Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

& NaturalProducts Pyrrolic and Dipyrrolic Degradation ProductsinPlants and Herbivores MarcelRitter,[a] Vincensius S. P. Oetama,[b] DanielSchulze,[c] Katrin Muetzlaff,[a] Anja K. Meents,[b] Raphael A. Seidel,[a] Helmar Gçrls,[c] MatthiasWesterhausen,[c] WilhelmBoland,[b] and Georg Pohnert*[a, b]

Abstract: The degradation of chlorophyll,the omnipresent uct familyisalso found in spit and feces of herbivores with green pigment, has been investigatedintensively over the specific metabolomic patterns reflecting the origin of the last 30 years resulting in manyelucidated tetrapyrrolic deg- samples. Based on chromatographicand mass spectrometric radationproducts. With acomparison to the degradation of evidenceaswell as on mechanistic considerations we also the structurally similarheme, we hereby propose anovel ad- suggest severaltentative new degradation products.One of ditionalchlorophyll degradation mechanism to mono- and them,dihydro BOX A, was fully confirmed as anovel natural dipyrrolic products.This is the first proof of the occurrence product by synthesis andcomparison of its spectroscopic of afamilyofmono- and dipyrrols in leaves that are previ- data. ously only known as degradationproducts.This prod-

Introduction regulated in .[6] In autumn,awell-studied chlorophyll degradation becomesvisible in abeautiful diversity of colors Chlorophyll (Chl) as amajor pigmentinplants, bacteria, and from senescent leaves.[7–11] However,Chl degradationcan also , is produced and degradedannually on a1012 kilogram be induced upon herbivore attack or pathogen infestation.[12, 13] scale.[1] The different Chl speciesexhibitaporphyrin structure In the first enzymatic degradation steps, the is with an additional 5-ringand Mg2+ as center . Except for cleaved by chlorophyllase leading to chlorophyllide,followed Chlc,one double bond in the d-ring is hydrogenated, resulting by ademetallation by Mg-dechelatase resulting in pheophor- in or dihydro .[2] The most abundantChla bidea (Scheme1).[14] As akey step, the macrocycle is opened exhibits four methyl, one ethyl and one vinyl substituents, as betweenthe Aand Brings by pheophorbideoxidaseleading well as amethyl and aphytol ester (Scheme 1).[3,4] to red chlorophyll catabolites(RCCs).[15] Adouble bond reduc- If the light absorption process in photosynthetic tissue tion at the bridge betweenthe Aand Drings leads to primary is overexcited,for example, under intensivelight conditions, or fluorescent chlorophyll catabolites, whichcan be further modi- the energy transfer is impaired, the absorbed energy can trig- fied by so far unknown or reactions to non-fluores- ger the production of reactive oxygen species (ROS).[5] Since cent chlorophyll catabolites (NCCs, Scheme 1).[16,17] Even if the these productsare cell toxic, the metabolism of Chl is highly pathwayand functions are not yet understood,structurally dif- ferentNCCs were detected especially in the vacuoles of plants [a] M. Ritter, K. Muetzlaff, Dr.R.A.Seidel, Prof. Dr.G.Pohnert and are considered end products of the chlorophyll degrada- FriedrichSchiller University Jena tion process.[18–22] Institute of Inorganic and Analytical Chemistry Lessingstr.8,07743 Jena (Germany) The furtherfate of Chl degradationproducts during abiotic E-mail:[email protected] decayaswellasherbivory remains open to alarge extent. Pre- [b] V. S. P. Oetama,A.K.Meents, Prof. Dr.W.Boland, Prof. Dr.G.Pohnert vious studies showedchlorophyll catabolite products in feces Max Planck Institute for Chemical Ecology of herbivores that are also known from enzymatic Chl degrada- Hans-Knçll-Str.8,07745Jena (Germany) tion in plants, forexample, chlorophyllide and pheophor- [c] D. Schulze, Dr.H.Gçrls, Prof. Dr.M.Westerhausen bide.[23,24] However,RCCs or similar compounds with an open FriedrichSchiller University Jena Institute of Inorganic and Analytical Chemistry macrocycle were not found so far in such samples since anaer- Humboldtstr.8,07743 Jena (Germany) obic conditions in the gut hinder potentialoxidative reactions Supporting information and the ORCID identification number(s) for the to open the macrocycle. As soon as the catabolites are outside author(s) of this articlecan be found under: the digestive system,the condition is drastically changed to https://doi.org/10.1002/chem.201905236. aerobic and the presence of light.[25,26] In this study,weaddress  2020 The Authors. Published by Wiley-VCH Verlag GmbH&Co. KGaA. the hypothesis, that under such conditions in leaves and feces, This is an open access article under the terms of the Creative Commons At- tributionLicense, whichpermits use, distributionand reproduction in any afurther Chl degradation to diversemono- or dipyrrolicprod- medium, provided the original work is properly cited. ucts occurs.

Chem. Eur.J.2020, 26,6205 –6213 6205  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

Scheme1.Enzymatic chlorophyll degradation in plants.

