The history of penicillin traces how observations of antibiotic activity in the mould Penicillium led to the development of penicillins, a family of widely used antibiotics. Ancient societies used moulds to treat infections, and many people observed the inhibition of bacterial growth by moulds. While working at St Mary's Hospital in London in 1928, Scottish physician Alexander Fleming was the first to show experimentally that a Penicillium mould secretes an antibacterial substance, which he named "penicillin". The mould was found to be a variant of Penicillium chrysogenum (now called Penicillium rubens), a contaminant of a bacterial culture in his laboratory. The work on penicillin at St Mary's ended in 1929. In 1939, a team of scientists at the Sir William Dunn School of Pathology at the University of Oxford, led by Howard Florey, which included Edward Abraham, Ernst Chain, Mary Ethel Florey, Norman Heatley and Margaret Jennings, began researching penicillin. They developed a method for cultivating the mould and extracting, purifying and storing penicillin from it, together with an assay for measuring its purity. "Penicillin" now became the name of the active ingredient in the mould juice. They carried out experiments on animals to determine penicillin's safety and effectiveness before conducting clinical trials and field tests. They derived penicillin's chemical formula and determined how it works. The private sector and the United States Department of Agriculture located and produced new strains and developed mass production techniques. During the Second World War penicillin became an important part of the Allied war effort, saving thousands of lives. Alexander Fleming, Howard Florey and Ernst Chain shared the 1945 Nobel Prize in Physiology or Medicine for the discovery and development of penicillin. After the end of the war in 1945, penicillin became widely available. Dorothy Hodgkin determined its chemical structure, one of the achievements for which she received the Nobel Prize in Chemistry in 1964. This led to the development of semisynthetic penicillins that were more potent and effective against a wider range of bacteria. The drug was synthesised in 1957, but cultivation of mould remains the primary means of production. It was discovered that adding penicillin to animal feed increased weight gain, improved feed-conversion efficiency, promoted more uniform growth and facilitated disease control. Agriculture became a major user of penicillin. Shortly after their discovery of penicillin, the Oxford team reported penicillin resistance in many bacteria. Research that aims to circumvent and understand the mechanisms of antibiotic resistance continues.
Early evidence
Many ancient cultures, including those in Australia, China, Egypt, Greece and India, independently discovered the useful properties of fungi and plants in treating infections. These treatments often worked because many organisms, including many species of mould, naturally produce antibiotics. However, ancient practitioners could not identify or isolate the active components in these organisms.
In 1895, Vincenzo Tiberio, an Italian physician at the University of Naples, published research on moulds initially found in a water well in Arzano; from his observations, he concluded that these moulds contained soluble substances having antibacterial action. A Pasteur Institute scientist, Costa Rican Clodomiro Picado Twight, similarly recorded the antibiotic effect of Penicillium in 1923. In these early stages of penicillin research, most species of Penicillium were non-specifically referred to as P. glaucum, so that it is impossible to know the exact species and that it was really penicillin that prevented bacterial growth. André Gratia and Sara Dath at the Free University of Brussels studied the effects of bacterial samples on other bacteria. In 1924, they found that dead Staphylococcus aureus cultures were contaminated by a streptomycete. Upon further experimentation, they showed that an extract of the streptomycete could kill not only S. aureus, but also Pseudomonas aeruginosa, Mycobacterium tuberculosis and Escherichia coli (E. coli). Gratia called the antibacterial agent "mycolysate". The next year they found a killer mould that could inhibit B. anthracis. Reporting in Comptes rendus des séances de la Société de Biologie et de ses filiales, they identified the mould as Penicillium glaucum. These findings, however, received little attention as the antibacterial agent and its medical value were not fully understood, and Gratia's samples were lost.
