Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 2002 Jul 1;542(Pt 1):237–244. doi: 10.1113/jphysiol.2002.017087

Secreted intestinal surfactant-like particles interact with cell membranes and extracellular matrix proteins in rats

Akhtar Mahmood 1, Michael J Engle 1, David H Alpers 1
PMCID: PMC2290410  PMID: 12096065

Abstract

Surfactant-like particles (SLP) are secreted from enterocytes basolaterally into the lamina propria, and reach the apical surface through the intercellular tight junctions. Interactions of SLP with apical and basolateral membranes and with extracellular matrix proteins were measured using a solid-phase binding assay and gel overlays. Small-intestinal SLP bound to basolateral membranes much more than to apical membranes, and more tightly to fibronectin than to laminin (affinity constant Ka = 1.23 × 10−2 μg vs. 0.67 × 10−2 μg; maximal number of binding sites 4.1 μg ml−1vs. 0.32 μg ml−1), but did not bind to collagen types I or IV. Small-intestinal SLP bound fibronectin more than colonic or gastric SLP. Binding to fibronectin was inhibited only partially by RGD peptide and gelatin, but not by heparin. An antibody against αv integrin also identified the fibronectin-binding component in SLP at ∼220 kDa, which is the expected size for integrin heterodimers. SLP binding to apical microvillous membranes was weaker and was inhibited by heparin. SLP bound more strongly to heparin itself, and this binding was inhibited by glucuronic acid and chondroitin sulfate. These data are consistent with the hypothesis that the time spent by secreted SLP in the lamina propria is prolonged by strong interactions with proteins in the basolateral membranes, and in the intestinal lumen by weaker interactions with apical membrane components, including heparin. These interactions may allow SLP the time to exert their functions in each tissue compartment.

Surfactant-like particle (SLP) is a specialised membrane that is secreted from enterocytes, colonocytes and gastric epithelial cells. It exhibits surfactant activity and contains dipalmitoylphosphatidylcholine and surfactant proteins, and is enriched with alkaline phosphatase (De Schryver-Kecskemeti et al. 1989; Eliakim et al. 1989, 1997). SLP is secreted basolaterally, but about 25 % moves through tight junctions and presents on the surface of the apical microvillous membrane (Alpers et al. 1995; Engle et al. 1995). It has a long residence time in this apical location (Alpers et al. 1995) and functions in part to bind luminal bacteria (Goetz et al. 1999; Mahmood et al. 2000). Furthermore, in the enterocyte it is involved in transcytosis of absorbed lipid droplets, and is delivered along with the lipid droplet into the intercellular space (DeSchryver-Kecskemeti et al. 1991). It has been demonstrated using immunocytochemistry that SLP is distributed within epithelial cells and in layers on the luminal side and beneath the basal membrane of mucosal cells (Zhang et al. 1996).

The factors that allow the retention of 75 % of the secreted SLP within the lamina propria have not yet been identified. Both pulmonary surfactant and SLP contain dipalmitoylphosphatidylcholine and specific amphipathic surfactant proteins (Eliakim et al. 1989, 1997). Pulmonary surfactant itself is associated with the cross-linked fibrin-fibronectin complex, and decreases the factor-XIIIa-mediated binding of fibronectin to fibrin (Elssner et al. 1999). This could provide a matrix for lung remodelling. The interaction required more than dipalmitoylphosphatidylcholine alone, as co-incubation with two commercially available surfactants, but not the phospholipid alone, reduced fibronectin incorporation into fibrin clots. In addition, the reaction was not dependent upon calcium, as would be expected for a fibronectin-integrin interaction (Elssner et al. 1999). Because SLP assume functions in the gastrointestinal tract that can be quite different from those of pulmonary surfactant in the lung, the question is raised whether they interact with fibronectin, but not perhaps by the integrin-independent mechanism reported in the lung.

Both intestinal mucosal cells and mesenchymal cells secrete extracellular matrix (ECM) proteins such as fibronectin and laminin, and other components of basement membranes (Simon-Assmann et al. 1986). Rat fetal intestinal cell lines (e.g. IEC-6) synthesise fibronectin and laminin near cell-cell adhesion areas, demonstrating that the resulting proteins are not just cell-surface associated (Quaroni et al. 1978; Scarpa et al. 1988). Fibronectin is also made in undifferentiated Caco-2 cells, a human colonic adenocarcinoma cell line that is a model for enterocyte differentiation (Vachon et al. 1995). However, the expression is nearly undetectable in differentiated cells. In rats the ECM proteins are distributed along the intestinal villus and at the epithelial-mesenchymal interface, although the precise pattern varies during postnatal development (Simon-Assman et al. 1986). In the adult rat, fibronectin and collagen type III are closely associated with not only the basement membrane, but also throughout the lamina propria, and appear to be integrated into the basal lamina of rat duodenal villi (Laurie et al. 1982).

