Wheat stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), is a devastating disease of wheat worldwide (Wellings, 2011). In China, the annual yield losses to stripe rust in wheat were estimated to be about 1 million metric tons (Chen et al., 2009). Cultivation of resistant varieties is the most effective, economical, and environmentally friendly way to control the disease. Although many resistance genes have been identified and utilized in wheat cultivars (Cheng et al., 2014; Lu et al., 2014; Han et al., 2015), the protection conferred has not been durable due to genetic variation in the pathogen population. One example is that of resistance gene Yr26 that was widely used in Chinese wheat breeding in recent years. New Yr26-virulent (CYR34, V26) races are now increasing and are causing unacceptable levels of disease on many of the cultivars with Yr26 (Han et al., 2015). Although resistances governed by quantitative trait loci (QTL) confer a broader-spectrum resistance, the level of resistance is not adequate to prevent significant crop losses, especially under severe epidemic conditions (Niks et al., 2015b). Thus, more durable control of stripe rust is urgently needed, and in addition to current exploration and identification of new resistance genes in wheat and its close relatives, a better understanding of non-host resistance (NHR) may offer opportunities in breeding for sustainable disease control.

In nature, a specific pathogen usually causes disease on a few plant species; that is, most plants are resistant to a wide range of phytopathogens. This form of disease resistance exhibited by all members of a plant species to all genetic variants of a non-adapted pathogen species [or possibly formae speciales (f. sp.)] is known as NHR. Due to its broad-spectrum effectiveness and durability NHR is of considerable interest for crop resistance improvement.

The genetic and molecular mechanisms underlying NHR remain largely unknown. Currently, the best-studied example of NHR is interaction between Arabidopsis and the non-adapted barley biotrophic fungal pathogen Blumeria graminis f. sp. hordei (Bgh), the causal agent of barley powdery mildew. Three NHR genes PEN1, PEN2, and PEN3 required for penetration resistance of Arabidopsis to Bgh have been isolated. Functional mutants of any one of the three PEN genes display increased penetration rates by Bgh. PEN1 (Collins et al., 2003; Lipka et al., 2005; Stein et al., 2006) encodes a membrane-associated syntaxin containing a SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) domain and is a member of a large family of proteins functioning in membrane fusion and secretion events. Cytological studies have demonstrated that PEN1 is involved in papilla formation. PEN2 encodes a glycoside hydrolase, which associates with the periphery of peroxisomes. PEN3 encode an ABC (ATP binding cassette) transporter that is localized to the plasma membrane. PEN2 and PEN3 may collaborate in transport of antimicrobial compounds to the apoplast. Importantly, the protein products of these PEN genes accumulate at sites of fungal penetration. Moreover, these genes also contribute to R gene-mediated resistance and cell death in response to both adapted and non-adapted pathogens (Johansson et al., 2014). While, for those plants that are evolutionary closely related to the natural host, the NHR is proposed to be predominantly mediated by multiple R genes that collectively confer resistance to all isolates of a pathogen species (Schulze-Lefert and Panstruga, 2011). For example, the NHR of Arabidopsis to non-adapted Albugo candida may result from recognition of pathogen effectors by multiple WRR4-like genes, which encode typical NB-LRR resistance proteins (Borhan et al., 2008).

There are many published reports on NHR to rust fungi. For example, NHR of Arabidopsis and Medicago to bean rust (Patto and Rubiales, 2014; Ishiga et al., 2015; Langenbach et al., 2016), barley and Brachypodium distachyon NHR to cereal rust pathogens (Zellerhoff et al., 2010; Dawson et al., 2015; Figueroa et al., 2015; Niks et al., 2015a). As the model of monocotyledonous plant, rice is unusual in not being affected by a rust pathogen (Ayliffe et al., 2011b). Several studies indicated that rust fungi have some potential to infect rice and trigger host defense responses such as production of reactive oxygen species and accumulation of Pathogenesis-related proteins (Ayliffe et al., 2011a; Li et al., 2012). Thus, it is of interest to characterize non-host interaction between rice and cereal rust pathogens and identify key genes involved in NHR. Proteomic studies revealed proteins that are involved in phytoalexin production, and glycerol-3-phosphate metabolism may have a role in rice NHR to P. triticina and Pst (Li et al., 2012; Zhao et al., 2014).

