The ultimate goal: Sequencing DNA
Following his success in sequencing RNA, Sanger began to turn his attention to how he might sequence DNA. This he saw as his 'ultimate goal'. Before his sequencing of RNA this had seemed an impossible project. One of the major challenges with DNA was its size and the fact that there were no small DNA entities comparable to tRNA or 5S ribosomal RNA on which he could develop a sequencing method. So far the only smallest form of DNA available was from the single-stranded bacteriophage phi X174. This had been purified in 1959 by Robert Sinsheimer at the California Institute of Technology. A year later some more DNA was purified from the lambda bacteriophage by Dale Kaiser and David Hogness at Stanford University. DNA sourced from bacteriophages was thought to be made up of approximately 5,000 nucleotides. This contrasted with other more interesting DNA which were thought to contain millions of nucleotides (Sinsheimer, 1959; Kaiser, Hogness, 1960; Sanger, Dowding, 1996).
In addition to the absence of any suitable DNA to experiment on, Sanger did not have access to any suitable enzyme to cut DNA. When working on RNA he had been greatly helped by the availability of ribonuclease T1 which cut the nucleic acid at very precise points. While scientists had discovered a number of enzymes for cutting DNA, deoxyribonucleases (DNAase), none of them had so far proved a reliable tool. (Sanger, Dowding, 1996).
In 1964 Sanger welcomed into his team Ken Murray, a British postdoctoral biochemist who soon joined in his efforts to find a way to sequence DNA. Early on Murray developed a two-dimensional fractionation technique for determining the sequence of tetranucelotides, a particular block of nucleotides made up of four base pairs (adenine (A) and thymine (T), cytosine (C) and guanine (G)). He obtained the tetranucleotides from DNA that he had partially digested with various enzymes. Yet, while Murray had managed to sequence the tetranucleotides, his method had a major downside: it could not pinpoint where on the original DNA nucleotide chain the tetranucleotides were located. This posed a major hurdle in terms of sequencing whole DNA (Murray, 1970; Finch, 2008).
Ken Murray. Credit: Geoffrey Argent Studio. The son of a Yorkshire miner and school caretaker, Murray had been forced to leave school at the age of 16 because, despite having a strong academic track record, his family could not financially support his higher education. Thereafter he worked as a laboratory assistant at Boots the Chemist and then at an outstation of Glaxo. The pharmaceutical company subsequently gave him a scholarship to undertake a part-time chemistry degree at the University of Birmingham. Murray subsequently did a doctorate and then joined J Murray Luck at Stanford University purifying and characterising histones, a group of proteins closely associated with DNA. Found in the nuclei of eukaryotic cells, histones play an important role in the packaging and ordering of DNA into structural units called nucleosomes and help in the regulation of genes. Murray's source for these histones was thymus glands from calves (Brammar, Gratzer, 2014).
In addition to Murray, Sanger had the assistance of his technician Bart Barrell and of Alan Coulson, another technician who joined him in 1967. Together they began looking for ways to improve the sequencing method. So far Sanger's approach had been to break down a large molecule into smaller fragments, using enzymes, and then to break these fragments down further and piece together the sequence from the fragments' overlaps. While this degradation technique had proven effective for sequencing insulin and short strands of RNA, it did not seem practical for sequencing such a large molecule as DNA.
Bart Barrell, c.1975. Credit: Bart Barrell, Laboratory of Molecular Biology (LMB). Barrell started working as Sanger's personal assistant in 1964 when he was 18 years old, just after leaving school. He developed a passion for natural history from an early age, and became fascinated by molecular biology while still at school. Before leaving school he presented a lecture on the structure of proteins and nucleic acids to the School Scientific Society. Barrell had turned down a County Major Scholarship to attend university because he wanted to start work in science immediately. Quickly impressed with Barrell's enthusiasm and enterprise, Sanger often discussed his ideas with him for advancing his sequencing methods. Barrell soon developed an expertise in sequencing and analysis and was entrusted by Sanger to apply it on ever more challenging molecules, leaving Sanger the freedom to dedicate his time to advancing the technology.
