Top

iBiology

Bringing the World's Best Biology to You

Membrane Rafts

Part 1: What are Membrane Rafts?

00:00:03.11 Hi, I'm Satyajit Mayor, from the National Centre for
00:00:07.28 Biological Sciences in Bangalore, here to talk to you
00:00:11.26 about membrane rafts as part of the iBioSeminars
00:00:17.01 series. I'm going to talk to you about membrane rafts in
00:00:20.00 three parts. The first part is where I will talk about what
00:00:25.20 are membrane rafts, and our understanding, at least our
00:00:29.18 colloquial understanding, of membrane rafts. In the
00:00:33.11 second part, I will try and describe our experiments that
00:00:36.25 have led us to investigate what membrane rafts may
00:00:41.20 look like in living cell membranes. And in the third part,
00:00:46.14 I'd like to tell you what in fact our investigation on trying
00:00:50.29 to understand these membrane rafts have revealed to
00:00:53.10 us in terms of trying to understand the structure of cell
00:00:57.00 membranes in a living, mobile, motile system. So without
00:01:06.11 further ado, let me move to the first part. What are
00:01:10.25 membrane rafts? Well, the cell membrane is where
00:01:17.20 membrane rafts are situated. The cell membrane, as
00:01:21.18 you all know, separates the inside of the cell from the
00:01:24.15 outside of the cell, and this barrier is what in fact allows
00:01:31.27 the cell to transact its interactions with the outside
00:01:35.00 world, without having to worry about the outside world,
00:01:38.03 and the cell is able to live within the membrane as a
00:01:42.26 distinct entity. The cell membrane has a particular
00:01:49.26 chemical composition. In fact, it is made up chiefly of
00:01:54.20 lipids. There are about 500 types of lipids. There are
00:02:00.00 several classes of these 500 types; these are defined
00:02:05.10 by the head group of the lipids: The
00:02:07.25 phosphatidylethanolamine is one type, the phosphatidylserine,
00:02:11.12 -choline, and the sphingomyelin is another type of lipid.
00:02:18.16 And the constituents of each class, or each type, of
00:02:21.16 lipid defer from each other by the nature of the
00:02:23.21 hydrophobic acyl chain that's attached to these lipids.
00:02:28.03 The length and the degree of unsaturation characterize
00:02:31.05 the differences between these different types of lipids.
00:02:36.07 An extremely important class of lipids is cholesterol,
00:02:41.06 because cholesterol, in most eukaryotic cells, is about
00:02:44.22 50% in terms of the total content of the phospholipids at
00:02:52.21 the cell surface, so in terms of mole-by-mole fraction,
00:02:57.03 cholesterol is often 50% of the total lipid composition of
00:03:01.12 a cell membrane. It's an extremely unusual lipid
00:03:06.11 because it has a planar structure, it in fact makes lipids
00:03:15.14 behave in a very funny way when associated with
00:03:17.04 them. But lipids in general, the different types of lipids,
00:03:22.00 have two distinct types of abilities to pack together.
00:03:26.29 First of all, they form bilayers, and in these bilayers, the
00:03:32.00 saturated acyl chain-containing lipids have a tendency
00:03:36.09 to pack a little tighter, and the lipids that have an
00:03:40.21 unsaturated chain, hydrocarbon chain, they tend to in
00:03:44.16 fact pack not as well and as compact with each other.
00:03:49.15 And cholesterol, on the other hand, is able to modify the
00:03:54.00 properties of these different types of lipid in different
00:03:56.27 directions. In fact, the roles that cholesterol plays in the
00:04:02.10 structure and organization of the membrane and
00:04:04.12 different lipids in the membrane is as yet quite
00:04:08.10 contentious. We're not very clear as to what exactly
00:04:13.14 cholesterol does the membrane, and I think the whole
00:04:15.17 field of membrane rafts is really a search to understand
00:04:21.24 what types of roles cholesterol, and what functions
00:04:24.12 cholesterol, might play in cell membranes. I think to
00:04:30.09 understand the cell membrane structure, it's important to
00:04:34.01 know the history of how the cell membrane structure
00:04:37.29 was elucidated. There's an excellent review by Michael
00:04:42.28 Edidin, it's a timeline in which he describes the different
00:04:48.00 discoveries which led to the final elucidation as we
00:04:53.10 know it today of the structure of the plasma membrane,
00:04:57.05 where he begins with the discoveries from Agnes
00:05:00.13 Pockels, where she takes oil droplets and puts them on
00:05:05.09 water and makes a unimolecular layer of oil on water; to
00:05:10.08 the discoveries from Langmuir, where he suggested, I
00:05:15.10 think, for the first time, that lipids might in fact form
00:05:18.01 bilayers and be in the form of a bilayer, which could
00:05:23.11 cover the hydrophobic parts of these lipids, to make
00:05:30.23 long bilayer, two-dimensional bilayer-like structures. The
00:05:39.09 debate in the cell biology literature was, how do you put
00:05:45.22 the proteins into this bilayer, because at the time it was
00:05:49.03 already known that the cell membrane constituted not
00:05:52.26 only of lipids but also of proteins. And there was a
00:05:56.13 debate between people who believed that the proteins
00:05:59.21 and lipids formed a trilamellar-like structure, and then
00:06:03.05 there was others who believed that it was simply a
00:06:06.24 bilayer, where proteins and lipids in fact interdigitated
00:06:11.03 between each other. But what I'd like to point out, that
00:06:16.05 the history of this field reveals that these was an
00:06:20.14 incredibly intertwined relationship between concepts
00:06:23.11 that were borrowed from studies on artificial
00:06:26.00 membranes, with experiments that were done on cell
00:06:28.16 membranes. And I think, together, the understanding of
00:06:33.00 the structure of the cell membrane was slowly
00:06:36.03 elucidated. A major breakthrough occurred when, I
00:06:40.15 think, Singer and Nicholson put together a fairly
00:06:46.18 revolutionary model at that time, the fluid-mosaic model,
00:06:49.23 which was derived from observations of electron
00:06:54.11 microscopic analysis of arrangement of proteins on
00:06:58.00 membranes and studies, in fact, some early studies from
00:07:01.22 Michael Edidin's work, where they looked at the
00:07:05.12 diffusion of proteins on cell membranes. These
00:07:13.02 observations were in fact coupled with an
00:07:16.10 understanding of how membrane proteins could be embedded in the
00:07:20.03 bilayer. They could be either peripheral or integral
00:07:23.22 membrane proteins, and I think understanding how
00:07:26.28 proteins were embedded in the bilayer, coupled with
00:07:29.00 these observations from electron microscopic analysis
00:07:32.22 of protein arrangements, led Singer and Nicholson to
00:07:36.06 propose that the cell membrane was in fact a sea of
00:07:42.29 lipids in which proteins were embedded as objects. And
00:07:49.07 in fact they were floating in this sea of lipids, which
00:07:51.26 were constituted of a bilayer-like structure. This model, I
00:07:56.28 think, has impacted our understanding of the cell
00:08:00.12 membrane structure quite substantially, and I think it's
00:08:03.00 worth understanding what the central tenets of this
00:08:05.24 model are today. The model states that the membrane
00:08:10.14 is an oriented, two-dimensional solution of amphiphatic
00:08:14.09 proteins (or lipoproteins), and lipids are in instantaneous
00:08:19.13 thermodynamic equilibrium in this system. So it's a model
00:08:23.03 about a passive fluid in which lipids are in fact
00:08:28.10 dissolved. The consequences of such a model are that,
00:08:33.26 unless otherwise acted upon, there is generally no long-
00:08:38.09 range order in a mosaic membrane within this lipid
00:08:43.05 matrix; and that the lipids of functional cell membrane
00:08:47.06 are in a fluid state rather than a crystalline state, in fact,
00:08:50.11 they are free to diffuse all over the place rather than be
00:08:52.08 gathered and stuck in one location. Of course, this is in
00:08:57.02 the absence of any external influences, so the lipids in
00:08:59.27 the cell membrane, left to themselves, behave as if they
00:09:07.17 are in a passive, thermodynamically equilibrated system.
00:09:14.03 And it is in this context that I think we should try and
00:09:16.16 understand what are membrane rafts and why did
00:09:20.23 people feel the need to propose structures such as
00:09:24.15 membrane rafts. In fact, today, there are over 2000 hits
00:09:28.24 that you can pick up in the PubMed, when you just
00:09:34.01 type "membrane rafts," you pick up about 2500
00:09:38.25 citations for lipids rafts. If you look at other descriptors
00:09:45.28 of these membrane domains, you may pick up more.
00:09:49.23 They're proposed to be involved in diverse cellular
00:09:52.06 processes: membrane trafficking, membrane sorting,
00:09:55.23 signaling, cell migration. They're hypothesized to be
00:10:00.10 involved in the nervous system, in immune function, in
00:10:04.25 nutrient uptake, in cell cycle, virus budding and entry,
00:10:08.18 pathogen biology, cell motility, what have you. And I
00:10:11.26 think they are thought to be involved in almost every
00:10:14.26 aspect of cellular function that involves membranes in a
00:10:19.02 cell. And of course, that affects diverse biological
00:10:25.27 phenomena. So what is understood by the term
00:10:29.22 "membrane rafts"? Membrane rafts... and here again I
00:10:34.16 think it's worth going to some sort of a summary of
00:10:39.28 perspective by some of the pioneers in the field. Kai
00:10:43.23 Simons, in a very influential paper in a Perspectives in
00:10:49.09 Nature, proposed that lipids rafts are regions of the cell
00:10:53.21 membrane where cholesterol and sphingolipids can
00:10:57.08 come together and thereby segregate in the plane of
00:11:03.16 the membrane, and attract other components to come
00:11:07.08 and concentrate in these regions. He also proposed
00:11:11.06 that there could be proteins that could stabilize these
00:11:15.18 segregated regions, proteins like caveolin, for example,
00:11:19.08 which can associate with the underside of the cell
00:11:23.06 membrane and stabilize these domains so that they
00:11:27.04 could hold a special lipid composition. So there are
00:11:34.13 three features of these types of domains. One is that
00:11:39.17 they're expected to be cholesterol and sphingolipid
00:11:41.27 enriched. A second feature that we understand in terms
00:11:45.17 of a descriptor for lipid rafts is that they are constituted
00:11:50.24 of membranes that are resistant to detergent extraction,
00:11:54.25 and we'll spend a little time about this notion (detergent
00:11:57.27 extraction) in the subsequent minutes of this part. And
00:12:03.29 the third idea about lipid rafts is that functions in the cell
00:12:11.11 that require the presence of cholesterol are also
00:12:15.12 thought to constitute functions that require membrane
00:12:22.29 rafts. So I think each aspect of this sort of trilogy of
