17]. Together, these differences suggest that Dgat1 and Dgat2 have non-redundant roles, but their specific cellular functions remain unclear.

Targeted studies investigating the role of Dgat1 and Dgat2 in intestinal lipid metabolism in mice indicated that these enzymes synthesize TAG for specific cellular fates within enterocytes and contribute to whole body physiology. Dgat1^−/− mice secrete TAG in smaller-sized CMs at a reduced rate and exhibit increased TAG storage in enterocyte CLDs compared to wild-type (WT) mice [18]. This intestine phenotype was shown to be critical for their resistance to diet-induced obesity and hepatic steatosis by using a mouse model where Dgat1 was only expressed in the intestine [19]. Mice with intestine-specific overexpression of Dgat1 (Dgat1^Int) secrete TAG from the intestine at a similar rate within similar sized CMs compared to WT mice [20]. Interestingly, we were barely able to detect TAG accumulation in enterocytes of Dgat1^Int mice by coherent anti-Stokes Raman scattering microscopy, but by biochemical analysis TAG levels were similar to WT mice [19, 20]. This result suggested that using a more sensitive technique to detect intracellular distribution of TAG, such as transmission electron microscopy (TEM) would be necessary to understand the role of Dgat1 and Dgat2 in synthesizing TAG for specific intracellular fates. Mice with intestine-specific overexpression of Dgat2 (Dgat2^Int) have higher intestinal TAG secretion rates but unchanged CM size and similar TAG accumulation in enterocytes compared to WT mice [20]. The higher TAG secretion rate seen in Dgat2^Int mice contributes to postprandial hypertriglyceridemia and an increased susceptibility to hepatic steatosis [20]. Together, these differences suggest that Dgat1 and Dgat2 have non-redundant roles in dietary fat absorption, but their specific cellular functions in this process requires further investigation.

A major difference within enterocytes of mice with varying levels of Dgat1 and Dgat2 is the storage of TAG within CLDs. Emerging evidence suggests that proteins present on CLDs play important roles in regulating CLD morphology and potentially the balance between TAG storage and secretion [2]. Both targeted and global proteomic approaches have been used to identify proteins associated with enterocyte CLDs that are important in their metabolism [21–25]. These proteins include enzymes involved in lipid synthesis, lipolysis and CM metabolism, as well as other diverse cellular pathways. This localization of these proteins may exert regulatory effects on CM and CLD metabolism directly or indirectly. Determining the differences in proteins that associate with enterocyte CLDs in models with varying levels of Dgat1 and Dgat2 (Dgat1^Int, Dgat2^Int, and Dgat1^−/− mice) will provide further insight into the mechanism through which Dgat1 and Dgat2 synthesize TAG for secretion and storage within enterocytes.

Based on our previous observations in the intestinal phenotypes of mouse models with varying Dgat1 and Dgat2 levels, we hypothesize that varying intestinal Dgat1 and Dgat2 levels alters TAG distribution in subcellular pools for CM synthesis as well as the morphology and proteome of CLDs. To test this hypothesis, enterocytes from the proximal region of the small intestine of mouse models with varying levels of Dgat1 and Dgat2 (Dgat1^Int, Dgat2^Int, and Dgat1^−/−) were analyzed 2 hours after a 200 μl oral olive oil gavage. An ultrastructural examination of enterocytes by TEM and a proteomic analysis of isolated CLDs using LC-MS/MS were performed. Identifying differences in subcellular TAG distribution and the proteins associated with CLDs in enterocytes from mouse models with varying Dgat levels will illustrate unique cellular contributions of Dgat1 and Dgat2 to the process of dietary fat absorption.

2. Methods

2.1 Diet and Animals

All procedures were approved by the Purdue Animal Care and Use Committee. Mice with intestine-specific overexpression of Dgat1 (Dgat1^Int), intestine-specific overexpression of Dgat2 (Dgat2^Int) and whole body Dgat1 deficient (Dgat1^−/−) mice were generated as previous described [16, 19, 20]. Male Dgat1^Int, Dgat2^Int and Dgat1^−/− mice, 4–6 months of age were used in this study. The mice were housed in a specific pathogen-free barrier facility with a 12 hour light/dark cycle (6AM/6PM) and fed a low-fat, rodent chow diet (PicoLab 5053, Lab Diets, Richmond, IN). On the day of the experiment the mice were fasted for 4 hours at the beginning of the light cycle and administered 200 μl of olive oil by oral gavage. No food was available after the gavage.

2.2 Intestinal TAG Secretion

Intestinal TAG secretion in these models of altered intestinal Dgat expression was assessed previously in our lab [19, 20]. Mice were fasted for 4 h starting at the beginning of the light cycle and administered 500 mg/kg Tyloxapol (T0307, Sigma-Aldrich, St. Louis, MO) via IP injection to block lipase activity in circulation. At 30 min post-Tyloxapol injection, mice were given a 200 μl oral olive oil gavage and blood was collected via the submandibular vein at 2 and 4 h post gavage. Plasma TAG concentration was determined using the Wako L-Type TG M determination kit (Wako Chemicals USA).

2.3 Transmission Electron Microscopy (TEM)

Mice were anesthetized using inhaled isoflurane 2 h after a 200 μl oral olive oil gavage. We performed whole mouse perfusion fixation by cardiac infusion of 2% glutaraldehyde and 2% paraformaldehyde in 0.1M sodium cacodylate (pH 7.4). The small intestine was harvested and divided into five equal length segments and labeled S1–S5 (proximal to distal) in relation to the stomach, with S2–3 representing the jejunum. Small pieces of the jejunum were stored in freshly prepared 2% glutaraldehyde in 0.1 M sodium cacodylate, pH 7.4 for at least 2 h at 4°C. Tissues were diced into 1 x 5 mm pieces then fixed with 1% osmium tetroxide in 0.1 M sodium cacodylate, pH 7.4, for 1 h at room temperature. Tissues were then washed repeatedly in distilled deionized water. Dehydration was completed with a graded series of ethanol and specimens were embedded in Embed 812 resin. Thick sections (0.5 μm) were stained with 1% toluidine blue and examined by light microscopy to confirm tissue orientation. Thin sections (80 nm) were cut on a Leica UC6 ultramicrotome and stained in 2% uranyl acetate and lead citrate. Images were acquired on a Tecnai T20 transmission electron microscope (FEI, Hillsboro, OR) equipped with a LaB6 source and operating at 100 kV. The quantitative measurements with acquired micrographs were made using ImageJ software (NIH, USA). Sixty intact enterocytes from the middle area of villi in each group were examined for quantitative measurements (20 cells/mouse; 3 mice/group). All the enterocytes selected for quantitative measurements had TAG present in the subcellular pools for both secretion (ER lumen and Golgi/secretory vesicles (SVs)) and storage (CLDs). The identification of these subcellular TAG pools is based on previously published electron microscopy work [26–30]. The TAG for secretion was assessed by measuring areas of expanded, TAG-containing ER vesicles and Golgi/SVs within an individual cell. The TAG for storage was assessed by counting the number of CLDs and measuring the diameter of CLDs within an individual cell. The amount of TAG stored in CLDs within an individual cell was indicated by the sum of the areas of identified CLDs. The area of a CLD was estimated from its diameter by applying the equation of (mean number of CLDs) x π x (mean radius)².

2.4 Enterocyte Isolation

The proximal third of the small intestine was excised 2 h after a 200 μl oral olive oil gavage and enterocytes were isolated as previously described [21, 24, 31]. Briefly, intestinal tissue was washed in cold tissue buffer (Hank’s Balanced Salt Solution containing 25 mM HEPES and 1% fetal bovine serum), then incubated in isolation buffer (Calcium and magnesium free Hanks’s Balanced Salt Solution with 1.5 mM EDTA) for 15 minutes at 37°C with rotation. The sample was vortexed briefly and the supernatant, which contained isolated enterocytes, was obtained. This isolation process was repeated, the supernatants were combined, and isolated enterocytes were pelleted.

