Synthesis of 1,3-Diynes Via Cadiot-Chodkiewicz Coupling of Volatile, in Situ-Generated Bromoalkynes

Abstract

A convenient Cadiot-Chodkiewicz protocol that facilitates the use of low molecular weight alkyne coupling partners is described. The method entails an in situ elimination from a dibromoolefin precursor and immediately subjecting to copper-catalyzed conditions, circumventing the hazards of volatile brominated alkynes. The scope of this method is described, and the internal 1,3-diyne products are preliminarily evaluated in ruthenium-catalyzed azide alkyne cycloadditions.

Graphical Abstract

In organic synthesis, 1,3-diynes have been attractive targets for several decades.¹ The structural motif appears in many natural products,^1b bioactive molecules,² intriguing materials³ and supra-molecular tools,^1a,4 and it has been exploited as a central functionality in several complexity-generating hexahydro-Diels-Alder cycloadditions.^5,6 The Glaser-Eglington-Hay coupling represents a straightforward synthesis of homo-coupled 1,3-diynes,⁷ but producing cross-coupled diynes is more complicated. Among methods for accessing the cross-coupled variant, the Cadiot-Chodkiewicz coupling⁸ stands out as the most widely employed, exemplified by its use in a number of natural product syntheses.⁹ Recent examples utilizing this coupling include polyyne natural products ivorenolide A^9a,b and 4,8-dihydroxy-3,4-dihydrovernoniyne (Figure 1).^9c This coupling strategy has also been employed to provide important intermediates en route to natural products, such as toward the selaginpulvilins^9d and the dictyodendrins.^9e

Figure 1.

1,3-Diynes in natural product synthesis.

During the course of our synthesis of gelsenicine,¹⁰ our strategy evolved to utilize a diene-diyne as a key precursor (Figure 2). The Cadiot-Chodkiewicz reaction appeared ideal for accessing this unsymmetrical diyne, although we were wary of potential issues. Namely, it is well appreciated that homocoupling can complicate these transformations, particularly in cases where the reactants have similar electronic characteristics (e.g., aryl-aryl, alkyl-alkyl).^1a,b,11 We considered two possible sets of coupling precursors (Figure 2a). In Case 1, a bromoalkyne would be coupled with propyne, but we anticipated this strategy would be complicated by the latter’s low boiling point (−23 °C).¹² Case 2, where we switch the bromination of the coupling partners, appeared more attractive from this perspective. Indeed, 1-bromopropyne is known, but previous reports¹³ indicated serious hazardous properties.¹⁴ To circumvent the handling of this compound, we conceived of an in situ approach involving direct elimination from a dibromoolefin and immediate coupling (Figure 2b). In several copper- and palladium-catalyzed reactions, 1,1-dibromo-1-alkenes have been used to generate alkynyl electrophiles, suggesting promise for this strategy.^15,16 Herein, we report the development of this elimination/coupling procedure, used successfully in our total synthesis effort,¹⁰ and we demonstrate its viability as a method to directly couple volatile alkyne substrates. We also describe preliminary reactivity of these diyne compounds in Ru-catalyzed azide-alkyne cycloadditions, highlighting that 1,2,3-triazoles can be obtained from these diynes in a chemo- and regioselective fashion.

Figure 2.

1,3-Diynes from volatile bromoalkynes - synthetic strategy.

There have been a few reports using 1,1-dibromoalkenes to generate 1-bromo-1-alkynes in situ as precursors for 1,3-diynes.^16a–c Unfortunately, these strategies require excess terminal alkyne or elevated temperature (70–100 °C), and none utilized a volatile haloalkyne intermediate. These cases contrasted with conditional stipulations in our desired coupling. First, the terminal alkyne needed to be used as the limiting reagent, chiefly because that coupling component had already been advanced multiple steps in our total synthesis campaign. Second, the temperature needed to be kept low so as to maximize the quantity of volatile alkyne in solution. Although we would consider all options, we felt that in general we wanted to steer away from Pd-based conditions that could result in byproduct formation from enyne reactivity.

