Ribosomal protein imbalance launches a C/EBP-based program to preserve tissue integrity

Genes encoding ribosomal proteins are expressed at rate limiting levels, rendering their biological function highly sensitive to the copy-number variation that results from genomic instability. Cells with a reduced number of ribosomal protein genes (RPGs) are eliminated, when intermingled with wild type cells, via a process known as cell competition. The mechanisms underlying this phenomenon are poorly understood. Here we report the function of a CCAAT-Enhancer-Binding Protein (C/EBP), Xrp1, that is critically required for the elimination of cells with a hemizygous RPG genotype. In such cells, Xrp1 is transcriptionally upregulated by an autoregulatory loop and is able to trigger cell elimination. Since genomic instability is likely to cause the loss of a haploinsufficient RPG, we propose a molecular model of how RPGs, together with a C/EBP-dependent transcriptional program, could preserve the genomic integrity of tissues.


Introduction
Multicellular organisms maintain genomic stability via the activation of DNA repair mechanisms to identify and correct damages present in their DNA, cell cycle arrest to prevent the expansion of DNA damaged cells, and finally programmed cell death to eliminate cells irremediably damaged (Ciccia and Elledge, 2010). The P53 transcription factor plays an evolutionary conserved role in the induction of apoptosis following DNA damage, however evidence points towards the existence of alternative routes for the induction of apoptosis in response to DNA damage (McNamee and Brodsky, 2009;Titen and Golic, 2008). It has been proposed that one of these routes relies on the detection of copy number reduction affecting haploinsufficient ribosomal protein genes (hRPGs) (McNamee and Brodsky, 2009).
Ribosomes are essential macromolecular machines that catalyze the synthesis of proteins in all cells; they consist of a set of ribosomal proteins (RPs) that surround a catalytic core of ribosomal RNAs (rRNAs). The coordinated function of RPs is well illustrated in D. melanogaster. In this model organism, the majority of RPGs is haploinsufficient and give rise to the same dominant phenotype referred to as the Minute phenotype. This phenotype is characterized by a general developmental delay and improper bristle development (Marygold et al., 2007). Haploinsufficient RPGs (hRPGs) are widely distributed across the genome, and owing to their dominant adult phenotype these loci have been used to probe the genetic consequences of diverse sources of chromosome damage (Dekanty et al., 2012). This suggests that hRPGs may be used to report loss of genetic integrity.
In addition, it has recently been shown that removal of one functional copy of a hRPGs activates a number of genes involved in the maintenance of genetic integrity (Kucinski et al. 2017). Furthermore, such prospective loser cells are eliminated when intermingled with wild-type cells, a process that is referred to as "cell competition" (reviewed by Baillon and Basler, 2014). This process occurs irrespectively of the presence of a functional p53 gene (Kale et al., 2015) and therefore provides a suitable assay to uncover molecular circuitries that can trigger apoptosis in destabilized genomes.

