Synergistic Peptide and Gold Catalysis: Enantioselective Addition of Branched Aldehydes to Allenamides.

The combination of a peptide catalyst and a gold catalyst is presented for enantioselective addition reactions between branched aldehydes and allenamides. The two catalysts act in concert to provide γ,δ-enamide aldehydes bearing a fully substituted, benzylic stereogenic center - a structural motif common in many natural products and therapeutically active compounds - with good yields and enantioselectivities. The reaction tolerates a variety of alkyl and alkoxy substituted aldehydes and the products can be elaborated into several chiral building blocks bearing either 1,4- or 1,5- functional group relationships. Mechanistic studies showed that the conformational features of the peptide are important for both the catalytic efficiency and stereochemistry, while a balance of acid/base additives is key for ensuring formation of the desired product over undesired side reactions.

Fully substituted, benzylic stereogenic centers are important building blocks of many natural products and biologically active compounds (Figure 1a). These structural motifs have become the target of a variety of synthetic methodologies.[ , , ] Among them, an attractive approach relies on amine catalyzed reactions between branched aldehydes and suitable carbon‐based electrophiles, which proceed through reactive enamine intermediates.[ , ] In recent years, the combination of a metal catalyst with an amine catalyst has expanded the scope to include otherwise unreactive electrophiles.[ , ] Interesting electrophiles to be used in these types of reactions are allenamides, which can act as synthetically versatile C‐2 or C‐3 synthons through synthetic elaboration of the enamide moiety formed after C−C bond formation (Figure 1b). Recently, two elegant studies by Mascareñas/López and González showcased the feasibility of reactions between branched aldehydes and allenamides catalyzed by prolinol silyl ethers and Au‐based catalysts. These reports also pointed out that dual catalysis is challenging since the organocatalyst and the metal catalyst must act in concert and not interfere with each other.[ , ] We reasoned that amine‐based catalysts, in which the amino group is shielded and therefore less prone to interact with the metal catalyst, may offer advantages in this reaction.

Figure 1

a) Selected therapeutically active compounds containing fully substituted benzylic stereogenic centers. b) The enantioselective addition of α‐branched aldehydes to allenamides. c) Tripeptide catalyst H‐dPro‐Pro‐Glu−NH2 and its conformation.

Our group developed the peptide H‐dPro‐Pro‐Glu‐NH2, which is a highly efficient and stereoselective organocatalyst for C−C bond formations that rely on the formation of enamine intermediates.[ , ] Detailed NMR spectroscopic analyses revealed that this peptide adopts a stable ground state conformation, in which the amine moiety forms a salt bridge with the glutamic acid side chain (Figure 1c). We envisioned that this intramolecular coordination would disfavour non‐productive interactions between the peptide and the metal center and that as a result, peptides of the type H‐dPro‐Pro‐Xaa would be efficient catalysts for the targeted transformations.

Herein, we report the stereoselective addition of branched aldehydes to allenamides, catalyzed by a synergistic combination of peptide and gold catalysts. The reaction proceeds under mild conditions, tolerates a broad range of alkyl‐aryl and alkoxy‐aryl aldehydes and provides the addition products with good to excellent enantioselectivities. The aldehyde‐enamide products proved to be synthetically versatile for further elaboration into chiral building blocks including diols, lactones and piperidines. We also show that the conformational properties of the tripeptide catalyst and a fine balance between the reaction components (peptide, gold, acid, and base) are key for the efficiency of the reaction.

We started our investigations to evaluate the duality of peptide and metal catalysts using the reaction between 2‐phenylpropanal (1  a) and allenesulfonamide 2  a as a model reaction (Table 1). Firstly, we tested a range of different Cu, Au, Pd, Rh, and Ir complexes known to activate allenamides and/or allenes and used H‐dPro‐Pro‐Glu‐NH2 (4  a) as peptide catalyst (Table 1, entries 1–5, Table S1). Many of the tested metal‐complexes did not allow access to the desired addition product or only in poor yields. Furthermore, dimerization of 2  a to allenamide dimer 6 was observed.[ , ] These findings underscored the challenge of identifying metal complexes that are compatible with amine‐based catalysts. The best results were obtained with gold carbene complex 5  a[ , ] that in combination with peptide 4  a provided a promising conversion of 62 % to γ,δ‐enamide aldehyde 3  a with an enantioselectivity of 47 % (Table 1, entry 2). N‐methyl morpholine (NMM) was included as a base since the trifluoroacetic acid (TFA) salt of the peptide was used.

