K. SIMONS
European Molecular Biology Laboratory
Postfach 10.2209, Meyer hofstr. 1, D-6900 Heidelberg,
FR Germany
References
One of the challenges of contemporary biology is to unravel how the molecular composition of the different cellular compartments is generated and maintained during the cell cycle. In animal cells most of the efforts have been directed towards the study of how newly synthesized proteins are transported to their correct cellular destinations, whereas the lipids, which make up the framework of the membranes in the cell, have been given much less attention.
This paper will focus on a working hypothesis for the generation and maintenance of the different protein and lipid compositions of the two cell surface domains in the polarized cells of simple epithelia. The epithelia lining the body cavities are composed of a single layer of polarized cells. The apical plasma membrane domains of the cells line the lumen of the cavity and the basolateral cell surface faces the underlying extracellular matrix and the blood supply.)1,2 Each cell in the layer is linked to its neighbours by intercellular junctions, including the tight junctions that form the permeability barrier between the cells.3 The tight junctions also define the boundary between the apical and the basolateral membrane domains. The sorting of newly synthesized surface glycoproteins defined for the two plasma membrane domains has been localized to the trans-Golgi network, the exit compartment of the Golgi complex.2,4,5 The sorting of newly synthesized lipids enroute for the epithelial cell surface also takes place intracellularly in the Golgi complex,6 raising the possibility that protein and lipid sorting are directly connected to each other. Our working hypothesis is that the transport machinery in the trans-Golgi sorts lipids and proteins into common carrier vesicles for delivery to the correct cell surface domain.
A number of studies indicate that the apical plasma membrane domain is enriched in glycosphingolipids whereas the basolateral domain has a correspondingly higher phosphatidylcholine content. The distribution of other phospholipids (such as phosphatidylserine and phosphatidylethanolamine), as well as of cholesterol, is similar in the two domains. The asymmetric distribution of lipids in the two surface domains appears to be restricted to the exoplasmic leaflet,2 where their intermixing is prevented by the tight junctions.7,8 The lipids in the cytoplasmic leaflet seem to diffuse freely between the two domains. Therefore, it follows that glycosphingolipids are exposed on the external surface of the apical membrane, whereas phosphatidylcholine would be exposed on the surface facing the basolateral milieu. The lipids common to both domains such as phosphatidylserine and phosphatidylethanolamine are primarily in the cytoplasmic leaflet. This predicted topology is consistent with the available data on lipid asymmetry in plasma membrane, choline-containing lipids and glycosphingolipids being generally exoplasmic and amino-containing phospholipids being cytoplasmic.9-13
Apical membranes may have such an unusually high content of glycolipids due to the stabilizing and protective function of these lipids.14,15,16 The apical membranes face the hazards of the external environment. In some epithelial cell types such as those lining the gall bladder and the bile duct, the apical surfaces even have to withstand solubilizing concentrations of the bile salt detergents. Glycosphingolipids are uniquely suited for a protective function because they can form intermolecular hydrogen bonds between the glycosyl head groups, the amide and hydroxyls of the sphingosine base and of the hydroxy fatty acid.14 To a lesser extent sphingomyelin also has this capacity to associate by intermolecular hydrogen bonds due to its ceramide constituent. This extensive intermolecular hydrogen bonding capacity is a characteristic feature that distinguishes sphingolipids from the major lipid family in animal cells, the glycerolipids. These cannot form interlipid hydrogen bonds between their diglyceride moieties. The ester and ether groups can function only as hydrogen bond acceptors, not as donors.
The sorting event in the trans-Golgi network need not be specific in both the apical and the basolateral directions. It is possible that only one direction is mediated by specific recognition of molecules to be transported (the signal-mediated pathway) and that the other route includes molecules in transit without specific signal recognition (the default pathway).17 For apical and basolateral membrane proteins, no conclusive answer is available yet, although several attempts have been made to localize the protein signals mediating sorting. One clue might be the tight exclusion of basolateral proteins from the apical side in MDCK cells,18 whereas the converse is not true.19,20 A small fraction of apical proteins are "missorted" to the basolateral side, possibly because the basolateral route operates by default. Biosynthetic protein transport from the Golgi to the fibroblast cell surface has been postulated to be a default pathway.21 The basolateral route could be the fibroblast homologue. Sorting in the apical direction would be signal mediated, and be specific for simple epithelia. In more complicated epithelial tissues such as liver, the situation is different. Each hepatocyte has several apical poles lining the bile canaliculi. Bartles et al.22 have shown that apical proteins are not sorted in the trans-Golgi network. There seems to be no apical route from the Golgi complex to the apical membrane in hepatocytes. Instead, apical proteins are delivered to the basolateral membrane, from where they are sorted to the apical side. The simplest interpretation of these findings is that the basolateral transport vesicles (the postulated default pathway) carry both apical and basolateral proteins. The apical sorting machinery seems to be lacking from the trans-Golgi network in hepatocytes.