The study reported here was undertaken to assess the interaction of SLP with fibronectin and other ECM proteins. It has been observed that in vitro, preferential binding of SLP to fibronectin in basolateral membranes occurs more than to laminin, and in brush-border membranes (BBMs) it binds preferentially to heparin. Binding to fibronectin was inhibited by gelatin and the RGD-containing peptide that identifies fibronectin binding to integrins, but was not inhibited by heparin. Binding to heparin was inhibited by both glucuronic acid and chondroitin sulfate. These findings may explain, at least in part, the distribution of SLP along the apical and basal surfaces of the epithelial cell in the gastrointestinal tract. Such a distribution is important for providing access to the lumen for bacterial binding or to the submucosa after SLP are delivered to that compartment following triacylglycerol absorption.

METHODS

Materials

SLP were prepared from rat small bowel and colon (Eliakim et al. 1989, 1997). Animals were anaesthetised with sodium pentobarbital (60 mg kg−1) administered i.p. until the animals were unresponsive to a forceps pinch in the foot. The thorax was then opened, a blunt needle was inserted through the left ventricle into the ascending aorta, the right atrium was opened, and the animal was perfused with normal saline until the liver was completely blanched; the animal was then killed by cervical dislocation. This protocol was approved by the Committee on Animal Experimentation at Washington University School of Medicine. Microvillous membranes and basolateral membranes were isolated from rat and human enterocytes, as described previously (Mahmood et al. 1993). Rabbit anti-small-bowel- and colonic-SLP antibodies have been characterised previously both by Western blot (Eliakim et al. 1989, 1997; Mahmood et al. 1993) and by immunocytochemistry (Zhang et al. 1996). Goat anti-human fibronectin and rabbit anti-human αv integrin antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Mouse monoclonal antibodies raised against heparin, purchased from Chemicon (Temecula, CA, USA), cross-reacted with dextran sulfate and heparan sulfate. Human fibronectin, laminin, GRGDNP peptide, GPenGRGDSPCA (cyclical), and rabbit anti-rat fibronectin antibodies were obtained from Life Technologies (Grand Island, NY, USA). Rabbit anti-rat fibronectin antibody was used for most experiments, as it cross-reacted well with the human fibronectin used as a positive control on Western blot and for enzyme-linked immunosorbent assay (ELISA). Porcine heparin, ratcollagen types I and IV, glucuronic acid, bovine chondroitin sulfate and gelatin were purchased from Sigma (St Louis, MO, USA). Rat small-intestinal mucins were isolated by the method of Ouwehand et al. (1995).

Binding studies

Binding assays to SLP, BBMs, basolateral membranes and to individual proteins (e.g. fibronectin) were performed using a modified ELISA assay (Hendriks et al. 1987), as described previously for binding of bacterial adhesins to SLP (Goetz et al. 1999). After binding of the acceptor membrane/protein, the plastic wells (BD Falcon 96 well plates, BD Biosciences, USA) were filled with 4 % bovine serum albumin and air-dried to coat the uncovered surfaces. The ligand was bound routinely in the presence of 5 mm calcium at pH 7.4, probed using specific antisera and second antibody coupled with biotin, and identified by the avidin-biotin system with horseradish peroxidase (Vector Laboratories, Carlsbad, CA, USA), detected with 3,3′-diaminobenzidine tetrahydrochloride (FAST DAB, Sigma). The data were calculated and plotted using the least-squares regression in Microsoft Excel. The ELISA for fibronectin itself was performed in the same way, except that the membrane was not added before the antibody against rat fibronectin, and a standard curve for human fibronectin was used to estimate levels of rat fibronectin. For pH maximum experiments, buffers were selected to provide intervals of 0.5 pH units from pH 3.0 to 9.0, including acetate (pH 3-5.5), phosphate (pH 6-7.5), and Tris-HCl (pH 8-9). After the solutions were made by calculation, they were measured by a pH meter, and were found to be within 0.1 pH unit of predicted values.

Electrophoretic studies

Western blots were performed as described previously (Mahmood et al. 1993). Antibodies used included those raised against rat small-intestinal SLP (1:4000 dilution), αv integrin (1:1000 dilution), human fibronectin (1:1000 dilution) and heparin (1:1000 dilution).

RESULTS

SLP binding to ECM proteins

SLP is a phospholipid-rich substance known to be secreted basolaterally from intestinal epithelial cells (Engle et al. 1995). Since pulmonary surfactant is known to interact with fibronectin in pulmonary cells (Elssner et al. 1999), experiments were carried out to determine whether fibronectin is also present as a complex with intestinal SLP. To this end, two separate SLP preparations were isolated from the rat jejunum and subjected to Western blot analysis for fibronectin (Fig. 1). In both examples, a reactive protein that migrated slightly more slowly (apparent Mr ∼ 220) than fibronectin was observed. This demonstrates that SLP interacted with either fibronectin or a fibronectin-like protein.

