Overexpression of barley BAX inhibitor 1 induces breakdown of mlo-mediated penetration resistance to Blumeria graminis

Abstract

Cell death regulation is linked to pathogen defense in plants and animals. Execution of apoptosis as one type of programmed cell death in animals is irreversibly triggered by cytochrome c release from mitochondria via pores formed by BAX proteins. This type of programmed cell death can be prevented by expression of BAX inhibitor 1 (BI-1), a membrane protein that protects cells from the effects of BAX by an unknown mechanism. In barley, a homologue of the mammalian BI-1 is expressed in response to inoculation with the barley powdery mildew fungus Blumeria graminis f.sp. hordei (Bgh). We found differential expression of BI-1 in response to Bgh in susceptible and resistant plants. Chemical induction of resistance to Bgh by soil drench treatment with 2,6-dichloroisonicotinic acid led to down-regulation of the expression level of BI-1. Importantly, single-cell transient overexpression of BI-1 in epidermal leaf tissue of susceptible barley cultivar Ingrid led to enhanced accessibility, resulting in a higher penetration efficiency of Bgh on BI-1-transformed cells. In Bgh-resistant mlo5 genotypes, which do not express the negative regulator of defense and cell death MLO, overexpression of BI-1 almost completely reconstituted susceptibility to fungal penetration. We suggest that BI-1 is a regulator of cellular defense in barley sufficient to substitute for MLO function in accessibility to fungal parasites.

Programmed cell death (PCD) in animals and plants is involved in many developmental processes and stress responses. Animal apoptosis is a morphologically and biochemically defined type of PCD irreversibly triggered by cytochrome c release from mitochondria via pores in the outer mitochondrial membrane. This process is regulated by members of the Bcl-2 protein family that either support PCD, such as pore-forming BAX, or inhibit PCD, such as Bcl-2. BAX, Bcl-2, and their relatives are not present in plants. However, screening in yeast identified another mammalian antagonist of BAX (1). This antagonist was designated BAX inhibitor 1 (BI-1), and functional plant homologues of BI-1 were identified recently (2–5). BI-1 can interact with Bcl-2 but not with BAX, and it is localized at the endoplasmic reticulum and the nuclear envelope rather than at mitochondria (1, 3).

The hypersensitive reaction (HR) of plants to avirulent pathogens restricts pathogen growth effectively and includes a characteristic PCD of one or a few cells at the site of pathogen invasion (6–9). Interestingly, although plant BI-1 expression can inhibit BAX-induced PCD in yeast and Arabidopsis, overexpression of Arabidopsis BI-1 in resistant RPM1 plants appeared not to interfere directly with the hypersensitive cell death reaction induced by avirulent Pseudomonas syringae pv. tomatae. However, because Arabidopsis BI-1 is expressed in response to multiple stress treatments, BI-1 might play a role in protecting plants from stress-induced metabolic perturbations (4) or types of PCD different from that induced by a bacterial pathogen.

The biotrophic powdery mildew fungus Blumeria graminis f.sp. hordei (Bgh) establishes a compatible interaction with barley by formation of haustoria in penetrated living cells. During pustule development and maintenance, a certain leaf area around the emerging fungal colony remains green while the rest of the leaf undergoes an early senescence process. This “green island” effect illustrates massive pathogen-induced changes of cell death regulation resulting in cell death suppression in invaded cells and leaf senescence. Although it is self-evident that the HR stops development of biotrophic pathogens, it is not unequivocally proven that cell death is required for resistance. Therefore, manipulation of plant cell death regulation could be a tool for understanding the role of cell death in pathogen defense.

