AUSTRALIAN FRONTIERS OF SCIENCE, 2005
Walter and Eliza Hall Institute of Medical Research, Melbourne, 12-13 April
The secret of success for bacteria
Dr Liz Harry, Institute for the Biotechnology of Infectious Diseases, University of Technology, Sydney
(Click on image for a larger version)
For an impatient person like myself, bacteria are really good to study if you are going to study reproduction, because they reproduce very, very quickly and you can knock things out in a few hours, which is great.
This panel at the bottom left of the slide shows the closest we will get to bacteria having sex. This is a photo taken years and years ago, showing ‘mating’, where cells exchange and donate DNA. So this is called the pilith and these are E. coli cells exchanging DNA.
What I am going to talk about is reproduction. The cells shown on this slide are all different. Bacteria have a very unusual ability, in that they can live all over the Earth. Why are they so successful? Why are they able to live everywhere? One of the reasons is that they are very adaptable and they reproduce very quickly. They come in all different shapes and sizes most of them, not all of them. We have to use the microscope to see them, so we ignore them a lot of the time until they cause problems. But a lot of the time they are very good for us, as well, and without them we wouldn’t be here.
What I want to talk about today is multiplication. How do they multiply? How do they actually manage to double their population in 10 minutes? That is very, very quick. And how are they then able to make sure that each daughter cell that is produced from every cell division gets the right amount of DNA and the correct DNA? They do this very faithfully.
(Click on image for a larger version)
I want to give you today a story about Bacillus subtilis, the organism we have been working on for a long time. Brett Neilan mentioned them. They are really one of the workhorses of bacteria. One of the reasons is that they sporulate, but also we know a lot about Bacillus. E. coli is the other workhorse; many reviews you read about bacteria just have E. coli, but we do know a lot about Bacillus as well. It is a Gram positive organism, so it is a good model for Bacillus anthracis, which causes anthrax it is very closely related and also for some Gram positive organisms which are very nasty, such as golden staph, which is causing a big problem in hospitals because it is antibiotic resistant. So this is our model.
Today I am going to tell you about part of our basic research that explains how a bacterial cell manages to divide at the right time and at the right place to ensure that the DNA gets equally divided when the daughter cells are produced.
Division basically involves the invagination of the cell envelope, which contains the cell wall and the cell membrane. So morphologically it is a very simple process, but in the past we have been very much behind eukaryotes in terms of looking at the cell cycle and how cell division is regulated. Apart from asking how this process is timed and spatially regulated, the other question is how it is coordinated with growth, with DNA replication and with segregation.
(Click on image for a larger version)
I have here a summary of the scope of my research, just to get you familiar with what we do. We have two big global projects. One of them, on proteomics, we have just started with Proteome Systems, and the story I am going to tell you today has actually led to an application of the regulation of cell division and what we know about that to a differential display proteomics approach to look for new drug targets, using a totally basic biology approach and B. subtilis as our model organism.
In the other global approach we are looking for protein interactions. This has involved the yeast-two hybrid global screen and it has brought up a very big network. We have seen those networks published with E. coli in yeast and we have a similar one for B. subtilis, where you get a lot of information about different proteins. We have connected up cell division with one of the proteins in chromosome segregation already, with this screen, but as with all global networks it takes a while to get to some biological significance. We are working on that at the moment.
Some of this work is also confidential, so what I would like to tell you about is a third project, on the regulation of cell division: how is cell division coordinated with things like DNA replication? This is work that is really within our lab, and it came from a collaboration with Gerry Wake. Shigeki Moriya is a visiting scholar in our lab from Japan who has also been working on this with me.
(Click on image for a larger version)
Bacteria are very simple organisms, so why haven’t we got very far with the cell cycle and cell division? It looks simple you can’t see much inside a bacterial cell; even with an electron microscope you just see a cell with a nucleus inside it, and there’s not much density difference. However, there are three challenges, and I think we can meet them now so things are going to change. The first is that there are overlapping cycles.
(Click on image for a larger version)
Most of you would be familiar with the G1/G2 SNM phases of the eukaryotic cell cycle, and that the cells usually need to progress through those different stages to get to the next stage. In bacteria, chromosome replication begins and as soon as it begins the origins are already segregating. So segregation occurs straight up. And the information for cell division to commit occurs soon after chromosome replication has started. So some DNA synthesis has occurred not all of it and certainly chromosome segregation is starting to occur by the time cell division occurs.
Each process overlaps, and within processes, chromosome replication begins and if the growth is fast, then another round will start before the first round is finished. So you get overlapping processes within particular processes, and also these processes overlap. It is very difficult to study in rich media; in fact, nobody looks at it in rich media. If you are looking at these processes you usually use minimal media.
(Click on image for a larger version)
The second thing is asynchrony. If you take a population of bacterial cells, they will be at all different stages of the cell cycle. It makes it very difficult to work out what the relationship between those processes is.
