Aryl Radical Enabled, Copper-Catalyzed Sonogashira-Type Cross-Coupling of Alkynes with Alkyl Iodides

Abstract

Despite the success of Sonogashira coupling for the synthesis of arylalkynes and conjugated enynes, the engagement of unactivated alkyl halides in such reactions remains historically challenging. We report herein a strategy that merges Cu-catalyzed alkyne transfer with the aryl radical activation of carbon–halide bonds to enable a general approach for the coupling of alkyl iodides with terminal alkynes. This unprecedented Sonogashira-type cross-coupling reaction tolerates a broad range of functional groups and has been applied to the late-stage cross-coupling of densely functionalized pharmaceutical agents as well as the synthesis of positron emission tomography tracers.

Graphical Abstract

INTRODUCTION

As one of the most valuable functional groups in organic chemistry, alkynes are ubiquitous in a diverse range of natural products, pharmaceuticals, agrochemicals, and organic materials.¹ In addition, alkynes have been widely used in Cu-catalyzed and strain-promoted azide–alkyne cycloaddition reactions, that is, click chemistry, for the conjugation of bioactive molecules.² Moreover, alkynes are vital synthetic precursors for the construction of a wide variety of functionalities, including alkenes, aldehydes, acids, and heterocycles.³ Recent work has shown that alkyne-tagged small molecules are promising agents for Raman imaging, as the Raman bands of alkynyl groups do not overlap with Raman scattering from endogenous molecules in live cells.⁴ Due to the importance of alkyne-containing molecules, the development of new methods that can efficiently install alkynyl groups into functionalized organic molecules has been the subject of numerous studies.^5,6

Sonogashira coupling reactions,⁷ in which aryl and vinyl (pseudo)halides react with terminal alkynes under Pd-catalyzed, Cu-cocatalyzed conditions, are among the most efficient and modular approaches for the synthesis of alkynes.⁸ Despite the success of Sonogashira coupling for constructing arylalkynes and conjugated enynes, the engagement of unactivated alkyl halides in such reactions remains a formidable challenge.⁹ This limitation is largely due to the sluggish oxidative addition of alkyl halides and the tendency of alkylpalladium intermediates to undergo the competing β-hydrogen elimination reactions.¹⁰ The classic ${\text{S}}_{{\text{N}}_2}$ reaction of acetylide ions with alkyl halides suffers from poor functional group tolerance. Few methods allow for the coupling of unactivated and β-hydrogen bearing alkyl halides with terminal alkynes (Figure 1a). Seminal work by Fu¹¹ and Glorius¹² showed that the use of sterically bulky N-heterocyclic carbene ligands facilitated the Pd-catalyzed Sonogashira coupling of unactivated alkyl halides. Hu¹³ and Liu¹⁴ have later reported that the analogous reactions could be catalyzed by Ni complexes using pincer and pyridine bisoxazoline ligands, respectively. Very recently, based on the pioneering work of Fu and Peters,¹⁵ and Hwang,¹⁶ the Lalic group reported a photoinduced, Cu-catalyzed approach for coupling unactivated alkyl iodides with terminal alkynes.¹⁷ Nonetheless, the generality of these elegant precedents remain rather limited, due in part to the requirement for strong bases, specially designed ligands, or the high reactivity of photoexcited metal species. Indeed, few of these methods have been applied to the cross-coupling of densely functionalized molecules.

Figure 1.

Development of a new approach for the cross-coupling of terminal alkynes with unactivated alkyl halides.

Our group has recently shown that merging the unique reactivity of aryl radicals with copper catalysis could enable the cross coupling of hitherto challenging alkyl electrophiles with organozinc reagents.¹⁸ This approach complements the elegant work by MacMillan¹⁹ and Leonori,²⁰ who have demonstrated that silyl radicals and α-amino alkyl radicals, respectively, could activate alkyl halides for Cu catalysis. More specifically, our strategy is inspired by the underexplored reactions between aryl radicals with alkyl iodides to generate alkyl radicals at rates approaching the diffusion-controlled limit (k = 10⁹ M⁻¹ s⁻¹).²¹ We reasoned that combining the unique ability of Cu catalysts to generate aryl radicals with arenediazonium salts, as in the Sandmeyer reactions,²² with the ability of aryl radicals to activate alkyl iodides²³ would provide a versatile platform for Cu-catalyzed cross-coupling reactions. This previously unexplored approach has promoted a Cu-catalyzed Negishi-type difluoromethylation reaction (Figure 1b).^18b We recently questioned whether this aryl radical activation strategy could be translated to the Sonogashira-type cross-coupling, thus enabling a general approach to the construction of C(sp³)–C(sp) moieties. We herein disclose the successful execution of these ideas and present a broadly applicable protocol for the cross-coupling of unactivated alkyl iodides with terminal alkynes under mild conditions (Figure 1c).

