Blood, 15 September 2001, Vol. 98, No. 6, pp. 1643-1643
Conjugal interactions between consenting spectrins: predicting
which mutations will lead to hereditary hemolytic anemias
The erythrocyte membrane skeleton comprises 
/
spectrin
heterodimers that must associate into elongated heterotetramers (and higher oligomers) in order to span between actin-rich junctional complexes. Mutations leading to weakened or flawed spectrin
self-association result in hereditary elliptocytosis (HE) or hereditary
pyropoikilocytosis (HPP), which in certain homozygous states can be
lethal. Because such mutations have been mapped to either the
NH2-terminus of -spectrin or the COOH-terminus of
-spectrin, the polypeptide's self-association domain has been
hypothesized to reside within these termini. Experimental support for
this hypothesis is already considerable.
Not surprisingly, some mutations within the spectrin self-association
domain do not cause hemolytic disease. In an effort to enable
prediction of the hemolytic consequences of novel point mutations
within these self-association sequences, Zhang and colleagues (page
1645) have undertaken to model the interaction using energy minimization and molecular dynamics computational strategies. As a
starting point, the authors assumed that the docking interface would
consist of 1 -helix from the NH2-terminus of
-spectrin nestled between 2 antiparallel -helices from the
COOHterminus of -spectrin. This assumption was strongly
supported by previous experimental data and by analogy with the known
triple helical motif that constitutes most of the structure of both and spectrin. Thus the self-association site has been envisioned to
mimic the coiled-coil triple helix of the basic spectrin repeat.
Confirmation of the derived model derives from multiple observations.
First, the computational methodology allowed prediction of the known
crystal structure of the 14th repeat unit of Drosophila -spectrin. Second, the modeling strategy computed a credible structure for the self-association complex of human spectrin that closely resembled the structure of the spectrin repeats, despite substitution of over 70% of the residues in Drosophila
spectrin with frequently nonconservative amino acids from human
spectrin. And most significantly, the altered structures that were
observed upon modeling 17 of the known mutant forms of human spectrin
predicted conformational deformations whose magnitude correlated
strongly with the severity of the consequent hereditary hemolytic diseases.
So where do we go from here? First, it would make sense to test whether
other mutants not provided by nature yield a hemolytic anemia whose
severity is predicted by the modeled degree of structural distortion.
Second, it would be very satisfying to exploit the methodology to
design therapeutic agents that might repair the flawed self-association
interactions in the hemolytic anemias. And finally, where possible,
similar methods should be applied to other membrane structural
interactions with the ultimate goal of developing a predictive model of
global membrane morphology and mechanical stability.
Philip S. Low
Purdue University
