Werner Arber

Darwinian evolution as understood by scientists of the 21st century

Category: Lectures

Date: 4 July 2007

Duration: 46 min

Quality: SD

Subtitles: EN

Werner Arber (2007) - Darwinian evolution as understood by scientists of the 21st century

After a short reminder of the historical development of evolutionary biology, elements to a molecular theory of Darwinien evolution will be presented. Biological evolution is driven by the availability of genetic variants in populations

It is a very great pleasure for me to introduce the next speaker, Dr. Werner Arber from University of Boston. The discoveries were made when Dr. Arber was working at University of Geneva in the Department of Molecular Genetics. And he received the Nobel Prize in Physiology or Medicine 1978 for the discovery of the restriction enzymes and their application to problems in molecular genetics. It’s actually one of the corner stones and starting points for the recombinant DNA technology. And the title of Dr. Arber’s work is ‘Darwinian Evolution as understood by scientists of the 21st century’. Please. I’m pleased to be with you. The important aspects of genetic engineering are actually reflected man-made contributions to biological evolution. Risks of doing so have been widely discussed and we realised rather rapidly already in the early 1970’s that a deeper insight into the Darwinian Evolution, now at the molecular level, would be appropriate to properly evaluate risks of genetic engineering, as compared to other risks which are and have always been in nature, namely by events of natural biological evolution. So this last topic I chose for you to show. And I start with a science historical view, starting with Charles Darwin, who in 1859 published his book in which he defends a theory of natural selection. That led to an insight into biological evolution. And interestingly just a few years later Gregor Mendel, absolutely independent of Charles Darwin, started the branch of science which nowadays is called genetics. It was done with plants, as you know, classical genetics didn’t immediately start, it started around 1900 and was developed. Then, around 1940 people realised that actually these two things have really tight connections, interconnections with each other. And it was clear that genetic variation, which is the driving force of biological evolution, is of course also the cause of genetic mutations here. And that led to the so called modern evolutionary synthesis. At that moment, genetics was not at the molecular level yet, it was just looking at phenotypes. It is microbial genetics which had its early development between 1940 and about ’55 with all the principle strategies of microbial genetics. And that then led finally by another kind of fusion with something which has its roots also in the 19th century, namely Friedrich Miescher’s work on nucleic acids. That then had its crown in 1953 by the double-helix structure published by Watson and Crick, as you know. Now, with this knowledge of microbial genetics and the double-helical structure, that led rather rapidly to molecular genetics and finally to genomics, proteomics, we are now here. And since a number of years there is another synthesis between molecular genetics and evolutionary biology, which I would call molecular evolution. And that's the topic of my contribution today. As we all know, neo-Darwinian evolution has three actually pillars. One is genetic variation, if there would be no genetic variations, there would be no way to get evolution. Natural selection is the way how organisms deal with the encountered environment, and there are not only physical or chemical constraints but also biological constraints. That means that all the organisms sharing the same ecosystem can mutually influence each other. Finally this influence can go to almost zero by isolation processes. We know reproductive isolation and geographic isolation. And that can modulate the process, while natural selection in fact gives the directions implurial on this tree of evolution. Of course always together with the available genetic variants. I’m a microbial geneticist, most evolutionary biologists work with higher animals. Many like to study the evolution from chimpanzee to human beings. That’s perfect but it’s very difficult to do molecular evolution experiments on that level. It’s much easier to do that with most simple organisms like bacterial cells, here you have a bacterial cell which is haploid and that is an advantage that any mutation that may occur gets rapidly manifested because of the haploidy. Here, symbolically, there’s a mutation having progeny, there is another mutation causing lethality, so there is no long term progeny. The drawing is a little bit wrong because the first mutation here should not show up when there are four cells but when there are about 400 cells. I couldn’t do that on that drawing, of course. And you see mutation rates of course are low and they must be low and I already mentioned that if there wouldn’t be – on the one hand, if it would be too low, no variation, there wouldn’t be any evolution. If it would be like here, as high as that, then, since many are lethal, the genetic stability would be in danger, and I think we couldn’t live under these conditions. Most of the evidence which I am going to show you today come actually from studies with such microbial systems. For those few of you who may not remember precisely how bacterial chromosome look like, it is an e-coli, very large circular molecule, drawn on a completely wrong scale, there is here symbolically one gene with a reading frame and several control signals for the expression of that information. You may have a mutation within the reading frame and that can alter the product of that gene, which is most of the time a protein, as you know. You