A critical role of STAYGREEN /Mendel’s I locus in controlling disease symptom development during Pseudomonas syringae pv. tomato infection of Arabidopsis

Production of disease symptoms represents the final phase of infectious diseases and is a main cause of crop loss and/or marketability. However, little is known about the molecular basis of disease symptom development. In this study, a genetic screening was conducted to identify Arabidopsis mutants that are impaired specifically in the development of disease symptoms (leaf chlorosis and/or necrosis) after infection with the bacterial pathogen Pseudomonas syringae pv. tomato ( Pst ) DC3000. An ethane-methanesulfonate (EMS)-induced Arabidopsis mutant ( noc1 ; for no chlorosis 1) was identified. In wild-type plants, the abundance of chlorophylls decreased markedly after Pst DC3000 infection, whereas the total amount of chlorophylls remained relatively unchanged in the noc1 mutant. Interestingly, noc1 mutant plants also exhibited reduced disease symptoms in response to the fungal pathogen Alternaria brassicicola . Genetic and molecular analyses showed that the nuclear gene STAYGREEN ( SGR or Mendel’s I locus) is mutated (resulting in the aspartate to tyrosine substitution at amino acid position 88) in noc1 plants. Transforming wild-type SGR cDNA into the noc1 mutant rescued the chlorosis phenotype in response to Pst DC3000 infection. The SGR transcript was highly induced by Pst DC3000, A. brassicicola, or coronatine, a bacterial phytotoxin that promotes chlorosis. The induction of SGR expression by coronatine is dependent on COI1, a principal component of the jasmonate receptor complex. These results suggest that pathogen/coronatine-induced expression of SGR is a critical step underlying the development of plant disease chlorosis.


INTRODUCTION
Tissue chlorosis and necrosis are among the most frequently observed disease symptoms associated with infection by diverse plant pathogens. These disease symptoms generally occur at the late stages of disease, but are highly relevant to agriculture because they cause actual damage to plant tissues, resulting in yield loss and/or poor marketability of crops. Symptomless infection is often associated with benign or symbiotic plant-microbe associations. Despite its importance in disease, the molecular basis of disease chlorosis and necrosis remain poorly understood.
Pseudomonas syringae pv. tomato (Pst) DC3000, the bacterial pathogen used in this study, is well known for causing localized necrosis and diffuse chlorosis in its hosts tomato and Arabidopsis (Ma and Cuppels 1991, Whalen et al. 1991, Katagiri et al. 2002. Many virulence genes are needed for Pst DC3000 to cause disease, among which two virulence systems have been studied extensively: the type III secretion system (T3SS) and the phytotoxin coronatine (COR). The T3SS delivers dozens of effector proteins into plant cells (Alfano and Collmer 2004, He et al. 2004, Buttner and He 2009). Many of these effectors suppress host immune responses (Boller and He 2009, Cui et al. 2009, Lewis et al. 2009) and some are also linked to production of disease necroses (Badel et al. 2003, DebRoy et al. 2004, Cohn and Martin 2005. COR, a molecular mimic of the plant hormone jasmonate, is not only involved in suppression of host immune responses, but also is important for the development of chlorosis symptoms (Feys et al. 1994, Mittal and Davis 1995, Bender et al. 1999, Kloek et al. 2001, Brooks et al. 2004, Block et al. 2005, Brooks et al. 2005, Melotto et al. 2008b). However, because effector-and COR-deficient bacterial mutants are generally affected in multiple steps of pathogenesis, it is often difficult to determine whether these virulence factors contribute to symptom development directly or indirectly through promoting bacterial colonization and/or multiplication.
An alternative approach to elucidate the molecular control of disease symptom production would be to isolate plant mutants that exhibit reduced disease necroses and/or chlorosis in response to pathogen infection. Indeed, numerous Arabidopsis constitutive defense mutants that show no or reduced disease symptoms to P. syringae infection have been isolated since the early 1990s (Bowling et al. 1994). However, similar to the situation with effector-or COR-defective Pst DC3000 mutants, bacterial populations in such plant mutants are often reduced compared to those in susceptible plants, making it difficult to conclude whether corresponding plant genes have a direct role in mediating symptom development or indirectly through affecting bacterial multiplication. There are a few exceptions: The ethylene-insensitive Arabidopsis, ein2, mutant, and the Arabidopsis mutant sgt1b (suppressor of G2 allele of skp1 b; affected in jasmonate signaling) exhibit reduced disease symptoms to Pst DC3000 infection without significantly affecting bacterial growth (Bent et al. 1992, Uppalapati et al. 2011 implicating the involvement of ethylene and jasmonate signaling in the production of Pst DC3000-elicited disease symptoms. How ethylene or jasmonate signaling lead to downstream disease symptom is not understood. To increase our understanding of the molecular basis of disease symptom development during Pst DC3000 infection of Arabidopsis, we conducted a genetic screen for Arabidopsis mutants that are reduced in disease symptom development, but not Pst DC3000 multiplication. In this paper, we report the identification and characterization of such an Arabidopsis mutant and cloning of the corresponding gene. Our results suggest that pathogen-responsive STAYGREEN (SGR)/NON-YELLOWING (NYE1)/Mendel's I locus plays a critical role in controlling disease chlorosis induced by Pst DC3000 and, interestingly, also by a fungal pathogen, Alternaria brassicicola.

