Peter Mitchell was an English biochemist with a rich uncle, a Cambridge thesis that was sent back and a stomach that finally stopped him. In 1961 he proposed that the energy of respiration is stored as a difference in acidity and electric charge across a membrane, and that the cell spends that difference to make ATP. For a decade or more the people whose work rested on a rival chemical theory told him he was wrong. He answered by restoring a decaying Regency mansion in Cornwall, endowing a private institute with family shares and, with Jennifer Moyle, measuring the protons that mitochondria push out when they breathe. He won the 1978 Nobel Prize in Chemistry alone. By then some of his numbers were being overturned, and his account of how ATP is made was on its way to being replaced. He was not an easy man to disagree with, and this page tries to show both what he got right and what it cost.
1920–1955
Uncle G’s shares and a thesis sent back
Peter Dennis Mitchell was born in Mitcham, Surrey, on 29 September 1920, the second son of a civil servant. His father, Christopher Gibbs Mitchell, supervised road-building across about a fifth of England for the Ministry of Transport and was responsible for the then-controversial introduction of one-way streets. His mother, Kate Taplin, a shy, musical rationalist and atheist, taught him to take responsibility for his own failings. The marriage was unhappy. By about 1931 his father was living in Birmingham for work and the couple were effectively separated. In 1951, the year Mitchell’s doctorate came through, Christopher Mitchell killed himself.
The money came from the Mitchell side. Christopher’s younger brother Godfrey bought an insolvent builder, George Wimpey & Co., in 1919 and made it one of the largest contractors in Europe: more than 300,000 houses by his death in 1982, ninety-eight airfields in the Second World War, Heathrow afterwards. Sir Godfrey, knighted in 1948 and known in the family as ‘Uncle G’, gave his nephew gifts of Wimpey shares. His biographers Prebble and Weber call them a major factor in Mitchell’s later life: economic freedom and, in time, the funding of his research and of his research institute. They also tie his lifelong confidence and entrepreneurial streak to this background, which is their interpretation, not a claim Mitchell made.
The scientist began unpromisingly. At Queen’s College, Taunton, he was an indifferent pupil outside mathematics and physics, and his Cambridge scholarship paper was, in his own word, dismal: he reached Jesus College in 1939 only because his headmaster, C. L. Wiseman, wrote that the result should be ignored. He took a Class III in Part I of the Natural Sciences Tripos and a II-1 in biochemistry. In 1942 James Danielli took him on as his only PhD student, taught him how membranes and transport were studied, and introduced him to David Keilin, the cytochrome man at the Molteno Institute, whom Mitchell said he came to love and respect more than any other scientist.
Cambridge never quite knew what to do with him. Danielli left, and Mitchell worked largely alone on a first thesis that opened with a philosophical discussion of directional processes, moved to a mathematical theory of diffusion inside cells and ended with modest experiments on bacteria taking up amino acids. The examiners, Ernest Gale and the physical chemist Sandy Ogston, sent it back. Ogston remembered finding its presentation, and Mitchell’s manner in discussion, ‘silly’, though he later thought referral had been ‘too hard’. Keilin, whose daughter remembered him as angry, is reported to have said that ‘the trouble is that Peter is too original for his examiners’. Mitchell wrote a second, conventional thesis on how penicillin alters nucleic-acid synthesis in bacteria, accepted in December 1950; the mechanism it proposed turned out to be wrong, but it turned his attention to phosphate crossing bacterial membranes. In 1955 the department did not keep him.
One lasting thing came out of those years. Marjory Stephenson steered her assistant, Jennifer Moyle, towards Mitchell’s bacterial work, and from 1948 the two worked together. In Bruce Weber’s account, both of them felt that Stephenson had read their complementary strengths correctly: Mitchell was the imaginative theorist and Moyle the meticulous experimentalist.
