Unit: Membrane Trafficking and Pathogenesis
Director: Chiara ZURZOLO
By using an imaging and FRET approach in living cells combined with a classical biochemical approach our laboratory has two major objectives :
- to understand the molecular requirements for apical sorting of GPI-anchored proteins in polarized epithelial cells.
- to characterize the exocytic and endocytic trafficking of the normal and abnormal prion proteins in epithelial and neuronal cells in order to identify the intracellular site of conversion.
One of the most important properties of epithelia is cell polarity, which results in the formation of distinct apical and basolateral domains of the plasma membrane. These distinct domains enable the cells to perform their vectorial activities. The mechanisms responsible for the generation and maintenance of epithelial polarity have been only partially characterized. Elucidation of the molecular basis of cell polarity is therefore a fundamental goal in mammalian cell biology. We have focused our attention on studying the mechanism of apical sorting of GPI-anchored proteins and the role of lipid microdomains (or rafts) and proteins and lipid segregation at the level of the TGN in two model epithelial cell lines, MDCK and FRT cells, which have different sorting phenotypes. In these cells we are also studying the exocytic and endocytic transport of the prion protein, PrPc, and its pathological mutants, in order to understand the role of intracellular trafficking and rafts in the function and the pathological conversion of this cellular protein to the infectious scrapie.
Two major projects are ongoing in my unit:
1) Mechanism of GPI-anchored protein sorting to the plasma membrane
2) Intracellular trafficking of PrPc and its mutants: correlation with TSE pathogenesis
1. Mechanism of GPI-anchored protein sorting to the plasma membrane
Rafts are membrane microdomains enriched in glycosphingolipids (GSLs) and cholesterol which have been proposed to function in the last Golgi cisterna, the TGN, as a sorting platform for the apical delivery of plasma membrane proteins. This is particularly evident in the case of GPI-anchored proteins that are sorted to the apical membrane in several epithelial cell lines and associate with rafts (or DRMs, Detergent-Resistant Microdomains) during their transport to the plasma membrane. Our knowledge of how this occurs is only rudimentary: e.g., what determines the association of these proteins with rafts, whether there are different kinds of rafts for different GPI-anchored proteins, or whether a TGN receptor that sorts apical and basolateral proteins into different post-Golgi vesicles is involved, are not yet known.
In this project we are using both microscopical and biochemical approaches to analyze the role of raft domains in sorting and trafficking of GPI-anchored proteins and to characterize the molecular components of the machinery.
For this purpose we had already obtained stable clones of FRT and MDCK cells expressing different chimeric proteins fused to GFP and set up an assay to study transport in basal conditions using a microscopy-based imaging system and deconvolution software. We had found that GFP-GPI is basolateral in FRT cells and apical in MDCK cells, whereas a GFP-TM construct containing the transmembrane and cytosolic domains of the LDL (low density lipoprotein) receptor is basolateral in both cells. We have further utilized this approach to characterize the kinetics of the basolateral and apical trafficking of these fluorescent markers and the characteristics of the post TGN carriers. We have followed the kinetics of emptying of the Golgi apparatus and plasma membrane arrival after block at 20°C and we have been able to determine the size and velocity of post-TGN intermediates arriving at the apical and basolateral plasma membranes in FRT and MDCK cells. This work is still in progress and has been done both using the CID at Pasteur and in collaboration with Dr Fogarty at the University of Massachussets (USA).
We and others have demonstrated that raft-association depends on the GPI anchor. However this event is not sufficient to determine apical sorting, which we believe occurs following "concentration and stabilization" of the protein in these microdomains probably through interactions via protein ectodomains. By using a biochemichal approach we found that only apical and not basolateral GPI-anchored proteins form high molecular weight oligomers. We believe that stabilization in rafts could occur because of protein oligomerization. We have now demonstated that imparing oligomerization of apical GPI-anchored proteins also affects sorting and therefore this step appears to be crucial for apical sorting (Paladino et al., JCB 2004). We also found that oligomerization occurs concomintantly to raft-association during the passage though the golgi apparatus. These data are in contrast with a recent paper which shows that GPI-proteins are sorted to the apical membrane via an indirect pathway, after their arrival to the basolateral membrane. In the last year we have demonstrated both by a biochemical and an imaging approach in living polarized cells that GPI-anchored proteins are sorted via a direct pathway to the apical membrane, thus finally solving a long debated issue in the field (this work is now close to submission (Paladino, Pocard et al in press JCB).