can have a mutation in one of the control signals for gene expression and then you may alter the quantity of that gene product. I define from now on a mutation as an alteration in the nucleotide sequence. I’m aware that would be changing one letter or a few letters or many letters and so on. I will give you examples later on. But I’m aware that this does not correspond to the original definition of mutation, because in classical genetics, of course, a mutation is defined as an altered phenotype, that also is transmitted to the progeny. That’s called a mutation. While in the molecular reverse genetics one defines an altered nucleotide sequence as a mutation. I’m aware, and we are all aware that this altered phenotype is due to an old nucleotide sequence but you couldn’t expect that all mutations, all novel alternations in nucleotide sequence give also an altered phenotype. Indeed we do know that if you take that definition which I just gave, only rarely a mutation, spontaneous mutation, is favourable or useful under the encountered living conditions and gives a selective advantage. Much more often a mutation is unfavourable, giving selective disadvantage. It can inhibit the life processes. If it’s traumatic it’s just lethal from one case. If it’s a little bit less traumatic, these mutants will not be able to overcome more favourable genetic structures. And very often also novel mutation has no immediate influence on the life processes. These are found, described as either silent or neutral mutations. And if you think about that, there is no good evidence that the spontaneous mutagenesis per se would have any directedness, it’s more random. And it is also clear that one should, because many are unfavourable and only a few favourable, that you shouldn’t have too many mutations in a genome per generation. As I mentioned, an E.coli in the order of one mutation in a few hundred cells. What approaches do we have to get insight into evolution at the molecular level, that means generation of genetic variance. First of all, increasing availability of nucleotide sequences, which can be for specific genes, groups of genes or even a part of a gene, a domain, or more and more we have entire genomes. And it is hard to compare these with our eyes, to see how homologous are these genes, which sometimes are related. But the computer, bioinformatics helps us, so with that one can see for example two organisms which are, let’s say, evolutionary relatively closely related, still there are differences within a gene and seeing how that functions have developed and so on. It would be however important also to be able to look at one really chosen specific mutational process, to see, compare the sequence just before and just afterwards. And with that knowledge, which can most easily be done with microbial genomes, and I will show you examples for that. And then we can later on having ideas. What kind of natural mechanisms do exist in nature to bring about genetic variation? So one can bring that knowledge on this term and that helps us really to have a deep insight into the processes. Before giving some specific example, I start giving you an overview of the results which come out from that study. You’ll see again, as before, here is the source of genetic diversity, which is in fact source of mutation. I noted here often during replication of DNA, novel mutations can occur. Mutagens can interact, chemical radiation mutagens and so on. Then there are recombinational reshufflings within the genome and horizontal gene transfer. I will give you later on just one example for that and two perhaps from this and one from these. So you will then see it. You have here natural selection over there, I mentioned that already, and you have isolation. So that kind of three pillars are in that drawing. What I additionally added here is if you, we know quite a number of specific reactions for this process, we know quite a number from this and from this. And if you think are there specific strategies in nature, not strategies, human reflected strategies, just that you see. And you see, one strategy is local change, one or a few nucleotides on the go, a change. A DNA arrangement is a segment of perhaps part of a gene or a gene or a few genes get rearranged, duplicated, deleted and so on and so forth, changing place within the genome. Acquisition means that by horizontal gene transfer part of a gene, a gene or a few genes, never very long segments, become transferred in another organism which may find their way into the genome of that recipient cell and giving to the genome additional genetic information for capacities which were not yet available before in that organism. I give you here an example for replication in fidelity. This is just text book knowledge, adenine, thymine, the normal standard base pairing and guanine, cytosine. Organic chemistry tells us that these forms here, with this hydrogen attached to that nitrogen, corresponds to the most frequently found structure of adenine. However, sometimes the hydrogen jumps for a short period of time down here, as shown here. Then, if at that moment the replication fork moves by, of course it will not be able to fit in a thymine into this structure. But by chance a cytosine would fit pretty well with hydrogen bonds. Now, these forms are rather unstable, short living, and switch back to the normal form. And that moment, when the replication fork has already moved a little bit further away, then a miss pairing is there. Now, this would be a tremendous source of mutagenesis and, thanks to the presence of several repair systems, which are found in bacteria, all over in higher organisms, in human beings, these mispairings at an early time can be discovered as not fitting to the right information and they can be restored. So that cytosine, which then would