Discovery of the properties of mould juice
While working at St Mary's Hospital, London in 1928, Alexander Fleming, a Scottish physician, was investigating the variation of growth in cultures of S. aureus, trying to replicate research from Trinity College Dublin. He spent the summer break with his family at his country home The Dhoon at Barton Mills, Suffolk. Before leaving his laboratory at the end of July, he inoculated several culture plates with S. aureus. He kept the plates aside on one corner of the table away from direct sunlight and to make space for his research student, Stuart Craddock, to work in his absence. He returned to his laboratory on 3 September. As he and Daniel Merlin Pryce, his former research student, examined the culture plates, they found one with an open lid and the culture contaminated with a blue-green mould. In the contaminated plate the bacteria around the mould did not grow, while those farther away grew normally, meaning that the mould killed the bacteria. Fleming photographed the culture and took a sample of the mould for identification. Fleming resumed his vacation and returned to St Mary's that month. He collected the original mould and grew it in culture plates. After four days he found that the plates developed large colonies of the mould. He repeated the experiment with the same bacteria-killing results. He concluded that the mould was releasing a substance that was inhibiting bacterial growth. On testing against different bacteria, he found that the mould could kill only certain Gram-positive bacteria. Staphylococcus, Streptococcus and diphtheria bacillus (Corynebacterium diphtheriae) were easily killed, but there was no effect on typhoid bacterium (Salmonella typhimurium) and the bacterium once thought to cause influenza (Haemophilus influenzae). He prepared a culture method from which he could obtain the mould juice, which he called "penicillin" on 7 March 1929, "to avoid the repetition of the rather cumbersome phrase 'mould broth filtrate'." In his Nobel lecture of 1945 he gave a further explanation, saying:
I have been frequently asked why I invented the name "Penicillin". I simply followed perfectly orthodox lines and coined a word which explained that the substance penicillin was derived from a plant of the genus Penicillium just as many years ago the word "Digitalin" was invented for a substance derived from the plant Digitalis.
After structural comparison with different species of Penicillium, Fleming believed that his specimen was Penicillium chrysogenum, a species described by an American microbiologist Charles Thom in 1910. Charles John Patrick La Touche, an Irish botanist, had recently joined St Mary's as a mycologist, and he identified the specimen as Penicillium rubrum, the identification used by Fleming in his publication. In 1931, Thom re-examined different Penicillia, including that of Fleming's specimen, and he came to the conclusion that Fleming's specimen was P. notatum, a member of the P. chrysogenum series. From then on, Fleming's mould was synonymously referred to as P. notatum and P. chrysogenum. To resolve the confusion, the Seventeenth International Botanical Congress held in Vienna, Austria, in 2005 formally adopted P. chrysogenum as the name. Whole-genome sequence and phylogenetic analysis in 2011 revealed that Fleming's mould belongs to P. rubens, a species described by Belgian microbiologist Philibert Biourge in 1923. The source of the fungal contamination in Fleming's experiment remained the subject of speculation for several decades. Fleming suggested in 1945 that the fungal spores came through the window facing Praed Street, but was disputed by his co-workers, who testified much later that Fleming's laboratory window was kept shut, and Fleming was unable to reach the window to open it. A consensus developed that the mould had come from La Touche's laboratory, a floor below Fleming's, and that spores had drifted in through the open doors. To achieve the antibacterial effect on the staphylococci cultures that Fleming observed, the mould had to be producing sufficient amounts of penicillin no later than when the bacterial growth was beginning to form visible colonies, because penicillin is only effective on bacteria when they are reproducing. Fortuitously, the temperature in the laboratory during that August was optimum first for the growth of the mould, below 20 °C, and later in the month for the bacteria, when it reached 25 °C. Fleming was a bacteriologist, not a chemist, so he left most of the chemical work to Craddock. In January 1929, Fleming recruited Frederick Ridley, a former research student of his with a background in biochemistry, to examine the chemical properties of the mould. Craddock and Ridley could not isolate penicillin, and before the experiments were over, both had left for other jobs. Fleming reported his findings to the British Journal of Experimental Pathology on 10 May 1929, and they were published in the next month's issue, but the article failed to attract much attention. Fleming was quite unsure of the medical application of his work and was more concerned with its application for bacterial isolation. The article also contained serious errors. Although Ridley and Craddock had demonstrated that penicillin was soluble in ether, acetone and alcohol as well as in water – information that would be critical to its isolation – Fleming erroneously claimed that it was soluble in alcohol and insoluble in ether and chloroform, which had not been tested. In fact, penicillin is soluble in ethanol, ether and chloroform. Further research was undertaken in the early 1930s by others including Harold Raistrick who confirmed the chemical instability of penicillin in 1932. While Fleming continued to send samples of the mould to researchers, the requests for samples tapered off.
