AUSTRALIAN FRONTIERS OF SCIENCE, 2008

The Shine Dome, Canberra, 21-22 February

Recognition of rust effectors in plant innate immunity
by Dr Peter Dodds

 

Peter Dodds is a senior research scientist with the CSIRO Plant Industry in Canberra, Australia. He has a PhD degree from the University of Melbourne, Australia on gametophytic self-incompatibility. He did postdoctoral work on pollen-pistil interactions at the Plant Gene Expression Center in Albany, California, and then returned to Australia as an ARC Postdoctoral Fellow at Plant Industry, analysing the evolution of disease resistance genes and specificity in flax rust disease. Peter has been project leader at Plant Industry since 2001, focusing on rust pathogen biology. His current research involves the identification of virulence and avirulence factors from rust fungi and investigating their role in disease, as well as the molecular basis of recognition between host resistance and rust avirulence proteins and the implications for host-pathogen co-evolution.

I am going to talk about the role of rust effectors. These are proteins which are secreted by rust fungi when they cause disease. We got into this from quite a different direction, actually. What we have been trying to do, over quite a number of years now, is to understand the plant immune system: how do plants recognise and respond to diseases, and what mechanisms do they have to do that?

It turns out that plants have evolved a number of ways of coping with diseases, which are in many ways analogous to the immune systems of animals – and humans as well. It turns out, however, that although there may be some similar components involved and similar concepts in terms of the recognition, really they operate in quite different ways.

The disease that we have been focusing on is a particular fungus called Melampsora lini, which is a rust disease infecting a flax plant. Flax is where linseed oil comes from – and also linen – so this is a common disease of linseed. It is also a very important disease of wheat.

(Click on image for a larger version)

There are actually three different rust species that infect wheat. Shown here is one of them, stem rust, which is a major disease of wheat. It can cause extreme losses of yield in wheat fields, and here you see a wheat field where there has been a complete loss of yield due to rust infection. As you can see here, however, what is happening is that the farmer is driving his tractor through the field to plough up the wheat, and as he goes through, there is a big cloud of all the rust spores that are being released from the infected wheat. So it really produces an enormous knock and is having a devastating effect on yields in agriculture.

The way in which these diseases are controlled in agriculture, in wheat and many other crops as well, is through breeding. Breeders go out and try and find plants which are resistant to disease, then they take those plants and cross them with the ones that are susceptible. And if there are genes which are controlling that resistance, then they are able to introgress those into cultivars which they can release commercially. So there is a genetic control of resistance. We have been trying to understand how that control of resistance is mediated.

(Click on image for a larger version)

These are some of the questions that we have been interested in. First of all: what are rust resistance genes, and how do they work? Breeders have been using these for probably the last hundred years, doing crosses from plants that are resistant to plants that are susceptible, and introgressing these resistance genes. But it is only in the last 10 years that we have actually started to understand what those genes are, and how they work.

The other really important question to ask is: how do rusts overcome resistance? I haven't said this yet, but when you introgress a resistance gene into a plant and put that out in the field, what you often find is that after a number of years – maybe 10 or 20 years – the rusts evolve and overcome resistance. Then you get new epidemics of disease, and very major effects on agriculture.

There are a couple of such examples at the moment. Two years ago in Australia there was a major outbreak of stripe rust, which is one of the three rust diseases that affect wheat. There was an epidemic of stripe rust, and that cost the industry about $100 million in terms of fungicides that had to be applied to protect the crops. On top of that, there were major yield losses as well.

The other major issue recently has been the emergence of a new strain of stem rust in Uganda, the Ug99 strain. That overcame the single resistance gene that was deployed all through Africa, India and the Middle East to protect wheat from stem rust, so that is causing a major crisis at the moment in Africa. And that rust strain is starting to spread out of Africa towards central Asia as well, so there is a global initiative going on now to try and breed new sources of resistance into the crop varieties that are used in that part of the world.

The third question we have had is: if we can answer the previous two questions, can we manipulate this system and try to engineer new types of resistance? So what I am going to show you today is some of the advances that we have made in terms of identifying what rust resistance genes are, how they work – we have got a few answers there – and also in terms of understanding how the rusts overcome resistance.

I will outline at the end a new strategy that we are working on at the moment to see if we can use this knowledge to try and engineer new resistance genes that we can use to protect plants against not only diseases such as rust but other diseases as well.

(Click on image for a larger version)

Before I go on, I am just going to talk a little bit about the biology of the rust infection.

