Send Orders of Reprints at Reprints@benthamscience.net Understanding Molecular Mechanisms of Durable and Non-durable Resis- Tance to Stripe Rust in Wheat Using a Transcriptomics Approach

Stripe rust of wheat, caused by Puccinia striiformis f. sp. tritici, continues to cause severe damage worldwide. Durable resistance is necessary for sustainable control of the disease. High-temperature adult-plant (HTAP) resistance, which expresses when the weather becomes warm and plants grow older, has been demonstrated to be durable. We conducted numerous studies to understand the molecular mechanisms of different types of stripe rust resistance using a tran-scriptomics approach. Through comparing gene expression patterns with race-specific, all-stage resistance controlled by various genes, we found that a greater diversity of genes is involved in HTAP resistance than in all-stage resistance. The genes involved in HTAP resistance are induced more slowly and their expression induction is less dramatic than genes involved in all-stage resistance. The high diversity of genes and less dramatic induction may explain durability and the incomplete expression level of HTAP resistance. Identification of transcripts may be helpful in identifying resistance controlled by different genes and in selecting better combinations of genes to combine for achieving adequate and durable resistance .


INTRODUCTION
Stripe rust (or yellow rust), caused by Puccinia striiformis Westend. f. sp. tritici Erikss. (Pst), is one of the most important diseases of wheat worldwide [1,2]. The disease continues to cause severe damage in many wheat-producing regions. Control is preferably through resistant cultivars. However, cultivars that are developed for resistance to stripe rust often become susceptible as virulent races continue to evolve in the pathogen populations and spread from one region to another. Race-specific resistance, which usually provides complete protection throughout the entire growth cycle and therefore referred as all-stage resistance, is usually not durable when conferred by a single gene. In contrast, hightemperature adult-plant (HTAP) resistance, which expresses when the weather becomes warm and/or plants grow older, has proven to be non race-specific and durable [1,3,4]. Plants with only HTAP resistance are susceptible to all races at the seedling stage, but become resistant or less susceptible in later growth stages when temperatures increase. Typical HTAP resistance is identified through a four-way (seedlinglow temperature, seedling-high temperature, adult plant-low temperature and adult plant-high temperature) test [1,5]. We routinely screen wheat germplasm and breeding lines for HTAP resistance by testing seedlings with various Pst races at a low temperature profile (diurnal temperature cycle *Address correspondence to this author at USDA-ARS, 361 Johnson Hall, Washington State University, Pullman, WA 99164-6430, USA; Tel: 509-335-8086; Fax: 509-335-9581; E-mail: xianming@wsu.edu changing from 4 o C at 2:00 a.m. to 20 o C at 2:00 p.m.) and then test adult-plants with selected races that are virulent in the seedling tests under a high temperature profile (diurnal temperature cycle changing from 10 o C at 2:00 a.m. to 30 o C at 2:00 p.m.). HTAP resistance has been successfully used in the US Pacific Northwest to reduce severe damage from stripe rust since Dr. Orville Vogel released wheat cultivars Gaines and Nugaines in the early 1960s. However, HTAP resistance is not complete and the disease level is influenced by growth stage, temperature, and inoculum load [1]. Therefore, the best approach is to combine genes for effective allstage resistance with those for HTAP resistance. Numerous genes conferring both types of resistance have been identified. Comparison of the gene structures and predicted proteins of cloned stripe rust resistance genes Yr18/Lr34 and Yr36 reported by Krattinger et al. [6] and Fu et al. [7], respectively, and unpublished Yr10 (Laroche, personal communication) and candidate Yr5 (Chen and associates, unpublished data), together with many race-specific resistance genes across many host species, leads to a hypothesis that resistances controlled by NBS-LRR genes may not be durable, and resistances controlled by non-NBS-LRR genes are more likely to be durable. These cloned genes are not the only genes that contribute to resistance; it is likely that each Yr gene regulates various defense and other functional genes to confer the final response phenotypes. Here, we compare our data in a series of transcriptomics studies previously published [8][9][10][11][12][13] and unpublished to understand molecular mechanisms of race-specific all-stage and nonrace-specific HTAP resistance with an ultimate goal of finding a molecular basis for durable resistance.

