Synthesis of Multisubstituted 1,2,3-Triazoles: Regioselective Formation and Reaction Mechanism

Abstract

A synthetically useful approach to functionalized triazoles is described via the reaction of β-carbonyl phosphonates and azides. 1,4- and 1,5-disubstituted and 1,4,5-trisubstituted triazoles can be regio- and chemoselectively accessed under mild conditions in good to excellent yields (31 examples, up to 99%). A mechanism is proposed that rationalizes the avoidance of the 4-phosphonate byproducts, which is aligned with crystallographic and experimental evidence.

1. Introduction

1,2,3-Triazoles, 5-membered heterocycles with three contiguous nitrogen atoms, have found applications in diverse areas such as organic synthesis, chemical biology, and material science.¹ Their reactivity and promising bioactivity are highly dependent on the pattern and nature of substituents.^1a,1b,2 Cu(I)- and Ru(II)-catalyzed 1,3-dipolar cycloadditions of azides with alkynes are conventionally used to synthesize 1,4- and 1,5-disubstituted triazoles (1,4-DTs and 1,5-DTs), respectively. To further achieve 1,4,5-trisubstituted triazoles (1,4,5-TTs), haloalkynes or aryl halides are introduced into the reaction to trap the in situ generated copper(I) triazolide intermediate for cascade coupling reactions. However, a well-designed ligand coordinating with a transition metal is paramount in minimizing the formation of undesired 1,4-DT byproducts (Scheme 1a).³

Scheme 1. Synthesis of Multisubstituted 1,2,3-Triazoles.

Highly strained cycloalkynes undergo azide–alkyne cycloadditions to furnish functionalized 1,2,3-triazoles under metal-free conditions.⁴ Further development led to the use of activated dipolarophiles such as enamines,⁵ enolates,⁶ or alkenes⁷ to accelerate the 1,3-dipolar cycloaddition with alkyl or aryl azides, as well as to synthesize a diazo imine intermediate using the Regitz diazo-transfer reagent for Dimorth cyclization. Nonetheless, the restriction of the activating groups (carbonyl or aryl moieties are required at the 4- or 5-position) limits the aforementioned protocols for generating 1,4,5-TTs (Scheme 1b,c).^5a−5d,8 Furthermore, a lack of regioselectivity is observed with asymmetrical ketones (Scheme 1d).^5f

Recently, 1,5-DTs were reported to be accessible via the reaction of β-ketophosphonates and azides mediated by KOH. However, to achieve good to excellent yields in the cyclization process, R¹ was limited to isopropyl-, t-Bu-, c-hexyl-, or aryl substituents (Scheme 1e).⁹ In contrast, with R¹ = Me, 1,5-DTs were obtained in very low yields. The innate steric substituent of R¹ was crucial to induce a syn-orientation of the alkoxy anion and the phosphonate of the triazoline intermediate to facilitate phosphonate elimination in the subsequent Horner–Wittig type process; otherwise, the formation of the undesired 4-phosphonated-1,5-trisubstituted triazole dominated owing to competing water elimination.^9b

In this study, we demonstrate that cesium carbonate serves as an effective base for promoting the reaction between β-carbonyl phosphonates and azides, yielding a variety of 1,4-/1,5-DTs and 1,4,5-TTs with good to excellent yields at room temperature (Scheme 1f). Meanwhile, X-ray crystallography and NMR spectroscopy analyses confirmed the involvement of a cesium-chelated intermediate in these [3 + 2] cycloaddition reactions, which significantly influences the chemo- and regioselectivity of the 1,2,3-triazole products.