In the present study we identified rice mutant crr1 (compromised resistance to rust 1) that allowed a high level of penetration rates by Pst and enhanced hyphal growth. The fungus was able to develop haustoria in mesophyll cells of the mutant, but failed to sporulate. Histological analysis revealed that hydrogen peroxide (H₂O₂) production and callose deposition were not affected in the crr1 mutant. Furthermore, upon infected by Pst crr1 showed strikingly enhanced expression levels of defense-related genes involved in the SA-, and JA-mediated defense pathways as well as phytoalexin synthesis. These observations suggested different molecular mechanisms underlying NHR to Pst in rice compared to the host resistance. Genetic analysis demonstrated that the phenotype of crr1 was conditioned by a recessive gene between markers ID14 and RM25792 at the end of rice chromosome 10. Characterization and genetic study of crr1 would provide new insights into NHR, and assist in breeding wheat cultivars with durable resistance to stripe rust.

Non-host resistance has potential to provide non-specific, durable resistance to diseases, and to offer alternative possibilities to traditional R gene-mediated resistance. Unfortunately, no rice gene underlying NHR to rust has been identified so far. In this study, we characterized and identified a gene conferring NHR to Pst using a rice mutant crr1. This study may contribute to our understanding of rice NHR to Pst and exploiting rice NHR to develop wheat lines with more durable resistance to Pst.

According to the two-layered paradigm of plant active immune system, there are two types of NHR: pre-invasion defense mediated by PAMP-triggered immunity (PTI) and post-invasion defense mediated by effectors-triggered immunity (ETI) (Gill et al., 2015). Our data showed that the frequency of stomatal penetration of Pst urediniospores on wild type rice (0.2%) was significantly reduced than that on wheat (over 25% in previous study) (Cheng et al., 2013). Similar studies demonstrated that 0.5% urediniospores of P. triticina formed ssv on Arabidopsis leaves (Shafiei et al., 2007). These results suggested that rice NHR to Pst might largely act at the pre-invasion stage. Histologic studies revealed H₂O₂ production and callose accumulation concentrated in and around guard cells following Pst inoculation, but no hypersensitive cell death was detected. Therefore, PTI mediated defense responses that precede host mesophyll cell invasion appear to play the predominant roles in blocking attempted Pst. Compared with intermediate hosts of Pst, such as barley and Brachypodium, rice is more distantly related to wheat and resistant to all rusts (Vogel et al., 2010; Mayer et al., 2011). Thus our results fit well with the molecular evolution model that PTI play a key role when pathogens attempt to infect more distantly related non-host species (Schulze-Lefert and Panstruga, 2011). Increased frequency of stomatal penetration and larger infection sites in crr1 may result from a disrupted defense step preceding mesophyll cell invasion.

Although crr1 plants exhibited more susceptibility to Pst, the H₂O₂ production was not affected relative to wild type. In addition, callose deposition increased as hyphal growth expanded in leaf tissues of the crr1 mutant. Similarly, the induction of defense related genes, such as genes involved in SA/JA mediated defense pathway and phytoalexin synthesis, were quicker and stronger in crr1 than that in wild type plants. Collectively, these findings suggest that increased development of Pst in apoplasts of crr1 plants triggers a more extensive defense response. Counter-intuitively, these defense responses seemed to have limited effects on preventing the growth of Pst. Similar findings were reported for non-host interaction between pen mutants and the barley mildew pathogen (Collins et al., 2003; Lipka et al., 2005; Stein et al., 2006). A natural variant of Arabidopsis accession Wa-1 which has compromised resistance to non-adapted wheat leaf rust pathogen exhibits increased SA and PR1 expression following P. triticina challenge (Shafiei et al., 2007). In these cases, there seems to be a “true” NHR which spanning multiple layers of defense. Crr1 mutants are indeed compromised for part of the defense (mainly related to ingress). However, the penetration rate is still very low and most defense responses still seem to be intact and in fact the pathogen cannot really proliferate/sporulate either.

