Session II: Pathogenicity of bacteria, fungi and viruses
Complex spatial changes in gene expression in response to virus invasion
of compatible hosts
Virus invasion of host plants is a multistage process. After multiplying in the initially infected cell, the virus moves symplastically to adjacent cells through plasmodesmata. Rapidly the infection will encounter parts of the vasculature where the virus can be loaded into the phloem for systemic translocation to distal regions of the plant. This is a passive process, which follows the demands of sink tissues for photosynthate generated in source tissues. The time taken for each of these stages is poorly resolved although estimates of the duration of virus multiplication in single cells in vivo are just a few hours. Hence, virus infection of plants is a completely asynchronous process, making the correlation of virus-induced changes in host functions with particularly stages in virus multiplication or movement very problematic. To overcome this difficulty we have applied a spatial analysis to advancing infection fronts such that, for example, changes close to the front can be considered as early events and those behind the front as late events.
Two experimental systems have been used to investigate changes in host gene expression induced by virus infection, namely pea cotyledonary tissues infected with several different viruses, and Cucumber mosaic virus (CMV)-infected cucurbit tissues. In both tissues, one striking effect of virus infection is the reduced accumulation of most host gene transcripts. A phenomenon akin to host gene "shut-off", originally observed for animal virus infections. A second common observation is the accumulation of hsp70 mRNA and protein at the extreme edge of the infection. In the pea system, we also see that the expression of actin and tubulin remain unchanged. In addition, in the cucurbit system, two further types of host gene upregulation were observed. The expression of HSP70 and NADP-dependent malic enzyme showed induction in apparently uninfected cells ahead of the infection. This response was more localized than the upregulation exhibited by catalase expression, which occurred throughout the uninfected regions of the tissue. Collectively, these experiments showed that virus infection induced immediate and subsequent changes in host gene expression and that it has the potential to give advance signaling of the imminent infection. The impact of these effects in relation virus replication and movement, and/or host defence will be discussed.
Induced resistance to plant viruses
Salicylic acid (SA) is part of a signal transduction pathway (STP) that induces resistance to viruses, bacteria and fungi. In tobacco and Arabidopsis the defensive STP branches downstream of SA. One branch induces PR-1 proteins and resistance to bacteria and fungi, while the other triggers induction of resistance to RNA and DNA viruses. Our initial evidence for the existence of the virus-specific branch was based on pharmacological data. Specifically, resistance to viruses can be activated using antimycin A and cyanide, or inhibited with salicylhydroxamic acid (SHAM), independently of the induction of PR-1 gene expression (Murphy et al. TIPS 4: 155, 1999). Recent work from Klessigs group appears to confirm the existence of the virus-specific branch. They have shown that in Arabidopsis, induction of resistance to turnip crinkle virus does not require the activity of the NPR1 gene, a regulator of PR gene expression (Kachroo et al. Plant Cell 12: 677, 2000).
Our results using antimycin A, cyanide and SHAM have suggested a role for the mitochondrial alternative oxidase (AOX) in resistance to viruses. This is because AOX activity is known to be stimulated by antimycin A and cyanide, but inhibited by SHAM. However, pharmacological experiments can never be taken as conclusive. Therefore, we are producing transgenic tobacco with modified levels of Aox gene expression and attempting to silence Aox gene expression using virus-induced gene silencing. These constructs will be used to assess definitively whether or not AOX plays a role in resistance to viruses and/or the virus-induced hypersensitive response.
In addition to this, we are investigating the resistance mechanisms that are induced by the virus-specific pathway. We have shown that multiple resistance mechanisms are induced by SA. Thus, in tobacco, SA induces resistance to the systemic movement of cucumber mosaic virus but has no effect on its replication or cell-to-cell movement. However, in the case of tobacco mosaic virus (TMV), SA appears to be able to inhibit the spread of virus out of the initially inoculated area of tissue. We have studied this further using a combination of methods that include analysis of viral RNA and protein accumulation in inoculated tissues and protoplasts, as well as in vivo imaging of green fluorescent protein-(GFP) tagged viruses in plant tissue. Our data indicates that SA engenders resistance to TMV by inhibiting its ability to spread between epidermal cell layer and by inhibiting its replication in the underlying mesophyll tissue. We believe that these effects may contribute to the reduction in the spread of virus observed in plants expressing systemic acquired resistance.
