Surfactant-like particles (SLP) are unilamellar secreted membranes associated with the process of lipid absorption and isolated previously only from the apical surface of enterocytes. In this paper, the intracellular membrane has been isolated from corn oil-fed animals, identified by its content of the marker protein intestinal alkaline phosphatase (IAP). Another brush-border protein, cubilin, and its anchoring protein megalin have been identified as components of extracellular SLP, but only cubilin is present to any extent in intracellular SLP. During fat absorption, IAP is modestly enriched in intracellular SLP, but full-length cubilin (migrating at 210 kDa in fat-fed mucosal fractions) falls by one-half, although fragments of cubilin are abundant in the intracellular SLP. Both IAP and cubilin colocalize to the same cells during corn oil absorption and colocalize around lipid droplets. This localization is more intense during feeding of corn oil with Pluronic L-81, a detergent that allows uptake of fatty acids and monoglycerides from the lumen, but blocks chylomicron secretion. Confocal microscopy confirms the colocalization of IAP and the ligand for cubilin, intrinsic factor. Possible roles for cubilin in intracellular SLP include facilitating movement of the lipid droplet through the cell and binding to the basolateral membrane before reverse endocytosis.
- fat absorption
- intracellular vesicle
enterocytes are polarized cells that process large amounts of triacylglycerols during fat-containing meals, packaging the fat into lipoprotein droplets that move across the cell to the basolateral membrane and intercellular space (14). However, the mechanism whereby this process occurs is still unclear. Triacylglycerol synthesis and intracellular movement initially proceed through vesicles of the smooth endoplasmic reticulum and fuse with Golgi vesicles before transcellular movement to the basolateral membrane. The process of fat absorption is accompanied by the localization of an intestinal alkaline phosphatase (IAP)-containing membrane surrounding the fat droplets intracellulary (20) as well as the basolateral secretion of an IAP-containing linear membrane (presumably the same or related to the intracellular membrane) (12). We have called this membrane surfactant-like particle (SLP), because it shares features with pulmonary surfactant (7). Partly by analogy with pulmonary surfactant, we have proposed that SLP arises intracellularly rather than being endocytosed from the apical brush-border membrane (13). The SLP moves to the basolateral membrane before discharging its lipid and itself into the basolateral space (12). To explain such intracellular movement, we searched for proteins involved in the process of vesicular movement.
Cubilin is a large (460 kDa) multiligand endocytic receptor that serves as the intrinsic-factor cobalamin receptor and is coexpressed with megalin, another large (600 kDa) multiligand endocytic receptor of the LDL-receptor family (6). Both proteins are glycoproteins. A long list of ligands has been identified for megalin, some of which are shared by cubilin (e.g., albumin, receptor-associated protein). Ligands for megalin include vitamin-binding proteins (e.g., transcobalamin, vitamin D binding protein), apolipoproteins (e.g., B, E), low molecular-weight peptides and hormones (e.g., parathyroid hormone, insulin), polybasic drugs, enzymes (e.g., amylase, lipoprotein lipase), and others (e.g., lactoferrin, plasminogen) (6). These proteins are important for the tubular reabsorption of proteins in the renal tubule. Cubilin contains no transmembrane sequence or other membrane-binding mechanism but interacts with megalin. This large complex is important for endocytosis and recycling of cubilin.
This article reports the isolation of intracellular SLP (iSLP) and compares it with extracellular SLP (eSLP) in terms of the enrichment of IAP and the presence of cubilin in those fractions. As expected, megalin is also present in these membranes but is not enriched so much as cubilin. After corn oil feeding, the cubilin content of eSLP falls dramatically, and that fall is not prevented by feeding the nonionic detergent Pluronic L-81, which blocks chylomicron secretion. The fall in cubilin concentration in intracellular SLP, however, falls less dramatically, especially after corn oil plus Pluronic L-81 feeding, allowing visualization of cubilin degradation products. On the other hand, megalin is not preserved as well as cubilin in iSLP. Moreover, confocal microscopy shows that during triacylglycerol feeding in rats, both IAP and cubilin colocalize in structures that appear to surround fat droplets. The presence of cubilin fragments in intracellular SLP is consistent with the hypothesis that SLP is involved in the process of intracellular triacylglycerol transcytosis.