Results and Discussion
In order to identify genes whose functions are necessary for the elimination of RPG heterozygous mutant cells, we performed a mosaic forward genetic screen using ethyl methanesulfonate (EMS) in D. melanogaster. We designed a mosaic system that allows direct screening for the persistence of otherwise eliminated loser clones (RpL19 +/-) through the larval cuticle (Fig. 1A,B). Our F1 genetic mosaic system enabled us to screen for a wide spectrum of suppressors, either dominant mutations (anywhere in the genome) or recessive mutations on the right arm of the third chromosome. In brief, the induction of a single somatic recombination event between two FRTs (FLP recognition targets) generates a RPG heterozygous mutant cell that becomes homozygous for the mutagenized right arm of the third chromosome (Fig.   1A). These loser cells/clones are induced at the beginning of larval development (L1).
If no suppressive mutation is present, these clones are efficiently eliminated over time such that they are not detectable any more by the end of the third instar larval stage (L3) when the screening is performed (Fig. 1B). This screen should achieve a high degree of specificity since a random mutagenic event is more likely to promote the elimination of a loser cell rather than suppressing it.
We screened 20,000 mutagenized genomes for the presence of mutations that would allow loser clones to persist. We retrieved 12 heritable suppressors (Supplementary Table S1) and focused our attention on three of the strongest suppressors that did not display an obvious growth-related phenotype. These suppressors did not belong to a lethal complementation group and the causative mutations were identified using a combination of positional mapping and whole-genome re-sequencing (see Experimental Procedures). Positional mapping placed the three suppressive mutations within a 100 kilobase interval that contains twelve genes. Sanger sequencing of the respective annotated exons did not reveal the presence of any mutations. We therefore complemented our initial mapping with whole-genome re-sequencing and identified three independent mutations in the introns of CG17836/Xrp1 (Fig. 1C, Supplementary Text and Supplementary Fig. S1,2).
A role for Xrp1 in loser cell elimination has been suggested by genetic association (Lee et al., 2016). Furthermore, Xrp1 expression is mildly upregulated in prospective loser cells (Kucinski et al. 2017). Despite these observations, the functional relevance of Xrp1 in cell elimination remains elusive. In order to confirm that these mutations affect the function of Xrp1 and no other unannotated gene we attempted to rescue these alleles with two different transgenes using a newly designed genetic set-up (see Fig. 1D). The WT genomic fragment restores the elimination of loser cells homozygous for the suppressive mutation Xrp1 08 while a mutated genomic fragment with a frame-shift mutation fails to do so (Fig. 1E). Taken together, these results indicate that Xrp1 is necessary for the elimination of cells impaired in ribosome biogenesis.
Using a transcriptional reporter for Xrp1 (Xrp1 02515 , containing a lacZ P-element in CG17836, Akdemir et al., 2007), we found that Xrp1 expression is upregulated in RPG +/cells, indicating that it might play an active role in the elimination of loser cells (Fig. 1F). In order to gain insights into this function we conditionally forced the expression of Xrp1 in the posterior half of the wing discs and observed a massive induction of apoptosis as revealed by anti-cleaved caspase 3 staining (Fig. 1G). We next introduced mutations into the Xrp1 coding sequence and selected for mutants where Xrp1 activity is impaired ( Fig. 2A). One of these mutants, Xrp1 61 , contains a frame shift mutation upstream of the Xrp1 basic region-leucine zipper domain (b-ZIP) ( Fig. 2B) that completely abrogates its function ( Fig. 2A). Xrp1 61 homozygous mutants are viable and display no obvious phenotypes. This null allele was then used to quantitatively assess the suppressive potential of loss of Xrp1 function on the elimination of RPG mutant cells (RpL19 +/-) and to compare it to the potential of other genetic alterations previously implicated in affecting cell competition (Fig. 2C, D, E).
We undertook a stringent comparative analysis based on the ratios between the areas of loser (RFP) and winner (GFP) clones ( Fig. 2E provides the numerical values of the Fig. 2C). We then analyzed the density distribution across individual samples of the same genotype (Fig. 2E, categorized by 0.1 increments of the ratio between GFP and RFP area). We reasoned that a genuine suppressor of RPG mutant cell elimination should not only increase the mean size of RPG mutant clones but also restore a normal distribution of RPG mutant clones. In the absence of competitive elimination the growth of RPG mutant clones is still affected but these clones should follow a size distribution that is comparable to the one followed by WT clones.