00:12:30.08 sphingolipid- and cholesterol-enriched domains,
00:12:32.24 detergent-resistant membranes, and a cholesterol-
00:12:35.13 sensitive functional context have provided, I would
00:12:42.25 argue, a colloquial understanding of what membrane
00:12:45.22 rafts are today. But, the structures again, I think, remain
00:12:54.27 illusive/elusive and definitely controversial. And I think
00:13:01.00 this seminar is about trying to bring to you some of those
00:13:03.24 controversies, and also bring to you a search for
00:13:08.06 membrane rafts in living systems, which we will talk
00:13:12.15 about in Parts 2 and 3 of this seminar. So, today, there
00:13:21.17 are many models for lipid rafts (besides the one
00:13:24.12 proposed by Kai Simons): That lipid rafts are constituted
00:13:27.14 of proteins that are shelled by a small boundary, lipid,
00:13:34.12 around them; another concept of membrane rafts is that
00:13:40.17 there are many different types of domains in cell
00:13:43.14 membranes, where in fact you have a large number of
00:13:48.00 different types of aggregates of different types of
00:13:51.10 composition, maintained in some fashion by cellular
00:13:58.14 membrane composition, particularly, by the cell
00:14:02.06 maintaining a particular lipid composition. So these are
00:14:05.20 the different types of ideas that are circulating about
00:14:09.26 what these lipid rafts might look like in cell membranes.
00:14:15.20 At this point, again, I think it's important to understand
00:14:18.14 why did Kai Simons and his coworkers feel the need to
00:14:24.09 propose a notion like lipid rafts. And I think an
00:14:28.23 understanding of that comes from, again, looking back
00:14:33.00 at some of the historical experiments that led to the
00:14:35.27 proposal of the raft hypothesis, and this comes from
00:14:39.02 studies looking at the membrane composition of the
00:14:45.05 different domains of an epithelial cell. An epithelial cell is
00:14:51.16 a polarized cell, it has an apical surface, and it has a
00:14:55.19 basolateral surface, and these two surfaces have
00:14:59.14 distinct lipid compositions. Now if the membrane of all
00:15:03.29 cells are a well-mixed system completely at
00:15:09.04 thermodynamic equilibrium, one won't expect, unless
00:15:15.01 there are physical barriers, to have a distinct
00:15:20.14 composition at one surface versus the other. And of
00:15:23.09 course, in epithelial cells we know there are barriers,
00:15:25.21 there are tight junctions, these little dark patches that I
00:15:28.04 draw here on both sides. These are tight junctions that
00:15:33.00 separate the apical surface from the basolateral
00:15:35.02 surface. But that means that one needs a mechanism
00:15:37.29 of generating a lipid composition from internal sources,
00:15:47.10 where molecules of different stripes can be delivered
00:15:51.19 preferentially to one surface versus the other. And that's
00:15:55.14 only possible, Kai Simons and Gerrit van Meer at that
00:16:00.24 time argued, when you can segregate these
00:16:04.22 components during the membrane trafficking process
00:16:08.17 from the Golgi to the cell surface. If you can segregate
00:16:11.17 components of one type from the other, deliver one set
00:16:15.04 of components, for example, cholesterol- and
00:16:17.16 sphingolipid-enriched components, to one domain and
00:16:21.03 not to the other, then automatically the cell is capable
00:16:24.08 of generating this heterogeneous composition in the
00:16:29.10 two surfaces of the cell. The other remarkable
00:16:34.20 observation they made was obtained by studying the
00:16:39.02 lipid composition of enveloped viruses. They found that
00:16:46.22 viruses that budded from the basal surface of cells had
00:16:51.18 a completely distinct lipid composition from those that
00:16:54.29 budded from the apical surface. In fact, some of the
00:16:59.09 virus that bud from the apical surface are the
00:17:03.12 hemagglutinin virus, and the VSVG of the VSV virus
00:17:09.29 buds from the basal surface. These viruses have very
00:17:14.27 different cholesterol and sphingolipid compositions. The
00:17:17.22 HA, or the hemagglutinin, influenza virus has a
00:17:25.26 cholesterol- and sphingolipid-enriched membrane,
00:17:29.18 whereas the VSV virus is in fact rather depleted of
00:17:33.16 these components. So together with these two ideas,
00:17:38.04 Kai Simons proposed that the membrane of cells must in
00:17:45.16 fact be partitioned in some form so that they can
00:17:50.20 segregate domains, and these domains can be dealt
00:17:53.14 with differently inside the cell. Another major player in
00:17:58.27 bringing about the raft hypothesis was the GPI-
00:18:02.20 anchored protein. So GPI-anchored proteins are
00:18:06.23 proteins that are in fact made in the endoplasmic
00:18:12.28 reticulum and attached to a preformed glycolipid, which
00:18:18.11 then brings the GPI-anchored protein attached to a lipid
00:18:24.05 anchor in the membrane of the endoplasmic reticulum,
00:18:28.14 after which the GPI anchor is in fact delivered to the
00:18:32.20 surface of a cell where it behaves like a lipid-anchored
00:18:36.21 component. This component in fact, when observed in
00:18:43.13 epithelial cells again, showed up at the apical surface
00:18:46.13 of these epithelial cells, and that again suggests that
00:18:52.16 perhaps these GPI-anchored proteins are also
00:18:55.13 associated with a specific domain in these epithelial
00:19:01.01 cells, during the trafficking process, and then they are
00:19:06.08 delivered to the cell surface in these specific domains,
00:19:10.06 and excluded from being delivered to the basolateral
00:19:12.16 surface. In fact, some of the early work from the
00:19:19.07 trafficking of these GPI-anchored proteins contributed
00:19:24.06 substantially to the raft hypothesis, because during the
00:19:28.09 trafficking process, Deborah Brown and Jack Rose, in
00:19:34.03 1992, in fact, published, I think, one of the classic
00:19:40.20 papers in this field, where they showed that, during the
00:19:44.10 trafficking of these GPI-anchored proteins, they were
00:19:46.28 associated with detergent-resistant membranes. And it
00:19:50.09 was the association with detergent-resistant membranes
00:19:53.20 that began in the Golgi, and then continued on till these
00:19:57.07 proteins were delivered to the cell surface. At about the
00:20:04.08 same time, Nigel Hooper and colleagues had also
00:20:06.15 discovered that a wide variety of glycosphingolipids and
00:20:10.12 GPI-anchored proteins were all insoluble in a cold Triton
00:20:16.06 (which is a detergent) extraction procedure. And putting
00:20:21.21 these two observations together, of cold Triton
00:20:24.10 insolubility and the association with this Triton-insoluble
00:20:29.14 domain during the trafficking process, Deborah Brown
00:20:33.01 and Jack Rose proposed that these detergent-resistant
00:20:36.04 membranes reflect the membrane domains that are
00:20:40.13 involved in trafficking of these specialized lipid
00:20:45.23 components to the cell surface. Coupled with that
00:20:52.04 notion of the detergent-resistant membranes, there were
00:20:57.01 studies in artificial membranes, where it was observed
00:21:01.13 that cholesterol-induced phase segregation of
00:21:06.14 homogeneously mixed lipid components, neutral
00:21:10.06 sphingolipids were also found to be clustered in these
00:21:13.21 artificial membranes. So putting all these observations
00:21:18.05 together, along with the notion that, if you crosslink
00:21:24.08 these lipid-anchored components at the cell surface,
00:21:26.23 one could often activate signaling mechanisms inside
00:21:29.26 the cell, and that these signaling mechanisms were
00:21:34.08 activated by crosslinking of these lipid components with
00:21:37.25 no obvious cytoplasmic connection. So the crosslinking
00:21:41.28 of these components with no obviously cytoplasmic
00:21:44.08 connection implied that there might be some trans-
00:21:46.27 bilayer connection between the outside, the
00:21:52.27 exoplasmic, leaflet of the cell and the inner leaflet of the
00:21:56.28 cell membrane. At this time, investigators studying the
00:22:06.14 signaling of these GPI-anchored proteins noticed that,
00:22:09.11 whenever they isolated these GPI-anchored proteins in
00:22:12.08 these Triton-insoluble structures from cells in which
00:22:15.28 signaling was occurring, they could find lipid-tethered
00:22:21.16 inner leaflet components that were in fact associated
00:22:27.13 with the GPI-anchored proteins, and these lipid-linked
00:22:31.19 inner leaflet components were part of the Src family,
00:22:35.00 tyrosine kinase complexes. IgE receptor, that is, the
00:22:41.28 immunoglobulin E receptor, was also found in
00:22:45.15 detergent-resistant membranes, and detergent-resistant
00:22:48.01 membranes were important for their signaling activity.
00:22:52.13 The B cell receptor was also found in these detergent-
00:22:56.03 resistant membranes, and these lipid domains that the B
00:23:02.05 cell receptor associated with were also thought to be
00:23:05.11 important for the signaling. Putting all these things
00:23:10.24 together, I think by the early to the mid-90s, there was
00:23:16.01 an operational definition of rafts developing. One, on
00:23:19.23 one hand, components that were associated with this
00:23:24.05 detergent-resistant membrane were thought to be part
00:23:26.21 of rafts, and on the other hand, if one was able to
00:23:30.08 perturb the lipid constituents in the cell membrane, for
00:23:33.19 example, cholesterol and sphingolipids, by using
00:23:38.04 cholesterol-chelating agents or metabolically depleting
00:23:43.20 the levels of these lipids in cells, then an operational
00:23:47.13 definition of a particular function in the cell could be
00:23:54.12 associated with that function in the cell requiring the
00:23:59.08 role of rafts, or requiring the rafts in that process. For
00:24:03.12 example, sorting of membrane proteins, if they were
00:24:08.19 associated with detergent-resistant membranes, and the
00:24:12.28 sorting was perturbed by cholesterol and sphingolipid
00:24:17.02 level perturbation, then it was thought that the sorting of
00:24:20.05 those membrane components was occurring by these
00:24:25.01 membrane rafts. So, in fact, this operational definition is
00:24:31.14 a very simple experimental tool. It's easy to use, and a
00:24:38.18 lot of people, a lot of investigators in the field, wanted to
00:24:42.13 know whether their favorite cellular process was
00:24:46.26 involving rafts by these criteria or not. And from the mid-
00:24:52.14 90s to, I'd say, today, a large number of researchers
00:25:01.08 have used these criteria to make claims about
00:25:04.05 membrane rafts in their favorite cellular process. But I
00:25:10.22 think this idea that components could be segregated in
00:25:14.08 cell membranes has also driven a lot of research in the
00:25:21.09 artificial membrane arena, where people have spent a
00:25:25.27 lot of time trying to investigate the phase behavior of