Michael Swann offered Mitchell a new Chemical Biology Unit in the zoology department at Edinburgh and, according to Weber’s biographical entry, Mitchell accepted on condition that Moyle came as his research associate. For the first time he ran his own group. He and Moyle studied how staphylococci take up phosphate and arsenate, found respiratory enzymes in the bacterial plasma membrane, and hardened a conviction that orthodox biochemistry, which treated the cell as a bag of enzymes in solution, had left out space. An enzyme plugged through a membrane moves a chemical group from one side to the other, so the reaction has a direction in space. Mitchell called this vectorial chemistry, against the ‘scalar’ chemistry of the test tube.
He turned the idea on the great unsolved problem of the 1950s, oxidative phosphorylation: how the oxidation of food along the respiratory chain pays for ATP, the source, as he put it, of about 95% of the energy of aerobic organisms. The reigning answer was Bill Slater’s chemical theory of 1953: an energy-rich chemical intermediate is formed in the chain and hands its energy to ADP. Nobody could find it. In Mitchell’s later words the field was ‘littered with the smouldering conceptual remains of numerous exploded energy-rich chemical intermediates’, and the philosopher Douglas Allchin counts at least sixteen published claims to have found one, none of which survived. Mitchell described his own entry as ‘simply a strategic conjectural one’: a chance that a membrane-based rationale might give the field a common framework.
He outlined the idea in Prague in August 1960 and in Stockholm a few weeks later, and published it in Nature in July 1961. The respiratory chain, he proposed, is built across a membrane so that oxidation delivers protons to one side. The membrane must be almost impermeable to protons, so that a gradient of acidity and of electric charge builds up across it. An ATPase in the same membrane can run backwards, letting protons flow through it to make ATP. No intermediate is needed. Mitchell himself later called these postulates ‘almost entirely hypothetical’ in 1961 and admitted that his first guesses at the proton numbers and the direction of proton pumping had to be revised after early experiments. The polarity error was caught because the proposal could be tested: Mitchell’s own early experiments, independently confirmed by Guy Greville, showed respiring mitochondria acidifying their medium, which implied the opposite orientation.
A shadow lies over the 1961 paper. In January 1961 the Oxford inorganic chemist R. J. P. (Bob) Williams had published a theory, submitted in 1960, in which protons held at high concentration within a membrane link the respiratory chain to ATP. On 24 February Mitchell wrote to ask what Williams meant; eleven letters passed in eight weeks and the exchange turned acrimonious. Mitchell sent his Nature paper off soon after the last of them and cited neither Williams’s paper nor the correspondence. His position was that the two theories differed in kind: his protons crossed a membrane between watery phases, while Williams’s stayed within it, so his own work owed Williams nothing. Williams held that the exchange fed Mitchell’s thinking. In a note to Mitchell in the 1970s he wrote that he regarded the 1961 publication as ‘a dishonorable act (that is a personal judgement)’ and urged him: ‘Just publish all our letters and let the world decide between us.’ Mitchell never agreed. Slater’s verdict was that both men brought in protons in 1961 but that Mitchell developed them into a coherent theory in 1966. Tom Blundell, in his foreword to Prebble and Weber’s biography, calls Mitchell’s handling of Williams ‘quite extraordinary’ and says publishing the letters would have been a fair response to Williams’s generosity. The ATPase mechanism that each man later proposed has not stood the test of time.
Edinburgh was going well when his stomach gave out. The university made Mitchell a senior lecturer in October 1961 and a reader in July 1962, a track that could lead to a chair. But gastric ulcers and, as he himself came to see, an inability to switch his mind off problems were wearing him down. The treatment on offer was a major operation with no guaranteed outcome; the alternative was complete rest and no university stress. He chose rest. He took leave from March 1963, resigned, and the university recorded the end of his readership in July. By his own account he then ‘withdrew completely from scientific research’ for two years.
