Evolution of a bistable genetic system in fluctuating and nonfluctuating environments

PNAS 2024 Vol. 121 No. 36 e2322371121 https://doi.org/10.1073/pnas.2322371121 1 of 12 RESEARCH ARTICLE | Significance Bacterial evolution experiments combined with mathematical modeling for studying adaptive strategies in variables natural environments are limited. We exploited the opvAB epigenetic switch, which generates bacterial subpopulations under bacteriophage infection and other selective pressures, to show experimental evidence that phenotypic heterogeneity is advantageous in changing and unpredictable environments. This study illustrates the potential of epigenetic mechanisms to generate variability in microbial pathogens, serving as an adaptive strategy in fluctuating environments. Author contributions: R.F.- F., J.C., and M.A.S.- R. designed research; R.F.- F., D.R.O., E.W., A.S., and M.A.S.- R. performed research; R.F.- F., D.R.O., E.W., A.S., J.C., and M.A.S.- R. analyzed data; and R.F.- F., J.C., and M.A.S.- R. wrote the paper. The authors declare no competing interest. This article is a PNAS Direct Submission. Copyright © 2024 the Author(s). Published by PNAS. This open access article is distributed under Creative Commons Attribution- NonCommercial- NoDerivatives License 4.0 (CC BYNC- ND). 1Deceased August 2, 2022. 2To whom correspondence may be addressed. Email: [email protected]. This article contains supporting information online at https://www.pnas.org/lookup/suppl/doi:10.1073/pnas. 2322371121/- /DCSupplemental. Published August 30, 2024. GENETICS Evolution of a bistable genetic system in fluctuating and nonfluctuating environments RocíoFernández- Fernándeza, DavidR.Olivenzab, EstherWeyerc,d, AbhyudaiSinghc,d , JosepCasadesúsb,1 , and MaríaAntoniaSánchez- Romeroa,2 Affiliations are included on p. 11. Edited by Eugene Koonin, NIH, Bethesda, MD; received December 21, 2023; accepted July 26, 2024 Epigenetic mechanisms can generate bacterial lineages capable of spontaneously switching between distinct phenotypes. Currently, mathematical models and simulations propose epigenetic switches as a mechanism of adaptation to deal with fluctuating environments. However, bacterial evolution experiments for testing these predictions are lacking. Here, we exploit an epigenetic switch in Salmonella enterica, the opvAB operon, to show clear evidence that OpvAB bistability persists in changing environments but not in stable conditions. Epigenetic control of transcription in the opvAB operon produces OpvABOFF (phagesensitive) and OpvABON (phageresistant) cells in a reversible manner and may be interpreted as an example of bethedging to preadapt Salmonella populations to the encounter with phages. Our experimental observations and computational simulations illustrate the adaptive value of epigenetic variation as an evolutionary strategy for mutation avoidance in fluctuating environments. In addition, our study provides experimental support to game theory models predicting that phenotypic heterogeneity is advantageous in changing and unpredictable environments. bacterial experimental evolution | epigenetic switch | adaptative strategies | fluctuating environments | phagebacteria interactions Microbial populations are faced with multiple challenges in natural environments and have evolved and adapted to survive in a wide range of natural conditions ( 1 , 2 ). An especially severe threat is the encounter with bacteriophages. Considering the abundance and ubiquity of these biological entities, one may thus understand the adaptive value of mechanisms that protect from bacteriophage infection ( 3 , 4 ). Indeed, bacteria have evolved a bewildering array of mechanisms of defense against bacteriophages ( 5 ). One such mechanism is the reversible (“phase-variable”) formation of bacterial subpopulations ( 6 ) able to survive in the presence of a virulent phage ( 7 – 12 ). This phase-variable system is an epigenetic switch that allows the existence of multiple stable phenotypic states in a genetically identical population. The occurrence of phase-variable phage resistance in unrelated bacterial species and the diversity of mechanisms involved suggest independent evolution, which in turn may be indicative of adaptive value. In Salmonella enterica , OpvA and OpvB proteins control O-antigen chain length in the lipopolysaccharide (LPS) ( 10 ), offering resistance to bacteriophages that use the O-antigen as receptor ( 11 ). The opvAB operon, regulated by the OxyR transcription factor and DNA methylation, governs opvA and opvB expression, leading to bistable switching between ON and OFF states ( 13 ). The bistable switch is lopsided toward the OFF state, and, although cells are predominantly in the OFF state, the encounter with a phage selects for the survival of OpvABON subpopulation ( Fig. 1A ). As soon as phage challenge ceases, phase variation rebuilds a mixed population made of OpvABOFF and OpvABON lineages ( 10 , 11 ). Because OpvAB ON cells are impaired for interaction with the animal host, phase variation of O-antigen length can be viewed as a tradeoff between pathogenic capacity and bacteriophage resistance. The OpvABON subpopulation of Salmonella preadapts to phage challenge survival, albeit at the expense of reduced virulence ( 11 , 14 ). Phenotypic switching has been proposed as a mechanism of bet-hedging to deal with fluctuating environments ( 15 – 21 ). Specifically, formation of OpvABOFF and OpvABON lineages may be seen as an example of bet-hedging ( 22 ), which allows Salmonella populations to prepare themselves for potential encounters with phages ( 11 ). Natural selection of phenotypic heterogeneity is a controversial notion in classical Darwinism as it involves group selection, which has been traditionally considered a weak evolutionary force ( 23 ), a criticism especially severe when formation of lineages is interpreted as a case of bethedging ( 24 ). This view is however countered by game theory models ( 25 – 27 ). Formation of distinct types of cells preadapted to different environments is also in agreement with OPEN ACCESS Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. 