00:25:28.27 these detergent-resistant membranes. And we'll get to
00:25:32.12 that in a little bit. The cellular context, again, of what
00:25:39.29 these detergent-resistant membranes look like has also
00:25:42.24 been investigated. And I think a lot of research now
00:25:48.01 indicates that perhaps these simple operational tools
00:25:52.10 may not be the best way to look for membrane rafts. But
00:25:57.03 if we examine the literature on this detergent-resistant
00:26:03.25 membrane composition, for example, it's a composition
00:26:09.05 where you have equimolar cholesterol, sphingolipid,
00:26:12.19 and phosphatidylcholine... if you examine this
00:26:15.19 composition in a wide variety of systems, in giant
00:26:19.09 unilamellar vesicles, in supported membranes, using
00:26:24.02 fluorescently labeled lipid probes that could partition into
00:26:27.19 different types of lipid phases, a whole wealth of
00:26:31.04 literature has emerged. And this has been driven by the
00:26:37.23 fact that this particular type of lipid composition supports
00:26:42.19 two types of immiscible lipid phases. It supports two
00:26:46.09 types of immiscible lipid phases in a fluid membrane,
00:26:50.12 where one of them is a liquid-ordered phase, and the
00:26:54.10 other is a liquid-disordered phase, and these two
00:26:58.00 phases coexist in the membrane. The liquid-ordered
00:27:00.29 phase is defined by a phase where there is limited
00:27:05.28 translational mobility, but there is no long-range order,
00:27:11.08 whereas a liquid-disordered phase is a phase there is
00:27:17.14 no short-range order nor limited translational mobility. So
00:27:21.14 these two phases coexist in cell membranes, and the
00:27:25.10 liquid-ordered phase is a phase that seems to
00:27:28.26 correspond to a phase with composition of that of the
00:27:32.05 detergent-resistant membrane. And soon enough,
00:27:37.07 Deborah Brown and Erwin London in the early 90s in
00:27:40.15 fact, showed that there was a connection between
00:27:44.04 liquid-ordered domains and detergent-resistant
00:27:46.26 membranes in artificial membranes, and made the link
00:27:51.08 with what the detergent-resistant membranes might be
00:27:54.22 revealing in a cellular context. So liquid-ordered
00:27:58.28 domains, which were correlated with detergent-resistant
00:28:03.04 membranes, became in fact a characteristic or a
00:28:08.20 property of membrane rafts present in cell membranes.
00:28:15.01 And if you look at the phase behavior of these raft
00:28:18.01 composition artificial membranes, one can see sort of
00:28:21.23 macroscopic segregation of lipid domains, in fact, you
00:28:24.20 can see regions of this artificial membrane which can
00:28:31.18 concentrate GPI-anchored proteins that have been
00:28:34.04 added to them, and in fact exclude phospholipids that
00:28:39.00 have been put into the same membrane. So, these
00:28:43.24 artificial membranes in fact revealed the capacity of
00:28:47.05 lipids to naturally segregate components when they
00:28:52.26 were introduced into the same membrane. The phase
00:28:57.19 behavior of raft composition in giant unilamellar vesicles
00:29:04.11 was another great push to understand cholesterol,
00:29:10.00 sphingolipid, and phosphatidylcholine composition
00:29:16.19 variations in these GUVs. And what people have found
00:29:21.07 is that, if you add cholesterol to these GUVs, you see a
00:29:27.12 dramatic homogenization of these liquid-ordered and
00:29:31.24 liquid-disordered membranes that could be present in
00:29:35.11 conditions where cholesterol is completely absent. So if
00:29:40.15 you have for example phosphatidylcholine, saturated
00:29:43.29 acyl chain-containing choline, and unsaturated acyl
00:29:48.19 chain-containing choline lipids, they segregate in the
00:29:52.17 membrane, and then if you add cholesterol, those
00:29:56.27 different domains seem to get mixed up. So it appears
00:30:01.26 then that cholesterol is able to modulate the nature of
00:30:06.05 these domains that are present in these artificial
00:30:09.02 membranes. And the lesson that we've learned from
00:30:13.17 these artificial membranes is that the role of cholesterol
00:30:19.06 is sort of rather an ambiguous one, where at one level,
00:30:24.22 it's able to promote the formation of domains under
00:30:27.20 certain conditions; at other level, it's able to mix
00:30:29.22 domains. And this could be because of different kinds
00:30:34.04 of chemical interactions that cholesterol has with other
00:30:40.03 lipids in the membrane. Cholesterol in fact is thought to
00:30:43.28 interdigitate between these acyl chains of lipids and fill
00:30:49.08 up spaces within lipids, so if lipids like unsaturated
00:30:53.01 chain-containing lipids are present, cholesterol might go
00:30:56.23 in and fill up spaces there. Alternatively, it could help in
00:31:01.08 stabilizing the saturated chains of phospholipids and
00:31:05.28 extending the chains out, and making them more rigid in
00:31:09.15 their structure. So cholesterol could have multiple roles,
00:31:13.16 and each of these roles could be exhibited in different
00:31:17.25 contexts in different times. Cholesterol could also form
00:31:21.03 condensed complexes with specific types of lipids, and
00:31:23.27 work from Harden McConnell's group has shown that
00:31:26.12 cholesterol and sphingolipids can form specific
00:31:30.18 complexes in artificial membranes, albeit these were
00:31:36.22 monolayer experiments. Other lessons that we've
00:31:41.23 learned from artificial membranes is that there could be
00:31:45.01 a trans-bilayer coupling of lipids across the outer and
00:31:52.16 inner leaflet. The charge of the lipid head group could
00:31:56.25 also affect the partitioning of a lipid component in a
00:32:00.03 liquid-ordered domain, with respect to that of a liquid-
00:32:03.09 disordered domain. But, I think the question remains,
00:32:09.26 are these detergent-resistant membranes preexisting
00:32:14.02 structures or functional domains in cell membranes, or
00:32:19.04 are they some artifact of the operational criteria that
00:32:22.26 people have used to isolate them? I think a very
00:32:26.21 important observation was made by Heerklotz and
00:32:31.12 Joachim Seelig, where they showed, using quantitative,
00:32:35.12 physicochemical studies, using isothermal calorimetry,
00:32:42.24 that when you added Triton to artificial membranes that
00:32:46.08 were homogeneous, you could in fact generate
00:32:49.27 domains that were relatively ordered, and this was
00:32:54.15 particularly true for membranes that contained
00:32:57.06 sphingomyelin. And they argued that this was because
00:33:00.12 Triton did not like to associate with sphingomyelins and
00:33:05.09 somehow pushed sphingomyelins away while
00:33:08.14 solubilizing other lipid components in the cell membrane.
00:33:11.29 So in fact they argued that the addition of the Triton, or
00:33:16.11 a detergent, to the cell membrane could influence the
00:33:20.28 organization of lipid components present in the absence
00:33:24.29 of the addition of Triton, and in fact, in some cases,
00:33:29.08 even generate ordered membrane domains when the
00:33:34.13 Triton was added into these systems. So I think the
00:33:38.11 detergent-resistant operational definition suffered a
00:33:42.23 major setback by some of these very detailed
00:33:46.21 observations made in artificial membrane studies.
00:33:51.09 Another, I'd say, setback to the idea of membrane rafts
00:33:56.19 in cell membranes was from experiments again by
00:34:01.09 Michael Edidin, where he noticed that, when you
00:34:04.06 removed cholesterol from cell membranes, the cell
00:34:06.22 membranes seem to suffer a disconnection from the
00:34:11.12 underlying cytoskeleton. In fact, the cortical actin that
00:34:15.16 was present beneath the cell membrane was disrupted
00:34:19.03 when cholesterol was taken away from the cell
00:34:21.22 membrane. And in fact, these PIPs, or phosphorylated
00:34:26.27 inositols or phosphatidylinositols, were selectively
00:34:36.08 removed from the bilayer by means as yet not well
00:34:41.21 understood, when cholesterol was removed from cell
00:34:43.23 membranes. Not only that, Antonella Viola and
00:34:49.11 colleagues noticed that when you deplete cells of
00:34:54.07 cholesterol, the intracellular stores of calcium were
00:34:58.08 removed, and the plasma membrane of cells tended to
00:35:02.00 be more depolarized. So these were artifacts
00:35:06.04 associated with these experimental ways of removing
00:35:10.07 cholesterol from cell membranes. Where does this leave
00:35:17.03 us in terms of what lipid rafts might be like? So are they
00:35:22.07 creations of these operational definitions, and I'd like to
00:35:25.02 argue that this is in fact the case, that the lipids rafts, as
00:35:30.20 observed by these different criteria of detergent
00:35:33.20 resistance, of cholesterol level perturbation, are in fact
00:35:40.23 in some sense an artifact of the operational definition
00:35:45.20 itself. And then where does this leave us with in terms
00:35:51.05 of the raft hypothesis? Here, I think the cells are telling
00:35:56.04 us a different story. The cells seem to be saying that
00:36:01.20 there is a definite need and a reason to expect
00:36:08.10 specialized membrane domains endowed with specific
00:36:11.08 function, especially in terms of the sorting and
00:36:14.02 segregation of components in the membrane to deliver
00:36:19.15 components to different aspects of the cell surface, to
00:36:24.09 engage with signaling across the bilayer of lipid-
00:36:27.08 anchored components. In fact, all these observations
00:36:31.18 suggest that there must be membrane domains which
00:36:35.02 exist in cell membranes, and these operational
00:36:39.22 definitions have merely been correlates of some kind,
00:36:42.28 but have not really revealed the true nature of these
00:36:46.12 membrane domains. So I think what is needed at this
00:36:50.21 stage is to revisit this raft hypothesis. And I think the
00:36:55.11 questions that we should be asking are: Can we
00:36:59.04 visualize specific regions of the cell membrane where
00:37:04.11 lipids and protein components are concentrated? Can
00:37:08.03 we literally look at these regions where components of
00:37:13.11 certain functional content and context are
00:37:17.11 concentrated? And can we observe them and can we
00:37:21.01 watch them functioning in cell membranes? And in fact,
00:37:27.09 is the membrane really a patchwork quilt, where lipids
00:37:32.19 are organized in some form or the other, perhaps with
00:37:36.00 the help of proteins? And I'm now at the end of this
00:37:41.16 part, where I'd like to make a break, and in Parts 2 and
00:37:51.16 3 of this iBioSeminar, I'm going to describe how
00:37:55.14 studying the endocytosis of GPI-anchored proteins has
00:37:58.20 in fact compelled me to reexamine the notion of
00:38:02.18 membrane rafts in a living cell. And in the Part 3, we
00:38:08.03 explore what these ideas have revealed to us about
00:38:13.02 membrane organization in a living cell. Thank you.
00:38:19.04

