3-Component reactions for accessing heterocycle-rich topologies: trapping of pyrrole-stabilized carbenes via net bimolecular C–H or N–H insertion

Abstract

The intermolecular trapping of free carbenes derived from diazo precursors is well documented in the literature. We have recently reported the generation of free carbenes derived from 2-alkynyliminoheterocycles, a process that is both 100% atom economical and metal free. Here, we expand the synthetic utility of heteroaryl-stabilized, free carbenes by demonstrating a host of intermolecular third-component trapping reactions. Terminal alkynes, N-Boc carbamates, and NH-amides are shown to be suitable third-component traps. Strategic selection of the three reaction partners (i.e., the 2-alkynyliminoheterocycle, the electron-deficient alkyne partner, and the trapping moiety) enables access to product diversity of variable topology and a high degree of heterocycle incorporation.

---

Introduction

Free carbenes and their metal-bound carbenoid counterparts have enabled synthetic methodology through a diverse display of reactivity, including: cyclopropanation, ylide generation, and X–H/C–H insertion.^1–6 Diazo compounds can produce free carbenes through ejection of molecular nitrogen under thermal^4,5 or photochemical^1–3 conditions (Fig. 1a). Davies and co-workers pioneered the broad application of intermolecular trapping of free carbenes to allow for rapid construction of highly functionalized and complex products.^7,8 However, the intermolecular trapping reactions of heteroaryl-substituted free carbenes are less explored and underutilized.⁹

Fig. 1 (a) Diazo compounds as free carbene precursor, (b) (3 + 2) cycloaddition of 2-alkynyliminoheterocycles A with electron-deficient alkynes B to produce a free carbene C, and (c) unexpected intermolecular trapping of free carbene D with terminal alkyne 2.

We recently described the formal (3 + 2) cycloaddition between 2-alkynyliminoheterocycles and electron-deficient alkynes to afford heteroaryl free carbene intermediates (Fig. 1b).¹⁰ In addition to the reactivity displayed by free carbenes C in the initial report (including, but not limited to, C–H insertion, (3,2)-sigmatropic rearrangement, 1,3-dipole formation, and carbene metathesis), we have demonstrated the ability to access complex polycyclic cyclopropanes via intramolecular capture of the carbene by tethered alkenes.¹¹ In exploring the reactivity of various electron-deficient alkynes, we observed an unexpected product 3 (34%) when reacting the allyl ether 1 with the terminal conjugated 1-yne-3-one derivative 2 (Fig. 1c).

This unexpected product had incorporated two equivalents of ynone 2, one that produced the carbene intermediate D and a second that had become incorporated through a formal C–H insertion event of the carbenic carbon atom into the terminal alkyne C–H bond of 2. This can be rationalized via a deprotonation–ion pair collapse (see intermediates D to E, Fig. 1c), a process precedented by work of Koenigs and coworkers for the capture of a different class of free carbene by a terminal alkyne.¹² The formation of 3 led us to initiate a study of additional examples in which various types of a third component might efficiently trap heteroaryl-substituted carbenes. We report here on the results of these investigations of three-component reactions, which provide structurally novel heterocyclic products in a single operation.

Results and discussion

Additional examples of these three-component reactions involving trapping by a terminal alkyne are shown in Fig. 2. Product 3 (cf.Fig. 1c) had incorporated two equivalents of the ynone 2, one functioning as the electrophilic alkyne (cf.B) and the second to trap the carbene D. Another example of this is seen in products 7a/b in which two copies of methyl propiolate (5a), another electrophilic terminal alkyne, have been engaged. We recognized that the versatility of this three-component process would be increased if three different molecules could be integrated into the products. We set out to achieve this outcome by using two different types of alkynes, one to function as the electrophilic moiety to drive carbene formation and the second, a terminal alkyne, to capture the carbene.

Fig. 2 Third-component trapping of free carbenes derived from alkynylpyridines 4a–d with terminal alkynes 6a–e. ^ap-Methoxyphenyl.

