Double-sieving-defective aminoacyl-tRNA synthetase causes protein mistranslation and affects cellular physiology and development

Abstract

Aminoacyl-tRNA synthetases (aaRSs) constitute a family of ubiquitously expressed essential enzymes that ligate amino acids to their cognate tRNAs for protein synthesis. Recently, aaRS mutations have been linked to various human diseases; however, how these mutations lead to diseases has remained unclear. In order to address the importance of aminoacylation fidelity in multicellular organisms, we generated an amino-acid double-sieving model in Drosophila melanogaster using phenylalanyl-tRNA synthetase (PheRS). Double-sieving-defective mutations dramatically misacylate non-cognate Tyr, induce protein mistranslation and cause endoplasmic reticulum stress in flies. Mutant adults exhibit many defects, including loss of neuronal cells, impaired locomotive performance, shortened lifespan and smaller organ size. At the cellular level, the mutations reduce cell proliferation and promote cell death. Our results also reveal the particular importance of the first amino-acid recognition sieve. Overall, these findings provide new mechanistic insights into how malfunctioning of aaRSs can cause diseases.

Figure 1. Drosophila PheRS loci and mutual stabilization of the two subunits.

(a) The P{EP}-element insertion allele G2060 is an α-PheRS mutant. (b) β-PheRS¹ deletes the first exon of β-PheRS and the neighbouring jar gene. The scheme is on the basis of the information available on FlyBase and is not to scale. (c) Expression of β-PheRS variants in β-PheRS¹ and wild-type fly heads. The lower band is wild-type β-PheRS, and the upper band is the myc-tagged β-PheRS. (d) β-PheRS protein levels in Kc cells upon knockdown of the subunits. The control (ctl) is Amp^R RNAi. α-PheRS¹, α-PheRS² and α-PheRS³ stand for three different dsRNAs against α-PheRS. Knockdown of either subunit downregulates β-PheRS levels. (e) Myc::α-PheRS protein levels in fly fat bodies upon knockdown of the subunits. ppl-Gal4 drove myc::α-PheRS expression and RNAi against either subunit. The control (ctl) is Cdk7 RNAi. Knockdown of either subunit downregulates Myc::α-PheRS levels. (f) Protein levels upon PheRS overexpression in the fat body dissected from wandering third instar larvae. ppl-Gal4 was used as driver and the control (ctl) is GFP overexpression. Higher levels of protein accumulated only when both subunits (PheRS) were co-overexpressed.

Figure 2. Sieving-defective mutations in Drosophila PheRS.

(a,c) Structure-based alignment of PheRS sequences using the PROMALS3D method. The sequences of prokaryotes and archaea/eukaryotes are shown in grey and blue, respectively. The C-terminal region of α-PheRS is listed in a, and the red Ala is the amino-acid residue substituted by Gly. Domain B3-4 of β-PheRS is shown in c. Residues crucial for editing, the ones at the entrance of the editing pocket (red) and the ones at the activation centre (black), are indicated. DROME, Drosophila melanogaster; HUMAN, Homo sapiens; PYRHO, P. horikoshii; THET8, T. thermophilus; ECOLI, E. coli. (b,d) Structure of the Drosophila PheRS catalytic module (first sieve) and domain B3-4 (second sieve) modelled by UCSF Chimera. The residues highlighted in red in b,d correspond to the red Ala in a,c, respectively. The catalytic module (b; domain B6-7 in light blue and domain A1-2 in grey) was modelled on the human PheRS (PDB: 3L4G). Domain B3-4 was modelled on P. horikoshii PheRS (d; PDB: 2CX1). (e,f) The Phe and Tyr aminoacylation activities of PheRS variants were determined in vitro. (e) PheRS variants are still active in Phe aminoacylation. However, while the enzymatic activity of the βA158W mutant is normal, the activity of the αA456G mutant is somewhat reduced. (f) Tyr mis-aminoacylation activity of PheRS variants. Wild-type PheRS produced very low levels of TCA-insoluble Tyr and the single mutant αA456G and βA158W produced only slightly elevated levels. In contrast, the αA456G, βA158W ‘double-mutant’ protein dramatically misacylated Tyr. Values are the means of three independent experiments. Error bars show s.d.

Figure 3. Expressed in the eye PheRS-sd mutations cause retinal defects.