Animals. Adult male Sprague-Dawley rats (150–180 g) were purchased from Sasco (Omaha, NE) and fasted overnight with access to water. At 9 AM, some animals were fed 2 ml of corn oil (2.28 mmol of triacylglycerol) by gastric gavage using an infant feeding tube (20). This dose maximally stimulates SLP production (11). Other animals received 2 ml of corn oil containing 97 mg of Pluronic L-81 (kindly provided by Dr. P. Tso, Univ. of Cincinnati) (27). At designated times after feeding or in the morning after fasting, animals were anesthetized with methoxyflurane, the abdomen opened, and the proximal small intestine removed and opened longitudinally (for eSLP isolation) or rinsed extensively with normal saline (for intracellular SLP isolation). Some sections were removed for fixation in buffered 10% formalin to process for immunocytochemistry with the biotin/streptavidin method (25). Others were fixed for transmission electron microscopy in 1% glutaraldehyde with 0.67% tannic acid and 4% p-formaldehyde containing 0.1 mM CaCl2 in 100 mM sodium phosphate buffer (pH 7.4). All protocols were approved by the Committee on Animal Research at Washington University School of Medicine.
Isolation of tissue fractions. After the bowel was opened longitudinally, the mucosal surface was washed with 20 ml of Tris · HCl buffer (10 mM, pH 7.4) containing 5 mM CaCl2 (solution A) as previously described (19). The mucosal surface was lightly scraped with Whatman no. 3 filter paper, and then the paper was soaked in solution A containing 0.1 mM phenylmethylsulfonylfluoride and 2 mM benzamidine and sonicated for 2 s at 80 W, followed by centrifugation at 600 g for 10 min to remove gelatinous surface material and remnants of paper. The resulting supernatant fraction was enriched in extracellular SLP from the apical surface of the enterocytes. SLPs were isolated on an NaBr gradient at a density of 1.07–1.08, as previously described (9). Microvillous membranes (MVM) were isolated from the remaining tissue after removing the mucosa with a glass slide by magnesium precipitation as described earlier (11).
Intracellular SLP was isolated by filling the proximal bowel loops with solution A for 10 min, discarding the luminal contents, and refilling the loops with 8–9 ml of a solution containing 10 mM EDTA (1), incubated for 10 min at 37°C, and draining the recovered enterocytes. The loops were filled again and retreated, and the underlying tissue was examined histologically to confirm that all of the cells had been removed. The luminal contents were centrifuged for 5 min at 1,500 g at 4°C, the supernatant was discarded, and the pellet was resuspended in solution A. After the cells were washed in solution A, they were homogenized with 10 strokes in a Dounce glass homogenizer, followed by gentle sonication for 1 min. The solution was centrifuged at 27,500 g for 15 min at 4°C to remove as much of the brush-border fragments as possible. Over 90% of IAP activity and >97% of sucrase activity were recovered in the pellet fraction. The supernatant was removed and applied to an NaBr gradient (0.49–1.46 M) used for isolating SLP. One-milliliter fractions were removed and assayed for IAP as described previously, reading the color in an ELISA reader at 410 nm (13).
Western blotting. Western blots were performed using the enhanced chemiluminescence system (Amersham, Arlington Heights, IL) as described previously (10). Intrinsic factor (IF) was overlaid on some Western blots to determine the presence of IF binding and the approximate size of the SLP protein to which it bound. For these experiments, 25 μg of protein were separated in 10% acrylamide gels, transferred to Immobilon-P membranes (Millipore, Bedford, MA), blocked overnight with 5% nonfat dry milk, 0.1% Tween 20 in Tris · HCl-buffered saline (pH 7.8), and washed extensively with the Tris-Tween buffer. Gels were overlaid with rat IF (25 μg in 15 ml of buffer) for1hat room temperature, washed with buffer five times, and then incubated with antibody against human IF, followed by second antibody (1:8,000).
