Multitrait diversification in marine diatoms in constant and warmed environments

Abstract

Marine diatoms are important primary producers that can undergo multitrait changes in response to environmental or demographic perturbations. Here, we investigated how flexible multitrait correlations are over hundreds of generations, and how lineage diversity and phenotype diversity are related. We used a perturbation-evolution experiment with Thalassiosira diatoms in which they were 'knocked-off' their ancestral multitrait phenotype peak using repeated population bottlenecks, and were then allowed to grow unhindered as large populations. We found that most multitrait phenotypes were isolated but that in 10% of cases, phenotypes were connected by viable trajectories. Following the bottleneck perturbation, evolving in a +4 degrees Celsius environment favoured the discovery of novel multitrait phenotypes, while evolving in the ancestral environment most often resulted in the ancestral phenotype re-evolving. Finally, both trait values and trait correlations evolved in this experiment, with the most drastic changes in trait correlations being associated with lower growth rates. Our findings demonstrate that demographic events that result in decreases in fitness, such as population bottlenecks, have the potential to play an important role in the realised connectivity of phenotypes, as well as in shaping the relationship between lineage and phenotypic diversity in diatoms.

Genus-Wide Transcriptional Landscapes Reveal Correlated Gene Networks Underlying Microevolutionary Divergence in Diatoms

“…We can experimentally investigate the contribution of chance events to molecular and trait variation in many microbes, including diatoms, with standard bottleneck experiments. Here, we generated trait variation in globally distributed marine diatom strains as a function of chance events in a constant environment using population bottlenecks (see Materials and Methods; samples from the same experiment were also taken by Hinners et al (2022) to measure whole-cell physiological traits only). Physiological assays for trait values reported in Hinners et al (2022) were previously developed and published in Argyle et al (2021a , 2021b) .…”

“…Here, we generated trait variation in globally distributed marine diatom strains as a function of chance events in a constant environment using population bottlenecks (see Materials and Methods; samples from the same experiment were also taken by Hinners et al (2022) to measure whole-cell physiological traits only). Physiological assays for trait values reported in Hinners et al (2022) were previously developed and published in Argyle et al (2021a , 2021b) . In this manuscript, we describe global shifts in transcript values and identify the expression of specific genes and pathways that drive most of the observed variation.…”

“…S1 ). We then measured changes to biogeochemically important whole-cell trait values such as cell size, before, during, and after recovery of bottlenecks ( Hinners et al 2022 ) (see overview for a synopsis of traits and findings in Hinners et al 2022 ). This approach enabled the examination of trait divergence without relying on (or allowing) divergent selection because of environmental differences.…”

“…b) Experimental design where 6 ancestral strains with 6 replicates each were bottlenecked and then full selected prior to RNA-Seq; b) was adapted from Fig. 1 from Hinners et al (2022) to indicate that this RNA-Seq data were generated from the same evolution experiment as the trait data in Hinners et al (2022) . c) Phylogenetic tree of diatoms in study and in MMETSP ( Keeling et al 2014 ) based on concatenated alignment of BUSCO Protista_83.hmm marker set.…”

“…there are many ways to have any given cell size. Because of this, measuring whole-cell trait values (as in Hinners et al 2022 ), such as cell size or elemental composition, may underestimate the genetic, epigenetic, or transcriptional variation present following bottlenecks and recovery. It is also necessarily biased in that only a small number of possible functional traits can be assessed.…”