Conservation of Nonsense-Mediated mRNA Decay Complex Components Throughout Eukaryotic Evolution

Abstract

Nonsense-mediated mRNA decay (NMD) is an essential eukaryotic process regulating transcript quality and abundance, and is involved in diverse processes including brain development and plant defenses. Although some of the NMD machinery is conserved between kingdoms, little is known about its evolution. Phosphorylation of the core NMD component UPF1 is critical for NMD and is regulated in mammals by the SURF complex (UPF1, SMG1 kinase, SMG8, SMG9 and eukaryotic release factors). However, since SMG1 is reportedly missing from the genomes of fungi and the plant Arabidopsis thaliana, it remains unclear how UPF1 is activated outside the metazoa. We used comparative genomics to determine the conservation of the NMD pathway across eukaryotic evolution. We show that SURF components are present in all major eukaryotic lineages, including fungi, suggesting that in addition to UPF1 and SMG1, SMG8 and SMG9 also existed in the last eukaryotic common ancestor, 1.8 billion years ago. However, despite the ancient origins of the SURF complex, we also found that SURF factors have been independently lost across the Eukarya, pointing to genetic buffering within the essential NMD pathway. We infer an ancient role for SURF in regulating UPF1, and the intriguing possibility of undiscovered NMD regulatory pathways.

Figure 1

NMD factors and PIKKs have been retained and lost throughout eukaryote evolution. (A) Key steps in the NMD process. (1) The transient SURF complex (SMG1 + UPF1 + eRFs) forms at a ribosome stalled at a stop codon. In the SURF complex, SMG8 and SMG9 prevent UPF1 activation by inhibiting its phosphorylation by SMG1. SMG1, SMG8 and SMG9 are also important for recruiting UPF1 into the SURF complex. (2) Following interaction between UPF1 and the core NMD factors UPF2 and UPF3, the SURF complex is remodelled into a transcript decay-initiating complex. This disrupts the SMG8/SMG9-mediated inhibition of SMG1, allowing it to phosphorylate and activate UPF1, which interacts with transcript decay-inducing factors, including SMG5-7 (3). (B) A total of 312 eukaryote genomes were surveyed for the presence or absence of genes encoding core NMD factors, PIKKs and subunits of the SURF complex. Numbers to the right indicate the number of individual species within a particular eukaryotic group that were examined. The summarized phylogeny to the left shows the relationships between the major eukaryotic groups studied. The percentage of species within each group where genes encoding these factors were found is represented by heat maps, as shown to the right. Where only a single representative genome was available for a particular group, only presence (grey box) or absence (white box) is shown. (C) Venn diagram showing the number of eukaryote genomes from which genes encoding components of the SMG1C (SMG1, SMG8 and SMG9) were found to be absent. (D) The coding capacity for the entire SMG1C has been lost at the base of major eukaryote lineages (represented by filled stars on the phylogeny). Independent losses of the SMG1C was also found within clades (open stars on phylogeny).

Figure 2

The fungal kingdom has retained SMG1. A total of 94 fungal genomes were surveyed for the presence or absence of genes encoding the individual members of the SMG1C (SMG1, SMG8 and SMG9). The summarized phylogeny to the left shows the relationships between fungal groups studied. The percentage of species within each fungal group where SMG1C-encoding genes were found is represented by the heat map, as shown to the right. Where only a single representative genome was available for a particular group, only presence (grey box) or absence (white box) is shown.

Figure 3

Plants show independent losses of SMG1. A total of 50 green plant genomes were surveyed for the presence or absence of genes encoding the individual members of the SMG1C (SMG1, SMG8 and SMG9). The phylogenies to the left show the relationships between the plants studied. (A) All major green plant groups have retained the gene encoding the PIK kinase SMG1. Loss of other components of the SMG1C vary between lineages. The percentage of species within each plant group where SMG1C-encoding genes were found is represented by by the heat map, as shown to the right. Where only a single representative genome was available for a particular group, only presence (grey box) or absence (white box) is shown. (B) Within the angiosperms (flowering plants), SMG1 loss is restricted to the Brassicas, which includes the model plant Arabidopsis thaliana. Gene presence (red box) or absence (white box) is shown for each species.

Figure 4

The domain structure of SMG1 proteins from the Amorphea and Diaphoretickes. The SMG1 kinase is characterised by three domains, the FAT domain (orange), the kinase domain (red) and the FATC, located at the C-terminus of the peptide (green). In addition, SMG1 contains a large C-insertion of approximately 1000 amino acids (yellow), situated between the kinase domain and the FATC. The structure of SMG1 is shown for the Excavata (N. gruberi), the Amorphea and the Diaphoretickes. Note that in the Diaphoretickes the C-insertion is consistently present, while in the Amorphea it is missing from SMG1 in several species.

Figure 5

Mapping of S/TQ dipeptide motifs in UPF1 proteins from across the eukaryotes. The UPF1 protein contains a number of domains. Towards the N-terminus is a RNA helicase motif (black). Near the centre of the peptide is a DEAD-like DNA helicase domain (light grey). C-terminal to the DNA helicase is a AAA domain (dark grey). Red arrow heads indicate the approximate positions of SQ and TQ dipeptides within the UPF1 peptide sequence, which are thought to be the target sites for phosphorylation by the SMG1 kinase. Representatives from the major eukaryote groups are shown, as indicated to the left. The presence or absence of the SMG1 gene in the genome of the relevant species is shown to the right. Orange bars at the bottom of the figure highlight S/TQ clusters maintained in species where loss of SMG1 is predicted to have been an ancient event.

Figure 6

Model depicting the proposed evolution of protein complexes that regulate UPF1 for NMD. The model predicts the co-existence of the SURF complex and an alternative UPF1-regulatory (AUR) complex in the last eukaryotic common ancestor (left). Each complex was maintained throughout the evolution of all major eukaryote lineages. The presence of both complexes allows for genetic buffering in the regulation of UPF1 activation. It also permits independent evolutionary trajectories for the two complexes, since loss of one regulatory complex can be buffered by the presence of the other complex (right). Since the identity of AUR is unknown, it is also possible that only SURF was present in the LECA, and different alternative mechanisms to activate UPF1 could have evolved independently. Additionally, AUR may activate UPF1 through phosphorylation, since phosphorylation of UPF1 has been shown to be important in two species lacking SMG1, but other mechanisms to activate UPF1 may also exist.