Comparative metabolomics of scab-resistant and susceptible apple cell cultures in response to scab fungus elicitor treatment

Apple scab disease caused by the fungus Venturia inaequalis is a devastating disease that seriously affects quality and yield of apples. In order to understand the mechanisms involved in scab resistance, we performed gas chromatography-mass spectrometry based metabolomics analysis of the cell culture of scab resistant cultivar ‘Florina’ and scab susceptible cultivar ‘Vista Bella’ both prior -to and -following treatment with V. inaequalis elicitor (VIE). A total 21 metabolites were identified to be altered significantly in ‘Florina’ cell cultures upon VIE-treatment. Among 21 metabolites, formation of three new specialized metabolites aucuparin, noraucuparin and eriobofuran were observed only in resistant cultivar ‘Florina’ after the elicitor treatment. The score plots of principal component analysis (PCA) exhibited clear discrimination between untreated and VIE-treated samples. The alteration in metabolite levels correlated well with the changes in the transcript levels of selected secondary metabolite biosynthesis genes. Aucuparin, noraucuparin and eriobofuran isolated from the ‘Florina’ cultures showed significant inhibitory effect on the conidial germination of V. inaequalis. The results expand our understanding of the metabolic basis of scab-resistance in apple and therefore are of interest in apple breeding programs to fortify scab resistance potential of commercially grown apple cultivars.

In the next step, we identified VIE-responsive metabolites that showed significant (p ≤ 0.05) differential concentrations between SR and SR0 and between SS and SS0. A total of 31 and 10 differentially accumulating metabolites were identified in 'Florina' and 'Vista Bella' cell cultures, respectively, as shown in Table 1. Statistically significant differentially accumulating metabolites detected from 'Florina' were further analyzed by employing Bonferroni correction (p ≤ 0.05) and their adjusted p-values were shown in Supplementary Table S2. Detected metabolites were further classified into specific metabolite classes such as amino acids, sugar alcohols, sugars, organic acids, vitamins, phenolics and secondary metabolites, on the basis of their chemical nature ( Table 2). Metabolite feature areas were normalized using the area of internal standard before performing comparative metabolomics.
As shown in Fig. 1 and Fig. 2, all of the differentially accumulating metabolites did not show distinct accumulation patterns in both SR and SS cell cultures during the time course studied. Interestingly, metabolites belonging to the same biosynthetic pathway often showed distinct accumulation patterns during different post elicitation time points. After VIE-treatment, most of the differential metabolites started enhanced accumulation. At 12 or 36 hpe most of the differential metabolites showed maximum accumulation in SR cultures, including the amino acids, organic acids and phenolics. Following the progression of elicitation, metabolite level first dramatically increased to attain a peak and then decreased thereafter at 72 hpe. Notably, a large number of differentially accumulating metabolites in 'Florina' cultures that showed elevated levels, have been identified in more than one time point following elicitation, suggesting activation of the metabolic pathways leading to the biosynthesis of these particular metabolites. tures both at the level of primary and secondary metabolites ( Fig. 1 and Fig. 2). In comparison, the numbers of metabolites showing a significant differential accumulation in VIE-treated SR compared to SR0 cultures were much lower in VIE-treated SS versus SS0 cultures. Among the 31 differentially accumulating metabolites detected from 'Florina' , 30 metabolites were up-regulated and one metabolite was down-regulated as shown in Fig. 1. On the contrary only nine metabolites were significantly up-and one metabolite was down-regulated in 'Vista Bella' cultures after VIE-treatment ( Fig. 1 and Fig. 2). Among the amino acids detected in 'Florina' , the levels of serine, tryptophan, phenylalanine, proline and aspartic acid were significantly up-regulated after VIE-treatment. Similarly, VIE-treatment triggers up-regulation of glucose, fructose and mannose contents; whereas sucrose content was down-regulated. Among the sugar alcohols, mannitol and sorbitol contents were significantly up-regulated. Pyruvic-, 3-phosphoglyceric-, citric-, succinic-, fumaric-, malic-, and malonic acid levels were up-regulated. Among the vitamins, only ascorbic acid was detected from 'Florina' cell cultures, whose level was significantly up-regulated after elicitor treatment. Phenolics and specialized biphenyl and dibenzofuran phytoalexins were the most abundant secondary metabolites detected in the VIE-treated 'Florina' cell cultures. Plant phenolics are known to play crucial role in pathogen defense 42 . Nine phenolics and three inducible phytoalexins (two biphenyl phytoalexins: noraucuparin and aucuparin; one dibenzofuran phytoalexin: eriobofuran) were detected. All these phenolic metabolites showed differential accumulation after the VIE-treatment. The content of these phenolics were first up-regulated after the VIE-treatment and thereafter declined to the basal level (Fig. 2). Interestingly, in our study, biphenyl and dibenzofuran phytoalexins were not detected from the untreated control cells and formed in 'Florina' cultures only after the VIE-treatment. On the contrary, in scab susceptible 'Vista Bella' cell cultures, only differential accumulation of caffeic-, chlorogenic-and ferulic acids were observed (Fig. 2). Catechin, salicylic acid, noraucuparin, aucuparin and eriobofuran were not detected from 'Vista Bella' cultures.

