The pepper MLO gene, CaMLO2, is involved in the susceptibility cell-death response and bacterial and oomycete proliferation




Loss-of-function alleles of the mildew resistance locus O (MLO) gene provide broad-spectrum powdery mildew disease resistance. Here, we identified a pepper (Capsicum annuum) MLO gene (CaMLO2) that is transcriptionally induced by Xanthomonas campestris pv. vesicatoria (Xcv) infection. Topology and subcellular localization analyses reveal that CaMLO2 is a plasma membrane-anchored and amphiphilic Ca2+-dependent calmodulin-binding protein. CaMLO2 expression is up-regulated by Xcv and salicylic acid, as well as abiotic stresses. Silencing of CaMLO2 in pepper plants confers enhanced resistance against virulent Xcv, but not against avirulent Xcv. This resistance is accompanied by a compromised susceptibility cell-death response and reduced bacterial growth, as well as an accelerated reactive oxygen species burst. Virulent Xcv infection drastically induces expression of the salicylic acid-dependent defense marker gene CaPR1 in CaMLO2-silenced leaves. CaMLO2 over-expression in Arabidopsis enhances susceptibility to Pseudomonas syringae pv. tomato and Hyaloperonospora arabidopsidis. Leaves of plants over-expressing CaMLO2 exhibit a susceptibility cell-death response and high bacterial growth during virulent Pst DC3000 infection. These are accompanied by enhanced electrolyte leakage but compromised induction of some defense response genes and the reactive oxygen species. Together, our results suggest that CaMLO2 is involved in the susceptibility cell-death response and bacterial and oomycete proliferation in pepper and Arabidopsis.


Plants have evolved sophisticated defense mechanisms to combat an abundance of microbial pathogens (Jones and Dangl, 2006). Plant resistance is specialized by and dependent upon a combination of two genes: one from the host (the resistance gene: R) and one from the pathogen (the avirulence gene: avr) (Flor, 1971; Hammond-Kosack and Jones, 1996). The R gene-dependent defense response is triggered by recognition within the host plants that leads to localized hypersensitive response cell death to restrict the invading pathogen (Chisholm et al., 2006).

In contrast to gene-for-gene resistance, durable broad-spectrum resistance to microbial pathogens may be conferred by mutations within single host plant genes. For instance, Arabidopsis lesion simulating disease resistance (lsd1) and enhanced disease resistance (edr1) mutants display broad-spectrum resistance to virulent bacterial and fungal pathogens, while the rice sekiguchi-lesion (sl) mutant exhibits resistance to the rice blast fungus Magnaporthe oryzae (Dietrich et al., 1994; Arase et al., 1997; Frye and Innes, 1998). Notably, broad-spectrum resistance to most races of the powdery mildew fungus Blumeria graminis f. sp. hordei (Bgh) is conferred by homozygous mutant (mlo) alleles of the mildew resistance locus O (MLO) gene in barley (Jørgensen, 1992; Büschges et al., 1997; Consonni et al., 2006).

The Arabidopsis and barley MLO genes encode calmodulin (CaM)-binding proteins with seven transmembrane domains. Mutations in MLO provide pre-invasive resistance at the cell wall that persists for a long time to combat powdery mildews (Büschges et al., 1997; Consonni et al., 2006). This immunity partially depends on the prototype family of soluble N-ethylmaleimide sensitive fusion protein (NSF) attachment protein receptors (SNAREs), syntaxins that include Arabidopsis PEN1 and the barley ortholog ROR2 (Collins et al., 2003; Consonni et al., 2006). In contrast, powdery mildew-resistant genotypes with a compromised MLO pathway exhibit enhanced sensitivity to necrotrophic Bipolaris sorokiniana toxins compared with wild-type barley (Kumar et al., 2001). Sensitivity is associated with a massive accumulation of H2O2. Therefore, a compromised MLO pathway is effective for control of biotrophic powdery mildew fungus but not for control of necrotrophic B. sorokiniana, which causes leaf spot blotch disease. Similarly, barley mlo mutants exhibit enhanced susceptibility to the rice blast fungus Magnaporthe oryzae (Jarosch et al., 1999).

MLO is temporarily induced in barley by Bgh, Magnaporthe oryzae, wounding, paraquat treatment, fungal-derived carbohydrate elicitors and leaf senescence (Piffanelli et al., 2002). Loss-of-function mlo mutants senesce more rapidly than wild-type MLO plants, indicating that MLO functions as a negative regulator of senescence-associated cell death (Piffanelli et al., 2002). A stimulus-induced imbalance between the MLO levels and cell-death signals may regulate plant cell death. The temporary nature of MLO induction probably dictates the cell response to subsequent events of pathogen attack (Stein and Somerville, 2002).

In addition to the distinct role of MLOs in plant defense, MLO proteins function in plant developmental processes such root thigmomorphogenesis (Chen et al., 2009) and pollen tube perception (Kessler et al., 2010). The exaggerated root-curling phenotypes of the mlo4 and mlo11 Arabidopsis mutants were shown to depend on auxin gradients, indicating that MLO4 and MLO11 co-function as modulators of touch-induced root tropism (Chen et al., 2009). AtMLO7/NORTIA (NTA), a member of the MLO family, regulates pollen tube reception in synergid cells of the female gametophyte in Arabidopsis (Kessler et al., 2010).

Topology analyses of the MLO family in plants suggests that MLO proteins function as seven-transmembrane domain-containing G protein-coupled receptors, which are known to link diverse extracellular stimuli to intracellular signaling cascades through heterotrimeric G proteins in vertebrates (Bockaert and Pin, 1999; Devoto et al., 1999). However, the barley MLO protein functions independently of the heterotrimeric G proteins (Kim et al., 2002b). MLO proteins were proposed to interact with CaM in vitro in a Ca2+-dependent manner (Kim et al., 2002a,b). OsMLO, an MLO homolog in rice, was identified by a yeast two-hybrid screen for rice CaM-binding proteins (Kim et al., 2002a), supporting the involvement of CaM in mlo resistance in barley (Kim et al., 2002b). A narrow region of the MLO C-terminus was identified as essential for CaM binding. CaM is reported to play a role in MLO function by supporting calcium cation interactions that influence plant defense and cell death (Kim et al., 2002a,b). The binding of Ca2+-attached CaM to the C-terminal CaM-binding domain of MLO is a common feature of MLO proteins. However, the precise mechanism of modulation of CaM by MLO proteins is poorly understood. The MLO CaM-binding proteins harboring seven transmembrane domains are unique to plants (Chen et al., 2006). The recessive mlo alleles are involved in plant resistance; however, the biochemical function of MLO in modulating Ca2+-loaded CaM remains elusive. Defining the physiological and biochemical roles of other members of the MLO protein family is necessary for an improved understanding of the specific function of MLO in regulating plant defense and cell death.

