The Penicillium Mould and the Science of Penicillin

Table of Contents

1. Penicillium: A Common Mould
2. Fleming’s Contaminated Plate of 1928
3. P. rubrum, P. notatum, P. chrysogenum — and P. rubens
4. What the Genes of Fleming’s Mould Show
5. The Beta-Lactam Ring
6. How Penicillin Breaks the Bacterial Wall
7. A Structure Argument Settled by X-Rays
8. Penicillinase: An Enzyme That Destroys Penicillin
9. Better Moulds: Strain Variation and the Cantaloupe Story
10. The Penicillium Genome
11. Key Research Papers
12. Connections

1. Penicillium: A Common Mould

Penicillium is a large genus of moulds — microscopic fungi that grow as fine threads and reproduce by releasing enormous numbers of dry spores into the air. The name comes from the Latin penicillus, a little brush or painter’s brush, because under the microscope the spore-bearing stalks branch at the tip into a tuft that looks like the bristles of a brush. To the naked eye the colonies are usually the familiar blue-green or grey-green fuzz seen on forgotten bread, fruit and other food.

These moulds are everywhere people live. The Dutch and Danish mycologists Jos Houbraken, Jens Frisvad and Robert Samson opened their 2011 study by describing Penicillium chrysogenum as “a commonly occurring mould in indoor environments and foods”. When a London team led by Daniel Henk sampled the air and dust in and around St Mary’s Hospital — the building where Fleming made his discovery — they found penicillin-family moulds still living there, more than eighty years later. The site’s page on mould and mycotoxins describes indoor moulds more generally, including the other, harmful substances some moulds make.

Why a mould would make an antibacterial

Moulds make many chemicals that are not needed for their own basic growth. Microbiologists usually explain compounds such as penicillin as weapons in competition: a mould growing on a crumb of food or a patch of soil shares that space with bacteria and with other fungi, and a substance that holds rivals back gives it room. This is an explanation, not something any single experiment has proved. Henk’s team did see one form of it in the laboratory: in their growth studies, P. chrysogenum significantly inhibited the growth of the other common species of the group. The same idea — looking in nature’s microbes for the chemicals they use against each other — drove later discoveries such as streptomycin from soil bacteria and the avermectins described on the site’s Satoshi Ōmura natural-products page.

Back to Table of Contents

2. Fleming’s Contaminated Plate of 1928

In 1928, at St Mary’s Hospital in London, the bacteriologist Alexander Fleming found that a mould had contaminated one of his culture plates of staphylococci. Around the mould colony the bacteria had failed to grow or had broken down, leaving a clear zone. Fleming grew the mould in liquid broth and found that the broth itself, with the mould filtered off, could stop the growth of several kinds of bacteria. He called this active “mould broth filtrate” penicillin, after the mould.

He published his findings in 1929 in the British Journal of Experimental Pathology (volume 10, pages 226–236), in a paper reprinted in 2001 by the Bulletin of the World Health Organization. Its title shows how Fleming himself first used the substance: “On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae”. Because penicillin held back many common bacteria but not the organism then called Bacillus influenzae, he could add it to culture plates to let that organism grow on its own — a laboratory tool. He reported that the filtrate killed sensitive bacteria and did not appear toxic to animals, and suggested it might be useful applied to infected areas, but the active substance itself was not isolated, and it was unstable and hard to concentrate.

The substance stayed a curiosity of the bacteriology laboratory for about a decade, until Florey and Chain chose it in 1938–39 for their survey of natural antibacterial substances. Their first report, in the Lancet of 24 August 1940, showed that a purified penicillin preparation protected mice from lethal infections. That story is told on The Oxford Penicillin Work, 1938–1941.

Back to Table of Contents

3. P. rubrum, P. notatum, P. chrysogenum — and P. rubens

Few famous organisms have changed names as often as Fleming’s mould. Fungi of this group look very alike, and the rules for telling species apart changed several times during the twentieth century. The name on a label therefore depended on which reference book the identifier used.

P. rubrum (1929)

Fleming’s 1929 paper named the mould Penicillium rubrum. For decades textbooks said that Fleming, not being a mycologist, had simply got the name wrong. Houbraken, Frisvad and Samson re-examined that claim in 2011. They point out that the identification was actually made by the mycologist Charles J. La Touche, who worked on the floor below Fleming, and that La Touche followed the standard monograph of the day, by the Belgian mycologist Philibert Biourge. Under Biourge’s scheme, they conclude, calling the mould P. rubrum was accurate at the time and not the mistake later writers described. They also note that the epithets rubrum (“red”) and rubens (“being red”) are easily confused.