(a) Eye surface phenotypes induced by ey-Gal4 (ey)-driven expression of PheRS-sd mutations. In the PheRS wild-type background, GFP expression served as control. Some ommatidia in αA456G-expressing eyes (arrowhead) are not properly organized, while βA158W-expressing ones are normal. Flies expressing GFP or αA456G in the gβPheRS-rescued β-PheRS^null background were similar to the ones in the PheRS⁺ wild-type background. Flies expressing αA456G in the gβA158W background gave rise to severe rough eye phenotypes. Flies were ~1-week old and the scale bar represents 200 μm. (b) Quantification of eye defects from the experiment shown in a. If ommatidia showed irregularities in any region of either eye, the animal was recorded as defective. Flies were ~1-week old. (c) Light microscopic images of semithin retina sections. Hexagonal units are individual ommatidia in which seven photoreceptor cells are visible in a given plane. Well-oriented ommatidia with normal complements of photoreceptors are seen in controls and in βA158W-expressing eyes. The expression of αA456G alone or αA456G together with βA158W resulted in reduced numbers of photoreceptor cells in individual ommatidia (indicated by arrowheads) and a misaligned pattern. Scale bar represents 10 μm.

Figure 4. Sieving-defective mutations induce impaired locomotive performance and advanced ageing.

(a,b) Negative geotaxis climbing assays were performed to study locomotive behaviour. For each genotype, three groups of male flies were tested at the indicated age. The assay was recorded as video, and the movies were analysed to determine the climbing index. (a) αA456G mutant flies displayed progressively decreasing climbing ability. Expression of the α-subunit (wild-type and αA456G mutant) was driven by Act-Gal4. (b) βA158W did not differ from the wild type during the first 5 weeks. Transgenes were in the β-PheRS^null background, rescued by gβPheRS or gβA158W under their endogenous promoter. Error bars show s.e.m. n=3, *P<0.05, **P<0.01, t-test. (c,d) Lifespan analysis of the male flies with the same genotype as in a,b. (c) αA456G mutant flies aged faster than wild type, and the differences became apparent already at early stages. P<0.001, log-rank test. (d) At younger age, gβA158W mutant flies aged similarly as the wild type, but at older age (more than 50 days old) the mutants aged faster. P<0.001, log-rank test. (e) Confocal microscopy images of the ventrolateral protocerebrum region of the adult fly brain. Embryos and larvae were kept at 18 °C to keep Gal80^ts active to repress Gal4. With this protocol, normal adults eclosed (while the ones expressing αA456G and βA158W were lethal without Gal80^ts) and these flies were then incubated at 29 °C for ~7 days, inactivating Gla80^ts and activating Gal4. Elav (green) marks neuronal cells, and anti-CP3 stains apoptotic cells (red). A few CP3-positive cells were detected in neurons expressing the αA456G mutant, and many more apoptotic neurons (arrowheads) were observed when both αA456G and βA158W variants were expressed. DNA is in blue. Scale bars represent 5 μm.

Figure 5. PheRS-sd mutants display reduced wing size.

(a) Adult wing phenotypes induced by PheRS-sd mutations. en-Gal4 (en) was used to drive transgene expression in the posterior compartment of developing wing discs. The dark dashed line indicates the border of the anterior/posterior compartment. In the PheRS wild-type background, wings expressing GFP and βA158W were normal. The ones expressing αA456G showed vein defects (arrowhead) and also reduced wing size in the posterior compartment. In the gβPheRS-rescued β-PheRS^null background, wings expressing GFP and αA456G, respectively, showed similar phenotypes as in the PheRS⁺ background. In contrast, wings expressing αA456G in a gβA158W background displayed severe size reduction with significant parts of the organ missing altogether. Scale bars represent 500 μm. (b,c) Quantification of wing sizes from the experiment shown in a. Twelve single female wings from different flies were analysed for each genotype. Wing size was measured by counting the pixels in an anterior (A; outlined in blue) and a posterior area (P; outlined in yellow) of each wing, and the ratio (P/A) was calculated. (b) In a PheRS⁺ wild-type background, αA456G mutants showed a size decrease, but βA158W did not. (c) In the rescued β-PheRS background, the αA456G mutant showed a size decrease and, when αA456G and βA158W are expressed, wing size was further decreased. Error bars represent s.e.m. n=12, n.s., not significant; ***P<0.001, t-test.

