Arabidopsis DMR6 encodes a putative 2OG-Fe(II) oxygenase that is defense-associated but required for susceptibility to downy mildew


*(fax +31 30 253 2837; e-mail


The Arabidopsis mutant downy mildew resistant 6 (dmr6) carries a recessive mutation that results in the loss of susceptibility to Hyaloperonospora parasitica. Here we describe the map-based cloning of DMR6 (At5g24530), which was found to encode a 2-oxoglutarate (2OG)-Fe(II) oxygenase of unknown function. DMR6 transcription is locally induced during infections with both compatible and incompatible H. parasitica isolates. High DMR6 transcript levels were also observed in constitutive defense mutants and after treatment with salicylic acid analog BTH, suggesting that DMR6 has a role during plant defense. Expression analysis of dmr6 mutants, using DNA microarrays and quantitative PCR, showed the enhanced expression of a subset of defense-associated genes, including DMR6 itself, suggesting dmr6-mediated resistance results from the activation of plant defense responses. Alternatively, resistance could be caused by the accumulation of a toxic DMR6 substrate, or by the absence of a DMR6 metabolic product that is required for H. parasitica infection.


In their natural environment, plants are challenged with a large variety of biotic and abiotic stresses. Despite their sessile nature they are able to cope with most forms of stress, including the attack of pathogens. However, plants are susceptible to a limited number of often specialized pathogen species. In such compatible plant–pathogen interactions one can envisage that the plant is actively involved in supporting the growth and development of the pathogen. Host proteins involved in establishing this basic compatibility can be considered susceptibility factors. The absence of such host susceptibility proteins, e.g. as a result of mutation, could lead to resistance or reduced susceptibility. This dependence on the host is particularly important for obligate biotrophic pathogens that require living plant tissue for their growth and reproduction. Most biotrophic fungi and oomycetes form specialized feeding structures, haustoria, within the host cells that they infect, which are thought to be important for nutrient uptake. In the rust pathogen Uromyces fabae, sugar and amino acid transporters, i.e. the d-glucose and d-fructose transporter, HXT1 (Voegele et al., 2001), and the amino acid transporters, AAT1 and AAT2 (Struck et al., 2002), are specifically localized to the haustorial membrane of the pathogen. If and how the plant contributes to the transport of nutrients over the plant cell membrane (the extrahaustorial membrane) is still an enigma. Besides the obvious feeding dependence, hardly anything is known about other aspects of disease susceptibility to fungal and oomycete biotrophs, i.e. the production of signals for pathogen development, the accommodation of infection structures and the vulnerability to suppression of plant defense responses.

Genetic studies on Arabidopsis have great potential to identify genes that are important for compatibility to biotrophic pathogens, i.e. the powdery mildew fungus Golovinomyces cichoracearum (previously known as Erysiphe cichoracearum) and the downy mildew oomycete Hyaloperonospora parasitica. Golovinomyces cichoracearum grows epicuticularly, forming haustoria from the outside into the epidermal cells. To identify the compatibility genes required for powdery mildew susceptibility, Vogel and colleagues have isolated 26 recessive Arabidopsis powdery mildew resistance (pmr) mutants (Vogel and Somerville, 2000). Four of the corresponding PMR genes have been cloned. PMR4 (GLS5 CalS12) encodes for a callose synthase (Nishimura et al., 2003). The pmr4 mutant can no longer induce a callose response, and shows enhanced activation of salicylic acid-dependent defense genes. PMR6 is a pectate lyase-like gene, and PMR5 is a gene of unknown function belonging to a large family of plant-specific genes. The cell-wall composition of both pmr5 and pmr6 mutant plants is altered, in particular the levels of pectin are increased (Vogel et al., 2002, 2004). PMR2 was identified as Atmlo2, which is an Arabidopsis ortholog of the Barley mlo gene (Consonni et al., 2006). Barley MLO encodes a plasma membrane protein with seven transmembrane domains (Buschges et al., 1997). The MLO protein is required for the successful entry of the powdery mildew pathogen Blumeria graminis f. sp. hordei into the host cell (Panstruga, 2005).