Isolation
In 1939, at the Sir William Dunn School of Pathology at the University of Oxford, Ernst Boris Chain drew the attention of the professor in charge of the school, the Australian scientist Howard Florey, to Fleming's largely forgotten 1929 paper. They decided that the study of antibacterial substances produced by micro-organisms might be a fruitful avenue of research. Florey led an interdisciplinary research team that included Edward Abraham, Mary Ethel Florey, Arthur Duncan Gardner, Norman Heatley, Margaret Jennings, Jean Orr-Ewing and Gordon Sanders. Each member of the team tackled a particular aspect of the problem in their area of expertise, with simultaneous research along different lines building up a complete picture. This sort of collaboration was practically unknown in the United Kingdom at the time. Three sources were initially chosen for investigation: Bacillus subtilis, Trueperella pyogenes and penicillin. "[The possibility] that penicillin could have practical use in clinical medicine", Chain later recalled, "did not enter our minds when we started our work on penicillin." The broad subject area was deliberately chosen as one requiring long-term funding. Florey approached the Medical Research Council (MRC) for support in September 1939. The secretary of the council, Edward Mellanby authorized the project, allocating £250 (equivalent to £14,000 in 2025) to launch the project, with £300 for salaries (equivalent to £17,000 in 2025) and £100 for expenses (equivalent to £6,000 in 2025) per annum for three years. "It seems to me", Mellanby wrote to Florey, "that the line of work you are suggesting will be interesting and may prove to be of practical importance." Florey felt that far more would be required. On 1 November 1939, Henry M. "Dusty" Miller Jr from the Natural Sciences Division of the Rockefeller Foundation paid Florey a visit. Miller encouraged Florey to apply for funding from the foundation and supported his application. "The work proposed", Florey wrote in the application letter, "in addition to its theoretical importance, may have practical value for therapeutic purposes." His application was approved, with the foundation allocating US$5,000 (£1,250) per annum for five years. The Oxford team's first task was to obtain a sample of penicillin mould. This turned out to be easy. Georges Dreyer, Florey's predecessor, had obtained a sample of the mould in 1930 for his work on bacteriophages, viruses that infect bacteria. Dreyer had lost interest in penicillin when he discovered that it was not a bacteriophage, but he had continued to cultivate it. Dreyer had died in 1934, but Campbell-Renton had continued to culture the mould and was able to supply it to the Oxford team. The next task was to grow sufficient mould to extract enough penicillin for laboratory experiments. The mould was cultured on a surface of liquid Czapek-Dox medium. Over the course of a few days it formed a yellow gelatinous skin covered in green spores. Beneath this, the liquid became yellow and contained penicillin. The team determined that the maximum yield was achieved in ten to twenty days.
The mould needs air to grow, so cultivation required a container with a large surface area. Initially, glass bottles laid on their sides were used. Most laboratory containers did not provide a large, flat area, and so were an uneconomical use of incubator space. The bedpan was found to be practical, and was the basis for specially-made ceramic containers fabricated by J. Macintyre and Company in Burslem. These containers were rectangular in shape and could be stacked to save space. The MRC agreed to Florey's request for £300 (equivalent to £15,000 in 2025) and £2 each per week (equivalent to £97 in 2025) for two women factory hands. In 1943 Florey asked for their wages to be increased to £2 10s each per week (equivalent to £99 in 2025). Heatley collected the first 174 of an order for 500 vessels on 22 December 1940, and they were seeded with spores three days later. Efforts were made to coax the mould into producing more penicillin. Heatley tried adding various substances to the medium, including sugars, salts, malts, alcohol and even marmite, without success. At the suggestion of Paul Fildes, he tried adding brewing yeast. This did not improve the yield either, but it did cut the incubation time by a third. The team also discovered that if the penicillin-bearing fluid was removed and replaced by fresh fluid, a second batch of penicillin could be prepared, but this practice was discontinued after eighteen months due to the danger of contamination. The mould had to be grown under sterile conditions. Abraham and Chain discovered that some airborne bacteria produced penicillinase, an enzyme that destroys penicillin. It was not known why the mould produced penicillin, as the bacteria penicillin kills are no threat to the mould; it was conjectured that it was a byproduct of metabolic processes for other purposes. The next stage of the process was to extract the penicillin. The liquid was filtered through parachute silk to remove the mycelium, spores and other solid debris. The solution was acidified by the addition of phosphoric acid for the dissociation process. Chain determined that penicillin was stable only with a pH of between 5 and 8, but the process required one lower than that. By keeping the mixture at 0 °C, he could retard the breakdown process. In this form the penicillin could be drawn off by a solvent. Initially ether was used, as it was the only solvent known to dissolve penicillin, but it is highly inflammable and toxic. At Chain's suggestion, they tried using the much less flammable amyl acetate instead, and found that it also worked.