2.5 CLD Isolation

CLDs were isolated from enterocytes using a previously validated sucrose gradient ultracentrifugation protocol that was modified for mouse enterocytes [24, 32, 33]. Enterocytes were lysed using ice cold lysis buffer containing sucrose (175 mM sucrose, 10 mM HEPES and 1 mM EDTA) and disrupted by passing through a 27 gauge 1 inch needle eight times. The 2 mls of cell lysate was transferred to a centrifuge tube and overlaid with 6 mls of sucrose free lysis buffer. The sample was centrifuged at 20,000 x g at 4°C for 2 h. Following the centrifugation, the sample was frozen at −80°C and then sliced into seven fractions of approximately 1 cm in length, with the top fraction containing isolated CLDs. A sample from the CLD enriched fraction was negatively stained with 2% w/v aqueous uranyl acetate and imaged using a Technai T20 transmission electron microscope (FEI, Hillsboro, OR). Protein concentrations of these fractions were determined using a BCA assay.

2.6 Protein Digestion and LC-MS/MS Analysis

CLD fractions were delipidated using 2:1 chloroform:methanol, followed by protein precipitation using ice cold acetone. In preparation for the digest, samples were denatured using 8M urea and 10 mM DTT for 1 h at 37°C, followed by alkylation (using a 2% 2–iodoethanol, 97.5% acetonitrile, 0.5% TEP solution). Peptides were digested using Trypsin/Lys-C Mix (Promega, Madison, WI, USA) using the Barocycler NEP2320 (Pressure Biosciences, Inc., South Easton, MA, USA) at 50°C under 20,000 psi for 120 cycles (2 h). One cycle consisted of 50 seconds at 20,000 psi followed by 10 seconds at atmospheric pressure. The digestion was quenched using trifluoroacetic acid. Injection volumes were adjusted to load 1 μg of sample onto the column for nanoLC-MS/MS analysis.

Peptides were separated on a nanoLC system (1100 Series LC, Agilent Technologies, Santa Clara, CA). The peptides were loaded on the Agilent 300SB-C18 enrichment column for concentration and the enrichment column was switched in-line with the analytical column after 5 min. Peptides were separated with the C18 reversed phase ZORBAX 300SB-C18 analytical column (0.75 μm × 150 mm, 3.5um) from Agilent. The column was connected to an emission tip from New Objective and coupled to the nano electrospray ionization (ESI) source of the high resolution hybrid ion trap mass spectrometer LTQ Orbitrap XL (Thermo Scientific). The peptides were eluted from the column using acetonitrile (ACN)/0.1% formic acid (mobile phase B). For the first 5 minutes, the column was equilibrated with 95% purified H₂O /0.1% formic acid (mobile phase A) followed by the linear gradient of 5% B to 40% B in 65 minutes at 0.3ul/min and from 40% B to 95% B in an additional 10 minutes. The column was washed with 95% of ACN/0.1% formic acid and equilibrated with 95% purified H₂O/0.1% formic acid before the next sample was injected. A blank injection was run between samples to avoid carryover and keep the system clean.

The LTQ-Orbitrap mass spectrometer was operated in the data dependent positive acquisition mode at a resolution of 30,000. Each full MS scan was followed by eight MS/MS scans where these abundant molecular ions were selected and fragmented by collision induced dissociation using a normalized collision energy of 35%.

2.7 Proteomic data analysis

The MS and MS/MS peak list files were analyzed using MaxQuant Version 1.5.2.8 [34]. To identify proteins, the MS/MS spectra were searched against the UniProt protein database, which combines the SwissProt (manually annotated and reviewed) and TrEMBL (automatically annotated and not reviewed) databases [35]. The following settings were used for conducting the search: trypsin and Lys-C digestion enzymes with a maximum of two missed cleavages, fixed modification of ethanolyl addition to cysteine, and variable modifications of N-terminal acetylation and oxidation of methionine. The MS tolerance was set at 4.5 ppm with a maximum of five modifications. A false discovery rate of 0.01 was used for protein and peptide identifications, and the spectra were searched against a reverse decoy database. The minimum peptide length was set at seven amino acids, and a minimum score of 40 was required for modified peptides. The MS/MS match tolerance was set at 40 ppm for protein identification, and at least one unique/razor peptide was required for identification.

The MaxQuant output was analyzed using the bioinformatics statistical analysis program Perseus 1.5.1.6. Potential contaminants, such as keratin, were removed, and then LFQ intensities transformed log2(x). Protein intensities falling below the level of detection were assigned a value of 12.0941. Only proteins identified in at least 4 of 6 biological replicates were considered present in a particular genotype and used for the analysis of relative protein levels across genotypes. The identified proteins were grouped into broad clusters based on Gene Ontology (GO) terms for biological process or molecular function using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) v 6.7 and the analysis tool in STRING version 10.0. Protein-protein interactions were visualized using STRING version 10.0 using the confidence view (medium confidence, score 0.400).

2.8 qPCR

Total RNA was extracted from jejunal mucosa with RNA STAT60 (Tel-Test, Friendswood, TX) and then DNase treated with a Turbo DNA-free Kit (Ambion, Austin, TX). cDNA was synthesized from 1μg DNase-treated RNA using the AffinityScript QPCR cDNA Synthesis Kit (Stratagene, La Jolla, CA). Total DNA was extracted from tissues using DNeasy Blood & Tissue Kit (Qiagen, Valencia, CA). SYBR green qPCR was performed using the Mx3000P QPCR System (Stratagene) and Brilliant III SYBR Green Master Mix (Stratagene). Primers used for determination of relative mRNA levels were produced by Integrated DNA technologies (Coralville, IA) and validated for efficiency and correct product size in cDNA from mouse intestinal mucosa (see Table 1). Mitochondrial content was assessed by determining the ratio of mitochondrial DNA to nuclear genomic DNA. Primers for mitochondrial specific gene are forward 5’-CCCATTCGCGTTATTCTT-3’ and reverse 5’-AAGTTGATCGTAACGGAAGC-3’. Primers for nuclear specific gene are forward 5’-TGGTAAAGCAAAGAGGCCTAA-3’ and reverse 5’-AGAAGTAGCCACAGGGTTGG-3’. The level of each gene was calculated with the comparative Ct method using WT mice as the reference group.

Table 1.

Primers used for QPCR

Gene	Primer Sequences
Cpt1α	F 5’-TGTGGTGTCCAAGTATCTGGCAGT-3’
	R 5’-AACACCATAGCCGTCATCAGAAC-3’
Acox1	F 5’-ATATTTACGTCACGTTTACCCCGG-3’
	R 5’-GGCAGGTCATTCAAGTACGACAC-3’
Plin2	F 5’-AAGAGGCCAAACAAAAGAGCCAGGAGACCA-3’
	R 5’-ACCCTGAATTTTCTGGTTGGCACTGTGCAT-3’
Plin3	F 5’-ATGGAATCCGTGAAACAGGGTGTG-3’
	R 5’-TGAGAGGTCCTGGAAGGAGTGAAT-3’
Seipin	F 5’-TTTGGCTTCGCTGAACAGAAGCAG-3’
	R 5’-ATAGCTGAAGAGCACGATGACGCT-3’
β-actin	F 5’-AGGCCCAGAGCAAGAGAGGTA-3’
	R 5’-GGGGTGTTGAAGGTCTCAAACA-3’

2.9 Statistical analysis

Values of quantitative data were reported as means ± standard error of the mean (SEM). Statistical analyses of data were performed in SAS (SAS Institute, Inc., Cary, NC). One-way analysis of variance (ANOVA) with Tukey’s HSD post hoc test was used for multiple comparisons. Type 3 tests of fixed effects and Kolmogorov-Smirnov tests were used for characterization of CLD size distribution. A p value < 0.05 denotes a statistically significant difference.

3. Results

3.1 Varying levels of intestinal Dgat1 and Dgat2 alters intracellular TAG distribution in enterocytes

Multiple subcellular TAG pools were identified in enterocytes of WT mice 2 hours after a 200 μl oral olive oil gavage using TEM (Figure 1A), and the ultrastructural morphology of these TAG pools were consistent with previously published work [26–30]. The subcellular TAG pools present in the lumen of the ER, Golgi and Golgi-derived SVs) are involved in CM synthesis and secretion, highlighting the complexity of this process. CMs are first assembled and expanded in the lumen of the ER (Figure 1A–➂, B and C), then transported to the Golgi (Figure 1A–➃ and D) and carried by SVs (Figure 1A–➄ and E) to the basolateral side of enterocytes for secretion. CM-sized lipid particles were also present in the intercellular space between adjacent enterocytes (Figure1A–➅ and E), showing that CMs were being secreted into the lymph. In addition, CLDs (Figure1A–➆ and F), which serve as a temporary TAG storage pool, were present in enterocytes. The subcellular location of TAG pools for secretion and storage were also distinct. TAG in the lumen of the ER was usually found in the apical region of enterocytes beneath the microvilli, whereas TAG in the Golgi and SVs was usually found in the central region of enterocytes just above the nucleus. CLDs of various sizes were primarily found in the cytoplasmic area above the nucleus and occasionally on the basolateral side of enterocytes. In addition, enterocyte CLDs were observed to have physical interactions with multiple organelles including the ER, nucleus, degradative vacuoles (possibly lysosomes) and mitochondria (Supplementary Figure 1).