RESULTS AND DISCUSSION

Mechanistically, we reason that Cu^I catalyst 1 reacts with terminal alkyne 2, in the presence of base 3, to form [Cu^I-acetylide] species 4 (Figure 2).²⁴ Subsequent single electron transfer (SET) from 4 to diazonium salt 5, followed by the extrusion of N₂, produces aryl radical 6 and [Cu^II-acetylide] intermediate 7.²⁵ The feasibility of this SET step was supported by the observation of an irreversible reduction peak at −0.26 V [vs SCE in dimethyl sulfoxide (DMSO)] for a diazonium salt and a peak at −0.96 V (vs SCE in DMSO) for an in situ generated Cu^I-acetylide species (Figure S1). Fast iodine abstraction from alkyl iodide 8 by the aryl radical 6 forms aryl iodide 9 and alkyl radical 10. The latter reacts with the [Cu^II-acetylide] species 7 to generate a formal organocopper(III) species 11,²⁶ which reductively eliminates to give the desired alkyne coupling product 12.²⁷ The reductive elimination reaction also regenerates the Cu^I catalyst, thus closing the catalytic cycle. We realized that a few challenges could be associated with this hypothesis. First, the formation rate of Cu^I-acetylide species 4 should be sufficiently high to match the fast reaction of a Cu^I complex with a diazonium salt. Additionally, the formation of dimerized alkynes under oxidative conditions, that is, Glaser coupling,²⁸ might compete with the cross-coupling pathway. Moreover, the known cross-coupling of the diazonium salts with the alkynes must be minimized.²⁹ We postulated that these challenges could be addressed by tuning the steric and electronic properties of the diazonium salts.

Figure 2.

Reaction design of aryl-radical-enabled, Cu-catalyzed Sonogashira-type cross-coupling.

Reaction Optimization.

A survey of a various combinations of different Cu catalysts, ligands, bases, and arenediazonium salts identified the optimized conditions, as shown in Table 1, entry 1 (see Supporting Information for other conditions studied). Thus, in the presence of [Cu(CH₃CN)₄]-BF₄ as the catalyst, 2,2′; 6′,2″-terpyridine (terpyridine) as the ligand, and potassium carbonate as the base, diazonium salt 14 facilitated the cross-coupling of alkyl iodide 13 with phenylacetylene to afford the desired alkyne product 15 in 91% yield at 50 °C. Consistent with our previous work on the aryl-radical-enabled difluoromethylation reaction,^18b an electron-rich and sterically hindered diazonium salt 14 was the optimal promoter for this alkynylation reaction, whereas electron-deficient or less sterically hindered ones were less effective (entry 2–3). We hypothesize that the steric hindrance of the thus-produced aryl radical prevented its direct coupling with the alkyne. Solvent screening revealed that this transformation was most efficient in DMSO, whereas the use of other standard organic solvents decreased the yield of the desired products (entries 4 and 5). Cu salts with weakly coordinating ligands were preferred for this coupling reaction, presumably by facilitating the formation of the copper acetylide species (entries 6 and 7). The commercially available terpyridine ligand was found to be the most effective ligand, while other bidentate ligands, including bipyridine and phenanthroline, had deleterious effects on the coupling reactions (entries 8 and 9). Previous work by Lalic suggested that the use of tridentate ligands could help prevent the undesired alkyne polymerization pathway.¹⁷ The base also played an important role in this reaction, with potassium carbonate being uniquely effective when compared with other commonly employed organic and inorganic bases (entries 10–12). Notably, although Cu-catalyzed Glaser coupling is known to occur under aerobic conditions, this aryl-radical-enabled protocol could afford the product in synthetic useful yield under an air atmosphere (entry 13). The desired coupling product could be formed in good yield with lower catalyst loading. (entries 14 and 15). As expected, no coupling products were obtained when the reaction was conducted without a copper catalyst, diazonium salt, or base (entries 16–18). The alkyl iodide remained intact in the absence of the diazonium salt, consistent with the Cu catalyst alone not being reactive toward an unactivated alkyl iodide.

Table 1.