COMPARISON OF TRANSCRIPTS INDUCED DURING RACE-SPECIFIC ALL-STAGE RESIS-TANCE MEDIATED BY YR5 AND RACE NON-SPECIFIC HTAP RESISTANCE MEDIATED BY YR39
Yr5 confers typical all-stage resistance to stripe rust, exhibiting infection type (IT) 0 (no visible symptoms) on leaves of the original donor, Triticum spelta album [14] and IT 1 (small necrotic flecks without uredinia) on the nearisogenic line (AvSYr5NIL) in the Avocet Susceptible (AvS) background [15]. In a study to determine genes involved in Yr5-controlled resistance [8,9], the Wheat GeneChip Affymetrix, Santa Clara, CA), which includes 61,127 probe sets representing 55,052 transcripts (www.affymetrix.com), was used to profile changes occurring in wheat near-isogenic lines (a BC 7 :F 4 line of AvS x Tsa and AvS) at 6, 12, 24 and 48 h post-inoculation (hpi) with race PST-78 that is avirulent on the Yr5 isoline and virulent on AvS. The microarray study identified 61 transcripts specific to Yr5-mediated resistance (Fig. 1). These transcripts are genes typically involved in signaling pathways and defense-related events known to occur during R-gene-mediated responses, including protein kinase signaling and production of reactive oxygen species, leading to hypersensitive responses. The gene expression pattern showed a peak at 24 hpi, which is correlated to haustorial formation. In this study, 19 transcripts were identified to be specifically induced for basal defense during the compatible interaction. However, due to lack of R-gene signaling, the response was weak. In contrast, Yr5-signalling resulted in a rapid and strong resistance response. Fig. (1). Comparison of numbers of transcripts associated with Yr5mediated race-specific all-stage resistance (61) and Yr39-mediated race non-specific high-temperature adult-plant resistance (99). Fourteen of the genes were common to both types of resistance.
In the study of Yr39-mediated HTAP resistance [10], the same Wheat GeneChip was used to identify genes induced in two selected F 7 recombinant inbred lines (RILs) from a cross between AvS and Alpowa inoculated with PST-78 urediniospores and mock-inoculated without urediniospores at the flag-leaf stage and grown at the high temperature profile (10-30 o C) after inoculation. The resistant RIL and Alpowa showed a typical HTAP resistance, IT 2-3 (necrotic stripes of 0.5-2.0 cm with occasional uredinia on the edges of necrotic stripes). Under the same temperature conditions, the susceptible RIL and AvS had IT 9 (uredinial stripes without chlorosis or necrosis). In this study, 99 induced transcripts were identified as HTAP resistance-specific (Fig. 1). This number is higher than that specifically involved in Yr5 resistance as discussed above. Transcript accumulation peaked at 48 hpi, which is later than the peak time (24 hpi) for Yr5-mediated all-stage resistance, but corresponded to the time point when rust hyphae were observed microscopically and were undergoing rapid increases in fungal biomass as detected by quantitative PCR assays in the compatible interaction. More than half (50.5%) of the annotated HTAP resistance transcripts were involved in defense and/or signal transduction, including R-gene homologs and transcripts associated with pathogenesis-related protein production, phenylpropanoid biosynthesis and protein kinase signaling. The identification of nine R-gene homologs leads to a hypothesis that these genes regulated by a master gene (Yr39) serve as secondary master genes regulating defense and other related genes contributing to HTAP resistance.
When we compared the Yr39-mediated HTAP resistance with the Yr5-mediated all-stage resistance, we found 14 genes involved in both types of resistance ( Fig. 1) [10]. These genes include WIR1A protein (involved in cell wall structure), beta-1,3-glucanase (a PR protein), phenylalanine ammonium lyase (a phenylpropanoid phytoalexin), peroxidase (involved in oxidative stress), protein kinase and calmodulin protein (involved in signal transduction), carbohydrate (related to transport), and blue copper-binding protein (related to electron transport) [10]. The common genes may be related to host cell death involved in both types of resistance. The putative functions of genes identified in Yr39-mediated resistance, but not in the Yr5-mediated resistance, included R proteins, UDP-glucosyl transferase and hydroxyanthranilate hydroxyl cinnamoyl transferase involved in phenylpropanoids, pleiotropic drug resistance/ABC transporter, putative disease resistance protein, latex protein allergen, receptor protein kinase involved in signal transduction, WRKY5 homolog involved in transcription, and amino acid/protein and ammonium/phosphate/ potassium involved in transportation. Some of the specific genes may explain the durability of the Yr39-mediated resistance, especially the R proteins. Three of the R protein genes are protein kinases with homology to RPG1 protein conferring durable resistance to stem rust in barley [16]. One R gene is a homolog of Cf2/Cf5 LRR disease resistance protein for Cladosporium fulvum resistance in tomato [17]. One putative R gene has homology with the putative stripe rust resistance protein Yr10 (http://pir.uniprot.org/uniprot/Q9FR 63). The remaining three R genes encode putative leucinerich repeat family protein, leucine-rich repeat transmembrane protein kinase, and NB-ARC domain containing protein. In addition, a homolog of Hm1 NADPH-dependent HC-toxin reductase protein was involved in the Yr39-mediated resistance. Hm1 conferring resistance to Cochliobolus carbonum in maize was the first cloned plant disease resistance gene [18]. The collective contribution of these R genes to the Yr39-mediated resistance may require different Pst genes for 61 99 14

Yr5
Yr39 recognition. The diverse R genes may regulate various defense genes involved in different abiotic stress response pathways. All of the R and defense genes make the Yr39mediated HTAP resistance diversely based, perhaps making it difficult for the pathogen to overcome.

