Plants produce cytokinin (CK) hormones for controlling key developmental processes like source/sink distribution, cell division or programmed cell-death. Some plant pathogens have been shown to produce CKs but the function of this mimicry production by non-tumor inducing pathogens, has yet to be established. Here we identify a gene required for CK biosynthesis, CKS1, in the rice blast fungus Magnaporthe oryzae. The fungal-secreted CKs are likely perceived by the plant during infection since the transcriptional regulation of rice CK-responsive genes is altered in plants infected by the mutants in which CKS1 gene was deleted. Although cks1 mutants showed normal in vitro growth and development, they were severely affected for in planta growth and virulence. Moreover, we showed that the cks1 mutant triggered enhanced induction of plant defenses as manifested by an elevated oxidative burst and expression of defense-related markers. In addition, the contents of sugars and key amino acids for fungal growth were altered in and around the infection site by the cks1 mutant in a different manner than by the control strain. These results suggest that fungal-derived CKs are key effectors required for dampening host defenses and affecting sugar and amino acid distribution in and around the infection site.
The role of plant-like hormonal compounds produced by fungal pathogens during infection has not been elucidated. Here we identified a conserved gene in most fungi, required for cytokinin production by the rice blast fungus and for its full virulence. Fungal-derived cytokinins are likely potent inhibitors of plant immunity. They are also needed to maintain elevated sugar contents at the site of infection and to drain or consume essential amino acids at, and around, the infection site. Thus, cytokinins represent the second example, after the bacterially-produced coronatine, of plant hormones hijacked by pathogens to successfully invade plant tissues. These findings also suggest that this invasion strategy could be widely conserved among fungi.
Citation: Chanclud E, Kisiala A, Emery NRJ, Chalvon V, Ducasse A, Romiti-Michel C, et al. (2016) Cytokinin Production by the Rice Blast Fungus Is a Pivotal Requirement for Full Virulence. PLoS Pathog 12(2): e1005457. https://doi.org/10.1371/journal.ppat.1005457
Editor: Jin-Rong Xu, Purdue University, UNITED STATES
Received: July 31, 2015; Accepted: January 25, 2016; Published: February 22, 2016
Copyright: © 2016 Chanclud et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work was partially supported by Agropolis and Cariplo Fondations under the reference « Rice Connections » 1201-001. This project is supported by the French ANR program “Investissement d’Avenir” ANR-10-LABX-0001-01 and a Canadian NSERC Discovery Grant. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Plant pathogens have evolved sophisticated strategies to manipulate host biological processes during infection [1,2]. (Hemi)biotrophic pathogens produce and secrete effector proteins and metabolites to hijack cellular metabolism of the infected tissues to their own benefit. For instance, the bacterial pathogen Xanthomonas oryzae pv. oryzae produces Transcription Activator Like effectors that specifically induce the expression of genes coding for sugar transporters and thus enhance bacterial nutrition . The virulence factors can also participate to the inhibition of plant defenses that lead to cell death, thus contributing to maintaining infected cells alive . Plant-pathogen interactions shaped the pathogen virulence arsenal and the host immune response system. A first layer of plant defenses is induced by the perception of pathogen- or microbe-associated molecular patterns, like flagellin from bacteria or chitin from fungi . These basal defense responses consist of an early accumulation of reactive oxygen species (ROS), a thickening of the cell wall, and production of metabolites/enzymes with antimicrobial activities. To limit these defenses triggered by chitin perception by the plant’s chitin receptor, fungal pathogens like Magnaporthe oryzae secrete chitin-binding effectors that enable escape from the host recognition system . Pathogens also interfere with other steps of plant immunity like signaling cascades following recognition and transcriptional regulators of host defenses [2,7].
Plant pathogens manipulate components of hormonal pathways, whether the corresponding hormones are involved in disease resistance (i.e. salicylic acid, jasmonic acid and ethylene ) or in the control of plant developmental processes (i.e. auxin, cytokinins and gibberellic acid [8–11]). Pathogens can affect hormonal homeostasis by targeting/secreting enzymes involved in hormone metabolism or by producing hormonal/mimicking compounds and thereby inhibit defenses, modify nutrient flows and/or induce symptom development. For instance, the fungal pathogens Ustilago maydis and Magnaporthe oryzae secrete respectively a chorismate mutase and a monooxygenase affecting salicylic acid or jasmonic acid homeostasis during infection and then contributing to their virulence [12,13]. Plant pathogens can also directly produce hormones or compounds with similar biological activities. The bacterial pathogen Pseudomonas syringae produces coronatine to mimic jasmonic acid, which actively opens stomata and counteracts salicylic acid accumulation [14,15]. As a consequence, these combined effects on several processes required for infection facilitate host invasion. In the case of pathogenic fungi, there is, however, no direct evidence that hormonal compounds produced are required for virulence.
Fungi and bacteria produce hormonal compounds that are chemically identical or very close to plant hormones involved in plant developmental processes such as cytokinins (CKs) [16–22]. CKs are adenine derivatives that differ in their side chains (aromatic or isoprenoid). In plants, isoprenoid CKs are mainly synthesized through a de novo biosynthesis pathway from adenosine phosphate. In this pathway, Isopentenyl transferase (IPT) enzymes perform the first step of biosynthesis. However, another minor CK biosynthesis pathway involving tRNA modification is also described in plants and yeast. This minor pathway requires tRNA-Isopentenyl Transferases (tRNA-IPT), enzymes that perform tRNA modification leading to CK production after tRNA degradation [23–25]. In plants, CKs were originally studied for their effects on cell division/differentiation . CKs also modulate nutritional source/sink distribution and programmed cell-death processes like xylem differentiation, senescence and hypersensitive response [27–30]. Tumor-forming pathogens are striking examples of microbes which are able to produce CKs, like the fungi U. maydis  and Claviceps purpurea  or the protist Plasmodiophora brassicae . Morrison et al., (2015a)  have recently shown that CK accumulation in U. maydis infected tissues is correlated to the virulence of this pathogen. However, in this latter case the evidence that these hormones are produced by the pathogen is still unclear. Likewise, the pathogenic bacterium Rhodococcus fascians secretes CKs to interfere with host CK signaling pathways by affecting the transcriptional regulation of key cell cycle genes leading to tumor development . In addition to having potential roles in the virulence of tumor-inducing pathogens, CKs have long been suspected to participate to virulence of pathogens that do not trigger tumors or organ deformations. However, it has not been demonstrated that the CKs produced by this type of pathogens are key virulence factors.
Because CKs delay plant senescence by limiting oxidative burst and maintaining photosynthesis activity , they are at the cross-road of several pathways of interest for manipulation by pathogens that need to keep host cells alive in order to drain nutrients for their own growth. Consistent with this role, CKs are accumulated in “green islands”, which are tissues corresponding to photosynthetically active zones maintained around pathogenic lesions caused by (hemi)biotrophic pathogens [36–39]. CK compounds have been measured in several different fungal species (even in non-plant interacting ones ) and CK secretion was mostly observed in the case of (hemi)biotrophic microbes [32,41]. For instance, Hu & Rijkenberg (1998) immuno-detected CKs in and around fungal hyphae of Puccinia recondita f.sp. tritici during wheat infection .
The hemibiotrophic fungus Magnaporthe oryzae responsible for the rice blast disease, also produces CKs in vitro  but the CK biosynthesis pathway in this fungus has not been established. Moreover, the role of CKs in the virulence of filamentous fungi that, like Magnaporthe, do not form tumors has never been demonstrated. In this study we identified a gene coding for a putative tRNA-Isopentenyl Transferase (tRNA-IPT) in the genome of M. oryzae and in all Ascomycete genomes tested. Mutants deleted for this gene, named Cytokinin Synthesis 1 (CKS1), were generated and were found to be impaired in CK production confirming that the tRNA degradation pathway is involved in fungal CK production in Magnaporthe as was suggested for other fungi by Morrison et al., (2015b) . Deleted cks1 mutants were not affected in their in vitro growth or asexual development in standard minimal conditions however they showed severely reduced virulence on rice. Remarkably, the CK-deficient mutant was not able to maintain the levels of several key nutrients at the infection site and induced early and strong plant defenses, suggesting that fungal CKs may contribute to metabolite mobilization and to rice defense inhibition. Our work confirms the view that, in M. oryzae, and possibly in many other fungi, this putative tRNA-IPT, contributes to CK production. Since CKS1 is required for the full virulence of Magnaporthe, we propose that CKs from fungal pathogens could act as effectors combining functions in defense inhibition and nutrient mobilization.
In silico analysis of M. oryzae genome revealed putative orthologous genes of yeast and plant cytokinin pathways
To investigate the role of CKs in the virulence of the blast fungus M. oryzae, homologs of plant and yeast genes involved in CK metabolism or signaling were searched in the M. oryzae genome using BLASTp. Potential orthologous genes of some important CK biosynthesis, degradation and signaling component coding genes were identified (S1 Table), further extending previous reports that the CK signaling and metabolic pathways are present in M. oryzae. As previously mentioned, independently of the de novo biosynthesis pathway, CKs can be also produced through tRNA degradation. In this second pathway the first step of CK production involves tRNA-Isopentenyl transferases (tRNA-IPT), as previously described in A. thaliana (AtIPT2 and AtIPT9) and Saccharomyces cerevisiae (MOD5) [23–25,44,45]. A putative tRNA-IPT protein, coded by MGG_04857, was identified in M. oryzae which shared 36% identity with AtIPT2 or MOD5 and the sequences of the different known interaction sites were highly conserved (S1A Fig). No other gene containing the coding sequence of the IPT domain was identified, suggesting that there is only one orthologous gene in M. oryzae. We modeled the three-dimensional structure of MGG_04857 (S1B Fig) by protein threading, a method based on fold recognition  and using the experimentally determined MOD5 structure as a template. The model produced suggested a conserved structure, but also confirmed a high conservation of primary sequences of interaction sites i.e. the ATP, DMAPP as well as tRNA binding sites (S1B Fig [47,48]) and indicated that MGG_04857 shows all the features of a bona fide tRNA-IPT. Taken together, these results support the hypothesis that the blast fungus can produce CKs , potentially perceive them, and suggest that MGG_04857 could act as a key enzyme in fungal CK production.