Principal component analysis (PCA) reveals metabolic alterations in VIE-treated 'Florina' cell cultures.
To assess the data reproducibility in the different biological replicates of the metabolites measurements, GC-MS data were subjected to principal component analysis (PCA). For PCA, only those 21 metabolites were included ( a marked in the Table 1) which showed differential accumulation only in the scab resistant cultivar 'Florina' but not in the scab susceptible cultivar 'Vista Bella' . PCA helps to reduce the dimensionality of complex data sets, which facilitate better visualization of the inherent patterns in the data. PCA analyses uses, linear orthogonal transformation of the original data variables to generate a new set of uncorrelated variables known as principal components (PCs) 43  Hierarchical clustering analysis of the metabolite profiles. In order to search for any probable discrepancies in the metabolite profiles of all the seven (0-72 h) sample groups, the 21 differentially accumulating metabolites used for the PCA analyses were organized and visualized by heatmap analysis tool of MetaboAnalyst 4.0 ( Fig. 4) 44 . The Heatmap offered excellent separation of the metabolite trend between the non-treated (0 h) and VIE-treated (6-72 h) samples.
Changes in metabolic pathways: metabolic pathway network. A simplified metabolic pathway network was reconstructed using key metabolic pathways, such as the shikimic acid pathway, the phenylpropanoid pathway, the glycolytic pathway, the pentose phosphate pathway, the biphenyl biosynthetic pathway, the flavonoids biosynthesis pathway, the tricarboxylic acid (TCA) pathway, and the amino acid biosynthetic pathway to show the regulated pattern diversity of each detected differentially accumulating metabolites from 'Florina'  with respect to their proportional incorporation into key metabolic pathways. As shown in Fig. 5, sucrose as the precursor for glucose and fructose was higher in non-elicited cells than that of VIE-treated cells. However, sucrose hydrolysis product, the glucose and the fructose were higher in the VIE-treated cells, suggesting that more active sucrose catabolism occurred in the VIE-treated cells. The level of mannose, sorbitol and mannitol were higher in the VIE-treated cells, suggesting that fructose is metabolized more towards these metabolites rather than re-entering into glycolytic pathway. The level of identified TCA cycle metabolites such as citric, succinic, fumaric and malic acids were up-accumulated in the VIE-treated cells suggesting higher turn-over number of TCA cycle. The metabolites derived from shikimate pathway showed an absolutely distinct accumulation pattern. The level of all the detected shikimate-derived metabolites such as caffeic acid, protocatechuic acid, catechin, chlorogenic acid, benzoic acid, ferulic acid, p-coumaric acid, trans-cinnamic acid and salicylic acid were higher in the VIE-treated cells. Aucuparin, noraucuparin and eriobofuran were synthesized only after the VIE-treatment. respectively. Combinations of aucuparin (5 µM) and noraucuparin (5 µM) resulted into relatively higher inhibition of conidial germination. For conidia germination assay, pH range of 5 to 5.6 was found to be optimum. These results clearly showed synergistic effect aucuparin, noraucuparin and eriobofuran on inhibition of V. inaequalis conidial germination. However the underlying mechanisms of actions for this inhibitory effect remain elusive.

