Fig 1.
Schematic detailing transcriptional and translational profiling of retrograde homeostatic signaling at the Drosophila NMJ.
(A) Schematic illustrating synaptic transmission at the Drosophila NMJ. Representative EPSP and mEPSP electrophysiological traces in wild type (w1118; BG57-Gal4/UAS-RpL3-3xflag, n = 6), GluRIIA mutants (w; GluRIIASP16; BG57-Gal4/UAS-RpL3-3xflag, n = 6), and overexpression of Tor in the postsynaptic muscle (Tor-OE: w;UAS-Tor-myc/+;BG57-Gal4/UAS-RpL3-3xflag; n = 6). Note that while mEPSP amplitudes are reduced in GluRIIA mutants, EPSP amplitudes remain the same as wild type because of a homeostatic increase in presynaptic release (quantal content). Tor-OE does not change mEPSP amplitude, but retrograde homeostatic signaling is induced, leading to increased EPSP amplitude and quantal content. Quantification of mEPSP amplitude (B), EPSP amplitude (C), and quantal content (D) for the indicated genotypes. (E) Schematic illustrating the putative role of protein synthesis in retrograde homeostatic signaling and the design of ribosome tagging to isolate postsynaptic RNA. (F) Schematic representing the workflow for transcriptional profiling, translational profiling using TRAP (translating ribosome affinity purification), and ribosome profiling. Student’s t test was used to compare GluRIIA and Tor-OE to wild type; ** = p<0.01.
Fig 2.
Development and validation of an optimized tissue specific ribosome profiling protocol in Drosophila.
(A) Schematic illustrating the ribosome affinity purification strategy. A tagged ribosome subunit (RpL3-Flag) is expressed and incorporated into ribosomes. Magnetic beads coated with anti-flag antibodies are used to immunoprecipitate ribosomes along with associated mRNAs. (B) Anti-flag immunoprecipitation from wild type (control), postsynaptic expression of RpL3-Flag (w;BG57-Gal4/UAS-RpL3-3xflag), and postsynaptic expression of RpS13-Flag (w;BG57-Gal4/UAS-RpS13-3xflag) in third-instar larval muscle. Samples were run on an SDS-PAGE gel and Commassie stained. The expected distribution of ribosomal proteins are present in RpL3-Flag and RpS13-Flag samples (noted by arrowheads), but not observed in wild-type controls. (C) Total RNA was extracted from anti-flag immunoprecipitations from wild type and RpL3-Flag larval muscle tissue and run on an agarose gel. Ribosomal RNA is present in RpL3-Flag RNA samples but absent in wild type samples. Total RNA extracted from wild type whole larvae was loaded to show the position of ribosomal RNA. (D) Workflow for the ribosome profiling strategy. (E) Representative RNA-seq mapping of the actin57B locus from transcriptional, translational (TRAP), and ribosome profiling. Note that ribosome profiling reads predominantly map to 5’UTR and coding regions, and are absent from the 3’UTR. RPM: reads per million total mapped reads. (F) Replicate ribosome profiling sequencing demonstrates highly reproducible results.
Fig 3.
Comparison of translational and ribosome profiling from Drosophila larval muscle.
(A) Plot of ribosome profiling RPKM as a function of transcriptional profiling RPKM for all muscle genes in wild type. Genes with high translation efficiency (TE; TE>2) or low TE (TE<0.5) are labeled in red and blue respectively. Genes with medium TE (TE between 0.5 and 2, indicated by the two dash lines) are labeled in grey. (B) Plot of translational profiling (TRAP) RPKM as a function of transcriptional profiling RPKM for all muscle genes in wild type. The same color coding scheme is used as in (A). (C) Graph showing percentage of total muscle genes that are in the high TE, medium TE or low TE group based on ribosome profiling or TRAP. Note that a lower percentage of genes are revealed to have high or low TE with TRAP compared to ribosome profiling. (D) Plot of translation efficiency defined by ribosome profiling as a ratio of TRAP of all genes in the three categories: high TE (ribosome profiling and TRAP TE average>2), medium TE (TE average between 0.5 and 2), and low TE (TE average<0.5). Note that ribosome profiling reveals higher TE for high TE genes, and lower TE for low TE genes compared to TRAP. *** = p<0.001; one-way ANOVA with post hoc Bonferroni’s test.
Fig 4.
Analysis of the transcriptome and translatome reveals dynamic translational regulation in Drosophila muscle.