Alan Coulson, 1987. Credit: Alan Coulson, LMB. Coulson joined Sanger in 1967 at the age of 20, shortly after completing a higher national diploma at Leicester Polytechnic. Sanger was very impressed with Coulson's skills and regarded him as his main collaborator in the laboratory, describing him as 'something of a wizard with sequences' (Sanger, 1988). They were both self-effacing and quiet so were well-matched in terms of their temperaments. Coulson carried out most of the work involved in growing bacteria in culture for multiplying bacteriohages, extracting their genetic material and producing radiolabelled DNA fragments, and was responsible for running these on the acrylamide gels to produce autoradiographs. Following Sanger's retirement, Coulson joined John Sulston and helped in the sequencing of the nematode worm C. elegans.
A possible new tool: DNA polymerase
For inspiration, Sanger looked to the work of Ray Wu and A Dale Kaiser, two molecular biologists based respectively at Cornell and Stanford Universities who had partially sequenced DNA from a lambda bacteriophage in 1968. What had helped them achieve this feat was the fact that this bacteriophage's DNA was a linear molecule with two single-stranded cohesive DNA fragments at the end. Wu and Kaiser had managed to sequence 12 nucleotides at the end of the two end DNA fragments. This was the first piece of DNA to have ever been sequenced. Wu and Kaiser's method involved a number of steps. Many of them were reminiscent of Sanger's own techniques. For example, they used the radioactive 32P isotope to label some of their nucleotides. They also broke the DNA molecule down with an enzyme. Analysis of the sequences was also carried out with the aid of ionophoresis on paper, and a two-dimensional electrophoresis method (Wu, Kaiser, 1968).
Ray Wu. Credit: Cornell University. Chinese born, Wu spent a year working with Kaiser in his laboratory at Stanford University between 1965 and 1966. He took the lead in developing the technique sequencing the lambda bacteriophage's DNA. Click here for more on Wu's work and life.
Armin Dale Kaiser. Credit: Dale Kaiser. Kaiser began working on the genetics of the lambda bacteriophages while doing a doctorate at California Institute of Technology, completed in 1955. He joined Stanford University in 1959, where he helped establish a new biochemistry department. Sanger would have become familiar with Kaiser between 1964 and 1965, because Kaiser spent a year in the LMB as a postdoctoral research fellow working alongside Hugh Huxley and Aaron Klug.
What was novel about Wu and Kaiser's method was that they made use of an enzyme known as DNA polymerase. First discovered in 1955, this enzyme is instrumental in DNA replication. In this process an enzyme, known as helicase, unwinds a double-strand of DNA into two single-strands to serve as templates. Following this, the polymerase combines the template strand with another strand of nucleotides, known as a primer, to generate a new double-strand of DNA. Both the template and primer are vital to this process, because the enzyme cannot perform synthesis independently (see figure 1).
Figure 1: Diagram showing the role of DNA polymerase in the duplication of DNA.
Wu and Kaiser had used the polymerase to generate small fragments of radiolabelled DNA (Wu, Kaiser, 1968). While intrigued by Wu and Kaiser's use of polymerase, Sanger had some reservations about adopting their method. Firstly, it entailed a lot of 'tedious' steps. Secondly, he doubted whether it could be used to sequence any DNA stretches longer than those found at the end of the lambda bacteriophage's genome (Sanger, 2001).
In addition to Wu and Kaiser's technique, Sanger began to look at one developed by Charles Weissmann and some colleagues at Zurich and Bristol University who had sequenced 175 nucleotides in RNA from the Qß bacteriophage. They did this by using a polymerase contained in the bacteriophage's own RNA. This polymerase had the advantage that it could synthesise complementary copies of the single-stranded RNA chain. With this the team were able to obtain homogeneous preparations of short, radioactive Qß RNA segments of any desired length. Weismann's group had also pioneered a short-term radioactive labelling technique (known as pulse-labelling) to produce radioactively labelled nucleotides, which they had used to deduce the RNA sequence. What was appealing about the team's work was that it offered an alternative approach to partial hydrolysis (Billeter, Dahlberg, et al, 1969; Sanger, 1980; Finch, 2008).
Soon after exploring the two methods, Sanger went to Sydney, Australia where he paid a short visit to the laboratory of Geoff Grigg, an Australian biochemist. One conversation he was to have in this laboratory was to prove invaluable to his thinking. This took place between himself, Grigg and John Griffiths, a British mathematician who had worked with Crick and was now spending a sabbatical in Grigg's laboratory. The discussion centred on a paper published by Paul Berg in 1969, which demonstrated how DNA polymerase I could be used to insert ribonucleotide triphosphates, RNA precursors, into a growing nascent DNA chain if magnesium was substituted with manganese in the polymerising reaction. This provided a means to substitute a specific deoxynucleotide with a specific ribonucleotide in the DNA chain. Such substitution opened up the possibility of cleaving a DNA chain at very specific points with an enzyme. If the method worked, Sanger would be able to obtain small enough DNA fragments with particular nucleotides on their end that would allow him to use the same sequencing methods as he had used with RNA. The key issue was how to obtain an appropriate template to kick-start the process. Griffiths suggested one solution might be to synthesise a primer from scratch. This primer would provide a tool to copy a DNA template. He hypothesised the primer could be just 6 to 10 nucleotides long (Grigg, 2001).