(Click on image for a larger version)
Those first two problems we partly solved by using B. subtilis. It provides us with a good life cycle. Normally, if there are enough nutrients it goes through this vegetative cycle and keeps dividing. When it runs out of nutrients, it senses this and undergoes sporulation to produce a mature spore, which can last probably for ever certainly a very long time.
When this spore germinates and grows out, when conditions are favourable again and we can mimic this in the lab the first cycle is synchronous in the population. And it is not complicated from previous divisions. This has enabled us to do experiments with germinated spores by which we can get clear-cut answers where previously they haven’t been able to be obtained.
(Click on image for a larger version)
The third thing, then, is an apparent lack of cellular organisation. Eukaryotes have organelles, a nucleus; prokaryotes just have a chromosome one chromosome, usually and a whole lot of ribosomes, a lot of proteins and whatever other chemicals. About 10 years ago I was involved in developing immunofluorescence procedures for bacteria, and with the event of the discovery of GFP and using those fusions we now know that the bacterial cell is highly organised at the level of protein localisation. Many (though not all) proteins that you look at have a specific localisation, and very important processes such as cell division, DNA replication and cytoskeletal processes involve proteins having a particular address in the cell. And often it is also a dynamic event.
(Click on image for a larger version)
So in cell division, as I will explain later, you have a Z ring, which is like a tubulin ring; the DNA replication factor is usually at midcell, so it is placed in a particular place. Until about two years ago it was believed that prokaryotes don’t have a cytoskeleton. It is just that you can’t see it with the EM. But if you label a specific protein that people suspected might be involved in cell shape control, you see a helix. People suggested it was actin, and it was crystallised in Cambridge and shown to be very, very similar to the three-dimensional structure of actin.
This has really been a cell biology revolution, and it is only since about 10 years ago that we have able to see these sorts of things in bacteria. So now we have those problems sorted.
(Click on image for a larger version)
I want to talk today about this first process in cell division, Z ring formation. FtsZ is a tubulin-like protein, and at a certain time prior to division it forms this ring by self-polymerisation. It has GTPase activity; it is very similar to tubulin in its biochemistry. You can see this by using an FtsZ-GTP fusion or you can now see it with immunofluorescence.
So to ask how a cell determines where and when it will divide you really have to ask the question: how does a cell determine where and when the Z ring will form?
What we wanted to answer is: how does the cell coordinate chromosome replication with cell division? We hypothesised that the Z ring would be controlled by the progress of the round of chromosome replication, and that is how there would be at least a connection between the two. Because Z is highly conserved it is found in mitochondria of primitive organisms and in chloroplasts and in almost all bacteria and it acts early, we thought that it would probably be involved in this link.
(Click on image for a larger version)
And that turns out to be true. We did this using the outgrown spores. We had a look at the first round of replication; we inhibited that round at various stages and asked the question: what happens to the timing of assembly and the positioning of the Z ring? We found a relationship between the early stages of replication and the ability of Z to form a ring precisely at midcell.
So in wild-type cells this is just a phase immunofluorescence image the Z ring forms at midcell, it forms at a certain cell length, and it normally forms after a significant amount of DNA synthesis. We don’t know how much, but it has to be some.
If we block the early stages of replication we do get Z rings, but they form at the wrong place. And that was reasonable; we expected that to happen. But there was something we did as a control which was a surprise. That is, there was another situation where we could get midcell Z rings: if we blocked replication at early stages but only certain treatments. That treatment happened to be in a thymine requiring cell. We left the thymine out of the medium so that the cells could initiate replication but they couldn’t make any DNA, they couldn’t synthesise, and we got mainly central Z rings in this case.
Most of the time, if you block the early stages of replication, theoretically blocking DNA synthesis in the same way, you will get acentral Z rings. The question was: why are we getting central Z rings in this case? What we suggested might be happening is based on the result you see in the upper panel here, which shows us that the positioning information for a Z ring to form occurs is there already at the very early stages of DNA replication, and normally it is blocked by the early stages of replication, to be relieved later on. And in this alternative case we are bypassing that block.
(Click on image for a larger version)
In other words, no checkpoint for bacteria had ever been identified, but we suggested that the early stages of DNA replication block midcell Z ring assembly, and this is relieved later in the round. So it is relieving a checkpoint, and this makes sure that Z rings don’t form too early that is, division doesn’t occur too early in the cell cycle. If that happened, the newborn cells would not have the right DNA in them.
(Click on image for a larger version)
To cut a long story short, we came up with a model. The reason we implicated the replisome in being involved in this checkpoint was that in minus-thymine conditions that gave rise to premature midcell Z rings, our results indicated the replisome was unstable. And so we developed the very simple hypothesis that, because the replication factor is at midcell and so is the Z ring, once the replication factory binds to the origin at midcell to initiate replication, it blocks the site for Z ring assembly, and only after a certain amount of replication does this resolve. And this had been shown in a Science paper, that you need a certain amount of replication for these two replisomes to resolve. And then a midcell Z ring site is unmasked.
These replisomes now represent two replication forks. At this point they are actually together.
