Mycorrhization Mitigates Disease Caused by “Candidatus Liberibacter solanacearum” in Tomato

Disease caused by the bacterial pathogen “Candidatus Liberibacter solanacearum” (Lso) represents a serious threat to solanaceous crop production. Insecticide applications to control the psyllid vector, Bactericera cockerelli Šulc (Hemiptera: Triozidae) has led to the emergence of resistance in psyllids populations. Efforts to select natural resistant cultivars have been marginally successful and have been complicated by the presence of distinct Lso haplotypes (LsoA, LsoB) differing in symptoms severity on potato and tomato. A potentially promising management tool is to boost host resistance to the pathogen and/or the insect vector by promoting mycorrhization. Here we tested the hypothesis that mycorrhizal fungi can mitigate the effect of Lso infection on tomato plants. The presence of mycorrhizal fungi substantially delayed and reduced the incidence of Lso-induced symptoms on tomato as compared to non-mycorrhized plants. However, PCR with specific Lso primers revealed that mycorrhization did not prevent Lso transmission or translocation to newly formed leaves. Mycorrhization significantly reduced oviposition by psyllids harboring LsoA and survival of nymphs from these eggs. However, mycorrhization had no effect on oviposition by psyllids harboring LsoB or the survival of nymphs from parents harboring LsoB. These findings indicate the use of mycorrhizal fungi is a promising strategy for the mitigation of disease caused by both LsoA and LsoB and warrants additional field testing.


Introduction
Diseases caused by "Candidatus Liberibacter solanacearum" (Lso) have had significant economic impacts on potato and tomato production in North and Central America and New Zealand [1,2]. Lso is a rod-shaped, Gram-negative bacterial species belonging to the α-proteobacteria class and the Rhizobiaceae family [3]. "Ca. Liberibacter" species are generally regarded as important emerging plant pathogens worldwide, and all of them rely on psyllid vectors for transmission [2]. Lso is related closely to "Ca. Liberibacter" species responsible for Huanglongbing (HLB), also known as citrus greening disease. These include: "Ca. L. asiaticus" (CLas), "Ca. L. africanus," and "Ca. L. americanus." plant productivity linked to enhanced nutrient uptake capacity and MIR associated with mycorrhization. We also investigate whether mycorrhization alters insect fertility or fitness that may be linked to a MIR effect. To test these hypotheses, a 2 × 3 factorial experiment was used to compare the health and growth of tomato plants with or without mycorrhization by R. irregularis, treated with B. cockerelli infected with either LsoA or LsoB, or receiving no insects. Symptom development and plant growth were measured over an 8-week period. Psyllid oviposition and nymph survival also were measured.

Plants
The tomato (Solanum lycopersicum L.) cultivar Moneymaker (Thompson & Morgan Inc., Jackson, NJ, USA) was used in this study. Seeds were surface-sterilized in diluted commercial bleach (3%) for 30 min and rinsed twice with sterilized water. Surface-sterilized seeds were pregerminated on water agar for 4-5 days. Pregerminated seeds were transferred to pots (10 × 10 × 10 cm) filled with an autoclaved mix (1:1, v:v) of turface (Turface Athletics MVP, Profile Products LLC, Buffalo Grove, IL) and sand (Brown Play Sand, Quikrete, Atlanta, GA) supplemented 1:100 g:g with slow-release rock phosphate. Plants were grown inside insect-proof cages (Bioquip Inc, Compton, CA) at room temperature with a photoperiod of 16h:8h (light:dark). Plants received 30 mL of sterile distilled water on alternating days and also 30 mL of 0.5 × Hoagland solution twice a week.

AM Fungi and Inoculation
The AM fungus utilized in this study is commercially distributed as Myke ® Pro Potato L (Premier Tech, Québec, Canada). One-hundred microliter of Myke ® Pro Potato L containing 10,500 viable spores/mL of R. irregularis were injected into the soil in the immediate vicinity of the root system of one-week-old plants. The control plants also received a 100 µL aliquot of a filtrate (obtained by 30 µm-sieving of the AMF inoculum in order to provide all nutritional and microbial components of the inoculum other than living AMF). Mycorrhization was quantified on 10 plants chosen randomly per inoculation treatment per experiment at three-and eleven-weeks after inoculation after trypan blue staining [36] as the percentage of the root length colonized [37].