Cornwall supplied the occupation. While being shown cottages near Bodmin he was taken up the drive of Glynn House, a Regency mansion empty for years and riddled with dry rot. ‘Only a lunatic,’ he concluded, ‘would want to become involved with it.’ He bought it anyway on 9 August 1962 for £2,817 10s, meaning to restore it and sell. Forty-five gallons of anti-rot fluid proved too little and four hundred were used. He acted as his own architect and master of works, labouring beside the builders’ foreman, and when the contractors quoted around £500 plus 15 per cent for a steel spiral staircase he taught himself the geometry of helices and had a Wadebridge firm make it for £140. He had already restored a derelict house at Carrington, bought from the Church of Scotland. He also acquired Glynn Farm with its eight Jersey cows, and the biography records that hand-milking them did wonders for his ulcers.
The research institute came out of the restoration. Mitchell was at first ambivalent about founding one, and his biographers judge that his health settled the question. He wrote to Moyle, still in Edinburgh, to ask whether she would join what both knew was a gamble. She recalled it as ‘obviously a gamble’ that would be ‘most exciting if it succeeded’, and felt ‘nothing would be lost’ by saying yes. They set up Glynn Research Ltd, a charitable company to promote fundamental biological research. Its endowment of about £250,000 came about equally from Mitchell and his elder brother Christopher John, and the money was Wimpey stock. Research began in 1964–65 with one technician, and the institute never afforded more than three scientists on its permanent staff. Until 1978 its council of management was Mitchell and Moyle alone.
Mitchell wanted to show that first-rate science could be done outside a university, and believed that small organisations were more alive, more congenial and for many purposes more effective than large ones. The biography describes the Glynn day: some hours of discussion between Mitchell and Moyle of the latest results, then experiments in the second half of the day, with Moyle at the bench while Mitchell built and maintained instruments. A visiting American, Peter Hinkle, noted that rat liver mitochondria were prepared only two or three times a week and that Mitchell discouraged weekend work.
Glynn’s first job was to make the field take him seriously, and Mitchell knew he did not yet know the field. He read the literature hard. In 1965 Bill Slater, author of the chemical theory he meant to replace, persuaded him, in Mitchell’s later words, to become more deeply involved in practical research on oxidative phosphorylation. Albert Lehninger had raised two objections to the 1961 proposal: it did not account for the number of ATPs made per oxygen atom, then believed to be three, and fragments of mitochondria that still made ATP should have no intact membrane. Mitchell and Moyle answered both. They showed that the fragments were closed vesicles turned inside out, which Slater called good evidence, and they set out to count protons.
The measurement is simple to describe. Rat liver mitochondria are suspended with a substrate to burn but no oxygen, so the respiratory chain is loaded with electrons and stalled. A small known amount of oxygen is added: a pulse of respiration. If chemiosmosis is right, the chain ejects protons through the inner membrane as it uses that oxygen, and a pH electrode in the surrounding medium records a pulse of acidity. The acid then fades, because protons leak back in. Mitchell and Moyle extrapolated the fade back to the moment of the pulse to estimate how many protons had really been ejected, and divided by the oxygen atoms added, giving a ratio they wrote →H+/O. In 1967 they reported that it was very close to 4 when electrons entered through succinate and 6 through β-hydroxybutyrate, constant between pH 5.5 and 8.5. That fits the three loops of Mitchell’s 1966 scheme, two protons per oxygen for each loop crossed: three for the β-hydroxybutyrate route, two for succinate. It also made the book-keeping neat. They had measured two protons passing through the ATPase per ATP, so six protons out per oxygen could pay for three ATPs.
