RESISTANCE (R) GENES EXPRESSION AND TRANSCRIPTION PROFILING Introduction :
In their constant struggle for survival, plants have developed a wide range of defence mechanisms to protect themselves against the attack of pathogens. While some of these resistance strategies rely on simple physical or chemical barriers, more sophisticated biochemical mechanisms based on gene-for-gene interactions between plants and their infectious agents have been reported. Plant disease resistance genes (R-genes) play a key role in recognizing proteins expressed by specific avirulence (Avr) genes of pathogens .R-genes originate from a phylogenetically ancient form of immunity that is common to plants and animals. However, the rapid
evolution of plant immunity systems has led to enormous gene diversification .Although little is known about these agriculturally important genes, some fundamental genomic features have already been described. It has been recently shown that proteins encoded by resistance genes display modular domain structures and require several dynamic interactions between specific domains to perform their function. Some of these domains also seem necessary for
proper interaction with Avr proteins and in the formation of signalling complexes that activate an innate immune response which arrests the proliferation of the invading pathogen .
R-genes can be functionally grouped in five distinct classes based on the presence of specific domains ,
the CNL class comprises resistance genes encoding proteins with at least a coiled-coil domain, a nucleotide binding site and a leucine-rich repeat (CC-NB-LRR);
the TNL class includes those with a Toll-interleukin receptor-like domain, a nucleotide binding site and a leucine-rich repeat (TIR-NB-LRR);
the RLP class, acronym for receptor-like protein, groups those with a receptor serine – serine – threonine kinase-like domain, and an extracellular leucine-rich repeat (ser/thr-LRR);
the RLK class contains those with a kinase domain, and an extracellular leucine-rich repeat (Kin-LRR);
the
‘Others’ class
includes all other genes which have been described as conferring
resistance through different molecular mechanisms, e.g. mlo and asc-1 .
R gene specificity
The proteins encoded by R genes are classified according to certain
structural
characteristics and localization in the plant cell. Leucine-rich repeats (LRRs)
Leucine-rich repeats (LRRs) are multiple, serial amino acid repeats (normally around 24 amino acids long) that contain leucines or other hydrophobic residues at regular intervals, along with regularly spaced prolines and asparagines . LRR domains are often involved in protein-protein interactions in recognition domains. Depending on where in the plant cell the R protein LRR reside, they have either cytoplasmic LRRs or extracytoplasmic LRRs. The role of resistance gene LRRs in recognition is supported by analysis of the Cf-4 and Cf-9 genes from tomato, in which the majority of amino acid substitutions between the two genes occur in these regions. The R proteins that have a cytoplasmic LRR domain also have a nucleotide-binding site (NBS) and some of them have a zipper-like domain of leucine molecules known as coiled coil, or have a domain of Toll/interleukin 1 receptor (TIR). The TIR domain
The TIR domain is so called because it contains similarity to the cytoplasmic signalling domain of the Drosophila melanogaster Toll protein and the interleukin-1 receptors of birds and animals. These proteins are key components of the innate immune system - exist in all multicellular organisms and relies on a set of germ-line encoded receptors that are expressed in a wide variety of cells, particularly those at host/environment boundaries. The NBS (NB) domain
Nucleotide-binding sites occur in diverse proteins with ATP- or GTP-binding activity, such as adenylate kinases, ATP synthase β-subunits, and the apoptosis proteins, APAF-1 (apoptotic protease activating factor) in humans, and CED-4 ( Caenorhabditis elegans death regulating protein 4) in the nematode C. elegans. NBS domains consist of a nucleotidebinding P-loop (kinase la) and two distally located domains, kinase 2 and kinase 3. The homology between NBS domains in R genes and those in apoptosis proteins has led to the proposed roles for these domains in plant defence, particularly as this sequence homology extends beyond the NBS domains.
Coiled coil (leucine zipper) domains
Coiled coil motifs have the amino acid pattern (hxxhxxx)n where h is a hydrophobic residue (A, F, I, L, M or V) and x is normally a polar or charged residue. These motifs are present in a number of R genes (e.g. RPM1, RPP8, Rx1, Mi and Prf ) and the patterns are consistent with them forming an amphipathic alpha-helical domain into a coiled coil structure. Coiled coils are often involved in protein homo/hetero dimerisation (particularly of eukaryotic transcription factors) -their presence suggests that dimerisation may be important for many R gene-encoded proteins. The organization and structure of resistance genes and their products.