Figure 1. Identification of subcellular TAG pools within enterocytes by TEM.

(A) An ultrastructural overview of enterocytes from the jejunum of WT mice 2 hours after a 200 μl oral olive oil gavage. At this time point during dietary fat absorption, TAG is found in several subcellular pools involved in either CM or CLD synthesis. CMs are assembled and further expanded in the ER lumen ➂, then transported to the Golgi ➃ and Golgi-derived SVs ➄, and finally secreted from enterocytes into the lymph ➅. CLDs serve as temporary TAG storage pools in enterocytes ➆. (B) TAG present in the lumen of the ER is surrounded by smooth ER membrane. (C) The smooth ER containing TAG (black arrows) is continuous with the rough ER membrane, which can be distinguished by the presence of ribosomes (white arrows). (D) TAG is present within stacks of the Golgi apparatus. (E) CMs are carried within SVs (black arrow) and secreted into the intercellular space (*). (F) CLDs are surrounded by a phospholipid monolayer.

To investigate the role of individual Dgats in directing synthesized TAG into specific subcellular pools, an ultrastructural examination of enterocytes from jejunum segments of WT, Dgat1^Int, Dgat2^Int and Dgat1^−/− mice was performed 2 hours after a 200 μl oral olive oil gavage. No abnormal villus morphology was observed in any of the mouse models, as assessed using light microscopy (Supplementary Figure 2). The pattern of TAG distribution along the villi was similar among the mouse models, with the greatest TAG content in the tip and the least in the crypt (Supplementary Figure 3). The abundance of intracellular TAG in the tip area made it difficult to clearly observe membrane structures in this region. Thus, the enterocytes in the middle area of villi, which were structurally intact and contained moderate amounts of intracellular lipids, were further examined.

Varying levels of intestinal Dgat1 and Dgat2 altered TAG distribution within enterocytes (Figure 2). Among these models, almost all enterocytes presented with TAG in at least one of the pools involved in the secretion process (ER lumen, Golgi, and/or SVs); however, not every enterocyte contained TAG stored in CLDs. Thus, more than fifty enterocytes representing the middle area of villi from each mouse were viewed to determine CLD presence. We found that when Dgat1 levels were increased (Dgat1^Int mice), the percentage of enterocytes containing CLDs was significantly lower than in WT, Dgat2^Int and Dgat1^−/− mice (Supplementary Figure 4).

Figure 2. Varying levels of intestinal Dgat1 and Dgat2 alters intracellular TAG distribution in enterocytes.

Representative TEM images of enterocytes from the jejunum of Dgat1^Int, WT, Dgat2^Int and Dgat1^−/− mice 2 hours after a 200 μl oral olive oil gavage. Representative CLD (*), TAG in ER lumen (black arrows), and TAG in Golgi/SVs (white arrows) are highlighted. Scale bar = 2 μm.

3.2 Varying levels of intestinal Dgat1 and Dgat2 alters the distribution of TAG in subcellular pools for CM synthesis and secretion

In our previous studies, the intestinal TAG secretion rates of WT, Dgat1^Int, Dgat2^Int and Dgat1^−/− mice were assessed by administering an IP injection of Tyloxapol (to block TAG clearance from circulation) and then measuring plasma TAG levels in response to an oral olive oil gavage. We found that Dgat1^Int, Dgat2^Int and Dgat1^−/− mice had unchanged, increased and decreased rates of intestinal TAG secretion compared to WT mice, respectively (Figure 3A) [19, 20].

Figure 3. Varying levels of intestinal Dgat1 and Dgat2 alters the distribution of TAG in subcellular pools for CM synthesis and secretion.

(A) TAG secretion was assessed by measuring plasma TAG levels in Dgat1^Int, WT, Dgat2^Int and Dgat1^−/− mice given an IP injection of Tyloxapol at 0, 2 and 4 hours after a 200 μl oral olive oil gavage (n = 6–16 mice/group). (B) Representative TEM images of the area above the nucleus in enterocytes from Dgat1^Int and Dgat2^Int mice 2 hours after a 200 μl oral olive oil gavage. A greater amount of TAG was found in the ER lumen of Dgat1^Int enterocytes, whereas a greater amount of TAG was found in Golgi/SVs of Dgat2^Int enterocytes. Representative CLD (*), TAG in ER lumen (black arrows), and TAG in Golgi/SVs (†) are highlighted. (C) TEM images highlighting distinct, small CLDs (*) and TAG within the ER lumen (black arrow) in enterocytes of Dgat1^Int mice 2 hours after a 200 μl olive oil oral gavage. (D) Mean area of TAG in the ER lumen and (E) mean area of TAG in Golgi/ SVs per cell were quantified using ImageJ. The value of each bar is the average of three biological replicates (20 cells/mouse, n=3 mice/group). Data are represented as mean ± SEM. Different letters denote significant differences, p < 0.05 (one-way ANOVA, Tukey HSD test), N.S = not significant.

To determine whether subcellular TAG pools involved in CM synthesis and secretion are altered by varying the levels of intestinal Dgat1 and Dgat2, we imaged enterocytes 2 hours after a 200 μl oral olive oil gavage using TEM and quantified the area of TAG in the lumen of the ER, as well as in the Golgi/SVs within individual cells from these models. In Dgat1^Int mice, the mean area of TAG in the ER is significantly greater than in WT mice (Figure 3B–D), while the mean area of TAG in Golgi/SVs is similar (Figure 3E). In Dgat2^Int mice, the mean area of TAG in the Golgi/SVs is significantly greater than in WT mice (Figure 3B and E), while the mean area of TAG in the ER is similar (Figure 3D). In Dgat1^−/− mice the mean area of TAG in the ER and Golgi/SVs trended lower than in WT mice (Figure 3D and E). Taken together, varying intestinal levels of Dgat1 and Dgat2 alters subcellular pools of TAG in the ER and Golgi/SVs as well as TAG secretion rate.

3.3 Varying levels of intestinal Dgat1 and Dgat2 alters enterocyte CLD morphology

We also identified alterations in CLD morphology in models with varying levels of intestinal Dgat1 and Dgat2. We found that enterocytes of WT, Dgat1^Int, Dgat2^Int, and Dgat1^−/− mice all contained CLDs of various sizes. However, Dgat1^Int and Dgat1^−/−, but not Dgat2^Int mice, had significantly altered CLD size distributions compared to WT mice (Figure 4A and B). In WT and Dgat2^Int mice the majority of CLDs were 1–3 μm in diameter. Dgat1^Int mice had a right skewed distribution in CLD size, with CLDs smaller than 1 μm in diameter observed most frequently. On the other hand, Dgat1^−/− mice exhibited the widest range of CLD sizes and were the only model that could synthesize CLDs greater than 7 μm in diameter (maximum ~10μm in diameter).

Figure 4. Varying levels of intestinal Dgat1 and Dgat2 alters enterocyte CLD morphology.

(A, B) Size distribution of CLDs within enterocytes of mice 2 hours after a 200 μl oral olive oil gavage. All of the CLDs identified from 60 enterocytes (20 enterocytes/mouse, n=3 mice/group) were examined by TEM and the diameter of CLDs were determined using ImageJ. Data is reported as % of all the CLDs measured per group. * denotes a statistically significant difference, p < 0.05 (Kolmogorov-Smirnov tests). (C) Mean number of CLDs and (D) mean size of CLDs per cell (diameter) were determined. (E) The amount of TAG stored in CLDs is estimated by the equation of (mean number of CLD) x π x (mean radius)². The value of each bar is the average of three biological replicates (20 cells/mouse, n=3 mice/group). (F) qPCR analysis of mRNA levels of genes involved in CLD metabolism. WT mice were the reference group, with their mRNA levels set as 1 (n=3–5 mice/group). Different letters denote significant differences, p < 0.05 (one-way ANOVA, Tukey HSD test), N.S = not significant.