Heatley was able to develop a continuous extraction process. The penicillin-bearing solvent was easily separated from the liquid, as it floated on top, but now they encountered the problem that had stymied Craddock and Ridley: recovering the penicillin from the solvent. Heatley reasoned that if the penicillin could pass from water to solvent when the solution was acidic, maybe it would pass back again if the solution was alkaline. Florey told him to give it a try. Sodium hydroxide was added, and this method, which Heatley called "reverse extraction", was found to work. The next problem was how to extract the penicillin from the water. The usual means of extracting something from water were through evaporation or boiling, but this would destroy the penicillin. Chain hit upon the idea of freeze drying, a technique recently developed in Sweden. This enabled the water to be removed, resulting in a dry, brown powder. Heatley developed a penicillin assay using agar nutrient plates in which bacteria were seeded. Short glass cylinders containing the penicillin-bearing fluid to be tested were then placed on the nutrient plates and incubated for 12 to 16 hours at 37 °C. By then the fluid would have disappeared and the cylinder surrounded by a bacteria-free ring. The diameter of the ring indicated the strength of the penicillin. An Oxford unit was defined as the purity required to produce a 25 mm bacteria-free ring. It was an arbitrary measurement, as the chemistry was not yet known; the first research was conducted with solutions containing four or five Oxford units per milligram. Later, when highly pure penicillin became available, it was found to have 2,000 Oxford units per milligram. Yet in testing the impure substance, they found it effective against bacteria even at concentrations of one part per million. Penicillin was at least twenty times as active as the most powerful sulfonamide. The Oxford unit turned out to be very small; treating a single case required about a million units. The Oxford team reported details of the isolation method in August 1941, with a scheme for large-scale extraction. In March 1942, they reported that they could prepare a highly purified compound. In 1943 Edward Abraham proposed a structure for penicillin that contained a beta lactam ring. This structure was confirmed in 1945 by Dorothy Hodgkin, using X-ray crystallography.
Trials Howard Florey's team at Oxford showed that Penicillium extract killed many kinds of bacteria. Gardner and Orr-Ewing tested it against gonococcus (against which it was most effective), meningococcus, streptococcus, staphylococcus, Bacillus anthracis, actinomyces and tetanus bacterium (Clostridium tetani) and the bacteria that cause gangrene. They observed bacteria attempting to grow in the presence of penicillin, and noted that penicillin was neither an enzyme that broke the bacteria down, nor an antiseptic that killed them; rather, it was a chemical that interfered with the process of cell division. Jennings observed that it had no effect on white blood cells, and would therefore reinforce rather than hinder the body's natural defences against bacteria. She also found that unlike sulphonamides, the first and only effective broad-spectrum antibiotic available at the time, it was not destroyed by pus. Medawar found that it did not affect the growth of tissue cells.
