Credit: © 2008 ACS

How do you study the structure and properties of a new protein if you only have extremely small amounts to begin with, and then it just does not crystallize? Make it for yourself, then crystallize the naturally occurring L-form with its mirror image D-form. That's the answer according to Stephen Kent and colleagues at the University of Chicago, in collaboration with biochemists at the University of Pennsylvania.

To study the antifreeze properties of a protein (sfAFP) isolated from the Canadian snow flea, it first had to be made1 by chemical synthesis because isolating it as a natural product yielded only microgram quantities and recombinant expression had proved difficult. A four-step 'native chemical ligation' method was used to join four peptide sequences together to form the 81-residue protein. Both the native L-sfAFP and its enantiomer D-sfAFP showed the same ability to inhibit recrystallization of ice, as would be expected because the surface of ice is achiral.

It is difficult to study the structure of many proteins because growing suitable crystals can be hard or impossible. To overcome this, a racemic crystal containing both enantiomers of sfAFP was grown2. It is believed that the two forms pair up to 'remove' the stereochemistry and enable access to more favourable crystal-packing motifs. This is the first time that racemic crystallization has been used to obtain crystals of proteins whose natural forms do not normally crystallize easily. Naturally obtained proteins do not contain both forms, so total chemical synthesis has to be used for this strategy.

The protein was found to have a unique fold, consisting only of stacked polyproline II helices. The structure was similar to a recently proposed model, and has implications for the antifreeze activity. An important application for antifreeze proteins could be for storing organs at low temperature, especially as the non-natural D-form is expected to biologically inert. If racemic crystallization can be applied to all proteins, it could prove to be a powerful tool in structural biology.