Molecular genetics of plant infection by the rice blast fungus Magnaporthe
The rice blast fungus elaborates a specialized infection structure called an appressorium in order to infect rice leaves. The appressorium is a dome-shaped cell which differentiates from a germ tube, shortly after conidial germination. M. grisea appressoria develop turgor pressure which is translated into mechanical force to breach the plant cuticle. This allows a narrow penetration peg to enter the leaf epidermis and colonise the tissue, later forming large bulbous infection hyphae.
We are investigating the process of appressorium-mediated infection in M. grisea and in particular the mechanism by which turgor is generated. M. grisea appressoria accumulate very high concentrations of glycerol which acts as an osmolyte, allowing the cell to take up water and develop hydrostatic turgor. Appressoria form in water on the leaf surface and therefore glycerol is synthesised from storage products in the spore. M. grisea conidia contain a number of storage compounds including glycogen, lipid and trehalose. We have used a combination of genetic, biochemical and cell biological methods to study the relative contribution of each storage product to glycerol generation. Trehalose degradation occurs rapidly during conidial germination. Trehalose is synthesised in M. grisea by trehalose-6-phosphate synthase encoded by the TPS1 gene. D tps1 mutants are extremely reduced in pathogenicity, due to a defect in cuticle penetration. The degradation of trehalose meanwhile appears to occur due to the activity of at least two trehalases, encoded by NTH1 and TRE1. D tre1 and D nth1 mutants are also affected in their ability to cause disease symptoms, although to a lesser extent than in D tps1. Genetic control of trehalose metabolism appears to be, at least in part, due to the action of cAMP-dependent protein kinase A (PKA), based on enzymatic assays. Glycogen and lipid stores are also degraded during conidial germination and can be observed accumulating in appressoria during their formation, before disappearing as turgor is generated. The movement of these reserves is controlled by the PMK1 MAP kinase pathway and degradation in the appressorium is regulated by PKA. We are currently identifying and characterising genes encoding enzymes involved in lipid and glycogen degradation in appressoria with the aim of determining the basis of turgor generation and apressorium function in M. grisea.
Pathogenicity and host range determinants in root-infecting fungi
Fungal pathogens are likely to require a core set of determinants that enable them to grow in plants, including factors such as PMK1-related MAP kinases. In addition to these general requirements additional genetic components will determine tissue and host specificity. The ability to tolerate antifungal secondary metabolites produced by the host plant represents one facet of specific interactions between fungal pathogens and their hosts. For example, detoxification of oat root saponins (avenacins) by the "take-all" pathogen Gaeumannomyces graminis var. avenae is required for successful infection. Like many plant secondary metabolites avenacins are localised in the vacuole, and so are likely to be released in response to infection by G. graminis, which is a necrotroph. The isolation and characterisation of saponin-deficient (sad) mutants of the diploid oat Avena strigosa has allowed us to confirm that avenacins are required for resistance of oat to both pathogens and non-pathogens. Avenacin biosynthesis is restricted to the genus Avena, and cereals other than oat are generally deficient in the ability to synthesise saponins of any kind. The isolation of saponin biosynthetic genes will allow us to address important questions about the evolution (or loss) of triterpenoid saponin biosynthesis in cultivated monocots and also has implications for novel disease control strategies. To this end we have recently isolated the Sad1 gene, which encodes b -amyrin synthase (the first committed enzyme in the saponin biosynthetic pathway) and isolation of other Sad genes is in progress.