Once back in England, Sanger began looking for a way to implement Griffiths' idea. The first challenge was deciding which sequence to synthesise. For inspiration, he turned his attention to some work that he and his team were carrying out on a filamentous bacteriophage (f1 phage). They had already determined the amino acid sequence in its coat protein. It was made up of just two amino acids, methionine and tryptophan. These were known to correspond with a unique nucleotide sequence which was easy to figure out. It was this sequence that Sanger decided to synthesise (Sanger, 1992).
The next hurdle was to synthesise a primer to match the DNA template from the f1 phage. This was important because polymerase cannot work without the presence of both a template and primer. Nucleotide synthesis, however, was extremely difficult. Up to now the only person to have achieved this feat was Har Gobind Khorana at the Massachusetts Institute of Technology (MIT). Click here to find out more about Khorana.
Not possessing the necessary chemical expertise, Sanger invited Hans Kossel, a German molecular biologist from the University of Freiburg, whom he met by chance at a conference, to join him at the LMB to collaborate on the project. Kossel had spent some time working alongside Khorana on nucleotide synthesis and had independently been toying with the idea of making a similar sequence to the one proposed by Sanger (Sanger, 1988; Sanger, 1992).
Hans Kossel. Credit: Raina Maier. Between between 1964 and 1967 Kossel worked in the laboratory of Khorana at the University of Wisconsin on the elucidation of the genetic code. Thereafter he worked in the newly founded Institute of Biology, Freiberg, focusing on the enzymatic hydrolysis of tRNAs and on the synthesis of oligonucleotides. He was one of the first scientists to suggest the possibility of enzymatic synthesis of DNA as a DNA sequencing method. Sanger invited Kossel to work on these areas with him, which he did for a short period in 1971 and 1973.
It took Kossel and his colleagues over a year of hard work to make the primer. The primer was eight nucleotides long (octanucleotide). Sanger and his team mixed the primer with the f1 DNA template, DNA polymerase I and some radioactive nucleoside triphosphates. This they did with the help of John Donelson, an American postdoctoral biochemist researcher. Donelson had arrived in Sanger's laboratory in 1970. He was particularly suited to the task, having spent some time working with Wu at Cornell University (Donelson, Wu, 1971).
To the team's delight the octanucleotide proved an efficient primer. Much to their surprise, however, it did not replicate the DNA sequence found in the phage's coat protein. Despite this setback, the primer proved sufficiently specific for Sanger's purposes. Importantly, it provided a suitably sized piece of radioactive DNA to start testing different sequencing techniques without the need to carry out partial hydrolysis (Sanger, 1992).
In 1973 Sanger and his team published a paper that indicated that they had sequenced 50 nucleotides in the f1 DNA. The sequencing had been determined with the two-dimensional homochromatography method Sanger had devised for sequencing 5S ribosomal RNA and another two-dimensional fractionation system, known as the wandering spot method. Developed by Victor Ling, another member of Sanger's team, the wandering spot technique made it possible to read off a sequence based on the relative position of specific spots (Sanger, Dowding, 1996).
To do such sequencing, the team had partially cut up the synthesised DNA fragments with the enzyme pancreatic ribonuclease. This had cleaved the molecule at specific points where the ribonucleotide cytosine was found, providing fragments which all ended with this ribonucleotide. The sequencing was then done by piecing together the fragments' overlaps. Sanger had used the same approach for both insulin and RNA (Sanger, Donelson, Coulson, et al, 1973; Sanger, 1992; Sanger, Dowding, 1996).
The sequencing of f1 DNA was a major step forward. Not only had Sanger managed to sequence a significantly longer piece of DNA than anyone else so far, he had demonstrated the possibility of using polymerase and the ribosubstitution method for DNA sequencing. Nonetheless, determining the sequence had involved many different steps and was hard-work. The need to carry out ribosubstitution would soon be made redundant with the availability of restriction enzymes from 1970. Such enzymes recognised and cleaved DNA at specific nucleotide sequences. This meant that scientists could now easily cut up large DNA molecules into fragments with specific ends that could function as a starting point for sequencing (Hutchison, 2007).
