Introduction

Indole has been recognized as a privileged structure, ranking 13th among the most frequently used 351 ring systems found in marketed drugs1,2. The substituted indoles are useful not only as medicines but also as agrochemicals and functional materials3,4,5. A number of substituted indoles have been synthesized via nucleophilic substitutions at the α-position of the indole. The moderately reactive (1H-indol-3-yl)methyl electrophiles 1, such as methylated gramine 1a (Y = +NMe3)6,7 and aryl sulfone 1b (Y = SO2Ar)8,9 with or without additives (Fig. 1a) have been used for the nucleophilic substitutions at the 3’-position. However, this approach suffers from the requirement of high temperatures and long reaction times, as well as limited substrate scope6,7,8,9. In addition, highly electrophilic vinyl iminium intermediate 2 is gradually generated; therefore, the resultant coexistence of substrate 1 and electrophile 2 leads to undesired dimerization and/or oligomerization6,7,8,9,10,11,12,13. Although the rapid generation of 2 from highly reactive (1H-indol-3-yl)methyl electrophiles 6 containing good leaving groups (Y = halogen, OSO2R) can potentially avoid the undesired dimerization/oligomerization, they have been rarely used in the substitution reactions at the 3’-position of the electron-rich indoles (Fig. 1b). The preparation of the highly reactive 6 is difficult due to concomitant undesired dimerization and/or oligomerization14. In fact, Eryshev et al. reported that (1H-indol-3-yl)methyl bromide (6a) could not be prepared via the bromination of indole-3-methanol using PBr3, most likely due to the instability of alkyl bromide 6a15. Cook et al. also reported an unsuccessful attempt toward the preparation of 6a via the bromination of 3-methyl-1H-indole using azobis(isobutyronitrile) and N-bromosuccinimide16. Rhee et al. also reported difficulty in isolating unprotected indole 6 with a good leaving group at the 3’-position17. Moreover, even in situ generation of 6 and its use in the following nucleophilic substitution without purification did not afford desired 317. They concluded that the preparation of (1H-indol-3-yl)methyl electrophile 6 is difficult unless 6 contains the electron-withdrawing group17.

Fig. 1: Nucleophilic substitutions at the 3’-position of substituted indoles.
figure 1

a Nucleophilic substitution using methylated gramine 1a and aryl sulfone 1b. b Nucleophilic substitution using (1H-indol-3-yl)methyl halide 6. c This study: rapid mild generation and nucleophilic substitution in a microflow reactor.

Full size image

Despite these reports, SciFinder search revealed five papers that reported the synthesis of (1H-indol-3-yl)methyl electrophile 6 (Table 1)15,18,19,20,21. Two among these also report spectral data for 619,21. However, based on the previous reports for similar compounds, some discrepancies were identified between the reported and expected data. Three of the five studies involved elemental analysis15, TLC analysis18, or no analytical data20. Besides the aforementioned five studies, five other studies reported nucleophilic substitution of 6, although the preparation procedure for 6 was not described22,23,24,25,26. As far as we could ascertain, the preparation of highly electrophilic (1H-indol-3-yl)methyl iodide (6c) and sulfonate 6d have not been reported. Based on SciFinder search, the alkyl bromide 6a is commercially available in a limited number of countries. However, despite our efforts to import 6a through trading companies, we were unable to find out a supplier. The alkyl halides 6b and 6c, and alkyl sulfonate 6d are not commercially available. Hence, it is an important pursuit to resolve the abovementioned contradiction in the synthesis of 6a and 6b.

Table 1 Previous reports for the synthesis and the nucleophilic substitution of (1H-indol-3-yl)methyl halide 6.
Full size table

Microflow technologies have garnered much attention owing to their advantages over conventional batch synthesis approaches27,28,29,30,31,32,33,34. In particular, the microflow reactor allows precise control of the reaction time and temperature, thus enabling the use of highly reactive and unstable species35,36,37,38,39. We have developed various efficient synthetic approaches using microflow technologies based on the rapid generation and reaction of unstable and highly active species before side reactions occur40,41,42. We anticipated that we could achieve nucleophilic substitution at the 3’-position of indoles while suppressing dimerization/oligomerization by employing microflow technology.

Herein, we report the examination of all the previously reported preparations of (1H-indol-3-yl)methyl halides 6a and 6b, and the structural revision of two reported indole derivatives. We also developed a highly versatile nucleophilic substitution at the 3’-position of indoles using microflow technologies (Fig. 1c).