Despite the abundance of information concerning barley MLO function in powdery mildew disease resistance (Büschges et al., 1997; Stein and Somerville, 2002; Devoto et al., 2003; Consonni et al., 2006, 2010), MLO function in response to hemibiotrophic bacterial pathogens is not fully understood. Recently, Arabidopsis AtMLO2 was shown to be the target of the Pseudomonas syringae type III effector HopZ2 that is required for its virulence function (Lewis et al., 2012). Here, we report the role of the pepper (Capsicum annuum cv. Nockwang) pathogen-induced MLO2 (CaMLO2) gene in susceptibility to hemibiotrophic bacterial pathogens. A CaMLO2 cDNA clone was derived from pepper leaves infected with avirulent Xanthomonas campestris pv. vesicatoria (Xcv) strain Bv5-4a. Similar to other plant MLO proteins, sequence analysis of CaMLO2 identified a small amphiphilic CaM-binding domain (CaMBD) at the C-terminus and seven transmembrane domains at the N-terminus. The CaMBD of CaMLO2 specifically interacted with CaM in a Ca2+-dependent manner in vitro. Subcellular localization of membrane-anchored CaMLO2 was determined via transient expression in Nicotiana benthamiana and biolistic bombardment in onion cells. CaMLO2 was differentially induced by virulent and avirulent Xcv as well as the oomycete pathogen Phytophthora capsici. Silencing of CaMLO2 in pepper plants conferred enhanced resistance to virulent Xcv infection, but not avirulent Xcv infection. Furthermore, ectopic expression of CaMLO2 in Arabidopsis conferred enhanced susceptibility to virulent Pseudomonas syringae pv. tomato (Pst) DC3000 and Hyaloperonospora arabidopsidis (Hpa) Noco2. Together, our data indicate that pepper MLO functions as a general negative regulator of plant immunity against hemibiotrophic bacterial pathogens, such as Xcv and Pst, as well as the biotrophic oomycete pathogen Hpa.


Isolation and sequence analysis of CaMLO2

The CaMLO2 gene was identified in a cDNA library constructed from pepper leaves inoculated with the avirulent Xanthomonas campestris pv. vesicatoria (Xcv) strain Bv5-4a using the macroarray method (Jung and Hwang, 2000). The 1885 bp full-length CaMLO2 cDNA contains an open reading frame (ORF) encoding a 61 kDa protein with 528 amino acid residues (Figure S1). Pepper CaMLO1 (GenBank accession number AAX31277), which was identified by Panstruga (2005), is 56% identical to CaMLO2 (accession number JN896629) (Figure S2). CaMLO2 was predicted to be homologous to MLO proteins. The CaMLO2 amino acid sequence is predicted to contain seven transmembrane domains distributed throughout the sequence, and a CaM-binding domain in the C-terminal region. The deduced amino acid sequence of CaMLO2 shares 87% identity with the tomato MLO homolog LeMLO1 (accession number AAX77013; Panstruga, 2005), and is 48–56% identical to Arabidopsis MLOs (AtMLO2, accession number Q9SXB6; AtMLO6, accession number Q94KB7; Devoto et al., 2003), turnip MLO (BrMLO1, accession number AAX77014; Panstruga, 2005), and barley MLO (HvMLO1, accession number P93766; Büschges et al., 1997). Over 30 plant genes encoding MLO homologs have been identified in dicot and monocot plant species, including barley, rice, Arabidopsis, grape vine, tomato and pepper (Panstruga, 2005). The phylogenetic tree revealed five sub-families within the dicot MLOs, and a single separate group containing the monocot MLOs from barley and rice (Figure S3a).

Comparative analysis of the CaMBDs from some reported CaM-binding proteins have enabled an investigation of whether multiple sequence motifs are required for CaM complex formation (Rhoads and Friedberg, 1997). A putative CaMBD was located within the last cytosolic region of CaMLO2, between Ala413 and Arg430 (Figure S3b). The consensus Ca2+-dependent CaM-binding motif, designated the 1-8-14 motif, contains three hydrophobic amino acid residues at positions 1, 8 and 14, and a basic amphiphilic α-helical region. Trp419, which is known to play a significant role in CaM binding to many CaM target proteins, is conserved in the CaMBD region of all MLO proteins (Figure S3b). A helical wheel projection revealed that the CaMBD of CaMLO2 is amphiphilic (Figure 1a). Taken together, these characteristics support typical CaM binding by CaMLO2.

Figure 1.

 CaM-binding activity of the CaM-binding domain (CaMBD) of CaMLO2.
(a) Helical wheel projection of the putative CaM-binding region of CaMLO2. Hydrophobic, hydrophilic/neutral and hydrophilic/charged amino acid residues are indicated by asterisks, dots and plus symbols, respectively. The dashed line indicates that the helix is amphipathic, with one side being hydrophilic and the other side hydrophobic.
(b) CaMBD of CaMLO2 directly binds to CaM in the presence of Ca2+. The GST-fused CaMBD of CaMLO2 was expressed in E. coli and purified using a glutathione–Sepharose 4B column. Biotin–CaM did not interact with GST (control). The presence of GST–CaMBD was confirmed by immunodetection using anti-GST antibody.

