Anole Annals

Anolis distichus is well-known in the anole world for the high degree of ecomorphological variation within the species, especially in dewlap color. In fact, there are 18 described subspecies! While there is some gene flow between various subspecies and populations, the phenotypic differences are maintained, which suggests strong selection. But the fine-scale genetic structure underlying these traits is not well understood. Anthony Geneva and colleagues decided to explore the genomic basis of adaptive divergence in a well-described hybrid zone between two A. distichus subspecies. The first, A. d. ignigularus, has a white dewlap, and occupies a dry forest habitat while the second, A. d. ravitergum, has a red dewlap and inhabits a wetter habitat. The two subspecies occur along a transect from dry to wet, and they hybridize in a narrow contact zone in the middle. These two subspecies provide a great system to explore the link between adaptive and genetic divergence.

Anolis distichus. Photo by Rich Glor

Geneva sequenced individuals using RNASeq across an environmental transect from wet to dry, including allopatric and sympatric populations of both species. He examined levels of divergence and introgression to explore which genomic loci might be the basis for the ecological adaptive divergence between these two species. He found a suite of candidate genes that differ between the two subspecies, as well as several that show signs of introgression between the two. Interestingly, several of the divergent genes are involved in two traits that likely are impacted the environment – insulin signaling, which may relate to metabolic differences between hot and cool climates, and vision, which may relate to differences in light availability and signal efficiency. Most of the introgressed genes, on the other hand, relate to conserved pathways, suggesting that these genes play similar roles in both subspecies.

Adpative divergence in anoles has been a topic of interest for a long time, and Geneva’s study provides and a valuable insight to the genetic basis of this interesting phenomenon.

On each of the Greater Antillean islands, habitat-specialist Anolis ecomorphs have independently evolved complex suites of shared phenotypes and behaviors. This remarkable convergence has motivated the work of generations of anolologists. With anoles entering the once-exclusive club of genome-enabled organisms, a new line of investigation has become possible: Is the convergence observed in anole ecomorphs caused by molecular convergence? Such convergence can take many forms, including shared changed at individuals sites, or shared changes in the rates of protein evolution of individual genes.

Russ Corbett-Detig of UCSC sought to answer this question using whole-genome sequence data from 12 species – four from each of the Trunk-Ground, Trunk-Crown, and Grass-Bush ecomorphs drawn from different islands and different evolutionary lineages. Accurately detecting molecular convergence is fraught and much recent research has focused on avoiding pitfalls that could lead to a positively misleading inference of convergence where none actually exists. Previous studies have trumpeted amazing cases of molecular convergence in a variety of animals, only to be later shown to be artifacts of data analysis.

Corbett-Detig did everything right. He used null models that account for the expected background levels of convergence caused by processes other than natural selection. He found no evidence of extra shared non-synonymous mutations in any of the three ecomorph groups. Similarly, he found no signal of shared changed in protein evolution in Trunk-Ground or Trunk-Crown but Grass-Bush anoles seemed to share elevated rates of changes in many genes. This result was exciting, but Corbett-Detig dug deeper and discovered a new way this type of analysis could be mislead – two of the four Grass-Bush anoles exhibited accelerated evolution across their entire genomes and, as a result, seemed to share faster rates at more genes than expected by chance. When Corbett-Detig corrected for this bias, the signal of convergence disappeared.

While this result was in one sense disappointing, it is also fascinating and suggests the evolutionary pathways to shared ecomorphological traits are numerous and strongly influenced by contingency. Furthermore, anole ecomorphs have evolved such a stunning set of similarities that other forms of convergence like genome structure, gene family expansion, or convergence in gene regulation may still hold the key to understanding the genetic basis the remarkable convergence of Anolis ecomorph classes.

Sexually antagonistic selection occurs when traits are beneficial for one sex, but detrimental to the other. This commonly occurs in species with sexual dimorphism, such that one trait is positively correlated with fitness in one sex, and negatively correlated with fitness in another. But in many organisms, the sexes do not become dimorphic until maturity – that is to say, juveniles all look pretty much alike, even when adults show clear differences between males and females. Which leads to the question: how does sexually antagonistic selection change over an organisms’ lifespan? Research from studies of Drosophila flies suggests that this is the case, but the question hasn’t been well-studied in vertebrates.

Everyone’s favorite anole, Anolis sagrei

Until now. In his Evolution talk, Aaron Reedy (University of Virginia) described his work testing whether sexually antagonistic selection changes over ontogeny using our favorite workhorse of evolutionary ecology, the brown anole (A. sagrei). Anolis sagrei are sexually dimorphic, with adult male body sizes up to 30% larger than females, but juveniles are monomorphic. Reedy and colleagues  sampled A. sagrei on several small islands in a Florida watershed four times a year, capturing thousands of adults and juveniles. They measured the body size of all lizards captured, and combined this morphological data with survivorship data to determine how selection was acting on body size in adults and juveniles.

