All cellular protein synthesis - which is indispensable for cell growth, proliferation, and viability - is carried out by ribosomes. The production of sufficient functional ribosomes is essential for development and homeostasis, and reductions in the abundance of functional ribosomes can lead to developmental diseases known as ribosomopathies.

Historically, translation has been viewed as a constitutive process, with ribosomes being machines with little to no independent influence on gene expression. However, translation is now known to be highly regulated, and growing evidence has shown that ribosomes can be heterogeneous in composition and may contribute to this regulation. In eukaryotes, cytoplasmic ribosomes consist of two large ribonucleoprotein subunits, 40S and 60S, which are composed of four different ribosomal RNAs (rRNAs) and 79 ribosomal proteins in total. Ribosome heterogeneity can arise from chemical modifications of rRNA, post-translational modifications of ribosomal proteins, incorporation of different ribosomal protein paralogues, or interaction with different ribosome-associated factors. This heterogeneity has been hypothesised to underlie the existence of “specialised ribosomes”, which could exist in specific tissues or developmental stages and may have the potential to preferentially translate specific subsets of mRNA transcripts. However, while many examples of ribosome heterogeneity have been described to date, there are very few examples to date where this has been conclusively linked to functional differences in development.

As a result of gene duplications, the Drosophila melanogaster genome encodes 14 pairs of ribosomal protein paralogues. Within these pairs, one gene (referred to henceforth as “canonical”) is usually expressed ubiquitously and is essential for viability: mutations in these genes cause homozygous lethality, while heterozygotes usually display the minute growth phenotype. At the outset of this project, little was known about the “non-canonical” paralogues, however a number of them have been shown to be expressed in a tissue-specific manner, frequently being enriched in the germline. The aim of this thesis is to systematically examine the duplication of ribosomal protein genes within the Drosophilidae, and the potential roles of the 14 “non-canonical” ribosomal protein paralogues in D. melanogaster.

To characterise the patterns of ribosomal protein gene duplication in Drosophila, I designed a pipeline in collaboration with Dr Daniel Gebert to identify duplications of “canonical” ribosomal protein genes in 12 Drosophila species, which span 70 million years of evolution. I identified 388 independent duplication events, with most taking place by retroposition. My results indicate that duplicates generally appear to be short-lived, although some have persisted for over 70 million years.

To examine the role of each of the 14 “non-canonical” ribosomal protein paralogues in D. melanogaster, I used CRISPR/Cas9 to systematically mutate each gene. Phenotypic analysis of the mutants revealed that none of the 14 “non-canonical” paralogues are required for viability or induced the minute phenotype. Given their tendency to be expressed in a germline-specific manner, I examined the fertility of the mutants. For most genes, fertility was unaffected in mutants, except for mutations in RpS5b, which led to female sterility. I show that loss of RpS5b results in strong activation of the Tor pathway and remodelling of germline metabolism. In addition, my experiments reveal that this germline stress response is transduced to the neighbouring somatic epithelium, leading to overgrowth, disorganisation, incomplete Notch activation, and non-autonomous activation of Tor kinase. I show that, in conjunction with the metabolic changes, these phenotypic alterations trigger the activation of the mid-oogenesis checkpoint, resulting in germline death and complete female sterility.

Interestingly, the “canonical” RpS5a and the “non-canonical” RpS5b genes have different and somewhat complementary expression profiles during germline development. Therefore, I set out to test whether the phenotypic alterations in RpS5b mutants result from either the loss of a “specialised” RpS5b-containing ribosome or, rather, from a lack of RpS5 protein in general. To do so, I used CRISPR/Cas9 and homology-directed repair to seamlessly replace the entire coding sequence of RpS5b with that of RpS5a, and vice-versa. These experiments definitively demonstrated that RpS5a and RpS5b proteins are functionally equivalent during oogenesis, revealing that the RpS5b mutant phenotype results from a general lack of RpS5 protein, rather than a potential functional divergence between these paralogous proteins. Altogether, my work highlights how ribosome abundance must be examined whenever investigating potential cases of “specialised” ribosomes.