P. notatum (1930)

The American mycologist Charles Thom received a culture of Fleming’s strain as he was finishing his own monograph on Penicillium in 1930, and re-identified it as Penicillium notatum, a species described earlier by the Swedish botanist Richard Westling. That is the name used throughout the Oxford work and the wartime production effort: the Oxford team grew “P. notatum”, and the Peoria laboratory in Illinois titled its 1944 survey of strains “Natural Variation and Penicillin Production in Penicillium notatum and Allied Species”.

P. chrysogenum

Later taxonomists judged P. notatum to be the same species as Penicillium chrysogenum, an older name, and placed the two in synonymy. From then on Fleming’s mould and the industrial penicillin moulds were usually called P. chrysogenum, and that is the name on the 2008 genome paper.

P. rubens (2011)

DNA changed the picture again. When Houbraken’s group compared the gene sequences of the most important penicillin-producing isolates, they found two clearly separate groups (clades), and showed that these were two species: P. chrysogenum in the strict sense, and Penicillium rubens, a species first named by Biourge in 1923. The two look much alike, but they make different side chemicals — P. chrysogenum produces secalonic acids D and F and/or a substance related to lumpidin, and P. rubens does not. Fleming’s original strain fell in the P. rubens group. So did the industrial strain whose genome had been sequenced, the wild “Wisconsin” strain from a mouldy cantaloupe that was its ancestor, and the strain first found to give satisfactory yields when grown submerged in liquid. In the authors’ words, Fleming’s strain “is not Penicillium chrysogenum but P. rubens”.

Back to Table of Contents

4. What the Genes of Fleming’s Mould Show

The same year, 2011, a separate team based at Imperial College’s St Mary’s campus — the hospital where Fleming had worked — published a population-genetics study of what it called “Alexander Fleming’s lucky fungus”. Daniel Henk, Matthew Fisher and their colleagues asked whether P. chrysogenum, long assumed to be a single species found all over the world, really was one species.

Their answer was no. Their analyses showed that the apparently ubiquitous mould is made up of at least two genetically distinct species, with only slight differences in how they grow. They found both species in air and dust samples collected in and around St Mary’s. They genotyped 30 genetic markers spread across the genome, and found that preserved fungal material from Fleming’s laboratory was nearly identical to strains derived from it in culture collections today — and that it belonged to the same species as the wild ancestor of the present-day industrial penicillin strains, not to the type species P. chrysogenum.

The study turned up two further surprises. First, the mould had long been regarded as asexual, but samples from around the world carried both mating-type genes in a near one-to-one ratio and showed signs of genetic recombination, the mark of a sexual cycle. Second, although the two species live side by side across the globe and diverged less than a million years ago, the team detected no hybrids between them. Their growth experiments suggested a reason: P. chrysogenum significantly inhibited the other species, and the authors proposed that this kind of competition may help keep the two species separate. Read together, the Henk and Houbraken studies agree on the practical point: the mould on Fleming’s plate and the moulds of the penicillin industry belong to the same lineage, now called P. rubens.

Back to Table of Contents

5. The Beta-Lactam Ring

The molecule the mould makes is small. Its working part is a compact chemical ring of four atoms — three carbon atoms and one nitrogen — called the beta-lactam ring. In penicillin this square ring is fused to a second, five-membered ring containing sulphur (a thiazolidine ring), and a side chain is attached to the pair. The side chain is what differs from one penicillin to another; the fused ring system is shared.

A four-membered ring is strained: its bonds are forced into angles much tighter than those atoms normally adopt. That strain makes the ring chemically reactive, and its reactivity is the source of penicillin’s power. As Maryna Lobanovska and Giulia Pilla put it in their 2017 review in the Yale Journal of Biology and Medicine, penicillin, like the other beta-lactam antibiotics, “contains a four-membered beta-lactam ring … which is responsible for the inhibition of transpeptidase”. The same strain also makes the ring fragile, which helps explain why Fleming and the Oxford team found penicillin so unstable in solution, and why a bacterial enzyme that breaks the ring can disarm it (section 8).