Avenacin detoxification is a determinant of host range for Gaeumannomyces graminis, and facilitates entry into the root through the avenacin-laden epidermal cell layer. However very little is known about other factors that are required by G. graminis and other soil-borne fungal pathogens for successful colonisation of root tissues. This is largely due to the genetic intractability of many of these fungi. G. graminis belongs to the family Magnaporthaceae, which includes other root-infecting fungi such as the turfgrass pathogen Magnaporthe poae and also the rice blast pathogen Magnaporthe grisea. M. grisea has recently emerged as a paradigm for molecular genetic dissection of factors determining fungal pathogenicity to leaves, and mutational analyses have identified a number of genes that are required by M. grisea for pathogenic differentiation and colonisation of leaf tissue. We have found that M. grisea can infect the roots of cereals and can be used for genetic analysis of fungal root infection. This has allowed us to distinguish mutants with leaf-specific, root-specific and general defects in ability to cause lesion formation on leaves and/or roots. For example, the melanin biosynthetic mutant alb1, which is non-pathogenic to barley leaves, is unimpaired in its ability to cause lesions on roots of barley or wheat. Conversely, a global regulator of nitrogen assimilation (NUT1) is essential for root infection but is dispensable for disease development on leaves, while the MAP-kinase PMK1 and the predicted ABC transporter ABC1 are required for infection of both tissues, and so represent general disease determinants. This system now offers considerable potential for comparative analysis and elucidation of the genetic components required for fungal infection, proliferation and disease development in leaves and roots. In a search for new mutants we have identified two restriction enzyme mediated integration (REMI) mutants of M. grisea that fail to give lesions on wheat and barley roots. The characterisation of the genetic defects in these and additional mutants isolated in further screening experiments will allow the identification of novel tissue-specific and general factors that are required for colonisation of plants by phytopathogenic fungi.
Raincoats and pathogens: the function of hydrophobins in Cladosporium
Hydrophobins are small proteins produced by most true filamentous fungi. They are usually secreted into the growth medium and assemble as rodlets on the outer surfaces of fungal structures to form a hydrophobic layer. All the hydrophobins so far studied are characterised by a conserved 8-cystein motif. These proteins have been grouped into two classes which differ in physical and chemical properties and in the spacing and arrangement of the cysteins. We have found that Cladosporium fulvum, the causal agent of tomato leaf mould, has at least 6 hydrophobin genes. We have named these HCf-1 to 6. HCf-1, -2, -3 and 4 are class I hydrophobins and their amino acid sequences are similar to each other. HCf-5 and 6 are class II hydrophobins and are similar to each other but to a lesser degree than the class I proteins. HCf-4 and HCf-6 are both "bimodular" hydrophobins and have an N-terminal domain of about the same size as the hydrophobin domain. The N-terminal domain of HCf-6 is extremely rich in glycine and asparagines residues. The function of these two domains is unknown. All HCf genes are expressed at different stages of development and in response to different nutritional status of the growth medium. We are currently creating mutants with altered levels of hydrophobins. We achieve this by deleting single hydrophobin genes (using homology-dependent gene deletion: "knockouts") or by reducing the overall expression of hydrophobins with similar sequence by homology-dependent gene silencing. These mutants enable us to test the function of the hydrophobins in development and infection. We have tested the possibility that because of surface localisation they mediate interactions between the fungus and the plant, perhaps by shielding the hyphae from detection and/or attack by the host. This does not seem to be the case in C. fulvum: so far we have found that HCf-1, HCf-2 and HCf-6 are not required for infection under laboratory conditions. However, HCf-1 and other class I hydrophobins, contribute significantly to the hydrophobicity of the conidia. This, in turn, appears to be an essential feature for dispersal mediated by water droplets.
Spanu, P. Gene 193, 89-96 (1997); Spanu, P. Physiological and Molecular Plant Pathology 52, 323-334 (1998); Hamada, W. & Spanu, P. Molecular & General Genetics 259, 630-638 (1998); Segers, G. C., Hamada, W., Oliver, R. P. & Spanu, P. D. Molecular & General Genetics 261, 644-652 (1999).
