COEVOLUTION OF VIRUS RESISTANCE AND VIRAL AVIRULENCE GENES

Avirulence genes in plant pathogens have been defined by their requirement for disease resistance in hosts containing corresponding R genes. Plant viruses evolve very rapidly owing to very short replication cycles, large numbers of genomes within each cell across many cells per host, and many hosts. For RNA viruses, the absence of a proofreading function in viral replicases may result in mutation rates as high as 104 per replication cycle per base. Viral genetic variation can result from several major genetic processes including mutation, recombination, and the acquisition of additional genomic sequence. As a consequence, resistance breaking viral genotypes are known for most host resistance, especially for genes showing HR. Avirulence determinants are typically identified by creating chimeric clones derived from viral genotypes with contrasting virulence and then testing for infectivity. Once an avirulence domain is identified, site-directed mutagenesis allows identification of specific point mutations responsible for virulence.

 Virtually any part of the viral genome can define an avirulence determinant. With respect to R genes that confer HR, avirulence factors include viral RNA polymerase subunits, movement protein, and CP. Several potyviral avirulence genes have been identified for dominant R genes that do not show HR. The CI and P3 proteins of Turnip mosaic virus serve as avirulence determinants for the Brassica napus R genes, TuRBO1 and TuRBO4/5 (96–98), while SMV HC-Pro and P3 are involved in overcoming Rsv1 in soybean. In contrast to the case for dominant genes where many different viral components have been identified as avirulence determinants, a pronounced trend is apparent viral factors that serve as the determinant for pathogenicity in resistance systems controlled by recessively inherited R genes. Of nine R gene studies to date, seven identify potyviral VPg as the pathogenicity determinant for recessive resistance, although the systems in question show diverse resistance phenotypes: Capsicum pvr1/pvr12 resistance to PVY is cellular, tobacco va resistance impairs the cell-to-cell movement, and Nicandra physaloides and Solanum commersonii affect long-distance movement. The eighth study, also focused on a potyviral system, PsbMV/pea, identified the P3-6K1 cistron as the pathogenicity determinant. Only one study to date has focused outside the Potyviridae. In this case, the 3_ untranslated region of the carmovirus Melon necrotic spotvirus (MNSV) genomic RNA defined the location of the viral determinant in the by interaction of MNSV with melon. In the eight cases where the viral elicitor is protein, host recognition of these viral proteins that serve as pathogenicity determinants is altered by amino acid substitutions that do not appear to significantly compromise the function of the protein in pathogenesis. For other microbial pathogens, there often appears to be a fitness penalty association with mutations from avirulence to virulence. Although this type of fitness/avirulence tradeoff has not been noted generally for plant viruses, there are specific examples where this occurs. Isolates of ToMV capable of overcoming Tm22 gene were found to multiply poorly on resistant plants. If the Tm22 resistant protein targets a domain of the viral MP such that this protein is mutated to overcome resistance, these mutations could result in diminished fitness.

