Variation of compounds in leaves of susceptible and resistant alternate hosts of Cronartium pini and C. ribicola

Leaf compounds may contribute to plant defense against Cronartium rusts. Secondary compounds are either natural or induced in leaves. We studied the variation of compounds in leaves of six alternate hosts of Cronartium pini and two of C. ribicola that represented either susceptible or resistant species to these rusts. Extracts from the plant leaves were analyzed using LC-MSMS (liquid chromatography tandem mass spectrometry) and compounds were compared between susceptible and resistant species of the same plant genera to identify significant differences between resistant and susceptible species. Also, LC–MS (liquid chromatography mass spectrometry) with external calibration was used to quantify 12 candidate compounds known from the literature. Among these compounds, the most abundant significant ones in C. pini -resistant Melampyrum pratense were chlorogenic acid and quercitrin, in Veronica chamaedrys ferulic acid, quercitrin and luteolin and in Impatiens glandulifera quercitrin, ferulic acid, kaempferol, rutin and hyperoside. In C. ribicola -resistant Ribes rubrum the most abundant significant compounds were caffeic acid, p-coumaric acid and quercitrin. Among all extracted leaf compounds, concentrations of three compounds were over 1000 times greater in rust-resistant M. pratense, three compounds in V. chamaedrys, eight compounds in I. glandulifera, and one compound in R. rubrum than in rust-susceptible species. Among the compounds, the most promising possibly linked to rust resistance were chlorogenic acid and quercitrin.


Introduction
Tree rusts of Cronartium are important pathogens of Pinus spp. in the northern hemisphere (Gäumann, 1959;Ziller, 1974). Cronartium pini (Willd.) Jørst. is a significant rust disease that kills Pinus spp. in Europe and Asia (CABI, 2020), while the rust is a quarantine species in North America (Kim et al., 2022). Cronartium pini causes severe damage especially on Pinus sylvestris L. in northern Fennoscandia (Kaitera, 2000;Samils et al., 2021;Wulff et al., 2012). In the 2000s, C. pini has caused severe losses especially on young Scots pine plantations in nutrient-rich soils (Wulff et al., 2012). The chemical compounds enriched in the wood after Cronartium infection have been investigated (Bullington et al., 2018;Kaitera et al., 2021). The results suggested that terpenes and resin acids are produced by the host to protect it from Cronartium rust. The rust spreads via alternate hosts, and over 50 susceptible species are known from 13 plant families (Kaitera et al., 2015;Kim et al., 2022). Important species belong especially to Orobanchaceae, Paeoniaceae and Balsaminaceae (Kaitera et al., 2015).
Another Cronartium, C. ribicola Fisch., is a serious pathogen of five-needle pines in North America (Zambino, 2010). It spreads via Ribes (Grossulariaceae), of which R. nigrum L. is a highly susceptible species, while R. rubrum L. is a resistant species. In Finland, nearly all R. nigrum cultivars are susceptible and R. rubrum cultivars are resistant to C. ribicola (Kaitera & Nuorteva, 2006).
Among plant genera, Melampyrum is one of the most susceptible ones to C. pini (Kaitera, 1999;Kaitera et al., 1999Kaitera et al., , 2012Kaitera et al., , 2015Kaitera et al., , 2017Kaitera et al., , 2018. M. sylvaticum L., M. nemorosum L., M. arvense L. and M. cristatum L. are highly susceptible species, while M. pratense L. is a resistant species (Kaitera, 1999;Kaitera & Nuorteva, 2003a, b;Kaitera et al., 1999Kaitera et al., , 2012. Other important alternate host genera are Pedicularis, Euphrasia, Impatiens and Veronica. Reasons for the differences in susceptibility and resistance are mostly unknown, but certain intrinsic compounds are likely to play a role. Terpenes are among such compounds (Bullington et al., 2018). They may be normally present or induced by pathogens and other factors. Variation in rust resistance among closely related species may be due to variation in leaf chemistry as was proposed for M. sylvaticum and M. pratense (Kaitera & Witzell, 2016). Especially chlorogenic acid was abundant in the resistant M. pratense, while the compound was lacking in M. sylvaticum. Chemical variation on M. pratense and M. sylvaticum in leaves of different ages, time of collection and among locations have not been studied.