In the presence of dimethyl acetylenedicarboxylate (5b, DMAD), the alkynylpyridine 4a, and 1-ethynyl-4-(trifluoromethyl)benzene (6a), could be transformed to the corresponding trapped product 7c (43%). Many of the subsequent experiments utilized DMAD as the electron-deficient alkyne because DMAD, unlike electron-deficient terminal alkynes, is, of course, unable to participate in C–H insertion chemistry. Formation of the 1-methoxycyclohexane derivative 7d (42%) shows that a bulky quaternized alkyl substituent on the terminus of the alkynylpyridine is tolerated. Phenylacetylene (6b) trapped the carbene to form 7e. However, this required the use of a large excess of 6b (1 : 1 cosolvent) to achieve the yield of 57%, implying that this least acidic of the terminal alkynes 6a–e is relatively slow at engaging the electron-rich carbene. 3-Ethynylpyridine (6c) was sufficiently electron-poor to participate as the trapping component to generate 7f (41%). The highest yielding alkyne trap was N-methyl-N-phenylpropiolamide (6d), leading to the amide derivative 7g (84%). We varied the loading of 6d from 1.2 to 5.0 equivalents and observed increased yields at higher loadings of the terminal alkyne (1.2 equiv.; 62%, 1.75 equiv.; 73%, and 5.0 equiv.; 84%).

We were also interested in whether a propiolate derivative trap would compete with the more electrophilic DMAD to produce the carbene. Using the indole-containing propiolate derivative 6e (along with the azetidinylated alkyne 4d), we observed the formation of 7h in 61% yield. It is notable that the methine hydrogen atom at C3 of the azetidine (vicinal to the carbene) did not interfere by way of a 1,2-insertion event into the initially propargylic C–H bond, a process that earlier was shown to be efficient (i.e., fast) for carbenes containing simple alkyl substituents.¹⁰ The ring strain associated with the formation of an exo-alkylidene azetidine likely reduces the rate of this intramolecular process. Notice that a similar factor was relevant to the compatibility of the cyclopropane ring during the formation of 7b described earlier.

In exploring a broader scope of functional groups that would serve as the carbene trapping agent (i.e., the third component), we found that N-Boc-carbamates were effective. This discovery greatly expanded the opportunities and utility of this methodology, given the large pool of valuable and commercially available amine-containing coupling partners. The results of carbamate trapping are shown in Fig. 3; DMAD was held constant as the electrophilic alkyne partner. N-Boc-benzylamine (8a) trapped the carbene to provide 9a (75%). Use of a 3-fluoro-4-pyridyl substrate 4e led to 9b (60%). The trapping with N-Boc-diphenylmethanamine (8b) to give 9c proceeded with lower efficiency (28%), showing that the steric bulk of the BocNH compound can be a limitation. An array of heterocycle-containing traps is tolerated: N-Boc-furfurylamine (8c) gave 9d (61%); the piperazine derivative 8d gave 9e (40%); the indole derivative 8e gave 9f (55%); and the amino pyrazole derivative 8f gave 9g (37%). In principle, N-Boc-propargylamine (8g) could have inserted at either the N–H or terminal alkyne C–H bond. We only observed products of NH trapping: namely, 9h (75%) and 9i (32%), the formation of which is perhaps a function of the relative acidity of the two potential sites of reaction.

Fig. 3 Third-component trapping by (a) N-Boc-carbamates (8a–g), (b) thiophen-2-ylmethyl phenylcarbamate (10), and (c) N-Fmoc-benzylamine (12). ^asee SI for reaction temperatures.

We sought to further capitalize on the carbamate substitution as a handle to append heterocycle functionality. Accordingly, the thiophen-2-ylmethyl carbamate 10 was found to trap efficiently, leading to adduct 11 (60%, Fig. 3b). Attempts to deprotect several of the N-Boc-carbamate products described in Fig. 3a under typical conditions (TFA/chloroform, rt) showed surprisingly complex behavior. This prompted us to examine the Fmoc-carbamate 12, which gave 13 (58%, Fig. 3c). In contrast to Boc-removal attempts under acidic conditions, this product could be smoothly deprotected to the secondary amine 14 when exposed to pyrrolidine in acetonitrile.