Results and Discussion

Synthesis of (1H-indol-3-yl)methyl halides

We examined all three previously reported syntheses of 6a and its analogues. The first report by Eryshev et al. includes the preparation of 6a via the Borodin–Hunsdiecker reaction15 (Fig. 2a). The reported yield of the alkyl bromide 6a was low (28%), and the structural confirmation was performed only by elemental analysis15. We examined the following reported procedure several times (for details, see Supplementary Information pages S5–S6). Red mercury (II) oxide was added to a solution of indole-3-acetic acid (8) in carbon tetrachloride and acetone. After being stirred at 55 °C for 10 min, bromine was added dropwise. As a result, a highly lachrymatory compound was generated, but the desired alkyl bromide 6a was not detected by 1H NMR spectroscopy and a large amount of red precipitate was generated. A detailed structural analysis of the product could not be performed because it was an inseparable mixture of many products. We speculated that the desired 6a formed in situ (it probably had lachrymatory nature), but the undesired polymerization of indole analogues underwent concomitantly that led to its precipitation. The second report by Scanlan et al. includes the synthesis of 6a by bromination with TMSCl and LiBr18. They reported only TLC analysis for structural confirmation of 6a (Fig. 2b). We examined the reported procedure several times (for details, see Supplementary Information pages S6–S7). However, the desired 6a could not be detected using 1H NMR spectroscopy and a large amount of red precipitate was generated most likely due to the undesired oligomerization/polymerization. The third report by Mekonnen Sanka et al. 19 includes the synthesis of 6a by the bromination of 7a using PBr3 (Fig. 2c), which has been reported to be impossible by Eryshev et al.15. The 1H NMR spectrum for 7a in their report was obviously different from those in another report43. Therefore, we purchased 7a and measured its 1H NMR spectrum (for details, see Supplementary Information page S7 and Supplementary Figure 11 in Supplementary Information page S33). Our observed spectrum was not consistent with that reported by Mekonnen Sanka et al.19, whereas, our observed spectrum was well consistent with those in another previous report43. Mekonnen Sanka et al. also reported the nucleophilic substitution of alkyl bromide 6a with an aryl piperazine 10 to obtain 3a19. Although we synthesized 3a using our developed method described later, the spectral data of 3a were not consistent with those reported by Mekonnen Sanka et al.19, but were consistent with those reported by Akkoc et al.44. These results suggest that the structures 6a and 7a reported by Mekonnen Sanka et al. are most likely incorrect.

Fig. 2: Examination of previously reported syntheses of (1H-indol-3-yl)methyl bromide (6a) and its derivatizations.
figure 2

a Preparation of 6a via the Borodin–Hunsdiecker reaction reported by Eryshev et al. b Preparation of 6a via bromination with TMSCl and LiBr reported by Scanlan et al. c Preparation of 6a via bromination with PBr3 reported by Mekonnen Sanka et al.

Full size image

Two papers have reported the syntheses of (1H-indol-3-yl)methyl chloride analogues. Degterev et al.20 reported that the reaction of 7-chloroindole (11) with N,N-dimethylmethyleneiminium chloride (12) afforded alkyl chloride 13, although no spectral data for 13 was reported. In contrast, the reaction of an indole analogue with N,N-dimethylmethyleneiminium salt generally provides a gramine analogue45. In fact, Faul et al. reported that the same reaction of 11 and 12 afforded 14 and not 1346. We examined the reaction according to the procedure reported by Degterev et al. and obtained a gramine analogue 14. However, alkyl chloride 13 could not be detected (Fig. 3a). The spectral data of our obtained 14 were consistent with those reported by Faul et al.46. The synthesis of 16 via chlorination of 7a using 15 (Fig. 3b) was reported by Jiang et al.21. We carried out the reaction according to the procedure reported by Jiang et al. The reaction proceeded well, and the spectral data of the obtained compound was consistent with those reported by Jiang et al.21. However, the structural determination by Jiang et al. had some concerns, especially in the 13C NMR data. (1) A signal at 174.1 ppm was observed, corresponding to a carbonyl carbon, although the proposed structure of 16 does not have a carbonyl group. (2) Six signals in the aromatic region (140.8–111.1 ppm), and two signals in the aliphatic region (63.2 and 45.9 ppm) were observed, although the proposed structure 16 has eight aromatic and one aliphatic carbons. Thus, we converted the obtained product to the corresponding benzamide 19 using p-bromobenzoyl chloride (18), and its structure was unambiguously determined by X-ray crystallography47. The analysis indicated that benzamide 19 has an oxyindole structure. Thus, we conclude that the chemical structure of the product from the reaction between 7a and 15 is not 16, but 17, which is consistent with the 13C NMR data.

Fig. 3: Examination of the reported syntheses of (1H-indol-3-yl)methyl chloride analogues.
figure 3

a Preparation of 13 using iminium chloride 12 reported by Degterev et al. b Preparation of 16 using 15 reported by Jiang et al.

