Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 May 9;15(5):e2000644.
doi: 10.1371/journal.pbio.2000644. eCollection 2017 May.

Phenotypic Heterogeneity Promotes Adaptive Evolution

Affiliations
Free PMC article

Phenotypic Heterogeneity Promotes Adaptive Evolution

Zoltán Bódi et al. PLoS Biol. .
Free PMC article

Erratum in

  • Correction: Phenotypic heterogeneity promotes adaptive evolution.
    Bódi Z, Farkas Z, Nevozhay D, Kalapis D, Lázár V, Csörgő B, Nyerges Á, Szamecz B, Fekete G, Papp B, Araújo H, Oliveira JL, Moura G, Santos MAS, Székely T Jr, Balázsi G, Pál C. Bódi Z, et al. PLoS Biol. 2017 Jun 20;15(6):e1002607. doi: 10.1371/journal.pbio.1002607. eCollection 2017 Jun. PLoS Biol. 2017. PMID: 28632738 Free PMC article.

Abstract

Genetically identical cells frequently display substantial heterogeneity in gene expression, cellular morphology and physiology. It has been suggested that by rapidly generating a subpopulation with novel phenotypic traits, phenotypic heterogeneity (or plasticity) accelerates the rate of adaptive evolution in populations facing extreme environmental challenges. This issue is important as cell-to-cell phenotypic heterogeneity may initiate key steps in microbial evolution of drug resistance and cancer progression. Here, we study how stochastic transitions between cellular states influence evolutionary adaptation to a stressful environment in yeast Saccharomyces cerevisiae. We developed inducible synthetic gene circuits that generate varying degrees of expression stochasticity of an antifungal resistance gene. We initiated laboratory evolutionary experiments with genotypes carrying different versions of the genetic circuit by exposing the corresponding populations to gradually increasing antifungal stress. Phenotypic heterogeneity altered the evolutionary dynamics by transforming the adaptive landscape that relates genotype to fitness. Specifically, it enhanced the adaptive value of beneficial mutations through synergism between cell-to-cell variability and genetic variation. Our work demonstrates that phenotypic heterogeneity is an evolving trait when populations face a chronic selection pressure. It shapes evolutionary trajectories at the genomic level and facilitates evolutionary rescue from a deteriorating environmental stress.

