Stereochemical basis of heat stability in bacterial ferredoxins and in haemoglobin A2

Abstract

MOST enzymes are quickly inactivated above about 55 °C but those from thermophile bacteria are stable for long periods at higher temperatures¹. We do not know why because so far their structures have proved too complex. For example although the tertiary and quaternary structures of the enzyme glyceraldehyde phosphate dehydrogenase from lobster muscle and from Bacterium stearothermophilus are alike their amino acid sequences differ by more than 130 out of some 330 positions which makes it hard to decide why the stearothermophilus enzyme is more stable. The electron transfer protein ferredoxin offers a better chance because its single polypeptide chain contains fewer than 60 residues; its structure is known and its heat stability and amino acid sequence have been determined in both mesophile and thermophile bacteria. We have built an atomic model of this protein, replaced its amino acid side chains in turn to correspond to the published sequences and searched for possible causes of the greater heat stability of ferredoxins from thermophile bacteria. We found that it arises mainly from external salt bridges linking residues near the amino terminus to others near the carboxy terminus. Haemoglobin A₂ a minor fraction of adult human haemoglobin which is a little more heat stable than the major fraction, haemoglobin A, seemed another good choice because its amino acid sequence differs from that of A at only 10 positions. The atomic model suggests that at only two of these positions are the replacements likely to contribute to the extra stability of haemoglobin A₂ one replacement providing an extra hydrogen bond between the α₁ and β₁ subunits and the other adding two non-polar interactions to a surface crevice within the β subunits. To account for the increased heat stability of the two proteins the extra bond energy provided by these interactions need not be larger than 10 kJ for ferredoxin or 5 kJ for haemoglobin A₂.