DNA methylation and the regulation of stimulus-specific memory formation

Abstract

Memory formation is a crucial task of the brain. It allows animals to dynamically respond to a changing environment by combining information about current and previous experiences. Thus, it promotes complex behaviours such as living in social groups and elaborate foraging tactics. The ability to form memories is present across the animal kingdom in a variety of forms. The honeybee - an eusocial insect - is capable of both ’simple’ associative and complex rule learning. Despite decades of research into the mechanisms of memory formation, however, much remains unknown. One little- understood aspect is the regulation of molecular mechanisms during memory formation. Understanding regulation of transcription is particularly important in this context, as transcription is a requirement for any stable memory. Epigenetic mechanisms regulate transcription by directly interacting with chromatin. Importantly, epigenetic mechanisms are conserved across species as distinct as humans and honeybees. This thesis aims to investigate one particular epigenetic mechanism - DNA methylation - and its role in honey bee memory formation by using behavioural, physiological and molecular assays.First, I studied the role of DNA methyltransferases (Dnmts), which catalyse DNA methylation, after olfactory reward learning. Bees were trained to associate an odour with a sugar reward. Dnmts were then blocked after conditioning using a pharmacological inhibitor. 24 hours later, bees were tested for memory retention and generalisation. Dnmt inhibition increased the generalisation to an odour that was not present during the training, thus impairing stimulus-specific memory. This effect was learning-dependent, as bees’ response to odours or sugar water alone did not change after treatment. Furthermore, this effect was robust against alterations in the training paradigm, but the directionality depended on the number of training trials.Next, I used Ca2+-imaging of the bees’ primary olfactory centre (i.e. antennal lobe, AL) to investigate whether Dnmts affect changes in AL processing established during memory formation. The AL is involved in odour discrimination and olfactory learning; both processes are also crucial for stimulus-specific memory formation. If Dnmts were inhibited after olfactory reward learning, the AL response to a new odour changed 48 hours later. This effect potentially serves as a functional explanation for the behavioural phenotype observed after Dnmt inhibition. Furthermore, this study provides the first in vivo evidence for Dnmt-mediated regulation of neuronal networks during memory formation.As DNA methylation regulates transcription, I next investigated the effect of Dnmt inhibition on gene expression. Using qPCR I studied 30 memory-associated genes. In response to Dnmt inhibition, 9 genes were upregulated after conditioning. For some of these genes I then investigated their methylation patterns using a bisulfite conversion based Mass spectrometry approach. Memory formation changed the methylation pattern compared to both unpaired (i.e. odour and sugar stimulation) and naive bees in these genes. Furthermore, I analysed the expression of Dnmts and Tet, which catalyses demethylation, after olfactory reward conditioning. Dnmt1b and Tet were upregulated after 1 hour, whereas Dnmt3 was only upregulated 5 hours after conditioning. Thus, Dnmts and Tet are likely active - in a specific temporal order - during a phase crucial for memory consolidation.Finally, I investigated genome-wide changes in DNA methylation and hydroxymethylation (i.e. intermediate of demethylation pathway) 24 hours after olfactory reward conditioning using MeDIP and hMeDIP sequencing. In order to connect and interpret the findings of my previous studies it is crucial to know which genomic regions Dnmts and Tet target during memory formation. Most memory-associated changes were located in the gene body; specifically, they were enriched in regionsflanking the transcription start and termination sites. This suggests a possible impact of DNA methylation and hydroxymethylation levels on transcription initiation and termination. Furthermore, transcription factors were enriched among the genes associated with changes in DNA methylation and hydroxymethylation. Thus, regulatory events are potentially multiplied by indirectly affecting genes targeted by these transcription factors.Overall, the results of this thesis reveal important novel aspects of DNA methylation and demethylation mediated regulation of memory formation in honey bees: (1) Dnmts are crucial for stimulusspecific memory formation and regulate cognitive, rather than perceptual, generalisation. (2) Dnmts regulate changes in odour processing in the primary olfactory centre during memory formation; likely promoting better discrimination between trained and novel odours. (3) Dnmts affect gene expression of a subgroup of memory-associated genes; but only in a few of these genes did methylation levels change as well, suggesting that Dnmts have a strong indirect effect on gene expression of memory-associated genes. This assumption is supported by the enrichment of transcription factors among genes associated with methylation or hydroxymethylation changes. (4) Dnmts and Tet are upregulated and thus likely active during a period crucial for memory consolidation. In conclusion, both DNA methylation and demethylation seem to play an important role in honey bee stimulusspecific memory formation by affecting the transcriptional landscape in neurons and consequently changing early odour processing and generalisation patterns.