Howard Temin

The Evolution of Retroviruses

Category: Lectures

Date: 2 July 1981

Duration: 35 min

Quality: HD MD SD

Subtitles: EN

Howard Temin (1981) - The Evolution of Retroviruses

Some of the Lindau meetings are remembered because of special jubilees. This was, e.g., the case for the 40th meeting in 1990. Then Count Lennart Bernadotte gave a remarkable open-minded personal history of the meetings

Thank you. Today I’m going to build on several of the previous lectures as it turns out. Conceptually I’m going to use the systems that Dr. Dulbecco talked about, namely the retroviruses. And then I am going to talk about some of the mechanisms involved in the processes he mentioned and show a parallelism between viruses which infect animal cells, vertebrate cells, and the elements that have been described by Dr. Arber. And then show how we can use these viruses to promote the kinds of transfers discussed by Professor Smith. I’m going to begin with the first slide to illustrate the nature of the biological material I am working with, namely the retrovirus, a virus which is produced by budding from an infected cell and exists as a particle about a hundred milimicrons the virus synthesizes double-stranded DNA, linear and circular molecules, and then the virus particle has specific mechanisms to integrate its DNA into that of the cell DNA with a very high efficiency. And then this integrated molecule of a provirus is the template for the formation of new viral RNA which starts the cycle over again upon infection of another cell. This is just an illustration of some of those caused by a particular species of chicken viruses, various kinds of leukaemias, sarcomas and carcinomas can be caused by these viruses. Since the viruses are very efficient in infection, the formation of the cancers is very efficient. So these viruses are the most powerful and efficient of all carcinogenic agents. The non-oncogenic viruses, they only specify viral proteins in their coding sequences. But the strongly oncogenic viruses have, either in addition or replacement, other sequences, and these sequences arise from the host cell, as Professor Dulbecco again described, and, when present in the virus, become, with the virus genes, highly efficient oncogenic agents. Now to look in more detail at the nature of the viral sequences which can change, if you will, the expression or activate the cell genes to make them oncogenic, techniques of DNA cloning are used. These are used to take DNA now from a chicken cell infected with an animal virus, digest this with the restriction enzyme, separate the restriction enzyme, cut DNA molecules. Then take DNA, cut with the restriction enzyme, from a bacterial virus related to those discussed by Professor Arber. Ligate together the chicken DNA molecules and bacterial virus DNA molecules and package together this new recombinant molecule in the bacterial virus proteins, making now a bacterial virus. Infect cells, E-Coli cells, grow up virus which will contain chicken DNA and Bacterial DNA, and then by hybridization techniques select the particular recombinant virus, bacterial virus, which contains the chicken DNA which in turn contains the animal virus. Now in this particular case, we have not shown the virus, the bacterial virus DNA would be off the slide, there would be cell DNA and then animal virus DNA. And the names are not important but they show a particular pattern. Now if one blows up the ends of the provirus as shown down here, and then carries out further restriction enzyme mapping, as you see just the greater density of names, one finds that a particular sequence is repeated. And if we look here we have Sac. Here we have Sac. Then HhI HhI HaeIII HaeIII and so forth. A pattern of restrictions on cleavage sites is the same on both ends. This pattern shown here also by crosshatching is known as the LTR or long terminal repeat and has been sequenced in this virus and in others, next slide, and shown to have some remarkable features. These are shown diagrammatically here. One of the most remarkable features is actually the way it’s made but this is an extremely complicated process and is related to these names, U3R and U5. For our purposes, what is important is that in the sequence of the long terminal repeat, our sequence is shown here in boxes which have been shown in these virus systems and in systems of genes from vertebrate cells to be the controlling signals for synthesis of RNA, the so-called Goldberg-Hugness or TATA Box, the CAT Box of -85, -24, and a capping sequence telling the cell polymerases to start here synthesis of viral RNA. So in the LTR the virus has control sequences which will tell the cell to make new virus. Similarly, in the LTR the cell, the virus has control sequences like this one telling the cell to add Poly(A) at the end of the viral RNA and then sequences like these to tell the cell to stop the synthesis. So these sequences at the ends of the virus are control sequences which tell the cell start synthesis / stop synthesis. Now if we look at the end of this entire sequence, we find at the very end three base pairs which are present as an inverted repeat, TGT and ACA. and carry out this sequencing at both ends of these viruses we find several things. Firstly at the end of each virus we have the same sequence, TGT TGT TGT at the left end, then going all the way to the right end the inverted repeat ACA ACA ACA. We find the same sequences inside, since the virus is present inside, but then on the outside we find a different sequence in each virus. In this one CCTTC GATAC AAAAT. So where the virus sits is different. But if we look at any one virus at the sequences on both sides, we find that five base pairs are the same. CCTTC CCTTC GATAC GATAC and AAAT AAAT and so forth through a whole collection of these. So that in the next slide, it is possible to draw a very simple general picture of the structure of a provirus. Now the particular numbers are true for this particular virus but the general feature is true for all retroviruses. The important features are a direct repeat of cell DNA indicated by the dark arrows and the numbers 12345, an inverted repeat of element DNA, in this case the virus, and then a larger direct repeat of viral sequences. This structure is exactly the same as the transposons that Professor Arber was talking about. They can also be described as elements that form direct repeats of cell DNA and have inverted repeats, and in some cases, as shown in the next slide, they can further have exactly the same kind of structure. In this case I am showing at the top, the structure we have just seen from the