Part 2: Looking for Functional Rafts in Cell Membranes

00:00:03.24 Welcome to the second part of the iBioSeminar. This
00:00:07.20 is titled, "Looking for functional rafts in cell
00:00:10.18 membranes." And I'm, as you can see, Satyajit
00:00:14.12 Mayor, from the National Centre for Biological
00:00:17.25 Sciences, talking about some work that we've been
00:00:21.00 doing in collaboration with a close colleague of mine,
00:00:25.02 Dr. Madan Rao, the Raman Research Institute, also
00:00:29.11 in Bangalore. He's a soft condensed matter physicist,
00:00:34.27 with whom we've been trying to explore the structure
00:00:38.14 of cell membranes, and in fact Parts 2 and 3 are
00:00:42.29 work that we've been doing in collaboration in an
00:00:45.19 interdisciplinary way between a cell biologist (myself)
00:00:50.02 and a physicist (Madan Rao). And the reason we
00:00:56.06 have engaged in this collaboration, I think, should
00:01:00.07 have in some ways been apparent from the lessons
00:01:04.13 we learned from Part 1, where we examined the
00:01:10.13 notion of membrane rafts and the operational criteria
00:01:14.19 that have been used to look for membrane rafts in
00:01:18.14 cells, and some of the problems associated with
00:01:22.15 those criteria to look for rafts in cell membranes.
00:01:26.18 Now, what we have then argued is that, while there
00:01:35.04 are problem associated with the operational criteria to
00:01:39.13 look for rafts, the cell membranes and cells
00:01:43.02 themselves seem to be telling us that membrane
00:01:47.08 segregation, or segregation of specific components
00:01:51.06 in the membrane of living cells, must occur. And the
00:01:57.01 need of the hour then is ways to look for these
00:02:01.27 segregated regions in cell membranes. So this part is
00:02:06.22 in fact our exploration of looking for these segregated
00:02:12.05 regions in cell membranes, potentially imbued with
00:02:17.15 function. Well, the story in my laboratory, anyway,
00:02:23.12 starts with trying to understand the endocytic
00:02:27.27 process, and in cells, one of the best-characterized
00:02:31.13 means of endocytosis is the one of clathrin- and
00:02:36.07 dynamin-dependent endocytosis, where these
00:02:39.15 components collaborate to make membrane pits, and
00:02:46.15 these pits then pinch off material at the cell surface,
00:02:50.19 bring them into the cell as endocytic vesicles, and
00:02:54.23 then these vesicles, which are composed of
00:02:59.18 components of the cell membrane encased in their
00:03:02.21 budding coats, these components then move into
00:03:09.02 the cell, to form endocytic structures, which contain
00:03:16.20 a fair representation of the components of the plasma
00:03:20.01 membrane. For example, in this slide, what you see
00:03:23.20 are transferrin (an iron-binding protein) bound to the
00:03:29.07 transferrin receptor, shown in the green fluorescent
00:03:34.09 image here, and a lipoprotein LDL, labeled in a red
00:03:39.19 dye, coming into endosomes inside the cell, wherein
00:03:44.21 they are present in the same endocytic organelle.
00:03:50.06 The questions that we have asked are in fact ones
00:03:53.23 about what happens to the endocytosis of the
00:03:59.23 membrane components in general. So if you look at
00:04:03.10 the underbelly of the plasma membrane, it looks like
00:04:08.25 something out of a movie, in fact it could be any kind
00:04:14.16 of a movie, could be a horror movie or a movie of
00:04:19.29 things on a Martian landscape, where you have
00:04:23.02 these fantastic organelles. In fact, this is a clathrin
00:04:25.23 basket that you see here, where the clathrin coats
00:04:29.10 the membrane and makes the membrane pucker in
00:04:32.24 to bring in components. And in this context, we want
00:04:37.17 to ask the question about what happens to
00:04:40.14 membrane lipids and lipid-anchored proteins when
00:04:42.22 they get taken up by the cell membrane. GPI-
00:04:49.00 anchored proteins, as we had discussed in the first
00:04:52.16 part, are an example of components of the
00:04:55.05 membrane that are in fact some of the original
00:04:59.09 reasons why the raft hypothesis was proposed,
00:05:01.27 because they are sorted to different surfaces of the
00:05:05.08 epithelial cell, the apical surface of an epithelial cell,
00:05:11.00 as opposed to being delivered to the basolateral. In
00:05:14.19 the endocytic process, we find that the GPI-
00:05:17.27 anchored proteins are also sorted. So at the cell
00:05:21.14 surface, if we label GPI-anchored protein, here we've
00:05:27.22 expressed a folate receptor. The GPI-anchored
00:05:32.13 proteins come in many stripes and flavors. There are
00:05:35.13 about 300 different types of GPI-anchored proteins.
00:05:39.22 Many of them are receptors for small molecules, for
00:05:44.09 example, the folate receptor is a GPI-anchored
00:05:47.04 protein in all mammalian systems. The folate receptor
00:05:52.22 brings in folates, and we can tag the folate receptor
00:05:56.25 with a small molecule ligand, the folate ligand,
00:06:00.29 attached to a fluorophore, so it becomes easy to
00:06:03.29 follow the folate receptor as it brings in this tagged,
00:06:08.19 fluorescent ligand into the cell. So at the cell surface,
00:06:15.12 if you take cells that express the folate receptor, and
00:06:18.11 we tag it with a small molecule ligand, at the cell
00:06:20.22 surface the folate receptor looks fairly uniformly
00:06:24.17 distributed. And in the same cell, if we incorporate an
00:06:29.16 exogenous phospholipid, now we've fluorescently
00:06:33.13 tagged a phospholipid, an NBD (which is a
00:06:38.01 fluorophore, a green fluorophore) tagged to a
00:06:42.10 sphingolipid, we find that this fluorophore is
00:06:47.01 incorporated in the cell membrane, and when it's
00:06:50.09 endocytosed, the two components of the membrane
00:06:53.09 (the NBD-labeled sphingomyelin, which is
00:06:55.25 incorporated at the cell surface; and the folate
00:06:58.13 receptor), they seem to segregate. The green dots
00:07:02.02 here indicate endosomes that are enriched for the
00:07:05.18 NBD lipid, and the red dots here indicate endosomes
00:07:09.02 that are enriched for the GPI-anchored folate
00:07:11.26 receptor. So this suggests that these two different
00:07:17.12 types of lipids are sorted at the cell surface and
00:07:22.00 brought into endosomes perhaps by distinct
00:07:26.06 mechanisms. In fact, at the cell surface, if we look by
00:07:33.26 an electron micrograph, and this is an electron
00:07:36.07 micrograph from a colleague of mine, Rob Parton,
00:07:39.26 from the University of Queensland, where he shows
00:07:42.21 that the GPI-anchored proteins are in fact coming
00:07:47.15 into the cell where these long, tubular invaginations
00:07:52.08 that have concentrated GPI-anchored proteins in
00:07:56.00 them. And work that we have done over the years
00:08:01.05 has revealed to us that GPI-anchored proteins are
00:08:05.14 endocytosed via a specialized endocytic pathway
00:08:09.09 that's called the "GEEC" pathway, for the "GPI-
00:08:12.17 anchored protein-enriched endosomal
00:08:15.09 compartments" that they make. And these pathways are in fact a
00:08:21.00 main means by which the cell seems to bring in a lot
00:08:24.21 of the fluid that is also endocytosed into cells. This
00:08:29.08 pathway requires the small molecule GTPase,
00:08:33.20 Cdc42, but does not seem to require a host of other
00:08:39.03 components that are necessary for the clathrin- and
00:08:43.10 dynamin-dependent endocytic machinery. Instead,
00:08:48.17 this pathway, which is a pinocytic pathway (bringing
00:08:51.10 in fluid into the cell as well), is rather lipid selective,
00:08:56.04 it's lipid dependent, it requires the small molecule
00:09:00.25 GTPase, and is in fact very sensitive to perturbations
00:09:07.18 of the actin cytoskeleton. But putting these
00:09:13.19 characteristics of a GPI-anchored protein together,
00:09:16.17 the fact that it's sorted into a specialized endocytic
00:09:21.07 pathway at the cell surface, while it lacks any kind of
00:09:24.13 cytoplasmic extension, is telling us something. It's
00:09:29.23 telling us that these components must be somehow
00:09:33.15 segregated in the plane of the membrane to form
00:09:37.02 regions or domains where these molecules can be
00:09:40.07 selectively endocytosed. So in other words, I think
00:09:44.28 it's providing us an opportunity to look for regions of
00:09:49.10 the cell membrane that could be enriched in these
00:09:52.18 components, which by all accounts, would
00:09:56.10 correspond to the functional definition of a raft:
00:10:02.02 regions of membranes imbued with function. So, this
00:10:08.10 is exactly what we've been trying to investigate. So
00:10:11.18 first of all, we know from work in our lab and other
00:10:15.27 laboratories that membrane rafts, if they exist, must
00:10:23.16 consist of some kind of a clustering of cholesterol
00:10:25.29 and sphingolipids; and that they must be platforms
00:10:30.05 which allow the segregation of specific proteins, so
00:10:34.12 that they can be sorted or generate signaling
00:10:39.13 functions in specific lipid contexts. So how do we
00:10:44.09 look for these rafts, so these components of
00:10:46.25 membrane which are functional? Well, one way that
00:10:52.19 people have looked for them are by this notion of this
00:10:56.11 detergent-resistant extraction method, the DRM
00:11:00.12 measurement. And as we discussed in the Part 1,
00:11:04.04 there are a lot of problems associated with that
00:11:06.06 operational definition, so for the purpose of this
00:11:10.25 presentation, I think we should ignore that criteria.
00:11:15.21 Other measurements that have been done to
00:11:19.25 investigate the segregation of components in cell
00:11:23.03 membranes are to ask questions about the diffusional
00:11:26.03 characteristics of molecules in the plane of the
00:11:28.05 membrane. And if molecules enter regions where the
00:11:33.23 local environment of the membrane is more liquid-
00:11:36.29 ordered, for example, one would expect that
00:11:39.26 components that are moving into these regions
00:11:43.29 would slow down in their mobility, and this is another
00:11:50.07 means to investigate whether there are regions in cell
00:11:53.07 membranes where there are lateral segregations of
00:11:57.00 components contributing to an ordering of lipids in
00:12:02.01 the membrane. We, in our laboratory, have not
00:12:06.24 looked at these diffusion measurements, but I'll come
00:12:09.16 back to them in Part 3 of this seminar. A method that
00:12:15.21 we have used is ones based on trying to investigate
00:12:19.22 the proximity between components in the cell
00:12:22.13 membrane and ask questions about can we observe
00:12:27.15 a coming together of these lipid molecules in
00:12:32.08 membranes of living cells when they are being
00:12:36.29 endocytosed. So to do that, we need to step back
00:12:45.03 and have, as I said, a proper probe. So these GPI-
00:12:49.21 anchored proteins is, by all accounts, a good way to
00:12:55.09 investigate the characteristics of segregation in cell
00:12:58.29 membrane, because they are sorted in the plane of
00:13:01.16 the cell membrane, and they are endocytosed by
00:13:04.05 these specialized pathways. So do they exhibit any
00:13:08.23 special segregated distribution in the plane of the
00:13:12.07 membrane? And if they do, what is the scale of those
00:13:18.09 segregated domains? So the first question we asked
00:13:21.12 was simply, by tagging the GPI-anchored proteins
00:13:26.19 with a fluorescent dye, again the folate receptor, we
00:13:31.01 tagged it with fluorescent folate and imaged it using a
00:13:36.05 confocal microscope, and what we found was that
00:13:39.28 the molecules look rather boring. They look as if
00:13:43.18 they're distributed uniformly on the surface of the cell,
00:13:47.05 with some small speckles of these components that
00:13:51.17 in fact correspond to little microvilli, which you can
00:13:54.22 see on the sides of these cells. So this doesn't look
