Separation of a cholesterol-enriched microdomain involved in T-cell signal transduction Yukiko Shimada 1 , Mitsushi Inomata 1 , Hidenori Suzuki 2 , Masami Hayashi 1 , A. Abdul Waheed 1 and Yoshiko Ohno-Iwashita 1 1 Biomembrane Research Group, Tokyo Metropolitan Institute of Gerontology, Itabashi-ku, Tokyo, Japan 2 Center for Electron Microscopy, Tokyo Metropolitan Institute of Medical Science, Bunkyo-ku, Tokyo, Japan Cholesterol is one of the major constituents of the plasma membrane, and is involved in the formation of the membrane bilayer. The distribution of choles- terol in the plasma membrane is not uniform, sug- gesting that cholesterol is also involved in the construction of functional membrane domains. One such functional membrane domain is called lipid rafts [1,2]. Lipid rafts are lateral lipid clusters formed of sphingolipids and cholesterol, in which particular molecules are concentrated to form platforms for intracellular transport and signal transduction. Cho- lesterol depletion reduces the association of these molecules with lipid rafts [3,4], indicating that choles- terol is necessary for the partitioning of these partic- ular molecules into functional domains in the plasma membrane. Several reports have suggested that lipid rafts are platforms for signal transduction in T-cells [5]. Lipid rafts obtained from resting T-cells are enriched in Src- family kinases, Lck and Fyn [6,7], and the linker for the activation of T-cells (LAT) [8]. Minor amounts of CD3f are associated with rafts, but the data concern- ing other T-cell receptor (TCR)⁄ CD3 constituents remains contradictory [6,7]. In addition, the partition- ing or recruitment of CD3e to lipid rafts after TCR stimulation with antibodies remains uncertain [9,10]. These inconsistent results might be due mainly to the different methods for isolating lipid rafts. Currently, Keywords raft; cholesterol; T-cell signalling; perfringolysin O Correspondence Y. Shimada, Biomembrane Research Group, Tokyo Metropolitan Institute of Gerontology, 35-2 Sakae-cho, Itabashi-ku, Tokyo 173- 0015, Japan Fax: +81 3 3579 4776 Tel: +81 3 3964 3241 extn 3063, 3068 E-mail: yshimada@tmig.or.jp (Received 29 June 2005, revised 15 August 2005, accepted 24 August 2005) doi:10.1111/j.1742-4658.2005.04938.x We isolated a cholesterol-enriched membrane subpopulation from the so-called lipid raft fractions of Jurkat T-cells by taking advantage of its selective binding to a cholesterol-binding probe, BCh. The BCh-bound mem- brane subpopulation has a much higher cholesterol ⁄ phospholipid (C ⁄ P) molar ratio ( 1.0) than the BCh-unbound population in raft fractions ( 0.3). It contains not only the raft markers GM1 and flotillin, but also some T-cell receptor (TCR) signalling molecules, including Lck, Fyn and LAT. In addition, Csk and PAG, inhibitory molecules of the TCR signalling cascade, are also contained in the BCh-bound membranes. On the other hand, CD3e, CD3f and Zap70 are localized in the BCh-unbound mem- branes, segregated from other TCR signalling molecules under nonstimulat- ed conditions. However, upon stimulation of TCR, portions of CD3e, CD3f and Zap70 are recruited to the BCh-bound membranes. The Triton X-100 concentration used for lipid raft preparation affects neither the C ⁄ P ratio nor protein composition of the BCh-bound membranes. These results show that our method is useful for isolating a particular cholesterol-rich membrane domain of T-cells, which could be a core domain controlling the TCR signalling cascade. Abbreviations C ⁄ P, cholesterol ⁄ phospholipid molar ratio; DRM, detergent-resistant membrane; LAT, linker for activation of T-cells; PAG, phosphoprotein associated with glycosphingolipid-enriched membrane microdomains; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; SM, sphingomyelin; TCR, T-cell antigen receptor. 5454 FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS detergent-resistant membranes (DRMs) have been assumed to represent lipid rafts in their biochemical aspects [2]. However, various biochemical methods and conditions are used for isolating DRMs, which gives rise to some conflicts concerning the molecules associ- ated with DRMs. The heterogeneity of lipid rafts has recently been discussed. It has been suggested that several types of lipid rafts with differing lipid and protein compositions perform different functions [11,12]. Fluorescent micro- scopic observation of GM1- and GM3-enriched rafts has shown that raft-associated proteins are also distri- buted asymmetrically in polarized cells [11]. Immuno- electron microscopy of peripheral blood T-cells shows distinct clustering and segregation of Lck and LAT on the inner leaflet of the plasma membrane [12]. These morphological analyses suggest the existence of raft subsets. There are some biochemical approaches to define the heterogeneity of raft-associated molecules by immunoisolation, providing some information based on protein–protein interactions in rafts [13–15]. How- ever, the heterogeneity of lipid rafts is less well under- stood in biochemical terms because the nature of lipid rafts remains unclear. For the biochemical characteri- zation of lipid rafts, a