Complex Evolutionary and Genetic Patterns Characterize the Loss of Scleral Ossification in the Blind Cavefish Astyanax mexicanus

Abstract

The sclera is the tough outer covering of the eye that provides structural support and helps maintain intraocular pressure. In some fishes, reptiles, and birds, the sclera is reinforced with an additional ring of hyaline cartilage or bone that forms from scleral ossicles. Currently, the evolutionary and genetic basis of scleral ossification is poorly understood, especially in teleost fishes. We assessed scleral ossification among several groups of the Mexican tetra (Astyanax mexicanus), which exhibit both an eyed and eyeless morph. Although eyed Astyanax surface fish have bony sclera similar to other teleosts, the ossicles of blind Astyanax cavefish generally do not form. We first sampled cavefish from multiple independent populations and used ancestral character state reconstructions to determine how many times scleral ossification has been lost. We then confirmed these results by assessing complementation of scleral ossification among the F1 hybrid progeny of two cavefish populations. Finally, we quantified the number of scleral ossicles present among the F2 hybrid progeny of a cross between surface fish and cavefish, and used this information to identify quantitative trait loci (QTL) responsible for this trait. Our results indicate that the loss of scleral ossification is common–but not ubiquitous–among Astyanax cavefish, and that this trait has been convergently lost at least three times. The presence of wild-type, ossified sclera among the F1 hybrid progeny of a cross between different cavefish populations confirms the convergent evolution of this trait. However, a strongly skewed distribution of scleral ossicles found among surface fish x cavefish F2 hybrids suggests that scleral ossification is a threshold trait with a complex genetic basis. Quantitative genetic mapping identified a single QTL for scleral ossification on Astyanax linkage group 1. We estimate that the threshold for this trait is likely determined by at least three genetic factors which may control the severity and onset of lens degeneration in cavefishes. We conclude that complex evolutionary and genetic patterns underlie the loss of scleral ossification in Astyanax cavefish.

On Intraspecific and Interspecific Variation in Teleost Scleral Ossification

“…This hybridization likely explains the variability in ossicle number observed in our samples, although a previous study of laboratory‐reared Chica individuals found no evidence for scleral ossicles in this group (O'Quin et al, ). Although the average size of many of these samples was below common length (data not shown), suggesting that individual differences in growth could account for these differences ossicle number (Franz‐Odendaal et al ., ; O'Quin et al, ), we found that body size accounted for only 45% of the variation in scleral ossicle number among these individuals. In addition, at the sizes sampled (4.3 cm on average), both ossicles, if present, should have started to ossify (Franz‐Odendaal et al, ); and, indeed, we found above average sized individuals with 0, 1, and 2 ossicles, suggesting that some of the variation in ossicle number found among these individuals is indeed genetic.…”

“…Although this distinction is subtle, we believe it is important for three reasons. First, this result was unexpected based on previous results that found only individuals with zero scleral ossicles from this cave (n = 3; O'Quin et al, ), suggesting that further sampling among additional populations may be useful for uncovering additional variation. Second, it was unexpected given our hypothesis that intraspecific variation would be limited by the epistatic nature of the genes that control scleral ossification which favors the presence of scleral ossicles (O'Quin et al, ; Lyon et al, ), suggesting that this epistasis has been overcome by considerable backcrossing, which is consistent with previous reports of introgression of cavefish alleles in this population (Strecker et al, ; Gross, ).…”

“…First, this result was unexpected based on previous results that found only individuals with zero scleral ossicles from this cave (n = 3; O'Quin et al, ), suggesting that further sampling among additional populations may be useful for uncovering additional variation. Second, it was unexpected given our hypothesis that intraspecific variation would be limited by the epistatic nature of the genes that control scleral ossification which favors the presence of scleral ossicles (O'Quin et al, ; Lyon et al, ), suggesting that this epistasis has been overcome by considerable backcrossing, which is consistent with previous reports of introgression of cavefish alleles in this population (Strecker et al, ; Gross, ). Third, we have found that a single population of fishes recapitulates much of the phenotypic diversity in scleral ossicle number (either 0, 1, or 2 scleral ossicles) found among most orders of teleosts (Franz‐Odendaal, ; Mok and Liu, ), suggesting that allelic variation in the genes that control scleral ossicle development are sorting among the members of this population.…”

“…We sampled multiple individuals of 10 riverine fish species from Kentucky in order to assess the level of intraspecific variation in scleral ossicle number found among teleosts. We hypothesized that intraspecific variation in these species would be low to nonexistent, with the exception of an outgroup species, Astyanax mexicanus , because most family‐level variation is low (Franz‐Odendaal, ; Mok and Liu, ) and because any phenotypic variation should be constrained by the complex genetic architecture underlying this trait (O'Quin et al, ; Lyon et al, ). Following a previous report by Franz‐Odendaal (), we also further examined patterns of interspecific variation in scleral ossicle number by using phylogenetically generalized least squares to evaluate the relationship between scleral ossicle number and depth preference in over 100 teleost species while also accounting for the phylogenetic relationships among these species.…”

“…In this species, surface morphs exhibit two scleral ossicles, whereas cave morphs generally exhibit none (Yamamoto et al, ; but see Dufton et al, ). Further sampling of cave morphs from different caves has revealed additional inter‐population level variation in scleral ossicle number (Dufton et al, ; O'Quin et al, ), although no variation within populations.…”