Alkyne Trifunctionalization via Divergent Gold Catalysis: Combining π-Acid Activation, Vinyl-Gold Addition and Redox Catalysis

Abstract

Here we report the first example of alkyne trifunctionalization through simultaneous construction of C-C, C-O, and C-N bonds via gold catalysis. With the assistance of a γ-keto directing group, sequential gold-catalyzed alkyne hydration, vinyl-gold nucleophilic addition and gold(III) reductive elimination were achieved in one pot. Diazonium salts were identified as both electrophiles (N source) and oxidants (C source). Vinyl-gold(III) intermediates were revealed as effective nucleophiles toward diazonium, facilitating nucleophilic addition and reductive elimination with high efficiency. The rather comprehensive reaction sequence was achieved with excellent yields (up to 95%), broad scope (>50 examples) under mild conditions (rt or 40 °C).

Graphical Abstract

INTRODUCTION

Alkynes represent an important group of compounds in chemical synthesis.^1–6 Over the past decade, alkyne functionalization has gained increasing attention for easy access to complicated molecular scaffolds.^{3, 7–10} Although alkyne difunctionalization is more commonly reported, its trifunctionalization has been rare.¹¹ Among the reported examples, rigorous degassing or harsh reaction conditions are generally required, which limits the application of these methods.¹² Moreover, previous literature mainly focused on C-O, C-X bond formation. Therefore, an efficient and practically useful approach for alkyne trifunctionalization involving C-C bond formation is highly desirable.

Throughout the first two decades of this century, gold catalysis has enjoyed the spotlight of organometallic chemistry, mainly due to its exceptional ability to activate alkynes and allenes under mild conditions.^13–17 The nucleophilic addition toward alkyne-gold π-complex followed by protodeauration has been applied as an efficient strategy in many important chemical transformations (Scheme1A, paths A and B).^{1, 18–22} Recently, with the joint efforts from numerous groups around the world, gold redox chemistry has gained increasing attention for its unique reactivity, arguably opening a new era of homogeneous gold catalysis.^23–28 The combination of π-activation and redox chemistry will intrinsically initiate new strategies for efficient and versatile transformations.

Scheme 1.

Representative reaction paths in gold catalysis

With the high oxidation potential between Au^I and Au^III (1.4 eV), strong oxidants are usually required in gold redox chemistry, leading to incompatibility with many nucleophiles for gold-activated alkynes.^29–37 Moreover, protodeauration has been a serious problem for any subsequent transformations using intermediates containing C-Au bonds.^38–41 Therefore, limited success was achieved combining gold π-activation and redox catalysis.^42–49 On the other hand, the good electron donating ability of gold cation toward C=C bond renders vinyl-gold a potentially good nucleophile, provided that protodeauration could be prevented. Although few examples have been reported so far, nucleophilic addition of vinyl gold provides an interesting new reaction path that further enriches the versatile reactivity of gold catalysis.^50–55 Clearly, the possibility of integrating all three gold reaction modes (π-activation, redox catalysis, and vinyl gold nucleophilic addition) is not only mechanistically fascinating but also practically appealing, featuring multiple functionalizations in one step. Herein, we report the concurrent construction of C-O, C-N, and C-C bonds from unactivated alkynes through the divergent gold catalysis. To the best of our knowledge, this is the first example of a transformation that successfully combines three gold reaction modes in a single step. The alkyne trifunctionalization products were obtained with high efficiency (up to 95%), broad substrate scope (>50 examples) under mild conditions (Scheme 1B).

RESULTS AND DISCUSSION

Reaction Development.

To achieve the integration of multiple gold reactivities, the compatibility of vinyl-gold nucleophilic addition (path C) is the most crucial factor. Due to the competing protodeauration, successful vinyl-gold nucleophilic addition examples are extremely rare in literature, especially in a catalytic fashion. Recently, with the introduction of γ-keto directing group (DG), our group reported the first intermolecular crossed aldol reaction using synergistic gold/iron catalysis (Figure 1A, rt, open air, neutral solution).⁵⁵ Fe(acac)₃ was identified as the critical co-catalyst to effectively prevent C-Au bond from protodeauration. The γ-keto directing group is also crucial as the reaction of phenylacetylene gave only hydration product 3a’ under identical conditions with no aldol product 2a’ observed.

Figure 1.

