Dual Agents: Fungal Macrocidins and Synthetic Analogues with Herbicidal and Antibiofilm Activities

Eight analogues of the bioherbicides macrocidin A (1) and Z (2) with structural variance in the size of the macrocycle, its para- or meta-cyclophane character, and its functional groups were synthesized on two modular routes and tested for herbicidal, antibiotic, and antibiofilm activities. Apart from the lead compounds 1 and 2, the structurally simplified dihydromacrocidin Z (3) and normacrocidin Z (4) showed high herbicidal activity in either thistles, dandelions or in both. The derivatives 2, 3, and dibromide 9 also inhibited the growth of Staphylococcus aureus biofilms by ca 70% when applied at subtoxic concentrations as low as ca 20 µM, which are unlikely to induce bacterial resistance. They also led to the dispersion of preformed biofilms of S. aureus, exceeding a similar effect by microporenic acid A, a known biofilm inhibitor. Compounds 3 and 9 showed no noticeable cytotoxicity against human cancer and endothelial cells at concentrations below 50 µM, making them conceivable candidates for application as anti-biofilm agents in a medicinal context.


Introduction
Macrocidins are polycyclic tetramic acid macrolactams (PTMs). Macrocidin A (1; Figure 1) was first isolated from the fungus Phoma macrostoma Montagne in 2003 by a Dow AgroSciences group headed by Graupner [1]. It was found to induce chlorosis in broadleaf weeds by a unique mode of action implying an interference with the phytoene synthase and desaturase in the chlorophyll and carotenoid biosynthesis [2]. So far, only two total syntheses of macrocidin A (1) have been published [3,4]. Macrocidin Z (2), which carries an E-alkene in lieu of the expoxide, was isolated from Phoma macrostoma cultures and synthesised in parallel by us only recently [5]. Although concentrated Phoma macrostoma cultures, formulated as broadcast granules, are being used as bioherbicides for environment-friendly weed management in the US and Canada, efforts towards the synthesis of simplified macrocidin derivatives with improved herbicidal properties were sporadically made (e.g., by Graupner's group [6] and Syngenta [7]), albeit without disclosing details. As 3-acyltetramic acids from a broad range of sources were found to have antibiotic or biofilm inhibitory effects [8][9][10][11][12] (e.g., by us in the case of macrocidin Z (2) [5]), we now synthesised analogues of the natural macrocidins with variation of the structural key features such as the size of the macrocycle, its paraor meta-cyclophane character, and its decoration with functional synthesised analogues of the natural macrocidins with variation of the structural key features such as the size of the macrocycle, its para-or meta-cyclophane character, and its decoration with functional groups other than epoxide. We tested them for herbicidal, antibacterial, and antibiofilm activities, and for cytotoxicity to human cells. Figure 1 depicts eight new derivatives 3-10 that were prepared on two efficient modular routes. They share the same tetramic acid derived from L-tyrosine, yet differ in their degree of resemblance to the lead macrocidins (1) and (2). The dihydro derivate 3 of macrocidin Z is the only one retaining the methyl group at an R-configured stereocenter. Like macrocidin Z (2), the derivatives 4 and 5 feature E-alkenes in lieu of the epoxide. Normacrocidin A (6) lacks the methyl group while retaining the epoxide, albeit with a nonnatural R,R configuration. In compound 7, the epoxide is replaced by a vicinal diol, in derivative 8 by a bromohydrin, and in 9 by a vicinal dibromide. Furthest from the natural leads 1 and 2 (in terms of structure) is derivative 10, which has a 13-membered instead of a 17-membered macrocycle comprising a meta-rather than a para-cyclophane. In Scheme 1, two alternative synthetic approaches to key intermediate 16, an N-Bocprotected normacrocidin Z, are illustrated. The known L-tyrosine derived tetramic acid 11 [4] was furnished with a 3-(hepta-2,6-dienoyl) residue by a one-pot reaction first with ketenylidentriphenylphosphorane, Ph3PCCO, to give the corresponding 3-acyl ylide (not shown), followed by a Wittig olefination of the latter with 4-pentenal to leave 3acyltetramic acid 12 in 55% yield [13]. An E-selective ring-closing metathesis reaction with a Grubbs type-II catalyst gave macrocycle 13 in 66% yield as a single stereoisomer. Its reduction with Wilkinson's catalyst and triethylsilane to afford a silyl enol ether (not shown), followed by its cleavage with KF, gave