00:13:58.12 like some big raft, or big segregated domain, in the
00:14:04.25 membrane which could correspond to a sorted site
00:14:09.05 for endocytosis. So the molecules at the scale of the
00:14:13.12 fluorescence micrograph, which is at the scale of
00:14:16.29 optical resolution, could be dispersed in a random
00:14:21.14 fashion, or they could be dispersed in a fashion
00:14:24.28 where the molecules are present as very small
00:14:28.12 aggregates in the cell membrane, way below the
00:14:31.00 resolution of light. Another analysis that we did was a
00:14:35.07 crude sort of immunogold electron micrographic
00:14:37.25 analysis, which also didn't reveal any large-scale
00:14:40.20 clustering of these proteins at the surface of the cell
00:14:45.02 membranes. The gold particles that you see are in
00:14:48.22 these golden arrows in fact look rather uniformly
00:14:53.04 dispersed on the surface of the cell. But below the
00:14:59.09 resolution of light, one could in fact have some
00:15:04.22 aggregation of these components taking place, and
00:15:07.15 to investigate that, different investigators have used
00:15:13.06 a variety of tools. So electron micrographic analysis
00:15:16.24 coupled with some statistical analysis of the
00:15:19.18 distribution of how the particles in the membrane has
00:15:23.19 been one very successful approach that Rob Parton
00:15:26.27 and John Hancock have pioneered. Chemical
00:15:30.13 crosslinking of components in situ is another
00:15:34.02 approach that Kurzchalia and Chiara Zurzolo have
00:15:41.08 utilized very effectively. And we, on the other hand,
00:15:45.06 have tried to use sort of a nonperturbing approach,
00:15:50.22 where we try and observe the proximity between
00:15:54.17 different GPI-linked proteins using Förster's
00:15:58.06 resonance energy transfer, or in other words, FRET,
00:16:03.28 which is a technique that allows you to look at
00:16:07.23 proximity between two fluorophores, and I will explain
00:16:11.10 some of the tenets of this technique in the next few
00:16:15.25 slides. So the FRET is a process in which energy is
00:16:21.02 transferred from a donor fluorophore that is excited by
00:16:27.27 an excitation light, and the energy is transferred to an
00:16:31.08 acceptor fluorophore, and that transfer of energy
00:16:35.14 results in the acceptor fluorophore emitting light of a
00:16:40.05 different wavelength. And the efficiency of this
00:16:44.25 transfer can be measured by many means, and the
00:16:48.27 efficiency of this transfer is exquisitely sensitive to the
00:16:52.22 distance between the two fluorophores. In fact, the
00:16:58.17 efficiency of transfer is sensitive to the sixth power of
00:17:03.21 the distance between between the fluorophores. So
00:17:05.11 if the fluorophore distances change by a factor of
00:17:10.21 25%, the efficiency of energy transfer decreases by a
00:17:17.08 factor of eight, so you have a very steep
00:17:19.21 dependence of the energy transfer efficiency with
00:17:23.08 the distance between the fluorophores. And the
00:17:26.02 scale that is used in this measurement is a
00:17:28.29 characteristic of the two chromophores that are
00:17:32.02 engaged in this energy transfer process. The scale is
00:17:36.04 called the Förster's radius, which is a characteristic of
00:17:41.14 the overlap of the donor and acceptor fluorophores,
00:17:47.17 and in fact serves as kind of a spectroscopic ruler to
00:17:54.17 measure distances in the nanometer scale in any
00:17:59.03 fluorescence experiment. We have in fact used a
00:18:05.24 variant of this type of FRET process to monitor the
00:18:11.04 distance between fluorophores. This variant is called
00:18:15.19 "homo-FRET," where we look at the energy transfer
00:18:19.06 between two like fluorophores. And the way one can
00:18:23.04 monitor the energy transfer between these like
00:18:26.15 fluorophores is by measuring the emission anisotropy,
00:18:31.25 or measuring the polarization of fluorescence
00:18:35.12 emission emitted by the donor and acceptor
00:18:40.14 fluorophores. I'll explain this in a little detail. So if you
00:18:47.15 have a fluorophore, for example, GFP (green
00:18:50.23 fluorescent protein), and excite it with polarized light,
00:18:55.22 the fluorophore is going to emit light in a polarized
00:19:02.12 fashion, and if you monitor the emission of
00:19:07.08 fluorescence, you will monitor a fairly polarized
00:19:11.19 emission of fluorescence, because GFP, being a
00:19:13.24 large molecule, doesn't allow the fluorophore in its
00:19:16.24 excited state to rotate very far before it emits. So the
00:19:20.14 only depolarization that one monitors is the
00:19:25.24 depolarization due to the arrangement of these GFP
00:19:30.20 fluorophores. If GFP was a very small molecule and
00:19:34.27 rotating very rapidly, the emission anisotropy that was
00:19:40.21 collected from the GFP fluorescence would be highly
00:19:44.02 depolarized. So rotation can depolarize the
00:19:47.28 fluorescence emission, and FRET can also
00:19:52.21 depolarize the fluorescence emission. So there are
00:19:56.19 two GFP fluorophores very close together, they can
00:20:00.13 transfer energy to each other, and in that transfer
00:20:02.14 process, the orientation information of the initial
00:20:10.15 excitation energy is lost, and the energy transfer
00:20:14.21 process itself leads to a depolarization of the
00:20:18.02 fluorescence from fluorophores that are engaged in
00:20:23.04 close proximity. So, by monitoring the extent of
00:20:28.08 depolarization due to these two different types of
00:20:33.01 processes (rotational and energy transfer), if one can
00:20:37.09 isolate these two different processes of
00:20:40.28 depolarization, one ends up with a very sensitive measure of the
00:20:47.23 FRET phenomena. So we've used this technique,
00:20:52.19 which is monitoring the anisotropy fluorescence from
00:20:57.04 these fluorescently tagged GPI-anchored proteins, to
00:21:02.15 ask whether they are present in close proximity to
00:21:06.13 their neighbors. So for this type of experiment, we
00:21:11.23 take Chinese hamster ovary cells and transfect them
00:21:16.07 with a fluorescent GPI-anchored protein (for example,
00:21:26.19 GFP tagged to a GPI-anchoring signal, or a
00:21:31.07 monomeric YFP tagged to a GPI-anchoring signal, or
00:21:35.10 our favorite molecule, the folate receptor) and then
00:21:40.26 excite these fluorophores with polarized light, and
00:21:44.13 then monitor the anisotropy of fluorescence emission.
00:21:47.18 So the first thing that we observe that the anisotropy
00:21:51.19 of the emission is uniform across a field of cells,
00:21:57.27 which have very widely varying fluorescence
00:22:04.06 intensity, indicating widely varying protein levels in
00:22:09.07 the cell membranes. So across a range of protein
00:22:13.11 expression in a cell membrane, we monitor a value of
00:22:16.06 anisotropy from different cell populations, and this
00:22:22.26 anisotropy is quite depolarized for the different GPI-
00:22:28.21 anchored proteins that we have transfected into
00:22:31.16 cells. And in fact that depolarization is sensitive to
00:22:41.22 the perturbation of cholesterol in the cell membrane,
00:22:44.08 so if you remove cholesterol from the cell membrane
00:22:46.21 using a variety of different means, we see a rise in
00:22:50.08 the fluorescence anisotropy, indicating that we are
00:22:54.01 somehow losing the FRET component of this
00:22:57.07 fluorescence anisotropy measurement. And the
00:23:01.19 fluorescence anisotropy now rises to a value which
00:23:04.19 corresponds to isolated fluorophores that are present
00:23:09.01 in cell membranes. So this sort of an experiment
00:23:14.00 immediately tells us that the GPI-anchored proteins
00:23:17.06 cannot be organized in a random arrangement in cell
00:23:21.19 membranes where they do not see their neighbors. In
00:23:24.16 fact, they must be organized in an arrangement
00:23:27.18 where there are optically unresolvable clusters of
00:23:33.07 proteins where this energy transfer process must be
00:23:36.01 taking place. And what we have been able to show
00:23:40.04 is that this FRET process arises from an extremely
00:23:46.07 high-density packing of these fluorescently tagged
00:23:51.00 GPI-anchored proteins in membranes of cells. This
00:23:54.25 high-density packing is something that we have been
00:23:58.01 able to uncover using a time-resolved fluorescence
00:24:01.23 anisotropy measurement. The details of these
00:24:06.13 measurements are quite complex, and I think they
00:24:11.24 are perhaps better referred to papers from our
00:24:16.23 laboratory. But what we've been able to show using
00:24:21.02 these time-resolved fluorescence anisotropy
00:24:23.04 measurements is that the GPI-anchored proteins, at
00:24:28.06 least the species that exhibit the FRET or close
00:24:31.17 proximity, are present in incredibly close distances
00:24:37.05 between each other. They're present in around 4
00:24:41.06 nanometer separation between each other. The two
00:24:43.20 GPI-anchored proteins, in this case let's say two GFP
00:24:47.18 molecules, which are 3 nanometers in size on their
00:24:50.18 own, are within 4 nanometers of each other. The
00:24:54.16 centers of these proteins are within 4 nanometers of
00:24:56.29 each other, so they're almost close packed, they're
00:24:59.06 almost touching. But that still doesn't tell us what size
00:25:06.18 these structures of these FRET-competent species
00:25:11.07 are. What is the size of their structures, are they a
00:25:15.05 few nanometers, are they tens of molecules, or are
00:25:18.22 they hundreds of molecules? Because all of that will
00:25:20.29 still fit within the optical resolution of a microscope,
00:25:26.07 which is about 300 nanometers. So some clues to
00:25:30.19 understanding the structure of these FRET-
00:25:35.06 competent species in cell membranes came in fact
00:25:38.15 from another type of experiment, where we were
00:25:41.15 monitoring the fluorescence anisotropy of these
00:25:44.21 folate receptors labeled with fluorescein tagged to
00:25:49.19 the folate ligand, and when we monitored the
00:25:53.20 fluorescence anisotropy of these folate receptors, the
00:25:59.09 fluorophores all seemed to get bleached, and when
00:26:02.03 they bleach, their fluorescence anisotropy simply
00:26:04.29 rises up the scale in very systematic steps, so if you
00:26:10.07 bleach the fluorophores by 50%, the fluorescence
00:26:13.12 anisotropy rises by a given value. If you bleach it
00:26:17.10 again by 50%, it changes by another systematic
00:26:20.27 value. And in fact our interpretation is that if you
00:26:25.29 have small clusters of these molecules, when they
00:26:29.07 bleach, within the clusters you see some change in
00:26:34.09 the distances between these different components.
00:26:39.06 So what we've tried to do is try and make theoretical
00:26:44.08 models for these structures. We've made two types
00:26:51.07 of models at two different scales. One at the scale
00:26:54.26 where the domains that these molecules form are
00:26:59.14 much larger than the Förster's radius, which is the
00:27:01.29 scale of our fluorescence energy transfer experiment.
00:27:06.05 And the other model is a model where the
00:27:09.28 fluorophores are distributed in a much smaller scale,
00:27:14.04 on the scale of the Förster's radius. But first let's look
00:27:17.21 at the model where we've looked at the change of
00:27:22.07 fluorescence anisotropy in a model where the
00:27:28.06 fluorophores are distributed in domains which are
00:27:30.16 much larger than the Förster's radius. So what we've
00:27:34.28 done is we've created a situation where we've
00:27:42.16 arranged fluorophores in these arrangements, and
00:27:45.25 then bleached them, and in each of these conditions
00:27:49.06 we've calculated the extent of anisotropy change
00:27:54.06 we'd expect from fluorophores that were in FRET
00:27:58.05 distances from each other. And what we find is in
