Single amino-acid mutation in a Drosoph ila melanogaster ribosomal protein: An insight in uL11 transcriptional activity

The ribosomal protein uL11 is located at the basis of the ribosome P-stalk and plays a paramount role in translational efficiency. In addition, no mutant for uL11 is available suggesting that this gene is haplo-insufficient as many other Ribosomal Protein Genes (RPGs). We have previously shown that overexpression of Drosophila melanogaster uL11 enhances the transcription of many RPGs and Ribosomal Biogenesis genes (RiBis) suggesting that uL11 might globally regulate the level of translation through its transcriptional activity. Moreover, uL11 trimethylated on lysine 3 (uL11K3me3) interacts with the chromodomain of the Enhancer of Polycomb and Trithorax Corto, and both proteins co-localize with RNA Polymerase II at many sites on polytene chromosomes. These data have led to the hypothesis that the N-terminal end of uL11, and more particularly the trimethylation of lysine 3, supports the extra-ribosomal activity of uL11 in transcription. To address this question, we mutated the lysine 3 codon using a CRISPR/Cas9 strategy and obtained several lysine 3 mutants. We describe here the first mutants of D. melanogaster uL11. Unexpectedly, the uL11K3A mutant, in which the lysine 3 codon is replaced by an alanine, displays a genuine Minute phenotype known to be characteristic of RPG deletions (longer development, low fertility, high lethality, thin and short bristles) whereas the uL11K3Y mutant, in which the lysine 3 codon is replaced by a tyrosine, is unaffected. In agreement, the rate of translation decreases in uL11K3A but not in uL11K3Y. Co-immunoprecipitation experiments show that the interaction between uL11 and the Corto chromodomain is impaired by both mutations. However, Histone Association Assays indicate that the mutant proteins still bind chromatin. RNA-seq analyses from wing imaginal discs show that Corto represses RPG expression whereas very few genes are deregulated in uL11 mutants. We propose that Corto, by repressing RPG expression, ensures that all ribosomal proteins are present at the correct stoichiometry, and that uL11 fine-tunes its transcriptional regulation of RPGs.


Introduction
Drosophila Minute mutants have been studied for almost a hundred years.They were first described for displaying thin and short bristles, i.e.Minute bristles, together with prolonged development [1].All Minute mutations are dominant and lethal when homozygous.The vast majority of them strongly impact viability and fertility, to the point that some could only be identified through transient aneuploidy experiments.Minute loci have been characterized over time and all but a few of Minute genes have been identified as Ribosomal Protein Genes (RPGs) [2,3].
A substantial part of RPs' contribution to cell metabolism has been attributed to their ability to alter ribosome behaviour with consequences on protein synthesis.However, some free RPs are also known for a long time to carry regulatory activities, consequently termed "extra-ribosomal functions" (for a review see [4]).This is notably the case for RPL12, aka uL11 following the new nomenclature proposed to avoid confusion between species [5].Indeed uL11 was shown in C. elegans and mammals to bind its own messenger RNA and inhibit its splicing, leading to aberrant transcripts targeted for degradation through nonsense-mediated mRNA decay [6,7].Furthermore, in S. cerevisiae, uL11 was shown to be required for the transcription of a subset of PHO pathway genes that are inducible under low phosphate conditions [8].
Drosophila melanogaster uL11 is encoded by a unique gene (uL11/RPL12/ CG3195/ FBgn0034968) located on the right arm of chromosome 2 at cytogenetic position 60B7.Three transcripts corresponding to uL11 encode the same 165 amino acid protein.uL11 expression is ubiquitous and described as "very high" to "extremely high" in all tissues, developmental stages, and cell lines [9].Two deletions encompassing the whole uL11 area have been described [i.e.Df(2R)bw VDe2 LPx KR and Df(2R)Exel6081].However, it was recently proposed that the cytological borders of the first one did not cover the uL11 region (FBrf0230794) and the second one was lost (FBrf0206661).Thus, there is no evidence that aneuploidy at this locus can be viable, and the locus is indeed described as haplolethal [4,10,11].Another RPG (eL39/RpL39), and several essential genes (eEF5, yki) are found in the vicinity of uL11 [11,12].For this reason, the genes responsible for this haplolethality remain uncertain.However, it is likely that uL11 contributes to this phenotype as no classical allele has been described yet.Moreover, we observed that ubiquitous RNAi-mediated uL11 inactivation is lethal during the first larval instar.
The uL11 protein forms, together with the ribosomal protein uL10, the basis of the P-stalk, a lateral protuberance of the 60S subunit which is a critical element of the ribosomal GTPaseassociated center known to interact with factors involved in translational elongation and termination [13].The uL11 protein consists of two domains connected with a hinge, a C-terminal globular domain anchored to a conserved region of the 28S rRNA and a flexible N-terminal domain.The N-terminal domain interacts with several translation factors, notably eEF2 [13,14].As a consequence, uL11 has been shown to play an important role at many steps of the translational cycle.In yeast, a deficiency for uL11 prevents the release of the ribosome associated protein Tif6, which is the last maturation step before the 60S subunit becomes functional and uL11 deficient ribosomes display a decrease in translational fidelity [15].In cultured human cystic fibrosis bronchial epithelial cells, depletion of uL11 reduces the rate of translational initiation and elongation [16].
We have recently established that Drosophila melanogaster uL11 interacts with the chromodomain of the epigenetic co-factor Corto [17].Chromodomains recognize trimethylated lysines on histones, and the interaction between uL11 and Corto was shown to require the trimethylation of uL11 on lysine 3 (uL11K3me3).Trimethylation of uL11 lysine 3 is very well conserved, existing also in S. cerevisiae, S. pombe and A. thaliana [18][19][20].In D. melanogaster, uL11 and Corto bind chromatin, colocalize at many sites on polytene chromosomes and are recruited on the hsp70 gene upon transcriptional activation [17,21].Overexpression of uL11 as well as that of the Corto chromodomain induces the transcription of many RPGs and Ribosomal Biogenesis genes (RiBis) suggesting that uL11 could globally regulate the level of translation through its transcriptional activity.Thus, our previous study led us to hypothesize that uL11K3me3 was involved in transcriptional regulation.
To test this hypothesis, we designed a CRISPR/Cas9 strategy for mutating the uL11 lysine 3 codon.In this study we have replaced the lysine 3 codon of uL11 with codons for amino acids that are not subject to methylation; alanine (uL11K3A) and tyrosine (uL11K3Y).The uL11 K3A mutant, but surprisingly not the uL11 K3Y mutant, displays Minute phenotypes i.e. longer development, low fertility, high lethality and thin bristles.As expected, uL11K3A and uL11K3Y proteins do not interact with the Corto chromodomain anymore.However, they still bind chromatin.RNA-seq analyses from wing imaginal discs show that Corto represses RPG expression whereas very few genes are deregulated in uL11 mutants.

