Phenotypic variation is ubiquitous in natural populations, yet many of the processes and mechanisms underlying such variation are still little understood. The fitness consequences of phenotypic variation are especially important at early-life stages. However, phenotypic variation at early life-stages is rarely considered for evolutionary diversification. Fishes are the most species-rich clade of vertebrates, and offer an array of skull and jaw morphologies that reflect their ecological speciality (1–3). One of the most famed examples of evolutionary diversification in fishes comes from the African rift-lake cichlids where adaptive radiations in feeding morphology has contributed to their rapid diversification (4–7). If variation in trophic structures exist early in life, then individuals may be able to specialize upon particular resources, resulting in increased phenotypic variation within a population (8), providing fuel for natural selection. For example, it is hypothesized that individual specialization in relation to different diets may reduce intraspecific competition (9, 10), which can over time facilitate divergence of trophic morphology and the evolution of resource polymorphism (i.e. discrete intraspecific morphs that have diverged upon different resources; (11–14)). Ultimately, resource polymorphism may facilitate sympatric speciation (14–16) and provide valuable insight into evolutionary processes underlying diversification. Glacial retreats have provided a window of opportunity to examine such processes in polymorphic Northern freshwater fishes, whereby their degree of phenotypic divergence often varies between different systems (17). What generates phenotypic variation within a single population is still poorly understood, although these processes may represent the first steps of divergence between morphs. By focusing on individual variation at very early life-stages (i.e. prior to onset of feeding), this study aims to understand which factors promote variability in trophic morphology within a single population of Arctic charr (Salvelinus alpinus), and potentially shed light on the mechanisms underlying early stages of evolutionary diversification.
The extent of intraspecific divergence can be biased along both genetic and developmental lines of least resistance (18, 19), expanding upon existing phenotypic variation, but limited by those available resources (20). Early development provides a more malleable phenotype than later in life (18), as gene expression is more dynamic (21–24) and can influence developmental trajectories (discussed in: (24–26)). Studying gene expression combined with morphology at early life-stages can thus provide insight into developmental processes underlying the initiation of phenotypic variation. Gene expression patterns have been linked to phenotypic divergence between different morphs in several taxa (26, 26–31), but patterns within morphs have been less well-studied. Variation in gene expression can also be associated with parental effects, whereby differential distribution of maternal resources (i.e. egg size) can contribute to trophic specialization (28, 32, 33).
Parental effects – when the phenotype of an individual is affected by the phenotype or environment of its parents (34) – are a common source of phenotypic variation at early life-stages both by genetic and non-genetic means (35, 36), and are crucial during early development (34). The extent to which offspring phenotype is influenced by maternal effects can depend on multiple maternally transmitted factors such as yolk, mRNA transcripts or other cytoplasmic factors packaged in the egg by the mother during oogenesis (11, 37). Studying early developmental stages can not only reveal the impact that parental effects might have on phenotypic variation of their offspring, but also gives insight into how phenotypes vary before exposure to external sources of environmental variation, such as diet. Importantly, trophic morphologies seen in developing fish embryos may reflect genetic and non-genetic parental effects and are determinant for survival of individuals at first feeding.
The propensity of Arctic charr for intraspecific diversity makes this species well suited for studying mechanisms that promote and/or precede the evolutionary origins of phenotypic diversity, especially associated with trophic polymorphism (38, 39). Although several studies have examined associations between parental effects, gene expression and phenotypic variation between established morphs (see references above), which are sometimes visible at early-life stages (40, 41), there are relatively few (if any) studies examining such associations within a morph, or a population. This study uses a single morph of Arctic charr (Vatnshlíðarvatn brown; (42)) to examine the association between early life-stage phenotypic variation, family (incl. direct genetic and/or maternal effects) and gene expression at hatching (H) and first feeding (FF). Previously, we showed that early life-stage gene expression was very dynamic at those developmental stages, and that there was a correlation between offspring size and the relative expression of two genes associated with craniofacial divergence between benthic and pelagic morphs of Arctic charr: Star and Sgk1 at H (24). Sgk1 was the only gene that showed a linear relationship, having higher expression in larger embryos at H. Such findings highlight the need to combine gene expression patterns with patterns of morphological variation to further our understanding of the mechanisms involved in early phenotypic divergence.
Here we combine our previous findings on dynamic early life-stage gene expression (24) with patterns of craniofacial morphology within an Arctic charr morph. We use acid-free double staining of cartilage and bone, coupled with geometric morphometrics, to test whether individuals from different families developed differently, resulting in different trophic morphologies at H and FF. We use eight growth-related genes chosen from the literature based on their involvement during early development, and six genes related to skeletogenesis based on findings in a previous study (27). Combining this previously collected gene expression data (24) with data on offspring morphology, we test the following predictions: 1) if there are genetic or parental effects in shape at early life-stages, we should see differences among families in craniofacial features; 2) if early life-stage variation is related to genes involved in growth and skeletogenesis of trophic structures, offspring craniofacial shape should covary with the expression of such genes; 3) if maternal investment (i.e. egg size) influences early life-stage phenotypes, female mean egg size or individual offspring size should correlate with offspring morphology; and finally 4) relative expression of the studied growth and skeletal related genes early in life can result in shape variation later in life.