Based on literature concerning plant resistance 12 compounds were selected to study concentration differences in the plant species: chlorogenic acid, caffeic acid, ferulic acid, p-coumaric acid, syringic acid, luteolin, kaempferol, myricetin, quercitrin, rutin, apigenin and hyperoside. Chlorogenic acid is a polyphenol and the ester of caffeic acid and quinic acid. In humans, it has been reported to have antioxidant, antibacterial, chemopreventive, antiviral and neuroprotective characteristics (Clifford et al., 2017;Magana et al., 2021). In the human diet coffee, fruits and vegetables are its major sources (Upadhyay & Rao, 2013). Chlorogenic acid has been shown to have bioactivity against various plant pathogens playing a defensive role against biotic and abiotic stresses (Kundu & Vadassery, 2019;Soviguidi et al., 2022). Caffeic acid is a hydroxycinnamic acid derivative and polyphenol. Like chlorogenic acid it also has many beneficial effects on human health. Caffeic acid is found at high levels in some herbs and fruits, and its bioactivity has been shown (Kiokias et al., 2020). Ferulic acid is a phenolic acid that can be found in the seeds of coffee, apple, artichoke, peanut and orange (Kiokias et al., 2020). p-Coumaric acid is found in many natural plants and organisms like fungi, peanuts, beans, tomatoes, carrots, basil and garlic (Kiokias et al., 2020). In addition, most fruits contain p-coumaric acid. Evidence for its high bioactivity has been reported. Syringic acid is a derivative of gallic acid. Its bioactivity for suppression of chronic diseases like human leukemia (HL)-60 and DV-145 human prostate carcinoma cells has been shown (Shahidi & Yeo, 2018). Luteolin, kaempferol, myricetin, quercitrin, rutin, apigenin and hyperoside are hydroxyflavones having potential bioactivity (Adamczak et al., 2020;Cirak et al., 2007;Dall'Agnol et al., 2003;Elansary et al., 2020;Gharibi et al., 2019).
In addition, an unbiased liquid chromatographytandem mass spectrometry (LC-MSMS) approach was chosen to characterize differences in the compound spectrum of the different plant species. Here all detectable compounds were characterized by their accurate mass and chromatographic retention time and differences in signal intensity (integrated ion counts of chromatographic peaks of the SICs) were compared between susceptible and resistant species. Data were collected in data dependent acquisition mode, in which the instrument control software recognizes signals and switches automatically to MSMS mode to select and fragmentate the corresponding ions. Accurate mass and MSMS spectra were then used to search different compound databases to identify compounds of interest.
The aim of this study was 1) to investigate the variation of compounds in leaves of alternate hosts species susceptible and resistant to C. pini and C. ribicola, 2) to compare compounds of resistant and susceptible species groups to one another, and 3) to characterize compounds that may be linked to rust resistance. These compounds may be important in developing control of rust diseases and might be utilized against other pathogens in the future.

Plant material
Circa 10-20 young leaves of eight species were harvested from 20 randomly selected wild or cultivated plants per species from the city area of Oulu. Resistant wild species of C. pini were M. pratense, Impatiens glandulifera Royle and Veronica chamaedrys L., and a resistant cultivated species of C. ribicola was Ribes rubrum. Susceptible wild species of C. pini were M. sylvaticum and V. longifolia L., while a susceptible cultivated species was I. balsamina L. Susceptible cultivated species of C. ribicola was R. nigrum. The cultivated plants were located in the Botanical Gardens of the University of Oulu (65°3,86 N, 25°27,79E). All plants of Melampyrum, Impatiens and Veronica were collected first into paper bags and transported to the laboratory prior to leaf collection. Ten leaves of Ribes spp. per species were collected directly into paper bags and transported similarly to the laboratory. The collection locations of the plants in Oulu were: M. sylvaticum and M. pratense (65°2,69 N, 25°28,04E), V. longifolia (65°2,27 N, 25°29,16E), V. chamaedrys (65°1,30 N, 25°25,96E), I. glandulifera (65°3,03 N, 25°25,20E), I. balsamina (garden plants grown from seed in the botanical Garden), R. nigrum and R. rubrum (cultivated plants in the botanical garden). The plants were collected in late June 2021. The habitats of the wild species were determined by the personnel of the Botanical Gardens. The leaves were collected mainly during flowering of the plants to ensure correct identification of the plants described in Hämet-Ahti et al. (1998).
