Antibiotic resistance is commonly touted as an example of “evolution happening before our eyes.” But, what portion of antibiotic resistance is actually due to chance mutations? Are there aspects of antibiotic resistance that are due to programmed genetic variants initiated by the organism itself? Here, I’m going to share a key paper that tries to answer these questions.

In 2019, a team of scientists discovered a protein called Mfd that when deleted makes it much harder for bacteria to “evolve” antibiotic resistance. (Ragheb et al. 2019) Mfd is a DNA translocase which had been purported to be involved with nucleotide excision repair, but the study found that the absence of this one protein decreased the mutation rate by 2- to 5-fold at certain genome positions. This suggests that Mfd is directing genetic changes at those specific locations.

A Protein Involved with Generating Specific Mutations

To give more detail, in the first experiment from Ragheb et al., the Luria-Delbruck Fluctuation Analysis was used to calculate the mutation rate for B. subtilis, P. Aeruginosa, S. Typhimurium, and M. tuberculosis in the presence and absence of Mfd. What was observed is that strains without Mfd had a 2-5-fold decrease in the mutation rates as measured by resistance to rifampicin, which is an antibiotic. This is fascinating because it suggests that the bacterial protein, Mfd, is actually causing mutations! The mutation isn’t due to some chance hit by ionizing radiation. The bacteria are actually mutating themselves.

Now for some important caveats:

• Just because the mutation rate as measured by rifampicin resistance increases doesn’t necessarily mean the mutation rate for the whole genome is affected. Selecting by rifampicin resistance means the calculated mutation rate is based on a region comprising 3582 bps (“rpoB DNA-Directed RNA Polymerase Subunit Beta [Bacillus Stercoris]”) in comparison to the full size of the genome which is 4,214,810 base pairs (Kunst et al. 1997). Therefore, the remainder of the genome could not have an increase in mutation rate. (Spoiler alert: this turns out to be the case!)
• It’s also unclear whether the heightened mutation rate in rpoB is a background process or if it’s triggered in response to antibiotic sensing. The Luria-Delbruck Fluctuation Analysis, famous for its role in distinguishing between such scenarios, can differentiate between these hypotheses, yet the text does not explicitly state which outcome was observed. For those not familiar with this experiment, the Luria-Delbruck Fluctuation Analysis has been interpreted as supportive of the Darwinian hypothesis, which suggests mutations occur randomly, with advantageous mutations naturally selected. This is as opposed to the Lamarckian hypothesis, which suggests mutations result from an organismal response to selective pressures.

The authors interpreted the absence of Mfd leading to a 2-5-fold reduction in mutation rates as modest, but they suspected that these differences might impact the kinetics (rate of resistance) and evolution of resistance. Thus, they looked at how resistance to antibiotics (rifampicin, phosphomycin, trimethoprim, kanamycin, and vancomycin), could evolve by starting off with lower concentrations of antibiotics and building up over 35-70 generations. These experiments showed a 6- to 21-fold difference for S. typhimurium and for B. subtilis. There was a 32-fold difference in the median antibiotic concentration that could be tolerated between strains containing and not containing Mfd. Mfd enables the bacteria to more quickly develop resistance to higher levels of a spectrum of antibiotics, all of which have different resistance targets in the genome. 

Another important observation is that when antibiotics were increased gradually over time, compared to the Luria-Delbruck Fluctuation Analysis where antibiotics were static and present at lethal levels to unprepared bacteria, a more pronounced difference between strains with and without Mfd was observed. The observation that a gradual accumulation resulted in a greater difference suggests that the bacteria are possibly sensing toxins and responding appropriately which there is not time for during the Luria-Delbruck analysis due to how this experiment is conducted.

How Is Mfd Generating Specific Mutations?

