Rapid assessment of the temporal function and phenotypic reversibility of neurodevelopmental disorder risk genes in Caenorhabditis elegans

Abstract

Recent studies have indicated that some phenotypes caused by decreased function of select neurodevelopmental disorder (NDD) risk genes can be reversed by restoring gene function in adulthood. However, few of the hundreds of risk genes have been assessed for adult phenotypic reversibility. We developed a strategy to rapidly assess the temporal requirements and phenotypic reversibility of NDD risk gene orthologs using a conditional protein degradation system and machine-vision phenotypic profiling in Caenorhabditis elegans. We measured how degrading and re-expressing orthologs of EBF3, BRN3A and DYNC1H1 at multiple periods throughout development affect 30 morphological, locomotor, sensory and learning phenotypes. We found that phenotypic reversibility was possible for each gene studied. However, the temporal requirements of gene function and degree of rescue varied by gene and phenotype. This work highlights the critical need to assess multiple windows of degradation and re-expression and a large number of phenotypes to understand the many roles a gene can have across the lifespan. This work also demonstrates the benefits of using a high-throughput model system to prioritize NDD risk genes for re-expression studies in other organisms.

Keywords: Caenorhabditis elegans; Auxin-inducible degradation; Habituation; Neurodevelopmental disorders; Phenotypic reversibility; Temporal windows of gene function.

Fig. 1.

A pipeline to assess the temporal requirements and phenotypic reversibility of neurodevelopmental disorder risk gene orthologs using the AID system and machine vision phenotypic profiling in Caenorhabditis elegans. (A) The auxin-inducible degradation (AID) system is a powerful approach that enables temporal and spatial control of protein depletion. CRISPR-Cas9 is used to tag the gene of interest (GOI) with the AID degron along with a fluorescent protein (FP) to visualize protein expression in vivo. In the presence of the small molecule indole-3-acetic acid (IAA), TIR1 (an E3 ubiquitin ligase) associates with the AID degron, recruiting endogenous proteosomes to degrade the ubiquitinated protein of interest. (B) Temporal degradation conditions were created by manually transferring animals on and off Petri plates containing IAA to inactivate or restore gene function at specific time points in development or adulthood. (C,D) The effects of protein degradation and re-expression across 30 morphological, locomotor, and sensory and learning phenotypes were objectively quantified in hundreds of animals simultaneously using a machine vision tracking system (C) throughout a short-term mechanosensory habituation paradigm (D).

Fig. 2.

The transcription factor EBF3•UNC-3 displays a reciprocal pattern of phenotypic induction and reversibility across development. (A) The transcription factor EBF3•UNC-3 acts to specify neuronal identity. (B) Continuous degradation of UNC-3 (blue) impaired the animals' ability to respond to mechanosensory stimuli compared to the no IAA control (black). Starting IAA exposure at L4 partially induced this impairment. (C) Ending IAA exposure after L4 (48 h post-hatch) partially rescued the impairment in response probability. (B,C) Data shown as mean±s.e.m. using plates as n (n=4 plates each with 40-100 worms per condition). Significant differences between individual components of the response curves across conditions are indicated in D. (D) Full phenotypic profile of UNC-3, indicating all phenotypes induced by continuous degradation (row 1), induced by adult-specific degradation (row 2), or observed following developmental degradation and adult re-expression (row 3). Cells represent directional t-statistics from comparisons to wild-type controls. Directional t-statistics are shown such that phenotypes higher than the no IAA control are progressively more yellow, and phenotypes lower than the no IAA control are progressively more blue. Only significant differences [false discovery rate (FDR)<0.01] are shown and t-statistics are capped at ±10 to aid visualization. The fourth row indicates the degree of phenotypic rescue (difference from continuous degradation vs adult re-expression), and the fifth row indicates whether that difference was significant (FDR<0.01). For the fourth row, progressively darker green indicates a higher amount of rescue and white indicates no rescue occurred or there was no effect observed with continuous degradation to rescue. See Materials and Methods and Table S1 for a complete description of all 30 quantitative phenotypic features. (E-G) Ending IAA exposure at L2 (24 h post-hatch, gray) almost completely rescued the impairment (E), with the level of phenotypic rescue decreasing with later onset of UNC-3 re-expression (F,G). Starting IAA exposure at L2 (yellow) induced an impairment level similar to the continuous degradation group (E), with the degree of phenotypic impairment decreasing with later onset of UNC-3 degradation (F,G). (E-G) Data shown as mean±s.e.m using plates as n (n=4 plates per condition each with 40-100 worms).