They predicted that juvenile males and females would experience concordant selection, while adult males and females would experience antagonistic selection. And this is exactly what they found: for juveniles, body size was correlated with survival in the same way between sexes. But in adults, this was not the case. In the first year of sampling, there was no selection on body size for adult females, but positive selection for males, such that bigger males survived better. Interestingly, during the second year of sampling, the relationship flipped – females experienced positive selection on body size, and males experienced negative selection. The reasons for this shift are uncertain, but the main point is clear – sexually antagonistic selection does indeed change over ontogeny. Reedy et al. are planning to follow up this great new research by expanding their study to look at more islands and more traits to get at the finer points of these selective differences, so stay tuned!

Oriol Lapiedra opened up the penultimate day of Evolution by discussing his results of a recent field experiment in the Bahamas. In this project, Lapiedra and colleagues evaluated how inter-individual variation in behavior – specifically risk-taking – influenced survival. To do this, the research team took advantage of a well-understood model system in evolutionary ecology: brown anoles (Anolis sagrei) on islands with and without anole-predators (curly-tailed lizards; Leiocephalus carinatus) in the Bahamas. Male and female brown anoles were collected and subjected to a behavioural trial which measured the amount of time it took for a lizard to leave a refuge after being exposed to a predator. These observations were used to quantify each individual’s propensity to take risks. For example, those individuals that left their refuge shortly after seeing a predator were interpreted as being more ‘risky’ than more conservative individuals. Following these trials, each lizard was x-rayed to assess morphology and individually tagged, before being released onto one of 4 predator-free islands or 4 predator-present islands, all of which were currently void of anoles.

Lapiedra et al. started with a priori hypotheses that overall survival would be lower on those islands with predators, and those that did survive would be individuals considered less risky. After waiting 4 months, the research team returned to the Bahamas to collect all lizards from each island and see which individuals had survived. The authors report that, as expected, overall survival was lower on islands with predators, and that there was a significant relationship between behaviour and survival such that high risk-taking individuals had much lower survival when predators were present. This suggests that under those biotic conditions, natural selection operates against those riskier phenotypes. On closer inspection, this relationship was largely driven by a strong relationship in females, with no significant relationship existing between risk-taking behavior and survival of males.

Lapiedra et al. then contrasted these results by independently assessing how morphology was related to survival. The authors found that both risk-taking behavior and morphology influenced survival, however – and, important to this study – the relative effect of an individual’s risk-taking behaviour was much more influential on survival.

At last night’s poster session, undergraduate Derek Briggs (U. Mass. Boston) presented findings from his senior capstone project in which he looked at several traits related to dominance and health. Using a dataset of x-rays and dewlap photos collected over a 4 year period from various urban and forest sites across Puerto Rico (by Kristin Winchell), Derek looked at body size, body condition, dewlap size, and injury rates (broken bones and missing digits) to see if there was a difference in frequency between urban and forest habitats.

Derek and his co-authors chose these traits because they thought they might be impacted by shifts in lizard density and distribution in the urban habitat which may lead to increased male-male competition. Specifically, in urban habitats, lizards tend to perch closer to one another because the potential perches are more clustered. This increase in local density could lead to increased encounter rates and fights over optimal perch sites, food resources, or mates. Derek hypothesized that this shift in distribution should lead to shifts in these traits, although he did not have a prediction about the direction of these shifts.

Derek Briggs with his poster.

Derek found that urban lizards were consistently larger than forest lizards in terms of snout-vent-length (SVL) but that body condition (mass~SVL) did not consistently differ between sites. Although all paired populations had significant differences in body condition, in some municipalities lizards were fatter in urban habitats and in some they were fatter in forests. In terms of dewlap size, Derek did not find any significant trends, although he still has quite a few dewlap photos to analyze still, so stay tuned!

In terms of injuries, Derek did not find significant differences between forest and urban animals for bone breaks or missing digits. However, these are rare events to begin with, so it is possible that a much larger sample size is needed to detect a difference. His findings do suggest a trend of more bone breaks in urban populations, and more missing digits in forest populations. He attributes this trend to either elevated male-male competition in urban habitats or differences in predator communities.

We look forward to seeing the full results from Derek’s honors thesis.

Kristin Winchell gives her talk on urban anoles at Evolution 2017.