We further analyzed changes in CLD morphology within individual cells containing CLDs. In Dgat1^Int mice, the mean number, diameter and total area of CLDs is similar to WT mice (Figure 4C, D, and E). In Dgat2^Int mice, the mean diameter of individual CLDs is similar to WT mice (Figure 4D); however, the mean number and total area of CLDs per cell are significantly greater than in WT mice (Figure 4C and E). In Dgat1^−/− mice, the mean number of CLDs is similar to WT mice (Figure 4C); however, the mean diameter and total area of CLDs per cell are significantly greater than in WT mice (Figure 4D and E). Even though varying levels of intestinal Dgat1 and Dgat2 alters the size or the number of CLDs per cell, intestinal mRNA levels of proteins involved in CLD metabolism or morphology (Plin2, Plin3, and Seipin) [21, 36] were similar in Dgat1^Int, Dgat2^Int and Dgat1^−/− compared to WT mice 2 hours after a 200 μl oral olive oil gavage (Figure 4F).

3.4 Varying levels of intestinal Dgat1 and Dgat2 alters the enterocyte CLD proteome

Since proteins associated with CLDs have been shown to regulate their synthesis and catabolism, we hypothesized that varying intestinal levels of Dgat1 and Dgat2 alters the proteins that associate with enterocyte CLDs. These alterations may contribute to the altered subcellular TAG distribution in enterocytes of these models. To determine this, CLDs were isolated from enterocytes of WT, Dgat1^Int, Dgat2^Int, and Dgat1^−/− mice 2 hours after a 200 μl oral olive oil gavage using density gradient ultracentrifugation [24, 32, 33]. This method enriches for CLDs, as demonstrated by immunoblotting of the obtained fractions for representative CLD, cytosolic, and membrane proteins [24]. A challenge in isolating CLDs from cells that make both CLDs and lipoproteins is the potential for ER lumenal lipids and lipoproteins to contaminate the CLD fraction. To minimize this, CLDs were isolated at a much lower speed than those used to isolate ER lumenal lipids or CMs [37–39]. In addition, using negative stain electron microscopy we observed that the isolated CLDs in WT, Dgat1^Int, and Dgat2^Int mice were similar in size, while the isolated CLDs from Dgat1^−/− mice spanned a greater size distribution, with the presence of small as well as larger CLDs compared to the other models (Figure 5A). This result is consistent with the ultrastructural observation in Figure 4A and the images also suggest a lack of broader contamination in the CLD enriched fraction.

Figure 5. Varying levels of intestinal Dgat1 and Dgat2 alters the enterocyte CLD proteome.

CLDs were isolated from enterocytes of Dgat1^Int, WT, Dgat2^Int, and Dgat1^−/− mice 2 hours after a 200μl oral olive oil gavage. (A) Negative stain EM of the intestinal CLD-enriched fraction from a representative sample for each mouse model. Scale bar = 0.5 μm. (B) Overlap of CLD-associated proteins for each mouse model, identified by proteomics. (C) Biological functions of CLD-associated proteins common to all four mouse models, classified based on GO terms (for biological process & molecular function). (D) Differentially expressed CLD-associated proteins common among all mouse models. Different letters denote significantly different relative protein levels (p < 0.05). (E) String map of all proteins with a GO term/functional classification related to lipid/lipoprotein metabolism. Black circles indicate proteins identified in all four genotypes.

Proteins present in the CLD fraction were identified using high resolution tandem mass spectrometry. A total of 158 proteins were identified in at least 4 of 6 biological replicates of at least one of the four mouse models. Of these 158 proteins, 53 of them were present in all four models (Figure 5B, Table 2), and the remainder were unique to one, two, or three genotypes (Figure 5B, Table 3). For further characterization, the gene functional classification tool DAVID and the protein analysis tool STRING were used to classify proteins into broad categories based on gene ontology (GO) terms for biological process or molecular function. Since several proteins had multiple functions that fell under more than one of the broad classifications, their functions were further investigated by performing individual protein searches using the UniProt database. The available information about the function of the protein was used to determine the most appropriate broad classification (Table 2 & 3, Function column).

Table 2.

CLD-associated proteins common among all genotypes.

Protein	Gene	UniProt ID	Function
Alpha-enolase	Eno1	P17182	Carbohydrate Metabolism
ATP synthase subunit alpha, mitochondrial	Atp5a1	Q03265	Carbohydrate Metabolism
ATP synthase subunit beta, mitochondrial	Atp5b	P56480	Carbohydrate Metabolism
Bifunctional ATP-dependent dihydroxyacetone kinase/FAD-AMP lyase (cyclizing)	Dak	Q8VC30	Carbohydrate Metabolism
Fructose-bisphosphate aldolase B	Aldob	Q91Y97	Carbohydrate Metabolism
Glyceraldehyde-3-phosphate dehydrogenase	Gm3839	P16858	Carbohydrate Metabolism
Malate dehydrogenase, cytoplasmic	Mdh1	P14152	Carbohydrate Metabolism
Malate dehydrogenase, mitochondrial	Mdh2	P08249	Carbohydrate Metabolism
Pyrethroid hydrolase Ces2e	Ces2e	Q8BK48	Carbohydrate Metabolism
Pyruvate kinase PKM	Pkm	P52480	Carbohydrate Metabolism
78 kDa glucose-regulated protein	Hspa5	P20029	Cell Stress Response
Apoptosis-associated speck-like protein containing a CARD	Pycard	Q9EPB4	Cell Stress Response
Bifunctional 3-phosphoadenosine 5- phosphosulfate synthase 2	Papss2	O88428	Cell Stress Response
Bifunctional epoxide hydrolase 2	Ephx2	P34914	Cell Stress Response
Glutathione S-transferase Mu 3	Gstm3	P19639	Cell Stress Response
Glutathione S-transferase P 1	Gstp1	P19157	Cell Stress Response
Hemoglobin subunit beta-1	Hbbt1	A8DUK4	Cell Stress Response
Protein disulfide-isomerase	P4hb	P09103	Cell Stress Response
Superoxide dismutase [Cu-Zn]	Sod1	P08228	Cell Stress Response
	Tpm1	G5E8R0	Cell Stress Response
Actin, cytoplasmic 2	Actg1	P63260	Cytoskeleton Organization
Annexin	Anxa2	B0V2N7	Cytoskeleton Organization
Tubulin alpha-1C chain	Tuba1c	P68373	Cytoskeleton Organization
Tubulin beta-4B chain	Tubb4b	P68372	Cytoskeleton Organization
Villin-1	Vil1	Q62468	Cytoskeleton Organization
Acylcarnitine hydrolase	Ces2c	Q91WG0	Lipid Metabolism
Apolipoprotein A-I	Apoa1	Q00623	Lipid Metabolism
Estradiol 17-beta-dehydrogenase 11	Hsd17b11	Q9EQ06	Lipid Metabolism
Fatty acid-binding protein, intestinal	Fabp2	P55050	Lipid Metabolism
Fatty acid-binding protein, liver	Fabp1	P12710	Lipid Metabolism
Microsomal triglyceride transfer protein large subunit	Mttp	O08601	Lipid Metabolism
NADH-cytochrome b5 reductase 3	Cyb5r3	Q9DCN2	Lipid Metabolism
Perilipin-3	Plin3	Q9DBG5	Lipid Metabolism
Creatine kinase U-type, mitochondrial	Ckmt1	P30275	Mitochondria/Redox
Cytochrome b5	Cyb5a	G5E850	Mitochondria/Redox
Dehydrogenase/reductase SDR family member 1	Dhrs1	Q99L04	Mitochondria/Redox
Histone H2A type 1-H	Hist1h2ah	Q8CGP6	Nucleosome Assembly
Histone H2B type 1-P	Hist1h2bp	Q8CGP2	Nucleosome Assembly
Histone H4	Hist1h4a	P62806	Nucleosome Assembly
Heat shock protein HSP 90-beta	Hsp90ab1	P11499	Protein Folding
14-3-3 protein zeta/delta	Ywhaz	P63101	Protein Localization/Transport
Alpha-actinin-4	Actn4	P57780	Protein Localization/Transport
Annexin A4	Anxa4	P97429	Protein Localization/Transport
Clathrin heavy chain	Cltc	Q5SXR6	Protein Localization/Transport
40S ribosomal protein SA	Rpsa	P14206	Translation
Elongation factor 1-alpha 1	Eef1a1	P10126	Translation
Elongation factor 1-delta	Eef1d	D3Z7N2	Translation
Heterogeneous nuclear ribonucleoproteins A2/B1	Hnrnpa2b1 O88569		Translation
Calmodulin	Calm1	P62204	Other (Calcium Mediated Signaling)
Galectin-2	Lgals2	Q9CQW5	Other (Carbohydrate Binding, Unknown Function)
Galectin-4	Lgals4	Q8K419	Other (Carbohydrate Binding, Unknown Function)
Heat shock cognate 71 kDa protein	Hspa8	P63017	Other (Protein Metabolism)
Sodium/potassium-transporting ATPase subunit alpha-1	Atp1a1	Q8VDN2	Other (Ion Transport)

Table 3.