Full size image

Although we examined all the previously reported syntheses of (1H-indol-3-yl)methyl halides and their reactions, we could not reproduce the reported results and could not confirm the generation of (1H-indol-3-yl)methyl halides. Additionally, our examinations unexpectedly led to revisions of previously reported structures of indole derivatives. These results clearly indicate the importance of the development of reliable and practical nucleophilic substitution approaches using highly active (1H-indol-3-yl)methyl electrophile. To achieve this goal, we examined the in situ generation of the (1H-indol-3-yl)methyl electrophile via the halogenation/sulfonylation of stable and readily available indole-3-methanol (7a) and its subsequent nucleophilic substitution with NaN3. The reaction time and temperature were precisely controlled using microflow technologies35,36,37,38,39.

Development of microflow nucleophilic substitution

First, we examined the activation reagents (Table 2, entries 1–11). When trivalent phosphorus reagents PBr3 and PCl3 were employed, the azidation proceeded smoothly in yields greater than 50% (entries 1 and 2). By contrast, none of the other reagents afforded satisfactory results (entries 3–11). Although quantitative data for the electrophilicity of the activating reagents are not available, we speculated that the electrophilicity of the reagents is important for obtaining good yields. On one hand, the use of highly reactive trivalent phosphorus electrophiles afforded relatively high yields (entries 1 and 2). On the other hand, the use of less reactive pentavalent phosphorus and carbon electrophiles, including POCl3, AcBr, AcCl, and Ac2O, resulted in the recovery of a large amount of alcohol 7a (entries 3–6) with concomitant generation of insoluble solids (entries 4 and 6). The use of sulfur electrophiles with medium levels of reactivity led to the recovery of alcohol 7a and/or the generation of insoluble solids probably due to undesired dimerization/oligomerization (entries 7–11). Then, we examined the activation time (Table 2, entries 1 and 12–14). Extension of activation time dramatically reduced the yield, along with the generation of insoluble solids (entry 1 vs. entries 12–14). When activation was carried out for 0.5 s, azide 3b was obtained in only 20% yield and 7a was not recovered (entry 14). We could not examine reaction times shorter than 0.02 s because neither the length nor the inner diameter of the reaction tube could be reduced further (for details, see Supplementary Information page S13). The use of a reduced quantity of PBr3 (0.35 equiv.) and activation time (0.02 s) improved the yield of 3b (84%, entry 15). When the reaction was performed at 0 °C, the yield was somewhat low (74%) (entry 16). When the highly electrophilic intermediate is gradually generated at lower temperatures, the substrate coexists with the electrophile for a longer time and causes undesired oligomerization. The key to suppressing the side reaction is that the electrophile is rapidly generated at a higher temperature and immediately used for the next reaction (details, see Supplementary Information page S15). To suppress undesired intermolecular reactions, such as dimerization/oligomerization of (1H-indol-3-yl)methyl electrophile, diluted conditions (0.0500 M for a solution of 7a) were examined (entry 17). The desired product 3b was obtained in excellent yields with sufficient reproducibility (93 ± 2%). It should be noted that the desired product was not detected and the precipitates that appear to be dimer/oligomers were generated in three independent experiments under batch conditions (entry 18), although the reaction mixture was vigorously mixed (1000 rpm) during the experiment (caution, care should be taken when performing the reaction under batch conditions because the reaction is exothermic and rapidly generates dangerous gas such as HBr). These results clearly indicated the instability of the highly active (1H-indol-3-yl)methyl electrophile. The rapid (0.02 s) and mild (25 °C) in situ generation of the extremely reactive (1H-indol-3-yl)methyl electrophile enabled rapid (0.1 s) and mild (25 °C) nucleophilic substitution.

Table 2 Optimization for in situ generation of (1H-indol-3-yl)methyl electrophile from 7a and its nucleophilic substitution with NaN3 in a microflow reactor.
Full size table

The substrate scope of the developed approach was examined (Fig. 4), after optimizing the base used, amount of PBr3, and temperature (for details, see Supplementary Information pages S14–S17). Hydrophilic nucleophiles were used in the form of aqueous solutions (Method A), whereas hydrophobic nucleophiles were used in CH3CN solutions (Method B). The use of secondary amines as nucleophiles afforded the desired tertiary amines 3a and 3c–3e in good yields (66–86% yields). Compound 3a was successfully synthesized in high yield (86%). The products 3f and 3g containing an electron-donating and an electron-withdrawing group, respectively, at the 5-position of the indole ring were obtained in high yields (81% and 83%, respectively). The product 3h with a bulky phenyl group at the 2-position of the indole ring was obtained without a significant decrease in the yield (71%). The use of amino-acid-derived secondary amines as N-nucleophiles afforded products 3i and 3j high yields (84% and 83%, respectively). When primary alkylamines were used as nucleophiles, tertiary amines 3k and 3l were obtained in 49% and 68% yields, respectively, via double indolylmethylation. When 2-bromoaniline with low nucleophilicity was employed, single indolylmethylation occurred to afford 3m in 50% yield. The use of S-nucleophiles, including sodium benzenesulfonate, alkyl thiol, and aryl thiol, afforded the desired products 3n–3q in good-to-excellent yields (77%–quant.). The use of Meldrum’s acid as C-nucleophile afforded the double indolylmethylated product 3o in good yield (86%). The indole analogues 3a–3r were soluble in the commonly used organic solvent such as ethyl acetate, dichloromethane, acetonitrile, and chloroform. As described above, nucleophilic substitutions of 1 (Fig. 1a) require high temperatures and long reaction times and involve undesired dimerization/oligomerization. By contrast, the developed approach enables rapid nucleophilic substitutions with a variety of N-, S-, and C-nucleophiles.

