Synthesis, Fungitoxic Activity against Botrytis cinerea and Phytotoxicity of Alkoxyclovanols and Alkoxyisocaryolanols

Clovane and isocaryolane derivatives have been proven to show several levels of activity against the phytopathogenic fungus Botrytis cinerea. Both classes of sesquiterpenes are reminiscent of biosynthetic intermediates of botrydial, a virulence factor of B. cinerea. Further development of both classes of antifungal agent requires exploration of the structure–activity relationships for the antifungal effects on B. cinerea and phytotoxic effects on a model crop. In this paper, we report on the preparation of a series of alkoxy-clovane and -isocaryolane derivatives, some of them described here for the first time (2b, 2d, 2f–2h, and 4c–4e); the evaluation of their antifungal properties against B. cinerea, and their phytotoxic activites on the germination of seeds and the growth of radicles and shoots of Lactuca sativa (lettuce). Both classes of compound show a correlation of antifungal activity with the nature of side chains, with the best activity against B. cinerea for 2d, 2h, 4c and 4d. In general terms, while 2-alkoxyclovan-9-ols (2a–2e) exert a general phytotoxic effect, this is not the case for 2-arylalkoxyclovan-9-ols (2f–2i) and 8-alkoxyisocaryolan-9-ols (4a–4d), where stimulating effects would make them suitable candidates for application to plants.


Introduction
Fungi of the Botrytis (Sclerotiniaceae) genus are involved in substantial crop losses in economically relevant cultivars. While most species are host specific, B. cinerea has a wide host range which involves hundreds of vascular plant genera [1]. Infection is usually promoted by high humidity conditions, involving a variety of pathways to achieve infection and colonization [2]. Chemical control based on synthetic fungicides [3] is still the prevalent protection option for many crops, but, their increasing impact on the environment [4], concerns on residue levels and effective delivery [5], and the ever-present acquired resistance phenomenon [6], are forcing a rethinking of the way strategies of chemical control are used [7].
On the other hand, in the context of integrated pest management [8], there is an increasing interest in the use of biopesticides [9], which comprises the use of living microorganisms and substances of natural origin, such as plant and microbial extracts, as well as natural products isolated from the abovementioned sources or closely related derivatives (selected examples can be found from plants [10][11][12], microorganisms [13][14][15][16] and endophytes [17][18][19]). Therefore, there is a need for the development of novel chemical control agents, capable of being used in conjunction with living microorganisms and with the ability to antagonize B. cinerea, which calls for aiming at selective targets in the fungal infection cycle. Evidence of the protective role of endophytic microorganisms against the attack of B. cinerea on relevant cultivars [20,21] stresses the need for chemical control agents with selectivity against the phytopathogen, as this would reduce the impact on the plant microbiome.
It has been reported that commercial fungicides have an impact on plant physiology [22]. For instance, the treatment of Vitis vinifera with fludioxonil and pyrimethanil, used in the control of B. cinerea, impairs photosynthesis [23,24], while mancozeb, an extensively used contact fungicide, induces oxidative damage on L. sativa [25]. Therefore, the development of bioinspired antifungal agents should be undertaken keeping in mind these undesired effects.
Botrydial is a low molecular weight phytotoxin [26], with a carbon skeleton not following the isoprene rule, which originates from farnesyl diphosphate by folding, cyclization, and a sequence of oxidative transformations [27][28][29][30]. Genes encoding for the relevant steps of the metabolic pathway to botrydial and related compounds have also been elucidated [31,32]. This secondary metabolite is part of the interaction of the fungus with the host plant, inducing cell death [33]. The modulation of its biosynthesis may lead to the control of the fungus through a selective mechanism, as symptoms of the disease can be decreased by inhibition of the production of the toxin in deleted mutants [34].
Our research group has approached the control of the fungus B. cinerea by developing nonphytotoxic analogs of the biosynthetic intermediates of the phytotoxin botrydial. A variety of compounds with different sesquiterpenic carbon skeletons have been prepared for this purpose [26].