RECESSIVE PLANT VIRUS RESISTANCE GENES CHARACTERIZED AT THE MOLECULAR LEVEL

Despite notable progress towards defining the elements that comprise dominant R gene-mediated defense responses in plants, little is known about the nature of plant susceptibility to disease. Owing to the relatively small number of proteins they encode, viruses completely depend on the host factors to complete their life cycle. Typical plant viruses encode 4 to 10 proteins that coordinate the complex biochemistry and intermolecular interactions required for viral infection cycles. Studying recessive virus resistance provides a unique opportunity to reveal host factors required for susceptibility and mechanisms of pathogenesis of the pathogen. Recent findings have confirmed early theoretical predictions that mutations of some host factors will result in recessively inherited resistance to plant viruses. The identification and characterization of host factors in which mutations interrupt viral pathogenesis will provide a new opportunity for understanding viral pathogenesis itself, as well as host responses; this is an approach that has been unavailable to date in the study of dominant resistance. Whether as a consequence of the economic importance of the Potyviridae, the relative prevalence of recessive resistance to this group of viruses, and/or the relative ease with which these viruses can be experimentally manipulated, studies of recessive R genes to date have focused largely on this viral family. Several host genes whose mutations impair the infection cycle of plant viruses, including BCTV,CMV, TEV, TuMV, TMV, TGMV, and TCV, have been identified and characterized in Arabidopsis. The translation initiation factor eIF4Ehas been identified repeatedly as a naturally occurring recessively inherited resistance locus in pepper pvr1, lettuce mo1, and pea sbm1 and has been implicated in barley as a candidate for rym4/5. The eIF4E isoform eIF(iso)4E also has been implicated in Arabidopsis and pepper resistance. The role of eIF4E and eIF(iso)4E in the potyvirus infection cycle is not known. However, the negative effects of mutations in these host factors on the infectivity of various potyviruses in various host plants imply that the effect of these host factors upon potyvirus infection cycle is probably conserved. The common feature linking pvr1/2, sbm1, and mo1 is that the viral avirulence determinants map to a specific region in the VPg, the protein covalently linked to the 5׳end of the viral RNA and perhaps mimicking the m7G cap of eukaryotic mRNAs. In eukaryotic cells, eIF4E binds to the m7G cap as the first step in recruiting mRNA into the translational preinitiation complex. A similar role for eIF4E might be predicted when potyvirus infects plant cells. Although eIF4E has never been shown to bind VPg in infection, VPg or its precursor VPg-Pro interacted with eIF4E or eIF(iso)4E in yeast two-hybrid and in vitro pull-down assays.

In the 1950s, pvr1 and pvr2 in pepper (Capsicum annuum and C. chinense) were initially considered allelic but then two loci were distinguished because of differences in resistance spectra. The allele formerly known as pvr21 is effective only against PVY-0, and pvr22 is effective against both PVY-0 and PVY-1. The allele pvr1 was relatively broad in effect, controlling TEV, PepMoV, and PVY). We now know that only one locus is involved, pvr1, at which at least three resistance alleles and two susceptibility alleles occur.

Point mutations in eIF4E that fall near critical positions for cap-binding function abolish interaction with TEV VPg and determine the range of isolates across three potyviral species that are controlled. Two of these alleles when homozygous block accumulation of the virus in protoplasts. The narrower spectrum allele retards movement of the virus through the plant but has no effect at the protoplast level. In a fourth case, Pepper veinal mottle virus, it appears eIF4E and probably eIF(iso)4E must be mutated to control the virus. In lettuce, mo11 and mo12 control common isolates of LMV. In the homozygous state, mo11 confers resistance, i.e., absence of LMV accumulation; mo12 results in reduced LMV accumulation and lack of symptoms. As observed in pepper, allelic variants of eIF4E, Ls-eIF4E0, Ls-eIF4E1, and Ls-eIF4E2 contained point mutations that result in predicted amino acid substitutions near the cap-binding pocket of the protein. In pea, sbm1 confers resistance to PSbMV pathotypes P1 and P4 as described above, now known to be a consequence of mutations in an eIF4E homolog.

Transient expression of susceptible-eIF4E in a resistant background complemented PSbMV infection by supporting both virus multiplication in primary target cells and cell-to-cell movement. Processes that account for cell-to-cell movement are not well understood, and therefore it is difficult to speculate on a plausible role for eIF4E in virus movement. Nevertheless, in both the pepper and pea systems, variants at an eIF4E locus result in inhibition of movement, as well as extreme resistance. Again, point mutations in the resistant eIF4E allele are located in and around the cap-binding pocket. Similar to pvr1, cap-binding ability of the eIF4E protein is abolished in the resistant eIF4E variant. Recently, rym4/5 for resistance to BaYMV in barley has also been shown to encode eIF4E. In contrast to dominant R genes where resistance to the same or closely related pathogens generally do not occur in syntenic positions, a recessive potyvirus R gene pot1from tomato was mapped to a collinear position with pepper gene pvr2. These results indicate that recessive R genes are highly conserved. Evidence to date indicates that, for the most part, dominant and recessive R genes may not be related mechanistically and evolutionarily.