Fig. 3.

EBF3•unc-3 shows diverse temporal patterns of phenotypic reversibility across morphological, locomotor and mechanosensory response phenotypes. The no IAA control group is depicted in black and continuous degradation group is depicted in blue for all panels. (A) Altered animal length could be partially rescued with early post-embryonic re-expression (starting at L2/24 h post-hatch) or fully induced with early post-embryonic degradation. (B,C) The degree of rescue and impairment of animal length increasingly diminished if UNC-3 was re-expression or degraded at L4 (48 h post-hatch) (B) or in adulthood (72 h post-hatch) (C). (D-F) Similarly, impairments in kink could be fully rescued with early post embryonic re-expression (D), but degree of rescue diminished with later re-expression (E,F). Degrading UNC-3 starting at L2 resulted in altered kink to a level similar to the continuous degradation control group (D); the degree of impairment lessened with later onset of degradation (E,F). (A-F) Data shown as mean±s.e.m. using plates as n (n=4 plates per condition each with 40-100 worms). Small points represent individual plate replicates and large points represent the mean±s.e.m. of plate replicates. n.s., not significant. (G-I) Impairments in response duration could not be rescued with UNC-3 re-expression across any of the tested temporal conditions. Degrading UNC-3 in early post embryonic development or development (L4) induced impairments similar to the continuous degradation condition (G,H), yet duration impairments were not strongly induced with UNC-3 degradation starting at 72 h post-hatch (I). (G-I) Data shown as mean±s.e.m. using plates as n (n=4 plates per condition each with 40-100 worms).

Fig. 4.

Degrading BRN3A•UNC-86 specifically impairs mechanosensory response probability and displays a reversibility window restricted to early post-embryonic development. (A) The transcription factor BRNA3•UNC-86 acts to maintain the expression of terminal identity genes in multiple neuron types. (B) Continuous degradation of UNC-86 (blue) specifically impaired response probability to mechanosensory stimuli compared to animals that were not exposed to IAA (black). Staring IAA exposure at L4 (48 h post-hatch, yellow) did not significantly induce phenotypic impairments. (C) Ending IAA exposure at L4 (gray) did not rescue impairments in response probability. (B,C) Data shown as mean±s.e.m. using plates as n (n=4 plates each with 40-100 worms per condition). Significant differences between individual components of the response curves across conditions are indicated in D. (D) Full phenotypic profile of UNC-86, indicating all phenotypes induced by continuous degradation (row 1), induced by adult-specific degradation (row 2), or observed following developmental degradation and adult re-expression (row 3). Cells represent directional t-statistics from comparisons to wild-type controls. Directional t-statistics are shown such that phenotypes higher than the no auxin control are progressively more yellow, and phenotypes lower than the no IAA control are progressively more blue. Only significant differences (FDR<0.01) are shown and t-statistics are capped at ±10 to aid visualization. The fourth row indicates the degree of phenotypic rescue (difference from continuous degradation vs adult re-expression), and the fifth row indicates whether that difference was significant (FDR<0.01). For the fourth row, progressively darker green indicates a higher amount of rescue and white indicates no rescue occurred or there was no effect observed with continuous degradation to rescue. See Materials and Methods and Table S1 for a complete description of all 30 quantitative phenotypic features. (E) Exposing animals to IAA starting at L2 (yellow) induces impairments in response probability. (F) Ending IAA exposure at L2 (24 h post-hatch, gray) enabled phenotypic rescue. (E,F) Data shown as mean±s.e.m. using plates as n (n=4-6 plates per condition each with 40-100 worms).

Fig. 5.

Ubiquitous degradation of the essential protein DYNC1H1•DHC-1 in adult animals reveals specific roles in mechanosensory responding and habituation. (A) The essential gene DYNC1H1•dhc-1 acts in cargo transport and stabilization of microtubule dynamics. (B-D) Starting IAA exposure in early adulthood (72 h post-hatch) did not impair response probability (B) but did deepen habituation of response duration (C) and decreased response speed (D) compared to the no IAA control animals (black). (E) Degrading dynein in adult animals (yellow) decreased average speed during the acclimation period and caused deeper habituation of response speed across the mechanosensory stimuli resulting in a lower average speed during the rest period post mechanosensory stimulation. (B-E) Data shown as mean±s.e.m. using plates as n (n=4 plates per condition each with 40-100 worms). (F) Full phenotypic profiles of DHC-1, indicating all phenotypes induced by adult-specific degradation starting at 3 days post-synchronization. Directional t-statistics are shown such that phenotypes higher than the no IAA control are progressively more yellow, and phenotypes lower than the no IAA control are progressively more blue. Only significant differences (FDR<0.01) are shown and t-statistics are capped at ±10 to aid visualization.

