by Frank Ryan
Over time, many researchers came to see bacteria as living organisms, classifying them according to the binomial Linnaean system; so, for example, the tuberculosis germ was labeled Mycobacterium tuberculosis, and the boil-causing coccoid germ was labeled Staphylococcus aureus. Oswald Avery, with his extremely conservative nature, kept his options open, eschewing the binomial system and referring to the TB germ as the “tubercle bacillus.” It is instructive for our story that Dubos, who came to know Avery better than any other colleague, would observe that “Fess” was similarly conservative in his approach to laboratory research. Science must adhere with a puritanical stringency to what can be logically observed and definitively proven in the laboratory.
In 1882, German physician Robert Koch discovered that Mycobacterium tuberculosis was the cause of the greatest infectious killer in human history—tuberculosis. Koch constructed a code of logic that would be applied to bugs when first determining if they caused specific diseases. Known as “Koch's postulates,” this was universally adhered to, and once a causative bug had been identified it was studied further under the microscope. Thus the bug was duly classified in a number of ways. If its cells were rounded in shape it was a “coccus,” if a sausage shape it was a “bacillus,” if a spiral shape it was a “spirochete.” Bacteriologists methodically studied the sort of culture media in which a bug would grow best—whether in agar alone, or agar with added ox blood, and so on. They also studied the appearance of the bacterial colonies when they were grown in culture plates—their colors; the size of the colonies; whether they were rough in outline or round and smooth; raised or flat; stellate, granular, or daisy-head. So the textbooks of bacteriology extended their knowledge base on a foundation of precise factual study and observation. And as understanding grew, this newfound knowledge was applied to the war against infection.
One of the useful things they learned about disease-causing, or “pathogenic,” bacteria was that the behavior of the disease, and thus of the bug itself in relation to its infected host, could be altered by various deliberate means: for example, through repeated cultures in the laboratory or by repeatedly passing generations of the bug through a series of experimental animals. Through such manipulations it was possible to make the disease worse or less severe by making the bug either “more virulent” or “attenuated.” Bacteriologists looked for ways to extrapolate this to medicine. In France, for example, the eminent Louis Pasteur used this principle of attenuation to develop the first vaccine to be used successfully against the otherwise universally fatal virus infection of rabies.
One fascinating observation that came out of these studies was the fact that, once a bug had been attenuated or been driven to greater virulence, the change in behavior could be “passed on” to future generations. Could it be that some factor of the bug's own heredity had been altered to explain the change in behavior?
Bacteriologists talked about “adaptation,” using the same term that was coming into vogue with evolutionary biologists when referring to evolutionary change in living organisms as they adapted to their ecology over time. While it was too early to be sure if bacterial heredity depended on genes, these scientists linked it to the physical appearance of bugs and colonies, or to the bugs’ internal chemistry, and even to their behavior in relation to their hosts. These were measurable properties: the bacterial equivalents of what evolutionary biologists were calling the “phenotype”—the physical makeup of an organism as opposed to what was determined by the hereditary makeup, or “genotype.”
Bacteriologists also came to recognize that the same bacterium could exist in different subtypes, which could often be distinguished from one another using antibodies. These subtypes were called “serotypes.” In 1921, a British bacteriologist, J. A. Arkwright, noticed that the colonies of a virulent type of dysentery bug, called Shigella, growing on the jelly-coated surfaces of culture plates, were dome-shaped with a smooth surface, whereas colonies of an attenuated, non-virulent, type of dysentery bug were irregular, rough-looking, and much flatter. He introduced the terms “Smooth” and “Rough” (abbreviated to S and R) to describe these colonial characteristics. Arkwright recognized that the R forms cropped up in cultures grown under artificial conditions, but not in circumstances where bacteria were taken from infected human tissues. He concluded that what he was observing was a form of Darwinian evolution at work.
In his words: “The human body infected with dysentery may be considered a selective environment which keeps such pathogenic bacteria in the forms in which they are usually encountered.”
Soon researchers in different countries confirmed that loss of virulence in a number of pathogenic bacteria was accompanied by the same change in colony appearance from Smooth to Rough. In 1923, Frederick Griffith, an epidemiologist working for the Ministry of Health in London, reported that pneumococci—the bugs that caused epidemic pneumonia and meningitis, which were of particular interest to Oswald Avery at the Rockefeller Laboratory—formed similar patterns of S and R forms on culture plates. Griffith was known to be a diligent scientist, and Avery was naturally intrigued.
Griffith's experiments also produced an additional finding, one that really shook and puzzled Avery.
When Griffith injected non-virulent R-type pneumococci from the strain known as type I into experimental mice, he included an additional ingredient in the injections, a so-called “adjuvant,” which usually pepped up the immune response to the R pneumococci. A common adjuvant for these purposes was mucus taken from the lining of the experimental animal stomach. But for some obscure reason, Griffith switched adjuvant to a suspension of S pneumococci, derived from type II, that had been deliberately killed off by heat. The experimental mice died from overwhelming infection. In the blood of these dead mice, Griffith expected to find large numbers of multiplying R-type I bacteria—the type that he had injected at the start of the experiment. Why then had he actually found S-type II? How on earth could adding dead bacteria to his inoculum have changed the actual serotype of the bacterium from non-virulent R-type I to highly virulent S-type II?