Ca2+-dependent binding of CaMBD to CaM

To determine the biochemical activity of CaMBD in CaMLO2, we performed a CaM-binding assay using the biotin–CaM overlay method (Figure 1b). We generated a GST–CaMBD fusion construct containing the presumptive binding site. The recombinant protein was expressed in Escherichia coli, separated by SDS–PAGE, and transferred to polyvinylidene difluoride membranes for use in immunoblotting and the biotin–CaM overlay assay. Expression of the GST recombinant protein was confirmed via immunoblotting using GST antibody. The GST fusion protein harboring CaMBD of CaMLO2 interacted with the biotin–CaM conjugate. Biotin–CaM bound to CaMBD in the presence of Ca2+ but not in the presence of a calcium chelator, EGTA, demonstrating that CaM binds to the predicted CaMBD in the C-terminal region of CaMLO2 in a Ca2+-dependent manner (Figure 1b).

Subcellular localization of CaMLO2

The topology of CaMLO2 was further analyzed using the dense alignment surface-transmembrane filter (DAS-TM) server ( The DAS transmembrane analysis predicted seven transmembrane regions within CaMLO2 (Figure 2a). To determine the subcellular localization of CaMLO2, we constructed a soluble-modified GFP (smGFP)-fused CaMLO2 ORF construct under the control of the CaMV 35S promoter (Kim and Hwang, 2011). Using confocal microscopy, control smGFP was uniformly observed throughout onion (Allium cepa) epidermal cells. However, the CaMLO2–smGFP fusion protein was primarily observed in the cell periphery (Figure 2b), indicating that the transmembrane domains of CaMLO2 may be responsible for CaMLO2 localization in the cell periphery. The 35S:CaMLO2:smGFP construct was also expressed in Nicotiana benthamiana leaves using the Agrobacterium-mediated transient system. As shown in Figure 2(c), GFP signals were detected exclusively in the plasma membrane of N. benthamiana leaves. To support this result, biochemical fractionation was used to separate the cytosolic and membrane fractions. Immunoblotting using antibody against GFP demonstrated the presence of CaMLO2–GFP in the membrane fraction and the total extract, but not in the soluble cytosolic fraction (Figure 2d). Collectively, the subcellular localization data indicate that the seven transmembrane domain-harboring CaMLO2 protein is localized in the plasma membrane.

Figure 2.

 Subcellular localization of CaMLO2.
(a) Topology of CaMLO2. The transmembrane region of CaMLO2 was predicted using the DAS-TM prediction server. The curves were obtained by pairwise comparisons of the proteins in the test set in an ‘each against the rest’ manner. Two cut-offs are indicated on the plots: ‘strict’ at a DAS score of 2.2 and ‘loose’ at a DAS score of 1.7. A hit at 2.2 is informative in terms of the number of matching segments, while a hit at 1.7 gives the actual location of the transmembrane segment.
(b) Transient expression of smGFP and CaMLO2:smGFP in onion epidermal cells 24 h after biolistic transformation, as observed by confocal laser scanning microscopy. Scale bars = 100 μm.
(c) Transient expression of CaMLO2:smGFP in Nicotiana benthamiana 3 days after agro-infiltration, as observed by confocal laser scanning microscopy. Scale bar = 100 μm.
(d) Immunoblot analysis of CaMLO2 transiently expressed in N. benthamiana leaves. Total (T), soluble (S) and membrane (M) fractions were subjected to anti-GFP immunoblotting. Equal loading of proteins was confirmed by Ponceau S staining.

Induction of CaMLO2 by biotic and abiotic stresses

Next, we investigated the transcriptional regulation of CaMLO2 expression in response to pathogen invasion, plant hormones and abiotic stimuli. CaMLO2 expression was distinctly induced in pepper leaves by Xcv infection (Figure 3a). Infection with the virulent strain Ds1 induced CaMLO2 expression in pepper leaves 20 and 25 h after inoculation. Notably, infection with the avirulent strain Bv5-4a was more effective at triggering CaMLO2 induction compared to infection with the virulent Xcv strain. As shown in Figure 3(b), CaMLO2 transcripts accumulated in pepper stems 3–7 days after inoculation with the oomycete pathogen Phytophthora capsici isolate S197. Application of exogenous salicylic acid (SA) caused a distinct increase in CaMLO2 expression in pepper leaves. By contrast, methyl jasmonate treatment did not induce CaMLO2 expression (Figure 3c). CaMLO2 expression was also up-regulated by methyl viologen, NaCl and drought stimuli (Figure 3d).

Figure 3.

 RNA gel-blot analyses of CaMLO2 expression in pepper plants at the six-leaf stage.
(a) Time courses of induction of CaMLO2 in leaf tissues after inoculation with virulent (compatible) strain Ds1 and avirulent (incompatible) strain Bv5-4a of Xanthomonas campestris pv. vesicatoria (108 cfu ml−1).
(b) Induction of CaMLO2 in stem tissues after inoculation with Phytophthora capsici isolate S197.
(c) Time courses of CaMLO2 expression in leaf tissues treated with salicylic acid (SA, 5 mm) and methyl jasmonate (MJ, 100 μm).
(d) Time courses of CaMLO2 expression in leaf tissues treated with methyl viologen (MV, 100 μm), sodium chloride (NaCl, 100 mm) or drought.
In each case, the gel blot was hybridized with a 32P-labeled full-length CaMLO2 cDNA probe. Equal loading (20 μg) was verified by visualizing rRNA on the gel stained with ethidium bromide.

Enhanced resistance of CaMLO2-silenced plants to Xanthomonas campestris pv. vesicatoria infection

To examine CaMLO2 loss-of-function in plants, we generated empty vector control (TRV:00) and CaMLO2-silenced (TRV:CaMLO2) pepper plants using virus-induced gene silencing (VIGS) (Choi et al., 2007; An et al., 2008; Hwang and Hwang, 2010). The 295 bp fragment within the 3′ region of the CaMLO2 cDNA (nucleotides 1476–1770, Figure S1) or the full-length CaMLO2 ORF (1587 bp) were cloned into the TRV vector (TRV:CaMLO2C or TRV:CaMLO2). Efficient silencing of CaMLO2C or CaMLO2 was confirmed by RT-PCR analyses of the CaMLO2 transcript levels in pepper plants infected by Xcv (Figures 4a and S4a). CaMLO2 expression was nearly abolished in the TRV:CaMLO2C and TRV:CaMLO2 pepper plants during both virulent and avirulent Xcv infection. Interestingly, however, induction of CaMLO1, a paralog of CaMLO2 (Panstruga, 2005), by Xcv infection was not compromised in the TRV:CaMLO2C pepper plants. These RT-PCR data indicate that CaMLO2 silencing was specific and successful.

