NMR-based metabolic study of leaves of three species of Actinidia with different degrees of susceptibility to Pseudomonas syringae pv. actinidiae

Abstract Bacterial canker of Actinidia, caused by the bacterium Pseudomonas syringae pv. actinidiae (Psa), is the most serious disease of these plants worldwide. Leaves of three species of Actinidia, namely A. chinensis var. chinensis, A. chinensis var. deliciosa and A. arguta, having different degrees of tolerance to Psa, were analyzed by Nuclear Magnetic Resonance spectroscopy. Aqueous extracts of leaves were studied and several metabolites, classified as organic acids, amino acids, carbohydrates, phenols and other metabolites, were identified by 1D and 2D NMR experiments and quantified. The metabolic profiles of these species were compared through univariate statistical analysis ANOVA and multivariate PCA. Levels of metabolites with known antibacterial activity, such as caffeic and chlorogenic acids, were observed to be higher in the A. arguta samples. Moreover, these metabolites have different Pearson correlation patterns among the three Actinidia species, suggesting a difference at the phenylpropanoid biosynthetic pathway. Graphical Abstract


Introduction
Kiwifruit (Actinidia Lindl. spp.) is an economically important horticultural crop worldwide for its quality and nutritional properties. This genus is native from China and half of the worldwide production is from this state. In order of production, Italy, New Zealand, Chile and Greece together are responsible for about 80% of the kiwifruit produced outside China. The most spread cultivar, 'Hayward', belongs to the Actinidia chinensis var. deliciosa (A. Chev.) A. Chev. species while 'Hort16A', belonging to the Actindia chinensis Planch. var. chinensis species and selected in New Zealand, was the first yellow-fleshed kiwifruit in production outside China and was the most important yellow-fleshed cultivar in international trade. Recently, interest focused on other species, namely Actinidia arguta (Siebold & Zucc.) Planch. ex Miq., characterized by small fruit, smooth skin and edible already at harvest time.
Actinidia cultivation is negatively affected by the spread of bacterial canker caused by Pseudomonas syringae pv. actinidiae (Psa). Psa is a destructive pathogen of kiwifruit orchards and causes economic losses in all major areas of kiwifruit production worldwide (Scortichini et al. 2012). The pathogen infects both green and yellow-fleshed varieties using leaves as the entry site to colonize plants thereby reaching the roots (Petriccione et al. 2013;Serizawa and Ichikawa 1993). This feature has important consequences for the prevention and control of the disease, since the effectiveness of contact bactericides is greatly reduced when Psa internally colonizes the plant and the damages caused are extensive and in the worst cases the entire crop can be lost. In particular, almost all of 'Hort16A' orchards have been wiped out and nowadays susceptibility to Psa is considered one of the most important characteristics in the selection of new kiwifruit cultivars (Ferguson 2015).
Currently, leaf application of active copper is considered the best method to protect Actinidia spp. from Psa, and the most widely used formulations are copper hydroxide (Cu(OH) 2 ) and copper oxychloride (Cu 3 Cl 2 (OH) 4 ). In general, an efficient spray coverage during periods of high risk of infection is necessary, especially regarding climatic conditions (temperature of 15 ± 3 C plus consistent rainfall), and to the phenological/growth stage of the vines as they are the most vulnerable after harvest, at leaf fall, at bud break and during following 2-4 weeks (Vanneste 2017).
Psa has strongly changed the varietal assortment of cultivated Actinidia and is likely to continue to do so in the future, since known products such as antibiotics, antimicrobial peptides, heavy metals, disinfectants, elicitors and biological control agents so far applied have led to variable results in the control of the disease, however without solving the problem. To contain the spread of the bacterium, development and marketing of Actinidia species which can better withstand Psa are the most sustainable strategy. In particular, A. arguta has proven to be more tolerant than both A. chinensis var. deliciosa and A. chinensis var. chinensis, with the latter being the most susceptible one.
Although there are some studies in which the relative scale of tolerance among the Actinidia species against Psa was evaluated (Scortichini et al. 2012;Datson et al. 2015), there is no knowledge on the mechanisms responsible for resistance to infections and diseases. This information is essential for strategies to be defined which may improve resistance in highly susceptible crop species.
Nuclear Magnetic Resonance (NMR) based metabolic profiling is useful for a better understanding of plant biochemistry because it can, for example, discriminate various cultivars or the same cultivar ailing from different pedoclimatic conditions Caligiani et al. 2014;Tomassini et al. 2016) and investigate their nutritional properties Cicero et al. 2015;Parra et al. 2018). NMR has further proven to be able to identify metabolites responsible for plant defense mechanisms against biotic and abiotic threats (Brasili et al. 2014;Brasili et al. 2016;Sciubba et al. 2015;Sampaio et al. 2016;Jafari et al. 2018).
The main objective of the present study is to characterize, by means of NMR spectroscopy, the metabolic profile of leaves of three species of Actinidia, namely A. chinensis var. deliciosa ('Hayward'), A. chinensis var. chinensis ('Hort16A') and A. arguta ('Arguta 87-14-15'), with a different degree of tolerance against Psa.

