p53 pulse modulation differentially regulates target gene promoters to regulate cell fate decisions

• DOI: 10.15252/msb.20188685
• Publication Date: 2019-09-26T08:41:33Z
• Authors: Marie D Harton; Woo Seuk Koh; Amie D Bunker; Abhyudai Singh; Eric Batchelor
• Keywords: Cell fate determination

Article26 September 2019Open Access Source DataTransparent process p53 pulse modulation differentially regulates target gene promoters to regulate cell fate decisions Marie D Harton Laboratory of Cell Biology, Center for Cancer Research, National Institute, Institutes Health, Bethesda, MD, USA This article has been contributed by US Government employees and their work is in the public domain USA. Search more papers this author Woo Seuk Koh Amie Bunker Abhyudai Singh Department Electrical Computer Engineering, Biomedical Mathematical Sciences, Bioinformatics Computational University Delaware, Newark, DE, Eric Batchelor Corresponding Author [email protected] orcid.org/0000-0003-3870-5615 Information Harton1, Koh1, Bunker1, Singh2 *,1 1Laboratory 2Department *Corresponding author. Tel: [phone]; E-mail: Molecular Systems Biology (2019)15:e8685https://doi.org/10.15252/msb.20188685 PDFDownload PDF text main figures. Peer ReviewDownload a summary editorial decision including letters, reviewer comments responses feedback. ToolsAdd favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract The tumor suppressor distinct cellular stresses. Although different stresses generate dynamics, mechanisms which cells decode dynamics genes are not well understood. Here, we determined individual how canonical vary responsiveness features dynamics. Employing chemical perturbation approach, independently modulated amplitude, duration, or frequency, then monitored levels promoter activation cells. We identified signal processing features—thresholding response amplitude modulation, refractory period duration dynamic filtering frequency modulation. showed that only affect activation, they also regulation downstream functions. Our study shows can responses, providing insight into perturbing be used fates. Synopsis Small-molecule oscillations generates responses. Oscillation characteristics control thresholds, period, filtering. controlled via microfluidic delivery small molecule inhibitor MDM2. pulses determines thresholds genes. A re-initiation established pulses. dynamically filtered promoters. Introduction Pulsatile have growing number important transduction pathways. In human cells, transcription factors NF-κB, as extracellular signal-regulated kinase (ERK), recent examples signaling molecules shown oscillate protein expression level activity (Lahav et al, 2004; Nelson 2011; Albeck 2013; Lee 2014; Kellogg Tay, 2015; Ryu 2015). relatively simple organism Saccharomyces cerevisiae, screen 10 with pulsatile (Dalal 2014). encode information about specific stimulus molecules, was observed change stimuli (Batchelor 2011). While regulatory responsible stimulus-specific shaping many systems Hao O'Shea, 2011), it remains challenge identify diverse output Given importance regulating stress decisions, developing methods precisely key may provide novel pharmacological interventions. mutated majority cancers thus robust cancer therapeutics. responds various signals subsequently several pathways, cycle arrest, apoptosis, senescence 2009; Vousden Prives, Purvis Lahav, 2013). Single-cell studies demonstrated undergoes complex, stimulus-dependent DNA double-strand breaks, increase series discrete fixed average frequency; ultraviolet radiation, single dose-dependent Alteration through inhibition E3 ubiquitin ligase MDM2 directly impacts patterns (Porter 2016) p53-mediated (Purvis 2012). altering 2012), clonal population exhibit same stimulus. Recent population-level analysis that, on average, leverage differences mRNA half-lives relative induce 2016; Hafner 2017); however, unclear produce variance outcomes within population. Previous from pathways elucidated molecular decoding temporal patterns, (Hansen 2013) stability (Hao Baltimore, Porter Zambrano 2017). For yeast factor Msn2, modes terms both variability threshold encoded protein's mammalian variation targets regulators, such p53, likely mechanism increased study, quantified changes two canonically regulated independent manipulation living (Fig 1A). found even when comparable driven dynamical input, displaying range sensitivities toward distinguish promoters, thresholding filtering, sets response. each affected its products functions arrest regulation. demonstrates respond differently suggesting an additional beyond instability ultimately impact decisions. Figure 1. system modulate monitor A. Schematic breaks. decoded functions, mediating B. Nutlin-3, prevents degradation Cellular exposure media without Nutlin-3 potentially six drug dosing regimens duration. C. Clonal lines expressing p53-Venus were engineered ("input") mCherry ("output") treatment. D. Representative phase contrast yellow fluorescent (indicating levels) images at indicated time points exposed "natural dynamics" regimen. E, F. traces (gray; E) mean (red; Heat map (F) alternative representation all (E). N = 51 Download figure PowerPoint Results strategy simultaneously track To determine parallel damage developed method absence extrinsic damage. employed based binds inhibits leading stabilization 1B). As transcriptionally upregulates MDM2, our disrupts negative feedback loop enable direct precise interrogate effects activation. rapidly enter removed addition growth medium respectively, ensuring wide manipulating 2012; 2016). alter fluorescence-based reporters. levels, well-characterized MCF7 breast carcinoma line expresses fusion to, but lower concentration than, endogenously expressed wild-type (Loewer 2010). quantify include transcriptional reporter construct reduced controlling red tagged nuclear localization sequence PEST 1C). focused genes, CDKN1A, could upregulated alone EV1A). These elements, p21 associated respectively. transcripts 2.66 2.79 h respectively 1A; 2016), opportunity promoter-specific previously due transcript decay rates 2016, Single copies reporters stably integrated genome parental line, verified qPCR, reporter. Click here expand figure. EV1. Characterization endogenous CDKN1A treatment μM Nutlin-3. Error bars SEM (n 2). Western blot untreated treated 400 ng/μl neocarzinostatin (NCS), 5 15 3 h. C–L. low-amplitude (C, D), high-amplitude (E, F), high-frequency (G, H), low-frequency/short-duration (I, J), long-duration (K, L) regimens. (gray) (red) G, I, K) maps (D, F, H, J, representations least 45 per condition. M. over 24 Line mean, box SD, bar 95% confidence interval (N cells). data available online live device exchange first produced natural containing followed 2.5 24-h frame, generated induction break-inducing agent EV1B; Using time-lapse fluorescence microscopy, oscillating 5.5-h maximum achieved breaks 1B D–F). concentrations Fig EV1C–F). modified timing washout EV1G–L). induced upon (Movies EV1–EV6). between EV1M), indicating no discernible alteration functioning p53-MDM2 results suitable tracking systematically alters either microscopy 2A–G). percentage twofold basal ("responding cells") regimen EV2A–C). Traces clustered k-means clustering generating clusters EV3), broad profiles across perturbation. All non-responding considered further analysis. 2. contrast, levels), B–G. "responding" (light gray) "not responding" (dark (B), (C), (D), (E), (F), (G) trace responding blue, show single-cell below course plot. EV2. Percentage modulations A–C. showing (A), (C) Data EV2A 5H. EV3. K-means reveals A–L. Partition (pink) (purple) promoter-mCherry four (A, B), low-frequency natural-frequency modulations. delineates complex sought formalize experimental computational model aid identifying biochemical parameters governing modeled Hill function models recapitulate other strength simplified determining altered amplitude. damage, high (coefficient ~ 70%) (Geva-Zatorsky 2006; Toettcher hypothesized thresholds. test hypothesis whether manipulate activate manner, analyzed variations full set perturbations (Figs 1 At degree perturbations, resulting maximal varied (e.g., EV4). EV4. Individual highly variable expressionDose–response curves representing rate (pink dots) total representative 15-h there deterministic accumulation, i.e., defining concentration, constructed population-averaged levels. response, resulted highest EV2B). Fitting values 2) monotonically increasing, nonlinear dependence production 3). dose–response (with fitting kmax 135, coefficient 7.5, K 407) having required higher compared 40, 7, 490; suggest contributing diversification 3. thresholdsDose–response (purple Values averaged Solid black represent best fit Dashed indicate half-maximal promoter. establishes much less characteristics—duration frequency—compared 2006). observation suggests greater selective pressure maintain properties expression. therefore might significantly short-duration (3 h) (8 EV1I–L, 2F G), characterizing (timing), magnitude (magnitude), initial accumulation (rate) 4A). particular increasing (2.6-fold) (1.6-fold) decreasing (1.4-fold) 4B–G). sensitivity (i.e., low, natural, frequencies) times corresponding equivalent cumulative EV5A). had elevated input inputs, than threefold EV5B). EV5C). suggested particularly sensitive 4. Responding characterized metrics (timing, magnitude, rate). Effects (B, E), G) (B–D) (E–G) short green) long durations. condition, interval. H. Mean MDM2-YFP (5.5 second (11 natural- 40 information: *P < 0.05 **P 0.01, two-sample t-test. EV5. Target comparing duration- frequency-modulated inputs. Rates calculated peak (tpeak) equal plus 5.5 (tX + τ), where τ period. B, Activation (B) next function, designed pulses, most greatly failed initiate altogether EV1K L). result prolonged