The HR is under genetic control, and some physiological features of HR resemble those of animal PCD (6–9). Moreover, BAX-induced cell death in tobacco is similar to HR (10). Leaf cell death control and defense control seem to be linked. For instance, cell death control mutants such as Arabidopsis lsd1 show both spontaneous cell death and broad-spectrum resistance (11). Also, Alvarez et al. (12) reported that onset of broad systemic acquired resistance in Arabidopsis is associated with development of microlesions. Barley lines carrying recessive mutant mlo alleles of the Mlo locus, similar to lsd1, show spontaneous leaf cell death and broad-spectrum resistance to Bgh (13, 14). Thus, the functional barley MLO protein is a negative control element of cell death and of defense responses to Bgh. Cell-survival mechanisms mediated by MLO probably negate plant defenses against Bgh, thereby allowing infection by the biotrophic fungus. However, mlo genotypes are highly susceptible to the hemibiotrophic pathogen Magnaporthe grisea and to necrosis-inducing culture filtrate from Bipolaris sorokiniana (15, 16). Kim et al. (17) recently demonstrated a link between MLO and calmodulin function. They suggested that negative MLO control of defense mechanisms against Bgh might be responsible for limited susceptibility to other pathogens, tagging MLO as a central modulator of antagonistic plant defense mechanisms. MLO structure is reminiscent of plasma membrane receptors that interact with heterotrimeric G proteins. However, MLO is not likely to depend on heterotrimeric G proteins (17) but possibly on small G proteins to fulfill its function in powdery mildew susceptibility. Down-regulation of the barley small GTP-binding protein RACB by RNA interference leads to enhanced penetration resistance to Bgh (18). This effect depends on Ror1, which is also required for mlo-mediated resistance (18, 19).

We present here a functional study on the implication of Bl-1 in disease resistance by an expression analysis of BI-1 in barley lines that are differently resistant to Bgh and by transient overexpression of BI-1 in barley epidermal cells during interaction with Bgh.

Materials and Methods

Plants, Pathogens, and Inoculation.

The barley (Hordeum vulgare L.) lines Ingrid, Pallas, and the corresponding backcross lines BCPMla12, BCPmlo5, and BCIngrid-mlo5 (I22) were obtained from Lisa Munk (Royal Veterinary and Agricultural University, Copenhagen). Their generation was described previously (20). Plants were grown in a growth chamber at 18°C with 60% relative humidity and a photoperiod of 16 h (60 μmol m⁻² s⁻¹ photon flux density). Barley powdery mildew fungus B. graminis (DC) Speer f.sp. hordei Em. Marchal, race A6 was inoculated onto barley primary leaves to give a certain density of conidia. We used 5 conidia mm⁻² for inoculation after chemical induction of resistance and macroscopic evaluation of induction success, 50 conidia mm⁻² for gene expression studies, and 150 conidia mm⁻² for gene function assessment on transformed leaf segments. Bgh was maintained on cv. Golden Promise under the same conditions.

Chemical Treatment.

2,6-Dichloroisonicotinic acid (DCINA, Syngenta AG, Basel), formulated as 25% active ingredient with a wettable powder carrier, was applied to 4-day-old barley seedlings of cultivar Pallas as a soil drench. The compound was used at a final concentration of 8 mg⋅liter⁻¹ soil volume. The suspensions were prepared with tap water. Soil drench with a wettable powder suspension served as a control.

RNA Extraction and Expression Analysis.

Total RNA was extracted from 8–10 primary leaf segments (5 cm long) by using RNA extraction buffer (AGS, Heidelberg) according to the manufacturer's instructions. For Northern blots, 10 μg of total RNA from each sample were separated in agarose gels and blotted by capillary transfer to positively charged nylon membranes. Detection of mRNAs was performed according to the DIG system user's guide with Digoxygenin-labeled antisense RNA probes (5). Before immunodetection, blots were washed stringently two times for 20 min in 0.1% (wt/vol) SDS/0.1× SSC at 68°C.

