Membranes

Here we go again. Before the exam we had discussed lipids, and their main functions: Energy storage, structural (membranes), and, to put it simple, participation as prosthetic groups and intra- and intercellular messengers. Today we will start discussing the complex arrangements formed by structural lipids, membranes.

Depending on the membrane we talk about, there are some conventions we will follow:

- The plasma membrane refers to main cellular membrane, the one that separates the cytosol from the rest of the universe.

- The inner (or cytosolic) layer of the membrane bilatyer referes to the lipid layer whose polar heads point towards the inside of the cell, towards the cytosol. The outer layer is the lipid layer from the bilayer that has its polar heads pointing towards the extracellular fluid.

As you probably know by now, membranes are used by cells to separate outside from inside. They are also used within a cell to separate different organelles from the cytosol. If it wasn't for transport proteins (next class), membranes would be virtually impermeable to polar solutes and ions, like glucose, Na+, Cl-, etc., etc. This is because the membrane is a lipid bilayer in which we have a highly hydrophobic core. Passing of bare polar solutes throught this 'solvent' is a really bad process energetically.

Appart from these compartimentalization function, membranes are the where many important biological processes take place. Among them, we have:

- Transport. As we said, membranes alone would not allow much to pass from one side to the other. Membranes are peppered with proteins that promote the transport or translocation of solutes from the outside to the inside and vice-versa.

- Energy conversion. Most of the energy-yielding processes in cells (oxidative phosphorylation , photosynthesis) take place in relatively large protein systems which form part of membrane systems (mitochondria and chloroplasts).

- Siganling. The cell has to know what is going on around it, and the only way to do this is by placing receptors on the membrane. This receptors will usually convert a chemical signal (a molecule binding) to a conformation change (a protein changing shape) which will cascade into the cell, producing a response from the cell.

Molecular composition of membranes

As we had discussed before, the main characteristic of the membrane is the precense of a lipid bilayer. Although you may think that the lipid part composes the vast majority of the membrane, it is actually a lot less. There is an amazing number of proteins and other small molecules associated with the membrane. The exact composition of the membrane will depend on the task the cell (or organelle) surrounded by it has to accomplish.

For example, red cells, which basically move around O₂ and CO₂ and have not organelles when mature, can be considered a membranous bag of hemoglobin - Therefore, a large percentage of their membrane mass will be lipids alone.

Neurons, on the other hand, have to communicate electric signals between cells, which, as we will see, is done by sodium transport proteins. The membrane of neurons has a large density of proteins, and thus a large percentage of their mass will be protein.

As for small molecules, the main one present in the membranes is cholesterol and other sterols - Other steroid-type molecules and lipid messengers/prosthetic groups will form part of a protein.

So what's the deal with cholesterol? Lets see. Cholesterol is a relativelly flat and rigid molecule that intercalates between fatty acid chains from phospholipids. If we put cholesterol in a membrane (say, of saturated phospholipids), two things will happen:

- First, since the contacts between fatty acid chains from different phospholipid molecules is slightly disrupted (i.e., we won't have as many hydrophobic interactions between fatty acid chains). This increases the fluidity of the membrane, i.e., it lowers the melting point of the membrane bilayer.

- In cases of very fluid bilayers, such as those composed by large ammounts of unsaturated phospholipids, intercalating cholesterol, which is rigid, will reduce the fluidity of the membrane, therefore increasing the melting point.

Lipid composition

If we consider only the lipid part, the ammounts of different types of lipids (phosphoglycerides, sphingolipids, phosphatidylinositol, etc., etc.) will also vary. Not only will the distribution of these lipids vary between different cell types, but it is also distributed asymetrically along the membrane.

This means that different different types of phospholipids will be distributed unevenly in the inner and outter membrane layers. For example, shpingolipids are concentrated in the outter layer, and phosphatitylinositol on the inner layer:

This makes perfect sense. If you remember, a many sphingolipids had oligosaccharides attached as their polar heads, and these oligosaccharides were used as signaling agents. By placing them on the outside, the polar heads (the oligosaccharides) will point towards the extracellular fluid, where the message they carry is read.

Also, if you remember, a phospholipase choped up phosphatidylinositol and releases inositol into the cytosol as a signal, and that is why phosphatidylinositol has higher concentration in the inner (cytosolic) membrane.