N-Boc-protected normacrocidin Z 16 in a 48% yield. Removal of the N-Boc-protecting group from compounds 16 and 13 using TFA left test candidates 4 and 5 as pure stereoisomers. For the synthesis of larger quantities of 16, another route was developed also starting from tetramic acid 11 in analogy to our syntheses of macrocidins A and Z [4,5] metathesis of 15 with Grubbs type-II catalyst afforded key intermediate 16 in 54% total yield over three steps. Enantiopure dihydromacrocidin Z (3) was obtained from N-Bocprotected macrocidin Z 17 by hydrogenation and removal of the N-Boc-protecting group from 18 in 99% yield over two steps [5]. Key intermediate 16 was also used to introduce further functionalizations, formally replacing the epoxide in macrocidin A (1). Its hydroxylation with AD-mix α afforded, after deprotection, diol 7 as an inseparable 2.3:1 mixture of two diastereomers (Scheme 2 shows major diastereomer). Alkene 16 was also converted to an inseparable mixture of two diastereomeric bromohydrins 8 with NBS and H2O in DMSO [14,15]. A side product (16%) of this reaction, carrying an additional bromo residue next to the enol group, could be separated. Upon treatment with KOtBu, the bromohydrins 8 were converted in 47% yield to a 9:1 mixture of expoxides 6 as indicated by 1 H NMR spectra. The configuration of the major isomer of 6 (shown in Scheme 2) was assigned by comparison with a mixture of isomers 6 obtained on a different route. Hence, we assume that the dr of precursor bromohydrins 8 was also 9:1 as for the epoxides 6. After futile attempts by Ramana et al. [16] and our group at direct epoxidation of alkenes such as 4, 16, or 17, this was the first time the epoxide function could be installed in the preformed macrocycle of a macrocidin precursor. Alkene 16 could also be brominated with bromine in CCl4 to give vicinal dibromide 9 as a mixture of two diastereomers with trans-positioned bromo residues as to NMR spectra. Finally, when heated in diethylaniline in a sealed tube, the para-cyclophane 16 underwent a Claisen rearrangement to afford meta-cyclophane 10 in 56% yield and as a 12.5:1 mixture of diastereomers according to NMR spectra. Like the stereopure derivatives 3-5, the diastereomeric mixtures of derivatives 6-10 were tested for bioactivity, although in the case of diol 7 the two diastereomers, present in similar proportions, might dilute or cancel each other out in terms of biological activities. It should be noted, that most derivatives shown in Schemes 1 and 2 were difficult to purify and analyze. Sometimes, only multiple Key intermediate 16 was also used to introduce further functionalizations, formally replacing the epoxide in macrocidin A (1). Its hydroxylation with AD-mix α afforded, after deprotection, diol 7 as an inseparable 2.3:1 mixture of two diastereomers (Scheme 2 shows major diastereomer). Alkene 16 was also converted to an inseparable mixture of two diastereomeric bromohydrins 8 with NBS and H 2 O in DMSO [14,15]. A side product (16%) of this reaction, carrying an additional bromo residue next to the enol group, could be separated. Upon treatment with KOtBu, the bromohydrins 8 were converted in 47% yield to a 9:1 mixture of expoxides 6 as indicated by 1 H NMR spectra. The configuration of the major isomer of 6 (shown in Scheme 2) was assigned by comparison with a mixture of isomers 6 obtained on a different route. Hence, we assume that the dr of precursor bromohydrins 8 was also 9:1 as for the epoxides 6. After futile attempts by Ramana et al. [16] and our group at direct epoxidation of alkenes such as 4, 16, or 17, this was the first time the epoxide function could be installed in the preformed macrocycle of a macrocidin precursor. Alkene 16 could also be brominated with bromine in CCl 4 to give vicinal dibromide 9 as a mixture of two diastereomers with trans-positioned bromo residues as to NMR spectra. Finally, when heated in diethylaniline in a sealed tube, the para-cyclophane 16 underwent a Claisen rearrangement to afford meta-cyclophane 10 in 56% yield and as a 12.5:1 mixture of diastereomers according to NMR spectra. Like the stereopure derivatives 3-5, the diastereomeric mixtures of derivatives 6-10 were tested for bioactivity, although in the case of diol 7 the two diastereomers, present in similar proportions, might dilute or cancel each other out in terms of biological activities. It should be noted, that most derivatives shown in Schemes 1 and 2 were difficult to purify and analyze. Sometimes, only multiple reversed-phase column chromatography runs led to