Interestingly, unlike loss of Xrp1, which fully rescues the elimination of RpL19 +/and RpL14 +/loser cells (Fig. 2 C,D,E,and Fig. S5), blocking apoptosis by means of overexpression of dIAP1 or p35, or by abrogating dronc or hid function does not fully suppress RPG +/cell elimination (Fig. 2C, D, E) suggesting that Xrp1 does more than merely induce apoptosis. Xrp1 may additionally hinder cells to progress through the cell cycle; indeed, Xrp1 expression has been reported to induce cell cycle arrest in cultured Drosophila cells (Akdemir et al., 2007). The co-overexpression of CycE (promotes cell-autonomously cell cycle entry (Neufeld et al., 1998) and dIAP1 (represses cell-autonomously apoptosis (Martin et al., 2009)  To further explore this notion we set out to identify direct genomic targets of Xrp1 by chromatin immunoprecipitation followed by deep sequencing (ChIP-seq) on wing imaginal discs (Schertel et al., 2015) (see Fig. 3A). ChIP-seq revealed a high number of Xrp1 binding sites spanning a wide range of affinity (see Fig. 3B, C), and establishes Xrp1 as a C/EBP transcription factor (see Fig. 3C, S4). Some of the top targets comprise a number of genes that are involved in cell cycle regulation, apoptosis, and other biological responses previously associated with the elimination of RPG +/mutant cells (Fig. 3B). Expression analysis at the mRNA (Fig. 3D) or protein level ( Fig. 3E and S6) reveals an up-regulation of genes involved in the induction of apoptosis (hid, Kale et al., 2015), the regulation of innate immunity (Dif, Meyer et al., 2014), and in the establishment of compensatory proliferation (Upd3, Kolahgar et al., 2015) as a response to forced Xrp1 expression. Most interestingly, Xrp1 is able to bind to its own promoter and activate its transcription (Fig. 3D, E).
The autoregulation of C/EBPs is a feature that is shared among the members of this family of proteins, and it is typically set in motion via mechanisms that are speciesspecific (Ramji and Foka, 2002). In Drosophila, the Xrp1 autoamplification loop could serve as a molecular switch to sustain high transcriptional input on Xrp1 targets to orchestrate the elimination of RPG +/mutant cells.
C/EBP transcription factors have been shown to prevent cell proliferation and induce apoptosis (Ramji and Foka, 2002). Of particular interest is the retention of C/EBP alpha in the nucleolus via binding to ribosomal DNA (Müller et al., 2010). The nucleolus is a distinct structure in the nucleus of eukaryotic cells that forms around the rDNA. It is the site of ribosome biogenesis and a major stress sensor organelle (James et al., 2014). Cells with only a single functional copy of the RpL19 gene (RpL19 +/-) have an enlarged nucleolus (Fig. 4A) since RPG insufficiency partially stalls ribosome assembly (Neumueller et al., 2013). Interestingly, Xrp1 binds with a high affinity to many rDNA loci in the genome of D. melanogaster (24 peaks mapped, Fig. 3B). The relationship of Xrp1 with the nucleolus and its functional role in the elimination of cells experiencing RPG imbalance points towards a mechanism whereby cells might use the nucleolus as a reservoir of Xrp1 and release it into the nucleus via a nucleolar stress related mechanism. Once in the nucleus, Xrp1 then switches on its own gene via an autoamplification loop and drive cell elimination (a graphical representation of this working hypothesis is presented in Fig. S7).
When intermingled with wild-type cells, cells having only one functional copy of a hRPG are eliminated in a Xrp1-dependent manner. In our experimental system the deletion of one copy of the RpL19 gene is catalyzed by the Flp/FRT recombination system which leaves no apparent lesion in the chromosomal DNA (Chen and Rice, 2003). Therefore, the trigger for cell elimination does not depend on DNA damage per se but lies within the unbalanced physiology of the cell. The protective function of Xrp1 at the tissue level and the overall benefit of detecting and eliminating cells with RP imbalance are better illustrated under stress conditions. Abrogation of Xrp1 function sensitizes animal to genotoxic stress such as UV-C (Fig. 4B, Table S2) or gamma rays (Akdemir et al., 2007). Furthermore genomic destabilization following the depletion of the spindle assembly checkpoint gene bub3 elevates the levels of Xrp1 expression in the cells that are not yet culled from the epithelia ( Fig. 4C and S6).
Hence this caretaker mechanism has the potential to preserve genomic integrity at the tissue level by eliminating viable cells that lost genomic integrity. The efficacy of this protective mechanism is probably best illustrated in the context of tumorigenesis for which genomic instability is regarded a major driving force (Hanahan and Weinberg, 2011). We therefore exploited the Flp/FRT recombination system to generate salvador -/mutant tumor clones. In this system, the loss of one functional copy of the RpL19 gene is sufficient to fully suppress the tumor growth of salvador -/mutant clones ( Fig. 4D; sav -/and sav -/-, RpL19 +/-). The specific growth arrest mediated by RpL19 +/is released upon abrogation of Xrp1 function ( Fig. 4D; sav -/-, RpL19 +/-, Xrp1 -/-), indicating that the protective function of RPGs haploinsufficiency can also operate within tumorous cells. Taken together these results implicate RPG haploinsufficiency as a potent tumor suppressor mechanism. In addition to the evidence that loss of Xrp1 function works as a suppressor of cell competition-driven elimination of both RpL19 +/and RpL14 +/loser cells (Fig. 2 C 4E) further supports the notion that haploinsufficiency at these loci can ensure a prompt response to genomic instability in order to prevent the initiation of tumorigenesis. Haploinsufficiency is the cornerstone of this mechanism since cell elimination is not triggered upon the loss of one functional copy of RpL3, a non-hRPG (Fig. 4F).