Resistance to Cucumber mosaic virus in Arabidopsis Conferred by RCY1

Extensive examination of 12 Arabidopsis ecotypes identified a CMV-Y-resistant ecotype, C24. The resistance response of C24 includes suppression of virus multiplication to a low level, the formation of necrotic lesions at the primary site of virus infection, and restriction of virus to the inoculated leaves. This resistance response in C24 is controlled by a single dominant RCY1 (resistance to cucumber mosaic virus strain Y) gene. The analysis of a series of chimeric viruses constructed from the avirulent isolate CMV-Y and the virulent isolate CMV-B2 revealed that the coat protein of CMV-Y serves as the avirulent determinant of resistance in C24. The RCY1 gene has been mapped in Arabidopsis within the MRC-5 region on chromosome 5, in which nine other defined resistance genes (RAC3, RPS4, HRT, TTR1, and five distinct RPP loci) are located. Fine mapping and sequence comparison of this region from C24 and a CMV-Y susceptible C24 mutant identified the RCY1 gene encoding 104-kDa CC-NBS-LRR type protein.

RCY1 is allelic to the resistance gene RPP8 against Peronospora parasitica in the ecotype Lansberg erecta and HRT against TCV in the ecotype Dijon-17. The RCY1-conferred resistance requires both salicylic acid and ethylene signaling but not jasmonic acid signaling. Table 4 shows some of the R gene and Avr factor pairs.

Resistance to Turnip crinkle virus in Arabidopsis Conferred by HRT

A single dominant gene, HRT, was identified for HR resistance to TCV. HRT is located on chromosome 5 and encodes a CC-NBS-LRR protein with striking similarity to the RPP8 gene family for resistance to the oomycete Peronospora parasitica. Despite very high sequence similarity, HRT and RPP8 specifically control only their cognate pathogens. Analysis of resistance in HRT-expressing transgenic plants indicated that HRT is necessary but generally insufficient for resistance. About 90% of the HRT-transformed Col-0 plants developed HR and

Expressed PR-1 after TCV infection yet remained susceptible to TCV. Full resistance to TCV required both HRT and a recessive allele rrt. Later experiments demonstrated that the HRT-/rrt-mediated response is dependent on EDS1 and independent of RAR1 and SGT1. In this system, TCV CP is the avirulence determinant recognized by HRT. A host protein, TIP (TCV interacting protein) that belongs to the NAC family of transcriptional activators is known to interact with TCV CP. Although the relevance of this interaction to the mechanism of resistance remains unclear, this interaction apparently functions to keep TIP out of the nucleus.

Resistance to Tomato mosaic virus in Tomato Conferred by Tm22

Tm22, the second tobamovirus R gene isolated, is one of the three R genes, Tm1, Tm2, and Tm22, used widely in tomato breeding to control Tomato mosaic virus (ToMV). The Tm1 gene from S. hirsutum confers extreme resistance and was mapped to chromosome 2. Tm2 and Tm22, considered to be alleles from S. peruvianum, are located close to the centromere of chromosome 7. Tm22, considered the more durable of the two alleles, was isolated by transposon tagging and encodes an 861 amino acid CC-NBS-LRR protein. The predicted protein from the susceptible allele tm2 also encodes a CC-NBS-LRR protein that appears comparable in most respects to the protein encoded by the resistance allele. Analysis of the nucleotide sequence of resistance-breaking virus isolates indicated that the MP protein is the avirulence factor in this resistance system. However, different mutations are required to overcome Tm2 and Tm22.

