During recent years, large transcriptome profiling has become a routine analysis in the study of plant responses against abiotic and biotic stresses and was considered as a promising tool on the way to improve crop performance
. Here, we report on a cytochrome P450 gene of barley which was identified in a macroarray-based study as differentially regulated during nonhost and host interactions with fungal isolates of the Magnaporthe species complex. We provide evidence that this gene is involved in execution of penetration resistance as an early response to the ectoparasitic growth of the pathogen on the leaf surface.
Starting in 2000, the Leibniz Institute of Plant Genetics and Crop Plant Research (IPK) at Gatersleben developed a macroarray on the basis of expressed sequence tags (ESTs) and made it available as a resource for the barley community to study gene expression in response to various treatments
. Using this array, we identified a set of barley genes whose transcripts accumulate to a higher dose or at earlier time points after inoculation with a Magnaporthe nonhost as compared to a Magnaporthe host isolate
. A prominent member among these genes was an EST clone identical to the barley MLOC_15925 cDNA which has a domain characteristic for cytochrome P450 enzymes (Additional file
1: Figure S2). Plant P450s are involved in primary and secondary metabolic processes and catalyze different monooxygenation and hydroxylation reactions
. The wide distribution of P450s in a plant species-specific manner implies that they are subject to high evolutionary pressure which led to diversification and finally to novel metabolic pathways and products with acquired new biological functions. Search for homology among plant P450s placed the MLOC_15925 protein into the CYP96 family (designated therefore CYP96B22, Figure
1) and suggested an enzymatic function as midchain alkane hydroxylase (MAH). The closest Arabidopsis homologue of CYP96B22 is CYP96A10 (At4g39490), which was shown to be a splicing variant of a monocistronic transcript, forming a dimer with CYP96A9
[29, 30]. It was suggested that the dimerization is required for the modification of hydrophobic substrates
. In Arabidopsis CYP96A10 transcripts were found less abundant than CYP96A9 transcripts
, accounting for CYP96A10 concentration as a rate limiting factor at least at the level of transcriptional regulation. Assuming a similar situation for barley, this could be interpreted as if CYP96B22 has gained a novel regulatory function in the resistance response of barley against Magnaporthe nonhost isolates. In this scenario, the transcriptional up-regulation of CYP96B22 will lead to enhanced formation of heterodimers with a putative barley protein homologous to CYP96A9 and finally to more production of a hydrophobic product.
So far the only CYP96 enzyme characterized in more detail is Arabidopsis MAH1 which provides secondary alcohols and ketones as building blocks for the cuticular wax layer (Additional file
1: Figure S1;
[16, 31]). This outermost surface layer is the first barrier which plant pathogens have to cope with during infection
. Additionally, the cuticle might also be a reservoir for the release of signal molecules which could be perceived by the attacked plant itself as damage-associated molecular patterns (DAMP,
) leading to the execution of defense responses. Alternatively, cuticle constituents could be used by pathogens as host identification factors which was shown e.g. in the barley-powdery mildew interaction for a C26-aldehyde
. Cuticle components may also function in bacterial pathogenicity, e.g. it was shown that the Arabidopsis CYP86A2, a plant P450 which is involved in cuticle development, is required for the induction of the bacterial type III secretion system
. On the other hand it was shown that free cutin monomers, most likely released by the activity of fungal derived cutinases, may act as endogenous DAMPs and induce plant defense responses
. In Arabidopsis the main epicuticular wax components are alkanes, secondary alcohols and ketones with a chain length of C29, all of which are synthesized via a MAH1-dependent pathway
. However, barley cuticles mainly contain a C26-primary alcohol, n-hexacosanol, whose synthesis does not require a MAH1 enzymatic conversion
. It might be, therefore, that MAH1-derived wax components in barley are less important against pathogen attack as structural barriers but instead act in signaling.