Conclusions
Here we have identified a CCAAT/Enhancer Binding Protein (C/EBP), named Xrp1, as an essential component for the elimination of cells with a reduced copy number of ribosomal protein genes when intermingled with wild type cells. We propose that the imbalanced production of ribosomal proteins triggers a C/EBP dependent transcriptional program that orchestrates the elimination of cells at the onset of genetic instability. Key to this mechanism is the observed haploinsufficiency at RPG loci that translate one-to-one a genetic imbalance into a protein imbalance. This resulting physiological readout at the level of ribosome biogenesis triggers a fail-safe mechanism that leads to the elimination of the impaired cell.
Mutagenesis studies in many diploid organisms indicate that the vast majority of genes are haplosufficient (Wilkie, 1994). Diploidy has been proposed to be evolutionarily selected for as a protective mechanism against the deleterious occurrence of somatic mutations (Orr, 1995). In this view, haploinsufficiency may be considered as an evolutionary accident. In D. melanogaster, haploinsufficiency is rare and the vast majority of haploinsufficient genes are RPGs (Cook et al., 2012). The haploinsufficiency at these loci is often attributed to the high cellular demand in RPs (Marygold et al., 2007). However, this is not an inherent feature of these genes since there are nine RPGs that are not associated with a Minute phenotype (RpSA,RpS2,RpS12,RpL3,RpL23,RpL26,RpL29,RpL30,RpL41) and for which no functional compensation is possible by paralogous genes (Marygold et al., 2007). The benefits emerging from RPG haploinsufficiency (Fig. 4F) appear to outweigh the costs (Orr, 1995) of maintaining it, as it provides a simple, yet effective, mechanism to protect the organism from the emergence of potentially deleterious cells. The basic elements of this mechanism (i.e. hRPGs (Barna et al., 2008)