Identification of the noc1 mutant
Approximately 10,000 EMS-mutagenized A. thaliana ecotype Columbia (Col-0) gl1 plants were screened for altered symptom development after the plants were dipped in a suspension containing 1x10 8 CFU/ml Pst DC3000 bacteria. One mutant isolated from this screen, noc1 (no chlorosis 1), was found to be defective in symptom development. In all experiments, noc1 leaves remained green whereas wild-type leaves began to show chlorosis between 48 and 72 hours after inoculation (Fig. 1A). In most experiments reduced severity of necrosis symptom was also observed in noc1 plants, although this phenotype was not as obvious as the lack-of-chlorosis phenotype. There were no noticeable differences between wild-type and noc1 plants in size, morphology, growth or development in the absence of pathogen inoculation. Most importantly, the reduction in disease symptoms in noc1 plants was not caused by reduced bacterial multiplication because Pst DC3000 populations in Col-0 gl1 and noc1 plants were similar at 1 and 3 days post-infection (dpi) (Fig. 1B). Thus, noc1 is a bono fide disease symptom mutant.
In addition to the noc1 mutant, we also isolated several other mutants that exhibited reduced disease symptom development during this study. However, these other mutants were dwarf and/or necrotic, suggestive of constitutive activation of non-specific disease resistance (Bowling et al. 1994). We did not conduct further characterization of such mutants.