Fig. 6.

Pan-neuronal degradation of DYNC1H1•DHC-1 is not lethal and causes multiple habituation impairments with distinct reversibility profiles. (A) Pan-neuronal degradation of dynein was achieved by crossing the dhc-1(ie28[dhc-1::degron::GFP]) strain with a strain where TIR1 expression is driven by the rab-3 promoter. (B) Continuous degradation of neuronal dynein (blue) impaired response probability compared to the no IAA control group (black). Re-expressing (gray) or degrading (yellow) neuronal dynein starting at L4 (48 h post-hatch) caused animals to be hyperresponsive to mechanosensory stimuli. (C,D) Impairments in response duration and response speed could be rescued with re-expression of dynein starting at L4 (gray). Starting degradation of neuronal dynein at L4 (yellow) did not induce significant impairments in response duration but did induce impairments in response speed. (B-D) Data shown as mean±s.e.m. using plates as n (n=4-6 plates each with 40-100 worms per condition). (E) Full phenotypic profile of DHC-1, indicating all phenotypes induced by continuous degradation (row 1), induced by adult-specific degradation (row 2), or observed following developmental degradation and adult re-expression (row 3). Cells represent directional t-statistics from comparisons to wild-type controls. Directional t-statistics are shown such that phenotypes higher than the no IAA control are progressively more yellow, and phenotypes lower than the no IAA control are progressively more blue. Only significant differences (FDR<0.01) are shown and t-statistics are capped at ±10 to aid visualization. The fourth row indicates the degree of phenotypic rescue (difference from continuous degradation vs adult re-expression), and the fifth row indicates whether that difference was significant (FDR<0.01). For the fourth row, progressively darker green indicates a higher amount of rescue and white indicates no rescue occurred or there was no effect observed with continuous degradation to rescue. (F) Exposing animals to IAA at L2 (24 h post-hatch) impaired response probability (yellow) and ending IAA exposure at L2 rescued impairments in response probability (gray). (G,H) Impairments in response duration and speed were rescued with re-expression of dynein starting at L2 (gray). Starting IAA exposure at L2 (yellow) did not affect response duration but did impair response speed. (F-H) Data shown as mean±s.e.m. using plates as n (n=6 plates per condition each with 40-100 worms).

Fig. 7.

Degrading and re-expressing dynein in neurons reveals distinct temporal functional windows for morphological and baseline locomotion features. The no IAA control group is depicted in black and continuous degradation group is depicted in blue for all panels. (A-D) Dynein is continuously required in neurons for normal aspect ratio (A), kinked body posture (B) animal speed (C) and body length (D). Degrading DHC-1 in neurons beginning at L4 (48 h post-hatch) induced impairment in animal aspect, kink and baseline speed that were similar to the continuous degradation control. Pan-neuronal re-expression of DHC-1 at L4 (gray) rescued impairments in all three phenotypes. (D) Dynein is continuously required in neurons for animal length, but re-expression at L4 can partially rescue impairment. (E) Continuously degrading DHC-1 did not affect animal curvature; however, a novel impairment in animal curvature occurred when dynein was re-expressed at L4. Beginning protein degradation at L4 caused animals to have a higher body curvature than the no IAA control and animals continuously exposed to IAA, whereas re-expressing pan-neuronal DHC-1 at L4 caused animals to exhibit a lower body curvature than controls. (A-E) Data shown as mean±s.e.m. using plates as n (n=4-6 plates per condition each with 40-100 worms). Small points represent individual plate replicates and large points represent the mean±s.e.m. of plate replicates. n.s., not significant.

Fig. 8.

Comparison of temporal profiles reveals shared phenotypic disruptions and prioritizing principles of phenotypic reversibility. (A) Number and kind of phenotypes that could be induced by continuously degrading each gene or reversed by re-expressing the gene 24 h (L2 stage) or 48 h (L4 stage) after synchronization (i.e. early or late in post-embryonic development). (B) Heatmap showing the phenotypes affected by continuous degradation of each gene. Altered phenotypes observed across all three genes are highlighted in pink. (C) Heatmap showing the phenotypes that could be rescued with protein re-expression starting at L2 (24 h post-synchronization). Phenotypes that were reversible across all three genes are highlighted in green. (B,C) Only significant differences (FDR<0.01) are shown.