Closely-related Xanthomonas citri subsp . citri isolates trigger distinct histological and

Citrus canker, caused by Xanthomonas citri subsp. citri (Xcc), has an important economic impact on the citrus industry. Extensive information is available about the disease but, nevertheless, the study of plant-pathogen interactions could provide new information in the understanding of citrus canker disease. A new isolate has been identified, Xcc AT, which has a high genetic similarity (> 90 %) to the virulent Xcc T strain based on genetic clustering analyses of the rep-PCR fingerprinting patterns, but it does not produce cankerous lesions in Citrus limon. In this study, we compared C. limon responses to Xcc AT and to the virulent Xcc T strain at both histological and transcriptional levels. Histologically, leaves inoculated with Xcc AT exhibited neither a typical disordering of the spongy mesophyll, nor a swelling of epidermis. A particular content (undetermined) was also found in mesophyll cells near the stomata, together with increased starch accumulation. The transcriptomic profiles were compared by cDNA-AFLP technique. A total of 121 fragments derived from transcript (TDF) were either specifically induced or repressed by the isolates, and 62 were sequenced. Analysis of global expression identified different classes of genes known to be involved in plant-pathogen interactions. This study constitutes the first approach of the specific interaction between the avirulent Xcc AT isolate and C. limon.


Introduction
Asiatic citrus canker, caused by the bacterial plant pathogen Xanthomonas citri subsp.citri (Xcc), is a widespread disease that affects most commercial citrus species and varieties.Xcc infects hosts through natural openings, such as stomata and lesions in plant tissue, producing pustule-like cankers on leaves, stems and fruits (Brunings and Gabriel, 2003) and defoliation, blemished fruit, premature abscission and fruit drop in severe infections, all of which results in crop loss.Because Xcc is a quarantine pathogen in many countries, the impact of the disease on commerce is the most serious consequence, as there are restrictions on interstate and international fruit trade from canker-affected areas are imposed (Gottwald et al., 2002;2009).
The disease is present in the Americas, where it first appeared in Brazil, and soon spread to other important citrus producing countries, such as Argentina, one of the world's largest lemon producers.Extensive information is already available about the disease, its epidemiology, the susceptibility of host genotypes and, in particular, the Xcc pathogen (Gottwald, 1993;Gra-ham et al., 2004;Van Sluys et al., 2002;Vojnov et al., 2010).However. the study of plant-pathogen interactions could provide new information in the understanding of citrus canker disease.Chiesa et al. (2013) identified a new Xcc isolate with a high genetic similarity (> 90 %) to the virulent Xcc T strain based on genetic clustering analyses of the fingerprinting patterns generated by BOX, repetitive extragenic palindromic (REP), and enterobacterial repetitive intergenic consensus (ERIC) polymerase chain reaction (PCR) (repetitive elements-based [rep]-PCRs).Named Xcc A T , this isolate does not produce cankerous lesions in a host-specific interaction with C. limon, and is deficient in extracellular polysaccharide (EPS) xanthan production and induces an oxidative burst and a defense response in C. limon, independent of supplementation with exogenous xanthan (Chiesa et al., 2013).
In this study, we compared C. limon responses to Xcc A T and the virulent Xcc T strain at both histological and transcriptional levels.The close genetic relationship between these two isolates could serve to understand the mechanism underlying resistance and susceptibility.

Plant material and inoculum preparation
Six-month plants of Citrus limon (L.) Burm cv.Frost Eureka Nuclear were used to reproduce canker disease under controlled growth conditions.Plant leaves were inoculated with a bacterial suspension of Xcc A T or virulent Xcc T strain with a concentration of 10 8 CFU mL −1 , using swab methods to mimic a natural Xcc infection process (Sendín et al., 2011).Inoculation with magnesium chloride (MgCl 2 ) was used as control.After inoculation, plants were covered with plastic bags and kept overnight, and then transferred to a growth chamber at 28 °C, with 70 % relative humidity and a 16 h photoperiod.