(A) Definition of number of genes encoded in the Drosophila genome and those expressed in the muscle transcriptome and translatome. (B) Heat map showing transcriptional levels of all annotated genes in the Drosophila larval muscle compared to those expressed in the central nervous system (CNS; [38]). These genes are grouped into four sections according to their expression status in muscle and CNS; the percentage of total genes is indicated above each section. (C) Heat map defining translation efficiency (TE) and transcription and translation expression levels (RPKM) of genes expressed in muscle. Genes are ordered according to translation efficiency, with a trend observed for genes with high translation efficiency having low transcriptional expression levels and vice versa. (D) Transcriptional expression levels of genes with low TE (TE<0.5, blue), medium TE (TE between 0.5 and 2, grey), and high TE (TE>2, red). The transcriptional expression levels of genes in the low TE group is significantly higher than that of the medium TE group, while transcriptional expression of the high TE group is significantly lower than that of the medium TE group (*** = p<0.001; one-way ANOVA with post hoc Bonferroni’s test). (E) Histogram of translation efficiency across all genes expressed in the muscle. The 100 genes with the lowest translation efficiency (blue) and highest translation efficiency (red) are indicated. (F) Histogram of translation efficiency for the 100 genes with the lowest translation efficiency. An enrichment in ribosomal proteins, indicated in blue, is observed. (G) Histogram of translation efficiency for the 100 genes with the highest translation efficiency. Genes in the most abundant functional class, encoding proteins involved in cellular structure, are indicated in red. (H) Graph showing the TE of the 100 genes with the highest or lowest TE in Tor-OE as a ratio of wild type. Note that the translation efficiency of ribosomal proteins in Tor-OE are significantly increased compared to wild type. *** = p<0.001; paired Student’s t-test. Additional details can be found in S3 and S4 Tables and S3 Fig.
Fig 5.
Analysis of transcriptional and translational profiling of GluRIIA expression in GluRIIA mutants and Tor overexpression in Tor-OE.
(A) Schematic illustrating the genomic GluRIIA locus in wild type and GluRIIASP16 mutants. Note that the 5’ region of GluRIIA is deleted in the GluRIIASP16 mutant, as well as the adjacent oscillin gene. Below: RNA-seq reads mapping to the GluRIIA locus from transcriptional and ribosome profiling in wild type and GluRIIASP16 mutants. The coverage graphs were divided into four sections corresponding to the regions indicated in the GluRIIA transcript. The numbers in each graph indicates the expression value of that region normalized to wild type transcriptional or ribosome profiling expression value. Note that no expression was detected by transcriptional or ribosome profiling in the deleted region in GluRIIA mutants, as expected. (B) Schematic illustrating the endogenous Tor mRNA transcript and the mRNA transcript transgenically expressed in Tor-OE (UAS-Tor). Both transcripts share the same coding sequence, but differ in their 5’UTR and 3’UTR sequences. Below are reads mapping to the indicated regions, divided into the three indicated sections. Note that both transcriptional and translational expression of UAS-Tor mRNA are significantly increased in Tor-OE, while transcription and translation of endogenous Tor mRNA is largely unchanged in Tor-OE. (C) Histogram of the distribution of gene translation changes in wild type versus wild type (black), which represents intrinsic variability, that of Tor-OE versus wild type (Red), and that of GluRIIA mutants versus wild type (blue). Note the shift in distribution observed in Tor-OE, suggesting a global increase in translation. (D) Cumulative percentage plot of distributions shown in (C), showing significant difference between Tor-OE versus wild type distribution compared to wild type versus wild type distribution. (p<0.001, Kolmogorov–Smirnov test).
Fig 6.
Few changes in postsynaptic transcription or translation are observed in GluRIIA mutants.
(A) Plot of transcriptional and translational expression levels of all genes in GluRIIA mutants (w1118;GluRIIASP16; BG57-Gal4/UAS-RpL3-3xflag) compared to wild type (w1118; BG57-Gal4/UAS-RpL3-3xflag), with near perfect correlations observed (indicated by r2 values). Right: Table showing the number of genes with significantly up-regulated (up arrow) or down-regulated (down arrow) transcription or translation efficiency (TE) in GluRIIA compared to wild type using the indicated cut off values. (B) Plot of transcriptional and translational expression levels of all genes in Tor-OE (w1118;UAS-Tor-myc/+;BG57-Gal4/ UAS-RpL3-3xflag) compared to wild type. Note that while moderate changes in transcription are observed, large differences in translation are found (indicated by r2 values). Right: Table showing the number of genes with significantly up-regulated (up arrow) or down-regulated (down arrow) transcription or TE in Tor-OE compared to wild type using the indicated cut off values. (C) Heat map indicating the 47 genes with significant increase in TE in Tor-OE, with the corresponding genes in GluRIIA mutants shown below. Note that no trend is observed in translational expression of these genes in GluRIIA mutants. TOP mRNAs are highlighted in red. Additional details can be found in S5 and S6 Tables.
Fig 7.
Increased cellular translation triggers adaptive responses in both transcription and translation.
(A) Diagram showing the number of significantly upregulated genes in transcription and translation in Tor-OE compared to wild type (p<0.05, fold change>3). (B) Pie chart illustrating the classes of differentially upregulated genes in Tor-OE compared to wild type. (C) Comparative fold changes for chaperones differentially upregulated in transcription and translation in Tor-OE compared to wild type. Note that all but one exhibit higher translational changes compared to transcriptional changes, implying an additional layer of regulation in translation efficiency in addition to the increased transcriptional expression. (D) Graphs showing RPKM values measured by transcriptional and ribosome profiling in wild type and Tor-OE for the representative heat shock protein Hsp23, the ubiquitin E3 ligase APC4, the RNA polymerase Rpl1, and the transcription factor myc. Read mapping of the indicated genes are illustrated below. ** = p<0.01, *** = p<0.001; Student’s t-test. Additional details can be found in S7 Table.