Already in 1938 when I started work in this field I was of the opinion that nucleic acids might be involved in the transmission of hereditary characteristics as had been suggested by the earlier work of Griffith on pneumococcal transformation; Avery's demonstration in 1944 that deoxyribonucleic acid (DNA) was the transforming factor seemed to me to settle the issue. Curiously enough Avery's work - so beautiful and, to me, so convincing - did not convince everybody. In the summer of 1946 I attended a symposium on nucleic acids held in Cambridge by the Society of Experimental Biology to which I agreed to contribute a paper on 'The structure and synthesis of nucleotides'. At that symposium I remember a violent argument between E. Stedman, who stoutly maintained that histones and not nucleic acids were the carriers of hereditary characteristics, and a number of others and notably Caspersson who, like me (and with better evidence), was a proponent of nucleic acid. I think it was really at that meeting that I first met and talked with the main operators in the biology and biochemistry of nucleic acids, and heard from Astbury about his X-ray studies. My interest was aroused, and I began - again for the first time -seriously to consider the chemical structural problems presented by the two types of nucleic acid - ribonucleic (RNA) and deoxyribonucleic (DNA). The time was in any case propitious, because we were just on the point of completing our first ATP synthesis and we now knew sufficient about phosphates and their behaviour to make nucleic acids, chemically speaking, a bit less daunting than they were to most people at that time. So we began to think about the problem a bit, and to study the behaviour of simple nucleotides alongside our coenzyme work.
I was, of course, familiar with most of the chemical literature on nucleic acids and their component nucleotides. As a result of his work extending over many years, P. A. Levene had substantially clarified the structure of the simple nucleotides and nucleosides which can be obtained by hydrolysis, and deduced correctly that the nucleic acids were made up of nucleotides linked together in some way through phosphate residues. But he had, based on analytical methods we now know to have been inaccurate, reckoned that only two nucleic acids existed - one from plants (RNA) and one from animals (DNA) and that each was composed of four nucleotides present in equal amounts. More unfortunate still, he supported the idea that the nucleic acids might be simply tetranucleotides which formed colloidal aggregates in solution. From the moment I read his claims and views I found myself in total disagreement. For one thing, there was already some evidence to suggest molecular weights of 500 000 or more for DNA, and, in any case, its general properties suggested strongly that it was a macromolecular substance like protein or one of the polymeric materials studied by Staudinger, and even then appearing in commerce in the form of synthetic rubber and synthetic fibres. Nor could I accept the idea that we were dealing with polymerised tetranucleotide units; on the evidence available I had doubts about the constant composition of both types of nucleic acids and the claimed existence of only two acids. These views were, of course, fully confirmed during the years that followed, but belief in the so-called 'tetranucleotide hypothesis' was, in my view, largely responsible for the slowness with which biochemists and biologists came to realise the importance of the nucleic acids in hereditary transmission. I have sometimes wondered whether the ready acceptance of the tetranucleotide hypothesis by many biochemists was not, perhaps, due to their belief that proteins with their manifold properties would be found to be responsible for all life processes, and they accordingly felt no need to look any further!
I have referred to the daunting nature of chemical studies in the nucleotide/nucleic acid field. At the time of which I am writing, and, indeed, throughout virtually all the work on nucleic acid components - and even on the basic purines and pyrimidines let alone their phosphorylated derivatives - poor solubility in organic solvents, difficulty of separation and purification, and lack of proper melting points or other reliable criteria of purity, made life very difficult for the chemist, and led to a lot of confusion and, indeed, errors in the literature. In those days we had only the early forms of chromatography available, and, apart from ultraviolet spectroscopy, and in the late stages of our work some minor applications of the newly developing technique of infra-red spectroscopy, we had to rely on the traditional methods of the organic chemist developed many years before for compounds quite different in type and in physical properties. Even in our own work we were led by such deficiencies into errors in the identification of simple nucleotides, at one point, by accepting the wholly invalid 'proof of the structure of benzylidene-adenosine due to Gulland, and had to do quite a bit of work before we realised the true position. As we went along several observations began to loom large in our thoughts. For one thing, we were struck by the astonishing difference in the ease with which RNA underwent hydrolysis as compared with DNA and probably related to this the ease with which one could purify deoxyribonucleotides as compared with ribonucleotides. To get a really pure specimen of yeast adenylic acid (adenosine 3'-phosphate) was a very difficult task indeed, since recrystallisation seemed often to operate in the reverse direction; this should have suggested phosphate migration to us as a possible reason, but I am afraid it did not immediately do so. What was, of course, clear, was that the difference between the stability of the two nucleic acids lay in the sugar portion of the molecule, and this also suggested to me that, while the stable DNA was involved in hereditary transmission, RNA probably had some quite different function in nature, where its impermanence would be an advantage. As far as we were concerned the final breakthrough came in 1949 when Waldo Cohn at Oak Ridge, Tennessee, applied ion-exchange chromatography to alkaline hydrolysates of yeast ribonucleic acid, and isolated, not just the nucleoside 3'-phosphates hitherto regarded on the basis of studies by Levene and others as the sole products, but also the 2'-phosphates. In Cambridge my colleague D. M. Brown drew my attention to some earlier and virtually unnoticed work by Fono on the hydrolysis of glycerophosphates, and suddenly the whole jigsaw fell into place, and we could explain all the hitherto puzzling facts about RNA and understand fully the differences in chemical behaviour between RNA and DNA. Moreover, we were able to establish conclusively by synthetic studies the reality of the alkaline degradation of RNA, first to four cyclic nucleotides and thence to an equilibrium mixture of the 2'- and 3'-phosphates, which undergo interconversion in an acid medium, and we were able to explain the nature of ribonuclease action and its significance for RNA structure. Having done this, Dan Brown and I were able to put forward definitive structures for the two types of nucleic acid as unbranched 3:5-linked polynucleotides in 1951 at the 75th Anniversary Meeting of the American Chemical Society (our detailed paper did not actually appear in print in the Journal of the Chemical Society until January 1952).