Asimilar degradation is reported for astructurally related Since the building blocks, and enzymatic degra- pigmentoflife, heme.[27] The enzymatic degradation through dation of heme and chlorophyll show ahighsimilarity,we to showsstriking similarities to the early searched for similar oxidative degradation pathways to pyrrols steps of Chl degradation (Scheme 2).[28,29] In mammals most of and dipyrrols for Chl. the bilirubin is conjugatedand excreted into the bile,[30] but an We investigated chlorophyll degradation products in sweet additional oxidative degradation occurs.ThisROS-mediated potatoleaves as wellasinthe spit and feces of the generalist breakdown leads to dipyrrolic propentdyopents(PDPs)[31] and herbivore Spodopteralittoralis. In vitro chlorophyll degradation monopyrrolic bilirubin oxidation end products (BOXes, was utilized to assign the structures. We could find PDPs and Scheme 2).[32–34] BOX Aand Bexhibit an intact vinyl and BOXesinthe completely new context of Chl degradation and differ only in the positionofvinyl and , whereas additionally detected new Chl catabolites, of which we proved BOX Cderives from the pyrrole. The dipyrrolic the structureofone of them by synthesis. PDPs also differ in the positionoftheir vinyl and methyl groups.Inaddition, two isomers in each structural group are observed(1for OH at vinyl pyrroleand 2for OH at propionic Results and Discussion acid pyrrole). Both isomersofeach group are in an equilibrium In vitro degradation and are provenintermediates to BOXes.[31] Thoseoxidative heme degradation products were quantified in nano- to micro- To identify pyrrolic and dipyrrolic Chl degradation products, molar concentrationswith a20–80 times higher abundance for we initially performed amodel oxidation in vitro. We used a PDPs compared to BOXesinbile,gallstones and the cerebro- procedure establishedfor the investigation of heme degrada- spinal fluid of stroke patients.[31,35,36] Heme degradation prod- tion to address the lower molecular weight fraction of Chl deg- ucts are involved in vessel constrictionsasstroke complication radationproducts.[31,32,39] Chl a,suspended in sodium hydroxide and act as effectors in the liver.[35–37] solution,was neutralized with hydrogen chloride before oxida- tion with hydrogen peroxide. Since the reaction is rather unspecific, diverseproductswith mono- and dipyrrolicstructures were found by UHPLC-MS in traces after one day.Prolonged incubation over sev- eral days did not lead to asubstantial increaseof these products.Potential reasons for the low detect- ed amounts are the insufficientsolubility of Chl in water,the high oxidative potential neededtobreak the macrocycle and, once the macrocycle is broken, a comparably easier further reaction of degradation products.Toavoid this, sodium hydroxide in metha- nol was used with 10%hydrogen peroxide, which re- sulted in higher amounts of degradation products. With this optimized method, we could prove by UHPLC/HR-MS and the addition of standards the oc- currenceofPDP B1/B2[31] as well as of BOX A, B, C[34,40, 41] and hematinic acid (HA)[34] as degradation products of Chla in vitro (Scheme 3, see Supporting Information, Figures S7–S12). The finding of BOX A seems surprising, considering that the corresponding Scheme2.Enzymatic (top) andoxidativeheme degradation by reactive oxygen species part of Chl holds an ethyl instead of avinyl group. To (ROS) to PDPs and BOXes. result in BOX A, the initial macrocycle opening has to

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6206  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

Scheme3.Overview of heme and chlorophyll degradation products, which are known from heme and chlorophyll (Chl) or introduced in this study.Hin red circle marks the products found in heme degradation and Cingreen circlemarks the products that we found in leavesand/or spit/fecesofherbivores (ROS =reactive oxygen species, PDPs =propentdyopents, MM=methylmaleimide, MEM =methyl ethyl maleimide, MVM=methylvinyl maleimide, HA=he- matinic acid, BOX =bilirubinoxidation end product). For simplification,the four known PDPs (Scheme 2) were considered collectively and only one molecule is shown. Accordingly,the new proposed but not fully characterizedPDPs representthe different isomers regarding vinyl/methyland OH-group position.The proposed structures are in accordancewith considerations on the possible degradation mechanismsinanalogy to heme degradation, high resolution MS data (Table S1) and retention time in RP-HPLC of PDPs (Figure S13). DihydroPDPs might arise from incorporation of the ethyl B-ring as shownoralternatively by the saturated d-ring of Chl. Sincethe macrocycle opening of Chl results in an group(Scheme 1) instead of acarbonyl group in heme degrada- tion (Scheme 2), we furtherproposePDPs with an aldehyde or acid group. As known from chlorophyllin, the additional five-ring EofChl can be oxidized and cleaved resulting in aformic acid group. This would lead to the formyl PDPs.For three of the proposed PDPspecies (dihydro,aldehyde and acid PDPs) the highest amounts were found inthe leavesand not in spit or feces (datanot shown).

occur at adifferent pyrrole bridge and not between rings A Following the published synthesis of BOX A, dihydro BOX A and Basinthe enzymatic ring opening or the oxidative path- was synthesized and fully characterized.[40,41] Startingfrom way known for heme degradation. methyl-(Z)-2-(3-bromo-4-methyl-5-oxo-1,5-di-hydro-2H-pyrrol-2- We tentatively identifieddihydro BOX Aand Bbased on ylidene)ethanoate(1)the synthesis of dihydro BOX Asucceeds their mass spectra and chromatographic retention times (Sup- through four synthetic steps with atotal yield of 64% porting information, Figures S10). We also detected products (Scheme 4). The introduction of an was performed that were in accordance with dipyrrolic structures, similar to with aSuzuki–Miyaura cross-couplingreaction and resulted in PDPs (Supporting information, Figure S13, Table S1). high yield utilizing apalladium complex with sterically de- manding1,1’-bis(diphenyl-phosphino)ferrocene (dppf) as cata- lyst.[43] Further conversion of coupling product 2 by cleavage Confirmation of proposed structures of the methyl ester,transformationtothe correspondingacyl To confirmthe structure of the proposed molecules, the degra- chloride 4 and subsequent formation of the amide functionali- dation products needed to be isolated in higher amounts. ty by reactionwith ammonia provided analytically pure dihy- Given the low yield, the degradation of chlorophyll a was not dro Z-BOXA. feasible and chlorophyllin wasused instead. Due to the better Structure and Z-configuration were confirmed by single-crys- water solubility of chlorophyllin, the oxidation was carriedout tal X-ray diffraction analysis. Structural parameters of the mo- in water without sodium hydroxide. Since chlorophyllin is a lecular structureshown in Figure 1are similar to those pub- mixture of different chlorins and porphyrins,[38,42] an even lished for Z-BOX Awith the exception of bond length and higher variety in productswas observed, which overall only led angle of the ethyl group. With abond length of 152.2(4) pm to asmall increase in the amount of degradation products. the C8 C9 bond exhibits atypical single bond length and the À After oxidation, achloroform extraction was performed to C3-C8-C9 bond angle of 112.4(2)8 is slightly compressed in obtain aBOXes-rich extract. The following preparative HPLC- comparison to the bond angle of 124.7(2)8 report- UV yielded fractionsofthe proposedmolecules, dihydro BOX A ed for Z-BOX A(see also Supporting Information). Similar to and B. With NMR and MS-MS experiments, the general BOX other BOXes, aggregation by formation of intermolecular N À structure wasproven, but due to overlaying signals of impuri- H···O hydrogen bridges is shown in the crystalline state, under- ties, asynthetic approachwas used to confirm the complete lining the ability of those compounds for strong non-covalent structure and configuration. interactions. With the addition of the syntheticstandard to the