We also found that amide and urea derivatives behaved well in the three-component reaction (Fig. 4). The secondary amide, N-methylbenzamide, was a particularly effective trap leading to the tertiary amide 15 (86%). The structure of 15 was validated by X-ray diffraction analysis (see p. S83, SI). The primary acetamide also participated to give 16 (54%). A dibenzoazepine-5-carboxamide derivative established that a urea would engage the carbene, here to produce 17 (62%).

Fig. 4 Amides and a urea trapping of the carbene produced from the alkynylpyridine 4a and DMAD (5b).

Iminoheterocycles other than pyridine have been shown to be effective carbene precursors.¹⁰ Here we used N-Boc-carbamates as the third component to establish the effectiveness of several different alkynyl-substituted heterocycles in this three-component approach (Fig. 5). The 2-alkynylthiazole 18 was trapped with N-Boc-benzylamine (8a), N-Boc-aniline (8h), and the N-Boc-valine derivative 8i to give 19 (76%), 20 (70%), and 21 (55%), respectively. The valine adduct 21 was formed as a 1.1 : 1.0 mixture of coeluting diastereomers, suggesting that it is unlikely that a high degree of stereoselectivity could be found in these NH insertion reactions. In comparison to its pyridine analogue 4a, the thiazole substrate 18 requires higher temperatures to achieve full conversion. The pyrimidine substrate 22 was even slower to react, although it still afforded the product 23 in respectable yield (54%). Lastly, we studied a non-aromatic iminoheterocycle, the dihydropyrrole derivative 24. Although this substrate generated the carbene more quickly, it proved to be less efficient in producing the three-component trapped product 25 (34%).

Fig. 5 Alkynyliminoheterocycles other than pyridine (thiazole 18, pyrimidine 22, and dihydropyrrole 24) can also participate in the three-component coupling.

We surveyed several other (modestly acidic) carbon species to explore whether we could identify any additional C–H insertion events analogous to the alkynylations described in Fig. 2. Somewhat surprisingly, we were unsuccessful in attempts to use any of cyclopentane-1,3-dione, dimethyl malonate, malononitrile, or HCN to achieve this outcome. An exception was the use of chloroform, a reactant that has been studied in some detail for its ability to engage N-heterocyclic carbenes (NHCs) through net C–H insertion.^13,14 Indeed the chloroform adduct 26 was isolated in 74% yield when 4a and 5b were heated in CHCl₃ at 100 °C (Fig. 6). The structure of 26 was validated by X-ray diffraction analysis (see p. S83, SI). When this reaction was performed in a 14 : 1 volume ratio of CDCl₃ : CHCl₃, the isotopomers 26 and 27 were formed in a 54 : 46 ratio, indicative of a substantial isotope effect (k_H/k_D ≈16) in cleavage of the C–H bond in the product-determining step. This, however, does not differentiate between a concerted or stepwise (via the ion pair F) process for interception of the free carbene. A reported DFT analysis¹⁴ of this mechanistic dichotomy for the reaction of NHCs with chloroform¹³ favored a stepwise, ion pair pathway. Interestingly, treatment of the chloroform adduct 26 with ⁿBuNH₂ resulted in clean formation of the dichloroalkene 28 (97%) via E2-elimination.

Fig. 6 Formal carbene insertion into the C–H bond of CHCl₃.

