Werner Arber (1984) - On the Unforeseeability of Foreseeable Manifestations of Life (German Presentation)

President of the Council, Chairman, ladies and gentlemen, Today I would like to shed some new light on an old question: The question whether life, our lives, is perhaps predetermined. Or whether external and internal influences can contribute certain aspects to life at any given time. I will try to lead this discussion from the standpoint of a molecular geneticist. I am, of course, aware of the fact that molecular genetics is not the be-all and end-all of life, and that it is in fact individual cells as well as cellular systems that make up life in its entirety. Yet I will attempt, based on genetic material and its products, to discuss to what extent this genetic material, the genome containing the genetic code, can predetermine manifestations of life, and whether this field of macromolecules, of biological activity, allows for a certain degree of flexibility, of said unforeseeability, an element of pure chance. I will spend the first short part of my lecture introducing the topic to those of you who are not abreast of modern molecular genetic developments before presenting some examples from my own area of research, and I shall then conclude by stating that the molecular interactions of biological macromolecules principally abide by the laws of nature, but that, depending on the observed reaction, smaller or larger scopes for contingency do exist, which is precisely where the element of chance can have an impact. Of course, I have to admit that even today there are many laws of nature that we are still unaware of and that perhaps individual aspects that we currently regard as being elements of chance may one day be reclassified as new laws of nature. This genetic material is deoxyribonucleic acid and in school we learn that DNA is a long chain molecule fundamentally made up of a sequence of individual building blocks. Here symbolized by letters, of which there are four different ones. And just like in our writing, the sequence in the linear order determines the information content of the genome. Here you can see a very short section of these very lengthy, threadlike molecules, of just 10 building blocks. Yet we need to bear in mind that even in the case of the simple Escherichia coli bacterial cell, the respective information content adds up to about 4 million, and in terms of real letters, is pretty much equivalent to the content of the Bible. A single human cell contains about one 1,000 times more information content, which would equate to a large library filled with 1,000 volumes of the Bible. The individual functional chapters of the genome are known as genes and what you see here is a very schematic diagram of a gene. The length of this segment, applying our writing metaphor again, would be anywhere between several lines of text to perhaps one page in this library. This is the content of a gene. The gene is made up of the part that later manifests itself as a gene product, a protein, but also of these parts before and after this sequence, which are essential for controlling how the gene is expressed. The question would therefore be whether or not this linear sequence of individual building blocks predetermines what this protein will look like. The answer is yes. We know that, of course. This protein will naturally not take such a strictly linear shape as depicted here, but will instead adopt a particular three-dimensional structure and then carry out its respective function, be it as an enzyme that triggers further biological reactions, or in the form of a structural component, for example a membrane. What I would like to discuss with you today: Do these green-coloured gene products function in a very particular way? Can we predict exactly what they will do? Now, my first example stems from a topic that I researched back in the day when I was a doctoral candidate, namely the question: When a bacterial virus infects a bacterial host cell, as you can see here in the case of the extensively researched lambda phage, then the first reaction will be that the viral particle attaches itself to the cell’s outer surface and injects its genetic material into the host cell. Shortly thereafter, following a predicted pattern, certain genes, not all of them, of this viral genome molecule are expressed, create proteins, and will then present the so-called early responses, which in turn will trigger further genes, leading to medium-early to late responses. Now, the interesting thing about this particular case is that, after half an hour, we see what has happened, that either the one shown on the left has occurred, or the one shown on the right. There are two possibilities. The first possibility, which is observed in about 70% of all cases with regard to this particular infection, is that the infecting virus replicates and creates progeny. As you can see here, after just half an hour this cell bursts and offspring, viral particles, are created, which can now in turn infect new cells. This is actually the normal life cycle of a virus. In 30% of the cases, something else happens to the infected cell. Here, the late responses are never expressed, but are instead repressed in advance by the early responses, so that in a later phase this genetic material of the virus is integrated into the cell, into the chromosome. From there it is passed on to the cell’s offspring. When this cell then divides, each of the two new cells contains this so-called provirus, so that we now have a whole population of clones, so-called lysogenic bacteria. Here you can see a first example of this predetermination. Of course these processes are regulated by genes. But if you, if it would be possible, at a time when t is 0, when the genetic material penetrates the infected cell, the cell could be asked: “What will happen to you?” Then all the cell could really reply is: integrating the viral genome and passing it on to my offspring.” You see, here is a process that cannot be clearly predicted, neither by us nor by anyone else, not even by the organisms involved. It would be going too far to explain this here, but you can see more or less how this decision comes about. What we have here in this early phase is a balance, at which point more of either one or the other gene product was formed and it all depends on what happens at this critical point in time whether the late responses are initiated, then that leads to lysis (the production of phages), or if they are