Vinyl-gold reactivity toward different electrophiles

Encouraged by this result, we set our goal to explore the possibility of integrating vinyl-gold chemistry with gold redox catalysis by conducting reactions of 1a and benzaldehyde under various reported gold redox conditions (Figure 1B). Initially, the ligand-assisted aryl iodide oxidative addition protocol, recently reported by Bourissou and Patel,^56–61 was tested. Unfortunately, dominant hydration product 3a was obtained with no vinyl gold addition product or reductive elimination product observed. Then photo-initiated diazonium activation was also performed, giving similar results with 3a as the dominant product. Both results highlighted the significance and necessity of preventing the rapid protodeauration to access any valid vinyl-gold reactivity. Interestingly, carefully monitoring the reaction under photocatalytic conditions revealed the formation of a new product 4a though in a very low yield. X-ray diffraction identified the structure of this compound as at result of potential vinyl-gold addition toward diazonium. Notably, aldehyde addition product 2a or 2aa were not obtained in all cases with diazonium presented, suggesting that diazonium salt might be a more reactive electrophile over aldehyde toward vinyl-gold.

To prevent protodeauration, Fe(acac)₃ co-catalyst strategy was applied to the reaction with diazonium salt. Indeed, addition of 10% Fe(acac)₃ raised the yield of 4a to 34%. Unfortunately, Fe(acac)₃ quenched the photocatalytic reactivity and gave almost no conversion over time, suggesting its poor compatibility with photo-promoted diazonium activation (see SI). Our group has recently reported the base-assisted diazonium activation as an alternative approach to achieve gold oxidation.^62–65 It is expected that conducting the reaction under basic conditions will not only help diazonium activation, but also reduce the rate of protodeauration. As expected, addition of Li₂CO₃ (1 eq) further improved the yield of 4a to 79% along with formation of 5a (reductive elimination product) for the first time.

These exploratory studies were inspiring to us in the following aspects: A) diazonium is a valid electrophile for vinyl gold addition (new C-N bond formation strategy) B) the integration of gold redox catalysis is theoretically feasible (formation of 5a). To develop this new reactivity, we first focused our efforts on the vinyl-gold(I) diazonium addition. Reactions between 1a and aryl diazonium salt were screened under various conditions (see details in SI). The combination of (p-CF₃-C₆H₄)₃PAuNTf₂ catalyst (5%) and 10% Fe(acac)₃ was identified as the optimal condition, giving the desired product 4a in 88% isolated yields (5% of 3a). The performance of some alternative conditions is summarized in Table 1.

Table 1.

Optimization of reaction conditions^a,b

entry	Variations	4a (%)	3a (%)	5a (%)
1	None [L=P(p-CF3-C6H4)3]	93(88)c	5	n.d.
2	L=PPh3	83	15	n.d.
3	L=JohnPhos	35	60	trace
4	L=IPr	0	80	n.d.
5	L=P(ArO)3	0	93	n.d.
6	L=PCy3	85	12	trace
7	L =P(p-F-C6H4)3	90	6	n.d.
8	Na2CO3 instead of Li2CO3	34	25	10
9	Other bases	<5	<5	n.d.
10	1.0 eq H2O	80	12	n.d.
11	5.0 eq H2O	78	20	n.d.
12	Other solvents (see SI)	<50	>40	<10
13	40 °C or higher	<78	>8	<15

^aStandard conditions: 5% (p-CF₃C₆H₄)₃PAuNTf₂, 10% Fe(acac)₃, 3 equiv. water and 0.2 mmol Li₂CO₃ were added to a THF solution (0.5 mL) of alkyne 1a (0.2 mmol) and aryl diazonium salt (0.4 mmol), and reaction was kept stirring at room temperature for 6 h.

^bConversion and yields were determined by ¹⁹F NMR and ¹H NMR spectroscopy using trifluorobenzene and 1,3,5-trimethoxybenzene correspondingly as internal standard.

^cIsolated yield

As shown in Table 1, the primary ligand on gold plays a pivotal role with P(p-CF₃-C₆H₄)₃AuNTf₂ giving the best yield, suggesting the balance of electron density on gold is critical (entries 1–7). The [L-Au-L’]⁺ type of catalysts, such as [LAu(TA)]⁺ and [LAu(CH₃CN)]⁺, were also tested, giving dominantly hydration product 3a (see SI). The choice of base was also proved to be vital. While Li₂CO₃ was the optimal base, Na₂CO₃ gave significantly reduced yield of 4a (entry 8), presumably due to the reaction with [Fe] co-catalyst under stronger basic conditions. Some alternative organic and inorganic bases were also tested, but gave poor performance (see SI). Since the reaction involved both alkyne hydration and vinyl-gold nucleophilic addition, the amount of water should be important to reach the ideal balance between vinyl gold addition and minimal protodeauration. After screening, 3 equivalents of water gave the best result with THF as solvent. Notably, conducting the reaction at elevated temperature (40 °C) gave arylhydrazone 5a as the minor product, suggesting the feasibility of the proposed alternative gold redox process (entry 13). With the optimal condition in hand, the substrate scope of alkynes and aryl diazonium salts were evaluated. The results are summarized in Scheme 2.