2 of 12 https://doi.org/10.1073/pnas.2322371121 pnas.org OpvABOFF OpvABON (Susceptible to phages) (Resistant to phages) dilution 1/100dilution 1/100dilution 1/100 dilution 1/100dilution 1/100dilution 1/100 dilution 1/100 dilution 1/100 dilution 1/100 dilution 1/100 Fluctuating environment Constant environment AB Switching frequency 6·10 -5 4·10 -2 C O N S T AN T E N VI R O NM E N T F L U C T UA TI N G E N VI R O NM E N T C E xper i me n t 1 E xper i me n t 2 E xperime n t 3 E xper i me n t 4 0510 15 20 25 30 35 40 45 50 0 Number of generations Proportionof OFFcells 0510 15 20 25 30 35 40 45 50 0 Number of generations Proport ion of OFFcells 0510 15 20 25 30 35 40 45 50 0 Number of generations proportionof OFFcells 0510 15 20 25 30 35 40 45 50 0 0.25 0.50 0.75 1.00 Number of generations P roport iono f OFF cells 0510 15 20 25 30 35 40 45 50 55 0 Number of generations Proportion of OFFcells 0510 15 20 25 30 35 40 45 50 55 0 Number of generations Proportion of OFFcells Number of generations Proportion of OFFcells Number of generations Proportion of OFFcells 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 D Long O-antigenShort O-antigen 0 x 0 x 0 x 0 x 0 x 0 x 0 x 0 x xp erimen t 1 E x per i men t 2 E x perime n t 3 E x per i men t 4 Number of generations Number of generations Number of generations Number of generations Number of generations Number of generations Number of generations Number of generations [phages][phages] [phages] [phages] [phages][phages][phages] [phages] Proportion of OFFcells Proportionof OFFcells Proportion of OFFcellsProportion of OFFcells 0510 15 20 25 30 35 40 45 50 55 60 65 0510 15 20 25 30 35 40 45 50 55 60 65 0 absence phage 0.25 0.50 0.75 1.00 0.25 0.50 0.75 1.00 Fig. 1. Evolution of the bistable opvAB genetic system in fluctuating versus constant environments. (A) S. enterica opvAB expression undergoes phase variation and produces a lineage of OpvABOFF cells with long O- antigens in the lipopolysaccharide that are susceptible to phages, and a lineage of OpvABON cells with shorter O- antigen chains resistant to phages. Switching from OFF to ON states occurs at lower frequencies than switching from ON to OFF. The frequency of OFF to ON transition was estimated to be 6 × 10−5 per cell and generation. The ON to OFF switching rate was around 1,000- fold higher: 4 × 10−2 per cell and generation. Proportion of OFF and ON cells in the bacterial population is around 99.8% and 0.2%, respectively (12). (B) Scheme of the laboratory evolution experiment to simulate fluctuating and constant environments. For fluctuating environment, presence and absence of bacteriophages were used to simulate a changing environment. A dilution of an overnight culture of 14028 opvAB::lacZ strain was performed in presence of phage and plated on LB supplemented with X- gal and agar with phages. Phage removal protocol was carried out to remove phage number and cease the challenge. Six alternating cycles of phages’ presence and absence were performed. For constant environment, cultures were achieved by diluting an overnight culture of 14028 opvAB::lacZ strain in presence of 9NA lysate, plating on LB supplemented with X- gal plates and phages each 24 h. (C and D) Proportion of OpvABOFF colonies of the strain opvAB::lacZ growing in fluctuating (C) and continuous presence of bacteriophage (D). Four independent laboratory evolution experiments were performed. Yellow shaded area under the curve represents the proportion of OFF cells, and blue shaded area above the curve shows the proportion of ON cells in the bacterial culture. Dotted gray shaded areas indicate the absence of phages in fluctuating environment. Simplified diagrams clarify the presence and absence of phages over generations. Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. PNAS 2024 Vol. 121 No. 36 e2322371121 https://doi.org/10.1073/pnas.2322371121 3 of 12 theoretical models by Richard Levins showing that polymorphism is advantageous in changing environments ( 28 , 29 ). Even though Levins’s studies were intended for genetic polymorphism, their conclusions could potentially be extended to any source of variation, including formation of epigenetic lineages. Natural selection operates on phenotypes, not on genotypes ( 30 , 31 ). To test Levins’s prediction that polymorphism increases the fitness of biological populations in changing environments but not in stable ones, we exploited the well-known opvAB epigenetic switch, which generates bacterial subpopulations differing in specific phenotypic traits. We conducted bacterial evolution experiments and applied mathematical modeling to shed light on the factors by which epigenetic mechanisms contribute to the ability of Salmonella to adapt and persist in natural environments. Results OpvAB Subpopulation Proportions Vary Depending on the Fluctuations of the Environmental Conditions. We exploited the opvAB gene expression system to investigate the variation of OFF and ON phenotypic states in fluctuating and nonfluctuating environments (Fig.1 A and B). We used a S. enterica strain harboring an opvAB::lacZ fusion (SV8011) to conduct four independent laboratory evolution experiments in different environmental conditions (Fig.1B). The strain SV8011 was grown with and without bacteriophages, and white (OpvABOFF) and blue (OpvABON) colonies were counted. In the fluctuating environment, alternate growth cycles with and without bacteriophages were repeated and colony counts were recorded (Fig.1 B and C). Phagefree medium was performed after removal of phage by treatment with purified LPS. In the constant environment, SV8011 was growth in presence of phages, and the bacteriophage removal step was omitted (Fig.1 B and D). The