UFMylation pathway, conservation of UBA5, and generation of fly Uba5 LoF alleles

(A) A diagram showing the UFMylation pathway. Details of the biochemical processes in the pathway are described in the main text. In UBA5 proteins, only the adenylation domains are shown in the diagram.

(B) A diagram of the UBA5:UFM1:UFC1 complex. In the complex, two copies of UBA5 form a homodimer that interacts with UFM1 via a trans-binding mechanism. The activation of UFM1 requires the adenylation domain of one UBA5 subunit and the UFM1-interacting sequence (UIS) of the other UBA5 subunit in the complex. The opposing protomer of the UBA5 homodimer also contributes a UFC1-binding sequence (UBS) that is required for UFM1 transthiolation.

(C) Alignment of the human UBA5 and fly Uba5 protein sequences. The functional domains of UBA5 are marked in colored boxes. The DEE44-associated variants are marked in the protein topology diagram and the protein sequence alignment (letters in red).

(D) Generation of the Uba5T2A-Gal4 allele and the uses of the allele in flippase (FLP)-mediated conversion. The expression of the GAL4 to drive a fluorescent protein allows assessment of gene expression, and humanization of the flies by expression of human UBA5 cDNA.

(E) Generation of Uba5 null allele by CRISPR-mediated indel formation.

(F) Loss of Uba5 causes lethality in early developmental stage. The lethality is rescued by a genomic rescue construct, the expression of FLP (Uba5T2A-Gal4 mutants only), and the expression of human UBA5 cDNA.

Uba5 is expressed in a subset of neurons and glial cells in fly CNS

(A) The expression of nuclear localized mCherry (mCherry.nls) driven by the Uba5T2A-Gal4 allele (Uba5T2A-Gal4 > mCherry.nls) shows that Uba5 is expressed in L3 larvae and adult flies.

(B and C) The larval CNS and adult brain of Uba5T2A-Gal4 > mCherry.nls animals were immunostained with a neuronal (Elav, Panel B) or glial marker (Repo, Panel C). Maximum projections of confocal z- stack images are shown. Single plane, high magnification images of the regions indicated by the dashed squares are shown on the right to visualize the colocalizations between mCherry and the immunostaining signals. Arrows indicate cells that colocalize both markers. Scale bar, 100 μm.

DEE44-associated variants exhibit different rescuing abilities in flies

(A) The DEE44-associated UBA5 variants rescued the lethality of Uba5 mutant flies with varying efficiency. Uba5T2A-Gal4/FM7 females were crossed with UAS-UBA5 males and the viability of Uba5T2A-Gal4/Y; UAS-UBA5/+ progenies were measured by Mendelian ratio and indicated by color codes: red, zero viability; yellow, partial viability (< 90% of expected number); green, full viability (90% and above).

(B) Three variants caused developmental delay in Uba5T2A-Gal4/Y; UAS-UBA5/+ flies. The embryos were collected within 6 hours and the number of eclosed adult flies was counted at the same time every day. Three replicates were performed in each group.

(C) Five variants caused reduced lifespan in Uba5T2A-Gal4/Y; UAS-UBA5/+ flies.

(D) Five variants caused progressive climbing defects in Uba5T2A-Gal4/Y; UAS-UBA5/+ flies. Flies were tested on Day 7 and Day 30. The climbing activity of CantonS wildtype flies is shown as reference. Numbers of animals (n values) in each group are indicated under the bars.

(E) Three variants caused a bang-sensitive phenotype in Uba5T2A-Gal4/Y; UAS-UBA5/+ flies. Flies were tested on Day 30. The bang-sensitivity of CantonS wildtype flies is shown as reference. Numbers of animals (n values) in each group are indicated under the bars.

(B-E) Flies were cultured under 25 °C. The results of DEE4 variant-expressing flies are compared with the result of reference UBA5-expressing flies. Results are presented as means ± SEM. Statistical analyses were performed via two-sided, unpaired Student’s t-test. ns, not significant;

*p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

Summary of phenotypes of humanized flies expressing UBA5 variants

Structural analysis of UBA5 variants

(A) Composite model of a UBA5 homodimer (green and blue) bound to ATP (grey sticks), UFM1 (magenta), and UFC1 (gold). The model was built using a series of UBA5 complex structures with UFM1 and UFC1 (PDB 6H77, 7NW1, and a modelled UBA5:UFC1 complex(Kumar et al., 2021; Soudah et al., 2019). Functional residues comprising the active site cysteines of UBA5 and UFC1, as well as the C-terminus of UFM1 are shown in yellow spheres. UBA5 variants are shown in red spheres and are labeled with their predicted structural effects.

(B) Close-up view of variants (red sticks) within the UBA5 active site (yellow sphere), ATP binding pocket, and homodimerization interface.

(C) Close-up view of variants (red sticks) expected to impact UBA5 protein stability (results shown in the following figures).

Preparation and stability of UBA5 variant proteins

(A) Coomassie-stained SDS-PAGE analysis of all purified UBA5 variant proteins.

(B) Thermal shift assay measuring the melting temperature (Tm) of all UBA5 variant proteins, with the exception of p.Gly168Glu and p.Cys303Arg which could not be produced. The p.Gln312Leu variant displayed two melting curves. Experiments were performed in triplicate over three biological replicates.

(C) Change in melting temperature for all UBA5 variants in the presence of 5 mM ATP. Upon ATP addition, the p.Gln312Leu variant transitioned to a single melting curve. Experiments were performed in triplicate over three biological replicates.

(B-C) Statistical analyses were performed via unpaired Student’s t-test. ns, not significant; *p<0.05;

**p<0.01; ***p<0.001; ****p<0.0001.

Measuring UFM1 activation and transthiolation with UbiReal

(A) Cartoon schematic illustrating the complexes formed during UFM1 activation and transthiolation, as well as their expected molecular weights. The fluorescent group attached to UFM1 is denoted by an orange star.

(B) Proof-of-concept UbiReal assay monitoring the fluorescence polarization (FP) of Alexa488- labeled UFM1 alone (species 1), following addition of UBA5 (species 2), and following addition of UFC1 (species 3).

(C) Fluorescence scan of samples described in (B) separated by SDS-PAGE, illustrating the formation of activated UFM1 complexes.

(D) UbiReal assay tracking UFM1 activation by reference and variant UBA5 proteins over time.

(E) Area Under the Curve quantification of UFM1 activation performed at 22 °C. Experiments were performed in triplicate over three biological replicates. Statistical analyses were performed using a Welch’s t test with comparison to the reference UBA5 data.

(F) As in (E), for reactions performed at 37 °C.

(G) UbiReal assay tracking UFM1 transthiolation for reference UBA5 and variants that showed little or no effect on activation.

(H) Area Under the Curve quantification of UFM1 transthiolation performed at 22 °C. Experiments were performed in triplicate over three biological replicates. Statistical analyses were performed using a Welch’s t test with comparison to the reference UBA5 data.

(I) As in (H), for reactions performed at 37 °C.

(E-F, H-I) Statistical analyses were performed via Welch’s t-test. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

Summary of protein stability and functions of UBA5 variants