Maintenance of the chlorophyll level in noc1 plants after infection with Pst DC3000
To quantify the chlorotic response to Pst DC3000, we conducted a chlorophyll abundance assay using leaf tissue infiltrated with 2x10 6 CFU/mL of Pst DC3000 and collected at 0, 24, 48, 72 and 96 hours post-inoculation (hpi). The results from one representative experiment are shown in Fig. 1C. Prior to inoculation with Pst DC3000, noc1 and wild-type plants had approximately equal amounts of total chlorophyll (25.2 mg/cm 2 and 28.6 mg/cm 2 , respectively). Wild-type plants began to lose chlorophyll by 24 hpi, with levels decreasing through 96 hpi. At 72 hpi, noc1 plants contained almost three times more chlorophyll than wild-type plants (27.0 mg/cm 2 in noc1 plants vs. 9.5 mg/cm 2 in wild-type plants). This experiment demonstrates that wild-type plants lose chlorophyll much faster than noc1 plants after Pst DC3000 infection.
The noc1 phenotype results from a single nucleotide change in AtSGR1 (At4G22920) To identify the NOC1 gene, noc1 plants were crossed with Ler plants and the F1 progeny were selfed to create an F2 population for mapping. The noc1 mutation shows normal Mendelian genetics and is recessive. We initially used bulk segregant analysis to analyze a pool of approximately 100 F2 individuals that exhibited the mutant phenotype (homozygous for the noc1 mutation). One marker, NGA107, located on the long arm of chromosome 4, showed linkage to the mutation. We tested a larger population of To determine whether the D88Y mutation in AtSGR is responsible for the noc1 phenotypes, we transformed noc1 plants with the full-length AtSGR cDNA from Col-0 gl1 cloned in pBAR-35S, which contains the CaMV 35S promoter and Basta (glufosinate)-resistance gene. Ten independent T2 lines were confirmed to exhibit Basta resistance and to harbor the AtSGR transgene. Three homozygous T3 lines were chosen for disease symptom observations after infection by Pst DC3000. In preliminary tests, all three were restored in chlorosis symptom development. We then focused on line #1 for further disease symptom and bacterial growth analyses. Results from this line are shown in Fig. 3. By 72-96 hpi, Pst DC3000-infected noc1/35S:AtSGR plants showed disease symptoms, including chlorosis, that were more pronounced than Pst DC3000-infected Col-0 gl1 plants ( Fig. 3A-C). Interestingly, in these experiments we noticed that noc1/35S:AtSGR plants did not support bacterial multiplication to the level observed in either noc1 or Col-0 gl1 plants at day 3 (Fig. 3E). This observation raised the possibility that accelerated disease symptoms may negatively affect Pst DC3000 growth and, accordingly, the noc1 mutant may allow better bacterial growth. However, we did not observe an enhanced Pst DC3000 growth at day 3 in noc1 plants (Figs. 1B and 4E). We commonly use a 3-day period for assessing Pst DC3000 multiplication in the laboratory; however, bacterial infection in the field involves longer durations. We therefore extended our multiplication assay to 6 days, which led to an interesting finding. Pst DC3000 populations declined in Col-0 gl1 and noc1/35S:AtSGR plants, as infected tissues senesced, whereas Pst DC3000 maintained a high population in the noc1 plants ( Fig. 3E). These results suggest that disease symptom development restricts Pst DC3000 persistence in infected tissues.

Effect of the noc1 mutation on A. brassicicola-induced chlorosis
To determine whether noc1 plants are also affected in disease symptoms caused by a fungal pathogen, Col-0 gl1 and noc1 plants were inoculated with spores of the necrotrophic fungus A. brassicicola. A necrotic lesion developed at the site of inoculation in Col-0 gl1, noc1, and complemented noc1/35S:AtSGR plants at 5 to 10 dpi. In some experiments, a chlorotic halo, surrounding the necrotic lesion, may also develop within 5-10 dpi in only Col-0 gl1 and noc1/35S:AtSGR plants (Fig. 4A). However, the chlorosis phenotype induced by A. brassicicola was variable between experiments and could not be quantified reproducibly. To quantify disease symptoms, we therefore measured necrotic lesion areas using Image J software. The area of necrotic lesion development was smaller in noc1 plants infected with A. brassicicola than that in Col-0 gl1 plants infected in the same manner ( Fig. 4A and 4B). Fungus-induced disease necroses were restored in noc1/35S:AtSGR plants infected with A. brassicicola ( Fig. 4A and 4B). Plants inoculated with buffer control (0.1% Tween 20) alone showed no signs of chlorosis or necroses (Fig.   4A). These results demonstrate that AtSGR1 is required for the development of disease symptoms caused by a necrotrophic pathogen.