Figure 1. Fibronectin detection in surfactant-like particles (SLP) in rat small-intestinal mucosa.

Figure 1

Open in a new tab

Human fibronectin (5 μg) or two separate preparations of rat SLP, designated SLP1 and SLP2 (14-18 μg protein) were resolved on gels, transferred to polyvinylidine (PVD) membranes, and incubated with goat anti-human fibronectin (FN) antibody (1:4000 dilution), as described in Methods.

The content of fibronectin in these freshly isolated SLP preparations, as assessed by ELISA, was 2 ± 0.6 ng (100 ng SLP protein)−1 (n = 4), slightly higher than that found in basolateral membranes (1.3 ± 0.4 ng (100 ng SLP protein)−1). This small amount of fibronectin did not appear to be antigenic when an antiserum was raised against rat small-intestinal SLP, because this antibody did not recognise human fibronectin on Western blot (data not shown). Such a reaction should have been expected if anti-rat fibronectin antibodies were present, because the antibody against human fibronectin recognises rat fibronectin (Fig. 1). Many attempts were made to remove the adherent fibronectin from freshly isolated membranes using, for example, changes in pH, salt concentration and antibody affinity columns, but none were successful.

To demonstrate a direct interaction between intestinal SLP and fibronectin, in vitro studies were performed. Table 1 shows that rat small-intestinal SLP bound to human laminin and fibronectin very efficiently. However, binding to collagen type I and type IV was seen only when relatively very large amounts of SLP were added. Binding of rat small-intestinal SLP to laminin exhibited a Ka of 0.67 × 10−2 μg, and a maximal number of binding sites (Bmax) of 0.32 μg ml−1. Comparable experiments using human fibronectin showed a Ka of 1.23 × 10−2 μg, and a Bmax of 4.1 μg ml−1.

Table 1.

Rat small-intestinal surfactant-like particle (SLP) binding to extracellular matrix (ECM) proteins

ECM protein SLP added (ng) SLP bound
(ng ± s.d.) (%)
Collagen type I 75 n.d. 0
750 142 ± 28 19
Collagen type IV 75 n.d. 0
880 54 ± 13 6
Laminin 90 87 ± 33 89
75 128 ± 3 17
Fibronectin 75 71 ± 18 95
600 93 ± 10 16
Open in a new tab

Wells were coated with 4-5 μg of ECM protein, air dried, blocked with with 4% bovine serum albumin, and dried again before adding rat small-bowel SLP. An antibody to SLP was added, followed by a biotinylated goat anti-rabbit IgG and streptavidin linked with horseradish peroxidase. The presence of antibody was revealed using 3,3′-diaminobenzidine and H2O2 at room temperature for 3–5 min. The amount of SLP bound was red from a standard curve using purified SLP. N = 4; n.d. = not detected.

Fibronectin bound best to small-intestinal SLP, and less well to SLP from the colon (Fig. 2A). To determine whether there were differences in the interactions of rat versus human intestinal SLP, human SLP were isolated from gastrointestinal tissues and subjected to similar binding experiments. In this case, the highest levels of fibronectin binding were seen with small-intestinal SLP when compared with SLP isolated from either the colon or the stomach (Fig. 2B). Although the degree of fibronectin binding to human and rat colonic SLP was comparable, rat small-intestinal SLP were able to bind more fibronectin than the comparable human preparation (compare Fig. 2A and B). The Bmax of fibronectin binding to human small-intestinal SLP was 1.1 μg ml−1. Binding of human fibronectin to rat small-intestinal SLP was maximal at pH 6.4-7.4, but the curve was very broad, and binding at pH 3.1 was still 50 % of maximal values (data not shown).

Figure 2. SLP binding to fibronectin.

Figure 2

Open in a new tab

A, rat small-intestinal and colonic SLP binding to rat fibronectin. Wells of the assay plates were coated with fibronectin (0.5 μg) and blocked with 4 % bovine serum albumin. SLP binding was studied by adding 5-60 ng of SLP protein from rat small intestine or colon, as described in Methods. Values are the mean ± s.d. of four determinations. B, human small-intestinal, colonic and gastric SLP binding to human fibronectin. Wells were coated and blocked as described in Table 1. An aliquot (5-50 ng) of human SLP protein isolated from small intestine, colon or stomach was added, as described in Methods. Values are the mean ± s.d. of four determinations.

Proteins responsible for binding of fibronectin to SLP

The specificity of SLP binding to fibronectin was examined using peptides that inhibit cell attachment to ECM proteins. SLP binding to fibronectin was inhibited by 37 % by the peptide GRGDNP, which inhibits cell attachment to fibronectin much more than to vitronectin (Table 2). However, the related cyclic peptide, GPenGRGDSPA, is a much more specific inhibitor for vitronectin-mediated cell attachment (Kumagai et al. 1991), and inhibited SLP binding to fibronectin by only 3 %. In addition, both gelatin and chondroitin sulfate inhibited SLP binding to fibronectin, by 14 % and 28 %, respectively (Table 2). These data were consistent with some, but not necessarily all, fibronectin binding to integrins in SLP.