Herbicidal Activity
Prior to screening them for antimicrobial effects, the new derivatives were tested for herbicidal activity against thistles and dandelions which had been found to be susceptible to the chlorosis-inducing natural macrocidins [2,6]. For both species, they were applied as max. 150 mM solutions to four pots with two plants each, and their bleaching, withering, and necrotizing effects were assessed after two and then after three to six weeks. None of the compounds reached the efficiency of the synthetic commercial herbicide diflufenican, which was used as a positive control. In line with literature, the lead compound macrocidin A (1) exhibited the highest maximum herbicidal efficiency of all tested compounds, causing 88% mortality of dandelions and 100% of thistles, three weeks after application of a 100 mM solution in a mixture of isopropanol/water = 1:1 + 0.25% Tween 20 ( Figure 2). Interestingly, the epoxide appeared not to be crucial for herbicidal activity, since macrocidin Z (2) still displayed a high efficiency of 88% mortality in thistles and of 50% in dandelions after 42 days at 100 mM. Contrary to an earlier assumption by Graupner, Bailey et al. [6], even derivatives with saturated backbones may show herbicidal efficiency, e.g., dihydromacrocidin Z (3) (38% mortality in thistles and dandelions). The α-methyl group seemed to be important, apparent from the lower figures for normacrocidin Z (4) when compared to 2 (dandelions: 0%, thistles: 63% mortality, 35 days after treatment with 150 mM) and for S,R,R-normacrocidin A (6) which was virtually inactive against both plants. The 13-membered macrocyclic meta-cyclophane 10 exerted a maximum herbicidal efficiency with 38% mortality in thistles, yet only 13% in dandelions after four weeks at 150 mM. Normacrocidin Z (4) and diol 7 also displayed a distinct specificity for thistles over dandelions (for pictures of treated plants cf. Figure S68 in the SI).

Herbicidal Activity
Prior to screening them for antimicrobial effects, the new derivatives were tested for herbicidal activity against thistles and dandelions which had been found to be susceptible to the chlorosis-inducing natural macrocidins [2,6]. For both species, they were applied as max. 150 mM solutions to four pots with two plants each, and their bleaching, withering, and necrotizing effects were assessed after two and then after three to six weeks. None of the compounds reached the efficiency of the synthetic commercial herbicide diflufenican, which was used as a positive control. In line with literature, the lead compound macrocidin A (1) exhibited the highest maximum herbicidal efficiency of all tested compounds, causing 88% mortality of dandelions and 100% of thistles, three weeks after application of a 100 mM solution in a mixture of isopropanol/water = 1:1 + 0.25% Tween 20 ( Figure 2). Interestingly, the epoxide appeared not to be crucial for herbicidal activity, since macrocidin Z (2) still displayed a high efficiency of 88% mortality in thistles and of 50% in dandelions after 42 days at 100 mM. Contrary to an earlier assumption by Graupner, Bailey et al. [6], even derivatives with saturated backbones may show herbicidal efficiency, e.g., dihydromacrocidin Z (3) (38% mortality in thistles and dandelions). The α-methyl group seemed to be important, apparent from the lower figures for normacrocidin Z (4) when compared to 2 (dandelions: 0%, thistles: 63% mortality, 35 days after treatment with 150 mM) and for S,R,R-normacrocidin A (6) which was virtually inactive against both plants. The 13-membered macrocyclic meta-cyclophane 10 exerted a maximum herbicidal efficiency with 38% mortality in thistles, yet only 13% in dandelions after four weeks at 150 mM. Normacrocidin Z (4) and diol 7 also displayed a distinct specificity for thistles over dandelions (for pictures of treated plants cf. Figure S68 in the SI).