COMPARISON OF TRANSCRIPTS IDENTIFIED FOR DIFFERENT GENES CONFERRING RACE-SPECIFIC ALL-STAGE RESISTANCE
In a study aimed at identifying common transcripts associated with race-specific all-stage resistance [13], genes Yr1, Yr5, Yr7, Yr8, Yr9, Yr10, Yr15 and Yr17 were selected because they are available in near-isogenic lines in AvS background. Due to budget limitations, we used a custom microarray instead of the Wheat Affymetrix GeneChip with a primary goal of identifying common transcripts. A total of 343 probes were selected based on their significant expression in the Yr5 and Yr39 studies [8,10]. An avirulent race was used to inoculate two-leaf seedlings of each of the Yr gene lines and PST-78 was used to inoculate the susceptible background line AvS, with the same race being used for each gene when possible. A mock-inoculation was also used for each of the lines. The inoculated seedlings were grown at the low-temperature profile described in the Yr5 study. Leaf samples were taken 24 and 48 hpi for RNA extraction and gene expression analysis. This study identified 28 genes significantly induced during the development of resistance phenotypes across all eight Yr genes [13]. Among these transcripts, those for putative blue copper-binding protein, heatstress transcription factor, pathogen-induced WIR1A protein, and ent-kaurene synthase transcripts were the most significant. Changed transcript levels were uniquely significant in each Yr gene line, indicating transcriptional events specific to particular Yr gene-mediated race-specific resistances. The results confirm the activity of known R-gene-mediated pathway race-specific resistance, including an oxidative burst that likely contributes to a hypersensitive response, as well as pathogenesis-related protein gene expression and activation of the phenylpropanoid pathway.

COMPARISON OF TRANSCRIPTS IDENTIFIED FOR DIFFERENT GENES CONFERRING RACE-SPECIFIC ALL-STAGE RESISTANCE AND RACE NON-SPECIFIC HTAP RESISTANCE IN ADULT PLANTS UNDER HIGH TEMPERATURES
Similar to the meta-analysis of transcripts for racespecific resistance mediated by various genes, we also conducted a study to identify common transcripts associated with race non-specific HTAP resistance mediated by different genes in comparison with all-stage resistance. Isogenic lines having Yr18, Yr29, Yr36 and Yr39 were selected as they were identified as single genes involved in race non-specific HTAP resistance. The isolines with the genes (Yr1, Yr5, Yr7, Yr8, Yr9, Yr10, Yr15 and Yr17) studied at the seedling stage [13] were also included. The custom microarray, experimental design, procedure, data collection and analyses were all as described for the race-specific resistance study [13], except that inoculated plants were grown in the high-temperature profile as described for the Yr39 HTAP resistance study [10]. Boot stage adult plants of NILs Yr1, Yr5, Yr7, Yr8, Yr9, Yr10, Yr15 and Yr17 were inoculated separately with appro-priate avirulent races (PST-21 for Yr8, Yr9, Yr10 and Yr17; PST-45 for Yr1 and Yr7; and PST-78 for Yr5 and Yr15) and virulent races (PST-17 for Yr1; PST-43 for Yr10; PST-45 for Yr17; PST-78 for Yr7, Yr8 and Yr9; an Australian isolate for Yr5; and no isolate virulent for Yr15); and those of the single gene lines for Yr18, Yr29, Yr36 and Yr39, together with AvS, were inoculated with PST-78 that is virulent on seedlings, but not adult-plants, with these genes.