Clovane and isocaryolane skeleton compounds are caryophyllene cyclization derivatives [35,36], which are reminiscent of botrydial biosynthetic intermediates such as 1( Figure 1). Within this context, 2-alkoxyclovan-9-ols, obtained by the cyclization of caryophyllene oxide in alcohols, catalyzed by TCNE [37], have shown different levels of activity, which are correlated with the nature of the substituent at C-2 [38]. On the other hand, the metabolism of selected 2-alkoxyclovan-9-ols has shown the preference of the fungus for the dealkylation of the side chain at C-2, yielding less active compounds [39,40]. Reported antifungal activities of clovane derivatives show a general decrease of the activity when further hydroxyl groups are introduced, or when the side chain at C-2 is absent [39,40]. On the other hand, QSAR calculations suggest an increase in the fungal activity if the lipophilicity of the side chain at C-2 is increased [38]. Evaluation of the antifungal activity of a homologous series of 2-alkoxyclovan-9-ol derivatives and some systematic variations of active 2-(aryl)alkoxyclovan-9-ols described before [38], should shed some light on the extent and limitations of this lipophylic effect.
8-Methoxyisocaryolan-9-ol, an isocaryolane derivative, has also shown an interesting level of antifungal activity against the phytopathogen B. cinerea [41]. Incubation of this compound and related substrates like isocaryolane-8,9-diol and isocaryolan-9-ol showed, on one hand, that these compounds interfered with the production of botryane metabolites, like dihydrobotrydial (Figure 1), a metabolic sink of botryane metabolism [42]. On the other hand, it was apparent that the presence of a methoxy group at C-8 was a key feature for the antifungal activity of isocaryolane derivatives; this ether was cleaved by fungal metabolism and antifungal activity decreased as the hydroxylation level increased [41,43]. Therefore, evaluation of a homologous series of 8-alkoxyisocaryolan-9-ol derivatives should shed some light on the suitability of isocaryolane compounds as antifungal leads.
Another aspect to consider in the development of clovane or isocaryolane derivatives as effective fungal leads, is the extent of the phytotoxic effect exerted by these compounds, which should be reduced as much as possible if the compounds of interest are to find practical applications in the field. Therefore, in this paper, we report the preparation of a homologous series of 2alkoxyclovan-9-ols (2a-2e), some related to 2-arylethoxyclovan-9-ols (2f-2i), as well as a homologous series of 8-alkoxyisocaryolan-9-ols (4a-4e). We also evaluate their antifungal properties against B. cinerea, and their phytotoxic activities on the germination of seeds and growth of radicles and shoots of L. sativa (lettuce) [44][45][46][47], a plant successfully used in the evaluation of the phytotoxicity of bioactive natural products.

General Experimental Procedure
Melting points were measured with a Reichert-Jung Kofler block and are uncorrected. Optical rotations were determined on a Perkin-Elmer 341 polarimeter. IR spectra were recorded on a Perkin-Elmer Spectrum BX FT-IR spectrophotometer. 1 H and 13 C NMR spectra were obtained on a Varian INOVA 400 NMR spectrometer using tetramethylsilane as an internal reference. NMR assignments were made by a combination of 1D and 2D techniques and by comparison with assignments available in the literature for previously described compounds, where appropriate. Mass spectra were recorded on a Finnigan Voyager spectrometer at 70 eV. High resolution mass spectra were recorded on a Micromass Autospec spectrometer at 70 eV or on a Waters Synapt G2 QTOF spectrometer in ESI mode. HPLC was performed with a Hitachi/Merck L-6270 apparatus equipped with a UV-VIS detector (L 4250) and a differential refractometer detector (RI-71). TLC was performed on Merck Kiesegel 60 F254 plates, with a 0.2 mm thick film. Silica gel (Merck) was used for column chromatography. Purification by HPLC was performed using a Si gel column (LiChrospher Si 60, 10 µm, 1 cm wide, 25 cm long).