By March 1940 the Oxford team had sufficient impure penicillin to commence testing whether it was toxic. Over the next two months, Florey and Jennings conducted a series of experiments on rats, mice, rabbits and cats in which penicillin was administered in various ways. Their results showed that penicillin was destroyed in the stomach, but that all forms of injection were effective, as indicated by assay of the blood. It was found that penicillin was largely and rapidly excreted unchanged in their urine. They found no evidence of toxicity in any of their animals. Had they tested against guinea pigs, research might have halted at this point, for penicillin is toxic to guinea pigs. At 11:00 am on Saturday 25 May 1940, Florey injected eight mice with a virulent strain of Streptococcus, and then injected four of them with the penicillin solution. These four were divided into two groups: two of them received 10 milligrams once, and the other two received 5 milligrams at regular intervals. By 3:30 am on Sunday all four of the untreated mice were dead. All of the treated ones were still alive, although one died two days later. Florey described the result to Jennings as "a miracle." Jennings and Florey repeated the experiment on Monday with ten mice; this time, all six of the treated mice survived, as did one of the four controls. On Tuesday, they repeated it with sixteen mice, administering different doses of penicillin. All six of the control mice died within 24 hours but the treated mice survived for several days, although they were all dead in nineteen days. On 1 July, the experiment was performed with fifty mice, half of whom received penicillin. All twenty-five of the control mice died within sixteen hours while all but one of the treated mice were alive ten days later. Over the following weeks they performed experiments with batches of 50 or 75 mice, but using different bacteria. They found that penicillin was also effective against staphylococci and gas gangrene. Florey reminded his staff that promising as their results were, a human being weighed 3,000 times as much as a mouse. The Oxford team reported their results in the 24 August 1940 issue of The Lancet, a prestigious medical journal, as "Penicillin as a Chemotherapeutic Agent" with names of the seven joint authors listed alphabetically. They concluded:
The results are clear cut, and show that penicillin is active in vivo against at least three of the organisms inhibited in vitro. It would seem a reasonable hope that all organisms in high dilution in vitro will be found to be dealt with in vivo. Penicillin does not appear to be related to any chemotherapeutic substance at present in use and is particularly remarkable for its activity against the anaerobic organisms associated with gas gangrene. The publication attracted little attention; Florey would spend much of the next two years attempting to convince people of the significance of their results. One reader was Fleming, who paid them a visit on 2 September 1940. Florey and Chain gave him a tour of the production, extraction and testing laboratories, but he made no comment and did not congratulate them on the work they had done. Some members of the Oxford team suspected that he was trying to claim some credit for it. Unbeknown to the Oxford team, their Lancet article was read by Martin Henry Dawson, Gladys Hobby and Karl Meyer at Columbia University, and they were inspired to replicate the Oxford team's results. They obtained a culture of Penicillium mould from Roger Reid at Johns Hopkins Hospital, grown from a sample he had received from Fleming in 1935. They began growing the mould on 23 September, and on 30 September tested it against viridans streptococci, and confirmed the Oxford team's results. Meyer duplicated Chain's processes, and they obtained a small quantity of penicillin. On 15 October 1940, doses of penicillin were administered to two patients with bacterial endocarditis at the Presbyterian Hospital in New York City, Aaron Alston and Charles Aronson. They became the first persons to receive penicillin treatment in the United States. The Columbia team presented the results of their penicillin treatment of the four patients at the annual meeting of the American Society for Clinical Investigation in Atlantic City, New Jersey, on 5 May 1941. Their paper was reported on by William L. Laurence in The New York Times and generated great public interest.
At Oxford, Charles Fletcher volunteered to find test cases for human trials. Elva Akers, an Oxford woman dying from incurable cancer, agreed to be a test subject for the toxicity of penicillin. On 17 January 1941, he intravenously injected her with 100 mg of penicillin. Her temperature briefly rose, but otherwise she had no ill-effects. Florey reckoned that the fever was caused by pyrogens in the penicillin; these were removed with improved chromatography. Fletcher next identified an Oxford policeman, Albert Alexander, who had a severe facial infection involving streptococci and staphylococci which had developed from a small sore at the corner of his mouth. His whole face, eyes and scalp were swollen to the extent that he had an eye removed to relieve the pain. On 12 February, Fletcher administered 200 mg of penicillin, following by 100 mg doses every three hours. Within a day of being given penicillin, Alexander started to recover; his temperature dropped and discharge