Almost all mature eukaryotic mRNAs contain a 3′ polyadenosine (poly(A)) tail, which promotes nuclear export of mRNAs, protects transcripts from exonucleolytic decay, and increases translation efficiency. As the poly(A) tail is involved in many steps of gene expression, its length is highly regulated. Poly(A) tail shortening, known as deadenylation, represses gene expression by initiating mRNA decay and inhibiting translation. Deadenylation is required for mRNA homeostasis; moreover, specific transcripts can be targeted for deadenylation, enabling gene expression to respond to external stimuli. In eukaryotes, cytoplasmic deadenylation is mainly carried out by two conserved multiprotein complexes: Pan2-Pan3 and Ccr4-Not. This thesis focuses on the in vitro characterisation of Pan2-Pan3. While Pan2-Pan3 and Ccr4-Not show poly(A) preference, it was unknown how the conserved exonucleases Pan2 and Caf1 recognise poly(A). In Chapter 2, structural and biochemical studies reveal that these DEDD-family deadenylases recognise the intrinsic stacked, helical structure of single-stranded poly(A) RNA. Recognition of this conformation in different biological contexts, enabled by the unique physicochemical properties of adenine, suggests a possible reason for poly(A) conservation. Chapter 3 examines the in vitro activity of Pan2-Pan3 and how it is regulated, particularly by the cytoplasmic poly(A)-binding protein Pab1. These studies elucidate domain requirements and species-specific features of Pab1-dependent regulation of Pan2-Pan3. My experiments further examine how Pan2-Pan3 may be recruited by sequence-specific RNA-binding proteins. Together, these results provide insights into Pan2-Pan3 deadenylation and how this dynamic process could be regulated and thereby influence post-transcriptional gene expression.