Discussion
The present GC-MS-based metabolomics analysis was performed to identify signature metabolites associated with apple resistance to scab fungus V. inaequalis. This study is the first report on the systemic identification of metabolic changes in response to VIE-treatment in the cell cultures of scab resistant cultivar 'Florina' and scab susceptible cultivar 'Vista Bella' , before and after treatment. In this study we have used apple cell culture as model system because cell cultures offer an uniform plant material for studying metabolic changes and typically plant cell cultures exhibit relatively less metabolic complexity due to lack of pigments and lignified tissues. Earlier plant cell cultures have been successfully used for analyzing plant-microbe interactions in apple and other plant systems 29,45,46 . The metabolomics data were then correlated with the expression patterns of genes linked to selected secondary metabolite biosynthesis to obtain a comprehensive insight of the changes occurring after VIE treatment.
From the metabolic point of view, VIE-treatment appears to cause massive metabolic re-programming in 'Florina' both at the level of primary and secondary metabolisms. In this study we have detected only 60 metabolites from the VIE-treated cell cultures of 'Florina' which is relatively less as compared to the previously reported metabolites from apple 38 . This is possibly explained by the fact that cell culture system used in this study is devoid of complex metabolites such as chlorophyll, pigments and lignified tissues, and thereby represents a metabolically simple system as compared to intact apple plants and fruits. The changes in the metabolites concentration suggested that six biochemical pathways were mainly altered by VIE-treatment. Considering the limited number of metabolites reported in this study, the most altered pathways were amino acid, sugar, sugar alcohols, organic acids, vitamin, phenolics and biphenyl-dibenzofuran phytoalexin biosynthesis. Earlier studies have shown that up-regulation of primary metabolites, mainly in the level of amino acids, sugar and sugar alcohols occur during plant-pathogen interactions 47 . Notably, VIE-treated 'Florina' cultures showed differential accumulation of selected organic acids. The elevated level of organic acids might be associated with the higher turnover number of glycolysis and TCA cycles, which generates more abundant biosynthetic intermediates. High organic acid concentrations are known to facilitate better ion absorption and thereby known to be associated with enhanced disease resistance 48 . Likewise, ascorbic acid is known to be associated with the superior defense responses in many plant species 49 .  Among secondary metabolites, phenolic and flavonoid contents were significantly up-regulated in the VIE-treated 'Florina' cell cultures as compared to 'Vista Bella' cultures. Similar array of polyphenols were previously detected in 'Florina' leaves 50 . However, quercitin and avicularin as detected earlier from 'Florina' leaves were not detected from VIE-treated 'Florina' cell cultures. Phenolic acids are known to be deposited along the cell wall to provide first line of defense against infection 51 . Previously it was reported that differential accumulations of flavonols (catechins and proanthocyanidins) were key factors in providing scab resistance in the apple cultivars 27,52,53 . In our study, high basal level of catechin content was also observed in the 'Florina' cell culture. Earlier finding showed that scab-resistant apple cultivars are particularly rich in the content of caffeic-, chlorogenic-, and ferulic acids and their concentration rapidly increases after the scab-infection as compared to susceptible cultivars 54 . It is known that benzoic acid serves as the precursor for biphenyl and dibenzofuran phytoalexins biosynthesis in apple 7 . Therefore, enhanced biosynthesis of benzoic acid in the VIE-treated 'Florina' cultures could be linked with bipehnyl and dibenzofuran phytoalexins formation in the 'Florina' . It has already been reported that trans-cinnamic acid serves as the precursor of many defense metabolites, such as phenylpropanoids, benzoates, lignans, and flavonoids 55 . It was worth speculating that after the VIE-treatment; first trans-cinnamic acid level is up-regulated, which in turn get metabolized to biphenyl and dibenzofuran phytoalexins through intermediate formation of benzoic acid. Moreover, considerable up-regulation of salicylic acid (SA) biosynthesis was also observed in the VIE-treated 'Florina' cell culture. The salicylic acid can induce local resistance in the form of hypersensitive reaction or long distance systemic acquired resistance (SAR) by triggering the production of pathogenesis-related proteins 56 . The crucial role of SA as signaling molecule in plant defense response has already been reported earlier 57 . VIE-treatment to 'Florina' cell culture led to the production of biphenyl and dibenzofuran phytoalexins which are known to inhibit the growth of V. inaequalis 7 . Interestingly, in our study, biphenyl and dibenzofuran phytoalexins were absent from the untreated controls cells and from the scab susceptible cultivar 'Vista Bella' . Notably, malusfuran, a glucosylated dibenzofuran phytoalexin detected earlier from the elicited cell culture of resistant apple cultivar 'Liberty' , has not been detected in this study. On the contrary, eriobofuran (free dibenzofuran) has been detected from the VIE-treated 'Florina' cell cultures. The glucosylated dibenzofurans are possibly broken-down to provide free dibenzofuran such as eriobofuran upon pathogen infection. An identical breakdown of phloridzin into phloretin was observed in several scab resistant apple cultivars 58 . During the present study, PCA was performed to evaluate the dynamics of the metabolic re-programming following VIE treatment (Fig. 3A). The metabolite profiles at 6, 12, 24, 36, 48 and 72 h reflected VIE-induced enhanced accumulation of metabolites as compared to non-treated 0 h control. Only, three metabolites, noraucuparin, aucuparin and eriobofuran were newly formed after VIE-treatment. The 21 detected metabolites that showed significant differential accumulation in SR cultivar 'Florina' indicates a strong defense response of SR cultivar to VIE-treatment. In addition, heatmap analyses (Fig. 4) demonstrated that the VIE-treated cells were the major sources of variance in the metabolic pool, which indicated significant metabolic re-programming by the VIE-treatment. We additionally examined the qRT PCR-based selected gene expressions of 'Florina' cultures before and after the VIE-treatment. The expression levels of all the seven genes studied were up-regulated after the VIE-treatment, thereby indicating up-regulation of phenylpropanoid, flavonoids and biphenyl biosyntheses after VIE-treatment. Enhanced expressions of PAL, C4H and 4CL genes are known to be associated with enhanced phenylpropanoid biosynthesis 59,60 . Likewise, high CHI and F3H expressions are linked to higher accumulation of catechin. Similar to PAL, a massive up-regulation of BIS3 expression was also observed. BIS3 is known to be the main BIS isoenzyme responsible for phytoalexin biosynthesis in apple shoots 7 . The expression of BIS3 preceded the accumulation of noraucuparin, aucuparin and eriobofuran, suggesting involvement of BIS3 in biphenyl-dibenzofuran biosynthesis. High AOX expression level in the VIE-treated 'Florina' cultures is probably associated with the enhanced biosynthesis of phenylpropanoids, as shown earlier in other plant species 61 . The elevated AOX expression is further known to be associated with the plant's ability to facilitate metabolic re-programming to cope-up with stress conditions 62 .
The inhibitory effect of noraucuparin, aucuparin and eriobofuran on conidial germination of V. inaequalis provides further support for our assumption that biphenyl-dibenzofuran phytoalexins play crucial role in scab resistance in SR cultivar 'Florina' . Here we observed that conidial germination is reduced by addition of these phytoalexins together (Supplementary fig. S2), suggesting a synergistic mode of actions of these phytoalexins. Together with the observation that these biphenyl and dibenzofuran phytoalexins were induced by VIE-treatment in 'Florina' , this result supports the hypothesis that these specialized phytoalexin accumulation in planta have a role in prohibiting the scab infection and disease progression. However, the exact mode-of-actions of these phytoalexin need to be further investigated for better understanding of resistance mechanism in apple.

