Limits...
Adaptive resistance in bacteria requires epigenetic inheritance, genetic noise, and cost of efflux pumps.

Motta SS, Cluzel P, Aldana M - PLoS ONE (2015)

Bottom Line: Here, using an efflux pump regulatory network (EPRN) model, we show that the following three mechanisms are essential to obtain adaptive resistance in a bacterial population: 1) intrinsic variability in the expression of the EPRN transcription factors; 2) epigenetic inheritance of the transcription rate of EPRN associated genes; and 3) energetic cost of the efflux pumps activity that slows down cell growth.Our results identify the molecular mechanism of epigenetic inheritance as the main target for therapeutic treatments against the emergence of adaptive resistance.Finally, our theoretical framework unifies known and newly identified determinants such as the burden of efflux pumps that underlie bacterial adaptive resistance to antibiotics.

View Article: PubMed Central - PubMed

Affiliation: Instituto de Ciencias Físicas, UNAM, Cuernavaca, Morelos, Mexico.

ABSTRACT
Adaptive resistance emerges when populations of bacteria are subjected to gradual increases of antibiotics. It is characterized by a rapid emergence of resistance and fast reversibility to the non-resistant phenotype when the antibiotic is removed from the medium. Recent work shows that adaptive resistance requires epigenetic inheritance and heterogeneity of gene expression patterns that are, in particular, associated with the production of porins and efflux pumps. However, the precise mechanisms by which inheritance and variability govern adaptive resistance, and what processes cause its reversibility remain unclear. Here, using an efflux pump regulatory network (EPRN) model, we show that the following three mechanisms are essential to obtain adaptive resistance in a bacterial population: 1) intrinsic variability in the expression of the EPRN transcription factors; 2) epigenetic inheritance of the transcription rate of EPRN associated genes; and 3) energetic cost of the efflux pumps activity that slows down cell growth. While the first two mechanisms acting together are responsible for the emergence and gradual increase of the resistance, the third one accounts for its reversibility. In contrast with the standard assumption, our model predicts that adaptive resistance cannot be explained by increased mutation rates. Our results identify the molecular mechanism of epigenetic inheritance as the main target for therapeutic treatments against the emergence of adaptive resistance. Finally, our theoretical framework unifies known and newly identified determinants such as the burden of efflux pumps that underlie bacterial adaptive resistance to antibiotics.

No MeSH data available.


Related in: MedlinePlus

Reversibility of the resistance phenotype.(A) This tracking plot shows that the expression of the activator increases while the antibiotic shocks are applied as in Fig. 2. Then, when the antibiotic is removed (indicated by the tilted black arrow), the expression of the activator decreases abruptly and eventually reaches its basal level. (B) Size of the population as a function of time for the same simulation as in A. After each antibiotic shock (small black arrows) the population size decreases exponentially and the recovery time becomes longer with each shock. After the antibiotic is removed (tilted black arrow) the population comes back again to its wild-type (WT) growth rate. To carry out the simulations in the antibiotic-free phase, every time the population reached the maximum size N = 5000, we took a random sample of 10% of the cells and made them grow without antibiotic, until the population reached again this maximum size, and so on. (C) Average transcription rate μβ = ⟨β0⟩ in the population as a function of time. Note that the average increases while the shocks are applied and then gradually comes back to small values when the antibiotic is removed. Error bars indicate the standard deviation. It can be observed that the standard deviation increases with the antibiotic stress. The panels below show the full distribution G (μβ, σβ) at three different times: before any antibiotic is introduced (circle); after several antibiotic shocks (star); after a long period of time without antibiotic (line). Time is measured generations, being one generation the time it takes for a cell with β0 = 1 to reach θF starting from F = 0.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4363326&req=5

pone.0118464.g003: Reversibility of the resistance phenotype.(A) This tracking plot shows that the expression of the activator increases while the antibiotic shocks are applied as in Fig. 2. Then, when the antibiotic is removed (indicated by the tilted black arrow), the expression of the activator decreases abruptly and eventually reaches its basal level. (B) Size of the population as a function of time for the same simulation as in A. After each antibiotic shock (small black arrows) the population size decreases exponentially and the recovery time becomes longer with each shock. After the antibiotic is removed (tilted black arrow) the population comes back again to its wild-type (WT) growth rate. To carry out the simulations in the antibiotic-free phase, every time the population reached the maximum size N = 5000, we took a random sample of 10% of the cells and made them grow without antibiotic, until the population reached again this maximum size, and so on. (C) Average transcription rate μβ = ⟨β0⟩ in the population as a function of time. Note that the average increases while the shocks are applied and then gradually comes back to small values when the antibiotic is removed. Error bars indicate the standard deviation. It can be observed that the standard deviation increases with the antibiotic stress. The panels below show the full distribution G (μβ, σβ) at three different times: before any antibiotic is introduced (circle); after several antibiotic shocks (star); after a long period of time without antibiotic (line). Time is measured generations, being one generation the time it takes for a cell with β0 = 1 to reach θF starting from F = 0.