00:28:01.16 fact that none of the model that have a size that's
00:28:09.12 much larger than the Förster's radius are able to
00:28:13.04 describe the data that we have generated from this
00:28:16.29 photobleaching perturbation of anisotropy. So for
00:28:21.06 example, if we look at the change in fluorescence
00:28:24.26 intensity as a function of the bleaching process, as
00:28:28.29 you move left across this screen, what we find is that
00:28:35.10 the fluorescence anisotropy creeps up the scale, and
00:28:40.03 the different models that we have for domains that
00:28:43.29 are much larger than the Förster's radius fail to fit the
00:28:49.09 observed experimental values of anisotropy. In fact,
00:28:53.13 the only model that fits this observed change is a
00:28:59.13 model where the density of the fluorophores in the
00:29:02.24 domains is very, very low. But the measured density
00:29:07.15 that we have for these fluorophores in these domains
00:29:11.24 is in fact that of molecular separations on the order of
00:29:15.05 4 nanometers, and if you put those numbers in, we
00:29:21.08 fail to fit the data that we get from the experiment.
00:29:25.06 But the model that best fits the experiment is a model
00:29:28.08 that looks like this: Where we have very tiny clusters,
00:29:32.26 three or four or maybe five molecules in size, which in
00:29:37.11 fact are present in a low abundance in the context of
00:29:45.06 a large number of monomers. So in this type of an
00:29:49.10 experiment, we find that the theoretical model
00:29:53.28 perfectly fits the experimental data, and using some
00:29:58.26 of the fit parameters, we have been able to come up
00:30:01.24 with the structure of these FRET-competent species
00:30:05.00 in the cell membrane, and it looks like this, that these
00:30:08.06 GPI-anchored proteins are arranged predominantly in
00:30:13.21 the form of monomers, and a small fraction between
00:30:16.13 20-40% are present as these little nanoclusters. Not
00:30:21.20 only that, experiments using this sort of a technique, I
00:30:26.00 think, have allowed us to ask questions about the
00:30:31.14 properties of these nanoclusters. One thing that we
00:30:36.05 have been able to show is that multiple GPI-
00:30:42.02 anchored species can cohabit one nanocluster. So if
00:30:46.12 you have many different GPI-anchored proteins, if
00:30:50.28 you monitor the fluorescence anisotropy for one of
00:30:53.28 these species, the fluorescence anisotropy in fact is
00:31:00.28 perturbed by the presence of the other species, and
00:31:04.20 that indicates that multiple or many different GPI
00:31:08.18 species can cohabit in the same little nanocluster. So
00:31:14.04 the GPI-anchored proteins in cell membranes then
00:31:17.21 are organized as monomers and mixed nanoclusters
00:31:22.06 in cell membranes. Another feature of this
00:31:26.22 organization is that it is flexible. If we selectively
00:31:33.15 perturb one of the GPI-anchored proteins by
00:31:37.00 crosslinking them with antibodies for example, the
00:31:41.27 other GPI-anchored proteins which do not get
00:31:45.23 crosslinked by an antibody (which is, let's say, able to
00:31:50.16 recognize the green GPI-anchored proteins versus
00:31:53.02 the red ones), the red ones now reorganize to a
00:32:00.04 distribution that would have corresponded to a
00:32:06.24 situation where they do not see the green ones. So
00:32:10.10 the GPI-anchored proteins seem to be flexibly
00:32:14.11 organized, and their organization can be perturbed or
00:32:17.14 changed and altered when you do something to
00:32:22.07 these proteins that moves them laterally away from
00:32:26.08 these preexisting structures. So they are, in other
00:32:32.00 words, flexibly organized as monomers and mixed
00:32:35.22 nanoclusters. They are concentration independent,
00:32:41.19 and I think this is a very important property of these
00:32:44.01 GPI-anchored proteins, that if you look at different
00:32:48.29 concentrations of GPI-anchored proteins in cell
00:32:51.10 membranes, the fraction of clusters that are present
00:32:55.01 is independent of the concentration of these GPI-
00:32:57.25 anchored proteins in the cell membrane. They're
00:32:59.21 cholesterol sensitive. And I think, given these
00:33:04.18 properties, people who are not interested in GPI-
00:33:07.13 anchored proteins might say, "Well, so what? You're
00:33:10.10 looking at GPI-anchored proteins and what you're
00:33:12.13 seeing is a characteristic of these molecules, and
00:33:15.08 maybe there's something special about GPI-
00:33:17.03 anchored proteins." But I think what I'd like to show
00:33:19.13 you now is some experiments that have been done
00:33:22.17 on other lipid-anchored molecules, for example the
00:33:26.18 Ras protein. The Ras is a inner leaflet lipid-anchored
00:33:32.05 molecule, and the Ras molecule is one of the key
00:33:36.19 players in oncogenic transformation of cells, and the
00:33:42.08 way it interprets growth factor signals from the
00:33:44.27 outside determines, in many cases, the ability of a
00:33:49.09 cell to proliferate or continue to grow or stop growing.
00:33:57.09 And the organization of that Ras molecule has been
00:34:02.22 a matter of great interest and contention, because it
00:34:05.29 seems to dictate the outcome of the signaling
00:34:09.00 process through growth factor signaling. And from
00:34:14.29 some experiments that John Hancock and his
00:34:17.29 colleagues have done, they have been able to show
00:34:21.13 that the Ras proteins are also clustered in the inner
00:34:25.12 leaflet of cells, in small, <10 nanometer clusters, 00:34:31.29 which are cholesterol dependent, concentration 00:34:34.12 independent, and in fact sensitive to perturbations of 00:34:38.03 the actin cytoskeleton. And the organization of these 00:34:41.17 components in the membrane are important for the 00:34:45.13 kind of oncogenic signaling that they give rise to. So 00:34:50.20 it is not only the Ras proteins that are present in 00:34:54.01 these small, nanocluster-type organizations. In fact, 00:34:59.06 the Ras proteins have told us that perhaps if you 00:35:03.23 crosslink GPI-anchored proteins, the organization of 00:35:08.03 the Ras proteins seem to be influenced by the 00:35:10.22 crosslinked GPI-anchored proteins. If you crosslink 00:35:13.07 GPI-anchored proteins, the clusters of the Ras 00:35:18.23 proteins seem to come and sit beneath these 00:35:21.10 clustered GPI-anchored flotilla that are present on 00:35:26.03 the exoplasmic side of the cell membrane. And this in 00:35:30.04 some ways seems to be revealing a coupling of the 00:35:33.22 outer leaflet of the bilayer, where the GPI-anchored 00:35:36.00 proteins are sticking their heads out, with the inner 00:35:38.12 leaflet of the bilayer, where the Ras molecules are 00:35:42.04 situated. Other glycolipids, which are lipids that have 00:35:49.20 different sorts of gangliosides or sialyl groups on 00:35:54.13 them, also exhibit a similar nanoclustered distribution, 00:35:58.11 which is sensitive to cholesterol and is concentration 00:36:02.29 independent. So together, it seems that multiple lipid- 00:36:07.06 anchored proteins and different types of lipids are 00:36:12.19 organized as monomers and nanoclusters, and this 00:36:16.25 sort of organization seems to be a very general 00:36:21.07 characteristic of membrane components in living 00:36:25.00 cells. So where does this leave us in terms of the 00:36:29.01 endocytosis of GPI-anchored proteins? So 00:36:33.07 specifically in the context of GPI-anchored proteins, 00:36:36.18 what we have tried to do is ask, what role do these 00:36:39.28 nanoclusters play in the endocytosis of these GPI- 00:36:43.27 anchored proteins. And what we find, and this is 00:36:48.08 experiments that we've done by perturbing the 00:36:51.19 organization of these GPI-anchored proteins by this 00:36:54.13 crosslinking trick, when we crosslink the GPI- 00:36:58.20 anchored proteins, we find that the red GPI- 00:37:02.10 anchored proteins (which are not crosslinked) 00:37:05.20 continue to be endocytosed by the distinct endocytic 00:37:08.24 pathway that GPI-anchored proteins seems to 00:37:10.26 marking, whereas the crosslinked GPI-anchored 00:37:13.06 proteins are segregated from this endocytic pathway 00:37:17.03 and move into other pathways of endocytosis. So it 00:37:21.04 appears the little nanoclusters of GPI-anchored 00:37:24.19 proteins act as sorting signals or signals for inclusion 00:37:31.29 and perhaps signals for exclusion from different 00:37:35.06 pathways of endocytosis. And this in some ways 00:37:42.12 brings us to formulate some kind of a hypothesis 00:37:46.08 about what kinds of arrangements these GPI- 00:37:49.22 anchored proteins may be getting into to mark 00:37:56.21 different endocytic routes or maybe generate new 00:38:00.11 endocytic routes. So one hypothesis that we have is 00:38:03.18 that these little nanoclusters of GPI-anchored 00:38:06.10 proteins may on their own not have very much 00:38:10.22 functional significance, but when they are brought 00:38:13.23 together and form larger-scale domains where many 00:38:17.28 of these nanoclusters come together and form little 00:38:22.27 rafts of nanoclusters, if you want to call them that, 00:38:25.18 those structures are functional entities which could 00:38:29.20 be endocytosed by these distinct endocytic 00:38:34.05 mechanisms. Well, the evidence that we have at the 00:38:39.21 end of this part of the lecture is that we have strong 00:38:48.07 proof that these nanoclusters of these lipid-anchored 00:38:51.06 molecules exist. Whether they form larger domains or 00:38:55.19 not remains to be seen. So in a way, the study of 00:39:03.29 these GPI-anchored proteins, lipid-anchored 00:39:06.19 components in the outer and the inner leaflet, have 00:39:10.26 revealed to us a working hypothesis of what could 00:39:15.00 be a functional domain in cell membranes. One 00:39:19.23 where, I'd imagine, that small clusters of GPI- 00:39:23.16 anchored proteins are somehow brought together or 00:39:27.20 induced into a larger domain, which is then capable 00:39:31.11 of specific sorting or signaling function. And this is a 00:39:37.09 very different model from the model of phase 00:39:41.04 segregation of components giving rise to a raft, 00:39:44.24 because it suggests that there are mechanisms 00:39:47.27 involved in maintaining a preexisting structure, 00:39:52.03 mechanisms involved in generating a new type of 00:39:55.06 structure which has specific function. And the 00:40:01.14 characteristics of the structures of GPI-anchored 00:40:05.12 proteins, I think, and these other lipid-anchored 00:40:09.10 molecules suggest a very different type of 00:40:14.18 mechanism to generate these structures, different 00:40:18.16 from an equilibrated phase separation-type 00:40:23.21 mechanism, where we'd expect the concentration of 00:40:27.16 the clusters to be sensitive in some form to the 00:40:30.27 concentration of the monomers in the membrane. So 00:40:38.21 the questions that these observations raise obviously 00:40:44.12 are: What makes these nanoclusters, and what 00:40:50.01 induces, if at all, the domains that are imbued with 00:40:56.09 function at a larger scale? Well, I think the similarities 00:41:03.18 in properties between all these different lipid- 00:41:06.12 anchored molecules suggest similar mechanisms for 00:41:10.04 the formation of these clusters of GPI-anchored 00:41:13.07 proteins, Ras proteins, and glycolipids. And it is 00:41:17.05 unlikely that a passive mechanism involving 00:41:20.13 thermodynamic phase separation is going to account 00:41:23.08 for these type of structures. Instead, I think we may 00:41:27.04 need a new proposal for understanding the formation 00:41:31.21 of these complexes in cell membranes, a new 00:41:34.13 proposal for the structure of the membrane, to 00:41:37.06 understand the formation of these structures in the 00:41:40.00 cell membrane. And Part 3 is going to try and explore 00:41:47.12 some ideas behind this new proposal for the structure 00:41:51.06 of the plasma membrane. 00:41:54.23