CLD-associated proteins unique to one, two, or three genotypes.

Protein	Gene	UniProt ID	Function
Dgat1Int
Cytochrome b-c1 complex subunit 1, mitochondrial	Uqcrc1	Q9CZ13	Carbohydrate Metabolism, Mitochondria/Redox
Guanine deaminase	Gda	Q9R111	Carbohydrate Metabolism
Diacylglycerol O-acyltransferase 1	Dgat1	Q9Z2A7	Lipid Metabolism
WT
Carbonyl reductase [NADPH] 1	Cbr1	P48758	Carbohydrate Metabolism
Dgat2Int
Aldehyde dehydrogenase X, mitochondrial	Aldh1b1	Q9CZS1	Carbohydrate Metabolism
Cytochrome c oxidase subunit 2	Mtco2	P00405	Carbohydrate Metabolism
Dolichyl-diphosphooligosaccharide--protein glycosyltransferase 48 kDa subunit	Ddost	O54734	Carbohydrate Metabolism
Ornithine aminotransferase, mitochondrial	Oat	P29758	Carbohydrate Metabolism
	Ces2g	E9PV38	Hydrolase
Peroxisomal multifunctional enzyme type 2	Hsd17b4	P51660	Lipid Metabolism
V-type proton ATPase catalytic subunit A	Atp6v1a	P50516	Nucleotide Biosynthesis
Zymogen granule membrane protein 16	Zg16	Q8K0C5	Protein Localization/Transport
T-complex protein 1 subunit delta	Cct4	G5E839	Protein Metabolism
Dgat1−/−
4-trimethylaminobutyraldehyde dehydrogenase	Aldh9a1	Q9JLJ2	Carbohydrate Metabolism
6-phosphogluconate dehydrogenase, decarboxylating	Pgd	Q9DCD0	Carbohydrate Metabolism
Aldehyde dehydrogenase family 16 member A1	Aldh16a1	D3Z0B9	Carbohydrate Metabolism
Glutamine--fructose-6-phosphate aminotransferase [isomerizing] 1	Gfpt1	P47856	Carbohydrate Metabolism
Glycerol-3-phosphate dehydrogenase [NAD(+)], cytoplasmic	Gpd1	E0CXN5	Carbohydrate Metabolism
Isocitrate dehydrogenase [NADP] cytoplasmic	Idh1	O88844	Carbohydrate Metabolism, Cell Stress Response
Cell death activator CIDE-B	Cideb	O70303	Cell Stress Response
Nucleophosmin	Npm1	Q5SQB0	Cell Stress Response, Cytoskeletal Organization
Peroxiredoxin-1	Prdx1	B1AXW5	Cell Stress Response
Peroxiredoxin-2	Prdx2	Q61171	Cell Stress Response
Peroxiredoxin-5, mitochondrial	Prdx5	G3UZJ4	Cell Stress Response
Adenylyl cyclase-associated protein 1	Cap1	P40124	Cytoskeleton Organization
Destrin	Dstn	Q9R0P5	Cytoskeletal Organization
Cytochrome P450 2D26	Cyp2d26	Q8CIM7	Mitochondria/Redox
Bifunctional purine biosynthesis protein PURH	Atic	Q9CWJ9	Nucleotide Biosynthesis, Mitochondria/Redox
Eukaryotic translation initiation factor 5A-1	Eif5a	J3QPS8	Protein Localization/Transport
Poly(rC)-binding protein 1	Pcbp1	P60335	RNA Processing
Small nuclear ribonucleoprotein Sm D3	Snrpd3	P62320	RNA Processing
Dgat1Int, WT
40S ribosomal protein S8	Rps8	P62242	Translation
Dgat1Int, Dgat2Int
Beta-1,4 N-acetylgalactosaminyltransferase 2	B4galnt2	Q09199	Carbohydrate Metabolism
Catalase	Cat	P24270	Cell Stress Response
Transmembrane 9 superfamily member 2	Tm9sf2	P58021	Unknown Function
WT, Dgat2Int
Heat shock protein HSP 90-alpha	Hsp90aa1	P07901	Protein Localization/Transport
WT, Dgat1−/−
L-lactate dehydrogenase A chain	Ldha	P06151	Carbohydrate Metabolism
NSFL1 cofactor p47	Nsfl1c	Q9CZ44	Cell Stress Response
Cofilin-1	Cfl1	F8WGL3	Cytoskeleton Organization
Creatine kinase B-type	Ckb	Q04447	Kinase, Transferase
Glutathione S-transferase	Gsta2	D3Z6A6	Transferase
40S ribosomal protein S14	Rps14	P62264	Translation
40S ribosomal protein S3	Rps3	D3YV43	Translation
	Fcgbp	E9Q9C6	Unknown Function
Dgat2Int, Dgat1−/−
14-3-3 protein epsilon	Ywhae	P62259	Protein Localization/ Transport
40S ribosomal protein S15	Rps15	P62843	Translation
40S ribosomal protein S18	Rps18	F6YVP7	Translation
Dgat1Int, WT, Dgat2Int
Nucleobindin-1	Nucb1	H3BK79	Calcium Ion Binding, Unknown Function
Apolipoprotein A-IV	Apoa4	P06728	Lipid Metabolism
Retinal dehydrogenase 1;Aldehyde dehydrogenase, cytosolic 1	Aldh1a1	O35945	Lipid Metabolism
Dgat1Int, Dgat2Int, Dgat1−/−
Protein disulfide-isomerase A3	Pdia3	P27773	Cell Stress Response
Protein disulfide-isomerase A6	Pdia6	Q3TML0	Cell Stress Response
Myosin-14	Myh14	K3W4R2	Cytoskeletal Organization
Myosin-9	Myh9	Q8VDD5	Cytoskeletal Organization
Cytochrome P450 2B10	Cyp2b10	Q9WUD0	Lipid Metabolism
Long-chain-fatty-acid--CoA ligase 5	Acsl5	Q8JZR0	Lipid Metabolism
Aconitate hydratase, mitochondrial	Aco2	Q99KI0	Mitochondria/Redox
Cytochrome b5 type B	Cyb5b	Q9CQX2	Mitochondria/Redox
Synaptic vesicle membrane protein VAT-1 homolog	Vat1	Q62465	Mitochondria/Redox
Ras-related protein Rab-2A	Rab2a	P53994	Protein Localization/Transport
Peptidyl-prolyl cis-trans isomerase A	Ppia	P17742	Protein Metabolism
40S ribosomal protein S3a	Rps3a	P97351	Translation
60S ribosomal protein L13	Rpl13	P47963	Translation
60S ribosomal protein L7	Rpl7	P14148	Translation
WT, Dgat2Int, Dgat1−/−
Fructose-1,6-bisphosphatase isozyme 2	Fbp2	P70695	Carbohydrate Metabolism
Transketolase	Tkt	P40142	Carbohydrate Metabolism
Heterogeneous nuclear ribonucleoprotein K	Hnrnpk	B2M1R6	Cell Stress Response
Purine nucleoside phosphorylase	Pnp	Q543K9	Cell Stress Response
Ubiquitin-like modifier-activating enzyme 1	Uba1	Q02053	Cell Stress Response
Plastin-1	Pls1	Q3V0K9	Cytoskeletal Organization
Leukotriene A-4 hydrolase	Lta4h	P24527	Lipid Metabolism
Guanine nucleotide-binding protein subunit beta- 2-like 1	Gnb2l1	P68040	Protein Localization/Transport
Transitional endoplasmic reticulum ATPase	Vcp	Q01853	Protein Localization/Transport
Leukocyte elastase inhibitor A	Serpinb1a	Q9D154	Protein Metabolism
Elongation factor 2	Eef2	P58252	Translation

The 53 proteins common among the four mouse models are involved in multiple biological processes (Figure 5C). The majority of these common proteins are involved in carbohydrate metabolism (19%), cell stress response (19%), or lipid metabolism (15%). The relative abundances of these common proteins were also compared among the genotypes, and four proteins were found to be differentially expressed among the models (Figure 5D). Hemoglobin subunit beta-1 (Hbbt), which was classified as a cell stress response-related protein [40], had significantly higher relative levels in Dgat2^Int enterocyte CLD fractions compared to Dgat1^−/− fractions. In addition, heterogeneous nuclear ribonucleoproteins A2/B1 (Hnrnpa2b1), which is involved regulating transcription/translation , was expressed at higher levels in Dgat1^−/− mice compared to all the other genotypes. Furthermore, 40S ribosomal protein SA (Rpsa), which also plays a role in translation [41], was expressed at higher levels in Dgat1^−/− compared to Dgat1^Int mice. Additionally, the carbohydrate metabolism-related protein malate dehydrogenase (Mdh1) [42] was expressed at higher levels in Dgat1^−/− compared to Dgat1^Int mice.