Scheme 2.

Reaction scope of difunctionalization^a,b

^aStandard conditions: 5% (p-CF₃-C₆H₄)₃PAuNTf₂, 10% Fe(acac)₃, 3 equiv. water and 0.2 mmol Li₂CO₃ were added to a THF solution (0.5 mL) of alkyne 1a (0.2 mmol) and aryl diazonium salt (0.4 mmol), and reaction was running at room temperature for 6 h. ^bIsolated yields. ^c 40 °C

An extensive survey of the reaction scope was presented with three variations: diazonium, directing group and alkyne. A series of aryl diazonium salts were prepared as coupling partners with alkyne 1a. In general, the reaction worked well for electron-deficient aryl diazoniums, affording aryl hydrazones 4a-4o in good to excellent yields. Electron-rich aryl diazoniums offered <30% yields (4p) due to reduced electrophilicity. Remarkably, with Fe(acac)₃ as co-catalyst, the reaction even tolerated weakly acidic functional groups, giving the desired addition product (4i) with little hydration byproducts observed. Variations on γ-keto directing groups have also been evaluated. Most aryl substituted carbonyl directing groups (DG) are tolerated, giving the desired product in good to excellent yields (4aa-4ai, 4an). Aliphatic carbonyl DG also worked though with slightly reduced yields (4aj,4ao). Impressively, a variety of substrates containing functional groups, such as cyclopropyl, alkynyl, and allyl groups, gave desired products in good yields (4ak-4am). This result highlighted the mild condition and excellent functional group tolerance of this method.

For many reported gold catalysis, activation of internal alkynes remains a formidable task due to reduced reactivity, especially under mild conditions. In our case, both aryl and alkyl substituted internal alkynes provided good to excellent yields (4ba-4bl, 5a). Notably, single regio-isomers consistent with the 5-exo-dig activation mechanism were obtained in all cases (no 6-endo-dig products observed). The room temperature activation of internal alkynes highlighted the reactivity boosting effect upon our strategic application of γ-keto directing group. Similar to our previous observation, extending the keto DG to δ-position (one extra carbon) led to dramatic reduction of coupling product yield (dominant hydration, see Supporting Information), emphasizing the choice of γ-keto directing group in achieving vinyl-gold addition. Having applied the gold/iron dual catalytic system effectively to vinyl-gold diazonium addition, we moved on to the more challenging alkyne trifunctionalization integrating redox catalysis.

As shown in Scheme 1A, the alkyne trifunctionalization can be achieved from two slightly different reaction sequences: 1) vinyl-Au^I oxidation to [vinyl-Au^III-aryl] followed by reductive elimination, 2) in-situ formation of [Ar-Au^III] π-acid for alkyne activation followed by reductive elimination.⁶⁶ Although the former strategy is extremely attractive for rapid, simultaneous bond construction, to the best of our knowledge, no catalytic reaction through this path has been realized yet, likely due to various competing side reactions. The vinyl-gold addition to diazonium shown above provided a good platform to explore this new transformation since diazonium could serve as potential oxidant for gold redox chemistry under proper conditions. Very recently, Hashmi group reported an interesting example of vinyl gold addition to diazonium for the formation of C-N bond, which strongly supported this hypothesis.⁶⁷ However, there are two major mechanistic concerns: A) protodeauration remains the key challenge for any reaction sequence involving Au-C bond, and B) gold(III) reductive elimination is a fast process due to its high redox potential. Based on this analysis, we proposed that in-situ formation of [Ar-Au^III] for alkyne activation is preferred in a practical sense for alkyne trifunctionalization. To testify this idea, L-Au^I-Cl type catalysts were used to react with diazonium salts with the hope to generate active [Au^III-Ar] prior to alkyne activation. After tuning several typical reaction parameters, trifunctionalization product 5a was obtained in excellent yields (94% isolated yields) with Cy₃PAuCl as catalyst, 3.0 eq of diazonium salt under basic conditions (2 eq Li₂CO₃) at 40 °C. Results from typical alternative conditions are shown in Table 2 (see detailed screening conditions in Supporting Information).

