How do we know the structure of cell membranes (part 5)

The fluid mosaic model

Combining these results (and many others not mentioned), Singer & Nicolson (1972) proposed a model of membrane structure that still holds up. Their model recognized that the main structural component of the membrane (the matrix into which all other components were incorporated) was a lipid bilayer. Most proteins associated with the membrane are embedded within it (Singer & Nicolson called these integral proteins); a smaller fraction of membrane-associate proteins were attached to either the inner or outer surface of the bilayer (referred to as peripheral proteins). The components – phospholipids and proteins – moved around along the surface (either inner or outer) by diffusion. Because proteins embedded in the matrix of lipids created a mosaic of proteins and lipids and the molecules of the membrane moved around each other like molecules of a liquid, Singer & Nicolson referred to their conception of membrane structure as the Fluid Mosaic Model.

Since 1972, specifics of the model have been altered (but I’m not aware of any major changes). In the mid-1970s, Joseph Schlessinger and his colleagues were measuring the diffusion rates of proteins and lipids in the membrane. (Recall that Frye & Edidin already had shown that molecules in the membrane moved around.) Schlessinger and coworkers used a technique known then as fluorescence photobleaching recovery (FRP), but which is now known as fluorescence recovery after photobleaching (FRAP). To use FRAP, fluorescent molecules are attached to other molecules on the surface of the cell – either a particular type of protein that’s common in the cell membrane, for example, or to the phospholipids; you can imagine a cell whose surface is glowing due to the fluorescent molecules. Then, scientists shine a laser at a small spot of the membrane. The energy from the laser excites the fluorescent molecules in the spot exposed to the beam; too much of this will cause the fluorescent molecules to stop fluorescing. They’re bleached, so there is no “glow” in the spot where the laser was aimed. FRAP measures how long it takes for the glow to return to the spot – this provides a measure of how fast unbleached molecules move into the bleached area. One of the interesting results of this work was that when you label proteins with the fluorescent molecules, the fluorescence of the bleached spot never quite returns to previous levels (see graph below). The scientists concluded that this meant not all proteins move – some of the proteins in the bleached area remained there. This wasn’t the case with lipids, so lipids move more freely than (at least some) proteins.

[Figure from Schlessinger and others (1976) PNAS 73(7): 2409-2413; click image for larger view.]

Subsequent research as shown that even membrane proteins that are able to move may still be restricted in some fashion (summarized by Jacobson and his coauthors in 1995). Also, researchers have found that certain types of lipids can aggregate together (along with certain proteins) to form rafts. The components of the raft stay together, but the raft as a unit will move around in the cell membrane (see atomic force micrograph below). Some details of rafts were summarized by Simons & Ikonen in 1997 (a little dated, I know).

[Atomic force micrograph showing membrane rafts “floating” in sea of other lipids. From D.L. Nelson & M.M. Cox (2005) Lehninger’s Principles of Biochemistry, 4^th Edition. WH Freeman & Co, p. 385.]

References:

Frye, L.D. and Edidin, M. (1970) “The rapid intermixing of cell surface antigens after formation of mouse-human heterokaryons.” Journal of Cell Science 7: 319-335

Jacobsen, K., Sheets, E.D., and Simson, R. (1995) “Revisiting the Fluid Mosaic Model of membranes.” Science 268: 1441-2

Schlessinger, J., Koppel, D.E., Axelrod, D., Jacobson, K., Webb, W.W., and Elson, E.L. (1976) “Lateral transport on cell membranes: Mobility of concanavalin A receptors on myoblasts.” PNAS 73(7): 2409-2413 [http://www.pnas.org/content/73/7/2409.full.pdf]

Schlessinger, J., Webb, W.W., Elson, E.L., and Metzger, H. (1976) “Lateral motion and valence of Fc receptors on rat peritoneal mast cells.” Nature 264: 550-2