Antimicrobial Activity
As 3-acyltetramic acids were frequently shown to have antimicrobial effects [8,9,12,[18][19][20], the new macrocidin derivatives were tested for activity against three different bacteria, namely the Gram-positive strain Staphylococcus aureus (SH1000) and the Gram-negative strains Acinetobacter baumannii, Escherichia coli with the wild-type strain K12 and the ΔTolC mutant (JW5503), which lacks the AcrAB−TolC efflux system. None of the macrocidin analogues displayed activity against the wild-type strain of E. coli. Weak activities against E. coli ΔTolC were found only for dibromide 9 (IC50 = 75 ± 15 µM), dihydromacrocidin Z (3) (IC50 = 82 ± 15 µM), and macrocidin Z (2) (IC50 ca 100 µM). These derivatives were also similarly active against S. aureus (cf. Table S1 for IC50 values and Figures S69-S70 for growth curves in the SI). In comparison, the clinically established antibiotic vancomycin was active with an IC50 of ca 12 µM against S. aureus, and the antibiotic erythromycin with a nanomolar IC50 value. None of the macrocidin derivatives showed a clear antibiotic effect on A. baumannii. Even when applied at the highest concentration of 100 µM, the compounds could not prevent the cultures from reaching an OD600 of at least 60-70% of the maximum value (cf. Figure S71 in the SI). Overall, the toxicity of macrocidin Z (2) and its new synthetic analogues 3-10 against bacteria is weak.

Antibiofilm Activity
The macrocidinoids 2-10 were tested for inhibitory effects on the formation of biofilms by Staphylococcus aureus and Pseudomonas aeruginosa bacteria, as well as for dispersive effects on preformed biofilms of S. aureus and the fungal species Candida albicans (cf. SI for data table S2). While all compounds 2-6 and 8-10 inhibited the formation of S. aureus biofilms by at least 75% relative to untreated controls (=0%) at the highest tested concentration of 250 µg/mL, only macrocidin Z (2), its dihydro derivative 3 and dibromide 9 caused a distinct biofilm inhibition of ca 70% when applied at subtoxic concentrations as low as 7.8 µg/mL (corresponding to 16 µM and 23 µM, respectively). These activities matched or even exceeded that of microporenic acid A (MAA), the known biofilm inhibitor [21] used as a positive control ( Figure 3A). A second group of moderately active inhibitors, comprising normacrocidin Z (4), diene 5 and phenol 10, led to an inhibition of biofilm formation of more than 30% when applied at a concentration of 15.6 µg/mL (corresponding to 43-48 µM). The derivatives 7 and 8 had little inhibitory effect at concentrations below 250 µg/mL.

Antimicrobial Activity
As 3-acyltetramic acids were frequently shown to have antimicrobial effects [8,9,12,[18][19][20], the new macrocidin derivatives were tested for activity against three different bacteria, namely the Gram-positive strain Staphylococcus aureus (SH1000) and the Gram-negative strains Acinetobacter baumannii, Escherichia coli with the wild-type strain K12 and the ∆TolC mutant (JW5503), which lacks the AcrAB−TolC efflux system. None of the macrocidin analogues displayed activity against the wild-type strain of E. coli. Weak activities against E. coli ∆TolC were found only for dibromide 9 (IC 50 = 75 ± 15 µM), dihydromacrocidin Z (3) (IC 50 = 82 ± 15 µM), and macrocidin Z (2) (IC 50 ca 100 µM). These derivatives were also similarly active against S. aureus (cf. Table S1 for IC 50 values and Figures S69-S70 for growth curves in the SI). In comparison, the clinically established antibiotic vancomycin was active with an IC 50 of ca 12 µM against S. aureus, and the antibiotic erythromycin with a nanomolar IC 50 value. None of the macrocidin derivatives showed a clear antibiotic effect on A. baumannii. Even when applied at the highest concentration of 100 µM, the compounds could not prevent the cultures from reaching an OD 600 of at least 60-70% of the maximum value (cf. Figure S71 in the SI). Overall, the toxicity of macrocidin Z (2) and its new synthetic analogues 3-10 against bacteria is weak.