When I think of Puerto Rico, the first thoughts that come to mind are of sunny beaches and lush rainforests. There are, however, also lots of urban habitats in Puerto Rico. San Juan, for example, has two million human residents, and also lots and lots of anoles. Doctoral candidate Kristin Winchell has been studying adaptation in urban anoles for several years. Last year, she published¹ her work showing that Anolis cristatellus in urban habitats have longer hindlimbs, bigger toe pads, and more lamellae than lizards in rural habitats.

A connection that was missing, however, was how the morphological shifts she documented related to performance differences in urban versus rural habitats. To get at this question, she conducted sprinting trials with different substrates to see how limb and toe characteristics affect sprinting capacity. Lizards in urban habitats use much smoother perches, such as fences and posts, and so the hypothesis was that the longer limbs and toe pad differences she detected improved sprinting performance on smoother substrates. She used three different substrates for sprinting trials – bark (rough surface), metal (smooth surface), and painted concrete (very smooth surface). She found that, overall, lizards sprinted more slowly on more slippery substrates. On average, lizards sprinted at 60% of their maximum capacity, indicating a strong performance hit when using slippery substrates.

Kristin confirmed that the urban anoles were better at sprinting on all substrates – including the slippery ones – than rural anoles. When she explored the results in greater detail, she found that only lamella number explained variation in sprint performance, with no appreciable effects of limb length or toe pad area. Kristin’s elegant study demonstrate how we can document evolution on recent timescales, and shows how urban environments provide strong selective pressures for the animals that live in them.

1. KM Winchell, RG Reynolds, SR Prado‐Irwin, AR Puente‐Rolón, LJ Revell. 2016. Phenotypic shifts in urban areas in the tropical lizard Anolis cristatellus. Evolution 70:1009-1022

As a lineage splits and diversifies, species’ traits diverge in different ways.  For example, as anoles diversified in the Caribbean, trunk-ground anoles’ bodies become muscular and stocky, trunk-crown anoles’ heads become long and thin, and grass anoles’ tails become long and slender. This process of adaptation to different environments seems simple and intuitive, but the evolution of traits is not so simple.

Most traits don’t evolve independently – changes in one trait are often correlated with changes in another trait, which can constrain a species’ response to selection. This correlation between traits is represented by the genetic variance-covariance matrix (G matrix). The size, shape, and orientation of the G matrix determine the speed and direction of morphological change, and defines the “line of least genetic resistance” along which a species can evolve. But of course, as species diverge and their traits shift, the correlations between these traits themselves may not stay constant – that is to say, the G-matrix itself can evolve. Which means that G represents both a constraint on evolutionary change, as well as a product of evolution itself. So does the G matrix evolve along with species divergence, or does it limit morphological evolution?

In his talk at Evolution 2017, Joel McGlothlin (Virginia Tech) described his efforts to address these question in anoles. As a poster child of adaptive radiation, Anolis provides an excellent opportunity to explore the dynamics of G matrix evolution and evolutionary constraint. To that end, McGlothlin and colleagues estimated G matrices for seven anole species (no easy task), including representatives from three ecomorph categories. He laid out the following question: has the G matrix evolved as Anolis diversified? Or do we see a signature of constraint conserved across anoles?

First, McGlothlin and colleagues found that the G matrix has indeed evolved in the course of Anolis diversification: the shape, orientation, and size of the G matrix was different for each species studied. More closely related species had more similar G matrices, and there was a weak link between ecomorph and G matrix structure, but overall, G was clearly different across the seven anole species. This suggests that trait correlations (and therefore species’ potential responses to selection) are not necessarily constant across the anole radation.

However, despite this overall divergence, one important aspect of the G matrix – its orientation – was similar across all anole species sampled. This suggests that the line of least genetic resistance has remained constant throughout the diversification of anole ecomorphs, and is deeply conserved. So even though individual species’ trait correlations have changed as anoles have diverged, the signature of morphological constraint has persisted. The study provides a fascinating illustration of the complexity of morphological evolution, and provides a fresh new link between micro- and macro- evolutionary processes in Anolis lizards.

When, why, and how does speciation take place? Travis Ingram, professor at the University of Otago in New Zealand, tackled this question in his talk at Evolution 2017 (and in this paper) by examining Anolis speciation in the context of anoles’ most enigmatic trait–the dewlap.

Anolis sagrei with its dewlap extended. Photo by Bonnie Kircher.