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1. Basic properties of the synthetic genetic circuits.
(A) Elements of the synthetic constructs. The synthetic constructs comprise the reverse-tetR trans-activator (rtTA-MF) gene and the C-terminally GFP-tagged PDR5 reporter gene (PDR5-GFP). The PDR5-GFP reporter gene is controlled by a synthetic tet-inducible tetreg promoter (Ptetreg2), activated by doxycycline (dox). Doxycycline forms a complex with the internally produced reverse-tetR transactivator protein (rtTAp). The complex activates the expression of the reporter gene through binding to the tetO2 sites of Ptetreg2. The two circuits differ only in the promoter of the rtTA-MF gene. In the case of the positive feedback (PF) circuit, the promoters of the rtTA-MF and the PDR5-GFP reporter gene are the same (Ptetreg2), generating a positive regulatory loop. The no positive feedback (noPF) circuit has no such loop, as the rtTA-MF is driven by a constitutive promoter (glyceraldehyde-3-phosphate dehydrogenase promoter [PGPD], glyceraldehyde-3-phosphate dehydrogenase [TDH3]). Transcription of the rtTA-MF gene and the PDR5-GFP reporter gene was terminated by using the terminator sequence of alcohol dehydrogenase 1 gene (tADH1) and cytochrome c (tCYC1), respectively. (B) Monitoring the fluorescence level of the reporter gene by flow cytometry. Fluorescence histograms of C-terminally GFP-tagged Pdr5p protein (Pdr5p-GFP) for selected PF and noPF strains, respectively. The noPF population showed low gene expression stochasticity, with unimodal expression distribution, and did not contain cells with extremely low or extremely high fluorescence levels. The difference in the mean target gene expression level was only 5%, and it was actually higher in the noPF-carrying strain, while the coefficient of variation differed by 150% between PF and noPF populations (Mann–Whitney U Test, p < 0.001). Measured fluorescence intensity was normalized to the forward scatter values (both log10-scaled). The inducer levels were 0.015 μg/ml (noPF) and 0.3 μg/ml (PF), respectively. (C) Evaluation of fluconazole minimum inhibitory concentration (MIC) values of noPF and PF strains. The figure shows the mean relative fitness of the ancestor PF (red lines) and noPF strains (blue line) as a function of fluconazole dosage. Absolute fitness was estimated by the increment of the optical density at 600 nm (OD600) after 72 h of growth of each strain at each fluconazole concentration (see Materials and methods). Relative fitness was calculated by normalizing the absolute fitness of each strain at each concentration to the noPF absolute fitness value in drug-free medium (see Materials and methods for details). Error bars indicate 95% confidence interval, based on 32 independent cultures grown and measured for OD600. MIC is defined at a 0.15 cutoff value of the increment of the OD600. (D) Fitness cost of phenotypic heterogeneity. The figure shows the mean relative fitness of the ancestor and the evolved strains in fluconazole-free medium. Evolved strains from the final day of the laboratory evolution experiment (Experiment A) were used for the analysis. Prior to evolution (ancestor), PF fitness is 7% lower than noPF fitness (Mann–Whitney U test, p < 0.001). Laboratory evolved strains (evolved) show fitness deficits in drug free medium, compared to their corresponding ancestors (11% in noPF, Mann–Whitney U test, p < 0.001; 16% in PF strains, Mann–Whitney U test, p < 0.001). Absolute fitness was estimated by the increment of the OD600 after 72 h of growth in drug-free medium (see Materials and methods). Relative fitness was calculated by normalizing the absolute fitness of each strain to the absolute fitness of the noPF ancestor. Error bars indicate 95% confidence interval, based on at least 140 independent cultures grown and measured for OD600. The data underlying Fig 1 can be found in S1 Data.
Fig 2
Fig 2. Impact of phenotypic heterogeneity on adaptive evolution.
Two complementary experiments were used to study the impact of phenotypic heterogeneity on adaptive evolution. Experiment A measured extinction rate of the evolving strains as a function of gradually increasing fluconazole dosage (for further details see S2A Fig), while experiment B aimed to maximize the fluconazole resistance increment during a fixed time period (for further details, see S2B Fig). (A) Outcome of Experiment A. The figure shows the extinction dynamics of independently evolving strains exposed to gradually increasing fluconazole dosages. There was a significant difference in the ratio of surviving strains between no positive feedback (noPF) and positive feedback (PF) populations, respectively (paired t-test, p < 0.05). 