Drosophila genetics
Drosophila melanogaster stocks and crosses were kept on standard yeast corn-meal medium (7.5% yeast, 9% cornmeal, 1% agar, 0.27% Moldex) at 25˚C.For all experiments, crosses were set up with similar densities to prevent confounding effects of overcrowding.w 1118 was used as the control strain.The corto 420 and corto L1 loss-of-function alleles were described in [17].
A 123 nucleotide-long single-stranded oligodeoxynucleotide (ssODN) carrying the lysine (AAA) to alanine (GCC) substitution flanked by two 60 nucleotide-long homology arms was used as a template for Homology-Directed Repair (HDR) (synthetized by Integrated DNA Technologies Inc) (S1 Table ).The uL11 region of the recipient line vasa-Cas9 (BL-51324) was sequenced in order to respect possible polymorphisms.To prevent base pairing with the sgRNA, the ssODN was designed to be homologous to the PAM carrying strand.

Fly transgenesis
Two hundred vasa-Cas9 embryos were injected with a mixture containing 100 ng/μL pU6-chiRNA:sgRNA and 100 ng/μL ssODN (BestGene Inc.).Transformant G0 flies (48 females and 44 males) were individually crossed to w 1118 ; In(2LR)Gla, wgGla1/SM5 flies (Gla/SM5).Among them, only 18 males and 11 females were fertile.Curly wing G1 siblings were individually crossed to Gla/SM5 flies.Once the G2 progeny born, G1 founding flies were harvested, genomic DNA extracted, and LNA allele-specific genotyping performed as described below.294 G1 individuals were genotyped to detect the presence of uL11 mutant alleles.Curly wing G2 offspring of G1 flies carrying a mutant allele of uL11 were then crossed with each other to establish mutant balanced strains.In order to eliminate potential unspecific mutations, ten balanced mutant females from the offspring were then individually backcrossed with w 1118 males and genotyped by High Resolution Melting Analysis (HRMA) after laying eggs.uL11 mutant females were kept and the whole procedure was repeated seven times.

Genomic DNA extraction
Genomic DNA was extracted by crushing single flies in 100 μL SB buffer (10 mM Tris pH 8.0, 1 mM EDTA, 25 mM NaCl, 200 μg/μL Proteinase K), followed by 30 min incubation at 37˚C.DNA was further purified by standard phenol-chloroform extraction followed by ethanol precipitation.