Antibiofilm Activity
The macrocidinoids 2-10 were tested for inhibitory effects on the formation of biofilms by Staphylococcus aureus and Pseudomonas aeruginosa bacteria, as well as for dispersive effects on preformed biofilms of S. aureus and the fungal species Candida albicans (cf. SI for data Table S2). While all compounds 2-6 and 8-10 inhibited the formation of S. aureus biofilms by at least 75% relative to untreated controls (=0%) at the highest tested concentration of 250 µg/mL, only macrocidin Z (2), its dihydro derivative 3 and dibromide 9 caused a distinct biofilm inhibition of ca 70% when applied at subtoxic concentrations as low as 7.8 µg/mL (corresponding to 16 µM and 23 µM, respectively). These activities matched or even exceeded that of microporenic acid A (MAA), the known biofilm inhibitor [21] used as a positive control ( Figure 3A). A second group of moderately active inhibitors, comprising normacrocidin Z (4), diene 5 and phenol 10, led to an inhibition of biofilm formation of more than 30% when applied at a concentration of 15.6 µg/mL (corresponding to 43-48 µM). The derivatives 7 and 8 had little inhibitory effect at concentrations below 250 µg/mL.
The dispersive effects on preformed biofilms of S. aureus were generally slightly less pronounced ( Figure 3B). Derivatives 6-8 were inactive at all concentrations up to 250 µg/mL. The most distinct effects of at least 35% dispersion over the concentration range from 250 µg/mL down to 15.6 µg/mL were again observed for derivatives 2, 3, and 9, which clearly outperformed MAA. The latter was active only at the highest two concentrations, such as the compounds 4, 5, and 10. In the case of C. albicans, compounds 7 and 10 proved inactive, whereas the other compounds, including MAA, showed weak dispersive effects at very high concentrations (250-125 µg/mL) with derivate 5 being the only one active at a lower concentration of 62.5 µg/mL (192 µM) (cf. Figure S72 in the SI). None of the tested derivatives inhibited the formation of P. aeruginosa biofilms.
This study demonstrated that macrocidin analogues may interfere with the formation and persistence of bacterial and fungal biofilms, depending on their structure and polarity. The strongest effects against S. aureus were found for the lipophilic and structurewise "simple" derivatives 2-5 and 9. Interestingly, the activity against biofilms decreased, or even disappeared, when hydroxy groups were introduced into the molecule, as the derivatives 7 and 8 exemplify. The dispersive effects on preformed biofilms of S. aureus were generally slightly less pronounced ( Figure 3B). Derivatives 6-8 were inactive at all concentrations up to 250 µg/mL. The most distinct effects of at least 35% dispersion over the concentration range from 250 µg/mL down to 15.6 µg/mL were again observed for derivatives 2, 3, and 9, which clearly outperformed MAA. The latter was active only at the highest two concentrations, such as the compounds 4, 5, and 10. In the case of C. albicans, compounds 7 and 10 proved inactive, whereas the other compounds, including MAA, showed weak dispersive effects at very high concentrations (250-125 µg/mL) with derivate 5 being the only one active at a lower concentration of 62.5 µg/mL (192 µM) (cf. Figure S72 in the SI). None of the tested derivatives inhibited the formation of P. aeruginosa biofilms.
This study demonstrated that macrocidin analogues may interfere with the formation and persistence of bacterial and fungal biofilms, depending on their structure and polarity. The strongest effects against S. aureus were found for the lipophilic and structurewise "simple" derivatives 2-5 and 9. Interestingly, the activity against biofilms decreased, or even disappeared, when hydroxy groups were introduced into the molecule, as the derivatives 7 and 8 exemplify.

Cytotoxicity
The macrocidinoids 2, 3, and 6-10 were submitted to provisional MTT tests for cytotoxicity/antiproliferative effects against human 518A2 melanoma cells, colon carcinoma cells HCT-116 wt and HCT-116 p53-/-, and KBV cervix carcinoma cells as well as hybrid endothelial EaHy cells. Gratifyingly, none of the compounds but 2 caused signs of toxicity or inhibition of proliferation in the tested cells when applied at concentrations as high as 50 µM (cf table S3 in the SI). This is far outside of any clinically relevant range and would bode well for a potential future application as biofilm interfering agents in a medicinal context. Even macrocidin Z (2) which was antiproliferative with IC50 concentrations of ca.