Several additional examples that show further diversity in the types of products that can be accessed using this three-component assembly are presented in Fig. 7. Electron-deficient alkyne partners other than DMAD can be used. The trifluoromethyl diynyl ketone 29 reacted with the alkynylpyridine 4h and N-Boc-benzylamine (8a) to afford the alkyne-substituted indolizine 30 (48%) as a single regioisomer (Fig. 7a). Each of the alkynylpyridine derivatives 4f and 4h engaged trifluorobutynoate 31 and the N-Boc-cyclopropylamine 32 as the second and third components to produce the trifluoromethyl substituted indolizine 33a (33%) and 33b (65%), respectively (Fig. 7b). Comparison of the outcome of 2-ethynylpyridine (4f) vs. that of its phenyl-substituted analogue 4h reveals that the terminal alkyne substrate is, overall, a less efficient participant in the three-component reaction (cf. also the marginal yield of product 9e, Fig. 3).

Fig. 7 Electron-deficient alkyne partners other than DMAD participate in three-component coupling reactions.

Finally, we probed whether the three-component reaction could be applied in the context of an even more complex molecular setting. We prepared the estradiol derivative 34, to which a requisite alkynyliminoheterocycle functionality is appended (Fig. 7c). When heated in the presence of the trifluorobutynoate 31 and N-methyl-N-phenylpropiolamide (6d), the (steroidylalkynyl)pyridine 34 was converted into the indolizine derivative 35 (48%) as a 10 : 1 coeluting mixture of epimers at the newly created (propargylic) stereogenic center (a DP4+ NMR analysis¹⁵ suggested that the major epimer has the R configuration; see SI, Section III). The level of diastereoselectivity (relative asymmetric induction) in this transformation is notable, as is the formation of a small amount of the cyclic enol ether 36 (9%), presumably arising from the ring-expansion/rearrangement indicated in species G.¹⁶

Conclusions

In conclusion, we have demonstrated that free carbenes generated by the formal (3 + 2) cycloaddition between 2-alkynyliminoheterocycles and electron-deficient alkynes can trap a variety of third components. Secondary N-Boc carbamates, amides, and ureas as well as acidic terminal alkynes are functionalities capable of efficiently participating in the third-component trapping event. Notably, the pK_a's of the CH or NH are similar across this series of effective third components, suggestive of a stepwise process to account for the net C–H/N–H insertion event.

Each of the three reaction partners – the alkynyliminoheterocycle, the electron-deficient alkyne, and the carbene trapping agent – can be varied, demonstrating the potential for formation of a diverse collection of poly-heteroaromatic products. The generic structure in Fig. 8 shows the topologies of pyrrole-templated compounds that can be accessed and summarizes the locations in the product at which heterocyclic moieties can be installed.

Fig. 8 Topologies of heterocycle-bearing, pyrrole derivatives accessible by reactions of pyrrole-stabilized free carbenes with third-component traps. Heterocycles can be integrated into the products at the positions designated as Het1/Het1′ (from the alkynyliminoheterocycle), Het2 (from the electron-deficient alkyne), and Het3 (from the carbene trapping agent).

Author contributions

Conceptualization: ALG; methodology and data curation: ALG, KBT; data interpretation, writing – original draft of manuscript and reviewing/editing final version: all authors; funding acquisition: TRH.

Conflicts of interest

There are no conflicts to declare.

Data availability

The NMR raw data can be accessed at: https://doi.org/10.6084/m9.figshare.29321516.

CCDC 2492945 and 2492946 contain the supplementary crystallographic data for this paper.^17a,b

Supplementary information: experimental procedures used to prepare each new chemical entity (NCE); line listings of spectroscopic characterization data for each NCE; copies of all ¹H and ¹³C 1D NMR spectra and selected sets of 2D NMR data; .zip file of Gaussian .out files of the conformers used in the DP4+ analysis of 35. See DOI: https://doi.org/10.1039/d5sc05345e.

Acknowledgements

This research was supported by a MIRA from the National Institutes of General Medical Sciences (R35 GM127097) of the U.S. National Institutes of Health (NIH). NMR spectra were recorded using an instrument partly funded by the Shared Instrumentation Grant program (S10 OD011952) of the NIH. High resolution mass spectral data were obtained in the University of Minnesota's Analytical Biochemistry Shared Resource Laboratory using instrumentation purchased with support from the NIH National Cancer Institute (P30 CA077598). DFT computations were done using resources available at the Minnesota Supercomputing Institute (MSI). Our coworker Jingyang Shi kindly provided the diyne 29.