To detect low-level transcripts, we used the One-Step RT-PCR kit (Qiagen, Hilden, Germany) for semiquantitative reverse transcription PCR following the manufacturer's instructions. We used a low cycle number of 20 that maintained different transcript levels during the exponential amplification phase but did not allow cDNA detection in agarose gels by ethidium bromide staining. Hence, cDNAs were separated in agarose gels, denatured, blotted on nylon membranes, and detected with specific nonradioactively labeled RNA probes by using standard protocols and stringent conditions. Hybridization, washing, and immunodetection were performed as described for Northern blotting. Primers were 5′-ccaagatgcagatcttcgtga-3′ (5′ primer) and 5′-ttcgcgataggtaaaagagca-3′ (3′ primer) for a 513-bp Ubi cDNA fragment (GenBank accession no. M60175) and 5′-atggacgccttctactcgacctcg-3′ (5′ primer) and 5′-gccagagcaggatcgacgcc-3′ (3′ primer) for a 478-bp BI-1 cDNA fragment (accession no. AJ290421).

Transient Transformation and Evaluation of Penetration Efficiency.

A transient transformation protocol originally developed for wheat was used to transform barley via biolistic delivery of expression vectors into epidermal cells of leaf segments as described by Schweizer et al. (21). In general, each shot consisted of 312 μg of 1.1-μm tungsten particles with 0.3 μg of pGFP (GFP the under control of the CaMV ³⁵S promoter) (21) together with 0.7 μg of empty vector or pBI-1 containing BI-1 under the control of CaMV ³⁵S promoter. BI-1 was cloned into the SalI site of pGY-1 via restriction sites linked to oligo DNA primers that were used to amplify BI-1. Resulting pBI-1 and antisense pasBI-1 were sequenced to confirm that the original ORF was unchanged. To induce posttranscriptional gene silencing via RNA interference, particles were coated with dsRNA as described (18, 22).

Leaf segments were bombarded with coated particles 4 h before inoculation with Bgh, race A6. Inoculation with 150 conidia mm⁻² led to ≈50% of transformed cells attacked by the fungus. The outcome of the interaction was evaluated subsequently by fluorescence and light microscopy. For each experiment, a minimum of 100 interaction sites was evaluated. Transformed GFP-expressing cells were identified under blue light excitation. Three different categories of transformed cells were distinguished: penetrated cells that contained a haustorium, cells that were attacked by a fungal appressorium but did not contain a haustorium, and cells that were not attacked by Bgh. Cells that contained more than one haustorium or that contained haustoria but less than fungi attacked were scored as one penetrated cell. Cells with multiple attack from Bgh without a haustorium were scored as one unpenetrated cell. Stomata and stomatal guard cells were excluded from the evaluation. Bgh was detected by light microscopy or by fluorescence staining of the fungus with 0.3% calcofluor (wt/vol in water) for 30 s.

Penetration efficiency was calculated as number of penetrated cells divided by number of attacked cells multiplied by 100 and used as a measure for resistance of bombarded cells.

Results

Characterization of the BI-1 Amino Acid Sequence.

The ORF of the barley BI-1 gene (GenBank accession no. AJ290421) encodes 247 aa (5). The deduced protein of ≈25 kDa is very similar to the homologue of rice (88% identical, 98% similar) and Arabidopsis (75% similar) and 53% similar to the human BI-1 protein (Fig. 1). The barley BI-1 amino acid sequence contains presumably seven putative transmembrane domains with the C terminus in the cytosol (TMpred prediction, www.ch.embnet.org/software/TMPRED_form.html; ref. 23).

Figure 1.

Comparison of deduced amino acid sequences of barley (H. vulgare, GenBank accession no. CAC37797), rice (Oryza sativa, accession no. Q9MBD8), Arabidopsis thaliana (accession no. Q9LD45), and human (Homo sapiens, accession no. AAB87479) BI-1 proteins. Black-shaded amino acids are identical in all sequences. Gray-shaded amino acids are identical only in plant homologues. Bars indicate the seven predicted transmembrane domains in HvBI-1.

Expression of BI-1 in Response to B. graminis.