This shows the different separation steps involved in the sequencing of f1 DNA. Credit: Sanger, Donelson, Coulson et al, 1973, figure 3.
The plus/minus technique
While Sanger and his team had managed to sequence a substantial portion of the f1 DNA, they had not completed its whole sequence. Moreover, their method was time-consuming and laborious. Eager to find a more rapid technique that would allow him to determine larger DNA sequences, Sanger continued looking for new avenues to tackle the problem.
In 1975 Sanger and his assistant Coulson published a paper (Sanger, Coulson, 1975) in which they outlined an approach radically different from the one before. What was novel about their technique was that was that it was not reliant on piecing together a sequence based on the examination of the overlaps in small DNA fragments. Furthermore, it enabled the direct visual scanning of a sequence. Sanger and Coulson had not developed the method overnight. Rather it gradually emerged as a result of some experiments, launched in April 1973, designed to test the action of different polymerases and the efficiency of different fractionation techniques (Sanger, Dowding, 1996; Garcia-Sancho, 2010).
These notes are from experiment D80. It was one of Sanger's first experiments where he began testing the possibility of copying with all four deoxytriphosphates. Sanger's DNA laboratory notebook, 1972. Credit: Wellcome Library, file SA//BIO, P/1/42.
In this note Sanger records starting 'a fairly ambitious experiment'. The aim of the experiment, labelled D93, was to find a way of extending DNA fragments, cleaved with a restriction enzyme, by the addition of all four of DNA's nucleotides, which were then to be separated with ionophoresis on gels and their sequences analysed. Results from this experiment were fairly promising. Sanger would continue to refine the method used in this experiment with the aid of Coulson and Barrell. The experiment laid the foundation for the plus and minus method. Sanger's DNA laboratory notebook, 1973-1974. Credit: Wellcome Library, file SA//BIO, P/1/43.
Eventually called the 'plus/minus' system, the technique consisted of a number of different steps. In the first instance a primer was synthesised. This was a complementary copy of a short length of the DNA sample under study. Once made, the primer was bound to a specific complementary region on a single-stranded DNA template so as to provide a starting point for DNA synthesis. The combined strands were then incubated with polymerase I (sourced from Escherichia-coli) so that it could add radioactively labelled nucleotides on to the 3' end of the primer. This process provided a random assortment of synthesised DNA fragments of varying lengths. Each was purified to remove excess nucleotides.
Following purification the samples were divided between two treatments. The first treatment, known as the 'minus' system, drew on the copying method devised by Wu and Kaiser. In this case the sample was incubated in a test tube with DNA polymerase I (sourced from Escherichia coli) together with just three out of four of DNA nucleotides (A, C, G, or T). One nucleotide was deliberately withheld from this incubation process. This took advantage of the natural biological process, whereby polymerase will carry on synthesising a DNA chain until it encounters a missing nucleotide. In the case, for example, where adenine is absent, polymerase will synthesise a string of nucleotides, coming to a halt just before this nucleotide. By leaving out a particular nucleotide, one at a time, the team were able to obtain DNA products with specific nucleotide endings.
In contrast to the 'minus' system, which relied on the removal of one specific nucleotide, the second treatment, the 'plus' system relied on the addition of one during the incubation process. Developed by Paul Englund between 1971 and 1972, the 'plus' system made use of a different polymerase, T4. This was sourced from the T4 bacteriophage. In normal circumstances this polymerase acts as a synthesising reagent. However, it will start to downgrade a DNA strand from its 3' end when it encounters an additional nucleotide. As in the case of the 'minus' system, the 'plus' system produced DNA fragments with specific nucleotide endings.
Coulson's note recording tests with T4 polymerase. Credit: Wellcome Library, Coulson papers, file: PP/COU, Notebook 'Plus and minus sequencing, T4 polymerase preparations', 1976.
Both systems relied on the preparation of four test tube reactions with the two different types of polymerase. Following treatment, the primers were separated from the template. This provided eight different DNA strands. Four of these came from the template, known as minus strands, and four came from the primer, known as plus strands. The end result was a series of overlapping fragments with each of the nucleotide endings (A, C, G, T).