Reaction Optimization for Aryl-Radical-Enabled Sonogashira Coupling of Alkyl Iodides^a

entry	variations	yield (%)b	entry	variations	yield (%)b
1	none	91	10	Na2CO3 as base	N.D.
2	4-methoxy-arenediazonium salt	30	11	K3PO4 as base	19
3	4-bromo-arenediazonium salt	N.D.	12	Et3N as base	N.D.
4	acetonitrile as solvent	69	13	under air	60
5	THF as solvent	15	14	20 mol % Cu catalyst and ligand	79
6	CuBr as catalyst	43	15	15 mol % Cu catalyst and ligand	65
7	CuI as catalyst	48	16	no Cu catalyst	N.D.
8	2,2’-bipyridine as ligand	12	17	no diazonium salt	N.D.
9	1,10-phenanthroline as ligand	8	18	no ligand	N.D.

^aReactions were conducted using 13 (0.050 mmol, 1.0 equiv), diazonium salt (0.10 mmol, 2.0 equiv), phenylacetylene (0.055 mmol, 1.1 equiv), potassium carbonate (0.15 mmol, 3 equiv), [Cu(CH₃CN)₄]BF₄ (25 mol %), and ligand (25 mol %) in DMSO at 50 °C.

^bYields were determined by GC with 1-ethylnaphthalene as the internal standard. N.D., not detected.

Substrate Scope.

With these optimized conditions in hand, we then explored the scope of this aryl-radical-enabled Sonogashira-type cross-coupling reaction. As shown in Table 2, this approach was broadly applicable to the coupling of a wide array of alkyl iodides with terminal alkynes. Secondary alkyl iodides appended to various four- to seven-membered rings efficiently reacted with a terminal alkyne, affording the corresponding coupling products in good to excellent yield (16 to 27, 52–85% yield). Functional groups including acetals, carbamates, esters, sulfonamides, and amides were well tolerated. Notably, a substrate possessing a tertiary amine, a traditionally challenging functional group for transition metal catalysis, delivered the alkynylated product in good yield (17, 73% yield). Heterocycles commonly encountered in medicinal chemistry, such as furan, thiophene, piperidine, azepane, and pyrrolidone, were compatible with this protocol. The coupling reaction is also tolerant of conjugated alkenes (21), which are known to rapidly react with aryl radicals,³⁰ highlighting the high selectivity of aryl radicals toward the activation of alkyl iodide bonds. Several medicinally relevant bicyclic and tricyclic compounds formed the desired products with good efficiency (24–27, 58–85% yield). Moreover, secondary acyclic alkyl iodides (28–31) and primary alkyl iodides bearing diverse functional groups, including alkyl bromides, silyl ethers, and hydroxyl groups, could be converted into the corresponding alkynes in good to high yield (32–37, 47–82% yield). A tertiary alkyl iodide derived from adamantane also participated in this transformation, albeit with moderate efficiency (38, 40% yield). Nonetheless, coupling of an acyclic tertiary alkyl iodide, tert-butyl iodide, failed to form the desired product (39), presumably due to the steric hindrance of the tert-butyl radical intermediate.

Table 2.

Substrate Scope of the Aryl Radical-Activated, Copper-Catalyzed Sonogashira-type Coupling of Alkyl Iodides^a

^aReactions were run with 0.20 mmol of alkyl iodides (1.0 equiv), 0.22 mmol of alkyne (1.1 equiv), 0.40 mmol of diazonium salt 14 (2.0 equiv), terpyridine (25 mol %), and [Cu(CH₃CN)₄]BF₄ (25 mol %) in of DMSO at 50 °C under Ar. Isolated yields were reported.

^bdr (diastereomeric ratio) of the starting alkyl iodides >20:1.

The compatibility of different alkynes with this Cu-catalyzed reaction was then studied. We were pleased to find that the efficacy of this protocol showed little dependence on the electronic properties of aryl substituents, as various phenyl acetylenes with electron-withdrawing, -neutral, and -donating groups reacted with an alkyl iodide to afford the corresponding alkynes in good yield (40–50, 62–85% yield). Interestingly, despite the known reactions between ferrocene and aryldiazonium salts,³¹ ethynylferrocene readily participated in this coupling reaction (47, 82% yield). Other conjugated alkynes, including an ethynylthiophene (48), an enyne (49), and a propiolate (50), could react smoothly, furnishing the conjugated alkyne products in good yield (58–75%). Moreover, coupling reactions proceeded with nonconjugated alkynes bearing various functional groups (hydroxyl, halide, sulfone, nitrile, nitro, acetal, amide, and sily; 51–61, 35–78% yield). Finally, the scalability of the reaction was evidenced by the preparation of 23 on a half gram scale (75% yield).

Late-Stage Cross-Coupling of Bioactive Molecules.

