Genomics of cellular proliferation in periodic environmental fluctuations

• DOI: 10.15252/msb.20177823
• Publication Date: 2018-03-06T01:11:04Z
• Authors: Jérôme Salignon; Magali Richard; Etienne Fulcrand; Hélène Duplus‐Bottin; Gaël Yvert
• Keywords: 0301 basic medicine; Medicine (General); Salinity; Saccharomyces cerevisiae Proteins; QH301-705.5; [SDV]Life Sciences [q-bio]; selection; Saccharomyces cerevisiae; yeast; stress; 03 medical and health sciences; R5-920; Methionine; Stress, Physiological; Gene Expression Regulation, Fungal; Biology (General); Selection, Genetic; Cell Proliferation; [SDV.GEN]Life Sciences [q-bio]/Genetics; Models, Genetic; fluctuating environment; Gene Expression Profiling; Articles; fitness; Mutation; Genetic Fitness; Gene Deletion

Article5 March 2018Open Access Transparent process Genomics of cellular proliferation in periodic environmental fluctuations Jérôme Salignon Laboratory Biology and Modeling the Cell, Ecole Normale Supérieure de Lyon, CNRS, Université Claude Bernard France Search for more papers by this author Magali Richard orcid.org/0000-0003-3165-3218 Etienne Fulcrand Hélène Duplus-Bottin Gaël Yvert Corresponding Author [email protected] orcid.org/0000-0003-1955-4786 Information Salignon1,‡, Richard1,‡, Fulcrand1, Duplus-Bottin1 *,1 1Laboratory ‡These authors contributed equally to work *Corresponding author. Tel: [phone]; E-mail: Molecular Systems (2018)14:e7823https://doi.org/10.15252/msb.20177823 PDFDownload PDF article text main figures. Peer ReviewDownload a summary editorial decision including letters, reviewer comments responses feedback. ToolsAdd favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract Living systems control cell growth dynamically processing information from their environment. Although single change have been intensively studied, little is known about how cells react fluctuating conditions. Here, we address question at genomic scale measuring relative rate (fitness) 3,568 yeast gene deletion mutants out-of-equilibrium conditions: oscillations between two In salt stress, fitness its genetic variance largely depended on oscillating period. Surprisingly, dozens displayed pronounced hyperproliferation under short stress periods, revealing unexpected controllers fast dynamics. We validated implication high-affinity cAMP phosphodiesterase regulator protein translocation mitochondria group. Periodic extracellular methionine, factor unrelated salinity, also altered but lesser extent different genes. The results illustrate natural selection acts mutations dynamic environment, highlighting unsuspected vulnerabilities molecular processes that are conserved across all eukaryotes. Synopsis Genome-scale analysis reveals can operate environments. environments, deviation time-average common, numerous deletions confer periods. Two screens performed inhomogeneity (deviation expectation) methionine availability. Fitness inhomogeneous hundreds genes, these genes differ surveys. Fast salinity provide profound advantage subset deletions. Gene same pathway or with similar alterations steady conditions response Introduction Cells modify themselves variation Interactions internal dynamics intracellular regulations external environment determine whether divides, differentiates, cooperates other dies. For some systems, usually model organisms, molecules involved signal transduction adaptation known. How they act motion, however, unclear, it difficult predict which ones may be essential upon certain frequencies fluctuations. addition, since most were conducted after occurrence, key missed. life therefore focus intense research, interplay proliferative remains poorly characterized. drives evolutionary selection, properties environments unknown. experimental data exist (Stomp et al, 2008; Bleuven Landry, 2016), scarce has mostly studied theoretical frameworks (Kussell Leibler, 2005; Cvijović 2015; Sæther Engen, Svardal 2015). Repeated stimulations consequences stimulus. First, small delay following stimulus become highly significant when cumulated over multiple stimuli. Second, given time depend past "remember", memory sometimes transmitted daughter (Hilker 2016). These features well illustrated study Razinkov