1
SUPPLEMENTAL RESULTS (TEXT)
Interspecific Comparison of Constitutive Ash Phloem Phenolic Chemistry Reveals Compounds Unique to Manchurian Ash, a Species Resistant to Emerald Ash Borer
Justin G.A. Whitehill1,2*, Stephen O. Opiyo3, Jennifer L. Koch4, Daniel A. Herms5, Donald F. Cipollini6, and Pierluigi Bonello1
1Department of Plant Pathology, The Ohio State University, 2021 Coffey Road, Columbus, OH 43210
2Michael Smith Laboratories, University of British Columbia, 301-2185 East Mall, Vancouver, BC, Canada V6T 1Z4
3Molecular and Cellular Imaging Center-Columbus, Ohio Agricultural Research and Development Center, 2021 Coffey Road, Columbus, OH 43210
4Northern Research Station, USDA Forest Service, 359 Main Road, Delaware, OH 43015
5Department of Entomology, The Ohio State University, Ohio Agricultural Research and Development Center, 1680 Madison Ave., Wooster, Ohio 44691
6Department of Biological Sciences, Wright State University, 3640 Colonel Glenn Highway, Dayton, OH 45435
Justin G. A. Whitehill ()
Michael Smith Laboratories
The University of British Columbia
301-2185 East Mall
Vancouver, BC V6T 1Z4
Canada
E-mail:
Simple Phenolics and Phenolic acids- Three compounds were classified as simple phenolics (1, 3, 11) and five compounds were classified as phenolic acids (2, 5, 8, 23, and 35). Compound (23) was unequivocally confirmed as 3-caffeoyl-quinic acid (chlorogenic acid). Simple phenolics hydroxytyrosolhexoside(1) and tyrosolhexoside(3) exhibited a [M-H]-, fragmentation patterns, and UV maxima consistent with identities from Eyles et al. 2007. Compound (11)exhibited a [M-H]-ofm/z 431 and UV spectral maximum at 275.3 nm. Fragmentation of [M-H]- resulted in a fragment ofm/z 299 corresponding to the loss of a pentose moiety (132 Da). Further fragmentation of m/z 299 resulted in a pattern matching that of tyrosolhexoside. Therefore, compound (11)was classified as tyrosolhexosidepentoside(Eyles et al. 2007; Vallverdu-Queralt et al. 2010). Compounds (2)and (5)both exhibited a [M-H]- of m/z 389 and similar UV maxima (276.5 and 280.1 nm, respectively). Fragmentation of m/z 389 resulted in a dominant fragment of m/z 329, indicating a neutral loss of 60 Da, corresponding to an acetate adduct. Fragment ions ofm/z 167 and m/z 161, combined with a dominant fragment ofm/z 329, identify the compound as vanillic acid hexoside acetate adduct (Eyles et al. 2007; Jimenez et al. 2010). Compound (8) had a [M-H]- of m/z 353 and UV maximumat 326.4 and shoulder (sh)at 295 nm. Fragmentation of m/z 353 led to daughter ions of m/z 191 (162 Da loss of caffeoyl moiety) andm/z 179, which is consistent with caffeoyl-quinic acid (Cardoso et al. 2005; Eyles et al. 2007). We identified compound (8)as caffeoyl-quinic acid A (Cardoso et al. 2005). Compound (35)had a [M-H]- of m/z 335 and UV maximum at 327.6 and shat 300 nm. Daughter ions of m/z 179 and m/z 135 were produced from the fragmentation of m/z 335. These data are consistent with those forcaffeoylshikimic acid (Fang et al. 2002; Lin and Harnly 2008).
Coumarins - Twelvecoumarinswere detected on each sampling date (6, 7,10, 12, 13, 16, 18, 19, 20, 21, 22, and 54). The coumarinsesculin(12), esculetin(13),fraxin(18), and fraxetin(22) were unequivocally confirmed by spiking samples with standards and comparing [M-H]-, retention time, and UV maxima. In our plant extracts, we were able to identify compounds (7, 10, 12, 13, 18, 20, and 54) in the negative ion mode, while compounds (6, 16, 19, 21, and 22) were only detected in the positive ion mode.
In the negative ion mode, compounds (7), (10), and (20) were identified as esculetinA, esculetin B, and esculetin C, respectively. Each compound had a [M-H]- of m/z 177, fragment ions of m/z 133 and m/z 89, as well as UV maxima consistent with or similar to the esculetin standard, but with differing retention times from the esculetin standard. Compound (54) gave a [M-H]-of m/z 545, which, following fragmentation, displayed a dominant signal of m/z 369 (neutral loss of 176). Fragmentation of m/z 369 gave fragments of m/z 207 and m/z 192 which are consistent with the fragmentation of the fraxin standard. The loss of 176 Da can correspond to several functional groups, therefore we identified compound (54) as a fraxin related compound.