Drosophila strains and cultures
Flies were grown on a standard cornmeal medium at 25°C unless otherwise specified.

Cloning of transgenes and transgenesis
The RpL19 3.08 kbp genomic rescue (2R:24967017..24970096 Dmel_r6.08) was amplified from a genomic DNA template, sequence confirmed, cloned within the NotI restriction site of the pUAST.attB and inserted into the attP landing site ZH-attP-86Fb (3R tester line) and ZH-attP-68E (3L tester line) (Bischof et al., 2007 (Bischof et al., 2013) and inserted into the attP landing site ZH-attP-68E. The Xrp1 mutated genomic rescue was generated by inserting 5bp (C> GATCCC at 3R:18925226 Dmel_r6.08) at the beginning of the second coding exon in the wild-type genomic fragment, which shifts the frame of all Xrp1 isoforms. Transgenesis was performed according to standard germ-line transformation procedures.

RPG loser clone induction and scoring
RpL19 +/loser clones in vivo screen: y, w, P{hs-FLP}; M{3xP3-RFP.attP}ZH-36B; P{FRT}82B mutagenized males were crossed to y, w, P{UAS-mCD8::GFP.L}LL4, M{RpL19 genomic}ZH-86Fb/ SM5a-TM6B tester virgin females. Parents were allowed to lay eggs for 24 hours and RpL19 +/loser clones were heat-shock induced for 30 minutes at 37°C, 24-48 hours after egg deposition. Progeny were screened at the end of the third instar larval stage when larvae stop feeding and move away from the food. No water was added nor was heat-shock applied to force the remaining larvae out of the food as it is routinely done. Special attention was given to the final larval density in the tubes since we noticed that it negatively influences loser clone elimination. In our hands optimal density for cell competition is achieved when the food is neither dry nor soggy; this proper balance is achieved when late third instar larvae climb up to 2/3 of the tube height without reaching the tube's cotton plug.
Consequently only such tubes were screened for the persistence of RpL19 +/-GFP positive clones through the larval cuticle of living larvae.
RpL19 +/loser clones for dissections: males of the appropriated genotype were crossed to the "3R" or "3L" tester virgin females. "3R" tester virgin females: y, w, loser clones were heat-shock induced for 15 minutes at 37°C, 44-52 hours after egg deposition. Yeast was added to the tubes 24 hours after the heat-shock was applied.
Progeny were screened at the end of the third instar larval stage when larvae stop feeding and move away from the food. No water was added nor was heat-shock applied to force the remaining larvae out of the food.

Mutagenesis and screen
EMS mutagenesis screens were performed according to standard procedure (Bökel, 2008). y, w, P{hs-FLP}; M{3xP3-RFP.attP}ZH-36B; P{FRT}82B starter line was first isogenized for the 3R cell competition screen. Isogenized males were fed with a 25 mM EMS, 1% sucrose solution and crossed to tester virgin females. RpL19 +/clones were induced in the resulting progeny. A total of 20,000 F1 larvae were screened for the persistence of RpL19 +/-GFP positive clones at the end of the third instar larval stage. 182 larvae showed persistence of GFP clones clearly above background noise.
125 of them gave rise to fertile adults and were further rescreened. 12 heritable suppressors were doubly balanced. For the Xrp1 "coding sequence directed mutagenesis" y, w; +; P{GSV6}Xrp1 GS18143 /TM3,Sb males were fed with a 50 mM EMS, 1% sucrose solution and crossed to tester virgin females y, w, P{ey-FLP}; P{Act>y+>GAL4-w}; M{3xP3-RFP.attP}ZH-86Fb. 10,000 F1 genomes were screened and 8 heritable suppressors were retrieved and balanced. A mutation in the Xrp1 coding region was identified in 5 of them. After the causative mutation was identified the upstream P{GSV6}Xrp1 GS18143 was removed using P element transposase and precise excision events were selected (direct sequencing of PCR amplicons) and recombined onto a P{FRT}82B chromosome for clonal analysis.
RpL19 knock-out was generated by mobilizing the P element P{lacW}RpL19 k03704 , imprecise excisions were selected based on the presence of the characteristic Minute bristle phenotype and the absence of the white + marker. The RpL19 IE-C5 1.09 kbp deletion (2R:24968426..24969517 Dmel_r6.08) was selected and characterized using direct sequencing of PCR amplicons. This specific excision removes all of RpL19's coding sequence and leaves neighboring genes unaffected.

Mapping the mutations
We initially mapped cell competition suppressors through meiotic recombinations coupled with DHPLC (Denaturing High-Performance Liquid Chromatography, Eliane Escher) for PCR amplicon analysis. The interval containing the suppressors Xrp1 08 and Xrp1 29 was narrowed down to a 106.5 Kb interval (3R:18872668..18979166 Dmel_r6.08). Sanger sequencing of the coding regions in this interval did not reveal the presence of any mutation. We then performed whole-genome sequencing on

Data access
ChIP-seq data and whole-genome resequencing data from this study will be submitted to the NCBI Gene Expression Omnibus data repository.

Antibodies
Inmunostainings on Drosophila wing discs were performed according to standard protocol. The following antibodies were used: rabbit anti-Cleaved-Caspase-3  (Letunic and Bork, 2007).