more sophisticated isolation method, one based on criteria other than detergent insolubility, is required. To understand lipid-based raft domains, we have focused on cholesterol as a major component of lipid rafts. Previously, we designed the novel cholesterol probes Ch and BCh [16,17] by modifying h-toxin (per- fringolysin O), a cholesterol-binding, pore-forming cytolysin produced by Clostridium perfringens [18]. Ch and BCh are noncytolytic derivatives of h -toxin that bind specifically and with high affinity to cholesterol in membranes [17–20]. Ch is produced by the limited pro- teolysis of h-toxin [16], and BCh by the biotinylation of Ch [17]. Their binding to artificial membranes is highly dependent on the cholesterol content of the membranes: they bind to liposomes with high choles- terol content but scarcely bind to liposomes containing less than 20 mol% cholesterol [18,21]. In intact cells, the depletion of cell cholesterol by approximately 30% abolishes their binding to plasma membranes [17–19]. This is in remarkable contrast to cell binding by filipin, another cholesterol-binding reagent. Filipin staining is significantly retained under the same depletion condi- tions [19]. Thus BCh binds to a specific population of cholesterol, while filipin binds indiscriminately to cell cholesterol. We have demonstrated that cell-bound BCh is predominantly recovered in raft fractions [18,19,22]. Electron microscopic observations showed that raft fractions prepared from BCh-bound platelets contain two populations of membrane vesicles, BCh-labelled and -unlabelled [19]. This observation implies that DRMs contain membrane subpopulations with different cholesterol enrichments, and that BCh could be a new probe to be used to isolate a particular lipid domain from raft fractions. In this study, we used the cholesterol probe BCh to isolate a cholesterol-enriched membrane domain from the so-called lipid raft fractions of T-cells. This partic- ular membrane domain can be prepared irrespective of the isolation conditions, and selectively retains signal- ling molecules such as Lck, Fyn, and LAT. The essential TCR-signalling molecules obtained in raft fractions, for example CD3f and Zap70, are not concentrated in the BCh-bound subpopulation under nonstimulated condi- tions; however, CD3f and Zap70 are recruited to the BCh-bound subpopulation after TCR stimulation. These results suggest that the so-called raft fractions consist of heterogeneous membrane groups, and that the cholesterol-enriched membrane domain isolated by BCh contains a core membrane domain for TCR signal transduction. Results BCh binds to a subpopulation of membranes in lipid raft fractions of Jurkat cells Jurkat cells incubated with BCh were treated with 1% (v ⁄ v) Triton X-100 solution for 15 min on ice and homogenized. The homogenate was ultracentrifuged in a sucrose density gradient and fractionated into 12 fractions from the top. Membrane-bound BCh was predominantly detected in fractions three to five (Fig. 1). These fractions correspond to one of the peaks of cholesterol, and contain typical raft marker proteins such as flotillin and Src-family kinases; thus they are identified as lipid raft fractions (raft frac- tions). BCh binding to membranes in the raft fractions was detected by immunoelectron microscopy (Fig. 2). BCh preferentially bound to some membrane vesicles, but not all membranes in the raft fractions, suggesting that raft fractions comprise at least two kinds of mem- brane groups as evaluated by BCh binding. Isolation of the BCh-bound membrane subpopu- lation from raft fractions and its evaluation We developed an isolation method for a raft subpopu- lation that binds BCh. The raft fractions were prepared from BCh-bound Jurkat cells and mixed with avidin- magnet beads on ice. BCh is a biotinylated probe. Ves- icles bound to BCh were retrieved with avidin-magnet Y. Shimada et al. Two raft subsets in T-cells FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS 5455 beads and separated from BCh-unbound vesicles that were recovered in the bead-unbound fraction. Total lipid rafts, and avidin-magnet beads-unbound and -bound fractions were subjected to SDS ⁄ PAGE and analysed by silver staining and western blotting (Fig. 3A,B). Almost all BCh was recovered in the avidin-magnet bead-bound fraction as determined by detection with antih-toxin antibody, indicating that most BCh-bound vesicles were recovered in the magnet bead-bound fraction (Fig. 3B). When raft fractions were prepared in the absence of BCh, no specific pro- teins so far tested were retained by the avidin-magnet beads by western blot analyses (Fig. 3B and data not shown). This indicates that membrane vesicles are not retained on the beads by nonspecific adsorption. These results show that our method is suitable for isolating membrane vesicles that selectively bind to BCh. BCh-bound vesicles are cholesterol-enriched and contain raft-marker proteins and some T-cell signalling molecules We analysed the cholesterol and phospholipid contents of bead-bound and -unbound fractions (Table 1, col- umns labelled 1% Triton). The total raft fractions con- tained about 30% of