In a previous study, we showed that a barley BI-1 homologue is expressed in early response to attack by Bgh (5). Here, we studied BI-1 expression during compatible and incompatible interactions of near-isogenic barley backcross Pallas (BCP) lines bearing no functional resistance gene, the major resistance gene Mla12 mediating the HR after race-specific recognition of AvrMla12 from fungal race BghA6 (6, 24), or the recessive null mutant mlo allele mlo5 mediating broad penetration resistance (13, 17, 19, 24). Expression of BI-1 was analyzed during the first 7 days after dense inoculation with conidia of BghA6. We selected the pathogenesis-related protein 1b gene to confirm defense-related gene expression in the near-isogenic lines. This marker gene was expressed in response to Bgh in all lines (not shown). Starting with the same RNA, we analyzed BI-1 expression by RT-PCR and cDNA blotting. We detected constitutive BI-1 expression in all lines. Expression tended to increase slightly with leaf age and changed remarkably in response to Bgh (Fig. 2). Bgh-induced expression of BI-1 occurred early in BCPMla12 and BCPmlo5, whereas BI-1 transcript accumulation was delayed in the susceptible parent Pallas. Expression of BI-1 at 1 day after inoculation was strongest in BCPMla12, closely correlating with the onset of HR (24).

Figure 2.

BI-1 expression in resistant and susceptible barley lines. cDNA gel blot analysis. cDNAs were synthesized by RT-PCR from total RNA. RNA was isolated from susceptible Pallas, resistant BCPMla12, or resistant BCPmlo5 at 0 (immediately before inoculation), 1, 4, and 7 days after inoculation with Bgh and in parallel from noninoculated control plants (Ø). RT-PCR for BI-1 was carried out with 20 cycles under specific conditions. We checked loading of RNA (0.5 μg) by rRNA staining with ethidium bromide in gels. Repetition of the experiment led to similar results.

Early expression of BI-1 in response to Bgh posed the question as to whether BI-1 is expressed in the leaf epidermis, the only tissue in direct contact with fungal infection structures. Therefore, RNA was isolated from stripped epidermal and remaining leaf tissue of BghA6-inoculated Pallas and BCPMla12 1 day after inoculation. RT-PCR and cDNA gel blots revealed that barley BI-1 was expressed mainly in mesophyll tissue when compared with Ubiquitin 1, which is constitutively expressed equally in epidermis and mesophyll tissue (Fig. 3). We detected a low level of BI-1 transcripts in epidermal tissue, whereas other genes showed epidermis-dominant expression in the same plants (data not shown; ref. 18).

Figure 3.

BI-1 is expressed in mesophyll tissue. cDNA gel blot analysis. RT-PCR analysis with RNA from Pallas (P) and BCPMla12 (P10) by 24 h after inoculation with BghA6 is shown. For extraction of total RNA, abaxial epidermal stripes (E, inoculated site of the leaves) were separated from the mesophyll and adaxial epidermis (M). Ubiquitin 1 (Ubi) was selected as a marker for tissue-unspecific gene expression. RT-PCR was carried out with 30 cycles under specific conditions.

Expression of BI-1 in Chemically Induced Resistance.

We investigated Bl-1 expression in plants that were treated with the resistance-inducing compound DCINA. Four-day-old plants were soil-drench treated with 8 mg of DCINA per liter of soil volume and inoculated 3 days later with a low density of Bgh conidia suitable to macroscopically estimate the efficacy of chemically induced resistance (CIR). Plants expressing CIR showed ≈70% less mildew colonies than control plants treated with the unloaded carrier substance (Fig. 4A). Northern blots and cDNA blots were carried out to compare BI-1 transcript accumulation during the onset and expression of CIR with the transcript accumulation of the CIR marker gene Bci4 (25). As expected, Bci4 expression was up-regulated in response to DCINA treatment. In contrast, BI-1 was down-regulated 1–3 days after chemical treatment (Fig. 4B). The low inoculation density used in this experiment was not sufficient to induce strong BI-1 expression in response to inoculation, whereas PR1b was clearly induced (data not shown). However, BI-1 expression recovered in chemically induced plants after inoculation.

Figure 4.