results of these experiments (Fig.1 C and D) reveal distinct patterns. In alternating phage presence (Fig.1C), the proportion of OpvAB OFF colonies decreases in phage presence, rebounding when absent. Conversely, the proportion of OpvABON colonies exhibits an opposite trend. In continuous phage exposure (Fig.1D), the proportion of OpvABON initially increases, but unexpectedly decreases with distinct speed and kinetics at each experiment after 5 to 10 generations, leading to nearextinction. OpvABOFF colonies then took over the culture after 40 to 50 generations. The numerical data were slightly different in each experimental replicate, but the overall phenomenon remained the same: OpvABON and OpvABOFF colonies are present throughout the experiment carried under fluctuating selection, whereas OpvABOFF colonies overtake the culture growing under continuous selection with a virulent phage. Importantly, this laboratory evolution experiment supports Levins’s prediction ( 28 , 29 ): Bistability is favored by maintaining different subpopulations in fluctuating environments. In contrast, when the environmental conditions are constant, phenotypic heterogeneity is lost and a monostable cell populational distribution becomes established. OpvABOFF Isolates from Cultures Grown in the Continuous Presence of Phage Are Predominantly Mutants. Since OpvABON colonies are made of phageresistant cells (Fig. 1A), survival upon bacteriophage infection explains their predominance over OpvABOFF colonies under fluctuating conditions (Fig.1C). In contrast, the number of OpvABON colonies progressively decreases under continuous selection, dropping to nearextinction; OpvABOFF colonies thus become the predominant class (Fig.1D). Because OpvABOFF cells have long O- antigen chains, the survival of OpvABOFF colonies from the constantenvironment evolutionary experiment in the presence of phage must be made possible by mechanism(s) unrelated to opvAB phase variation. One potential mechanism could be mutation. To investigate whether mutational genetic adaptation is responsible for the survival of OpvABOFF cells in the presence of phage, we subjected twenty phageresistant OpvABOFF isolates to whole DNA genome sequencing. All of them showed mutations in their genomes (356 different mutations were detected). Two thirds mapped to coding sequences (97.5% in known genes and 2.5% in genes annotated as hypothetical proteins) and one third to intergenic regions. As a result, 72 genes encoding proteins with known function were affected (Fig. 2 and SI Appendix, Table S1). More than 50% (38 out of 72 genes) of these genes were altered in more than one phageresistant isolate (Fig.2A) and nearly 87% (33 out of 38 genes) presented more than one mutation in the same gene, predominantly substitutions (Fig.2 B and C). All these observations suggest that these mutations were not randomly distributed. The analysis of mutated genes reveals a high prevalence of genes related with phage infection and subsequent bacterial lysis (Fig. 2D). In addition, a number of mutations in genes associated with virulence and cell permeability were identified. Other biological processes compromised in phageresistant isolates were related to translation, carbohydrate metabolism, and cell wall and membrane biogenesis (Fig.2D). Most of these processes have been also reported to be altered in phageresistant isolates from other Gramnegative bacteria, such as Escherichia coli or Vibrio alginolyticus (32–34), suggesting their robust implication in mechanisms of phage resistance. To demonstrate their involvement in phage resistance of twenty phageresistant OpvABOFF isolates a susceptibility assay based on EBU (Evans Blue Uranine) plate and growth curves in presence/absence of phage were performed (SIAppendix, Fig.S3), confirming the phage resistance phenotype of these OpvABOFF isolates. In summary, DNA sequencing of OpvABOFF isolates from cultures grown in the continuous presence of bacteriophages demonstrated the existence of mutations related to phage resistance, supporting the hypothesis that prolonged exposure to phages influences the genetic adaptation and favors a more stable phenotype over bistability. Mathematical Modeling of the Bistable opvAB Epigenetic System. To further explore the evolution of the bistable gene expression in opvAB system under diverse environmental conditions, first, we constructed a mathematical model for simulating the opvAB operon subjected to epigenetic control by DNA methylation. Second, we evaluated how the rates between Salmonella OpvABOFF and OpvABON cell subpopulations evolved in contact with bacteriophages using different model assumptions. Interaction dynamics in Salmonella cell population containing the opvAB operon in presence of phage is illustrated in Fig. 3A . To start with, we simulated the evolution of a bacterial population in an environment where phages were always present ( Fig. 3 B and C ) or where absence and presence of phages fluctuated, by sequentially removing and reintroducing phages over several intervals ( Fig. 3 D and E ). The predicted fraction of ON and OFF cell states was plotted as a function of time ( Fig. 3 B and D ). In the constant presence of bacteriophage, the proportion of OFF cells was predicted from the initial concentrations using Eq. 9 in SI Appendix , Fig. S1 . The model is able to match the initial rapid decrease of the proportion of OFF cells after the introduction of bacteriophages and the subsequent gradual recovery to the initially observed proportions of OFF/ON subpopulations (majority of OFF cells), as shown experimentally ( Fig. 3B and SI Appendix, Fig. S2 ). When the proportion of mutants was Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. 