Membrane lipid dynamics

Although there are thousands (millions!) of non-covalent interactions between the lipids that compose the membrane bilayer and mke it hard to break up, several studies have shown that the bilayer system is extremely dynamic. A phospholipid molecule can travel laterally across the membrane (that means, across the membrane layer it is inmersed) at high speeds, and therefore the whole lipid layer is like a 'sea' of lipids. This is called latteral difussion:

Also, we have to remember that the acyl chains of phospholipids are composed mainly of straight chain hydrocarbons with many single bonds. Therefore, as we increase the temperature the acyl chains will start rotating around the single bonds more rapidly, giving a 'disordered' apereance to the bilayer - The membrane becomes more fluid, which not only increases the winding of the acyl chains, but also increases lateral difussion:

One process that is 'forbiden' withouth helper molecules is the flip-flop movement of phospholipid molecules. This refers to the translation of one phospholipid unit from the outter to the inner layer of the membrane, or vice versa:

It is pretty easy to figure out why it is not very common: To accomplish this, we need to take the polar head of the phospholipid from one side to the other, which requires draging it through the very non-polar 'solvent' created by the fatty acid chains - It is a very unfavourable event, with a large and positive DG barrier.

However, the phospholipids themselves are synthetized in proteins attached to the inner membrane, and therefore we need something that will move them from the inside of the membrane to where they belong, which may depend on the the type of phospholipid and the type of cell.

This is accomplished by two types of membrane proteins (see below). Ones are called the flipases (very original name...). These proteins just distribute the excess of one type of phospholipid on one side of the membrane evenly through both sides of the membrane, and the type of transport they perform is called facilitated difussion or passive transport (more tomorrow): They just redistribute things evenly.

However, we know that there is more of certain phospholipid on one layer than the other. There is a second class of membrane proteins called the phospholipid translocases, which will take the synthetized phospholipid and put it where it belong, irrespective of the differences in concentration. This type of transport is called active transport, and requires energy (burning ATP). more next class...

Membrane proteins

The other big (huge) components of the bilipidic membranes are proteins, also called in this case membrane proteins. As we said, the proteins in the membrane are responsible for a large variety of processes. The most important are energy generation (ATP - oxidative phosphorilation and photosynthesis), and transport of polar solutes. Other proteins amplify signals from the environment which tell the cell to do certain things (for example, stress due to the attack of a bug may trigger the generation of some nasty chemical to fend off the bug - that kind of signaling).

Membrane proteins are divided in two groups, integral proteins (which are surrounded by lipids), and periphera proteins (which rest on either side of the membrane).

Integral proteins

These proteins are inmersed in the membrane. They are tightly bound to the lipids surrounding them through hydrophobic interactions. The interactions are so many, that it is very hard to separe them from the lipids. If we break up the membrane and try to isolate the integral proteins, we will never get rid of lipids surrounding them, except if we treat them with non-polar organic solvents (which disolve lipids), we treat them with soaps (which take the place of the lipids), or with denaturing agents like urea or guanidine hydrochloride (which disrupt hydrophobic interactions).

In order to have favourable hydrophobic interactions with the bilayer, the residues that are stuck in the membrane have to be highly non-polar. In general, integral proteins are amphiphiles, which means that they have a polar section (which is usually in contact with the aqueous solution surrounding the membrane) and a highly non-polar section (or sections) which intercalate with the membrane bilayer.

Among integral proteins we have two different types. One will have part of it buried in the membrane and the other poking out of it. The second class, which we will discuss in some detail, have three sections: One pokes to the extracellular fluid, another is buried in the inside of the membrane, spanning across the whole membrane bilayer (30 Å), and the third will poke into the cytosol. These integral proteins are called transmembrane proteins, and most ion channels and pumps are like this.

Since integral proteins bind tightly to lipids, once we remove the lipids surrounding them they become very unstable. The hydrophobic sides of the proteins will tend to find other hydrophobic surfaces to bind to (hydrophobic collapse). This usually means that several molecules ot the integral protein will come together, or aggregate, and precipitate out of solution. This makes structural studies (X-ray, NMR) of these proteins extremelly difficult, and most of the structural data we have comes from simple models and computer simulations.

In spite of this, there are some things we can hypothesise about integral proteins (which from now on will be transmembrane proteins for us). First, since protein sequencing is a lot easier than X-ray or NMR studies, we can get information about the amino acid composition and sequence of these proteins relativelly easy.

Most of the times, these type of analysis show segments of polypeptide which are hydrophilic, and therefore we can identify them as being the sections which interact with the cytosol or the extracellular fluid.

Second, if we label the proteins with polar molecules that cannot go through the membrane and then we isolate and sequence the protein, we can identify which amino acids are laying on the outside of the cell, because the labelling agent will only label amino acids on the outside of the cell.

Finally, if we look at the hydrophobic regions of the molecule, we can identify segments of protein that could travel across the membrane. A simple but useful method to do this are hydrophaty plots. Each amino acid has a hydrophobic parameter, which is proportional to the DG of transfering it from non-polar solvents to water: Polar amino acids, like lysine or arginine, will have large negative hydrophaty numbers, because it is very favourable to transfer them from non-polar solvents to water.