from his suppurating wounds declined. By 17 February, his right eye had become normal. However, the researchers did not have enough penicillin to help him to a full recovery. Penicillin was recovered from his urine, but it was not enough. In early March he relapsed, and he died on 15 March. Because of this experience and the difficulty in producing penicillin, Florey changed the focus to treating children, who could be treated with smaller quantities of penicillin. Subsequently, several patients were treated successfully. The second was Arthur Jones, a 15-year-old boy with a streptococcal infection from a hip operation. He was given 100 mg every three hours for five days and recovered. Percy Hawkin, a 42-year-old labourer, had a 100-millimetre (4 in) carbuncle on his back. He was given an initial 200 mg on 3 May followed by 100 mg every hour. The carbuncle completely disappeared. John Cox, a semi-comatose 4-year-old boy was treated starting on 16 May. He died on 31 May but the post-mortem indicated this was from a ruptured artery in the brain, and there was no sign of infection. The fifth case, on 16 June, was a 14-year-old boy with an infection from a hip operation who made a full recovery. In addition to increased production at the Dunn School, commercial production from a pilot plant established by Imperial Chemical Industries became available in January 1942, and Kembel, Bishop and Company delivered its first batch of 910 litres (200 imp gal) on 11 September. Florey decided that the time was ripe to conduct a second series of clinical trials. Ethel Florey was placed in charge, but while Howard Florey was a consulting pathologist at Oxford hospitals, and therefore entitled to use their wards and services, Ethel, to his annoyance, was accredited merely as his assistant. Doctors tended to refer patients to the trial who were in desperate circumstances rather than the most suitable, but when penicillin did succeed, confidence in its efficacy rose. Ethel and Howard Florey published the results of clinical trials of penicillin in The Lancet on 27 March 1943, reporting the treatment of 187 cases of sepsis with penicillin. It was upon this medical evidence that the British War Cabinet set up the Penicillin Committee on 5 April 1943. The committee consisted of Cecil Weir, Director General of Equipment, as chairman; Alexander Fleming; Howard Florey; V. D. Allison, another one of Fleming's former research students; Sir Percival Hartley, the head of the MRC; and representatives from pharmaceutical companies. This led to the mass production of penicillin by the next year.
Developing industrial production
Knowing that large-scale production for medical use was futile in a laboratory, the Oxford team tried to convince the war-torn British government and private companies to engage in mass production, but the initial response was muted. Dr Blount, director of research at Glaxo Laboratories, wrote to Florey at Oxford in September 1940 but received no reply. It appeared that Florey had already appealed for assistance to two British pharmaceutical companies but had been turned down by them, and had become disillusioned with the British pharmaceutical industry. In April 1941, Warren Weaver met with Florey, and they discussed the difficulty of producing sufficient penicillin to conduct clinical trials. Weaver arranged for the Rockefeller Foundation to fund a three-month visit to the United States for Florey and a colleague to explore the possibility of production of penicillin there. Florey and Heatley left for the United States by air on 27 June 1941. Knowing that mould samples kept in vials could be easily lost, they smeared their coat pockets with the mould. Florey met with neurophysiologist John Fulton, who introduced him to Ross Harrison, the Chairman of the National Research Council (NRC). Harrison referred Florey to Thom, the chief mycologist at the Bureau of Plant Industry of the United States Department of Agriculture (USDA) in Beltsville, Maryland, and the man who had identified the mould reported by Fleming. On 9 July, Thom took Florey and Heatley to Washington, D.C., to meet Percy Wells, the acting assistant chief of the USDA Bureau of Agricultural and Industrial Chemistry and as such the head of the USDA's four laboratories. Wells sent an introductory telegram to Orville May, the director of the UDSA's Northern Regional Research Laboratory (NRRL) in Peoria, Illinois. They met with May on 14 July, and he arranged for them to meet Robert D. Coghill, the chief of the NRRL's fermentation division, who raised the possibility that fermentation in large vessels might be the key to large-scale production. On 17 August, Florey met with Alfred Newton Richards, the chairman of the Committee for Medical Research (CMR) of the Office of Scientific Research and Development (OSRD), who promised his support. On 8 October, Richards held a meeting with representatives of four major pharmaceutical companies: Squibb, Merck, Pfizer and Lederle. Vannevar Bush, the director of OSRD was present, as was Thom, who represented the NRRL. Richards told them that antitrust laws would be suspended, allowing them to share information about penicillin. This was not legalized until 7 December 1943, and it covered only penicillin and no other drug. OSRD arranged with the War Production Board (WPB) for them to have priority for equipment for laboratories and pilot plants.