Stripe rust infection type data observed 20 days after inoculation were as expected for each compatible or incompatible Yr gene-Pst race combination, except the Yr8, Yr10 and Yr17 single gene lines in the presumed compatible interactions. Adult plants of these three lines exhibited resistance in the test with virulent races. The phenotypes of Yr8 and Yr17 confirmed the presence of HTAP resistance (Chen and associates, unpublished data), while that of the Yr10 line was surprising. Because these lines have both all-stage and HTAP resistance, they were not included in the analyses of comparing transcripts of HTAP resistance with all-stage resistance.
For the race-specific adult-plant resistance gene lines, two probes [ribosomal protein L2 (Ta.28514.1) and mitogenactivated protein kinase (Ta.236.1)] were down-regulated when compared to their mock-inoculated checks and no probes were down regulated across the all-stage resistance gene lines compared to their compatible race inoculations.  [8,13] are also given in (Table 1).  is virulent on seedlings of the HTAP resistance gene lines, was used in inoculation. The Yr36 line was included in the study, but no significantly induced genes were identified. e The Yr8 near-isogenic line (AvSYr8NIL) also has a gene for race non-specific HTAP resistance linked to the race-specific all-stage resistance gene Yr8.
In general, transcripts detected in the seedling lowtemperature tests were also significant in the adult-plant high-temperature tests for race specific all-stage resistance.
Based on common and unique transcripts identified in the Yr gene-mediated resistances, a dendrogram was constructed to show their relationships (Fig. 2). Yr5 was more closely related to Yr17; Yr8 was more closely related to Yr10; and Yr18 was more closely related to Yr39 than to other genes. Yr29 was more distantly related to all of the other genes. Yr36 was not included in the dendrograms as none of the transcripts for other genes was significantly changed in expression levels; this could indicate that it utilizes signaling and defense pathways that are different from those identified for the other genes. Although this hypothesis needs to be tested, the results may be in agreement with the previous finding that Yr36 is a very old gene that is not present in common wheat cultivars [7].
After various comparisons, five genes were clearly identified to be involved in race-specific all-stage resistance controlled by Yr1, Yr5, Yr7, Yr9 and Yr15 and only one gene was commonly expressed in HTAP resistances mediated by different genes (Yr18, Yr29, Yr36 and Yr39) ( Table 2). The annotation of the five transcripts specific to all-stage resistance provided additional evidence for classic R-gene mediated pathways being involved in race-specific resistance. The five genes commonly involved in all-stage resistance included a hydroxyproline-rich glycoprotein, a NB-ARC domain containing protein, a protein kinase and two functionunknown genes. In addition to the separate analysis at the seedling stage [13], we found another NB-ARC protein and a protein kinase for defense signaling. Also, the hydroxyproline-rich glycoprotein is involved in cell wall strengthening like the WIR1A protein [19]. Among the five genes, three were first identified in Yr5-mediated all-stage resistance and two were first identified in the Yr39-mediated HTAP resistance. The transcript with significant changes across all HTAP resistances is a nonclathrin coat protein.
The function of nonclathrin coat protein-related resistance is not clear, but such proteins have been reported to bind to the cytoplasmic dilysine motif of membrane proteins of the early secretory pathway [20]. The nonclathrin coat protein identified in this study may be involved in transporting antifungal substances across the cell membrane to directly contact Pst haustoria or hyphae. The lack of many shared transcripts in HTAP resistance compared to all-stage resistance leads to a hypothesis that diverse genes and biochemical pathways are used by HTAP resistance controlled by different genes. Together with diverse genes and pathways identified for Yr39mediated HTAP resistance [10], we conclude that highly diverse genes and biochemical pathways are the molecular basis for the race non-specificity and durability of HTAP resistance. Fig. (2). Hierarchical clustering (Euclidean metrics, complete linkage, distance indicated at each branch) of shared and unique data from ( Table 1).

PERSPECTIVES AND CONCLUSIONS
The transcriptomics studies conducted so far have identified genes involved in Yr5-mediated all-stage resistance and Yr39-mediated HTAP resistance; and common genes involved in all-stage resistance and HTAP resistance mediated by different Yr genes. The results have provided some insights for understanding the molecular mechanisms of racespecific resistance compared to race non-specific resistance. In particular, the studies have linked a large number of genes with diverse functions to race non-specificity and durability of HTAP resistance. The data of these studies lead to several hypotheses to be tested and more studies to be conducted for a better understanding of various types of resistance and how to utilize the basic information to achieve more sustainable and better control of stripe rust.