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6207  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

(Scheme3).[31,34] These metabolites were quantified in amounts 1 in the pg mgÀ dry mass range (Figure 2). In comparison, the Chl content in the sweet potatoleaves was reported in low 1 [44] mgmgÀ dry mass. Similar to the observations made during the Chl a in vitro degradation, once the macrocycle is opened, the reactivity for further oxidative transformationsishigher, potentially resulting in faster furtherreactions. Additionally, low molecular weight degradation products might be taken up or metabolized by the organism. We alsodetected LC-MS- evidencefor unidentified dihydro BOXesaswell as novel PDPs that wereinitially found in the in vitro degradation.Intotal, mass spectra HRMS and retentiontimes pointed towards five new PDPs, and two new BOXes(Supporting Information Fig- ure S13, TableS1). 1 PDPs occurred overall in 0.1 to 1pgmgÀ dry mass, whereas BOX Aand Bcould only be observed in traces. This is in ac- cordancewith the overall lower concentration of those prod- ucts in heme degradation.There, PDPs as intermediates usually occurred in 20–80 times higher amounts than BOXes.[31,36] In Scheme4.Synthesis of dihydro Z-BOX Athrough afour-stepprocedure(THF contrast, BOX Coccurred in higher amounts compared to PDPs = tetrahydrofurane, DCM = dichloromethane, DMF = N,N-dimethylform- in the leaves. This discrepancy points towards the existence of amide,dppf = 1,1’-bis(diphenylphosphino)ferrocene). so far unknown PDP speciesasdegradation intermediates. Peaks with matching masses of the five proposed PDPs that could act as such intermediates(Supporting Figure S13) were found in the investigated samples.Due to the lack of stand- ards, only the peak areas obtained from LC-MS can be com- pared, but the overall amounts of the proposed molecules can 1 be estimated in the ng to high pg mgÀ dry mass range. In principle, it cannot be excluded that plant-derived heme or bilins contribute to the observed break down products as well. Since the relative amountsofthe detected products are similar to those observed in the in vitro Chl oxidationexperiments we conclude that heme can, if at all, only contribute to aminor extendtothe detected catabolic products. Further,the amount of clearly Chl derived products (green dots in Scheme 3) compared to those that could be derived from Chl as well as heme (green and red dots in Scheme 3) are similar, pointingalso to asubstantial or exclusive Chl-origin. Figure 1. Molecular structure and numbering schemeofdihydro BOX A. The ellipsoidsrepresent aprobabilityof30%,Hatoms are shownwith arbitrary According to literature, maleimides and hematinic acid were radii. found in senescent and radish leaves,aswell as lake sediments, which are considered first hints on low molecular weightoxidative degradation products.[44–47] In our study,HA in vitro chlorophyll a degradation the structure of the degrada- was found in similar quantities as PDPs whereas maleimides tion product could be proven (Supporting Information, Fig- could not be detected. ure S10). This finding also supportsthe potential existence of dihydro BOX Band dihydro PDPs. Induction The potentialinductionofChl degradation products in re- Products of oxidative chlorophyll degradationinleaves sponse to herbivores and pathogens was investigated in detail To search for oxidative degradation products of Chl in vivo, in sweet potato leaves. To mimic natural stress conditions, sweet potato leaves (Ipomoea batatas)were extracted and ana- leaves were either infested with the generalist herbivore Spo- lyzed by UHPLC-MS. The analysisrevealed the occurrence of doptera littoralis or infected with the pathogenic fungus Alter- Chl degradation products with identicalchromatographicand naria brassicicola. After 30 min of herbivore feeding or 48 hof mass spectrometric properties compared to those described fungal inoculation, the leaves were harvested,dried and ex- above.Severalstructures are shared with heme degradation tracted. We observed no significant up-regulation of Chl catab- products.Weunambiguously identified PDPs B1, and B2 as olites compared to the untreated leaves (Figure 2). Previous well as BOX Cbycomparison with authentic standards studies reported decreasedChl concentration after herbivore

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6208  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

Figure 2. Comparison of different proposed chlorophyll degradation products in leaves (green), spit (yellow) and feces (brown). Leaves without induction are labelledwith „ “, those induced with S. littorals or A. brassicicola are labelled with „+“. Feces were analyzed directlyafter collection (0 days) andafter seven À days in dry (7d) and wet conditions(7w). Note the different scales of the y-axis in D. One-way-Anova with Bonferroni post-hoc test was carried out for statisti- cal analysis. Letters indicatestatistical difference between the indicated groups aand b(*p=0.05, ** P= 0.05, *** P =0.005).

