Metabolomic analysis reveals that the accumulation of specific secondary metabolites in Echinacea angustifolia cells cultured in vitro can be controlled by light

Guarnerio et al. Plant cell Reports

Department of Biotechnology, University of Verona

ESM 2

Supplemental Report: annotation of MS signals

According to fragmentation data for the standards and/or data widely reported in the literature,the following basic rules were considered for the putative annotation of compounds. Neutral losses:162 = hexose;146 = deoxyhexose; 176 = glucuronic or galacturonic acid;86 =malonyl residue; m/z=179, MS/MS=135:caffeic acid.Various adducts were annotated on the basis of the table provided by Huang et al. (1999), and (with regard to chloride adducts) on the basis of the relative abundance of M+35 and M+37 adducts (McLafferty and Tureček 1993).

Phenylethanoid glycosides (PhGs)

Various hydroxytyrosol-based PhGs were found. Hydroxytyrosol-hexoside (id 40, m/z=315) was annotated on the basis of MS/MS =153 and MS3=123 [M-H-CH2O]-, which differentiates it from dihydroxybenzoic acid (ms/ms=153, ms3=109 [M-H-COO]-).

The presence of a fragment with m/z= 315 was taken to indicate PhGs: echinacoside (id 5) (Ma et al., 2008), the marker PhG of Echinacea, and verbascoside (id 36) were the more abundant and their respective isomers (id 56 and 255) were also present. The existence of two verbascoside isomers is reported byRyan et al. (1999)and Ghisalberti (2000). Another very abundant PhG is id20, with m/z = 477 (315+162), annotated as derhamnosyl verbascoside, already found in Echinacea (Cheminatet al. 1988). Ids 120 (m/z=769) and 142 (m/z=799), characterized by the neutral loss of 146 and 176 respectively, were annotated as rhamnosyl and glucuronylverbascoside.Id 214, with m/z=637, giving a fragment with m/z = 461 (loss of 176), eluting about 4-5 min after verbascoside, could be a methylated form of verbascoside (e.g. leucosceptoside A, Sasaki et al. 1989).Id 213, with m/z= 639 and fragment with m/z=477 [M-H-162]-and ms3 fragments with m/z= 314.8, 160.7 and traces of signals at m/z=179 and 135, eluting 4 min before id 20, could correspond to a glucosylated form of id 20, i.e. to hydroxytyrosol-caffeoyl-hexosyl-hexoside (plantamajoside, Rønsted et al. 2000).

Caffeoylquinic acids (CQA)

Two different caffeoylquinic acids (ids 3and 31)were annotated on the basis of a parent ion of m/z=353, ms/ms fragment with m/z= 191 and very weak ms/ms ion at m/z 179. According to Clifford et al. (2005), this fragmentation should correspond to 1-CQA or 5-CQA, which are distinguishable only by their retention times.

Four molecules (ids 1, 12, 18, 138) with m/z=515 and at ms/ms=353 elute at later retention times. Ids 1, 12 and 138 give a ms3 base peak at m/z=190.8, id 18 at m/z=172.7. Moreover, id 18 shows strong ms/ms ions at m/z= 299 and 203. According to Clifford et al. (2005), id 18 should correspond to 1,4-dicaffeoylquinic acid,and ids 1 and 12, with weak m/z=335 and 179 ions, to 1,5dicaffeoylquinic acids;id 138, with strong m/z=179 fragment, and much lower retention time, to 1,3-dicaffeoylquinic acid.Id 21 (m/z=677) gives a ms/ms base peak at 515 [M-H-162]-, and weak peaks at 353 [M-H-324]- and 485 [M-H-192]-; the fragmentation of the m/z = 515 fragment is very different from that of the dicaffeoylquinic acids discussed above, since it gives a ms3 base peak at 323 instead of 353; this fragmentation suggests that id 21 could be a caffeoylquinic acid hexosidewith an additional hexosyl or caffeoylresidue, not linked to quinic acid.

Id 205 (m/z=633) is an unknown derivative of caffeoylquinic acid, giving rise to a ms/ms peak with m/z=353 [M-H-280] and a ms3 peak with m/z=191.

Id 45 has been annotated as coumaroylcaffeoylquinic acid based on m/z=499 and ms/ms peaks at m/z=353 [M-H-146], 337 [M-H-162], 191 (quinic acid), 179 (caffeic acid) and 163 (coumaric acid); id 103, with m/z=367 and ms/ms base peak with m/z=191 (quinic acid) and weak peak with m/z=193 (ferulic acid) has been annotated as feruloylquinic acid; id 54 (m/z=529) has been annotated as feruloylcaffeoylquinic acid based on its ms/ms peaks at 367 [M-H-162], 353 [M-H-176], 191 (quinic acid) and ms3peaks at m/z=191 (quinic acid), 193 (ferulic acid) and 179 (caffeic acid).

