Introduction

The presence of weed infestations exerts a high strain on food production around the globe by depleting resources for the crops and facilitating the transmission of diseases . Although herbicides remain the most effective solution for weed control due to the associated efficiency and simplicity, they face multiple challenges, such as the emergence and growth of resistant weed populations. It is therefore essential that crop protection research acts rapidly to provide farmers with new solutions that enable them to fight back against resistant weed species . Nevertheless, discovering novel and commercially viable modes of action within the timeframe needed to significantly impact the control of resistant weeds is a demanding task. Thus, we analyzed several herbicidal modes of action with emphasis on the structural diversity of small-molecule ligands. In this context, acyl-acyl carrier protein (acyl-ACP) thioesterase inhibitors have shown a remarkable variability. Fatty acid thioesterase (FAT) enzymes represent a family of proteins exclusively found in higher plants. They mediate the release of fatty acids from the plastids to the endoplasmic reticulum, where they are utilized for the synthesis of acyl lipids that are essential components for various physiological and defensive processes . As this enzyme target does not exist in other kingdoms, structure–activity relationship (SAR) studies on selective inhibitors reduce the prevalence of undesired effects, such as toxicity in mammals . Despite being employed in the field for over three decades, the mode of action of preemergence herbicide cinmethylin (1, Scheme 1) has remained unknown until 2018. At that time, the binding affinity to enzyme targets, e.g., acyl-ACP thioesterases, belonging to the protein family of FATs, was demonstrated by using co-crystallization, fluorescence-based thermal shift assays, and chemoproteomics techniques . Likewise, methiozolin (2) is a recently assigned FAT inhibitor that has shown good results in selectively controlling grass weeds in both cool and warm seasons . Recently, it has been shown that several herbicides bearing a gem-dimethylbenzylamide motif, e.g., cumyluron (3a) and oxaziclomefone (3b), previously exhibiting an unknown mode of action, control weeds due to the inhibition of FAT . In search for further chemical entities that can control resistant weed species via the inhibition of FAT, we were interested in exploring a compound class containing a 1,8-naphthyridine core that was first reported by BASF, e.g., compound 4 .

Scheme 1: Selected known inhibitors 1–3 of acyl-ACP thioesterases (belonging to the protein family of FATs) and new lead structures 4–7a.

In contrast to bicyclic cinmethylin (1) and methiozolin (2), substituted 1,8-naphthyridine 4 does not contain any stereocenters but still displays promising efficacy against grass weeds. Further considering the rather low molecular weight (220 g/mol) and structural simplicity, compound 4 is a highly attractive initial lead structure with ample space for structural variations. By formally replacing one pyridine moiety of 1,8-naphthyridine 4 by a five-membered thiazole unit, we have identified thiazolo[4,5-b]pyridine 5 as a strong inhibitor of acyl-ACP thioesterase, which has further been confirmed via an X-ray co-crystal structure . Additionally, greenhouse trials have shown that thiazolopyridine 5 and a large number of closely related analogues display excellent control of grass weed species in preemergence applications . Independently, researchers at Syngenta have shown that the pyridine unit in the 1,8-naphthyridine scaffold can also be formally substituted by an isothiazole group, as can be seen in isothiazolo[3,4-b]pyridine 6 .

Thus, several bicyclic heteroaromatic motifs containing two nitrogen atoms serve as structural surrogates of cinmethylin (1), bearing a substituted 7-oxabicyclo[2.2.1]heptane scaffold . Based on the findings outlined above and based on other plant-specific modes of action, it is plausible that FAT inhibitors encompass a broader range of structural motifs . Herein, we present our approach to complement heteroaromatic lead structures 4–6 by introducing a nonaromatic motif via preparation of the novel 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7.