Light microscopy and cDNA-AFLP analysis
For histological analyses, samples of inoculated leaves were taken at 0, 3 and 7 days post-inoculation (dpi).Cross sections of these samples were stained with safranin, methyl blue and iodine (D`Ambrogio de Argüeso, 1986), mounted on slides in a solution of glycerin / water (50/50) and observed in a light microscope.

Results and Discussion
Symptomatology and histology reveal differences in the interaction between C. limon and the closelyrelated Xcc A T and Xcc T Differential symptomatology was observed in young C. limon leaves treated with avirulent (Xcc A T ) or virulent (Xcc T) isolates.Seven days post-inoculation (7 dpi), Xcc T produced the typical circular water soaked lesions, in contrast with leaves inoculated with Xcc A T , where circular dark spots were observed (Figure 1).
The differential symptomatology observed in C. limon after inoculation with Xcc T or Xcc A T was histologically analyzed by light microscopy.In leaf samples inoculated with Xcc T, it was observed that cells around the stomata were disorderly, intercellular space in the spongy mesophyll region had diminished, and leaf epidermal cells had swollen by 3 dpi (Figure 2A, B).By 7 dpi typical hyperplasia and hypertrophy of mesophyll cells were observed, with a consequent rupture of abaxial epidermis and bacteria oozing at the site of infection (Figure 2D, E).In contrast, in leaf samples inoculated with Xcc A T (Figure 2G, H, J), as well as in control leaves (Figure 2M, N, P, Q), the spongy mesophyll did not show disordering and epidermal cells had not swollen either by 3 dpi nor 7 dpi.However, the mesophyll cells near the stomata showed a particular content by 7 dpi in leaves inoculated with Xcc A T (not determined) (Figure 2K).The starch accumulation was higher in guardian cells of stomata in leaves inoculated with Xcc A T , than in leaf tissue inoculated with Xcc T (Figure 2C, F, I, L).The highest amount of starch was observed at 3 dpi, but it dropped at 7 dpi (Figure 2I, L).In control leaves, cumuli of starch were not observed (Figure 2O, R).This reaction has been cited as a general plant response to different pathogen infections (Bertani et al., 2014;Pardo et al., 2012).Some evidence suggests that the wild Xcc strain presents in its  genome a gene coding for a protein similar to natriuretic plant proteins, which are involved in the regulation of cell growth and homeostasis (Gottig et al., 2008).This protein was shown to alter host physiological responses by means of starch granule degradation in the stomatal guard cells of the plant, causing stomatal openings thereof (Gottig et al., 2008).It is possible that this protein is defective or absent in Xcc A T , leading to starch accumulation in stoma cells which then hamper the entry of the pathogen.