BI-1 expression is repressed during chemical induction of resistance. (A) Chemically induced resistance in barley cultivar Pallas to Bgh. Barley first leaves treated with DCINA show less powdery mildew pustules than control leaves treated with the carrier substance wettable powder as a soil drench. (B) RNA (Bci4, 10 μg of RNA) and cDNA blots. RNA was extracted 0, 1, 2, and 3 days after soil drench treatment (dpt) with DCINA or the carrier substance wettable powder and additionally 1 and 4 days postinoculation (dpi, corresponding to 4 and 7 dpt). RT-PCR (Ubi, BI-1) was carried out with 20 cycles under specific conditions. Repetition of the experiment led to similar results.

BI-1 Overexpression Induces Accessibility to Bgh.

For gene function assessment, we performed transient BI-1 overexpression in barley epidermal cells by biolistic transformation and subsequent microscopic analysis of the interaction of Bgh with transformed cells (Fig. 5 A and B; refs. 18 and 21). In six independent experiments, overexpression of BI-1 in susceptible barley cultivar Ingrid resulted in significantly enhanced penetration efficiency (PE) of Bgh. The average PE was significantly enhanced from 47% to 72% (165% of the controls) on cells expressing BI-1 compared with control cells (Fig. 5C). In independent experiments using an antisense construct, the average PE of Bgh on pasBI-1-bombarded cells was reduced relatively by 12% compared with cells bombarded with an empty vector (Fig. 5D). However, the PE on cells transformed with pasBI-1 was not significantly different from that on controls. We obtained similar results, i.e., weak but not significant induction of resistance, by cobombardment of pGFP together with dsRNA of BI-1, which should induce sequence-specific RNA interference and thus down-regulation of BI-1 (18, 22). On Ingrid, coexpression of sense or antisense BI-1 with GFP did not change the number of transformed GFP-expressing cells per shot, indicating that BI-1 did not alter cell survival during the first 2 days after transformation (not shown).

Figure 5.

Overexpression of BI-1 induces supersusceptibility. (A) GFP-expressing cell that was penetrated by Bgh. The fungus formed a haustorium with finger-like protuberances (arrow) and elongated secondary hyphae on the leaf surface (arrowheads). Surface structures were stained with calcofluor and visualized by UV-light excitation. GFP and calcofluor images were merged in photoshop. (B) GFP-expressing cell that was attacked (arrow) but not penetrated by Bgh. (C) Average penetration efficiency of Bgh in six independent experiments with Bgh on barley cultivar Ingrid. PE of Bgh was enhanced significantly (P < 0.01, Student's t test) in cells that were bombarded with pBI-1 compared with cells that were bombarded with empty control pGY1. (D) Penetration efficiency of Bgh on cells that were bombarded with antisense-BI-1 (pasBI-1) was evaluated in independent experiments with Bgh. PE of Bgh was reduced nonsignificantly (P > 0.05) on cells that were bombarded with pasBI-1 compared with cells that were bombarded with empty control pGY1. Columns represent average values of independent experiments. Bars represent standard errors.

Induction of supersusceptibility by the putative cell death guard protein BI-1 is reminiscent of MLO function in barley (17, 26). Therefore, we transformed epidermal cells of Bgh-resistant mlo5 barley to test whether BI-1 interferes with mlo5-mediated resistance. In our experimental system, mlo5 genotypes in the background of cv. Pallas or cv. Ingrid were slightly accessible to Bgh. In seven independent experiments, we found that penetration efficiencies in control GFP cells ranged from 0% to 11% (minimum–maximum). Strikingly, BI-1 overexpression reconstituted accessibility in mlo5 barley close to a level that is typically seen in susceptible (Mlo) lines. Average penetration efficiency of Bgh on Ingrid-mlo5 and Pallas-mlo5 leaf segments was enhanced from 4% to 23% and from 6% to 33%, respectively (Fig. 6). This is equivalent to relative 520% and 510% of controls, respectively.

Figure 6.

Overexpression of BI-1 induces breakdown of mlo5-mediated penetration resistance. Penetration efficiency of Bgh was evaluated in three to four independent experiments with Bgh on barley cultivar Ingrid-mlo5 or Pallas-mlo5. PE of Bgh was enhanced significantly (P < 0.05) in cells that were bombarded with pBI-1 compared with cells that were bombarded with empty control pGY1. Columns represent average values of at least three independent experiments. Bars represent standard errors.