Water for the chromatography was produced in house with a Synergy UV instrument (Millipore, cat.no SYNSV0000), equipped with a LC-PAK Polisher (Cat.No. LCPAK 0001) cartridge for the final purification step. Acetonitrile and formic acid were OPTIMA LCMS grade (Fisher Chemical, code A955-212 and A117-50, respectively).

Pretreatment and extraction of leaves
In the laboratory, healthy green leaves of the plants without any sign of fungal or insect damage, were separated from the rest of the plant material in a laminar cabin with sterile tweezers. Then the leaves were air-dried for ca. a week in a laminar cabin in open paper bags and stored at -20 °C prior to analysis. The leaf samples were crushed manually inside a paper bag until a powdery consistency was achieved. About 15 mg of each plant material was weighed into Eppendorf vials. Methanol, containing 5 mg/l of internal standard (ampicillin), was used as an extraction solvent. The solvent was cooled to + 4 °C before usage. 1 ml of the solvent was added to vials which were kept at + 4 °C for one hour. The samples were shaken using Eppendorf MixMate (5 min, 1400 rpm) after which they were centrifuged at + 4 °C (5 min, 12,000 rpm, Hettich Mikro 200). The supernatant was transferred to another vial and the residue was extracted with 0.5 ml of pure methanol (without internal standard). The supernatants were combined and filtered through a disposable syringe filter (pore size 0.2 µm, Pall Corporation). The extracts were stored at -20 °C.
Chemical analysis 20 biological repeats i.e. all collected leaves from individual plants per species, 160 samples in total, were analyzed. In the LC-MS approach compounds were characterized by retention time, accurate mass and peak area. This dataset was used to obtain maximal data points for quantitation and quantify the 12 candidate compounds with external calibration. Additionally, the same samples were analyzed with an LC-MSMS approach to obtain data for the identification of unknown compounds. 5 µl sample aliquots were eluted from a Waters Aquity Premier HSS T3 column (2.1 × 100 mm, 1.8 µm, Part No. 186009468)) with a gradient made with 0.1% formic acid in water and acetonitrile from 3 to 70% over 14 min, column temperature 40 °C (Waters Aquity UPLC-system comprising column oven (186015010), binary, high pressure mixing pump module (186016007), and autosampler (186015001)). The detector was a Q-Exactive plus orbitrap mass spectrometer in biopharma configuration (Thermos Fisher Scientific) operated in negative polarity at resolution set to 70,000 A in m/z range from 115 to 1200. For the MSMS acquisition, the same conditions were used with the addition of DDA-controlled fragmentation with stepped collision energy (nce) of 25 and 35 and fragment acquisition with m/z range 200 to 2000. The DDA data were processed with Compound Discoverer (Thermo) using standard settings for natural compound analysis. Mz vault, mz cloud, Chem-Spider and mass lists were applied in compound identification.
For the quantitation of the candidate compounds a calibration curve with 8 levels from 1 µg/ml to 1 ng/ ml was established, data processing was done with the X-calibur and its Quanbrowser option (Thermo). As quality controls two pools comprising 30 µl aliquots of 50 different samples were applied. The compounds, known to occur in leaves of M. pratense and M. sylvaticum (Kaitera and Witzell 2017), were: chlorogenic acid, caffeic acid, syringic acid, p-coumaric acid, rutin, hyperoside, ferulic acid, quercitrin, myricetin, luteolin, apigenin and kaempferol.