Part 3: Making Rafts in Living Cell Membranes

00:00:03.05 Welcome to this third part of the iBioSeminar on rafts
00:00:08.29 in cell membranes. I'm Satyajit Mayor from the
00:00:14.12 National Centre for Biological Sciences, working in
00:00:19.11 close collaboration with my physicist colleague,
00:00:21.29 Madan Rao, trying to make sense of what we see in
00:00:27.04 living cell membranes. So if you remember, or in case
00:00:33.20 you haven't seen the previous part, what we had
00:00:39.00 observed was that many different lipid-anchored
00:00:43.09 proteins are organized as small nanoclusters in
00:00:47.10 membranes of living cells in a manner that seems
00:00:51.14 consistent with their being formed by some active
00:00:55.12 process, because they exhibit a concentration-
00:00:59.04 independent clustering phenomena in living cell
00:01:06.19 membranes. We had also proposed that these
00:01:12.27 nanoclusters of lipid-anchored molecules may come
00:01:16.23 together and form regions in cell membranes that we
00:01:22.09 believe are functional domains, which can perform
00:01:28.03 sorting and signaling functions under regulatory
00:01:32.29 control of the cell's machinery. Well, the questions
00:01:39.27 that we were left with at the end of the last part were
00:01:46.05 questions about what makes these nanoclusters of
00:01:50.20 these different lipid-anchored molecules in cell
00:01:52.19 membranes, and what induces their forming
00:01:57.04 functional domains. And I think if we think can
00:02:00.00 address those two issues, we would be well on our
00:02:03.17 way to understanding functional domains in
00:02:06.20 membranes of living cells. As you will appreciate,
00:02:12.06 coming to this point has required a development of
00:02:16.12 new technology, new tools, to image processes in
00:02:19.15 living cells at the scales at which they demand.
00:02:24.05 We've had to develop new types of microscopy to
00:02:28.09 be able to study and elucidate the structure of these
00:02:32.29 little nanoclusters in living cell membranes. And in
00:02:36.16 fact, understanding and beginning to observe the
00:02:44.00 formation of these functional, larger-scale domains
00:02:48.12 also has required new developments in microscopy.
00:02:55.23 So what we have been able to do is image the
00:03:05.23 signals that we had observed from these nanoscale
00:03:08.15 clusters, the FRET signals, the fluorescence or
00:03:12.15 Förster's resonance energy transfer signals, from
00:03:16.08 these nanoscale structures, we've begun to image
00:03:18.18 them at much higher resolution, using a high-
00:03:23.03 resolution FRET imaging technique, where we're
00:03:27.01 able to map the FRET, using this "homo-FRET"
00:03:32.02 technique that we had developed in the previous
00:03:36.17 iBioSeminar segment, at much higher spatial
00:03:41.27 resolution. So the spatial resolution that we have
00:03:46.12 now achieved is about 300 nanometer resolution,
00:03:50.24 and at this resolution, if we begin to image the FRET
00:03:55.06 signals coming from these lipid-anchored proteins,
00:03:58.20 specifically GPI-linked proteins at the surface of
00:04:01.25 living cells, we see the following. So in living cell
00:04:07.15 membranes, here is the fluorescence intensity
00:04:11.13 distribution of these GPI-linked molecules at the
00:04:16.03 surface of a mammalian cell (a Chinese hamster
00:04:20.04 ovary cell, in this case), and if you notice, the
00:04:24.10 distribution in most parts looks quite uniform and flat.
00:04:28.20 There are some regions which correspond to
00:04:30.16 microvilli that have slightly higher intensity than the
00:04:35.20 surroundings, and there are other regions which
00:04:38.20 correspond to edges of cells which look rather
00:04:44.25 undifferentiated. But if you now switch our gaze to
00:04:49.24 the FRET signal, and here the code is red for very
00:04:58.07 low FRET and blue for very high FRET
00:05:03.22 (corresponding to high level of anisotropy and low level of
00:05:08.19 anisotropy), the low level of anisotropy, which is the
00:05:12.22 blue signals, indicate regions where there is
00:05:16.06 enormous amount of FRET going on, and the high
00:05:19.09 level of anisotropy indicates where very low levels of
00:05:24.02 FRET is taking place. And what you immediately see
00:05:27.07 is that there are different shades of FRET distributed
00:05:33.07 throughout the cell surface. There is a region of low
00:05:39.18 FRET and a region of high FRET, which correspond
00:05:44.28 to regions where these clusters of these GPI-
00:05:48.17 anchored proteins are enriched or depleted. If you
00:05:53.25 look more closely, for example at the microvilli, we
00:05:58.07 find that the microvilli are regions where the clusters
00:06:03.25 of the GPI-anchored protein seem to be specifically
00:06:06.28 depleted. The same seems to be the case at tips of
00:06:11.21 lamellipodia, so tips of lamellipodia we find that these
00:06:17.01 GPI-anchored proteins also look quite heavily
00:06:21.05 depleted. So these are edges of cells where the cells
00:06:24.24 begin to extend, protruding actin-driven membranes,
00:06:33.18 where at the very tips you see a very high anisotropy
00:06:38.14 that corresponds to the exclusion of these
00:06:41.17 nanoclusters. In fact, in living cells, when we
00:06:45.26 observe cells performing their natural ruffling
00:06:50.25 functions, we find that the cells seem to be
00:06:58.07 dynamically excluding the clusters from the edges
00:07:01.26 that are ruffling. In fact, ruffling cells, if you look at
00:07:07.17 the arrows in the movie, you find that the arrows
00:07:13.06 indicate moments in time when, in ruffling cells, the
00:07:20.09 ruffles have extended out, and the GPI-anchored
00:07:24.16 protein clusters as detected by the FRET signals are
00:07:29.06 completely depleted from these ruffling edges. Other
00:07:35.20 regions in the cells are where the nanoclusters of
00:07:40.21 these molecules are enriched, and they seem to be
00:07:44.03 enriched specifically in the flat regions of cells, or in
00:07:48.17 regions just behind the tips of lamellipodia, the
00:07:52.21 lamellum, where the distribution of actin is likely to be
00:08:00.06 very different from the tip of the lamellipodia, or in
00:08:06.11 fact even microvilli. So it appears from these sorts of
00:08:11.21 observations that the GPI-anchored protein
00:08:16.00 nanoclusters are, in many cellular environments,
00:08:23.20 correlating with the underlying cytoskeletal matrix. If
00:08:29.12 you just look at the flatscape of a cell, where the
00:08:34.11 membrane is relatively flat at least to the eye of the
00:08:39.26 microscope, we find that there are a patchwork of
00:08:44.08 domains where GPI-anchored proteins are extremely
00:08:48.20 enriched, these very deep blue regions, and GPI-
00:08:51.17 anchored proteins are not as enriched in these
00:08:54.00 slightly yellow regions. But these patchworks are
00:08:57.00 regions which are showing a kind of variegated
00:09:01.26 pattern, which when analyzed by some statistical
00:09:06.28 means, shows a completely nonrandom segregation
00:09:13.03 of clusters. So in other words, it appears that the
00:09:18.10 membrane is not a homogenous, well-mixed
00:09:22.03 membrane where clusters are present, sprinkled
00:09:25.28 uniformly over the surface of the membrane, but in
00:09:28.08 fact, there are regions where clusters are selectively
00:09:30.24 concentrated, and they're selectively concentrated
00:09:34.08 or excluded from regions of membrane that seem to
00:09:39.01 have distinct underlying cytoskeleton. Experiments
00:09:44.02 that we have begun to do suggest that the
00:09:49.15 localization of these nanoclusters correlate with
00:09:53.08 specific types of the cytoskeletal organization,
00:09:57.08 specifically the actin organization. In fact, perturbing
00:10:00.20 the actin cytoskeleton using drugs that can impair
00:10:06.19 the polymerization or prevent the depolymerization of
00:10:10.03 actin dramatically affect the distribution of these
00:10:13.25 nanoclusters in cell membranes. So it seems that we
00:10:21.04 may have something that motivates the formation of
00:10:24.15 these regions in cell membranes where nanoclusters
00:10:27.06 are concentrated. In fact, the regions where
00:10:32.14 nanoclusters are concentrated correlate with regions
00:10:37.13 of actin where the actin cytoskeleton is likely to be
00:10:45.13 arranged in a horizontal fashion beneath the plasma
00:10:47.29 membrane. And regions where these clusters are
00:10:51.07 excluded seem to be correlating with regions where
00:10:54.07 the cytoskeleton is effectively organized vertical to
00:10:59.17 the plane of the membrane. And this may be in some
00:11:05.17 ways a mechanism that the cell uses to regulate the
00:11:10.24 local composition of these clusters in different
00:11:17.05 regions of the cell membrane. And then it is quite
00:11:22.10 likely that the sorting and signaling functions that we
00:11:26.13 have ascribed to different types of membrane
00:11:30.03 domains could be realized through this kind of
00:11:34.09 organization of membrane components in the cell.
00:11:39.02 The experiments that I've just shown you are of
00:11:42.10 course extremely new and perhaps almost "hot off
00:11:51.01 the bench," but they are not unsupported by other
00:11:56.14 literature that has been accruing over the past
00:12:02.22 several years, specifically from work from Ken
00:12:06.15 Jacobson's group, Mike Sheetz's group, and Aki
00:12:10.21 Kusumi's group in Japan, where they have argued
00:12:14.02 now that the membrane of a living cell is in fact
00:12:19.07 supported by, it's sitting on top of, a cytoskeleton.
00:12:23.27 And the cytoskeleton in fact not only, as I indicated
00:12:28.01 to you, influences the local distribution of these
00:12:31.08 nanoclusters in the membrane, it also influences the
00:12:35.28 mobility of molecules on the cell surface. So when
00:12:39.01 Aki Kusumi looks at the movement of a single
00:12:45.17 particle tagged to a fluorophore on the cell
00:12:51.17 membrane, or tagged to a gold particle, if he looks at
00:12:56.07 a lipid or if he looks at a membrane protein, they
00:13:00.02 exhibit sort of a rather characteristic pattern of
00:13:05.15 diffusion when observed at a relatively slow frame
00:13:11.07 rate. When Aki Kusumi looks at the membrane at an
00:13:17.01 extremely fast frame rate, they picture that they
00:13:20.23 observe is quite different. It seems that the particles
00:13:24.21 that move spend times in small confined regions in
00:13:28.19 the cell membrane and then jump to other regions in
00:13:31.16 the cell membrane, as if they were exhibiting some
00:13:35.27 sort of a hop diffusion mechanism. And this type of
00:13:41.27 characteristic is only observed when the frame rate
00:13:45.00 for observation is in fact 40,000 times per second,
00:13:52.00 achievements that are, at the moment, at the edge
00:13:54.10 of technology today. These regions can be mapped
00:13:59.20 out, and when they are identified by computer
00:14:04.10 programs that can detect breaks in the diffusional
00:14:09.03 characteristics within different regions, or can identify
00:14:12.09 these little domains, they show a fairly interesting
00:14:17.13 pattern of adjacent regions that are organized in
00:14:23.05 some fashion. And when, again, Aki Kusumi's group
00:14:29.01 looks at the underside of the plasma membrane, very
00:14:31.29 close to the lipid bilayer, they find that the
00:14:35.29 cytoskeleton seems to be also arranged in a pattern
00:14:40.06 that corresponds to these domains in which these
00:14:43.28 lipids seem to be confined or these proteins seem to
00:14:47.29 be confined during the diffusional process. So,
00:14:55.13 putting these different observations together, it
00:14:58.16 seems as if the membrane of a living cell is not simply
00:15:04.27 the lipid bilayer situated on its own, but it's in fact a
00:15:11.11 bilayer that's in close conversation with the
00:15:14.08 underlying cytoskeleton, and that conversation
00:15:17.27 seems to be generating new structures, rafts, if you
00:15:22.09 want to call them that, or structures where molecules
00:15:26.10 can be confined by the scaffolding of this
00:15:29.04 cytoskeleton for periods of time necessary for
00:15:32.07 engaging signaling output. And we believe that
00:15:39.13 those kinds of organizational forms can also
00:15:47.03 generate budding pits at the cell surface. Of course,
00:15:54.24 these are speculations, and I think it's important to
00:15:59.00 be able to clearly state what we know and what is a
00:16:05.02 proposal. So what we know is that we can see small
00:16:09.06 clusters of these lipid-anchored proteins in
00:16:11.14 membranes whose distribution and whose existence,
00:16:16.18 as well, seems to be very dependent on the
00:16:21.03 presence of the specific organization of the
00:16:25.03 underlying cortical actin, and that it appears a two-
00:16:31.00 way conversation between the lipids in the bilayer
00:16:35.04 and the cortical cytoskeleton. And this obviously
00:16:40.00 implies that the membrane of a living cell is not a
00:16:46.08 passive fluid mosaic model, but in fact it's an
00:16:51.21 extremely active system, which is, I guess, a walking,
00:16:58.24 talking composite between the bilayer and the
00:17:03.04 underlying cortical cytoskeleton which it is
00:17:05.13 associated with, and there are multiple things going
00:17:08.16 on in the membrane that could influence the local
00:17:10.29 organization of components. There's trans-bilayer flip-
00:17:15.14 flop mechanisms, which can change the local
00:17:17.21 composition of lipids. There are protein-protein
00:17:20.10 interactions whose aggregation can again change
00:17:23.24 the nature of the lipids surrounding these protein-
00:17:26.24 protein aggregates. There could be synapses in the
00:17:30.05 membrane, which could bring other cells close to the
00:17:35.03 membrane, which again can influence the local lipid
00:17:38.17 composition. There is, as I think, determining
00:17:42.09 cytoskeletal interactions, cytoplasmic proteins like
00:17:45.29 negatively charged molecules like MARCKS proteins
00:17:49.04 which also can influence the local lipid composition.
00:17:53.05 And finally, the well-characterized clathrin- and
00:17:58.10 caveolae-coated regions of the membrane, I think,
00:18:00.28 represent one example of how proteins in fact, and
00:18:06.03 their dynamics, can influence the local structure and
00:18:09.19 composition in cell membranes. So I think together
00:18:13.29 putting many of the pieces of this puzzle, where we
00:18:18.27 have a bilayer consisting of a variety of different
00:18:23.09 membrane components, it seems that the cell is
00:18:26.26 using special organizing principles from these active
00:18:32.01 polymers of the cytoskeleton to locally control the
00:18:39.10 nature, composition properties, of the plasma
00:18:44.09 membrane in a way that I think is going to keep us
00:18:47.28 busy in trying to understand for many decades to
00:18:51.17 come. But I think this is just the beginning of a new
00:18:54.28 way of looking at the cell membrane, and for all the
00:18:59.24 young researchers out there, I think this is a very
00:19:02.29 important area of research because, although there
00:19:07.09 has been work in this area for such a long time, I
00:19:10.25 think we know very little about how the local
00:19:13.25 composition of the cell membrane is regulated. And
00:19:18.13 of course, what I've described to you today is in fact
00:19:22.20 a proposal, and like many, these proposals are open
00:19:27.27 for debate, challenge, and investigation. And I think
00:19:32.01 this investigation has to be a multidisciplinary one,
00:19:36.16 where we are using tools from every resource
00:19:40.25 possible, we're using tools of physics, of the latest
00:19:45.04 physics, of the physics of active systems; chemistry,
00:19:50.22 where we try to understand how the chemistry of the
00:19:53.20 membrane is coupling to the activity of components
00:19:57.29 in the cell; and also exploring all the tools that the
00:20:03.09 biological system allows us, the tools of genetics and
00:20:09.06 imaging, to bring together and to bring to bear on this
00:20:14.13 fundamental problem of how membranes of living
00:20:17.27 cells are organized. And I'd like to invite all the young
00:20:21.28 researchers out there to this open arena, which I
00:20:29.15 think is just the start of a new adventure in cell
00:20:34.25 biology. Thank you.
00:20:38.14