Schlessinger, J., Axelrod, D., Koppel, D.E., Webb, W.W., and Elson, E.L. (1977) “Lateral transport of a lipid probe and labeled proteins on a cell membrane.” Science 195: 307-9

Simons, K. and Ikonen, E. (1997) “Functional rafts in cell membranes.” Nature 387: 569-72

Singer, S.J. and Nicolson, G.L. (1972) “The fluid mosaic model of the structure of cell membranes.” Science 175: 720-731

How do we know the structure of cell membranes (part 4)

A sea of lipids - evidence for a fluid membrane

While researchers accumulated support for the idea that proteins were embedded in the lipid bilayer and projecting out beyond the inner and outer surfaces, a classic experiment by Frye & Edidin (1970) was adding to our knowledge of what a membrane was like. Citing the ability of animal cell membranes to move (as with the formation of pseudopods by the amoebas that you might remember from junior high biology class), Frye & Edidin hypothesized that cell membranes must be fluid – the components of the membrane must have some freedom of movement relative to other components.

To test this, they fused mouse and human cells together to form cell hybrids and heterokaryons. Each type of cell has distinct molecules on its surface (called surface antigens) that would identify it as being either a mouse cell or a human cell. Scientists can make antibodies to bind to specific antigens. Frye & Edidin attached fluorescent molecules to the antibodies so that they could track where the antibodies were (one color for antibodies that bound to mouse antigens, and another color for antibodies that bound to human antigens). Since the antibodies would bind to the antigens, knowing where the antibodies were also told them where the antigens were. Right after fusion of the two cells, the antigens were found only on half the cell.

To help visualize this, imagine the blue circle below is a human cell. If scientists attached antibodies to antigens on the cell's surface and attached a blue molecule to the antibodies, then the human cell would look blue. The presence of the blue color indicates that human antigens are also present. The same would be true for a mouse cell, except the scientists attach a red molecule to the antibodies bound to mouse antigens. Right after fusing the human cell and the mouse cell, you'd have something that looked like the image on the right: all the human antigens in the membrane that came from the human cell and all the mouse antigens in the membrane that came from the mouse cell.

But over time (it took about 40 minutes at 37 °C) the antigens from the mouse half would spread out across the whole cell and the human antigens did the same. Frye & Edidin ended up with a mosaic of mouse and human antigens over the surface of the heterokaryon (shown on the far right, below). One likely explanation is that the antigens in the membrane moved, producing the mosaic pattern that they observed.

One of the things that makes this a classic experiment is that instead of simply concluding, “Yep, our initial hypothesis was right,” Frye & Edidin considered other explanations for how these mosaics had formed. They offered four possibilities: (1) the surface antigens of the hybrid cell were being rapidly removed from the membrane and then replaced by newly-made copies of the antigen molecule, (2) new copies of the antigens were added to the membrane from a pre-existing pool of antigen molecules stored in the cell, (3) the antigen molecules moved (by diffusion) from place to place within the membrane [this was their favored hypothesis], and (4) antigen molecules were taken out of the membrane, into the center of the cell, and then put back into the membrane somewhere else.

They excluded the other possibilities by altering the conditions of the experiments and seeing what happened. For example, if explanation (1) is correct, then new antigen molecules must be built inside the fused cell. Frye & Edidin added a chemical to a batch of the fused cells that would prevent the synthesis of new antigen molecules. This should have prevented the mosaic pattern from forming, but it formed anyway. This suggests that explanation (1) is incorrect. By eliminating possibilities, you become more confident in the possibilities that remain. This is a hallmark of good science: test multiple, competing hypotheses.

At the end of the day, what we learned from Frye & Edidin's experiment is that the cell membrane is fluid - the parts move around, they're not fixed in place. In a future post, I'll describe how all of the results described so far were pulled together to produce one coherent model of what cell membranes are like.