Cytotoxicity
The macrocidinoids 2, 3, and 6-10 were submitted to provisional MTT tests for cytotoxicity/antiproliferative effects against human 518A2 melanoma cells, colon carcinoma cells HCT-116 wt and HCT-116 p53−/− , and KBV cervix carcinoma cells as well as hybrid endothelial EaHy cells. Gratifyingly, none of the compounds but 2 caused signs of toxicity or inhibition of proliferation in the tested cells when applied at concentrations as high as 50 µM (cf. Table S3 in the SI). This is far outside of any clinically relevant range and would bode well for a potential future application as biofilm interfering agents in a medicinal context. Even macrocidin Z (2) which was antiproliferative with IC 50 concentrations of ca. 15 to 30 µM in cells of colon carcinoma HCT-116 and cervical carcinoma KBV warrants a more in-depth study of its applicability as an anti-biofilm agent.

Discussion
We synthesised eight derivatives of the natural tetramic acids macrocidin A (1) and Z (2) on two efficient modular routes, which allow the introduction of various functionalities and scaffold modifications on a few key intermediates. The double bonds of intermediates 13 and 16 were converted in good yields to bromides, diols, bromohydrins and saturated bonds using standard reactions. A thermal Claisen rearrangement opened an easy access to a 13-membered macrocycle featuring a meta-cyclophane motif. For the first time, we could introduce the epoxide in a macrocidin precursor with a preformed macrocycle. The derivatives and the natural lead compound macrocidin Z (2) were tested for herbicidal and antimicrobial activity, as well as for biofilm interference. The incentive for this extended bioscreening were the frequent reports on the high incidence of antimicrobial and antifungal effects by 3-acyltetramic acids in general. Interestingly, the structurally simple compounds macrocidin Z (2) and dihydromacrocidin Z (3) showed a high herbicidal and antibiofilm activity. Normacrocidin Z (4) was selectively herbicidal against thistles, and dibromide 9 displayed an S. aureus biofilm dispersing effect surpassing even that of the known biofilm inhibitor microporenic acid A. With the exception of diol 7, which was moderately herbicidal against thistles, hydroxy groups on the alkyl backbone of the macrocycle, appear to be generally detrimental to both herbicidal and antibiofilm effects. Contrary to suppositions in the scant literature on macrocidinoids, the epoxide function is obviously not crucial to either activity. The observed distinct and strain-specific effects of the active macrocidinoids 2, 3, and 9 on the biofilms of S. aureus are all the more interesting, as their direct antibacterial activities (i.e., toxicities against bacteria) are rather limited. Thus, their application as biofilm inhibitors would probably not induce bacterial resistance. The at best marginal cytotoxicities of compounds 3 and 9 in human cancer and endothelial cells indicate that these compounds would presumably be well tolerated also by higher organisms. Compound 2 might indeed pose a toxicity problem which should be clarified prior to further tests as a biofilm inhibitor.