Production of MGG_04857 (CKS1- Cytokinin Synthesis 1) mutants and their growth in vitro
To test the involvement of the putative tRNA-IPT coded by the MGG_04857 gene in fungal CK production, knock-out mutants of this gene, later called cks1 (see below), were generated by homologous recombination in the M. oryzae GY11 genetic background (S2 Fig). In addition, a complemented strain (cks1CKS1) was built by transformation of the cks1 mutant strain with a construct carrying the genomic sequence of MGG_04857 under control of its own promoter. Gene disruption and complementation were confirmed by PCR on genomic DNA and qRT-PCR and showed that the expression of MGG_04857 was similar between wild type and cks1CKS1 isolates, but not detected in the cks1 mutants (S2E Fig).
Since yeast MOD5 mutants are known to display altered phenotypes , we measured the growth of cks1 mutants under normal and environmental stress. The in vitro growth of the cks1 mutant was not different from the wild type or cks1CKS1 strains under standard minimal growth conditions (S3A Fig). Moreover, the development of infection structures like the appressorium was not impaired by the cks1 mutation on glass slides (S3B Fig) or on rice leaf surface (S3C Fig). This suggests that the early steps of fungal growth are not significantly modified by the cks1 mutation. By contrast, the cks1 mutant showed slight but significant and reproducible reduced growth when grown under 1mM H2O2 (S4A Fig). This phenotype could be complemented by the addition of the CK kinetin to the medium (S4B Fig), suggesting that this defect could be related to a reduced CK production due to the MGG_04857 deletion.
MGG_04857 (CKS1- Cytokinin Synthesis 1) is required for CK biosynthesis in M. oryzae
To investigate the role of MGG_04857 in CK production, CK levels were determined by liquid chromatography-positive electrospray ionization tandem mass spectrometry (LC(ESI+)-MS/MS) [19,50] both in the culture supernatant and mycelia of the different strains (Table 1).
Four major isoprenoid CKs, cisZR (cis-zeatin riboside), iPR (isopentenyl adenosine), cisZNT (cis-zeatin nucleotide) and iPNT (isopentenyladenine nucleotide) were detected in mycelia and supernatant of the wild type GY11. The riboside forms, cisZR and iPR, were the most abundant CKs in mycelia while the nucleotide precursor forms, cisZNT and iPNT, were the main secreted ones. The other CK forms, trans-zeatin and dihydrozeatin were not detected. Thus, in the minimal medium, cisZR and cisZNT are the major CKs produced and secreted by the wild-type strain. This is consistent with CK measurements made on several other fungi where cis-zeatin forms were the main CKs detected . The cks1 mutant was not able to produce and/or secrete any detectable CKs whereas the complemented strain, cks1CKS1, produced CKs at similar levels as the wild type strain GY11 did. Thus, MGG_04857 appears to be required for CK biosynthesis in the rice blast fungus (therefore called CKS1 for Cytokinin Synthesis 1) and is likely coding for an active tRNA-IPT protein, although this activity needs to be established.
M. oryzae cks1 mutants have reduced virulence on rice
To test for involvement of fungal CKs in the interaction between rice and the blast fungus, we inoculated mutant and control strains on the susceptible rice cultivar Nipponbare. The cks1 mutant strain was less virulent than cks1CKS1 or GY11 wild-type strains as shown by a reduction of disease symptoms (Fig 1A).
Nipponbare plants were inoculated with Magnaporthe cks1 mutants and control strains to evaluate virulence. The results for control GY11 and the complemented strain were similar and are only shown for symptoms. (A) Disease symptoms were observed 6 days after inoculation. Grey spots represent susceptible lesions whereas brown spots represent failed penetration events. The number of susceptible lesions per leaf (generalized linear model, p-value = 0.02) and size of lesions (mixt model, p-value = 0.003) were measured as shown in (B) and (C) respectively (for more details see materials & methods). The values represent the mean and SD from four biological replicates each composed of 6 plants. The percentage of spores from the cks1 and cks1CKS1 that penetrated the leaf was measured (D) under the microscope at different time points (hpi: hour post-inoculation). The data presented is the mean and SD of three biological replicates (>100 infection sites/replicate). A t-test was used to compare the penetration of the mutant and control strains, *, p-value < 0.01; **, p-value < 0.002; ***, p-value < 3.10−5. All experiments were repeated three times with similar results and one representative experiment is shown here.
The CK-deficient strain produced less grayish, sporulating lesions per leaf (Fig 1B) and these lesions were smaller than those caused by control strains (Fig 1C). These results suggest that leaf penetration (reflected by lesion number) and invasion (reflected by lesion size) of the CK-deficient strain are both impaired, although not completely abolished. Impaired penetration was confirmed by microscopic observation of individual interaction sites which showed that cks1 failed to penetrate as efficiently as cks1CKS1 during the early times of infection (<24 h, Fig 1D). This is likely due to defective steps after appressorium formation since this developmental stage was unaffected by the cks1 mutation (S3C Fig).
The reduced virulence of cks1 mutant is linked to the absence of cytokinins
As a test for biological activity of fungal-produced CKs, we performed qRT-PCR and quantified the expression of rice Response Regulator (RR) genes OsRR6 and OsRR1 (Fig 2) that were previously described to be transcriptionally regulated by CKs [51,52] and can thus be used as CK bio-sensors .
The transcriptional regulation of CK marker genes (OsRR6 (Os04g57720) and OsRR1 (Os11g04720) as named by Pareek et al., (2006) was evaluated by quantitative RT-PCR using the Actin gene for normalization. Nipponbare plants were inoculated with spore suspension (in gelatin 0.5%) of either the cks1 mutant (black bars) or cks1CKS1 control strain (white bars) and gene expression was measured at 2, 4 and 6 hours post inoculation (hpi), before penetration of the leaf tissues but at stages where there was no significant difference in growth of the cks1 mutant compared to the complemented strain (S3C Fig). The values presented are the Log2 ratios (infected/not infected) of the means calculated from four independent replicates. Uninfected plants were sprayed with gelatin 0.5% but without spore suspension. This experiment was repeated three times and showed similar results. A t-test was used to compare the means of expression quantified in cks1 (black bars) and cks1CKS1 (white bars) inoculated plants. *: values significantly different at p-value < 0.05.
During early contact of fungal conidial spores with the plant surface (2h, 4h, 6h), and at a stage where all strains showed similar growth in vitro or on the plant surface (S3C Fig), the two CK response markers tested had significantly lower expression in plants inoculated with cks1 than with cks1CKS1 strains. Similar results were found for all other RR genes tested (S5A Fig). This result supports the hypothesis that fungal CKs affect host CK signaling pathway and is consistent with the report made by Jiang et al., (2013)  that conidia, the first M. oryzae cells in contact with the host, can produce CKs.
To test if CKs could restore the virulence of the mutant strain, we exogenously applied the CK kinetin at 24h after inoculation (Fig 3) since the delay of penetration of the mutant is noticed at this time (Fig 1D). Kinetin was applied in the same conditions as previously described . Kinetin treatment fully restored the virulence of cks1 since the number and the size of lesions caused by cks1 were similar to those caused by the cks1CKS1 complemented strain (Fig 3). Similar results were obtained with an exogenous application of cis-zeatin (S6 Fig), which is the major cytokinin produced by M. oryzae (92% of CKs in supernatant; Table 1). These data strongly support the hypothesis that the reduced virulence of the cks1 mutant is directly due to a defect in CK production.
One day after inoculation with the cks1 mutant or the cks1CKS1 control strain, plants were treated with 50 μM of the CK compound, kinetin, or buffer alone. Kinetin alone without infection had no visible effect on leaf aspect. (A) The symptoms were observed 6 dpi. The number of lesion per leaf (B) and the size of lesions (C) were measured. The values represent the mean and SD of three biological replicates of 10 individuals. The entire experiment was repeated 3 times with similar results. The different letters indicate significant differences between values (p-value < 0.01) as estimated by a generalized linear model (B) or a mixed model (C) using ANOVA analysis (See Materials and Methods).
The cks1 mutant induces a stronger oxidative burst than control strains
To evaluate the effect of fungal-derived CKs on early host basal defense responses, we measured ROS accumulation. Compared to cks1CKS1, the accumulation of H2O2 (as revealed by DAB staining) was much more pronounced in response to the cks1 mutant (Fig 4A) and extended well beyond penetration sites (punctuated brown spots in complemented mutant controls) since the DAB staining was visible throughout all leaf tissues.