who manipulated stability GAL1 mRNA transcripts participated short-term "memory" galactose exposures: resulted was negligible one galactose-to-glucose changes (Razinkov 2013). Other memorization effects observed bacteria during repeated lactose glucose transitions, due both conferred persistent expression long-term (Lambert Kussell, 2014). high concentrations best-studied mechanisms adaptation. When increases abruptly, size immediately reduces triggers large translation programme (Uesono Toh-e, 2002; Warringer 2010) turnover mRNAs (Miller 2011) re-defined, calcium accumulates cytosol activates calcineurin (Ariño 2010), osmolarity sensors activate high-osmolarity glycerol MAPK (Hohmann, 2009; Ariño intracellularly as harmless compensatory solute 2009), membrane transporters extrude excess ions 2010). Via widespread adaptation, participate transition salt. What happens case less clear, investigated adaptive response. example, transitions 0 0.4 M NaCl showed activation efficient transient each except range ~8-min where sustained severely hampered (Mitchell tolerance contribute specific regimes Yeast respond modulating import biosynthesis. This relies pathways stress. It described (Thomas Surdin-Kerjan, 1997), not context sought systematically search Identifying such done applying mutant periodically testing effect mutation averaged time. words, does (proliferation wild type) match average alternating condition? problem temporal heterogeneity equivalent homogenization commonly encountered physics spatial heterogeneity, microscopic heterogeneities materials macroscopic stiffness conductivity (Hassani Hinton, 1998). A homogeneous (averaged time) implies (i) occurs rapidly compared frequency changes, (ii) affect lag phase (iii) mutated relevant mechanisms. contrast, indicative role study, present responding reveal identify unexpectedly major periods Results Genomic profiling rates measured experimentally contribution thousands an oscillated used collection ~5,000 non-essential individually deleted (Giaever 2002). Since every barcoded synthetic DNA tag inserted genome, abundance pooled cultures estimated parallel sequencing barcodes (BAR-Seq) (Smith Robinson set up automated robotic platform culture library serial dilutions. Every 3 h (average division time), populations transferred standard medium containing (S) (N) 0.2 NaCl. culturing either maintained N, S exposed N 6, 12, 18, 24 42 (Fig 1A). regime run quadruplicates account biological technical variability. duration experiment days, sampled times 0, 24, 48 sequencing. After normalization filtering, examined Figure 1. Experimental design. Populations strains cultured media (no salt), (salt) various Allele determined BAR-Seq compute mutant. Time course population, shown six mutants. Relative corresponds median log2(y/y0) values ± SD (n = replicate cultures, condition day 3: n 3), y normalized number reads, y0 0. Conditions: (blue), (yellow), 6-h (NS6, hatching). Generalized linear models (predicted value SE) fitted (B), coloured condition: (blue); (yellow); NS6 predicted null (grey) complete (red). ***P < 10−8. n.s., non-significant, based GLM (see 4). (w) computed (B). Bars, mean SEM, (S, NS6) according condition. Grey dashed line: expected additivity (geometric weighted spent medium). Scatterplot showing (y-axis, regime) (x-axis, geometric S). Deviation diagonal reflects inhomogeneity. Red dots: 456 (FDR 0.0001, see Correlation estimates (w). Each dot (N, NS6), (x-axis) individual assays (one co-cultured WT cells, y-axis). Whole data: 52 R, Pearson coefficient; grey line, x; red regression. Validation counting. One graph shows GFP-tagged wild-type flow cytometry. Median cultures). Download figure PowerPoint Protective diverse contributions differed way controlled regimes. Differences visible among inhibiting promoting NBP2 negative HOG (Mapes Ota, 2004) MOT3 transcriptional having functions osmotic (Montañés 2011, 3; Martínez-Montañés As Fig 1B, improved (condition regime, mot3Δ/Δ if exposures beneficial had no positive effect. benefit clearly nbp2Δ/Δ cells. apparent protective Rim101 alkaline required proper accumulation Ena1p transporter Na+ extrusion (Marqués Eight covered our experiment. Not