Life's Greatest Secret

by Matthew Cobb

  With the help of NIH colleagues and highly skilled visitors such as Marianne Grunberg-Manago, Nirenberg’s lab was soon able to use a variety of techniques to synthesise triplets with known composition, and then put them through the Millipore ‘plater’ device to demonstrate which amino acid they coded for. One of Nirenberg’s trusted technicians, Norma Heaton, recalled that during this period the atmosphere in the laboratory was ‘intense … busy … crowded … competitive’. She described how they would suck up radioactive reagents with their mouths (‘that would never be allowed today’, she said) and that members of the lab would crowd round the data coming out of the radioactivity counter, eager to know what new codon had been discovered, ‘Then you would hear this shout, like “Oh, we discovered a new one.”’52

  Heaton also gave some insight into the prosaic work that takes place in a laboratory – exactly what has to be done to obtain the data that is interpreted to make scientific breakthroughs. The routine she described resembles the precise, repetitive gestures of a worker on a production line, which, in a way, is what she was:

  Initially we used single platers. It was a little round, stainless steel tube, just big enough to hold the Millipore filter, and about so high, and it screwed onto a base. You had a glass Erlenmeyer flask connected to a vacuum, and then you had a rubber gasket at the top, and you plunked this thing down.

  Then you had the vacuum on, and you took one of the test tubes that had your experiment in it, and you would precipitate the complex with TCA. Then you would pour it through the plater and the precipitated complex would be collected on the filter.

  Then you would unscrew it, take out the Millipore filter with forceps, and put it in order onto a piece of aluminium foil. Initially, we used what was called a Nuclear-Chicago planchet counter. You placed the dried filter onto little copper or aluminium planchets, about so big around, they had a little, tiny lip, and you would put the filter on that, and then you would stack them up and you would put them into the Nuclear-Chicago, and they would drop down and as they went across, the level of radioactivity would be counted. …

  It all had to be timed. When you got good at this, you knew how many seconds it took you to unscrew this single plater, take it out, put it down, set it up with a new Millipore. I think I got so I could do it every thirty seconds, or maybe every twenty-five seconds.53

  As the data came tumbling out of the Nirenberg laboratory, Crick, like so many before him, tried to find the reason why some amino acids were coded by more than one codon – the logic behind the degenerate nature of the code – simply by thinking about it. He wondered how the codon on the messenger RNA molecule bound with a complementary set of bases – what he called an anticodon – on the small transfer RNA (tRNA) molecule.* It was not clear whether there was one tRNA molecule per RNA codon (so sixty-four different versions), or one molecule per amino acid (in which case there would there would be twenty), or some intermediate situation. There was some experimental evidence that the tRNA that attaches to phenylalanine could recognise both the UUU and the UUC codons; to explain this curious phenomenon, Crick resorted to the precise molecular modelling that had preoccupied him during the race to discover the structure of the DNA molecule at the beginning of 1953. He came up with the idea that there was a degree of what he called molecular wobble in the binding of the third base in the RNA codon with its equivalent in the tRNA anticodon. Crick provided a masterly survey of the situation at the time, and then concluded with a smile: ‘In conclusion it seems to me that the preliminary evidence seems rather favourable to the theory. I shall not be surprised if it proves correct.’54

  It was correct – we now know that most organisms have more than twenty but less than sixty-four tRNAs (for example, there are forty-eight tRNA ‘anticodons’ in humans, but only thirty-one in bacteria), which is explained by the wobble in the anticodon’s ability to recognise more than one base in the third position of the codon.

  By the middle of 1965, Nirenberg’s group had identified the function of fifty-four out of the sixty-four RNA codons; at around the same time, Khorana confirmed these data by using synthetic codons of known sequence.55 None of the theoretical schemes that had been so carefully developed over the previous decade proved to be correct. The genetic code is highly redundant, so that in many cases a base in a codon can alter without changing the amino acid that is being coded for. Most of these silent changes in DNA occur in the third base – this was the reason why theoreticians had wondered whether in fact the code was basically a doublet code. In some cases the third base in the codon provides no additional information because all four alternatives code for the same amino acid, as a result of the wobble in codon–anticodon binding.

  Three interlinked issues remained: understanding which way the genetic message is read, and finding out how the cell knows where the genetic message begins and ends. A sequence of bases can be read in either direction, with completely different meanings: a DNA codon reading AGG codes for serine, whereas GGA codes for proline. Furthermore, because of the complementary nature of the two DNA strands a given stretch of DNA contains four possible alternatives – in this example, AGG and GGA on one strand, and TCC and CCT on the complementary strand. The genetic code seemed to be becoming even more complicated, but this mystery was soon solved.

  By 1963 a series of experiments using radioactively labelled mRNA showed that for each gene only one strand was read by the cell’s machinery, while in 1965, Ochoa’s group confirmed that the message was read by the ribosome in the same direction as mRNA was synthesised from the DNA strand, in what is known as the 5′ → 3′ direction (pronounced ‘five prime’ and ‘three prime’).56 These numbers refer to the way that ribose molecules, which are composed of five carbon molecules in a ring, are chained together in RNA and DNA by phosphate molecules attached to the fifth carbon of one ribose molecule, the third carbon of the next, the fifth carbon of the following, and so on (bases – A, C, G and T/U – are attached at the first carbon of each ribose).