attack andinfection with A. brassicicola.[13] Our study does, in the feces after 7days (Figure 2D). In accordance, also con- however, not show an increased concentration of the defined centrationsofBOX Aand Bwere low in the spit samples.Dihy- oxidative degradationproducts indicating alternative degrada- dro BOX Bwas not observed at all, which fits to the overall tion pathwaysbeing induced. proposed pathway by known PDPs from whichdihydro BOX B cannotbegenerated. Spit To verify if the identified chlorophyll degradation products are Feces found in the spit of insectsand if their formationiscontinued during digestion, we investigated awell-established plant/her- Feces samples of Spodopteralittoralis were analyzed to moni- bivore combination. Spit and feces of the tobaccohorn worm tor the production of Chl catabolites during and after plant di- Spodopteralittoralis that was feeding on sweet potatoleaves gestion in the insect.Since it is known from literature, that a were extracted and investigated with UHPLC-HRMS. The con- gut protein directlyscavenges an early metabolite of tent of Chl degradationproductswas then compared to that Chl,[23,24,48] the hypothesis was, that asubstantial degradation is in the intact leaves. occurring by microbes after excretion. Therefore, samples were Herbivore spit contained significantly higher amounts of not only analyzed directly after excretion, but also incubated PDPs and BOX Ccompared to leaves (Figure 2). This fits to the for one week at dry and wet conditions. In the initial samples, overall idea of astepwise degradation, since those products the overall amountswererather low compared to leaves and can only occur after at least one breakage of afive-membered spit. After one week, trends of increased amounts can be ob- ring. This could occur already in the leaves or in the spit of served(Figure 2). Nevertheless, since the differences are not herbivores. Asmall amount of dihydro BOX Awas also found significant,furtherstudies need to be carried out to investigate in the spit. This compound was not detected in leaves and just this into more detail.

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6209  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

To study whether those findings are specific for S. littoralis, ed as strong (s), medium (m), weak (w) or broad (br). Mass spectra the feces of two other herbivores (Helicoverpa armigera and were recorded on aFinnigan MATSSQ 710 or ThermoFinnigan Chelymorphaalternans)were also extractedand analyzed. MAT95XL. Elemental analyses were performed using aLeco Since both species were also reared on sweet potatoes, the CHNS-932 Elemental Analyzer. amountsofdegradation products found can be compared. For quantification and reaction monitoring aDionex UltiMate 3000 The feces of H. armigera showed no significant differences to UHPLC (Thermo Fisher Scientific, Leicestershire, UK) equipped with an Acquity UPLC BEH C18 column (1.7 mm, 100”2.1 mm) was S. littoralis whereas C. alterans had overall higheramountsof used. The UHPLC was coupled to aQ-Exactive Plus Orbitrap mass degradation products indicating acertain species specificity spectrometer (Thermo Fisher Scientific, Leicestershire, UK) and ioni- (Figure 2). zation was carried out with electrospray in positive ion mode. Sol- vent Acontained 2% acetonitrile in water with 0.1%formic acid and solvent B100%acetonitrile with 0.1%formic acid. The follow- Conclusions ing gradient (time, vol% solvent B) was used:0.0 min, 0%;0.5, 0%; 1.0 min, 18%; 8.0 min, 18 %; 9.0 min, 100%; 10.9 min, 100%; We present so far unknown chlorophyll degradation products 1 11.0 min, 0%;13.0 min, 0% with aflow rate of 0.4 mLminÀ .Exter- in plants and herbivores.Mono-and dipyrrolic compounds, nal calibration using standards was performed for quantification. known from heme degradation,were for the first time con- For preparative separation aHPLC (Shimadzu LC-8A, Kyoto, Jap) nected to chlorophyll degradation andquantified in biological with aHTEC C18-column (5 mm, 250x 16 mm, Macherey–Nagel, 1 samples. They occur in the range of pgmgÀ dry weight with a Dueren, Ger) equipped with aSPD-10AVUV–Visdetector measur- species-specific distribution that suggestsenzymatic involve- ing at 220 nm was used. Solvent Acontained 2% acetonitrile in ment in their formation. Aseries of novel molecules was pro- water and solvent B100%acetonitrile. The gradients (time, vol% posed as extension to the existing degradation products and solvent B) were as follows:0min, 16%; 50 min, 16 %, 51 min, 100%, 56 min, 100%; 58 min, 16%; 66 min, 16 %with aflow rate tentatively found in the investigatedsamples. Thestructure of 1 of 6mLminÀ . dihydro BOX A, asofar unreported natural product, was provenbysynthesis. Overall, this study opens anew field of re- Oxidative degradation:The oxidative degradation of chlorophylla was carried out based on published protocols for bilirubin.[32,39] To search on the degradation of chlorophyll to smaller molecules. optimize the process, small test reactions of chlorophylla were car-