Coghill made Andrew J. Moyer available to work on penicillin with Heatley, while Florey left to see if he could arrange for a pharmaceutical company to manufacture penicillin. As a first step to increasing yield, Moyer replaced sucrose in the growth media with lactose. An even larger increase occurred when Moyer added corn steep liquor, a byproduct of the corn industry that the NRRL routinely tried in the hope of finding more uses for it. The effect on penicillin was dramatic; Heatley and Moyer found that it increased the yield tenfold. At the Yale New Haven Hospital in March 1942, Anne Sheafe Miller, the wife of Yale University's athletics director, Ogden D. Miller, was succumbing to a streptococcal septicaemia contracted after a miscarriage. Her doctor, John Bumstead, was also treating John Fulton for an infection at the time. He knew that Fulton knew Florey, and that Florey's children were staying with him. He went to Fulton to plead for some penicillin. Florey had returned to the UK, but Heatley was still in the United States, working with Merck. A phone call to Richards released 5.5 grams of penicillin earmarked for a clinical trial, which was despatched from Washington, D. C., by air. The effect was dramatic; within 48 hours her 41 °C (106 °F) fever had abated and she was eating again. Her blood culture count had dropped from 100 to 150 bacteria colonies per millilitre to just one. Bumstead suggested reducing the penicillin dose from 200 milligrams; Heatley warned him not to. Heatley subsequently came to New Haven, where he collected her urine; about 3 grams of penicillin were recovered. Miller made a full recovery, and lived until 1999.
Deep submergence
Until May 1943, almost all penicillin was produced using the shallow-pan method pioneered by the Oxford team, but NRRL mycologist Kenneth Bryan Raper experimented with deep submergence production, in which penicillin mould was grown in a vat instead of a shallow dish. The initial results were disappointing; penicillin cultured in this manner yielded only three to four Oxford units per cubic centimetre, compared to twenty for surface cultures. He got the help of U.S. Army's Air Transport Command to search for similar mould in different parts of the world. Good moulds were found in samples from Chongqing, Bombay and Cape Town. The best sample, however, was from a cantaloupe sold in a Peoria fruit market in 1943. The mould was identified as Penicillium chrysogenum and designated as "NRRL 1951" or "cantaloupe strain". The spores may have escaped from the NRRL. Between 1941 and 1943, Moyer, Coghill and Raper developed methods for industrialized penicillin production and isolated higher-yielding strains of the Penicillium fungus. To improve upon that strain, researchers at the Carnegie Institution of Washington subjected NRRL 1951 to X-rays to produce a mutant strain designated X-1612 that produced 300 milligrams of penicillin per litre of mould culture, twice as much as NRRL 1951. In turn, researchers at the University of Wisconsin used ultraviolet radiation on X-1612 to produce a strain designated Q-176. This produced more than twice the penicillin of X-1612, but in the form of the less desirable penicillin K. Phenylacetic acid was added to switch it to producing the highly potent penicillin G. This strain could produce up to 550 milligrams of penicillin per litre. Pfizer was a small New York company that specialised in making citric acid, for which it had developed deep submergence techniques. This involved converting molasses to citric acid by fermenting it in a large tank in which it was stirred and the pH was carefully controlled. Pfizer's vice president, John L. Smith, whose daughter had died from an infection, put all of Pfizer's resources into the development of a practical deep submergence technique. The company invested $2.98 million in penicillin in 1943 and 1944 (equivalent to $55 million in 2025). Pfizer scientists Jasper H. Kane, G. M. Shull, E. M. Weber, A. C. Finlay and E. J. Ratajak worked on the fermentation process while R. Pasternak, W. J. Smith, V. Bogert and P. Regna developed extraction techniques.
Now that they had a mould that grew well submerged and produced an acceptable amount of penicillin, the next challenge was to provide the required air to the mould for it to grow. This was solved using an aerator, but aeration caused severe foaming of the corn steep. The foaming problem was solved by the introduction of an anti-foaming agent, glyceryl monoricinoleate. The technique also involved cooling and mixing.