Researchers, including Avery himself, had previously shown that S and R types were determined by differences in the polysaccharide capsules coating the cell bodies of the bugs. Griffith's findings suggested that the test bacteria, initially R-type pneumococci, had changed their polysaccharide coats inside the infected bodies of the mice to that of the virulent strain. But they could not have achieved this by just flinging off the old coat and putting on the new one. The coat was determined by the bacteria's heredity—it was an inherited characteristic. Further cultures of the recovered bacteria confirmed that the S-type bred true. There appeared to be only one possible explanation: adding the dead S bacteria to the living R bacteria had induced a mutation in the heredity of the living R-type bacteria, so they literally transformed into S-type II.
In the words of Dubos: “[At the time] Griffith took it for granted that the changes remained within the limits of the species. He probably had not envisaged that one pneumococcus type could be transformed into another, as this was then regarded as the equivalent of transforming one species into another—a phenomenon never previously observed.”
It is little wonder that Avery was astonished by Griffith's findings. Like Robert Koch before him, Avery subscribed to the view that bacterial strains were immutable in terms of their heredity. The very concept of a mutation—that heredity was capable of an experimentally induced change—was a highly controversial issue within biology and medicine at this time. To understand why, we need to grasp the concept of what a mutation means.
By the late nineteenth century, Darwinian theory had entered a crisis. Darwin himself had been well aware that natural selection relied on some additional mechanism, or mechanisms, capable of changing heredity, so that natural selection would have a range of “hereditable variation” to choose between. Generations later, in the opening chapters of his innovative book Evolution: The Modern Synthesis, Julian Huxley put his finger on the
nub of the problem. “The really important criticisms have fallen upon Natural Selection as an evolutionary principle and centered round the nature of inheritable variation.” In 1900, a Dutch biologist, Hugo de Vries, put forward a novel mechanism that would be capable of providing the necessary variation: the concept of a random change in a unit of inheritance. Opportunity for change exists when genes are copied during reproduction, when a random change in the coding of a gene might arise from an error in copying the hereditary information. De Vries called this source of hereditary change a “mutation.” It was only with what Julian Huxley termed “the synthesis” of Mendelian genetics—the potential for change in the inherited genes through mutation—and Darwinian natural selection operating on the hereditary choices presented within a species, that Darwinian theory became credible again to the great majority of scientists.
In time, Griffith's finding would be confirmed to be what Avery was now wondering about: it was a mutation. Geneticists would show that the change from the R to the S strain of pneumococcus involved the transfer of a gene from the dead S-type II bacteria to the living R-type I bacteria, which was incorporated into subsequent bacterial reproductive cycles, transforming the cells of the R-type I bacterium into the cells of the S-type II bacterium. It was indeed the bacterial equivalent of a change of species. And Griffith was proven right in inferring that Darwinian natural selection had operated even in the short time frame of the infection of a cohort of laboratory mice.
Griffith's experimental findings galvanized bacteriologists and immunologists around the world. His discovery was confirmed in several different research centers, including the Robert Koch Institute in Berlin, where the pneumococcal types had first been classified. The news was inevitably a hot topic of discussion in Avery's department, as Dubos would recount: “but we did not even try to repeat them at first, as if we had been stunned and almost paralyzed intellectually by the shocking nature of the findings.”
At first Avery simply couldn't believe that bacterial types could be transformed. Indeed, he had been one of the authoritative figures who had settled the fixity of bacterial reproduction being true to type years before. But from 1926, Avery encouraged a young Canadian physician working in the Rockefeller Laboratory, M. H. Dawson, to investigate the situation. According to Dubos, Dawson, unlike Avery, was convinced from the start that Griffith's conclusion must be correct because he believed that “work done in the British Ministry of Health had to be right.”
Dawson began by confirming Griffith's findings in laboratory mice. His results suggested that the majority of non-virulent bacteria—the R types—had the ability in certain circumstances to revert to the virulent S type. By 1930, the young Canadian was joined by a Chinese colleague, Richard P. Sia, and between them they took the experimental observations further by confirming that the hereditary transformation could be brought about in culture media, without the need for passage through mice. At this stage, Dawson left the department, and Avery encouraged another young physician, J. L. Alloway, to take the investigation further. Alloway discovered that all he needed to bring about the transformation was a soluble fraction derived from the S pneumococci by dissolving the living cells in sodium deoxycholate, then passing the resultant solution through filters to remove the bits of broken-up cells. When he added alcohol to the filtered solution, the active material precipitated out as sticky syrup. Throughout the laboratory this sticky syrup was referred to as the “transforming principle.” So the work continued, experiment following experiment, year by year.