repressed, then that does not happen; instead, there is now a chance for the viral genome to be integrated. Here in this diagram I would like to show you a second example of how certain life processes are unforeseeable. If these lysogenic bacteria are cultured repeatedly, then we have a probability of 1:10,000 per generation that one cell will suddenly start producing phages. The genome is certainly there, it was simply being repressed, and this repression can abruptly cease under certain physiological conditions, so that this one cell enters the lytic phase while the other 9,999 cells continue to reproduce. On the other hand, it cannot be predicted when and to what extent this single cell will enter the induction stage. Yet this is important for its offspring, of course. This cell will then no longer be able to reproduce. The cell solution contains free viruses. Now, it is possible, and this is where it gets interesting, to influence this probability of induction by external means. In this particular case, for example, through exposure to ultraviolet light, you can, instead of just a ten-thousandth, stimulate the induction process in virtually the entire population at a given point in time. By exposing it to ultraviolet light. And, this effect, under certain conditions, can also be achieved by raising the temperature. This shows that this cell itself not only observes internal laws, but that it perceives external influences, environmental factors, and that the cell reacts to this. I consider the third type of unforeseeability to be as follows: in this already-described induction phase, the genetic material that was incorporated into the genome of the host cell is cut back out at a certain stage, and this process is normally very precise. In a few cases, however, this excision process does not occur at the correct site, and I have another diagram to illustrate this point again more clearly. So, when this viral genome, here coloured red, is inserted into the blue bacterial genome, which is, of course, is drawn too short here, the viral genome can be fully excised from the lysogenic cell during induction, without something becoming lost. This is a site-specific recombination, a reversal of the original insertion. Yet, however, in rare cases, it is possible that during a recombination process, which is currently regarded as being illegitimate, because we cannot predict where it will take place, here, here or somewhere else, so that a part of the recombined viral genome, here coloured red, remains in the bacterial chromosome after excision, yet that a different part of the host bacterium’s chromosome contains the viral genome responsible for galactose fermentation. And the site at which the recombination process occurs, meaning the amount of lost viral material and the amount of newly-added bacterial material, differs from case to case. It is not pre-programmed, but on the other hand, something left to chance. This results here in a hybrid genome, containing both viral and bacterial genetic material, which is why it is of such great importance for the environment, for the population of micro-organisms. Because such viruses, like the ones drawn here, can of course infect other bacteria and thereby genetic material, transferring individual gene from a previously infected host cell to a new host cell and introducing new traits. At this point I would like to remind you that it was precisely this natural process that for scientists served, about ten years ago, as a model aiming to develop the strategy of in vitro recombination, in which genetic material from a given source, from living cells, is introduced into what is known as a natural vector, such as a viral genome molecule, in vitro, that is in a test tube, and then reintroduced into bacterial cells, where this genetic material can then be replicated, which enables scientists to conduct detailed molecular studies. Researchers have therefore actually adopted this strategy by observing nature, it was not newly invented by science. Now, on the other side, these bio transducing phages are then created. What I have depicted in this diagram, when this process occurs, which could theoretically also occur, in that case, this genome, which has not lost any viral material and contains a large quantity of biological material from the host cell, would be too big to be packaged into viral particles, and that is surely the reason why we will never be able to observe these phages. Because there is a spatial limitation in the structure of the heads of these viruses. I would now like to shortly talk about the mechanisms, to return to them once again. There is, when inserting the viral material into the bacterial chromosome, a preferred site, which in fact lies between the genes that are responsible for galactose fermentation of the E. coli cell and those that are responsible for biotin production. However, it has been observed that the E. coli cell features numerous additional, secondary insertion sites, even though it is much less likely that these sites are used. So if the insertion occurs at variable sites, we can predict that analogously an illegitimate excision will not result in lambda gal or lambda bio, but will instead create other lambda hybrids containing genetic material from the host cell that of course have genes from a different source than the large bacterial chromosome, so that in this natural process it would, in principal, be possible under natural conditions to create bacteria that contain genes from the bacterial chromosome and can, in turn, pass these genes on to others as gal or bio. Probability plays an important role here, of course, because the likelihood of this insertion occurring at other sites is significantly smaller than at the main site of insertion. I will now put aside this diagram of the lambda phage and shortly turn to the restriction enzymes. You have already heard in the lecture presented by my colleague, Dr. Feinendegen, that restriction enzymes can be found in many bacterial cells whose main task is to watch for any foreign material infiltrating the cell and if yes to destroy it as quickly as possible. Because it is in the cell’s inherent interest to preserve its genome, not receive and foreign genetic material, and by no means become infected by viruses, because, as we have seen, viral proliferation increases the likelihood of the cell’s death. Now what I have drawn here is a simplified diagram of the mechanism of action of a restriction enzyme called EcoK found in the