Fig. 4: Scope of the developed microflow nucleophilic substitution.
figure 4

Unless stated otherwise, a solution of substrate 7 (0.0500 M, 1.00 equiv.) in CH3CN, a solution of PBr3 (0.00875 M, 0.350 equiv.) in CH3CN, and a solution of nucleophile (0.0375 M, 3.00 equiv.) and NEt3 (0.0375 M, 3.00 equiv.) were introduced at 25 °C with syringe pumps A, B, and C, respectively. The nucleophile is described in green. Hydrophilic nucleophiles were used in the form of aqueous solutions (Method A), whereas hydrophobic nucleophiles were used in CH3CN solutions (Method B).

Full size image

A plausible reaction mechanism is shown in Fig. 5. It was reported that coupling between PBr3 and alcohol 7a is fast and the corresponding phosphite 21 is generated through intermediate 2048,49,50. Gerrard reported that very fast proton trapping by 21 is key for enhancing the electrophilicity of the phosphorus center in 2148. It is conceivable that the elimination of the extremely electrophilic (1H-indol-3-yl)methyl cation species 2 from 22 occurs rapidly. Previous studies have indicated that the second and third reactions are slower than the first reaction from 2248. However, these second and third elimination reactions of 2 via 23 may be facilitated by the electron-donating ability of the indole ring. Reportedly, proton trapping by the intermediate 20, which leads to the electrophilic activation of the phosphorus center in 20, is rather slow48. However, we could not exclude the possibility of generating 2 via the activation of 20. In fact, Hudson reported that dealkylations of ROPBr2 and (RO)2PBr like 20 that afford 2 are also possible50. We speculated that the entire process shown below is rapid and the rapid generation of 2 in a microflow reactor avoided any undesired dimerization/oligomerization. This is a significant advantage over the reported approach based on the gradual generation of 2 from 1 (Fig. 1a). We attempted to detect the generation of 2 by in-line IR and to compare in-line IR spectra with their predictions by DFT calculations. However, the elucidation of the reaction intermediate was difficult because the characteristic IR absorption of indole was easily changed due to the influence of the association state51. We also tried to detect the generation of 2 by reactions in an NMR tube; however, the efforts were futile by its instability (details, see Supplementary Information page S18–S22).

Fig. 5: Plausible mechanism.
figure 5

Plausible mechanism for rapid and mild generation and nucleophilic substitution of extremely active (1H-indol-3-yl)methyl electrophile 2.

Full size image

In conclusion, we examined all the previously reported syntheses of (1H-indol-3-yl)methyl halide 6. We could not reproduce the reported syntheses and revised several reported structures of indole analogues. To develop a reliable and practical synthetic approach for nucleophilic substitutions at the α-position of the indole ring, we examined the rapid (0.02 s) and mild (25 °C) generation of the highly reactive (1H-indol-3-yl)methyl electrophile that enabled the rapid (0.1 s) and mild (25 °C) nucleophilic substitution. Eighteen unprotected indole analogues were successfully synthesized using the developed microflow nucleophilic substitution with various N-, S-, and C-nucleophiles. Extending the residence time of the electrophile (from 0.02 to 0.5 s) dramatically decreased the yield from 77 to 20%. Moreover, comparable batch conditions resulted in a 0% yield. These results clearly indicate the instability of the highly active (1H-indol-3-yl)methyl electrophile. Microflow technology realized the in situ preparation and use of such a highly unstable species for nucleophilic substitutions. This study offers a solution for a general and important problem in nucleophilic substitution at the α-position of electron-rich aromatic rings.

Methods

General techniques

See Supplementary Information (page S3).

Synthesis of indole analogues

See Supplementary Information (pages S5–S10).

Optimization of nucleophilic substitution

See Supplementary Information (pages S10–S18).

Typical procedure and compound characterization data

See Supplementary Information (pages S22–S32).

NMR chart

See Supplementary Figures 11–56.