In their BioEssays article entitled ''Evolution of the Spectrin Repeat,'' Pascaul et al. describe the results of phylogenetic tree building for the repetitive protein unit found within the human representatives of the cytoskeletal proteins β£-actinin, β£-spectrin, and β€-spectrin. Their results corroborate the previously proposed block duplications for these proteins identified using dot plots and lend further support to the model, which proposed a common origin for these proteins along with dystrophin from an β£-actininlike ancestor. What warrants discussion, and is the purpose of this letter, concerns the significance of identifying such block duplications.
In two contemporaneous papers, we show that the use of such trees can successfully be extended to the analysis of all the completed β£and β€-spectrin sequences, including the novel β€-spectrin isoform β€ Heavy -spectrin (β€ H ). Our analysis also detected traces of block duplications that are similar in location and scope to the other analyses, and we were also able to detect a single duplication event that extended the β€-spectrin repeat array by 13 units to the 30 seen in β€ H . Furthermore, our analysis of the β€ H Src Homology 3 domain shows that is most closely related to that found in β£-spectrin lending further credence to the common origins of β£and β€-spectrin. We also proposed that β£and β€-spectrin originated through DNA rearrangement; however, we suggest that this event was a single unequal exchange between the homodimeric spectrin ancestor and a pleckstrin homology (PH) domain containing protein. This parsimonious model achieves the same results as the ''promoter insertion'' model of Pascual et al., but it also explains the origin of the C-terminal PH domain that all β€-spectrin genes encode.
Rather than limiting such analyses to attempts to try and piece together the exact sequence of events leading to the current structure of such repeat arrays, we feel that the identified block duplications should be seen as an example of the types of mechanism at work during the evolution of tandem arrays of protein motifs. These mechanisms-unequal crossing over and gene conversion-are essentially driven by DNA identity, as they require misalignment during meiosis. Given that such mechanisms were operational in these tandem arrays, they would have led to the phenomenon known as concerted evolution, whereby the arrays would tend to stay relatively homogeneous with the occasional sweep-through of variants. Accordingly, we would expect that successive rounds of homogenization would often erase the evidence for the earliest duplications leaving only traces of the most recent events. Furthermore, while there is substantial heterogeneity between repeat lengths today, the dystrophin arrays are thought to have had a 109-amino acid progenitor, while the spectrin arrays are based on 106 amino acids and such