total cellular cholesterol (Fig. 1). Eighty per cent of the cholesterol in the raft fractions was retrieved in the BCh-bound membrane fraction (bead-bound fraction), which corresponds to 24% of total cellular cholesterol. The cholesterol ⁄ phospholipid (C ⁄ P) molar ratio of the BCh-bound membrane frac- tion is approximately 1.0, which is much higher than that ( 0.3) of the BCh-unbound membrane fraction (bead-unbound fraction) (Table 1). This clearly indi- cates that total raft fractions contain two distinctly dif- ferent subpopulations of membranes with respect to cholesterol enrichment. Approximately 40% of total raft protein was recov- ered in the BCh-bound membrane fraction (Table 1). Silver staining shows that the BCh-bound membrane fraction contains several distinctly different proteins from the unbound membrane fraction (Fig. 3A). To determine the protein profiles of raft subpopulations, these two fractions were analysed by western blotting (Fig. 3B,C). The raft marker protein flotillin was recovered almost exclusively in the bead-bound frac- tion, indicating that this molecule is predominantly localized in BCh-bound membranes (Fig. 3B). The majority of GM1 ganglioside, a raft marker lipid, is also localized in BCh-bound membranes as judged by the binding of cholera toxin (Fig. 3B, CTX). It has been reported that several proteins participa- ting in T-cell signalling are enriched in lipid rafts even under nonstimulated conditions [9,23]. We found that Src-family kinases (Lck and Fyn) and LAT recovered in total raft fractions were also associated with BCh-bound membranes (Fig. 3B,C). On the other hand, neither Zap70 nor CD3f were detected in the BCh- bound membrane fraction, but in the BCh-unbound membrane fraction (Fig. 3C). A small amount of CD3e was partitioned to the raft fractions, all of which was recovered in the BCh-unbound membrane fraction Fig. 1. BCh binds to lipid rafts in Jurkat cells. BCh-bound Jurkat cells (1 · 10 7 ) were treated with 1% Triton X-100, homogenized, and subjected to sucrose density gradient centrifugation. The resulting gradients were fractionated from the top (0.4 mL each; total 12 fractions). The distributions of cholesterol and cell-bound BCh in the gradient fractions were analysed. The BCh detected in fractions 10 and 11 probably represents a toxin liberated during membrane homogenization. Total, BCh in the total lysate before sucrose-density gradient fractionation. The results are representa- tive of seven independent experiments. Fig. 2. Immunoelectron microscopic observation of BCh in rafts. The raft fractions were prepared from BCh-bound Jurkat cells. BCh was immunolabelled with antibiotin and 10 nm protein-A gold and observed by negative staining. Arrows indicate BCh-bound vesicles. Two raft subsets in T-cells Y. Shimada et al. 5456 FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS (Fig. 4). The adaptor protein Grb2 was also detected in the unbound membranes. It is worthy to note that PAG and Csk, which negatively control TCR signalling, were found in BCh-bound membranes (Fig. 3C). Thus, mole- cules participating in T-cell signalling exhibit clear localizations between BCh-bound and -unbound mem- branes, suggesting that these two membrane domains play different roles in T-cell signalling. The results also show that raft fractions prepared by the conventional method contain at least two subpopulations of mem- branes that are distinct from each other in their molecu- lar components. Triton X-100 concentration does not affect the partitioning of signalling molecules into BCh-bound membranes It has been reported that molecular species and con- tents recovered in raft fractions depend on detergent concentrations. We examined the effect of Triton A B C Fig. 3. Molecular components in isolated BCh-bound and -unbound vesicles. Raft fractions were prepared from BCh-bound Jurkat cells as described. The BCh-bound membrane fraction was retrieved with avidin-conjugated magnetic beads, and then the total raft fraction, BCh- unbound membrane fraction and BCh-bound membrane fraction were subjected to SDS ⁄ PAGE. (A) Proteins were analysed by silver staining. M, Molecular mass markers (kDa). Open and filled triangles show bands that differ between the BCh-bound and -unbound membrane frac- tions, respectively. (B) Proteins were visualized by western blotting and probed with anti-(h-toxin) Ig or Igs against raft-associated molecules (+BCh). In parallel experiments, raft fractions were prepared in the absence of BCh and subjected to fractionation with magnet-beads (–BCh). (C) Blots were probed with antibodies against T-cell signalling-related molecules. The results are representative of seven independent experiments. Table 1. Comparison of cholesterol and phospholipid contents of raft fractions prepared with 1% or 0.2% (v ⁄ v) Triton X-100. Jurkat cells (10 7 cells) were incubated with either 1% or 0.2% Triton X-100 and subjected to sucrose density gradient centrifugation as described in Experimental Procedures. Raft fractions (total) were separated into BCh-bound (bound) and –unbound (unbound) membrane fractions. Lipids were extracted by the method of Bligh and Dyer. Cholesterol, phospholipids, and proteins in 1 mL raft fractions (0.83 · 10 7 cells equivalent) were determined. Data are means ± SD of values from three independent experiments. 