We addressed the biological function of CYP96B22 in a transient assay using BSMV as a tool to knock-down gene expression (Figure
4). The usefulness of BSMV-mediated VIGS for functional genomics in wheat and barley has been demonstrated in several recent publications
[37, 38] and we already adopted the assay for the study of barley-Magnaporthe interactions
. Since it was shown for wheat that BSMV-infection prior to inoculation with M. oryzae host isolates may lead to decreased penetration
, we used in all of our VIGS-experiments plants inoculated with BSMV without silencing construct as an internal control (Figure
4). Doing so and comparing these control plants with those showing the highest reduction in CYP96B22 transcript abundance, a significant decrease in the frequency of papillae at sites of attempted penetration by the Magnaporthe nonhost isolate CD180 was observed (Figure
4). Since this latter result went along with more epidermal cells showing deposition of autofluorescent material at entire cell walls, this is indicative of ineffective papillae which could not block Magnaporthe penetration
[11, 23, 40]. The more rapid transition of fungal invasion across cell walls became even more obvious in the barley host interaction with M. oryzae isolate TH6772 where CYP96B22 silencing led to an increase of invasive hyphae in attacked epidermal cells. Similar findings for decreased penetration defense and more frequent entry of pathogens into epidermal cells, were also reported for Arabidopsis pen-mutants (for review see
). In the light of this finding CYP96B22 could be regarded as being a barley PEN-gene. However, neither in Arabidopsis pen-mutants nor in our experiments with CYP96B22-VIGS, nonhost pathogens were able to sporulate, indicating effective post-penetration resistance mechanisms. Disabling this second line of defense by blocking the EDS1-PAD4-SAG101 signaling complex renders Arabidopsis fully susceptible even to non-adapted pathogens
. In barley, the deposition of autofluorescent material seems to be a cytological marker for this back-up strategy in interactions with Magnaporthe.
Nonhost resistance is generally considered to be determined by several quantitative trait loci all of which must act in common to prevent infection by a would-be pathogen
. Only the detailed monitoring of minor cytological changes in BSMV-VIGS treated barley plants enabled us to assign a function in penetration resistance against Magnaporthe nonhost isolates to the barley gene CYP96B22. It would be advisable, therefore, to refrain from macroscopic evaluation in nonhost analysis after knock-out or knock-down of candidate gene expression. Since the infection process of the adapted powdery mildew fungus on barley plants after knock-down of CYP96B22 expression was not altered (Figure
5), we could rule out the possibility of an entire collapse of penetration resistance in these plants. The finding that CYP96B22 may act differently in barley powdery mildew vs. Magnaporthe interactions was not astonishing, since similar observations had been made before for barley MLO or ROP genes
[10, 44]. This ambivalence in gene regulation was interpreted as part of a dedicated response of barley to biotrophic or hemi-biotrophic pathogens. Interestingly, CYP96B22 was found to be down-regulated in barley after inoculation with Puccinia triticina and Puccinia hordei. However a functional verification of this result e.g. in a BSMV-VIGS assay has not been reported yet.
An unexpected finding of the present study was the sensing of the ectoparasitic growth of Magnaporthe germ tubes at the leaf surface by barley plants. Using the M. oryzae mutant Δpmk1, we undoubtedly verified that penetration is not necessary to trigger CYP96B22 expression (Figure
3). We showed also for Magnaporthe nonhost isolates, for which no Δpmk1 mutants existed, that they were able to cause the same phenotype by using a fungicide leading to non-functional appressoria and thereby preventing penetration (Figures
3). Other studies also indicated a signal exchange between pathogens and plants before penetration, however the chemical nature of the signal itself was never elucidated
[45–47]. Most recently it was shown for Arabidopsis that effectors of Colletotrichum higginsianum are secreted before penetration
. The homology of CYP96B22 to MAH1 enzymes and its implication in synthesis of cuticular waxes leads us to speculate that MAH1-dependent waxy components might be deliberated during the mucilage-facilitated attachment of spores and germ tubes to the leaf surface. Thereafter they are sensed as DAMPs and lead to increased defense responses associated with penetration resistance as e.g. the formation of effective papilla. In the host situation, M. oryzae isolates might then secrete effector molecules which negate these defense reactions