Insects
Currently, there are four different B. cockerelli haplotypes in the United States [38], and the Northwestern haplotype was used in this study. Colonies of B. cockerelli harboring LsoA or LsoB were previously characterized and maintained as described by Mendoza-Herrera et al. [17]. Diagnostic PCR analyses to detect Lso in B. cockerelli were performed for all colonies as described previously [39,40], and confirmed infection status. Three weeks after AM fungus inoculation, three adult male and three adult female psyllids from the confirmed LsoA-or LsoB-harboring colonies were placed on a single leaf inside an "organza-bag" [15] in the middle tier of the shoot system and adult psyllids were removed two days later. Uninfested plants were maintained similarly, but no psyllids were added.

Experimental Design
The experiment consisted of a 2 × 3 factorial completely randomized design with two inoculation treatments: non-mycorrhized plants (NM) and mycorrhized plants (Ri) and three infestation treatments: psyllids harboring LsoA, psyllids harboring LsoB, and uninfested plants (no psyllids). There were 12 replicates of each inoculation treatment, totaling 72 plants (one plant per pot). The entire experiment was replicated twice between May 2015 and June 2016, and all experiments were teminated 8 WAI before the plants outgrew the insect-proof cages.

Assessment of the R. irregularis Effects on Disease Development and Plant Growth
After infestation, the disease symptom development was monitored until 8 WAI. The incidence and severity of foliar disease symptoms were scored 3, 6, 8 WAI based on the 0 to 4 scale, as described in Table 1. For incidence, a rating of 1-4 was scored as symptomatic, and severity estimates were made only on symptomatic plants. Growth measurements were taken at experiment termination on 10 randomly chosen plants per treatment and included root and shoot dry weights, plant height, and number of green leaves.

Detection of Lso on Top-Tier Leaves
In order to detect Lso movement from the site of infection, tissues from newly expanded leaves (mid-vein) were sampled from the top-tier of leaves at 3 and 6 WAI. DNA was extracted as described previously [15]. PCR-based detection was carried out using the Lso TX 16/23 primers to amplify 383 bp of the 16S-23S rDNA intergenic region, as described previously [41][42][43]. This primer set cannot discriminate between LsoA and LsoB. Positive PCR controls consisted of psyllid and tomato genomic DNA previously authenticated as infected by Lso, whereas the negative control was genomic DNA from an Lso-free tomato [15]. The tomato elongation factor-1 (EF1) gene primer set EF1 F/EF1 R was used to control for false negatives in plant samples [40].

Assessment of AM Effect on B. cockerelli Oviposition and Nymph Survival
Two days after infestation, the adult psyllids were removed from the leaves, and eggs were counted. One week later, living nymphs were counted on alternate days. Manipulation of plants for nymph counting (including removal and replacement of organza bags) sometimes resulted in injury or breakage of the leaf. Data from leaves that became injured were excluded. Young adults were removed as they emerged.

Statistical Analyses
For all analyses, data from both experiments were pooled. For percent root colonization, data were log-transformed to ensure normality. A general linear model (ANOVA) and Fisher's protected LSD were used to analyze transformed root colonization data and disease severity data. For the growth data, missing data were input using the multivariate singular value decomposition (SVD) method. Logistic regression was performed on growth data and showed that all growth variables were significant predictors of treatment effects. Growth data were analyzed using a generalized linear mixed model (GLMM) with Lso treatment, mycorrhizal treatment, and Lso × mycorrhizal treatment as fixed effects, and a Student's t was used for multiple comparisons. For psyllid life stages (eggs vs nymphs), means were compared at each observation date using the same fixed effects and multiple comparison statistic, and a multivariate analysis of variance (MANOVA) was used for comparisons across dates.

Estimation of Root Colonization by AM Fungi Rhizophagus Irregularis
Root colonization by R. irregularis was assessed twelve weeks after the inoculum was applied to the plants. Data for root length colonization was slightly, but significantly different among treatments, e.g., 82 ± 0.1 ≥ 75 ± 0.1 ≥ 69 ± 0.1%, for plants infested with LsoB-infected psyllids, no-psyllids, and LsoA-infected psyllids, respectively.