Figure 4.

 Enhanced resistance of CaMLO2-silenced pepper plants to infection by virulent (compatible) strain Ds1 of Xanthomonas campestris pv. vesicatoria (Xcv).
(a) RT-PCR analysis of CaMLO2 expression in empty vector control (TRV:00) and CaMLO2-silenced (TRV:CaMLO2C) leaves 1 day after inoculation with virulent (compatible) strain Ds1 and avirulent (incompatible) strain Bv5-4a of Xanthomonas campestris pv. vesicatoria (Xcv).
(b) Bacterial growth in leaves 0 and 3 days after inoculation with Xcv (5 × 104 cfu ml−1).
(c) Real-time RT-PCR analysis of expression of CaMLO1, CaMLO2, the defense-related SA-dependent gene CaPR1 and the JA-dependent gene CaDEF1 in leaves after inoculation with Xcv (107 cfu ml−1).
Values in (b) and (c) are means ± standard deviations from three independent experiments. Asterisks indicate significant differences, as determined by Student’s t test (*< 0.05).

We further analyzed the effect of CaMLO2C or CaMLO2 silencing on Xcv growth in pepper leaves (Figures 4b and S4b). Proliferation of virulent Xcv Ds1 was significantly decreased in CaMLO2C- or CaMLO2-silenced leaves compared to leaves harboring the empty vector control. However, CaMLO2C or CaMLO2 silencing did not significantly affect avirulent Xcv growth. Antioxidants such as ferulic acid in monocotyledonous cell walls are known to fluoresce intense blue under long-wavelength UV light (Rudall and Caddick, 1994). Antioxidant phenolic compounds were detected as a light blue color under UV light in empty vector leaves inoculated with virulent (compatible) Xcv (107 cfu ml−1) (Figure S4c). By contrast, CaMLO2-silenced leaves did not exhibit any disease symptoms and phenolic antioxidants under visible light and UV conditions, respectively, 6 days after inoculation with compatible Xcv. Avirulent Xcv infection caused localized hypersensitive response cell death, as visualized under visible light and/or UV conditions.

Using quantitative real-time PCR, we next investigated whether silencing of CaMLO2C or CaMLO2 affects defense-related genes CaPR1 and CaDEF1 (defensin) as well as CaMLO2 and CaMLO1 (Figures 4c and S4d). Similar to RT-PCR (Figures 4a and S4a), CaMLO2 induction by Xcv infection were significantly compromised in CaMLO2-silenced plants; however, CaMLO1 expression was not compromised in CaMLO2-silenced plants (Figures 4c and S4d). Upon infection with virulent Xcv but not with avirulent Xcv, CaPR1, a representative SA-dependent marker gene, was more significantly up-regulated in CaMLO2C- or CaMLO2-silenced leaves compared to empty vector control leaves. By contrast, CaMLO2C and CaMLO2 silencing did not induce CaDEF1, a jasmonic acid (JA)-dependent gene that was significantly up-regulated in empty vector control pepper leaves infected by virulent Xcv (Figures 4c and S4d). Notably, the transcript levels of CaPR1 and CaDEF1 were similar in the empty vector control and silenced leaves during avirulent Xcv infection, indicating that silencing of CaMLO2C or CaMLO2 did not alter expression of these genes during avirulent Xcv infection of pepper. Together, these results indicate that silencing of CaMLO2C or CaMLO2 confers enhanced basal defense to virulent Xcv infection. Additionally, the defense marker genes CaPR1 and CaDEF1 are antagonistically regulated by silencing of CaMLO2C or CaMLO2 in pepper plants.

Induction of H2O2 accumulation by virulent Xcv infection was more pronounced in CaMLO2-silenced pepper leaves compared to empty vector control leaves (Figures 5a,b and S5a,b). This effect was not observed in plants infected by avirulent Xcv. In addition, avirulent Xcv infection caused a distinct induction of cell-death response phenotypes. These phenotypes were similar in both empty vector control and CaMLO2-silenced leaves, as observed by trypan blue staining (Figures 5c and S5c). CaMLO2 silencing also significantly induced electrolyte leakage from leaf tissues during virulent Xcv infection, but not during avirulent Xcv infection (Figures 5d and S5d). This indicates that CaMLO2 silencing confers enhanced resistance to virulent Xcv infection.

Figure 5.

 Quantification of ROS and cell death in empty vector control (TRV:00) and CaMLO2-silenced (TRV:CaMLO2C) pepper leaves infected with virulent (compatible) strain Ds1 and avirulent (incompatible) strain Bv5-4a of Xanthomonas campestris pv. vesicatoria (Xcv).
(a) Detection of H2O2 in leaf tissues stained with DAB 1 day after inoculation with Xcv (107 cfu ml−1). Scale bars = 500 μm.
(b) Quantification of H2O2 in leaf tissues at various time points after inoculation with Xcv (107 cfu ml−1).
(c) Detection of cell death in leaf tissues stained with trypan blue 1 day after inoculation with Xcv (107 cfu ml−1). Scale bars = 200 μm.
(d) Electrolyte leakage from leaf discs at various time points after inoculation with Xcv (107 cfu ml−1).
Values in (b) and (d) are means ± standard deviations from three independent experiments. Asterisks indicate significant differences, as determined by Student’s t test (*< 0.05).