Morphology. For immunocytochemistry, tissues were embedded in paraffin and immunostaining was performed using the standard avidin-biotin-peroxidase complex method as reported previously (24). Slides were counterstained with hematoxylin. Endogenous peroxidase was quenched by pre-treatment with 1% H2O2 in methanol for 20 min. The primary antiserum was applied after dilution for 1 h at 37°C, followed by extensive washing with PBS. Sections were incubated again for 1 h at 37°C with biotinylated goat anti-rabbit immunoglobulin G (1:200), followed by streptavidinperoxidase for 30 min at 37°C (Vector Laboratories, Burlingame, CA). Between applications, sections were extensively washed with PBS. Presence of antibody was revealed using 3,3′-diaminobenzene (Fast DAB, Sigma, St. Louis, MO) and H2O2 at room temperature for 3–5 min.
For confocal immunofluorescence, slides were prepared as described previously (24). After overnight incubation with the first primary antibody (rabbit anti-rat IAP) at 4°C, secondary antibody labeled for red fluorescence with Cy3 (Jackson Immunoresearch Laboratories, West Grove, PA) was added for 1 h at room temperature. The second primary antiserum (goat anti-human IF) and the secondary antibody labeled for green fluorescence with FITC (Sigma, St Louis, MO) were each added for 1 h at room temperature. The secondary antisera were both used at 1:100 dilution. Preparations were imaged with an upright microscope equipped with a Zeiss 63 × 1.4 numerical aperture, planapochromat objective, and a Bio-Rad MRC1024 confocal adapter (Hercules, CA). A krypton-argon laser was used with epifluorescence filter sets designed for Texas Red (SR101 and Cy3) and fluorescein (FITC-VVA and FMI-43). Ten to forty images separated by 0.25–0.5 μm were obtained for three-dimensional reconstructions.
For electron microscopy, tissues were postfixed in phosphate-buffered 1% osmic acid, dehydrated in graded concentrations of acetone, and infiltrated with Spurr's acetone as an intermediary solvent. Tissues were embedded in Spurr's resin (Electron Microscopy Sciences, Washington, PA) and viewed in a JEOL-100C transmission electron microscope.
Antisera. Anti-rat intestinal alkaline phosphatase was reported previously (11) and used at 1:200 for immunocytochemistry and at 1:3,000 for Western blotting. Goat anti-human IF antibody was raised against recombinant protein produced in baculovirus-infected Sf9 cells (25) and used at 1:4,000 for Western blots. Anti-rat intestinal IF-Cbl receptor complex was provided by the generosity of Dr. B. Seetharam (Medical College of Wisconsin, Milwaukee, WI) (23) and was used at 1:200 for immunocytochemistry and at 1:3,000 for Western blotting. Megalin was purified from rabbit kidney (5) and antibody raised by Cocalico Biologicals (Reamstown, PA). The antibody was generously provided by Dr. Z. Chen (Washington University School of Medicine) and was used at 1:8,000 for Western blotting. Rat cubilin was purified from kidney membranes as described by Birn et al. (3). Although the initial preparation of cubilin showed that >90% of the protein was a 460-kDa peptide, the unused peptide developed multiple smaller fragments on storage (see Figs. 2 and 3), and the antibody produced by Cocalico Biologicals recognized mostly lower molecular weight peptides on Western blotting (see Fig. 3), presumably due to partial proteolysis during antibody production. This property was used to identify the proteolytic fragments of cubilin. The antibody was used at 1:2,000 for Western blotting.
Identification of cubilin in SLP. Eliakim et al. (11) previously demonstrated that rat small intestinal SLP is enriched in IAP but not in other brush-border hydrolases. The small intestine is presumed to play a role in intracellular transport of absorbed triacylglycerol based on studies in rats. The SLP was observed to surround the fat droplet during fat absorption and move through the cell before entering the paracellular space by an exocytic process (7, 20). Hence, our current hypothesis is that proteins involved in intracellular movement of the lipid droplet might be present in intracellular SLP. Two intestinal apical plasma membrane proteins known to be involved in transcytosis are megalin and cubilin (6), and they might be components of SLP. Because SLP content isolated from the apical surface of the enterocyte increases after fat feeding, the presence of cubilin in isolated rat brush borders was examined at indicated times after corn oil feeding, using antiserum against the rat intestinal 460-kDa protein. In the fat-fed intestine (Fig. 1A), the major size in eSLP (and MVM as well) was 210 kDa, and smaller peptides were found as well. This size is about half of the full-length cubilin (460 kDa), but one that has been attributed to proteolysis (22), and found consistently in apical brush border preparations, even in the presence of protease inhibitors (29). The absence of the 460-kDa peptide in the zero-time sample might be due either to variability in eSLP preparation or because the zero time is taken from animals immediately after fat feeding, not fasting.