In contrast to the powdery mildews, downy mildews do penetrate into the host cell tissue and grow intercellularly, forming haustoria in mesophyll and epidermal cells (Koch and Slusarenko, 1990). A loss-of-susceptibility mutant screen in Arabidopsis has resulted in the identification of 20 downy mildew resistant (dmr) mutants, and eight of these that corresponded to six different loci, dmr1dmr6, were studied in more detail (Van Damme et al., 2005). The dmr3, dmr4 and dmr5 mutants showed elevated levels of PR-1 gene expression in the absence of pathogen infection, indicating that these are enhanced defense-response mutants. The dmr1, dmr2 and dmr6 mutants are postulated to be mutants in which cellular processes that are required for downy mildew infection are disturbed. The corresponding Arabidopsis DMR gene products could contribute to the H. parasitica infection process and promote disease susceptibility. Here we describe the cloning and characterization of DMR6 that encodes an oxidoreductase of the 2-oxoglutarate (2OG)-Fe(II) oxygenase superfamily. Although mutation of DMR6 leads to downy mildew resistance, expression of the gene is induced during plant defense. Furthermore, dmr6 mutants express enhanced levels of a subset of defense-associated genes, indicating that DMR6 negatively affects plant defense.


DMR6 encodes a putative 2OG-Fe(II) oxygenase

Previously dmr6-1 was identified as a recessive trait mediating H. parasitica resistance, and was mapped near marker nga139 on chromosome 5 to a region predicted to encompass 74 genes (Van Damme et al., 2005). Fine mapping reduced the dmr6-1 interval to a chromosomal region between the markers IND_K16H17 and CAP_At5g24590 located in the genes At5g24420 and At5g24590, respectively. Comparative analysis of the coding sequences within this chromosomal region in dmr6-1 and the parental line, Ler eds1-2, revealed a single point mutation in the second exon of At5g24530. This single base change of G to A, typical for an EMS mutation, changes TGG (trp codon) to TGA (premature stop codon) at nucleotide position 691 of the genomic sequence (Figure 1a). The At5g24530 gene is predicted to encode a 2OG-Fe(II)-dependent oxygenase with a mass of 39.4 kDa, of which the biological function has not yet been described. The premature stop codon truncates the predicted oxidoreductase enzyme of 342 amino acids, at position 141 before the conserved catalytic domain, as defined by Pfam03171, suggesting that dmr6-1 is a null allele.

Figure 1.

 Mutations in the DMR6 gene and their effect on susceptibility to Hyaloperonospora parasitica.
(a) The DMR6 gene contains four exons that form a coding sequence of 1026 bases. The two alleles are indicated: dmr6-1 with a base change in exon 2 and dmr6-2 with a T-DNA insertion in intron 2.
(b) Quantification of sporangiophores of H. parasitica isolate Waco9 on the dmr6-1 mutant, compared with its parental line Ler eds1-2, and on the dmr6-2 mutant (FLAG 445D09 T-DNA line), compared with its parental line Ws-4.
(c) Restoration of susceptibility by complementation with the At5g24530 cDNA under the control of the 35S promoter in the dmr6-1 mutant. H. parasitica sporangiophores per seedling were quantified on Ler eds1-2, dmr6-1 and three complementation lines, #122, #132 and #211.

A second allele, dmr6-2, was identified in a T-DNA insertion line (FLAG 445D09) from the mutant collection of INRA, Versailles (Samson et al., 2002). The presence and location of the T-DNA insert in the second intron of At5g24530 (Figure 1a) was confirmed by PCR and sequence analysis. FLAG 445D09 lines homozygous for the T-DNA insertion showed strongly reduced susceptibility to H. parasitica isolate Waco9, whereas the parental line (Ws-4) was highly susceptible (Figure 1b). RT-PCR, using a DMR6 and T-DNA primer, on cDNA synthesized from dmr6-2 mRNA resulted in the amplification of a DNA fragment (data not shown). This indicates that the second intron, containing the T-DNA, is not spliced out correctly in dmr6-2. However, the correctly spliced DMR6 transcripts could still be detected in dmr6-2 by quantitative PCR (Q-PCR), although at a strongly reduced level (∼ 11-fold, data not shown). This suggests that dmr6-2 is not a complete null allele. Nevertheless, the dmr6-2 mutant is nearly as resistant to H. parasitica as dmr6-1.

