Synthesis and Herbicidal Activity Against Buffelgrass (Cenchrus ciliaris) of (±)-3-deoxyradicinin

A novel synthetic strategy for obtainment of (±)-3-deoxyradicinin (2) is reported. This synthetic methodology is more efficient than those previously reported in the literature and also shows higher versatility towards the introduction of different side-chains at both C-7 and C-2. The obtained compound (±)-2 shows phytotoxicity against the grass-weed buffelgrass comparable to that of the natural phytotoxin radicinin (1). Therefore, (±)-2 can constitute a more practical synthetic alternative to 1 as bioherbicide for buffelgrass control.


Introduction
Fungal phytotoxins constitute an essentially endless source of bioactive metabolites. Their large structural diversity reflects in a wide variety of bioactivities and potential applications in both medicine and agriculture. One of the most appealing application of fungal phytotoxins is in the development of bioherbicides, which show lower or nil toxicity, and then a lower environmental and ecological impact, than the traditional synthetic pesticides [1][2][3]. For this reason, these are now emerging as a very promising tool for weed management. The search for new and selective bioherbicides usually starts with the isolation of phytotoxins from weed pathogens [3]. This strategy has been followed also for the search of selective compounds for the control of buffelgrass (Pennisetum ciliare or Cenchrus ciliaris), a perennial grass-weed native to Africa, the Mediterranean area, and Middle-East [4], which has become highly invasive in North America and Australia where it is a menace for the native vegetation, infesting roadsides and urban landscapes and promoting wildfires. Buffelgrass control is thus an emerging concern in those countries, where it is now carried out by broad-spectrum herbicides which, however, can heavily damage non-target native plants [5]. Looking for novel bioherbicides for this weed control, the production of phytotoxic metabolites by the buffelgrass foliar pathogens Cochliobolus australiensis and Pyricularia grisea has been recently investigated [6][7][8]. From the liquid culture of C. australiensis, the known phytotoxin radicinin (1) ( Figure 1) and a new dihydropyran-pyran-4,5-dione named cochliotoxin were isolated together with their close metabolites 3-epi-radicinin, radicinol and 3-epi-radicinol [6]. Radicinin (1) was first isolated from the fungus Stemphylium radicinum [9] and its structure was elucidated some years later by chemical and spectroscopic studies [10] while the absolute configuration was established by X-ray analysis of its 4-p-bromobenzoate [11]. Compound 1 has been also found to be produced by many other fungal species [11][12][13][14][15][16][17]. Phytotoxins 1 and cochliotoxin, other fungal species [11][12][13][14][15][16][17]. Phytotoxins 1 and cochliotoxin, together with the other compounds isolated from P. grisea [8], were assayed for their phytotoxic activity on host and non-host indigenous plants. Radicinin (1) demonstrated high target-specific toxicity on buffelgrass, low toxicity to native plants, and no teratogenic, sub-lethal, or lethal effects on zebrafish (Brachydanio rerio) embryos [18]. Previous studies had also shown its antifungal, insecticidal and plant growth regulatory activity, as well as antibiotic activity against Gram-positive bacteria, including Staphylococcus aureus [19,20]. More recently, the ability of 1 to inhibit Xylella fastidiosa, the causal agent of many devastating plant diseases, including Pierce′s Disease of grapevine, phony peach disease, alfalfa dwarf disease, plum leaf scald, citrus variegated chlorosis, and leaf scorch of almond, coffee, elm, oak, oleander, pear, and sycamore was reported [21]. Although the high and specific bioactivity of 1 against buffelgrass makes this phytotoxin very promising for the development of bioherbicides for the control of this weed, the low quantity of 1 available from fungal sources constitute a real bottleneck for such development, do not allowing open field studies [22]. Furthermore, the large scale production of fungal natural products requires the use of fermenters where the sticking of fungal mycelium to walls and blades often results as the most critical problem, strongly affecting the production yield [23]. For these reasons, we considered worthwhile to develop an efficient synthetic strategy for the obtainment of 1 or some its active analogues. In particular, we faced the synthesis of racemic (±)-3-deoxyradicinin (2) ( Figure 1) the immediate biosynthetic precursor of 1 [24] with the aim to define the effect on phytotoxicity of both the absolute configuration and the presence of the hydroxy moiety at C-3. In fact, a recent structure-activity relationship study carried out on 1 by some of us demonstrated that the presence of an α,β-unsaturated carbonyl group at C-4 is essential for the phytotoxic activity and that the double bond of the side chain at C-7 also play a role to impart activity [25]. However, the effect on phytotoxicity of the hydroxy group at C-3 was not ascertained with certainty. The absolute configuration at C-3 seemed to have some role, because 3-epi-radicinin was found to be less active, but the effect of the absolute stereochemistry at C-2 was not investigated [25]. Our aim was then to develop an efficient synthetic strategy to obtain (±)-2 in high yield, starting from commercially available products. The synthetic strategy was also designed to be versatile for the synthesis of analogues of 2 with different side-chains at C-7. Moreover, the transformation of 2 to 1 has been reported in the literature [26], therefore, a novel approach to (±)-2, will also constitute a formal synthesis of (±)-1.