References:

Frye, L.D. and Edidin, M. (1970) “The rapid intermixing of cell surface antigens after formation of mouse-human heterokaryons.” Journal of Cell Science 7: 319-335

How do we know the structure of cell membranes (part 3)

Two broad models for the relationship between membrane proteins and the lipid bilayer

As I noted in an earlier post, cell membranes are composed mostly of lipids (primarily phospholipids) and some protein. The question is how are the proteins arranged relative to the bilayer of phospholipids? Over the years, there were two broad models of membrane structure. In a 1969 review, Stoeckenius and Engelman identified them as the Danielli (or bilayer) model and the Subunit model. The Danielli model was an extension of the lipid bilayer hypothesis proposed by Gorter & Grendel. The model attempted to incorporate membrane proteins into the idea of a lipid bilayer by proposing that there would be layers of protein associated with the inside and outside of the bilayers (Danielli and Davson, 1935). This was sometimes referred to as the “protein sandwich” model, with the lipid bilayer sandwiched between two layers of protein. In the figure below (taken from Danielli & Davson’s 1935 paper), the spheres represent proteins and the rectangles topped with half circles are phospholipids. I’m not really that clear on Danielli’s evidence for this model – he refers to other research he conducted saying, “…in a number of egg cells proteins are present of such a surface activity such that an adsorbed layer [of proteins] is bound to be present…”

[Click figure for larger version]

The subunit model seems to have grown out of discoveries related to the protein coats of viruses. According to Stoeckenius & Engelman, the exact nature of the subunit was never clearly defined. Apparently, it was assumed to be some combination of lipids and proteins stuck together to form repeating modular structures, so they were referred to as lipoprotein subunits. It’s not clear to me that the subunit model was incompatible with the idea of a lipid bilayer, but it probably was incompatible with the notion of a protein sandwich. One example of a subunit model (shown below) was given by Benson (1966). Notice that his model has a bilayer with a protein woven in amongst the lipids. Perhaps I’m projecting onto this something that Benson did not intend, but he seems as close to the currently accepted structure as Danielli was. At least he recognized that the proteins were in the bilayer. On the other hand, the idea that membranes were composed of self-assembling lipoprotein subunits was incorrect.


[A portion of Figure 4 from Benson (1966) Journal of the American Oil Chemists’ Society 42: 265-270; click figure for larger image.]


From a sandwich to a mosaic

By the mid-1960s, a number of researchers were clearly unsatisfied with the protein sandwich. In 1966, the Proceedings of the National Academy of Sciences (USA) published similar studies by Wallach & Zahler and Lenard & Singer. Both studies used spectroscopic methods (including infrared spectra, fluorescence spectra and optical rotator dispersion) to examine the shape of proteins associated with cell membranes. Without getting in over my head, the basic idea of all spectroscopic methods is to shine electromagnetic radiation (such as visible light or infrared radiation) at something and analyze the pattern of radiation absorbed or transmitted. Different types of radiation provide different ways of “seeing” a molecule. In this case, the scientists found particular structures in the proteins that suggested extensive interaction between hydrophobic parts of proteins with the hydrophobic parts of lipids in the bilayer. This implied that parts of proteins were in the bilayer. Both groups of researchers explicitly recognized that this contradicted the “protein sandwich” model of membrane structure. Continuing spectroscopic work (Glaser and colleagues, 1970) led to the suggestion that most of the lipids in a membrane weren’t interacting with proteins. As a result, the researchers envisioned membranes as having “a mosaic pattern…In this scheme, globular protein molecules…are interspersed in a matrix consisting of the remaining lipids in a form similar to that of a discontinuous bilayer.” They provided a schematic drawing of this vision (see below). Today, we might summarize this by saying that islands of protein float in a sea of lipids.

[From Glaser and others (1970) PNAS 65(3): 721-8; click for larger version.]