To corroborate that At5g24530 is required for H. parasitica growth, the dmr6-1 mutant was transformed with the cDNA sequence of At5g24530 under the control of the 35S promoter. In multiple independent dmr6-1 transformants, overexpression of At5g24530 was confirmed by Q-PCR. Three independent T3 lines, #122, #132 and #211, that overexpress DMR6 (Figure S1) were almost fully restored for H. parasitica susceptibility (Figure 1c). The complementation data, together with the identification of two independent dmr6 mutants, clearly demonstrates that a functional At5g24530 gene is required for susceptibility of Arabidopsis to H. parasitica.

DMR6 transcription is activated during H. parasitica infection

To investigate whether DMR6 transcription is altered during infection with H. parasitica, relative transcript levels were determined by Q-PCR. Arabidopsis Ler seedlings were sprayed with either a compatible (Cala2) or incompatible (Waco9) isolate, and DMR6 transcript levels were measured at different time points after inoculation on three biological replicates. As shown in Figure 2, transcript levels increased in both the compatible and incompatible interactions after 1 day post inoculation (dpi), suggesting that DMR6 activation is a general response to H. parasitica that is not compatibility specific. At the early time points (1–3 dpi) DMR6 was more highly expressed in the incompatible interaction than in the compatible interaction, although this was only significant at 3 dpi. The early activation of DMR6 suggests that it is activated as part of the defense response, which is in general more strongly activated in incompatible interactions. DMR6 expression levels did not increase in the incompatible interaction after 3 dpi. This correlates well with the arrest of pathogen growth that occurs between day 1 and 3. In contrast, in the compatible interaction at 4–5 dpi, when the pathogen has colonized the leaf, the level of DMR6 transcript was elevated by almost 40-fold, whereas in the incompatible interaction it was elevated by 20-fold.

Figure 2.

 Transcript levels of DMR6 after inoculation with a compatible (Cala2) or incompatible (Waco9) Hyaloperonospora parasitica isolate. Transcript levels were determined at different days post inoculation (dpi), as indicated. DMR6 transcript levels were measured in three independent biological replicates. Bars represent the mean induction relative to mock-treated plants, with error bars representing the standard deviation.

To study the localization of DMR6 expression during H. parasitica infection, transgenic lines were generated containing a construct with the DMR6 promoter linked to the uidA (GUS) reporter gene (ProDMR6:GUS). In each of the backgrounds tested, Col-0, Ler eds1-2 and dmr6-1, five independent ProDMR6:GUS lines were analysed. As the localization of GUS expression in Col-0 and Ler was essentially the same as that observed in Ler eds1-2, only results from this latter line are shown. H. parasitica hyphae were stained with trypan blue, and GUS activity was visualized using magenta-Xgluc as a GUS substrate, resulting in a magenta precipitate. GUS activity was specifically detected in cells containing haustoria or directly surrounding the intercellular hyphae, in both the compatible and incompatible interaction (Figure 3a–f), indicating that H. parasitica-induced DMR6 expression is strictly localized to sites that are in direct contact with the pathogen. Infection sites were smaller, and the number of cells showing GUS activity was lower in the incompatible interaction compared with the compatible interaction at 2 dpi (Figure 3a,c). However, our Q-PCR data showed that the overall level of DMR6 transcript is higher in the incompatible interaction at 2 dpi (Figure 2). As plants were inoculated with an equal dose of spores resulting in a similar number of infection sites, this suggests that the DMR6 transcript level is relatively higher in incompatible infection sites at 2 dpi. Interestingly, GUS activity was higher in cells containing the first formed haustoria than in cells with newly or recently formed haustoria in the compatible interaction (Figure 3c,e), indicating that DMR6 becomes activated following haustoria formation by the pathogen. This could also explain why DMR6 transcript levels are higher in the compatible interaction at later stages of the infection (4–5 dpi), as more tissue is colonized by the pathogen. No difference was observed in the localization of DMR6 expression in response to H. parasitica infection between Ler eds1-2 and dmr6-1 plants (Figure 3a–f), although GUS activity was generally higher in the dmr6-1 mutant background.

Figure 3.