Enhanced susceptibility of CaMLO2-OX Arabidopsis plants to Pseudomonas syringae and Hyaloperonospora arabidopsidis infection

To investigate CaMLO2 gain-of-function in planta, we ectopically expressed CaMLO2 in the Arabidopsis ecotype Columbia (Col-0) using the floral-dipping method (Clough and Bent, 1998). CaMLO2 over-expressing (OX) Arabidopsis lines 1, 2 and 4, which constitutively express CaMLO2, were used in this study (Figure 6a). As shown in Figure 6(b), CaMLO2-OX Arabidopsis leaves exhibited symptoms of high susceptibility 2 days after inoculation with Pseudomonas syringae pv. tomato (Pst) DC3000 (107 cfu ml−1). However, disease symptoms were not observed in wild-type leaves. Consistent with these disease phenotypes, CaMLO2 over-expression significantly enhanced Pst DC3000 growth in transgenic plants compared to wild-type plants (Figure 6c). By contrast, no significant difference in Pst DC3000 (avrRpm1) growth was found between wild-type and CaMLO2-OX transgenic leaves (Figure 6c).

Figure 6.

 Enhanced susceptibility of CaMLO2-OX transgenic Arabidopsis to Pseudomonas syringae pv. tomato (Pst) DC3000 infection.
(a) RT-PCR analysis of ectopic over-expression of CaMLO2 in wild-type (WT) and CaMLO2-OX transgenic plants. ACT1 was used as an internal control.
(b) Disease symptoms developed on leaves 2 days after inoculation with Pst DC3000 and DC3000 (avrRpm1) (107 cfu ml−1).
(c) Bacterial growth in the leaves 0 and 3 days after inoculation with Pst DC3000 and DC3000 (avrRpm1) (105 cfu ml−1). Values are means ± standard deviations from three independent experiments. Statistically significant differences according to Fisher’s protected LSD test (< 0.05) are indicated by different letters above the bars.

We next investigated whether CaMLO2 over-expression triggers ROS accumulation and electrolyte leakage in leaves of wild-type and transgenic Arabidopsis lines 1, 2 and 4 infected with Pst DC3000 (Figure 7). During virulent Pst DC3000 infection at the early stage (0.5 and 1 days after inoculation), H2O2 accumulation was compromised in CaMLO2-OX leaves. However, the levels of H2O2 induction at later stages (2–5 days after inoculation) were similar in wild-type and CaMLO2-OX leaves (Figure 7a). By contrast, avirulent Pst DC3000 (avrRpm1) infection drastically induced H2O2 accumulation in both wild-type and CaMLO2-OX leaves at all time points after infection (Figure 7a). These ROS bursts indicate that the susceptibility cell-death response caused by virulent Pst DC3000 infection does not trigger H2O2 induction in CaMLO2-OX leaves. Interestingly, virulent Pst DC3000 infection significantly promoted ion leakage from CaMLO2-OX leaves in comparison to wild-type leaves 3–5 days after inoculation, consistent with the timing of appearance of the susceptibility cell-death response (Figure 7b). During avirulent Pst DC3000 (avrRpm1) infection, the ion leakage levels were similar in wild-type and CaMLO2-OX leaves. These results strongly support the notion that the high levels of ion leakage from CaMLO2-OX leaves may result from the susceptible but necrotic cell-death response at the late infection stage.

Figure 7.

 Quantification of H2O2 (a) and electrolyte leakage (b) from leaf tissues of wild-type and CaMLO2-OX transgenic Arabidopsis plants at various time points after inoculation with Pst DC3000 and DC3000 (avrRpm1) (107 cfu ml−1).
Values are means ± standard deviations from three independent experiments. Statistically significant differences according to Fisher’s protected LSD test (< 0.05) are indicated by different letters above the bars.

We further investigated whether CaMLO2 over-expression regulates defense response genes in the three CaMLO2-OX Arabidopsis lines (Figure 8). Induction of the SA-dependent pathogen-related 1 (PR1) gene, the oxidative stress-induced marker gene glutathione S-transferase 1 (GST1) and the JA-dependent gene plant defensin 1.2 (PDF1.2), but not respiratory burst oxidase protein homolog D (RbohD) and the JA- and/or fungal-responsive gene vegetative storage protein 2 (VSP2), was significantly compromised in CaMLO2-OX leaves of all tested transgenic lines during virulent Pst DC3000 infection (Figure 8). By contrast, induction of these defense response genes in wild-type and CaMLO2-OX plants following infection with avirulent Pst DC3000 (avrRpm1) was not significantly different, with the exception of the compromised induction of GST1 in the CaMLO2-OX plants 2 days after inoculation. These results indicate that CaMLO2 over-expression does not enhance induction of these defense response genes during avirulent Pst infection.

Figure 8.

 Real-time RT-PCR analyses of the expression of defense-related genes in wild-type and CaMLO2-OX Arabidopsis plants 1 and 2 days after inoculation with Pseudomonas syringae pv. tomato (Pst) DC3000 or Pst DC3000 (avrRpm1) (107 cfu mL−1).
PR1, pathogenesis-related protein 1; RbohD, respiratory burst oxidase protein D; GST1, glutathione S-transferase; PDF1.2, plant defensin 1.2; VSP, vegetative storage protein. Expression of the internal control Actin was normalized to 1. Values are means ± standard deviations (= 3). Statistically significant differences according to Fisher’s protected LSD test (< 0.05) are indicated by different letters above the bars. n.d., not determined.

We also investigated whether CaMLO2 over-expression alters Hyaloperonospora arabidopsidis (Hpa) infection. CaMLO2-OX plants were more susceptible to Hpa infection compared to wild-type plants (Figure 9a). More prolific hyphal growth was observed on the cotyledons of CaMLO2-OX plants compared to wild-type plants (Figure 9b). The number of sporangiophores and spores produced on the cotyledons were also greater in CaMLO2-OX plants than in wild-type plants (Figure 9c,d). These results indicate that CaMLO2 is involved in oomycetal hyphal growth and the production of sporangiophores and spores.

Figure 9.

 Increased susceptibility of CaMLO2-OX plants to Hyaloperonospora arabidopsidis isolate Noco2.
(a) Disease symptoms 5 days after inoculation (5 × 104 conidiospores per ml). Scale bars = 1 mm.
(b) Trypan blue staining of oomycete hyphal structure 3 days after inoculation. Scale bars = 500 μm.
(c) Quantification of asexual sporangiophores on at least 50 cotyledons of wild-type and CaMLO2-OX transgenic plants 5 days after inoculation.
(d) Mean number of spores on the cotyledons of wild-type and CaMLO2-OX transgenic plants.
Values in (c) and (d) are means ± standard deviations from three independent experiments. Different letters indicate statistically significantly different values (Fisher’s protected LSD test, < 0.05).