Statistical analysis
The data set for negative ionization, filtered for features (compounds) with intensity counts over 10 6 in any of the individual samples, was used in the statistical analyses. About 40 features (compounds) with the highest ratios at minimum 25 (e.g. M. pratense:M. sylvaticum > 25) distinguishing the susceptible and resistant plant groups at p < 0.001 (Compound Discoverer) were listed with the identification suggested by the data base search. Identification leading to compounds non-existing in plant leaves (e.g. fluorine compounds) were left unnamed like compounds without identification. The concentrations of the 12 compounds selected prior to the chemical analysis were compared between resistant and susceptible species with Welch two-sample t-test with unequal variances using the R program (R Core Team 2019, Version 3.6).

Concentrations of pre-selected compounds in the samples
Concentrations of chlorogenic acid were significantly 570 times higher in samples of M. pratense compared to those in M. sylvaticum (p < 0.0001, Table 5). They were also insignificantly two times higher in samples of V. chamaedrys compared to samples of V. longifolia (Table 6), 17 times higher in samples of I. glandulifera compared to samples of I. balsamina (Table 7), and 5 times higher in samples of R. rubrum compared to samples of R. nigrum (Table 8).
The mean concentration of caffeic acid in samples of M. sylvaticum was 0.04 ng/mg, whereas samples of M. pratense did not contain measurable amounts (NF, not found; Table 5). Concentrations of caffeic acid were insignificantly 1.1 times higher in samples of V. longifolia compared to those in V. chamaedrys and 1.8 times higher in samples of I. balsamina compared to those in I. glandulifera (Tables 6 and 7). However, concentrations of caffeic acid were significantly 21 times higher in samples of R. rubrum than in samples of R. nigrum (p < 0.0001, Table 8).
Concentrations of syringic acid were insignificantly 3.5 times higher in samples of M. pratense compared to those in M. sylvaticum (Table 5). The mean concentrations in samples of V. chamaedrys and I. balsamina were 0.28 ng/mg and 0.12 ng/ mg (Tables 6 and 7). Samples of V. longifolia, I. glandulifera, R. rubrum and R. nigrum did not contain any syringic acid (Tables 6, 7 and 8).
Concentrations of p-coumaric acid were insignificantly 1.1 times higher in samples of M. pratense compared to those in M. sylvaticum, and 1.5     (Tables 5 and 7). However, concentrations were significantly 3.6 times higher in samples of V. longifolia than in samples of V. chamaedrys (p < 0.0001, Table 6), and 44 times higher in samples of R. rubrum than those in R. nigrum (p < 0.0001, Table 8).
Concentrations of rutin were significantly 13 times higher in samples of M. sylvaticum compared to those in M. pratense (p < 0.0001), 6 times higher in samples of I. glandulifera than those in I. balsamina (p < 0.0001), and 7 times higher in samples of R. nigrum than those in R. rubrum (p < 0.0001; Tables 5,7 and 8). Samples of V. chamaedrys and V. longifolia did not contain any rutin ( Table 6).
Concentrations of hyperoside were significantly 1.8 times higher in samples of M. sylvaticum compared to those in M. pratense (p < 0.001), and 35 times higher in samples of I. glandulifera compared to those in I. balsamina (p < 0.0001; Tables 5 and 7). Concentrations were insignificantly 1.4 times higher in samples of R. rubrum than those in R. nigrum (Table 8). The mean concentration of hyperoside in samples of V. chamaedrys was 1.17 ng/mg, whereas samples of V. longifolia did not contain measurable amount ( Table 6).
Concentrations of ferulic acid were significantly 20 times higher in samples of M. sylvaticum compared to those in M. pratense (p < 0.0001; Table 5). I. balsamina were also significantly higher compared to those in V. longifolia and I. glandulifera (p < 0.0001, Tables 6 and 7). Concentrations were insignificantly 1.7 times higher in samples of R. nigrum than those in R. rubrum (Table 8).

Concentrations in samples of V. chamaedrys and
Concentrations of quercitrin were significantly 18 times higher in samples of M. pratense compared to those in M. sylvaticum (p < 0.0001; Table 5), 3.5 times higher in samples of I. glandulifera compared to those in I. balsamina (p < 0.0001; Table 7), and 4.7 times higher in samples of R. rubrum compared to those in R. nigrum (p < 0.01; Table 8). The mean concentration of quercitrin in samples of V. chamaedrys was 0.50 ng/mg compared to NF for V. longifolia (p < 0.0001; Table 6).