Table 2.

Inhibition of rat small-intestinal SLP binding to fibronectin

Inhibitor added (μMg) SLP bound
(μg) (ng ± s.d.) (%) (% inhibition)
None 78 ± 8 87 —
RGDNP (20) 48 ± 7 55* 37
GPenGRGDSPA (20) 74 ± 9 84 3
Gelatin (20) 66 ± 3 75** 14
Heparin (20) 78 ± 8 87 0
Chondroitin sulfate (20) 55 ± 3 63** 28
Open in a new tab

Wells were coated with 0.5 μg well−1 of human fibronectin, air dried, blocked with 4% bovine serum ilbumin, air dried again, and small-bowel SLP (88 ng) was added with or without inhibitors. See legend of lable 1 for further details. The amount of SLP bound was determined as described in Table 1. Student's t test I two-tailed) was used for statistical analysis.

*

P± 0.01

**

P± 0.05.

To identify integrins in SLP, a Western blot was performed using an antibody against αv integrin. This integrin was chosen because it is a component of the heterodimers that bind both fibronectin and vitronectin. Rat small-intestinal SLP contained small amounts of αv integrin, and the reactive band was found most intensely near the top of the gel (apparent Mr ∼ 220), consistent with the size of an intact integrin heterodimer (Fig. 3).

Figure 3. Western blot of integrin αv in rat small-intestinal SLP.

Figure 3

Open in a new tab

Fifteen micrograms of rat SLP or brush-border membrane (BBM) proteins, or 5 μg human fibronectin were separated on a 10 % SDS-PAGE gel, transferred to PVD membranes and incubated with antibodies raised against integrin αv (1:1000 dilution), and developed as described in Methods. Note that integrin was detected only in the SLP preparation.

SLP binding to intestinal membrane proteins

Integrins are widely distributed in human and rodent enterocytes (Simon-Assmann et al. 1994). Both α- and β-subunits are found in the villus and crypt cells in the adult rat intestine (Rao et al. 1994). Basolateral membranes contain many integrins, and would be expected to bind very well to a membrane such as SLP that contains fibronectin and perhaps other ECM proteins (Keely et al. 1995). Table 3 shows that under the conditions used, both rat and human basolateral membranes bound rat SLP well. Similar to the situation with isolated SLP (Fig. 3), fibronectin isolated from freshly prepared basolateral membranes was bound to SLP proteins in a gel overlay system with an apparent Mr similar to that of αv integrin heterodimers. However, this finding was not observed with fibronectin using microvillous membranes (data not shown). Taken together, these data suggest that fibronectin is bound in vivo to integrin heterodimers in small-intestinal SLP.

Table 3.

Rat small-intestinal SLP binding to intestinal membranes and proteins

Protein Inhibitor added SLP added SLP bound
(μg) (μg) (ng) (ng ± s.d.) (%)
BLM, rat (l) none 125 99.6 ± 4.9 80
BLM, human (1.2) none 180 74.6 ± 14.2 41
BBM (10) none 1000 29 ± 4.5 3
heparin (50) 1000 6 ± 0.9 0.6
Heparin (0.4) none 92 51 ± 8.4 55
chondrotin sulfate (20) 92 23 ± 3.7 25
heparin (10) 92 26 ± 2.9 28
glucuronic acid (20) 92 24 ± 5.3 26
Mucins (5) none 145 5 ± 0.86 3
Open in a new tab

Wells were coated with the indicated protein, and small-bowel SLP was added in concent rations that did not satutate binding capacity, but produced easily detectable binding. Where indicated, inhibitor was added to the binding reaction. Binding was determined by the avidin-biotin system, using horseradish peroxidase as the detection system. N = 4 for all conditions; BLM = basolateral membranes, BBM = brush border membranes.

Although SLP is secreted basolaterally, about 25 % of the membrane passes intercellularly and localises to the apical cell surface (Engle et al. 1995). Moreover, SLP is present in abundance bound to the apical surface of enterocytes, as that is the source of purified SLP (De Schryver-Kecskemeti et al. 1989). Therefore, we examined binding of rat small-intestinal SLP to proteins that might be present at the apical surface of the enterocyte. Table 3 shows that SLP bound to BBMs, but much less than to basolateral membranes, only 3 % versus 80 %. Because of the lower level of binding, a higher concentration of SLP was used to demonstrate clear binding over background. The binding of SLP to BBM was largely reversed by the addition of heparin, a surface component of the intestinal BBM. SLP bound to heparin itself, a surface proteoglycan, and this binding was inhibited by the addition of heparin, glucuronic acid and chondroitin sulfate (Table 3). The capacity for rat small-intestinal SLP binding to heparin was low, with a Bmax of about 0.3 μg heparin (ng SLP protein)−1 (Fig. 4). SLP did not bind to intestinal mucins. These data suggest that SLP binding to apical BBMs could be mediated in part by heparin or heparin-like proteoglycans, but the membrane containing heparin could be either the SLP itself or the BBM, or both.