Total Synthesis of (± )-3-deoxyradicinin (2).
To the best of our knowledge, only two approaches to total synthesis of (±)-2 have been reported in the literature so far. However, the first procedure, reported by Kato et al. in 1969 [26], appears unpractical, providing (±)-2 in five steps and <10% overall yield, but from not commercially available precursors which, in turn, require some additional steps for preparation [27]. On the other hand, the second one by Suzuki et al. [24], although leading to (±)-2 in only four steps from triacetic lactone methyl ether, still provides only 1% of overall yield. In addition, both approaches appear to be poorly versatile to other pyrones differently substituted at C-7.
We then decided to apply a different approach, starting from commercially available 2,2,6trimethyl-4H-1,3-dioxin-4-one (3) (Scheme 1). Our aim was then to develop an efficient synthetic strategy to obtain (±)-2 in high yield, starting from commercially available products. The synthetic strategy was also designed to be versatile for the synthesis of analogues of 2 with different side-chains at C-7. Moreover, the transformation of 2 to 1 has been reported in the literature [26], therefore, a novel approach to (±)-2, will also constitute a formal synthesis of (±)-1.

Total Synthesis of (±)-3-deoxyradicinin (2)
To the best of our knowledge, only two approaches to total synthesis of (±)-2 have been reported in the literature so far. However, the first procedure, reported by Kato et al. in 1969 [26], appears unpractical, providing (±)-2 in five steps and <10% overall yield, but from not commercially available precursors which, in turn, require some additional steps for preparation [27]. On the other hand, the second one by Suzuki et al. [24], although leading to (±)-2 in only four steps from triacetic lactone methyl ether, still provides only 1% of overall yield. In addition, both approaches appear to be poorly versatile to other pyrones differently substituted at C-7.
Accordingly, the dioxinone 3 was enolized with lithium diisopropylamide (LDA) at −78 °C in THF and trapped with trimethylsilyl chloride (TMSCl), providing the silyl dienolether 4 in 63% yield after high vacuum distillation (pressure lower than 1.0 mmHg) of the crude product. This yield may be considered satisfactory, taking into account that the purification of 4 by distillation shall be carried out on strictly controlled temperature conditions because it decomposes when heated at temperatures higher than 50 °C. The silyl enol ether 4 was reacted with crotonaldehyde in Mukaiyama conditions in the presence of TiCl4 in dichloromethane at −78 °C, affording the alcohol 5a in 61% yield. Oxidation of 5a with Dess-Martin periodinane gave rise to the corresponding ketone 6a in practically quantitative yield. The latter compound spontaneously eliminated acetone and cyclized upon heating in toluene, providing the hydroxy α-pyrone 7a in 96% yield. Finally, acylation of 7a with crotonyl chloride, followed by in situ intramolecular Michael addition of the hydroxy moiety, provided the desired deoxyradicinin (±)-2 in 50% yield (overall yield 18%). Notably, such synthetic strategy is versatile towards the synthesis of radicinin analogues with different substituents at C-7 and C-2. In fact, by simply employing a different aldehyde in the Mukaiyama reaction with compound 4 can be introduced a different side-chain at C-7, while the use of a different α,βunsaturated acid chloride in the last step allows to insert a different substituent at C-2. This potentiality was verified by the synthesis of the methyl-substituted dihydropyran-pyran-4,5-dione (±)-8 which was obtained following the same synthetic procedure described for (±)-2 (Scheme 1). In fact, the Mukaiyama reaction of 4 with acetaldehyde provided alcohol 5b which, after oxidation to 6b and subsequent cyclization, provided the pyrone 7b. The acylation with crotonyl chloride finally provided (±)-8 in similar (17%) overall yield.
With (±)-2 in hand, we considered worthwhile to try its conversion to (±)-1 following the oxyacetylation procedure described in the literature by Kato et al. [26] (Scheme 2). Accordingly, (±)-2 was treated with lead tetracetate in acetic acid, stirring at 100 °C for 2h. Contrary to what reported Scheme 1. Synthesis of (±)-3-deoxyradicinin (2) and analogue (±)-8. Accordingly, the dioxinone 3 was enolized with lithium diisopropylamide (LDA) at −78 • C in THF and trapped with trimethylsilyl chloride (TMSCl), providing the silyl dienolether 4 in 63% yield after high vacuum distillation (pressure lower than 1.0 mmHg) of the crude product. This yield may be considered satisfactory, taking into account that the purification of 4 by distillation shall be carried out on strictly controlled temperature conditions because it decomposes when heated at temperatures higher than 50 • C. The