AtSGR is highly induced during Pst DC3000 and A. brassicicola infection
Previous studies have shown that SGR expression is critical for initiation of developmentally regulated chlorophyll degradation in a number of plant species (Armstead et al. 2006;Armstead et al. 2007;Park et al. 2007;Ren et al., 2007). The requirement of AtSGR for disease chlorosis suggests that the expression of AtSGR might be induced during pathogen infection. To examine this possibility, we collected total RNA from H 2 0-and Pst DC3000-inoculated noc1 and Col-0 gl1 plants at 36, 48 and 60 hpi and performed northern blot analyses using an AtSGR-specific probe. We found that the AtSGR expression is strongly induced by Pst DC3000, but not H 2 O, in both Col-0 gl1 and noc1 plants at all sampled time points (Fig. 5A). Thus, AtSGR expression is regulated during Pst DC3000 infection and the noc1 mutation does not significantly affect this expression. RT-PCR analysis was also performed with RNA from A.
We next investigated whether specific virulence factors of Pst DC3000 could induce the expression of AtSGR. COR is a well-known bacterial virulence factor that promotes the development of disease chlorosis in plants (Uppalapati et al. 2005;Uppalapati et al. 2007;Ishiga et al. 2009). Recent studies have demonstrated that COR mimics the active form of the plant hormone jasmonate and directly targets the jasmonate receptor complex in which the COI1 F-box protein is a principal component (Katsir et al., 2008;Melotto et al., 2008a;Fonseca et al., 2009;Sheard et al., 2010). To determine whether COR could induce expression of AtSGR, we treated 8-day-old, Col-0 gl1 seedlings with buffer or 10 µM COR. Seedlings were collected at 3 hours post-treatment (hpt) and total RNA was isolated for each group. RT-PCR analysis showed that, like JAZ9 (a known

Second, previous reports show that bacterial phytotoxin COR induces chlorosis in plants
and that COR-deficient mutants have reduced ability to cause chlorosis (Ma and Cuppels, 1991;Whalen et al. 1991;Uppalapati et al. 2005;Uppalapati et al. 2007). Consistent with these observations, we found that COR was sufficient to activate AtSGR expression, providing a molecular basis for the involvement of COR in disease chlorosis.
Additionally, COR-treated coi1 mutants showed no induction of AtSGR expression (Fig.   6), suggesting that the jasmonate receptor complex is required for COR-induced expression of AtSGR. Interestingly, we found a putative MYC2-binding motif (ACGTG) AtSGR expression. It should be pointed out, however, that in Arabidopsis, purified COR was shown to induce purpling, instead of chlorosis (Bent et al., 1992). The exact reason for this phenomenon is not known. Because COR structurally and functionally mimics jasmonate, which is known to induce anthocyanin production in Arabidopsis (Ellis and Turner 2001), it is possible that the chlorosis induced by COR in Arabidopsis is masked by purpling associated with anthocyanin production. In any case, although COR is sufficient for induction of AtSGR, we found that it is not required for AtSGR induction during infection with Pst DC3000 (Fig. 6). Thus, it is likely that COR is not the only virulence factor in Pst DC3000 that is involved in the development of disease chlorosis, but that the action of additional virulence factors (possibly type III effectors) may also be involved in the induction of AtSGR and contribute to the production of disease chlorosis.
Indeed, disease chlorosis is often observed in Arabidopsis mutants that allow high multiplication of COR-deficient Pst DC3000 (Melotto et al. 2006, Zeng andHe 2010).
Third, in addition to affecting disease chlorosis, the noc1 mutation also reduced disease necrosis caused by Pst DC3000 and A. brassicicola infection, although this effect is less obvious compared to that of disease chlorosis (Figs. 1 and 4). This is an interesting finding in light of a recent study that investigated the effect of AtSGR overexpression and RNAi-mediated suppression on the hypersensitive response (HR) elicited by Pst DC3000 (avrRpm1) in Arabidopsis (Mur et al. 2010). It was found that increased and decreased AtSGR expression, respectively, accelerated and suppressed the kinetics of HR-associated cell death in resistant Arabidopsis plants. Mur and colleagues postulate that some phototoxic chlorophyll catabolites contribute to HR cell death in resistant plants (Mur et al. 2010). If so, we speculate that such chlorophyll catabolites could also contribute to the formation of disease necrosis in susceptible Arabidopsis plants infected by Pst DC3000 or A. brassicicola, as observed in our study (Figs. 1 and 4).
The stay-green phenotype was first described in 1866 by Gregor Mendel in differentiating yellow and green cotyledon color in segregating populations of pea (Mendel 1866). SGR orthologues have been cloned from a wide range of dicot and monocot species (Armstead et al. 2006;Armstead et al. 2007;Jiang et al. 2007;Park et al. 2007;Sato et al. 2007;Aubry et al. 2008;Barry et al. 2008). Consequently, we hypothesize that pathogen-induction of SGR genes may be a common mechanism underlying disease chlorosis across a wide spectrum of plant-pathogen interactions. As such, further study of SGR genes and their regulation could lead to transgenic plants with not only controlled senescence, but also disease symptom expression, thereby benefiting agriculture.

