Note pubs.acs.org/jnp
Detection of a Toxic Methylated Derivative of Phomopsin A Produced by the Legume-Infesting Fungus Diaporthe toxica Svenja Schloß,† Thomas Hackl,‡ Corinna Herz,§ Evelyn Lamy,§ Matthias Koch,⊥ Sascha Rohn,† and Ronald Maul*,†,∥ †
Hamburg School of Food Science, Institute of Food Chemistry, University of Hamburg, Grindelallee 117, 20146 Hamburg, Germany NMR Facility, Department of Chemistry, University of Hamburg, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany § Institute for Prevention and Cancer Epidemiology, Molecular Preventive Medicine, University of Freiburg−Medical Center, Elsässerstraße 2, 79110 Freiburg, Germany ⊥ Department 1, Analytical Chemistry; Reference Materials, Bundesanstalt für Materialforschung und -prüfung (BAM), Richard-Willstätter-Straße 11, 12489 Berlin, Germany ‡
S Supporting Information *
ABSTRACT: Phomopsin A (PHO-A), produced by the fungus Diaporthe toxica, is a mycotoxin known to be responsible for fatal liver disease of lupin-fed sheep. The full spectrum of the toxic secondary metabolites produced by D. toxica is still unknown. PHO-A and the naturally occurring derivatives B−E have been subject to several studies to reveal their structures as well as chemical and toxicological properties. In this work, a methylated derivative (1) of PHO-A isolated from lupin seeds inoculated with D. toxica is described. It was characterized by high-resolution mass and NMR data and shown to be the N-methylated derivative of PHO-A. 1 is cytotoxic against HepG2 cells.
T
the prevailing environmental conditions. Additionally, fungi are known to be very adaptable, leading to the colonization of different environments and the expression of different metabolite profiles. The identification of derivatives of known mycotoxins is essential for a better understanding of the toxin contamination. While establishing a quantitative LC-MS/MS method for PHO-A, first unpublished indications were obtained leading to the discovery and structure elucidation of a novel phomopsin derivative (1), produced by D. toxica as described in the present work. The presence of 1 was primarily observed in an extract of a fungal culture of D. toxica analyzed by high-resolution mass spectrometry (HRMS). Moreover, the compound was observed in a presumably pure and commercially available standard of PHO-A. The mass signal of PHO-A (m/z 789.2864) was accompanied by an additional signal (m/z 803.3020) at the same retention time and a similar isotopic pattern originating from the characteristic isotopic distribution of 35Cl/37Cl. The mass difference of +14.0156 Da indicated a modification of PHO-A with a methyl group (Figures S1 and S2, Supporting Information). Further investigations of the fragmentation pattern revealed the same mass difference in all observed fragments (Figures S3 and S4, Table S1, Supporting
he sweet lupin (Lupinus spp.), a protein-rich legume crop, is gaining importance in the food industry, especially as an alternative to genetically modified soy beans.1 The revival of legumes, demonstrated by the U.N.’s designation of 2016 as the International Year of Pulses, requires a renewed assessment of health and toxicological aspects. Similar to Fusarium infections in cereals, lupins can be afflicted by Diaporthe toxica, a parasitic and saprophytic fungus, causing blights, crop losses, and severe toxin contaminations.2 Sheep are prone to lupinosis, a disease that manifests in adverse effects such as lethargy, weight loss, and jaundice.3 These symptoms are triggered by antimicrotubular properties of secondary metabolites of D. toxica, the phomopsins.4,5 Phomopsins are a group of hexapeptide macrocyclic mycotoxins.6−8 A comprehensive review of phomopsins has already been published by the European Food Safety Authority in 2012.9 Several studies have been carried out to gather information about the occurrence of phomopsins and to improve analytical techniques.10,11 It has been recently shown that D. toxica has the potential to grow and produce toxins on other legumes, e.g., peas and beans.12 The main toxin, phomopsin A (PHO-A), is often accompanied by small amounts of phomopsin B (PHO-B). Co-occurrence of the toxic phomopsins C, D, and E has also been reported.13−15 The full profile of phomopsins, however, has not yet been comprehensively investigated. The biosynthesis of fungal secondary metabolites is influenced by genetic diversity and © 2017 American Chemical Society and American Society of Pharmacognosy
Received: July 15, 2016 Published: June 14, 2017 1930
DOI: 10.1021/acs.jnatprod.6b00662 J. Nat. Prod. 2017, 80, 1930−1934
Journal of Natural Products
Note
Figure 1. Assumed location of the modification of phomopsin A with an additional methyl group based on mass spectrometric data.