Figure 4. Rat small-intestinal SLP binding to heparin.

Figure 4

Open in a new tab

Wells of the assay plates were coated with heparin, blocked with 4 % bovine serum albumin, and rat small-intestinal SLP binding was determined by adding 167 ng of SLP protein per well. Values represent the mean ± s.d. of four determinations.

Heparin was identified by Western blot only in rat SLP that had been freshly isolated from the small intestine (Fig. 5), but not in the apical brush border or basolateral membranes, suggesting that the heparin on SLP was interacting with other proteoglycans on the surface of the epithelial cell. However, the heparin was detected most abundantly at an apparent Mr much larger (60-70) than that of the Sigma purified heparin (average Mr ∼ 6), suggesting the formation of a heteromeric complex. Heparin did bind to isolated apical BBMs, but at a much lower capacity than to SLP, 20 pg heparin (μg brush-border protein)−1.

Figure 5. Western blot of heparin in rat intestinal membranes.

Figure 5

Open in a new tab

Rat small-intestinal SLP, basolateral membrane (BLM) and BBM (43-48 μg) were separated on SDS acrylamide gels, transferred to PVD membranes, and incubated with a mouse monoclonal antibody raised against human heparin (1:1000 dilution), as described in Methods.

DISCUSSION

SLP, a membrane enriched in phospholipid and alkaline phosphatase, is secreted basolaterally from the epithelial cells lining the gastrointestinal tract (Engle et al. 1995). Most of the SLP is recovered in the lamina propria tissue space, where it enters a pool of SLP (Yamagishi et al. 1994). Some SLP passes through the intercellular space and onto the apical surface of the cell, where its turnover is slow (Alpers et al. 1995). Moreover, the SLP remains associated with the apical membrane of its tissue of origin; it has been shown that small-intestinal SLP isolated from the surface of the enterocyte contains only intestinal alkaline phosphatase, and colonic and gastric SLP contain only tissue non-specific alkaline phosphatase (Eliakim et al. 1989, 1997). Given the location, we hypothesised in the present study that SLP would bind to components of the lamina propria and apical enterocyte membrane. Indeed, we found that SLP bound to the protein components of the lamina propria (fibronectin and laminin) and to a heparin-binding component of the apical BBM. Figure 6 shows a schematic representation of the proposed extracellular distribution of intestinal SLP. The relatively greater binding of small-bowel SLP to fibronectin may reflect the fact that the secretion of this type of SLP is primarily basolateral. No data are available on the direction of secretion of colonic SLP in vivo, but it seems likely that levels of secretion might be lower than in the small bowel, as both blood and lymphatic flow are lower in the colon. After SLP is secreted from the cell, the alkaline phosphatase is removed by the abundant serum phospholipase D, but SLP is found localised beneath the basement membrane of the enterocyte (Zhang et al. 1996).

Figure 6. Schematic diagram of the proposed disposition of secreted rat intestinal SLP.

Figure 6

Open in a new tab

Following basolateral secretion from the enterocyte about 75 % of the SLP is recovered in the lamina propria, where it binds to fibronectin via interactions with integrins and other protein(s). SLP may be retained in the lamina propria by binding of this fibronectin to cellular integrins. SLP also move apically through the tight junctions (Engle et al. 1995), and the 25 % of SLP that is apically recovered remains associated with the apical membrane. This association may be due to interactions with heparin or other uronic acids (see text for details).

Sources of fibronectin might be either from the enterocyte or mesenchymal cells in the lamina propria. Indeed, fibronectin and laminin are made in rat fetal duodenal (IEC-17) cells and are found in areas of cell-cell adhesion (Scarpa et al. 1988). However, fibronectin production is greatest when the enterocyte is relatively undifferentiated, and becomes almost undetectable in differentiated Caco-2 cells, a model for differentiated fetal-like ileal cells (Levy et al. 1994; Vachon et al. 1995). This suggests that fibronectin is made in crypt epithelial cells and not in the cells lining the mature villus. In addition to the enterocyte, in the adult animal mesenchymal cells appear to be good sources of fibronectin (Simon-Assmann et al. 1986). Thus, there appear to be multiple potential sources for the production of fibronectin in the intestinal mucosa.