1% Triton 0.2% Triton Total Unbound Bound Total Unbound Bound Cholesterol (nmol) 28.0 ± 3.5 3.0 ± 1.1 23.0 ± 3.5 48.0 ± 9.2 15.0 ± 0.1 33.0 ± 9.1 Phospholipids a (nmol) 33.0 ± 1.5 10.0 ± 1.8 23.0 ± 3.2 79.0 ± 9.4 47.0 ± 3.1 32.0 ± 6.2 C ⁄ P ratio 0.80 ± 0.05 0.30 ± 0.04 1.00 ± 0.02 0.60 ± 0.05 0.30 ± 0.03 1.00 ± 0.13 Protein (lg) 47.4 ± 4.6 28.7 ± 1.4 19.3 b ± 2.7 51.0 ± 3.0 38.5 ± 1.5 12.5 b ± 4.5 a Phospholipids were determined as the amounts of inorganic phosphorus. b Amount of proteins in the BCh-bound membrane fraction was estimated by subtracting the amount in the unbound fraction from that in the total raft fractions. Y. Shimada et al. Two raft subsets in T-cells FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS 5457 X-100 concentration on the partitioning of signalling molecules into BCh-bound membranes. Total raft frac- tions prepared with 0.2% (v ⁄ v) Triton X-100 contained about twofold more membranes than those prepared with 1% Triton X-100 as judged by lipid content (Table 1). However, the amount of membranes with a high C ⁄ P ratio recovered in the BCh-bound membrane fraction did not increase much by preparation at the lower Triton X-100 concentration. This is in contrast to a remarkable increase in membranes with a low C ⁄ P ratio recovered in the unbound fraction (Table 1). Although higher amounts of CD3e and CD3f were recovered in total raft fractions prepared at lower Tri- ton X-100 concentration, these molecules were found exclusively in the BCh-unbound membrane fraction regardless of Triton X-100 concentration (Fig. 4). Thus, the Triton X-100 concentration did not affect such characteristics of BCh-bound membranes as C ⁄ P ratio and associated molecular species. Lipid composition of membrane subpopulations Lipid extracts from total raft fractions, and from BCh-bound and -unbound membrane fractions were analysed by TLC (Fig. 5A). In comparison with total cell lipid extracts, the raft fractions were rich in choles- terol and sphingomyelin (data not shown). The BCh- bound membrane fraction contained cholesterol at a level more than twofold that of the unbound fraction, a finding consistent with its higher C ⁄ P ratio. It is noteworthy that the PS ⁄ PI intensity ratio was remark- ably different between the BCh-bound and -unbound membranes, at ratios of 10 : 1 and 1 : 2, respectively. In the former membranes, PS is a major component, and PI is a minor one. This relationship is reversed in the latter membranes. Gangliosides were also analysed by TLC (Fig. 5B). The total raft fractions contained GM1 and GM3 as the major components, with GM1 enriched in the BCh-bound membrane fraction pre- pared with 1% (v ⁄ v) Triton X-100 (ratio of GM1 in the BCh-bound membrane to that in the BCh-unbound membrane ¼ 2 : 1). When prepared at lower (0.2%) Triton X-100 concentrations, the amount of ganglio- sides recovered in the BCh-unbound fraction increased, while the level in the BCh-bound fraction was unchanged (ratio of GM1 in the BCh-bound mem- brane to that in the BCh-unbound membrane ¼ 1 : 2). Recruitment of Zap70 and CD3d to the choles- terol-enriched membrane subpopulation after anti-CD3 stimulation It has been reported that when activated by stimuli such as the anti-CD3 Ig, T-cell rafts undergo dynamic chan- ges in their size and molecular composition [23]. We analysed initial changes in the components of choles- terol-enriched subpopulations upon T-cell activation. After activation, the amounts of CD3e and CD3f recovered in raft fractions were much increased. We found that parts of Zap70 and CD3f were recruited to BCh-bound vesicles upon T-cell activation with anti- CD3 Ig (Fig. 6). The phosphorylated form of Zap70 was detected in raft fractions from stimulated cells, and a part of it was associated with BCh-bound vesicles. Lck and LAT in raft fractions were associated exclu- sively with BCh-bound vesicles regardless of activation. These results suggest that the TCR signalling initiation machinery is formed in cholesterol-enriched membrane domains. Some proteins, such as moesin, remain associ- ated with BCh-unbound membranes even after activa- tion, suggesting that the recruitment is a specific feature of some signalling molecules. Discussion Lipid rafts are defined as lateral clusters of cholesterol and sphingolipids; however, the biochemical definition of lipid rafts remains obscure. Regardless of detergent type and concentration, DRMs have been assumed to represent lipid rafts in biochemical aspects. But deter- gent conditions is the most critical factor in influencing the partitioning of molecules to raft fractions [6,7,9,23,24]. To analyse the nature of lipid rafts, it is necessary to establish a particular probe to isolate lipid rafts regardless of the preparation conditions. In this study, we tried to isolate a cholesterol-enriched Fig. 4. Protein partitioning under different detergent conditions. BCh-bound Jurkat cells were treated with either 1% Triton X-100 or 0.2% Triton X-100, and