Haustoria are special structures of biotrophic fungus that suppress host immune system and take up nutrients from host cells. Formation of haustoria is usually regarded as the symbol of parasitism establishment (Yi and Valent, 2013). Haustoria have been observed by fluorescence microscopy in several studies concerning rice-rust fungus interactions, while the details of haustoria is not clear yet (Ayliffe et al., 2011a; Yang et al., 2014). In the present study, we observed haustoria formation in mesophyll cells in rice mutant by TEM. This finding suggested that rust fungi have the potential to absorb nutrients and infect rice. However, only a limited number of haustoria were observed. Although urediniospores have stored nutrients for germination and initial infection, it was unbelievable that such a few hautoria can support the considerable colonization that encompassed hundreds of host cells in crr1. This suggests the possibility that hyphae are able to take up nutrients from the intercellular spaces. Although it is generally believed that the apoplast of plants leaves is a relatively nutrient-poor environment, a considerable number of microbes do derive nutrients from it. In a study of interaction between barley and P. hordei, sucrose and glucose were found in apoplast at much lower concentrations in infected than in healthy leaves, and uptake of hexoses by intercellular hyphae was suggested as the cause of the reduction (Tetlow and Farrar, 1993). A sucrose transporter SRT1 from Ustilago maydis was shown to take up sucrose from intercellular spaces of maize leaves allowing the hyphae to grow along the phloem (Wahl et al., 2010). Thus, we postulated that hyphae of Pst may obtain nutrients from the apoplast in leaves of the crr1 rice mutant. These nutrients absorbed from the intercellular spaces support limited hyphal growth, but they may be insufficient to support sporulation.

Using BSA and CIM strategies we located the crr1 locus between id14 and RM25792 at the end of chromosome 10. This region contains about 40 annotated genes, and there is no previously annotated host or non-host resistance gene in the interval. Recently, three genes involved in Brachypodium NHR to Pst were identified by flanking sequence isolation from mutant (An et al., 2016). Among them, Bradi5g17540 encodes a BAP28 domain containing protein, Bradi5g17540 encodes a MYB transcription factor and Bradi5g11590 encodes a lipoxygenase. However, none of them is present in the candidate region of crr1. Another gene conferring non-host resistance to P. striiformis was identified in barley, and designated as Rps6 (Andrew et al., 2016; Li et al., 2016). Rps6 was mapped on the long arm of barley chromosome 7H, which has no collinearity with rice chromosome10, suggesting that it is different with crr1 locus.

Several studies have demonstrated the feasibility of transferring single NHR-related genes across plant species to create durable, broad-spectrum resistance (Brutus and He, 2010; Wen et al., 2011; Schoonbeek et al., 2015; Langenbach et al., 2016). The mechanisms of rice resistance to Pst can be used to improve wheat resistance. Since it is not feasible to transfer gene through homoeologous recombination by hybridization, transgenic strategy might be a good alternative. Thus, further studies based on more comprehensive efforts will be needed to clone the corresponding rice gene conferring NHR to stripe rust fungus.

ZK and JZ conceived and designed research. JZ and YY performed the genetic mapping and analyzed the data. JZ and DY performed the histological observation and gene expression analysis. GZ, HZ, and YC screen and identified the mutant. JW and KZ developed genetic population and molecular marker. MJ performed the TEM experiment. LH and ZK contribute comments during manuscript preparation. JZ wrote the manuscript and ZK revised the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.