Fibronectin is known to bind to both hydrophilic and hydrophobic surfaces (Detrait et al. 1999), and to certain proteins such as collagen, fibrin, gelatin, glycosaminoglycans and cell receptors, including integrins (Haas et al. 1984). Fibronectin binding to SLP was blocked by GRGDNP peptide, but only partially. Furthermore, specificity was demonstrated by the lack of inhibition with the cyclic RGD peptide GPenGRGDSPCA, which inhibits cell attachment to vitronectin, but not fibronectin (Kumagai et al. 1991). Consistent with this observation is the presence of αv integrin, a fibronectin-binding integrin, in secreted SLP. Moreover, fibronectin already bound to isolated SLP migrated at an apparent Mr of ∼220, reflecting either the apparent Mr of rat fibronectin itself, or fibronectin that has not been displaced completely from the integrin heterodimer during SDS-PAGE. It seems likely that binding of fibronectin to SLP is the result of binding to heterodimeric integrins, at least in part. The presence of mRNA encoding both α- and β-integrins in the mature rat enterocyte (Rao et al. 1994), and the wide spatial distribution among cell membranes of epithelial cells (Keely et al. 1995) is consistent with the finding of αv integrin in SLP.

There are other examples of fibronectin binding to cell membranes that is not mediated by integrins. Two fibronectin binding proteins (47 and 65 kDa) that are not inhibited by RGD peptides or by an antibody raised against the cell binding domain in fibronectin are secreted from Borrelia burgdorferi (Probert et al. 1998). Other Borrelia species express similar proteins. An 85 kDa protein that binds to the gelatin and collagen binding domain of fibronectin is expressed by Mycobacterium bovis (Peake et al. 1993). This binding was enhanced by the addition of heparin, perhaps by increasing accessibility to this binding site (Khan et al. 1988).

The purpose of SLP binding to fibronectin seems to be the retention of SLP in the lamina propria, although the functional significance of prolonged residence time in that tissue compartment is not clear. One possibility is to modify the response to tissue injury, as found in the lung for pulmonary surfactant. The bronchial lavage from rats exposed to oxygen damage contained increased levels of fibronectin as well as pulmonary surfactant (Bhalla et al. 1999). It was suggested that fibronectin plays a role in the response to injury, particularly as the increased fibronectin secretion fell by 48 h after injury. Pulmonary surfactant and two commercial surfactant products also decreased factor XIIIa-mediated fibronectin binding to fibrin, but dipalmitoylphosphatidylcholine did not (Elssner et al. 1999). Thus, the proteins of pulmonary surfactant appear to play a role in fibronectin interactions after injury. Future studies will be needed to determine whether a similar role is mediated by intestinal SLP.

The binding of SLP to apical microvillous membranes was weaker than to basolateral membranes, but this binding may serve to keep the membrane associated with the apical portion of the cell, consistent with the long residence time in this compartment in vivo (Alpers et al. 1995). The uropathogenic Escherichia coli binds to both small-intestinal and colonic SLP, whereas microvillous membranes do not (Goetz et al. 1999; Mahmood et al. 2000), consistent with a role for SLP in providing a site for bowel colonisation of these organisms. At least some of the SLP binding to the microvillous membrane may be mediated by heparin. This proposed role of heparin would be similar to that suggested for binding cholesterol esterase (Bosner et al. 1988), a mechanism thought to enhance cholesterol absorption by providing a kinetic advantage for the released cholesterol (Bosner et al. 1989). In that case, the binding of the esterase was inhibited by heparin, but the heparin content of brush borders was measured by a method that is used to detect uronic acids, not the monoclonal antibody against heparin used in the current studies. Other brush-border proteins or uronic-acid-containing structures may play a role in binding SLP, as suggested by the fact that chondroitin sulfate inhibits the binding of SLP to both heparin and fibronectin. Further studies will be needed to explore the possibilities of other ECM or matrix components that might interact with SLP.

Acknowledgments

This work was supported in part by NIH grant DK14038 and a DDRCC core centre grant DK57524. We thank Shrikant Anant for helpful suggestions.