then raft fractions were obtained by sucrose density gradient centrifugation as described. BCh-bound vesicles were retrieved by avidin-magnetic beads. The total raft fraction, BCh-unbound membrane fraction and BCh-bound mem- brane fraction were subjected to SDS ⁄ PAGE. Blot membranes were probed with anti-CD3e, anti-CD3f and anti-LAT Igs. The results are representative of five independent experiments. Two raft subsets in T-cells Y. Shimada et al. 5458 FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS membrane domain from the so-called lipid raft fractions using a cholesterol-binding probe, BCh. The choles- terol-enriched membranes obtained from Jurkat cells by our method contain particular proteins for T-cell signal- ling in addition to so-called raft marker molecules. Gen- erally, membranes are more resistant to lower detergent concentrations. We examined the effect of detergent concentration on the features of the cholesterol-enriched membranes isolated by BCh. Increased amounts of pro- teins and lipids were partitioned to the total raft frac- tions prepared under lower detergent concentrations (Table 1). Obviously, much higher amounts of CD3e and CD3f were recovered in the total raft fractions (Fig. 4). However, we found that the detergent concen- tration scarcely altered the molecular species associated with BCh-isolated membrane domain. Our study sug- A B Fig. 5. Analysis of lipid compositions of raft subpopulations. BCh-bound Jurkat cells were treated with Triton X-100 and subjec- ted to sucrose density gradient centri- fugation as described. The raft fractions were further fractionated with avidin mag- netic beads. Lipids from the total raft frac- tion, and the BCh-unbound and BCh-bound membrane fractions were extracted by the method of Bligh and Dyer with slight modifi- cation. (A) After Bligh–Dyer separation, the lower phase was concentrated and analysed by HPTLC. Phospholipids were visualized with 3% cupric acetate ⁄ 8% phosphoric acid solution. (B) The upper phase was applied to a Bond Elute packed column. Ganglio- sides were eluted with methanol, separated by HPTLC and detected with resorcinol- hydrochloric acid-reagent. M, Marker lipids. The results are representative of two independent experiments. Y. Shimada et al. Two raft subsets in T-cells FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS 5459 gests that our method for isolating a particular choles- terol-enriched domain provides a useful tool for analy- sing functional membrane domains concerned with signal transduction. Our study clearly shows that so-called lipid raft frac- tions comprise two subpopulations that differ in cholesterol content or cholesterol distribution. The two membrane subpopulations separated by using BCh each has a distinctive lipid composition. As expected, the BCh-bound subpopulation comprises cholesterol- rich membranes with a high C ⁄ P ratio of 1.0, and characterized as PS-rich membranes containing GM1 and GM3. Because BCh was first bound to the cell surface and then the cells were subjected to fraction- ation, it is expected that the BCh-bound membrane vesicles are derived from the plasma membrane. Generally, PS distributes in the inner leaflet of plasma membranes in mammalian cells. Although there is insufficient information about the inner leaflet of lipid microdomains, the inner leaflet of BCh-bound mem- brane subdomains could have a PS-enriched environ- ment. On the other hand, the BCh-unbound membranes in raft fractions might be derived from any of the following origins: PM-derived cholesterol-poor membranes, intracellular cholesterol-poor membranes or intracellular cholesterol-rich membranes. Because the BCh-unbound membranes are cholesterol-poor on average (C ⁄ P ¼0.3), intracellular cholesterol-rich membranes are expected to comprise a minor popula- tion, if any. To evaluate the contribution of intracellu- lar cholesterol-rich membranes to TCR signalling, we incubated membranes of total raft fractions with BCh after detergent extraction and analysed their content. We found that a majority of CD3 in the raft fractions was recovered in the BCh-unbound fraction after this treatment (data not shown), suggesting that CD3 localized in intracellular cholesterol-rich membranes might represent a small population. As neither endo- plasmic reticulum nor lysosomal marker proteins (cal- nexin, nor Lamp-1, respectively) were detected in raft fractions, it is unlikely that these intracellular organ- elles contaminate the raft fractions. However, judging from the observation that the PS ⁄ PI profile of the BCh-unbound subpopulation is similar to that of the endoplasmic reticulum of BHK21 cells and rat liver cells [25], it is possible that the BCh-unbound sub- population includes membranes of intracellular origin. Thus our study clearly shows the existence of hetero- geneous subpopulations with quite different lipid profiles in raft fractions. To evaluate the functional meaning of these hetero- geneous subpopulations in raft fractions, we next examined the differential distribution of TCR signalling molecules between the cholesterol-enriched subpopula- tion (BCh-bound subpopulation) and the