  The direction in which the genetic message was read was now known, and it seemed logical to assume that there had to be some way in which the message set its reading frame, enabling it to be read correctly. To everyone’s surprise, in 1966 it was found that the sole RNA codon for methionine – AUG – also acts as a start codon for protein synthesis if it is at the beginning of a sequence.57

  A final enigma was the presence of three RNA codons that did not appear to code for any amino acids – nonsense codons. The first of these (UAG) was a mutation that had been identified in 1962 in phage and given the enigmatic name ‘amber’, apparently after the German translation of the name of one of the American scientists involved, Bernstein. The other two nonsense codons were identified by the Brenner laboratory, and given colour names to go with the original – ochre (UAA) and opal (UGA).58 In 1964, Brenner’s group showed that the amber codon – UAG – coded for stop. This also provided evidence to support one of the widespread assumptions of molecular genetics, namely that the gene and the protein it creates are colinear. By creating amber mutants at different points in the gene coding for the phage head protein, they showed that the lengths of the corresponding protein fragments in each mutant were correlated with the position of the amber mutation. At almost exactly the same time, a group at Stanford led by Charles Yanofsky also provided evidence of colinearity between the DNA sequence and protein structure, in Escherichia coli.59

  All of the assumptions that had underpinned the work on the genetic code over the previous years had proved correct: the code was universal, it possessed redundancy, its fundamental unit – the codon – was composed of three bases, the message was read in a particular direction, there was a reading frame, gene and protein were colinear, and the sequence contained simple instructions to start and stop reading. The detail had often been wrong, but as researchers groped towards the truth, using a mixture of theoretical insight and experimental ingenuity, the central principles they had clung to were all shown to be true.
/>
  It was not until 1967 that the last of the sixty-four words in the genetic code was read, appropriately in an article co-signed by Francis Crick.60 This was the opal codon – UGA. Like the amber and ochre codons, it read ‘stop’.

  *

  The June 1966 Cold Spring Harbor symposium was entirely devoted to the genetic code. The opening talks were by Nirenberg, Matthaei and Khorana. By choosing these speakers, the organisers – who included Watson and Crick – recognised the way in which the field had been transformed in the previous few years. The sessions looked at the direction in which the code was read, the role of ‘punctuation’ (initiation and stop, rather than commas), control of gene expression, a detailed exploration of the structure and function of tRNA, and a discussion of mutation and errors (under the title ‘Infidelity of information transfer’). None of the talk titles at the meeting even mentioned ‘information’, and all the presentations were framed in terms of biochemistry, not information theory. It was biology, and not mathematics, that dominated the celebration of the cracking of the code. The final two talks outlined two of the future developments that would preoccupy the field, down to the present day: the use of DNA sequences to establish patterns of evolution, and the origin and evolution of the genetic code.61

  The published proceedings of the meeting opened with an article by Crick that was a magisterial overview of the previous fifteen years’ work, entitled ‘The genetic code – yesterday, today, and tomorrow’. Crick looked back to the earliest ideas of Caldwell and Hinshelwood in 1950, Dounce in 1952 and above all Gamow in 1953, when the ‘coding problem’ suddenly came into focus with the discovery of the double helix structure of DNA. Crick did not refer to Schrödinger’s code-script idea – he clearly did not feel that this insight had any direct effect on subsequent events. Nor did he cite Boivin’s insight into the relation between DNA, RNA and protein; he may not have known of it.

  Crick’s opening words accurately summed up the situation: ‘This is an historic occasion.’62 To all extents and purposes, the genetic code was known. Cracking the code underlined the power of experimentation: none of the attempts to work out the code theoretically had got it right. Crick argued that it had been inevitable that they would fail:

  We can see now, from the known code, that it would have been almost impossible to have deduced it correctly at the time.

  None of the hypothetical codes dreamt up by the theoreticians were correct, because they made assumptions that were logical, rigorous and hopelessly wrong. The physicists’ appetite for elegance and the biochemists’ naive assumptions about natural selection led them to assume that the code had to be extremely economical, that it would look as though it had been designed along logical principles. But that is not how biology works. The genetic code is a product of biology and is messy, illogical and inelegant. It is highly redundant, but to bewilderingly varied degrees: one amino acid (leucine) has six codons, whereas another (tryptophan) has only one. Explaining this pattern on the basis of chemical, physical or mathematical principles has so far proved difficult. Whatever logic there may have been has been overlain by billions of years of evolution and chance events. As Jacob put it in 1977, natural selection does not design, it tinkers with what is available.63