Resistance to Tomato spotted wilt virus in Tomato Conferred by Sw-5

Economic considerations have promoted the goal of TSWV-resistant tomato varieties in plant breeding programs for nearly 70 years. Early genetic studies reported five genes, Sw-1a, Sw-1b, sw-2, sw-3, and sw-5, from two species, Solanum pimpinellifolium and Solanum lycopersicum, all of which were overcome quickly. Sw-5, introgressed from Solanum peruvianum into tomato, has demonstrated broad and stable resistance. In resistant genotypes, local necrotic lesions develop on inoculated tissue, and systemic movement of the virus is restricted. The Sw-5 locus was isolated by positional cloning and sequenced, revealing that the resistance allele encodes a CC-NBS-LRR R protein. Sw-5 is remarkably similar to the tomato Mi gene for nematode resistance with the exception of four heptad amphipathic leucine zippers at the N terminus. This pronounced similarity suggests that Sw-5 and Mi may share a common signal transduction pathway. Sw-5 and its paralogs were mapped to tomato chromosome 9 and chromosome 12 with other fungal, viral, and bacterial R genes. A comparative analysis with the genus Capsicum, which is considerably diverged from Solanum within the tribe Solanae, indicated that paralog position was largely conserved between these genera. In Capsicum, monogenic dominant TSWV resistance conferred by Tsw showed identical resistance phenotype and strain-specificity to Sw5, but no cross-hybridization with Sw5 was detected. When resistance-breaking TSWV strains were analyzed, avirulence determinants mapped to different subgenomic RNAs.

Resistance to Potato Virus X in Potato Conferred by Rx1 and Rx2

The Rx loci in potato, Rx1 on chromosome V and Rx2 on chromosome XII, confer resistance to PVX in the absence of necrotic cell death. Rx-mediated resistance results in a very rapid arrest of PVX accumulation in the initially infected cell. In contrast to HR-associated resistance, Rx-mediated resistance is active in protoplasts. When protoplasts isolated from resistant (Rx) and susceptible (rx) potato genotypes were inoculated with PVX and TMV, Rx protoplasts showed<100-fold less PVX RNA accumulation, relative to a positive control using TMV. When TMV was coinoculated with PVX, TMV RNA accumulation was also reduced to a level comparable to PVX in resistant protoplasts, demonstrating that once induced, the resistant response can target viruses other than the elicitor virus. Rx1, isolated from tetraploid potato by map-based cloning, encodes a 107.5-kD CC-NBS-LRR protein. Rx1 and Rx2 show the same specificity for the PVX CP, extremely similar nucleotide sequence, and similar linkage with resistance to Globodera. Transgenic experiments demonstrated that the response to PVX in Rx-containing genotypes can be altered depending on the mode of expression of the viral CP.

Transgenic potato or tobacco plants expressing Rx show extreme resistance against PVX. When the PVX CP is constitutively expressed in the same plants, HR is observed, indicating that the amount of CP in the plant cell determines the macroscopic host response. Constitutive gain-of-function Rx mutants in which cell death is activated in the absence of viral CP were obtained by random mutagenesis. Sequence analysis revealed that most of the constitutive gain-of-function mutations occurred in or near the conserved NBS-LRR sequence motifs. It is not clear whether this phenotype is resulted from release of negative regulation by the LRR and adjacent sequences or introduction of an incompatibility between the domains such that they are no longer held inactive. In experiments designed to determine the biochemistry of Rx function, segments of the protein were expressed independently in an elegant system where phenotypic response could be easily assayed. PVX CP-dependent HR was observed after fragments of Rx (CC and NBS-LRR domains) and PVXCP were expressed transiently in N. benthamiana via agroinfiltration.

These results indicate that a functional Rx protein can be reconstituted through physical interactions between domains, even when the domains are expressed in different molecules. Furthermore, PVX CP disrupted the interaction between these Rx derived domains. The current model suggests that CP recognition induces sequential conformational changes in Rx, disrupting intramolecular interactions, thereby activating Rx-mediated signaling. Experiments using virus-induced gene silencing (VIGS) showed that Rx-mediated resistance does not require EDS1 and RAR1. Bieri et al., however, have shown that silencing Rar1 actually reduces the levels of Rx. Therefore, Rar1 is likely a cochaperone required to varying degrees by different R proteins. Silencing of tobacco MAP kinase kinase kinase (MAPKKK) interferes with the function of the Rx gene. Similar to results described above for the N gene, silencing SGT1 also compromised Rx-mediated resistance, and HSP90 is required, presumably acting as a cochaperone to stabilize Rx.