ried out in 5m NaOH/H2O, in EtOH and in 2m NaOH/MeOH with

different amounts of H2O2 (1–10%) and monitored by UHPLC- Experimental Section HRMS. In the final procedure, chlorophylla (3.45 mg, 3.86 mmol) was suspended in 2.5 mL of 2m NaOH/MeOH and stirred for 24 h. Materials and methods The pH was adjusted to 7.5 with HCl before H2O2 was added to a Chemicals:All solvents and compounds were purchased and used final concentration of 10%and stirred for 48 h. After extraction without further purification:Chlorophylla (purified from Anacystis with chloroform to isolate aBOXes rich fraction, the water phase nidulans)and chlorophyllin (commercial grade) from Sigma–Aldrich was subjected to solid phase extraction (30 mg Oasis hydrophilic (Munich, Ger), NaOH from Carl Roth (Karlsruhe, Ger), 30%HCl from lipophilic balanced cartridges;Waters, Manchester,United King- Merck (Darmstadt, Ger) and 50%H2O2 from VWR (Darmstadt, Ger). dom). After elution with water and 20%ACN/water,the fractions HPLC gradient grade acetonitrile was obtained from VWR and were dried under an N2-flow and analyzed with UHPLC-HRMS. water was purified with TKA microPure (Thermo Electron. Niederel- Standards (PDPs, BOXes) were added to the extract to proof the bert, Ger). ULC gradient grade water and acetonitrile was pur- occurrence of the compounds. chased from Fisher Scientific (UK) and ULC formic acid was ob- For preparative scale preparation of degradation products, the tained from Biosolve B.V.(Valkenswaard, NL). For synthesis of dihy- degradation of chlorophyllin (1 g, 1.38 mmol) was carried out in dro BOX Aall chemicals were used as purchased. Methyl (Z)-2-(3- water (0.7 L) at pH 7.5 with an overall concentration of 1% H O bromo-4-methyl-5-oxo-1,5-dihydro-2H-pyrrol-2-lidene) ethanoate 2 2 for 1d.Analogous to chlorophyll, achloroform extraction followed (1) was prepared according to aliterature procedure.[40] Solvents by asolid phase extraction (6 gcartridge) were carried out. The were dried by standard methods if necessary.Tetrahydrofuran fractions were dried with an evacuated centrifuge and preparative (THF) and toluene were distilled under nitrogen from sodium/ben- HPLC-UV–Vis was carried out to isolate the product. zophenone. Dichloromethane (DCM) was distilled under nitrogen from calcium hydride. N,N-Dimethylformamide (DMF) was dried Purification of standards:PDPs, BOX Cand hematinic acid were pu- over calcium hydride and after filtration distilled under nitrogen. rified from bilirubin oxidation according to published proto- [31,34] [40,41] Deionized water was degassed by nitrogen sparging at reflux tem- cols. BOX Aand Bwere synthesized as published. perature. Merck silica gel 60 (0.040–0.063 mm) was used for Synthesis and analytics of dihydro BOX A column chromatography. Methyl (Z)-2-(3-ethyl-4-methyl-5-oxo-1,5-dihydro-2H-pyrrol-2-ylidene) Instrumentation:Toremove solvents and to dry samples, an evac- ethanoate (2): Asolution of methyl-(Z)-2-(3-bromo-4-methyl-5-oxo- uated centrifuge (Christ SpeedVac RVC2–25) at 40 8Cand afreeze 1,5-di-hydro-2H-pyrrol-2-ylidene)ethanoate (1)(500 mg, dryer (Christ Alpha 1–2 LD) were used. 2.03 mmol), potassium ethyltrifluoroborate (332 mg, 2.44 mmol), NMR data (1H; 13C) was collected on a400 MHz Bruker AvanceIor [1,1’-Bis(diphenylphosphino)-ferrocene]palladium(II) dichloride

AvanceIII spectrometer using the residual solvent resonance of complex with dichloromethane (83 mg, 0.10 mmol) and K2CO3 1 13 1 the solvents [D1]CDCl3 ( H d=7.26; C d=77.16), [D6]DMSO ( H (842 mg, 6.10 mmol) in degassed toluene/water (25 mL, 4:1v/v) 13 1 13 d=2.50; C d=39.52) or [D8]THF ( H d=1.72, 3.57; C d=25.31, was heated to 858Cfor 20 hinanitrogen atmosphere. After cool- 67.39) as internal standard for referencing. Chemical shifts (d)are ing to room temperature, ethyl acetate (50 mL) was added and the reported in parts per million (ppm). ABruker ALPHA Platinum-ATR solution was washed with water (2”20 mL) and brine (20 mL). The spectrometer was used to record IR spectra, intensities are report- combined aqueous layers were extracted once with ethyl acetate

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6210  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

(20 mL), the organic extracts then combined and dried over 163 (46), 148 (12), 137 (49), 122 (23), 108 (10); Elemental analysis

Na2SO4.The solvent was removed and the crude product purified (%): calcd for C9H12N2O2 (180.21): C59.99, H6.71, N15.55;found:C by column chromatography (silica gel, ethyl acetate/n-heptane, 1:1 60.01, H6.83, N15.88. v/v) to provide 2 as apale yellow solid (374 mg, 94%). CCDC 1945467for dihydro BOX Acontain the supplementary crys- 1 HNMR (400 MHz, CDCl3,297 K): d=8.99 (s, 1H,NH), 5.37 (s, 1H, tallographic data for this paper.These data are provided free of

CH), 3.77 (s, 3H,COOCH3), 2.41 (q, J=7.6 Hz, 2H,CH2CH3), 1.92 (s, charge by The Cambridge Crystallographic Data Centre. 13 1 3H,CH3), 1.13 (t, J=7.6 Hz, 3H,CH2CH3)ppm; C{ H} NMR