be at that wrong place, would be removed and the thymine filled in. And interestingly the repair systems do even know which is the parental and which is the newly synthesised strand. So it’s a wonderful system. However, these repair systems are not 100% active. They leave an occasional number of mispairings there and that’s of course an important source for biological evolution. And we think that this and other processes which I will discuss later on, have been in long evolutionary times fine tuned to do this job at the right frequency. So here you see that nature uses an intrinsic property of matter, of the nucleotides, to make tautomeric forms as a source of local mutagenesis. There are other sources of local mutagenesis, which I have no time to describe. I now talk on DNA rearrangements. We do know of course generic combination, which is very important in higher organisms, animals during meiosis, but these enzymes do not their job during mitosis, that would be horrible if they would do it all the time. So you see, all of these rearrangement enzymes which are encoded in the genome, are kept at a very low frequency of expression if they are not needed. And for short periods of time, when they are appropriate to be used, they show up and do their job. The same as in bacteria, bacteria also have these kind of things and they can repair, for example radiation effects, cutting DNA fragments. Then there is transposition of mobile genetic elements. I give you one example of that, there is site specific recombination processes, I will talk briefly on one specific aspect of this and still other less well-known for many of these processes. The molecular mechanisms have been extremely well studied. Now, this is an example of a transposable genetic element. An E.coli bacteria, which has a size of a chromosome of almost 5,000 kilobase pairs. There is a number of, symbolically here in black, so called mobile genetic elements, IS element, insertion sequence elements. I have here drawn bacterial cells which in addition to its single chromosome has relatively small plasmid. That small plasmid in this particular case is in fact the genome, 90kb genome of bacterial virus P1, bacterial phage P1. And this phage can reside during long periods of time in the cell. Very occasionally only, all of a sudden the expression of these viral genes can be activated, then viral production occurs, the cell will die and liberate the progeny of several hundred viruses. That’s called the lysogenic cell. An experiment which we have done some decades ago actually was looking for lethal mutations. Think about how difficult it is to study lethal mutations. With this system you can do it. Because here you have about 50 really densely packed genes for bacterial virus production. And during that maintenance, you need only very few genes for plasmid replication. Then you can expect that sometimes spontaneous mutations occur in that cell, which may hit important viral genes which are important for viral reproduction. And some of these mutations may be insertions jumping, an IS element jumping from the chromosome over here to insert into that genome of the P1 virus. We have grown for about three months by diluting every day just in room temperature these cells, and after three months we looked for lethal mutations. That’s an easy task because we just irradiate the cells if their active virus is coming out, that’s easy to show, if no virus come out it’s a lack of virus reproduction. And then we had a big surprise, we could see that among all those isolated, no independent mutants, having undergone a lethal mutation, 95% were due to that transpositional event, inserting somewhere in that genome an IS element, which before was in the bacterial chromosome, 95%, only 5% were local mutations. So transposition of these mobile elements is a major source of lethal mutations. Having these mutations, one can see where did they insert. Is that the reproducible effect or not, is it more random? Look first, this is a linear outline of that P1 genome and obviously there are hot regions where we have many insertions, each point is independent, here is another region, here we know that there are quite important viral genes. If we would have isolated several thousand, we would certainly have some insertions there. But these insertions of IS elements kind of have a funny distribution. There is some randomness but certainly not full random. So particularly in the small segment we had many IS2s and three IS30, the blue ones are other, these are either IS1, 2, 5 or gamma delta. Then we sequenced some of these insertions here and surprisingly IS30 is a very site specific process, mediated by the enzyme transposase, and in IS2 we had not two of the analysed going in the same sequence. And even these sequences where they inserted, they do not show any obvious homology. So we conclude that some of these elements are very strict, they go and look at sequence, do it very site specifically there, although we do know from other studies that IS30 can occasionally go at other sites also. While IS2 prefers certainly this region and this region but with the insertions within this preferred large region are random. And we sub cloned some of these segments here and in the sub clone, it the plasmid they are still attracting IS2. So something must be there, I still don’t know what, so it’s interesting. You see each of these transposable elements has its own strategy. Now, I show you another, this is now a well-known integration of bacterial phage lambda. Which is here shown in its circular form after infection. And can site specifically incorporate into the genome, that’s also then a lysogenic cell, in contrast to the P1 lysogen, which you’ve seen before, which has the provirus as a plasmid, here the provirus lambda is carried as a part