Plasma membrane proteins may be involved in defense mechanisms against biotic stresses (Veronese et al., 2006; An et al., 2008). These proteins usually contain one or more transmembrane domains within their N-terminus. MLO proteins, part of a plasma membrane-anchored protein family harboring seven transmembrane domains, are found exclusively in plants (Devoto et al., 1999, 2003; Chen et al., 2006). The barley mlo mutant exhibits durable broad-spectrum resistance to the obligate biotrophic fungus Blumeria graminis f.sp. hordei (Jørgensen, 1992). As a result, the mlo allele is widely adopted in breeding programs (Piffanelli et al., 2002; Stein and Somerville, 2002). Numerous genetic and cytological studies of barley mlo mutant plants suggest that the MLO proteins play a regulatory role in cell death and plant defense (Wolter et al., 1993; Freialdenhoven et al., 1996; Devoto et al., 1999). Recently, Humphry et al. (2010) found that MLO genes are co-expressed with a battery of genes involved in plant defense, suggesting that MLO genes have an authentic function in plant immunity.

CaMLO2 expression was strongly induced in pepper plants during the incompatible interaction with the hemibiotroph Xanthomonas campestris pv. vesicatoria (Xcv) and following exogenous SA application. By contrast, CaMLO2 silencing in pepper plants antagonistically regulated expression of SA-dependent CaPR1 (PR-1) and JA-dependent CaDEF1 (defensin). In CaMLO2-silenced pepper leaves during virulent Xcv infection, CaPR1 expression was induced but CaDEF1 expression was compromised. These findings support a putative role for CaMLO2 in plant SA-dependent defense. The antagonistic relationship between the SA and JA signaling pathways is thought to enable plants with the ability to prioritize one pathway over another to effectively fine-tune the defense response to the invading pathogen (Beckers and Spoel, 2006). The increase in CaMLO2 transcripts following treatment with methyl jasmonate and drought also provides indirect evidence for an additional role for CaMLO2 in leaf senescence. More recently, jasmonic acid was shown to play a central role in programmed cell death, defense and leaf senescence (Reinbothe et al., 2009).

In this study, we provide experimental evidence for the Ca2+-dependent CaM-binding ability of the membrane-anchored CaMLO2. Plant cells produce Ca2+ signals of varying amplitudes, frequencies and durations. These signals are produced in response to a variety of external stimuli as a means to control many cellular processes (Knight et al., 1997; Trewavas and Malho, 1998). CaM proteins play a pivotal role in Ca2+ signal transduction (Bouche et al., 2005). Mechanistic knowledge regarding CaM-binding proteins related to the plant defense response may be crucial to understanding the precise role of CaM in plants. Ca2+-dependent CaM-binding motifs comprise two major groups: the 1-8-14 and 1-5-10 motifs, in which the numbers indicate the positions of conserved hydrophobic residues (Rhoads and Friedberg, 1997). CaM binds to the predicted CaMBD located within the C-terminal region of CaMLO2 in a Ca2+-dependent manner, demonstrating that CaMLO2 is a potential CaM-binding protein. CaM is known to be required for full MLO activity (Kim et al., 2002b). The MLO–CaM interaction was previously demonstrated in vivo using FRET microscopy (Bhat et al., 2005). The MLO–CaM complex interaction is constitutive and dynamic in the absence of the pathogen. Furthermore, pathogen entry to the host cells triggered an increase in the MLO–CaM interaction, suggesting that the pathogen-triggered increase in the MLO–CaM interaction requires successful fungal invasion into plant cells (Bhat et al., 2005).

Plant MLO proteins are reminiscent of the most abundant class of seven-transmembrane domain-containing receptors in metazoans. This metazoan group, designated G protein-coupled receptors (GPCRs) (Horn et al., 1998; Devoto et al., 1999), constitute a superfamily of seven-transmembrane domain-containing receptors that receive a signal through the extracellular portion and transduce that signal to the cytosol. A putative Arabidopsis GPCR, GCR1, is the only plant GPCR that is known to share sequence signatures with mammalian GPCRs (Plakidou-Dymock et al., 1998; Chen et al., 2004). This suggests that plants lack a known GPCR family. However, it is possible that plants may have evolved a distinct class of GPCRs. Topology analysis revealed that CaMLO2 contains seven transmembrane domains located throughout the sequence, and smGFP-tagged CaMLO2 was detected at the plasma membrane in onion cells. The only abundant seven-transmembrane domain family in higher plants is an MLO family (Devoto et al., 2003).