REFERENCES

  1. Alpers DH, Zhang Y, Ahnen DJ. Synthesis and parallel secretion of rat intestinal alkaline phosphatase and a surfactant-like particle protein. American Journal of Physiology. 1995;268:E1205–1214. doi: 10.1152/ajpendo.1995.268.6.E1205. [DOI] [PubMed] [Google Scholar]
  2. Bhalla DK, Gupta SK, Reinhart PG. Alteration of epithelial integrity, alkaline phosphatase activity, and fibronectin expression in lungs of rats exposed to ozone. Journal of Toxicology and Environmental Health. 1999;56:329–343. doi: 10.1080/009841099157647. [DOI] [PubMed] [Google Scholar]
  3. Bosner MS, Gulick T, Riley DJS, Spilburg CA, Lange LG., III Receptor-like functions of heparin in the binding of neutral lipids. Proceedings of the National Academy of Sciences of the USA. 1988;85:7438–7442. doi: 10.1073/pnas.85.20.7438. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bosner MS, Gulick T, Riley DJS, Spilburg CA, Lange LG., III Heparin-mediated binding of pancreatic lipase and uptake of hydrolysed triglyceride in the intestine. Journal of Biological Chemistry. 1989;254:20261–20264. [PubMed] [Google Scholar]
  5. Detrait E, Lhoest JB, Bertrand P, van den Bosch de Aquilar P. Fibronectin-pluronic coadsorption on a polystyrene surface with increasing hydrophobicity: relationship to cell adhesion. Journal of Biomedical Materials Research. 1999;45:404–413. doi: 10.1002/(sici)1097-4636(19990615)45:4<404::aid-jbm16>3.0.co;2-4. [DOI] [PubMed] [Google Scholar]
  6. De Schryver-Kecskemeti K, Eliakim R, Caroll S, Stenson WF, Moxley MA, Alpers DH. Intestinal surfactant-like material: a novel secretory product of the rat enterocyte. Journal of Clinical Investigation. 1989;84:1355–1361. doi: 10.1172/JCI114306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. DeSchryver-Kecskemeti K, Eliakim R, Green K, Alpers DH. A novel intracellular pathway for rat intestinal digestive enzymes (alkaline phosphatase and sucrase) via a lamellar particle. Laboratory Investigation. 1991;65:365–373. [PubMed] [Google Scholar]
  8. Eliakim R, DeSchryver-Kecskemeti K, Nogee L, Stenson WF, Alpers DH. Isolation and characterization of a small intestinal surfactant-like particle containing alkaline phosphatase and other digestive enzymes. Journal of Biological Chemistry. 1989;264:20614–20619. [PubMed] [Google Scholar]
  9. Eliakim R, Goetz GS, Rubio B, Chailley-Heu B, Shao J -S, Ducroc R, Alpers DH. Isolation and characterization of surfactant-like particle in rat and human colon. American Journal of Physiology. 1997;272:G425–434. doi: 10.1152/ajpgi.1997.272.3.G425. [DOI] [PubMed] [Google Scholar]
  10. Elssner A, Mazur G, Vogelmeier C. Inhibition of XIIIa-mediated incorporation of fibronectin into fibrin by pulmonary surfactant. American Journal of Physiology. 1999;276:L625–630. doi: 10.1152/ajplung.1999.276.4.L625. [DOI] [PubMed] [Google Scholar]
  11. Engle MJ, Grove ML, Becich MJ, Mahmood A, Alpers DH. Appearance of surfactant-like particle in apical medium of Caco-2 cells may occur via tight junctions. American Journal of Physiology. 1995;268:C1401–1413. doi: 10.1152/ajpcell.1995.268.6.C1401. [DOI] [PubMed] [Google Scholar]
  12. Goetz GS, Mahmood A, Hultgren SJ, Engle MJ, Dodson K, Alpers DH. Binding of pili from uropathogenic Escherichia coli to membranes secreted by human colonocytes and enterocytes. Infection and Immunity. 1999;67:6161–6163. doi: 10.1128/iai.67.11.6161-6163.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Haas R, Culp LA. Binding of fibronectin to gelatin and heparin: effect of surface denaturation and detergents. FEBS Letters. 1984;174:279–283. doi: 10.1016/0014-5793(84)81173-4. [DOI] [PubMed] [Google Scholar]
  14. Hendriks HG, Koninkx JF, Draaijer M, van Dijk JE, Raaijmakers JA, Mouwen JM. Quantitative determination of lectin binding capacity of small intestinal brush border membranes. An enzyme linked lectin sorbent assay (ELLSA) Biochimica et Biophysica Acta. 1987;905:371–375. doi: 10.1016/0005-2736(87)90465-2. [DOI] [PubMed] [Google Scholar]
  15. Keely PJ, Wu JE, Santoro SA. The spatial and temporal expression of the α-2, β-1 integrin and its ligands, collagen I, collagen IV, and laminin, suggest important roles in mouse mammary morphogenesis. Differentiation. 1995;59:1–13. doi: 10.1046/j.1432-0436.1995.5910001.x. [DOI] [PubMed] [Google Scholar]
  16. Khan MY, Jaikaria NS, Frenz DA, Villanueva G, Newman SA. Structural changes in the NH2-terminal domain of fibronectin upon interaction with heparin. Journal of Biological Chemistry. 1988;263:11314–11318. [PubMed] [Google Scholar]