The control that mattered was the uncoupler. Dinitrophenol and other uncouplers sever respiration from ATP synthesis. If, as the chemical theory held, an energy-rich intermediate lay between the chain and the ejection of protons, an uncoupler that destroyed it should shrink the acid pulse. If an uncoupler merely opened a leak in the membrane, the pulse should decay faster but start the same size. Mitchell and Moyle added dinitrophenol, FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone) and gramicidin. The pulses decayed faster; the initial →H+/O did not change. They concluded that uncoupler-sensitive chemical intermediates do not exist between the respiratory chain and the proton-translocating mechanism. A companion paper measured the membrane’s proton conductance, only 0.45 micromho per square centimetre of membrane near pH 7, low enough to account for the tight braking of respiration when no ATP is being made, and showed that the same uncouplers raised it greatly.
The result did not end the argument. It met Lehninger’s objection, but a chemical theorist could treat proton movements as a side reaction of the intermediate, and Prebble and Weber note that the experiments ‘did not distinguish’ one theory from the other. Britton Chance, whose apparatus could time such events, found proton ejection lagging respiration by about a second; Mitchell, using his own methods, found them synchronous. Firmer support came from André Jagendorf, whose chloroplasts made ATP in the dark when soaked in acid and moved into alkali (1965–66). Slater’s laboratory repeated that but could not make mitochondria do the same; Mitchell, Moyle and Robert Reid did coax a brief burst of ATP from acid-treated mitochondria in 1966. By 1968, in Weber’s account, Mitchell had evidence for each of his three pillars: proton-tight membranes, proton ejection by the respiratory chain, and ATP made by a proton gradient. After that, the field could no longer ignore him.
‘Ox-phos wars’ was the biochemists’ own name for the fight, and Prebble and Weber call the debate often heated and sometimes acrimonious. Many of Mitchell’s opponents had spent careers hunting the intermediate; he called the field one ‘afflicted with various sorts of religious fervor’. Slater, the clearest case, began as an ally. In a review of oxidative phosphorylation written in 1964–65 he called Mitchell’s theory ‘beautifully simple’ but thought the real process more complicated than it allowed. After reading Mitchell’s privately printed ‘Grey Book’ of 1966 he wrote that it ‘has made me mad’, and that he was not answering Mitchell’s letters, ‘since I see no profit in discussion with someone who is so obsessed with his theory’. At a meeting in Warsaw that year Brian Chappell saw Slater so furious that he was hopping on one foot; Mitchell told Jagendorf that Slater ‘seems to have declared war on some of our experiments’. In 1967 Slater wrote that the theory was ‘at the moment very harmful’, that its predictions were being confused with fact, and that ‘Peter’s little grey book is becoming like Mao’s red book.’ Williams, an ally on some points, thought Mitchell’s polemic ‘wild’ and him ‘such a difficult man to deal with’, though some of his arguments were ‘very instructive’.
Mitchell gave as good as he got. He told Sergio Papa that the technical quality of some critics’ experiments was ‘second-rate—mostly, I guess, done by students with little experience’, then apologised for an outburst ‘loaded with sadness’. He wrote that purely conformational explanations, which can accommodate any proton count, ‘encourage sloppy experiments’. He published a chart naming the biochemists who had come over to his view, which at least one senior colleague thought in bad taste. Prebble and Weber add that he was adept at finding alternative readings of other people’s results, a source of considerable irritation. In 1974 Efraim Racker, anxious that the quarrelling was making the field look chaotic to grant panels, proposed that its leaders sign a joint statement. Mitchell declined to sign: ‘Such authoritarian techniques belong (perhaps unhappily) to politics and not to western science.’ Racker, for his part, said gently that Mitchell’s ideas had not been treated too kindly. In a 1975 lecture he said, tongue in cheek, that Mitchell’s replacement of the intermediate by a ‘hypothetical’ proton gradient and an ‘imaginary’ membrane potential had sounded to the establishment like ‘the pronouncements of a court jester or of a prophet of doom’.
