Improved synthesis of 6-[(ethylthio)methyl]-1H-indazole

Introduction

Although natural compounds bearing an indazole structure are rare [1], the chemistry of synthetic indazoles is well developed. A number of indazole derivatives possess important pharmaceutical activities, and they are dopamine antagonists, anti-inflammatory, analgesic, or antypyretic agents [2, 3]. Other derivatives have herbicidal, bactericidal, and fungicidal properties and are also important in the dyes industry [4–6]. Indazoles and the analogous pyrazoles have also been used as ligands in coordination chemistry since the discovery of the tris(pyrazolyl)borate ligand by Trofimenko in the late 1960s [7]. This family of tridentate ligands, also known as scorpionate ligands, has been developed in increasingly wide fields of chemistry, ranging from bio-inorganic, organometallic, and coordination chemistry to catalysis and material sciences [8]. Modification of the functional groups connected to the pyrazolyl moiety has been extensively studied in order to control or modify the steric and electronic environment surrounding the metal center. However, the development of scorpionate ligands using indazole analogues is very limited [9]. Only a few examples of tripodal ligands based on indazole subunits are known. A tris(indazolyl)borate ligand has been recently prepared with a perfluorinated indazole [10] and 6-[(ethylthio)methyl]indazole has been used as building block for the tripodal anchoring group of a surface-mounted molecular motor (Figure 1) [11–14].

Figure 1

Structure of the molecular motor using 6-(ethylthio)methyl-tris(indazolyl)borate ligand as tripodal anchoring group [11].

In this communication, we present the improved synthesis of an indazole 5 functionalized with a thioether group at the 6-position. The corresponding tris(indazolyl)borate ligand was successfully used in a family of ruthenium-based single molecular rotary motors [11]. This tripodal ligand is also of broad interest to anchor various metallic complexes on metallic surfaces through its sulfur atoms. Such a bifunctional molecule that combines coordinating sites and anchoring groups for the covalent attachment of rigid complexes onto surfaces is an active target of research because it is the first step toward the creation of molecular devices. In particular, the mobility of a molecule can be efficiently restricted by attaching it to a surface through three points of attachment, i.e., using a tripodal ligand that prevents translation [15–17].

The thioether-functionalized tris(indazolyl)borate ligand derived from 5 suffers from its poor accessibility [18] because six steps are needed from the commercially available 3-amino-4-methylbenzoic acid and the best overall yield is as low as 15% (Scheme 1). This six-step sequence also suffers from its poor reproducibility, with the overall yield sometimes dropping below 5% for no apparent reason. In the new synthetic sequence presented here, the overall yield is improved by a factor of 3 and the reproducibility is systematic.

Scheme 1

Original synthesis of 6-(ethylthio)methyl indazole [18].

Reagents and conditions: (A) SOCl₂, EtOH, reflux, N₂, 16 h, 98%; (B) Ac₂O, AcOK, isopentyl nitrite, reflux, N₂, 16 h, then (C) HCl, 60°C, 1 h, 64% (overall yield for steps b and c); (D) LiAlH₄, THF, 0°C, N₂, 2 h, 93%; (E) MsCl, Et₃N, EtOAc, 0°C, N₂, 12 h, then (F) EtSH, KOH 1.5 eq., reflux, 1 h, EtOH-THF (1:1), N₂, 25% (overall yield for steps E and F).

In a first step, 3-amino-4-methylbenzoic acid was quasi-quantitatively esterified by reaction with ethanol in the presence of thionyl chloride. Conversion of 1 to the corresponding indazole was performed by reaction with isopentylnitrite as described by Jacobson [19, 20]. 6-(Ethoxycarbonyl)-1H-indazole (2) was obtained with a 64% yield. The ester function was then reduced with LiAlH₄ with a yield of 93%, and the resulting alcohol 3 was activated by treatment with mesyl chloride in the presence of triethylamine. The nucleophilic substitution of the mesylate product by the in situ generated potassium ethanethiolate gave the target thioether 5 but with a low and poorly reproducible yield of 10%–25%. This uncontrolled reactivity arises from the low and erratic conversion of 6-(hydroxymethyl)indazole to its O-mesylate derivative due to the competitive reaction of mesyl chloride with the NH function of the indazole system.

In this work, it was sought to circumvent this inefficient step by employing a different strategy. In the first instance, compound 3 was allowed to react with ⁿBuLi followed by the reaction with two equivalents of mesyl chloride with the idea to substitute both the oxygen of the alcohol function and the nitrogen of the indazole moiety. Unfortunately, a 1:1 mixture of the N-mesylated compound (Scheme 2, 3·NMsOH) with another N-mesylated compound with the substitution of the OH group by a chloride atom (3·NMsCl) was obtained. This pattern had been described in the literature for similar substrates [21]. The presence of the latter product can be explained by the formation of the dimesylated species followed by the nucleophilic substitution of the O-mesylate by the chloride atom coming from mesyl chloride.

Scheme 2

Mesylation reaction of 3.

Changing the base and the mesylating agent changes the proportions of the products obtained as described in the text.

Replacement of ⁿBuLi by NaH resulted in a 1:3 mixture of the dimesylated species (3·NMsOMs) with the mono N-mesylated compound (3·NMsOH). Surprisingly, this test reaction did not give any trace of the chloride derivative previously observed. These proportions were obtained from ¹H-NMR spectra for which the integrations of the different CH₂ groups in the 4- to 5-ppm region allowed us to characterize each molecule. Unfortunately, the instability of mesylates on silica gel and alumina prevented their separation by column chromatography.

Reaction of 3 with NaH followed by treatment of the mixture with two equivalents of dimesyl anhydride (MsOMs), a chloride-free mesylating agent, resulted in the formation of the dimesylated compound 3·NMsOMs quantitatively. Reaction with different nucleophiles including not only ethanethiol but also poorly reactive 4-phenylphenol very efficiently displaces the O-mesylate group; unfortunately, the final deprotection of the N-mesyl group by treatment with HCl is not selective, and the different groups attached at the benzylic position were lost, yielding 3 as the main product. Attempted deprotection under basic conditions such as ammonia in methanol at room temperature gave similar results.

As an alternative strategy, bromination of alcohol 3 was attempted (Scheme 3). Reaction with hydrobromic acid in hot acetic acid [22] gave the brominated derivative 6 with a yield of 89%. As shown on Scheme 3, this step is not only more efficient than mesylation, but the subsequent nucleophilic displacement of the bromide by the thiolate nucleophile is also very efficient. Reaction of 6 with ethanethiolate generated in situ from ethanethiol in the presence of DBU gave the target compound 5 with a 70% isolated yield. The choice of the base is also an important parameter because reaction in the presence of NaH gave only 56% of conversion and the use of K₂CO₃ or Cs₂CO₃ gave no conversion at all.

Scheme 3

Improved synthesis of 6-[(ethylthio)methyl]-1H-indazole via the formation of 6-(bromomethyl)-1H-indazole.

Reagents and conditions: (A) 33% HBr in acetic acid, AcOH, 120°C, N₂, 1 h, 89%; (B) EtSH, DBU, THF, reflux, 1 h, N₂, 70%.

In summary, the overall yield for the six-step synthesis of 6-[(ethylthio)methyl]indazole (5) was improved from 5%–15% to 42%. This threefold increase in the overall isolated yield and improved reproducibility of 5 gives this precursor to the corresponding tris(indazolyl)borate ligand a growing interest for surface-immobilized metal complexes.

Experimental details