Caffeoyl-hexaric acids (CHA)

Many signals produced a 209 fragment in ms/ms and/or ms3.Id 89is very hydrophilic,and is a M+97 adduct of a molecule with m/z= 371, undergoing a neutral loss of 162; this molecule should therefore be a tetrahydroxyhexanedioic acid hexoside.Six other signals(ids6, 14, 41, 42, 80, 105),had m/z=533 (209+324) and a ms/ms base peak=371 [M-H-162]. A 191 peak was also present. Fragmentation of the 371 base peak gave rise, in addition to the 209 base peak, to weaker peaks with m/z=191 and 353. Other peaks with m/z=146.7 and 128.8 and lower intensities were also present. These six molecules were annotated as dicaffeoyl hexaric (2,3,4,5-tetrahydroxy-hexanedioic)acids, since peaks with m/z=209, 191, 147, 129 are reported in the Human Metabolomics Database (Wishart et al., 2009) for the fragmentation of galactaric acid (HMDB00639), whereas for hydroxyferulic acid, which also has m/z=209, very different fragmentation is reported (Simirghiotis et al. 2009). Dicaffeoylhexaric acids have already been found in other Asteraceae such asSmallanthus sonchifolius(Takenaka et al. 2003) and Eupatorium perfoliatum (Maas et al. 2009). The six molecules could differ in the hexaric acid and/or the position of the caffeoyl groups.

Ids 4 and 7, with m/z=695, ms/ms base peak at 533 [M-162], and higher retention time, should correspond to tricaffeoylhexaric acids, as reported by Takenaka et al. (2003). Another five molecules with m/z = 857 gave rise to ms/ms ions with m/z=695, 533. Two of these (ids 15 and 114) had very high retention times (36-39 min), and the fragmentation of the base peak with m/z=533 is very similar to that of the above molecules, suggesting that ids 15 and 114 are tetracaffeoylhexaric acids. Id 81 strictlycoelutes with id 7 and seems to be its M+162 adduct. Id97elutes earlier than the tricaffeoylhexaric acids, which suggests the presence of an additional sugar residue.Id 154 is characterized by the loss of 324/342 amu (absence of the m/z 695 fragment), and in ms3 by the presence of 191 as base peak. Its low retention time suggests high hydrophylity, with the possible presence of two sugar units.Id 66 (m/z=1019) has a further 162 amu residue.

Caffeoyltartaric acids (CTA)

Id 47 (m/z=311) gives fragments 179 (caffeic acid) and 149 (which shows the fragmentation of tartaric acid standard) and has therefore been annotated as caffeoyltartaric acid. Id 61 (m/z=473) gives an ms/ms base peak at m/z = 311 [M-H-162]-, with the same fragmentation of caffeoyltartaric acid. The much higher retention time indicates this is a dicaffeoyltartaric acid, although the presence of 341 and 219 fragments (not typical of 2,3-dicaffeoyl-tartaric acid, Buiarelli et al. 2010) suggests that the two caffeoyl residues could be linked together.

Id 71 hasm/z=325 and ms/ms fragments with m/z= 193(ferulic acid) [M-H-132]. It has been annotated as feruloyltartaric acid, due to the presence of weak ms/ms ions with m/z=131 and 103 characteristic of tartaric acid.

Other caffeic acid derivatives

Various signals generating fragments with m/z=179 and 135 were annotated as caffeic acid derivatives. Id 74, m/z=341 and neutral loss of 162 amu is composed by a caffeic acid and a hexose residue.Id 125 is the formic acid adduct of a molecule with m/z=341, whose fragmentation is very different from that of id 74, giving rise to many fragments typical of sugar fragmentation. Since the retention time is much higher than that of sucrose, we annotated id 125 as a another caffeic acid-hexose compound, in which the link between the two moieties should be different from that of id 74.

Id 34 and id 189, with m/z=473, slight different retention time and very different abundance, giving rise to a m/z=341 ms/ms base peak with a neutral loss of 132, were annotated as caffeic acid-hexose-pentose.

Id 70 and id 192, with m/z=457 and base ms/ms peak 341 [M-H-116] are other caffeic acid-glucose derivatives in which the loss of 116 could correspond to a malic acid or a deoxypentose residue.