Conclusions
This is the first report to use non targeted GC-MS metabolomics analyses of apple cell cultures treated with elicitor prepared from the scab fungus V. inaequalis. Our results showed that the developed metabolomics method has excellent sensitivity and specificity to analyze metabolites from apple cell cultures. Further, our results indicated that apple cell cultures could be used as a model system to understand apple-Venturia interactions in terms of metabolic re-programming. We were able to identify significant changes in the secondary metabolites in 'Florina' cell cultures, indicating metabolic re-programming of specific biosynthetic pathways after VIE-treatment. Based on these results we can conclude that biphenyl and dibenzofuran biosynthesis pathway plays a crucial role in providing scab resistance in the 'Florina' cell cultures. In the present study, we analyzed metabolic re-programming in the V. inaequalis elicitor treated apple cell culture system. However, future metabolomics analysis using scab-infected apple plants is required to get a complete picture of metabolic basis of scab resistance in apple. Our results would provide a strong basis for the future metabolomics analysis of apple plants to decipher marker metabolites associated with the scab resistance in apples.

Chemicals.
Solvents used for extraction were of high-performance liquid chromatography grade. 2-phenylphenol (internal standard for gas chromatography) was purchased from SRL chemicals (India). N-methyl-N-(trimethylsilyl)-trifluoroacetamide (MSTFA), pyridine and all metabolite standards were obtained from Sigma.