(Table 1). Some correlation signals in the heteronuclear spectra were very weak or absent, attributed to the dynamic behavior of the molecule. Additionally, the NMR spectra revealed a nonphomopsin-like major impurity of approximately 50% (signals between 3.0 and 3.8 ppm). An impurity is also visible in the LC-DAD analysis of the isolated substance 1 as well as in isolated PHO-A (Figure S12, Supporting Information). Thus, the impurity is most likely a result of the isolation process and does not specifically interfere with further activity testing of 1. In order to establish the assignment, additional spectra were recorded in DMSO-d6. The 1H NMR spectra in both solvents exhibited a singlet integrating to 6 at 2.74 ppm (broadened, in MeOH-d4) and at 2.28 ppm (in DMSO-d6). The corresponding 13 C shifts were at 43.4 ppm (MeOH-d4) and 42.9 ppm (DMSO-d6), being revealed by HSQC correlation peaks and in agreement with two N-bound methyl groups. This was unequivocally confirmed by a correlation between both methyl groups in the HMBC spectra observed in both solvents (Figure S12, Supporting Information). Thus, 1 is the N,N-dimethyl derivative of PHO-A. In addition to solvent effects the difference in line shape and 1H shift could indicate different flexibility and conformation of the macrolactam substructure. Also, the olefinic proton of the maleic acid residue bound via an amide (cis/trans) shows different dynamic behavior in both solvents. While in DMSO-d6 two signals at 7.17 and 7.18 ppm in a ratio of 0.6 to 0.4 are observed in MeOH-d6, the same proton shows only an average signal at 7.53 ppm. The NMR spectra revealed a non-phomopsin-like major impurity of approximately 50% (signals between 3.0 and 3.8 ppm). However, all 1H and most 13C signals were assigned and have comparable shifts to previously published NMR data of PHOB.17 Although no elucidation of the absolute configuration of all stereochemical centers has been carried out, according to the chromatographic behavior, MS and NMR data, and the fact that an identical microorganism produces both molecules, we assume that the novel PHO-A derivative 1 possesses an analogous configuration to PHO-A. For obtaining information on the toxicological potential of 1, an initial in vitro screening of its cytotoxicity was performed. Table 2 indicates PHO-A and 1 impact viability as well as vitality of HepG2 cells from human liver in a concentrationdependent manner. At concentrations exceeding 0.4 μM, the methylated form of PHO-A induced more cell toxicity as
Information). In-source formation of a methyl adduct in the absence of methanol or a chemical transformation of PHO-A with a mass gain of one methyl group under the prevailing mild conditions was considered unlikely. The multiple reaction monitoring transitions for 1 were optimized and added to an existing HPLC-MS/MS method used for quantifying PHO-A.12 The comparison of commercially available PHO-A standard solutions with extracts obtained from infested lupin seeds, peas, and beans showed varying ratios of the amounts of PHO-A and 1 (Figure S5, Supporting Information). Should the observed mass spectrometric signal of 1 have been an artifact formed during ionization, the ratio would have been invariable, substantiating the identity of a new substance. As consecutive mass fragments may be formed by serial cleavage of PHO-A, and the mass difference from compound 1 is present in all observed fragments, the modification has to be located in the fragment of m/z 226 of PHO-A.16 The chemical nature and position of the additional methyl group were specified by an H/ D exchange experiment. The number of exchangeable protons in PHO-A and 1 can provide evidence if the additional methyl group in 1 is replacing an acidic proton of the hydroxyl or amine group or is attached to the carbon backbone (Figure 1). The HRMS spectrum of PHO-A and 1 shows a multitude of signals due to a dynamic H/D exchange (Figure S6, Supporting Information). The assignment of the mass signal to their probable isotopes (Tables S2 and S3, Supporting Information) is done by taking into account the distinguishable mass differences of 1.9971 Da between 35Cl and 37Cl and 2.0126 Da between H2 and D2. The mass difference between the calculated m/z and the actual values of their monoisotopic m/z of PHO-A as well as the novel metabolite 1 was