 Analysis of a promoter DMR6-GUS fusion (ProDMR6:GUS) in transgenic Ler eds1-2 (left panels) and dmr6-1 (right panels) seedlings. GUS activity was visualized with Magenta-Xgluc as substrate, and Hyaloperonospora parasitica growth was visualized with trypan blue staining. The scale bars in the upper six panels (a–f) correspond to 10 μm; scale bars in the lower panels correspond to 100 μm.
(a) and (b) GUS activity in Arabidopsis cells that are in contact with the incompatible H. parasitica isolate Emoy2 in Ler eds1-2 and dmr6-1 transgenic lines, respectively at 2 days post inoculation (2 dpi).
(c) and (e) GUS activity in the infected cells of the Ler eds1-2 transgenic lines with the compatible H. parasitica isolate Cala2, at 2 and 3 dpi, respectively.
(d) and (f) GUS activity of the dmr6-1 transgenic lines with the compatible H. parasitica isolate Cala2, at 2 and 3 dpi respectively, is restricted to the infected cells.
(g) No GUS activity is detected in Ler eds1-2 transgenic lines 3 days after mock treatment.
(i) GUS activity is detected in the leaf primordia and shoot apical meristem in Ler eds1-2 transgenic seedlings 3 days after BTH treatment.
(h) GUS activity in dmr6-1, 3 days after mock treatment, in the leaf primordia and shoot apical meristem.
(j) GUS activity is higher in the true leaves, in addition to the apical meristem and leaf primordia, in the dmr6-1 transgenic lines 3 days after BTH treatment.

DMR6 expression is defense associated

The elevated DMR6 transcript levels during both compatible and incompatible H. parasitica interactions suggest that DMR6 transcription becomes activated as part of a defense response. As the activation of many defense genes is strongly impaired in plants unable to accumulate or respond to salicylic acid (SA), we tested DMR6 levels in the sid2 and npr1 mutant. sid2-1 (SA induction-deficient; Wildermuth et al., 2001) does not accumulate SA, whereas the npr1-1 (non-expressor of PR genes) mutant (Cao et al., 1994) shows strongly impaired defense gene expression in response to SA. The induction of DMR6 in response to a compatible (Waco9) or incompatible (Cala2) isolate was compared with wild-type Col-0 at 1 dpi in three independent biological replicates. Figure 4 shows that induction of DMR6 is higher in the incompatible than in the compatible interaction, confirming the data from Figure 2. Interestingly, the difference in DMR6 expression between compatible and incompatible is larger in Col-0 than in Ler, which could be because of the difference in the R gene that mediates resistance: RPP2 in Col-0 versus RPP5 in Ler. DMR6 transcript levels were not significantly altered in the sid2-1 and npr1-1 mutants, as compared with Col-0 upon inoculation with the incompatible isolate Waco9 or compatible isolate Cala2. This indicates that SA accumulation and NPR1 function are not important for the early transcriptional activation of DMR6 in response to H. parasitica infection. The responsiveness of DMR6 to SA or its analog BTH was tested on the Ler eds1-2 and dmr6-1 reporter lines containing the ProDMR6:GUS construct. No GUS activity could be detected in untreated Ler eds1-2 plants (Figure 3g). Interestingly, untreated dmr6-1 plants show GUS activity in the shoot apical meristem and leaf primordia (Figure 3h), indicating that dmr6-1 mutants have constitutively enhanced DMR6 transcription in these tissues. After BTH treatment, GUS activity was also detected in Ler eds1-2, and was primarily localized in the same tissues (the leaf primordia and shoot apical meristem) as in untreated dmr6-1 mutants (Figure 3i), indicating that DMR6 expression in these tissues is particularly sensitive to BTH. dmr6-1 shows an increase in GUS activity upon BTH treatment, with the GUS activity no longer localized strictly to the shoot apical meristem but also detected in true leaves (Figure 3j), indicating that in the dmr6-1 mutant DMR6 expression is more sensitive to BTH.

Figure 4.

 Transcript levels of DMR6 in defense regulatory and constitutive defense mutants.
(a) Fold induction of DMR6 transcript levels in npr1-1 and sid2-1, inoculated with either a compatible or incompatible isolate, relative to Col-0.
(b) Transcript levels of DMR6 in the constitutive defense mutants dmr3, dmr4 and dmr5, relative to the parental line Ler eds1-2. DMR6 transcript levels were measured in three independent biological replicates. Bars represent the mean fold change compared with the parental lines, with error bars representing the standard deviation.

