Non-Sequence Data: Plasticity of gene expression and thermal tolerance: implications for climate change vulnerability in a tropical forest lizard

Study System: The slender anole is a small (<2.5 g) diurnal, semi-arboreal predator that is ubiquitous throughout the lowland tropical rainforests of Panama. Slender anoles are an ideal model organism to test hypotheses about gene expression and plasticity in tropical ectotherms because they live in lowland closed-canopy forests where they are thermoconformers (i.e., slender anoles do not behaviorally thermoregulate, Logan et al. 2020), are short-lived for a vertebrate (>95% annual mortality, Andrews & Nichols 1990) , and are easy to maintain in captivity (Stapley et al. 2011; Stapley et al. 2015; Stapley 2018).

Field-active body temperatures and environmental temperatures: During July 2019, we captured 284 slender anoles by hand or catch pole from Soberanía National Park, Panama, and measured field-active body temperatures. During the same season, we recorded environmental temperatures using 90 data loggers positioned randomly over a large portion of our field site. We then used field-active body temperature and environmental temperature data to ensure that the thermal conditions of our gene expression, physiological, and behavioral plasticity experiments were ecologically realistic. See Supplementary Materials for details on our field-active body temperature and environmental temperature methods.

Phenotypic plasticity, growth, and energetics under long-term warming We used a greenhouse experiment to assess the potential for phenotypic plasticity under long-term warming. We captured 40 lizards (equal sex ratio) in June of 2019 and transported them back to the Smithsonian facility in Gamboa, Panama. After acclimation to laboratory conditions for 48 hours, we measured standard morphological traits (mass and SVL) as well as a suite of physiological and behavioral traits. We measured voluntary thermal maxima (VTmax; an index of heat tolerance), critical thermal minima (CTmin; an index of cold tolerance), and body temperatures in a laboratory thermal gradient following Logan et al. (2020) (see Online Supplementary Materials for detailed methods). We calculated the mean, minimum, and maximum body temperatures chosen in the gradient as different aspects of thermal preference that might display plasticity. After measuring phenotypes, we randomly assigned 10 males and 10 females to either a control or warm greenhouse (total of 20 lizards per greenhouse). Lizards were placed into 23 cm x 23 cm x 46 cm mesh cages (one individual per cage) which were themselves placed inside the greenhouse. In the control greenhouse, we set the thermostat to 24°C for the first five days, 25°C for next five days and 26°C for the remaining 21 days, adjusting the temperature to recreate the natural forest thermal regime as closely as possible. In the warm greenhouse, we ramped the thermostat from 24 to 30°C over a period of 14 days and then held the thermostat constant at 30°C for the final 17 days. Here, we were mimicking the gradual onset of a heat wave. Hereafter, we refer to the treatments of the long-term experiment as either long-term warming or control. We monitored the temperatures experienced by lizards in the greenhouses by taking 549 surface body measurements during the study period using a Fluke infrared temperature gun. We used surface body temperature as our estimate of lizard temperature because the measurement of surface temperature does not require the handling of lizards and stress from handling can affect experimental results (Foss et al. 2012). We verified that surface temperatures closely approximated cloacal temperatures (figure A1 available online) by measuring both of these variables on a subset of lizards during the second week of the experiment (n = 120, r = 0.88). After four weeks in the greenhouses, we remeasured CTmin, VTmax, and thermal preference in a laboratory thermal gradient for 35 individuals (2 lizards from the control treatment died and 3 lizards from the heat treatment died during the experiment). We also calculated growth in terms of both SVL and mass. We then used residual body mass from a linear regression of body mass on SVL as an index of body condition (Logan et al. 2012). Finally, we dissected all individuals and weighed organs associated with energy storage and reproduction, including visceral fat bodies, livers, and gonads. We analyzed growth and organ mass using linear regressions, including sex, treatment, and sex by treatment interactions as predictor variables. Body size covariates were included in models when appropriate. Differences between initial and final values (plasticity) for thermoregulatory and thermal tolerance traits were determined using a repeated-measures analysis of variance (ANOVA). Prior to analyses, we ensured that all variables fit the assumptions of statistical tests. All statistical analyses were completed in JMP v. 12.0 (JMP 2019). Modeling the impact of thermal tolerance plasticity on activity time under climate warming We modeled how phenotypic plasticity of thermal tolerance might alter potential activity time under climate warming. We used data and equations from Neel et al. (2021) to predict lizard body temperatures based upon projected environmental temperatures by the end of the century, assuming a 3 °C increase by the year 2100 (IPCC 2018). We assumed a uniform increase in temperature of 0.0365 °C per year over that time period. We then projected future activity levels by assuming that lizard activity would cease at mean environmental temperatures exceeding VTmax. We projected activity time using values of VTmax for either the control (VTmax = 29.1 °C) or warm (VTmax = 29.7 °C) treatment. Activity time was expressed as the percent change in activity time relative to 2019, which was assigned a value of 100%.