Twenty ways to lose your bladder: common natural mutants in zebrafish and widespread convergence of swim bladder loss among teleost fishes

Abstract

Convergence is an important evolutionary phenomenon often attributed solely to natural selection acting in similar environments. The frequency of mutation and number of ways a phenotypic trait can be generated genetically, however, may also affect the probability of convergence. Here we report both a high frequency of loss of gas bladder (swim bladder) mutations in zebrafish and widespread convergent loss of gas bladders among teleost fishes. The phenotypes of 22 of 27 recessive lethal mutations, carried by a sample of 26 wild-caught zebrafish, involve loss or noninflation of the gas bladder. Nine of these bladderless mutations showed no other obvious phenotypic abnormalities other than the lack of an inflated gas bladder. At least 19 of the 22 bladderless mutations are genetically distinct, as shown by unique morphology or complementation. Although we were not able to obtain eggs for all 21 required crosses to demonstrate definitively that the remaining three mutations are different from all other bladderless mutations, all available evidence suggests that these mutants are also distinct. At least 79 of 425 families of extant teleosts include one or more species lacking a gas bladder as adults. Analysis of the trait's phylogenetic distribution shows that the gas bladder has been lost at least 30-32 times independently. Although adaptive explanations for gas bladder loss are convincing, a developmental bias toward bladderless phenotypes may also have contributed to the widespread convergence of this trait among teleosts. If gas bladder development in teleosts is as vulnerable to genetic perturbation as it is in zebrafish, then perhaps a supply of bladderless phenotypes has been readily available to natural selection under conditions for which it is advantageous not to have a gas bladder. In this way, developmental bias and selection can work together to produce widespread convergence.