A translation-independent function of PheRS activates growth and proliferation in Drosophila

ABSTRACT Aminoacyl transfer RNA (tRNA) synthetases (aaRSs) not only load the appropriate amino acid onto their cognate tRNAs, but many of them also perform additional functions that are not necessarily related to their canonical activities. Phenylalanyl tRNA synthetase (PheRS/FARS) levels are elevated in multiple cancers compared to their normal cell counterparts. Our results show that downregulation of PheRS, or only its α-PheRS subunit, reduces organ size, whereas elevated expression of the α-PheRS subunit stimulates cell growth and proliferation. In the wing disc system, this can lead to a 67% increase in cells that stain for a mitotic marker. Clonal analysis of twin spots in the follicle cells of the ovary revealed that elevated expression of the α-PheRS subunit causes cells to grow and proliferate ∼25% faster than their normal twin cells. This faster growth and proliferation did not affect the size distribution of the proliferating cells. Importantly, this stimulation proliferation turned out to be independent of the β-PheRS subunit and the aminoacylation activity, and it did not visibly stimulate translation. This article has an associated First Person interview with the joint first authors of the paper.

The mean value of intensities of the fluorescent signals in the anterior and the posterior compartment were compared after applying the same settings (that did not produce a signal for the secondary antibody) for all image acquisitions.Pixel intensities in the two compartments of the wing discs were then compared.Driving the expression of only a wild-type copy of α-PheRS, the anti α-PheRS antibody produced a mean pixel intensity of 68 ± 9 in the posterior compartment and a mean pixel intensity of 37 ± 7 in the anterior compartment (n=4).The intensity was therefore 1,8 times as high in the posterior.The expression of a missense mutant of α-PheRS alone gave similar results.Posterior mean pixel intensity was 60 ± 13, the anterior one 34 ± 6; posterior levels were thus 1,7 times as high (n=8).Under the assumption that our primary antibody has 0%, 25% and 33%, respectively, cross reactivity, we measured and calculated an increase of the wild-type α-PheRS levels of 80%, 107%, and 120%, respectively.We have no evidence for high cross-reactivity of this antibody but cannot measure it because cells lacking α-PheRS are dead.For the mutant α-PheRS levels, the corresponding increase is 70%, 93%, and 105%, respectively.

Fold protein expression difference upon o/e of !-PheRS
-1 -2 -3 -4 -5 Phe frequency of occurrence Over-expression of one subunit of PheRS might affect the expression of the (ab) 2 PheRS complex that is required to load phenylalanine (Phe) onto the tRNA Phe producing the Phe-tRNA Phe needed for translation.Even though a general change in translation was not observed when translation was assessed in situ (Fig 3E), it is still conceivable that this might lead to a change in levels of Phe-tRNA Phe and uncharged tRNA Phe in the cell and that this might specifically affect translation of one or a few mRNAs with an extreme (probably more likely a high) frequency of Phe codons.To seek such hypothetical events that might be meaningful for growth and proliferation control, we systematically assessed the effect of overexpression of the a-PheRS subunit on the abundance of polypeptides with different and extreme frequencies of Phe codons using a proteomics approach.Towards this goal, the wild-type a-PheRS open reading frame was over-expressed under the strong a-tubulin84B-Gal4 >UAS control.First instar larvae were harvested, proteins were extracted and quantified using mass spectrometry.Experiments were carried out in biological triplicates and the identification of peptides and proteins was by Maxquant.The Maxquant Label-Free Quantification (LFQ) values were used to calculate the expression levels.
The expression levels of these proteins were then compared to control larvae that expressed only the a-tubulin84B-Gal4 driver.Neither size nor p-value filters were applied to correlate this data with the Phe frequency of these proteins.To determine the Phe codon frequency, the UniProtKB 2020_05 database was used to determine the length and the number of Phe for each of the 22,045 Drosophila melanogaster proteins in the database.From this, the Phe frequency was established for each protein.
A) The 2,986 proteins, for which LFQ enrichment could be determined, were used to analyze a possible correlation of enhancement of expression upon o/e of a-PheRS with relative abundance of Phe in the protein.This correlation analysis was performed in Excel.The result shows no or only a very weak negative correlation for this data set (-0.0029 for the linear expression difference and -0.0304 for the log2 fold expression difference).B) PheRS codon frequency does not change highly between the significantly under-or overexpressed proteins upon α-PheRS o/e.C) α-PheRS, with a log2 expression change of 0.652, does not show up in the significantly over-expressed proteins.