Each DNA fragment was then placed alongside each other in separate lanes on acrylamide gel ready for ionophoresis. Following the application of an electrical charge the DNA fragments would migrate from the top to the bottom of the gel, with the bigger fragments moving slower than the smaller ones. In this context the gel acted like a sieve which helped sort out the fragments by their size. Once this was completed, an X-ray film was overlaid on the gel for a period of time, usually overnight, and then developed to produce an autoradiograph, which revealed the position of the radioactive tagged nucleotides in the fragments. The final image showed a series of dark bands in a ladder-line pattern which highlighted where the fragments overlapped (Hutchison, 2007; Finch, 2008).
The use of acrylamide gel for sequence analysis marked a significant change. Previously Sanger had typically carried out such work through the use of two-dimensional paper chromatography. This, however, was time-consuming because it involved lots of cutting out of bands. One of the attractions in using the acrylamide gel system was that it provided a one-dimensional reading of a sequence. The decision to use gels for fractionating DNA had initially seemed a crazy idea to Sanger. While gels were commonly used to separate intact proteins as well as DNA and DNA fragments, they had never before been used to separate individual nucleic acids (Sanger, 1992).
This shows John Donelson who played a pivotal role in the development of the acrylamide gel system for the plus and minus system. Credit: John Donelson.
When starting to develop the plus and minus method Sanger and Coulson largely carried out fractionation with homochromatography. However, they soon turned to Donelson to help with the process, which he did using an acrylamide gel-based system that he was developing for some other experiments. At first this system had appeared unpromising because initial results proved only marginally better than those obtained from homochromatography. Some progress was made through the adoption of larger gels. This produced sharper bands in autoradiographs. Yet teething problems continued. On a number of occasions the larger fragments ran faster than the smaller ones. This completely upset the sequence. Moreover, when testing the system with DNA synthesised with the ribosubstitution method for which a sequence was already known, they kept getting inverted sequences. This they attributed to a secondary structure in the DNA template. Eventually, after a lot of trial and error, the problem was resolved by the addition of 8M urea, a chemical, and the application of a high voltage, the combination of which made the gel become hot. After all these changes the team began to get much clearer results than with paper chromatography (Sanger, 1992; Sanger, Dowding, 1996).
The combination of the plus and minus method in tandem with the acrylamide gel-based fractionation system was a major step forward. Now a sequence could be read directly from an autoradiograph. This was done by scanning the picture from the bottom to the top, noting where the dark bands, the product of radioactivity, appeared. All of this could be done with the naked eye. Sanger considered this new approach one of the best ideas he ever had. The new method not only allowed researchers to directly scan a DNA sequence, but also to determine much longer stretches than ever before, 50 nucleotides at a time. It also marked a departure from the use of degradation to a copying procedure. (Sanger, 1988; Finch, 2008).
phi X 174 bacteriophage
Soon after developing the plus and minus method, Sanger's team started using it to determine the sequence of a single-stranded circular DNA from the phi X 174 bacteriophage. This was part of an ongoing project in Sanger's laboratory. The work involved sequencing larger and larger pieces of the bacteriophage's DNA. Initially this was done using Sanger's earlier sequencing methods, but by 1975 the team began largely using the plus and minus method. Much of the work was undertaken by Elizabeth Blackburn and Gillian Air. Another important person on the team was Clyde Hutchison. He had worked on phi X 174 bacteriophage since the 1960s, having studied it as part of his doctorate under the supervision of Robert Sinsheimer at Caltech. Hutchison not only had expertise in the bacteriophage's genetics, but brought with him a mutant strain of the bacteriophage which enabled the team to do some comparative sequence analysis. Later on some additional researchers joined the project: Nigel Brown, John Fildes, Pat Slocombe and Michael Smith.
By the end of 1976 the team had determined virtually the whole sequence of the bacteriophage's DNA. It was over 5,000 nucleotides long. Much of the sequencing had been carried out with the plus and minus technique. It had proven a highly effective tool for the rapid sequencing of long stretches of DNA. Within a short time the whole sequence had been completed, published in 1977. It contained 5,386 bases. The achievement marked a major breakthrough: it was the first complete genome to be sequenced (Sanger, Air, Barrell, et al 1977; Finch, 2008).