In an effort to demonstrate the applicability of our protocol to the synthesis of drug-like molecules, we tested the cross-coupling of a wide range of bioactive molecules using this Sonogashira-type transformation (Table 3). Alkyl iodides derived from various monosaccharides (glucose, xylose, ribose, fructose, and galactose) were smoothly converted to the corresponding alkyne derivatives in moderate to good yield (66–71, 40–75% yield). Notably, a single diastereomer was formed for the alkynylation of a glucose derivative (66), whereas the previous synthesis of an analogous alkyne molecule required a five-step synthesis.³² In addition, a variety of steroid derivatives successfully underwent coupling reactions (72–75, 41–70% yield). Three alkynes derived from bioactive molecules—norethindrone (76), caffeine (77), and clodinafop (78)—were successfully coupled under the standard conditions. A benzoquinone-containing drug agent (idebenone), a xanthine derivative (proxyphylline), and an estrogen receptor modulator (ospemifene) were all smoothly converted to the desired products (79–81).

Table 3.

Late-Stage Cross-Coupling of Bioactive Molecules^a

^aSee Supporting Information for detailed conditions.

^b(diastereomeric ratio) of the starting alkyl iodides >20:1.

Empaglifloizin, a glucoside compound and one of the top-selling small molecule pharmaceuticals in 2021,³³ formed the corresponding alkynylated analogue in 50% yield (82). The compatibility of unprotected triol groups further highlights the mild conditions of this protocol. Densely functionalized pharmaceutical agents including tedizolid (83), which consists of a tetrazole, a pyridine, and an oxazolidone ring, as well as ticagrelor (84) which incorporates a thioether, a diol, and a triazolo[4,5,-d]pyrimidine scaffold, were converted to their alkynylated analogues in 66% and 45% yield, respectively. Considering the value of unnatural amino acids in chemical biology and drug discovery, we also applied this protocol to the synthesis of two alkyne-containing analogues of methionine and alanine (85 and 86). Moreover, we developed a one-pot protocol for the synthesis of terminal alkynes, given their wide range of applications in click chemistry. Thus, the coupling of trimethylsilylacetylene with the alkyl iodide derived from proxyphylline under the standard conditions, followed by the in situ deprotection of the silyl group, afforded the alkyne derivative (87) in overall 65% yield. Finally, to demonstrate the potential of this method for the conjugation of functionalized alkynes and alkyl iodides, an alkyl iodide derived from empaglifloizin successfully coupled with the alkyne derived from clodinafop, affording the desired product (88) in 62% yield. It is worthwhile mentioning that this coupling product was not obtained under the previously reported photoinduced or Ni-catalyzed conditions, further highlighting the potential utility of this aryl-radical-enabled protocol for the synthesis of previously inaccessible alkyne molecules.

Mechanistic Studies.

Mechanistic studies shed light on this aryl-radical-enabled, Cu-catalyzed protocol (Figure 3). The addition of 2 equiv of TEMPO, a radical-trapping agent, completely suppressed the formation of the coupled product, whereas the TEMPO trapping product (89) was isolated in 55% yield (Figure 3a). Additionally, the coupling of a cyclopropyl-containing alkyl iodide 90 afforded mainly the ring-opened product 91, confirming the intermediacy of a cyclopropylcarbinyl radical, which is known to undergo rapid ring opening to the homoallylic radical (rate constant k = 2.7 × 10¹¹ s⁻¹) (Figure 3b).³⁴ The alkynylation of alkene-containing alkyl iodide 92 afforded the cyclized product 93 in 52% yield, consistent with a 5-exo-trig radical cyclization pathway (Figure 3c). These results together support the formation of alkyl radicals in these coupling reactions.

Figure 3.

Mechanistic studies support the proposed aryl radical activation mechanism. (a) TEMPO trapping experiment; (b) radical clock experiment; (c) ring closure experiment; and (d) reactivity of Cu-acetylide species.

Finally, the involvement of Cu–acetylide species in this reaction was studied (Figure 3d). Using a stoichiometric amount of Cu-acetylide complex 94, no reactions occurred with alkyl iodide 13, regardless of the presence of terpyridine as the ligand. In addition, although the diazonium salt 14 could undergo SET with 94, as indicated by the immediate effervescence of the reaction mixture, the coupling product 15 was not formed in the absence of terpyridine. On the contrary, 94 smoothly transferred the alkyne group to the alkyl iodide 13 in the presence of both the diazonium salt and terpyridine to afford the alkyne product 15 in 43% yield. These findings are consistent with the involvement of a ligand-bound Cu-acetylide species that transfers the alkyne group via the aryl radical activation pathway.

Synthetic Application for Positron Emission Tomography.