The avirulent Xcc A T elicits a distinct gene expression pattern in the interaction with C. limon
To find out what genes are involved in the interaction between C. limon and the new Xcc A T isolate, a transcript profile was attained through cDNA amplified fragment length polymorphism (cDNA-AFLP) technique, as a first approach to elucidate certain of the mechanisms involved in this specifically induced interaction.This study was carried out 48 h post-inoculation (hpi), comparing the gene profile induced by the avirulent Xcc A T isolate with those produced by the virulent strain (Xcc T) or mock inoculation with MgCl 2 .Out of 752 fragments derived from transcripts (TDFs) observed in silver-stained cDNA-AFLP gels, 121 TDFs were specifically induced or repressed by virulent (Xcc T) and avirulent (Xcc A T ) isolates.Most of the up-regulated TDFs were observed when the plant was challenged with the virulent isolate Xcc T (47 %); the reverse was true for down-regulated TDFs, as their lowest number was recorded in this case (6 %) (Table 1).By contrast, in the interaction with the Xcc A T isolate, the numbers of up-regulated (26 %) and down-regulated (21 %) TDFs were similar (Table 1).A total of 62 differential TDFs sequences were successfully obtained.Table 1 shows the similarities of TDFs differentially expressed in response to either Xcc T or Xcc A T infection.The expression level of eight TDFs randomly selected from Table 1 was validated by qRT-PCR (Figure 3).As much as 18 % of the TDFs did not show any significant similarity to known proteins.The next most abundant group was composed of TDFs with significant similarity to resistance genes (15 %) and genes involved in plant metabolism (15 %).The other TDFs were identified as being similar to genes involved in transcription and translation (13 %), energy (10 %), pathway signaling (5 %), structural proteins (5 %) and transport mechanisms (2 %).Other TDF groups were linked with citrus EST (11 %) or uncharacterized proteins (8 %) in the database.
Some TDFs differentially expressed in the interactions are involved in the same mechanisms; for example, genes encoding proteins of the light-harvesting complex of photosystem II (PSII) (Table 1, Cit 26, Cit 32 and Cit 89).It has been shown that the PSII system plays an important role in preventing the accumulation of reactive oxygen species (ROS) (Asada, 1999), and it also seems to participate in both compatible and incompatible plant pathogen interactions (Baldo et al., 2010).
In the interaction with the virulent isolate, up-regulated genes involved in cell cycle, growth and differentiation (and related with hyperplasia and hypertrophy) were identified, as has been previously described (Cernadas et al., 2008).
Interestingly, one of the TDFs induced by Xcc T showed sequence similarity to an MYB transcription factor (Cit 102).Baldo et al. (2010) also found a TDF homologue to MYB 11 in Malus domestica, which was overexpressed in susceptible cultivars after inoculation with Erwinia amylovora.In the same way, two MYB genes in Arabidopsis thaliana were specifically induced by Type 3 effector proteins, and they negatively regulated phenylpropanoid metabolism (Preston et al., 2004).They are widely distributed in plants and have been reported to be involved in the ABA-response and to have interactions with other transcription factors (Ambawat et al., 2013).
In this interaction, one TDF similar to a lipoxygenase (Cit 29) was also induced.Lipoxygenase (LOX) has been identified as generating highly cytotoxic compounds like unsaturated aldehydes and lipid hydroperoxides, which interact with transition metals to produce corrosive radicals (Brash, 1999).These compounds could produce damage in the membrane by an alternative way to the oxidative burst produced during hypersensitivity reactions (HR) of an incompatible interaction (Adam et al., 1989;Porter et al., 1995;Mittler et al., 1996).However, the LOX mechanism has been proven to be associated with plant-pathogen interaction, and has been observed in both compatible and incompatible interactions (Jalloul et al., 2002).In concordance with Chiesa et al. (2013), we could say that it is possible that Xcc T inactivates ROS mechanism; nonetheless, the plant can activate another defense pathway, such as the LOX mechanism.
As expected, many of the TDFs expressed differentially during the interaction between C. limon and Xcc A T were found to be related in some way to resistance and plant defense response.Xcc A T induced a TDF similar to plant receptor-like serine threonine kinase (RLSTK) (Cit 34).This receptor recognizes molecules of pathogens and activates signals triggering plant defense responses; such is the case of rice bacterial blight resistance gene product XA21 (Song et al., 1995) and FLS2, a flagellin perception gene in A. thaliana (Gómez-Gómez and Boller, 2000).Another of the TDFs induced by Xcc A T was Cit 11, which shows similarity to the enzyme aminocyclopropanecarboxylate oxidase (ACC oxidase), which in turn catalyzes the cleavage of 1-aminocyclopropane-1-carboxylic acid (ACC) into ethylene.It is widely known that ethylene, together with other compounds such as salicylic acid, phytoalexins or reactive oxygen species (ROS), forms part of the complex plant defense mechanism (Joshi and Nayak, All data were normalized using a plant cytochrome oxidase (COX) gene.The mean expression value was calculated for every TDF with three replications.The corresponding signal in a cDNA-AFLP analysis is included for each target.(Cernadas et al., 2008).Interestingly, Xcc A T induced the expression of a TDF similar to a PRP38 splicing factor (Cit 108).Recently, new findings have provided insights into the roles of alternative splicing in the regulation of expression of plant defense genes (Marone et al., 2013).This factor has been generally associated with abiotic stresses such as cold and salinity (Forment et al., 2005;Iida et al., 2004); however, this is the first report of this gene in association with a biotic factor.
On the other hand, in the interaction between C. limon and Xcc A T a TDF similar to a gene coding for SNARE proteins (Cit 90) was suppressed.One of the primary roles of SNARE proteins is to mediate fusion between cellular transport vesicles and the cellular plasma membrane for exocytosis.A study using inhibitors of vesicle trafficking has shown that this inhibition delays pustule development induced by Xcc in sweet orange Pera leaves (Cernadas et al., 2008), so the suppression of this gene would be consistent with the lack of symptoms in Xcc A T inoculated leaves.
In conclusion, unlike the virulent type, the avirulent isolate Xcc A T did not prevent the development of physical barriers, such as starch accumulation or cell wall thickening, both events associated with plant defense mechanisms.Differences in gene expression patterns of infected plants, together with the close genetic relationship between these two isolates, could serve as the basis for developing different hypotheses related to the resistance or susceptibility mechanisms responsible for the different infection patterns that these two isolates exhibit.However, a high-throughput sequencing approach will be needed to get a more complete picture of differences in the plant molecular response to both Xcc isolates.