animal virus. In this case I am showing the structures of cellular movable genetic elements like transposons but in this case from eukaryotic organisms, Ty1 from yeasts and copia from drosophila. You can immediately see a great similarity in the structures of these elements, especially a repeat of cell DNA, in this case the large terminal repeat in both cases and less clear the inverted repeats at the same. But it’s clear from these comparisons that there is an enormous sequence similarity not in the actual nucleotide sequence, but in the pattern of the nucleotide sequence between retroviruses, cellular movable genetic elements of eukaryotes and transposable elements of viruses. We can then suppose, in the next slide, following, I’m sorry, furthermore, not only are the patterns the same, but the actual nucleotide sequences are very similar. These are the same three elements from vertebrates, drosophilae, yeast, the sequences at the end are exactly the same and then all of these sequences are similar. This great similarity of pattern of sequence and of actual sequence is consistent with the hypothesis that the animal virus has evolved from the cellular movable genetic element. In other words, that the animal virus is a transposon which has escaped to be a free living virus. If we started with an insertion sequence, and this insertion sequence had certain signals for control of synthesis of RNA and in the insertion sequences made a transposon, or a protein like DNA polymerase which could be a precursor of the virus reverse transcriptase. showing in the open box the insertion sequence. Now if this transposon transposed around another cellular gene which could be the precursor of the viral internal protein. And then there was a deletion, giving rise to a new transposon which now would have two viral proteins. And then there was a mutation for another aspect of viral function, primer binding site. Then this transposon could transpose and transpose again around a cellular sequence which could be the precursor for another virus protein. And then by the same process a transposon could be formed which would be exactly the same as a viral provirus. Then from such a transposon, transcription would give viral RNA which then could be packaged and used as a virus. Now once we had reached this stage, actually, because the structures of the provirus and the product of transposition are exactly the same, one cannot tell, by looking in the cell at the structure of the DNA, whether this DNA resulted from infection or from transposition. And in fact now in animal cells there are many structures like these, the so called A-particles, the 30S-particles, and there are new ones being described all the time. The question is what are these? Are they proviruses, defective proviruses or are they transposons or are they a little bit of both? There is no way to tell from looking at them which they are because the structures are exactly the same and there are many of them in cells. Now this structure and this way of formation then indicates in the next slide, please, a way in which the strongly transforming viruses could have been formed. This is for a particular virus, the names are not important, this would be the proto-oncogene, which Professor Dulbecco spoke about yesterday. And in the virus the proto-oncogene would now be an oncogene in the virus flanked by the LTRs. The way in which it might have occurred is by the same series of transpositions and deletions, which could happen all at the same time or could happen in two successive cycles. These kind of results, and the fact that it appears a wide variety of cellular DNA can be carried in a virus between the LTR as a provirus and then be packaged in a virus indicated to us that it might be possible for us to put DNA inside a virus LTR and have the virus carry that DNA into other cells. In otherwise, to make the virus into a specialized transducing virus for any sequence we might choose to do. So what I have been telling you up to now has been natural evolution of viruses, one might say that researchers have selected some of these highly oncogenic viruses, but at least they were unplanned. Now I’m going to tell you how we have carried this evolution a step further by duplicating in the test tube the same kind of processes that we imagine have gone rise in the evolution of these viruses. and had been sequenced and that was easy to select for. And the gene we selected was that coding for thymidine kinase and coded for by the herpes simplex virus type 1. This had been cloned and sequenced by others and is shown here as a 3.4 kilobase pair BamHI fragment. This fragment codes for a messenger RNA which is coding for the thymidine Kinase protein. It is possible by the use of the appropriate antibiotics and nucleosides to select for the presence of an active protein. Now at the end of this there is a region which has the end of the virus protein, ending right here at this point. And then later on is the signal here the same as present in the retrovirus for the Poly(A) addition and the end of the messenger RNA synthesis. And there is in between these two a HAHA1 site, so that it will be important later that we are able to separate the end of RNA synthesis from the end of protein synthesis. This is done right here, where we have taken, by a very complicated process of genetic engineering, and removed all of the sequence between here, which is just at the end of the protein but not at the end of the RNA synthesis, and here. And then put this back here, giving rise to a shortened thymidine kinase fragment. First we will look at this fragment, the next slide. I will indicate how this fragment from herpes simplex virus was inserted into the retrovirus spleen necrosis virus. The herpes simplex thymidine kinase was present in a plasmid, and with BamHI the fragment was cut out from the plasmid. The animal virus had previously also been cloned into a plasmid, and it’s actually a circularly permuted fragment of the virus, but still a complete virus. And this was opened with the enzyme BamHI and the herpes simplex virus was ligated into it giving rise to a virus with thymidine kinase integrated into it. So by doing this, were now not showing the whole plasmid, but by doing the experiment shown, we can then insert the thymidine kinase in different orientations to the right or to the left in different positions in the retrovirus. Now we have a plasmid growing in E-Coli, but what were interested in is vertebrate cells. And we want to be able to ask about the activity of the virus and the activity of the thymidine kinase. So what we do is cut out the DNA