To further investigate the correlation between the transcriptional activation of DMR6 and plant defense, DMR6 expression levels were analysed in the dmr3, dmr4 and dmr5 mutants (Van Damme et al., 2005), which have constitutively activated defense responses in the absence of pathogen infection. In all three dmr mutants high basal levels of DMR6 expression were observed compared with the parental line Ler eds1-2 (Figure 4b). This again supports the idea that expression of DMR6 is associated with plant defense. DMR6 expression was highest in dmr3 (approximately 97-fold induction) and dmr4 (approximately 56-fold induction), and was somewhat lower in dmr5 (approximately 34-fold induction). The combined data show that DMR6 expression is activated during various defense responses that are either pathogen- or chemical-induced.

DMR6 negatively affects expression of defense associated genes

To investigate whether dmr6 mutants have altered expression of more genes than DMR6 itself, changes in gene expression between dmr6-1 and its parental line were analyzed using DNA microarrays. Probes were synthesized from RNA extracted from the above-ground parts of healthy 14-day-old seedlings and were hybridized on 25k CATMA arrays (Allemeersch et al., 2005; Hilson et al., 2004). A total of 59 different CATMA probes were identified that showed significant differences in hybridization (P < 0.05, fold change > 2) between dmr6-1 and the parental line. The 59 CATMA probes corresponded to 57 unique AGI-IDs (Arabidopsis Genome Initiative Identification), of which 50 (including DMR6) showed an increased transcript level and seven showed a reduced transcript level in dmr6-1 compared with Ler eds1-2 (Tables S1 and S2). Several genes that show induced expression in dmr6-1 have previously been described to be associated with plant defense e.g. ACD6 (Lu et al., 2005), PR-4 (Potter et al., 1993) and PR-5 (Uknes et al., 1992). As the samples used for array hybridization were derived from a single biological replicate, strong conclusions cannot be drawn from this experiment. In order to obtain statistically sound data, a selection of genes, including the defense associated genes ACD6, PR-1, PR-2, PR-4 and PR-5, were verified by Q-PCR on three biological replicates. Expression levels were measured in the dmr6-1 mutant, the dmr6-2 mutant and three dmr6-1 complementation lines (data is shown for one representative line #122). All genes tested were more highly expressed in both dmr6-1 and dmr6-2 compared with the parental lines (Figure 5a,b). Moreover, the complemented dmr6-1 mutant (with the 35S::DMR6 construct) showed significantly lower expression of defense associated genes than the mutants, demonstrating that the loss of a functional DMR6 gene is responsible for the elevated expression levels of a number of defense-associated genes. It is unclear why the defense gene expression levels are still higher in the complemented mutant than in the parental line. The difference in defense gene expression between dmr6-2 compared with Ws-4 is larger then the difference between dmr6-1 compared with Ler eds1-2, and could be the result of differences in the genetic background of the two dmr6 mutants. The enhanced expression of DMR6 in dmr6-1 (Figure 3h) was also detected in the microarray experiment and confirmed by Q-PCR (Figure 5c). Interestingly, DMR6 transcript levels in dmr6-1 (approximately 11-fold induction) were not as high as in dmr3, dmr4 and dmr5 (97-, 56- and 34-fold induction, respectively; Figure 4b). The transcript data clearly indicate that DMR6 has a negative effect on the expression of the subset of the defense-associated genes tested.

Figure 5.

 Relative transcript levels of 10 defense-associated genes in dmr6-1 and dmr6-2. Gene IDs are as indicated.
(a) Transcript levels of nine defense-associated genes in dmr6-1 and dmr6-1 complemented with a 35S::DMR6 construct, relative to the parental line Ler eds1-2.
(b) Transcript levels of nine defense-associated genes in dmr6-2 relative to the parental line Ws-4.
(c) Transcript levels of DMR6 in dmr6-1 and dmr6-1 complemented with a 35S::DMR6 construct relative to Ler eds1-2. Gene transcript levels were measured in three independent biological replicates. Bars represent the mean fold change relative to the parental lines, with error bars representing the standard deviation.


The DMR6 locus was identified in a genetic screen for loss of susceptibility to the downy mildew pathogen H. parasitica (Van Damme et al., 2005). Map-based cloning identified DMR6 as At5g24530, which encodes an oxidoreductase, for which no biological function has been described. Overexpression of At5g24530 restored susceptibility to H. parasitica in the dmr6-1 mutant, thereby confirming that this gene is required for susceptibility to H. parasitica. An additional allele, dmr6-2, was identified in the line FLAG 445D09 (Samson et al., 2002) that has a T-DNA insertion in the At5g24530 gene. Both the dmr6-1 and dmr6-2 mutants exhibit reduced susceptibility to H. parasitica.