When Alloway left the department in 1932, Avery began to devote some of his own time to the pneumococcus transformation, in particular aiming to improve the extraction and preparation of the transforming substance. Frustration followed frustration. He focused on its chemical nature. Discussion took place with other members of the department, ranging from the “plamagene” that was thought to induce cancer in chickens (now known to be a retrovirus), or to the genetic alterations in bacteria that were thought to be caused by viruses. According to Dubos, Alloway suggested the transforming agent might be a protein-polysaccharide complex. But by 1935, Avery was beginning to think along other lines. In his annual departmental report that year, he indicated that he had obtained the transforming material in a form that was essentially clear of any capsular polysaccharide. In 1936, Rollin Hotchkiss, a biochemist who had now arrived to work in the department, wrote a historic comment in his personal notes:
“Avery outlined to me that the transforming agent could hardly be a carbohydrate, did not match very well with a protein and wistfully suggested it might be a nucleic acid!” At this stage, Dubos, who many years later would write a book about Avery and his work, dismissed this as no more than a surmise. There were good reasons for his caution.
That year, few researchers throughout the world believed that the answer to heredity lay with nucleic acids. These chemical entities had been discovered by a Swiss biochemist, Johann Friedrich Miescher, back in the late 1800s. Fascinated by the chemistry of the nucleus, Miescher had broken open the nuclei of white blood cells in pus, and subsequently the heads of salmon sperm, to discover a new chemical compound which was acidic to pH testing, rich in phosphorus, and comprised of enormously large molecules. After a lifetime of experimentation on the discovery, Miescher's pupil Richard Altmann would introduce the term “nucleic acid” to describe Miescher's discovery. By the 1920s, biochemists and geneticists were aware that there were two kinds of nucleic acids. One was called “ribonucleic acid,” or RNA, which contained four structural chemicals: guanine, adenine, cytosine, and uracil, or GACU. The other was called “desoxyribonucleic acid,” or DNA, which was a major component of the chromosomes. They had deciphered its four bases—three identical to RNA, guanine, adenine, and cytosine, but with the uracil replaced by thymine—making the acronym GACT. They knew that these four bases consisted of two different pairs of organic chemicals; adenine and guanine being purines, and cytosine and thymine being pyrimidines. They also knew that they were strung together to form very long molecules. At first they thought that RNA was confined to plants, while DNA was confined to animals, but by the early thirties this was dismissed when it was found that both RNA and DNA were universally distributed throughout the animal and plant kingdoms. Still they had no knowledge of what nucleic acids actually did in the nuclei of cells.
A distinguished organic chemist based at the Rockefeller Institute, Phoebus Aaron Levene, proposed that the structures of DNA and RNA were exceedingly boring—they formed groups of four bases that repeated themselves in the identical repetitive formation throughout the molecule, like a four-letter word, repeated ad nauseam. This was called “the tetranucleotide hypothesis.” Such a banal molecule couldn't possibly underlie the exceedingly complex basis of heredity. In the words of Horace Freeland Judson, “the belief was held with dogmatic tenacity that DNA could only be some sort of structural stiffening, the laundry cardboard in the shirt, the wooden stretcher behind the Rembrandt, since the genetic material would have to be protein.”
Proteins are lengthy molecules made up of smaller organic chemical units known as amino acids. There are 20 amino acids in the makeup of proteins, reminiscent of the number of letters that make up alphabets. If genes were the hereditary equivalents of words, only the complexity of proteins could fashion the words capable of spelling out the narratives. Chemists, and through extrapolation geneticists, not unnaturally assumed that only this level of complexity could possibly accommodate the incredible memory template that the complexity of heredity demanded—a line of thought that Judson labeled “The Protein Version of the Central Dogma.”
This was the contentious zeitgeist that Avery now confronted. As early as 1935, in his annual reports to the Board of the Institute, he indicated that he had growing evidence that the “transforming substance” appeared free of capsular polysaccharide and it did not appear to be a protein.
Further progress on this line of research appeared to drag. In part this was b
ecause Dubos, working in the same department, had made a breakthrough in his search for antibiotic drugs. In 1925, Alexander Fleming, at St. Mary's Hospital in London, had discovered a potential antibiotic, penicillin, but he had been unable to take his work to the stage of useful production for medical purposes. Now, working on the philosophical principle encapsulated by the biblical saying “dust to dust,” Dubos had pioneered the search for microbes in soil that would potentially attack the polysaccharide coat of the pneumococcus. By the early 1930s, he was making progress. From a cranberry bog in New Jersey he found a bacillus that dissolved the thick polysaccharide capsule that coated the pneumococcus with its armor-like outer covering. Dubos went on to extract the enzyme that the cranberry bog bacillus produced. He and Avery had reported their discovery in a paper in the journal Science in 1930. In a further series of papers, the two scientists would report further experiments, all aimed at extrapolating the discovery to human trials of the cranberry bog enzyme in treating the potentially fatal pneumonia and meningitis caused by the pneumococcus.
But their research encountered difficulty after difficulty. In part these arose from a predictable ignorance in a field of such pioneering research. A more personal, and devastating, problem arose when, under the stress of it all, Avery developed thyrotoxicosis—a debilitating autoimmune illness in which his thyroid gland became overactive.