Escherichia coli K-12 cell. This is a pretty complex enzyme. It is, when created as a gene product, inactive for the time being, meaning that it cannot keep watch for foreign material. It requires the presence of the co-factor S-Adenosyl methionine which, when present, binds itself to the enzyme and activates the enzyme, and only then can the enzyme form a specific bond with DNA, making it possible to recognize a very particular sequence in the order of these building blocks. This sequence, in this case, is AAC, an arbitrary sequence of six other base pairs, there must be six of them, and then GT GC. This is recognized. This is where another complication occurs. It is of course clear that, for purely statistical reasons, the E. coli cell’s DNA chain molecule that is four million base pairs long must be repeated at the end of this sequence. And when this enzyme is present, that is dangerous. That is presumably why the bacterial cell developed a technique to protect its own sequence, thereby preventing it from being classified as foreign. And this occurs through methylation. Both this adenine and this adenine feature a methyl group, which is also attached by the EcoK enzyme complex. It has been discovered that if one strand of the DNA carries a methyl group and the other does not, the enzyme will introduce a second methyl group to form this structure, so that foreign material is not detected. It would only be recognized if such a sequence has absolutely no methyl groups. That is, in neither strand any of these methyl groups. In that case, the eco enzyme would send out a signal that this sequence has been recognized. This is foreign genetic material, please destroy. And now what is interesting: I have brought along a small DNA molecule. You see how this mechanism works. In this particular case, I am going to assume that the recognized sequence is located here. The enzyme binds to this site, and the interesting thing is that the cleavage does not occur here at this location, but that the enzyme, which has been observed under the electron microscope, starts to traverse this DNA in a complicated manner before it is finally decided here at this location a specific site at which to cut the strand - I need to add a small co-factor here to make this model work yet the cleavage occurs down here. This is truly spectacular. The enzyme, instead of migrating to the gap, stays here at this fixed location. This phenomenon can be observed under the microscope after the split. Yet the split takes place somewhere else. And the interesting thing about this process is that if you were to observe it 1,000 times, the cleavage will never occur at the same site. It always occurs at another site. Which laws determine where the cut occurs, we do not know. The process appears to comprise an element of chance. For these types of DNA fragments, which could potentially be reused, it is important to know whether the cleavage always occurs at the same site or in different locations. There is, just to make sure that you do not misunderstand me, I must add here that there are other enzymes, such as EcoRI, which might be assumed that the cleavage should always take place exactly there. Here you can see a comparison with what we were looking at earlier. This is the recognition site. The cleavage occurs somewhere else, either to the left or to the right, each time differently. The second type of enzyme always cuts at the same site. Here this is strictly regulated. And there is a third type that stretches out its arm across around 25 to 27 of these building blocks and then cuts the strand there. This shows the flexibility of nature, as we can see. I will now leave the restriction enzymes behind and move on to the process of transposition, which has been a personal interest of mine for many years. The genome, we know today, is not as stable as we believed for a long time and a major factor is transposition. Here only in a schematic diagram. Here we have an element, known as an IS element in bacteria, coloured blue here, which forms part of this long chain molecule. And here you can see a very schematic reading frame, which can result in a gene product, a so-called transposase that together with host proteins in the host cell forms a complex and then in a process, the details of which are not fully understood, that is capable of translocating a new copy or even, this is not quite certain, such as the one you see here, to a new target sequence. I would like to thank here our Chairman for touching upon this topic earlier. A very interesting question now becomes apparent: Are there rules that stipulate where the insertion takes place? I would just like to add that several types of such IS elements exist. My list only includes some of the elements found in K-12 strands of the E. coli cell. Other bacteria contain very different elements. As you see, here we have S1, 2, 3 and so on. Question: Are there even more elements of which we do not know how many copies exist per genome? Meaning per bacterial chromosome. Here that number can vary anywhere between one and perhaps twelve here. The question is how we can verify how they work, and at this point I would like to discuss some of individual experiments that we have been conducting over the past few years. We took a plasmid, which is a small, additional DNA molecule in the large bacterial chromosome. This P1 plasmid is, in turn, the genome of a virus, and it is relatively easy to prove that a mutation has occurred here due to such an insertion sequence being skipped, because viral particles can now no longer be created by induction. We can prove this, and if we were to wait long enough for such a spontaneous mutation to occur and then ask ourselves: Are spontaneous mutations normally the result of point mutation, the substitution of individual building blocks or by the skipping of IS elements, then the answer would in fact be: In this experiment, the likeliest reason for isolated, spontaneous mutations is transposition. That is, not for instance replication errors or other types of nucleotide substitution. We then, on this long genome consisting of 90,000 base pairs, mapped the individual mutations and observed, here, this is figuratively a schematic restriction map. For the specialists among you, the length of the entire genome is simply presented in a linear manner. Here was a point for each independently mapped insertion mutation, and as you can see, they are not random. This means we do not have an arbitrary distribution. That