Discussion

We have shown here that the putative cell death regulator BI-1 is involved in the powdery mildew resistance of barley. We demonstrated pathogen-induced BI-1 expression, DCINA-induced BI-1 repression, and supersusceptibility to Bgh induced by BI-1 overexpression. Most importantly, barley BI-1, when overexpressed in mlo5 barley, was sufficient to induce accessibility to Bgh. Thus, BI-1 is a suppressor protein of mlo-mediated penetration resistance.

Stress response and cell death regulation in plants is not well understood. Many approaches to isolate pro- or antiapoptotic homologues of the animal Bcl-2 family from plants failed, and because the Arabidopsis genome sequence is now available, one may assume that plants lack homologous proteins. Nevertheless, there are some similarities of plant and animal PCD that indicate common elements in both systems. For instance, reactive oxygen intermediates, cysteine proteases, DNA degradation, and some morphological changes seem to take part in animal and plant PCD. Identification of some plant homologues of animal cell death suppressors, for instance BAG (Bcl-2 associated athanogene), DAD (defender against apoptotic death), and BI-1, indicates common elements of negative cell death control for eukaryotes. BI-1 proteins are highly conserved among humans, animals, and plants. Barley BI-1 is very similar to rice and Arabidopsis BI-1 proteins that were shown to inhibit BAX function in yeast and Arabidopsis (2–4). However, the mechanism by which BI-1 inhibits PCD is unknown. Further homologues of BI-1 have been identified in Arabidopsis, and it will be interesting to see whether they have cell death-suppressing capacity (8).

Bgh-induced expression of barley BI-1 correlated with early defense against Bgh in resistant barley as well as with pathogen development in susceptible barley. Therefore, we speculate that BI-1 is generally involved in cell survival at sites of fungal attack. BI-1 might be involved in both restriction of HR-associated cell death and fungus-induced cell survival. Cell survival and Bgh resistance are antagonistically regulated in barley. This is most prominently demonstrated by the fact that loss of MLO function leads to both Bgh resistance and spontaneous cell death (14). The functional Mlo gene is expressed in response to pathogens, wounding, reactive oxygen intermediates, and during leaf aging (27). A similar expression profile was found for BI-1 (refs. 4 and 5; Fig. 2; unpublished results). Because BI-1 overexpression induced susceptibility even in a mlo-mutant genotype, BI-1 similar to MLO might support a survival pathway, which negatively interferes with penetration resistance. The fact that overexpression of BI-1 mediated susceptibility independent from MLO, although they share no sequence similarity, shows that BI-1 acts independently or downstream from MLO.

The CIR response to Bgh correlates with enhanced epidermal cell death, papillae formation, and highly localized H₂O₂ accumulation (24, 28). The onset of CIR was accompanied by down-regulation of BI-1, whereas overexpression of BI-1 resulted in induced susceptibility (Figs. 4–6). This finding supports the notion that BI-1 is a negative regulator of penetration resistance to Bgh and strengthens the hypothesis that cell death control and plant defense against biotrophic pathogens are negatively linked. Accordingly, resistance, induced in barley by DCINA, is not effective against the toxin-producing and thus cell death-promoting fungal pathogen B. sorokiniana (J. Kumar and K.-H.K., unpublished results). Transient expression of antisense-BI-1 or BI-1-specific RNA interference did not influence fungal penetration efficiency strongly. This can be explained by the tissue-specific expression pattern of BI-1 that is weakly expressed in the epidermis (Fig. 3). Therefore, single-cell epidermal BI-1 silencing may not be sufficient to abrogate BI-1 function. Alternatively, BI-1 protein might show a low turnover rate explaining insufficiency of transient gene silencing.