cholesterol- poor subpopulation (BCh-unbound subpopulation). Under nonstimulated conditions, transducer molecules, for example Fyn and Lck, were detected in the choles- terol-enriched subpopulation of lipid rafts. Flotillin and LAT, which are abundant in raft fractions, were also colocalized with these Src-kinases, accumulating in the BCh-bound membrane fraction. On the other hand, CD3e, CD3f and Zap70, main components of the TCR signalling initiation machinery, were mainly partitioned to the BCh-unbound subpopulation, segregated from other signalling molecules such as LAT, Lck and Fyn. However, upon stimulation with the anti-CD3e Ig, these molecules were recruited to the BCh-bound mem- branes in raft fractions. The segregation from and Fig. 6. Recruitment of signalling molecules to BCh-bound mem- branes upon T-cell stimulation. Jurkat cells were either stimulated with anti-CD3e for 10 min at 37 °C (+ anti-CD3e) or kept nonstimu- lated without antibody addition (– anti-CD3e). Cells were then incu- bated with 10 lgÆmL )1 BCh in NaCl ⁄ P i ⁄ BSA on ice, treated with 0.2% (v ⁄ v) Triton X-100 and subjected to sucrose density gradient centrifugation as described. The raft fractions were collected and separated into subfractions with avidin-magnetic beads. The total raft fraction, and the BCh-unbound and BCh-bound membrane fractions were analysed by western blotting. p-Zap70, Phospho- Zap70. The results are representative of seven independent experiments. Two raft subsets in T-cells Y. Shimada et al. 5460 FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS recruitment of these molecules to BC h-bound mem- branes could contribute to the on ⁄ off switching of TCR signalling. In addition to signalling machinery molecules, PAG and Csk, which are known to be negative regulators of TCR signalling [26], were also detected in BCh-bound membranes. By association with phosphorylated PAG, Csk negatively regulates Src-kinases [27], maintaining the ‘off state’ of signalling under nonstimulated conditions. Experiments in which b-cyclodextrin is used to remove cholesterol have provided controversial results [28,29], and the role of cholesterol in T-cell signalling remains unclear. However, our results imply that phase separation of the plasma membrane depending on cholesterol content might be involved in segregating signalling molecules from each other to maintain the ‘off’ state of T-cell signalling. Taken together, the BCh-bound cholesterol-enriched subpopulation contains both activator and inhibitor molecules for TCR signal transduction, and is likely to play an indispensable role in controlling the on ⁄ off of the signalling cascade. Using BCh, it is possible to isolate particular functional membrane domains regardless of the preparation conditions. At present, the function of the BCh-unbound raft subpopulation with a lower cholesterol content is unclear. However, the BCh-unbound region might also play an import- ant role in TCR signalling as it contains receptor molecules for TCR signalling under nonstimulated conditions. We propose that not total DRMs, but the BCh-bound cholesterol-enriched subpopulation will provide an opportunity to elucidate the structure– function relationship of lipid rafts in signal trans- duction. Experimental procedures Materials Anti-(h-toxin) serum was produced as described previously [20]. Anti-PAG IgG was raised in rabbits using a synthetic antigen peptide corresponding to the C-terminal 15 residues (ESISDLQQGRDITRL) with a cysteine residue added to the N terminus as a site for conjugation to a carrier protein. Dynabeads M-280 conjugated with streptavidin were from Dynal (Oslo, Norway). Jurkat cells were obtained from ATCC. Anti-Lck, anti-Fyn and anti-CD3f IgGs were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-Zap70, anti-Csk, anti-moesin, anti-Grb2 and anti-(flo- tillin-1) IgGs were from BD Bioscience (San Jose, CA, USA). Anti-(phspho-Zap70) IgG was from Cell Signaling Technology, Inc (Beverly, MA, USA). Anti-CD3e IgGs were from Santa Cruz Biotechnology and R & D Systems, Inc (Minneapolis, MN, USA). Anti-LAT IgG was from Upstate Biotechnology (Lake Placid, NY, USA). Cholera toxin B subunit-peroxidase conjugate was from Sigma (St. Louis, MO, USA). Preparation of BCh h-Toxin was overexpressed in Escherichia coli and purified from the periplasmic fractions using a DEAE–Sephacel col- umn [30]. A nicked h-toxin (Ch) was obtained by limited proteolysis with subtilisin Carlsberg [16]. BCh was prepared from Ch as described previously [17]. Preparation of detergent-insoluble, low density membrane fractions (raft fractions) Jurkat cells (1 · 10 7 cells) were incubated with 10 lgÆmL )1 BCh in NaCl ⁄ P i containing 1 mgÆ mL )1 BSA (NaCl ⁄ P i ⁄ BSA) for 5 min on ice, washed twice with NaCl ⁄ P i , and incubated with 1% or 0.2% (v ⁄ v) Triton X-100 in TN buf- fer (25 mm Tris ⁄ HCl pH 6.8, 150 mm NaCl) containing 2mm phenylmethanesulfonyl fluoride, 200 lm leupeptin, 25 lgÆmL )1 aprotinin and phosphatase inhibitor cocktail set II (Calbiochem) for 15 min on ice. Then the cells were homogenized with a Potter–Elvehjem homogenizer, and the homogenate was mixed with an equal volume