Concentrations of myricetin were insignificantly 1.9 times higher in samples of M. sylvaticum compared to those in M. pratense, and 1.7 times higher in samples of V. longifolia than those in V. chamaedrys (Tables 5 and 6). Samples of I. glandulifera, I. balsamina, R. rubrum and R. nigrum did not contain any myricetin (Tables 7 and 8).
Concentrations of luteolin were significantly 15 times higher in samples of M. sylvaticum compared to those in M. pratense (p < 0.001; Table 5), and 4 times higher in samples of V. chamaedrys compared to those in V. longifolia (p < 0.001; Table 6). The mean concentration of luteolin in samples of I. glandulifera was 0.35 ng/mg ( Table 7). Samples of I. balsamina, R. rubrum and R. nigrum did not contain any luteolin (Table 7 and 8).
Concentrations of apigenin were insignificantly 1.4 times higher in samples of M. sylvaticum compared to those in M. pratense, and 14 times higher in samples of V. chamaedrys compared to those in V. longifolia (Tables 5 and 6). The mean concentration in samples of R. nigrum was 20.25 ng/mg being significantly higher compared to R. rubrum (p < 0.0001, Table 8). Samples of I. glandulifera, I. balsamina and R. rubrum did not contain any apigenin (Tables 7 and 8).
Concentrations of kaempferol were significantly three times higher in samples of I. glandulifera than those in I. balsamina (p < 0.0001; Table 7). The mean concentration of kaempferol in samples of R. nigrum was 0.53 ng/mg being significantly higher compared to R. rubrum (p < 0.01; Table 8). Concentrations were insignificantly 1130 times higher in samples of V. chamaedrys than those in V. longifolia, and 1.1 times higher in samples of M. pratense compared to those in M. sylvaticum (Tables 5 and 6).

Discussion
The highest concentration of chlorogenic acid (5-O-caffeoylquinic acid) was found in M. pratense (182.48 ng/mg, Table 5). R. rubrum (58.78 ng/mg) and R. nigrum (11.98 ng/mg) contained also significant amounts of chlorogenic acid (Table 8). In the other studied plant species the concentrations were low (Tables 6 and 7). It is well known that chlorogenic acid plays a defensive role against biotic and abiotic stresses in plants. Petkovsek et al. (2009) noticed seasonal changes in phenolic compound concentrations in the leaves of scab-resistant and susceptible apple cultivars. The mean chlorogenic acid concentrations found from the resistant and susceptible species were 365-3262 ng/mg and 184-500 ng/mg, respectively. They discovered that concentrations of total phenolics as well as single phenolic compounds, like chlorogenic acid, were statistically significantly higher in resistant than in susceptible apple varieties during the growing season. The concentration of chlorogenic acid was also higher in leaves of rustresistant M. pratense compared to those of rust-susceptible M. sylvaticum in a recent study (Kaitera & Witzell, 2016). The role of chlorogenic acid in plant response to abiotic stresses (heavy metal, UV light, heat, cold, salinity and drought) has also been extensively studied (Soviguidi et al., 2022). Our results clearly support the significant role of chlorogenic acid in plant defense mechanism against biotic stresses.
In this study, the highest concentrations of quercitrin were measured from I. glandulifera (5493.58 ng/ mg) and I. balsamina (1567.80 ng/mg) ( Table 7). The concentration in the resistant species I. glandulifera was significantly higher than in the susceptible species I. balsamina. Also, high concentrations were found in R. rubrum (451.34 ng/mg) and R. nigrum (96.02 ng/mg; Table 8). The concentrations of quercitrin in other tested plant species were low, but statistically significantly higher in the resistant species, M. pratense and V. chamaedrys, compared to the susceptible species M. sylvaticum and V. longifolia (Tables 5  and 6). Elansary et al. (2020) studied polyphenols of Frangula alnus Mill. and Peganum harmala L. leaves and their bioactivity. They studied the concentrations of several polyphenolic compounds and reported that quercitrin was the main flavonoid in F. alnus (11,323 mg/kg). Leaf extracts of both species showed cytotoxic effects against Jurkat, MCF-7, HeLa and HT-29 cancer cells. They concluded that the polyphenolic composition of leaves including quercitrin, trifolin and cymaroside play a significant role in the bioactivity of these plants.