Figure S1 :
Figure S1: Measurement of additional expression of α-PheRS in the posterior wing disc compartment.The mean value of intensities of the fluorescent signals in the anterior and the posterior compartment were compared after applying the same settings (that did not produce a signal for the secondary antibody) for all image acquisitions.Pixel intensities in the two compartments of the wing discs were then compared.Driving the expression of only a wild-type copy of α-PheRS, the anti α-PheRS antibody produced a mean pixel intensity of 68 ± 9 in the posterior compartment and a mean pixel intensity of 37 ± 7 in the anterior compartment (n=4).The intensity was therefore 1,8 times as high in the posterior.The expression of a missense mutant of α-PheRS alone gave similar results.Posterior mean pixel intensity was 60 ± 13, the anterior one 34 ± 6; posterior levels were thus 1,7 times as high (n=8).Under the assumption that our primary antibody has 0%, 25% and 33%, respectively, cross reactivity, we measured and calculated an increase of the wild-type α-PheRS levels of 80%, 107%, and 120%, respectively.We have no evidence for high cross-reactivity of this antibody but cannot measure it because cells lacking α-PheRS are dead.For the mutant α-PheRS levels, the corresponding increase is 70%, 93%, and 105%, respectively.

Figure S2 :
Figure S2: No or only minor negative correlation between protein expression levels andPhe frequency under high a-PheRS over-expression conditions.Over-expression of one subunit of PheRS might affect the expression of the (ab) 2 PheRS complex that is required to load phenylalanine (Phe) onto the tRNA Phe producing the Phe-tRNA Phe needed for translation.Even though a general change in translation was not observed when translation was assessed in situ(Fig 3E), it is still conceivable that this might lead to a change in levels of Phe-tRNA Phe and uncharged tRNA Phe in the cell and that this might specifically affect translation of one or a few mRNAs with an extreme (probably more likely a high) frequency of Phe codons.To seek such hypothetical events that might be meaningful for growth and proliferation control, we systematically assessed the effect of overexpression of the a-PheRS subunit on the abundance of polypeptides with different and extreme frequencies of Phe codons using a proteomics approach.Towards this goal, the wild-type a-PheRS open reading frame was over-expressed under the strong a-tubulin84B-Gal4 >UAS control.First instar larvae were harvested, proteins were extracted and quantified using mass spectrometry.Experiments were carried out in biological triplicates and the identification of peptides and proteins was by Maxquant.The Maxquant Label-Free Quantification (LFQ) values were used to calculate the expression levels.The expression levels of these proteins were then compared to control larvae that expressed only the a-tubulin84B-Gal4 driver.Neither size nor p-value filters were applied to correlate this data with the Phe frequency of these proteins.To determine the Phe codon frequency, the UniProtKB 2020_05 database was used to determine the length and the number of Phe for each of the 22,045 Drosophila melanogaster proteins in the database.From this, the Phe frequency was established for each protein.A) The 2,986 proteins, for which LFQ enrichment could be determined, were used to analyze a possible correlation of enhancement of expression upon o/e of a-PheRS with relative abundance of Phe in the protein.This correlation analysis was performed in Excel.The result shows no or only a very weak negative correlation for this data set (-0.0029 for the linear expression difference and -0.0304 for the log2 fold expression difference).B) PheRS codon frequency does not change highly between the significantly under-or overexpressed proteins upon α-PheRS o/e.C) α-PheRS, with a log2 expression change of 0.652, does not show up in the significantly over-expressed proteins.
Marked in red are proteins with a Phe frequency over the 75% quantile of all analyzed proteins and in blue the proteins with a Phe frequency under the 25% quantile of all analyzed proteins.Psi has a role in tissue growth.It is also underexpressed in β-PheRS o/e.The effect is not specific for α-PheRS o/e Lipid storage droplet-2 (Lsd-2) encodes a protein associated with lipid droplets.It acts as a barrier for lipases (such as the product of bmm) and thus prevents the mobilization of lipid stores.It is involved in regulation of lipid storage amount and energy homeostasis and acts in concert with the product of Lsd-1.