Other amino acids, like luecine, phenylalanine, etc., will have large positive hydrophaty numbers, because it is very unfavourable to tranfer them from non-polar solvents to water.

All these numbers are tabulated. If we now take our transmembrane protein and we plot the hydrophaty number versus the residue number, we can identify regions which are hydrophobic and regions that are hydrophilic.

In order to get decent results, we usually average the hydrophaty indexes of a number of amino acids for each amino acid we are evaluating (3 before and 3 after, or 10 before and 10 after - These is called a window). In the example shown above (ovine rhodopsin), we can easily identify seven regions in the primary structure of ~ 25 to 30 amino acid residues each that have high hydrophaty indexes.

Why ~ 25 amino acids each? What protein structural motif can span 30 Å with 25 amino acids? If you think a little, the only one is an a-helix. An a-helix is perfect for this: First, 25 residues are 30 Å long. Second, if we use hydrophobic residues in the a-helix, all the hydrophobic side chains will be poking into the membrane, and we will have a favourable interactions between the helix and the membrane. Third, in an a-helix, all the h-bonds are within itself - in this way, we avoid having the polar regions of the peptide bond (the -CO-NH-) interacting with the lipids. a-helices are commonly found in transmembrane proteins. An X-ray structure for a protein from similar protein (bacterial rhodopsin - the photosynthetic reaction center) was solved, and here's how it looks:

Here's the same thing, as a CHIME-PDB file. What about using b-sheets? This is harder, because if you remember, there was usually one strand of the b-sheet exposed to something - lipid in this case - which will not be able to form h-bonds. The only case in which we have b-sheets in transmembrane proteins is in b-barrels, because all the h-bonds in these structural motifs are satisfied internally.

All the interesting proteins we will discuss next class are transmembrane proteins...

Peripheral proteins

The other type of membrane proteins are the peripheral proteins. As the name says, these are either to one side or the other of the membrane bilayer. Their attacment to the membrane can be varied. The thing can be simply attached to the lipid surface or to an integral protein trough non-covalent interactions. These will most likely be salt-bridges and h-bonds, because all the polar heads of the lipids (or the polar segments of the integral proteins) is what faces the extracellular fluid or the cytosol.

Another way is through lipid, non-covalent, anchors. In this case, a specific amino acid residue is esterified (or amidated) by a chain of a lipid moiety. Some examples are:

This little lipid 'tail' of the protein will intercalate with the rest of the membrane lipids, and acts like an anchor for the protein in the lipidic 'sea'. The hydrophobic tail can be a normal fatty acid (as above), or a prenyl residue - A derivative of isoprense. Common ones are farnesyl and geranylgeranyl - Depending on the sequence of amino acids in the C-terminal end of the peptide, we will get a farnesyl or a geranylgeranyl residue:

Finally, we can have a covalent anchor between the peripheral protein and the lipid bilayer that keeps it close to the membrane. It is more or less a variation of the previous one, but here we have a oligosaccharide chain (a glycan) that connects a specific residue in the protein to a phospholipid of the membrane, usually a phasphatidylinositol lipid moiety. I think you'll get a kick out of the name: glyccosilphosphatidylinositol-linked proteins...

In all peripheral proteins, it is relativelly easy to isolate them - Since most of the protein surface is in contact with aqueous solvent (or with the polar ends of the membrane), the isolated protein will be relativelly stable in water. In cases of anchored proteins, disruption of the membrane with detergents will give us the protein. For covalently bound proteins, phospholipases work (they are actually what the cell uses to chop them off...).

This is how the whole enchilada (phospholipids, cholesterol, peripheral and integral proteins) looks like:

In the same way that different phospholipids are in an asymetrical disposition through the membrane, integral proteins are distributed asymetrically. Furthermore, certain integral protein will always have its polar section pointing out from the same side of the membrane (that is, wee cannot have the same protein pointing towards the extracellular fluid somewhere and to the cytosol somewhere else).

Fluid mosaic membrane bilayer model

If we now take all this stuff into account, we can come with a model for basically all lipid bilayers. We already discussed that the lipids are in a dynamic environment. The proteins that are inmersed in them will also be mobile, except in cases were they are anchored to other cytoskeletal proteins (spectrin). The whole thing looks like a really thick plate of Campbell's Vegetable soup, in which the membranes proteins (the noodles, potatoes, etc., etc.) float in the lipid solvent (the broth of the soup).

Since we can consider the membranes as bidimensional structures (very thin layers that span long areas), the bilayer looks like a mosaic. Since it is fluid, this model is called the fluid mosaic bilayer model, from which you probably heard already...

Next time we will start looking at different transmembrane proteins, particularly those in charge of solute transport across membranes.

Prepared by Guillermo Moyna, 1999.