leads us to the conclusion, in these bacterial structures, spontaneous mutations are most often triggered by transposition and do not occur at random. This process is regulated by enzymes. The question is, how do the enzymes control this process? We took a closer look at this insertion hot spot. This is what Christian Sengstag and Patrick Caspers have been doing for the past few years. The answer is depicted here: This is a magnified view of 1,756 base pairs and from these red-coloured IS2 elements, there are the IS2 that occur most frequently, we mapped nine and from the green-coloured IS30 all three, and the answer therefore is as follows: The IS30 element is an element that is capable of a very particular sequence and always inserts at the same site, between the same base pairs. Whether in one direction or the other does not matter. In three independent tests we located it at the same exact site on the entire genome. In the case of IS2, another IS element, which as you see, occurs here more frequently, selects now in detail a different insertion site every time. The sites are never repeated. And this entire segment has been sequenced and we know all the building blocks, the insertion sites are never the same. No rules here. What, this is the question that we are concentrating on, what compels IS2 to enter into this region but not in the other? I cannot give you an answer yet. We have a working hypothesis, namely what you see right here, that this could possibly also be an activation site, and that then the insertion takes place somewhere else. At a given unmeasured distance, and at a different site each and every time. Yet the distance is not arbitrary, so that the activation sequence could perhaps be hidden somewhere in this area. That is another reason why further research is necessary. Now, this non-arbitrariness of the site selection process is significant, because what we have here is a recombination of non-homologous genetic material, meaning the different material sources are not directly related to each other. This fact plays an important role in the biological evolution of micro-organisms, because namely DNA segments from random sources can now form new combinations. Without this material being a priori identical or related, so that these processes, which I cannot explain in detail here, can, indeed through new bonds, can result in new functions. I believe, as do many of my colleagues, that this is a fundamental factor of biological evolution and I have taken the liberty of sketching a theoretical branch on the tree of evolution, which refers to these bacteria only. In the vertical gene flow, meaning from generation to generation, the bacteria can, of course, result in internal restructuring within each individual bacterial strain, and new composites will trigger new gene functions. Yet what is highly significant in this regard is that with the help of both natural vectors of viral genomes as well as conjugative plasmids that exchange material via cell-cell contact, different bacterial strains can exchange packaged informational material. Each package is an individual gene or part of a gene. It is never the whole thing, only short segments which can then be reinserted into the infected cell where they can stimulate new responses, so that this leads us to the conclusion that this evolutionary process is in fact more than just a branch on the tree, but is an entire network with cross connections. That means that for the evolution of a new bacterium in the far future not only this branch or this twig will be of importance, but the current gene pool in its entirety. That should perhaps give us food for thought with regard to preserving our gene pool. I would now in conclusion like give one last example, this time from the field of higher cells. Namely with regard to the immune system, where for many years, it has been known that the genes for antibodies in the embryonic cell are not present as functioning genes from the very start, as we know, it is not until the differentiation stage that these cells are recombined into functional cells capable of producing antibodies for the immune system. Here in the case of immunoglobulin lambda 1, for others it is even more complex than depicted here. It is comparable to taking a segment from a random corner of this long library, then choosing a second segment from a different page and gluing the two together. The interesting thing is that this process of gluing together does not always occur at precisely the same location, only approximately, so that the minor inaccuracy in turn results in a large number of variation possibilities. We have learned to see that through the recombination of, say, about 1,000 different sub-units could give rise to millions of new gene products, each featuring its own specificities. A tremendous diversity. This combination is in turn naturally regulated by enzymes, but it includes a small scope for contingency, just like the ones we have seen in our other examples. Now, that was the last of my examples. A short summary: I have tried to show that at the level of these basic macromolecular interactions, processes controlled by enzymes, that these enzymes of course abide by the laws of nature and obey the instructions encoded in the genome, but however, that many of these processes contain a certain scope for contingency right from the start, and that the enzymes do not do what we expect them to do in a repeatable manner. But that there are instead are alternatives, this element of chance that I described and that by extrapolating this observation to a whole cell or to an organism, we could speculate that the sum of all of these countless contingencies presents sufficient scope for variation in life and, as I have tried to demonstrate, that external factors also influence the actual processes. That we will never be able in advance to predict all of the molecular processes of a living being, and probably, I would like to extrapolate, much less what that organism will do over the long course of its life. That, I believe, is very reassuring, as it could allow us to overcome the fear and unease we feel at the thought of life no longer being interesting once we have discovered all the laws of nature, and that right at birth, each living being could basically receive a certificate stating exactly what will become of this particular life. From the perspective of a molecular geneticist I would reply: That is not what nature intended, nor is it possible. I thank you.