Being able to compare the DNA sequence from the mutant strain with the normal bacteriophage had also yielded some unexpected results. During this research the team had found that the 5,326 bases coded for ten genes, and that four of these overlapped. This suggested that the same string of DNA was being used to specify two different proteins. Which protein it coded for depended on the context in which it was read. Such findings challenged a hypothesis, first put forward in 1941, that genes were arranged in a linear order along the genome with one DNA sequence coding for one protein sequence. The discovery also destroyed any idea that the sequence alone contained all the information about an organism, even one as a simple as a bacteriophage. Clearly a DNA sequence was not sufficient without other extra-sequential information. One of the questions this posed was what counts as a gene and when (Sanger, Dowding, 1996; Brown, 2003).
An alternative approach: Maxam-Gilbert
Around the same time that Sanger and Coulson developed the their plus and minus method, another sequencing technique was emerging across the Atlantic. Developed at Harvard University by the American biochemist Walter Gilbert and his doctoral student Allan Maxam, this technique became available the same year as the plus and minus method. The technique shared many similarities with the one developed by Sanger and Coulson. Not only did it make use of radioactively labelled DNA fragments generated with restriction enzymes, it also deployed acrylamide gels to separate the nucleotides in the fragments and used autoradiographs to deduce the sequence. Where the two methods diverged was how they generated segments with particular nucleotide endings. In contrast to Sanger and Coulson, who achieved this by harnessing the copying mechanism of polymerase, Maxam and Gilbert did this by partial degradation which they accomplished with the help of one of four base-specific chemicals (Sanger, 1988).
By the late 1970s the Maxam-Gilbert technique had become more popular than the plus and minus method, particularly among scientists based in the United States. In part this was because it produced bands for every sequence position. The other attraction was that it could be used for sequencing double-stranded DNA. By contrast the plus and minus method was only applicable to single-stranded DNA (Hutchison, 2007; Garcia-Sancho, 2010).
Sanger would later admit that he was not 'altogether overjoyed' by the appearance of the Maxam-Gilbert technique, but considered it an 'ingenious' method. He himself had considered devising a similar chemical method, but had rejected this in favour of enzymes based on his previous work on proteins and RNA, which had shown enzymes to have greater specificity than chemical reagents. His earlier experiences had led him to assume that any rapid degradative sequencing method would need a chemical with very specific reactions. Maxam and Gilbert's method, however, showed that the specificity of chemical reactions did not matter and that strong data could be obtained without specific reactions (Sanger, 1988).
Walter Gilbert. Credit: National Library of Medicine.
In fact Sanger was not totally satisfied himself with the plus/minus system. While much better than his previous methods, it still had some problems. One difficulty was that bands on the autoradiographs varied in strength and were sometimes totally absent. Moreover, he and Coulson had found that the bands tended to get bunched together on a relatively small area of the gel when just one incubation was carried out. To some extent this problem could be resolved by carrying out preliminary incubations, where samples were incubated with a nuclease for varying lengths of time. Such incubations provided bands of varying sizes with a wider distribution on the gel. They, however, were time-consuming so were not ideal (Sanger, Dowding, 1996).
After making little headway with other methods, Sanger and Coulson began to investigate the possibility of using dideoxynucleotides (dideoxys), a DNA nucleotide that can inhibit the action of DNA polymerase and block its attachment of nucleotides to a growing DNA chain. Sanger and Coulson's decision was inspired by some experiments conducted by Arthur Kornberg and colleagues in 1969. They had shown it was possible to produce DNA chains with the thymine (T) ending by adding dideoxynucleotide triphosphate (ddTTP), an analogue of a normal DNA polymerase substrate that they had synthesised, to the incubation of a template and primer with polymerase (Atkinson, Deutscher, Kornberg, 1969).
Sanger did not expect the dideoxy method to work any better than the plus and minus system, but he was nonetheless eager to try it so wrote to Kornberg to obtain some ddTTP. Kornberg, however, had used up all of the ddTTP he had synthesised. Soon after this, however, Sanger by chance come across Klaus Geider at a meeting in Germany. This was a stroke of luck because Geider had synthesised some ddTTP in his laboratory at the Max Planck Institute and was willing to share a sample. Much to Sanger and Coulson's surprise the addition of Geider's ddTTP in their first experiment produced much better autoradiographs than anything they had seen before. These displayed 'sharp bands of equal intensities extending over a long sequence'. Just as important as the clear results, use of ddTTP rendered preliminary incubations unnecessary (Sanger, 1988; Sanger, Dowding, 1996).