So does this work? Can we suggest that the replisome physically blocks the site for Z ring assembly? One of the things we wanted to look at was: if it does, we would suggest that the replisome is as precisely located at midcell as the Z ring. We had reason to believe that this is not the case, but we had to do the experiments.
(Click on image for a larger version)
As you can see, when we label the PolC with GFP and the Z ring with Z YFP and we can do this in the same cells, as well as separate cells you can see that the scatter about midcell, which is 0.5, is 2 per cent for the Z ring it is fairly precisely located or placed but PolC, which is one of the sub-units of the replication factory, actually is far more scattered than the Z ring. This suggests, then, that the replisome at midcell is not likely to act as a physical block to Z ring formation.
We did another experiment. One of the reason we thought of for why these are scattered was that perhaps it just doesn’t localise very well in the first place, but the other possibility is that it moves about. That is, the Z ring stays put and the PolC or the replication factory moves about.
(Click on image for a larger version)
This turns out to be true. Again if we use germinated spores we can just look at one round, because you can slow that round down and you are looking at the same cycle in the cells.
So when I press the button here, you will see that the PolC moves around and you will also see that in this particular movie the one spot quite often becomes two.
(Click on image for a larger version)
What I told you before with this model is that the replisome moves about and it is less scattered so it can’t physically block the midcell site for Z ring assembly. It was suggested in a Science paper that not only is the replisome stationary but it resolves after 80 per cent of the round. That was just a suggestion, and we don’t think this is true now because they seem to be resolving throughout the round of replication.
(Click on image for a larger version)
And most of the movies I think about 90 per cent of them actually showed that this spot, shown here as I repeat this PolC movie, actually came apart into two spots very, very often.
So it is not that simple model. And we weren’t surprised, because most of the models that you think up are usually useful for experimental testing, not for telling yourself that you are correct.
(Click on image for a larger version)
So what are we left with? We still believe there is a blocker; we have done some more experiments to suggest that this is a checkpoint. At the moment we are not sure what this is, and this is what we are wanting to find out.
(Click on image for a larger version)
I can show you some preliminary results but I can’t reveal all the results at this point because they are confidential. At the moment we have two types of DNA replication mutants as examples. Almost all of them fall into this category. If you block initiation of replication the very first stage of replication you usually get an acentral Z ring. This is seen in the two cells in the mid-top panel, one of them on its side but recognisable by its horseshoe shape. And in the right-hand top panel you see the chromosome, in red. So the chromosome tends to stay central, and these are forming to one side of the chromosome.
That’s what usually happens. As I told you before, if you leave out the thymine, sometimes they form central. So you can relieve what we think is a checkpoint.
But recently we have identified a whole lot of DNA replication mutants, and we have looked to see whether, under any of the conditions where these mutations take effect, we can get midcell Z rings by completely blocking initiation. And one of them actually does give us midcell Z rings.
So now I think we are able to get to the point where we can show what the direct connection is between DNA replication and cell division. This is a highly controversial area, and no-one has yet ever shown a direct relationship between the two. That is what we are interested in doing, because some sort of relationship must exist.
(Click on image for a larger version)
I come now just very briefly to the other little interesting thing that we have done. I told you before that division starts with Z rings, and initially it was thought that these are all distributed throughout the cytoplasm and when you get a Z ring it is a very obvious band by fluorescence.
(Click on image for a larger version)
However, lately we have seen something else, just by looking more. It is really interesting, when you think you know what you are looking for, how biased you can be. We have changed the cell biology conditions a little bit, because students year after year would say to me, ‘I’m seeing something there other than a ring.’ And one of them did look a bit further and saw that in short cells, specifically, instead of this obvious band you saw a sort of a spiral shape, as in the bottom left-hand panel of this slide. If you look at this with YFP you tend to see a better resolution of it.
The next question was: does it move? And we find that it does, as in this little movie that I am showing now. So, rather than Z forming from just a monomer or polymer in the cytoplasm, we think that it actually forms a spiral. We certainly see these in E. coli as well as Bacillus. What we think then happens is that somehow it goes to a ring. And the big question is: how does it do this?
In terms of linking replication with division and how division is controlled in time and space, this process now is what is going to become the most interesting: how do these sub-units arrange themselves or manage to get from a spiral to a ring? It is a very dynamic process, it is not as simple as just all reorganising themselves so that there is nothing in this bottom right-hand panel. The exchange is something like eight seconds within the ring. We don’t know what the exchange is within the spiral. So it is a very interesting kinetic problem. And we also see little spirals off the ring. There is a constant movement and exchange of these Z molecules, and we would like to know how that occurs.
(Click on image for a larger version)
These are my two collaborators. And Margaret Migocki has done most of the movies that I am showing you, and Phoebe Peters has been doing the spiral work. And the ARC has basically funded all of this work.
I should say that I have just moved to IBID, the Institute for Biotechnology of Infectious Diseases, and it is great to be somewhere new and working on infectious diseases.