of 80% (w ⁄ v) sucrose and overlaid with 2.4 mL 35% (w ⁄ v) sucrose and 1.3 mL 5% sucrose in TN buffer. The gradients were centrifuged at 250 000 g for 18 h at 4 °C in a SW55 rotor. After centrifugation, fractions (0.4 mL each) were collected from the top. Lipid extraction and lipid composition analysis Total lipids in the detergent-insoluble membrane fraction were extracted by the method of Bligh and Dyer [31] with slight modification. Cholesterol was quantified by a Determi- ner cholesterol assay kit (Kyowa Medex, Japan). For the analysis of phospholipid compositions, 2D TLC was carried out as follows. Samples were applied to an HPTLC plate (Merck) at the lower left-hand corner. The plate was chroma- tographed in the first dimension with chloroform ⁄ methanol ⁄ acetic acid ⁄ formic acid ⁄ water (35 : 15 : 6 : 2 : 1, v ⁄ v ⁄ v ⁄ v ⁄ v), and then with hexane ⁄ diisopropyl ether ⁄ 80% phosphoric acid (65 : 35 : 2, v ⁄ v ⁄ v) at the same direction. The third chromatography was performed in the second dimension with ethyl acetate ⁄ isopropanol ⁄ water (50 : 35 : 15, v ⁄ v ⁄ v) at a rotation of 90 °C from the first direction. Phospholipids were visualized by treatment with 3% cupric acetate ⁄ 8% phosphoric acid solution [20]. For ganglioside analysis, the upper phase from the Bligh–Dyer separation was applied to a Bond Elute packed column equilibrated with chloro- form ⁄ methanol ⁄ water (3 : 48 : 47, v ⁄ v ⁄ v). The column was washed with excess distilled water, and the bound materials Y. Shimada et al. Two raft subsets in T-cells FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS 5461 were eluted with 3 mL methanol [32]. Sample solutions were applied to an HPTLC plate and the plate was developed with acetonitrile ⁄ isopropanol ⁄ 2.5 m ammonium hydroxide con- taining 10 m m KCl (10 : 65 : 25, v ⁄ v ⁄ v). Gangliosides were detected by resorcinol ⁄ hydrochloric acid reagent. Each spot was scanned and quantified by the image analysis program macscope. Electron microscopy The detergent-insoluble membrane fractions were prepared from BCh-bound Jurkat cells, adsorbed to nickel grids, and immunolabelled with a rabbit antibiotin IgG and protein-A coupled to 10-nm colloidal gold particles as described [19]. After negative staining with 1% (w ⁄ v) uranyl acetate, raft membranes were analysed in a JEM-1200EX electron microscope (JEOL, Tokyo, Japan). Western blotting Proteins were separated by SDS ⁄ PAGE, transferred to an Immobilon-P membrane and visualized using ECL plus (Amersham Bioscience, Piscataway, NJ, USA). Others Proteins were analysed using a bicinchoninic acid protein assay kit (Pierce, Rockford, IL, USA). Phosphorus assays were performed by the method of Fiske and Subbarow [33]. Acknowledgements We thank Dr H. Waki for technical advice and gifts of authentic gangliosides. We thank Dr S. Iwashita for critical reading of the manuscript and helpful discus- sion. We thank Dr M.M. Dooley-Ohto for reading the manuscript. This work was supported by a Grant-in- Aid for Science Research from the Japan Society for the Promotion of Science (to Y. O I.). References 1 Simons K & Ikonen E (1997) Functional rafts in cell membranes. Nature 387, 569–572. 2 Brown DA & London E (2000) Structure and function of sphingolipid- and cholesterol-rich membrane rafts. J Biol Chem 275 , 17221–17224. 3 Hooper NM (1999) Detergent-insoluble glycosphingo- lipid ⁄ cholestrol-rich membrane domains, lipid rafts and caveolae. Mol Membr Biol 16, 145–156. 4 Ostermeyer AG, Beckrich BT, Ivarson KA, Grove KE & Brown DA (1999) Glycosphingolipids are not essen- tial for formation of detergent-resistant membrane rafts in melanoma cells. Methyl-beta-cyclodextrin does not affect cell surface transport of a GPI-anchored protein. J Biol Chem 274, 34459–34466. 5 Alonso MA & Millan J (2001) The role of lipid rafts in signaling and membrane trafficking in T lymphocytes. J Cell Sci 114, 3957–3965. 6 Montixi C, Langlet C, Bernard AM, Thimonier J, Dubois C, Wurbel MA, Chauvin JP, Pierres M & He HT (1998) Engagement of T cell receptor triggers its recruitment to low-density detergent-insoluble mem- brane domains. EMBO J 17, 5334–5348. 7 Xavier R, Brennan T, Li Q, McCormack C & Seed B (1998) Membrane compartmentation is required for efficient T cell activation. Immunity 8, 723–732. 8 Zhang W, Trible RP & Samelson LE (1998) LAT palmitoylation: its essential role in membrane micro- domain targeting and tyrosine phosphorylation during T cell activation. Immunity 9, 239–246. 9 Drevot P, Langlet C, Guo XJ, Bernard AM, Colard O, Chauvin JP, Lasserre R & He HT (2002) TCR signal initiation machinery is pre-assembled and activated in a subset of membrane rafts. EMBO J 21, 1899–1908. 10 Giurisato E, McIntosh DP, Tassi M, Gamberucci A & Benedetti A (2003) T cell receptor can be recruited to a subset of plasma membrane rafts, independently of cell signaling and attendantly to raft clustering. J Biol Chem 278, 6771–6778. 11 Gomez-Mouton C, Abad JL, Mira E, Lacalle RA, Gal- lardo E, Jimenez-Baranda S, Illa I, Bernad A, Manes S & Martinez AC (2001) Segregation of leading-edge and uropod components into specific lipid rafts during T cell polarization. Proc Natl Acad Sci USA 98, 9642–9647. 