A major reactive band was found most intense at the top of the 10% gel in fasting MVMs, consistent with the expected molecular size of 460 kDa (data not shown). When other tissue fractions were examined, the band was not present consistently, with smaller bands from 50 to 70 kDa especially in the eSLP samples, as seen in Fig. 1. However, the large 460-kDa band was found reproducibly when tested against purified rabbit kidney cubilin, using the antibody raised against rat intestinal cubilin (see Fig. 2B) and antibody against rabbit kidney cubilin (see Fig. 3A). Thus proteolysis seems the most likely explanation for loss of the 460-kDa band in tissue preparations with long preparation times, such as eSLP.
eSLP maintains high levels of IAP and SLP protein at 5 h after fat feeding (28). In the present experiments, IAP activity once again peaked at 5 h after corn oil feeding (3-fold increase), but the 210-kDa cubilin content followed the pattern seen in the brush border, declining to almost undetectable levels by 5 h after corn oil feeding (Table 1, Fig. 1A). On the other hand, during the period from 3 to 5 h after feeding, IAP-specific activity in MVMs increased ∼20% at 1 and 3 h after fat feeding, which is consistent with our previously published results (11).
One possible explanation for this lack of coordinated expression was that the distribution of cubilin in brush borders and eSLP along the length of the intestine might not reflect the content in iSLP in the underlying enterocytes. In the rat, unlike humans, the distribution of cubilin is maximal in the midsmall intestine (17). Therefore, eSLP was isolated from fasting animals from the proximal, middle, and distal thirds of the intestine. Figure 1B shows that the most immunoreactivity was found in eSLP isolated from the middle third of the fasted intestine as expected and also that smaller fragments of immunoreactive cubilin were detected. In most of the experiments, the 210-kDa cubilin was found almost exclusively as the large form, especially after fat feeding. Because there was not much 460-kDa peptide in most samples and because we were interested in observing the smaller fragments of cubilin, we continued to analyze the samples on 10% gels.
We examined content of IAP and the 210-kDa cubilin in the iSLP isolated from the enterocytes from these three intestinal segments (Table 2). There was a modest (20–30%) increase in the content of IAP in iSLP but only in the proximal 2/3 of the intestine at 5 h after corn oil feeding. These changes are lower than the threefold increase seen in IAP in eSLP (Table 1), in confirmation of our previous data (19). Compared with the 90% loss of cubilin in the MVM and eSLP after corn oil feeding, ∼40% of the 210-kDa cubilin content of iSLP was preserved in iSLP (Table 2).
Cubilin and megalin in iSLP. Cubilin does not contain a transmembrane segment but is bound to plasma membranes by its interaction with the 600-kDa multiligand endocytic receptor megalin. Therefore, megalin content in rat intestinal membranes was determined (Fig. 2A). Megalin was detected in abundance in kidney MVM, as expected, and was also easily seen in eSLP from fasting animals, although less than in kidney MVMs. Moreover, the largest immunoreactive proteins seen in the kidney were absent from eSLP, a situation similar to the presence of 210-kDa cubilin seen in eSLP compared with the expected primary translation product (460 kDa). The megalin antibody did not recognize cubilin (lane 1). Surprisingly, megalin was nearly absent in iSLP, both when animals were fed corn oil alone and with Pluronic L-81, a nonionic detergent that permits uptake of fatty acids and monoglycerides but blocks secretion of chylomicrons and eSLP and enhances IAP-specific activity in eSLP (20).
When samples from the same experiment were probed with antibody against the receptor for the IFCbl complex, the 210-kDa form of cubilin was readily seen in iSLP 5 h after corn oil, although the content was considerably below that seen in eSLP from fasting animals (Fig. 2B). As was the case with megalin, no enrichment of cubilin was seen in intracellular membranes from fat-fed animals treated with Pluronic L-81. Moreover, the lower molecular weight fragments that react with the antibody in the eSLP sample were not seen in the intracellular fractions.
