The genomes of plants are marked by reoccurring events of whole-genome duplication. These events are major contributors to speciation and provide the genetic material for organisms to evolve ever greater complexity. Duplicated genes, referred to as paralogs, may be retained because they acquired new functions, or their gene products are in a dosage balance. Regulatory DNA elements - some of which are conserved across species and hence called conserved non-coding sequences (CNSs) - that control expression of duplicated genes are thus under similar purifying selection. In the present study, I have performed in-depth analyses of paralogous genes in Arabidopsis thaliana, their expression profile, their sequence conservation, and their functions, in order to investigate the relationship between gene expression and retention of paralogous genes.

Paralogs with lower expression than their duplicate were found to be under less purifying selection. A gene ontology (GO) term enrichment analysis showed that paralogs with similar expression levels were enriched in GO terms related to macromolecular complexes, whereas paralogs with different expression levels were enriched in terms associated with stress responses. Compared to homologs in the close relative Arabidopsis lyrata, the lower expressed genes had less similar expression patterns than their higher expressed paralogs. Also, lower expressed genes had greater tissue specificity and fewer CNSs in promoters, introns, and 3’ untranslated regions. These results suggest that a concurrent purifying selection acts on coding and non-coding sequences of paralogous genes in A. thaliana.

Mutational analyses of the promoters from a paralogous gene pair were performed in transgenic A. thaliana plants. The results revealed a 170-bp long DNA sequence that forms a bifunctional cis-regulatory module; it represses gene expression in the sporophyte while activating it in pollen. This finding is important for many aspects of gene regulation and the transcriptional changes underlying gametophyte development.

In conclusion, the presented thesis suggests that, following whole-genome duplication events, multiple evolutionary processes drive the retention of paralogs