32% of the noPF populations failed to adapt to the final employed fluconazole dosage (224 μg/ml), while the same figure was as low as 12% for PF (chi-squared test, p < 0.05). Evolved strains from the final day of Experiment A were used for further genomic and functional analyses. (B) Outcome of Experiment B. Ten independent populations of noPF (blue circles), PF (red circles), noPF mutator (blue triangle), and PF mutator (red triangle) strains were each allowed to evolve in parallel using an established automated evolution protocol (see Materials and methods). The figure shows the distribution of the level of resistance (i.e., the highest fluconazole dosage the strain could grow) as a function of time (transferring steps). The median line of the ten independent populations in each strain are color coded as follows: blue continuous line denotes noPF, red continuous line denotes PF, blue dashed line denotes noPF mutator, and red dashed line denotes PF mutator lines. At the end of Experiment B (i.e., at the last four transferring steps), there was a significant difference in the resistance level between PF nonmutator and noPF nonmutator strains (t-test, p < 0.05) and also between PF mutator and noPF mutator strains (t-test, p < 0.05), respectively. (C) Properties of C-terminally GFP-tagged PDR5 gene (PDR5-GFP) expression after evolution. The figure shows the mean and coefficient of variation (CV) of PDR5-GFP expression distribution in the noPF and PF strains. The values were normalized to the corresponding ancestors, respectively. Mean expression level did not change significantly in the evolved PF strains (Mann–Whitney U test, p = not significant), while it increased by an average 17% in the noPF strains (Mann–Whitney U test, p < 0.001). The CV showed an average 54% increment in the evolved noPF strains (Mann–Whitney U test, p < 0.001), while the same increment was only 7% in the evolved PF strains (Mann–Whitney U test, p < 0.05). Error bars indicate 95% confidence interval. AN, ancestor; EV, evolved. (D) C-terminally GFP-tagged Pdr5p protein (Pdr5p-GFP) fluorescence distribution after laboratory evolution. The fluorescence histograms show the Pdr5p-GFP distribution of one representative laboratory evolved PF (red) and noPF (blue) strain, respectively, and the corresponding ancestors (black). For the characteristics of the distribution in all evolved strains, see S6 Fig. The data underlying Fig 2 can be found in S1 Data.
Fig 3
Fig 3. Outcome of the promoter-swap experiment.
(A) The promoter-swap experiment. The promoter controlling the reverse-tetR trans-activator (rtTA-MF) gene in three selected evolved positive feedback (PF) strains was swapped for the corresponding promoter present in the ancestor no positive feedback (noPF) strain. As a result of eliminating the positive feedback loop (denoted as a dashed line), the PF evolved strains exhibited a C-terminally GFP-tagged Pdr5p protein (Pdr5p-GFP) fluorescence distribution reminiscent of the noPF strains. (B) Evidence for synergism between adaptive mutations and phenotypic heterogeneity. The figure shows the relative increase in the minimum inhibitory concentration (MIC) of three selected evolved PF strains (EV-1, EV-2 and EV-3), prior to (red full circle) and post (red empty circle) swapping. The median MIC values of the genotypes are normalized to that of the corresponding ancestor strains. After controlling for the level of MIC reduction in the ancestor strains, the promoter-swap procedure resulted in a 2.1–5.5-fold reduction in the resistance level of the evolved PF strains (Mann–Whitney U test, p < 0.05). Four independent cultures per strain were grown and measured for MIC (empty circles). The data underlying Fig 3 can be found in S1 Data.
Fig 4
Fig 4. Phenotypic heterogeneity modulates the effect of mutations on resistance level.
(A) Distribution of C-terminally GFP-tagged Pdr5p protein (Pdr5p-GFP) fluorescence level at three different expression states by flow cytometry. The histograms show the fluorescence distributions of Pdr5p-GFP, as follows: positive feedback (PF) with high phenotypic heterogeneity (HH), no positive feedback (noPF) with adjusted expression level (AE), and PF with high expression level (HE). The inducer levels were set to 0.2 μg/ml doxycycline (AE in noPF strain), 0.3 μg/ml doxycycline (HH in PF strain), and 3 μg/ml doxycycline (HE in PF strain), respectively. (B) Comparison of HH, AE, and HE. The figure shows the minimum inhibitory concentrations (MIC) of the ancestor (AN) and three evolved (EV) PF strains (EV-1, EV-2, and EV-3) in three different expression settings. The settings were HH, AE, and HE. In the ancestor, the AE and HH settings reached the same resistance level, as expected (Mann–Whitney U test, p = not significant), while the HE setting reached higher MIC level than HH (Mann–Whitney U test, p < 0.05). In the evolved strains (EV-1, EV-2, and EV-3), however, the MIC level of the HH setting was somewhat higher than that of AE (Mann–Whitney U test, p < 0.05) or HE (Mann–Whitney U test, p < 0.05). Four independent cultures per strain were grown and measured for MIC (empty circles). The data underlying Fig 4 can be found in S1 Data.
Fig 5
Fig 5. Genomic analysis.