Plant material, mutagenesis and growth conditions
Approximately 1g of Arabidopsis thaliana ecotype Columbia-0 gl1 seeds was mixed with 100 ml of distilled water and 250 µl of ethylmethanesulfonate (EMS). The mixture was incubated overnight at room temperature in the dark with gentle agitation. The seeds were washed six times with 500 ml of distilled water, resuspended in 300 ml of 0.1% agarose and sown onto a soil mixture (equal portions of Baccto high-porosity professional plant mix, perlite and vermiculite, covered with a thin layer of fine vermiculite). The flats were covered with lids and incubated in the dark at 4˚C for three days. The flats were then transferred to a growth chamber [20˚C with 12 hours of fluorescent light (100 µEinsteins/m 2 /sec) and 12 hours of darkness]. The plants were self-fertilized to create a population of M2 plants.
The inoculated plants were incubated in high (80-90%) humidity conditions for 96 hours and screened for a lack of symptom development.

Bacteria enumeration in inoculated leaves of noc1 mutants and wild-type Col-0 gl1
plants Four to five week-old plants were used for bacteria enumeration. Pst DC3000 was grown in low-salt Luria-Bertani broth to the mid-to-late logarithmic phase at 28˚ C. Bacterial cultures were pelleted and resuspended in sterile water to a final OD 600 of 0.2 [equivalent to 1 x 10 8 colony-forming units (CFU)/ml] for dip-inoculation. Fully expanded leaves were dip-inoculated with bacterial suspensions. Plants were placed in trays with standing water and covered with plastic wrap to maintain high humidity. During the experimental period there was no obvious tissue desiccation in inoculated plants. Bacteria enumeration followed the protocol described by Katagiri et al. (2002). P values were derived from multiplication data utilizing Microsoft Excel software for t-test statistical analysis.

Alternaria brassicicola infection
We inoculated 10 leaves of each plant with 10 µl of A. brassicicola spores at a concentration of 6.4 x 10 5 spores/ml suspended in 0.1% Tween 20. For control inoculations we used 0.1% Tween 20. Plants were covered with humidity domes and kept at high humidity throughout the infection process.
For quantification of disease symptoms caused by fungal infection, the areas of necrotic lesions caused by Alternaria inoculation were measured on 30-infected plant leaves. The average lesion area was calculated from 30-50 lesions at 5 and 8 dpi using Image J software. Older leaves were excluded from the sample set to avoid senescence-associated chlorosis and necrosis.
AtSGR expression was measured in A. brassicicola-infected tissue at 0 and 5 days postinoculation (dpi). RNA was isolated from leaf tissue using an RNAeasy Kit (Qiagen, www.qiagen.com). Semi-quantitative RT-PCR was performed as described by the manufacturer (Takara, www.takara-bio.com).

Chlorophyll extraction and quantification
For chlorophyll abundance assays we infiltrated leaf tissue with 2 x 10 6 CFU/ml Pst