Id 46 and id 87 (m/z=741) with the loss of 418and id 58 (m/z=711) with the loss of 388, generate a ms/ms base peak with m/z=323 containing caffeic acid.

Miscellaneous

Many signals were found at retention times comprised between 2.6-3 min. Among these, different sugars and their adducts were annotated as hexose (id 84,m/z=179 and its Cl adduct id 76), dihexoses (id 9, m/z=341, and many adducts), trihexoses (id 44 and many adducts).

Id 38 (m/z=447) and id 73 (m/z=437) are the formic acid and the 35Cl adducts of id 155 (m/z=401), a trihydroxyflavone pentoside (ms/ms base peak with m/z=269).

Id 17 (m/z=621) and id 23 (m/z=579) are very abundant molecules that give rise to a ms/ms base peak with m/z=417, with the loss respectively of 204 (acetylhexose) and 162 (hexose). The m/z=417 fragment (id 149) has a ms3 base peak with m/z=181. At the same retention time of id 17 there is a weak signal with m/z=665, and with m/z=689 [M+Na] in positive mode. As we found previously (Strazzer et al. 2011), the loss of 44 from malonylhexoside is common in negative mode. Id 17 and id 23 should therefore be the malonylhexoside and the hexoside of the unknown molecule with m/z=418.

Id 100 (m/z=299) has been annotated as hydroxybenzoic acid-hexose, for the presence of 137 and 93 fragments in ms3.

References

Buiarelli F, Coccioli F, Merolle M, Jasionowska R, Terracciano A (2010) Identification of hydroxycinnamic acid–tartaric acid esters in wine by HPLC–tandem mass spectrometry. Food Chem 123:827–833

Cheminat A Zawatzky R, Becker H, Brouillard R. (1988) Caffeoyl conjugates from Echinacea species: structuresand biological activity. Phytochemistry 27:2787-2794

Clifford MN, Knight S, Kuhnert N (2005) Discriminating between the Six Isomers of Dicaffeoylquinic Acid by LC-MS. J Agric Food Chem 53:3821- 3832

Ghisalberti EL (2000) Lantana camara L.(Verbenaceae). Fitoterapia 71:467-486

Huang N, Siegel MM, Kruppa GH, Laukien FH (1999) Automation of a Fourier transform ion cyclotron resonance mass spectrometer for acquisition, analysis, and e-mailing of high-resolution exact-mass electrospray ionization mass spectral data. J Am Soc Mass Spectrom 10:1166–73

Ma C, Hattori M, Chen H, Cai S, Daneshtalab M (2008) Profiling the Phenolic Compounds of Artemisia pectinata by HPLC-PAD-MSn. PhytochemAnal 19:294–300)

Maas M, Petereit F, Hensel A (2009) Caffeic Acid Derivatives from Eupatorium perfoliatum L.. Molecules 14:36-45

McLafferty FW, Tureček F (1993) Interpretation of mass spectra. University Science Books (ISBN 0-935702-25-3)

Rønsted N, Göbel E, Franzyk H, Jensen SR, Olsen CE (2000) Chemotaxonomy of Plantago. Iridoid glucosides and caffeoyl phenylethanoid glycosides. Phytochemistry 55:337-348

Ryan D, Robards K, Prenzler P, Jardine D, Herlt T, Antolovich M (1999) Liquid chromatography with electrospray ionisation mass spectrometric detection of phenolic compounds from Olea europaea. Journal of Chromatography A, 855:529–537

Sasaki H, Nishimura H, Chin M, Mitsuhashi H (1989) Hydroxycinnamic acid esters of phenethylalcohol glycosides from Rehmannia glutinosa var. Purpurea. Phytochemistry 28:875-879

Simirgiotis MJ, Caligari PDS, Schmeda-Hirschmann G (2009) Identification of phenolic compounds from the fruits of the mountain papaya Vasconcellea pubescens A. DC. grown in Chile by liquid chromatography–UV detection–mass spectrometry. Food Chem 115:775–784

Strazzer P, Guzzo F, Levi M (2011) Correlated accumulation of anthocyanins and rosmarinic acid in mechanically stressed red cell suspensions of basil (Ocimum basilicum). J Plant Physiol 168:288-293

Takenaka M, Yan X, Ono H, Yoshida M, Nagata T, Nakanishi T (2003) Caffeic Acid Derivatives in the Roots of Yacon (Smallanthus sonchifolius) J Agric Food Chem 51: 793-796

Wishart DS, Knox C, Guo AC, et al.(2009) HMDB: a knowledgebase for the human metabolome. Nucleic Acids Res 37(Database issue):D603-610.

1