In order to analyse the role of the GPI anchor and of the protein ectodomain in the sorting event we have transfected different GPI-anchored and transmembrane (TM) proteins in MDCK and FRT cells. By chemical cross-linking and co-immunoprecipitation protocols we have been able to demonstrate that two different apical GPI-proteins are in the same complex, from which TM proteins are excluded. This fact seems to be independent from the signal for GPI anchor attachment (Tivodar et al., in preparation).
A parallel approach to analyse proximity of GPI and TM proteins at the cell surface and in the TGN (during their sorting) is by analysing Fluorescence Energy Transfer (FRET). To this aim we will use both classical FRET (using the CFP/YFP pair) and anisotropy studies (using GFP). In the last year we have prepared CFP/YFP and GFP constructs fused to different GPI anchors (from PLAP, DAF, FR, and PrP). We have transfected them in pairs and obtained stable clones. We are now in the process of setting up the FRET experiments and preliminary results are quite encouraging (Pocard et al in preparation).
Another approach that we have undertaken to understand the role of rafts in the sorting of GPI-anchored proteins has been the analysis of the lipid composition of the rafts associated with apical and basolateral GPI-anchored proteins. Indeed one possibility to explain the diffent sorting could be that different GPI proteins associate with different lipid environments. By labelling sphingolipids with H3 sphingosine and following the different proteins with specific antibodies we have purified on sucrose density gradients the DRMs (detergent resistant membrane domains) associated with them and analysed the lipid content. We found that there are no qualitative differences in the lipid composition (sphingolipids, phospolipids and cholesterol) of DRMs associated with apical or basolateral GPI-anchored proteins, suggesting that the different sorting is not dependent on different apical and basolateral rafts. We believe that the affinity of these different proteins for the raft domains and possibly differences in the protein ectodomains might have an important role. This work has been submitted (Tivodar, et al., submitted).
On the same line we began to analyze the role of the N- and O-glycans of the protein ectodomains in their apical soprting. We are using two GPI-anchored model proteins and we are currently performing different point mutations and deletions to remove the sugars. These constructs will be stably expressed in MDCK cells and analysed for their sorting, trafficking, DRM-association and oligomerization (Catino et al in preparation).
We have also started the analysis of the molecular machinery involved in the apical sorting of GPI-anchored proteins, and began to look at the role of the SNAREs. We have now established RNAi in transient and stable transfected cells and we are looking at the effect of interference of the apical SNAREs (syntaxin 3 and TiVamp) on GPI protein sorting (Pocard, in preparation).
2. Intracellular trafficking of PrPc and its mutants: correlation with TSE pathogenesis
Transmissible spongiform encephalopathies (TSE) are fatal neurodegenerative disorders of humans and animals that are unique because they can be either infectious, genetic or sporadic in origin. All three forms result from a post-translational alteration in the conformation of a host-encoded membrane glycoprotein called PrPC, which denotes the cellular isoform of the prion protein. Conversion of this normal isoform from an a-helical, readily proteolysed surface protein, to the scrapie isoform, PrPSc, a highly protease resistant, ß-pleated sheet-containing hydrophobic aggregate is not only diagnostic of TSE infection, but is widely taken to be the defining infectious event and source of the subsequent neuropathology.
It is our contention that the conversion of prion protein from PrPC to the pathogenic PrPSc isoform is strongly influenced by the intracellular trafficking of the protein in the cells. Epithelial cells are very similar to neuronal cells, but are better characterized for their intracellular trafficking and endocytosis, so they represent an ideal model to study the intracellular trafficking of PrPc and its mutants. Furthermore, in neuroblastoma cells PrPC associates with TritonX-100 (TX100)-insoluble membrane microdomains and the depletion of cholesterol reduces PrPC degradation and inhibits PrPSc generation. By extraction in cold TX100 we found that PrPC is mainly insoluble, indicating that association with detergent-resistant microdomains (DRMs) is not sufficient for apical sorting of GPI-proteins. We have further characterized the DRM- asssociation of PrPC during pulse chase experiments and found that one of the immature PrPC isoforms already associates with DRM in the endoplasmic reticulum. Perturbation of this association by depleting the cells of cholesterol leads to protein misfolding, therefore suggesting that DRM association is necessary for the correct folding of PrP (Sarnataro, Campana et al. MBC 2004). Because TSE is a result of PrP misfolding this issue appears to be very relevant for understanding the mechanisms of PrP misfolding (Campana et al; TCB 2005).