Rhizophagus Irregularis Delays the Development of Symptoms Associated with Lso Infection and Promotes Plant Growth under Disease Stress
Plants that were not infested with psyllids remained healthy during the entire experiment and displayed no disease symptoms. For both mycorrhized and non-mycorrhized plants infested with psyllids no disease symptoms were observed at 3 WAI (Table 2). By 6 WAI, both mycorrhized and non-mycorrhized plants displayed disease symptoms (e.g., distorted growth, especially shortened internodes or stunting), regardless of whether the psyllids harbored LsoA or LsoB. Wilt symptoms were observed first on non-mycorrhized plants (~3 days after 3 WAI). Disease incidence based on symptoms at 6 WAI was 100% for non-mycorrhized plants treated with Lso-infected psyllids compared to 42% for mycorrhized plants treated with Lso-infected psyllids ( Table 2). By 8 WAI, disease incidence based on symptoms for mycorrhized plants was 58%, and these results were consistent between experiments. However, there were no differences in the mean disease severity indices for non-mycorrhized versus mycorrhized plants that displayed disease symptoms (Table 2). These data indicate that mycorrhization reduced the incidence of disease, but once mycorrhized plants acquired symptoms associated with Lso infection, the symptoms developed to the same level of severity. Notably across both experiments, four of 14 mycorrhized plants that did not have any symptoms at 6 WAI, developed advanced symptoms by 8 WAI. In general, the symptoms because of LsoB were more severe than those because of LsoA (Table 2).  Lso was detectable via PCR by 3 WAI, and about half of the symptomatic plants were Lso positive. By 6 WAI, 75% of the symptomatic plants tested Lso positive via PCR, but only 50% of symptomless, psyllid infested plants were Lso positive (data not shown).
The analysis indicated a significant Lso treatment-by-mycorrhization treatment interaction effect (p < 0.05), but only for shoot dry weight biomass production, indicating that mycorrhization improved shoot biomass production of the uninfected plants, but not the Lso infected plants, which were significantly smaller than the uninfected plants (groups 'a' and 'b' vs. group 'c' in Figure 1A). For all other variables there was a significant main effect for mycorrhization treatment (p < 0.05), which resulted in greater root biomass, plant height, and leaf number in mycorrhized plants (NM vs. RI in Figure 1B-D). There also was a significant main effect for Lso treatment for these three variables (p < 0.05). Multiple comparisons indicated LsoB-infected plants had less root growth, were shorter, and produced/retained fewer leaves than the uninfected plants, and were shorter and produced fewer leaves than the LsoA-infected plants.

Effect of Mycorrhization on the Development of Nymphs from Parents Harboring Different Lso Haplotypes
Mycorrhization by R. irregularis also influenced psyllid oviposition success, but the effect of mycorrhization differed based on the Lso haplotype the psyllids harbored. In the absence of mycorrhization, psyllids harboring LsoA laid significantly more eggs than psyllids harboring LsoB. However, mycorrhization significantly reduced the number of eggs laid by psyllids harboring LsoA, but had no effect on the number of eggs laid by psyllids harboring LsoB (Table 3). All eggs hatched 9 DAI regardless of the treatments. As expected based on egg number, there were more nymphs from LsoA-infected B. cockerelli parents on non-mycorrhized plants than mycorrhized plants. However, there was a significant decrease in the number of living LsoA-infected nymphs from 9 DAI to 11 DAI on mycorrhized plants, but no such decline was observed on non-mycorrhized plants. Number of hatched eggs from LsoB-infected parents resulted in statistically similar nymph populations at 9 and 11 DAI, regardless of the mycorrhization treatment.

Effect of Mycorrhization on the Development of Nymphs from Parents Harboring Different Lso Haplotypes
Mycorrhization by R. irregularis also influenced psyllid oviposition success, but the effect of mycorrhization differed based on the Lso haplotype the psyllids harbored. In the absence of mycorrhization, psyllids harboring LsoA laid significantly more eggs than psyllids harboring LsoB. However, mycorrhization significantly reduced the number of eggs laid by psyllids harboring LsoA, but had no effect on the number of eggs laid by psyllids harboring LsoB (Table 3). All eggs hatched 9 DAI regardless of the treatments. As expected based on egg number, there were more nymphs from LsoA-infected B. cockerelli parents on non-mycorrhized plants than mycorrhized plants. However, there was a significant decrease in the number of living LsoA-infected nymphs from 9 DAI to 11 DAI on mycorrhized plants, but no such decline was observed on non-mycorrhized plants. Number of hatched eggs from LsoB-infected parents resulted in statistically similar nymph populations at 9 and 11 DAI, regardless of the mycorrhization treatment. The findings indicate that mycorrhization significantly reduced oviposition by psyllids harboring LsoA and survival of nymphs from these eggs. However, mycorrhization had no effect on oviposition by psyllids harboring LsoB or the survival of nymphs from parents harboring LsoB.