Our optimization studies are presented in Table 1. Our strategy here involved first testing with hydrocinnamaldehyde-based dibromoolefin 2a, whose in situ elimination would be more straightforward to monitor by conventional methods (i.e., TLC). NaH¹⁷ did not induce the elimination (entry 1), while TBAF•3H2O in DMF¹⁸ was incompatible with the coupling. We found that LiHMDS in THF¹⁹ could induce the elimination at the desired lower temperature, and the desired 1,3-diyne product was observed in low yield when this was subjected to Cu(I) coupling conditions (CuCl, NH₂ OH•HCl, DMF,²⁰ entry 3). Palladium/copper mixed systems^11c,21 were not as effective (entries 4, 5). However, we found that using aqueous conditions with Cu(I)^1d in the coupling process were ultimately the most promising (entries 7, 8).

Table 1.

Reaction optimization.

entry	equiv 1	base (equiv)	catalyst mol%	mol% NH2OH•HCI	equiv n-BuNH2	solvent	time (h)	isolated yield (%)
R = −CH2CH2Ph (a)
1	—	NaH (2.0)	—	—	—	—	—	—b
2	2.0	TBAF•3H2O (10)	CuCl (20)	30	30	DMF	24	0
3	2.0	LiHMDS (2.0)	CuCl (20)	30	2	DMF	24	21
4	2.0	LiHMDS (2.0)	CuCl (2)	0	2c	DMF	24	0
			Pd(dba)2 (4)
5	2.0	LiHMDS (2.0)	CuCl (20)	30	2	THF	18	14
			Pd(dba)2 (4)
6	2.0	LiHMDS (2.0)	CuCl (50)	150	10	THF	18	28
7	2.0	LiHMDS (2.0)	CuCl (20)	30	10	H2O/THF	18	49
8	2.0	LiHMDS (2.0)	CuCl (20)	30	20	H2O/THF	3	80
R = −CH3 (b)
9	2.0	LiHMDS (2.0)	CuCl (20)	30	20	H2O/THF	18	44
10	2.5	LiHMDS (2.5)	CuCl (20)	30	20	H2O/THF	1	27
11	1.0	LiHMDS (1.0)	CuCl (20)	30	20	H2O/THF	1	57
12	1.5	LiHMDS (1.5)	CuCl (20)	30	20	H2O/THF	1	69
13	1.5	LiHMDS(1.5)	CuCl (40)	60	20	H2O/THF	1	72

^aElimination conditions: NaH: THF, 0 to 23 °C, 24 h; TBAF•3H₂O: DMF, 60 °C, 3 h; LiHMDS: THF, −78 to 0 °C, 1.5 h.

^bBromoalkyne intermediate 2a was not observed; coupling was not attempted.

^cEt₃N used as base instead of n-BuNH₂.

With this lead, we then switched to the volatile 1-bromopropyne (Table 1, entries 9–13). Because the dibromoolefin is also relatively volatile, we found that the simplest protocol was to synthesize this directly from acetaldehyde using the zinc-based dibromoolefination conditions originally described by Corey and Fuchs (eq 1).^22,23

Simply passing the reaction mixture through a plug of silica gel to remove the phosphine byproducts and removing the volatiles afforded the crude dibromoolefin in sufficient purity to be advanced directly. Testing this compound in the above optimized conditions showed they were not as effective. We hypothesized that using a significant excess of the in situ generated bromopropyne was not advantageous, and eventually we found that 1.5 equiv was the optimal amount to employ with an increased loading of CuCl/NH2OH•HCl, to ensure efficient reaction completion.

(1)

With this protocol fully optimized, we proceed to evaluate this coupling of in situ generated 1-bromopropyne with a range of terminal alkynes (Scheme 1). Aryl alkynes were generally effective. Electron-neutral aryl alkyne coupling partners afforded the 1,3-diyne product in good yields (7a,b). Comparatively, terminal alkynes with more electron-rich arenes were lower yielding (7d,e), while electron-deficient arenes were tolerated unless the withdrawing functionality was sensitive to reduction (i.e., 7f). Ortho-substituted aryl bromide 7j could be obtained, a compound with a convenient coupling handle for potential further manipulations.

Scheme 1.

Coupling of in situ-generated 1-bromopropyne. ^a CH₂Cl₂ was added to the coupling reaction to solubilize the reactants. ^b Coupling stirred 4 h at 0 °C and 2 h at 23 °C. ^c Gram scale.