Figure 6. PheRS-sd mutations induce apoptosis and reduce cell proliferation.

(a) Confocal microscopy images of wing discs from third instar wandering larvae. The posterior compartment is marked with GFP in green. CP3 is visualized in red with anti-CP3 antibodies. The anterior wing disc compartment (GFP-negative) serves as internal control. All flies shown are in the β-PheRS^null background, rescued by gβPheRS or gβA158W. A few CP3-positive cells were observed in αA456G mutant wing discs; however, many more cells underwent apoptosis when αA456G and βA158W mutants were expressed. Some CP3-positive cells are pointed out by arrowheads. Note that CP3-positive cells (red) always overlap with GFP (green). DNA is shown in blue. Scale bars represent 50 μm. (b) Quantification of apoptotic wing discs. No apoptosis was detectable in wing discs from controls and wing discs expressing only βA158W. A small number of wing discs expressing αA456G showed apoptotic cells. Apoptosis was most prominent in wing discs that expressed αA456G and βA158W mutant protein. (c,d) Wing discs were subjected to FACS analysis. The GFP signal was used to distinguish cell populations that express the transgenes. (c) The αA456G mutation reduces cell numbers. Cell numbers were determined relative to the cell numbers in the anterior (GFP-negative) area. Error bars indicate s.e.m. n=3, **P<0.01, t-test. (d) The αA456G mutation does not reduce cell size. Cell size was determined by Forward Scatter. Cells from the GFP-positive compartment are smaller than the ones from the GFP-negative compartment in the control. In αA456G mutants, their sizes are similar. Figure shows representatives of three independent experiments.

Figure 7. Sieving-defective mutations induce ER stress.

An Xbp1-eGFP reporter was used to monitor the expression of the spliced Xbp1 isoform in PheRS-sd mutant wing discs using immunostaining. While no signal was observed in controls, GFP expression was detected both in αA456G mutants and in cells expressing the mutant forms αA456G and βA158W. Note that the GFP signal is always restricted to the posterior compartment. DNA is in blue and GFP is in green. Scale bars represent 25 μm.

Figure 8. PheRS-sd mutant flies are sensitive to non-cognate amino acids.

(a) Scheme illustrating the experimental procedure. The ppl-Gal4 driver was used to drive PheRS-sd mutations in the larval fat body. Embryos were aged to 24 h after egg laying. First instar larvae with the correct genotype were selected and transferred to different amino-acid-rich food. When the larvae developed into pupae, the number of pupating individuals was recorded to calculate the pupariation rate. (b–e) PheRS-sd mutant flies are sensitive to the non-cognate amino acid Tyr. Larvae were raised on standard food (b), Ala-rich food (100 mM, c), Phe-rich food (100 mM, d) and Tyr-rich food (100 mM, e). No significant differences in pupariation rate were observed between control and PheRS-sd mutants on standard food and Ala- or Phe-rich food. On Tyr-rich food, the pupariation rate of the αA456G mutant was reduced to ~60%, and the one of larvae expressing αA456G and βA158W PheRS to ~30%. Error bars represent s.e.m. n=6. **P<0.01, *P<0.05, t-test.

Figure 9. PheRS-sd mutations induce Phe-to-Tyr substitutions.

(a) The FtoY reporter was generated to analyse the Phe-to-Tyr substitution in vivo. Third instar larvae were dissected and stained with X-gal. In driver-only larvae and in LacZ^Y503F mutants, no clear staining was detected, while a strong signal was observed in LacZ wild-type larval guts. PheRS-sd mutations restore β-galactosidase activity of LacZ^Y503F mutants. In the β-PheRS-rescued β-PheRS^null background, the αA456G mutant produced low levels of blue signal in the gut, and when αA456G and βA158W PheRS were expressed, stronger staining of additional cells become apparent (arrowheads). Scale bar represents 100 μm. (b) Quantification of number of guts that stained positive for X-gal. (c) Quantification of the blue staining region of the gut. The same size area of the gut was chosen for each sample and the blue pixels within this area were determined to calculate the ratio. Error bars indicate s.e.m. n=4, *P<0.05, t-test.