Genotyping by locked nucleic acid allele-specific qPCR
Forward allele-specific primers with 3' end matching either wild-type (LNAWT) or mutated 3 rd codon (LNAK3A) of uL11 (AAA or GCC, respectively) and a Locked Nucleic Acid (LNA) nucleotide [24] at the second position of the mismatch codon were used in combination with a reverse primer (CRISPR1_R) to amplify a 219 nucleotide fragment (S1 Table ).25 μL reactions were set to contain 5 to 15 ng of genomic DNA, 0.5 μM forward and reverse primers, 0.4 nM dNTP, 0.75 μL SYBR green (Diagenode), and 2.5 units of DreamTaq polymerase (Thermo Fisher Scientific) in TMAC buffer (67 mM Tris pH 8.8, 6.7 mM MgCl2, 16.6 mM ammonium sulfate, 0.5 mM tetramethylammonium chloride, 0.17 mg/mL BSA) [25].0.5 ng of plasmid containing the uL11 coding region in which the AAA lysine 3 codon was replaced by GCC was used as positive control.qPCR reactions were carried out in a CFX96 system (BioRad) [95˚C 3 min; 40 cycles (95˚C 20 s, 64˚C 20 s, 72˚C 30 s)].To confirm the presence of the mutated allele, a 1.5 kb region centred on the lysine 3 codon was sequenced.

High Resolution Melting Analysis (HRMA)
Genomic DNA was analysed by HRMA as described in [26].Briefly, oligonucleotides uL11-HRMA-F and uL11-HRMA-R (S1 Table ) were used to amplify a 173 bp region centred on the uL11 lysine 3 codon.PCR reactions were performed with SsoFast TM EvaGreen1 Supermix (Bio-Rad) in 20 μL reactions containing 2 to 15 ng genomic DNA and 0.5 μM each oligonucleotide.Cycles were carried out in a CFX96 system (BioRad) [98˚C 3 min; 40 cycles (98˚C 2 s, 57.3˚C 15 s)].Thermal melting profiles were obtained in the same device by increasing temperature from 75 to 95˚C using a temperature increment of 0.2˚C.They were normalized as described by [27].

uL11K3me3 antibodies
Polyclonal anti-uL11K3me3 antibodies were generated in rabbit using a peptide corresponding to the first 16 amino acids of uL11 with methylated lysine 3 [PPK(me3)FDPTEVKL-VYLRC] (Eurogentec).The serum was first loaded on a uL11K3me3 peptide affinity column which allowed to retain anti-uL11K3me3 and anti-uL11 antibodies.After elution, they were separated by passage through an unmethylated uL11 peptide affinity column.Specificity of the antibodies was checked by dot blot (S1 Fig).

Analysis of mutant life history traits
uL11 wild-type or mutant chromosomes were balanced with CyO,Dfd-EYFP (from strain BL-8578) or SM5.About 100 females and 60 males were placed in laying cages on agarose plates (2% agarose, 5% vinegar, neutral red) supplemented with yeast.To measure embryonic lethality, 100 embryos were collected from each laying cage, transferred on new agarose plates and emerging first instar larvae were counted.To measure larval and pupal lethality, 100 embryos were collected and transferred into yeast cornmeal medium tubes at 25˚C.Pupae and adults were then counted.Three independent experiments were performed and results were pooled.To measure developmental time, first instar larvae were collected and transferred into yeast cornmeal medium tubes at 25˚C (50 to 100 larvae per tube).Vials were checked from 9 days after egg laying until no more adults emerged.Statistical significance was assessed by Chi-2 tests.

Measure of bristle length
Adult bodies free of wings, legs and heads were aligned on agar cups.Images were captured using a Leica Model MZ FLIII microscope equipped with a Leica Model DC480 camera.Scutellar bristles were measured using the ImageJ segmented line tool.Normality was checked by Shapiro-Wilk tests and homogeneity of variances by F tests.Student's t-tests were then set taking into account homo-or heteroscedasticity of variances.

Measure of wings
Adult flies were kept in 70% ethanol for 48 h and transferred into PBS glycerol (1:1 v/v).Wings were dissected and mounted on glass slides, dorsal side up, in Hoyer's medium.Slides were scanned with a Hamamatsu Nanozoomer Digital Slide scanner, running the Nanozoomer software with a 20x objective and an 8 bit camera.Wing pictures were separately exported into TIF format using NDP.view and the 5 x lens.Measurements of wing length were performed as described in [28].

Plasmids
uL11 was amplified from w 1118 embryonic cDNAs and subcloned into pENTR/D-TOPO (Invitrogen) [17].pENTR-uL11 K3A and pENTR-uL11 K3Y were obtained by site-directed mutagenesis using the oligonucleotides described in [17] and in S1 Table, respectively.The cDNAs were then transferred either into the pAWM or the pAWH Gateway1 Drosophila vectors allowing expression of fusion proteins with a C-terminal Myc or HA tag under the control of an actin promoter in S2 cells.