The relationship between virulence and reactive oxygen species accumulation was evaluated in the cks1-infected leaves. (A) The oxidative burst was detected 48 h after inoculation using DAB stain that turns brown upon reaction with H2O2. Brown spots correspond to sites where the wild-type blast fungus penetrated (see inlet), whereas a global browning was visible with infection with cks1 mutant. This experiment was repeated two times and gave similar results. (B, C) DPI, an NAD(P)H oxidase inhibitor partially restores the virulence of the cks1 mutant. One day after inoculation (once appressorium formation was initiated), plants were treated with DPI (0.5μM diluted in DMSO as previously described . The symptoms were observed 6 dpi (B) and the number of lesions per leaf was measured (C). The letters indicate significantly different values according to a generalized linear model and ANOVA analysis (p-value < 0.04), see Materials and Methods.
In order to test if the cks1 mutant was able to infect the host more efficiently if ROS production was impaired, we treated inoculated plants with DPI (Diphenylene Iodonium), an inhibitor of flavor-enzymes like the NAD(P)H oxidase involved in H2O2 production . The virulence of cks1 mutant was partially restored when plants were treated with DPI (Fig 4B and 4C). This suggests that the capacity of the cks1 mutant to invade the host cell is restored, although only partially, when ROS production is reduced.
The cks1 mutant induces an early and strong transcriptional response of rice defense markers
We measured the expression of some well-established rice defense-marker genes during infection with the different strains. It first appeared that the defense-marker genes were differentially expressed earlier (2 to 6 hpi; Fig 5A and S5B Fig) in plants infected with cks1—before fungal penetration that mostly occurs after 24h (Fig 1D). After penetration of the fungus (>24 hpi) and very strikingly, many defense-markers tested also showed a stronger induction in plants infected with cks1 as compared to those infected with the control cks1CKS1 (e.g. CHI, PBZ1, PR10 and PR5 in Fig 5B).
The transcriptional regulation of defense-marker genes was evaluated upon inoculation with cks1 mutant and complemented control strain. Nipponbare plants were inoculated with spore suspension (in gelatin 0.5%) of either the cks1 mutant or cks1CKS1 control strain. Gene expression (normalized by plant Actin gene) was measured at different times after inoculation. PBZ1 is a classical disease-related marker coding for a PR10 protein , PR5 and PR10 are classical disease-related markers , CHI and CHI7 are chitinases . (A) Gene expression was measured before the first indications of fungal penetration (< 6hpi). (B) Kinetin (50 μM) was applied (KIN) or not (mock) at 24 hpi and gene expression was also measured at 48 hpi. A t-test was used to compare the means between cks1 and cks1CKS1; for one given gene * indicate significant differences between cks1 and complemented strain; for (A) *, p-value < 0.04 and **, p-value<0.002 and for (B) *, pvalue<0.03; **, p-value<0.008. The values presented are the means calculated from four independent replicates. The experiments were repeated twice with similar results.
In the next step, we tested whether this over-induction of defense-markers by cks1 was affected when the cks1 virulence was complemented by exogenous application of the CK kinetin. This was the case for most genes that initially showed an over-induction (e.g. CHI, PBZ1 and PR5; Fig 5B). It indicates that complementation by exogenous cytokinin treatment reverts penetration and growth of the fungus as well as plant it brings back defense-marker expression to normal levels. Altogether these results support the hypothesis that the loss of virulence and the over-induction of defense associated with the cks1 mutation are linked to an absence of CK production by the mutant strain.
The virulence of the cks1 mutant can be restored in immuno-depressed plants
The enhanced oxidative burst and defense-genes expression is consistent with the reduced virulence phenotype observed with the cks1 mutant. This suggests that fungal CKs normally produced by M. oryzae may inhibit key plant defense reactions. To further support this hypothesis, we tested the capacity of the cks1 isolate to infect rice mutants which are defective for the chitin receptor, CEBiP, and the master transcriptional regulator NH1 that are both known to be immuno-depressed [57,59].
The fungal cks1 mutant was more virulent on cebip and nh1 rice mutants than on the wild-type plants (Fig 6). The restoration of virulence of the cks1 mutant on rice mutant plants impaired for defense responses strongly supports the hypothesis that the cks1 mutants have the capacity to infect rice as long as defenses are inhibited.
The cks1 strain is more virulent on rice mutants deficient for basal defenses. KO-cebip (A) and KO-nh1 (B) rice mutants and control plants (WT) , all in Nipponbare genetic background, were spray-inoculated with cks1. Symptoms were measured 6 days after inoculation on three replicates containing 6 plants. The values are the mean and SD from three biological replicates. A t-test was used to compare the percentage of susceptible lesions of cks1 on immune-deficient mutant and respective control plants, p-value < 0.02.
Elevated fertilization can restore virulence of the cks1 mutant
CKs affect different key metabolic processes in plants, like Calvin-Benson or tricarboxylic acid cycles and mediate source/sink modifications . Therefore, we hypothesized that the lack of virulence of the CK-deficient strain could be also due to a reduced capacity to exploit or drain nutrient resources. Under this hypothesis, we reasoned that high levels of fertilization could enhance cks1 virulence. Indeed high fertilization levels were previously shown to increase amino-acid and sugar contents in rice leaves  as well as rice blast susceptibility . The protocol described in  was used to fertilize plants 24h before inoculation with cks1 (Fig 7). Under high fertilization regime, plants were more susceptible to the cks1 mutant (Fig 7A), supporting the hypothesis that this mutant is able to infect rice tissues but likely requires external complementation with essential nutrients.
Plants were fertilized (high fertilization) or not (low fertilization) 24h before inoculation. Fertilization was done as in Ballini et al, 2013  to test the effect of plant nutritional status on cks1 virulence. (A) Symptoms 6 days after inoculation and the number of lesion per leaf in plants inoculated with cks1 mutant or cks1CKS1 under low or high nitrogen fertilization. Three biological replicates composed of 10 plants were analyzed per strain/condition. The different letters indicate significant differences between values (p-value < 0.03) as estimated by a t-test. (B) The expression of defense-marker genes was measured 48hpi and first normalized with Actin in Nipponbare plants inoculated with cks1CKS1 (white bars) or with cks1 (black bars), fertilized (High Nitrogen, HN) or not (Low Nitrogen, LN) 24h before inoculation. For each gene, the mean and the SD of relative expression obtained from 4 biological replicates (each of 3 plants) are presented. A t-test was done on raw data to compare relative expression in cks1 and cks1CKS1 inoculated plants. *, p-value<0.04; **, p-value<0.003; ***, p-value<0.0005.
We then evaluated if the over-induction of defense-markers, observed in plants infected with the cks1 mutant (Fig 5B), was still visible when the virulence was reversed by fertilization. Quite remarkably, the over-induction of defense-marker genes tested did not revert under high fertilization (Fig 7B) in cks1-infected plants, despite the fact that the virulence of this mutant strain was restored (Fig 7A). The data strongly support the idea that the over-induction of defense-markers by cks1 is not due to an arrest of growth itself (which would subsequently trigger enhanced plant defense) but rather to another defect that leads to reduced virulence, likely in CK production. The virulence of the cks1 mutant despite high expression of defense is likely compensated by high nutrient availability under high fertilization.
The cks1 mutation affects sugar and amino acid contents in and around rice infected tissues
Previous metabolomic analysis of plants infected by the rice blast fungus showed that some amino acids (glutamate and aspartate) as well as sugars (fructose and glucose) are nutrients which are drained towards the infection site . We therefore tested whether CKS1 is required to efficiently maintain nutrient levels and/or modify nutrient fluxes at and around the penetration site. In order to test this hypothesis, after local deposition of a droplet of spore suspension of the different strains on the leaf surface, we measured the levels of different sugars and amino acids in the blast-infected and the neighboring non-infected zones (Fig 8 and S7 Fig).
Metabolomic analysis of plants infected with cks1 and complemented mutant strains. Glucose, aspartate and glutamate contents were quantified (nmol/mg of fresh weight) as well as other sugars and amino acids, during infection (times are indicated), at the site of inoculation corresponding to the “infected zone” and one centimeter apart on both sides (upper and lower) with respect to the inoculated zone (S7 Fig). Only the lower “non-inoculated, neighboring part of the leaf” is shown but all data relative to the “upper non-infected zone” can be found in S7 Fig. For more details see also Materials and Methods. A t-test (*, p-value < 0.05) was used to compare amino acid contents in leaf fragments from plants inoculated with the cks1 (black bars) and cks1cks1 control complemented strain (white bars). The dashed arrows point to the significant changes (*t-test, p-value < 0.05) of amino-acid contents in cks1-infected plants between 24 and 48 hpi. For each time point four replicates composed of three leaf fragments were analyzed, mean and SD are indicated.
During the infection by the cksCKS1 control strain, glucose (Fig 8A) and fructose (S7 Fig) contents progressively increased in and around the infected zone, which is consistent with previous observations that the accumulation of sugar is associated with successful pathogen invasion . Initially, glucose and fructose contents were significantly higher 48 hpi in plants infected with the cks1 mutant compared to plants infected with the control strain, an observation that can be related to enhanced induction of defense (see Discussion). By contrast, at later time points (after 48hpi), soluble sugars contents were significantly lower in plants infected with the cks1 mutant. Given the known effects of CKs on maintaining photosynthesis active , this observation supports the idea that CKs produced by Magnaporthe could contribute to maintain sugar production during infection.