silyl enol ether 4 was reacted with crotonaldehyde in Mukaiyama conditions in the presence of TiCl 4 in dichloromethane at −78 • C, affording the alcohol 5a in 61% yield. Oxidation of 5a with Dess-Martin periodinane gave rise to the corresponding ketone 6a in practically quantitative yield. The latter compound spontaneously eliminated acetone and cyclized upon heating in toluene, providing the hydroxy α-pyrone 7a in 96% yield. Finally, acylation of 7a with crotonyl chloride, followed by in situ intramolecular Michael addition of the hydroxy moiety, provided the desired deoxyradicinin (±)-2 in 50% yield (overall yield 18%). Notably, such synthetic strategy is versatile towards the synthesis of radicinin analogues with different substituents at C-7 and C-2. In fact, by simply employing a different aldehyde in the Mukaiyama reaction with compound 4 can be introduced a different side-chain at C-7, while the use of a different α,β-unsaturated acid chloride in the last step allows to insert a different substituent at C-2. This potentiality was verified by the synthesis of the methyl-substituted dihydropyran-pyran-4,5-dione (±)-8 which was obtained following the same synthetic procedure described for (±)-2 (Scheme 1). In fact, the Mukaiyama reaction of 4 with acetaldehyde provided alcohol 5b which, after oxidation to 6b and subsequent cyclization, provided the pyrone 7b. The acylation with crotonyl chloride finally provided (±)-8 in similar (17%) overall yield.
With (±)-2 in hand, we considered worthwhile to try its conversion to (±)-1 following the oxyacetylation procedure described in the literature by Kato et al. [26] (Scheme 2). Accordingly, (±)-2 was treated with lead tetracetate in acetic acid, stirring at 100 • C for 2h. Contrary to what reported in the literature, where obtainment of only 3-acetoxy radicinin was claimed, we got 6:4 trans:cis mixture of (±)-radicinin acetate 9a and 3-epi-(±)-radicinin acetate 9b (19% overall) together with the unsaturated pyranopyrandione 10 (74%) as the main product of the reaction. The latter probably formed from (±)-9 through spontaneous acetic acid elimination. The low yield obtained, the low diastereoselectivity observed and, most importantly, the large formation of 10 as a by-product in the present reaction makes necessary a further optimization of such synthetic transformation, possibly trying different reagents for the α-oxidation of 3-deoxyradicinin (2). This type of study is out of the scope of the present research and will be developed successively. However, the availability of the radicinin derivative 10 prompted us to test also this compound in the phytotoxicity assays on C. ciliaris (vide infra). in the literature, where obtainment of only 3-acetoxy radicinin was claimed, we got 6:4 trans:cis mixture of (±)-radicinin acetate 9a and 3-epi-(±)-radicinin acetate 9b (19% overall) together with the unsaturated pyranopyrandione 10 (74%) as the main product of the reaction. The latter probably formed from (±)-9 through spontaneous acetic acid elimination. The low yield obtained, the low diastereoselectivity observed and, most importantly, the large formation of 10 as a by-product in the present reaction makes necessary a further optimization of such synthetic transformation, possibly trying different reagents for the α-oxidation of 3-deoxyradicinin (2). This type of study is out of the scope of the present research and will be developed successively. However, the availability of the radicinin derivative 10 prompted us to test also this compound in the phytotoxicity assays on C. ciliaris (vide infra).

Phytotoxicity Assays on C. ciliaris.
Compounds (±)-2 and 10 were tested by buffelgrass (C. ciliaris) leaf puncture assay at 2.5 × 10 −3 M as reported in Materials and Methods section and their activity was evaluated in comparison with that showed by the natural metabolite radicinin (1) previously reported by Masi et al. [18] (Figure 2). As inferred from Figure 2, the activity of compound (±)-2 is comparable to that of the natural radicinin (2S,3S)-1, while the activity of the pyranopyrandione 10 is slightly lower. The lack of the hydroxy group on C-3 in 2 seems to suggest, contrary to previous observations [25], that this moiety is not essential for the compounds phytotoxicity. Moreover, the comparable activity of racemic 2 and non-chiral compound 10 with optically active (2S,3S)-1, highlight that the absolute configuration of the methyl group at C-2 also has a minor influence on phytotoxicity. As far as absolute configuration at C-3 is concerned, previous results have shown that in the presence of a hydroxy moiety on this stereocenter its stereochemistry results very important to impart a strong activity considering the Scheme 2. Acetoxylation of (±)-3-deoxyradicinin (2) with Pb(OAc) 4 (bold and dashed lines in figure refer to the relative configuration).