2. Results and Discussion

The base required for the efficient generation of 1,4,5-TT 3aa from α-benzyl-β-ketophosphonate 1a and phenyl azide 2a was systematically studied at room temperature (Table 1). Amines such as DBU, TMG, TEA, and piperidine, along with K₂CO₃, were ineffective in CH₃CN (Table 1, entries 1–5). Whereas KOH in CH₃CN was suitable for preparing 1,5-DTs,^9a it resulted in a very low conversion (12%; entry 6); conversely, Cs₂CO₃ more than doubled the yield (32%; entry 7). A change to polar aprotic solvents increased the solubility of Cs₂CO₃ and significantly enhanced the efficacy of the heterogeneous reactions (entries 11–13), with DMSO being more effective (95%) than DMF (73%). While Cs₂CO₃ was observed to be superior to other alkali carbonates (entries 12–15), providing triazole product 3aa in good isolated yields (95% conversion), almost no product was obtained using strong bases such as NaHMDS or KHMDS in THF (entries 19 and 20).

Table 1. Optimization of Reaction Conditions.

entry	solvent	base	concna (%)	entry	solvent	base	concna (%)
1–4e	MeCN	DBU, TMG, TEA, or PPRb	n.r.	12	DMSO	Cs2CO3	95
				13	DMSO	K2CO3	73
5	MeCN	K2CO3	n.r.	14	DMSO	Na2CO3	6
6	MeCN	KOH	12	15	DMSO	Li2CO3	<1
7	MeCN	Cs2CO3	32	16	DMSO	CsF	72
8	THF	Cs2CO3	21	17e	DMSO	DBU	9
9	DCM	Cs2CO3	n.r.	18	DMSO	Cs2CO3c	79
10	MeOH	Cs2CO3	n.r.	19e	THF	NaHMDS	n.r.
11	DMF	Cs2CO3	73	20e	THF	KHMDS	4

^aThe conversion was determined from the crude product through the integration ratio of ¹H NMR.

^bDBU = 1,8-diazabicyclo[5.4.0]undec-7-ene, TMG = tetramethylguanidine, TEA = triethylamine, PPR = piperidine.

^c1.2 equiv of Cs₂CO₃ was used.

^eThe reaction is homogeneous.

We next focused on the scope and limitations of the reaction under the optimized conditions (Table 2). The reactions of 1a with aryl azides 2a–f at room temperature furnished the corresponding 1,4,5-TTs 3aa–af in good to excellent yields (69–96%; Table 2a). While both electron-deficient and -rich aryl azides were accommodated, the latter were less effective, but this was readily ameliorated by warming the reaction to 60 °C (cf. 3aa and 3ad). Neither benzyl azide 2h nor alkyl azide 2i underwent cyclization with phosphonate 3a, showing no reaction at room temperature for 24 h or at 60 °C for 6 h. Additionally, extended heating led to the gradual decomposition of phosphonate 3a.

Table 2. Synthesis of Diverse 1,4,5-TTs, 1,4-DTs, and 1,5-DTs^a.

^aThe reaction was completed at rt in 0.5 h unless the reaction time is indicated.

^bThe reaction was performed at 60 °C.

^cThe reaction was performed at 1.02 mmol scale based on 1a.

^dStarting material 1c was recovered in 13% yield.

In contrast to previous literature precedence,⁹ the nature of R¹ in phosphonates was no longer limited to Me; n-pentyl, t-butyl and aryl groups were tolerated and provided the desired TTs 3be–je in good yields (77–99%; Table 2b). The presence of electron-donating or -withdrawing group on the benzoyl group is favored for the reaction (3ee–je) and demonstrates good reactivity toward the aryl azide 2e. However, a bulky substituent (R¹ = t-Bu, 3ce) slowed the formation of planar Z-enolate, thereby decelerating the rate of the 1,3-dipolar cycloaddition process. The chemoselectivity of the substituents at the 4-position of triazole (R²) was systematically examined by the reaction of α-substituted-β-ketophosphonates 1k–q with p-nitrophenyl azide 2e (Table 2c). Terminal alkenyl, alkynyl, and ester groups were tolerated under the mild conditions used herein, furnishing the desired TTs 3me–pe in good to excellent yields (87–96%). The formation of 3qe and 3re (Table 2g) individually shows that the orientation of R¹ and R² can be switched by the judicious choice of reaction partners. The demand for electrophilic dipoles and the preference for electron-rich dipolarophiles indicate that the 1,3-dipolar cycloaddition between azides and β-ketophosphonates is dominated by the LUMO_dipole–HOMO_{dipolarophile} interaction.¹⁰