There was one downside, however. The only chain terminator that had been synthesised so far was ddTTP and this was only capable of producing DNA fragments ending with thymine. Unable to persuade anyone else to produce other chain-terminating inhibitors for the three other nucleotides, adenine (A), cytosine (C) and guanine (G), Sanger and Coulson set about preparing them. This was a major challenge because neither of them had any experience in nucleotide synthesis. Sanger's main expertise was in synthetic organic chemistry. Such skills had served him well when making amino acid derivatives thirty years earlier when working on insulin. They, however, were useless for synthesising nucleotides. In the end Sanger and Coulson called upon the support of Mike Gait, an expert on nucleotide synthesis. Gait had joined the LMB in 1975 after spending two years as a postdoctoral researcher in Khorana's laboratory at MIT. Additional advice was sought from Bob Shepherd, head of the LMB's peptide chemistry group from 1971 (Sanger, Dowding, 1996; Finch, 2008).
While new to nucleotide synthesis, Sanger relished the opportunity. In part this was because he thought it would entail doing a lot of crystallisation, something he had enjoyed doing since his days of working alongside Ordish, his schoolmaster at Bryanston. As he recalled, 'the first sign of a pure compound was to obtain pure, well-shaped crystals, and it was always exciting to come into the lab in the morning to find a flask full of beautiful crystals, though more often all one found was an oily mess at the bottom.' (Sanger, Dowding, 1996).
Nucleotide synthesis, however, proved much less fun than Sanger expected. To his great disappointment he discovered that nucleotide products did not crystallise. Instead, he had to run the products on a chromatography system to test their purity, a much less enjoyable exercise. Progress was slow and did not always proceed smoothly. Coulson recalled one occasion where Sanger dropped to the floor one of their synthesised products which they had taken three months to prepare. Fortunately Coulson managed to rescue it by scraping it off the floor. After a year of hard labour Sanger and Coulson had finally succeeded in their objective. Looking back on this work Sanger would comment that what they had done was 'not exactly chemical synthesis' because they had not purified or crystallised the product; nonetheless the impure product 'seemed to do the trick' (Finch, 2008).
Sanger and Coulson's new approach incorporating dideoxys was published in December 1977. It was a much more accurate and rapid approach than the plus and minus system. Now up to 120 nucleotides could be read in one go (Sanger, Nicklen, Coulson, 1977). Following publication, Sanger and Coulson continued refining the method, introducing the use of thinner acrylamide gels and the replacement of the radioactive label 32P with 35S. These changes provided much clearer band patterns than before and higher resolution images. All of this made it easier to read the sequence. Moreover, it allowed 465 nucleotides to be read at a time. Sequencing capacity was also increased through the introduction of gels with narrower lanes. Further improvements were made with the introduction of hand-held repetitive pipetting devices which made it possible to do sequencing reactions in 96-well plates manually. On a single day one person could typically run 8 gels, each with 12 sequence ladders. It was not possible, however, to do such work more than twice a week (Sanger, Coulson, 1978; Finch, 2008).
Diagram showing the key steps in the dideoxy sequencing system.
One of the first applications the dideoxy method was used for was to recheck the phi X 174 bacteriophage DNA sequence previously determined with the plus and minus method. The new results indicated that the team needed to make 30 revisions to their original findings (Sanger, Coulson, 1978 ; Finch, 2008).
Sanger receiving the Nobel Prize, 1980. Credit: Fred Sanger.
In 1980 Sanger was awarded his second Nobel Prize in Chemistry. With his characteristic modesty Sanger would always tell people 'it was much more difficult to get the first prize than to get the second one because if you’ve already got a prize, then you can get facilities for work, and you can get collaborators, and everything is much easier." (Sanger, 2001a).
Sanger is the only Briton to have won two Nobel Prizes and only scientist to have won the Prize twice for chemistry. Half of the second Nobel Prize went to Walter Gilbert. It was awarded on the basis of 'their contributions concerning the determination of base sequences in nucleic acids'. Paul Berg received the other half of the prize in recognition 'for his fundamental studies of the biochemistry of nucleic acids, with particular regard to recombinant DNA.'
Over the next few years, the dideoxy method, also known as Sanger's method, would quickly overtake the Maxam-Gilbert sequencing technique in terms of popularity. In part this was because it was perceived to be much safer on account of it using fewer toxic chemicals and lower amounts of radioactivity. Sanger's method was also seen to be more natural becasue it shared many similarities with the natural workings of DNA, harnessing the duplicating action of polymerase to produce DNA fragments for analysis as opposed to degrading DNA with chemicals. This was highly attractive to molecular biologists who were at the forefront of adopting the technique. Moreover, it slotted easily into their daily laboratory practices. The rising dominance of the Sanger method would mean that it would subsequently become the technique of choice for automation (Hutchison, 2007; Garcia-Sancho, 2012). Click here to see Sanger celebrating his second Nobel Prize.