Table 2.

Alkyne trifunctionalization with gold redox catalysis^a,b

entry	variation	5a (%)	3a (%)
1	none	95 (94)	2
2	No Li2CO3, Ru(bpy)3PF6, Blue LED	0	48
3	PPh3AuCl	90	7
4	JohnPhosAuCl	trace	35
5	IPrAuCl	10	50
6	(ArO)3PAuCl	0	0
7	(4-CF3C6H4)3AuCl	85	12
8	3.0 eq H2O	88	8
9	Other solvents (see SI)	<89	>8
10	Na2CO3 instead of Li2CO3	30	15
11	Other Bases (see SI)	<10	< 15
12	50 °C or higher	<67	>20

^aStandard conditions: 5% Cy₃PAuCl, 1 equiv. water and 0.4 mmol Li₂CO₃ were added to a MeCN solution (0.6 mL) of alkyne 1a (0.2 mmol) and aryl diazonium salt (0.6 mmol), and reaction was running at 40 °C for 3 h.

^bConversion and yields were determined by ¹⁹F NMR and ¹H NMR spectroscopy using trifluorobenzene and 1,3,5-trimethoxybenzene correspondingly as internal standard.

Firstly, the typical photocatalytic diazonium activation was not suitable for this transformation (entry 2), forming dominatly ketone 3a as the hydration product. Secondly, our previously developed base-promoted conditions for thermal activation of diazonium proved essential in this transformation. While Cy₃PAuCl appeared to be the optimal catalyst (entries 3–7), the choice of solvent, base and amount of water also had huge impacts on the reaction performance (see detailed screening in Supporting Information). Higher reaction temperature (>50 °C) significantly raised the hydration rate. Nevertheless, under the optimal condition, the trifunctionalization of alkyne was successfully achieved with high efficiency. Based on this optimal condition, reaction scope was explored. The results are summarized in Scheme 3.

Scheme 3.

Reaction scope of alkyne tri-functionalization^a,b

^aStandard conditions: 5% Cy₃PAuCl, 1 equiv. water and 0.4 mmol Li₂CO₃ were added to a MeCN solution (0.6 mL) of alkyne 1a (0.2 mmol) and aryl diazonium (0.6 mmol), and reaction was running at 40 °C for 3 h. ^bIsolated yields.

The reaction scope was also evaluated through three variations: diazonium salts, directing groups, and internal alkynes. A variety of aryl diazonium salts were tested as reaction partner with alkyne 1a. Compared to alkyne difunctionalization, this reaction exhibited a broader scope with respect to, presumably owing to the higher reactivity of vinyl gold(III) intermediates. In general, the reaction worked well for both electron-deficient and electron-rich aryl diazonium salts. The reaction of electron-deficient diazonium salts afforded aryl hydrazones 5a-5i with good to excellent yields, while electron-rich diazonium salt gave a moderate yield (5j). Interestingly, ortho and meta substituted aryl diazonium salts, which were unsatisfactory substrates in base promoted gold-redox process, worked well in this transformation (5k-5n). Di-substitued aryl diazonium salts also exhibit decent reactivity (5o-5r). Furthermore, substrates with diverse γ-carbonyl DG including both aryl (including heterocycles) and aliphatic substituents displayed good to excellent yields (5aa-5an). Notably, although Ar-Au^III is a superior π-acid which can activate alkenes, great chemoselectivity was observed with carbon-carbon double bond remaining intact in this transformation (5al). Some other ‘vulnerable’ functional groups such as cyclopropyl and alkynyl also survived under the reaction condition, exemplifying mildness and robustness of this transformation. Significantly, when treating internal alkynes with aryl diazonium 1a, azo compounds containing a quaternary carbon center were formed (5ba-5bq). The electron-deficient aryl-substituted internal alkynes dominantly went through 5-exo-dig cyclization, providing azo compounds in good to excellent yields (5ba-5bg, 5bj-5bm). In contrast, less electron-deficient aryl-substituted internal alkynes afforded azo products in slightly reduced yields as 5:1 and 3.5:1 regioisomers (1,4-diketones from 5-exo addition vs 1,5-diketones from 6-endo addition), respectively (5bh and 5bi). Electron-rich aryl-substituted internal alkyne (5bo) gave exclusively 1,5-diketone in a good yield. Impressively, internal alkynes with cyclic ketone directing group reacted with diazonium salts, giving desired product 5bq in modest yield but with good diastereoselectivity. This result featured stereoselectivity improvement of vinyl gold nucleophilic addition by the 5-exo-cyclization mechanism besides the aforementioned ‘boosting effect’. On the other hand, the low diastereoselectivity of product 5bp was likely an outcome of α-carbon epimerization.