Results and Discussion

Although the 2,3-dihydro[1,3]thiazolo[4,5-b]pyridine scaffold looks relatively simple at a first glance, it displays a very different reactivity compared to the parent naphthyridine series. Likewise, 1,8-naphthyridines are easily accessed in high yield and on a multigram scale via Friedländer synthesis . This was in clear contrast to the intermediate thiazolo[4,5-b]pyridines and the desired 2,3-dihydro[1,3]thiazolo[4,5-b]pyridine that we wanted to access, with approaches to prepare the thiazolo[4,5-b]pyridines using the Friedländer synthesis often being met with failure or disappointingly low product yield . We thus particularly emphasized on upscaling, facile workup, and a robust yield for each step. This was due to the potential need for the preparation of multigram quantities of the most active compounds for advanced biological testing. Pleasingly, our four-step approach using a potassium O-ethyl dithiocarbonate-mediated formation of thio intermediates 11a–c (thiol–thione tautomers) with subsequent sulfur removal using iron powder in acetic acid proceeded smoothly to afford thiazolopyridines 12a–c in good yield. This allowed us to circumvent the previously employed alkylation–oxidation–reduction sequences (Scheme 2) . Thereupon, we recognized that we could introduce two halogen atoms in the halogenation step and carry one through to the end of the synthetic route, which enabled us to introduce a methyl substituent in this position. Likewise, methyl-substituted thiazolo[4,5-b]pyridines 5, 15a, and 15c were synthesized using an optimized Suzuki coupling and served together with compounds 12a–c as key intermediates to explore different reagents and conditions to prepare 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7a–c and 13a–c via a late-stage reduction (Scheme 2 and Table 1).

Scheme 2: Preparation of 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7a–c and 13a–c via iron-mediated sulfur removal and subsequent reduction. dppf = 1,1'-bis(diphenylphosphino)ferrocene, SPhos = 2-dicyclohexylphosphino-2',6'-dimethoxybiphenyl, APDTC = ammonium pyrrolidinedithiocarbamate.

Table 1: Preparation of 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7a–c via reduction of the thiazole moiety: optimization of the reaction conditions.a

entry R1 reagents solvent T, °C T, h 7a–c, %b 17a–c, %b 18a–c, %b 1 CH3 H2, Pd/C, 2–20 bar MeOH 20 4 — — — 2 F H2, Pd/C, 4–35 bar MeOH 20 6 — — — 3 F N2H4, Pd/C EtOH 80 6 — — — 4 F B2(OH)4 H2O 80 6 — — — 5 F Bu4NBH4 THF 20 5 — — — 6 F Bu4NBH4 THF 70 5 5 — 48 7 H Bu4NBH4 dioxane 80 5 8 — 59 8 F NaCNBH3, AcOH MeOH 60 5 4 — 15 9 F SiCl3H toluene 110 3 — — 35 10 CH3 SiEt3H dioxane 80 3 — — 43 11 F NH3-BH3 toluene 80 4 5 4 12 12 F NH3⋅BH3, B(C6F5)3 toluene 80 4 16 33 18 13 CH3 NH3⋅BH3, B(C6F5)3 toluene 80 7 8 29 30 14 CH3 NH3⋅BH3, B(C6F5)3 toluene 45 6 4 51 3 15 H NH3⋅BH3, B(C6F5)3, HCO2H toluene 45 7 64c — 2 16 F NH3⋅BH3, B(C6F5)3, HCO2H toluene 45 5 66c — 5 17 CH3 NH3⋅BH3, B(C6F5)3,
HCO2H toluene 45 7 59c — 4

aAll reactions in the optimization phase were carried out using 0.2 mmol of 5, 15a, and 15c, respectively. bDetermined by analytical HPLC. cIsolated yield after silica gel column chromatography.