Results and discussion
The comprehensive metabolic profile analysis of Actinidia leaves was carried out by 1 H NMR spectroscopy of hydroalcoholic extracts. The extracts of the three species showed quantitative differences but no qualitative ones (Figures S1 and S2).
Resonance assignment was carried out through bidimensional TOCSY ( Figure S3), HSQC ( Figure S4) and HMBC ( Figure S5) experiments and confirmed by literature data (Fan 1996;Wishart et al. 2013). 1 H chemical shifts, multiplicity and 13 C chemical shifts are summarized in Table S1. A total of 31 metabolites were identified and 26 were quantified and the quantitative analysis is reported in Table 1 with the statistical significance assessed by one-way ANOVA.
Several molecules were observed to differentiate among the three species. In greater detail, regarding the aminoacid content, Ala, GABA and Tyr are more abundant in A. arguta, average in 'Hayward' and in lower quantity in 'Hort16A'. Observing the organic acids, A. arguta has a higher amount of LA, QA and MA than the other two species. The carbohydrate F is present in higher levels in A. arguta and 'Hayward' than in 'Hort16A' while S is more present in 'Hayward', followed by A. arguta and then by 'Hort16A'. Regarding the total phenol content of the three species, while A. arguta and 'Hayward' have similar levels (both higher than the one of 'Hort16A'), the molecules of this class are differently distributed. In fact, 'Hayward' has a greater amount of the flavonoids EC and Kam, while A. arguta is richer in terms of Caf and CGA.
Multivariate PCA, performed as an explorative statistical analysis, provided a model whose first component explained 44% of the overall variance and the second one another 16%. In the Score plot (Figure 1) it is possible to observe three spontaneous groupings, each corresponding to one Actinidia species. It is interesting to observe that to greater values of PC1 correspond the species with greater resistance against Psa attacks while PC2 differentiates 'Hayward' from the other two.
To identify the molecular species responsible for the observed groupings, it is necessary to observe the Loading plot (Figure 1)  Plants usually react to external attacks and pathogens producing secondary metabolites to contrast infections. These substances can counteract the spread of the disease, but it is necessary that the host organism recognizes when it has been affected. One of the most dangerous aspects of Psa is its ability to suppress these cellular recognition mechanisms (Ichinose et al. 2013), and therefore the presence of pre-formed metabolites with antibacterial properties becomes one of the most effective factors for plant defenses. Nevertheless, the ability to synthesize these molecules after infection is also of paramount importance and this fact is frequently employed in several scientific fields Tocci et al. 2010;Valletta et al. 2016;Badiali et al. 2018).
Among the molecular species identified and quantified in Actinidia leaves, several are known to possess antibacterial activity such as GABA, CGA, Caf, QA, EC, Afz and Kam.
GABA is an aminoacid strongly involved in the recognition of extraneous agents (Kinnersley and Turano 2000) and is also known that some GABA-containing preparations can stimulate the natural plant defenses (Dagorn et al. 2013).
QA is an intermediate of the phenylpropanoid pathway able to be transformed into several different secondary metabolites reacting to external stimuli (Qudsia et al. 2010).
Caf and CGA, as well as the other metabolites of the phenylpropanoid pathway, are known both for their antioxidant and antibacterial activities (Kabir et al. 2014;Shadle et al. 2003).
Flavonoids such as EC, Afz and Kam are secondary metabolites with known antibacterial activity (Cushnie and Lamb 2011) that are naturally occurring in many plants, but whose synthesis is greatly enhanced during external attacks (Brasili et al. 2016) and are among the most common substances involved in the defense against biotic stress (Petrussa et al. 2013).
One hypothesis for the greater tolerance against Psa of A. arguta is its higher content of GABA with respect to the amount measured in the other two species. This hypothesis is compatible with the ability of GABA to bolster plant immunity defenses.
Even if no comparative studies were performed in this work, another possible explanation could be the greater antibacterial activity against Psa of phenylpropanoids (Caf and CGA) compared to the one of flavonoids (EC, Afz and Kam) since the formers are present in higher amount in the species with the greater tolerance.
Since the ability to produce consistent amounts of defensive metabolites in response to external attacks is of paramount importance, it is interesting to study the Pearson correlation matrices of the molecules involved in plant defense (Figure 2, Figure S6, S7, S8).
It is notable that while the amount of flavonoids is higher in 'Hayward' than in A. arguta (Table 1), the significant covariances of Afz and Kam are more numerous, in greater strength and, in some instances, with a different sign in the more tolerant species A. arguta, indicating a different activity of the metabolic pathway involved in their biosynthesis, namely the phenylpropanoid one. Of interest is also the presence of correlations between the flavonoids and GABA only in A. arguta, which suggests a greater link between the mechanisms regulating the plant recognition of external threats (GABA one) and the production of metabolites with a more direct defensive role, such as the ones belonging to the phenylpropanoid pathway, in response to those stresses.

Conclusions
This study confirms the validity of NMR as a powerful tool for the investigation of the metabolic network involved in plant defense against biotic stresses. This technique, applied to the study of kiwifruit leaves and supported by univariate and multivariate statistical techniques, has shown how a variety of a more Psa-tolerant species of Actinidia is richer in terms of GABA, caffeic acid and chlorogenic acid. Moreover, the capability of A. arguta to produce Afz and Kam is different to the one observed in the other two species. This experimental approach allows to identify which metabolites are involved in the tolerance to external attacks of plant species, independently by both the nature of the plant and the type of threat.

Disclosure statement
The authors declare no competing financial interest.

Funding
The present work has been carried out under the project "Development of high-resolution NMR methodological platforms for the determination in various plant organs of metabolites involved in susceptibility or resistance to pathogens and insects", funded by the Italian Ministry of Agriculture and Forestry Policy [grant number CRAM 3.00.08.00].