12 Schade AE & Levine AD (2002) Lipid raft heterogeneity in human peripheral blood T lymphoblasts: a mechan- ism for regulating the initiation of TCR signal transduc- tion. J Immnol 168, 2233–2239. 13 Harder T & Kuhn M (2000) Selective accumulation of raft-associated membrane protein LAT in T cell recep- tor signaling assemblies. J Cell Biol 151, 199–207. 14 Munoz P, Navarro MD, Pavon EJ, Salmeron J, Mala- basi F, Sancho J & Zubiaur M (2003) CD38 signaling in T cells is initiated within a subset of membrane rafts containing Lck and the CD3-zeta subunit of the T cell antigen receptor. J Biol Chem 278, 50791–50802. 15 Radeva G & Sharom FJ (2004) Isolation and characteri- zation of lipid rafts with different properties from RBL-2H3 (rat basophilic leukaemia) cells. Biochem J 380, 219–230. 16 Ohno-Iwashita Y, Iwamoto M, Mitsui K & Ando S (1986) Cold-label hemolysin produced by limited proteolysis of theta-toxin from Clostridium perfringens. Biochemistry 25, 6048–6053. 17 Iwamoto M, Morita I, Fukuda M, Murota S, Ando S & Ohno-Iwashita Y (1997) A biotinylated perfringolysin O derivative a new probe for detection of cell surface cholesterol. Biochim Biophys Acta 1327, 222–230. Two raft subsets in T-cells Y. Shimada et al. 5462 FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS 18 Ohno-Iwashita Y, Shimada Y, Waheed AA, Hayashi M, Inomata M, Nakamura M, Maruya M & Iwashita S (2004) Perfringolysin O, a cholesterol-binding cytolysin, as a probe for lipid rafts. Anaerobe 10, 125–134. 19 Waheed AA, Shimada Y, Heijinen HFG, Nakamura M, Inomata M, Hayashi M, Iwashita S, Slot JW & Ohno- Iwashita Y (2001) Selective binding of perfringolysin O derivative to cholesterol-rich membrane microdomains (rafts). Proc Natl Acad Sci USA 98, 4926–4931. 20 Ohno-Iwashita Y, Iwamoto M, Mitsui K, Ando S & Nagai Y (1988) Protease-nicked theta-toxin of Clostri- dium perfringens, a new membrane probe with no cytolytic effect, reveals two classes of cholesterol as toxin-binding sites on sheep erythrocytes. Eur J Biochem 176, 95–101. 21 Ohno-Iwashita Y, Iwamoto M, Ando S & Iwashita S (1992) Effect of lipidic factors on membrane cholesterol topology – mode of binding of theta-toxin to cholesterol in liposomes. Biochim Biophys Acta 1109, 81–90. 22 Shimada Y, Maruya M, Iwashita S & Ohno-Iwashita Y (2002) The C-terminal domain of perfringolysin O is an essential cholesterol-binding unit targeting to choles- terol-rich microdomains. Eur J Biochem 269, 6195– 6203. 23 Roper K, Corbeil D & Huttner WB (2000) Retention of prominin in microvilli reveals distinct cholesterol-based lipid micro-domains in the apical plasma membrane. Nat Cell Biol 2, 582–592. 24 Schuck S, Honsho M, Ekroos K, Shevchenko A & Simons K (2003) Resistance of cell membranes to different detergents. Proc Natl Acad Sci USA 100, 5795–5800. 25 Renkonen O, Gahmberg CG, Simons K & Kaariainen L (1972) The lipids of the plasma membranes and endo- plasmic reticulum from cultured baby hamster kidney cells (BHK21). Biochim Biophys Acta 255, 66–78. 26 Brdicka T, Pavlistova D, Leo A, Bruyns E, Korinek V, Angelisova P, Scherer J, Shevchenko A, Hilgert I, Cerny J, Drbal K, Kuramitsu Y, Kornacker B, Horejsi V & Schraven B (2000) Phosphoprotein associated with glycosphingolipid-enriched microdomains (PAG), a novel ubiquitously expressed transmembrane adaptor protein, binds the protein tyrosine kinase Csk and is involved in regulation of T cell activation. J Exp Med 191, 1591–1604. 27 Davidson D, Bakinowski M, Thomas ML, Horejsi V & Veillette A (2003) Phosphorylation-dependent regulation of T-cell activation by PAG ⁄ Cbp, a lipid raft-associated transmembrane adaptor. Mol Cell Biol 23, 2017–2028. 28 Marwali MR, Rey-Iadino J, Dreolini L, Shaw D & Takei F (2003) Membrane cholesterol regulates LFA-1 function and lipid raft heterogeneity. Blood 102, 215–222. 29 Tu X, Huang A, Bae D, Slaughter N, Whitelegge J, Crother T, Bickel PE & Nel A (2004) Proteome analysis of lipid rafts in Jurkat cells characterizes a raft subset that is involved in NF-kappaB activation. J Proteome Res 3, 445–454. 30 Sekino-Suzuki N, Nakamura M, Mitsui K & Ohno- Iwashita Y (1996) Contribution of individual trypto- phan residues to the structure and activity of theta-toxin (perfringolysin O), a cholesterol-binding cytolysin. Eur J Biochem 241, 941–947. 31 Bligh EG & Dyer WJ (1959) A rapid method of total lipid extraction and purification. Can J Biochem Physiol 37, 911–917. 32 Figlewicz DA, Nolan CE, Singh IN & Jungalwala FB (1985) Pre-packed reverse phase columns for isolation of complex lipids synthesized from radioactive precur- sors. J Lipid Res 26, 140–144. 33 Fiske CH & Subbarow Y (1925) The colorimetric deter- mination of phosphorus. J Biol Chem 66, 375–400. Y. Shimada et al. Two raft subsets in T-cells FEBS Journal 272 (2005) 5454–5463 ª 2005 FEBS 5463 . AA, Hayashi M, Inomata M, Nakamura M, Maruya M & Iwashita S (2004) Perfringolysin O, a cholesterol-binding cytolysin, as a probe for lipid rafts. Anaerobe. Gomez-Mouton C, Abad JL, Mira E, Lacalle RA, Gal- lardo E, Jimenez-Baranda S, Illa I, Bernad A, Manes S & Martinez AC (2001) Segregation of leading-edge and uropod