Polysome fractionation
Cells were harvested at 50% confluence and washed in Schneider medium at room temperature to remove the fetal bovine serum.They were then resuspended in ice-cold lysis buffer (20 mM Hepes pH 7.5, 250 mM KCl, 10 mM MgCl2, 5 mM DTT, 1 mM EDTA, 0.5% NP-40) supplemented with EDTA-free protease inhibitor cocktail (Roche Diagnostics) and 40 U/mL Ribolock RNAse Inhibitor (ThermoFisher).For EDTA treatment, the same buffer adjusted to 25 mM EDTA and without MgCl2 was used.After centrifugation at 500 g for 5 min to pellet nuclei, supernatants were layered onto 10 to 50% sucrose gradients in polyribosome buffer (20 mM Hepes pH 7.5, 250 mM KCl, 20 mM MgCl2, 2 mM DTT), supplemented with EDTA-free protease inhibitor cocktail and 40 U/mL Ribolock RNAse Inhibitor.Gradients were centrifuged at 39,000 rpm for 165 min at 4˚C in a Beckman SW41-Ti rotor.Optical density at 254 nm was monitored using a density gradient fractionator (Teledyne Isco, Lincoln, NE).
Western blots were performed according to standard protocols using mouse anti-HA (Sigma F2411; 1/1000) as primary antibodies and anti-mouse (Sigma NA931; 1/20000) as secondary antibodies.They were revealed using the Supersignal™ West Pico PLUS Chemiluminescent Substrate (Thermo Scientific) as described by the supplier.

Puromycin assays
Puromycin assays were adapted from [29] with the following modifications: 20 third instar larvae were turned inside-out and incubated for 1 h at 25˚C under gentle rotation in Schneider's medium supplemented or not with 10 mg/mL cycloheximide (Sigma).Puromycin (antpr1, InvivoGen) was then added at a final concentration of 0.28 mg/mL and incubation was continued for 2 h.Total proteins were extracted in a buffer containing 30 mM Hepes pH 7.4, 0.1% NP40, 150 mM NaCl, 2 mM Mg(OAc) 2 supplemented with phosphatase and protease inhibitors (Roche) (adapted from [30]).60 μg of protein extracts were deposited on a 12% acrylamide gel.
Western blot were performed according to standard protocols using mouse anti-puromycin (Kerafast, 3RH11; 1/500) or mouse anti-H3 (Diagenode; C15200011; 1/1000) as primary antibodies, and anti-mouse (Sigma; NA931; 1/20000) as secondary antibodies and revealed using the Supersignal™ West Pico PLUS Chemiluminescent Substrate (Thermo Scientific).Puromycin and H3 signals were measured using ImageJ.The puromycin signal (signal in the samples treated with CHX and puromycin minus signal in the untreated sample) was normalized towards the H3 signal.Statistical significance was assessed by Student's t-tests.

RNA-seq, bioinformatic analyses and RT-qPCR
Wing imaginal discs of third instar female larvae (one disc per larva) were dissected by batches of 50 in ice-cold PBS and frozen in liquid nitrogen.150 discs (three batches) were pooled.Total RNAs were extracted using RNeasy kit (Qiagen).
RT-qPCR were performed on wing imaginal disc cDNAs as described in [28].Expression levels were quantified with the Pfaffl method [33] and normalized to the geometric mean of two reference genes, GAPDH and Spt6, the expression of which did not vary in the mutants.Sequences of primer pairs are listed in S1 Table.

CRISPR/Cas9 editing of the uL11 lysine 3 codon
uL11 is located within a cluster of highly transcribed genes, many of which are also essential (including eEF5, RpL39/eL39, yki. ..) [11,12].This cluster is indeed part of the 1.6% haplolethal regions of the euchromatic Drosophila genome [3].Furthermore, uL11 is bordered by two small intergenic sequences (465 and 620 bp, respectively) that might contain regulatory elements (Fig 1).Thus, the insertion of a selection cassette within this locus could disrupt gene expression and impede viability.We therefore chose to edit the uL11 lysine 3 to alanine (K3A) by a single step CRISPR/Cas9 mediated HDR using a single-stranded oligodeoxynucleotide donor template (ssODN) [23,34].
To recover the successful HDR events, we set up a molecular screening protocol based on an allele specific amplification strategy [34].Discriminating power was increased by the substitution of the penultimate nucleotide of the screening primers with a locked nucleic acid (LNA) (S2 Fig) .The presence of a single LNA sufficiently improved specificity to allow us analyzing pools of flies for the presence of a single allele copy (S2 Fig) .294 G1 individuals were tested for the presence of a mutated allele carrying the K3A substitution.Mixtures of genomic DNA from pools of 4 to 5 individuals were prepared and the uL11 locus was amplified with either the lysine codon (LNA-WT) or the alanine codon (LNA-K3A) matching primer.While most genomic DNA mixtures displayed amplification kinetics similar to the one of the negative control, six of them exhibited faster amplification (ΔCt between 2 and 7).We thus repeated the experiment on individual genomic DNAs from 6 positive pools.Ten genomic DNA originating from three independent G0 founding flies, two males and one female, exhibited quicker amplification with the LNA-K3A primer than with the control primer (ΔCt>5) (S2 Fig) .Sequencing the uL11 locus confirmed that these flies were heterozygous for the recombinant allele uL11 K3A .Three uL11 K3A mutants, uL11 K3A-12 , uL11 K3A-43 and uL11 K3A-6 , coming from three different founders, were obtained (S3 Fig) .Unless specified, analyses were performed with uL11 K3A-43 .
To detect other mutations potentially resulting from non-homologous end joining (NHEJ) events, we also performed HRMA of a qPCR amplicon centred on the uL11 lysine 3 codon.Denaturation kinetics of these PCR products were analyzed individually for the 294 G1 flies.Among them, 36 denaturation profiles differed from the wild-type control.Sequencing of the amplicons confirmed the presence of a mutation at the uL11 locus in each of these 36 samples.Consistently, the 10 uL11 K3A mutants identified with the allele-specific amplification strategy were also recovered by HRMA.Seven additional alleles were thus identified, that all carry a mutation impairing the lysine