Atkinson, M R, Deutscher, M P, Kornberg, A (1969) 'Enzymatic synthesis of deoxyribonucleic acid. XXXIV. Termination of chain growth by a 2',3'-dideoxyribonucleotide', Biochemistry, 8/12: 4897-904. Back
Billeter, M A, Dahlberg, J E, Goodman, H M, Hindley, J, Weissman C (1969) 'Sequence of the first 175 nucleotides from the 5' terminals of Qß RNA synthesised in vitro', Nature, 224, 1083-86. Back
Brammar, W J, Gratzer, W B (2014) 'Sir Kenneth Murray', Biographical Memoirs of Fellows of the Royal Society, 60: 331-48. Back
Brown, A (2003) In the Beginning was the Worm: Finding the secrets of life in a tiny hermaphrodite, London. Back
Coulson, A (2015) email to L Marks, 7 July. Back
Donelson, J E, Wu, R (1971) 'Nucleotide sequence analysis of deoxyribonucleic acid', Journal of Biological Chemistry, 247/14: 4654-60. Back
Finch, J (2008) A Nobel Laureate on Every Floor, Cambridge, 2008. Back
Garcia-Sancho, M (2010) 'A new insight into Sanger's development of sequencing: From proteins to DNA, 1943-1977', Journal of the History of Biology, 43: 265-23. Back
Garcia-Sancho, M (2012) Biology, Computing and the History of molecular Sequencing: From Proteins to DNA, 1945-2000, Basingstoke. Back
Grigg, G (2001) 'Discovering DNA sequencing', Australian Biochemist, 31/1: 14-16. Back
Hutchison, C (2007) 'DNA sequencing: Bench to bedside and beyond', Nucleic Acids Research, 35/18: 6227-37. Back
Kaiser, A D, Hogness, D S (1960) 'The transformation of Escherichia coli with deoxyribonucleic acid isolated from bacteriophage lambda-dg', Journal of Molecular Biology, 2: 392-415. Back
Murray, K (1970) 'Nucleotide “maps” of digests of deoxyribonucleic acid', 118: 831-41 Back
Sanger, F (1980), 'Determination of nucleotide sequences in DNA', Nobel Lecture, 431-47. Back
Sanger, F (1988) 'Sequences, sequences, sequences', Annual Review of Biochemistry, 57: 1-29. Back
Sanger, F (1992) 'A life of research on the sequences of proteins and nucleic acids: Dr Fred Sanger in conversation with George Brownlee, Biochemistry Society archives. Back
Sanger, F (2001) 'The early days of DNA sequences', Nature Medicine, 7/3: 267-8. Back
Sanger, F, Air, G M, Barrell, B G, Brown, N L, Coulson, A R, Fiddes, J C, Hutchison, C A, Slocombe, P M, Smith, M (1977), 'Nucleotide sequence of bacteriophage phi X174 DNA', Nature, 265: 687-95. Back
Sanger, F, Coulson A R (1975) 'A rapid method for determining sequences in DNA by primed synthesis with DNA polymerase', Journal Molecular Biology, 94, 441-48. Back
Sanger, F, Coulson A R (1978) 'The use of thin acrylamide gels for DNA sequencing', FEBS Letters, 87/1: 107-10. Back
Sanger, F, Donelson, J E, Coulson, A R, Kossel, H, Fischer D (1973) 'Use of DNA polymerase I primed by a synthetic oligonucleotide to determine a nucleotide sequence in phage f1 DNA', Proc Nat Acad Sci USA, 70/4. 1209-13. Back
Sanger, F, Dowding, M (1996) Selected Papers of Frederick Sanger (with commentaries), London. Back
Sanger, F, Nicklen, S, Coulson, A R (1977), 'DNA sequencing with chain-terminating inhibitors', Proc Natl Acad Sci USA, 74/12: 5463-7. Back
Sinsheimer, R L (1959) 'A single-stranded DNA from bacteriophage phi X174', Journal Molecular Biology, 1:43. Back
Wu, R, Kaiser, A D (1968) 'Structure and base sequence in the cohesive ends of bacteriophage lambda DNA', Journal Molecular Biology, 35: 523-37. Back