Table 1

Addition reaction between aldehydes 1  a and 1  b and allenesulfonamide 2  a catalyzed by a combination of a peptide and a metal complex.

Entry	4	5	1	Base	Solvent	Conv. [%][a]	ee (%)[b]
1	4  a	‐	1  a	NMM	CH2Cl2	<5	–
2	4  a	5  a	1  a	NMM	CH2Cl2	62	47
3	4  a	5  b	1  a	NMM	CH2Cl2	22	45
4	4  a	5  c	1  a	NMM	CH2Cl2	<5	–
5	4  a	5  d	1  a	NMM	CH2Cl2	<5	–
6	4  b	5  a	1  a	NMM	CH2Cl2	43	34
7	4  c	5  a	1  a	NMM	CH2Cl2	16	25
8	4  d	5  a	1  a	NMM	CH2Cl2	59	69
9	4  d	5  a	1  a	NMM	CHCl3	64	80
10[c]	4  d	5  a	1  a	NMM	CHCl3	81	79
11[c]	4  d	5  a	1  a	DMAP	CHCl3	(80)	83
12	4  d	5  a	1  b	NMM	CHCl3	(80)	88
13[d]	4  e	5  a	1  b	iPr2NEt	CHCl3	(80)	95

Unless otherwise stated, reactions were carried out using an equal stoichiometry of aldehyde 1  a/1  b and allenesulfonamide 2  a at a concentration of 0.2 M in CH2Cl2. [a] Conversion to 3  a/3  b estimated by 1H NMR spectroscopic analysis of the crude mixture using 1,3,5‐trimethoxybenzene as internal standard. Yields in parentheses. [b] Enantiomeric excess determined by chiral stationary phase SFC. [c] 2 equiv. aldehyde used. [d] 1.5 equiv. aldehyde used, at 0.05 M. NMM=N‐methylmorpholine, DMAP=4‐(dimethylamino)pyridine, Dipp=1,6‐diisopropylphenyl.

‐

–

–

–

Variations of the peptide catalyst revealed that the stereochemistry, the CO2H group and its position as well as the length of the side chain of the C‐terminal amino acid are important for high yield and selectivity (Table 1, entries 2, 6, 7; Table S2). The DLL stereochemistry and the glutamic acid residue (peptide 4  a), features that allow for the intramolecular salt bridge, provided the best results and the lowest amount of side products such as the allenamide dimer 6. Installation of an alkyl chain at the C‐terminus of the peptide (4  d) improved the enantioselectivity of the reaction further (Table 1, entry 8). It is possible that the alkyl chain increases the local hydrophobicity and thereby stabilizes the salt bridge and improves the selectivity of the reaction. Examination of other key reaction parameters revealed that the use of chloroform as solvent increased the enantioselectivity and an excess of the aldehyde improved the yield (Table 1, entries 9 and 10, Tables S5 and S6). TFA, the counterion of the peptide, proved to be important for the reactivity and enantioselectivity and better suited than other acids (e. g., benzoic acid, H2ClCCO2H, Ph2P(O)OH; Table S7). Finally, a screen of different bases, including DBU, DABCO, and Et3N, identified that DMAP is slightly superior to NMM (Table 1, entry 11, Table S8). Under these optimized conditions, γ,δ‐ensulfonamide aldehyde 3  a was obtained in 80 % yield with an enantioselectivity of 83 %.

We then allowed aldehyde 1  b, substituted with a methoxy instead of a methyl group, to react with allenesulfonamide 2  a and obtained the product with an enantioselectivity of 88 % ee (Table 1, entry 12). This stereoselectivity could be further increased to 95 % by using tripeptide 4  e that has an aspartic acid in place of a glutamic acid residue as well as through fine‐tuning of the reaction conditions (aldehyde:allenamide=1.5 : 1, 0.05 M, iPr2NEt as base) (Table 1, entry 13; Tables S10‐S16).

With optimized reaction conditions in hand, we proceeded to evaluate a scope of the reactions between allenamides and alkyl‐aryl aldehydes (Scheme 1) and alkoxy‐aryl aldehydes (Scheme 2). Pleasingly, the products from reactions with branched alkyl‐aryl aldehydes were obtained with good yields and enantioselectivities (3  a, 3  c–3  f, 60–80 % yield, 78–83 % ee; Scheme 1). Even carbamate protecting groups (Boc and Alloc) at the nitrogen in the allenamine were tolerated and provided the enamide products 3  g and 3  h with comparable stereoselectivity. Branched allensulfonamides resulted in only traces of product (3  i), most likely due to the increased steric requirement.

Scheme 1

Scope of the peptide and gold catalyzed addition of branched alkyl‐aryl aldehydes to allenamides. Isolated yields, enantiomeric excess determined by chiral stationary phase SFC.

Scheme 2

Scope of the peptide and gold‐catalyzed addition of branched alkoxy‐aryl aldehydes to allenesulfonamides. Isolated yields, enantiomeric excesses determined by chiral stationary phase SFC.

Also aldehydes bearing a methoxy group and different electron rich and electron poor aryl moieties at Cα formed γ,δ‐ensulfonamide aldehydes with good yields (58‐80 %) and excellent enantioselectivities (90‐95 % ee) (3  b, 3  j–3  l; Scheme 2). Reactions with aldehydes bearing sterically more demanding alkoxy substituent proceeded with moderate to good yields but the high enantioselectivities were retained (3  m‐3  o; Scheme 2). Single crystals of the major enantiomer of 3  b suitable for X‐Ray diffraction could be grown and revealed the absolute stereochemistry at the fully substituted stereogenic center to be (S).[ , ]

We envisioned that the combination of enamide and aldehyde moieties would provide versatile sites for derivatisation of the addition products into a range of different chiral building blocks (Scheme 3). Indeed, ozonolysis of 3  a, followed by reduction afforded 1,4‐diol 7, and acid catalyzed hydrolysis, followed by reduction afforded 1,5‐diol 8. Here, the enamide functional group serves as a C‐2 or C‐3 synthon, respectively. Reductive amination of 3  b under neutral conditions proceeded to afford amino enamide 9. Reduction of 3  b also proceeded smoothly to enamide alcohol 10, which upon ozonolysis and oxidation afforded lactone 11 through the corresponding intermediate lactol. Finally, piperidine 12 was accessed through acid promoted hydrolysis of 3  a in the presence of benzyl amine, followed by reductive cyclisation using Na(CN)BH3. Such 3,3‐disubstituted piperidines have been previously studied as opioid agonists but synthetic access towards them has been limited. In all reactions, only minor changes in the enantioselectivity were observed. These results highlight the versatility of γ,δ‐enamide aldehydes as building blocks to access a broad range of different chiral compounds.

Scheme 3

Derivatisation of products 3  a and 3  b: (a) O3, CH2Cl2/MeOH, −78 °C, then NaBH4. (b) p‐TSA, CH2Cl2, 0 °C, then NaBH4. (c) i) BnNH2, Na2SO4, CH2Cl2, rt, ii) NaBH4, MeOH. (d) NaBH4, MeOH, 0 °C. (e) i) O3, CH2Cl2, −78 °C, then SMe2, ii) CrO3/H2SO4, Me2CO, rt. (f) BnNH2, p‐TSA, CH2Cl2, 0 °C, then NaCN(BH3), MeOH, rt. Enantiomeric excesses determined by SFC or HPLC.

NMR spectroscopic studies with peptides 4  d and 4  e in deuterated chloroform, the solvent used for the reactions, revealed characteristic Nuclear Overhauser effects (NOEs) that are closely related to those previously observed for 4  a. These signals are typical for a type I β‐turn (Figure 1c and 2a). Most importantly, these studies support that the conformation of 4  d and 4  e is similar to that of 4  a. The stereochemical outcome of the addition reaction is therefore consistent with reaction of the Au‐activated electrophile at the upper, less sterically hindered face of the enamine, whereby the aromatic residue is in conjugation with the proline nitrogen (Figure 2b).

Figure 2

a) NOEs observed in the NMR spectra of peptides 4  d and 4  e. b) Stereochemical model of C−C bond formation.