Microorganism and Antifungal Assays
The culture of the B. cinerea strain UCA 992 employed in this work was isolated from Domecq vineyard grapes, Jerez de la Frontera, Cádiz, Spain. This culture of B. cinerea has been deposited at the Mycological Herbarium Collection (UCA), Facultad de Ciencias, Universidad de Cádiz. Antifungal bioassays were performed by measuring radial growth on agar medium in a Petri dish in the presence of test compounds. Test compounds were dissolved in EtOH to a final compound concentration in the culture medium of 10 −4 M. Solutions of test compounds were added to glucose-malt-peptone-agar medium (61 g of glucose-malt-peptone-agar per L, pH 6.5-7.0). The final EtOH concentration was identical in both the control and treated cultures. The medium was poured into 6 or 9 cm diameter sterile plastic Petri dishes and a mycelia disk of B. cinerea, 5 mm in diameter and cut from an actively growing culture, was placed in the center of the agar plate. Radial growth was measured for five days. The fungal growth inhibition percentage was calculated as ((D1 − D2)/D1) × 100, where D1 is the average diameter of the control colony and D2 is the average diameter of the treated colony. Every concentration was evaluated in triplicate. Fungal growth inhibition percentages are given as averages +/− standard deviations 2.5. Bioassay for the Germination and Growth of Lettuce Seeds L. sativa (cv. Grand Rapids) seeds were purchased from Isla Pak, RS, Brazil. All undersized and damaged seeds were discarded. The bioassay was conducted in 100 mm Petri dishes containing Whatman #1 filter paper (90 mm) as a support. L. sativa seeds (25 per dish), controls, and test compound solutions (10 mL) at concentrations of 10 −3 and 10 −5 M were placed in the dishes. All solutions were prepared with deionized water and the pH values, buffered with 10 mM 2-(N-morpholino)ethanesulfonic acid (MES), were adjusted to 6.0-6.5 with NaOH solution. Concentrations lower than 10 −3 M were obtained by serial dilution. Dishes were wrapped with Parafilm to reduce evaporation and incubated in the dark at 25 • C in an environmental chamber. After 5 days, germinated seeds were counted (a seed was considered to be germinated when the radicle was at least 0.2 mm long) and the lengths of the radicle and shoots were measured using a pachymeter [43]. Dishes were kept at 4 • C during the measurement process to prevent subsequent growth.
Osmotic pressure values were measured on a micro-osmometer and ranged between 30 and 38 mOsmolar [45]. The experiments were carried out in triplicate.
The effects of the test compounds on L. sativa germination and growth are given as percentage differences from the control, calculated from the differences (in cm) between mean values obtained upon the addition of test compounds and mean values obtained for the control: (seeds grown without addition of tested compounds)/mean values for control × 100. Thus, zero represents the control, positive values represent stimulation of the studied parameter, and negative values represent inhibition. The data were evaluated using Student's t-tests and the differences between the experiment and control were significant at a value of p ≤ 0.0517.

Results and Discussion
The chemical transformations carried out for the preparation of 2-alkoxyclovan-9-ols (2a-2e) and 2-arylethoxyclovan-9-ols (2f-2i) are summarized in Figure 2. Compounds 2a-2g were obtained by the treatment of caryophyllene oxide with a catalytic amount of tetracyanoethylene (TCNE), dissolved in the appropriate alcohol, at room temperature [37]. On the other hand, compounds 2h and 2i were obtained in a modified procedure, by the treatment of caryophyllene oxide with a catalytic amount of TCNE and 1.3-1.4 equivalents of the appropriate alcohol, dissolved in propan-2-one, at room temperature. These compounds were identified by their 1 H (see Table 2) and 13 C (see Table 3) NMR spectra, combined with 1D and 2D NMR techniques and by comparison with authentic samples, where appropriate. All synthesized compounds (2a-2i) displayed the characteristic pattern of 1 H and 13 C NMR signals of clovan-9-ols, bearing different chains at carbon C-2. The compounds2a [37], 2c [48], 2e [48], and2i [24] were previously described in the literature. A revised assignment of 1 H and 13 C NMR spectra for compound 2i together with selected HMBC correlations can be found in the Supporting Information section.