Mentions: So far, the difference between genetic and epigenetic inheritance consists on one hand, in the time scales at which these two processes produce phenotypic changes in the population, and on the other hand in the parameters they affect. Genetic inheritance affects the pump efficiency εI whereas epigenetic inheritance affects the transcription rate β0. Another important difference is that changes caused by genetic mutations are very unlikely to be reversible whereas epigenetic changes are much more likely to be reversible [31, 38]. To test if our model can reproduce the experimentally observed reversibility, we replicate the simulations described in the previous section (where levels of resistance are ramped higher) for the mixed scenario. But now, after several rounds of selection, we remove the external antibiotic and allow the cells to grow and divide without stress. (The antibiotic is removed by setting Iext = 0 in Eq.(5) of the SI.) Fig. 3A shows that the concentration of the activator abruptly decreases when the antibiotic is removed (indicated by a big black arrow), eventually reaching its basal level of expression. Fig. 3B shows the size of the population as a function of time across the successive antibiotic shocks, and then in the free medium. The population size decreases exponentially each time an antibiotic shock is applied. However, after some time, the surviving cells grow and divide restoring the population size to the maximum size N = 5000. Note that when the antibiotic is removed the population growth returns to its wild-type behavior. These results are qualitatively similar to those observed experimentally [1, 3–8, 9]. However, this fact does not necessarily mean that the cells return to their wild type levels of susceptibility. For the cells that have reversed back to a sensitive phenotype could still have very high values of transcription rate, β0, and this high rate would imply that as soon as the antibiotic is applied again, the activator and the efflux pumps will be produced rapidly and at high concentrations. At this stage, most of the cells would be able to survive easily almost any antibiotic shock making the system non-reversible.


Adaptive resistance in bacteria requires epigenetic inheritance, genetic noise, and cost of efflux pumps.

Motta SS, Cluzel P, Aldana M - PLoS ONE (2015)

Reversibility of the resistance phenotype.(A) This tracking plot shows that the expression of the activator increases while the antibiotic shocks are applied as in Fig. 2. Then, when the antibiotic is removed (indicated by the tilted black arrow), the expression of the activator decreases abruptly and eventually reaches its basal level. (B) Size of the population as a function of time for the same simulation as in A. After each antibiotic shock (small black arrows) the population size decreases exponentially and the recovery time becomes longer with each shock. After the antibiotic is removed (tilted black arrow) the population comes back again to its wild-type (WT) growth rate. To carry out the simulations in the antibiotic-free phase, every time the population reached the maximum size N = 5000, we took a random sample of 10% of the cells and made them grow without antibiotic, until the population reached again this maximum size, and so on. (C) Average transcription rate μβ = ⟨β0⟩ in the population as a function of time. Note that the average increases while the shocks are applied and then gradually comes back to small values when the antibiotic is removed. Error bars indicate the standard deviation. It can be observed that the standard deviation increases with the antibiotic stress. The panels below show the full distribution G (μβ, σβ) at three different times: before any antibiotic is introduced (circle); after several antibiotic shocks (star); after a long period of time without antibiotic (line). Time is measured generations, being one generation the time it takes for a cell with β0 = 1 to reach θF starting from F = 0.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4363326&req=5