Unravelling the molecular mechanisms of pathogenicity in the soft rot
pathogen Erwinia carotovora: a new approach
Erwinia carotovora subsp. atroseptica (Eca) is a commercially important bacterial pathogen causing blackleg and soft rot diseases of potato. E. c. subsp. carotovora (Ecc), however, causes soft rot of a wide range of plants, including potato. To identify genes involved in the plant-bacterium interaction we are, a) studying gene expression in vitro and in planta (cDNA-AFLP), b) producing physical maps of the Eca and Ecc genomes (BAC libraries), c) producing sequence information on targeted regions of the genome (including the hrp cluster) and d) undertaking functional analysis. Following cDNA-AFLP, differentially amplified cDNAs are being isolated, sequenced and their expression confirmed by Northern blots. This is the first report of cDNA-AFLPs applied to bacteria and opens many new areas of investigation. We are also creating physical genetic maps of Eca and Ecc from contiged Bacterial Artificial Chromosome (BAC) libraries. By positioning known pathogenicity genes or subspecies-specific genes (cDNA-AFLP) onto the Eca map, clones rich in pathogenicity genes are being identified. Random sequencing and a 1x sequence of targeted clones have identified a number of novel pathogenicity and host range related sequences. The maps are also being compared with the Escherichia coli genome to examine similarities in the genomic organisation of these closely related plant and human pathogens, and to identify further novel subspecies-specific sequences. Functional analysis on the Eca hrp cluster suggests that the Type III system is functional, although its role is yet to be confirmed.
Sequencing of the complete genome of Ralstonia solanacearum opens
new avenues towards the understanding of pathogenicity determinants
Ralstonia solanacearum is a Gram negative beta-proteobacterium. This plant pathogen has an unusually wide host range causing "bacterial wilt" disease in over 200 plant species belonging to more than 40 families. Strain GMI1000 causes disease on various solanaceous crops, including potato, tomato and eggplant, and on the model plant Arabidopsis thaliana; it also induces a hypersensitive response on tobacco. GMI1000 is classified as a wide host-range race 1 isolate.
Pulsed field gel electrophoresis revealed that the genome of GMI1000 is composed of two replicons of 3.6 and 2.2 megabase pairs. In order to learn the complete sequence of this genome, clones carrying 2 to 10Kb inserts in addition to a BAC library were sequenced. Data representing over 9 equivalents of the total genome have been generated and assembled into two circular scaffolds of 3.6 and 2.1 Mb. "Final sequencing and annotation of the sequence are in progress. Presently, 65% of the total genome has been analysed permitting the prediction of close to 3000, proteins, 630 of which have no homology with previously described proteins. Based on these preliminary data, there is no clear functional repartition of genes between the two replicons.
With an average G+C content of 67%, the observed composition of the genome fits with previously published data. Frequently, local variations ranging from 49% to over 72% G+C has been observed in regions spanning 1.5 kb to over 20Kb . These heterogeneous regions are homologous to genetically mobile elements, suggesting that numerous horizontal gene transfers might have occurred in this soil organism that is naturally competente for transformation by naked DNA. No clear pattern appears in the partitioning of these regions over the two replicons. Proteins encoded in these regions are often associated with adaptation of the bacterium to life in particular environmental conditions or ecological niches including pathogenicity. This will be illustrated by presenting our functional analysis of a locus encoding the PopP protein, a structural homologue of the YopP/YopJ pathogenicity effector protein from Yersinia.
As the study of plant-pathogen interactions in the form that we know it today is only about 100 years old, virtually all informative data on these interactions has been gathered within a single human lifetime. This fact allows a clear appreciation of the incredible advances in techniques, knowledge, and understanding that have occurred in this field as it has developed. Three of the most conceptually influential phenomena, the hypersensitive response, the gene-for-gene hypothesis, and induced resistance were first reported over 40 years ago. However, as the molecules, and ultimately the genes, associated with these phenomena have been discovered, the conceptual framework surrounding them has changed. This change is, perhaps, most dramatic for the gene-for-gene theory for which the ever-increasing information on the nature, function and evolution of resistance genes, plus the identification of other genes required for resistance, have provided a substantial depth of understanding. Whether our increased knowledge of the hypersensitive response has resulted in a similar increase in clarity of understanding is more debatable. Also notable is the fact that despite the research emphasis on resistance to disease in the 20th century, there are still few plant-pathogen systems in which the features that stop pathogen growth are unequivocally known, or in which the factors that condition susceptibility have been identified. Arguably the most conceptually confused phenomenon in the current literature, despite a steady increase in knowledge, is nonhost resistance. However, current and developing tools or techniques applicable at the cell or molecular levels hold enormous future promise for rapidly delivering data, and hopefully knowledge, on the still unknown features of plant-pathogen interactions.