The spectroscopy data of compound 2g were very similar to those of 2f. A comparison of the 1 H NMR spectra of both compounds showed as the main difference, the deshielding of the signal corresponding to H-2' from 2.79 ppm in 2f to 4.03 ppm in 2g (Table 2). This, together with the presence in the 13 C NMR spectrum of two methylenes, one of them bearing an oxygen function, and five aromatic methines, in addition to the fifteen characteristic carbons of the clovane skeleton (Table 3) [36]. Extension of this procedure to higher alcohols led to inseparable mixtures. Alternatively, the treatment of caryophylladienol 3 with a catalytic amount of Sn(OTf) 2 , dissolved in the appropriate alcohol (from methanol to pentan-1-ol), led to the preparation of a series of (8R,9R)-8-alkoxyisocaryolan-9-ols (4a-4e), with yields ranging from 34% to 42%, together with variable amounts of isocaryolanone 5 and aldehyde 6 ( Figure 3). These compounds were identified by their 1 H and 13 C NMR spectra (see Table 5), combined with 1D and 2D NMR techniques and by comparison with authentic samples, where appropriate. While compound 4a was previously described by us [36], compound 4b was reported as rumphellol B, a metabolite of the gorgonian coral Rumphella antipathies [49].
(8R,9R)-8-Alkoxyisocaryolan-9-ols 4c to 4e displayed the characteristic pattern of 1 H and 13 C NMR signals of this family of compounds (Table 5). So, compound 4c showed signals at δ H 3.56 (1H, dd, J = 10.8, 6.3 Hz), corresponding to H-9β, a multiplet between 3.40-3.28 ppm, assigned to the methylene C-1 of the lateral chain at C-8, and three singlet signals integrating for three protons at δ H 0.97, 0.96, and 0.79 ppm, characteristic of the methyl groups C-13 to C-15. In addition, it can be observed there was a triplet signal integrating for three protons at δ H 0.88 (J = 7.4 Hz), which was assigned to the methyl group of the chain at C-8, and two doublets, at δ H 1.87 and 0.89 (J = 13.0 Hz), corresponding to each of the protons on carbon C-12. These data led to the assignment of the structure for compound 4c as (8R,9R)-8-propoxyisocaryolan-9-ol. All the carbons and their associated proton signals were assigned using the COSY, HSQC and HMBC spectra.
HREIMS of compounds 4d and 4e, obtained as colorless oils, indicated the presence of molecular ions at m/z 294.2592 and 308.2733, which corresponded with the molecular formulas of C 19 H 34 O 2 and C 20 H 36 O 2 , consistent with the structures proposed respectively, for compounds 4d and 4e.
The 1 H NMR signals of compounds 4d and 4e resembled those of the previously described (8R,9R)-8-alkoxyisocaryolan-9-ols 4a to 4c. On the other hand, their 13 C NMR and HSQC spectra showed significant differences, in addition to the characteristic 15 carbons of the isocaryolane skeleton. While compound 4d presented three methylenes, one of them bearing an oxygen function, and a methyl group, compound 4e possessed a methyl group and four methylenes, one of them bearing an oxygen function, which is consistent for the presence of a butoxy chain for compound 4d and a pentoxy chain in compound 4e. These observations, together with a careful analysis of the COSY and HMBC experiments, led to the identification of compound 4d as (8R,9R)-8-butoxyisocaryolan-9-ol and of compound 4e as (8R,9R)-8-pentoxyisocaryolan-9-ol.
