Powdery mildew infection of Arabidopsis thaliana



The powdery mildew Golovinomyces orontii is an obligate fungal biotroph that alters plant cellular architecture and metabolism to acquire all of its nutrients, including carbon and nitrogen, while limiting plant defensive responses including cell death.  G. orontii is uniquely suited to spatially and temporally resolved analyses as (i) its structural development and progression of disease on A. thaliana is well-defined, limited to epidermal cells, visible by light microscopy, and quantifiable, and (ii) host responses are typically focused at the site of infection.  Much work on powdery mildew-host interactions has focused on early stages of the interaction including factors required for penetration and/or formation of the haustorial complex, the fungal feeding structure.  My laboratory focuses on the later phase of a compatible interaction, the extensive external growth and reproduction of the powdery mildew, shown below at 5 days post infection (5 dpi).  Analysis at this later infection stage allows us to identify factors required to support the sustained extensive external growth of the fungi and the development of asexual fungal reproductive structures  known as conidiophores, as well as factors that limit host resistance responses.

Arabidopsis defense responses mediated by the phytohormone salicylic acid are important in limiting the extent of powdery mildew growth (Dewdney et al. Plant J 2000; Wildermuth et al. Nature 2001).  To define the transcriptional impact of SA in this system, replicated global expression profiling was performed and two novel statistical methods developed in collaboration with Terry Speed (UC Berkeley, Statistics) were employed.  The multivariate empirical Bayes statistic was designed for the analysis of replicated time series expression data and takes into account replication, moderation, and the relationship between samples over time (Tai and Speed, Annals of Statistics 2006; Chandran et al. Plant Physiology 2009).  The innovative adaptive model building procedure for cis-acting regulatory element identification uses multivariate time series expression data and the relative distance between cis-acting regulatory elements to identify individual and interacting motifs of significance within and across principal components (Zhang, Wildermuth, and Speed, Annals of Applied Statistics 2008). 

We found SA synthesis via ICS1 has an extensive transcriptional impact, altering the expression of ~4% of profiled genes, and identified genes with distinct and previously unresolved patterns of SA-impacted or SA-independent temporal expression (Chandran et al. Plant Physiology 2009; Zhang, Wildermuth, and Speed, Annals of Applied Statistics 2008).  This allowed us to identify known and novel SA-impacted processes and process components as well as to elucidate the regulatory circuitry of the host response.  For example, cis-acting regulatory elements bound by ethylene responsive factors (ERFS) were associated with cross-talk between SA and ethylene/jasmonic acid (JA) signaling pathways in powdery mildew, consistent with findings for the tomato ERF Pti4 (Gu et al., Plant Cell 2002) and AtERF1 (unpublished).  Furthermore, the enhanced sensitivity achieved using replicated time series data with these powerful statistical methods allowed us to identify regulators with altered expression in whole leaves that were previously below selection thresholds.  For instance, we identified PUX2, a plant ubiquitin regulatory X domain-containing protein, as an SA-induced novel regulator of powdery mildew resistance, with pux2 mutants exhibiting enhanced resistance to powdery mildew (Chandran et al., Plant Physiology 2009).

To further resolve the processes, components, and regulators of the powdery mildew (PM) interaction, we wanted to specifically examine the cells at the site of infection (the infected epidermal cell containing the fungal feeding structure and the neighboring epidermal and underlying mesophyll cells).  We developed such a system using lasermicrodissection (LMD) to isolate specific cells of interest for global expression profiling (Inada and Wildermuth, Planta 2004; Chandran et al., Proceedings of the National Academy of Sciences USA 2009).  Though LMD had been extensively used to study cancer progression in humans, its application to fragile plant leaf material with vacuoles is quite recent.  We invested considerable effort developing and verifying our tissue preparation, cell isolation, mRNA amplification, Affymetrix GeneChip expression profiling, and analysis methods to ensure we obtained the highest quality data possible.  For the community, we detailed our experimental and statistical methods in the extensive supplemental material provided with our 2009 PNAS paper and in two invited book chapters for Methods in Molecular Biology: Plant Immunity.

Postdoctoral researcher Dr. Divya Chandaran is isolating the mesophyll cells underlying the infected haustoria-containing epidermal cell from a 10 micron section of tissue prepared from an infected Arabidopsis leaf at 5 dpi.