(101 MHz, CDCl3,297 K): d=171.9, 167.9, 150.6, 146.3, 131.9, 93.1, 51.8, 17.8, 14.1, 8.6 ppm; IR (ATR): n=3251 (m;N H), 1728 (m;C= Collection of biological samples 1 À O), 1680 (s;C=O, Amide I), 1637 (s;C=C) cmÀ ; MS (DEI): m/z (%)= Plant and insect material: Ipomoea batatas Lam. cultivar Tainong 195 (100) [M]+,163 (84), 148 (20), 135 (23), 108 (18); Elemental 57 was grown for three weeks under long day conditions (16 h analysis (%): calcd for C10H13NO3 (195.22): C61.53, H6.71, N7.18; 2 1 light with 100 mmolmÀ sÀ ,8hdark) at 28 C(day) and 25 C found:C61.47, H6.71, N7.11. 8 8 (night) at 70%relative humidity. (Z)-2-(3-ethyl-4-methyl-5-oxo-1,5-dihydro-2H-pyrrol-2-ylidene)ethanoic Spodoptera littoralis (Boisd.,Lepidoptera, Noctuidae) larvae were acid (3): To asolution of 2 (300 mg, 1.54 mmol) in THF (6 mL) was provided by Bayer Cropscience Germany and reared on an artificial added an aqueous solution of LiOH (2m,3equiv) at 0 C. After stir- 8 diet.[49,50] The larvae were separated and starved 24 hprior to each ring at room temperature for 27 h, diluted hydrochloric acid was feeding assay. added at 08Ctill apHvalue of 2was reached. Ethyl acetate Feces and spit collection:Feces were collected from five different (20 mL) was added, the organic layer separated, and the aqueous th layer extracted with ethyl acetate (3”10 mL). The combined organ- sets of Spodoptera littoralis larvae. Each set comprised ten 4 instar larvae which were feeding on two Ipomea batatas plants (3 weeks- ic layers were washed with brine, dried over Na2SO4 and the sol- vent removed in vacuo to provide 3 as acolorless solid (273 mg, old). After five hours of feeding, the larvae were removed and the 98%). feces collected after different incubation times (0, 1, 3, 5, and 1 7days). For the first measured time-point (0), the feces were col- HNMR (400 MHz, [D6]DMSO, 297 K): d=12.57 (s, 1H,COOH), 9.50 lected immediately after removal of the larvae. The remaining (s, 1H,NH), 5.42 (s, 1H,CH), 2.43 (q, J=7.6 Hz, 2H,CH2CH3), 1.83 (s, 13 1 feces were subsequently placed in two different conditions:low 3H,CH3), 1.05 (t, J=7.6 Hz, 3H,CH2CH3) ppm; C{ H} NMR (dry) and high humidity (wet). High humidity (see above in plant (101 MHz, [D6]DMSO, 297 K): d=171.3, 167.8, 149.3, 146.5, 130.5, and insect material) mimics the natural conditions of the herbivory 94.4, 16.9, 14.0, 8.1 ppm; IR (ATR): n=3409 (m;O H), 3287 (m;N À À plant system. Low humidity incubation (258C, 30%humidity,18 H), 3173–2187 (m, br;OH···O), 1725 (m;C=O), 1660 (s;C=O, À 1 + hours light) presents the laboratory condition of the insect room. Amide I), 1633 (s;C=C) cmÀ ; MS (DEI): m/z (%)=181 (54) [M] , Another set of feces was collected from Helicoverpa armigera and 164 (13), 137 (100), 122 (30), 108 (18); Elemental analysis (%): Chelymorpha alternans reared on I. batatas plants. Spit was collect- calcd for C9H11NO3 (181.19): C59.66, H6.12, N7.73;found:C59.68, ed from 5th instar S. littoralis larvae reared on I. batatas plants. The H6.13, N7.79. larvae were gently squeezed on the second segment of the dorsal (Z)-2-(3-ethyl-4-methyl-5-oxo-1,5-dihydro-2H-pyrrol-2-ylidene)ethanoyl part to regurgitate the spit. The spit was pooled from 100 larvae. chloride (4): To asolution of 3 (250 mg, 1.38 mmol) in anhydrous The feces and spit were kept in 20 8Cuntil sample preparation. À DCM (8.5 mL) were added oxalyl chloride (227 mg, 1.79 mmol) and Fungal infected leaves collection: Alternaria brassicicola was anhydrous DMF (one drop). The mixture was stirred for 2h at grown for two weeks at 228Conpotato dextrose agar medium as room temperature in anitrogen atmosphere. The solvent was re- described in literature.[51] Subsequently,spore suspensions were moved by vacuum distillation and the residue dried in vacuo, pro- freshly prepared by adding 5mLofsterile H2Otothe fungal myce- viding 4 as apale brown solid, which was used for the next step lium and gently scraping off the spores and hyphae with afollow- without further purification (271 mg, 98%). ing filtration step through asterile nylon membrane (75 mmpore 1 HNMR (400 MHz, [D8]THF,297 K): d=9.79 (s, 1H,NH), 5.72 (s, 1H, size). The spore density was determined using ahemocytometer 6 1 CH), 2.48 (q, J=7.6 Hz, 2H,CH2CH3), 1.89 (s, 3H,CH3), 1.12 (t, J= and adjusted to 2”10 sporesmLÀ with sterile H2Ocontaining 13 1 7.6 Hz, 3H,CH2CH3)ppm; C{ H} NMR (101 MHz, [D8]THF,297 K): 0.01%Tween-20. A. brassicicola was kindly provided by the Jena d=172.6, 164.5, 154.6, 147.5, 133.9, 97.6, 18.0, 14.1, 8.4 ppm; MS Microbial Resource Center. + (DEI): m/z (%)=199 (11) [M] ,164 (100), 136 (19), 108 (10); HRMS For the co-cultivation with A. brassicicola,single leaves (n=10–12) + (EI): m/z calcd for C9H10ClNO2 :199.0400;found:199.0402 [M] . of 3week-old I. batatas plants were cut and washed in tap water (Z)-2-(3-ethyl-4-methyl-5-oxo-1,5-dihydro-2H-pyrrol-2-ylidene)ethan- to remove the milky latex juice exuding from the cut stem. The de- amide (dihydro Z-BOX A):Asolution of 4 (257 mg, 1.29 mmol) in an- tached leaves were each placed on round filter paper soaked in hydrous THF (25 mL) was cooled to 08C. Gaseous ammonia was 1mLsterile H2Oinasquare petri dish and infected with 5 mLofA. 6 1 bubbled through the solution for 0.5 hat08Cand was maintained brassicicola spore suspension (2”10 sporesmLÀ )orsterile water for additional 0.5 hwhile warming up to room temperature. The containing 0.01%Tween-20 as control. The plates were then solvent was removed in vacuo and the residue washed with water sealed with parafilm and placed in humid conditions (see above) (3”5mL). Purification by recrystallization from methanol provided for 48 h. pale yellow crystals of dihydro Z-BOX A(164 mg, 71%). Leaf collection upon herbivory:Five Spodoptera littoralis larvae 1 (4th instar) were placed on 3leaves of asingle sweet potato plant HNMR (400 MHz, [D6]DMSO, 297 K): d=9.66 (s, 1H,NH), 7.64 (s, (3 weeks-old) and allowed to feed for 30 minutes. Afterwards the 1H,NH2), 7.22 (s, 1H,NH2), 5.58 (s, 1H,CH), 2.39 (q, J=7.6 Hz, 2H, treated leaves from 5plants in total (n=15) were harvested and CH2CH3), 1.82 (s, 3H,CH3), 1.07 (t, J=7.6 Hz, 3H,CH2CH3)ppm; 13 1 immediately frozen in liquid nitrogen until further extraction. C{ H} NMR (101 MHz, [D6]DMSO, 297 K): d=170.9, 168.1, 146.3,

