Guppies as a model system
Guppies have become a model system in ecology and evolutionary biology
due to repeated, independent adaptation of natural populations to
distinct predation environments (Endler, 1995; Haskins, Haskins,
McLaughlin, & Hewitt, 1961; D. Reznick, Butler IV, & Rodd, 2001). In
Trinidad, high- and low-predation population pairs from different river
drainages represent independent evolutionary lineages (Barson, Cable, &
Van Oosterhout, 2009; Fraser, Künstner, Reznick, Dreyer, & Weigel,
2015; Gilliam, Fraser, & Alkins-Koo, 1993; Willing et al., 2010) in
which colonization of low-predation environments has led to parallel,
adaptive changes in life history traits, morphology, and behavior
(Endler, 1995; Magurran, 2005; D. A. Reznick, Bryga, & Endler, 1990; D.
Reznick et al., 2001; D. N. Reznick, 1997). For example, low-predation
guppies shoal less tightly (Huizinga, Ghalambor, & Reznick, 2009;
Magurran & Seghers, 1990b, 1991), escape more slowly (Ghalambor,
Reznick, & Walker, 2004), are slower to re-commence movement following
a predator encounter (Elvidge, Ramnarine, & Brown, 2014; Harris,
Ramnarine, Smith, & Pettersson, 2010), and perform fewer predator
inspections (Magurran & Seghers, 1990a, 1994) than their high-predation
counterparts.
Building on decades of work comparing high- and low-predation
populations from the wild, more recent laboratory breeding designs have
disentangled genetic and environmental influences of predation on
phenotypic differences. These studies demonstrate that a combination of
genetic and environmental influences shape guppy life history (Torres
Dowdall et al., 2012), morphology (Fischer, Soares, Archer, Ghalambor,
& Hoke, 2013; C. A. Handelsman, Ruell, Torres-Dowdall, & Ghalambor,
2014; Ruell et al., 2013; Torres-Dowdal, Handelsman, Reznick, &
Ghalambor, 2012), physiology (Fischer, Harris, Hofmann, & Hoke, 2014;
Corey A. Handelsman et al., 2013), and behavior (Fischer, Ghalambor, &
Hoke, 2016b; Huizinga et al., 2009; Torres-Dowdal et al., 2012). Yet
despite many years of work at the level of organism level phenotypes,
the mechanisms underlying adaptive phenotypic differences in guppies
remain largely unexplored. A single study characterized brain gene
expression in multiple populations from the same lineage during the
earliest stages of adaptation (~3 generations after
colonization of low-predation environments) and found a negative
relationship between phenotypic plasticity and adaptive divergence
(Ghalambor et al., 2015).
In the current study, we compare the effects of genetic background
(high- versus low-predation populations) and developmental environment
(rearing with and without predator cues) on brain gene expression
patterns in two parallel, independent evolutionary lineages of guppies.
These lineages diverged at least 600,000 years ago, with subsequent,
more recent colonization of low-predation environments by high-predation
ancestors within each river drainage (Fajen & Breden, 1992; Willing et
al., 2010). We here compare whole-brain gene expression patternswithin and between lineages based on evolutionary history
with and rearing with predators. We focus on gene expression
specifically in the brain because the previous gene expression study in
guppies used brain tissue (Ghalambor et al., 2015) and because of
previous evidence for behavioral plasticity in our focal populations
(Fischer et al., 2016b).
To test our hypotheses we (1) quantify the extent of overlap in genes
exhibiting significant expression plasticity as well as genetic
expression divergence within each drainage, (2) examine associations in
the direction of plastic and genetic expression differences, (3) ask
whether expression plasticity itself evolves, and (4) assess whether
parallel phenotypic adaptation across lineages relies on shared gene
expression mechanisms. First, if developmental plasticity does indeed
predict genetic divergence, we expect a within-lineage association in
gene identity and/or expression direction among those genes showing
plastic and genetic expression differences. Second, if gene expression
plasticity itself evolves, we expect the extent and/or direction of
plastic responses to differ between genetic backgrounds. Finally, if
only a few gene expression configurations (i.e. mechanistic paths) can
give rise to shared adaptive behavioral phenotypes, we expect parallel
adaptation to low-predation habitats to be characterized by parallel
evolution in a set of genes that are largely shared between lineages. In
contrast, if transcriptional mechanisms are flexible, we expect gene
expression divergence in largely non-overlapping gene sets. Taken
together, our results allow us to assess the flexibility of
transcriptional mechanisms of adaptation across timescales – from
developmental plasticity, to genetic divergence, to parallel adaptation
across lineages.