Previous reports indicated that fused bicyclic triazoles were obtained using highly strained cycloalkyne species^11a or extreme conditions (heat/sealed at 80 °C for an extended time).^11b−11d Under our developed conditions, the exposure of 2e to cyclic β-ketophosphonates comprising a cyclopentyl, cyclohexyl, or cycloheptyl moiety formed 4,5-fused triazoles 3se, 3te, or 3ue, respectively, in high yields (87–92%; Table 2d). Upon deprotonation, these cyclic phosphates consistently yield Z-enolates. By replacing Cs₂CO₃ with 5 M KOH (aq) as the base in a DMSO solution, the efficient synthesis of bicyclic triazole 3te was achieved, yielding 87% within 30 min. 1,4-DTs 3ve–xe (Table 2e) and 4-fluorinated 1,4,5-TTs 3ye and 3ze (Table 2f) were successfully synthesized by the reactions of 2e with appropriately substituted 2-diethoxyphosphorylacetaldehydes 1v–x and α-fluoro-β-ketophosphonates 1y–z, respectively. Although fluorinated heterocycles are crucial to the pharmaceutical and agricultural industries,¹² few reports describe the synthesis of fluorinated triazoles.¹³ Our developed synthetic method conveniently and effectively produces this important class of compounds without stringent conditions or the need for excess fluoride sources.

To gain insight into the role of the cesium cation, the reaction of 1a with 2a in DMSO-d₆ was continuously monitored by ¹H NMR (500 MHz) and ³¹P NMR (202 MHz) spectroscopy (Figure 1a–c). The signal α-H^b (δ = 3.98 ppm) clearly identifies 1a as an β-ketophosphonate. After being treated with Cs₂CO₃(s) for 1 h, the signal attributed to α-H^b disappeared, indicating 1a is readily deprotonated and converted to the cesium bound carbanionic intermediate M (Figure 1b, at 1 h).^14a Upon introduction of azide 2a, the signals of intermediate M steadily disappeared and were replaced by that of triazole 3aa and phosphate P (at 24 h). The ³¹P NMR spectra also supported the transformation of phosphonate 1a (δ = 24.0 ppm) to phosphate P (δ = 0.6 ppm) through the chelated intermediate M (δ = 41.8 ppm) (Figures 1c and S1 and 2 in SI).^14b−14d In contrast, other alkali carbonates (Li₂CO₃, Na₂CO₃, or K₂CO₃) or even stronger bases (KOH, DBU, NaHMDS, or KHMDS) did not generate the corresponding carbanion chelate M sufficiently, and thus, 3aa was not formed effectively (Table 1 and Figure 2)

Figure 1.

Reaction was monitored via NMR in 0.68 mL of DMSO-d₆ with 1a (0.136 mmol), 2a (0.163 mmol), and Cs₂CO₃ (0.272 mmol). (a) Proposed cesium-chelated intermediate M. (b) ¹H NMR and (c) ³¹P NMR spectra of triazole 3aa formation.

Figure 2.

Reaction was monitored via NMR in 0.68 mL DMSO-d₆ with 1a (0.136 mmol) and base (0.251 mmol). (a) ¹H and (b) ³¹P NMR spectra for the base treatment of β-ketophosphonate 1a after 1 h.

