The mechanisms underlying seasonal timing of breeding: a multi-level approach using bi-directional genomic selection on timing of egg-laying

With climate change being one of the major threats to current biodiversity, it is essential for species to adapt sufficiently in order to survive. Some species adapt their phenology faster, in reaction to increasing temperatures, compared to others, resulting in mismatched timing. For many seasonal breeding avian species in temperate zones, such as the great tit, the reproductive period is short and coincides with warmer temperatures and increased food supplies required for successful rearing of offspring. Therefore, seasonal breeders time their reproductive cycle to the changing seasons in order to maximize reproductive success and offspring survival. With springs getting warmer earlier in the year, it is of importance for great tit females to start laying earlier to be able to raise their offspring in an optimal period (i.e. sufficient food abundance). However, females show large variation in timing of breeding, which lies in the underlying physiology: different cues are used and translated by a cascade of neuro-endocrine processes along the hypothalamic-pituitary-gonadal-liver (HPGL) axis into a laying date. Natural selection could act on this variation between females, but it is still unclear on which of the compartments (brain, ovary, liver) of the HPGL axis cues act and thus where the variation in timing between females arises. It is of importance to understand how the components of the physiological mechanism contribute to genetic variation in timing before one is able to understand how natural selection can act on timing of reproduction.

In this thesis, the main aim was to explore the molecular basis of the physiological mechanisms underlying avian seasonal timing of breeding. A promising way to do this is by comparing (extremely) early and (extremely) late laying females. In Chapter 2, I describe a large-scale selection experiment, where we created selection lines for early and late egg-laying using genomic selection. In this chapter we show that genomic selection on a complex trait such as timing of breeding is possible, because we find that the early and late selection line birds differ genomically and that this difference increases over the generations. In addition, we find that F3 generation birds differ also phenotypically, with a significant average difference in egg-laying dates of ~10 days between selection lines. 

By housing pairs of the selection lines in climate-controlled aviaries and in outdoor aviaries for two consecutive years and in contrasting environments (either artificial or semi-natural), I was able to determine that temperature has a direct effect on timing of breeding instead of via food phenology and that females laid on average earlier in the warm environment (Chapter 3). Further, because we obtained two laying dates per female, we evaluated whether our selection on laying date also changed the birds’ phenotypic plasticity and found early selection line females to initiated egg laying consistently ~9 days earlier compared to late selection line females in outdoor aviaries, but no difference in the degree of plasticity. This suggests that while natural selection may lead to a change in phenotype in the average environment it is unlikely to result in a correlated response on the degree of plasticity in timing of breeding.

I also aimed to determine whether individual differences in timing of breeding in females are reflected in differences in their molecular biology and if so where. In Chapter 4 we generated comprehensive RNA expression data from a set of three tissues important in the neuro-endocrine cascade (HPGL axis) underlying avian seasonal timing of breeding, from three different time points and from two temperature treatments and two selection lines for breeding time. Time was the strongest driver in this study, but we found an interesting interaction between time and temperature in hypothalamus, with several genes involved in circadian rhythms differentially expressed. Even though the hypothalamus has been considered the final integration point of environmental cues and guide top down hormonal regulation and in this way direct ovarian function to time breeding, we find evidence for downstream regulation of timing of breeding in Chapter 5. Differences in key reproductive candidate gene expression between phenotypically early and late laying females were found exclusively in the ovary and liver. This also suggests that adaptation in the HPGL axis to changing environments might be downstream.

The effects of the environment need to be translated into gene transcription (Chapter 4 and 5), for which DNA methylation is a likely key regulator. Therefore, in Chapter 6, we investigated in great tits whether methylation changes were tissue-specific or tissue-general and whether such methylation changes were associated with expression changes within and between tissues. Overall, we found a positive correlation between changes in DNA methylation in red blood cells and liver, both genome-wide as well as for the sites within the promoter region or transcription start site (TSS) separately. Within the TSS of genes, hyper-methylation over time in red blood cells was highly correlated with a decrease in the expression of the associated gene in the ovary. Tissue-general changes in DNA methylation could potentially be informative for changes in gene expression in inaccessible tissues.

I explored the molecular basis of the physiological mechanism underlying seasonal timing of breeding in an avian model species; the great tit. I looked at the phenotype, investigated candidate gene and genome-wide gene expression. In addition, we looked at DNA methylation (in relation to gene expression). The main conclusions are that (1) genomic selection is possible in wild populations, (2) temperature directly influences timing of breeding and (3) that timing of breeding is regulated downstream in the HPLG axis. However, we are only scratching the surface of this complex trait and further studies (also considering other ‘endo-phenotypes’ and their interactions, see Chapter 7) are necessary in order to make predictions about whether birds in general, and great tits specifically, will adapt to rapidly changing environments.