Paul Boyer’s rival proposal, from 1964, was that energy is held neither in a chemical intermediate nor in a proton gradient but in the changing shape of proteins, like winding a spring. Many former chemical theorists took refuge in it. At a Bristol meeting in 1973 Boyer was conciliatory; Mitchell replied that the differences should be stated ‘without blurring the edges’. History split the prize. The ATP synthase does work by conformational change, as Boyer said, and he shared the 1997 Nobel Prize in Chemistry with John Walker and Jens Skou. Protons driving it, as Mitchell said, is also accepted, and textbooks teach what Weber calls the Mitchell–Boyer mechanism. Mitchell’s own picture, in which protons take part directly in the chemistry at the catalytic site, was wrong, and Blundell calls his reaction to Boyer’s model ‘difficult to understand’.
He was not simply stubborn, and some of his boldest guesses held. An experiment that persuaded many outsiders was Racker and Walther Stoeckenius’s in 1973: bacteriorhodopsin, a light-driven proton pump from bacteria, was put into vesicles with an ATPase and made ATP when lit, with no respiratory chain and no mitochondrial intermediate. In 1975 Mitchell proposed the Q cycle, in which coenzyme Q ferries extra protons across the membrane; Weber calls it an ‘extraordinary feat of imagination’ well beyond the data then available, and it is essentially accepted today.
1974–1986
Getting the arithmetic right
The bitterest fight was over a number. From 1974 Lehninger’s group at Johns Hopkins and Mårten Wikström in Helsinki reported more protons per oxygen than Glynn had. Lehninger’s explanation was that phosphate in the Glynn preparations re-entered mitochondria with protons and hid some of those ejected. In 1976 Lehninger’s group, with M. D. Brand and B. Reynafarje, used the oxygen-pulse method itself and found about 3 protons per site when phosphate movement was blocked and 2 when it was not: the Glynn figure, reproduced, with a reason for it. After three years of debate Moyle and Mitchell published a new programme of measurements and concluded that their numbers were right and the phosphate adequately controlled. Mitchell told Lehninger they disagreed over interpretation, not raw data, and in the biography’s account Lehninger felt ignored.
The stakes were Mitchell’s mechanism, not just the arithmetic: bigger numbers undercut his looped chain and the ‘ligand conduction’ that explained it. Prebble and Weber judge that for Mitchell, in later years, theory tended to take primacy over experiment, and that he looked for other readings of results until a revision was unavoidable. From the late 1970s until 1986 he was also in a public argument with Wikström over whether cytochrome oxidase pumps protons at all. It does. By 1985–86 Mitchell and his colleagues had concluded that the respiratory chain ejects not six protons per oxygen but at least eight, and he had withdrawn the calcium–phosphate proposal he had used to explain away Lehninger’s calcium data. He went on to argue that his basic scheme could absorb the higher numbers. Meanwhile the proton cost of making ATP itself had risen, in Weber’s account, from the two his first theory required to three or four.
The strain showed. A failed ear operation in 1972 left Mitchell almost completely deaf; the ulcers that had ended his university career still flared, and there were fits of depression. Around 1977, according to Prebble and Weber, he met the wife of another biochemist at a Biochemical Society dinner, was attracted to her, and installed her in a bungalow he had restored on the south Cornish coast. She left within a week. His wife Helen forgave him, though the marriage never regained its earlier warmth. That summer he had a nervous breakdown, which he described to Hinkle as ‘hysterical dissociation’, and his doctors ordered rest. The biography treats the episode as a fulcrum: a man who had been deeply creative for three decades was afterwards mostly defending positions.
Glynn’s future wobbled too. In October 1976 the institute’s council resolved to end biochemical research from October 1977 and turn to the social behaviour of Homo sapiens. Mitchell’s Nobel autobiography is plain about the interest: communication between people, the spread of violence, and what he called the flagrant abuse and basically dishonest manipulation of the monetary system by governments. Redundancy notices were planned for three staff. Lars Ernster travelled to Glynn to argue against it, and Mitchell later wrote that his persuasion changed the decision. Bioenergetics went on, with a behavioural-biology project that never found funding.