145.9, 129.8, 96.9, 17.1, 14.0, 8.1 ppm; IR (ATR): n=3346 (m;NH2), Sample preparation:All the obtained samples were dried for 2d

3153 (m;NH2), 1696 (m;C=O, Amide I), 1664 (s;C=O, Amide I), with afreeze dryer and ground in liquid N2.Three times around 1 + 1610 (s;NH2,Amide II) cmÀ ; MS (DEI): m/z (%)=180 (100) [M] , 20 mg were extracted with an adapted Bligh and Dyer method:

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6211  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

After the addition of 0.2 mL CHCl3 and 0.4 mL MeOH the mixture [13] A. K. Mukherjee, M. J. Carp, R. Zuchman, T. Ziv,B.A.Horwitz, S. Gep- was shaken for 5min, which was repeated after the addition of stein, J. Proteomics 2010, 73,709–720. [14] T. Suzuki,Y.Shioi, Photosynth. Res. 2002, 74,217 –223. 0.2 mL CHCl3 and after 0.2 mL H2O. After centrifuging, both super- natants were combined and dried with an evacuated centrifuge. [15] S. Hçrtensteiner,F.Vicentini, P. Matile, New Phytol. 1995, 129,237 – 246. After adding 70 mLH2O, shaking and centrifuging, the supernatants [16] M. Oberhuber,J.Berghold, K. Breuker, S. Hortensteiner,B.Krautler, Proc. were measured twice with UHPLC-MS. To evaluate the influence of Natl. Acad. Sci. USA 2003, 100,6910–6915. the drying process, spit was extracted without drying and com- [17] B. Kräutler,B.Jaun, P. Matile, K. Bortlik, M. Schellenberg, Angew. Chem. pared to extracts of dried samples. By comparing both results with Int. Ed. Engl. 1991, 30,1315 –1318; Angew. Chem. 1991, 103,1354 – normalization to the dry mass, amean value of 34%ofthe degra- 1357. dation products were observed after drying, compared to direct [18] H. Vergara-Domínguez, J. J. Ríos,B.Gandul-Rojas, M. Roca, Food Chem. measurements of the liquid samples. This indicates that some com- 2016, 212,604 –611. pounds either degrade further or cannot be eluted again after [19] B. Christ, I. Süssenbacher,S.Moser,N.Bichsel, A. Egert, T. Müller,B.Kräu- drying. tler,S.Hçrtensteiner, Plant Cell 2013, 25,1868 –1880. For external calibration, standards of PDP A1/A2/B1/B2, BOX A/B/C, [20] C. Vergeiner, S. Banala,B.Kräutler, Chem.Eur.J.2013, 19,12294 –12305. dihydro BOX Aand HA were measured three times at concentra- [21] S. Moser,T.Müller,A.Holzinger,C.Lütz, B. Kräutler, Chem. Eur.J.2012, 18,10873 –10885. tions between 0.01 mm and 5 mm. [22] J. Berghold, T. Müller,M.Ulrich, S. Hçrtensteiner,B.Kräutler, Monatsh. Data analysis:The mean peak area of two measurements was nor- Chem. 2006, 137,751–763. malized on the weighted dry mass and mean values and standard [23] A. Badgaa, A. Jia, K. Ploss, W. Boland, J. Chem. Ecol. 2014, 40,1232 – deviation was calculated from the three biological replicates. For 1240. statistical analysis, one-way-Anova with Bonferroni post-hoc test [24] A. Badgaa, R. Büchler,N.Wielsch,M.Walde, R. Heintzmann, Y. Pauchet, was used. A. Svatos, K. Ploss, W. Boland, J. Chem. Ecol. 2015, 41,965 –974. [25] H. Chen,E.Gonzales-Vigil, C. G. Wilkerson, G. A. Howe, Plant Physiol. 2007, 143,1954 –1967. [26] X. Tang, D. Freitak, H. Vogel,L.Ping, Y. Shao, E. A. Cordero, G. Andersen, Acknowledgements M. Westermann, D. G. Heckel, W. Boland, PLoS One 2012, 7,1–9. [27] G. A. Hendry,O.T.Jones, J. Med. Genet. 1980, 17,1–14. Financialsupport by the German Research Foundation (DFG) [28] C. Li, R. Stocker, RedoxRep. 2009, 14,95–101. within the framework of the ResearchGroup FOR1738 is ac- [29] D. E. Baranano, M. Rao, C. D. Ferris, S. H. Snyder, Proc. Natl. Acad.Sci. knowledged. D. S. is grateful to the Graduate Academy of the USA 2002, 99,16093–16098. [30] S. Erlinger,I.M.Arias, D. Dhumeaux, Gastroenterology 2014, 146,1625 – Friedrich Schiller University Jena for financial support. Finn 1638. Gude is acknowledged for experimental support. [31] M. Ritter,R.A.Seidel,P.Bellstedt, B. Schneider,M.Bauer,H.Gçrls, G. Pohnert, Org. Lett. 2016, 18,4432 –4435. [32] K. R. Kranc, G. J. Pyne, L. Tao, T. D. Claridge, D. A. Harris, T. A. Cadoux- Conflict of interest Hudson, J. J. Turnbull, C. J. Schofield,J.F.Clark, Eur.J.Biochem. 