These membranes were isolated from animals after fat feeding, a stimulus that enhances secretion of pancreatic enzymes, perhaps explaining the absence of the 460-kDa full-size cubilin from the Western blots. Isolation of iSLP takes longer than isolation of eSLP, so that continuing proteolysis (with pancreatic or intracellular enzymes) might explain the lower cubilin content of iSLP. To test that iSLP might contain proteolytic fragments of cubilin that escaped detection by antibody against the intact protein, the same samples were probed in a fresh Western blot using the antibody raised against rat cubilin that detected the lower molecular weight bands much better than the intact 210-kDa protein. Figure 3A shows that this antibody readily detects the smaller protein fragments between 43 and 90 kDa. No cross-reacting proteins were identified in the homogenate of rat liver, a tissue that does not contain cubilin (data not shown). iSLP content of these degraded peptides was comparable with that found in eSLP, and the content of cubilin-reactive bands in the Pluronic L-81 sample was nearly identical to that of the 5-h corn oil-fed sample.
These data are consistent with the presence in the iSLP of cubilin that is degraded during purification. However, intracellular degradation of cubilin also may occur in vivo. Such degradation, coupled with the relative lack of the anchoring protein megalin, may explain the near absence of the 220-kDa cubilin in eSLP after fat feeding. Although the amount of the larger 210-kDa cubilin peptide falls during fat feeding in MVM and eSLP (Table 1), it is clear that the total of cubilin peptides in the membranes (larger plus smaller) is considerably greater and perhaps does not fall at all. However, because the reactivity with antibody could not be easily normalized for the wide range of peptide sizes found, we did not attempt to quantify the total cubilin peptides following fat feeding in either eSLP or MVM.
An experiment was performed to document that the 210-kDa reactive band in SLP was indeed cubilin. iSLP was separated by PAGE and then overlaid with IF (Fig. 3B, lane 2). IF bound to the gel over an area corresponding to an apparent molecular mass of 120–220 kDa. In a parallel experiment, eSLP was isolated from fat-fed animals treated with Pluronic L-81, a detergent that prevents the secretion of SLP from the enterocyte (27) (Fig. 3B, lane 3). As noted in Fig. 2B using iSLP, cubilin antibody identified a discrete band of ∼200 kDa, consistent with the presence of cubilin in SLP. The variability in the presence of lower molecular weight immunoreactive peptides is presumably due to differing degrees of proteolysis in the preparations.
Immunolocalization of IAP and cubilin in rat intestinal mucosa. The presence of cubilin in both extra- and intracellular SLP suggested that this large molecule, known to participate in transcytosis, might identify the membrane surrounding fat droplets. This membrane has been previously associated with IAP (8). Immunocytochemical reaction to both proteins was found along the brush-border (microvillous) membrane along nearly the entire villus (Fig. 4A), as demonstrated in earlier studies (20, 23). Use of preimmune rabbit serum produced no brown reaction product (data not shown). After corn oil feeding in rats, fat droplets were not seen with any frequency until 5 h after the gavage feeding (26). When the detergent Pluronic L-81 was fed along with the corn oil, more of the lipid was retained within the enterocytes. Intestinal tissue from these animals was examined by immunocytochemistry, using consecutive slices for IAP and for IF-receptor (cubilin) localization. Figure 4B demonstrates that intracellular IAP and cubilin could be identified in the jejunum by 3 h after fat feeding, and the pattern of retention showed that the same cells contained both proteins, a pattern most strikingly evident near the villus tips where most of the fat droplets are seen more readily at 5 h after feeding. Unlike humans, in rodents, IF-receptor activity is distributed throughout the midsmall intestine, whereas IAP is most abundant in the duodenum but easily detected in the jejunum (17). Thus the proximal jejunum offers a good site for detecting both proteins. Subsequent images show the outer half of the villi so that the relationship of the proteins to the fat droplets can be most easily seen.