We have recently tested DRM-association and its effect on folding of different PrP mutants. We found that the glycosylation mutant A182T seems to be strongly associated with DRMs in the endoplasmic reticulum. By using different drugs affecting cholesterol and sphingolipid intracellular content it appears that the mutant and the wilde-type (wt) proteins reside in different membrane domains. Thus we are analyzing the lipid composition of DRMs associated with the normal and mutated forms of PrP in the DRMs. Preliminary results indicate that there are no qualitative differences between the two PrP forms. Another approach that we are setting up to define the membrane environment of PrPC and PrP mutants is a Fluorescence Resonance after Photobleaching (FRAP) assay. To this aim we have made GFP fusion proteins with PrPwt and mutants and are analysing the characteristics of these fluorescent mutants after bleaching (Campana, in preparation).
We have also analyzed the exocytic pathway of the single and double glycosylation mutants and of a mutant lacking the GPI anchor. Interestingly the glycosylation mutant A182T, which is pathological (i.e., found in familial forms of CJD), accumulates in the
endoplasmic reticulum where it is DRM-associated. We have also found that this mutant is misfolded and is degraded only in part by the proteasome. However the proteasome is not involved ion the degradation of the misfolded (scrapie-like) form of the mutant which instead accumulates when the association with rafts is perturbed. We therefore propose a role for rafts in protection from misfolding in the case of the wt protein as well as of pathological mutants ( Campana et al. 2005; Campana et al. submitted)
In the last year we have charachterized the intracellular localization of PrPSc in infected neuronal cell lines (N2a and GT1) and have started to establish primary neuronal cell lines (cortical and DRG derived neurons) in order to study the trafficking and localization of PrPC in a neuronal cell system (Casanova and Schiff).
We are also trying to infect FRT and MDCK cells in which we have characherized PrPC trafficking. We are currently setting up the best condition for infection and the best cell line to infect to attempt to identify the site of conversion. In order to continue these studies and to try to identify the site of conversion in living cells a fluorescent microscope with a CCD camera to use live imaging in the P3 laboratory will be set up in 2006. In the meantime we have prepared and stably co-transfected PrP-YFP- and -CFP-linked constructs in FRT cells in order to study whether there will be an effect of the mutant form on the normal trafficking or conformation.
We have also made some progress in our understanding of the internalization route followed by PrP. It has been shown that PrPC can be internalized either via clathrin-coated pits or via caveolae. Following endocytosis, PrPC is recycled back to the cell surface after processing. Study of the endocytic recycling pathway is of interest because it may be the route along which the critical conversion of PrPC to PrPSc takes place, and because one physiological function of PrPC might be to facilitate uptake of an as yet unidentified extracellular ligand, by analogy with receptors responsible for uptake of transferrin and low-density lipoprotein.
To test the involvement of caveolae and coated pits in the endocytosis of PrPC isoforms we have used FRT cells which do not express caveolin and do not have caveolae, as well as FRT cells transfected with cav1, which do form caveolae. We found that PrPC is internalized in both cell lines, mainly through long uncoated invaginations. We also found that PrPC internalization is slowed down by caveolin 1 suggesting a role for this molecule as a negative regulator of internalization. By using cholesterol depletion we found that PrPC internalization is dependent on rafts. However PrPC uptake is not completely blocked by cholesterol depletion suggesting that another mechanism (raft-independent) is involved. Thus, we have analysed the involvement of different factors (Dynamin II, EPS15 and CDC42 ) involved in clatrin-dependent and independent endocytosis. By the use of GFP-linked dominant negative mutants and by a biochemical approach we have shown that two pathways, (raft-dependent and clathrin-dependent) are concomitantly involved in PrPC internalization. This work is now close to submission (Sarnataro, Casanova et al.).
Keywords: intracellular trafficking, rafts, protein sorting, prion, conversion site