Plant material and elicitor-treatment. One scab resistant (SR) apple (Malus domestica) cultivar 'Florina'
and one scab-susceptible (SS) cultivar 'Vista Bella' was used in this study to develop cell suspension culture. Cell suspension cultures were developed as described before 2 . Cell cultures were grown in liquid LS medium containing 2 µM 2,4-D and 1 µM NAA at 26 °C in the dark on an orbital shaker at 120 rpm. An elicitor prepared from apple scab fungus, V. inaequalis was used in this study. V. inaequalis strain (MTCC No.: 1109) was purchased from microbial type culture collection and gene bank (MTCC), Chandigarh, India. V. inaequalis elicitor (VIE) was prepared by homogenizing the fungal hyphae as described before 63 . Briefly, 10 g of ground fungal mycelium was added to 1 L of acidified water (pH 2.0). Water extract was then boiled for 1 h on a hot plate, cool down to room temperature and filter sterilized. After filtration, the pH of the fungal extract solution was adjusted to 5.0 and the final volume was adjusted to 1 L by adding distilled water. This solution was used as the VIE. Elicitor-treatment was performed by adding 2.5 mL of the VIE (~70 mg of fungal polysaccharide) to the 50 mL of seven-day-old cell apple suspension cultures of both 'Florina' and 'Vista Bella' . Upon elicitor-treatment, cells were harvested at defined time points: 0, 6, 12, 24, 36, 48, and 72 hpe. In control treatment, similar volume of sterile distilled water was added in lieu of the VIE. All experiments were performed with at least three biological repeats.
Extraction of polar metabolites. The VIE-treated apple cell cultures were harvested by vacuum filtration at defined time points (0-72 h) and kept in hot air oven at 60 °C for 4 h. Dried cell mass (2 g) was crushed in liquid nitrogen and fine powdered samples were used for metabolite extraction. The extraction of polar metabolites for GC-MS analyses was performed following the protocol described before 64 with suitable modifications. An extraction mixture was prepared by adding methanol/water/chloroform in the ratio of 2.5:1:1 (v/v/v) and stored at −20 °C. Pre-cooled extraction mixture (1 mL) was then added to 200 mg powdered samples in a 1.5 mL micro centrifuge tube and vortexed vigorously at room temperature for 2 min. In order to identify extraction efficiency, 50 µL of 2-phenylphenol (from 2 mg.mL −1 methanol stock) was spiked in the extraction mixture as the internal standard (IS) and vortexed again for 1 min. The extracts were then centrifuged at 14000 g for 5 min. The resulting supernatant (0.8 mL) was transferred into a new 1.5 mL tube and then 0.4 mL of water was added to the supernatant, whole mixture was vortexed for 10 s and subsequently centrifuged at 14000 × g for 5 min. The polar upper phase (methanol/water) was transferred to a new micro-centrifuge tube and then dried out in a vacuum concentrator (Labconco, Centrivap; USA) at 20 °C for 2 h followed by 12 h freeze drying in a lyophilizer. Finally dried material was subjected to double derivatization for GC-MS analyses 65  Initial temperature of 80 °C for 1 min, followed by temperature increase to 220 °C at the ramp rate of 10 °C.min −1 , followed by temperature increase to 310 °C at the ramp rate of 20 °C.min −1 and finally a 10 min hold at 320 °C. Total run time calculated was 39 min. Helium was used as carrier gas at a flow rate of 1 mL.min −1 . The inlet temperature and interface temp was set 280 °C. The MS unit was tuned to its maximum sensitivity and the mass range for total ion current was m/z 80-700, and the detector voltage was set at 1700 V. Each sample was replicated three times. Scan was started after solvent delay of 7 min with scan frequency of 4 S −1 (2.0 HZ).

Metabolite identification.
Metabolites were identified by comparing the mass-to-charge ratios and abundance of each metabolite detected against a standard NIST-17 mass spectral library (National Institute of Standards and Technology), and our in-house mass spectral database that include several secondary metabolites, amino acids, organic acids, and sugar standards. Metabolite identity was reported only when the matching value of the mass spectra comparison was more than 70 percent, and an increase in the area of the corresponding peak was observed when spiking the sample with the corresponding pure standard. Each mass spectrum was carefully analyzed for co-elution detection. Co-elution was not detected in any of the identified peaks.
Metabolite data pre-processing and statistical analysis. Raw  , and alternative oxidase (MdAOX) were evaluated using the gene-specific primers (Supplemental Table S3). All samples were normalized using apple actin gene. Scaling of expression level was performed in relation to the respective mRNA expression levels in the control (0 h) cells, which were set to 1. Three technical repeats were performed. Published mathematical model was used for the estimations of efficiency and gene expression levels 66 .
Conidia germination inhibition assay. V. inaequalis mycelium was grown on potato dextrose agar plates covered by cellophane paper and conidia were harvested as described before 67 . Conidia were suspended in autoclaved water in a concentration of 1 × 10 5 conidia.mL −1 . Conidia germination inhibition assay was performed as described earlier 68 . An aliquot (40 µL in 50% methanol) comprised of aucuparin, noraucuparin and eriobofuran, 5 µM each, either individually or in combinations, was mixed with 10 µL of conidial suspension (1 × 10 5 conidia. mL −1 ). After mixing well, 10 µL was immediately plated on 2% water-agar slides. In control slides conidia were incubated in presence of 50% methanol. After 24 h incubation at 20 °C the number of germinated conidia of both treated and control slides were counted under microscope. The inhibition percentage of conidial germination was calculated using following the formula:

Data Availability Statement
All data generated and analyzed in this study are included in this manuscript either as main data or as supplementary information files.