To better understand the role of CLDs in intestinal lipid metabolism in these models with varying levels of intestinal Dgat1 and Dgat2, further analysis was performed on all of the identified proteins with biological process or molecular function GO terms related to lipid metabolism. Overall, 13% of all of the proteins identified in this study were associated with lipid/lipoprotein metabolism. The majority of these proteins have been previously identified as CLD-associated proteins in various cell types, and those identified in enterocyte models are highlighted (Table 4, Previously Identified column). In addition, just under half of these lipid metabolism related proteins were common to WT, Dgat1^Int, Dgat2^Int, and Dgat1^−/− mice, while the remainder were unique to one, two, or three of these models (Table 4). A few of these differentially expressed proteins, including Dgat1, Lambda-crystallin homolog (Cryl1), and Aminopeptidase B (Rnpep), were identified for the first time as being associated with CLDs in these models. Predicted interactions between the identified lipid metabolism proteins were visualized by performing a STRING analysis. The STRING network shows several protein clusters, including one large cluster containing proteins involved in fatty acid modification as well as TAG and CM synthesis (Figure 5E).

Table 4.

CLD-associated proteins involved in lipid/lipoprotein metabolism.

Protein	Gene	Previously Identified	Dgat1Int	WT	Dgat2Int	Dgat1−/−
Acylcarnitine hydrolase	Ces2c	[24, 55]	✓	✓	✓	✓
ATP synthase subunit alpha, mitochondrial	Atp5a1	[24, 52, 54–56, 72, 73]	✓	✓	✓	✓
ATP synthase subunit beta, mitochondrial	Atp5b	[22, 24, 52, 54–56, 72, 73]	✓	✓	✓	✓
NADH-cytochrome b5 reductase 3	Cyb5r3	[22–24, 52, 54, 55, 72–74]	✓	✓	✓	✓
Bifunctional epoxide hydrolase 2	Ephx2	[24, 54, 55]	✓	✓	✓	✓
Fatty acid-binding protein, liver Fabp1		[24, 54, 55]	✓	✓		✓
Fatty acid-binding protein, intestinal	Fabp2	[24, 55]	✓	✓	✓	✓
Estradiol 17-beta- dehydrogenase 11*	Hsd17b11	[22, 24, 53–56, 72–77]	✓	✓	✓	✓
Microsomal triglyceride transfer protein large subunit*	Mttp	[22–24, 55]	✓	✓	✓	✓
Perilipin-3*	Plin3	[22–24, 53, 55, 56, 72–75, 78–81]	✓	✓	✓	✓
Retinal dehydrogenase 1	Aldh1a1	[24, 55]	✓	✓	✓
Apolipoprotein A-IV*	Apoa4	[22, 24, 55, 73]	✓
Long-chain-fatty-acid--CoA ligase 5*	Acsl5	[23, 24, 55]	✓		✓	✓
Cytochrome b5 type B	Cyb5b	[52, 54, 55, 72–75, 77, 79, 81]	✓		✓	✓
Cytochrome P450 2B10	Cyp2b10	[24]	✓		✓	✓
Leukotriene A-4 hydrolase	Lta4h			✓	✓	✓
Aminopeptidase B	Rnpep				✓	✓
Lambda-crystallin homolog	Cryl1				✓	✓
Diacylglycerol O-acyltransferase 1	Dgat1		✓
Peroxisomal multifunctional enzyme type 2	Hsd17b4	[55, 72, 73]			✓
Cell death activator CIDE-B*	Cideb					✓
Cytochrome P450 2D26	Cyp2d26	[24, 55]				✓

^*Proteins validated as being associated with CLDs using immunohistochemistry &/or Western blot [22–24] represent enterocyte models

3.5 Varying levels of intestinal Dgat1 and Dgat2 impacts enterocyte mitochondrial biology

Fatty acid oxidation (FAO) takes place mainly within mitochondria and has the potential to reduce the availability of fatty acids for TAG synthesis and incorporation into CMs and CLDs [43–45]. Therefore, we determined mRNA levels of genes involved in FAO and the ratio of mitochondrial DNA to genomic DNA (indicator of mitochondrial content) in enterocytes of WT, Dgat1^Int, Dgat2^Int, and Dgat1^−/− mice 2 hours after a 200 μl oral olive oil gavage. We found similar mRNA levels of FAO genes (Cpt1 and Acox) in mice with varying intestinal Dgat1 and Dgat2 levels (Figure 6A). In Dgat2^Int mice, however, mitochondrial content was significantly lower than that observed in all the other mouse models (Figure 6B). In addition, we frequently observed round, swollen-looking mitochondria in Dgat1^−/− mice by TEM (Figure 6C and D). Taken together, varying levels of intestinal Dgat1 and Dgat2 affects enterocyte mitochondrial morphology and content without affecting mRNA levels of FAO genes in response to acute fat challenges.

Figure 6. Varying levels of intestinal Dgat1 and Dgat2 impacts enterocyte mitochondrial biology.

(A) qPCR analysis of mRNA levels of genes involved in FAO. (B) qPCR analysis of the ratio of mitochondrial DNA to nuclear DNA for the assessment of mitochondrial content. WT mice was the reference group, with its DNA level set as 1 (n=3–4 mice/group). Different letters denote significant differences, p < 0.05 (one-way ANOVA, Tukey HSD test), N.S = not significant. (C) Representative TEM images of mitochondria (black arrow) within enterocytes of WT and Dgat1^{− /−} mice 2 hours after a 200 μl oral olive oil gavage. (D) Representative TEM images of normal mitochondria (left) and round, swollen mitochondria (right) found in Dgat1^−/− mice 2 hours after a 200 μl oral olive oil gavage.

4. Discussion

In the present study, we investigated whether varying intestinal Dgat1 and Dgat2 levels alters TAG distribution in subcellular pools for CM synthesis as well as the morphology and proteome of CLDs. The multi-step and dynamic nature of this process makes it challenging to determine distinct roles of Dgat enzymes in CM and CLD synthesis; however the results presented here highlight that Dgat1 and Dgat2 play non-redundant, but coordinated roles in dietary fat absorption. A proposed model for the roles of Dgat1 and Dgat2 in enterocytes during dietary fat absorption was generated based on these as well as previous results (Figure 7). Overall, we found that Dgat1 preferentially synthesizes TAG that supports the synthesis of TAG-rich CMs, whereas Dgat2 preferentially synthesizes TAG that is incorporated into nascent CMs and CLDs.

Figure 7. Summary of results and proposed model of cellular roles of Dgat1 and Dgat2 in synthesizing TAG for CM and CLDs within enterocytes.