The DMR6 protein is a member of the 2OG-Fe(II) oxygenase superfamily of oxidoreductases. The catalytic domain, as defined by Pfam PF03171, is located from amino acid 188 to 288 in DMR6, and contains the three Fe(II) ion-binding residues (His212, Asp214 and His269) (Roach et al., 1995). Molecular oxygen is reduced at the Fe(II) ion, where it reacts with 2-oxoglutarate and a specific substrate through the incorporation of one atom of oxygen in each compound. The 2OG-Fe(II) oxygenases are widespread in bacteria and eukaryotes (Aravind and Koonin, 2001). In plants, these enzymes catalyze different hydroxylation and desaturation steps. Examples are the gibberellin 20-oxidases, gibberellin 3β-hydroxylases and gibberellin 2-oxidases in the biosynthesis of gibberellins (Hedden and Phillips, 2000), ACC oxidase in the final step in ethylene synthesis (Wang et al., 2002), and flavanone 3β-hydroxylase (F3H) in the biosynthesis of flavanoids, cathechins and anthocyanidins (Lukacin and Britsch, 1997).

Expression analysis by Q-PCR and promoter GUS lines demonstrated that DMR6 is strongly induced during the interaction with both compatible and incompatible H. parasitica isolates. The activation of DMR6 is locally induced within cells that are in direct contact with the pathogen. Induction of DMR6 at early stages of infection (1 dpi), in both compatible and incompatible interactions, was independent of SA and NPR1 function. However, under other conditions, SA and NPR1 were found to be important for DMR6 expression, as observed in the DNA microarray data by Mosher et al. (2006), in which DMR6 expression was measured as being derepressed in sni1 (suppressor of npr1, inducible), as well as being BTH non-responsive in the npr1 mutant. As SNI1 is thought to function as a negative regulator of systemic acquired resistance (SAR), DMR6 expression can be considered as being SAR induced. This is supported by the fact that DMR6 expression is induced by BTH, as shown in the ProDMR6:GUS lines. DMR6 is induced to higher levels in incompatible than in compatible interactions with H. parasitica during early time points post inoculation. In general, defense responses are more strongly induced during incompatible interactions, leading to higher transcript levels of defense genes that are often dependent on the accumulation of SA (Cao et al., 1994; Lamb et al., 1992; Tao et al., 2003). High basal expression levels of DMR6 in the constitutive defense mutants, dmr3, dmr4 and dmr5, in the absence of pathogen infection support the link between DMR6 transcription and plant defense. Interestingly, an ortholog of DMR6 in wheat, which was described as an F3H gene, showed a similar pattern of expression during compatible and incompatible interactions with the Hessian fly (Giovanini et al., 2006), suggesting that the monocot DMR6 ortholog has a similar function.

It is unlikely that DMR6 has direct antimicrobial activity or has a role in the biosynthesis of antimicrobial compounds, as plants that lack a functional DMR6 gene have reduced susceptibility. Moreover, overexpression of DMR6 in the dmr6-1 mutant restores susceptibility. The constitutive activation of defense-associated genes including PR-genes in dmr6-1 and dmr6-2 strongly suggests that DMR6 negatively affects the expression of these genes. The enhanced expression of defense-associated genes could be responsible for the observed dmr6-mediated resistance. The question remains why dmr6-1 mutant seedlings are only resistant to H. parasitica and Colletotrichum higginsianum (O’Connell and Panstruga, 2006), and not to Pseudomonas syringae, Golovinomyces orontii (Van Damme et al., 2005) and the white rust oomycete, Albugo candida (E. Holub, personal communication). If the enhanced defense is causing the resistance, one would expect dmr6 mutants to be resistant to a broad range of pathogen species. However, many constitutive defense mutants that show this type of broad-range resistance show dwarfism, e.g. dmr3 and dmr4 (Van Damme et al., 2005). dmr6 mutants do not show dwarfism and this could indicate that constitutive defense activation is less strong. This is supported by the fact that a number of defense genes are expressed to a much higher level in dmr3 and dmr4 than in dmr6. Possibly, the subset of defense genes and their level of expression in the dmr6 mutants only provide protection against H. parasitica and C. higginsianum. The mechanism of upregulation of this subset of genes through mutation of DMR6 is unknown.