The antifungal properties of compounds 2a-2i and 4a-4e, at a 10 −4 M dose, which is a lower dose than the one previously reported for 2-alkoxyclovan-9-ols [38], were determined against the growth of B. cinerea using the poisoned food technique [50]. The commercial fungicide Euparen ® (dichlofluanid) was used as a standard for comparison in this test. Several levels of inhibition were observed (Figure 4). For compounds 2a-2e, where the side chain at C-2 varies from methoxy to pentoxy substituents, a tendency of increasing activities was observed, reaching a maximum for (2S,9R)-2-butoxyclovan-9-ol (2d) (Figure 4). These results were consistent with the suggested predictions resulting from a QSAR model published previously for this class of compounds [38]. See a discussion on the correlation between the antifungal activity and log P below. On the other hand, the antifungal activities of clovan-9-ol derivatives O-substituted at C-2, do not depend only on the lipophilic character of the side chain [38]. For instance, as we have described in our previous work [38], compounds including nitrogen atoms in this side chain at C-2, such as (2S,9R)-2-(2 -(p-nitrophenoxy)ethoxy)clovan-9-ol (2i), display good levels of activity at doses above 50 ppm. In order to evaluate the different contributions of the structural components of the aryl substituent on the antifungal activity, compounds 2f-2i were prepared, as described above, and their antifungal properties were compared with those of compounds 2a-2e (Figure 4). Compounds 2f and 2g, which do not present a p-nitro group, displayed antifungal activities lower than those of (2S,9R)-2methoxyclovan-9-ol (2a), the less active compound on the homologous series described above, while p-nitro containing compounds showed increased levels of activity. The most active compound of this series, (2S,9R)-2-(2 -(p-nitrophenyl)ethoxy)clovan-9-ol (2h), presented an activity profile comparable to the one of (2S,9R)-2-butoxyclovan-9-ol (2d), the most active compound in the above described homologous series (Figure 4), which is an improvement on previously described data [38], where a higher concentration was used (50 ppm). Compounds 2a and 2i, also evaluated in a previous study [38], are included here for comparison purposes.
Most of the active compounds in this study (2d, 2h, 4c and 4d) present a calculated [51] log p near to the upper limit for this parameter, established by the Lipinsky "rule of five" for the oral availability of drugs-most molecules which show activity via oral administration have log p < 5, molecular weight <500 and should contain no more than 10 hydrogen bond acceptors and less than 5 hydrogen bond donors- [52] (compound 2d, log p = 4.499; compound 2h, log p = 4.824; compound 4c, log p = 4.310 and compound 4d, log p = 4.870). Compounds with longer lipophilic aliphatic chains like 2e and 4e show a decreased activity, presenting log p values above five (compound 2e, log p = 5.004; compound 4e, log p = 5.370). For 2-alkyloxyclovan-9-ols 2a-2c, a roughly linear correlation between log p values and activity ( Figure 4) can be established (compound 2a, log p = 3.061; compound 2b, log p = 3.437; compound 2c, log p = 3.939). This is not the case for either 2-arylalkyloxyclovan-9-ols (2f-2i) or 8-alkyloxyisocaryolan-9-ols (4a-4d), where additional factors, like electronic distribution, must be taken in account.
On the other hand, the introduction of an aromatic moiety in the side chain of clovane derivatives (compounds 2f-2i), sharply increased the inhibitory effect on germination at the highest concentration used (10 −3 M). Furthermore, this effect is exacerbated when either a nitro or an alkoxy moiety is connected to the aromatic ring (see compounds 2g and 2h). Noteworthily, the simultaneous presence of both moieties, arranged in a 1,4 substitution pattern, led to a cancellation of the above mentioned effects, as germination inhibition by compound 2i is similar to that shown by (2S,9R)-2-(2 -phenylethoxy)clovan-9-ol (2f).