References:
Benson, A.A. (1966) “On the orientation of lipids in chloroplast and cell membranes.” Journal of the American Oil Chemists’ Society 42: 265-270

Danielli, J.F. and Davson, H. (1935) “A contribution to the theory of permeability of thin films.” Journal of Cellular and Comparative Physiology 5: 495-507

Glaser, M., Simpkins, H., Singer, S.J., Sheetz, M., and Chan, S.I. (1970) “On the interactions of lipids and proteins in the red blood cell membrane.” PNAS 65(3): 721-8
[http://www.pnas.org/content/65/3/721.full.pdf]

Gorter, E. and Grendel, F. (1925) “On bimolecular layers of lipoids on the chromocytes of the blood.” Journal of Experimental Medicine 41: 439-443

Lenard, J. and Singer, S.J. (1966) “Protein conformation in cell membrane preparations as studied by optical rotatory dispersion and circular dichroism.” PNAS 56(6): 1828-1835
[http://www.pnas.org/content/56/6/1828.full.pdf]

Stoeckenius, W. and Engelman, D.M. (1969) “Current models for the structure of biological membranes.” Journal of Cell Biology 42: 613-646

Wallach, D.F.H. and Zahler, P.H. (1966) “Protein conformations in cellular membranes.” PNAS 56(5): 1552-9 [http://www.pnas.org/content/56/5/1552.full.pdf]

How do we know the structure of cell membranes (part 2)

The lipid bilayer

The first research to produce evidence that the lipids of a cell membrane are in two layers (a lipid bilayer) is attributed to Gorter & Grendel (1925). They arrived at this conclusion by measuring how much lipid was present around red blood cells (which they called chromocytes), and determined that it was sufficient to surround the red blood cell twice. The basics of their approach are straightforward: they took red blood cells obtained from various animals (humans, rabbits, dogs, goats, guinea pigs, and sheep) and mixed the cells with acetone. Any lipids in the cells would dissolve in the acetone, leaving the rest of the cell contents behind. Having extracted the lipids from the cells, the acetone could be evaporated off, leaving only the lipids. With a little more work, it’s possible to determine how large a surface those lipids would occupy. If you know how many red blood cells were used to extract the lipids and you know the surface area of an average red blood cell, then you can figure out the total surface area of all the red blood cells. By comparing the surface area of the red blood cells to the surface area covered by the extracted lipids, Gorter & Grendel found that the area covered by the lipids was twice the surface area of the red blood cells. Looking at their original data below, it’s the last three columns on the right that are most important. The third column from the right [“Total surface of the chromocytes (a)”] is the surface area of the red blood cells. The next column is the surface area of the lipids extracted from the red blood cells [“Surface occupied by all the lipoids of the chromocytes (b)”]. The last column is the ratio of the two – and column (b) is approximately twice column (a).

[click for larger image]

Their conclusion is straightforward: “It is clear that all our results fit in well with the supposition that the chromocytes are covered by a layer of fatty substances that is two molecules thick.” Other work by later scientists would reaffirm this conclusion. However, Gorter & Grendel went beyond this, suggesting the way these two layers of lipids would be oriented: “We therefore suppose that every chromocyte is surrounded by a layer of lipoids, of which the polar groups are directed to the inside and to the outside… On the boundary of two phases, one being the watery solution of hemoglobin [the fluid inside the red blood cell], and the other the plasma [the fluid in which your red blood cells float], such an orientation seems a priori to be the most probable one.”

To understand what Gorter & Grendel are talking about, consider the phrase ‘oil and water don’t mix’. They don’t mix because water is a polar molecule and oils (e.g., olive oil) are non-polar. Polar molecules have an asymmetrical distribution of electrical charge – this is due to an asymmetrical distribution of bonding electrons. As a result, parts of the molecule are more negatively charged than other parts. For non-polar molecules, electrical charge is more or less symmetrically distributed around the molecule. Polar molecules don’t interact very strongly with non-polar molecules, so the two types of substances don’t mix. Since oils are a type of lipid, we can generally conclude that lipids and water don’t interact very well. However, one particular group of lipids, known as phosopholipids, do interact with water. In fact, phospholipids interact with both water and other lipids. Phospholipids are able to interact with water because part of a phospholipid molecule is polar (the head) and part is non-polar (the tails). Because of this property, phospholipids mix well with both water and other lipids. Phospholipids are often represented like the drawing below (with the circle symbolizing the polar head and the two lines representing the two fatty acid – non-polar – tails):


When placed in water, phospholipids tend to form structures that minimize the contact between the molecules of water and the non-polar tails, and maximize the contact between the water molecules and the polar heads. So, Gorter & Grendel are hypothesizing that the two layers of lipids suggested by their data are arranged in the manner shown below.


References:
Gorter, E. and Grendel, F. (1925) “On bimolecular layers of lipoids on the chromocytes of the blood.” Journal of Experimental Medicine 41: 439-443

There's also a nice teaching module designed to help students arrive at the idea of a lipid bilayer on there on, just by considering the data of Gorter & Grendel:

An Emerging Model of Cell Membrane Structure (Laura Martin) – http://cnx.org/content/m15250/latest/ Created October 15, 2007; Connexions web site. Version 1.2.

How do we know the structure of cell membranes (part 1)

Why would anyone care about the structure of a cell membrane? For nutrients to enter the cell (and for wastes to leave it) they will have to cross the cell membrane. Knowing about the composition and structure of the cell membrane is the first step in understanding how things can pass through it, and the ability of a cell to regulate what enters and leaves is fundamental to a host of important physiological processes: generation and conduction of nerve signals and urine production are two examples.

We all learned that animal cells are surrounded by a membrane, variously referred to as the plasma membrane, the cell membrane, or just the membrane. Anyone who has taken a class in Anatomy and Physiology (or Cell Biology, or even most General Biology courses) has learned about the structure of cell membranes: what they are made of and how the components are arranged. If you remember anything about cell membranes, it’s probably a few key phrases like “lipid bilayer” or “fluid mosaic model”. In a typical A&P class, these details probably are stated matter-of-factly. But why do these phrases capture the essence of cell membrane structure, and where did they come from? How do we know what we think we know about the structure of cell membranes?

The composition of cell membranes

It now seems obvious that cell membranes are composed primarily of lipids; it was not always so apparent. The idea of a lipoidal cell membrane seems to originate from the research of C.E. Overton in the 1890s (De Weer, 2000). Although the details were never published, Overton apparently deduced that lipids were a primary component of cell membranes from his research on the ability of various substances to pass through cell membranes. He found that membranes were more permeable to substances that were lipid soluble (meaning that they dissolved in lipids). In effect, things passed through the membrane by dissolving through it. By 1935, the lipoidal composition of membranes was entrenched enough for two researchers to assert the fact without attribution or reference. They wrote: “There is now a considerable body of evidence supporting the view that living cells are surrounded by a thin film of lipoidal material.” (Danielli and Davson, 1935). Proteins were also known to be associated with membranes. But how were these lipids and proteins arranged? I'll begin to address that in subsequent posts.

References:
Danielli, J.F. and Davson, H. (1935) “A contribution to the theory of permeability of thin films.”
Journal of Cellular and Comparative Physiology 5: 495-507

De Weer, P. (2000) “A century of thinking about cell membranes.” Annual Review of Physiology
62: 919-926


There's also a nice teaching module designed to lead students to the recognition that membranes must be primarily composed of lipids by showing them data from R. Collander (1937) Trans. Faraday Soc. 33:985-990. The data relate the solubility of a solute in oil to how fast it moves into an algal cell:

Deducing the Biochemical Composition of the Cell Membrane - Reasoning from Quantitative Evidence (by Laura Martin) - http://cnx.org/content/m15249/latest/ Created October 15, 2007; Connexions web site. Version 1.2