In the positive ion mode, compound (6) exhibited a [M-H]-of m/z 385, with a dominant MS2 fragment of m/z 209. Further fragmentation of m/z 209 was consistent with the fraxetin standard. Therefore, compound (6) was identified as fraxetin related compound. Compound (16) and (19) were identified as fraxidin A and B due to a [M-H]-of m/z 223, fragmentation pattern, and UV spectra that were all consistent with the fraxidin standard, except for a difference in retention time. Compound 21 was identified as mandshurin as all data regarding compound structure were consistent with the identification proposed by Eyles et al. 2007.
Lignans - Four compounds were classified as lignans (25, 28, 37, and 39) and each displayed similar UV maxima consistent with those published in the literature and the pinoresinol(39)standard. Compound (25)had a [M-H]- of m/z717. While our literature search revealed no compounds matching m/z 717 in, we did note a mass difference of 36 Da between compound (25)and pinoresinoldihexoside(Eyles et al. 2007), which likely corresponds to the presence of two water residues. Further investigation of compound (25)revealedsimilar RT and UV spectra as that of Eyles et al. 2007, as well as identical MS2 spectra (m/z 519 and m/z 357) corresponding to the successive losses of two hexoside (162 Da) residues from the parent ion. Based on this evidence we tentatively identified compound (25)as pinoresinoldihexoside + 2 H2O. Compound (28)displayed a [M-H]- of m/z 571 with daughter ions of m/z373, 535, and 343. The daughter ions matched to (+)-1-hydroxypinoresinol-4’-O-glucoside (Guo et al. 2007) along with similar UV spectra. We noted a mass difference of 36 Da, corresponding to the loss of two water residues, between the parent ion (m/z 571) and (+)-1-hydroxypinoresinol-4’-O-glucoside, therefore we called compound (28)(+)-1-hydroxypinoresinol-4’-O-glucoside + 2 H2O. Compound (37)had a [M-H]- of m/z 357 and dominant daughter ion fragments of m/z 151 and m/z 136 along with a UV maximum at 276.5 nm. These data were consistent with the pinoresinol standard except for a slight shift in the retention time and difference in UV maxima. We, therefore, tentatively identified this compound as pinoresinol A.
Monolignols - Compound (15) was classified as a monolignol and its identity was confirmed by comparison against the syringin standard. Compound (15) was detected in positive ion mode but not in negative ion mode, although it was previously detected in that mode by Eyles et al. 2007. In the positive ion mode, compound (15) had a [M-H + Na]+of m/z 395 and UV maximum at 264.7 nm. Fragmentation of the parent ion led to daughter ions of m/z 233 and m/z 217 and correspond to sodium adducts of the major fragments detected by Eyles et al. (2007) in the negative ion mode.
Secoiridoids- Twelve compounds were classified as secoiridoids(9, 27, 33, 36, 38, 44, 48, 49, 51, 52, 56, and 57). Oleuropein(51) was identified by comparison with the standard. Compound (9) gave a [M-H]- of m/z 601 and was found to co-elute with compound (11). Fragmentation of m/z 601 led to dominant daughter ions of m/z 403 and m/z 223 which gave the identification as elenolic acid derivative 1 and was described previously by Eyles et al. (2007). Compounds (27) and (33)shared similar UV maxima, [M-H]- and fragmentation patterns, and matched to ligustroside(57)(Eyles et al. 2007). Therefore compounds (27) and (33)were tentatively named ligustroside A, and ligustroside B, respectively. Compound (36) gave a [M-H]- of m/z 555 and dominant daughter ion of m/z 393 (loss of 162 Da) corresponding to the aglycone resulting from the loss of a hexose moiety. These data were consistent with the identity of 10-hydroxyoleuropein (Hosny 1998). Compound (38) had a [M-H]-of m/z 525 and a dominant, decarboxylated MS2 fragment of m/z 481. MS2 fragments of m/z 389 and m/z 319 combined with the parent ion of m/z 525 were previously identified as being characteristic of dimethyoleuropein(Savarese et al. 2007). Compounds (44)and (48)displayed UV maxima, [M-H]-, and fragmentation patterns that matched the oleuropein standard but with different retention times, and were therefore identified as oleuropein A and oleuropein B. Compound (49) exhibited a [M-H]- of m/z 509 and fragmentation gave a dominant [M-H]- of m/z 347, corresponding to the loss of a hexose (162 Da). A shoulder at 278 nm was in agreement with the UV spectra observed for ligustroside and led to the putative identification of compound (49) as demethylligstroside(Takenaka et al. 2000). Compound (54) gave a [M-H]- of m/z 793.9 and fragmentation led to dominant daughter ions of m/z 403 and m/z 223, corresponding to an elenolic acid derivative similar to compound (9), which led to the identification of compound (52) as elenolic acid derivative 2. Compound (56) had a [M-H]- ion of m/z 553 and dominant daughter fragments of m/z 391, m/z 321, and m/z 289. These fragments display a mass difference of 14 Da from the [M-H]- (m/z 539) and main fragments (m/z 377, m/z 307, and m/z 275) of oleuropein, indicating the addition of a methyl group to oleuropein. Compound (56) also had a similar UV maximum to oleuropein and we therefore gave it a tentative identification of oleuropein related compound 2. Compound (57) wasidentified as ligustroside and was described previously by Eyles et al. 2007.