Lethality and developmental delay of uL11 mutants
We first examined the lethality of uL11 homozygous mutants.Whereas we obtained many homozygous uL11 K3Y adults, very few uL11 K3A homozygotes emerged and almost all of them were males.To follow the lethality of uL11 K3A mutants during development, we compared the number of embryos, larvae and pupae to the one of the w 1118 control.During embryogenesis, they did not display more lethality than the w 1118 control with the exception of uL11 K3A /uL11 + whose lethality is slightly higher (Chi 2 test, p<0.05) (Fig 3).By contrast, during larval life, lethality was very high for uL11 K3A /uL11 + , and uL11 K3A /uL11 K3A (Chi 2 test, p<0.001), but did not increase neither for uL11 K3Y /uL11 + nor for uL11 K3Y /uL11 K3Y larvae.We did not observe any lethality during the pupal life for all genotypes.Similarly, developmental time from egg deposition to adult emergence was considerably extended for uL11 K3A /uL11 + (up to 48 h) and uL11 K3A /uL11 K3A (up to 96 h) and for a second uL11 K3A mutant, uL11 K3A-12 , either heterozygotes or homozygotes (S4 Fig) .However, the developmental time of uL11 K3Y /uL11 K3Y flies was unaffected (Fig 3).
To summarize, the K3Y mutation had no effect on these life history traits whereas lethality and developmental time were increased both in heterozygous and homozygous K3A mutants, which characterized this allele as dominant.

Bristle and wing size of the uL11 mutants
Minute mutants have thinner and shorter bristles as compared to wild-type flies.As it also seemed to be the case for uL11 K3A mutants either heterozygotes or homozygotes (Fig 4 and S4 Fig) , we measured the anterior and posterior scutellar bristles of uL11 K3A and uL11 K3Y mutants.In males, scutellar bristles were indeed significantly shorter in heterozygous and homozygous uL11 K3A mutants as compared to control flies but unaffected in homozygous uL11 K3Y mutants (Fig 4).In females, we also observed shorter bristles in heterozygous and homozygous uL11 K3A mutants while bristles of uL11 K3Y homozygous mutants were only slightly affected (S5 Fig) .Furthermore, heterozygous and homozygous uL11 K3A males have shorter wings while those of uL11 K3Y homozygous males were unaffected (Fig 4).Similarly, heterozygous uL11 K3A females exhibited shorter wings whereas wings of homozygous uL11 K3Y females were unaffected (S5 Fig).
These results confirmed that uL11 K3A exhibits characteristics of a dominant allele and showed that the severity of the phenotypes depended on the mutation.The K3A mutation was highly detrimental while the K3Y mutation had almost no impact on the size of scuttelar bristles and wings.Minute mutants are known to be poorly fertile and viable, to exhibit developmental delay and have shorter and thinner bristles, all phenotypes that we observed in uL11 K3A mutant flies.Moreover, Minute alleles are dominant which is also the case of the uL11 K3A allele.All these data characterized the uL11 K3A mutant as a Minute mutant.