We are currently conducting studies to test a hypothesis that ABC transporter proteins are involved in nonrace specific HTAP resistance. We are in the process of obtaining the full-length sequence for the identified ABC transporter gene, and so far have 5,754 bp of the genomic sequence from the Yr39 donor, Alpowa, using a PCR based genome walking technique (Clontech GenomeWalker TM Universal Kit, 638 904). Thus far, the ABC transporter-like wheat gene has more than 75% identity in genomic sequence to a rice gene, Os01g42410, with the greatest differences occurring in intron regions. Sequence conservation of the ABC transporter gene in Alpowa (Yr39) also appears to be high across different wheat cultivars after alignment of the newly acquired sequence with Chinese Spring genomic sequences (cerealsdb.uk.net) failed to detect significant differences, supporting the high up-regulation of the gene in the Yr18 line presented above, as Chinese Spring has Yr18 [6,21]. Comparison of the ABC transporter gene in Alpowa with Yr18/Lr34 [6,21] shows that they have low nucleotide sequence similarity (39%), confirming our initial hypothesis that they are different genes. Future goals include obtaining the fullgenomic sequence of the gene in addition to the 5' untranslated region from Alpowa and other wheat cultivars to identify functional domains and single nucleotide polymorphisms (SNPs) potentially influencing HTAP resistance in Alpowa and other wheat cultivars.
The identification of the nine R-protein genes involved in Yr39-mediated HTAP resistance was initially a surprise to us, as R proteins are largely believed to be involved in recognition of pathogen effectors, leading to race-specific resistance. However, the high number of such types of genes leads us to believe that these genes collectively contribute to race non-specificity and therefore durability. Our hypothesis is that when up-regulated by the Yr39 master gene, these genes serve as secondary master genes regulating other defense or functionally related genes to operate the entire defense machinery against Pst infection and growth in the plant tissue. These R proteins may recognize different effectors, which may make it difficult for the fungus to change to nonrecognition. In regard to the ABC transporter gene, we are currently obtaining full-length sequences of these R genes to characterize them among wheat genotypes and to determine their functions for the HTAP resistance phenotype and their roles of being regulated or regulating in the total network of defense pathways.
With a primary goal of identifying transcripts commonly involved in all-stage resistance or HTAP resistance controlled by different genes, the custom microarray was constructed using genes identified in the Yr5 and Yr39 studies to represent those involved in either type of resistance. However, such a cost-saving approach did not allow us to identify transcripts uniquely involved in resistance mediated by individual Yr genes. Therefore, we still do not have the majority of the genes identified for all Yr genes studied, except for Yr5 and Yr39. Using the Wheat Affymetrix GeneChip, which continues to have new genes added, is still a useful highthroughput technique to identify possible genes involved in resistance to stripe rust and other diseases in wheat. Alternately, transcriptome sequencing [22,23], which can be used to determine numbers of transcripts for genes, should be useful to study genes involved in different types of resistance.
HTAP resistance has two components: temperature sensitivity and developmental stage dependence. These two components are not equally required for resistance. Among cultivars with a broad-sense HTAP resistance, which can be determined by a virulent race in a seedling test under low temperatures and in a field or greenhouse test with the same race under high temperatures, resistance in some cultivars is more temperature sensitive whereas others are more plant-stage dependent. In our studies, Yr39 is more typical of HTAP, where the maximum expression of resistance is in flag leaves and under high-temperatures. In contrast, both Yr18 and Yr36 can express resistance even in the seedling stage when under high temperatures [6,7]. Thus it is likely that some transcripts involved in HTAP resistance respond more to temperatures, some more to growth stage, and others more or less neutral. In our studies to date this issue has not been addressed. Identification of genes responding to different environmental and growth stage conditions may allow choice of durable resistance genes that are more suited to specific regions in order to more effectively diversify resistance.
In comparison of a pair of two resistance genes which regulate different transcripts with another pair of two resistance genes which regulate commonly shared transcripts, the former pair of resistance genes when in combination may lead to more durable resistance than the latter pair of genes as the combined resistance conferred by the former pair is based on more diverse defense pathways. Furthermore, the correct combinations of genes may provide higher levels of resistance. In this way, transcriptomics studies of resistance genes will not only provide an understanding of the basic mechanisms of resistance, but will allow for immediate ap-plication in selecting resistance genes for precision breeding. The results may revealed which genes are more likely to be durable and which are not. It will also add to the biotechnological tool box for identification of the same or different genes. As the technology is advancing, transcriptomics testing should become less expensive and higher throughput. It will become more feasible to use transcriptomics approaches for identifying genes and developing markers for different types of resistance.

CONFLICT OF INTEREST
The author(s) confirm that this article content has no conflicts of interest.