of the genome. When you shine light on that, UV light or even spontaneously, it can excise again precisely and go back and replicate and produce progeny virus. But sometimes, if one does that, the recombination doesn’t precisely occur, this is with this illegitimate excision occurring with a frequency of about 10 to the -4, as compared to the normal excision. And this gives rise to transducing viruses. Lambda inserts closely here to the galactose fermentation markers and therefore in this lambda transducing phage that can transduce by horizontal gene transfer, just by virus infection another bacterial cell and provide it with the ability to ferment the sugar galactose. This in fact for many of you - you know this is text book knowledge and I just summarise now what we know for horizontal gene transfer. The example which I just gave you is the one of a natural gene vector, lambda, having incorporated as part of its genome galactose fermentation markers. Other viruses, P1 among them, has another strategy to transfer host genes, symbolically shown here where these phages do not recombine with the viral genome host, but take a fragment of the host genome, about a headful of DNA from that host genome, incorporate it into a viral particle and transfer it horizontally to another recipient cell. There are basic knowledge bacterial genetics, is that by cell-cell contact in conjugation between here also two somewhat different bacterial strains, transfer can occur. And finally it is known, and that was already shown in 1944 by Avery and his collaborators, that free DNA, well purified from any attached protein, can provide genetic information to other pneumococcal cells, that’s called transformation. That was to prove that DNA is genetic information. Gene transfer, of course, if it should give some genetic addition to the recipient cell, the genes have to be incorporated and inherited later on. So I summarise the possibilities that can happen depending on the particular case by homologous recombination, by heterologous, for example here insertion of an IS element and so on. Or, as I showed you before, in the case of lambda by site specific integration. Or finally by maintaining as an autonomous plasmid. So there are various ways in nature to do that. Now, there are a number of factors limiting gene acquisition. First of all, surface compatibility, think on conjugation, not all different bacteria strains can make these close pairings. Or a virus must absorb on a particular surface. If they can do that, the DNA penetrates and encounters – practically bacterial strains have 1 or even a few systems for restriction modifications. These are, one can say simple immune systems, identifying foreign DNA and differentiating it from the cell’s own DNA, which is marked in particular methylation activities that cells own. And the restriction enzyme then cuts the DNA, that DNA is rapidly degraded, within a few minutes by exonucleases. But during that short period of time, some cut segments succeed to incorporate in the genome. So that keeps horizontal gene transfer low in order to ensure a certain stability. And it also contributes to the strategy of acquisition in small steps, small segments, because that’s important, because if let’s say half a chromosome would be incorporated into a cell having its own chromosome and these are genetically relatively unrelated, the functional compatibility would certainly be destroyed at that selection level. So all of that goes into the thing that per acquisition, you can acquire foreign DNA but usually in small steps. I come now, that leads me to discuss the qualitative differences between the 3 strategies which I explain, here local change DNA rearrangement, DNA acquisition. Local change, just a few nucleotides, is an improvement of available biological functions, that’s the main function of that. It’s probably one of the very essential thing that organisms can develop what they already have. DNA rearrangement within the genome gives rise to duplications, which can then be giving, in the long term possibilities, for maintaining an important gene and the duplicate can be modified slowly by local change and other aspects. But also the transposition, as we have seen, or site specific reshufflings and so on, can give rise to what we call gene fusions. This is that there is a recombination within the reading frames of two different genes and bringing functional domains together. And we do know from sequence analysis that many genes share some important function domain with other genes. And all these processes can by chance contribute to bring these things together. Or, what I define as operon fusion is to provide to a reading frame, an alternative segment for control of gene expression which may increase or decrease the frequency of gene expression and the efficiency. So all of that is good. Obviously all of these alternations are submitted to natural selection, if they are favourable in the environmental conditions in which the organisms grow, they are maintained and may even overgrow the parental population. If not, they are rapidly eliminated again. The final search strategy of acquisition, I consider it as a sharing in successful developments made by others, that’s a beautiful strategy, just get something from another organism who in long evolutionary development has now a functional thing, like antibiotic resistance. We microbiologists have learned a lot on that strategy just by studying the medical problem of antibiotic resistance which spreads around horizontally. Now, the evolutionary tree - it was mentioned yesterday, I think, in a talk - shouldn’t be seen just as a tree as such, but you may draw symbolically between any branch horizontal connectors in which at one time some genetic information can by horizontal transfer be transferred. So