One name, many molecules

The penicillin first extracted from the mould and purified at Oxford and in wartime America is the molecule now called benzylpenicillin, or penicillin G. Today, Lobanovska and Pilla note, the word “penicillin” is used generically for a family of molecules built on the beta-lactam structure with the same kind of antibacterial activity, classified by the chemical groups attached to the ring. The same core also appears in other antibiotic families discovered later, such as the cephalosporins, carbapenems and monobactams.

Back to Table of Contents

6. How Penicillin Breaks the Bacterial Wall

Most bacteria are wrapped in a cell wall made of peptidoglycan, a mesh of long sugar chains tied together by short chains of amino acids. Lobanovska and Pilla describe its main job as preserving the cell’s integrity and shape. The wall works like the tough outer casing of a tyre: inside it, the bacterium’s contents press outward, and without a strong, continuously maintained wall the cell would swell and burst.

The mesh only becomes strong once its strands are cross-linked. The cross-links are made by enzymes called transpeptidases, also known as penicillin-binding proteins (PBPs). A transpeptidase grips the last two units of a short peptide chain — two molecules of the amino acid D-alanine — and joins that chain to a neighbouring strand.

Penicillin works by imitation. Its beta-lactam ring is shaped enough like the end of that D-alanine pair that the transpeptidase takes hold of it instead. When it does, the strained ring springs open and the enzyme’s active site becomes chemically bonded to the penicillin, so the enzyme can no longer build cross-links. In Lobanovska and Pilla’s words, penicillin is able to “bind irreversibly the active site of the transpeptidase, preventing the enzyme from cross-linking the peptidoglycan strands”. A growing bacterium keeps adding new wall material that can no longer be tied together, the wall weakens, and the cell breaks open.

Why it spares human cells, and why some bacteria escape

Human and animal cells have no peptidoglycan wall and no transpeptidases of this kind, which is why penicillin can kill bacteria while leaving the body’s own cells largely untouched — the property Fleming first noticed and the Oxford mouse tests confirmed. The mechanism also explains its limits. Penicillin acts on bacteria that are actively building wall. And, as the 2017 review explains, many Gram-negative bacteria carry an extra outer membrane that acts as a selective barrier and blocks penicillin from reaching its target, which is why the original penicillin worked best against Gram-positive organisms such as streptococci and staphylococci. The site’s pages on penicillin and strep A and the antibiotic resistance animation show these ideas in practice.

Back to Table of Contents

7. A Structure Argument Settled by X-Rays

For several years after it became a medicine, nobody knew for certain what penicillin looked like as a molecule. Chemists could measure which atoms it contained, but not how they were joined. During the war, chemists in Britain and America worked on the question in parallel, and two rival structures emerged.

One side favoured the beta-lactam structure: the strained four-membered ring fused to a sulphur-containing ring. According to the Science History Institute’s biography of Florey and Chain, this structure was proposed by Chain and Abraham and also by the American chemist R. B. Woodward; Lobanovska and Pilla write that Abraham first proposed it in 1943. The other side, led by the eminent Oxford organic chemist Robert Robinson, favoured an oxazolone structure, which had no four-membered ring. Both camps had chemical evidence they found persuasive, and the question could not be settled by the chemistry of the day.

It was settled by physics. Dorothy Crowfoot Hodgkin, an Oxford crystallographer, worked with Barbara Low to determine the structure of penicillin by X-ray crystallography — shining X-rays through crystals of penicillin salts and working back from the pattern of scattered rays to the positions of the atoms. By 1945 their analysis showed the beta-lactam ring. Hodgkin described the work in 1949 in “The X-ray analysis of the structure of penicillin”. It was one of the first times X-ray crystallography had decided the structure of a natural molecule that chemistry alone could not, and Hodgkin went on to solve the structure of vitamin B12, work for which she received the 1964 Nobel Prize in Chemistry.

Knowing the structure did not make penicillin easy to manufacture by chemistry. The Science History Institute notes that a complete chemical synthesis came only in 1957, and that fermentation — letting the mould make the molecule — remained the commercial route.

Back to Table of Contents

8. Penicillinase: An Enzyme That Destroys Penicillin

Bacteria began showing their counter-measures before penicillin had treated its first patient. In December 1940, only months after the mouse results appeared in the Lancet, Edward Abraham and Ernst Chain published a one-page letter in Nature titled “An enzyme from bacteria able to destroy penicillin”. They had found that an extract of a strain of Escherichia coli inactivated penicillin. They named the enzyme penicillinase.