2000, 267,7094–7101. [33] W. L. Wurster,G.J.Pyne-Geithman, I. R. Peat, J. F. Clark, ActaNeuroch. The authors declare no conflict of interest. Suppl. 2008, 104,43–50. [34] M. Ritter,S.Neupane, R. A. Seidel,C.Steinbeck,G.Pohnert, Org. Biomol. Chem. 2018, 16,3553 –3555. Keywords: bilirubin oxidation end products · chlorophyll [35] R. A. Seidel, T. Claudel, F. A. Schleser,N.K.Ojha, M. Westerhausen, S. degradation · natural products · porphyrinoids · Nietzsche, C. Sponholz,F.Cuperus, S. M. Coldewey,S.H.Heinemann, G. propentdyopents Pohnert, M. Trauner,M.Bauer, J. Hepatol. 2017, 67,272 –281. [36] A. Joerk, M. Ritter,N.Langguth, R. A. Seidel, D. Freitag, K.-H.Herrmann, A. Schaefgen,M.Ritter,M.Günther,C.Sommer, D. Braemer,J.Walter,C. [1] G. A. F. Hendry,J.D.Houghton,S.B.Brown, New Phytol. 1987, 107, Ewald, R. Kalff, J. R. Reichenbach, M. Westerhausen, G. Pohnert, O. W. 255–302. Witte, K. Holthoff, Circ. Res. 2019, 124,101 –114. [2] and :Biochemistry Biophysics, Functions [37] A. Joerk, R. A. Seidel, S. G. Walter,A.Wiegand,M.Kahnes,M.Klopfleisch, and Applications (Eds.:B.Grimm, R. J. Porra, W. Rüdiger,H.Scheer), K. Kirmse, G. Pohnert, M. Westerhausen,O.W.Witte, K. Holthoff, J. Am. Springer Netherlands, 2006. Heart Assoc. 2014, 3,1–13. [3] I. Fleming, Nature 1967, 216,151–152. [38] A. Mortensen,A.Geppel, Innovative Food Sci. Emerging Technol. 2007, 8, [4] R. B. Woodward,W.A.Ayer,J.M.Beaton, F. Bickelhaupt, R. Bonnett, P. 419–425. Buchschacher,G.L.Closs,H.Dutler,J.Hannah, F. P. Hauck,S.Itoˇ ,A.Lan- [39] R. A. Seidel,M.Kahnes, M. Bauer,G.Pohnert, J. Chromatogr.BJ.Chro- gemann, E. Le Goff, W. Leimgruber,W.Lwowski, J. Sauer,Z.Valenta, H. matogr.B2015, 974,83–89. Volz, Tetrahedron 1990, 46,7599 –7659. [40] M. Klopfleisch,R.A.Seidel,H.Gçrls, H. Richter, R. Beckert, W. Imhof, M. [5] A. S. Brandis,Y.Salomon, A. Scherz, Chlorophylls Bacteriochlorophylls Bio- Reiher,G.Pohnert, M. Westerhausen, Org. Lett. 2013, 15,4608 –4611. chem. Biophys. Funct. Appl. (Eds.:B.Grimm, R. J. Porra, W. Rüdiger,H. [41] R. A. Seidel, B. Schowtka, M. Klopfleisch,T.Kühl, A. Weiland,A.Koch, H. Scheer), Springer Netherlands 2006,pp. 461–483. [6] A. Tanaka,R.Tanaka, Curr. Opin. Plant Biol. 2006, 9,248–255. Gçrls, D. Imhof,G.Pohnert, M. Westerhausen, TetrahedronLett. 2014, [7] U. Eckhardt, B. Grimm, S. Hçrtensteiner, Plant Mol. Biol. 2004, 56, 55,6526 –6529. 1–14. [42] S. Chernomorsky, Mutat. Res. 1994, 324,177–178. [8] B. Kuai, J. Chen,S.Hçrtensteiner, J. Exp. Bot. 2018, 69,751–767. [43] G. A. Molander,M.D.Elia, J. Org. Chem. 2006, 71,9198 –9292. [9] S. Hçrtensteiner,B.Kräutler, Biochim. Biophys.Acta Bioenerg. 2011, 1807, [44] M. Li, G. Y. Jang, S. H. Lee, M. Y. Kim, S. G. Hwang, H. M. Sin, H. S. Kim, J. 977–988. Lee, H. S. Jeong, FoodSci. Biotechnol. 2017, 26,97–103. [10] L. Guyer,S.S.Hofstetter, B. Christ, B. S. Lira, M. Rossi, S. Hçrtensteiner, [45] S. Naeher,P.Schaeffer,P.Adam, C. J. Schubert, Geochim. Cosmochim. Plant Physiol. 2014, 166,44–56. Acta 2013, 119,248 –263. [11] B. Kräutler, Chem. Soc. Rev. 2014, 43,6227 –6238. [46] Y. Suzuki, K. Tanabe, Y. Shioi, J. Chromatogr.A1999, 839,85–91. [12] S. Jung, Plant Physiol.Biochem. 2004, 42,225–231. [47] Y. Suzuki, Y. Shioi, Plant Cell Physiol. 1999, 40,909–915.

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6212  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim Full Paper Chemistry—A European Journal doi.org/10.1002/chem.201905236

[48] B. Mauchamp, C. Royer,A.Garel, A. Jalabert, M. Da Rocha, A. M. Grenier, [51] J. M. Johnson,M.Reichelt, J. Vadassery,J.Gershenzon, R. Oelmüller, V. Labas, J. Vinh, K. Mita,K.Kadono, G. Chavancy, Insect Biochem. Mol. BMC Plant Biol. 2014,162. Biol. 2006, 36,623–633. [49] R. Bergomaz, M. BopprØ, J. Lepid. Soc. 1986, 40,131 –137. Manuscript received:November 19, 2019 [50] M. Heyer,S.S.Scholz, D. Voigt, M. Reichelt, D. Aldon, R. Oelmüller,W. Revised manuscript received: January 21, 2020 Boland,A.Mithçfer, PLoS One 2018, 13,e0197633. Acceptedmanuscript online:January 23, 2020 Version of record online:April 28, 2020

Chem. Eur.J.2020, 26,6205 –6213 www.chemeurj.org 6213  2020 The Authors. Published by Wiley-VCH Verlag GmbH &Co. KGaA, Weinheim