Linking Phylogeny and Ecology to Enigmatic Freshwater Mussel Shell Morphologies

Abstract

Virtually all organisms rely on external morphology to interact with the environment, which makes observable differences between species, the fundamental unit of biodiversity, an exceptionally useful means of classification. Consequently, high morphological variation within-species regularly leads to ecological and phylogenetic hypotheses that do not reflect evolutionary history. Studies describing phylogenetic splits between species that are nearly indistinguishable are becoming more common and not only do species with a high degree of morphological similarity create headaches for biologists, but it illustrates a need to combine multiple lines of evidence to discover species, understand their evolution, and inform their conservation. I investigated patterns of interspecific and intraspecific morphological trait diversity in a group of North American freshwater mussels to better understand how biogeographic history and environment influence speciation and ecological interactions. In my first chapter, I worked to reconstruct phylogenetic relationships in the Pleurobemini, a tribe of mussels that contains multiple species with shell morphologies that heavily overlap. Adding to the confusion, several species also exhibit a pattern of shell morphology that is linked to hydrologic gradients. Historically, these shell morphology differences caused back-and-forth taxonomic changes that have become even more complex as molecular sequencing has revealed unrecognized species diversity. Recent genetic studies have attempted to provide some clarity to this situation, but taxonomic changes have thus far remained incomplete. In an effort to help rectify the taxonomy of some pleurobemines, I conducted extensive geographic sampling of the genus Fusconaia and some closely related species as the first step to building a backbone phylogenetic framework. I used a combination of DNA barcoding, traditional shell morphometrics, and low coverage whole genome sequencing, also known as genome skimming, to overcome previous sampling limitations. With my barcoding data, I constructed a taxonomically comprehensive phylogeny and identified samples that were good candidates for genome skimming. I extracted mitochondrial genomes and developed a set of target-capture probes from genome skimming data then conducted an in-silico target-capture sequencing experiment with hundreds of ultraconserved elements (UCEs). I obtained key pieces of information from maximum likelihood and species tree phylogenetic analyses of mitochondrial genomes and UCEs; specifically I found mitonuclear discordance. While molecular evidence can be strongly supportive of taxonomic hypotheses, a purely molecular approach is insufficient to make robust determinations. Before making taxonomic determinations, I conducted morphometric analyses of five poorly sampled lineages and confirmed strong correlations between shell inflation and stream size in the species Fusconaia flava. After accounting for variation in hydrologic habitat, morphometric analyses showed mean shape was different at the species level but also identified strong morphological overlap at the individual level. Overall, the phylogenetic discordance I found between two specific lineages most closely fit a pattern of mitochondrial introgression but conducting additional testing for hybridization went beyond the capabilities of my datasets. Although my analyses had almost an adverse effect by deepening the taxonomic mystery, I demonstrated genome skimming may be a useful approach for future studies. For my second chapter, I used both tree-based and non-tree-based methods to more completely explore species boundaries between species in the Fusconaia flava complex. I generated additional genomic data using the “3RAD” restriction-site associated DNA sequencing approach and explicitly sampled individuals I previously identified as possible hybrids as well as species I did not expect there to be gene flow between. I obtained thousands of single nucleotide polymorphisms (SNPs) that allowed me to reconstruct phylogenetic relationships, perform population-level genomic analyses, and generate a time-calibrated phylogeny. Using several approaches including a sparse non-negative matrix factorization method, principal component analysis, and the program TreeMix, I was able to infer population splits and estimate admixture. Based on the admixture coefficients I estimated, I used a phylogenetic network approach called SNaQ to test for hybridization and reticulate evolution. To estimate divergence times, I used the SNAPP module implemented in BEAST to generate a time-calibrated species tree. Lastly, I conducted population genomic analyses focusing on populations in the Red River in Oklahoma and tested for an isolation-by-distance effect. What I discovered was that the F. flava species complex appears to be highly structured and gene flow does not appear to be widespread. Most importantly, I was able to rule out hybridization and incomplete lineage sorting as the mechanisms causing mitonuclear discordance and instead my analyses strongly suggested a pattern of mitochondrial introgression was causing mitonuclear discordance in two Fusconaia lineages. My population genomic analyses also suggested isolation-by-distance was unlikely in Red River Fusconaia and instead aligned more closely with patterns of reproductive isolation. Although I did identify isolated instances of probable hybridization, gene flow appears to be spatially limited. These low levels of gene flow seemingly align with my estimated divergence times which show speciation likely occurred during the Pliocene and early Pleistocene epochs. In my third chapter, I wanted to understand if there was an underlying significance to a pattern of shell shape variation known as Ortmann’s Law of Stream Position. Ortmann’s Law of Stream Position is a hypothesis that posits laterally compressed shells found in small creeks become more inflated (wider) as the stream transforms into a large river. I previously confirmed shell inflation increases as a function of flow rate but did not investigate changes in other traits like shell thickness. From a functional standpoint, I wanted to identify if shell thickness and shell width were correlated and if stouter shells found in large rivers are more advantageous by acting as a sort of anchor, whereas compressed shells allow the mussel to move more freely. To investigate shell thickness, I digitized museum specimens with micro-computed tomography (micro-CT scanning) and extracted hydrologic data for each collection locality. In conjunction with my micro-CT scanning study, I conducted a fluvial experiment with wild Fusconaia individuals from the Gasconade River, USA to identify how small river and large river analogs respond to different flows. Using historical hydrology measurements from the Gasconade River, I scaled flume conditions to mimic natural high and low flow events and quantified locomotion by measuring valve movement with Hall sensors. Across six trials, I found statistically significant evidence that mussels from small river habitats make more locomotory movements than mussels from large river habitats. Interestingly, I also found large river mussels appeared to move primarily during high flow which suggests the objective of their movements may not be to burrow but to stabilize themselves during high flows. My hypotheses are also supported by my comparison of shell thickness which showed median thickness is highly correlated with shell width which indicates wider shells are also denser. Consequently, denser-shelled mussels appear to be able to withstand higher dislodgement forces whereas thinner-shelled mussels are able to displace sediments more efficiently to take refuge in the streambed. Finally, in my fourth chapter I synthesized the results of my first and second chapters to formally revise the taxonomy of two Fusconaia lineages. I resurrected the species Fusconaia hebetata and formally described a new species which I named Fusconaia vaughnae. This research is a significant contribution to systematic malacology and is part of a growing body of work designed to more completely understand freshwater mussel evolution. I have demonstrated how phylogenetic frameworks underpin ecological hypotheses, which is important for understanding the origin and function of anatomical features. Holistically, I have demonstrated that underlying patterns of morphological variation are more complex than we typically perceive them to be. Freshwater mussels in particular are a fascinating study system that can offer a glimpse into past speciation events and life in riverine environments. However, in the face of extinction from the worst effects of climate change and habitat destruction, mussels also serve as a sort of “canary in the coal mine” and remediating taxonomic issues surrounding Fusconaia will undoubtably facilitate more accurate taxon-based conservation of a group of mussels containing cryptic diversity and endangered species.