Phenylethanoids- Twelve compounds were identified as phenylethanoids(26, 32, 34, 40, 41, 42, 43, 45, 46, 50, 53, and 64). Verbascoside(45) was confirmed by direct comparison with a standard. Compounds (26) and (43) each exhibited a [M-H]- of m/z 639 and fragmentation led to successive fragments of [M-H – 162]-m/z 477, which yielded characteristic fragments (m/z 315 and m/z 135) of another phenylethanoid, calceolarioside. Compounds (26) and (43) also had UV maxima at 329.9 nm and a shoulder at 295nm, which led to the tentative identification of compounds (26) and (43) as lugrandoside A and B, respectively. Compounds (42) and (50) were described previously by Eyles et al. (2007) and were identified as calceolarioside A and B, respectively. Compound (32) had a similar UV spectrum (maximum at 326.4 nm and a shoulder at 290 nm), [M-H]-(m/z 477), and dominant daughter fragments (m/z 315 and m/z 135) to compounds (42) and (50) and was, therefore, identified as calceolarioside C. The characteristics of these three compounds were consistent with those found in the literature for calceolariosides(Eyles et al. 2007; Guo et al. 2007). Compounds (34) and (41) exhibited a [M-H]- of m/z 785 and dominant daughter fragments of m/z 623 and m/z 461, which are consistent with the previously described forsythoside A-O-glucoside(Guo et al. 2007). Therefore, we identified compounds (34) and (41) asforsythoside A-O-glucoside – A and forsythoside A-O-glucoside – B, respectively. Compound (40) had a [M-H]- of m/z 507 along with notable daughter fragments (m/z 475, 323, 179, and 161) and UV spectrum that matched to the published description of -methoxylferruginoside(Guo et al. 2007). Finally, we identified compounds (46), (53), and (64) as verbascoside A, B, and C, respectively, as each compound displayed similar UV spectra, [M-H]-, and fragmentation patterns as the verbascoside standard, with varying retention times.
Flavonoids- Of the five flavonoids detected, the identities of compounds (55), (65), and (66) were confirmed by comparison with standards, while compound (58) had a [M-H]- of m/z 431 with a dominant daughter ion of m/z 269. Subsequent fragmentation of m/z 269 and the UV spectrum were identical to the apigenin standard. Therefore, we identified compound (58) as apigeninglucoside. Compound (61) gave a [M-H]-of m/z 447 and daughter fragments of m/z 285, 255, 327, 227, and 211. UV maxima at 350 and 264 nm led to the tentative identification of compound (61) as kaempferolgalactoside(Ye et al. 2005).
Coumarins-secoiridoids- Compound (47) was the only coumarin-secoiridoid detected, and had a [M-H]-of m/z 725 and a dominant daughter ion [M-H – 386]- of m/z 339. A neutral loss of 386 Da indicates a secoiridoidal group. The daughter ion of m/z 339 fragmented in the same manner as the esculin standard. Therefore, compound (47) was identified as escuside, which is a combination of a secoiridoid and a coumarin (Iossifova et al. 2002).
Unknowns- Ten of the initial 66 compounds selected based on HPLC-PDA chromatograms (Figs. 1-3) were not tentatively identified because they could not be matched reliably to known species using MS data, or a putative identity was not found in the literature. However, each compound could be sorted into a respective compound class based on its UV spectrum and retention time. Therefore, compound (4) was identified as unknown phenolic acid 1; (14), (59), and (60) as unknown coumarin 1, 2, and 3; (17) as unknown monolignol 1; (24) and (30) as unknown lignan 1 and 2; (31) and (62) as unknown secoiridoid 1 and 2; and (63) as unknown phenylethanoid 1. Grouping compounds into their respective classes allowed quantification based on equivalent standards and UV spectra.