  17. Kumagai H, Tajima M, Ueno Y, Giga-hama Y, Ohba M. Effect of cyclic RGD peptide on cell adhesion and tumor metastasis. Biochemical and Biophysical Research Communications. 1991;177:74–82. doi: 10.1016/0006-291x(91)91950-h. [DOI] [PubMed] [Google Scholar]
  18. Laurie GW, Leblond CP, Martin GR. Localization of type IV collagen, laminin, heparin sulfate proteoglycans, and fibronectin to the basal lamina of basement membranes. Journal of Cell Biology. 1982;95:340–344. doi: 10.1083/jcb.95.1.340. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Levy P, Loreal O, Munier A, Yamada Y, Picard J, Cherqui G, Clement B, Capeau J. Enterocytic differentiation of human Caco-2 cell line is correlated with down-regulation of fibronectin and laminin. FEBS Letters. 1994;338:272–276. doi: 10.1016/0014-5793(94)80282-3. [DOI] [PubMed] [Google Scholar]
  20. Mahmood A, Mahmood S, DeSchryver-Kecskemeti K, Alpers DH. Characterization of proteins in rat and human intestinal surfactant-like particle. Archives of Biochemistry and Biophysics. 1993;300:280–286. doi: 10.1006/abbi.1993.1039. [DOI] [PubMed] [Google Scholar]
  21. Mahmood A, Engle MJ, Hultgren SJ, Goetz GS, Dodson K, Alpers DH. Role of surfactant-like particles as a potential reservoir of uropathogenic Escherichia coli. Biochimica et Biophysica Acta. 2000;1523:49–55. doi: 10.1016/s0304-4165(00)00095-7. [DOI] [PubMed] [Google Scholar]
  22. Ouwehand AC, Conway PL, Salminen SJ. Inhibition of S-fimbria mediated adhesion to human ileostomy glycoproteins by a fraction isolated form bovine colostrum. Infection and Immunity. 1995;63:4917–4920. doi: 10.1128/iai.63.12.4917-4920.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Peake P, Gooley A, Britton MJ. Mechanism of interaction of the 85B secreted protein of Mycobacterium bovis with fibronectin. Infection and Immunity. 1993;61:4828–4838. doi: 10.1128/iai.61.11.4828-4834.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Probert WS, Johnson BJB. Identification of a 47 kDa fibronectin-binding protein expressed by Borrelia burgdorferi isolate B31. Molecular Microbiology. 1998;30:1003–1015. doi: 10.1046/j.1365-2958.1998.01127.x. [DOI] [PubMed] [Google Scholar]
  25. Quaroni A, Isselbacher KJ, Ruoslahti E. Fibronectin synthesis by epithelial crypt cells of rat small intestine. Proceedings of the National Academy of Sciences of the USA. 1978;75:5548–5552. doi: 10.1073/pnas.75.11.5548. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Rao M, Manishen WJ, Maheshwari Y, Sykes DE, Siyanova EY, Tyner AL, Weiser MM. Laminin receptor expression in rat intestine and liver during development and differentiation. Gastroenterology. 1994;107:764–772. doi: 10.1016/0016-5085(94)90125-2. [DOI] [PubMed] [Google Scholar]
  27. Scarpa S, D'Orazi G, Modesti A, Valia S, Zucco F, Stammati A. Expression and synthesis of fibronectin and laminin by an intestinal epithelial cell line. Tissue and Cell. 1988;20:305–312. doi: 10.1016/0040-8166(88)90066-3. [DOI] [PubMed] [Google Scholar]
  28. Simon-Assmann P, Kedinger M, Haffen K. Immunocytochemical localization of extracellular-matrix proteins in relation to rat intestinal morphogenesis. Differentiation. 1986;32:59–66. doi: 10.1111/j.1432-0436.1986.tb00556.x. [DOI] [PubMed] [Google Scholar]
  29. Simon-Assmann P, Duclos B, Orian-Rousseau V, Arnold C, Mathelin C, Engvall E, Kedinger M. Differential expression of laminin isoforms and alpha 6-beta 4 integrin subunits in the developing mouse and human intestine. Developmental Dynamics. 1994;201:71–85. doi: 10.1002/aja.1002010108. [DOI] [PubMed] [Google Scholar]
  30. Vachon PH, Simoneau A, Herring-Gillam FE, Beaulieu J-F. Cellular fibronectin expression is down-regulated at the mRNA level in differentiating human intestinal epithelial cells. Experimental Cell Research. 1995;216:30–34. doi: 10.1006/excr.1995.1004. [DOI] [PubMed] [Google Scholar]
  31. Yamagishi F, Komoda T, Alpers DH. Secretion and distribution of rat intestinal surfactant-like particles following fat feeding. American Journal of Physiology. 1994;266:G944–952. doi: 10.1152/ajpgi.1994.266.5.G944. [DOI] [PubMed] [Google Scholar]
  32. Zhang Y, Shao J -S, Xie Q, Alpers DH. Immunolocalization of alkaline phosphatase and surfactant-like particle proteins in rat duodenum during fat absorption. Gastroenterology. 1996;110:478–488. doi: 10.1053/gast.1996.v110.pm8566595. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES