Models of the Plasma Membrane

The second law of thermodynamics would ensure that order would deteriorate if all the components of a living organism were not confined within a small enough space to allow feedback and regulation.

Life as we know it is cellular. Small quantities of cytoplasm are bound in sacs of plasma membrane. One of the tasks of cell biology has been to figure out how these remarkable membranes work.

We know that membranes have certain properties.

1. They act as a barrier between the cell and its environment, allowing a complex organized system to exist inside the cell.
2. They permit the passage of selected substances into and out of the cell.
3. They flex, bend and flow to allow the cell to change shape.

As we have accumulated evidence about the cell membrane, biologists have tried to develop models to explain how the boundaries of cells are constructed, and how they function.

The earliest models of the membrane were very simple. Evidence from experiments with cells late in the 19th century suggested that lipid soluble substances entered cells faster than water soluble substances (a fact very important both to the study of poisons and to the development of drugs). The first model of membranes simply suggested that they were layers of lipid (fat). These early models did not account for how a lipid layer could remain stable in contact with the water of the cell and the environment, since lipids are repelled by water. Presumably, the lipids would all have to be bonded together to maintain their integrity, but that would have seriously restricted the flexibility of membranes.

Chemical analyses of isolated membranes in the early twentieth century revealed that they were largely made of phospholipids and cholesterol. It was known that phospholipids were amphiphilic - i.e. they had one end (the head) that was water soluble and two "tails" that were lipid soluble. Experimentally, phospholipids could be made to form a double layer of molecules in water, with the water soluble heads of each layer oriented outward toward the water and the lipid soluble tails oriented inward toward each other. This bilayer would then be stable.

Chemical analyses also revealed a great deal of protein in membranes. Though membranes were more permeable to lipids than to water, they still absorbed water faster than a pure phospholipid layer should have. Since most proteins are water absorbent, a new model was developed. It suggested that the phospholipid bilayer was coated on both sides with water-soluble proteins, in a kind of lipoprotein sandwich. This model, proposed in 1934, was called the Davson-Danielli model (after the scientists who described it).







When the first good electron micrographs revealed that membranes indeed had a three layered structure (dark-light-dark), this was taken as confirmation of the Davson-Danielli model, which remained the accepted view until about 1970. Note that in the micrograph to the left, you can see two adjacent membranes, and each has a light layer sandwiched between two dark layers.




Meanwhile, however, advances in many areas of biology and chemistry had begun to uncover evidence incompatible with the Davson-Danielli model. Biologists had long known that membranes had some properties of fluids. They had watched amoebas move and form food vacuoles for several centuries, and they knew that many other cell processes showed similar membrane behaviour. The Davson-Danielli model did not make it clear how such fluidity could occur without tearing or breaking of bonds.

Microsurgical methods reinforced the idea that the membrane was a fluid. If a cell membrane is pushed with a probe, it bends like the surface of a balloon, and springs back when released. If it is penetrated, however, the membrane simply conforms around the probe. When the probe is withdrawn, the membrane reseals as if it were a liquid flowing into itself.

New chemical methods, especially methods for analysing the structure of proteins, revealed that the proteins of membranes were highly variable in both quantity and type. It was hard to believe that they could be critical to the basic structure of membranes and still be so variable. New methods also showed that the proteins in membranes, rather than being essentially hydrophilic, were largely lipophilic and hydrophobic. If they had formed an outer layer on the membrane, these hydrophobic proteins would have been located between the hydrophilic heads of the phospholipids and the water of the cytoplasm or the environment. This would have created a chemically unstable membrane.

Scanning electron microscopes revealed an additional problem. Using "freeze-fracture" techniques, biologists were able to split cell membranes along the lipid layer, which forms a zone of weakness in a frozen cell.  They could then coat the membrane with platinum, and get a "three dimensional" view of its surface texture. In the image, you can see the inside of some membranes, as if one layer of phospholipids had been peeled back. What they saw was a smooth plane with numerous bumps sticking out. The bumps were very variable from membrane to membrane, and were about the right size to be proteins. Since the proteins showed up in the middle of the membrane, they could not merely be a layer coating the outside. They must be imbedded right in the phospholipid layer. Presumably, the water soluble portions of the proteins occurred outside the membrane itself, and the lipid soluble parts were located in the double layer of lipid tails.

New techniques for tagging proteins and hybridizing cells have really confirmed the modern model of membranes (see diagram at right). Fluorescent tags, that give off coloured light under ultraviolet illumination, can be attached to the proteins of a cell membrane. When two cells are tagged so that their membrane proteins glow with different colours, and the cells are then forced to fuse into one large cell, the fate of their proteins can be seen by the differences in colour. Immediately after fusion, all the tags of one colour will be on one side of the cell. After about an hour, they will be equally mixed with the proteins of the other colour. Thus, the proteins must have flowed from their original non-random distribution to their final random one. This is clear evidence that a membrane behaves more like a liquid, in which molecules move freely, than it does like a solid.










All the current evidence is compatible with the Fluid Mosaic Model of plasma membranes, which is generally accepted by all biologists. This model suggests that the membrane is fundamentally a phospholipid bilayer, as originally supposed sixty-five years ago. The molecules of this bilayer are not bonded to one another, but float freely in two dimensions. In the third dimension (perpendicular to the membrane surface) they are held in place by the opposing repulsions and attractions to water. Any molecule that moves relative to the plane of the membrane will either expose its lipophilic tail to a watery environment, or its hydrophilic head to the lipid layer in the middle. Repulsions will then tend to push it back where it came from. A membrane is thus liquid in two dimensions, but an elastic solid in the third.

Proteins are thought to float freely in this fluid bilayer, also held in place by their lipophilic sections (orange in diagram) which are attracted to the fatty middle layer of the membrane. This accounts for their high mobility, and for the fact that they are so visible in freeze-fracture images.

According to the Fluid Mosaic Model, the basic structure of the membrane is provided by the phospholipid molecules. Varying the lipid tails can give membranes different properties. For example, unsaturated fatty tails make a membrane more liquid, while the addition of cholesterol to the fatty layer makes them more viscous and more repellent to water.

Proteins are responsible for the special characteristics of different types of membranes, controlling their ability to transport molecules, to receive chemical messages, to attach to adjacent cells, etc. The Fluid Mosaic Model solves both requirements #1 and #3 at the beginning of this page. As you will see in the next page, it also makes possible a solution to requirement #2.