By 5 h after corn oil feeding, lipid droplets can be easily identified (Fig. 4C), and cells in the outer third of the villus showed both proteins appearing to surround lipid droplets. A thin rim of positive stain can be seen just beneath the basal membrane of the enterocytes, consistent with the secretion of SLP into the lamina propria after fat feeding, as described previously for IAP alone (31). The lipid droplets were quite large in some cells, and it might be argued that the stain only appears to surround the droplet but, in fact, is pushed into proximity by the limited intracellular space. To demonstrate that this is likely not the reason for the peridroplet localization, the bottom half of Fig. 4C was examined at higher magnification. It is clear that some cells showed staining around droplets without much staining for the IF receptor in the remainder of the cell.
When the receptor localization was examined in cells fed corn oil plus Pluronic L-81 for 5 h, the retention of droplets within enterocytes was more intense and the IF-receptor was more generally present in all cells in the outer half of the villus (Fig. 4D). In these adjacent sections from the villi of another animal fed corn oil and Pluronic L-81, individual lipid droplets can be seen surrounded by immunoreactive stain using antibodies against both IAP and IF receptor. In addition, the rim of stain beneath the basal membrane of the enterocytes seen in animals fed only corn oil (Fig. 4C) was not seen when Pluronic L-81 was added. These findings are consistent with retention of SLP within the enterocyte after Pluronic L-81 treatment. To further demonstrate the colocalization of the two proteins around the lipid droplets, intestine from animals fed just corn oil for 5 h was examined by confocal microscopy (Fig. 4E). For this double-labeled experiment, rabbit anti-rat IAP and goat anti-human IF antibodies were used. The staining from the IF-receptor ligand, IF, was not so intense as that against the receptor itself, because IF is degraded intracellularly. However, only rabbit antibodies were available against the IF-receptor and IAP. Even though the intensity of stain for IF was less, it was clear that IAP and IF, representing the location of its receptor, did colocalize around lipid droplets.
The data shown in Fig. 4 appear to localize both IAP and the IF receptor to the peridroplet region, an effect enhanced by preventing secretion of lipid droplets with Pluronic L-81. Previously, this treatment was demonstrated to produce a proliferation of membranous structures around the fat droplets (8). Samples from the intestines of rats fed corn oil plus Pluronic L-81 were examined again by transmission electron microscopy (Fig. 5). Confirming the earlier results, samples from Pluronic L-81-treated rats showed the accumulation of electron-dense membranes surrounding the lipid droplets. These membranes are the presumed site for the localization of both IAP and the IF receptor cubilin. These membranes could not be appreciated in the control fat-fed tissue.
The isolation of iSLP allows confirmation of some of the assumptions made in earlier work from our laboratory. We had shown an early peak of IAP synthesis before the peak of IAP-specific activity in the brush-border membrane, consistent with delivery of IAP to the basolateral space independent of the MVM (2). In addition, the SLP on the surface of the enterocyte was stable and did not appear to reenter the enterocyte by endocytosis (2, 28). In cultured Caco-2 cells, IAP appeared basolaterally earlier than apically (30), and apical IAP increased by treatment that opened tight junctions (12). These data were consistent with the concept that iSLP was the precursor of eSLP, rather than the reverse. The data in the present paper show that iSLP has a protein composition that differs from that of eSLP, in that megalin is present only at very low levels in iSLP isolated before and during fat feeding but is present in eSLP isolated after fasting. The fall in cubilin content of eSLP during fat feeding (Fig. 1) is consistent with the interpretation that newly synthesized iSLP resulting from fat feeding might be diluting the preexisting eSLP.
This study demonstrates that cubilin, as well as IAP, is a component of SLP, both extra- and intracellularly. Megalin is a component of eSLP but is present in very low amounts in iSLP. Although IAP is enriched modestly in these intracellular membranes following fat feeding, the 210-kDa form of cubilin (the predominant form found in this study) is not. Both eSLP and iSLP were isolated following corn oil feeding for many of the experiments, and this feeding increases pancreatic secretion. The presence of the 210-kDa form of cubilin as the dominant large form in eSLP is probably related to the activity of pancreatic proteases to which the membrane is exposed, despite the inclusion of reversible inhibitors of pancreatic proteases in the preparation of iSLP. In fact, the dominant forms of cubilin in isolated iSLP are peptides, representing either large fragments of cubilin, aggregated smaller peptides, or a combination of both. Aggregation of repeating units of cubilin has been well demonstrated (26). These peptides have been produced by proteolysis, presumably, although whether this is due to continued activity of pancreatic proteases or to intracellular proteases is not certain.