A different kind of R gene named RPW8 has been found in Arabidopsis. RPW8 is different in that it confers resistance to a broad range of powdery mildew pathogens instead of a specific pathogen race . The R proteins that have an extracytoplasmic LRR domain contain a transmembrane region, Cytoplasmic domain that acts as a protein kinase. R proteins predicts a role -signal transduction. Examples of R Genes
In 1992, the first R gene, the maize Hm1 gene-makes resistant to race 1 of the fungus Cochliobolus carbonum, which causes a leaf spot disease on susceptible corn varieties.
Race 1 of C.carbonum, produces a hostspecific toxin, the HC toxin - infect the corn varieties that lack the Hm1 gene and are susceptible to the fungus. Resistant to race 1, expression of the Hm1 gene results in the production of an enzyme called HC toxin reductase. Enzyme
reduces and thereby detoxifies the HC toxin and in that way keeps the plants free from infection by the fungus. Pto gene of tomato, - resistance to the bacterial speck-causing strains of P. syringae pv. tomato that carry the avirulence gene avrPto. The protein encoded by the Pto R gene appears
to be a serine – threonine protein kinase, an enzyme suspected to play a role in signal transduction leading to the hypersensitive response. Other R genes isolated from plants include the tomato Cf2, Cf4, Cf5, and Cf9 genes, which confer resistance to the leaf mold-causing fungus Cladosporium fulvum races 2, 4, 5, and 9 that carry the avirulence genes avr2, avr4, avr5, and avr9, respectively. Tobacco N 1 gene, which confers resistance to TMV; The flax L6 gene, which confers resistance to the rust fungus Melampsora lini race 6 carrying the avr6 gene; The rice Xa21 gene, which confers resistance to many races of the leaf-spotting bacterium Xanthomonas oryzae;
How Do R Genes Confer Resistance?
Elicitor molecule produced by an avr gene of the pathogen is recognized by a specific plant receptor encoded by an R gene. Following recognition of the elicitor by the receptor molecule, one or more kinase enzymes may become activated and amplify the signal by phosphorylating, and thereby energizing, other kinases and other enzymes. This leads to a cascade of biochemical reactions that, hypersensitive response and, thereby, localized host resistance at the point of attack by the pathogen.
Evolution of R Genes
During the evolutionary race for survival of the plant from the pathogen, a resistance (R1) gene evolved, e.g.,by modification of one of the general resistance genes, and that gene allowed the plant to recognize one of the initial steps of infection by the new pathogen (race 1) To resist infection, such an individual plant and its progeny (variety 1) were selected for survival and so the plant and the R1 gene survived and multiplied. Modified receptor 1 product of the R1 gene recognizes specifically a particular compound (elicitor 1) produced by a pathogen gene, - avirulence ( avr1) gene. Pathogens carrying this avr1 gene (race 1) cannot survive on such R1 gene-carrying plants. If, a mutation affects the avr1 gene of race 1 of the pathogen, the gene and the avirulence are destroyed. New offspring of the pathogen become virulent again, capable of attacking the sofar resistant variety 1 of the plant. This new virulent pathogen population could be called race 2. The host plant (variety 1) is now susceptible to race 2, which infects and may kill many plants. Other Plant Genes for Resistance to Disease
Nonhost plant carries numerous R gene-coded receptors - responsible for recognition production of signal molecules, some of which trigger a cascade of localized reactions, leading to the hypersensitive response and, through the plant-induced death of the affected cell, to localized resistance. During development of the hypersensitive response, signal molecules that transmit the alarm to other cells and to most distal parts of the plant. There, they trigger the activation of additional defense response genes called systemic acquired resistance genes. These genes mobilize the host defenses throughout the plant and are effective against new infections by the same pathogen and also against infections by unrelated pathogens.Numerous “minor” genes for resistance- may affect superficial or internal, structural or biochemical defenses, preexisting or induced on or after infection. Such minor genes are probably quite numerous in all plants. They are triggered into action by signal compounds produced by the pathogen or by the infected cells.
