Listeriolysin O, the major virulent determinant of Listeria monocytogenes, is known for forming pores on cholesterol-rich membranes. In the present study, we reveal its other facet, rafts clustering. By immunofluorescence microscopy, we show that the glycosylphosphatidylinositol-anchored proteins CD14 and CD24, which normally exhibit uniform distribution on J774 cells, undergo clustering upon treatment with LLO. The non-raft marker transferrin receptor is unaffected by such treatment. Rafts clustering might explain the induction of tyrosine phosphorylation observed on LLO-treated cells.
Fifteen years after the formulation of the rafts hypothesis [1–3], considerable debate still persists and the exact structure of rafts remains unclear. Evidence for the existence of lipid rafts was initially based on resistance of the protein and lipid components to solubilization by non-ionic detergents. However, because of potential artifacts inherent in detergent extraction, the precise size and composition of the native rafts is still unclear.
Recent characterizations of rafts relying on fluorescence resonance energy transfer techniques have also led to conflicting outcomes. For instance, some studies indicated clustering of GPI (glycosylphosphatidylinositol)-anchored proteins in resting cells [4,5], whereas other studies have failed to detect such clustering [6,7]. The consensus view from these investigations is that native rafts are rather small and/or dynamic, thus explaining why they have eluded detection by conventional light microscopy. So far, microscopical visualization of rafts has only been accomplished after cross-linking raft components with lectins and/or antibodies. This allowed raft and non-raft components to segregate into detectable micron-sized patches . Even here there appears to be a remarkable lack of consensus on the composition of such clusters since some studies have indicated that raft subtypes exist and that different GPI-anchored proteins tend to segregate into distinct lipid rafts, thus undercutting co-clustering by antibodies . This obstacle notwithstanding, rafts clustering still stands out as one of the most practical prospects for visualizing rafts in living cells. The development of new tools that allow clustering of rafts irrespective of their subtypes should be an important step towards fully exploiting such prospects. In the present study, we demonstrate that LLO (listeriolysin O) is a potent aggregator of lipid rafts which, in contrast with antibodies, can induce co-clustering of different rafts-associated proteins.
LLO is a pore-forming toxin produced by Listeria monocytogenes that opens up the phagosomal membrane, thus enabling the bacteria to escape into the cytosol where it replicates. LLO accomplishes this by binding to cholesterol, then oligomerizing into large complex pores .
LLO is known to trigger several signalling pathways in a variety of host cell types. Although some of these signals are due to the influx of Ca2+ through pores , other mechanisms of signal induction do exist. If preincubated with cholesterol, LLO loses its cytolytic activity but nonetheless binds to host cells and triggers signalling [10,12]. To understand how LLO interacts with cells to induce signalling, we investigated the potential involvement of lipid rafts. A HA (haemagglutinin)-tagged LLO (HA–LLO) or its cholesterol-inactivated form (CL–HA–LLO) was used to this end.
LLO induces clustering of CD14 and CD24 but not the TFR (transferrin receptor)
To characterize the LLO-rafts association, J774 cells treated with HA–LLO or CL–HA–LLO were subjected to detergent extraction. Both forms of the toxin were found to co-partition with several rafts-associated proteins into detergent resistant membranes (results not shown). Next, we analysed the membrane distribution of LLO on cells in relation to raft marker proteins CD14 and CD24 as well as TRF as a nonrafts marker. Under basal conditions, both raft markers exhibited even membrane distribution (Figures 1A and 1E) as shown before. The TFR, however, displayed a clustered distribution pattern (Figure 1I) even at basal condition, probably due to its concentration in coated pits.
Clustering of GPI-anchored proteins by CL–HA–LLO and induction of tyrosine phosphorylation
Subsequently, cells were incubated with either HA–LLO or CL–HA–LLO, fixed, before staining for the bound toxin as well as CD14, CD24 or TFR. Figure 1 displays cells treated with CL–HA–LLO. The toxin was distributed into distinct clusters on cell membranes. Interestingly, CD14 and CD24 exhibited clustering on such cells. Such clusters overlapped extensively with those of the toxin (Figures 1B–1D and 1F–1H). In contrast, the LLO clusters were distinctly segregated from those of TFR. Similar results were obtained using active HA–LLO (results not shown). Taken together, these results demonstrate that LLO in its active or cholesterol-inactivated forms specifically binds to and aggregates lipid rafts, thus co-clustering different rafts-associated proteins.
Induction of signalling
We then evaluated whether rafts aggregation by LLO translates into signalling. To that end, lysates of J774 cells stimulated with HA–LLO and CL–HA–LLO were analysed for tyrosine phosphorylation by immunoblotting. As shown in Figure 1(M), the stimulated cells showed a significant increase in tyrosine phosphorylation. The model in Figure 2 illustrates how LLO aggregates rafts together with rafts-associated kinases and adaptor molecules thus inducing signals.
Model for rafts aggregation and signal induction by LLO
Significance of rafts aggregation on the interaction of L. monocytogenes with its host
The significance of LLO is underscored by the fact that mutant strains lacking LLO are avirulent in mice and incapable of generating protective immunity [12,13]. Signals triggered by LLO strongly influence the course of infection, e.g. facilitating uptake by epithelial cells  or inducing pro-inflammatory cytokines and chemokines in macrophages, which facilitate spreading by recruiting more potential host cells (J. Jablonska and S. Weiss, unpublished work). Since rafts represent a ubiquitous target for LLO on all host cells, it is therefore capable of acting as a pleiotropic pseudocytokine/chemokine that triggers various host responses.
Potential applications of LLO in studying rafts
The lack of efficient tools for studying rafts is largely to blame for the polarized opinions on the existence and role of rafts. Because of their limited size, rafts, unless physically clustered, have so far eluded detection by conventional microscopy and fluorescence resonance energy transfer [7,14]. Currently, the commonly used agents for clustering are cholera toxin (CT-B) and/or antibodies. These reagents, however, have their limitations. First, several results indicating the presence of raft subtypes with distinct protein and GM1 compositions [9,15] imply that CT-B or antibodies might aggregate only a subset of rafts. This is, however, not the case for LLO that clusters rafts by oligomerizing cholesterol, the dynamic ‘glue’ of all rafts.
Secondly, the restricted number of raft molecules that can be engaged by antibodies or CT-B limits the size of clusters. For instance, although CT-B engages five GM1 molecules, cross-linking by an antibody is nonetheless required for measurable raft clustering to be induced . In contrast, due to the extensive oligomerization of LLO monomers (up to 40–80), the raft clusters formed thereof are large and can be visualized easily (Figure 1).
In conclusion, LLO is a tool with great potential that not only can be used to visualize rafts but also to identify putative raft components as well as signalling pathways mediated through rafts. The use of CL–LLO allows such studies without damaging cellular membranes.
Structure Related to Function: Molecules and Cells: A Focus Topic at BioScience2004, held at SECC Glasgow, U.K., 18–22 July 2004. Edited by D. Alessi (Dundee, U.K.), T. Cass (Imperial College London, U.K.), T. Corfield (Bristol, U.K.), M. Cousin (Edinburgh, U.K.), A. Entwistle (Ludwig Institute for Cancer Research, London, U.K.), I. Fearnley (Cambridge, U.K.), P. Haris (De Montfort, Leicester, U.K.), J. Mayer (Nottingham, U.K.) and M. Tuite (Canterbury, U.K.).