UV-C treatment
L1 larvae (yw and Xrp1 61 ) were exposed to different intensity of UV-C (254 nm) on apple-agar plates and then transferred to standard cornmeal medium. Percentage of eclosed animals was scored as well as the percentage of adults exhibiting obvious phenotypes (bristle number/morphology, eye/notum/wing malformations and melanotic masses).

RPGs cancer-deletion profiling
The RPG cancer-deletion analyses was performed for each of the 79 human RPGs using the TUMORSCAPE query function based on the GISTIC analysis (Mermel et al., 2010). Deletions for RPGs were significantly detected within the 14 cancer subtypes that regroups more than 80% (2549/3131) of the cancer samples on which the GISTIC analysis is based. This analysis identifies genes that are significantly deleted across the datasets and may therefore underestimate the extend of RPGs deletion due to intra-tumour heterogeneity. Cancer subtypes with GISTIC-based RPG      Supplementary Table S3.
(B) Protective function of Xrp1. Log 2 ratios of the viability and the mutagenicity of Xrp1 null animals relative to WT exposed to increasing intensity of UV-C (254 nm).
As the dose increases, the detrimental effects of UV-C exposure in Xrp1 61 animals increases. Black bars represent "viability", black dashed bars represent "mutant phenotype", the baseline is set at 1, below the baseline means reduction and above the baseline means increase. Data breakdown is presented in Table S2.

Xrp1 intronic mutations:
The intronic mutations we identified in our screen are all substitutions of single nucleotides. These nucleotides are conserved within the Drosophila genus and inspection of the alignment revealed an embedment of these nucleotides in conserved putative motifs (supplementary Figure S1). Of particular interest are the polypyrimidine motifs containing the nucleotide mutated in Xrp1 20 and Xrp1 08 . These motifs flank the alternative first exon and are potential splice regulators. The CTCTCT motif present near the 5' splice site of Xrp1 has been identified as a putative intronic splicing enhancer (ISE) in Drosophila (Brooks et al., 2011). This motif is one of the two ISEs that is conserved in vertebrates and is predicted to be a binding site for the polypyrimidine-tract binding protein (PTB) splicing regulator (Brooks et al., 2011). The presence of these motifs prompted us to investigate the consequences of the Xrp1 08 allele on exonic junctions. The most prominent effects of this allele are a strong and consistent reduction in the expression of two similar Xrp1 transcripts, RC and RE (supplementary Figure S2). These transcripts differ only in the composition of their 5'UTRs. They share the same transcriptional start site and contain the same protein-coding sequences that code for the short isoform of this gene (supplementary Figure S1).

Figure S1 Alignment of Drosophila Xrp1 sequences
The mutations retrieved from the EMS screen are highlighted in red. All these alleles are single base pair substitutions of conserved nucleotides. The nucleotide substitutions are indicated on top of the Drosophila melanogaster sequence (Dmel).
Conservation extends beyond these nucleotides and seems to affect intronic motifs.
These motifs are capitalized and colored in red. The allele Xrp1 20 disrupts the repetition of the conserved hexanucleotide pyrimidine motif TCTDTB (D stands for A, G or T and B stands for C, G or T). The allele Xrp1 08 disrupts the conserved putative intronic splice enhancer (ISE) CTCTCT and Xrp1 29 disrupts a conserved GATA motif

Figure S2 Intronic mutations reduce Xrp1 transcripts E and C abundance
Xrp1 transcripts are designated with a single letter from A to G on the left side. Xrp1 08 leads to a constitutive downregulation of the transcripts E and C, 9-folds and 6-folds respectively. (Left) Primers used for the qRT-PCR are numbered from 1 to 8. The exonic junctions amplified from each of these primer pairs is color-coded (1-2: brown, 3-4: orange, 1-6: red, 3-6: red, 5-6: purple and 7-8: magenta). The transcripts containing these exonic junctions are indicated in each colored box. (Right) qRT-PCR results are normalized to wild type. Error bars indicate standard deviations. Note that E and C transcript levels (red bars) diminish as the number of Xrp1 mutant copies increases.

Table S1 Complementation matrix of the EMS induced suppressors
Green color indicates complementation, red indicates non-complementation, ND