TRANSCRIPTION PROFILING Research hypotheses can now be addressed in the context of entire developmental or biochemical pathways under a large array of experimental or field conditions. Access to complete genomic sequences, coupled with rapidly accumulating data related to RNA and protein expression patterns, have made it possible to determine comprehensively how genes contribute to complex phenotypes. The genetic inheritance of transcript profiles can be used to map the chromosomal location of relevant, regulatory loci. In addition, comparison of gene expression networks among model and crop species can be used to infer function as well as evolutionary history, based on expression patterns shared with genes of known function. Interesting topics for plant pathologists include the mechanisms underlying gene-for gene resistance and basal defense, host vs non host resistance, biotrophy vs necrotrophy, or pathogenicity of vascular vs nonvascular pathogens, among many others. Examples of how transcript profiling has been utilized effectively to drive biological discovery in host-pathogen interactions. With the increasing availability of a large number of microarray data sets in public repositories, it is also beneficial to use comparative meta-profiling strategies to draw conclusions that bridge multiple experiments for profiling. Expression profiling once thought to be available only to large research groups with substantial infrastructure, can now be used by small groups and individual laboratories. A myriad of microarray platforms now exist for most organisms through commercial and public sources. Off-the-shelf, high-density DNA arrays are now available for barley, wheat, rice, maize, sugarcane, grape, citrus, poplar, tomato, Arabidopsis, Brassica, and cotton, among others. In addition to using these arrays to monitor the expression of thousands of transcripts in parallel, many projects are accessing this technology for both host and pathogen through communitydriven development of multispecies arrays, such as soybean/ Phytophthora sojae (root rot)/ Heterodera
glycines (soybean
cyst
nematode), Medicago truncatula/Medicago
sativa/
Sinorhizobium meliloti, Fusarium graminearum (scab) , and rice /Magnapothe grisea (rice blast).
Also, many companies now collaborate with investigators to produce custom arrays (or the oligonucleotides for spotting) at a reasonable cost. This enabling technology provides the opportunity to investigate the regulation of entire pathways in both hosts and pathogens under uniform experimental conditions.
Profiling plant R gene-mediated responses to pathogenic bacteria.
A number of studies have yielded global views of the transcriptional response of plants undergoing R gene – mediated defense against pathogenic bacteria. Mysore et al. utilized Gene Calling technology in tomato to define the gene expression changes in resistance to bacterial speck, and the specific contributions of the R gene Pto and the gene Prf , which is required for Pto function. They defined an early role for Prf in the response pathway, and also identified
changes dependent on Prf but not Pto, suggesting a distinct, independent role for Prf in pathogen recognition. In a subsequent study that used a variety of transcript profiling techniques and resources (including subtractive-suppressive hybridization and cDNA microarrays), it was shown that over expression of Pto induced gene expression changes similar to those observed during immune responses in animals. To gain insight into the molecular basis of tomato resistance to Xanthomonas axonopodis pv. vesicatoria strains expressing the effector Avr Rxv, which is
governed by three, nondominant resistance genes, Bonshtein et al. used a similar profiling approach . Others have obtained global transcript profiles of plants undergoing R gene – mediated responses toward identifying genes that might enhance resistance when expressed highly in genetically engineered or selectively bred plants. In a study of interactions of Arabidopsis with P. syringae, Tao and colleagues examined global gene expression patterns in responses of susceptible plants to a virulent strain (a compatible interaction), responses of plants carrying either of two R genes ( Rpm1 or Rps2) to corresponding avirulent strains (incompatible interactions), and plants exhibiting nonhost resistance to a strain that normally infects bean (incompatible interaction). The analysis showed overall strong similarity among the responses mediated by the two different R genes and in nonhost resistance, despite genetically well-defined differences in respective signaling pathways. The study also revealed that the differences among responses in incompatible and compatible interactions were largely quantitative, analagous to observations of Caldo and associates with barley- Blumeria interactions and Eulgem and colleagues on Arabidopsis-Peronospora interactions. Profiling fungal gene expression within the host
In addition to differential gene expression occurring within host cells, it is necessary to consider the gene expression of pathogens within the infection site as well. A study to monitor expression of genes in Colletotrichum graminicola, causal agent of maize anthracnose stalk rot,