Silencing of CaMLO2 in pepper plants conferred enhanced resistance against virulent Xcv infection, but not avirulent Xcv infection. By contrast, CaMLO2 over-expression in Arabidopsis resulted in enhanced susceptibility to both the hemibiotrophic pathogen Pseudomonas syringae pv. tomato (Pst) DC3000 and the biotrophic oomycete Hyaloperonospora arabidopsidis (Hpa) Noco2. Silencing of CaMLO2 compromised the necrotic phenotype associated with susceptibility cell death in response to virulent Xcv infection. This suggests that CaMLO2 is crucial for the induction of susceptibility cell death in pepper plants. In CaMLO2-silenced plants, early immune responses, such as H2O2 accumulation and increases in ion leakage, were elevated in leaf tissues during virulent Xcv infection. In comparison to empty vector control plants, virulent Xcv infection drastically induced expression of the SA-dependent pathogenesis-related marker gene CaPR1 in CaMLO2-silenced plants. In the case of barley mlo resistance to Bgh, MLO loss-of-function enhanced the cell wall-restricted H2O2 burst, as well as the pathogen-induced expression of PR genes (Peterhansel et al., 1997; Piffanelli et al., 2002). Collectively, these results suggest that CaMLO2 plays a pivotal role in abolishing stress-induced cell death and the H2O2 burst at the infection site. The stress-induced disruption in the balance between MLO levels and cell death-triggering signals is thought to dictate whether a cell dies (Stein and Somerville, 2002). This proposal is well supported by experimental data for CaMLO2 over-expression in Arabidopsis. During virulent Pst DC3000 infection, transgenic plants over-expressing CaMLO2 exhibited faster initiation of susceptibility cell death than wild-type plants. Interestingly, this susceptibility cell-death response did not trigger H2O2 induction, but instead triggered electrolyte leakage in CaMLO2-OX leaves. These data suggest that CaMLO2 over-expression confers enhanced susceptibility to Pst DC3000 infection, leading to the susceptibility cell-death response without ROS bursts. More recently, Lewis et al. (2012) identified Arabidopsis thaliana MLO2 as a target of the Pseudomonas syringae type III effector HopZ2, and found that MLO2 negatively contributes to resistance against Pst DC3000 in Arabidopsis and is required for HopZ2 virulence function. In compatible plant–microbe interactions, susceptibility cell death occurs relatively late during the course of infection. Host-controlled cell death is intimately associated with the onset of susceptibility cell death and symptom development (Greenberg and Yao, 2004). Taken together, the experimental evidence from the CaMLO2 loss-of-function and gain-of-function analyses suggests that CaMLO2 contributes to the susceptibility cell-death response and bacterial and oomycete proliferation in plants, but does not function in the incompatible hypersensitive response. However, molecular mechanisms underlying the lack of CaMLO2 function in the hypersensitive response against avirulent pathogenic bacterial infection require further investigatation.

Experimental procedures

Plant materials and pathogen inoculation

Pepper (Capsicum annum L., cv. Nockwang) and Arabidopsis thaliana ecotype Columbia (Col-0) plants were grown in a plastic tray containing soli mix (Soil conditioners perlite and vermiculite and loam soil, 1:1:3 v/v/v) at 26°C under a 14 h light (130 μmol photons m−2 sec−1)/10 h dark photoperiod and 60% relative humidity.

Xanthomonas campestris pv. vesicatoria (Xcv) strains Ds1 and Bv5-4a, whicha re virulent and avirulent to pepper plants, respectively, were cultured in YN broth (5 g yeast extract and 8 g nutrient broth per liter). The bacterial suspension was infiltrated into the leaves of pepper plants at the six-leaf stage using a needleless syringe. The infected plants were placed in a growth room at 28°C following incubation in a dew chamber for 16 h at 100% relative humidity to promote bacterial infection.

The virulent isolate S197 of Phytophthora capsici was grown on oatmeal medium (30 g oatmeal, 2% agar and 1 L H2O). The zoospores were harvested in sterile tap water and adjusted to a final concentration of 105 zoospores per ml. The stems of pepper plants at the six- or eight-leaf stage were wounded by making 0.5–1 cm longitudinal slits in the stem. A small quantity of cotton soaked in a zoospore suspension was attached to the stem wounds, and taped on to maintain moist conditions.

Pseudomonas syringae pv. tomato (Pst) DC3000 and DC3000(avrRpm1) was grown in King’s B broth (10 g peptone, 15 g glycerol, 1.5 g K2HPO4 and 5 mm MgSO4 per liter) containing 50 μg ml−1 rifampicin. Leaves of Arabidopsis plants were infiltrated with Pst DC3000 using a needless syringe.

Spores of Hyaloperonospora arabidopsidis (Hpa) isolate Noco2 were propagated and maintained by weekly sub-culture on 7-day-old Arabidopsis seedlings. A spore suspension (5 × 104 conidiospores ml−1) in sterile tap water containing 0.05% Tween 20 was sprayed onto 7-day-old seedlings. Inoculated plants were covered with plastic to maintain moisture, and were grown at 16°C in a plant growth chamber. The number of sporangiophores and spores on the cotyledons were counted to assess the disease severity at 7 days after inoculation.

Isolation and sequence analysis of CaMLO2 cDNA

We isolated pathogen-inducible cDNAs from a pepper cDNA library using a differential hybridization procedure (Jung and Hwang, 2000). One of the sequenced cDNA clones, the MLO homolog CaMLO2, was selected and inserted into the pCR2.1-TOPO vector (Invitrogen,

Amino acid alignment of CaMLO2 and other plant MLO homologs was performed using Clustal X (Thompson et al., 1997). Phylogenetic analysis of CaMLO2 and related plant MLO proteins was performed using MEGA version 4 (Tamura et al., 2007). The MLO proteins in other plants were identified using the BlastP algorithm at the NCBI website ( Accession numbers used in phylogenetic analysis are listed in Appendix S1.

A helical wheel projection was constructed using binding site analysis in the Calmodulin Target Database (; Yap et al., 2000).

CaM binding assay

The CaMBD of CaMLO2 was introduced and expressed in E. coli BL21 (DE3). Expression of the GST fusion proteins was induced by application of 0.3 mm isopropyl thio-β-d-galactoside for 12 h at 16°C. The bacterial cells were harvested and resuspended in 1× PBS buffer (0.14 m NaCl, 2.7 mm KCl, 10 mm Na2HPO4, 1.8 mm KH2PO4, pH 7.3). Recombinant GST–CaMBD fusion protein was purified using a glutathione–Sepharose 4B column (GE Healthcare,

The CaM-binding ability of the recombinant protein was assessed using an Affinity CBP fusion protein detection kit (Stratagene, A duplicate blot was probed using a biotinylated CaM conjugate as the primary probe, in the presence of 1 mm CaCl2 or 10 mm EGTA for determination of Ca2+-dependent CaM binding. After washing for 20 min, the blotted membrane with PBST (1× PBS with 0.1% Tween 20), the membrane was probed using streptavidin alkaline phosphatase as the secondary detection reagent. Bound biotin–CaM was visualized using the chemiluminescent substrate CDP-STAR® (Sigma-Aldrich,