The advantage of Cs₂CO₃ over K₂CO₃ was evident when dimethyl (2-oxopropyl)phosphonate 4a was used as the dipolarophile. Under the Cs₂CO₃-mediated conditions, 1,5-DTs 5 were formed exclusively in most cases with good to excellent yields (71–91%). In contrast, the chemoselectivity between 5 and undesired 4-phosphonated TT 5′ was reported low or even reversed (5′ae) in the presence of K₂CO₃ (Table 3).^9b The existence of the cesium-chelated intermediate was studied via the crystallization of 1a and 4a individually with Cs₂CO₃ to form 6 and 7, respectively (as Cs atoms were severely disordered, hydrogen atoms in structure 6 could not be defined). The cesium enolate crystal structure of 6 showed a Z/E ratio of 2:1, while the Z conformation was observed exclusively for 7 (Figure 3). The individual exposures of 6 and 7 to 2a in DMSO gave only 3aa and 5aa, respectively.

Table 3. Comparison between K₂CO₃ and Cs₂CO₃.

Figure 3.

Single crystal X-ray structures of cesium enolate (a) 6 (CCDC 2156078) and (b) 7 (CCDC 2156080).

A reaction mechanism is proposed, as depicted in Figure 4. In the absence of the chelation effect, the equilibrium between enolates E-A and Z-B is governed by the steric effect of R¹ and R², as well as by the interaction between the alkoxy anion and the phosphonate. Either the E- or Z-olefinic anion (E-A or Z-B), acting as a dipolarophile, competitively reacts with an azide to form 3′ or 3, respectively, via the triazoline intermediate A′ or B′. After protonation and deprotonation, the orientation of A′ with R² = H tends to eliminate hydroxide and yield the 4-phosphonated triazole 3′. The substituted R² occupations (R² ≠ H) is bias against the previous pathway and equilibrates the Z-B conformer that invariably forms 3. However, in the presence of Cs₂CO₃, the enolate of β-ketophosphonate would preferentially bind in the Z-conformation as cesium enolate intermediate M, irrespective of the nature of R². The resulting syn-orientation of the alkoxy anion and phosphonate facilitates the formation of the oxaphosphetane B′′ and accelerates the HWE-type elimination to form 3.

Figure 4.

Proposed mechanism for the formation of 1,2,3-triazoles in the presence or absence of Cs₂CO₃.

3. Conclusion

We found that Cs₂CO₃ in DMSO is a unique system to facilitate the formation of 1,2,3-triazoles from β-carbonyl phosphonates under mild conditions. The reaction is highly regioselective, with the formation of expected substituted triazolyl products in high yield. We demonstrated the cesium-chelated Z-enolate as an efficient dipolarophile in the [3 + 2] cyclization with an applied azide dipole. The protocol allows access to not only 1,4- or 1,5-DTs but also 1,4,5-TTs with various functional groups in good to excellent yields. This protocol successfully provided access to 1,4,5-TTs containing diverse functionalities at positions 4 or 5, which were previously not readily accessible.

4. Experimental Section

General Procedure for the Synthesis of Multisubstituted 1,2,3-Triazoles

α-Substituted-β-ketophosphonate 1 (1.0 equiv, 0.3 M) and cesium carbonate (2.0 equiv) were mixed in DMSO for 10 min, and azide 2 (1.2 equiv, 0.3 M) in DMSO was injected into the resulting reaction. After the reaction was completed according to thin-layer chromatography (TLC), the solution was diluted with EtOAc (10 mL) and washed with brine (15 mL × 3) to remove DMSO. Occasionally, the product might be found in the aqueous layer. Then the aqueous layer could be extracted with additional EtOAc if needed. The organic layers were all combined, dried over anhydrous MgSO₄, filtered, and concentrated under reduced pressure to give the crude residue. The residue was purified by flash column chromatography on silica gel to give the desired product.

Data Availability Statement

The data underlying this study are available in the published article and its Supporting Information.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.joc.3c02836.

• Experimental procedures; FT-IR spectra, NMR spectra, mass spectra, and crystal data for all new compounds (PDF)

The authors declare no competing financial interest.