Dear Science,

An MD I know was telling me about antibiotic resistance, and how antibiotics only last a couple of years before they are no longer effective. I was wondering: If bacteria become resistant to an antibiotic, when we stop using an antibiotic for a decade or so, will the bacteria "forget" the resistance? Could we go back to antibiotics from the '60s or '70s and treat our antibiotic-resistant diseases of today?

Future Nobel Prize Winner

Your MD friend doesn't have it quite right. Most antibiotics last for decades; in fact, for most bacteria (aside from some key exceptions), just about any antibiotic will kill them off. Only in bacteria constantly bombarded by antibiotics does resistance stick around.

Antibiotics—at least the key chemicals at the head of all the variants we use in clinics—almost all started as natural products made by one kind of bacteria (or fungus) to kill off another, usually in the soil beneath our feet. The resistance genes likewise come from microorganisms that have been fighting these genetic cold wars for eons. But being resistant to antibiotics is a bitch—it wastes energy that could be better spent growing, dividing or digesting. For most bacteria, including pathogenic bacteria, being resistant is a significant selective disadvantage because it's so wasteful.

Bacteria have a few tricks to get around this. The resistance genes are often turned on only in the presence of antibiotics. Doctors can get around this by using a huge dose of antibiotics early; the bacteria die before the resistance genes can be turned on. The next trick is tougher: Bacteria can trade genes on little mobile bits of DNA called plasmids. In any given hospital, nursing home, or sickly patient, one can find plasmids ready to make antibiotic-resistance proteins in huge quantities. By gobbling up these plasmids, bacteria can quickly learn to be resistant to antibiotics. Just as quickly, bacteria can ditch the plasmid when the antibiotic goes away. This is how a given patient starts out with E. coli that can be killed off by anything and leaves the hospital with E. coli that can be killed only by the newest and toughest antibiotics.

For this sort of antibiotic resistance, your idea is right. If we stop using an antibiotic for a good amount of time, the bacteria ditch the resistance plasmids. Eventually, the plasmids themselves start to disappear.

Much more rarely, bacteria can learn to be resistant in their core genes to antibiotics. Some Staphylococcus aureusa nasty bacteria that causes skin infections (among other kinds of infections)—have changed the genes used to make its walls. Building the new walls doesn't cost much more than the old—and the new walls are impervious to our best antibiotics. For this sort of resistance, no amount of time not using antibiotics would help. We're stuck with these mutants.

Prescribingly yours,


Send your science questions to