(A) Distribution of mutational events across the evolved strains. Four independently evolved positive feedback (PF) and four no positive feedback (noPF) strains were subjected to whole-genome sequencing, respectively. The figure shows the number and types of detected mutations per strain. There was no significant difference in the number of mutations between noPF and PF strains (Mann–Whitney U test, p = not significant). (B) List of genes with nonsynonymous, synonymous, and intergenic mutations across the evolved strains. As expected from the mode of action of fluconazole, many of the mutated genes are involved in ergosterol biosynthesis (ERG25) and regulation (ROX1) and membrane transport (PDR5, NFT1). (C) Synergism between phenotypic heterogeneity and PDR5 mutation. The figure shows the relative fluconazole minimum inhibitory concentration (MIC) of the PF and noPF strains with a specific PDR5 mutation (His595Asp) inserted. AN, ancestor; AN*, ancestor with a single PDR5 mutation inserted. The median MIC values of the genotypes were normalized to that of the corresponding ancestor strains. Insertion of the mutation resulted in an 85% decline in fluconazole susceptibility when phenotypic heterogeneity was high, but its beneficial effect was reduced otherwise (Mann–Whitney U test, p < 0.05). Four independent cultures per strain were grown and measured for MIC (empty circles). (D) Comparison of high phenotypic heterogeneity (HH), adjusted expression level (AE), and high expression level (HE) in a PDR5 mutant strain. The figure shows the MIC values of AN and AN* in three different expression settings. The expression settings were as follows: HH, AE, and HE. In the ancestor, the AE and HH settings reached the same resistance level, as expected (Mann–Whitney U test, p = not significant), while the HE setting reached higher MIC level than HH (Mann–Whitney U test, p < 0.05). In AN*, however, the MIC level of the HH setting was somewhat higher than that of AE (Mann–Whitney U test, p < 0.05) or HE (Mann–Whitney U test, p < 0.05). Four independent cultures per strain were grown and measured for MIC (empty circles). The data underlying Fig 5 can be found in S1 Data.
Fig 6
Fig 6. Comparison of phenotypic heterogeneity and constitutively high gene expression.
(A) High expression level (HE) of PDR5 imposes a fitness cost. The figure shows the mean relative fitness of four genotypes at different expression states in drug-free medium. The genotypes studied were the ancestor (AN) positive feedback (PF) and three evolved (EV) PF strains (EV-1, EV-2 and EV-3). The three expression states were as follows: low expression level (LE), HE, and the original high phenotypic heterogeneity (HH) setting (for further details see S8 Fig). Absolute fitness was estimated by the increment of the OD600 after 72 h of growth in drug-free medium. Relative fitness was calculated by normalizing to the absolute fitness of the corresponding strain in the HH expression state. Error bars indicate 95% confidence intervals, based on growth measurement of at least 60 independent cultures per expression state. (B) Phenotypic heterogeneity provides fitness advantage under fluconazole stress. The figure shows the relative fitness of four genotypes as a function of increasing fluconazole concentrations. The genotypes studied were the AN PF and three EV PF strains (EV-1, EV-2 and EV-3). Green line indicates LE, blue line indicates HE, while the red line indicates intermediate expression level with HH. Fitness was estimated by the OD600 values after 72 h of growth and was normalized to the fitness in the HH expression setting at each fluconazole dosage. The continuous lines connect the median relative fitnesses, based on growth measurement of eight independent cultures for each expression setting. * indicates significant fitness difference between HE and HH expression settings (Mann–Whitney U test, p < 0.05). In both panel A and B, the inducer levels were set to 0.015 μg/ml (LE), 0.3 μg/ml (HH), and 3 μg/ml (HE), respectively. For further details, see the main text. The data underlying Fig 6 can be found in S1 Data.

Similar articles

See all similar articles

Cited by 28 articles

See all "Cited by" articles

References

    1. Jablonka E, Lamb MJ. Précis of Evolution in Four Dimensions. Behav Brain Sci. 2007;30: 353–365. 10.1017/S0140525X07002221 - DOI - PubMed
    1. Schlichting C, Pigliucci M. Phenotypic Evolution: A Reaction Norm Perspective. Sinauer; 1998.
    1. West-Eberhard MJ. Developmental Plasticity and Evolution. OUP; USA; 2003.
    1. Balázsi G, van Oudenaarden A, Collins JJ. Cellular Decision-Making and Biological Noise: From Microbes to Mammals. Cell. 2011;144: 910–925. 10.1016/j.cell.2011.01.030 - DOI - PMC - PubMed
    1. Tyedmers J, Madariaga ML, Lindquist S. Prion Switching in Response to Environmental Stress. PLoS Biol. 2008;6: e294 10.1371/journal.pbio.0060294 - DOI - PMC - PubMed

Publication types

LinkOut - more resources

Feedback