Positron emission tomography (PET) is a non-invasive imaging technique that allows tracking radiolabeled drug molecules inside living subjects.³⁵ Central to this technology is the development of novel PET tracers—targeted molecules that consist of positron-emitting isotopes.³⁶ Due to the short half-lives of such isotopes, including ¹⁸F (t_1/2 = 109.8 min), synthesis of PET tracers requires the reactions to be completed within a short period and conducted at the late stage of the overall synthesis process. Given the fast reaction kinetics of our alkynylation reaction (<20 min) and its broad functional group tolerance, we hypothesize that this protocol could offer a unique opportunity to the rapid conjugation of radio-labeled molecules for the development of PET tracers.

Thus, we aimed to develop a radio-conjugation approach that can couple ¹⁸F-labeled alkynes with alkyl iodides. Such an indirect method for ¹⁸F-radiolabeling reactions could circumvent a harsh step to carry out direct 18F-fluorination on complex molecules, which is incompatible with many sensitive functional groups.³⁵ In addition, the use of a communal ¹⁸F-labeled prosthetic group could allow for rapid assembly of a variety of target molecules and their corresponding derivatives without the need of ad hoc and lengthy syntheses of complex precursor molecules in the absence of their PET imaging utility. The major challenge for the radio-conjugation is to form the reactive Cu-¹⁸ F-acetylide intermediates using ¹⁸F-alkynes, the concentration of which is typically in the range of picomolar to nanomolar. In comparison, high concentration of alkynes (~0.2 M) is employed under catalytic conditions to facilitate the formation of such intermediates. In addition, the trace Cu-¹⁸ F-acetylide complex needs to rapidly participate in the ensuing SET and alkyne transfer steps to form the ¹⁸F-labeled products.

¹⁸F–alkyne 95 was chosen as the model substrate due to its ease of preparation; treatment of its NO₂ precursor (96) with ¹⁸F-fluoride in DMSO afforded 95 with a high radiochemical yield (76%, see Supporting Information for details). Under conditions similar to the catalytic reactions, 95 successfully coupled with primary and secondary alkyl iodides to afford the ¹⁸F-labeled alkyne products with moderate to good radiochemical yields (Table 4, 97–101, 15–36%). In addition, as high molar activity (A_m) radiotracers are often needed for imaging low abundance of receptors in vivo, we measured the A_m of compound 100, which was synthesized in >99% radiochemical purity. Gratifyingly, tracer 100 exhibited A_m activity of 87 GBq/μmol (2.36 Ci/μmol), which is sufficient for imaging purposes. Given these promising proof-of-concept results, we anticipate that this aryl radical strategy could be applied to the rapid installation of other radioisotopes for the synthesis of novel radiotracers. These studies are currently carried out in our laboratories.

Table 4.

Synthesis of ¹⁸F-Labeled Alkynes^a

^aReaction conditions: alkyl iodide (0.01 mmol), ¹⁸F-alknye (0.5–1 mCi), [Cu(CH₃CN)₄]BF₄ (1.6 mg), terpyridine (1.2 mg), K₂CO₃ (2.8 mg), 2,4,6-trimethylbenzenediazonium tetrafluoroborate (4.7 mg), DMSO (0.5 mL), 60 °C, 20 min; Radiochemical yield (RCY) and product identity were determined by radio-high-performance liquid chromatography (HPLC).

^bReaction conditions: alkyne (2 mg), K₂CO₃/kryptofix 222 (1/5 mg), DMSO (0.4 mL), 150 °C, 10 min; Radiochemical yield of the isolated product was reported as decay-corrected, and the isolation was performed by radio-HPLC.

CONCLUSIONS

We report herein a widely applicable approach for the Cu-catalyzed Sonogashira-type cross-coupling of unactivated alkyl iodides with alkynes. This method harnessed the iodine abstraction ability of aryl radicals to allow the participation of unactivated alkyl iodides in Cu-catalyzed Sonogashira coupling reactions. This aryl-radical-enabled coupling reaction demonstrated high functional group tolerance and empowered the late-stage coupling of densely functionalized molecules. Mechanistic studies were consistent with the transfer of alkyne groups from Cu-acetylide species to alkyl radicals, which were generated via an aryl radical activation pathway. Moreover, given the fast kinetics of the aryl radical-enabled coupling reaction, this protocol has also been applied to the conjugation of [¹⁸F]-labeled molecules for the synthesis of novel PET tracers. Given the rich medicinal potential of functionalized alkynes as well as the broad availability of alkyl halides and terminal alkynes, we anticipate that this protocol will find wide application in drug development. From a broader perspective, we envision that this aryl radical activation strategy will be of great utility in both modern synthetic chemistry and pharmaceutical research.