Whilst several synthetic approaches towards 2,3-dihydro-1,3-benzothiazoles involving the hydrogenation of 1,3-benzothiazoles have been described , the corresponding preparation of 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines remained unexplored to our great surprise. Thus, we investigated the conversion of [1,3]thiazolo[4,5-b]pyridines 5 (R1 = F), 15a (R1 = H), and 15c (R1 = CH3) into 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7a–c thoroughly, with the aim to establish a practicable and robust synthetic route enabling us to carry out a broad SAR study. Initial attempts to prepare 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7a and 7b using hydrogen and palladium on charcoal under elevated pressure did not show any conversion of the starting material (Table 1, entries 1 and 2). Correspondingly, [1,3]thiazolo[4,5-b]pyridine 5 remained unchanged upon application of methods that had been successfully utilized in the hydrogenation of 1,3-benzothiazoles, involving diboronic acid or hydrazine hydrate as key reagents in protic solvents at an elevated temperature (Table 1, entries 3 and 4). Whilst tetrabutylammonium borohydride at room temperature did not lead to a conversion of the starting material 5 either, a trace amount of desired 2,3-dihydro[1,3]thiazolo[4,5-b]pyridine 7b was formed at elevated temperature, accompanied by disulfide 18b as the main product (Table 1, entries 5 and 6). This result indicated that borohydride reagents were able to activate the thiazole moiety in [1,3]thiazolo[4,5-b]pyridines, leaving the pyridine unit unchanged. While sodium cyanoborohydride afforded a comparable result, albeit with lower conversion, the use of silane reagents at elevated temperature led to the cleavage of the thiazole ring, furnishing disulfides 18b and 18c exclusively (Table 1, entries 7–10). Interestingly, the reaction of 5 with ammonia borane at elevated temperature in toluene furnished three reaction products with a low yield since 2,3-dihydro[1,3]thiazolo[4,5-b]pyridine 7b was formed together with disulfide 18b and aminoborane 17b (Table 1, entry 11). We thus evaluated B(C6F5)3 as a nonmetallic catalyst to activate ammonia borane in the reductive hydrogenation of the C=N-bond in [1,3]thiazolo[4,5-b]pyridines 5 and 15c. In line with reports on the hydrogenation of quinolines and indoles , pyridines , and imines , the reactions of 5 and 15c with ammonia borane (3 equiv) in the presence of a catalytic amount of B(C6F5)3 in toluene as an aprotic solvent at a temperature of 80 °C afforded aminoboranes 17b (R1 = F) and 17c (R1 = CH3) as main products along with desired target compounds 7b and 7c (Table 1, entries 12 and 13). However, disulfides were still formed as side products in a significant amount. Pleasingly, the undesired thiazole cleavage could successfully be minimized by reducing the reaction temperature to 45 °C, furnishing aminoborane 17c in 51% yield (Table 1, entry 14). The borane group could be cleaved off easily by the subsequent treatment of 17c with formic acid in acetonitrile, affording 2,3-dihydro[1,3]thiazolo[4,5-b]pyridine 7c as the only reaction product. By applying this optimized two-step procedure to [1,3]thiazolo[4,5-b]pyridines 5, 15a, and 15c, the desired 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7a–c were prepared in good yield (Table 1, entries 15–17, 59–66% isolated yield), enabling us to investigate the biological profiles as well as the reactions with acyl chlorides to form amides 16a–f (Scheme 2) . These acylations proceeded cleanly under mild conditions, using the corresponding acyl chloride together with triethylamine as a suitable base in DCM.

Furthermore, we evaluated the tolerance of [1,3]thiazolo[4,5-b]pyridines 12a–c, containing a bromine atom, towards the optimized B(C6F5)3-mediated reduction. In good accordance with the results obtained for dimethylated [1,3]thiazolo[4,5-b]pyridine 15c, the corresponding aminoboranes 17d and 17e were formed when 6-bromo[1,3]thiazolo[4,5-b]pyridine 12c was treated with ammonia borane in toluene at 45 °C in the presence of a catalytic amount of B(C6F5)3 (Scheme 3). Whilst diphenyl analogue 17f was isolated as a side product upon arylation with B(C6F5)3, debromination was only observed in traces. Both aminoboranes 17d and 17e were then cleaved separately in clean conversions using formic acid to afford the desired substituted 6-bromo-5-(2-tolyl)-2,3-dihydrothiazolo[4,5-b]pyridine (13c, 54% combined isolated yield). As shown for N-acylated target compounds 16a–f, the acylation of 6-bromo-2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 13a–c, affording target compounds 14a–c, proceeded under mild conditions with a suitable acyl chloride reagent and triethylamine as base in DCM. It was not necessary to add a further base to activate the thiazoline nitrogen atom. Following the aforementioned two-step procedure, more than 15 unprecedented 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines bearing different substituents were obtained for biological and biochemical tests.

