Penicillin: The Mold That Won a War

A leaky window in St. Mary’s

In the summer of 1928, Alexander Fleming went on holiday and left some petri dishes stacked on a bench in his Paddington lab. The window was open. Spores drifted in from a downstairs mycology lab. One of them landed on a plate of Staphylococcus aureus — a common skin bacterium and a known killer.

When Fleming came back two weeks later, the plate was mostly green. A Penicillium notatum mold had grown across it. But around each spot of mold, there was a perfectly clean ring where the staphylococci had been wiped out.

He published the observation in 1929. Nobody much cared.

For the next twelve years, the discovery sat largely forgotten. Fleming couldn’t purify the active compound. The yields were tiny. He had no biochemistry team. He had no funding. He moved on to other work.

Then a war started.

Oxford, 1940

By 1940, German aircraft were bombing London nightly and an Australian named Howard Florey, working with Ernst Chain at the Sir William Dunn School of Pathology in Oxford, decided to try Fleming’s mold again. They suspected that whatever it was killing bacteria with might be the most important molecule in medicine — if anyone could isolate it.

They could.

Their first patient was Albert Alexander, an Oxford policeman in his mid-40s. He had developed a runaway infection — wartime accounts blamed a rose-thorn scratch, but more recent historical work points to shrapnel wounds from a German bombing raid on his police station the previous November. Whatever the entry point, the infection ate through his scalp and face, into his eye and lungs. He was dying. On 12 February 1941, Florey’s team injected him with their entire stockpile of purified penicillin.

He turned the corner within 24 hours. Fever broke. Wounds healed. He was speaking, eating.

Then they ran out.

To stretch the supply, Florey’s team had been recovering penicillin from Alexander’s urine each morning and re-injecting it. Even so, the supply gave out before the infection was fully cleared. The bacteria came back. He died on 15 March 1941.

Albert Alexander is the reason every modern intensive-care unit holds antibiotics in stock by the kilogram. He proved the molecule worked. The only failure was scale.

A mold on a melon

Florey’s team needed industrial production. The British pharmaceutical industry was busy making mustard gas antidotes. So in July 1941 Florey flew to Peoria, Illinois, where the U.S. Department of Agriculture ran a lab specializing in industrial fermentation.

The Peoria team tried thousands of Penicillium strains from around the world. The winner came from a moldy cantaloupe a lab assistant named Mary Hunt — known to colleagues as “Moldy Mary” — picked up in a Peoria market. That single strain, NRRL 1951, produced more than 100 times the penicillin yield of Fleming’s original.

Every penicillin tablet, capsule, and IV bag in the world today descends from a mold on a fruit Mary Hunt bought because it looked unappetizing.

By D-Day, 6 June 1944, Allied production was 2.3 million doses per month. Lives saved on the Normandy beaches alone are estimated in the tens of thousands. By war’s end, every Allied field hospital had penicillin. German hospitals did not.

This is the molecule that won the war as much as any other single thing.

How it kills

Bacteria look fragile — single cells, no skeleton, suspended in fluid — but they aren’t. They’re held together by a wire mesh called peptidoglycan, a tough polymer lining the outside of every bacterial cell. Peptidoglycan is what stops them from bursting in their own internal pressure, which is high enough to inflate a soccer ball.

Building peptidoglycan requires an enzyme called DD-transpeptidase. The enzyme cross-links the mesh, stitching short peptide chains together every few nanometers. Without it, the wall is loose and the cell pops.

The active piece of penicillin is a small four-membered ring called the β-lactam. Its shape is almost exactly the same as the natural substrate the DD-transpeptidase enzyme grabs from the cell to cross-link the mesh.

When penicillin floats by, the enzyme grabs it instead. But once the enzyme’s active site latches onto the β-lactam, the ring snaps open and bonds covalently to the enzyme’s serine residue. The enzyme is now permanently disabled — it can never let go.

A bacterium that needs to divide has no working cross-linker. Its new wall sections are weak. It absorbs water, swells, ruptures.

We don’t have peptidoglycan. We have no β-lactam target. Penicillin slides through us doing nothing. From a side-effect perspective, it’s almost the ideal drug: enormously toxic to bacteria, almost invisible to us.

The exception is allergic reaction. About 1% of people develop antibodies against the β-lactam ring itself; in a few of them, the reaction is severe. But the molecule’s pharmacology is otherwise close to magic.

The arms race we are losing

Bacteria evolve.

The most common defense they evolve is β-lactamase — an enzyme that cuts the β-lactam ring before it can react with DD-transpeptidase. Penicillin-resistant Staphylococcus aureus was reported within four years of widespread human use. By 1960, hospitals were already dealing with strains penicillin couldn’t touch.

The response was chemistry: modify the molecule. Methicillin (1959) was a β-lactamase-resistant penicillin. Bacteria responded with MRSA — methicillin-resistant Staphylococcus aureus. Vancomycin worked, until VRSA — vancomycin-resistant S. aureus — appeared in 2002. Carbapenems worked, until carbapenem-resistant Enterobacteriaceae — CREs, with mortality rates approaching 50% — became routine in some hospitals.

Each new antibiotic class buys us, on average, around ten years before resistance is widespread. The pipeline of new classes has been nearly empty since the 1980s, because antibiotics are not a profitable target for pharma — they get used briefly, then resistance kills the market.

This is why the World Health Organization, the CDC, and every infectious-disease body on the planet now ranks antimicrobial resistance among the top global health threats. We are slowly losing the chemical war Florey started in 1941.

A statistic that almost no one tracks

Before penicillin, a strep throat could kill a child. A scratched knee could kill a farmer. Pneumonia killed about 30% of hospitalized cases. Childbirth carried a 1% maternal-death rate, mostly from infection. A compound fracture meant a 25% chance of dying from the resulting sepsis.

After penicillin, those numbers dropped to near zero — within a single human generation.

If you had to pick the single discovery that did the most to extend human lifespan, you’d be choosing between sanitation, vaccines, and antibiotics. Each contributed about 10 years to global life expectancy in the 20th century. Penicillin alone is responsible for a few of those years.

The catch is that lifespan gain doesn’t show up as a graph that goes up. It shows up as the absence of a death that would otherwise have happened. Your grandmother who survived her first appendicitis at 19. Your great-uncle who didn’t die from his cut hand in 1947. The numbers are large; the individual stories are invisible.

Penicillin’s defining feature is that you almost certainly owe at least one ancestor’s life to it. You just can’t tell which one.

1945

In December 1945, Fleming, Florey, and Chain shared the Nobel Prize in Physiology or Medicine. In his acceptance lecture, Fleming made a prediction that turned out to be exact:

“The time may come when penicillin can be bought by anyone in the shops. Then there is the danger that the ignorant man may easily under-dose himself and by exposing his microbes to non-lethal quantities of the drug, make them resistant.”

He was talking about us. Eighty years later, that resistance is now a measurable cause of about 1.3 million deaths a year worldwide.

The molecule still works. We’re just running out of ways to give it long enough to do its job.

---

The penicillin G molecule has its own 3D page in the library — rotate it, find the small four-membered β-lactam ring that opens DD-transpeptidase like a key, download the structure file. It also has its own bouncing wallpaper.