News of the 1978 Nobel Prize in Chemistry reached Glynn from Reuters, not Stockholm; asked for his reaction, Mitchell said ‘Astonishment’. The prize, worth £84,000, was his alone, for ‘his contribution to the understanding of biological energy transfer through the formulation of the chemiosmotic theory’. The wording was cautious. In the biography’s account the chairman of the Nobel Committee for Chemistry wrote later that the committee ‘was well aware of the fact that Mitchell was wrong about how the proton gradient is created (proton pumps rather than redox loops) and how it is utilized to make ATP’, which is ‘why the academy’s citation read rather vaguely’. Some thought it premature. Others thought it should have been shared with Racker, Chance or Williams, and one prominent American bioenergeticist wrote that he was sorry ‘those who labored so hard to bring substance to its shadow should not have been similarly recognized’.
In his lecture, given on 8 December with Helen and Jennifer Moyle in Stockholm, Mitchell opened by celebrating David Keilin, who never won a prize, and praised the ‘altruism and generosity’ of former opponents who had promoted his hypothesis to a theory. He thanked Slater for drawing him into practical work, admitted that his first guesses at stoichiometry and polarity had needed revision, and judged the dissent of Azzone, Lehninger and Wikström over proton numbers no ‘serious threat’ to a ‘relatively hard experimental datum’. Within a decade he had given way on the numbers. He said afterwards that the money arrived when he had a large overdraft, because so much of his family money had gone into Glynn, and that being a laureate made his provocative remarks harder to get away with.
The Nobel citation names one person. Mitchell’s autobiography for the Foundation calls Moyle his ‘former research colleague’ whose ‘enthusiastic support’ he enjoyed during the restoration, and says the two of them founded Glynn Research Ltd; his lecture thanks his ‘research associate’ for ‘constant discussion, criticism and help’. Prebble and Weber put it more bluntly: it is very questionable whether he would have made significant progress without her. Almost none of his theoretical papers carried her name, virtually all of his later experimental papers did, and her careful experiments, in their judgement, were essential to persuading biochemists. Mitchell himself wrote that she had done the greater part of the experimental work at the bench. In 1982 Mitchell asked Lord Swann to put her name forward for a public honour and wrote to the permanent secretary at the Department of Education and Science that the part he had played in the new theory ‘has tended to overshadow the immensely important role of Dr. Moyle’s experimental contributions to our joint efforts—to my mind, quite unfairly’. The attempt failed, as did his attempt to obtain an honorary degree for her. She retired in 1983 after some thirty-five years working with him. The biographers add a caution of their own: many of the quantitative results of those Glynn experiments have not stood the test of time, though they sustained a line of inquiry that mattered.
Glynn ran on dwindling Wimpey income. By the mid-1980s Mitchell had handed the research directorship to Peter Rich, a Cambridge bioenergeticist who brought in outside grants (the sources differ on the exact year of Mitchell’s retirement), and he spent the rest of his working life refining his ATPase mechanism and chasing money for the institute. His other obsession was money itself: in 1977 he had struck one-ounce silver bullion pieces called Glynns, inscribed ‘freedom to save, liberty to earn’, a commercial failure. He died at Glynn on 10 April 1992. In 1996 the institute’s research operations moved to University College London as the Glynn Laboratory of Bioenergetics. Weber’s verdict is that what began as an attempt to do major research outside the universities ended up absorbed back into the university system.
What lasted is a picture in which a membrane, a proton gradient and an ATP-making enzyme are one circuit, together with the Q cycle. What did not last is his count of six protons per oxygen, his looped chain as the way the gradient is made, and his account of the ATPase. The ledger of credit stays untidy: Williams’s grievance unresolved, Boyer and Walker prized for the enzyme, Moyle without a prize at all. For a different account of how a cell is organised, resting on different premises and different experimental support, see Gilbert Ling; the two are not members of one school.