For most amino acids, there were no strong differences between the tissues infected by cks1 mutant and those infected by cks1CKS1 except for aspartate and glutamate (Fig 8A). Away from the infected zone, the concentration of these two amino acids transiently decreased at 48hpi in plants infected with the control strain and increased at the site of infection at 72 hpi, suggesting that these amino acids can be drained towards the infection site or accumulated. By contrast, their level remained almost stable during infection with the cks1 mutant. This suggests that the cks1 mutant strain is not able to drain or consume aspartate and glutamate as efficiently as the control strain during infection.
Identification of CKS1, a conserved gene required for fungal cytokinin production
The pathogenic fungus Magnaporthe oryzae produces and secretes CKs  but its biosynthesis pathway had remained unknown. Moreover, the involvement of CKs in virulence of pathogenic fungi that do not induce tumors was still undetermined. Recently, a cluster including two genes (including one coding for a IPT-LOG) involved in the de novo CK biosynthesis pathway, was characterized in the ergot fungus Claviceps purpurea . In mutants deleted for these two genes, CK production was partially affected but virulence was not. In the present study, we identified a gene in the rice blast fungus, CKS1, required for CK biosynthesis. The protein encoded by this gene presents all the features of a tRNA-IPT enzyme, the type of which is known in plants and yeast, and suspected in many fungi, to perform the first step of one of the CK biosynthesis pathways [25,40,45]. Phylogenetic analysis of tRNA-IPT protein sequences suggests that this gene is highly conserved among Ascomycete fungi (S8 Fig) and beyond [47,65]. Our work sets the basis for functional analysis of this pathway in several other plant associated fungi known to produce CKs [41,66,67].
We generated a cks1 mutant strain and demonstrated that this strain does not produce any of the CK types secreted by the wild type GY11 and cks1CKS1 complemented strain (Table 1). Moreover, the CKS1 protein is the only one found to contain an IPT domain in the rice blast fungal genome (S1 Table). These results suggest that the CK biosynthesis pathway controlled by CKS1 is probably the only one in the rice blast fungus. Nucleotide forms, which are known to be precursors to riboside, free base and glycosylated CKs , seem to be the major type of CKs secreted by control strains (cks1CKS1 and GY11). This contrasts with the yeast Δmod5 mutant which was found to still produce CKs, and suggested that tRNA turnover is not mainly involved in yeast CK production . However, the Δmod5 mutant was grown on medium composed of yeast extract that already contains hormonal compounds. Thus, the free CK production by yeast observed in that study could have come from the recycling of CK compounds provided by the medium.
Like the yeast Δmod5 mutant, the M. oryzae cks1 mutant had no obvious pleiotropic effects under standard growth conditions in minimal medium (S3 Fig). By contrast, the growth of cks1 was affected under oxidative stress and this could be reverted by exogenous CKs (S4 Fig). This result suggests that CKs play a role in fungal processes, like in yeast, for which MOD5 has primary roles in translation and is required for antifungal drug resistance . These processes may participate to the loss of virulence of the cks1 mutants (see below).
In plants, CK signaling is mediated by a multistep phosphorelay system involving Histidine Kinase Receptors, Histidine Phospho-transfer proteins and Response Regulators [53,69]. This kind of transduction system is widespread among organisms [70–73]. Several studies mentioned its involvement in osmoregulation in yeast and in hyphal growth of Neurospora crassa . Based on protein sequence homology, putative orthologous genes to those of the plant CK signaling pathway were found in the M. oryzae genome (S1 Table). Two of these genes, MoSLN1 (coding for an histidine kinase receptor) and MoSSK1 (coding for type-A response regulator), were previously shown to be required for full virulence of M. oryzae [75,76]. This suggests that CKs could also be perceived by the blast fungus in order to trigger a signal potentially required for its virulence. However, the involvement of these proteins in CKs perception and/or in CKs signaling transduction in response to plant or fungal-derived CKs remains to be established.
Cytokinins are required for fungal virulence
In addition to becoming deficient in CK production, the Magnaporthe cks1 mutant was less virulent than wild-type GY11 or cks1CKS1 control strains since its capacity for penetration and invasion was strongly impaired (Fig 1). Mutations affecting fungal invasion are still scarce in M. oryzae as most mutations affect appressorium formation and penetration only . Only very recent studies reported the role of protein effectors in virulence of Magnaporthe during host invasion . The characterization of the CK deficient cks1 strain demonstrates that CKs play a key role in virulence of the rice blast fungus. Testing whether CKs play similar roles in other pathogenic fungi is now possible since CKS1 homologs exist in most of them (S8 Fig).
An exogenous supply of kinetin (Fig 3) or cis-zeatin (S6 Fig) post inoculation restored the virulence of the CK-deficient strain, suggesting that the lack of virulence of these mutants was due to their inability to produce CKs. Moreover, exogenous application of kinetin reverted part of the over-induction of defense by the cks1 mutant (Fig 5B). This suggests that although other defects due to the deletion of the tRNA-IPT gene may exist, they are not responsible for this over-induction of defense. The demonstration that these fungal-derived CKs are secreted in planta is technically challenging because of the presence of plant CKs and the difficulty to localize such small metabolites. However the observation that the expression of plant CK responsive genes is differentially affected by the cks1 strain before penetration in the plant tissue (Figs 2 and 5A) suggests that the CKs produced by Magnaporthe are also secreted and detected by the plant. The plant receptors and pathways engaged in CK detection remain to be identified and CK mutants in rice will have to be produced to address this question.
Fungal cytokinins affect plant defense and nutrient fluxes and fungal physiology
Several observations support the hypothesis that the impaired virulence of cks1 is the consequence of an inability of the fungus to manipulate the plant defense pathways and metabolic fluxes rather than a consequence of a self-triggered growth arrest (caused by the cks1 loss-of-function) that would in turn trigger enhanced defense. First, enhanced defense is already visible before fungal penetration (Fig 5A and S5B Fig) at a time where the cks1 and control strains were indistinguishable in terms of growth (S4C Fig). Second, reduced fungal growth could be partially restored by manipulating the plant in three ways: (i) by reducing plant defense chemically (Fig 4) (ii) by impairing immunity genetically in two independent mutants (Fig 6), (iii) by modifying plant fertilization (Fig 7A). Third, the restoration of virulence under high fertilization (Fig 7A) did not affect the over-induction of defense responses (Fig 7B). This indicates that the arrest of fungal growth can be compensated by high N-fertilization, although the associated over-induction of defense response cannot, presumably because CKs are still not produced. For these reasons, we propose that the impaired virulence of the cks1 mutant is the consequence of the absence of CK production that would normally dampen defenses and modify nutrient fluxes for the pathogen’s benefit. Since the cks1 mutant is also more susceptible to oxidative stress in vitro (S4 Fig), this may further reduce its global capacity to grow into plant tissues, especially if the plant ROS production is enhanced (Fig 4A). In that sense the cks1 mutant is similar to the des1 mutant which is required for dampening ROS-mediated plant defense .
Possible effects of fungal CK on the plant metabolism
The transient and elevated sugar content in cks1 compared to the cks1CKS1 strain at 48 hpi (Fig 8A) is consistent with previous studies in other pathosystems that showed that an early and strong accumulation of soluble sugars, providing energy for the establishment of host defenses [63,79,80]. By contrast, the lower contents of glucose, fructose and sucrose found after 72 hpi with the cks1 strain suggests that photosynthesis is reduced in plants infected by the CK-deficient strain. We propose that Magnaporthe-derived CKs contribute to prevention of photosynthesis breakdown during infection process, for instance by limiting oxidative stress generated through photorespiration and, in consequence, allowing the establishment of the biotrophic phase. Among the different amino acids quantified, the contents of two key amino acids (glutamate and aspartate), which are known to be essential for M. oryzae , were differently affected between mutant and control strains following infection (Fig 8A). In Arabidopsis, CKs were described to alter transcription of genes like glutamate dehydrogenase, asparagine synthetase and aspartate aminotransferase . The fungal-derived CKs could also alter transcription of these genes to re-channel these amino acids towards fungal hyphae. Furthermore, CKs were previously shown to modify amino acid uptake through fungal cell membranes ; therefore, cks1 mutants could also be affected in their capacity to import these molecules into fungal hyphae. Altogether, these possible effects of CKs could explain why aspartate and glutamate contents were differently affected during infection by cks1 strain and the complemented cks1CKS1 strain. These results suggest that fungal CKs could be involved in pathogen nutrition during infection as hypothesized for many plant/fungi interactions by Greene (1980) .
Our results also suggest that the lack of virulence of cks1 mutants is partially due to an inability to limit plant defense responses like the oxidative burst (Fig 4) and transcription of defense-related genes (Fig 5A). The enhanced oxidative burst could lead to the stronger induction of defense markers observed during mutant infection and could participate to the strengthening of the cell wall to limit fungal penetration [83,84]. This is consistent with the ROS scavenging activity of CKs demonstrated in transgenic tobacco by Pogány et al., (2004). Quite paradoxically, exogenously applied CKs have been shown to enhance, in combination with salicylic acid, rice defense marker genes expression and phytoalexin biosynthesis [43,86]. This synergistic effect depends on key defense transcriptional regulators like OsWRKY45, a pivotal factor in biotic and abiotic stress responses . Similarly, in Arabidopsis, specific recognition of bacterial CKs by plant CK receptors leads to a stronger induction of plant defenses and host resistance, involving plant CK Response Regulators and the transcription factor TGA3 . In this context, how the CKs produced by Magnaporthe oryzae can act as negative regulators of defenses remains to be elucidated. A tight temporal and spatial production of CKs by the blast fungus could be central to avoiding enhanced activation of defenses.