Different patterns were shown for the effect of clovane derivatives on radicle and shoot growth ( Figure 6). On one hand, alkoxyclovanols 2a-2e showed a general inhibitory effect on shoot growth at the highest concentration evaluated (10 −3 M), with a maximum inhibitory effect observed for (2S,9R)-2-propoxyclovan-9-ol (2c). A different response was found for this homologous series (2a-2e) on radicle growth, where a progression on the length of the alkyl chain leads to activity changes from inhibitory to stimulatory. A maximum stimulatory effect was found for (2S,9R)-2-butoxyclovan-9-ol (2d), the most effective alkoxyclovanol of the homologous series 2a-2e, for the inhibition of the growth of B. cinerea (10 −4 M dose, see Figure 4). On the other hand, while there was almost no effect of (2S,9R)-2-(2 -phenoxyethoxy)clovan-9-ol (2g) on shoot growth, and a moderate inhibitory effect on radicle growth, there was a general stimulatory effect, on both shoot and radicle growth, for compounds 2f, 2h, and 2i, where 2i showed the greatest stimulatory effect. Consequently, given that (2S,9R)-2-(2 -(p-nitrophenyl)ethoxy)clovan-9-ol (2h) showed an inhibitory effect on the growth of B. cinerea of a similar magnitude to that of the most effective alkoxyclovanol of the homologous series (2d) (10 −4 M dose, see Figure 4), this made compound 2h the most potentially useful of the clovane derivatives evaluated so far, as the inhibition of germination was compensated with stimulation of both radicle and shoot growth. Compound 2d, with a comparable inhibitory effect on B. cinerea growth (see Figure 4), showed inhibitory effects on both germination and shoot growth and only a moderate stimulatory effect on radicle growth (see Figure 6). 14 of 18 cinerea growth (see Figure 4), showed inhibitory effects on both germination and shoot growth and only a moderate stimulatory effect on radicle growth (see Figure 6).
On the other hand, different trends could be found for the isocaryolane derivatives evaluated. Along the homologous series 4a-4e, an increase of the stimulation of germination could be found at the highest dose evaluated (10 −3 M), as discussed above ( Figure 5). Additionally, also at the 10 −3 M dose, different tendencies could be found for radicle and shoot growth ( Figure 6); while a general stimulatory effect was apparent for radicle growth, with a maximum effect found for (8R,9R)-8-propoxyisocaryolan-9-ol (4c), a general inhibitory effect on shoot growth was observed. Given that (8R,9R)-8propoxyisocaryolan-9-ol (4c) had inhibitory activity on the growth of B. cinerea of a similar magnitude to the most effective alkoxyclovanol, compound 2d (see Figure 4), this made compound 4c the most potentially useful of the isocaryolane derivatives evaluated so far.
Supplementary Materials: The following are available online at www.mdpi.co assignments of 1 H and 13 C NMR spectra for compound 2i, together with correlations. Table S1. Figures S1-S20a. 1 H-and 13 C NMR spectra of compounds 4b-4e. On the other hand, different trends could be found for the isocaryolane derivatives evaluated. Along the homologous series 4a-4e, an increase of the stimulation of germination could be found at the highest dose evaluated (10 −3 M), as discussed above ( Figure 5). Additionally, also at the 10 −3 M dose, different tendencies could be found for radicle and shoot growth ( Figure 6); while a general stimulatory effect was apparent for radicle growth, with a maximum effect found for (8R,9R)-8-propoxyisocaryolan-9-ol (4c), a general inhibitory effect on shoot growth was observed. Given that (8R,9R)-8-propoxyisocaryolan-9-ol (4c) had inhibitory activity on the growth of B. cinerea of a similar magnitude to the most effective alkoxyclovanol, compound 2d (see Figure 4), this made compound 4c the most potentially useful of the isocaryolane derivatives evaluated so far.
On the other hand, in the dose range 10 −3 M to 10 −5 M, different phytotoxicity profiles can be associated to the structural classes mentioned above. In general terms, 2-alkoxyclovan-9-ols (2a-2e) exert a phytotoxic effect, especially on shoot growth (see Figure 6). 2-Arylalkoxyclovan-9-ols (2f-2i) are phytotoxic on germination (see Figure 5), but stimulate shoot and radicle growth (Figure 6), which would make them suitable candidates for post-germination applications, with compound 2h being the best candidate. Finally, 8-alkoxyisocaryolan-9-ols (4a-4d) inhibit shoot growth ( Figure 6), but stimulate germination ( Figure 5) and radicle growth (Figure 6), which would make them suitable candidates for both pre-and post-germination applications, with compounds 4c and 4d being the best candidates.