Subcellular localization and immunoblot analyses of CaMLO2

The CaMLO2 coding region was introduced into the pBIN35S-GFP vector to generate a C-terminal smGFP-tagged fusion protein under the control of the CaMV 35S promoter. For biolistic transformation, onion (Allium cepa) epidermis was bombarded with gold particles coated with plasmid DNA using a Bio-Rad ( PDS-1000/He particle delivery system. Bombarded onion specimens were incubated for 24 h on 0.5× MS (Murashige and Skoog) agar medium, and observed using an LSM 5 Exciter confocal laser-scanning microscope (Zeiss,

For Agrobacterium-mediated transient expression of CaMLO2 in N. benthamiana, the binary vector pBIN35S:GFP was constructed, containing a C-terminal smGFP-tagged CaMLO2 fusion protein. The plasmids were transformed into Agrobacterium tumefaciens strain GV3101 by electroporation. The overnight culture was centrifuged and the harvested cells were diluted to an OD600 of 1.0 using infiltration buffer (10 mm MES, 10 mm MgCl2, pH 5.7). Silencing inhibitor, p19, was co-infiltrated to stabilize exogenous expression. Three days after infiltration, the epidermal cells were observed using a confocal laser scanning microscope.

Protein expression was confirmed via immunoblotting using anti-GFP rabbit polyclonal antibody (Santa Cruz Biotechnology, Total protein extracts were prepared by grinding 0.5 g leaf tissue in 2 ml buffer [10 mm Tris/HCl pH 7.0, 0.33 m sucrose, 1 mm EDTA, and plant protease inhibitor cocktail (Roche,]. The insoluble debris was pelleted by centrifugation at 20 000 g for 20 min at 4°C. To separate the soluble and membrane fractions for biochemical fractionation, the total protein fraction was diluted with CaCl2 to 20 mm, and centrifuged for 1 h at 100 000 g at 4°C. The fractionated proteins were subjected to SDS–PAGE, and immunoblotting analysis was performed using a standard method to detect GFP-tagged CaMLO2.

RNA gel-blot and real-time RT-PCR analyses

Total RNA was extracted from pepper and Arabidopsis plants using Isol-RNA lysis reagent (5 Prime, according to the manufacturer’s instructions. RNA gel-blot and real-time RT-PCR analyses were performed as described previously (Choi and Hwang, 2011; Choi et al., 2012). Gene-specific primer pairs used for the real-time RT-PCR analysis are listed in Table S1.

Virus-induced gene silencing (VIGS)

pTRV (tobacco rattle virus)-based VIGS was performed to generate CaMLO2 knockdown (TRV:CaMLO2) pepper plants (Choi et al., 2007). The CaMLO2 ORF was inserted into pTRV2 to generateTRV2:CaMLO2. The 295 bp 3′ region of CaMLO2 was also PCR-amplified using specific primers and inserted into pTRV2 to generate TRV2:CaMLO2C. The gene-specific primers are listed in Table S1. pTRV1 was combined with pTRV2 empty vector or pTRV2 carrying CaMLO2 or CaMLO2C (1:1 ratio) in Agrobacterium tumefaciens GV3101, and co-infiltrated into fully expanded cotyledons of pepper plants. Co-infiltrated plants were grown in a plant growth room at 25°C under a 14 h light/10 h dark photoperiod.

Arabidopsis transformation

Transgenic Arabidopsis plants expressing CaMLO2 were generated using the 35S:CaMLO2 binary vector and the floral dipping method (Clough and Bent, 1998). Transformants were selected on 0.5× MS agar plates containing 50 μg ml−1 kanamycin. Three transgenic Arabidopsis lines (numbers 1, 2 and 4) were confirmed by RT-PCR analysis using CaMLO2 gene-specific primers.

Trypan blue and DAB staining

Trypan blue staining was performed to visualize cell death and oomycete hyphae. Leaves inoculated with pathogens were boiled for 10 min in lactophenol/trypan blue solution (10 ml lactic acid, 10 g phenol, 10 ml glycerol, 10 mg trypan blue and 10 ml distilled water) diluted with an equal volume of ethanol. Stained leaves were de-stained in 2.5 g ml−1 chloral hydrate solution overnight. De-stained leaf tissues were mounted in 60% glycerol and photographed with a light microscope.

For DAB staining to visualize H2O2, inoculated leaves were detached and placed in 1 mg ml−1 solution of 3,3′-diaminobenzidine (DAB, pH 3.8) (Sigma) for more than 12 h. To develop the brownish coloration of the DAB polymer, stained leaves were boiled in 100% ethanol for 10 min.

Quantification of hydrogen peroxide and ion conductivity

H2O2 levels in leaf tissues were quantified using the xylenol orange assay (Gay et al., 1999; Kim and Hwang, 2011). The xylenol orange assay reagent was freshly prepared: 200 μl solution (25 mm FeSO4, 25 mm (NH4)SO4 in 2.5 m H2SO4) was added to 20 ml of 125 μm xylenol orange in 100 mm sorbitol. Eight leaf discs (0.8 cm in diameter) were floated on 1 ml of distilled water in a microtube for 1 h. After centrifugation for 1 min at 12 000 g, 100 μl of the supernatant was immediately added to 1 ml of xylenol orange assay reagent. The mixture was incubated for 30 min at room temperature. H2O2 amounts were measured by absorbance at 560 nm using a DU 650 spectrophotometer (Backman Coulter,

Cell death based on electrolyte leakage from leaf tissues was quantified by measuring ion conductivity (Mackey et al., 2002). At various time points after bacterial infiltration, eight leaf discs (1.4 cm in diameter) were excised, and washed for 30 min in 20 ml distilled water. After incubating the leaf discs for 3 h in 20 ml of distilled water, the ion conductivity was measured using a Sension7 conductivity meter ((HACH,


We thank S.P. Dinesh-Kumar (Department of Plant Biology, University of California at Davis, CA) for providing the vectors pTRV1 and pTRV2, and U. Bonas (Deparement of Genetics, Martin Luther Universität, Germany) for Agrobacterium tumefaciens strain GV3101. This work was supported by the Next-Generation BioGreen21 Program (Plant Molecular Breeding Center, grant number PJ008027), Rural Development Administration, Republic of Korea.