User:Kinkreet/Protein Science/Membrane Proteins

Membrane proteins are found in many essential metabolic pathways. For example, integral photosynthetic complexes are found in photosynthesis and respiration, for transport as well as for ATP synthesis. Other more general uses for membrane proteins include transporters, pumps and channels that regulate movement of ions and nutritions in and out of the cell, plus wastes and toxins out of the cell.

Membranes are made up of a huge variety of lipid molecules, which are amphiphilic, having both hydrophobic and hydrophilic properties. Likewise, membrane proteins have the same amphiphilic properties: the peripheral part of the protein is hydrophilic, while the integral region is hydrophobic. Some integral membrane proteins might not have any proteinous portion inside the membrane, and is lipid-modified so that the lipid inserts into the membrane. There are four main types of lipid modifications; 3 for membrane proteins with their peripheral region inside the cell, and one for membrane proteins with peripheral regions outside the cell. Inner Outer
 * 1) Palmitoylation on an internal cysteine or serine, much like the post-translational modifications found in Wnt and Shh ligands
 * 2) N-myristoylation on N-terminal glycine
 * 3) Farnesylation (or geranylgeranylation) on C-terminal cysteine
 * 1) GPI anchor on C-terminus.

Monotopic membrane proteins have parts of its proteinous part inserted into the membrane, but it does not pass through the membrane.

β-barrel membrane proteins are commonly found on the outer membranes, but not the inner membranes, of bacteria, chloroplast and mitochondria; it is thought to facilitate transport down a concentration gradient. Common β-barrels are porins which include FepA, OmpLA, maltoporin, TolC and α-Hemolysin toxin. In the amino acid sequence of a β-barrel protein, the hydrophobic residues are spaced apart so that most of them point towards the membrane lipid environment, whereas most of the hydrophilic residues point towards the hydrophilic core. Because of its spiralling structure and cross-links between strands, there is not a lot of flexibility in these proteins, thus their functions are less versatile than α-helical proteins. However, its inflexibility means it is easier for structural determination.

As mentioned, α-helical membrane proteins are more flexible and have a bigger diversity in function, and thus are also found to be more abundant. There are six major types of α-helical membrane proteins: Type I - VI.

The internal hydrogen bonds between groups on the peptide backbone of the alpha helices make the helices generally hydrophobic, however, hydrophobicity also depends on the side chains, as they are displayed to the outside of the backbone, and may mask the hydrophobicity of the backbone. Therefore, transmembrane domains must contain a high proportion of hydrophobic residues. A typical transmembrane domain can have a range of residues, from 20-30 residues. This variation occurs due to varying thickness of the membrane and the orientation of the alpha helix (perpendicular, slanted or bent).

A hydrophobicity scale, developed by Engelman, represents an amino acid's approximate hydrophobicity in the form of a number. A positive number means it is hydrophobic, and a negative score implies it is not hydrophobic. This scale is determined from the free energy required to remove any amino acid side chain from a lipid environment into an aqueous environment. (Therefore, negative scores means this is favourable)

Although most of the transmembrane residues have to be hydrophobic, some can also contain hydrophilic and charged residues; often these residues are important for function. In a study by von Heijne, it is found that 7-8 leucine residues (out of the 20-30 residues in the TM domain) are all that is required to form a transmembrane domain.

von Heijne created a peptide with an N-terminal transmembrane domain, immediately followed by a proline, a random coil, and then a variable region followed by a marker on the C-terminus. He would then mix this construct with liposomes and observe for insertion. The transmembrane domain always inserts, the proline is added to facilitate the turning of the random coil towards the membrane, the random coil exists outside the liposome. If the variable region (~20 residues in length) is capable of insertion into the membrane of the liposome, it would reinsert along with the C-terminal marker. Therefore, any variable regions capable of insertion would reinsert into the membrane and the marker would end up inside the liposome. The marker is an enzyme (oligosaccharyl transferase), which activity can be detected if it is inside the lumen of the liposome. The liposome contains an in vitro transcription-translation system which will produce the proteins required for enzymatic activity detection.

von Heijne altered the number of leucine regions; he observed that a gradual increase in the number of leucine eventually gave rise to a helical structure; 7-8 leucines are required for this to occur.