Discussion
Efforts to manage diseases of solanaceous plants caused by Lso have been complicated by the presence of two Lso haplotypes that differ in the intensity of disease symptoms they cause. In this study, we examined whether mycorrhization mitigates disease symptoms in tomato caused by both LsoA and LsoB haplotypes. To avoid bias because of vector genotype, all B. cockerelli insects used in this study were from the U.S. Northwestern haplotype. Disease symptoms developed in 100% of non-mycorrhized, psyllid-treated plants. As reported previously [17], disease symptoms in non-mycorrhized plants infected with LsoB were significantly more severe than in plants infected with LsoA. In the previous study, advanced symptom development was associated with higher Lso titer as measured via quantitative PCR and a distinct Lso distribution.
Our study demonstrated that mycorrhization consistently reduced the incidence of disease. Whereas 100% of non-mycorrhized plants developed significant disease, at least 40% of the mycorrhized plants were still without discernable disease symptoms at 8 weeks, and this was consistent in two separate experiments. The benefits of mycorrhization are particularly important for LsoB-infected tomatoes that typically do not survive beyond 8 weeks under greenhouse conditions [17]. We still do not know why some of the mycorrhized plants developed disease symptoms and 40% of the population did not. Our data indicate that mycorrhization did not prevent Lso infection, because Lso was detectable in the non-symptomatic plants. However, as reported previously [15], conventional PCR was not sensitive enough to detect Lso in the newly expanded leaves of all symptomatic plants, and we detected Lso in only~50% of the Ri-treated non-symptomatic plants. Future studies should investigate whether the reduced disease severity observed in AM inoculated plants is associated with a reduction in Lso titer or a modification of Lso distribution.
Previous reports suggested that mycorrhization may mitigate disease via improvements in plant vigor and MIR-mediated effects on disease progress, psyllid fitness, or both [35]. As expected, mycorrhization promoted shoot biomass production in the absence of Lso infection and root biomass, plant height, and leaf production/retention in the absence or presence of mycorrhization [22,23]. These findings are consistent with the hypothesis that mycorrhization contributed to plant vigor and the mitigation of LsoA-and LsoB-mediated disease symptoms [44]. Enhanced uptake of phosphorous is one of the major benefits provided by mycorrhizae to plants [45], and indeed injection of potassium phosphate (in combination with other compounds) into CLas-infected citrus trees suppressed HLB disease symptoms and reduced CLas titer [46].
A growing body of evidence suggests that mycorrhization elicits a MIR response that primes JA-dependent defense responses in plants [35]. Consistent with this hypothesis, mycorrhization has been shown to mitigate plant health problems caused by necrotrophs and chewing insects, whereas mycorrhized plants may be more susceptible to biotrophs and potentially phloem-feeding insects targeted by salicylic acid (SA) -regulated defenses [28,34]. This pattern correlates with an activation of JA-dependent defenses and repression of SA-dependent ones [35]. JA-dependent defense responses also may reduce insect growth, development, and survival on plants [47]. However, insect behaviors and development on plants are likely to be complicated by the effect of endosymbionts/plant pathogens on insect physiology as well as on JA and SA-regulated plant defenses once transmitted to the plant. Although not investigated in this no-choice experiment, previous studies demonstrated a positive correlation between plant mycorrhization and phloem-feeding performance [27,28,[30][31][32]35,48]. We investigated whether mycorrhization would influence other psyllid behaviors. Our study documented MIR-mediated changes in psyllid oviposition behavior, but this effect was strongly dependent on the Lso haplotype harbored by the parent. On non-mycorrhized plants, the number of eggs laid by females harboring LsoA was significantly greater than that observed for females harboring LsoB. On mycorrhized plants, the number of eggs laid by females harboring LsoA was no different from the low number of eggs laid by females harboring LsoB on either mycorrhized or non-mycorrhized plants. Mycorrhization also reduced the survival of nymphs hatching from the eggs of LsoA-infected parents. Nymph numbers from the eggs of LsoB-infected parents were generally low and not significantly different on non-mycorrhized and mycorrhized plants. We do not know the mechanism underlying the effects of MIR on the ovipositioning behavior of parents or the survival of nymphs harboring LsoA. We hypothesize that feeding nymphs might be directly affected by MIR induced production of volatile organics compounds, phytoalexins, or other defense-related molecules produced by plants in response to insect herbivory [49]. We speculate that different Lso haplotypes induce distinct responses in plants and insects, which influence the observed differences in psyllid behavior between LsoA and LsoB parents.
These results constitute the first report that mycorrhization mitigates disease on tomato Lso-caused by both LsoA and LsoB. These results reinforce the idea that mycorrhization confers resistance to phloem-limited pathogens [28,34]. However, Lso-mediated disease involves complex interactions between pathogen, vector, and host, and it is still unclear to us how mycorrhization is affecting these interactions. Future experiments under greenhouse conditions and ultimately in the field will be required to determine whether mycorrhization can protect plants sufficiently to improve marketable yield and to dissect whether there are differences in the outcome of yield trials because of different Lso haplotypes.