Werner Arber (1984)

On the Unforeseeability of Foreseeable Manifestations of Life (German Presentation)

Werner Arber (1984)

On the Unforeseeability of Foreseeable Manifestations of Life (German Presentation)

Comment

Werner Arber has a long record of involvement with the Lindau Meetings, beginning in 1981 when he gave his first lecture there. After some years he even became a member of the organizational council(das Kuratorium), where he has served since the early 1990’s. When you listen to the present lecture you can hear a speaker who was exceptionally well organized and in complete control of the lecture, which has a well-defined introduction, some topical examples and an interesting ending. The subject matter is whether our lives are fully determined by genes or not. This question derives from the work that is contained in the first part of the citation by Karolinska Institutet when Werner Arber received his Nobel Prize. This happened about 20 years after he, as a young research associate in Geneva, made the initial discovery leading to the concept of restriction enzymes. He is not the only Nobel Laureate who has been rewarded for work partly done as a PhD student and post doctorate research associate, but he belongs to a clear minority. For the introduction to the lecture in Lindau, he brought pedagogical molecular models to explain about the structure of the genes in DNA and how they lead to the production of specific proteins. As one example, he then described how viruses that enter bacteria sometimes have their DNA split up and can add their own genes to the DNA of the bacteria. This is an inherently unpredictable process in nature but is useful in the laboratory because it can be influenced by external conditions. As a conclusion, Arber discussed the evolution of micro-organisms, which through precisely the probability element in the process of modifications of their DNA may have been able to reach the diversity we find on Earth today

Anders Bárány

Cite


Specify width: px

Share

COPYRIGHT

Cite


Specify width: px

Share

COPYRIGHT


Related Content