When you look at an infected plant, what you see is that the plant is covered in orange or red or brown pustules. But that doesn't really tell you a lot about what is going on. What is actually happening is happening inside the leaves.

When a spore infects a plant, first it lands on the leaf surface. The spores are usually spread by wind, and can actually be spread very long distances. You saw the large cloud of rust spores that was emerging from the field as it was being ploughed. Some of the introductions of rust into Australia have actually come from South Africa – they have been blown across on the trade winds from South Africa and landed in Australia.

When a spore lands on a leaf, it germinates and produces a small germ tube such as you can see on this slide; it actually enters the leaf through some of the natural openings in the leaf, called stomates. (This is what the leaf uses for gas exchange to allow photosynthesis to occur.) As it penetrates the leaf, it grows hyphae which spread through the spaces between the cells in the leaf, and they form specialised structures called haustoria. So these are actually formed inside the plant cell, and this is the major source of nutrient for the rust. It is able to extract nutrients from the plant cell and use them to continue growing throughout the leaf, and eventually it differentiates and forms more spores, which are then released from the infected plant. So when you have a heavy infection, you can really get enormous numbers of spores.

(Click on image for a larger version)

Shown here are the sorts of numbers that lead to that big cloud of rust spores. And here we see an infected flax plant.

(Click on image for a larger version)

The really critical part of this infection process is the formation of the haustoria, so here we have an electron micrograph picture of a haustorium. (This is actually from an infected wheat leaf.)

What you see here is the plant cell wall marking the outside of the plant cell; up at the top, almost off the picture, is where the rest of the fungus is. The haustorium has penetrated through the plant cell wall. It has a long necklike structure and then a large body which actually exists inside the plant cell.

This body is surrounded by two membranes. The inner membrane is the haustorial membrane, and around the outside of that is the plant cell membrane – the plant has a cell wall with a membrane just inside that, and then that continues around the outside of the haustorium. We know from various biological studies that this is a major site of nutrient uptake for the rust: there are amino acid transporters and sugar transporters that are expressed along this membrane, and these are responsible for extracting those nutrients from the host cell.

It also has a really major effect on the host cell metabolism and cell biology. One of the things you really notice in this is that the host cell nucleus migrates in the host cell and comes to sit right next to the haustorium. So it appears that the haustorium is actually manipulating the cell biology of the host cell and changing the sorts of genes that are expressed in that host cell and the sorts of metabolism that is occurring in the cell.

That is very important for the rust to establish disease. But it also turns out to be very important for the plant to establish resistance.

(Click on image for a larger version)

This is another diagram looking at the same sort of thing: you have an infection process, with the formation of a haustorium, and in a susceptible plant that leads to further differentiation of the fungus and you get a production of large numbers of pustules.

(Click on image for a larger version)

But in a resistant plant you have a recognition event, and that always occurs at the site of the first haustorium formation. The major response that the plant has on recognising the presence of the haustorium is that it induces a 'hypersensitive' response. This is really just a localised cell death. Basically, what it is doing is that it is killing off this cell so that the rust is not able to extract nutrients and then can't grow and spread through the plant.

So if you look at a resistant plant – actually, the one shown here has been given a very heavy infection – you see a lot of very small necrotic spots, representing sites of attempted infection by the rust where the plant has actually mounted a successful defence response.

In trying to understand what is actually going on here – what are the components in the plant which are recognising the rust, what are the rust factors that they are actually recognising – we have really focused on the haustorium structures. It turns out that you can isolate these haustoria free from the rest of the plant and the rest of the fungus from an infected plant. So if you take an infected plant, grind it up (in a blender, essentially) and put it through a mesh and then over a column which has a protein on it which will bind to some of the polysaccharides on the surface of the haustorium, you can isolate very pure preps of these structures. Then you can look at what sorts of genes are being expressed in those structures, and what role they are playing in disease.

 

(Click on images for larger versions)

Essentially, to summarise a whole lot of experimental work: this is the model that we have come up with, in terms of what is happening during disease and during resistance.

At the left of this slide you can see the rust fungus growing in the intercellular space. When it contacts a plant cell, shown here in green, it forms a haustorium. As I said, there are two membranes that surround this, the inner membrane being the fungal cell membrane and the plant cell membrane going outside that.

Inside the haustorium a whole suite of 'effector' proteins are being synthesised, and they are secreted proteins: they are actually secreted from the haustorium and cross its membrane into the extrahaustorial matrix.