Penicillinases belong to the family of enzymes now called beta-lactamases. As the name says, they attack the beta-lactam ring: they break it open by hydrolysis (adding water across the strained bond), and once the ring is open the molecule can no longer bind the transpeptidase. Lobanovska and Pilla describe this as one of the two main reasons many Gram-negative bacteria resist penicillin, the other being their outer membrane.

At first penicillinase looked like a laboratory finding about bacteria that penicillin did not treat anyway. It became much more than that. After penicillin came into wide use, the 2017 review notes, penicillinase-producing strains also emerged among Gram-positive species, above all Staphylococcus aureus, where penicillin-resistant strains were seen in hospital patients by 1942. That shift drove the search for semisynthetic penicillins built to resist the enzyme — methicillin, oxacillin and dicloxacillin among them — and later for beta-lactamase inhibitors. The longer story of resistance is told on Wartime Mass Production and the Antibiotic Era and on the site’s page on antibiotic resistance and MRSA.

Back to Table of Contents

9. Better Moulds: Strain Variation and the Cantaloupe Story

Fleming’s own strain was a poor producer by later standards. Once the work moved to the United States in 1941, the Northern Regional Research Laboratory of the US Department of Agriculture in Peoria, Illinois, began looking for better moulds as well as better ways of growing them. Penicillin-producing moulds vary naturally from strain to strain, and the Peoria team — Kenneth Raper, Dorothy Alexander and Robert Coghill — published a systematic survey of that variation in 1944, “Natural Variation and Penicillin Production in Penicillium notatum and Allied Species”. Lobanovska and Pilla describe the search for more potent strains as one of the three main streams of the wartime project, with Norman Heatley working closely with the Department of Agriculture on it.

The mouldy cantaloupe

The most famous result of that search came from a market fruit. Robert Gaynes, writing in Emerging Infectious Diseases in 2017, reports that a superior strain isolated from a mouldy cantaloupe produced six times more penicillin than Fleming’s strain. Houbraken’s 2011 study lists this wild “Wisconsin strain” as isolated from a mouldy cantaloupe in Peoria and calls it the parent of most high-yielding penicillin-producing strains — the strain whose descendant was sequenced in 2008, and which, like Fleming’s, is P. rubens. In other words, a large share of the penicillin made since the war descends from a mould on a cantaloupe.

Who found that melon is a different matter. A popular story credits a Peoria laboratory worker, remembered as “Moldy Mary” (Mary Hunt), with bringing it in from a local market after being sent out to collect mouldy produce. The story is widely repeated, but it does not appear in the sources used for this page, and it is best read as legend rather than documented fact. What the scientific record does show is the result: a natural strain from a piece of fruit, far better at making penicillin than the mould on Fleming’s plate.

From wild strain to industrial strain

The cantaloupe strain was a starting point, not an end. The 2008 genome paper describes industrial penicillin production as resting on “an unprecedented effort in microbial strain improvement”, the many rounds of selection that followed. Growing conditions mattered too: the Peoria laboratory found that adding phenylacetic acid, the chemical precursor of penicillin G’s side chain, raised production, and the move from surface culture to submerged culture in deep tanks is described on the wartime production page.

Back to Table of Contents

10. The Penicillium Genome

In 2008 a team led by Marco van den Berg and Roel Bovenberg, writing in Nature Biotechnology, published the complete genome sequence of the penicillin mould. They sequenced the strain Wisconsin54-1255, a laboratory descendant of the cantaloupe strain, and reported a genome of 32.19 million DNA letters (32.19 Mb), in which they identified numerous genes responsible for key steps in penicillin production.

The team then compared the sequenced strain with a high-producing industrial strain, measuring which genes were switched on, and grew both with and without phenylacetic acid, the side-chain precursor of penicillin G. In the high-producing strain they found increased activity of genes for making the three amino acids from which the mould assembles penicillin — valine, cysteine and alpha-aminoadipic acid — and of genes for microbodies, the small compartments inside the fungal cell where parts of the process take place. Some gene products, they reported, directly controlled how much beta-lactam the mould turned out. Genes predicted to encode transporters, the proteins that move molecules across membranes, were strongly over-represented among those switched up when penicillin G production was stimulated; the authors noted that many of the transport steps for penicillins and their building blocks were still not understood at the molecular level.

