232

15

The Molecules of Life

odic table of the elements in chemistry. Indeed, the dehydron concept is needed to

computationally fold a peptide chain ab initio.

Examination of protein–protein interaction interfaces fully bears out the dehy-

dron interpretation. Appropriate complementarity is achieved by overexposed apolar

groups and dehydrons (rather than H-bond acceptors and donors, or positively and

negatively ionized residues, although these may play a minor rôle). One also notes

that each subunit of haemoglobin, a very stable and soluble (i.e., nonsticky) protein,

has just three dehydrons: Two are at the interface with the other subunits, and one is

the bond connecting residues 5 and 8 (i.e., flanking the sickle cell anaemia mutation

site at residue 6). In contrast, the prion protein, which is pathologically sticky, has

an extraordinarily high density of dehydrons (mean rhoρ is only about 11).

There are also evolutionary implications. It has long been realized that the evolu-

tion of proteins via mutations in their corresponding genes is highly constrained by

the need to maintain the web of functional interactions. There is a general tendency

for proteins in more evolved species to be able to participate in more interactions;

they have more dehydrons. For example, mollusk myoglobin is a perfectly wrapped

protein and functions as a loner. Whale myoglobin is in an intermediate position, and

human myoglobin is poorly wrapped, hence sticky, and operates together with other

proteins as a team. Although the folds in a protein of a given function are conserved

as species diverge, wrapping is not (even though the sequence homology might still

be as much as 30%). Structural integrity becomes progressively more reliant on the

interactive context as a species becomes more advanced.

A corollary is that the proteins of more complex species are also more vulnerable

to move into pathological states. The prion diseases form a good example; they are

unknown in microbes and lower animals. Moreover, they mainly attack the brain,

the most sophisticated and complex organ in the living world.

15.5.3

Protein Structure Determination

High-throughput methodology (also called structural genomics) comprises the

following steps:

1. Select the gene for the protein of interest.

2. Make the corresponding cDNA.

3. Insert the cDNA into an expression system.

4. Grow large volumes of the protein in culture (if necessary with appropriate iso-

topic labelling of C and N).¹⁴

5. Purify the protein (using affinity chromatography).

14 For some of the problems associated with the production of recombinant proteins, see Protein

production and purification, Nature Methods 5 (2008) 135–146.