from the bacterial plasmid, ligate it together with itself and then reintroduce this DNA into the animal cells and look for expression of these genes in the animal cells. So were started with animal viruses, went to bacteria, grew them in bacteria, and manipulated them in bacteria. Then isolated the DNA which still remains the DNA specifying the animal viruses, and reintroduced that DNA back in the animal cells. Now these are mouse cells and these are rat cells which are thymidine kinase negative. The plasmid itself can transform them, and in most cases, at least in these cases, the insertion of the herpes thymidine kinase into the virus DNA does not affect its activity. However here these are chicken cells where were looking at the replication of the virus, the virus itself replicates with a high efficiency as a result of the transfection, but the introduction of the thymidine kinase reduces or abolishes the infectivity. So it was necessary to make some modifications. And so to do this, we made these deleted viruses, and this is just to give you a feeling, nothing more than that, of the way one those these kinds of genetic engineering. But here we take one of these plasmids, now we’ve gone back to bacteria again, with the thymidine kinase inserted. And we now want to make a deletion of the virus from here to here, and we want to substitute this thymidine kinase for a thymidine kinase where we have deleted some of the control sequences. So to do this we digest this plasmid with HindIII and separate the two molecules and digest each of them, this one with BglII just isolating this fragment, this one with SacI discarding this fragment and keeping this fragment. And then digest this one with SacI and BamHI, just keeping this fragment. Then we can mix together these three fragments in a test tube with an enzyme and ligate them together and then by selecting get this modified plasmid. Then we can cut this plasmid and ligate and get a structure which, when introduced into animal cells again, will replicate. So first, we asked about the biological activity of these new structures which are deleted in virus and deleted in thymidine kinase. These are again in mouse and rat cells. We see that deletion of the end of the protein reduces the activity very much, deletion of the end of the RNA does not reduce the activity in the plasmid, but when we put this one into a deleted virus, we see there is some reduction. But generally we can see that we still have enough thymidine kinase with the modified virus and the modified gene to use. Now since we have modified the virus as I showed you, we cannot expect to get infectivity directly. The virus has lost three kilobase pairs for example. So the experiment becomes even more complicated that we take from bacteria cloned DNA of a complete virus, and so we transfect cells with a mixture of the two kinds of bacterial DNA: one bacterial DNA which specifies a complete virus known as a helper, the other the bacterial DNA which specifies the modified virus carrying the modified thymidine kinase gene. And when this entire mixture is put onto cells, virus will be produced. And now we are in the realm of animal virology again and can assay the virus which is produced. these complicated mixtures of the cloned DNA from bacteria. And we are assaying, here are the virus titers, the helper virus which grows to a very high titer, and this is just a virus, it’s not of particular interest. But these, and we can concentrate on this one, this is an assay for the thymidine kinase gene in the virus. And we see that, as the whole virus is produced, the defective virus, which is now transducing the thymidine kinase gene, is produced to quite respectable titers. Now to show it is important to get these high titers of production to remove the sequences specifying the end of the RNA. Just if we compare this point with this one, these are DNAs which are otherwise the same, except this one has been deleted for the sequences for termination of RNA synthesis and this one has them. Presumably the competition between this defective virus and the helper is too much, so there is little, and these are other, other kinds of these recombinant viruses. But you can see one can get a lot of this virus and to establish that this is still the virus which we started with from the bacteria, one can take this virus and infect chicken cells with it, (next slide), extract the DNA and study the structure of the DNA made in the chicken cells. So we now are looking at DNA from chicken cells which was infected by virus produced by other chicken cells which were infected by virus produced by still other chicken cells, which were transfected by DNA grown in E-Coli. And we find by hybridization, these are the experimental ones, that we have the same size of DNA as we had for bacteria, and if we look for the thymidine kinase gene, the thymidine kinase gene is still stably carried in these new recombinant viruses, as it was stably carried in the bacteria. Furthermore (the next slide), since we have a relatively high titer of these viruses, we can use them in transformation experiments to transform essentially an entire culture of cells. So this is an example of rat cells in the selective medium when they are uninfected all of the cells are killed. However if these cells were infected with these recombinant viruses and then selected, essentially all of the cells survive. Because essentially all of the cells have been infected by the transducing virus and since the virus is very efficient at introducing its genetic material, the thymidine kinase is now part of the virus genetic material and so is integrated efficiently into the sensitive cell, thereby the sensitive cell is made resistant. How the retro viruses behave as transposable elements which have an additional virus phase. This explains their ability to cause cancer at such a high efficiency because they can take these proto-oncogenes, activate them by putting them within their own control sequences. And then very efficiently integrate the modified sequences into a new cell. Furthermore I have shown you how we now, using the techniques of genetic engineering, can introduce any other cloned sequence into the virus. The virus, essentially, doesn’t care what’s in between the LTRs, whatever is between the LTRs over quite a wide range of size will now be just as efficiently introduced into any sensitive cell. If as in this case, the new DNA is DNA which codes for a protein which can be made essential to the cell, this can be used to protect the cell, this can be used in other ways which I think are obvious to all of you. Thank you.