4 of 12 https://doi.org/10.1073/pnas.2322371121 pnas.org modeled, the mutant proportion shows a short delay before a rapid increase approaching 1 ( Fig. 3C ). This trend correlated with the previously observed experimental data, which showed the prevalence of OFF mutant cells in a constant environment. Thus, the model reproduces experimental data demonstrating that constant environment conditions favor the selection of mutations and A CD B Fig. 2. Description of the mutations detected in OpvABOFF phageresistant isolates. (A and B) Identification of altered genes found in the DNA sequencing of OpvABOFF phageresistant isolates. For each affected function, predominancy in phageresistant isolates (out of 20) (A, orange) and type of mutations (B, substitutions (violet), insertions (yellow), and deletions (green)) found in each gene are represented. (C) Analysis of types of most common mutations found in the 20 OpvABOFF phage resistant isolates. Altered functions shown are shared by at least 30% of the mutants (6 out 20). Types of mutations are identified. Colors identified the number of phageresistant isolates. (D) Clusters of biological process classification and distribution (frequency/incidence) of mutated genes in sequenced phageresistant isolates. Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. PNAS 2024 Vol. 121 No. 36 e2322371121 https://doi.org/10.1073/pnas.2322371121 5 of 12 A F G HI J BD CE Fig. 3. Modeling the bistable opvAB system in diverse environmental conditions. (A) Flow chart of populations and interaction dynamics used for model. Uninfected cells replicate and switch between OFF and ON states. Wildtype cells permanently mutate and become resistant to the bacteriophage particles even in the OFF state. Susceptible OFF cells are infected by bacteriophage particles and continue to replicate until they become late infected and release new bacteriophage particles by bursting. (B) Proportion of OpvABOFF cells predicted by the model with best fitting parameters from the constant environment. (C) Proportion of mutant cells predicted by the model using the best fitting parameters from the constant environment. (D) Proportion of OpvABOFF cells predicted by the model from the fluctuating environment. (E) Proportion of mutant cells predicted by the model using the best fitting parameters from the fluctuating environments. Predicted proportion of cells in the ON state was calculated by 1- proportion of OFF cells. (F) Proportion of OpvABOFF cells and proportion of mutant cells predicted by the model using different intervals of phage absence and presence in the simulations from the fluctuating environment, assuming 0% mutant survival. (G) Proportion of OpvAB OFF cells predicted by the model using different intervals of phage absence and presence in the simulations from the fluctuating environment, assuming 0% mutant survival. In these simulations it is assumed 5 generation intervals in presence of phage but different phage absence interval times. (H–J) Proportion of OpvABOFF cells and proportion of mutant cells predicted by the model using 5 and 10 generation intervals in the simulations from the fluctuating environment, assuming 5%, 50%, and 100% mutant survival (H, I, and J, respectively). Yellow shaded area under the curve represents the proportion of OFF cells and blue shaded area above the curve shows the proportion of ON cells in the bacterial culture along time. Red shaded area under the curve represents the proportion of mutant cells over time. Dotted gray shaded areas indicate the absence of phages in fluctuating environments. Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. 6 of 12 https://doi.org/10.1073/pnas.2322371121 pnas.org decrease phenotypic diversity ( Figs. 1D and 3 B and C and SI Appendix, Fig. S2 ). For the fluctuating environment, the model was roughly able to predict the shifting behavior of the OFF and ON proportions. Expectedly, the simulation illustrates the increase in the proportion of OFF cells as bacteriophage particles were removed and the upsurge in the ON proportion as phage particles were reintroduced ( Fig. 3D ). However, the prediction of population proportions in fluctuating conditions oscillated close to 0.5 values. This slightly differs from our previous experimental observations where the amplitude of oscillations varies between experiments ( Fig. 1C ). This discrepancy could be due to assuming cyclical fluctuations in intervals of five generations in presence and absence of phages for modeling the fluctuating environment and it will be discussed later. As a consequence of the procedure for bacteriophage removal, the number of generations (which depends on initial and final cell numbers in the culture at the end of the phage-challenge step) during each cycle of phage presence and absence varies across different experiments ( Fig. 1C , shaded intervals). We ran several simulations changing the number of generations occurring in the presence and/or absence of phages ( Fig. 3 F and G ). For each scenario, OFF and ON subpopulations’ size fluctuates based on the point of occurrence of phages and the duration spent recovering in culture without phage. In general, the fluctuations in the proportions of ON and OFF population (amplitude of the oscillations) became much larger as frequency decreased. Therefore, in fast-changing environments, the population evolved to a bistable phenotype ( Fig. 3 F , Top ), whereas in slow-changing environments the tendency is for one phenotype to be predominant ( Fig. 3 F , Bottom ). When we conducted simulations with asymmetric cycles (phage-free intervals longer or shorter than intervals with phage presence), we observed that the amplitude of oscillations seems to correlate with the length of the interval without phage ( Fig. 3G ). The longer the culture is free of phages, the greater the amplitude of oscillations, and the higher the probability that one subpopulation dominates the culture for an extended period. These phenomena could be similarly observed in each of the replicates undergoing fluctuating environments ( Fig. 1C ). Hence, as the simulation indicates