The variety of domains found in 3-D structures of proteins is limited when compared to the numbers of possible structures. The structures of many domains appear to be conserved by evolution.

• The major driving force for these conserved foldings: burying and clustering of hydrophobic side chains, to minimize their contact with water

Structural classifications of domains are important because:

• Key utility of classifications: demonstrate evolutionary conservation of protein structure.

• Proteins can be grouped into families according to similarities in domain structure and amino acid sequence.

• All members of a family probably descended from a common ancestral protein, one of only a few thousand that may have been present in the primitive organisms of three billion years ago

Numerous databases exist for the classification of these structures. Three examples:

• SCOP (Structural Classification of Proteins)

• CATH (Class, Architecture, Topology, Homologous superfamily)

• DALI/FSSP/DDD (Fold classification based on Structure-Structure alignment of Proteins)

For example, THE SCOP database uses the following classifications:

• Species: a distinct protein sequence and its natural or artificial variants

• Protein: similar sequences of essentially the same function that originate from different biological species

• Family: proteins of rerlated sequences but different functions

• Superfamily: proteins families sharing common structural and functional features

Classifications by all of the above databases were based on the following assumptions, generally accepted at the time of their creation:

• sequences of proteins performing the same function diverged with biological speciation

• a protein has only one native structure

• homologous proteins fold into similar structures

• protein structure is conserved by evolution to a greater extent than sequence

• a particular protein fold can have evolved more than once

While these assumptions are still true in the main, exceptions to all of them are now known.

Look next at a few illustrations of the kinds of similarities that are observed. The most common types are

• All a: almost entirely a-helices and loops

• All b: only b-sheets and nonrepetitive structures linking them

• a/b: supersecondary structures like the bab fold, in which the two types of structure alternate

A coiled-coil is a pair of amphipathic a-helices, interacting through their hydrophobic faces. This kind of motif also is sometimes called a zipper, because of its resemblance to a partially closed zipper. The zippers in the pictures below have been excised from yeast GCN4 protein.

A Leucine Zipper

The hydrophobic groups (blue in the right-hand picture) interdigitate, and tend to be the isobutyl side chains of leucine residues. These structures often are called leucine zippers. Their van der Waals interactions with each other stabilize the zippers.

Another common domain is the four-helix bundle:

Four-Helix bundle in Human Growth Hormone (1hgu)	Four-Helix bundle in Cytochrome b (1lm3)

• As the cytochrome b picture shows, four-helix bundles are held together by hydrophobic intereactions among leucine and isoleucine side chains, creating a strongly hydrophobic enviroment in the interior of the bundle.
  • In this case that environment is the perfect place for a heme.

• The human growth hormone bundle has two pairs of parallel helices joined antiparallel

• The cycochrome bundle is made of antiparallel helices

Many proteins that are all, or chiefly, b-sheet, have two sheets packed face-to-face, in what is called a b-sandwich.

The sandwich typically is stabilized by hydrophobic interactions between side chains on the facing sheets.

Here are two examples:

Transthyretin (1eta)	Glycosyl Asparaginase (1pgs)
Side Views

Each of these proteins has two such domains. Note also in the side view of glycosyl asparaginase that facing sides of the sheets interdigitate liphophilic residues.

Here is another intriguing structure, built largely of b-sheet, pectate Lyase C from Erwinia chrysanthemi

Pectate Lyase (1air)	1air Turned 90o	Jmol

The fold is a b-helix. A group led by Professor Bonnie Berg of MIT found this fold to be extremly common in the toxins and surface proteins of human disease organisms such as cholera, ulcers, malaria, pneumonia, and Lyme disease [PNAS, 2001, 98, 14819]

A particularly striking all b motif is the b-barrel, which we previouslly described in triose phosphate isomerase.

Here is one from the E-coli outer membrane (1by3):

• The barrel is composed of 22 antiparallel b-strands

• It forms a trans-membrane channel through which iron, bound to the carrier ferrichrome, enters from the surrounding medium

The second example is rather more fearsome: a hemolysin from the bacterium Staphylococcus aureus:

The protein is assembled from seven chains, which combine when they encounter an appropriate cell surface receptor.

The barrel is very hydrophobic, punches through the cell membrane, and allows the indiscriminate flow of ions in and out of the cell, destroying it.

We have already seen one example of an a / b protein, the TIM barrel of triose phosphate isomerase:

Triose Phosphate Isomerase	Glycolate Oxidase

• Not surprisingly, the catalytic site of both enzymes is located in the middle of the barrel.