surprisingly, regulators pathway, decreased increased 1B Appendix S1). consistent need functional cost maintaining required. However, (Appendix RIM21, DFG16 RIM9 code units transmembrane sensing complex (Obara 2012), rim21Δ/Δ dfg16Δ/Δ rim9Δ/Δ Similarly, Rim8 Rim20 mediate Rim101p repressor (Xu Mitchell, 2001; Herrador 2010); rim8Δ/Δ rim101Δ/Δ whereas rim20Δ/Δ did not. only example displaying differences. lacking HST1- HST3 NAD(+)-dependent histone deacetylase (Brachmann 1995) grew S, hst1Δ/Δ tolerated better than hst3Δ/Δ 1B). Thus, Widespread then asked mutants, matched S. tested statistical significance quantified expectation. inference, exploited full count data, replicated populations, fitting generalized included non-additive term associated obtained discussed above 1C. Overall, period, many ~2,000 because 2,497 false discovery (FDR) Table At stringent FDR listed significant. quantification, Qian al (2012) 1D) plotted function 1E). nbp2Δ/Δ, often good agreement. Highlighting revealed surprising trend: majority increase (expected > 1), high. annotations corresponding higher-than-expected enriched members cAMP/PKA S2), estimate parallel, important limitations: estimation indirect distinguished possible interactions pool. validate observations competition assays. strain, counted cytometry (Qian 2012; Duveau previous reports Venkataram 2016; 1F, unambiguously several 1G). Impact If average) attributable dynamics, should oscillations. Our four larger 6 h. ratio expectation deviates Plotting distribution oscillation that, expected, period 2A). closely three highest extreme so 2B). 2. Proliferative depends Violin plots indicated Traces labels, Top, 0.0001. cin5Δ/Δ, srf1Δ/Δ yor029wΔ/Δ pool regimes, BAR-Seq. 3). population Bars: 95% CI bootstrap intervals. result extremely fit short-period suggested differences To case, Distinguishing non-genetic presence replicates design grown agrees (Martin Lenormand, 2006). Remarkably, even fast-oscillating slow-oscillating 2C). exert additional selective pressures level whole vs. related conditions, deletions, was. Inhomogeneity 3A–D, dots). Interestingly, particular very together distinct advantageous blue Annotations osmosensing although category display behaviour. defective components osmostress outproliferate likely do trigger costly unnecessary process. 3. respect A. correlated fitness. Red, (> 1.05). Blue, 1.045) moderate (< 1.05), those annotated relation light name. Black dots, examples osmosensing/response, B. difference Colours, panel (A). C, D. Observed (panel C) D). identity. Some (w 1) penalized 1). phenomenon special × interaction called antagonistic pleiotropy (AP) 2012). anticipate will impact alternates favourable unfavourable especially necessarily homogenized Using test 4), found statistically AP 0.01, S3 S3). coded subunits chromatin-modifying Set1/COMPASS S2 S4). direction 4A–C). 33 (vhr1Δ/Δ rim21Δ/Δ), changed visualize periodicity dependence clustered them inhomogeneities 4B C). highlighted five behaviours: could strongly favour (e.g. cin5Δ/Δ) mainly oca1Δ/Δ), mildly rim101Δ/Δ) decrease csf1Δ/Δ) depending (vhr1Δ/Δ). generally asymmetric dependency 4. Oscillations classified ("pos"itive advantageous, "always" oscillations, "neg"ative disadvantageous). Hierarchical clustering representative clusters (others) cultures. Salinity heighten made observation exceed fall below 2B), behaviour "transgressivity" hereafter. By using available values, detected 55 significantly higher maximum 23 lower minimum 5A, 0.03, Importantly, transgressivity studying assays, pde2Δ/Δ, tom7Δ/Δ, trm1Δ/Δ yjl135wΔ/Δ 5B–E). implications spectrum hyperproliferative clones experienced repetitive Strikingly, processes: (pde2Δ/Δ), into (tom7Δ/Δ), autophagy (atg15Δ/Δ), tRNA modification (trm1Δ/Δ), phosphatidylcholine hydrolysis (srf1Δ/Δ) signalling (ssk1Δ/Δ, ssk2Δ/Δ); previously 5. Extreme emerging (NS6) (y-axis). Violet, 78 0.03). B–E. (BAR-Seq, left, 1B) (flow cytometry, right, (hatching). F–G. Complementation Diploid homozygous pde2 tom7 (strains GY1821 GY1804, respectively) complemented integration HO locus GY1929 GY1921, respectively). Strains (strain GY1961), S′ (0.4 NaCl; orange) SEM Tom7p necessary limit mentioned above, impairing cA