and our experimental data show, the period of environmental changes can modulate the OFF and ON subpopulations proportions generated by an epigenetic switching system. As previously described, mutant cells become predominant under a constant phage pressure, consequently, genetic adaptation appears to be a better strategy in constant environments. Surprisingly, we found a similar behavior in the experiment number four grown in fluctuating environments: The OFF cells take over the culture after 30 to 35 generations leading to the establishment of a single subpopulation ( Fig. 1C ). We wondered how mutant survival affects the behavior on OFF/ON proportions predicted by the model in changing environments. The initial assumption in our model was that no mutants survived at the dilution events, preventing the recovery of the initial OFF/ON proportions. In fluctuating environments, the jump in proportions that occurs after the first interval is due to the assumed loss of the mutant population ( Fig. 3D ). At the end of the first interval (5 generations in the presence of phages), the proportion of mutants is about 10% ( Fig. 3E ), so a gap occurs when the mutant concentration is set to 0 for the next initial conditions. When we considered 10 and 20 generation intervals, the proportion of mutants at the end of the initial interval was close to 25% and 75%, respectively ( Fig. 3F ), causing a much larger gap in the predicted proportions ( Fig. 3F ). Because total extinction is a strong assumption, the same interval scenarios were also tested with 5%, 50%, and 100% mutant survival rates ( Fig. 3 H –J ). When a survival rate of 5% was tested, no significative difference compared to the 0% survival was found ( Fig. 3 F and H ); however, a 50% mutant survival rate showed distinctive patterns of fluctuations in the different interval scenarios and the ON/OFF proportions noticeably began to approach the steady state much more rapidly ( Fig. 3I ). Because there was partial survival, the gaps became smaller at the end of the first interval, but then greater at the subsequent dilutions where the mutant population was higher than in the 0% survival cases. The increase in the mutant proportion during the intervals of phage presence is also more noticeable. It is worth highlighting that unlike what happens with the predicted OFF/ON proportions in 10 generations intervals and 0% mutant survival ( Fig. 3F ) where the ON subpopulation takes over the culture during phage presence intervals, when a 50% mutant survival is tested, the OFF subpopulation becomes predominant after 40 to 50 generations ( Fig. 3I ). A 100% survival mutant rate produced very similar results to the constant environment case ( Fig. 3J ), and OpvABOFF subpopulation takes over the culture more rapidly. These model parameters would explain the results obtained for the experimental replicate number 4 grown in fluctuating conditions ( Fig. 1C ). To clarify the disparity in Fig. 1 , Experiment 4 under fluctuating conditions and experimentally demonstrate the information predicted by the mathematical model, we calculated the proportion of ON and OFF cells in three different laboratory evolution experiments with alternate cycles with and without bacteriophages (SI Appendix, Fig. S4 ). First, a culture of SV10150 strain (a phage-resistant mutant isolate) was grown (SI Appendix, Fig. S4A ); second, the behavior was analyzed in a culture of SV8011 strain (wild type) with the incorporation of bacterial cells from a culture of SV10150 (phage-resistant mutant strain) during the first contact with phages (SI Appendix, Fig. S4B ); and, third, with the incorporation of bacterial cells from a culture of SV10150 (phage-resistant mutant strain) after the first contact with phages, when phages are removed from culture (SI Appendix, Fig. S4C ). These evolution experiments simulate the behavior of cell subpopulations when the emergence of mutations occurs at different times for the bacterial evolution. In the case of cultures from a phage-resistant mutant isolate, it does not undergo changes in the proportion of OpvABON and OpvABOFF cells (SI Appendix, Fig. S4A ). This was expected since SV10150 is a strain containing phage-resistant mutations and does not need to select subpopulations to adapt to stress by phage challenge: All cells are in state OFF. In experiments with the artificial incorporation of cells containing phage-resistance mutations to bacterial culture (SI Appendix, Fig. S4 B and C ), when the culture is exposed to phages, as soon as phage-resistant mutants “appear,” they saturate the culture because they have a low fitness cost able to grow better and compete with wild type cells to take over the population: The ON subpopulation is no longer selected and predominant even under fluctuation conditions. Therefore, the discrepancy observed in Fig. 1 Experiment 4 under fluctuating conditions is very probably a consequence of the appearance of any mutations that conference phage resistance. In summary, our model accurately predicts how the fluctuations in the presence/absence of phages and the survival rate of mutants significantly affect the structure of bacterial cell populations. Response of Epigenetic for opvAB Expression to Multiple Selective Pressures. Microorganisms face the challenge of adapting to multiple selection pressures. In natural environments, where the presence of a selective pressure eliminated the mutants, the maintenance of the ON/OFF populations would offer the higher Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. PNAS 2024 Vol. 121 No. 36 e2322371121 https://doi.org/10.1073/pnas.2322371121 7 of 12 evolutionary process. The expression of the opvAB operon reduces the length of the O- antigen and, consequently, the absorption of bacteriophages, becoming resistant to these phages. Notably, the acquisition of phageresistance by phase variation of O- antigen chain length involves an