Phytotoxicity Assays on C. ciliaris
Compounds (±)-2 and 10 were tested by buffelgrass (C. ciliaris) leaf puncture assay at 2.5 × 10 −3 M as reported in Materials and Methods section and their activity was evaluated in comparison with that showed by the natural metabolite radicinin (1) previously reported by Masi et al. [18] (Figure 2). in the literature, where obtainment of only 3-acetoxy radicinin was claimed, we got 6:4 trans:cis mixture of (±)-radicinin acetate 9a and 3-epi-(±)-radicinin acetate 9b (19% overall) together with the unsaturated pyranopyrandione 10 (74%) as the main product of the reaction. The latter probably formed from (±)-9 through spontaneous acetic acid elimination. The low yield obtained, the low diastereoselectivity observed and, most importantly, the large formation of 10 as a by-product in the present reaction makes necessary a further optimization of such synthetic transformation, possibly trying different reagents for the α-oxidation of 3-deoxyradicinin (2). This type of study is out of the scope of the present research and will be developed successively. However, the availability of the radicinin derivative 10 prompted us to test also this compound in the phytotoxicity assays on C. ciliaris (vide infra). figure  refer to the relative configuration).

Phytotoxicity Assays on C. ciliaris.
Compounds (±)-2 and 10 were tested by buffelgrass (C. ciliaris) leaf puncture assay at 2.5 × 10 −3 M as reported in Materials and Methods section and their activity was evaluated in comparison with that showed by the natural metabolite radicinin (1) previously reported by Masi et al. [18] (Figure 2). As inferred from Figure 2, the activity of compound (±)-2 is comparable to that of the natural radicinin (2S,3S)-1, while the activity of the pyranopyrandione 10 is slightly lower. The lack of the hydroxy group on C-3 in 2 seems to suggest, contrary to previous observations [25], that this moiety is not essential for the compounds phytotoxicity. Moreover, the comparable activity of racemic 2 and non-chiral compound 10 with optically active (2S,3S)-1, highlight that the absolute configuration of the methyl group at C-2 also has a minor influence on phytotoxicity. As far as absolute configuration at C-3 is concerned, previous results have shown that in the presence of a hydroxy moiety on this stereocenter its stereochemistry results very important to impart a strong activity considering the As inferred from Figure 2, the activity of compound (±)-2 is comparable to that of the natural radicinin (2S,3S)-1, while the activity of the pyranopyrandione 10 is slightly lower. The lack of the hydroxy group on C-3 in 2 seems to suggest, contrary to previous observations [25], that this moiety is not essential for the compounds phytotoxicity. Moreover, the comparable activity of racemic 2 and non-chiral compound 10 with optically active (2S,3S)-1, highlight that the absolute configuration of the methyl group at C-2 also has a minor influence on phytotoxicity. As far as absolute configuration at C-3 is concerned, previous results have shown that in the presence of a hydroxy moiety on this stereocenter its stereochemistry results very important to impart a strong activity considering the significant reduction in phytotoxicity showed by (2S,3R)-3-epi-radicinin in comparison with (2S,3S)-1 [25].
The obtained results highlight that derivatives like (±)-2 and 10, much simpler to be prepared than optically active radicinin (1) can be considered as valid alternative to this phytotoxin as possible bioherbicides against buffelgrass. Moreover, the possibility to employ, with similar results, racemic or non-chiral compounds in place of enantiopure ones is a great advantage for the preparation of the large amounts of compounds necessary for the field test of bioherbicides. In fact, it is obviously much simpler to prepare on large scale a racemic compound than an optically active one, which usually require to employ expensive chiral reagents or catalysts.