used laser capture microscopy (LCM) in combination with fluorescent AmCyan protein-tagging, to produce samples for microarray analysis. This approach identified over 8000 genes as significantly expressed in C. graminicola as early as 2 days after inoculation, which is an early stage of infection with relatively little fungal biomass accumulation. Studies with another comparable host-pathogen interaction at 2 days after inoculation, but not employing LCM, led to the identification of only about 900 expressed fungal genes. Thus, LCM clearly provided for greater enrichment of fungal mRNA and an order of magnitude increase in power to detect fungal mRNA transcripts. Comparison of gene expression between in vitro grown cultures and in planta grown cells showed significant up-regulation of secreted proteins, perhaps signifying the production of effectors and other proteins required for pathogenicity. It would be interesting to determine how the maize cells collected in these samples were also responding to C. graminicola in these samples. Both and colleagues found many pattern of coordinate expression among B. graminis hordei genes in defined metabolic pathways from cDNA microarray profiling
experiments monitoring the infection cycle in barley. This allowed an assessment of the metabolic status of the fungus during asexual development as it infected the host plant. Genes encoding several glycolytic enzymes are significantly up-regulated as mature appressoria form, and in the infected epidermis, which contain fungal haustoria. Concomittantly, host plants show up-regulation of sugar transport and utilization-related genes after powdery mildew infection, providing a source for nutrient acquisition. Obligate biotrophic rust fungi infect host tissue via intercellular mycelia that form haustoria within the living plant cells. Jakupovic and colleagues identified genes expressed during biotrophic growth of the bean rust Uromyces fabae by EST sequencing of a haustorium-specifc cDNA library. Several of the Uromyces ESTs were identical to the in planta induced genes ( PIGs) identified in earlier studies. Virus-encoded sequences were identified, providing evidence for two RNA mycoviruses in U. fabae. Subsequent microarray experiments revealed many cDNAs that were significantly expressed in rust infected leaves as compared to germinated urediniospores, suggesting a shift in rust gene expression between germination and the biotrophic stage of development. To monitor the ascomycete B. cinerea, a broad-spectrum plant pathogen, in real-time infection conditions, Gioti and associates infected Arabidopsis leaves with B. cinerea and assayed transcript accumulation using a custom
macroarray. Seven percent of B. cinerea genes were differentially expressed during infection, and 27 genes were significantly up-regulated in planta. Two of the genes, trichodiene oxygenase
and pentalenene synthase, had already been associated with fungal pathogenicity, whereas eight have unidentified functions. The 27 genes were clustered into three groups; the first group showed maximal expression at the early stage following fungal penetration, the second showed maximal expression at the outset of the colonization of plant leaves, and the third showed maximal expression when the colonization of plant leaves was completed. A gene homologous to FKBP12 proteins was identified from cluster three and was confirmed to be a pathogenicity determinant via gene disruption. The genomes of several filamentous fungi have recently been completely sequenced, making possible genome-wide expression analysis. Guldener and colleagues took advantage of the genome sequence of F. graminearum, the causal organism of Fusarium head blight of wheat and barley, to design a wholegenome (18 K) Affymetrix Gene Chip. To establish a baseline set of gene expression data, F. graminearum Gene Chips were interrogated with RNA isolated from fungus grown in culture under three nutritional regimes, in addition to in planta growth in infected barley (12). During the barley infection time course 7132 Fusarium probe sets were called present, even though the fraction of fungal transcripts in the total RNA from infected plants is quite low, notably during the early stages of infection. APPLICATIONS OF TRANSCRIPTIONAL PROFILING: AN EXPANDING RANGE OF POSSIBILITIES Dissection of Changes in Gene Expression Levels
One of the temptations of whole-genome expression platforms is to simply generate data for discovery purposes. While this may be a valid approach for open technologies in which the data can be used for genome annotation, it is harder to justify for microarrays and other closed technology platforms. Despite the ease of producing reams of data, it will be meaningless unless experiments are properly designed with the appropriate biological materials and replicates. The extraction of meaningful data requires analytical strategies and the interpretation depends on close interactions among biologists, computer scientists, and statisticians. The detection of differential expression among two types of tissues differing by some experimental variable is one of the most basic questions addressed with transcriptional analysis. Typically, a user-defined cut-off or threshold for the ratio of expression levels in the two tissues is used to identify differentially expressed genes. The underlying assumption is that genes with