expensive payoff for a pathogen as Salmonella: opvAB expression impairs serum resistance and reduces proliferation within macrophages (10, 11). Thus, we hypothesized that mutants resistant to phages would experience a permanent reduction in virulence; on the contrary, the transient phase variation of opvAB gene expression would offer bacterial populations the versatility needed to confront a variety of adverse agents, such as phages and their hosts. To demonstrate this, we examined survival in guinea pig serum (GPS) and intramacrophage replication in cultures containing cells that could reversibly switch (SV8011 strain) or be permanently locked in the ON (SV6401 strain) or OFF states (SV6413 strain, OpvABOFF lineage phageresistant with altered LPS profile) (Fig.4). Survival in GPS was analyzed for 90 min (Fig.4A) and the rate of intramacrophage replication was measured after 18 h of infection in J774 mouse macrophages (Fig.4B). Our results showed an increase of killing by serum and an impaired intramacrophage proliferation of those strains in which phase variation mechanism is abolished (SV6401 and SV6413 strains) compared to wild type (SV8011). Interestingly, when SV8011 cultures that were previously in contact with phage (thus containing the majority of OpvABON cells but with active reversible switching) were analyzed, the serum resistance and the ability to proliferate within macrophages increased compared to the strains with inactive opvAB phase variation (Fig.4 A and B). These observations support our hypothesis that the reversible switching of opvAB gene expression may be required for adaptation to changing AB CD Fig. 4. Role of reversible switching of opvAB expression for adaptation to diverse selective pressures. (A) Survival rate in presence of 30% of guinea pig serum (GPS) for 90 min. Survival rate is relative to 0 min in response to GPS. Survival of the wild type strain (SV8011) is considered 100%. (B) Intramacrophage proliferation rate for 18 h inside J774 murine macrophages cell line. Proliferation rate is relative to 0 min in contact with macrophages. Survival of the wildtype strain (SV8011) is considered 100%. (C) Dynamic analysis of proportion of OpvABON and OpvABOFF cells in presence of GPS that have previously been in contact with phage. Proportions were investigated in a wildtype strain (SV8011) before adding bacteriophage, after 10 generations with 9NA phage, immediately after adding 30% GPS, and after 90 min of incubation in presence of GPS. (D) Dynamic analysis of proportion of OpvABON and OpvABOFF cells inside macrophages of the wildtype strain (SV8011) that have been in contact with bacteriophage before invasion of macrophages. Proportions shown correspond to the wildtype strain before adding bacteriophage, after 10 generations in presence of 9NA phage, 30 min inside macrophages, and after 18 h of proliferation in macrophages. Strains used for these studies were the following: SV8011, as wildtype strain harboring an opvAB::lacZ fusion, containing active epigenetic switching, that is, the majority of OpvABOFF cells and minority of OpvABON cells (around 0.2%); SV8011 + phage is a culture that was previously in contact with 9NA phage for 10 generations, containing the majority of OpvABON cells and minority of OpvABOFF phageresistant cells, probably as mutants; SV6413 is a strain containing an opvAB deletion and an altered LPS profile, which exhibited phage resistance and bacterial cultures only show OFF phenotype; SV6401, a strain that remains blocked in the OpvABON state and, thus, bacterial population only display ON cell phenotype. SV6413 and SV6401 strains present permanent lipopolysaccharide modifications and are phageresistant. SV6796 strain is 14028ΔwzzST ΔwzzfepE and is sensitive to serum. SV6063 is an avirulent strain which has a deletion of SPI- 2 genes that makes it incapable of proliferating inside macrophages.The experiment was performed in triplicate and averages and SD are shown. Statistical indications: ns, not significantly different, ****P < 0.0001, ***P < 0.001, **P < 0,01, *P < 0.05. Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. 8 of 12 https://doi.org/10.1073/pnas.2322371121 pnas.org environments. To confirm this hypothesis, we exposed wildtype cultures to bacteriophages for 10 generations prior to monitoring the proportion of OpvABON and OpvABOFF subpopulations in the presence of serum (Fig.4C) or before growth within macrophages (Fig. 4D). Highly dynamic OFF/ON proportions and drastic changes were observed: The initial proportion of OpvABON cells was very small (0.2%) in the absence of phage, but, as soon as bacteriophages were added, the OpvABON proportion increased rapidly, reaching values between 45% and 70%. However, after 90 min of serum treatment or 18 h of intramacrophages replication, the OFF/ON proportions returned to the initial prephages levels and bacterial cultures were again enriched in OpvABOFF cells (serum resistant and increased intramacrophage proliferation). Accordingly, survival in response to a new additional selective pressure (exposure to serum or macrophages) is made possible by opvAB phase variation, which allows the fast switching of opvAB from ON to OFF better states adapted to the new challenges. These results highlight the importance of the reversible switch between ON and OFF expression, as a method for dealing with rapidly varying environments without involving potentially random mutation. Rapidly and specially unpredictably fluctuating environments could select epigenetic gene expression mechanisms for reversible phenotypic plasticity. Discussion Prokaryotes are constantly exposed to changing environmental conditions. Adaptation to these environments is facilitated by prokaryotic biochemical machinery which produces different phenotypic variants. Both mutational and nonmutational mechanisms can contribute to generate this phenotypic variability ( 14 , 35 – 39 ). In