In the dmr6 mutants, the substrate of the putative DMR6 encoded 2OG-Fe(II) oxygenase is expected to accumulate, and could either be directly toxic to the pathogen or act indirectly by stimulating the expression of host defense genes. Possibly, the accumulated substrate in the dmr6 mutant triggers and enhances the expression of DMR6. Also, in wild-type plants, in reponse to H. parasitica infection or BTH treatment, DMR6 substrate levels could be increased, resulting in the transcriptional activation of DMR6 as part of the plant defense response. In addition, this would explain the high sensitivity of dmr6-1 mutants to BTH-induced DMR6 expression, as dmr6 mutants may be unable to metabolize the BTH-induced substrate, resulting in enhanced activation of DMR6. In this scenario, the function of DMR6 would be to control DMR6 substrate levels during plant defense, thereby acting as a negative regulator.

Alternatively, it could be that the product of the DMR6 enzyme is either negatively regulating defense-associated gene expression or positively affecting susceptibility to H. parasitica. However, it is difficult to understand why plants would activate DMR6 transcription during defense if the product of the DMR6 enzyme or DMR6 itself would be beneficial for the pathogen. In this scenario, plants would only maintain DMR6 if it had an evolutionary advantage, e.g. an important role in plant defense. However, this is unlikely as dmr6 mutants are not impaired in basal, R gene-mediated or BTH-induced resistance.

Genetic studies using defense regulatory mutants will reveal whether the enhanced defense gene expression is an essential component of dmr6-mediated resistance, or a secondary effect. Future studies aimed at identifying the DMR6 substrate and product, as well as genetic suppressor studies, will increase our understanding of the role of the DMR6-encoded 2OG-Fe(II) oxygenase in both defense and susceptibility to downy mildew.

Experimental procedures

Plant growth conditions and H. parasitica inoculations

Plants were grown on potting soil in a growth chamber (Snijders, at 22°C with 16 h of light (100 μE m−2 sec−1) and a relative humidity of 75% before inoculation. Hyaloperonospora parasitica isolates Waco9 (kindly provided by Dr M. Aarts, WUR, and Cala2 (kindly provided by Dr E. Holub, Warwick HRI, were maintained on Arabidopsis accessions Ws-0 and Ler, respectively. Inocula (4 × 105 spores ml−1) were transferred weekly to 10-day-old healthy seedlings (Holub et al., 1994) by use of a spray gun. Seedlings were air-dried for approximately 45 min and incubated under a sealed lid at 100% relative humidity in a growth chamber at 16°C with 9 h of light per day (100 μE m−2 sec−1). Sporulation levels were quantified 6 dpi by counting the number of sporangiophores per seedling (n ≥ 40).

dmr6 mutant backgrounds

dmr6-1 was identified previously and crossed to the Col-0 rpp2 mutant FN2 (Sinapidou et al., 2004) to create a mapping population (Van Damme et al., 2005). For all other experiments, a dmr6-1 line was used that was back-crossed twice to the parental line Ler eds1-2 (Parker et al., 1996). dmr6-2 was identified among the segregating offspring of the T-DNA insertion line FLAG 445D09 in the Ws-4 accession (Samson et al., 2002). The T-DNA insertion was confirmed by PCR using a primer designed in the At5g24530 gene, LP primer (5′-CAGGTTTATGGCATATCTCACGTC-3′), in combination with the T-DNA right-border primer, Tag3′ (5′-CTGATACCAGACGTTGCCCGCATAA-3′) or RB4 (5′-TCACGGGTTGGGGTTTCTACAGGAC-3′). The exact T-DNA insertion in the second intron of At5g24530 was confirmed by sequencing of amplicons generated with the T-DNA primers from both the left and right borders, in combination with the gene-specific primers LP or RP (5′-ATGTCCAAGTCCAATAGCCACAAG-3′).