Pyridine 7k was formed in 69% yield, and enynes were also effective (7l). Alcohols and ethers were successful alkyne coupling partners, where the products could be formed in generally good yields. Notably, the transformation also performs reasonably consistent upon scaleup; ether 7p was formed in 76% yield on gram-scale.²⁴ Propiolates and silylalkynes were ineffective in this coupling procedure (7t,u).²⁵

We also evaluated our coupling procedure using in situ generated 1-bromobutyne. 1-Bromobutyne is relatively volatile (bp = 90–91 °C), but it is not reported to have similar hazardous properties as those of 1-bromopropyne. However, the method can serve as a useful strategy to circumvent handling 1-butyne (bp = 8 °C). Scheme 2 illustrates examples of couplings originating from 1,1-dibromobutene. Aryl and heteroaryl alkynes were again successful. Propargylic alcohols were coupled effectively to afford products 8d-f in excellent yields. Benzyl ether 8g was also formed, albeit in diminished yield compared to the related propynyl example.

Scheme 2.

Coupling of in situ-generated 1-bromobutyne.

1,3-Diynes have been shown to be precursors to thiophenes and pyrroles, among other transformations.^1c Toward further expanding the potential applications of these compounds, we sought to investigate 1,3-diyne reactivity in metal-catalyzed azide-alkyne cycloadditions.²⁶ While a few cases using other catalyst systems for intermolecular azide-alkyne cycloadditions with internal alkynes have been reported,²⁷ ruthenium-catalyzed systems have been by far the most prominent.²⁸ Internal 1,3-diynes can pose intriguing questions regarding cycloaddition reactivity, with respect to both chemoselectivity and regioselectivity.²⁹ In a standard 1,3-diyne cycloaddition, there can be up to eight possible products. Selective reactivity has been successfully addressed to some extent with terminal 1,3-diynes using Cu catalysis via acetylide formation,³⁰ but that mechanistic strategy cannot be extended to internal alkynes. Other selectivity modes, however, may be achievable.³¹

When propargylic alcohol 7o was subjected to Ru-catalyzed conditions as described by Fokin, an approx. 5:1 ratio of triazole products 9 and 10 were observed (Scheme 3). Cu-catalyzed cycloaddition attempts using this diyne were unsuccessful.^29a Interestingly, in the Ru case only the alcohol-adjacent alkyne was reactive; the other alkyne remained intact. Our observation matched hypotheses regarding cycloaddition regioselectivity for internal alkynes with Ru catalysis,^32,33 where hydrogen bond donors are presumed to enable catalyst coordination with the ligating chloride to favor the regioselectivity outcome for the observed triazole isomer (intermediate 11). Notably, the alcohol imparting chemoselectivity between two alkynes has not been observed in this transformation. When benzyl ether diyne 7q was utilized, which lacks this hydrogen bond donor, even better regioselectivity was observed (12). The alkyne chemoselectivity was somewhat compromised, however, in that the second alkyne became competitive in cycloaddition as triazole 12 was being formed. Regioselectivities for both additions were excellent (>19:1); origins of these effects are presently unclear. Mechanistically, the most nucleophilic car bon of an alkyne precursor is expected to become C-4 in the triazole,³⁴ but assessments of relative nucleophilicities of the alkyne carbons were inconclusive.³⁵ To our knowledge, ethereal groups have not been invoked as catalyst directing moities in these cycloadditions, but they could be influential here. In this preliminary study, it appears that synthetically useful levels of selectivity can be attained; factors that impact chemo- and regioselectivity in these azide-alkyne cycloadditions appear to be somewhat interconnected and merit further investigation.

Scheme 3.

1,3-Diynes in Azide-Alkyne Cycloadditions

In summary, we have developed a useful addition to the Cadiot-Chodkiewicz alkyne-alkyne cross coupling. The in situ generation of the reportedly hazardous 1-bromo-1-propyne intermediate allows an efficient and convenient workaround to using volatile unit alkynes as coupling partners. We have also showed that this method can be readily scaled for applications in complex molecule synthesis. Preliminary studies in azide-alkyne cycloadditions demonstrate that synthetically useful chemo- and regioselectivities can be achieved. Further investigations of the synthetic utility of these 1,3-diynes are currently underway.