CKs play a key role in plant-microorganism communication  and it seems to be particularly true in plant symbiotic relations with fungi and bacteria [90,91]. Our work, showing that Magnaporthe requires CK production to be fully virulent, extends the key role that these hormones play in the interactions between plants and microbes to pathogenic fungi that do not trigger organ deformation. Our work suggests that CKs produced by M. oryzae act like classical effectors during host invasion as they significantly reduce defenses (e.g. ). Moreover, our study shows that fungal CKs can divert and attract plant nutrients essential for fungal growth, much like the TAL effectors from bacteria . Therefore, CKs, like other hormones, could represent metabolic effectors with several biological functions. Given their central role in several metabolic processes, plant hormones represent ideal factors to be acquired as effectors by pathogens. Accordingly, evolution probably led the rice blast fungus to include CKs into its weapons for successful colonization of rice plants.
Materials and Methods
In silico analysis on M. oryzae and protein modelling
Based on published studies on plant CK metabolisms (cited in S1 Table), sequences from plant proteins were used to perform BLASTp on the Magnaporthe proteome at http://www.broadinstitute.org/annotation/genome/magnaporthe_grisea/Blast.html?sp=Sblastp. We used an E-value of 1e-3, with the comparison Matrix BLOSUM62 and gapped alignment. Similarly, BLASTp were performed on the yeast proteome available at http://www.yeastgenome.org/cgi-bin/blast-sgd.pl with default parameters. Hit proteins from yeast and M. oryzae were used to BLASTp back on Magnaporthe proteome to ensure Best Blast Mutual Hits were identified.
Protein structure predictions were realized with the on-line platform I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER/). The primary sequences of proteins of interest were submitted (At2g27760 and MGG_04857) and secondary structures were predicted. Based on secondary structure predictions and primary sequences, the most similar proteins whose 3D structure was determined by NMR or X-ray crystallography were used as template for the model prediction. The model presented in S1B Fig was obtained with the MOD5 yeast protein (PDB accessions: 3epjA and 3ephA) as a template. Similar models were obtained with 3foz (E. coli) and 3a8t (Humulus lupulus) proteins as templates. Afterwards, alignments between the protein of interest and templates were generated by the use of different threading programs (MUSTER, FFAS-3D, SPARKS-X, HHSEARCH I, Neff-PPAS, HHSEARCH, pGenTHREADER, wdPPAS, PROSPECT2). Finally, the 10 best alignments were used to generate five structural models characterized . The quality of structures used as templates were evaluated by Qmean server as well as the models obtained (http://swissmodel.expasy.org/qmean/cgi/index.cgi).
Statistical analysis and experimental design
We analyzed results obtained for symptom quantification (Figs 1B, 3B and 4C) using a generalized linear model with a quasi-Poisson error structure. Significance was determined using a Chi² test. In each experiment, three biological replicates composed of 10 plants were analyzed per strain/condition. The size of the lesion trait was analyzed using a mixed model, with the “leaf” factor as random error, to compare the invasion of the different strains in the different conditions (Figs 1C and 3C). Boxplots represent data distribution using the median (indicated by the black line) and approximate quartiles. Each experiment was replicated at least three times. Gene expression data were analyzed using a Student t-test on four biological replicates, with each replicate composed of five to six plants (Figs 2, 5, 7B and S5). A student t-test was also used to analyze data presented in figs 1D, 6, 7A, 8, S3, S4, S6 and S7.
Transformation was performed as described by Ribot et al., (2013). Protoplasts of GY11 were prepared as described previously . For the knock-out cks1 mutant, 1.2kb upstream and 1.2 kb downstream regions of the gene of interest were amplified by PCR using genomic GY11 DNA (100 ng) as template. The strategy used for constructing the gene replacement cassettes is derived from Kämper (2004)  and presented in the S2 Fig. Primers used are shown in S2 Table. Growing colonies were transferred to Tanaka-Hygromycin or Tanaka-Basta plates for assessing resistance. Resistant colonies were further grown on rice agar media for 7–10 days at 26°C, purified by single-spore isolation, tested for resistance, and stored at -20°C. At least 2 independent transgenic fungal lines were isolated for each plasmid. Resistant colonies were characterized by PCR using a Phire Plant Direct PCR kit (Thermo Scientific, Waltham, MA, USA).
Measurement of cytokinin content
Fungal isolates were grown in 50mL of minimal Tanaka liquid medium without yeast extract (10g/L Glucose, 2 g/L NaNO3, 2 g/L KH2PO4, 0.5 g/L MgSO4-7H2O, 0.1 g/L CaCl2-2H2O, 4 mg/L FeSO4-7H2O, 1mg/L Thiamine, 5μg Biotin and microelements as Tanaka-B medium ) on rotary shaker for 10 days at 26°C. Yeast extract was excluded because this compound already contains hormones, including CKs, which leads to misinterpretation between CKs really produced compared to those that are just taken from the media and metabolized by the fungus. Fungal cultures were centrifuged for 10 min at 2000×g, the supernatant was collected and quantified. Mycelia were rinsed with sterile Tanaka liquid medium, centrifuged for 10 min at 2000×g and pressed with absorbent paper to accurately quantify the fungal biomass. Samples were then frozen and lyophilized. CK extraction and measurements were performed as previously described [19,50].
Fungal and plant growth
Fungal isolates were grown on rice flour agar for spore production . For the determination of interaction phenotypes and cytology analysis, a suspension of fungal conidiospores (5×104 sp/mL) was spray-inoculated on the leaves of 3-week-old plants. For gene expression analysis, an inoculum of 2×105 sp/mL was used. Radial mycelial growth was measured on minimal Tanaka solid medium (20g/L agar added to the recipe mentioned above), during 13 days at 26°C. Nipponbare plants (O. sativa ssp. japonica) were grown during three weeks as described previously . In standard conditions, nitrogen fertilization was performed for three weeks and inoculation was done 4 days after fertilization. For the high/low nitrogen experiments, plants were fertilized for two weeks as in standard experiment, and in the third week, plants were fertilized (or not) one day before infection, as described in .
Measurements of amino acids and sugar content
Amino acids and sugar contents were quantified as described in Gravot et al., (2010)  in leaf tissue of plants locally inoculated with a drop (15μL) of inoculum at 20 000 sp/mL. One centimeter corresponding to the inoculation site, and one centimeter above and below was sampled to quantify amino acids on 4 replicates composed of three leaf fragments (see S7 Fig for a picture of the experimental setting).
RNA extraction and quantitative RT–PCR analysis
RNA extraction was performed as mentioned in Delteil et al, 2012 . Quantitative PCR was performed using LC 480 SYBR Green I Master Mix (Roche, Basel, Switzerland) and a LightCycler 480 instrument (Roche). Amplification was performed as follows: 95°C for 10 min; 40 cycles of 95°C for 15 s, 60°C for 20 s and 72°C for 30 s; then 95°C for 5 min and 40°C for 30 s. In Figs 5 and 7, in order to compare genes with different expression levels, for a given gene, all values were normalized using the average value of this gene across the different conditions.
Cytokinin and DPI treatments
Kinetin (Sigma) and cis-zeatin (OlChemIm) were diluted in 50% ethanol to prepare a stock solution of 50 mM. The solution sprayed contained 50 μM of CKs (replaced by ethanol 50% in mock solution), Tween at 0.02% final, diluted in water. Kinetin was applied to plants as described before  with slight modifications. Diphenylene iodonium (DPI; SIGMA D23926) treatment was performed at 0.5 μM final concentration, prepared from a 50× stock solution diluted in 50% DMSO. Mock treatment was performed with the equivalent volume of 50% DMSO/water. The volume sprayed was calculated to saturate the leaf surface (5mL for 10 three week-old plants).
S1 Table. Putative genes involved in CK metabolisms and signaling pathways, in Oryza sativa, Arabidopsis thaliana, Magnaporthe oryzae and Saccharomyces cerevisiae.
CK biosynthesis in rice and Arabidopsis is well described. Based on homology between primary protein sequences some putative orthologous could be identified in M. oryzae and yeast by BLASTp. By the same way, primary sequences of plant enzymes already described to be involved in degradation of CKs were used to find orthologous in fungi. Finally, in plants, CK transduction signaling is performed by a multistep phosphorelay system, involving Histidine Kinase (HK) receptors, Histidine Phosphotransfer Proteins (HPt) and Response Regulators (only RR-A and RR-B are shown). Multistep phosphorelay is also already described in yeast and is conserved in M. oryzae. IPT: Isopentenyl Transferase; tRNA-IPT: tRNA-Isopentenyl Transferase; LOG: Lonely Guy; CYP735A: Cytochrome P450 monooxygenase; AK: Adenosine Kinase; CK-N-GT: Cytokinin N-Glucosyltransferase; ZOGT: Zeatin-O-glucosyltransferase; CKX: Cytokinin Oxydase. a BLASTp with AK from A. thaliana; b identified by BLASTp from yeast and plant enzymes; c found annotated in yeast genome and publications associated (www.yeastgenome.org)
S2 Table. Primers used for RT-qPCR gene expression studies and for PCR genotyping.
