The structure of the mouse ADAT2/ADAT3 complex reveals the molecular basis for mammalian tRNA wobble adenosine-to-inosine deamination

Abstract

Post-transcriptional modification of tRNA wobble adenosine into inosine is crucial for decoding multiple mRNA codons by a single tRNA. The eukaryotic wobble adenosine-to-inosine modification is catalysed by the ADAT (ADAT2/ADAT3) complex that modifies up to eight tRNAs, requiring a full tRNA for activity. Yet, ADAT catalytic mechanism and its implication in neurodevelopmental disorders remain poorly understood. Here, we have characterized mouse ADAT and provide the molecular basis for tRNAs deamination by ADAT2 as well as ADAT3 inactivation by loss of catalytic and tRNA-binding determinants. We show that tRNA binding and deamination can vary depending on the cognate tRNA but absolutely rely on the eukaryote-specific ADAT3 N-terminal domain. This domain can rotate with respect to the ADAT catalytic domain to present and position the tRNA anticodon-stem-loop correctly in ADAT2 active site. A founder mutation in the ADAT3 N-terminal domain, which causes intellectual disability, does not affect tRNA binding despite the structural changes it induces but most likely hinders optimal presentation of the tRNA anticodon-stem-loop to ADAT2.

Sequence-independent substrate selection by the eukaryotic wobble base deaminase ADAT2/3 involves multiple protein domains and distortion of the tRNA anticodon loop

“…In the bacterial enzyme TadA, extensive remodeling of the anticodon loop is provoked by favorable base-specific interactions, resulting in splayed-out anticodon loop nucleosides bound to a rather rigid protein surface (Losey et al, 2006). Since base-specific recognition does not take place in ADAT2/3 (Auxilien et al, 1996; Ramos-Morales et al, 2021), we looked for other features that could evoke such a non-canonical anticodon loop rearrangement. We identified a eukaryote-specific molecular “RY-gate” formed by residues arginine R159 Tb ADAT2 and tyrosine Y205 Tb ADAT2 , which may provide a certain control mechanism for the tRNA to overcome before fully entering the catalytic pocket (Fig.…”

“…5b, Extended Data Fig. 4a and Extended Data Table 2) (Liu et al, 2020; Ramos-Morales et al, 2021). The earlier observation that long variable arms in tRNA Val or tRNA Ser do not affect inosine formation, is in accordance with our data as the map does not show density in the region where the variable arm would be positioned (Achsel and Gross, 1993; Auxilien et al, 1996; Ragone et al, 2011; Spears et al, 2011a).…”

“…( a ) Two views of the ADAT3 N fitted in the post-processed EM map, in blue, with ‘a’ and ‘b’ regions as the two contact points of our model with tRNA, as indicated in C. ( b ) ADAT3 N including linkers (residues 4-172); surface model colored by vacuum electrostatic potential, positive charges in red and negative charges in blue, with D-Arm of tRNA in a black cartoon representation. Residues with a role in tRNA binding are annotated: from Ramos-Morales et al, 2021 in olive and from Liu et al, 2020 in yellow. ( c ) Schematic of the tRNA with interaction contacts.…”

“…In accordance with prior propositions, we suggest an initial binding step in which RNA is promiscuously captured via electrostatic interactions by these two RNA binding domains. Mutation or deletion of either the charged KR-motif of ADAT2 C (Ragone et al, 2011) or putative RNA binding residues in ADAT3 N (Ramos-Morales et al, 2021), lead to tRNA binding defects that point towards an initial and simultaneous RNA binding of both regions (Fig. 6b).…”

“…5). As a relic of its origin, in extant ADAT2/3 enzymes, the ADAT2 enzyme is still able to form homodimers in vitro , yet these are inactive (Ramos-Morales et al, 2021; Spears et al, 2011b; PDB: 3DH1), reflecting the evolved dependency on ADAT3 for enzyme activity.…”