Tertiary Structure

Tertiary structure is the complete three-dimensional arrangement of one polypeptide chain: how its helices, strands and loops pack and where every side chain ends up in space. Where secondary structure is local, tertiary structure is global and driven mainly by burial of hydrophobic side chains away from water, reinforced by salt bridges, long-range hydrogen bonds and disulfide cross-links. It is the level at which a binding site exists, because a receptor contact surface is usually built from residues that are far apart in the sequence and brought together only by the fold.

Chain length decides whether the concept applies at all. Insulin, at 51 residues held by three disulfides, has a genuine and well-characterised fold; so do the disulfide-rich defensins and conotoxins, where the cystine framework is what holds the shape. Most therapeutic peptides are too short: a 30-residue chain has too little surface to bury a hydrophobic core, so it samples many conformations and acquires a shape only on binding, an induced fit rather than a pre-existing structure. Structures are determined by X-ray crystallography, solution NMR and now cryo-electron microscopy, increasingly of the peptide already bound in its receptor.

The practical consequence is where a design constraint must come from. With no stable fold of its own, affinity has to be bought by pre-organising the molecule chemically, through cyclisation, stapling or engineered disulfides, rather than by adjusting a fold that does not exist.

The error to avoid is treating a rendered three-dimensional image as an observed fact. Many published peptide structures are computational predictions, or the conformation seen in a crystal or bound to a receptor, none of which is the free molecule in plasma; an elegant picture can conceal a molecule that is disordered most of the time.