Somehow – and we don't know how this happens – some of those proteins, at least, end up inside the plant cytoplasm. So they cross the plant membrane, and there are a couple of options as to how they may do that. They may cross it directly, they may be taken up through the plant retrograde transport system, there may be a specific process that is induced by the haustorium to allow this to happen, or they may just be proteins that are intrinsically able to cross membranes. We don't know how that happens, but they get inside the plant cell and presumably what they are doing there is bringing about all the changes in the host cell which are necessary for infection.

So there are a lot of things that need to happen to make this a good site for the rust to exchange nutrients. You have got to change the metabolism of the host cell, you have got to make lots of amino acids and sugars available for uptake by the haustorium, you have got to turn off some of the basal defence responses so that you are not inducing any of the resistance responses.

This is something that is happening in a susceptible interaction, when you get disease and it leads to the formation of rust pustules.

 

(Click on images for larger versions)

There is actually a cost involved in doing this. We find that when you hit a resistant plant, it is these proteins which are absolutely critical for infection because they are bringing about all those changes in the host cell, but it is these proteins which are actually triggering the defence response when we see a resistant plant. So these are recognised by the plant immune system, and it turns out that the resistance genes that breeders use to confer resistance to different diseases are the recognition components of the plant immune system.

(Click on image for a larger version)

We have been able to isolate quite a number of those 'resistance genes' from flax over the last 10 to 12 years, and they all have a common structure. This is replicated in other plants for resistance genes against many other diseases as well, so they really represent a common component of the plant immune system.

They have a typical structure, as illustrated in this slide. There are three major parts to the resistance protein. There are two domains at one end of the protein which are involved in signalling, so they are actually inducing a resistance response, and there is a second part of the protein which contains a leucine-rich repeat which is involved in recognition. So this is the region of the protein which is actually recognising the effector proteins from the rust fungus.

This is actually inducing a very effective resistance response. In the case illustrated here, if you look at the difference between resistant and susceptible plants you can see a really dramatic difference between the two responses. So the ability to recognise specific effectors from the rust and then signal a defence response really gives you a very effective resistance mechanism against disease.

(Click on image for a larger version)

A lot of you would have some familiarity with the human immune system, but perhaps not very much with the plant immune system, so I just want to briefly compare the two systems. Shown here are some of the features which really distinguish them.

In animals and in humans you have an acquired immunity system. This is mediated by specialised cells which circulate through the body, through the blood system. So there is a specialised line of cells which contribute to immunity, and they have what is called an adaptive recognition system: when they are exposed to a pathogen, they are able to generate new recognition capacities which give better protection against that pathogen. This is the basis of vaccination.

The first key difference in plants is that the immune system is cell autonomous, so there is no specialised immune system and there is nothing circulating through the plant. Every cell in the plant has to be able to react independently to infection. The second key feature is that there is no adaptive recognition: it is pre-determined recognition. So there is a genetic system, with resistance genes – and this is why breeders are able to do crosses and find new resistance genes in that way, because all of the recognition capacity is pre-determined in the genetic structure of the plant.

It turns out that animals have an underlying immune system which acts in addition to the acquired immune system. This is called innate immunity, and in many ways it is analogous to the plant innate immunity system as well, because again it is cell autonomous and it is non-adaptive. The key difference between the systems is that the innate immunity system involves recognition of some really generic pathogen molecules, generally called pathogen-associated molecular patterns, PAMPs. These are things like lipopolysaccharides. All of the Gram-negative bacteria which can cause disease have a cell wall which contains lipopolysaccharides, and this is a very generic recognition mechanism that can recognise microbes – which are not necessarily going to cause disease, but they may. And the outcome of the action of the innate immunity system in many animal cells is to stimulate and attract some of the circulating immune system.

But in plants this innate immune system, which in animals is really quite restrictive – there are only a few of these things around and they only recognise a very small number of different PAMPs – has really been elaborated.

There are, again, two levels of recognition in plants. There is one which involves PAMP recognition, similar to the innate immunity in animals, but this gives a very weak and non-specific type of a resistance response. So this is involved in preventing infection by opportunistic microbes, rather than things that are actually pathogens. Anything that is a pathogen contains a whole lot of effectors, such as the ones I talked about in rust which are being delivered into plant cells, and one of the main functions of those effectors is to turn off this PAMP recognition. But then again there is the effector recognition system in plants, which is mediated by the resistance genes. This provides very strong and highly specific resistance.

