Alexander Iain Tulloch Lapresa

Sex is ubiquitous among eukaryotes, yet the long-term advantage of sex has yet to be theoretically determined. Crucially, genes are not selected individually, but collectively as part of a genome. Individual genomes produce through mutation both beneficial and deleterious alleles of these genes, which often must be transcribed (even relative to environmental conditions) to generate a fit phenotype. Sexual reproduction is the mechanism separating these genes from each other, both bringing beneficial alleles together and making the removal of deleterious alleles more efficient. Any benefit of sexual reproduction must come from this generation of adaptive allele combinations. Meiosis is the process whereby alleles on homologous chromosomes are exchanged. To do so, double strand breaks (DSBs) in the chromosome are repaired through homologous recombination as either crossovers (CO) or non-crossovers (NCOs), both events potentially forming new allele combinations. Multiple methodologies indicate the rate of meiotic crossovers varies across the genome in all species, forming a “recombination landscape”. In many species this recombination landscape changes across different environments, which may have adaptive potential. The aim of my thesis was to study a possible explanation for changes in meiotic landscapes across environments, namely a possible association between transcription and meiotic recombination. To test this hypothesis, we attempted to induce differential meiotic recombination landscapes from changes in gene transcription in both Saccharomyces paradoxus and Arabidopsis thaliana. In Chapter II, my coauthors and I review the role of the histone methylation mark H3K4me3 as a target for catalysis of DSBs during meiosis across eukaryotes. In all eukaryotes H3K4me3 is associated with transcribed gene sequences, which in most studied species means meiotic double strand breaks, and by consequence crossover formation is recruited to transcribed DNA sequences. As the expression of genes is often dependent on environmental context, H3K4me3 marks may at least partially explain changes in recombination landscape. Whether an association of transcription and meiotic recombination has adaptive benefits derived from associating meiotic recombination to DNA sequences under selection, as well as its specific association to transcription start sites, remains to be theoretically explored. Chapter III describes the new CREPES algorithm used to detect meiotic recombination using variant positions in short-read sequences. CREPES classifies reads (or paired-end reads) as either recombinant, non-recombinant, or uninformative. To validate this method, we applied CREPES to 250 bp paired-end reads produced from bulk meiotic progeny of a Saccharomyces paradoxus cross between parents diverging at 1.59 % of sequence positions. We chose to count only reads that contained at least two consecutive variant positions from each parental type, avoiding possible sequencing errors. As these reads are relatively short, our method cannot distinguish COs from NCOs, for which gene-conversion events span ~ 2kbp. Classification of reads as informative or non-informative, and the filtering out of optical duplicates reduced the effective coverage from ~1,200X to ~320X. Despite ~85% of the genome being covered by informative reads, we could only detect a maximum of ~ 50% of the expected recombination events detected previously in comparable crosses. This likely (but unverified) lack of sensitivity could be due to a lack of information at the ~15% of the genome showing high levels of homology, particularly given that there may be a bias in recombination towards homologous sequences. To explore the positions of meiotic recombination, we divided the reference genome in 500-bp windows and divided the recombinant paired-read coverage by the total informative coverage. Correlations within samples were low at the 500-bp window level, but increased with window size, making the study of recombination at the gene level and above a feasible exercise. Despite the limitations of CREPES, it may be possible to detect changes in recombination landscape particularly using windows larger than 2.5kbp. Chapter IV uses the method developed in Chapter III to explore the changes in recombination landscape at the gene level between different environments. We grew the S. paradoxus strain from Chapter III in glucose at 30 °C, glucose at 25 °C and galactose at 30 °C to induce differential transcription. RNAseq confirmed differential transcription between the samples, however also surprisingly yielded transcriptional signatures of early meiosis in Glucose at 30 °C. We then induced sporulation via nitrogen and carbon starvation in KAc and sequenced the progeny in bulk. To analyse recombination, we estimated the ratio of recombinant read coverage at 2.5-kbp windows centred at the transcription start site of genes. Although we could detect significant differences in Spearman correlation of window-specific recombination rates between conditions and between high and global transcription windows, only a low correlation was found between recombination rate and transcription, and no effects attributable to differential transcription. There are changes in meiotic recombination landscape associated to the somatic growth environment, and a weak general association between transcription and recombination, but no association between changes in transcription and recombination between the different treatments. In Chapter V, we studied the effect of hormone application on the recombination rate in Arabidopsis thaliana using “traffic lines”. To track recombination in regions associated with environmentally induced genes, we chose A. thaliana traffic lines containing DsRed and GFP fluorophores. Eight lines were selected with fluorophore-coding genes spaced between four and nineteen centimorgans apart. In three of the lines, the fluorophore-delimited region spanned methyl-jasmonate responsive genes, two responsive to salicylic acid, alongside another three selected for other criteria. We sprayed heterozygous lines with solutions of methyl jasmonate and salicylic acid over the course of flowering, estimating recombination rates within the eight windows flanked by markers. Pilot testing of differential transcription using qPCR showed differential induction of a series of target genes in wild-type plants. To detect overall transcriptional response of the plants to the hormone treatments, we sequenced total RNA from flower meristems above the first open flower, including all meiocytes. No significant changes in transcription rate attributable to hormone application were observed despite visible phenotypic differences in the plants. As could be expected from this result under our hypothesis, we detected no significant changes in recombination in any of our samples. The amount of variation in recombination rate that we observed was comparable with previous studies. More targeted sampling methods and smaller windows of analysis may be necessary to detect an impact of transcription effects on meiotic recombination. In conclusion, we did not detect changes in the flower transcriptome nor in meiotic crossover rates attributable to hormone application and therefore could not reach any significant conclusion on the association between transcription and recombination in A. thaliana. In conclusion, the hypothesis laid out in Chapter II has not yet been properly tested. This is due to a combination of incomplete recombination data from CREPES (described in Chapter III), inappropriate environmental conditions (in Chapter IV), and a combination of excessively large analysis window sizes and inappropriate tissue sampling (in Chapter V). Despite these setbacks, Chapter IV did show genome-wide recombination rates change in different environments, warranting further exploration of changing recombination landscapes. To do so would require studying recombination at a higher resolution either using marker-assisted selection on a small number of genes, ChIPseq associated to meiotic DSBs, or whole-genome sequencing of individual progeny.