Topology
A number of online prediction tools are available, such as TMHMM. This gives us information about which residues are likely to be transmembrane. However, this does not tell us the orientation of the protein relative to the membrane, a feature called topology. However, von Heijne found that membrane proteins tend to have a higher frequency of positively-charged residues on the cytoplasmic side than the non-cytoplasmic side. This is commonly known as the positive-inside rule. He determined this because he found that changing the number of positively-charged residues on either side can change the topology of the protein. Using this rule, we can predict the full topology of the protein. However, these predictions are only predictions, and are often wrong in some parts. The best α-helical transmembrane topology prediction are achieved using the hidden Markov models with evolutionary information. However, β-barrels should be predicted using TMHMM. This is because the hydrophobic residues are spaced apart and there should not be a block where there is high hydrophobicity. The TMHMM program will simply predict it to be on the outside. This is not exactly incorrect, as β-barrels are often secreted and then inserted into the membrane after.

To predict β-barrels, a program such as BOMP (β-barrel Outer Membrane protein Predictor) can be used.

There is an exception where a protein can exist in both orientations, these are known as dual-topology proteins. There are two possible ways this can occur: one where there are two different genes encoding for different proteins with the same function (this is not the case), or that one gene encodes for one protein, but because the charge difference between the two sides are small that it can be inserted in two topologies.

As with all prediction tools, caution must be used and all predictions should be experimentally verified.

PhoA only folds into its active conformation while in the periplasm, and GFP only folds into its active conformation in the cytoplasm. Using these, and other molecules, as tags, and observing its conformation (for example, whether it fluoresces, we can determine the empirical topology of the protein.

Membrane protein variations
Most transmembrane domains of membrane proteins are made up of the conventional α-helices, with 3.6 amino acid residues per turn, and hydrogen bonding between residues n and n+4. However, 310(n, n+3) and π (n, n+5) helices can also be found. These helices are sub-optimal and bulge out from the rest of the helix, thus exposing residues out of line so it cannot form hydrogen bonds; this results in instability of that region. This instability often have a functional role in the membrane protein, for example in bacteriorhodopsin.

Proline also serves a role in generating variation. Proline cannot hydrogen bond and can induce kinks in the helix, it can further induce local instability.

Helices can pack together if they have sides which are hydrophobic and/or are small. Glycine is both hydrophobic and small, and thus can pack tightly against other helices. Therefore, helices with GXXXG have a flat and hydrophobic surface and is can pack against each other tightly.

Features
Another feature observed in transmembrane proteins are that near the interface between the membrane and the lumen/cytoplasm/extracellular space, there are a relatively higher proportion of aromatic residues, mostly tryptophan and sometimes tyrosine. This is known as the aromatic belt.

Proteins and lipids evolved together, and so the lipids do not always play a passive, structural role, but often serves a function also. For example, phosphatidylethanolamide (PE) is involved in the correct folding of LacY of lactose permease; cardiolipin stabilizes interactions between the subunits of the Fdh-N trimer (it was shown by crystal structure that the protein subunits do not make physical contact with each other)

Cholesterol
Cholesterol rigidify a membrane (does not make it less fluid, only more resistant to outside pressure), it also reduces permeability and increases mechanical durability. It is involved in post-Golgi protein sorting, lipid raft assembly and function. It has also shown to have a role in the function and organisation of G-protein coupled receptors (GPCRs). Cholesterol is able to have this effect either by interacting with the proteins themselves, or by altering the physical lipid environment near the proteins. In a study of β2-Adrenergic Receptor, cholesterol can act increases thermostability. A tyrosine residue is important for interaction. This feature is also found in a range of GPCRs, where interaction with cholesterol incraesed its affinity for some ligands.

Classes of membrane proteins
Topology

Positive inside rule (reducing environment in the cytosol)

Hydrophobicity analysis

Helix variations

Membrane proteins and lipids

How are multipass transmembrane proteins, barrels inserted into the membrane.