Impact of the uL11 mutations on translation
The strategic location of uL11 protein at the basis of the P-stalk in the GTPase-associated center of the ribosome suggests that its mutation might have a detrimental impact on translation.To test this hypothesis, we assessed the level of global translation in the uL11 mutants.In order to label neo-synthesized proteins, third instar wild-type or mutant larvae were incubated with puromycin.Puromycin intake was normalized to histone H3 levels.A significant decrease in global translation level was observed in uL11 K3A homozygous larvae as compared to wild-type larvae, whereas it was unmodified in uL11 K3Y homozygous, in agreement with the absence of Minute phenotypes (Fig 5).uL11 K3A heterozygous larvae, which however present Minute phenotypes, did not exhibit any decrease in the global translation level.On the one hand, the phenotypes of uL11 K3A / uL11 + flies are not as severe as those of uL11 K3A / uL11 K3A , and a low decrease in global translation might be undetectable in our assays.On the other hand, the amount of protein synthesis is known to vary depending on the proliferative activity of the tissue or even depending on the cell type [35].For example, it is visibly impacted by heterozygous RPG mutations in clonal analyses of cell competition in Drosophila larvae wing imaginal discs [36,37].A low decrease of protein synthesis in larval proliferative tissues might be averaged in our global analyses.
We then asked whether the uL11K3A protein would retain the ability to associate with translating ribosomes.We generated stable cell lines expressing either uL11K3A-HA or uL11-HA under the control of the Actin promoter.Cytoplasmic extracts were purified from both genotypes and lysates were loaded onto sucrose gradients for fractionation.As a control, an extract of each genotype was supplemented with 25 mM EDTA, a concentration that disrupts the interaction between ribosomal subunits and mRNA.The resulting fractions were analyzed by western blot to reveal the presence of uL11K3A-HA or uL11-HA (Fig 5).These experiments revealed an enrichment of uL11-HA in the 60S fraction (large ribosomal subunits), in the 80S fraction (ribosomes) and in polysomes, but not in the 40S fraction (small ribosomal subunit), as expected.The same pattern was observed for uL11K3A-HA.Furthermore, EDTA treatment triggered the relocation of uL11-HA and uL11K3A-HA towards lighter fractions, confirming that the sedimentation profiles truly resulted from their association with polysomes (Fig 5).All these data showed that uL11K3A-HA was efficiently incorporated into translating ribosomes.Nevertheless, the decrease in puromycin incorporation observed in uL11 K3A mutants suggests that the yield of translation might be altered.

Interaction of uL11 mutant proteins with Corto and chromatin
Overexpression of uL11 as well as overexpression of the Corto chromodomain enhances the transcription of RPGs and RiBi genes [17].As both proteins interact together and bind chromatin, we hypothesized that their transcriptional activity were linked.We have previously shown that the trimethylation of uL11 on lysine 3 mediates the interaction between uL11 and the chromodomain of the epigenetic cofactor Corto (CortoCD).We first assayed the physical interactions between CortoCD and the mutant uL11 proteins.To do this, we co-transfected Drosophila S2 cells with pA-FLAG-CortoCD and either pA-uL11 K3A -Myc or pA-uL11 K3Y -Myc.The FLAG-tagged chromodomain was immunoprecipitated using anti-FLAG antibodies.Contrarily to uL11-Myc that co-immunoprecipitated with FLAG-CortoCD, neither uL11K3-A-Myc nor uL11K3Y-Myc co-immunoprecipitated with it, corroborating our previous results showing that uL11K3 trimethylation mediates the interaction with CortoCD (Fig 6).
We next asked whether the uL11 mutant proteins were still able to bind chromatin.To address this question, we performed Histone Association Assays (HAA) [32].Chromatin was extracted from S2 cells permanently expressing uL11-HA, uL11K3A-HA or uL11K3Y-HA and purified by immunoprecipitation with an anti-Histone H3 antibody.As expected, Histone H2B was co-immunoprecipitated with H3.Not only the wild type protein, but also uL11K3A and uL11K3Y mutant proteins were found associated with chromatin suggesting that their ability to modulate transcription might be kept (Fig 7).