General Information
IR spectra were recorded with a PerkinElmer Spectrum 100 FT-IR spectrophotometer (PerkinElmer, Rodgau, Germany) with ATR sampling unit. Optical rotations were measured at 589 nm (Na-D line) on a PerkinElmer 241 polarimeter (PerkinElmer, Rodgau, Germany); [α] D values are given in 10 −1 deg cm 2 g −1 . High resolution mass spectra were obtained with a UPLC/Orbitrap MS system in ESI mode (ThermoFisher Scientific, Bremen, Germany). NMR spectra were recorded with a Bruker Avance III HD 500 spectrometer ( 1 H NMR: 500 MHz and 13 C NMR: 125 MHz) (Bruker, Karlsruhe, Germany). Chemical shifts are given in parts per million, relative to the residual solvent peak as an internal standard, and coupling constants (J) are quoted in Hz. Most tetramic acids were measured in CDCl 3 and also in CD 3 OD where they usually exist as a single (enol) tautomer. Quaternary C-atoms of tetramic acids were sometimes difficult to spot in JMOD or 13 C NMR spectra. For these, more signals cropped up in HMBC and/or HSQC correlation spectra and were considered for peak assignment. In CDCl 3 solution, signals of virtually all C-atoms of tetramic acids were visible yet split up in multiple, difficult to assign sets for individual tautomers both in 1 H and JMOD/ 13 C NMR spectra. In line with literature, we assume the tautomers with exocyclic C-C double bond as drawn for the 3-acyltetramic acids in Figure 1, to be the major tautomer [17]. For the purification of synthetic products, chromatography silica gel 60 (40-63 µm) or silica gel RP18 (40-63 µm) were used. Analytical thin layer chromatography (TLC) was carried out using Merck silica gel 60 F 254 pre-coated aluminum-backed plates. Analytical HPLC was performed on a Shimadzu Nexera XR (Shimadzu GmbH, Duisburg, Germany) using a Knauer Eurospher II C18-column (150 × 4 mm) (Knauer GmbH, Berlin, Germany). Chiral HPLC was performed on a Beckmann System Gold Programmable Solvent Modul 126 using a Phenomenex Lux ® Amylose-1-HPLC-column (100 × 4.6 mm) (Phenomenex Ltd., Aschaffenburg, Germany). All air-and water-sensitive reactions were carried out under a dry argon atmosphere. 1-hydroxyhepta-2,6-dien-1-ylidene)-3,  5-dioxopyrrolidine-1-carboxylate (12). Tetramic acid 11 [4] (1.90 g, 5.50 mmol, 1.10 eq) in dry THF (305 mL) was treated with ketenylidenetriphenylphosphorane (1.66 g, 5.50 mmol, 1.10 eq) in dry THF (140 mL) over 20 min while refluxing. After stirring for 2 h, KOtBu (0.62 g, 5.50 mmol, 1.10 eq) was added. The solution was stirred for a further 20 min, before 4-pentenal (0.42 g, 5.00 mmol, 1.00 eq) in dry THF (65 mL) was added over a period of 15 min. Stirring at reflux was continued for 4 h and for 17 h at room temperature. The solvent was concentrated under reduced pressure and the crude product was dissolved in CH 2 Cl 2 (300 mL). It was washed with sat. NH 4 Cl solution (200 mL). The aqueous phase was extracted with CH 2 Cl 2 (3 × 100 mL), the combined organic phases were washed with brine (300 mL) and dried over Na 2 -1(18),6,8,12,16(17 197.4, 196.8, 164.3, 156.6, 149.8, 134.5, 131.9,  125.8, 125.70, 118.4, 102.2, 83.5, 67.6, 62.1, 35.0, 32.7, 32.22, 28.7, 28.3, 26.9 ppm; Significant signals minor tautomer: 13 -1-carboxylate (14). 6-Heptenoic acid (2.11 mL, 15.6 mmol, 1.00 eq) in dry CH 2 Cl 2 (78 mL) was treated with EDC·HCl (3.59 g, 18.7 mmol, 1.20 eq) and DMAP (0.38 g, 3.12 mmol, 0.20 eq) at 0 • C. The solution was stirred for 20 min, before tetramic acid 11 (5.93 g, 17.2 mmol, 1.1 eq) was added at room temperature. After stirring for 4 h, the reaction was quenched with 0.5M H 2 SO 4 (250 mL). The organic phase was separated and the aqueous phase was extracted with EtOAc (3 × 150 mL). The combined organic phases were washed with brine (200 mL) and dried over Na 2 SO 4 . After removal of the solvent under reduced pressure, the crude product was purified by column chromatography (silica gel 60, 10% EtOAc in hexanes → 15% EtOAc in hexanes → 20% EtOAc in hexanes → 25% EtOAc in hexanes) to obtain 14 as an orange resin (5.64 g, 12.4 mmol, 79%).  . Bromohydrins 8 and a side product with an additional bromo substituent were obtained separately and only partially deprotected. The mixture of bromohydrin 8 and its N-Boc-protected derivative was dissolved in dry CH 2 Cl 2 (9 mL) and treated with TFA (900 µL). The solution was stirred for 15 min at room temperature and toluene (100 mL) was added. The mixture was concentrated under reduced pressure and toluene (50 mL) was added again. Removal of the solvent under reduced pressure gave 8 as a yellowish foam and as a mixture of two inseparable diastereomers of initially unknown dr according to 13 2 mM). Spraying was done with 0.4 mL/100 cm 2 . Effects were assessed after 14 days and a second time after three to six weeks when mortality was fully developed. A mortality factor, i.e., the percentage of eventually dead plants, was calculated according to the Henderson-Tilton method [22] using the formula:

(S,Z)-tert-Butyl-2-(4-(allyloxy)benzyl)-4-((E)-
with T b = number of plants before treatment (=8); T a = number of vital, treated plants at the end of observation period; C b = number of vital, untreated control plants at beginning (=8); C a = number of vital, untreated control plants at the end of observation period.

Antimicrobial Activity
The antibacterial activities were determined by the so-called broth microdilution method [23]. In brief: all cultivations were done in standard microbiological media such as TSB medium (tryptic soy broth) for S. aureus (SH1000) and LB medium (lysogeny broth) for Escherichia coli (ATCC25922) and at 37 • C (only A. baumannii was cultivated at 30 • C). The overnight cultures of the bacterial test strains were diluted to an OD 600 of 0.1 and further incubated until an OD 600 = 0.5 was reached. These cultures were used as working cultures. They were diluted to obtain an OD 600 of 0.1, determined in 45 µL of the bacterial suspension in each well of a 384-well plate, or in 90 µL of the bacterial suspension in each well of a half-area 96-well plate. Compound solutions were prepared in separate 96-well compound plates starting from 10 mM stock solutions in DMSO. The compound concentrations were adjusted so that the maximum DMSO concentrations in the assay plates were 1%, assuring no interference with growth from the solvent. The respective volumes of the compound solutions were added to the microbial suspensions with the 96channel semi-automated pipettor CyBio Selma (Analytik Jena). The OD 600 was determined directly after compound addition and subsequently after 1, 3 and 24 h using the Epoch 2 microplate reader (BioTek Instruments).

Antibiofilm Activity
Staphylococcus aureus DSM 1104 from a stock kept at −20 • C was precultured in 25 mL CASO (casein-peptone soymeal-peptone) medium in a 250 mL flask at 37 • C and shaken (100 rpm, 20 h. The OD 600 of the culture solution was adjusted to 0.001 McFarland standard. The solution was incubated in 96-well microtiter plates (TPP tissue culture ref. No. 92196) for 18 h at 37 • C with 150 µL of serially diluted test compounds (250-2 µg/mL) in CASO with 4% glucose broth. Compounds showing high activities (e.g., 2, 3, 9) were diluted in the range of 10-0.3 µg/mL. The inhibition of biofilm formation was evaluated by staining with 150 µL of 0.1% crystal violet (CV; Thermo Fisher, Waltham, MA, USA) following previously established protocols [21,24]. Briefly, the supernatant of the 96-well plate was discarded and the wells were washed once with PBS buffer. The remaining biofilms were stained with 0.1% CV at room temperature for 15 min, washed three times with PBS buffer, and finally dissolved in 150 µL ethanol (95%). The absorbance of the resulting solution at 530 nm was quantified using a plate reader (Synergy 2, BioTek, Santa Clara, CA, USA). Methanol (2.5%) and Standard deviations (SD) of two repeats with duplicates were 10% or less. Effects on the biofilms and SD values are shown in Table S1 in the Supporting Information (SI).
The precultured bacterial suspension of S. aureus strain DSM 1104 was adjusted to 0.001 McFarland standard at OD 600 and incubated in 96-well tissue microtiter plates for 18 h in 150 µL CASO with 4% glucose broth. The supernatant of the 96-well plate was removed and the remainder was washed with 150 µL PBS buffer. The test compounds were serially diluted in 150 µL of fresh media (CASO with 4% glucose broth) to concentrations of 250-2 µg/mL, and added to the wells. The plates were incubated for a further 24 h at 37 • C. Staining of the preformed biofilm and of the controls was carried out as described above [21,24]. The SD of two repeats with duplicates were 15% or less. SD values are shown in Table S1 in the ESI.