In MDCK cells, specificity in the apical direction might be aided by glyco-sphingolipid-protein interactions. Sphingolipid clustering in the luminal (exoplasmic) leaflet of the trans-Golgi network is postulated to form the budding site for an apical membrane vesicle.6,23 The self-association could be mediated by interlipid hydrogen bonding. This asymmetric sphingolipid microdomain is assumed to be the starting point for inclusion of associating apical proteins that bind directly to the glycosphingolipids or indirectly via interactions to a trans membrane sorting protein. Such a sorting protein should bind to both glycosphingolipids and to the apical proteins. Moreover, this protein is assumed to have another function. Its cytosolic domain interacts with a cytosolic protein coat to induce the curvature leading to vesiculation. According to this model, exclusion of glycerolipids in the exoplasmic leaflet could result from their inability to form interlipid hydrogen bonds with the sphingolipids. The basolateral transport vesicles are predicted to form from membrane regions depleted of apical components and these should, therefore, be enriched in phosphatidylcholine as a consequence of lipid asymmetry. The asymmetry of the lipids facilitates the lateral separation of the apical and the basolateral precursor domains in the trans-Golgi network. Sorting would, according to this view, be the formation of microdomains mimicking the properties of the membranes of their destination. This hypothesis makes several predictions that can be tested. First, there should be specific interactions between glycosphingolipids and apical proteins or between glycosphingolipids and the putative bridging protein. Second, for exclusion of phosphatidylcholine to occur, the luminal leaflet of the membrane segment that forms the apical transport vesicle has to be almost covered by sphingolipids. The sphingo- to phospholipid ratio of the apical transport vesicles would depend on the size of the vesicle; the larger the vesicle, the nearer the ratio between the surface areas of the luminal and the cytoplasmic leaflets will be to one. Furthermore, if sphingolipid clustering in the trans-Golgi network were a prerequisite for apical delivery, then presumably sphingolipid recycling between the apical membrane and the trans-Golgi would be necessary to replenish the sphingolipids lost in 4 h vesiculation event. Also, the putative sorting protein has to cycle between the apical membrane and the TGN to be able to perform sorting. Sorting proceeds for hours in virus-infected MDCK cells which synthesize only virus proteins and not host proteins.
1. RODRIGUEZ-BOULAN, E. 1983. In Modern Cell Biology. B.H. Satir, Ed.: Volume 1: 119-170. Alan Liss, New York.
2. SIMONS, K. and S.D. FULLER. 1985. Ann. Rev. Cell Biol. 1: 243-288.
3. DIAMOND, J.M. 1977. Physiologist 20: 10-18
4. GRIFFITHS, G. and K. SIMONS. 1986. Science 234: 438-443.
5. HUGHSON, E., A. WANDINGER-NESS, H. GAUSEPOH, G. GRIFFITHS and K. SIMONS In press. M.
Schwartz, Ed.: Elsevier, Holland.
6. VAN MEER, G., E.H.K. STELZER, R.W. WIJNAENDTS-VAN-RESANDT and K. SIMONS. 1987. J. Cell Biol. 105: 1623-1635.
7. DRAGSTEN, P.R., R. BLUMENTHAL and J.S. HANDLER. 1981. Nature 294: 718-722.
8. VAN MEER, G. and K. SIMONS. 1988. EMBO J. 5: 1455-1464.
9. BRETSCHER, M.S. 1972. J. Mol. Biol. 71: 523-528.
10. BRETSCHER, M.S. and M.C. RAFF. 1975. Nature 258: 43-49.
11. GAHMBERG, C.G. and S. HAKOMORI. 1973. J. Biol. Chem. 248: 4311-4319.
12. OP DEN KAMP, J.A.F. 1979. Ann. Rev. Biochem. 48: 47-71.
13. THOMPSON, T.E. and T.W. TILLACK. 1985. Ann. Rev. Biophysics and Biophysical Chem. 14: 361-386.
14. PASCHER, I. 1976. Biophys. Acta 455: 433-451.
15. CURATOLO, W. 1987a. Biochim. Biophys. Acta 906: 111-136.
16. CURATOLO, W. 1987b. Biochim. Biophys. Acta 906: 137-160.
17. PFEFFER, S.R. and J.E. ROTHMAN. 1987. Annual Reviews of Biochemistry 56: 829-852.
18. FULLER, S.D. and K. SIMONS. 1986. J. Cell Biol. 103: 1767-1779.
19. BALCAROVA-STANDER, J., S.E. PFEIFFER, S.D. FULLER and K. SIMPONS. 1984. EMBO J. 3: 2687-2694.
20. PFEIFFER, S., S.D. FULLER and K. SIMONS. 1985. J. Cell Biol. 101: 470-476.
21. WIELAND, F.T., M. L. GLEASON, T. A. SERAFINI and J.E. ROTHMAN. 1987. Cell 50: 289-300.
22. BARTLES, J.R., H. M. FERACCI, B. STIEGER and Q.L. HUBBARD. 1987. J. Cell Biol. 105: 1241-1251.
23. SIMONS, K. and G. VAN MEER. In press. Biochemistry.