Transcriptomic analysis of uL11 and corto mutants
We then performed a transcriptomic analysis from wing imaginal discs of third instar larvae either trans-heterozygous for two loss-of-function alleles of corto (corto L1 /corto 420 ) or homozygous for the uL11 alleles (uL11 K3A or uL11 K3Y ).The w 1118 line was used as reference.Total numbers of reads are shown in S2 Table .Differential analyses were performed to obtain adjusted p-values associated to expression fold-changes for the three genotypes as compared to the reference.Taken a log 2 (fold-change) < -0.50 or > 0.50 and an adjusted p-value < 0.05, we found 458 down-regulated and 481 up-regulated genes in corto L1 / corto 420 mutants (S3 Table ).Strikingly, 241 of these deregulated genes were also deregulated in wing imaginal discs overexpressing the Corto chromodomain [21] and all of them were deregulated in the same direction.Notably, 55.5% of the genes corresponding to the GO Cellular Component term "cytoplasmic translation" (66/119) were up-regulated in corto L1 /corto 420 mutants and most of them (54/66) were also up-regulated in wing imaginal discs overexpressing the Corto chromodomain (S4 Table ).These results indicated that CortoCD overexpression behaved as a dominant negative allele and confirmed that Corto down-regulates RPGs, directly or indirectly.
Using the same cutoffs, we found 143 down-regulated and 251 up-regulated genes in the uL11 K3A mutant.Down-regulated genes were enriched in GO terms related to transcription factor activity and sequence-specific DNA binding whereas up-regulated ones were enriched in GO terms glutathione metabolic process, telomere maintenance and DNA recombination (S4 Table ).Only few genes were deregulated in uL11 K3Y (39 down-regulated and 45 up-regulated).Up-regulated genes were enriched in the category "glutathione metabolism" (KEGG pathway, Benjamini adjusted p-value 5.90E-03) as for uL11 K3A .Fifty-two deregulated genes were shared with uL11 K3A (S3 Table ).Most of them were deregulated in the same direction (29 genes up-regulated and 17 genes down-regulated).Hence, contrarily to uL11 overexpression [17], RPGs and RiBi genes were not deregulated in uL11 mutants.Interestingly, among the 251 genes up-regulated in uL11 K3A , 82 were up-regulated in other RPG mutants [38,39] (S3 Table ).
Very few deregulated genes were shared between the uL11 mutants and corto (69 shared by uL11 K3A and corto, 29 shared by uL11 K3Y and corto, 14 shared by the three genotypes) (Fig 8).We chose some genes to analyze their expression by RT-qPCR in the three mutants: CG13516, one of the few genes deregulated in the three genotypes and the most down-regulated gene in uL11 K3Y , Hsp67Bc, encoding a small heat-shock protein involved in cold stress tolerance [40], which is up-regulated in corto L1 /corto 420 and uL11 K3Y , and GstE6, encoding a Glutathione Stransferase up-regulated in uL11 K3A and uL11 K3Y .The RNA-seq data were confirmed by RT-qPCR for these three genes in the three mutants, except for Hsp67Bc for which the increase in corto L1 /corto 420 was not significant (Fig 8).

Discussion
We have previously shown that ribosomal protein uL11 interacts with the chromodomain of the Drosophila Enhancer of Trithorax and Polycomb Corto when tri-methylated on lysine 3 (uL11K3me3) [17].uL11, Corto and RNA Polymerase II co-localize at many sites on polytene chromosomes and overexpression of uL11 induces the transcription of many RPGs and RiBi genes.These data have confirmed that Drosophila uL11 is involved in transcription [21] and further suggest that the lysine 3 supports this extra-ribosomal activity.In the aim of testing this hypothesis, we generated mutant alleles of uL11 using the CRISPR/Cas9 technology.By introducing a template to promote Homology-Directed Repair, we obtained a mutant in which the lysine was replaced by an alanine.However, we also obtained mutants harbouring indels probably obtained by Non-Homologous End Joining.Strikingly, the lysine 3 codon of uL11 was either deleted or substituted by another amino acid in all recovered mutants.Hence, it seems that a strong selection pressure occurs to maintain the uL11 ORF, consistent with the haploinsufficiency of this RPG.

A single amino acid substitution generates a Minute phenotype
The uL11 K3A allele is almost totally lethal at the homozygous state.The few escapers are only males that hatch with a developmental delay larger than two days as compared to wild-type flies raised in the same conditions.Interestingly, uL11 K3A heterozygotes are also delayed but slightly less (about one day).In addition, uL11 K3A heterozygous females are frequently sterile making it necessary to carry out crosses of heterozygous uL11 K3A males with wild-type females and to genotype the offspring in order to maintain the stock.Hence, the uL11 K3A allele clearly appears dominant, which is also visible for other phenotypes, such as the shorter and thinner macrochaetes.Reduced viability, notably of females, delayed development, thin bristles and dominance are signatures of the Minute mutations that have been shown to correspond to RPG deletions [2].The associated phenotypes are thought to reflect a defect in RPs' stoichiometry resulting in a decreased capacity for protein synthesis.In accordance, tissues that are the most dependent on translation are likely to be the most affected.For instance, high ribosome biogenesis level is suggested to be necessary for the maintenance of germinal stem cells in the Drosophila ovarium, which could explain the reduced fertility of Minute females [41,42].Similarly, macrochaetes are described to require a very high amount of protein synthesis over a short developmental period [2].Unexpectedly, replacement of a single lysine by an alanine in the N-terminal tail of uL11 induces a Minute phenotype.The global level of translation in this mutant is decreased although the mutated protein can be efficiently incorporated into translating ribosomes.However, translation speed or accuracy might be altered.Whatever the exact origin, the Minute phenotype of the uL11 K3A mutant might be due to an alteration in translation.In yeast, uL11 loss-of-function has been shown to halve the speed of translation and to cause increased amino acid mis-incorporation and termination codon readthrough [15].Another possibility would be that uL11K3A ribosomes display altered affinity for specific mRNA, as has been described for ribosomes lacking RPL38/eL38 in mice [43].
Surprisingly, the uL11 K3Y mutant does not display any Minute phenotype.The N-terminal extension of uL11 is supposed to be unstructured as neither the first 6 nor the first 9 amino acids were resolved in the structure of the D. melanogaster and S. cerevisiae 80S ribosome, respectively [14,44].This region is composed of hydrophilic amino acids and could thus be use as a platform for protein interactions.Addition of methyl groups on lysines, even if it has no effect on the overall charge of the residue, increases its hydrophobicity.Trimethylation of uL11 lysine 3 might thus modulate the activity of the N-terminal tail [45].However, in S. cerevisiae and pombe, deletion of Rkm2, the methyltransferase responsible for uL11 lysine 3 methylation has little impact on ribosome assembly and function or on cell viability [19,46] suggesting that the methylation of lysine 3 is not crucial for basic translational activity.In the uL11 K3Y mutant, the tyrosine, a hydrophobic residue, might mimic the effect of methyl groups.If this is the case, in-depth comparison of translation between uL11 K3Y and uL11 K3A mutants, including sequencing of polysomal mRNA, should permit to elucidate the role of the uL11 Nterminal tail methylation in translation.