Mechanistic studies and functional group conversions.

Under photochemical (blue LED) or thermal conditions, arylgold(III) species were formed by treating aryldiazonium with gold(I) choloride complex.^{66, 68} To gain more mechanistic insight, we monitored the reaction with nESI-MS (Figure 2A). Diagnostic signals of [Au^III-Ar] catalyst and key intermediates A and B were all successfully captured, whereas their structure compositions confirmed by CID (MS-MS, see details in SI). These HRMS results greatly supported the proposed reaction sequence with formation of [Au^III-Ar] followed by alkyne π-activation, vinyl-gold nucleophilic addition toward diazonium salts and final Au^III reductive elimination.

Figure 2.

Derivatization and proposed mechanism

Apart from its mechanistic novelty, the resulting 1,4-diketone products contain versatile functional groups, which make them interesting building blocks for complex molecular skeleton construction. It is well known that 1,4-diketone is a good synthon for the synthesis of hetereocycles through double condensiaiton by reacting with amine or hydrazine.^{69, 70} This approach was further confirmed by the example shown in Figure 2B: treating 4a with Lawesson’s reagent produces hydrozone substituted thiophene 6a, a corrector of a mutant cystic fibrosis transmembrane conductance regulator (CFTR), in excellent yield.⁷¹ To explore the synthetic applications of this new method, we put our efforts on exploring new transformations that are uniquely associated with the resulting products from the following three perspectives: 1) selective carbonyl reduction over hydrazone; 2) effective protocol for directing group removal, and; 3) selective hydrazone reduction to N-N bond.

For effective transformations of the resulting products, one practical concern is the selective reduction of ketone (C=O) over C=N and N=N, or vice versa. After exploring multiple typical reduction conditions, we are pleased to identify Pd/C/H₂ as an optimal condition for selective imine-ketone reduction (5a to 6b, 90%). In addition, NaBH₄ was identified as the optimal reagent for ketone reduction over imine, giving 6c in 93% yields. Interestingly, the resulting diol 6c could be readily coverted into substituted THF 6d, bearing imine moiety with no hydration. The result greatly enriched good synthetic potential of this method by introducing multiple functional groups in one simple step. Notably, the hydrazones (6a-6d) were all assigned as E isomer based on X-ray single crystals of starting material 5a. Although the γ-ketone directing group is an useful synthetic handle as demonstrated above, it could be more attractive if an effective directing group removal strategy could be developed to further extend the synthetic applications. After extensive condition screening, aqueous HCl (3N) was identified as the practical condition for γ-ketone dissociation, giving hydrazones (7a-7c) in good to excellent yields. Notably, this method not only gives an effective approach to prepare hydrazones bearing more reactive aldehyde (traditional condensation will not work), but also allows the rapid construction formation of alkyl-aryl hydrazone (7b, 7c), which could be readily converted into substituted cinnoline 8 using reported conditions (two steps, high yields). The hydrazones (7b-7c) were assigned as Z configuration based on literature reported.⁷²

Finaly, while the reduction of N-N to NH₂ are well-studied process with numerous conditions reported,^73–75 it is important to develop practical conditions for selective N=N hydrogenation to N-N single bond with C=O presented. After exploring multiple typical reduction conditions, we are pleased to discover that the azo compounds bearing tertiary carbon (such as 5ba, prepared from internal alkynes) could be readily reduced to N-N single bond product 7d without further reduction of C=O (92%). This example highlightes the diverse substitution pattern (two different aryl groups)of the reporting method and provides a practical synthesis of substituted hydrazine and amine derivatives.

CONCLUSION

In summary, we disclosed herein a divergent gold catalysis for alkyne trifunctionalization. The overall sequence combined gold π-activation, vinyl gold nucleophilic addition, and gold(III) reductive elimination. Aryl diazonium salts were identified as the effective electrophile and oxidant. Besides the ability to construct complex molecular motifs within simple steps, this new method demonstrated a new strategy to apply vinyl-Au^III as effective nucleophiles for new bonds construction. This new design principle of gold catalysis could certainly be applied to other gold promoted chemical transformations, which is currently under investigation in our lab.