The presence of megalin in eSLP was an interesting finding. This protein appears to direct cubilin to the apical surface membrane (6), but SLP is produced not from apical membranes, peaking in protein-specific activity on the apical surface of the enterocyte well before brush-border membranes are maximally labeled (2). Megalin binds the chaparone protein receptor-associated protein and helps to deliver cubilin to the apical membrane (4). Megalin is localized to the brush border, endocytic vesicles, membrane-recycling compartment, and dense apical tubules (6). Its presence in eSLP does not appear to result from movement of iSLP to the surface of the cell but may involve transfer from the adjacent brush-border. Its presence may explain the localization of cubilin to eSLP, but cubilin can still localize to the kidney brush border membrane in the absence of megalin (16). Megalin is important in initiating endocytosis, but eSLP does not appear to be endocytosed, and endocytosis is not felt to be important in the process of fat absorption (21). Thus the role of megalin in eSLP is not clear.
The lack of megalin in iSLP may be related to the directed intracellular localization of iSLP to the lipid droplet rather than to the apical membrane, because new SLP is synthesized during fat feeding. The presence of cubilin in iSLP may be related to the hydrophobic nature of cubilin and to the fact that some cubilin may escape capture by megalin during synthesis. The NH2-terminal region of cubilin is thought to interact with the outer leaflet of the brush-border membrane (30) and, by analogy, may also interact with the surface of the lipid droplet as it moves through the cell. However, on the basis of current information, there is no way to know whether cubilin interacts with the membrane of SLP or with another protein component of SLP.
The data presented in this paper raise the possibility that the composition of the membrane surrounding the fat droplet is not a constant one but may vary in protein content as it traverses the cell. Cubilin may play a role in directing the iSLP to the newly forming lipid droplet or in acting as one of the proteins that mediates transcytosis of the droplet. Because cubilin binds proteins other than intrinsic factor (e.g., albumin, transferrin), additional functions are possible. Inside the cell, cubilin in iSLP may not bind much intrinsic factor, as IF bound to brush-border cubilin is directed to the lysosome, where it is degraded by the lysosomal protease cathepsin L (15). However, some IF can be detected surrounding lipid droplets (Fig. 4E), an unusual location for lysosomes. Thus cubilin in iSLP probably does retain the ability to bind proteins, as the small repeat units of cubilin have been shown to do (30).
eSLP is secreted basolaterally after its synthesis, and ∼25% of this membrane moves through the tight junction to reside on the surface of the enterocyte (12). eSLP contains more cubilin and much more megalin than iSLP. Because eSLP does not undergo endocytosis from the apical surface of the enterocyte, the nearly complete lack of megalin in iSLP and the marked drop in the 210-kDa cubilin found in eSLP after fat feeding (Fig. 1A) suggest that after fat feeding, some of the iSLP also moves to the apical surface of the cell where it dilutes the eSLP already there. It is not clear whether the lack of large cubilin or the presence of cubilin fragments in eSLP after fat feeding reflects the action of pancreatic proteases in vivo or is merely a reflection of proteolysis that occurs during processing of the membranes.
The observation that cubilin is present in SLP not only extends the possibility for additional functions for cubilin but also provides an additional marker for SLP, both extra- and intracellularly. The role of SLP in fat absorption may be investigated further in animals in which the genes for IAP and/or cubilin have been knocked out. In megalin-knockout mice, absorption of glucose and amino acids are not affected, but lipid absorption was not studied in this model (18). The availability of other knockout models that are less lethal than the megalin one may help to advance knowledge about the role of SLP in fat absorption.
This work was supported, in part, by National Institutes of Health Grants AM-14038 and DK-33487.
We thank Dr. S. Anant for help in the review of the manuscript and in preparation of the figures.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2003 the American Physiological Society