MoCKS1: Magnaporthe oryzae Cytokinin Synthesis 1; OsRR1, 2, 3, 6, 10 corresponding to the annotation of response regulator coding genes published by Pareek et al., (2006) . PBZ1: Probenazole-inducible gene; CHI and CHI7 coding for chitinases; PR: Pathogenesis Related genes. The primers 1F, 2R, 1R, 3F and 3R are those used in S2 Fig
S1 Fig. The MGG_04857 protein is similar to a tRNA-Isopentenyl transferase.
(A) Primary sequence alignment of MOD5, AtIPT2 and the orthologous protein from M. oryzae. MOD5 and AtIPT2 are tRNA-isopentenyl transferases involved in CK biosynthesis in yeast  and Arabidopsis , respectively. Binding sites (underlined) are conserved in the putative tRNA-IPT from M. oryzae (MGG_04857, CKS1). The ATP binding site, from the amino acid 18 to 26 (GSTGTGKS), is conserved as well as tRNA binding sites (DAMQ 43–46 and T111) and DMAPP binding site (249–265). The percentage of identity and similarity between these proteins is higher than 30% and 50%, respectively (between MOD5 and AtIPT2: 32% identities, 55% positives (e-value 7.1e-27); MOD5 and putative tRNA-IPT from M. oryzae: 36% identities, 60% positives (8.6e-52); AtIPT2 and putative tRNA-IPT from M. oryzae: 36% identities, 54% positives (4e-40). (B) The putative tRNA-IPT from Magnaporthe has a predicted structure similar to the yeast MOD5 protein. MOD5 structure (light blue) was determined by X-ray diffraction . The MOD5 structure used for the alignment corresponds to the 3ephA and 3epjA accessions in the Protein Data Bank (http://www.rcsb.org/pdb/home/home.do). The predicted structural model of tRNA-IPT from M. oryzae (MGG_04857; dark blue) was obtained by threading on the on-line platform I-TASSER, C-score 0,02 (Cf Materials and Methods). The quality scores of the structural model obtained for MGG_04857 and MOD5 by the Qmean server were 0.610 and 0.656 respectively. Conserved substrate binding sites are indicated by red arrows.
S2 Fig. Characterization of fungal cks1 mutants and complemented strains.
(A) Genomic structure of MoCKS1 (MGG_04857) gene. The cks1 strain was generated by homologous recombination between the endogenous CKS1 gene and a PCR fragment containing the hygromycin resistance gene. The knock-out mutation was complemented by a construct containing the genomic sequence of MoCKS1 under its own promoter thus corresponding to the cks1CKS1 control isolate (genomic sequence 1100nt upstream and 1646nt downstream the ATG). The position of primers used for genotyping and the length of PCR products are indicated. Genotyping was established by PCR (1.2% agarose gel). The genotype of the strains is indicated and * corresponds to negative controls. (B) PCR products were obtained with primers 1F-1R showed in (A) on both sides of the insertion site. (C) PCR products obtained with primers 1F-2R, demonstrate the presence of the hygromycin resistance gene replacing CKS1 endogenous gene. (D) PCR products obtained with primers 3F-3R showing the presence of MoCKS1 genomic sequence under its own promoter. (E) In vitro relative expression of MoCKS1, MGG_04857, in cks1 strains (cks1 n°3, cks1 n°4), potential complemented strains cks1CKS1 (cks1CKS1 4–1 and cks1CKS1 4–4; generated from cks1 n°4 strain) and WT strain (GY11). The expression of MoCKS1 was normalized with the expression of the MG4 constitutive gene. The cks1 n°3 and 4 strains do not express MoCKS1. We chose cks1 n°4 as mutant strain and cks1CKS1 4–4 as control complemented strain. In all experiments presented, these strains were named cks1 and cks1CKS1. GY11 corresponds to the wild type genetic background used to generate fungal mutants.
S3 Fig. Early developmental of cks1 is not modified in vitro or on the plant surface.
(A) The mycelial growth of the different strains was initiated from a fungal disc of 1cm of diameter from a first plate where the fungi reached maximum growth. The diameter of mycelia was measured during 13 days on minimal medium. The values are the mean and SD of 5 replicates per strain. (B) The development of the appressorium was measured on glass slides at the indicated time points for each strain and (C) on the plant leaf surface. Plants were inoculated and the frequency of spores showing complete appressorial development was measured. There was no significant difference between cks1 (black bars) and cks1CKS1 (white bars) strains as estimated with a t-test.
S4 Fig. The cks1 mutant is less tolerant to oxidative stress.
The cks1 mutant and complemented strain were grown on minimal medium containing H2O2 and radial growth was measured as a read-out of fungal fitness. (A) The cks1 mutant is hypersensitive to oxidative stress. (*: P<0.001; t-test comparing cks1 and complemented mutant strains of 5 replicates). (B) The effect of 1 mM H2O2 was tested in the presence of 25μM of kinetin.
S5 Fig. The cks1 mutant triggers different host CK signaling and defense transcriptional responses during early stages of infection.
(A) The transcriptional regulation of CK marker genes (OsRR2 (Os12g04500), OsRR3 (Os01g72330) and OsRR10 (Os02g35180) as named by Pareek et al., (2006)  was evaluated by quantitative RT-PCR using the Actin gene for normalization. (B) The expression of defense marker genes was also measured: Os01g03390 (BB trypsin inhibitor), Os11g37970 (HEL protein), Os07g48020 (Peroxidase). Nipponbare plants were inoculated with spore suspension (in gelatin 0.5%) of either the cks1 mutant (black bars) or cks1CKS1 control strain (white bars) and gene expression was measured at 2, 4 and 6 hours post inoculation (hpi), before penetration of the leaf tissues. The values presented are the Log2 ratios (infected/not infected) of the means calculated from four independent replicates. Uninfected plants were sprayed with gelatin 0.5% but without spore suspension. This experiment was repeated twice and showed similar results. A t-test was used to compare the means of expression quantified in cks1 (black bars) and cks1CKS1 (white bars) inoculated plants. *: p-value < 0.05; **: p-value < 0.03; ***: p-value < 0.001.
S6 Fig. The reduced virulence of cks1 strain is restored by exogenous application of cis-zeatin.
Plants were treated with 50μM of cis-zeatin or buffer alone 24h after inoculation with cks1 mutant and cks1CKS1 complemented strain. The symptoms were observed 6dpi (A) and the number of lesion per leaf is shown (B). The values represent the mean and SD of three biological replicates of 10 individuals. The different letters indicate significant differences between values obtained by t-test (p-value<0.04).
S7 Fig. Accumulation of amino acids at the infected site and around.
Sugar and amino acid contents were quantified (presented in nmol/mg of fresh weight), during infection (times are indicated), at the site of inoculation corresponding to the”infected zone” and one centimeter apart (respectively named “lower” and “upper non-infected zones”). For more details see Materials and Methods. A T-test (*, p-value < 0.05) was used to compare amino acid contents in leaf fragment from plants inoculated with the cks1 (black bars) and cks1CKS1 (white bars).
S8 Fig. The tRNA-IPT protein is highly conserved across Ascomycetes.
Phylogenetic tree based on the primary sequence of putative orthologous tRNA-IPT proteins identified by BLASTp in Ascomycetes. Multiple protein sequences alignment was performed with MUSCLE and alignment curation with Gblocks. The phylogenetic tree was obtained using PhyML . The phylogenetic tree was rooted with the farthest species, Pyronema omphalodes. The different classes of Ascomycetes and bootstrap values are indicated.
We particularly thank Sophien Kamoun for his scientific advice, Henri Adreit and Romain Gallet for their help for spore isolation and statistical analysis respectively, and Marc-Henri Lebrun and Florian Frugier for fruitful discussions. We thank Nathalie Marnet and Solenne Berardocco, from the IGEPP's Metabolic Profiling and Metabolomic Platform (P2M2; Rennes-France), for their technical support in amino acid and sugar measurements.
Conceived and designed the experiments: JBM EC. Performed the experiments: EC VC AK AD CRM JBM. Analyzed the data: EC AG TK NRJE AK JBM. Contributed reagents/materials/analysis tools: EC AK AG VC. Wrote the paper: EC JBM TK AK NRJE AG.
- 1. Kamoun S. (2007) Groovy times: filamentous pathogen effectors revealed. Curr. Opin. Plant Biol. 10, 358–365 pmid:17611143
- 2. Macho A.P. and Zipfel C. (2015) Targeting of plant pattern recognition receptor- triggered immunity by bacterial type-III secretion system effectors. Curr. Opin. Microbiol. 23, 14–22 pmid:25461568
- 3. Verdier V. et al. (2012) Transcription activator-like (TAL) effectors targeting OsSWEET genes enhance virulence on diverse rice (Oryza sativa) varieties when expressed individually in a TAL effector-deficient strain of Xanthomonas oryzae. New Phytol. 196, 1197–1207 pmid:23078195
- 4. Hogenhout S.A. et al. (2009) Emerging Concepts in Effector Biology of Plant-Associated Organisms. Mol. Plant-Microbe Interact. 22, 115–122 pmid:19132864
- 5. Zipfel C. (2009) Early molecular events in PAMP-triggered immunity. Curr. Opin. Plant Biol. 12, 414–20 pmid:19608450
- 6. Mentlak T. et al. (2012) Effector-mediated suppression of chitin-triggered immunity by Magnaporthe oryzae is necessary for rice blast disease. Plant Cell 24, 322–35 pmid:22267486
- 7. Pitzschke A. et al. (2009) MAPK cascade signalling networks in plant defence. Curr. Opin. Plant Biol. 12, 421–426 pmid:19608449
- 8. Robert-Seilaniantz A. et al. (2007) Pathological hormone imbalances. Curr. Opin. Plant Biol. 10, 372–9 pmid:17646123
- 9. Chen Z. et al. (2007) Pseudomonas syringae type III effector AvrRpt2 alters Arabidopsis thaliana auxin physiology. Proc. Natl. Acad. Sci. U. S. A. 104, 20131–20136 pmid:18056646
- 10. Thatcher L.F. et al. (2009) Fusarium oxysporum hijacks COI1-mediated jasmonate signaling to promote disease development in Arabidopsis. Plant J. 58, 927–939 pmid:19220788
- 11. Verhage A. et al. (2010) Plant immunity: it’s the hormones talking, but what do they say? Plant Physiol. 154, 536–40 pmid:20921180
- 12. Djamei A. et al. (2011) Metabolic priming by a secreted fungal effector. Nature 478, 395–8 pmid:21976020
- 13. Patkar R.N. et al. (2015) A fungal monooxygenase-derived jasmonate attenuates host innate immunity. Nat. Chem. Biol.