Pfizer opened a pilot plant with a 7,600-litre (2,000 US gal) fermentor in August 1943 and Ratajak delivered the first penicillin liquor from it on 27 August. The one tank was soon producing half the company's output. Smith then decided to construct a full-scale production plant. The nearby Rubel Ice plant was acquired on 20 September 1943 and converted into the first deep-submergence production plant, with fourteen 130,000-litre (34,000 US gal) tanks. The work was carried out in five months under the leadership of John E. McKeen and Edward J. Goett, and the plant opened on 1 March 1944.
Worldwide production
Australia
In mid-1943 the Australian War Cabinet decided to produce penicillin in Australia. Colonel E. V. Keogh, the Australian Army's Director of Hygiene and Pathology, was placed in charge of the effort. Keogh summoned Captain Percival Bazeley, with whom he had worked at the Commonwealth Serum Laboratories (CSL) before the war, and Lieutenant H. H. Kretchmar, a chemist, and directed them to establish a production facility by Christmas. They set off on a fact-finding mission to the United States, where they visited NRRL and obtained penicillin cultures from Coghill. They also inspected the Pfizer plant in New York and the Merck plant at Rahway, New Jersey. A production plant was established at the CSL facilities in Parkville, Victoria, and the first Australian-made penicillin began reaching the troops in New Guinea in December 1943. By 1944, CSL was producing 400 million Oxford units per week (enough for 400 treatments), and there was sufficient penicillin production to allocate some for civilian use. Wartime production in Australia was in bottles and flasks, but Bazeley made a second tour of facilities in the United States between September 1944 and March 1945 and was impressed by the progress made on deep submergence technology. In 1946 and 1947 he created a pilot deep submerged plant at CSL using small 45-litre (10 imp gal) tanks to gain experience with the technique. Two 23,000-litre (5,000 imp gal) tanks became operational in 1948, followed by eight more, giving CSL a capacity of 230,000 litres (50,000 imp gal) . During the 1950s and 1960s, CSL produced semisynthetic penicillin as well. Penicillin was also produced by F.H. Faulding in South Australia, Abbott Laboratories in New South Wales and Glaxo in Victoria. By the 1970s there was a worldwide glut of penicillin. Glaxo ceased production in 1975 followed by CSL in 1980.
Canada During his visit to North America in August 1941, Howard Florey approached the Connaught Laboratories at the University of Toronto, where he met with the director, R. D. Defries, and Ronald Hare. Florey was rebuffed; Defries argued that the laboratories did not have the space, and he expressed his belief that constructing facilities to culture penicillin would be a waste as it would soon be synthesised. The results of clinical trials caused a change of heart, and in August 1943 the Canadian government asked the Connaught Laboratories to initiate mass production of penicillin. The Spadina Building was purchased by the University of Toronto for the purpose, and refurbished at a cost of Canadian $1.2 million (equivalent to Canadian $22 million in 2025), split equally between the university and the government. Penicillin was initially cultured in 200,000 bottles occupying 740 square metres (8,000 sq ft) of air-conditioned laboratory space. Production was switched to the deep submergence method in November 1945.
Continental Europe A translation of the Oxford team's 1941 report reached Germany via Sweden the following year. On 6 December 1943, the Reich Health Ministry ordered the medical community to conduct research into penicillin and other antibiotics. Three vials of penicillin captured by the Afrika Korps reached Germany in 1943 and one was sent to Heinz Öppinger at Hoechst in Frankfurt, who began conducting experiments with moulds. Penicillin was produced there in 300-litre batches, and Öppinger developed a rotating drum for a deep-tank fermentation process. Research was also carried out by Schering in Berlin using a sample of Fleming's mould, which they failed to cultivate; their efforts to determine the chemical structure of penicillin were also unsuccessful. Maria Brommelhues at IG Farben's Bacteriological Laboratory in Elberfeld catalogued different species of penicillin. Hitler's personal physician, Theodor Morell, treated Hitler with penicillin for injuries sustained in the 20 July 1944 assassination attempt. Information about penicillin research in Germany was gathered by the Manhattan Project's Alsos Mission and forwarded to Florey in the UK. Much of Germany's penicillin came from Czechoslovakia, where research was carried out at Charles University in Prague and the Fragner Pharmaceutical Company by a team that included chemist Karel Wiesner. Work was also conducted in secret in France and at the Delft University of Technol