Map-based cloning

INDEL- and CAPs-based markers were used for fine mapping. Primers were designed around the flanking sites of known insertion and deletion polymorphisms between Col-0 and Ler, as obtained from the Cereon database administered by Monsanto (Jander et al., 2002). The dmr6-1 interval mapped between the INDEL-marker IND_K16H17 within gene At5g24420 (forward primer 5′-TGGGGTTGTGGTTTATTCTGTTGAC-3′ and reverse primer 5′-TGGCCAATAGTAGTTGATACGCAAGA-3′) and CAPs-marker CAP_At5g24590 (forward primer 5′-GCATCATTTGTACCGTACTGAGTC-3′ and reverse primer 5′-TAGTGGATACTCTGTCCCTGAGGT-3′, restriction enzyme PdmI). Candidate genes were sequenced to identify the dmr6-1 mutation.

Complementation lines

The At5g24530 coding sequence of Col-0 was PCR-amplified with primers that included restriction sites that were used for directional cloning; a forward primer (5′-TTCTGGGATCCAATGGCGGCAAAGCTGATATC-3′) containing a BamHI restriction site near the start codon (ATG), and a reverse primer (5′-GATATATGAATTCTTAGTTGTTTAGAAAATTCTCGAGGC-3′) containing an EcoRI site after the stop codon (TTA). The fragment was cloned directionally between the 35S promoter and the Nos terminator, and the 35S-DMR6-Tn fragment was cloned into the pGreenII0229 (Hellens et al., 2000). Complementation lines were generated by transforming dmr6-1 plants by the floral-dip method with the Agrobacterium tumefaciens strain C58 (Clough and Bent, 1998). d,l-Phosphinothricin-resistant (BASTA, 300 μm) seedlings were selected and analyzed for overexpression of DMR6 and H. parasitica susceptibility.

Promoter DMR6:GUS transgenic lines

A 2.5-kb fragment of the DMR6 promoter was amplified using proDMR6F (5′-GACTCTGTCTGAGTCTGCAGTCCCAAACCATG-3′) and proDMR6R (5′-GCCGCCATTGGATCCCAGAAAATTGAAGAAG-3′), generating PstI and BamHI restriction sites, respectively. The two restriction sites allowed cloning of the fragment into the pGREENII0229G plasmid (Hellens et al., 2000) in front of the GUS gene. ProDMR6:GUS transgenic lines were generated by transforming Ler eds1-2; Col-0; and dmr6-1 plants by the floral-dip method (Clough and Bent, 1998) with the A. tumefaciens C58 strain. ProDMR6:GUS transgenic (T3) seedlings were infected with H. parasitica and at 3 dpi the seedlings were vacuum infiltrated with Magenta-Xgluc solution [50 mm NaPO4 (pH 7.0), 0.5 mm K3FE(CN)6, 0.5 mm K4FE(CN)6, 0.1% Triton X-100, 0.5 mg ml−1 Magenta-GlcA; Duchefa,] and stained overnight at 37°C. Seedlings were cleared with 70% EtOH. H. parasitica was stained in lactophenol (1:1:1:1 volume of lactic acid/glycerol/phenol/H2O) containing 1 mg mL−1 trypan blue, by boiling for 1–2 min and destaining overnight in chloral hydrate. Trapped air bubbles were removed by 1 min of speed vacuum infiltration. Hyaloperonospora parasitica growth and GUS expression was visualized by interference contrast microscopy.

DNA microarray and Q-PCR analysis

Total RNA was extracted with an RNeasy kit (Qiagen, and treated with the RNase-free DNase set (Qiagen). RNA was quantified using a UVmini-1240 spectrophotometer (Shimadzu, All procedures regarding the microarray analysis are detailed in Appendix S1. For Q-PCR, cDNA was synthesized with SuperScript-III reverse transcriptase (Invitrogen, and oligo(dT)15 (Promega, from total RNA. Cycle thresholds were determined per transcript in triplicate in multiple biological replicates using the ABI PRISM 7700 sequence detection system (Applied Biosystems, using SYBR Green I (Applied Biosystems) as the reporter dye. The data are normalized using Arabidopsis ACT2 levels (At3g18780). The used primer sets for the transcripts generating 99–101 base-pair fragments are listed in Table S3.


We thank F. Kindt and R. Leito for artwork, P.J. Weisbeek for his support and for critically reading the manuscript, and E. Holub for testing Albugo candida on the dmr mutants. This research was initiated with a fellowship of the Royal Netherlands Academy of Arts and Sciences to G. Van den Ackerveken. R. Huibers was co-financed by the Centre for BioSystems Genomics (CBSG) which is part of the Netherlands Genomics Initiative/Netherlands Organisation for Scientific Research.