Qualitative and Quantitative Differences in Phenolic Profiles- On both sampling dates, we detected a total of 30 compounds in black ash cv. ‘Fallgold’, 17 in blue ash, 22 in European ash, 16 in green ash cv. ‘Patmore’, 23 in seedling-propagated green ash, 27 in Manchurian ash cv. ‘Mancana’, and 23 in white ash cv. ‘Autumn Purple’(Table 1 and S1; Figures 1 – 3). The following compounds were only detected in June: (4), (6), (8),(11), (17), (26), (30), (37), (40), (41), (43), (48), (49), (55), and (64). Compounds detected only in August include: (9), (14), (20), (46), (47), (52), (53), (54), (56), (59), (60), (62), and (63).
Qualitative patterns of phloem chemistry among the taxareflected their phylogenetic relationships (Table 1, 2, S1, and S2; Figs. 1 – 4, and Figs. S1, S2, S4, and S5). There was a high degree of overlap in the phenolic profiles of the species belonging to the section Meloides, with green ash cv. ‘Patmore’ and seedling-propagated green ash exhibiting complete overlap, and both green ash taxa sharing 94% of their compounds with white ash cv. ‘Autumn Purple’. There was also considerable overlap in the phenolic profiles of taxa belonging to the section Fraxinus with black ash cv. ‘Fallgold’ and European ash that shared 77.8% and 66.7% of their compounds, respectively, with Manchurian ash cv. ‘Mancana’.
There was considerably more qualitative variation in the phenolics of more distantly related taxa. Manchurian ash cv. ‘Mancana’ shared only 41% of its compounds with blue ash, a member of the section Dipetalae, 30% of compounds with white ash cv.’ Autumn Purple’, 30% with seedling-propagated green ash, and 26% with green ash cv. ‘Patmore’, all of which belong to the section Melioides. Blue ash shared only 24% of its compounds (1, 15, 36, and 39) withtaxa in the Melioides section.
Seven metabolites that were detected in seedling-propagated green ash[(2), (5), (52), (53), (55), (62), and (65)]were not detected by HPCL-UV in green ash cv. ‘Patmore’. LC-MS full scan mode data did detect the presence of [M-H]-ions corresponding to compounds (5), (53), and (55) in green ‘Patmore’ extracts, although at much lower quantities than in seedling-propagated green ash.
The phenolic chemistry of blue ash was unique, localizing to its own PCA quadrant (Figs. 1, S1, and S4) and forming its own cluster (Figs. S2 and S5). Blue ash phloem is characterized primarily by the presence of hydroxycoumarins (Table 1 and S1), of which esculin(12) is particularly characteristic (Table 2 and S2; Figure 2, S3, and S4). Only a few compounds were unique to the resistant Manchurian ash cv. ‘Mancana’, and concentrations of individual phenolics tended to be similar to those of susceptible black ash cv. ‘Fallgold’ (Table 2 and S2; Fig. 3). Compounds (25), pinoresinoldihexoside, and (60), unknown coumarin 3, were only identified in Manchurian ash cv. ‘Mancana’ phloem. Trees belonging to the Melioides, green ash cv. ‘Patmore’, seedling-propagated green ash, and white ash cv. ‘Autumn Purple’ were very different from the other species under investigation, but were virtually indistinguishable from one another qualitatively (Table 1, 2, S1, and S2; Fig. 1 and S1). Overall, concentrations of compounds tended to be higher in August than in June across all species (Table 2 and S2, respectively).
Because the PCA for June showed similar trends and relationships as August, we report results for only the August sample (Table 1 and 2). The first two principal components accounted for 92% of the variation and revealed four clusters, one in each of the quadrants (Fig. S1). Bootstrap analysis supported the four clusters (Fig. S2). A separate cluster analysis based on the individual 66 compounds revealed that seven metabolites, esculin(12), syringin(15), fraxin(18), mandshurin(21), calceolarioside A (42), calceolarioside B (50), and oleuropein(51), clustered separately with low AU values (Fig. S3). Furthermore, a biplot analysis exploring the relationships among the species and the metabolites that were responsible for the individual clusters (Fig.S4) revealed esculin(12) to be the major variable separating blue ash from the other species; syringin(15) separated green ‘Patmore’, seedling-propagated green ash, and white ash cv. ‘Autumn Purple’ from the other taxa; fraxin(18), mandshurin(21), and oleuropein(51) separated European from Manchurian ash cv. ‘Mancana’ and black ash cv. ‘Fallgold’; and calceolarioside A (42) and calceolarioside B (50) separated the Manchurian and black ash cultivars.
Based on the results of the PCA, cluster analysis, and biplot analysis, we explored how removing the seven major metabolites affected the species groupings. The subsequent PCA revealed results similar to the first, but with important differences. The first two principal components accounted for 91% of the variance and produced three clusters consistent with the three sections of Melioides, Dipetalae, and Fraxinus (Fig. 4). To test the statistical reliability of the clusters found in this PCA plot, we again used bootstrap analysis and obtained three highly supported clusters, again representing the three Fraxinus sections (Fig. S5).