Site-specific profiling increased our sensitivity dramatically, allowing us to identify transcriptional regulators hidden in whole leaf global expression analyses with a high success rate (Chandran et al., PNAS 2009).  It also allowed us to identify novel and known processes mediating the sustained growth and reproduction of the powdery mildew.  For example, seven transcriptional regulators associated with the cell cycle had altered expression at the site of infection at 5 days post infection, and cell cycle-associated cis-acting regulatory elements were enriched in genes with altered expression. 

Induced, localized host endoreduplication and biotrophic nutrient acquisition

Further experimental work was then performed to investigate whether the host cell cycle is manipulated as part of the late stage sustained growth and reproduction of the powdery mildew.  This resulted in our finding that host endoreduplication, an atypical cell cycle where DNA replication occurs in the absence of cell division, is dramatically induced at 5 dpi in leaf mesophyll cells underlying the infected epidermal cell that contains the fungal feeding structure (Figure 2) (Chandran et al. PNAS 2009).  Abrogation of this response results in reduced external growth of the powdery mildew.  As endoreduplication is often associated with specialized cells/organs with enhanced metabolic capacity, this suggests that induced host endoreduplication adjacent to the site of fungal nutrient acquisition might serve to accommodate the increased metabolic demands imposed by the biotrophic fungus. 

Figure 2. DNA content was markedly increased in mesophyll (M) cells underlying the infected epidermal (E) cells containing the haustorial complex (HC) at 5 days post infection, as indicated by red stars.  DAPI-stained nuclei are shown on the right (Chandran et al. PNAS 2009).

Induced, highly localized endoploidy at and/or adjacent to sites of sustained biotroph-plant nutrient exchange have now been reported for a diverse set of interactions (Wildermuth, Current Opinion in Plant Biology 2010).  These include parasitic interactions with powdery mildews and root nematodes, and symbiotic interactions with arbuscular mycorrhizal fungi and endosymbiotic rhizobia.  Where reported, when endoreduplication is compromised in these systems, growth and reproduction of the biotroph is reduced.  This supports the hypothesis that induced localized host endoreduplication may act as a common mechanism to support the enhanced metabolic demand imposed by these biotrophs. 

Detailed re-analysis of published transcriptome data for these systems as well as developmental systems with similar characteristics supports the hypothesis that endoreduplication serves to enhance metabolic capacity and identified specific processes and genes likely to be responsible for this enhancement (Wildermuth, COPB 2010).  For example, up-regulation of genes involved in fermentation including pyruvate decarboxylase and use of the pyruvate dehydrogenase bypass suggest these pathways are favored when flux through pyruvate is elevated as we would expect in these cells.  Remarkably, despite the dramatic differences in time scale, I found preferentially up-regulated ploidy-impacted processes associated with induced endoreduplication are similar to those retained following ancient whole genome duplication (WGD) events and follow the tenets of the Gene Balance Hypothesis and Metabolic Control Theory.  In yeast, others have shown that the retention of fermentation genes following ancient WGD would have conferred an immediate selective advantage in a glucose-rich environment (as discussed in Wildermuth COPB 2010).

As part of this work, data from my laboratory indicates that the transcription factor MYB3R4, a known activator of mitosis, also regulates endoreduplication.  Our data building on the work of others suggests MYB3R4 can function as a mitosis activator or repressor (resulting in endoreduplication) depending on its phosphorylation status (Chandran et al., PNAS 2009).  Furthermore, our analysis of the above systems indicates MYB3R proteins may be a common control point for induced endoreduplication.  There is also evidence for a similar duality in function for the Drosophila MYB transcription factor protein, a component of the Myb-MuvB (MMB)/dream complex, with extension to similar complexes in humans and C. elegans that modulate cell cycle progression. 

Current and Future Work

Current and future efforts in the lab are focused on specifically defining the regulation and function of powdery-mildew induced enhanced ploidy associated with the sustained and extensive external growth and reproduction of powdery mildew, focused on its role in metabolism.  Though enhanced ploidy has long been associated with enhanced metabolic capacity, the specifics of how ploidy results in enhanced metabolic capacity are largely undefined.

Currently funded by the National Science Foundation IOS-09581000.

See Publications to download cited articles.




2 hpi


5 hpi

Haustorial Complex

24 hpi

Growth and


72+ hpi