Videos in this Talk
  • Part 1: What are Membrane Rafts?
    Part 1: What are Membrane Rafts?
    Audience:
    • Student
    • Researcher
  • Part 2: Looking for Functional Rafts in Cell Membranes
    Part 2: Looking for Functional Rafts in Cell Membranes
    Audience:
    • Student
    • Researcher
  • Part 3: Making Rafts in Living Cell Membranes
    Part 3: Making Rafts in Living Cell Membranes
    Audience:
    • Student
    • Researcher
Speaker: Satyajit 'Jitu' Mayor
Total Duration: 1:40:52
Recorded: May 2007
All Talks in Cell Biology

Talk Overview

The plasma membrane demarcates the inside of the living cell from the outside, serving both as the ultimate physical barrier and location where the cell transacts its business with the outside world. The study of the structure of cell membranes continues to be fertile ground for research for many biologists today; the idea of “Membrane Rafts” has only heightened interest in this area. The main focus of this three part series is to explore the concept of “Membrane Rafts”. In the first part, we will examine what is popularly understood by the term “Membrane Raft” and ask why this concept was proposed in the first place. This requires an appreciation of the most dominant picture of a biological membrane, “the fluid-mosaic” model. “Membrane Rafts” were hypothesized in response to this model to account for observations of segregation of membrane components made in cell systems. Different proposals to account for the segregation of membrane components and many methodologies used to identify membrane rafts, including the pervasive detergent insolubility method, are also discussed in this lecture. This part is aimed at a general audience or for an advanced undergraduate-level student.

In this part, I focus on our laboratory’s ongoing preoccupation with rafts where we study the characteristics of a popular “raft-marker”, a class of cell surface lipid-tethered proteins, the GPI-anchored proteins. These lipid-tethered molecules are sorted at the cell surface into a specialized endocytic pathway, suggesting their segregation at the cell surface. Here I describe collaborative interdisciplinary work from my laboratory and that of my physicist colleague, Prof. Madan Rao, where we explore the organization of lipid-tethered proteins in living cells using new biophysical tools developed for this purpose. These studies have compelled us to refine the notion of membrane rafts as functional lipid-assemblies consisting of nanoscale clusters and monomers. This organization appears to be characteristic of many other putative “raft” lipids and lipid-anchored proteins at the inner and outer leaflet of cell membranes.

Speaker Bio

Satyajit 'Jitu' Mayor

Satyajit ‘Jitu’ Mayor

Satyajit “Jitu” Mayor was born in Baroda, India and received his MSc in Chemistry from IIT, Mumbai. He obtained his Ph.D. in Life Sciences from the Rockefeller University, New York, working in the Laboratory of Molecular Parasitology with Prof. George Cross where he studied the biosynthesis of GPI-anchors in the African sleeping sickness parasite, Trypanosoma… Continue Reading

Playlist: Membranes and Organelles

Related Resources

Related Articles

Singer, S.J. & Nicolson, G.L. The fluid mosaic model of the structure of cell membranes. Science 175, 720-31 (1972).

Silver, B.L. The physical chemistry of membranes: An introduction to the structure and dynamics of biological membranes (Allen & Unwin and The Solomon Press, 1985).

Israelachvili, J.N. Intermololecular surface forces (Academic Press, London, 1992).

Yeagle, P.L. The membranes of cells (Academic Press, Inc, San Diego, 1993).

Simons, K. & Ikonen, E. Functional rafts in cell membranes. Nature 387, 569-72 (1997).

Edidin, M. Lipids on the frontier: a century of cell-membrane bilayers. Nat Rev Mol Cell Biol 4, 414-8 (2003).

Fielding, C.J. (ed.) Lipid Rafts and Caveolae: From Membrane Biophysics to Cell Biology (Wiley-VCH Verlag GmbH & Co, Weinheim, 2006).

Associated Reading

London, E. & Brown, D.A. Insolubility of lipids in triton X-100: physical origin and relationship to sphingolipid/cholesterol membrane domains (rafts). Biochim Biophys Acta 1508, 182-95 (2000).

Lichtenberg, D., Goni, F.M. & Heerklotz, H. Detergent-resistant membranes should not be identified with membrane rafts. Trends Biochem Sci 30, 430-6 (2005).

Edidin, M. The state of lipids rafts: From model membranes to cells. Ann Rev Biophys & Biomolec Struct 32, 257-283 (2003).

Simons, K. & Vaz, W.L. Model systems, lipid rafts, and cell membranes. Annu Rev Biophys Biomol Struct 33, 269-95 (2004).

Mayor, S. & Rao, M. Rafts: scale-dependent, active lipid organization at the cell surface. Traffic 5, 231-40 (2004).

Kusumi, A. et al. Paradigm shift of the plasma membrane concept from the two-dimensional continuum fluid to the partitioned fluid: high-speed single-molecule tracking of membrane molecules. Annu Rev Biophys Biomol Struct 34, 351-78 (2005).

Rao, M. & Mayor, S. Use of Forster’s resonance energy transfer microscopy to study lipid rafts. Biochim Biophys Acta 1746, 221-33 (2005).

Hancock, J.F. Lipid rafts: contentious only from simplistic standpoints. Nat Rev Mol Cell Biol 7, 456-62 (2006).

Reader Interactions

Leave a Reply

Your email address will not be published. Required fields are marked *