(A) Results summary: In Dgat1^Int mice, greater amounts of TAG in the ER lumen, similar amounts of TAG in Golgi/SVs, a trend toward smaller CLDs, and a similar intestinal TAG secretion rate were found compared to WT mice. In Dgat2^Int mice, similar amounts of TAG in the ER lumen, greater amounts of TAG in Golgi/SVs, an increased number of CLDs, and a higher intestinal TAG secretion rate were observed compared to WT mice. In Dgat1^−/− mice, a trend towards less TAG in the ER lumen and Golgi/SVs and an increase in CLD size were found in enterocytes compared to WT mice. Dgat1^−/− mice also secrete TAG from the intestine at a slower rate within smaller sized CMs compared to WT mice. Some of the CLD-associated proteins are present in all genotypes, whereas others, such as ApoAIV and Cideb, are unique to certain genotype(s) and may be contributing to the altered subcellular distribution of TAG observed in these models. (B) Proposed model: A proposed model of coordinated, non-redundant roles of Dgat1 and Dgat2 in CM synthesis was generated by combining the results of the current study with previous knowledge of the CM synthesis pathway. CM assembly in enterocytes requires two distinct pools of TAG within the lumen of the ER: ApoB-containing nascent CMs and ApoB-free LLDs. In this model, Dgat1 is proposed to preferentially synthesize TAG for incorporation into ApoB-free LLDs, which determines the availability of TAG for nascent CM expansion and thus the size of CMs. Dgat2 is proposed to preferentially synthesize TAG for incorporation into ApoB-containing nascent CMs, which likely regulates the rate of intestinal TAG secretion. In addition, TAG storage in CLDs and its subsequent catabolism provides substrates for re-synthesis of TAG at the ER membrane that may contribute to CM synthesis and secretion at later times. Based on the results from the current study, Dgat1 is proposed to restrict CLD growth in enterocytes by preferentially synthesizing TAG that is delivered into the ER lumen, thus limiting TAG storage in CLDs. Dgat2, on the other hand, is proposed to preferentially increase TAG storage in CLDs, which may provide fatty acids for re-synthesis of TAG at the ER membrane for incorporation into CMs at later time points. Differential associations of proteins with enterocyte CLDs in these models may serve to regulate the localization and activity of proteins involved in dietary fat absorption.

4.1 Varying levels of intestinal Dgat1 and Dgat2 alters the distribution of TAG in subcellular pools for CM synthesis and secretion

Consistent with earlier ultrastructural studies of enterocytes after consumption of dietary fat [26–29], we observed TAG present along the secretory pathway (in the ER, Golgi and SVs) in enterocytes from WT mice as well as mice with altered intestinal Dgat levels. CMs are first synthesized in ER lumen by a two-step process that involves two distinct pools of TAG: apolipoprotein B (ApoB)-containing nascent CMs and ApoB-free lumenal lipid droplets (LLDs) [46] (Figure 7B). ApoB-containing nascent CMs are generated by incorporation of a small amount of TAG onto ApoB, the main structural protein of CMs. ApoB-free LLDs are used in the second step of CM synthesis for expansion of nascent CMs, yielding TAG-rich pre-CM particles. ApoB-free LLDs were initially identified in mice deficient in ApoB by TEM [30], and have a distinct appearance similar to what was observed predominantly in the Dgat1^Int mice in this study (Figure 3B and C). ApoB-containing nascent CMs are challenging to visualize due to their poor lipidation, however only particles containing ApoB are able to be transported from the ER to the Golgi for secretion [47]. Pre-CM particles are transported to the Golgi by pre-CM transport vesicles and ultimately to SVs for exocytosis from the basolateral side of enterocytes into lymph [48]. By quantitatively assessing TEM images we observed alterations in TAG distribution within these subcellular locations in enterocytes of mouse models with varying intestinal Dgat1 and Dgat2 levels (Figure 3). The proposed roles of Dgat1 and Dgat2 in synthesizing TAG for CM synthesis and secretion are discussed below.

We propose that Dgat1 drives the synthesis of TAG for incorporation into ApoB-free LLDs based on the results from Dgat1^Int and Dgat1^−/− mice (Figure 7). First, the accumulation of TAG in the ER in Dgat1^Int enterocytes suggests more ApoB-free LLDs than ApoB-containing particles are produced when Dgat1 is overexpressed (Figure 2 and 3). Although the presence/absence of ApoB on these particles was not assessed, the result mimics the accumulation of LLDs seen in enterocytes of mice with intestine-specific deficiency of ApoB [30] and in response to pharmaceutical inhibition of protein synthesis [26], where only LLDs can be synthesized. However, the greater amount of TAG present in the ER lumen in Dgat1^Int enterocytes has little effect on the amount of TAG reaching the Golgi/SVs within enterocytes, the size of CMs in the blood, or intestinal TAG secretion rate, since these are all similar to WT mice [20]. This suggests the presence of other mechanisms that determine which or how many ApoB-free LLDs are able to fuse with ApoB-containing nascent CMs within the ER. Second, significant reductions in the size of CMs secreted into the blood are seen in Dgat1^−/− compared to WT mice [20]. This was supported by a trend, but not significant reduction in TAG content in the ER and Golgi/SVs within enterocytes of Dgat1^−/− compared to WT mice in this study (Figure 3D and E). This suggests that the absence of Dgat1 leads to a defect in LLD synthesis that prevents the synthesis of TAG-rich CMs. Taken together, our results support a role of Dgat1 in synthesizing TAG that is incorporated into LLDs, which determines the availability of TAG for nascent CM expansion and contributes to the synthesis of TAG-rich CMs.

We propose that Dgat2 drives the synthesis of TAG for incorporation into ApoB-containing nascent CMs, as supported by the results from Dgat2^Int and Dgat1^−/− mice (Figure 7). First, since lipid particles must acquire ApoB before being transported from the ER to Golgi, the increased TAG content in the Golgi/SVs in Dgat2^Int enterocytes suggests that Dgat2 is involved in synthesizing TAG for incorporation into ApoB-containing nascent CMs. In addition, TAG present in the Golgi/SVs within enterocytes positively correlates with CM secretion from enterocytes [26, 29, 30], so the increased TAG content in the Golgi/SVs of Dgat2^Int enterocytes is also consistent with the increased intestinal TAG secretion rate in this model [20]. This increased TAG secretion in Dgat2^Int mice is likely due to an increased number of CMs being secreted since there are no differences in CM size compared to WT mice [20]. Second, mice lacking Dgat2 die within a few hours of birth [17, 49]. This supports an essential role of Dgat2 in the synthesis and secretion of ApoB-containing lipoproteins from enterocytes for dietary fat absorption and survival. Furthermore, Dgat1^−/− mice, which only express Dgat2, have nascent CMs present in the Golgi of enterocytes and are capable of secreting CMs into blood [18]. Taken together, these results suggest that Dgat2 plays an important role in driving the synthesis of ApoB-containing nascent CMs that fuse with LLDs synthesized by Dgat1, and together these enzymes regulate the rate of intestinal TAG secretion.

4.2 Varying levels of intestinal Dgat1 and Dgat2 alters enterocyte CLD morphology

Varying levels of intestinal Dgat1 and Dgat2 affects the number and size of CLDs in enterocytes after dietary fat is consumed, which may in part be due to their contribution to CM synthesis and secretion. The CLD morphology observed in enterocytes of Dgat1^Int and Dgat1^−/− mice (Figure 4 and 7) support the hypothesis that Dgat1 restricts and Dgat2 promotes CLD expansion reported in other models [12, 50]. In primary hepatocytes, which express both Dgat1 and Dgat2, inhibition of Dgat2 reduced CLD size whereas inhibition of Dgat1 increased CLD size [12]. In addition, the overexpression of Dgat2 in HEK293T cells resulted in increased CLD size [50]. This observation is thought to be due to the localization of Dgat2 to CLDs and its ability to synthesize TAG directly at the CLD surface [50, 51]. Although we did see larger CLDs in Dgat1^−/− mice, we did not see an increase in CLD size when Dgat2 was overexpressed in the intestine. We hypothesize that this is due to the combination of high endogenous levels of Dgat1 and high transgenic levels of Dgat2, which is different than in the liver or other cell types that have been used to assess the roles of these enzymes. In this situation the presence of high amounts of Dgat1 can direct synthesized TAG to LLDs and thus restrict CLD expansion. In addition, high amounts of Dgat2 can synthesize TAG for incorporation into both nascent CMs and CLDs, which may also limit CLD expansion.

4.3 Varying levels of intestinal Dgat1 and Dgat2 alters the enterocyte CLD proteome

Consistent with other CLD proteomic studies, proteins involved in diverse cellular functions were found to be associated with enterocyte CLDs among the mouse models. Interestingly, the greatest percentage of the proteins identified in all four mouse models were classified as playing roles in carbohydrate metabolism or cell stress response (Figure 5C). It is unknown whether or not these proteins play functional roles at the CLD surface, are being sequestered at this location to regulate their function/activity in the cell, or are contaminants resulting from the lipid droplet isolation procedure. However, proteins involved in these processes have been identified in CLD proteomics studies from multiple cell types [22–24, 52–56].

