Identifying regulators of associative learning using a protein-labelling approach in Caenorhabditis elegans

(A) List of C. elegans lines used in this study. Lines are marked as ‘not applicable’ for ‘number of times backcrossed with N2’ when (A) the line itself is N2 (i.e., wild-type) or (B) the strains correspond to transgenic animals with a wild-type background. (B) Quantifying biotinylated protein levels between non-transgenic animals and TurboID C. elegans. This excel sheet contains six tables, each corresponding to an individual western blot. These blots are Figure 1C (red, rows 1–8), Figure 1—figure supplement 1 (yellow, rows 9–16), & Figure 1—figure supplement 3 (green, rows 17–47). Each table compares biotinylated protein signal between experimental groups, which are non-transgenic/Non-Tg (from N2) or TurboID/TbID-expressors (from YLC207). Rows corresponding to Figure 1C & Figure 1—figure supplement 3 involve high-salt control and trained groups prepared for mass spectrometry experiments (i.e., with biotin exposure limited to memory encoding), whereas Figure 1—figure supplement 1 compares untreated versus biotin-treated animals, Each row corresponds to a unique biological replicate of western blot and mass spectrometry data as annotated under the ‘Replicate’ column. For each sample, the sum of areas containing bands within an entire lane is used as a readout for biotin-tagged protein signal intensity from whole worm lysates. Non-transgenic/Non-Tg C. elegans and TurboID/TbID were treated with biotin during salt associative learning, to generate high-salt control and trained groups for each line. (C) Protein lists for all assigned hits detected in TurboID, trained worms. C. elegans peptides were assigned corresponding protein identities by MASCOT and then assigned protein lists were generated for each biological replicate by subtracting proteins also present in ‘non-transgenic, trained’ worms and/or in ‘TurboID, control’ worms within the same replicate. Each protein list in rows 9–378 are defined in this table by its corresponding biological replicate, the mass spectrometer used, and the inclusion of data from unbound and bound peptide samples. Peptide samples generated for biological replicate 3 were run in both mass spectrometers used in this study, the Q Exactive or the Exploris, and the corresponding protein list is named ‘3 a’ and ‘3b’ respectively. (D) Protein lists for all unassigned hits detected in TurboID, trained worms. Protein identities were determined through bulk BLAST searching C. elegans peptides detected by mass spectrometry, rather than MASCOT, and then unassigned protein lists were generated for each experimental group in each biological replicate. Proteins also in ‘non-transgenic, trained’ and/or ‘TurboID, control’ lists for the same biological replicate were subtracted from the corresponding ‘TurboID, trained’ list. Each protein list in rows 9–587 are defined in this table by its corresponding biological replicate, the mass spectrometer used, and the inclusion of data from unbound and bound peptide samples. Peptide samples generated for biological replicate 3 were run in both mass spectrometers used in this study, the Q Exactive or the Exploris, and the corresponding protein list is named ‘3 a’ and ‘3b’ respectively. (E) Cellular component GO terms associated with assigned hits in the learning proteome. All (1010) assigned protein hits detected in TurboID, trained worms were separated into 10 clusters by k-means clustering on STRING (version 12.0) and GO term analysis was conducted for each cluster using ClueGO (version 2.5.10) on Cytoscape (version 3.10.0). GO terms were then sorted into ‘plasma membrane & extracellular matrix (ECM)’, ‘cilia & dendrites’, ‘cell body & cytoplasm’, ‘nucleus’, ‘Golgi apparatus’, ‘cytoskeleton’, ‘mitochondria’, ‘axon & vesicles’, ‘pre-synapse’, and ‘other’ (i.e., non-informative terms) categories due to recurring and/or similar terms. Colour-coding is based on the GO Group. Each category is separated by underlined and bolded titles corresponding to the cellular component in rows 1–285. Consolidated lists of genes/proteins are detailed from rows 287–465. (F) Biological process and molecular function GO terms associated with assigned hits in the learning proteome. All (1010) assigned protein hits detected in TurboID, trained worms were separated into 10 clusters by k-means clustering on STRING (version 12.0) and GO term analysis was conducted for each cluster using ClueGO (version 2.5.10) on Cytoscape (version 3.10.0). GO terms were then sorted into ‘protein synthesis’, ‘protein degradation’, ‘insulin & protein kinases’, ‘G protein signalling’, and ‘neurotransmission’ categories due to recurring and/or similar terms. Colour-coding is based on the GO Group. Each category is separated by underlined and bolded titles corresponding to the biological processes listed above in rows 1–109. Consolidated lists of genes/proteins are detailed from rows 111–162. (G) Identifying neuron classes represented within the learning proteome in this study. This table presents neurons that express ≥1 gene encoding proteins found in either the trained or control lists. For neurons present in both datasets, fold-change values were calculated as the ratio of the number of genes expressed in the trained condition to those in the control. Neurons are ranked in descending order of fold-change. Values under the '#Proteins from trained (normalised)' column are normalised based on the fold-difference between the number of proteins in the trained vs control protein lists (normalisation factor = ~1.8). Classifications for neuron type are assigned to each neuron class based on Pereira et al., 2015. The CeNGEN database was used to identify gene expression patterns Taylor et al., 2021 and includes all 302 neurons in the C. elegans hermaphrodite nervous system. Only 128 neuron classes are listed here since neurons that do not express any proteins from trained or control groups are not shown. (H) Molecular pathways for proteins detected in TurboID, trained C. elegans only. Proteins unique to trained animals that express TurboID were identified by subtracting those also in ‘TurboID, control’ and/or ‘non-transgenic, trained’ protein lists. Gene names are listed with assigned hits in bold. Each gene/protein has been grouped based on ‘Biological Process’ GO term annotations provided by STRING (version 12.0) and information available on WormBase (version WS290) based on data accessed during September-November 2023. (I) Statistical analyses for learning assays performed in this study. Data from statistical analyses are separated by figure with appropriate underlined and bolded titles, including Figure 1B (red, rows 1–7), Figure 4 (yellow, rows 9–31), Figure 5 (green, rows 33–55), Figure 6 (blue, rows 57–87), Figure 6—figure supplement 1 (purple, rows 89–191), Figure 6—figure supplement 2 (pink, rows 193–247), and Figure 6—figure supplement 2 (red, rows 249–287). Statistical analyses were performed on GraphPad PRISM 8 using two-way ANOVA and Tukey’s multiple comparisons tests (alpha value = p < 0.05). Columns left-to-right list the groups being compared, the mean difference (Diff.), the 95% confidence interval of diff., whether the difference is significant, the degree of significance (****≤0.0001; ***≤0.001; **≤0.01; *≤0.05; ns = non-significant), and the adjusted p value.

These plasmids encode Prab-1743 3::V5::TurboID::gpd-2 3’ UTR and Prab-3::kin-2(ce179)::SL2::tag-RFP::gpd-2 3’ UTR, 1744 respectively. The folder contains an image of each plasmid map, as well as their DNA sequences 1745 in.fasta and.gb formats.