For the V. chamaedrys and V. longifolia species pair the most significant compounds were phloretin 2'-O-(6"-O-acetylglucoside), okanin 3,4,3'-trimethyl ether 4'-glucoside and luteolin 4'-methyl ether 7-(4G-rhamnosylneohesperidoside) ( Table 2). Methylated okanin derivatives can be found from Bidens torta Sherff (McCormick et al., 1984). McCormick et al. (1984) determined the structures of four methylated chalcones including okanin 3,4,3'-trimethyl ether 4'-glucoside. Rao et al. (2020) studied the response of phenolic compounds in rice to different growing conditions. Luteolin 4'-methyl ether 7-(4G-rhamnosylneohesperidoside) was one of the compounds determined from different rice varieties and growing locations. They discovered that the effect of cultivation environment on the concentration and antioxidant activity of this compound varied between rice varieties indicating the influence of both genetics and environment on the compound. Earlier, luteolin was reported to be richer in leaves of rust-susceptible M. sylvaticum compared to those of rust-resistant M. pratense (Kaitera & Witzell, 2016). Phloretin can be found in apple tree leaves. Antifungal activity of phloretin against several plant pathogenic fungi has been reported (Shim et al., 2010). Phlorizin, a glucoside of phloretin, is also present in the apple tree (root bark, shoots, leaves) and experimental evidence suggests that it plays a significant role in apple tree physiology (Ehrenkranz et al., 2005). For the bioactivity of phloretin 2'-O-(6"-O-acetylglucoside) we couldn't find any literature.
For the R. rubrum and R. nigrum species pair the most interesting compounds were ( +) -maackiain 3-O-glucoside, gliricidol, quercetin 3-(2"-p-coumarylglucoside) and quercetin 3-(2Gal-apiosylrobinobioside) (Table 4). ( +) -Maackiain 3-O-glucoside, also called sophojaponicin, belongs to the pterocarpans group of compounds. It has been isolated from the roots of Cicer judaicum Baksier, which is an annual herb from the Middle East (Stevenson & Veitch, 1996). Gliricidol is a flavonoid found from the methanolic extract of Gliricidia sepium (Jacq.) Steud. bark (Rastrelli et al., 1999). It has shown bioactivity against Artemia salina L. larvae. For the two quercetin derivatives we couldn't find any literature about their bioactivity in plants, but generally, the quercetin compounds are known to have many possible health effects on humans. Quercetin compounds were also rich in leaves of Melampyrum spp. in a recent study (Kaitera & Witzell, 2016). In conclusion, our quantitative results of the pre-selected compounds revealed two compounds, chlorogenic acid and quercitrin, whose concentrations differ significantly between rust-resistant and susceptible plant species. The literature also supported the probable bioactivity of these compounds against rust diseases. From the discovery approach, we could find additional compounds with a putative role in the plant defense against rust disease. It is also known that in infected wood of mature P. sylvestris, C. pini induced a 1.3-108 fold increase in concentrations of monoterpenes, resin acids and several sesquiterpenes compared to control wood . In P. albicaulis Engelm. seedlings, terpene concentrations were higher in C. ribicola -resistant trees compared to susceptible ones (Bullington et al., 2018). Also C. quercuum f.sp. fusiforme Burds. & G.A.Snow -susceptible P. elliotii Engelm. trees contained lower amounts of some monoterpenes than resistant ones (Michelozzi et al., 1991). Therefore, monoterpenes are important compounds in Cronartium resistance to Pinus spp. Further research is needed to describe the temporal and spatial variation of the compounds in alternate host plants of Cronartium. Inoculation tests in controlled environment should be done to study the induced chemical changes in alternate hosts due to rust infections. Also the effect of leaf extracts and individual compounds of extracts of rust-resistant species should be tested against Cronartium rusts in controlled experiments. images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.