General Experimental Procedures
1 H (400 MHz) and 13 C (100 MHz) NMR spectra were recorded in CDCl 3 or DMSO-d6 on a Varian INOVA 400 spectrometer, using TMS as an internal standard (see Supplementary Materials). GC/MS spectra were recorded on a HP 6890 gas chromatograph, equipped with a HP-5975 mass spectrometric detector. HRESI mass spectra were recorded on Agilent Technologies 1100 LC/MS TOF instrument. Analytical TLC was performed on silica gel 60 Macherey-Nagel sheets. The spots were visualized by exposure to UV radiation (254 nm) and/or by spraying the plates with a potassium permanganate solution. Column chromatography was performed using silica gel (Merck, Kieselgel 60, (60-230 mesh). THF was freshly distilled before its use on sodium benzophenone ketyl under nitrogen atmosphere. Diisopropylamine (DIPA) and CH 2 Cl 2 were dried by distillation over calcium hydride and stored under a nitrogen atmosphere. Trimethylsilyl chloride (TMSCl), crotonaldehyde, and crotonoyl chloride were distilled and stored under argon atmosphere before their use. Commercially available n-butyllithium (Aldrich) was a 2.5 M solution in hexanes. The other analytical grade solvents and commercially available reagents were used without further purification.

((2,2-Dimethyl-4-methylene-4H-1,3-dioxin-6-yl)oxy)-trimethylsilane (4)
A solution of DIPA (5.4 mL, 1.1 eq) in THF was stirred under nitrogen atmosphere and cooled to 0 • C, then n-butyllithium 2.5 M (15.4 mL, 1.1 eq) was slowly added in about 10 min and the mixture was stirred at the same temperature for another 45 min. Afterwards, the pale yellow solution was cooled to −78 • C and 2,2,6-trimethyl-4H-1,3-dioxin-4-one 3 (4.65 mL, 35 mmol) was added in 5 min. After 1h, also freshly distilled TMSCl (5.33 mL, 1.2 eq) was added in about 10 min; the dark orange mixture was stirred for 1 h and then slowly heated to room temperature. The mixture was filtered under reduced pressure, the solid was washed with hexanes and the filtrate concentrated by distillation under a reduced pressure. The crude product was purified by distillation under a high vacuum, using a rotary pump (pressure less than 1 mmHg) and heating to 50 • C. The pure silyl dienolether 4 [28] (4.76 g, 63%) corresponded to the collected fraction having bp = 32-34 • C.

Leaf Puncture Bioassays
The synthetic derivatives (±)-2 and 10 were assayed at 2.5 × 10 −3 M for phytotoxicity on leaves of buffelgrass (C. ciliaris) following a procedure previously reported [22,25] and their activity compared with that of natural 1 [18]. Accordingly, compounds were first dissolved in MeOH (final concentration 4%) and stock solutions using sterile distilled water were then prepared. An incision of ca. 3 mm was made on the adaxial surface of each leaf section of 3 cm with an insulin needle. The leaf sections were placed in groups of six on the surface of a water-saturated filter paper in each of four petri dishes. Five leaf sections in each petri dish were tested with the solution containing the compound, while one leaf section was used as a negative control (4% MeOH only). A droplet (10 µL) of the appropriate solution was applied over each needle incision using a micropipette. The dishes were sealed with parafilm and incubated at 24 • C for 3 days in a temperature-regulated chamber under a photoperiod of 14-10 h (light/dark). After three days of treatment, necrotic lesion development was evaluated by removing the petri dish cover, placing a glass disc on the leaf sections to flatten them into a single plane, and photographing each dish with its leaf sections. Each acquired image was then analyzed with the software ImageJ to measure the necrotic area caused by the solution.

Statistical Analyses
Statistical analyses were carried out using the GraphPad Prism 8 software. Data were represented as the mean ± standard deviation and analyzed for statistical significance using ordinary one-way or two-ways analysis of variance (ANOVA) and multiple comparisons. For all test, P < 0.5 was considered to indicate a statistically significant difference.

Conclusions
A novel efficient synthetic strategy for obtainment of (±)-3-deoxyradicinin (2) has been developed. This synthetic methodology not only is more efficient than those previously reported in the literature, but it also shows higher versatility towards the introduction of different side-chains at both C-7 and C-2 as shown with the synthesis of the analogue (±)-8. The obtained compound (±)-2 shows phytotoxicity against the grass-weed buffelgrass comparable to that of the natural phytotoxin radicinin (1). Therefore, (±)-2 can constitute a more practical synthetic alternative to 1 as bioherbicide for buffelgrass control. In fact, while 1 is available only in very small amount by fungal culture, (±)-2 can be obtained in much larger quantities by organic synthesis thus allowing to carry out extensive studies on its phytotoxic activity and even open field tests. These results are then of interest for both synthetic and natural products chemistry as well as for the nowadays forefront research on bioherbicides for weed control in agriculture.