this work, we investigated the adaptative significance of phe - notypic heterogeneity in bacterial survival using a well-characterized mechanism that generates phenotypic variation: the epigenetic switch based on the bistable expression of opvAB system ( 10 , 11 , 40 ). Here, we setup laboratory evolution experiments to monitor the formation of OpvABOFF and OpvABON lineages in contact with bacteriophages. In alternate cycles of phage presence/absence, opvAB bistability persists beyond 50 generations, supporting opvAB phase variation as a strategy used by Salmonella to survive potential phage encounters. By contrast, continuous phage pressure appears to select for the mutants in genes related to new phage particle assembly and bacterial lysis, as well as genes associated with virulence and cell permeability ( 32 – 34 ) ( Figs. 1 and 2 ). Our results provide clear evidence that in changing environments phenotypic diversity achieved by epigenetic mechanisms is beneficial for the population, whereas, under a constant environment, the maintenance of more than one phenotype appears to be expensive for the population and epigenetic polymorphism is lost, and adapted mutants are selected instead. These findings are in consonance with our mathematical predictions ( Fig. 3 ) and those made by R. Levins in the 60s ( 28 , 29 ) based on game theory, who predicted that genetic polymorphism is advantageous in changing environments but not in stable environments. Therefore, this study extends the Levins’s predictions made for genetic polymorphism to epigenetic polymorphism ( 41 – 45 ). Increased phenotypic heterogeneity arising from cellular differentiation into distinct phenotypes can improve the population’s adaptability and proliferation in dynamic environments, thereby enhancing the robustness of the bacterial population to environmental stress. Bacterial adaptation to environmental changes can be facilitated by phenotypic adaptation on a short timescale and by tuning via mutations in the long run ( 36 ). Both genetic and epigenetic variations are not mutually exclusive and can compete as two strategies of adaptation to fluctuating environments. Using computational approaches, we simulated the epigenetic switching based on opvAB system under a range of factors (selection pressure, fluctuation frequency, mutation step-size) to mimic different environmental conditions ( Fig. 3 ). The model simulations seem to adjust quite well to experimental data and predict conditions that favor adaptation by epigenetic switching over mutation depending on the duration and strength of selection pressure. Epigenetic switching is selected over genetic adaptation in fast-changing environments since this strategy seems to minimize adaptation time when the environment fluctuates frequently. These results highlight the importance of epigenetic switching to module the structure of bacterial cell subpopulations in changing environments ( 16 , 17 , 20 , 21 ). Microbial populations can evolve a combination of genetic (selection of beneficial mutations) and epigenetic (generation of nonmutational variation that allows adaptation to environmental challenges without mutation) strategies to survive and thrive in dynamic and unpredictable environments. Unlike genetic adaptation, which is a mutational and mostly irreversible mechanism, reversible epigenetic switching provides a high adaptability to adverse agents in natural environments ( Fig. 5 ). Evolution laboratory experiments using multiple selective pressures (exposure to phage, GPS, or macrophages) in cultures containing different proportions of OpvABON and OpvABOFF cells and even cultures with cells blocked in one state (ON or OFF) revealed that only cells that can switch between the OFF and ON states, thanks to the reversibility of opvAB phase variation, exhibited the highest survival to added environmental challenges ( Fig. 4 ). These results bring to light the advantages of nonmutational versus mutational adaptation mechanisms in unpredictable natural environments ( Fig. 5 ). By switching between phenotypic states, microbial populations can quickly respond to changes in their surroundings, allowing them to take advantage of beneficial conditions and evade unfavorable ones ( Fig. 5A ). On the contrary, in the case of genetic adaptation, after an environmental change, selection gradually shifts the entire population toward the new optimal phenotype through a mutation or the accumulation of more than one mutation. However, this new phenotype is likely not well-adapted to a new different challenge and the entire bacterial population could be killed ( Fig. 5B ). Therefore, the possibility of quickly reverting phenotypes provides an added value to epigenetic switching as essential tools for adaptation to changing environments. Microorganisms have developed mechanisms to adapt to changing environmental conditions. Examples include encounters with bacteriophage and antimicrobial compounds, avoiding elimination in hosts that can mount an immune response following infection, and adapting to nutrient variability. Understanding these mechanisms may have potential applications in the fields of biotechnology and biomedicine, particularly in the development of strategies to combat pathogenic strains. Combining theoretical and experimental studies, we have identified how diverse adaptative strategies result in optimal bacterial growth in specific environments. Particularly, our research highlights the role of epigenetic changes in increasing phenotypic diversity and flexibility, enabling bacteria to better respond to environmental fluctuations. Focusing on phage interactions, we show how bacterial responses to threats like bacteriophages can model broader microbial information-processing strategies. This work illustrates the adaptive value of epigenetic variation as a strategy for mutation avoidance in changing and diverse environmental conditions. Phenotypic heterogeneity generated by phase variation may thus allow for enough flexibility to adapt to rapidly changing environments. Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235. PNAS 2024 Vol. 121 No. 36 e2322371121 https://doi.org/10.1073/pnas.2322371121 9 of 12 Materials and Methods Bacterial Strains and Bacteriophages. S. enterica strains listed in SIAppendix, TableS5 belong to serovar Typhimurium and derive from the mousevirulent strain ATCC 14028. To simplify, S. enterica serovar Typhimurium is often abbreviated as S. enterica. Bacteriophages 9NA (46, 47) was kindly provided by Sherwood Casjens, University of Utah, Salt Lake City. Media and Growth Conditions. Bertani’s lysogeny broth (LB) was used as standard growth medium. Solid media contained agar at 1.5% final concentration. Cultures were grown at 37 °C. Aeration of liquid cultures was obtained by shaking at 200 rpm in an Infors Multitron shaker. Simulation of Fluctuating Environments. Presence and absence of bacteriophages were used as conditions to simulate a fluctuating environment (Fig.1B). The first day of the assay, an overnight culture of opvAB::lacZ strain (SV8011) was diluted 1:100 in 5 mL of LB and 100 µL of 9NA lysate (109 to 1012 PFU/ mL). After 24 h of incubation at 37 °C with shaking, an aliquot was plated on LB supplemented with X- gal (40 µg/mL) and 100 µL of 9NA bacteriophage. Before the following dilution step, phages were removed using the procedure detailed in the section “Bacteriophage removal procedure.” After phage removal, a dilution was made into 9NA bacteriophagefree LB and incubated for 24 h at 37 °C with shaking. The following day, cells were plated on LB supplemented with X- gal plates (9NA bacteriophagefree). The following dilution and plating were made in presence of 9NA phage, and so on. Fluctuating conditions were maintained for 7 d: One day adding 9NA phages to cultures and plating on plates with phage, and the following day reducing the number of phages in cultures and plating on phagefree plates. This assay was performed in quadrupled. Simulation of Constant Environments. To simulate a continuous environment (Fig.1B), an overnight culture of 14028 opvAB::lacZ strain (SV8011) was diluted 1:100 in 5 mL of LB with 100 µL 9NA bacteriophage (109 to 1012 PFU/mL). The starting culture was incubated 24 h at 37 °C with shaking at 200 rpm. The following day, an aliquot of this culture was plated on LB supplemented with X- gal plates. These plates had been previously spread with 100 µL of 9NA bacteriophage. Plates were incubated 24 h at 37 °C. These cultures were used for making a new dilution, also in presence of phages. For seven days, dilutions and plating were made in presence of 9NA bacteriophage. This assay was performed in quadrupled. Bacteriophage Removal Procedure. We have finetuned a protocol to reduce the number of bacteriophages in cultures. First, to remove bacteriophages in suspension, cultures containing phages were centrifugated for 10 min at 4,500 Bacteriophages Serum or macrophages Bacteriophages Serum or macrophages Selection of the OpvAB ON subpopulation Resuscitation of OpvAB OFF subpopulation Reconstruction of the original population OpvABON OpvABOFF Selection of the resistant subpopulation REVERSIBLE, NON-MUTATIONAL ADAPTATION BACTERIAL SUBPOPULATIONS OpvABOFF OpvABON Susceptible to phages and resistant to serum and macrophages Resistant to phages and susceptible to serum and macrophages Mutant Resistant to phages OpvABOFF OpvABON Epigenetic switching frequency 6·10-5 4·10-2 wild type mutant 10-12 Mutation rate MUTATIONAL ADAPTATION New population composed by phage-resistant mutant FREQUENCY OF SWITCHING BETWEEN SUBPOPULATIONSENVIRONMENTAL CONDITIONS Selection of the resistant subpopulations Selection of the OpvAB OFF subpopulation Resuscitation of OpvAB ON subpopulation and reconstruction of the original population Bacteriophages Serum or macrophages Dead cell Susceptible to phages, serum or macrophages Selection of the mutant subpopulation No resuscitation of wild type subpopulation Selection of the resistant mutant cells mutant wild type No selection. No resistant subpopulations. All cells disappear. No resucitation of subpoulations. . DEAD CULTURE . Selection of the resistant subpopulation AB Fig. 5. Microbial strategies involved in adaptation to diverse and changing environments controlled by opvAB system. Phenotypic variants in a bacterial population can be fashioned by epigenetic (A) and/or genetic mechanisms (B). (A) Transcription of the S. enterica opvAB operon produces lineages of OpvABOFF and OpvABON cells. In the presence of phages, the OpvABOFF subpopulation will be eliminated, while the OpvABON subpopulation will survive. When the phage challenge ceases, OpvABOFF cells produced by phase variation will survive, and the bistability of OpvAB will be regained. The presence of both OpvABOFF and OpvABON subpopulations will sustain preadaptation to future challenges. In presence of serum or macrophages, the OpvABOFF subpopulation will be now the survivor, allowing the subsistence of the bacterial culture. (B) Genetic adaptation involves a beneficial mutation to resist the phage challenge. In the presence of phages, mutant cells will survive and take over the culture; but when phage attacks stop, wildtype cells will not be restored since bistability was lost. The lack of phenotypic switching in culture will reduce the ability of bacterial population to adapt to new and future challenges, where the acquired mutations may not result advantageous. One notable distinction between these strategies is that while epigenetic mechanisms enable reversible switching between different states, genetic adaptation has an irreversible impact in a short time (frequency of switching for both mechanisms is detailed in legend). Downloaded from https://www.pnas.org by 150.214.182.235 on January 27, 2025 from IP address 150.214.182.235.