Extra-ribosomal activities of uL11
Regulation of translational capacity might be indirect and due to extra-ribosomal activities of free RPs, notably in transcriptional regulation (for a recent review see [47]).Overexpression of uL11 increases the transcription of many RPGs and RiBi genes [21].Although it cannot be ruled out that artificial overexpression induces a non-specific response, it is striking that overexpression of RPGs is also observed in loss-of-function corto mutants that encode a direct partner of uL11.Corto, by repressing RPG expression, could ensure that all ribosomal proteins are present at the correct stoichiometry, thus preventing ribosomal stress.As the uL11 K3Y mutant displays neither Minute phenotypes nor a global decrease in translation, we assumed that solely its transcriptional activity would be affected.However, only very few genes are deregulated in this mutant questioning the existence of a proper transcriptional activity for uL11.To reconcile these findings, it is tempting to speculate that uL11, by physically interacting with Corto on chromatin, fine-tunes its transcriptional regulation of RPGs.In a context where uL11 no longer interacts with Corto, i.e. in the uL11 lysine 3 mutants, the expression of these genes does not vary.Going with that model, if other ribosomal proteins were now out of stoichiometry (for example in another Minute mutant), one would expect that the extra uL11 would accumulate, bind Corto, and repress the expression of other RPG genes to restore balance.
Alternatively, the absence of obvious phenotypes and the fact that only few genes are deregulated in uL11 K3Y mutant could mean that the transcriptional activity of uL11 is triggered under stress conditions which is the case for many extra-ribosomal functions of RPs.For example, in S. cerevisiae uL11 has been shown to regulate the PHO pathway in low phosphate conditions potentially at the transcriptional level [8].It would thus be interesting to test whether the Drosophila uL11 mutants, and especially uL11 K3Y , display an altered resistance to stresses that affect ribosome biogenesis, for instance by raising them with specific food diets.
3 codon: a single (K3Y) or double (P2LK3E, P2QK3R) amino acid substitutions, a single (ΔK3) or double (ΔK3F4) amino acid deletion, and an insertion of 2 or a deletion of 4 nucleotides (F+2 and F-4, respectively) (S3 Fig).Preliminary observations revealed that the mutants could be dispatched into two groups depending on the severity of their phenotypes.The first group contains the K3A, ΔK3, ΔK3F4, F+2 and F-4 alleles, and the second group the K3Y, P2QK3R and P2LK3E alleles.We choose to focus on two representative alleles, K3A and K3Y (Fig 1 and S2 and S3 Figs).They were introduced into the same

Fig 5 .
Fig 5. Translational rate of uL11 mutants.A-Western blot showing puromycin incorporation in larvae of the two mutants as compared to wild-type larvae, in absence (-) or presence (+) of cycloheximide (CHX), an inhibitor of translation.Puromycin incorporation was revealed with an antipuromycin antibody.Histone H3, revealed with an anti-panH3 antibody, was used as loading control.MWM: molecular weight marker.B-Quantification of the puromycin signal in the 4 genotypes without (-) or with (+) CHX treatment.The puromycin signal was normalized to the H3 signal.Student's t-tests were performed to compare puromycin incorporation in mutant and wild-type larvae.� p-value < 0.05; ns: non significant.C, C', D, D'-Polysome fractionation: cytoplasmic lysates (C, D) and EDTA-treated lysates (C', D') from S2 cells expressing uL11-HA (C, C') or uL11K3A-HA (D, D') were fractionated by centrifugation onto a sucrose gradient.Optical density at 254 nm was monitored during fractionation (top panels).The peaks observed in the gradient correspond to the different ribosomal complexes: 40S subunit, 60S subunit, 80S monosome, polysomes.Proteins extracted from fractions were analyzed by Western blotting with anti-HA antibody (lower panels).A vertical line indicates that different wells from the same gel were juxtaposed in the image for clarity.Images are representative for three obtained replicates.https://doi.org/10.1371/journal.pone.0273198.g005