Several proteins associated with lipid and lipoprotein metabolism were found to associate with CLDs in the current study, and some were differentially expressed among mice with altered intestinal Dgat levels. We have selected a few of these proteins to highlight below. They include: Cideb, ApoAIV, Dgat1, Cryl1, and Rnpep. Among them, only the localization of Cideb and ApoAIV to CLDs has been validated using additional methods (Western blot or immunofluorescence) [22, 24, 57, 58]. Although validation of the localization of the other proteins to enterocyte CLDs is still needed, the findings suggest that these lipid related proteins may be uniquely contributing to the altered susceptibilities to obesity and related metabolic diseases in these mouse models.

Interestingly, we found that Cideb uniquely associates with enterocyte CLDs in Dgat1^−/− mice, while ApoAIV was identified in all of the models except Dgat1^−/− mice. Cideb is a member of the CIDE protein family, which has been shown to regulate CLD size in multiple cell types [59, 60]. Cideb has been show to localize to CLDs in mouse liver and when exogenously expressed in Caco-2 cells [58]. Cideb-deficient mice exhibit a decreased TAG secretion rate, decreased CM size, and increased TAG storage within enterocyte CLDs, suggesting Cideb is important for CM lipidation. Since Dgat1^−/− and Cideb-deficient mice exhibit a similar phenotype, this suggests that Cideb localization to CLDs may prevent it from promoting CM lipidation at the ER membrane. Consistent with previous studies that validated the localization of ApoAIV to enterocyte CLDs [22, 24], ApoAIV was identified on enterocyte CLDs of WT, Dgat1^Int, and Dgat2^Int mice in the current study. However ApoAIV was not identified on enterocyte CLDs of Dgat1^−/− mice, which exhibit decreased TAG secretion. Since ApoAIV is known to play a role in regulating CM metabolism, it is possible that ApoAIV plays a functional role at the CLD surface that promotes CM secretion. Taken together, the differential association of these proteins that regulate CM metabolism with enterocyte CLDs in Dgat1^−/− mice compared to the other models suggests that they may contribute to the altered TAG storage and secretion in this model.

Dgat1 was only found to associate with CLDs in enterocytes of Dgat1^Int mice. Dgat1 is thought to localize exclusively to the ER membrane and has not previously been identified in other CLD proteomic studies. However, other TAG synthesis enzymes that were initially thought to be ER resident proteins (Mgat2 [18], Dgat2 [7, 50], and several glycerol-3-phosphate pathway enzymes [7]) have also been found to localize to CLDs and play a role in local CLD growth. Since Dgat1 is a multi-spanning integral membrane protein [61], it is unlikely that it directly associates with the CLD monolayer, and thus may only be identified in Dgat1^Int mice due to its overexpression. However we cannot rule out the possibility that Dgat1associates with CLDs through an alternative domain or mechanism, as is the case for Dgat2 [50]. Alternatively, it could be misfolded or non-functional at the CLD surface and targeted to CLDs for degradation (as is the case for apoB [62]) as a way to regulate Dgat1 activity at the ER.

Two lipid related proteins, Cryl1 and Rnpep, were only found to associate with enterocyte CLDs in models with a higher than normal ratio of Dgat2 to Dgat1 (Dgat2^Int and Dgat1^−/− mice). Cryl1 shares sequence homology with 3-hydroxyacyl-CoA dehydrogenase, an enzyme that catalyzes the third reaction of fatty acid beta oxidation [18, 63]. Human CRYL1 mRNA is expressed at low levels in the small intestine and Cryl1 has been detected at moderate levels in mouse large intestine [64], but very little is known about its role in intestinal lipid metabolism. Therefore, further investigation of this protein’s function in enterocytes is needed. In addition to exhibiting aminopeptidase activity, Rnpep also hydrolyzes the lipid mediator leukotriene A₄ (LTA₄) [65], resulting in the formation of the pro-inflammatory LTB₄ [66]. Interestingly, we also identified Leukotriene A-4 hydrolase (Lta4h) associated with enterocyte CLDs in all except the Dgat1^Int model, and this protein has also been identified in other CLD proteomics studies (see Table 4). The presence of these enzymes at the CLD surface suggests that an acute high dietary fat load has the potential to induce an inflammatory response, consistent with previous studies [67].

4.4 Varying levels of intestinal Dgat1 and Dgat2 impacts enterocyte mitochondrial biology

The changes in mitochondrial biology observed in enterocytes of Dgat2^Int and Dgat1^−/− compared to WT mice have the potential to contribute to the alterations in intestinal TAG metabolism observed in these models. We found that Dgat2^Int mice have significantly reduced mitochondrial content, but similar mRNA levels of genes involved in FAO in response to an acute fat challenge. Interestingly, we previously found that Dgat2^Int mice have reduced mRNA levels of FAO genes in a fed state after chronic high fat diet feeding [20]. We are uncertain why we observed differences in the chronic high fat diet fed state but not after an acute high fat challenge. However, these results highlight the potential for FAO to alter the availability of fatty acids for TAG synthesis and incorporation into CMs and CLDs. In addition, we found that Dgat1^−/− mice have altered mitochondrial morphology but no change in mitochondrial content in enterocytes. The poor membrane integrity of mitochondria in Dgat1^−/− enterocytes indicates possible impaired mitochondrial function, which may contribute to or result from the abnormal TAG storage in CLDs observed in this model. Although FAO was not directly measured in these studies, these observations suggest that regulation of FAO is important in these models. Future studies are needed to further assess the role of FAO in intestinal TAG metabolism.

4.5 Intestinal DGAT1 and DGAT2 in mice versus humans

Although it is known that DGAT1 is the major intestinal DGAT enzyme in both mice and humans, it is still unclear whether or not DGAT2 plays an important physiological role within this tissue in humans. Dgat2 is present at significantly lower levels compared to Dgat1 in the intestine of mice, and whether DGAT2 is expressed in human intestine is questionable [6]. One of the reasons for this is that a DGAT1 loss of function mutation has been identified in humans and observed to cause congenital diarrheal disorder resulting in severe outcomes [68, 69]. Dgat1^−/− mice, on the other hand, have no obvious increase in fecal fat or diarrhea, and their intestine phenotype contributes to their beneficial resistance to diet induced obesity [16, 18, 19]. In addition, gastrointestinal side effects observed in humans with DGAT1 mutations are also observed in humans treated with DGAT1 inhibitors, leading to the conclusion that these effects may be target specific. More recently, however, the DGAT1 inhibitor pradigastat has been well tolerated in overweight/obese individuals and patients with familial chylomicronemia syndrome [70, 71], suggesting these inhibitors may still be beneficial for some patients. Although mice and humans may differ in terms of the relative contributions of DGAT1 and DGAT2 to dietary fat absorption, knowledge gained from investigating the altered intestinal lipid metabolism in models with varying intestinal levels of Dgat1 and Dgat2 in mice will contribute to novel knowledge of additional players in the process of dietary fat absorption that may be significant in human health and disease.

5. Conclusion

Overall, the results from this study show that Dgat1 and Dgat2 have non-redundant roles in intestinal lipid metabolism that are required for proper regulation of dietary fat absorption. The ultrastructural analyses of enterocytes suggest that Dgat1 preferentially synthesizes TAG for incorporation into ApoB-free LLDs, promoting CM expansion. The results also support that Dgat2 preferentially synthesizes TAG for incorporation into ApoB-containing nascent CMs and CLDs. The defect in CM synthesis and secretion due to the deficiency of Dgat1 highlights the importance of the coordinated functions of Dgat1 and Dgat2 in the process of dietary fat absorption. Furthermore, varying levels of intestinal Dgat1 and Dgat2 impacts the proteins that associate with CLDs, which may play a role in regulating both TAG storage and secretion. Further investigation of these identified proteins may lead to the identification of novel regulators of dietary fat absorption, which have the potential to serve as therapeutic targets for the treatment of obesity, diabetes, and cardiovascular disease.

Highlights.

• Dgat1 and Dgat2 preferentially synthesize TAG for distinct subcellular pools within enterocytes.

• Dgat1 and Dgat2 regulate CM assembly and secretion in enterocytes.

• Dgat1 and Dgat2 alter the morphology and proteome of enterocyte CLDs.

• Dgat1 and Dgat2 function coordinately to regulate dietary fat absorption.

Abbreviations

ApoB

apolipoprotein B

CLD

cytoplasmic lipid droplet

CM

chylomicron

Dgat

diacylglycerol acyltransferase

FAO

fatty acid oxidation

LLD

lumenal lipid droplet

SV

secretory vesicle

TAG

triacylglycerol

TEM

transmission electron microscopy