Ingram posited that we can think of relationships between speciation rates and the value of particular traits in two ways. One possibility is that the value of a particular trait in a lineage influences the probability that that lineage speciates, trait evolution facilitating speciation. Conversely, particular traits may be especially likely to diversify at speciation events, in response to speciation.  Ingram tested these two hypotheses in Anolis, crowd-sourcing photographs of outstretched anole dewlaps  to quantify dewlap size and ending up with analyze-able dewlap size information for 184 species from across the whole clade.

Ingram detected no relationship between speciation rates and dewlap size,  indicating no evidence for dewlap-size-dependent speciation in anoles (possibility 1 above). However, probing a bit further, Ingram considered why bigger dewlaps may be related to speciation rates–what if a bigger dewlap allows for greater pattern complexity, allowing more species to coexist by accessing more axes along which their dewlaps can diverge? Quantifying dewlap complexity as the number of colours on a dewlap, Ingram did find a relationship between size and complexity, but curiously, more complex dewlaps were linked to lower, and not higher, speciation rates. Why remains a mystery. Suggesting evidence for speciational evolution (possibility 2 above), 34% of dewlap size evolution was associated with speciation events. Intriguingly, this pattern was driven almost entirely by mainland and not island anoles.

In sum, though the precise processes linking speciation and dewlap evolution remain rather enigmatic, it seems to me that Ingram’s macroevolutionary approach has given us a number of directions in which to take microevolutionary and behavioral ecological studies to understand why dewlaps vary in the ways that they do!

Sexual dimorphism, or phenotypic differences between the sexes, is characteristic of nearly all animal species. Males and females often differ in size, shape, color, and many other morphological and behavioral phenotypes. This dimorphism can often make it difficult to study selection on various phenotypic traits – how do you measure selection on a trait accurately when that trait may be expressed differently in each sex?

Anolis sagrei exhibits sexual dimorphism. (Photo by Bob Reed)

In a talk at the annual Evolution meeting, Robert Cox and Joel McGlothlin help us answer this question. Using dewlap and skeletal measurements – which differ widely between males and females – and data from breeding experiments on Anolis sagrei, they examine the quantitative genetic architecture of these sexually dimorphic traits. Using a matrix-based model, which accounts for genetic correlations between and within sexes, Cox and McGlothlin are also able to see how these sexually dimorphic traits react to a variety of selection regimes, including selection that acts in opposite directions in males and females. In addition, using these simulations, they are able to estimate how different traits can be evolutionarily constrained: genetic correlations between the sexes appear to constrain selection on skeletal phenotypes, but not dewlap-related phenotypes.

These methods are likely to be extremely useful to anyone hoping to measure selection in natural population of anoles, or any other sexually dimorphic species. Sex differences often play an important role in how an organism can evolve in the wild, and introducing them into the way we quantify selection and its response is a key contribution to understand this process. I encourage anyone interested in the details of this method to check out the recent paper by the authors below for more details!

Cox, R. M., Costello, R. A., Camber, B. E., & McGlothlin, J. W. (2017). Multivariate genetic architecture of the Anolis dewlap reveals both shared and sex‐specific features of a sexually dimorphic ornament. Journal of Evolutionary Biology.

We all wish anoles were invincible, but, sadly, they aren’t. Sofia Prado-Irwin’s poster at the Evolution 2017 meeting discussed one of anoles’ putative foes–the adenovirus. Adenoviruses infect a wide diversity of hosts, from amphibians to mammals, and though they are well characterized in captive and domesticated populations, we know very little about their evolution in the wild.

Sampling opportunistically from deceased animals in a breeding colony of Anolis sagrei as well as from one fecal sample, Prado-Irwin (Harvard University) was able to examine the prevalence of adenovirus in lizards caught on six different Bahamian islands. In particular, she was curious about three questions:

• Was the mortality of animals in the breeding colony associated with adenovirus?
• Is adenovirus present in anoles in the wild?
• Does adenovirus coevolve with its hosts? In other words, does the phylogeny of A. sagrei from these 6 islands match the phylogeny of those animals’ viruses? Or perhaps, instead, the geographic distance between hosts’ islands explain how strains of adenovirus are related to one another?

Extracting genomic DNA and then amplifying virus-specific genomic regions, Prado-Irwin was able to show that adenovirus was certainly found in wild as well as lab-housed animals. However, mortality was unlikely to be due solely to the virus–only 23% of the deceased animals were infected. Finally, there was no evidence for for the adenovirus phylogeny matching either the lizard hosts’ phylogeny or tracking their geographic distribution. Instead, adenoviruses seem to shift hosts readily, with some A. sagrei adenovirus protein sequences being more closely related to mammalian adenovirus strains than to other anole strains! In a nutshell, virus evolution is complicated, and much remains to be learned about these submicroscopic maybe-destroyers of our favourite lizards.