pone.0118464.g003: Reversibility of the resistance phenotype.(A) This tracking plot shows that the expression of the activator increases while the antibiotic shocks are applied as in Fig. 2. Then, when the antibiotic is removed (indicated by the tilted black arrow), the expression of the activator decreases abruptly and eventually reaches its basal level. (B) Size of the population as a function of time for the same simulation as in A. After each antibiotic shock (small black arrows) the population size decreases exponentially and the recovery time becomes longer with each shock. After the antibiotic is removed (tilted black arrow) the population comes back again to its wild-type (WT) growth rate. To carry out the simulations in the antibiotic-free phase, every time the population reached the maximum size N = 5000, we took a random sample of 10% of the cells and made them grow without antibiotic, until the population reached again this maximum size, and so on. (C) Average transcription rate μβ = ⟨β0⟩ in the population as a function of time. Note that the average increases while the shocks are applied and then gradually comes back to small values when the antibiotic is removed. Error bars indicate the standard deviation. It can be observed that the standard deviation increases with the antibiotic stress. The panels below show the full distribution G (μβ, σβ) at three different times: before any antibiotic is introduced (circle); after several antibiotic shocks (star); after a long period of time without antibiotic (line). Time is measured generations, being one generation the time it takes for a cell with β0 = 1 to reach θF starting from F = 0.
Mentions: So far, the difference between genetic and epigenetic inheritance consists on one hand, in the time scales at which these two processes produce phenotypic changes in the population, and on the other hand in the parameters they affect. Genetic inheritance affects the pump efficiency εI whereas epigenetic inheritance affects the transcription rate β0. Another important difference is that changes caused by genetic mutations are very unlikely to be reversible whereas epigenetic changes are much more likely to be reversible [31, 38]. To test if our model can reproduce the experimentally observed reversibility, we replicate the simulations described in the previous section (where levels of resistance are ramped higher) for the mixed scenario. But now, after several rounds of selection, we remove the external antibiotic and allow the cells to grow and divide without stress. (The antibiotic is removed by setting Iext = 0 in Eq.(5) of the SI.) Fig. 3A shows that the concentration of the activator abruptly decreases when the antibiotic is removed (indicated by a big black arrow), eventually reaching its basal level of expression. Fig. 3B shows the size of the population as a function of time across the successive antibiotic shocks, and then in the free medium. The population size decreases exponentially each time an antibiotic shock is applied. However, after some time, the surviving cells grow and divide restoring the population size to the maximum size N = 5000. Note that when the antibiotic is removed the population growth returns to its wild-type behavior. These results are qualitatively similar to those observed experimentally [1, 3–8, 9]. However, this fact does not necessarily mean that the cells return to their wild type levels of susceptibility. For the cells that have reversed back to a sensitive phenotype could still have very high values of transcription rate, β0, and this high rate would imply that as soon as the antibiotic is applied again, the activator and the efflux pumps will be produced rapidly and at high concentrations. At this stage, most of the cells would be able to survive easily almost any antibiotic shock making the system non-reversible.

Bottom Line: Here, using an efflux pump regulatory network (EPRN) model, we show that the following three mechanisms are essential to obtain adaptive resistance in a bacterial population: 1) intrinsic variability in the expression of the EPRN transcription factors; 2) epigenetic inheritance of the transcription rate of EPRN associated genes; and 3) energetic cost of the efflux pumps activity that slows down cell growth.Our results identify the molecular mechanism of epigenetic inheritance as the main target for therapeutic treatments against the emergence of adaptive resistance.Finally, our theoretical framework unifies known and newly identified determinants such as the burden of efflux pumps that underlie bacterial adaptive resistance to antibiotics.

View Article: PubMed Central - PubMed

Affiliation: Instituto de Ciencias Físicas, UNAM, Cuernavaca, Morelos, Mexico.

ABSTRACT
Adaptive resistance emerges when populations of bacteria are subjected to gradual increases of antibiotics. It is characterized by a rapid emergence of resistance and fast reversibility to the non-resistant phenotype when the antibiotic is removed from the medium. Recent work shows that adaptive resistance requires epigenetic inheritance and heterogeneity of gene expression patterns that are, in particular, associated with the production of porins and efflux pumps. However, the precise mechanisms by which inheritance and variability govern adaptive resistance, and what processes cause its reversibility remain unclear. Here, using an efflux pump regulatory network (EPRN) model, we show that the following three mechanisms are essential to obtain adaptive resistance in a bacterial population: 1) intrinsic variability in the expression of the EPRN transcription factors; 2) epigenetic inheritance of the transcription rate of EPRN associated genes; and 3) energetic cost of the efflux pumps activity that slows down cell growth. While the first two mechanisms acting together are responsible for the emergence and gradual increase of the resistance, the third one accounts for its reversibility. In contrast with the standard assumption, our model predicts that adaptive resistance cannot be explained by increased mutation rates. Our results identify the molecular mechanism of epigenetic inheritance as the main target for therapeutic treatments against the emergence of adaptive resistance. Finally, our theoretical framework unifies known and newly identified determinants such as the burden of efflux pumps that underlie bacterial adaptive resistance to antibiotics.

No MeSH data available.


Related in: MedlinePlus