Older Than the Drugs: How Bacteria Evolve Antibiotic Resistance

When Alexander Fleming accepted the Nobel Prize in 1945, he used part of his lecture to warn that bacteria could be trained to resist penicillin in the laboratory, and that careless use of the drug would do the same thing in the world. He was right within a few years. Resistant staphylococci were common in hospitals before the decade was out. The pattern has repeated for every antibiotic since: introduction, a few good years, then the steady return of infections the drug no longer touches. It is worth being clear about why this happens, because the reasons are evolutionary rather than accidental, and they explain why the problem cannot be engineered away with one clever molecule.

Resistance is not invented, it is selected

The common picture of resistance has bacteria “learning” to survive a drug, as if exposure teaches them. That gets the biology backwards. In any large bacterial population there is already genetic variation, thrown up by mutation and shuffled by other processes. Most of those variants are useless or harmful. A few happen to blunt the effect of a given antibiotic. When the drug arrives, it kills the susceptible cells and spares the rare resistant ones, which then multiply into the space left behind. The drug does not create resistance. It selects for it. A bacterial population can run through enough generations in a single day to do in a hospital ward what would take a vertebrate species millennia.

That resistance predates the clinic is not a guess. In 2011 a team led by Vanessa D’Costa recovered DNA from 30,000-year-old permafrost in the Canadian Arctic and found genes encoding resistance to beta-lactams, tetracyclines, and glycopeptides, including a version of the gene that today defends bacteria against vancomycin (D’Costa et al., 2011). The microbes that make antibiotics, mostly soil bacteria and fungi, have been producing these compounds for hundreds of millions of years, and their neighbors have been defending against them for just as long. The collection of resistance genes in the environment, sometimes called the resistome, is ancient. Human medicine simply gave those genes a reason to spread.

Four ways to survive a poison

The molecular tactics fall into a handful of categories, and knowing them makes the clinical news easier to read (Davies and Davies, 2010).

The most direct is to destroy the drug. Beta-lactamases are enzymes that cut the chemical ring at the heart of penicillins and related antibiotics before the drug can act. There are now thousands of variants, and the extended-spectrum and carbapenemase versions chew through some of the drugs that were held in reserve for exactly this situation.

A second tactic is to change the target so the drug no longer fits. Many antibiotics work by jamming a specific bacterial machine, a particular enzyme or a piece of the ribosome. A small change to that machine can keep it working while spoiling the drug’s grip. The classic example, discussed below, is the altered penicillin-binding protein that defines methicillin-resistant Staphylococcus aureus.

The third tactic is to pump the drug out. Efflux pumps are membrane proteins that recognize an antibiotic and expel it before it accumulates to a lethal level. Because some pumps handle many different compounds, a single one can grant low-level resistance to several drug classes at once.

The fourth is to keep the drug out in the first place, by reducing the permeability of the cell envelope. On its own this rarely gives full resistance, but combined with an efflux pump or a modifying enzyme it can push a borderline strain over the line.

Sharing the answer: horizontal gene transfer

If bacteria could only pass resistance to their own descendants, the problem would still be serious but slow. What makes it fast is that bacteria swap genes sideways, between unrelated cells and even across species, through horizontal gene transfer. There are three main routes. In conjugation, one cell builds a bridge to another and copies across a plasmid, a small loop of DNA that often carries several resistance genes together. In transformation, a cell takes up loose DNA from its surroundings. In transduction, a virus that infects bacteria accidentally ferries bacterial genes from one host to the next.

Conjugation does most of the clinical damage, partly because the plasmids involved frequently bundle resistance to multiple drugs onto one transferable unit, and partly because stress, including antibiotic exposure, can switch transfer on. Mobile elements called transposons and integrons act as cassettes that collect resistance genes and slot them into plasmids and chromosomes, so a gene that arose in a harmless soil bacterium can end up, a few transfers later, in a serious human pathogen. This is why resistance to a brand-new drug can appear in a clinic that has never used it. The gene was already circulating somewhere in the microbial world and only needed a vehicle.

A worked example: methicillin-resistant Staph

Staphylococcus aureus shows the whole process in fast motion (Chambers and DeLeo, 2009). Penicillin worked beautifully against it in the 1940s, then penicillin-destroying enzymes spread on plasmids and undid that within a decade. Chemists answered in 1959 with methicillin, a penicillin redesigned to resist those enzymes. Resistant isolates turned up in 1960, the year of its introduction, almost immediately.

The trick this time was not a drug-destroying enzyme but a replacement part. MRSA carries a gene called mecA, sitting on a mobile chromosomal cassette known as SCCmec that the bacterium acquired by horizontal transfer. The gene encodes an alternative penicillin-binding protein, PBP2a, which does the cell-wall-building job that beta-lactam antibiotics normally block, but with such low affinity for those drugs that they cannot stop it. Because nearly all beta-lactams attack the same machinery, one acquired gene confers resistance to almost the entire class at once. MRSA spent decades as a mostly hospital problem, then in the 1990s new community-associated strains appeared in people with no hospital contact, carrying smaller, more mobile cassettes. The pathogen kept finding new versions of the same answer.

Why the arms race is hard to end

There is a hopeful-sounding idea that resistance carries a cost, so that removing a drug should let susceptible strains outcompete resistant ones and restore the drug’s power. Sometimes this happens. Often it does not, and the reason is itself evolutionary. Resistance can impose a fitness penalty at first, a slightly slower growth rate from running an efflux pump or making an extra enzyme. But bacteria acquire compensatory mutations that offset the cost without giving up the resistance, so the resistant strain becomes nearly as fit as its ancestor and has little reason to disappear when the drug is withdrawn (Andersson and Hughes, 2010). The genie does not go back in the bottle on its own.

The scale of the result is now measurable. A systematic analysis estimated that in 2019 bacterial antimicrobial resistance was directly responsible for about 1.27 million deaths worldwide and was a contributing factor in nearly 5 million more, which places it among the leading causes of death globally (Murray et al., 2022). Those numbers come not from a failure of chemistry but from the ordinary working of natural selection on organisms that reproduce in minutes and trade genes freely.

None of this means the situation is hopeless. It means the countermeasures have to respect the biology. Using antibiotics sparingly slows the selection. Combining drugs makes simultaneous resistance less likely. Surveillance catches new resistance genes while they are rare. What the history rules out is the comfortable hope of a permanent fix. Bacteria were solving the antibiotic problem long before we posed it, and they will keep solving each new version of it. The realistic goal is to stay a step behind for as short a time as possible.