- 14. Melotto M. et al. (2008) Role of stomata in plant innate immunity and foliar bacterial diseases. Annu. Rev. Phytopathol. 46, 101–122 pmid:18422426
- 15. Uppalapati S.R. et al. (2007) The phytotoxin coronatine contributes to pathogen fitness and is required for suppression of salicylic acid accumulation in tomato inoculated with Pseudomonas syringae pv. tomato DC3000. Mol. Plant. Microbe. Interact. 20, 955–965 pmid:17722699
- 16. Greene E.M. (1980) Cytokinin production by microorganismes. Bot. Rev. 46, 25–74
- 17. Gruen H.E. (1959) Auxins and Fungi. Annu. Rev. Plant Physiol. 10, 405–440
- 18. Hedden P. et al. (2001) Gibberellin Biosynthesis in Plants and Fungi: A Case of Convergent Evolution? J. Plant Growth Regul. 20, 319–331 pmid:11986758
- 19. Kisiala A. et al. (2013) Bioactive cytokinins are selectively secreted by Sinorhizobium meliloti nodulating and nonnodulating strains. Mol. Plant. Microbe. Interact. 26, 1225–31 pmid:24001254
- 20. Özcan B. and Topçuoglu S.F. (2001) GA 3, ABA and Cytokinin Production by Lentinus tigrinus and Laetiporus sulphureus Fungi Cultured in the Medium of Olive Oil Mill Waste *. Turk J Biol 25, 453–462
- 21. Strzelczyk E. et al. (1994) Cytokinin-like substances and ethylene production by Azospirillum in media with different carbon sources. Microbiol. Res. 149, 55–60
- 22. Strzelczyk E. and Pokojska-Burdziej A. (1984) Production of auxins and gibberellin-like substances by mycorrhyzal fungi, bacteria and actinomycetes isolated from soil and the mycorrhizosphere. Plant Soil 81, 185–194
- 23. Dihanich M.E. et al. (1987) Isolation and characterization of MOD5, a gene required for isopentenylation of cytoplasmic and mitochondrial tRNAs of Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 177–184 pmid:3031456
- 24. Golovko A. et al. (2002) Identification of a tRNA isopentenyltransferase gene from Arabidopsis thaliana. Plant Mol. Biol. 49, 161–169 pmid:11999372
- 25. Miyawaki K. et al. (2006) Roles of Arabidopsis ATP/ADP isopentenyltransferases and tRNA isopentenyltransferases in cytokinin biosynthesis. Proc. Natl. Acad. Sci.
- 26. Riou-Khamlichi C. et al. (1999) Cytokinin Activation of Arabidopsis Cell Division Through a D-Type Cyclin. Science (80-.). 283, 1541–1545
- 27. Carimi F. et al. (2003) Cytokinins: new apoptotic inducers in plants. Planta 216, 413–421 pmid:12520332
- 28. Fosket D.E. and Torrey J.G. (1969) Hormonal control of cell proliferation and xylem differentiation in cultured tissues of Glycine max var. Biloxi. Plant Physiol. 44, 871–880 pmid:5816361
- 29. Peleg Z. et al. (2011) Cytokinin-mediated source/sink modifications improve drought tolerance and increase grain yield in rice under water-stress. Plant Biotechnol. J. 9, 747–758 pmid:21284800
- 30. Vescovi M. et al. (2012) Programmed cell death induced by high levels of cytokinin in Arabidopsis cultured cells is mediated by the cytokinin receptor CRE1/AHK4. J. Exp. Bot. 63, 2825–2832 pmid:22312114
- 31. Morrison E.N. et al. (2015) Phytohormone Involvement in the Ustilago maydis–Zea mays Pathosystem: Relationships between Abscisic Acid and Cytokinin Levels and Strain Virulence in Infected Cob Tissue. PLoS One 10, e0130945 pmid:26107181
- 32. Hinsch J. et al. (2015) De novo biosynthesis of cytokinins in the biotrophic fungus Claviceps purpurea. Environ. Microbiol. 17, 2935–2951 pmid:25753486
- 33. Devos S. et al. (2006) A hormone and proteome approach to picturing the initial metabolic events during Plasmodiophora brassicae infection on Arabidopsis. Mol. Plant. Microbe. Interact. 19, 1431–1443 pmid:17153927
- 34. Pertry I. et al. (2009) Identification of Rhodococcus fascians cytokinins and their modus operandi to reshape the plant. Proc. Natl. Acad. Sci. U. S. A. 106, 929–34 pmid:19129491
- 35. Wingler A. et al. (1998) Regulation of Leaf Senescence by Cytokinin, Sugars, and Light. Plant Physiol. 116, 329–335
- 36. Walters D.R. and McRoberts N. (2006) Plants and biotrophs: a pivotal role for cytokinins? Trends Plant Sci. 11, 581–586 pmid:17092762
- 37. Angra-Sharma R. and Sharma D.K. (1999) Cytokinins in pathogenesis and disease resistance of Pyrenophora teres-barley and Dreschslera maydis-maize interactions during early stages of infection. Mycopathologia 148, 87–95 pmid:11189749
- 38. Behr M. et al. (2012) Remodeling of cytokinin metabolism at infection sites of Colletotrichum graminicola on maize leaves. Mol. Plant. Microbe. Interact. 25, 1073–82 pmid:22746825
- 39. Angra R. and Mandahar C.L. (1991) Pathogenesis of barley leaves by Helminthosporium teres I: Green island formation and the possible involvement of cytokinins. Mycopathologia 114, 21–27
- 40. Morrison E.N. et al. (2015) Detection of phytohormones in temperate forest fungi predicts consistent abscisic acid production and a common pathway for cytokinin biosynthesis. Mycologia 107, 245–257 pmid:25572099
- 41. Murphy A.M. et al. (1997) Comparison of cytokinin production in vitro by Pyrenopeziza brassicae with other plant pathogens. Physiol. Mol. Plant Pathol. 50, 53–65
- 42. Hu G.G. and Rijkenberg E.H.J. (1998) Ultrastructural localization of cytokinins in Puccinia recondita f.sp. tritici-infected wheat leaves. Physiol. Mol. Plant Pathol. 52, 79–94
- 43. Jiang C.-J. et al. (2013) Cytokinins act synergistically with salicylic acid to activate defense gene expression in rice. Mol. Plant. Microbe. Interact. 26, 287–96 pmid:23234404
- 44. Sakakibara H. (2006) Cytokinins: activity, biosynthesis, and translocation. Annu. Rev. Plant Biol. 57, 431–49 pmid:16669769
- 45. Armstrong D.J. et al. (1969) Cytokinins: distribution in species of yeast transfer RNA. Biochemistry 63, 504–511
- 46. Zhang Y. (2008) I-TASSER server for protein 3D structure prediction. BMC Bioinformatics 9, 40 pmid:18215316
- 47. Golovko A. (2000) Cloning of a human tRNA isopentenyl transferase. Gene 258, 85–93 pmid:11111046
- 48. Pai D.A. et al. (2014) Correction for Pratt-Hyatt et al., Mod5 protein binds to tRNA gene complexes and affects local transcriptional silencing. Proc. Natl. Acad. Sci. 111, 2397–2397
- 49. Suzuki G. et al. (2012) A yeast prion, Mod5, promotes acquired drug resistance and cell survival under environmental stress. Science 336, 355–9 pmid:22517861
- 50. Farrow S.C. and Emery R.J.N. (2012) Concurrent profiling of indole-3-acetic acid, abscisic acid, and cytokinins and structurally related purines by high-performance-liquid- chromatography tandem electrospray mass spectrometry Concurrent profiling of indole-3-acetic acid, abscisic acid,. Plant Methods 8, 42 pmid:23061971
- 51. Brandstatter I. and Kieber J.J. (1998) Two Genes with Similarity to Bacterial Response Regulators Are Rapidly and Specifically Induced by Cytokinin in Arabidopsis. Plant Cell 10, 1009–1019 pmid:9634588
- 52. Jain M. et al. (2006) Molecular characterization and differential expression of cytokinin-responsive type-A response regulators in rice (Oryza sativa). BMC Plant Biol. 6, 1 pmid:16472405
- 53. Pareek A. et al. (2006) Whole-genome analysis of Oryza sativa reveals similar architecture of two-component signaling machinery with Arabidopsis. Plant Physiol. 142, 380–97 pmid:16891544
- 54. Chi M.-H. et al. (2009) A novel pathogenicity gene is required in the rice blast fungus to suppress the basal defenses of the host. PLoS Pathog. 5, e1000401 pmid:19390617
- 55. Desikan R. et al. (1996) Generation of active oxygen in elicited cells of Arabidopsis thaliana is mediated by a N A D P H oxidase-like enzyme. FEBS Lett. 382, 213–217 pmid:8612756
- 56. Nakashita H. et al. (2001) Characterization of PBZ1, a probenazole-inducible gene, in suspension-cultured rice cells. Biosci. Biotechnol. Biochem. 65, 205–208 pmid:11272832