Scheme 3: Evaluation of potential side reactions in the borane-mediated preparation of 2,3-dihydro[1,3]thiazolo[4,5-b]pyridine 13c.

Converting the thiazole moiety into a thiazoline unit had a measurable impact on several physicochemical parameters, such as LogP and water solubility. Whilst thiazolo[4,5-b]pyridine 5 afforded a moderate water solubility of 49 mg/L, paired with a LogP of 2.28 (pH 2.3), the corresponding 2,3-dihydro[1,3]thiazolo[4,5-b]pyridine 7b had a higher water solubility of 173 mg/L and a lower LogP of 1.59 (pH 2.3). However, the lipophilicity of the new 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines was highly dependent on the substituents. For example, the brominated analogs 13b and 13c showed considerably higher LogP values of 2.88 (i.e., 13b) and 3.17 (i.e., 13c). We were thus curious to see how the structural change from a heteroaromatic thiazole unit to a partially saturated thiazoline moiety affected the in vitro and in vivo efficacy of the target compounds.

All compounds that were prepared to explore the SAR of substituted 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines, i.e., 13a–c, and 7a–c, the acylated analogues 14a–c and 16a–f, as well as selected aminoboranes 17d and 17e, were tested for target affinity in dedicated in vitro tests, as well as for herbicidal effects in vivo upon preemergence application to plants. Based on our experience with thiazolopyridine-based FAT inhibitors , five representative grass weeds (ALOMY, ECHCG, LOLRI, POAAN, and SETVI) were chosen as model plants to assess initial preemergence activity using a dose rate of 320 g/ha, whereas in vitro tests were carried out using FAT A, isolated from duckweed (Lemna paucicostata, LEMPA, Lp). As outlined in Table 2, entries 19 and 20, cinmethylin (1) and methiozolin (2) proved to be suitable commercial reference compounds. They showed good and broad control of grass weeds in various test systems, albeit with incomplete control of the commercially important grass weed LOLRI, paired with insufficient control of ECHCG by methiozolin (2) in our greenhouse tests. Furthermore, we used thiazolo[4,5-b]pyridine 5 as a strong internal standard to assess how modification of the thiazole moiety would affected the biological activity. It is worth noting that we emphasized investigating the preemergence efficacy as this application type is still important for cereals.

Table 2: Preemergence in vivo efficacy screening of 2,3-dihydro[1,3]thiazolo[4,5-b]pyridines 7a–c and 13a–c as well as of N-acylated analogs 14a–c and 16a–f against selected monocotyledon weeds, and binding affinity to FAT A from LEMPA.

entry compound R1 R2 pI50a ALOMYb,c ECHCGb,c POAANb,c SETVIb,c LOLRIb,c 1 7a H H 5.9 5 5 5 5 5 2 7b F H 6.1 5 5 5 5 5 3 7c CH3 H 6.3 5 5 5 5 5 4 13a H H 5.7 4 5 5 5 3 5 13b F H 7.3 5 5 5 5 5 6 13c CH3 H 4.5 1 5 1 5 1 7 14a H Ac 5.5 4 5 5 5 5 8 14b F iBtrd 4.7 3 3 3 5 2 9 14c CH3 Ac 4.3 3 3 3 5 1 10 16a H Ac 5.5 4 5 5 5 5 11 16b H iBtr 4.7 3 3 5 5 4 12 16c F Ac 5.3 4 4 5 5 4 13 16d F iBtr 4.7 3 5 5 5 4 14 16e CH3 Ac 4.1 3