  Much of the work that contributed to the code brought worldwide acclaim. The Nobel Prize committee repeatedly rewarded those who had made the essential breakthroughs. Joshua Lederberg, George Beadle and Ed Tatum won Nobel prizes in 1958 for their work on microbial genetics, which had transformed the way that genes were understood and could be studied, and Watson, Crick and Wilkins won in 1962 for the double helix structure of DNA. In 1965, Jacob, Monod and Lwoff were awarded the prize for their work on the repressor and the genetic regulation of protein synthesis, and in 1968 Nirenberg, Khorana and Robert Holley won for their work on the genetic code, nucleotide synthesis and tRNA structure respectively. To celebrate, Nirenberg’s lab hung a banner across a corridor that read ‘UUU are great Marshall’.64 In the following year, the founders of the phage group – Delbrück, Hershey and Luria – were awarded the prize for their work on virus genetics, through which they had set the stage for many of the fundamental discoveries of molecular genetics. Other participants in the race either did not win the prize (Benzer, Matthaei) or were awarded one for other work (Brenner, Ochoa). By the end of the 1960s, a period had closed. Many of the leading figures, such as Crick, Benzer and Nirenberg, turned their attention to neurobiology, while Brenner became interested in developmental biology. In 1968, phage group member Gunther Stent gave an overview of the period under the elegiac title ‘That was the molecular biology that was’.65

  *

  The discoveries made in this period of science have transformed the whole of biology, and enabled us to make massive advances in the development of new medical treatments. As well as producing a revolution in our knowledge, the twenty-two years that separated Avery and Schrödinger from Crick, Nirenberg, Jacob and Monod also produced a revolution in our thinking. Everyone now knows that genes contain information, and that they function as part of complex networks, controlling the production of proteins and the activity of other genes. The informational and cybernetic theories that flourished in the late 1940s and 1950s have left little direct trace in the thinking of today’s scientists, but their influence remains, in terms of the metaphors and concepts that we all use to think about some of the most fundamental features of life. The way in which we now think about genes and how they work would have been incomprehensible to a scientist from the 1930s. Cracking the genetic code was a leap forward in humanity’s understanding of the natural world and our place within it, akin to the discoveries of Galileo and Einstein in the realm of physics, or the publication of Darwin’s On the Origin of Species. These comparisons are not the fruit of hindsight, they were made at the time.66

  The discovery of the code did not occur through the genius of a single person or even a handful of thinkers. Instead, it required brilliant insights, audacious experiments and above all a lot of hard work by a great many people. The range of techniques that was used was enormous, with particular emphasis on the interface between physics and biology. Above all, cracking the code represented the triumph of experimentation over theory. None of the purely theoretical approaches was able to uncover the secret, which finally revealed itself through the probing of the experimenters, not the pencil-chewing of the theoreticians.

  Unlike previous collective scientific breakthroughs, such as the Manhattan Project, or subsequent ones, such as the landing of men on the Moon, the Large Hadron Collider or the Human Genome Project, there was no concerted organisation of the research. Institutions such as the Medical Research Council in the UK, the NIH in the US, and the Institut Pasteur and the Centre national de la recherche scientifique (CNRS) in France all supported key researchers, but the work had not been coordinated; there had been no committee or council that oversaw the project. There had been funding for the wartime work that led to cybernetics and information theory, but at the time the potential implications of that research were unknown, beyond their immediate practical use in the war effort. Whereas Cold War governments poured substantial amounts of money and resource into nuclear research and rocket science, and the private sector spent billions developing computers and new medicines, the amount of money that flowed into molecular biology and cracking the code was tiny, and several researchers – Avery, Watson and Crick, Nirenberg – initially received minimal support for their work.

  Apart from the meetings at Cold Spring Harbor, there had not even been a regular meeting-place for researchers to discuss progress, and the clubby nature of those meetings had excluded the two unknowns who ultimately made the breakthrough. Equally, there was no undisputed leader driving the project forward with a clear vision. Gamow had inspired the RNA Tie Club, but his broader influence was less strong. Crick was omnipresent, but he did not have the power to promote or reorient areas of research or particular researchers and he had no control over fundin
g.

  Above all, unlike many other great examples of collective research, the solution to the problem was not known in advance – this was not like the Manhattan Project, for which the end point was clear from the outset. The research into the genetic code was not an engineering problem: it was pure research, which could not simply be cracked by hard work, and the researchers involved could not be given a small part of the problem to solve, knowing that it would contribute to the eventual outcome. Crick was not the Oppenheimer of the genetic code. His intelligence, criticism and encouragement did much to create the essential insights that shaped how the science developed, but although he came up with the idea that ‘the precise sequence of the bases is the code which carries the genetical information’, the thirteen-year-long campaign to crack the code had not been led by him, nor did he foresee even the outline of the eventual solution.

  The path taken by the scores of researchers who had been involved in the coding problem was as unpredictable and on occasions as illogical as the code itself seems to be. Cracking the code was an example of what Jacob called ‘night science’, in which intuition and audacious guesswork accompany strict logic.67 In 1966, John Cairns, the director of Cold Spring Harbor Laboratory, accurately summarised the significance of what had just taken place:

  The effort that has gone into this decipherment, the strange sense of urgency, and the remarkable variety of approaches that have together led to the solution, must be without parallel in the history of biology.68

  * Eck began his paper by accepting the slight possibility that the hereditary material was made of nucleoproteins, not DNA. This was the latest doubt about the genetic role of DNA that I have found.