Figure 1 −
Figure 1 − Differential symptomatology induced in Citrus limon by Xcc T and Xcc A T .Foliar symptoms were observed at 7 days post-inoculation (dpi) in leaves inoculated with either Xanthomonas citri subsp.citri isolates, Xcc T (left) or Xcc A T (middle), and in mock inoculation with MgCl 2 (right).

Figure 2 −)
Figure 2 − Histopathologic analysis of Citrus limon leaves inoculated with virulent (Xcc T) or avirulent (Xcc A T ) Xanthomonas citri subsp.citri isolates on the 3 and 7 day post-inoculations (dpi).Crosssections of leaves inoculated with Xcc T (A, D) and Xcc A T (G, J), and control leaves (M, P); a detailed close-up of stomata is included as insert in A. Paradermal views of leaves inoculated with Xcc T (B, E), Xcc A T (H, K), or control leaves (N, Q).Histochemical Lugol test in leaves inoculated with Xcc T (C, F), Xcc A T (I, L) or control leaves (O, R).Arrows indicate stomata (st) or starch accumulation (sg); scale bars: 15 µm.

Figure 3 −
Figure 3 − Quantitative real time retrotranscription-PCR (RT-PCR) analysis of randomly selected transcript derived fragments (TDF) expressed diferentially during Xcc/ Citrus limon interaction.Relative expression of 8 randomly selected TDF (Cit 27, 64, 63, 3, 102, 62, 25 and 35) was quantified through quantitative real time RT-PCR using the 2 −ΔΔCt method to validate the expression level observed in cDNA-AFLP analysis.All data were normalized using a plant cytochrome oxidase (COX) gene.The mean expression value was calculated for every TDF with three replications.The corresponding signal in a cDNA-AFLP analysis is included for each target.

Table 1 -
Similarities of sequences of Citrus limon differentially expressed in response to either Xanthomonas citri subsp.citri isolates (Xcc T or Xcc A T ) infection.
*Those TDFs (fragments derived from transcripts) that not have similarities with known sequences were excluded from the table.

December 2016 2011
). ACC oxidase induction could indicate that ethylene increases with the interaction between C. limon and this avirulent isolate.The activation of this gene in citrus plants was previously described in sweet orange Pera inoculated with X. axonopodis pv.aurantifolii (Xaa) pathotype C, which does not produce canker in this genotype.