in some way I see interesting philosophical value in that drawing. Up to very recently we considered that if I’m up here, I’m linked with many others of those other organisms in the far past, evolutionary. But if you think about that evolution goes on and there is steady re-possibility of horizontal gene transfer, that we are also interlinked with many other organisms in our future. And I come to summarise what I said. I started to explain to you tautomeric forms and there are many other aspects of non-genetic elements contributing to spontaneous mutagenesis. But there are also evolution gene products which in the case of transposition, that example we just showed you, the IS elements, I consider as active variation generators. They generate variations driven by these enzymes, which are available at very low rates only. The repair systems which I mentioned, which are helping to keep the mutagenesis by for example tautomerism low, these I consider as modulators of the frequency of genetic variation. And I think these enzymes, together with the non-genetic elements really, which are also part of nature, are learning to us that natural reality takes actively care of the biological evolution. Evolution should not any longer be considered as due to accidents, due to errors, as you find in most text books, that’s a wrong attitude towards nature. I think we should learn from the molecular knowledge and see in evolution, in the generation of genetic variants a wonderful activity which is part of nature in which we live. With that I think I consider this as an expansion of the Darwinian Theory to the level of molecular processes. Very briefly, I was really, just in the lecture which we heard before the break, I’ve seen also that - you have to have proteins at the right moment in sufficient quality available, if they are not needed they should be removed. Here again, we have to promote genetic variation but also to limit, so that in order to have certain genetic stability but still allow a steady biological evolution, and we do think that that fine tuning was brought about in long evolutionary period by what we call second order selection. That means those organisms which had the capacities, they were able to come to what they are today. We only study evolution as it occurs today and I must just confess, I am not sure how the first living being on our planet came about. Now, you can ask, well, what you hear here, is that only true for bacteria? I believe no, I think, in order not to use too much time, I think I will go rather rapidly through that. For an evolutionary fitness I consider that organisms should be equipped with evolution genes for each strategy to generate genetic variance, the intrinsic properties of nature are available anyhow. Then you think that bacteria have learned to have cell differentiation also. It’s known that within certain colonies there is some cooperation that can give rise to multicultural organisms. And then bacteria come back as symbionts, endosymbionts and I attach much importance to that. Once you live for years and years in another organism, there is some chance during that cohabitation to make gene transfer and I think good examples are mitochondria and other organelles and so on and so forth. One should see that - I better not go further. We do see up to some 10 or 20 years ago, I believed that in my genome there are only genes of importance for my own personal life, from fertilisation to my death, I then realised that in my genome and in bacterial genomes and any other genome, there are other genes which are not serving the purpose of that individual life but serving for biological evolution, namely those evolution genes. It is also clear to anybody who looks into that, that some gene products serve both purposes. For the individual life it’s the fulfilment of that life and for the evolution genes it’s expansion of life, that’s the source of biodiversity. We suffer loss of biodiversity, the hope is that long term - not next year or in 10 years – but in long term, loss of biodiversity will be reconstituted. Not the same biodiversity, another one, but nature actually is able to propagate and produce biodiversity as long as living conditions exist on our planet. I had thought to just remind you that there are some medical implications of what I told you, we are aware that some unfavourable mutations in the germline give rise to inherited disease, chromosomal abnormalities. Then, at the somatic level there are, besides unfavourable, for example cancer, somatic mutations, there are also favourable aspects, like the immune system of higher animals is precisely using what we learn from bacteria, all these strategies are there. We also know that individual responses to some target specific therapies depend really on different alleles which are carried in different human beings. And, last but not least, infectious diseases by micro-organisms, pathogenicity in cohabitation and so on. And I mentioned already the microbial resistance to antibiotics. With these few remarks for those medical doctors among you, I’d like to close here. Thank you for your attention.


After a short reminder of the historical development of evolutionary biology, elements to a molecular theory of Darwinien evolution will be presented. Biological evolution is driven by the availability of genetic variants in populations. The occasional generation of genetic variants is brought about by cooperative actions of products of evolution genes and of non-genetic elements.

Three qualitatively different natural strategies (local sequence change, intragenomic rearrangement of DNA segments, and DNA acquisition by horizontal gene transfer) contribute each with a few specific mechanisms to the overall mutagenesis. World view aspects and medical implications will be mentionned.