The genome closed a loop that began with a stray mould on a culture plate. The sequenced strain was published as P. chrysogenum; three years later Houbraken’s group re-identified it, like Fleming’s strain, as P. rubens. A substance first seen as a clear ring around a mould colony in 1928 can now be traced to the genes that make it, in a lineage of moulds that runs from St Mary’s Hospital and a Peoria cantaloupe to the fermentation tanks of the modern antibiotic industry.

Back to Table of Contents

Key Research Papers

1. Fleming A. On the antibacterial action of cultures of a penicillium, with special reference to their use in the isolation of B. influenzae. 1929. Bull World Health Organ. 2001;79(8):780-90. (Reprint of the 1929 paper, Br J Exp Pathol 10:226–236.) PubMed PMID: 11545337
2. Chain E, Florey HW, Gardner AD, Heatley NG, Jennings MA, Orr-Ewing J, Sanders AG. Penicillin as a chemotherapeutic agent. The Lancet. 1940;236(6104):226-228. DOI: 10.1016/S0140-6736(01)08728-1
3. Abraham EP, Chain E. An enzyme from bacteria able to destroy penicillin. Nature. 1940;146(3713):837. DOI: 10.1038/146837a0
4. Houbraken J, Frisvad JC, Samson RA. Fleming’s penicillin producing strain is not Penicillium chrysogenum but P. rubens. IMA Fungus. 2011;2(1):87-95. PubMed PMID: 22679592
5. Henk DA, Eagle CE, Brown K, Van Den Berg MA, Dyer PS, Peterson SW, Fisher MC. Speciation despite globally overlapping distributions in Penicillium chrysogenum: the population genetics of Alexander Fleming’s lucky fungus. Mol Ecol. 2011;20(20):4288-301. PubMed PMID: 21951491
6. Lobanovska M, Pilla G. Penicillin’s Discovery and Antibiotic Resistance: Lessons for the Future? Yale J Biol Med. 2017;90(1):135-145. PubMed PMID: 28356901
7. Hodgkin DC. The X-ray analysis of the structure of penicillin. Adv Sci. 1949;6(22):85-9. PubMed PMID: 18134678
8. Raper KB, Alexander DF, Coghill RD. Penicillin: II. Natural Variation and Penicillin Production in Penicillium notatum and Allied Species. J Bacteriol. 1944;48(6):639-59. PubMed PMID: 16560880
9. Gaynes R. The Discovery of Penicillin—New Insights After More Than 75 Years of Clinical Use. Emerging Infectious Diseases. 2017;23(5):849-853. DOI: 10.3201/eid2305.161556
10. van den Berg MA, Albang R, Albermann K, Badger JH, Daran JM, Driessen AJ, Garcia-Estrada C, Fedorova ND, Harris DM, Heijne WH, Joardar V, Kiel JA, Kovalchuk A, Martín JF, Nierman WC, Nijland JG, Pronk JT, Roubos JA, van der Klei IJ, van Peij NN, Veenhuis M, von Döhren H, Wagner C, Wortman J, Bovenberg RA. Genome sequencing and analysis of the filamentous fungus Penicillium chrysogenum. Nat Biotechnol. 2008;26(10):1161-8. PubMed PMID: 18820685

PubMed Topic Searches

1. Penicillium rubens
2. Penicillium chrysogenum and penicillin biosynthesis
3. Penicillin-binding proteins, transpeptidase and beta-lactams
4. History of the penicillin structure and Hodgkin’s X-ray work
5. Penicillin history at Oxford

Further Reading

• Science History Institute, “Howard Walter Florey and Ernst Boris Chain” (scientific biography). sciencehistory.org

Back to Table of Contents

Connections

• Howard Florey and Ernst Chain — Turning Penicillin into a Medicine
• Howard Florey and Ernst Chain: Lives and Careers
• The Oxford Penicillin Work, 1938–1941
• Wartime Mass Production and the Antibiotic Era
• Pharmacology: Notable Doctors
• Alexander Fleming
• Satoshi Ōmura: Natural Products Legacy
• Mould and Mycotoxins
• Cantaloupe
• Strep A: Penicillin and Antibiotic Treatment
• Staphylococcus aureus: Antibiotic Resistance and MRSA
• Antibiotic Resistance (Animation)

Back to Table of Contents