Some of the Lindau meetings are remembered because of special jubilees. This was, e.g., the case for the 40th meeting in 1990. Then Count Lennart Bernadotte gave a remarkable open-minded personal history of the meetings. In particular he then described how the meetings were transformed from meetings for specialists to meetings mainly for young researchers and students. But the medicine meeting of 1981 must also be remembered, although for a quite different reason. It was a meeting gathering almost 30 Nobel Laureates, among them, for the first time, two of the inventors of genetical engineering, Werner Arber and Hamilton Smith, and, also for the first time, two of the practitioners of the technique, used in attempts to understand cancer, Renato Dulbecco and Howard Temin. As if this was not enough, the two scientists who unravelled the structure of genes, Francis Crick and James Watson, was also there, as was one of their strong competitors, Linus Pauling. If a time machine was available and if it could only be used for one return journey, the Lindau meeting of 1981 would be a good choice. Science fiction aside, through the tape recordings of the lectures, we at least get a glimpse of the kind of subjects that probably were up for discussion at the two closed sessions where only “Nobel Laureates, assistants and students” were allowed to participate. Howard Temin only came to the Lindau meetings twice before he passed away, only about 60 years old. As a student of Renato Dulbecco, his lecture is quite different from that of his 20 year older teacher. Temin seems to have mastered the techniques of genetic engineering and during his lecture he treats the building blocks of DNA with the same brilliance as a professional juggler. With my own background as a theoretical physicist, I cannot judge to what extent young researchers (“assistants”) and students of 1981 were able to follow his show! Anders Bárány