MLO was detected mainly in the plasma membrane (29), whereas BI-1 was visualized as a fusion with GFP in endomembranes, particularly in ER and the nuclear envelope (refs. 1 and 3 and unpublished results). These findings question a physical interaction of MLO and BI-1. We speculate that BI-1 is a mesophyll teammate of epidermis-expressed MLO (27). Interestingly, although MLO is epidermis-specific, mlo mutants show spontaneous cell death especially in the mesophyll (30). Possibly, MLO-dependent signal exchange between epidermal and mesophyll tissue is required for BI-1 function in cell death control. Future studies need to be done to show whether overexpression of BI-1 in the mesophyll prevents barley from pleiotrophic mlo effects that have agronomic impact (13).

The role of BI-1 in susceptibility of wild-type barley is not yet clear. BI-1 overexpression induced supersusceptibility to Bgh (Fig. 5C), similar as MLO overexpression does (17). More importantly, BI-1 overexpression induced breakdown of mlo5-mediated penetration resistance. Sanchez et al. (4) reported that AtBI-1 overexpression in RPM1-Arabidopsis was insufficient for suppression of HR triggered by avirulent P. syringae pv. tomatae. Plant BI-1, when expressed in human fibrosarcoma cells, induced apoptosis-like PCD instead of preventing it, possibly by competing with the functional mammalian BI-1 (31). Our finding that barley BI-1 is able to substitute for the cell death suppressor MLO as a suppressor of Bgh penetration resistance supports the idea that BI-1 inhibits a specific, although unidentified, type of endogenous plant PCD. Interestingly, besides BI-1, two antioxidants, a soybean ascorbate peroxidase and a tomato glutathione S-transferase, as well as nuclear AtEBP, a protein that is known to be ethylene responsive, are able to suppress BAX-induced cell death in yeast (32–34). Barley and Arabidopsis BI-1 genes are responsive to pathogen challenge and wounding, which are both associated with oxidative stress. Together, BI-1 might be a redox-responsive cell death regulator involved in senescence processes similar as suggested for MLO (27). BAX and oxidative stress lead to pore formation in mitochondrial membranes finally triggering PCD (8). It was speculated that BI-1 could interfere with such a pore formation or forms by itself cell death-antagonistic ion channels (1, 4, 8). Recently, it was shown that antisense down-regulation of tobacco BI-1 accelerated cell death in BY-2 cells on carbon starvation (35). Mitochondria integrate diverse cell death signals including carbon starvation in plants such as in animals, and association of heterologously expressed BAX with plant mitochondria induces HR-like PCD (8, 10). However, it is not understood how cytochrome c release from mitochondria contributes to PCD in plants. Also, the role of cytochrome c in triggering plant defense is unclear. Based on its cellular localization, BI-1 should not directly interact with mitochondrial membranes. Possibly, BI-1 controls cellular levels of reactive oxygen intermediates, which accumulate both upstream and downstream of mitochondrial pore formation (8). In barley, this assumption is supported by the fact that resistance to Bgh is closely linked to H₂O₂ accumulation, whereas successfully invaded cells are completely bare of H₂O₂ (24).

The ambivalence of the MLO function in different pathosystems requires breeders to take the different infection strategies of plant parasites into account when they produce transgenic, pathogen-resistant plants. Because mlo5 barley is highly susceptible to M. grisea and to toxins of B. sorokiniana (15, 16), one may look at functional MLO as a resistance factor to hemibiotrophic and necrotrophic fungi. In the same direction, BI-1 could contribute to resistance against necrotrophic pathogens, as shown for the heterologous expression of cell death suppressors such as Bcl-2 or p35 in tobacco or tomato, respectively (36, 37).

This study demonstrates the contribution of BI-1 to regulation of plant defense to a pathogen. The fact that BI-1 proteins are able to suppress cell death in animals and plants as well as penetration resistance in barley indicates conserved overlapping pathways that regulate PCD and defense responses, possibly in all higher eukaryotes.

Abbreviations

BCP

backcross Pallas

Bgh

Blumeria (Erysiphe) graminis f.sp. hordei

BI-1

BAX inhibitor 1

CIR

chemically induced resistance

DCINA

2,6-dichloroisonicotinic acid

PCD

programmed cell death

HR

hypersensitive reaction

PE

penetration efficiency