differential expression are somehow involved in the condition that distinguished the tissues. The statistical methods for identifying such genes have been much better developed in recent years and are able now to identify up- or down-regulated genes with statistical significance. The end product is a list of candidate genes believed to be involved in the phenotype of interest; these genes must then be validated using much more time-consuming functional studies. The integration of pathway information could lead to the association of pathways with a process when genes in that pathway are over represented in the differentially expressed genes. Although for most organisms few data are available describing the pathways and related genes, such data may be generated empirically by the application of pattern discovery methods. These methods include the numerous clustering techniques designed to construct groups of genes with related patterns within the data set. This simplifies and structures the data based on inherent patterns rather than imposing assumptions made a priority. Ultimately, it may be possible to reconstruct or model complex signaling pathways by combining interferences made from transcriptional profiling data with biochemical and metabolic data. Categorization of Tissues Based on Expression Patterns
Expression
profiling
provides
a
comprehensive
approach
for
the
molecular
characterization of tissues, treatments, or cell types. The state of the transcriptome represents a phenotype that provides a clear physiological picture of cellular activity. Class prediction methods are statistical techniques that can be used to classify expression profiles from different samples into known groups. The use of microarray phenotypes for tissue classification is most widely and successfully used in cancer research; the molecular data can distinguish tumors more reliably than other approaches, resulting in more accurate disease diagnoses. Comparisons among different samples of the same cancer type reveal distinct subgroups, provide a molecular classification of the cancer type, and can determine the stages of progression of the disease. Hierarchical clustering analysis of the array data is used to sort specimens. These studies have defined candidate marker genes that can discriminate between normal and diseased tissues. The combined sets of diagnostic marker genes may be used to develop specialized or customized arrays that contain only the diagnostic genes of specific interest. However, while the idea of customized arrays was pertinent when array densities were low and most arrays were homemade, this strategy may be less important as costs decrease for high-density commercial arrays for which uninformative genes can be ignored.
In plants, this type of classification based on transcriptional profiles could be applied to the sorting of mutants based on perturbations in distinct signaling pathways. This strategy does not require optimal microarray probe design or even that the probes identify known genes. The microarray elements must serve as molecular markers, providing detectable signals and behaving independently. Moreover, complete coverage of all genes by the technology is not critical, as long as the genes that are represented provide enough resolution for diagnosis or identification. Every informative array element or probe will provide an additional dimension for the analysis and for maximum resolution and significance; these probes should outnumber the distinct pathways or mutants under analysis. Application of Technologies to Diverse Genotypes
Natural variation in gene expression levels between closely related plant varieties can be treated as a genetic polymorphism. Microarrays or other methods can be used to describe patterns of gene expression among individuals in a mapping population. Each pattern constitutes a molecular phenotype. Transcript abundance levels differing in the parents of a mapping population and segregating among the progeny can be mapped and characterized as quantitative traits. These expression profiles may be more easily interpreted or quantified than some visible phenotypes. Differences in expression of a given gene may result either from allelic differences in its promoter or from effects of distal regulatory loci. In both cases, the variation is due to genetic differences that can be subjected to genetic analysis. In parallel, the individuals in the population can be genotyped using standard molecular techniques. With molecular phenotypic and genotypic data, expression level differences can be mapped using approaches based on quantitative traits, and with these data, quantitative phenotypic measurements may be associated with genetic markers. Accessions of Arabidopsis are rich in genetic variation for many traits, and the analysis of this natural variation using quantitative methods may provide more insight into plant signaling and gene function than classical mutagenesis studies. This is because of the complexity of variation found between ecotypes and because variation in the genetic background may increase the penetrance of certain weak alleles or promote novel phenotypes resulting from gene interactions. Another important point is that alterations in the transcriptional activity of a gene may have more significant effects than polymorphisms that alter the protein sequence. Substantial variation in gene expression has been demonstrated between primate species and