The authors don’t exactly know how Mfd is generating specific mutations or if it is doing so directly, but they used Sanger sequencing to identify genetic variants which occurred during this time course within the rifampicin and trimethoprim resistance targets (rpoB and folA respectively). Analyzing the sequences from every time point revealed genetic changes that enabled resistance. Without Mfd only about a half to a third of the resistance changes were observed, and there was also a delay in the changes arising without Mfd. Adding to the evidence of Mfd’s key role in developing specific resistance changes, second and third genetic changes were rarely observed in the respective targets without Mfd. Taken together this means that Mfd has some promutagenic function that might be critical for the generation of additional genetic changes.

Then to see if there were mutations outside of these regions the authors performed whole genome sequencing (WGS) on six randomly chosen replicates from the time course experiment. The WGS confirmed the Sanger sequencing results: without Mfd there were many fewer mutations in the rpoB locus. Importantly they did not find additional mutations outside of rpoB for any of the evolved rifampicin resistance strains. This is a critical finding, which strongly suggests that this mechanism is NOT introducing nonspecific mutations across the genome, but making specific mutations in known antibiotic targets to confer resistance. 

Accordingly, the results from WGS for the strains with trimethoprim resistance were different. They found mutations outside the coding region of folA, in its promoter region, as well as a recurring mutation in the dnaQ gene which generated a hypermutator phenotype:

We found that three out of six WT sequenced trimethoprim-evolved strains contained a point mutation in the dnaQ gene (all strains had the same dnaQ(I33N) mutation) none of the Δmfd strains contained any mutations in the dnaQ gene.

Consistent with the hypermutator mutation they found 600 accumulated mutations across the genome, which they note will need to be studied to see if they are playing a functional role or are just hitchhiker mutations. The observation that a very specific dnaQ mutation was generated multiple times independently suggests that the Mfd protein plays a critical role in which specific regions of the genome will change in response to the presence of toxins (antibiotics).

An Icon of Evolution

Antibiotic resistance is an icon of observable evolution; however, the extent to which chance mutations underpin the phenomenon remains uncertain. Darwinian evolution hinges upon random events, such as mutations induced by ionizing radiation or mistakes from DNA copying machinery. Yet, the pivotal findings of this investigation highlight that a substantial fraction of antibiotic resistance stems not from stochastic mutations but rather from bacterial pre-programmed mutation-inducing mechanisms, facilitated by the protein Mfd, which plays a strategic role in directing the specific genomic regions where changes will occur. This in turn confers antibiotic resistance. 

This demonstrates that directed mutagenesis is critically timed to antibiotic resistance and thereby calls into question the common portrayal of antibiotic resistance as an example of Darwinian evolution unfolding in real time. For now, the burden of proof has increased for those that rely on any chance-based theory of origins, because now they must explain the emergence of this programmed mechanism. And they have lost some of the mutations they were relying on to do so, as this research shows that much of what they’ve claimed were truly random mutations are now clearly occurring under programmed control.

Accordingly, these findings are more consistent with the theory of intelligent design. ID predicts that things that look like evolution are actually the result of design mechanisms built into organisms — in this case, mechanisms that harness stochasticity toward a functional end, rather than the chance-driven process of Darwinian evolution.

References

• Kunst, F., N. Ogasawara, I. Moszer, A. M. Albertini, G. Alloni, V. Azevedo, M. G. Bertero, et al. 1997. “The Complete Genome Sequence of the Gram-Positive Bacterium Bacillus Subtilis.” Nature 390 (6657): 249–56.
• Ragheb, Mark N., Maureen K. Thomason, Chris Hsu, Patrick Nugent, John Gage, Ariana N. Samadpour, Ankunda Kariisa, et al. 2019. “Inhibiting the Evolution of Antibiotic Resistance.” Molecular Cell 73 (1): 157–65.e5.
• “rpoB DNA-Directed RNA Polymerase Subunit Beta [Bacillus Stercoris].” Accessed April 13, 2024. https://www.ncbi.nlm.nih.gov/gene/83883335.