References

1. H. Meerwein, H. Rathjen and H. Werner, Die Methylierung von RH-Verbindungen mittels Diazomethans unter Mitwirkung des Lichtes, Ber. Dtsch. Chem. Ges. A/B, 1942, 75, 1610–1622.
2. Z. Yang, M. L. Stivanin, I. D. Jurberg and R. M. Koenigs, Visible light-promoted reactions with diazo compounds: A mild and practical strategy towards free carbene intermediates, Chem. Soc. Rev., 2020, 49, 6833–6847.
3. Z. Zhang and V. Gevorgyan, Visible light-induced reactions of diazo compounds and their precursors, Chem. Rev., 2024, 124, 7214–7261.
4. E. Buchner and L. Feldmann, Diazoessigester und Toluol, Ber. Dtsch. Chem. Ges., 1903, 36, 3509–3517.
5. A. Ford, H. Miel, A. Ring, C. N. Slattery, A. R. Maguire and M. A. McKervey, Modern organic synthesis with α-diazocarbonyl compounds, Chem. Rev., 2015, 115, 9981–10080.
6. C. Ebner and E. M. Carreira, Cyclopropanation strategies in recent total syntheses, Chem. Rev., 2017, 117, 11651–11679.
7. S. R. Hansen, J. E. Spangler, J. H. Hansen and H. M. L. Davies, Metal-free N–H insertions of donor/acceptor carbenes, Org. Lett., 2012, 14, 4626–4629.
8. I. D. Jurberg and H. M. L. Davies, Blue light-promoted photolysis of aryldiazoacetates, Chem. Sci., 2018, 9, 5112–5118.
9. R. S. Sheridan, Heteroarylcarbenes, Chem. Rev., 2013, 113, 7179–7208.
10. Q. Xu and T. R. Hoye, Free carbenes from complementarily paired alkynes, Nat. Chem., 2024, 16, 1083–1092.
11. A. L. Guzman, A. N. Mann and T. R. Hoye, Alkynes to (free) carbenes to polycyclic cyclopropanes, J. Am. Chem. Soc., 2024, 146, 28642–28647.
12. S. Jana, C. Pei, C. Empel and R. M. Koenigs, Photochemical carbene transfer reactions of aryl/aryl diazoalkanes—experiment and theory, Angew. Chem., Int. Ed., 2021, 60, 13271–13279.
13. A. J. Arduengo, J. C. Calabrese, F. Davidson, H. V. Rasika Dias, J. R. Goerlich, R. Krafczyk, W. J. Marshall, M. Tamm and R. Schmutzler, C−H insertion reactions of nucleophilic carbenes, Helv. Chim. Acta, 1999, 82, 2348–2364.
14. F. Perez, Y. Ren, T. Boddaert, J. Rodriguez and Y. Coquerel, A stable N-heterocyclic carbene organocatalyst for hydrogen/deuterium exchange reactions between pseudoacids and deuterated chloroform, J. Org. Chem., 2015, 80, 1092–1097.
15. N. Grimblat, M. M. Zanardi and A. M. Sarotti, Beyond DP4: an improved probability for the stereochemical assignment of isomeric compounds using quantum chemical calculations of NMR shifts, J. Org. Chem., 2015, 80, 12526–12534.
16. A. Padwa, D. J. Austin, Y. Gareau, J. M. Kassir and S. L. Xu, Rearrangement of alkynyl and vinyl carbenoids via the rhodium(II)-catalyzed cyclization reaction of α-diazo ketones, J. Am. Chem. Soc., 1993, 115, 2637–2647.
17. (a) CCDC 2492945: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2pp3l3; (b) CCDC 2492946: Experimental Crystal Structure Determination, 2025,  DOI:10.5517/ccdc.csd.cc2pp3m4.

---