(Click on image for a larger version)

The second key question I was interested in was: how is it that the rusts are able to escape resistance? When we look at the effector proteins in the rust, what we find is that they are highly variable in sequence: when you look at different rust strains you find different variants. They have been under very strong selection, obviously, to avoid recognition. And when you look at some of the strains of rust which are able to escape resistance, you find that they have different surface properties in the proteins.

This information has come from a collaboration we have had with Bostjan Kobe's group at the University of Queensland – he was able to crystallise some of the avirulence proteins, the effector proteins.

(Click on image for a larger version)

Shown here is the structure of two of the effector proteins that we have found in flax rust. These are two different variants: one is recognised by a resistance protein and the other is not. Essentially, they have identical structures; they only differ in some of the surface properties of the protein. On this slide they are colour coded to represent different chemical properties that are exposed on the surface of the protein, so different colours represent different chemical properties, and there are a couple of areas where you see substantial differences in the chemical properties on the surface of those proteins.

(Click on image for a larger version)

We have done extensive analysis of these proteins, and basically exchanged those residues between these two proteins, and we find that all of them contribute in some way to the recognition of the protein.

(Click on image for a larger version)

Here is a model that we have for how that recognition occurs. It shows the recognition domain of the resistance protein, which is modelled on a related structure, and the structure of the rust effector that it recognises.

This is a docking model of those two proteins, showing how they might come together. We can see that there is a very large surface area where they are able to physically interact, and it is the combination of chemical residues that are exposed on the surface of the two proteins that determines whether or not a recognition event is occurring there.

(Click on image for a larger version)

If the rust wants to escape recognition, what it has got to do is to change some of those surface properties on the effector proteins.

What happens in nature is that you get what we call a co-evolutionary arms race. You might start off with a situation where you have an effective resistance gene; the rust evolves to overcome resistance by changing its effector proteins; and then you get concomitant selection on the host to develop new resistance proteins which are able to recognise the new variants of the avirulence proteins.

That is what happens in a natural system. In agriculture what happens is that you put out a variety which has an effective resistance gene; the rust evolves and changes its effector; and this resistance gene is no longer effective. And because this is in agriculture and it relies entirely on breeding, the development of new resistance proteins never happens. So breeders always have to go out and try and find a new source of resistance from wild populations.

So what we want to try and do is to replicate the evolutionary process in the lab, and see if we can generate new variants of resistance proteins.

(Click on image for a larger version)

This slide sets out the strategy that we have in mind for doing that. It is basically just replicating what happens in nature in terms of evolution.

We would start off with an effective resistance gene that we know works, and generate a large pool of mutagenic variants of that resistance protein. And then what we can do is select for the ability of those proteins to interact with a rust effector protein. So we start off with a situation of resistance that doesn't recognise the effector protein; we generate a whole lot of variants, and see if we can pull one out that actually does recognise the effector.

Then we take that protein and put it back into transgenic plants, and look to see whether or not we can get resistance. So this is a proof of concept experiment at this stage that we are conducting in the flax system.

(Click on image for a larger version)

What we would like to be able to do is to apply that concept to important crops such as wheat. In wheat there is much less known, but there are about 50 stem rust resistance genes, for example, that have been defined genetically in wheat, and we have some projects to try and clone some of those.

There is a stem rust genome project which is under way at the moment, being conducted by some collaborators of ours, Les Szabo and Christine Cuomo.

What we want to be able to do is to be able to predict all of the secreted effectors from the stem rust haustoria, and then use these as targets for recognition by the host.

(Click on image for a larger version)

To sum up quickly: what we have been finding over the last few years is that rust haustoria deliver a whole suite of different effector proteins into plant cells. And some of the questions we still have are: how are they transported into the plant, and what is their function in disease?

Also, what we realise now is that the resistance genes are encoding immune receptors that recognise those effectors. So we really need to understand what is the structural basis of that interaction – I showed you a model, but we need to test that experimentally.

Other questions are: how does the protein interaction activate the resistance response, and, lastly, can we engineer new resistance genes that we can use to protect agricultural crops?

(Click on image for a larger version)

Lastly, I want to acknowledge some of the people who have been involved in this work. I work at CSIRO, in Plant Industry; Jeff Ellis is the 'big boss' there and he established the flax rust system in Plant Industry and has been a tremendous mentor to me over the last 10 years, while I have been there. And then there are some of our collaborators, especially Bostjan Kobe's group, which has done the structural analysis; and the genome sequencing groups at the University of Minnesota and the Broad Institute.