Gastropod evolutionary rates and phylogenetic relationships assessed usingpartial 28S rDNA and histone H3 sequences

Sequence data for two segments of 28S and Histone H3 from 36 gastropod taxa , a chiton, two bivalves and Nautilus are used to test recently published m orphology-based phylogenetic hypotheses of gastropod relationships. Statist ical results suggest that the accuracy of the available hypotheses could be improved. The data support the monophyly of the Patellogastropoda (true li mpets), Euthyneura and the 'higher' vetigastropods and the polyphyly of the 'Cocculiniformia'. The division of the gastropods into two major clades (E ogastropoda and Orthogastropoda) as has been proposed on morphological grou nds is not supported, and neither the Caenogastropoda nor Heterobranchia is well supported. Within the Euthyneura, opisthobranchs are paraphyletic wit h respect to the pulmonates. The hot vent taxon, Depressigyra, groups with the lower vetigastropod Pleurotomaria in some analyses. Much of the variabi lity in the 28S rDNA segments lies in discrete areas of the sequence. Foron e of the segments, corresponding to positions 691-942 of the mosquito Aedes albopictus 28S sequence, the variable regions represent known expansion re gions (D4 and D5). For the other segment, corresponding to positions 2259-2 538 of the A. albopictus sequence, the variable area, which is found in the patellogastropods, vetigastropods and Nautilus, represents an unreported e xpansion region. The data show marked variability in the rate of evolution in both segments of the 28S rDNA, whether or not the expansion regions are included. The variability is largely clade specific. Rates are high in the patellogastropods, vetigastropods, the lower heterobranch 'Heterostropha' ( Cornirostra and Philippea), Depressigyra and the deep sea cocculinid limpet Coccopigya and substantially lower in other taxa. Rate variation in the hi stone H3 data is less extreme. The correlation between evolutionary rates i n the two 28S rDNA segments is very high, andis also significant for the th e pairing of each of the 28S rDNA segments with H3. The rate variability ma y be due to differential selection but no causative factor has been identif ied. The histone H3 data have high codon usage bias. For all amino acids en coded by multiple codons, at least some triplets occur at a frequency of le ss than a quarter of their expected usage. For all three-, four-and sixfold degenerate amino acids, the most abundant triplet occurs at least twice as frequently as expected. Despite the usage bias, there is a large amount of apparent homoplasy in synonymous alternatives at both the first and third codon positions.