among fish populations, suggesting that natural selection may act as, or more, effectively on transcriptional than translational differences. In plants, most such studies will first be carried out in Arabidopsis due to the experimental advantages of this model plant; there is little doubt that gene expression analysis ultimately will be used to characterize and to map complex phenotypes in many plant species. Which technology platforms will be used for studies of natural variation in gene expression? All of the platforms described above will measure variation in expression, but some will also be sensitive to genotypic differences that could interfere with measurements of expression. For example, the oligos used in some microarray platforms are short enough to be sensitive to sequence polymorphisms within the homologous region of the transcript. The short probes (25-base oligos) used on Affymetrix arrays will be most sensitive to single nucleotide polymorphisms (SNPs); one base difference in the length of the oligo is enough to substantially diminish hybridization. Because Affymetrix uses 10 or more probes for each gene, differences in hybridization intensity among the probes may be attributed to genomic polymorphisms. In fact, some research groups have exploited this property using labeled genomic DNA to identify SNPs or insertion/deletion events. An early and elegant study demonstrated polymorphic hybridization to Affymetrix microarrays due to strain-specific differences in yeast ( Saccharomyces cerevisiae) used Affymetrix arrays to assess the polymorphisms in the Landsberg ecotype of
Arabidopsis by hybridization of genomic DNA to the array designed from the Columbia genome. In contrast to the 25-mer oligos, long oligos (70-mers) are more tolerant to polymorphisms, presumably because the additional nucleotides provide greater stability. This has been demonstrated in experiments using RNA from Arabidopsis thaliana, Arabidopsis arenosa, and Brassica oleracea. Whole-genome long-oligo arrays could be used to analyze gene expression in a wide variety of related species with smaller genotypic effects on hybridization. This reduced sensitivity to SNPs means that long-oligo microarrays will not be useful for distinguishing expression levels of alleles or closely related gene families. Measurement of Allele-Specific Differences
Beyond simply measuring expression level differences among homozygous inbred lines, an additional challenge for gene expression technologies will be to characterize and quantify subtle allele-specific differences in expression at heterozygous loci. Hybrid vigor is a wellcharacterized but poorly understood trait that is important to modern agriculture; one possible explanation for hybrid vigor is transgressive variation in expression. Expression differences for a
particular allele in a hybrid compared with the parental lines result either from imprinting (; a cis effect) or trans-acting regulatory elements encoded in the two genomes. Imprinting is generally associated with monoallelic expression, so biallelic nonparental expression is indicative of transacting regulation of expression. To put it differently, the promoter and other adjacent regulatory elements for a given allele are identical in the F 1 hybrid and parental lines, so any differences in expression for a specific allele between an inbred parent and the F 1 hybrid must result from the interchromosomal effects in the hybrid. Similar intergenome effects may alter gene expression patterns in polyploids. Draft sequences of rice indica and japonica varieties have been published, and these data create a unique opportunity for large-scale measurements of differential expression in closely related varieties and hybrids, because the sequence of alleles from each variety will be known and may be used for measurements of allele-specific expression levels. Sequence based-measurements of gene expression such as LongSAGE or MPSS are sensitive to single nucleotide polymorphisms and therefore could be used to globally quantify allele-specific expression. However, the sequence of both alleles must be known to ensure a specific match for the tag. For microarrays, a priori knowledge of SNP locations enables the use of short oligonucleotides, such as those present on the Affymetrix arrays, to measure differential expression between alleles. This type of analysis was performed using human genes and demonstrated that a significant proportion of the alleles that were examined were differentially expressed. Differential display-type methods, which distinguish genes based on restriction site polymorphisms, can be used to screen for allele-specific expression differential-display approaches are advantageous when the sequence of one or both alleles is unknown. This approach has been used to identify allele-specific differences in expression for small numbers of maize genes. Whole-genome analyses of allele-specific expression in plants will require gene sequences from multiple varieties and may require specialized microarrays that detect SNPs to distinguish alleles.