- 57. Delteil A. et al. (2012) Building a mutant resource for the study of disease resistance in rice reveals the pivotal role of several genes involved in defence. Mol. Plant Pathol. 13, 72–82 pmid:21726398
- 58. Kaku H. et al. (2006) Plant cells recognize chitin fragments for defense signaling through a plasma membrane receptor. Proc. Natl. Acad. Sci. 103, 11086–11091 pmid:16829581
- 59. Kishimoto K. et al. (2010) Perception of the chitin oligosaccharides contributes to disease resistance to blast fungus Magnaporthe oryzae in rice. Plant J. 64, 343–54 pmid:21070413
- 60. Otani Y. (1959) Studies on the relation between the principal components of the rice plant and its susceptibility to the blast disease. Annu. Phytopathol. Soc. Japan 16, 97–102
- 61. Ballini E. et al. (2013) Diversity and genetics of nitrogen-induced susceptibility to the blast fungus in rice and wheat. Rice 6, 32 pmid:24280346
- 62. Parker D. et al. (2009) Metabolomic analysis reveals a common pattern of metabolic re-programming during invasion of three host plant species by Magnaporthe grisea. Plant J. 59, 723–37 pmid:19453445
- 63. Kocal N. et al. (2008) Cell wall-bound invertase limits sucrose export and is involved in symptom development and inhibition of photosynthesis during compatible interaction between tomato and Xanthomonas campestris pv vesicatoria. Plant Physiol. 148, 1523–1536 pmid:18784281
- 64. Rivero R.M. et al. (2009) Cytokinin-dependent photorespiration and the protection of photosynthesis during water deficit. Plant Physiol. 150, 1530–40 pmid:19411371
- 65. Persson B.C. et al. (1994) Synthesis and function of isopentenyl adenosine derivatives in tRNA. Biochimie 76, 1152–1160 pmid:7748950
- 66. Vizarova G. (1975) Contribution to the Study of Cytokinin Production by Phytopathogenic Fungi. Biol. Plant. 17, 380–382
- 67. Ng P.P. et al. (1982) Cytokinin production by ectomycorrhizal fungi. New Phytol. 91, 57–62
- 68. Laten H.M. and Zahareas-doktor S. (1985) Presence and source of free isopentenyladenosine in yeasts. Proc. Natl. Acad. Sci. U. S. A. 82, 1113–1115 pmid:3883351
- 69. Tsai Y.-C. et al. (2012) Characterization of genes involved in cytokinin signaling and metabolism from rice. Plant Physiol. 158, 1666–84 pmid:22383541
- 70. Chang C. and Stewart R.C. (1998) The two-component system. Regulation of diverse signaling pathways in prokaryotes and eukaryotes. Plant Physiol. 117, 723–731 pmid:9662515
- 71. Lohrmann J. and Harter K. (2002) Plant two-component signaling systems and the role of response regulators. Plant Physiol. 128, 363–369 pmid:11842140
- 72. Saito H. et al. (2001) Histidine Phosphorylation and Two-Component Signaling in Eukaryotic Cells. Chem. Rev. 101, 2497–2509 pmid:11749385
- 73. Stock A.M. et al. (2000) Two-Component Signal Transduction. Annu. Rev. Biochem. 69, 183–215 pmid:10966457
- 74. Alex L.A. et al. (1996) Hyphal development in Neurospora crassa: Involvement of a two-component histidine kinase. Proc. Natl. Acad. Sci. U. S. A. 93, 3416–3421 pmid:8622950
- 75. Motoyama T. et al. (2008) Involvement of putative response regulator genes of the rice blast fungus Magnaporthe oryzae in osmotic stress response, fungicide action, and pathogenicity. Curr. Genet. 54, 185–195 pmid:18726099
- 76. Zhang H. et al. (2010) A two-component histidine kinase, MoSLN1, is required for cell wall integrity and pathogenicity of the rice blast fungus, Magnaporthe oryzae. Curr. Genet. 56, 517–528 pmid:20848286
- 77. Li G. et al. (2012) Genetic control of infection-related development in Magnaporthe oryzae. Curr. Opin. Microbiol. 15, 678–84 pmid:23085322
- 78. Dong Y. et al. (2015) Global Genome and Transcriptome Analyses of Magnaporthe oryzae Epidemic Isolate 98–06 Uncover Novel Effectors and Pathogenicity-Related Genes, Revealing Gene Gain and Lose Dynamics in Genome Evolution. PLOS Pathog. 11, e1004801 pmid:25837042
- 79. Morkunas I. and Ratajczak L. (2014) The role of sugar signaling in plant defense responses against fungal pathogens. Acta Physiol. Plant. 36, 1607–1619
- 80. Bolton M.D. (2009) Primary metabolism and plant defense—fuel for the fire. Mol. Plant. Microbe. Interact. 22, 487–497 pmid:19348567
- 81. Sakakibara H. et al. (2006) Interactions between nitrogen and cytokinin in the regulation of metabolism and development. Trends Plant Sci. 11, 440–8 pmid:16899391
- 82. LeJohn H.B. and Stevenson R.M. (1973) Cytokinins and magnesium ions may control the flow of metabolites and calcium ions through fungal cell membranes. Biochim. Biophys. Res. Commun. 54, 1061–1066
- 83. Durner J. et al. (1997) Salicylic acid and disease resistance in plants. Trends Plant Sci. 2, 266–274
- 84. León J. and Lawton M.A. (1995) Hydrogen Peroxide Stimulates Salicylic Acid Biosynthesis in Tobacco ‘. Plant Physiol. 108, 1673–1678 pmid:12228572
- 85. Pogány M. et al. (2004) Juvenility of tobacco induced by cytokinin gene introduction decreases susceptibility to Tobacco necrosis virus and confers tolerance to oxidative stress. Physiol. Mol. Plant Pathol. 65, 39–47
- 86. Akagi A. et al. (2014) WRKY45-dependent priming of diterpenoid phytoalexin biosynthesis in rice and the role of cytokinin in triggering the reaction. Plant Mol. Biol. 86, 171–83 pmid:25033935
- 87. Cheng H. et al. (2015) The WRKY45-2—WRKY13—WRKY42 Transcriptional Regulatory Cascade Is Required for Rice Resistance to Fungal Pathogen. Plant Physiol. 167, 1087–1099 pmid:25624395
- 88. Choi J. et al. (2011) Cytokinins and plant immunity: old foes or new friends? Trends Plant Sci. 16, 388–94 pmid:21470894
- 89. Giron D. et al. (2013) Cytokinins as key regulators in plant-microbe-insect interactions: connecting plant growth and defence. Funct. Ecol. 27, 599–609
- 90. Drüge U. and Schonbeck F. (1993) Effect of vesicular-arbuscular mycorrhizal infection on transpiration, photosynthesis and growth of flax (Linum usitatissimum L.) in relation to cytokinin levels. J. Plant Physiol. 141, 40–48
- 91. Frugier F. et al. (2008) Cytokinin: secret agent of symbiosis. Trends Plant Sci. 13, 115–20 pmid:18296104
- 92. Ribot C. et al. (2013) The Magnaporthe oryzae effector AVR1-CO39 is translocated into rice cells independently of a fungal-derived machinery. Plant J. 74, 1–12 pmid:23279638
- 93. Böhnert H. et al. (2004) A Putative Polyketide Synthase / Peptide Synthetase from Magnaporthe grisea Signals Pathogen Attack to Resistant Rice. Plant Cell 16, 2499–2513 pmid:15319478
- 94. Kämper J. (2004) A PCR-based system for highly efficient generation of gene replacement mutants in Ustilago maydis. Mol. Genet. Genomics 271, 103–10 pmid:14673645
- 95. Ou, S.H. (1985) Rice Diseases.
- 96. Berruyer R. et al. (2003) Identification and fine mapping of Pi33, the rice resistance gene corresponding to the Magnaporthe grisea avirulence gene ACE1. Theor. Appl. Genet. 107, 1139–47 pmid:12838393
- 97. Faivre-rampant O. et al. (2008) Characterization of the model system rice—Magnaporthe for the study of nonhost resistance in cereals. New Phytol. 180, 899–910 pmid:19138233
- 98. Gravot A. et al. (2010) Diurnal oscillations of metabolite abundances and gene analysis provide new insights into central metabolic processes of the brown alga Ectocarpus siliculosus. New Phytol. 188, 98–110 pmid:20862781
- 99. Zhou C. and Huang R.H. (2008) Crystallographic snapshots of eukaryotic dimethylallyltransferase acting on tRNA: insight into tRNA recognition and reaction mechanism. Proc. Natl. Acad. Sci. U. S. A. 105, 16142–16147 pmid:18852462
- 100. Dereeper a. et al. (2008) Phylogeny.fr: robust phylogenetic analysis for the non-specialist. Nucleic Acids Res. 36, 465–469