We next aimed to further understand the driving forces and mechanism behind the catalytic efficiency of the reaction. Here, we focused on the addition of alkyl‐aryl aldehydes to allenamides and its catalytic mixture of peptide 4  d⋅TFA, gold carbene complex 5  a, and DMAP. We started by mixing these three components in a 1 : 1 : 1 ratio and analysed the mixture by 1H NMR spectroscopy. The spectrum shows one major set of signals relating to the peptide catalyst and two sets of signals in a ratio of ∼10 : 1 relating to the gold catalyst (Figure 3). The chemical shifts of the signals relating to the peptide indicate that peptide 4  d exists in an equilibrium between the TFA‐salt and the “desalted” form, a finding that was corroborated by a titration of the peptide‐TFA salt with DMAP (Figure S5). The major set of signals for the gold species relates to Au‐DMAP complex 13 and the minor set of signals to Au‐TFA complex 14. In addition, a small amount of the salt formed by DMAP and TFA is visible. These assignments were supported by comparison with authentic samples by 1H and 19F NMR spectroscopy and, in case of 13, further confirmed by crystal structure analysis (Figures S2‐S4).[ , ] These experiments demonstrate that an equilibrium exists at the gold center, between the DMAP and TFA ligands. No complexation is, however, observed between the peptide and the gold center in the catalytic mixture. In contrast, a complex mixture was observed between 4  d and 5  a in the absence of DMAP and TFA (Figure S3). Together with the intramolecular salt bridge within the structure of the peptide, these additives thus help to separate the peptide and gold catalysts and disfavour non‐productive interactions. These findings suggest that acid and base additives are important for the efficiency of the catalytic reaction.

Figure 3

1H NMR spectrum of a 1 : 1 : 1 mixture of 4  d⋅TFA, 5  a and DMAP.

A series of experiments in which TFA and/or DMAP were excluded or added in excess to the reaction mixture further supported this conclusion (Table 2). In the absence of TFA hardly any product formed and in the absence of DMAP or both additives only small amounts of γ,δ‐enamide aldehyde 3  a formed along with considerable amounts of allenamide dimer 6 (Table 2, entries 1–3). Addition of a two‐fold excess of DMAP with respect to the Au‐complex and the peptide prevented product formation (Table 2, entry 4) Thus, the absence of DMAP favours dimerization of 2  a to form 6 over reaction to form 3  a. An excess DMAP or the absence of TFA has an inhibitory effect.

Table 2

Addition reaction between aldehydes 1  a and 1  b and allenamide 2  a catalyzed by a combination of a peptide and a metal complex.

Entry	Deviation from conditions	3  a [%][a]	6 [%][a]
1	No TFA, no DMAP	10	21
2	No DMAP	9	29
3	No TFA	6	–
4	Additional DMAP (5 mol%)	–	–

Reaction conditions: aldehyde 1  a: allenamide 2  a=2 : 1, concentration=0.2 M. [a] Yield estimated by NMR using 1,3,5‐trimethoxybenzene as internal standard.

–

–

–

These findings support the following catalytic cycle (Scheme 4): The peptide catalyst forms an enamine intermediate with the aldehyde that then reacts in the stereochemistry determining step with an electrophilic π‐complex (I) formed through ligand exchange of the Au‐DMAP (or Au‐TFA) complex with the allenamide. Upon C−C bond formation, vinyl gold‐iminium intermediate II is formed, which after protodeauration and hydrolysis furnishes the γ,δ‐enamide aldehyde (3) and regenerates the gold and peptide catalysts, respectively.

Scheme 4

Proposed mechanistic cycle. [Au]=(IPr)Au+.

This mechanistic rationale is in agreement with the observed effects of TFA and DMAP on the reaction. The reduced reaction speed upon increasing the concentration of DMAP likely stems from the equilibrium between the electrophilic π‐complex (I) and Au‐DMAP complex 13. The strong coordinating ability of DMAP likely favours gold‐complex 13 over intermediate I. Increasing the amount of DMAP therefore shifts the equilibrium towards 13 and thus decreases the amount of intermediate I and therefore the overall reaction efficiency. Lowering the concentration of DMAP has the opposite effect, shifting the equilibrium to favour intermediate I. Upon formation of I, two reaction pathways can take place (A (green) and B (red) in Scheme 4). In pathway A, I reacts with the enamine, leading to product 3. Alternatively, in pathway B, reaction of I with a second equivalent of allenamide 2 takes place, leading to allenamide dimer 6 or higher oligomers. The DMAP thus plays a crucial role in modulating the amount of I, acting to perfectly balance the production of intermediate I, so that pathway A is favoured over other undesired reaction pathways. TFA has a similar effect as increasing its amount should facilitate decomplexation of the DMAP ligand from the gold center by protonation of the pyridine nitrogen. A lower amount of DMAP or a higher amount of TFA should therefore increase the reaction rate. This expectation proved to be correct when performing the reaction with either 3 mol% of DMAP or 7.5 mol% of TFA. In both cases the reaction proceeded faster without sacrificing the enantioselectivity, although with reduced conversion to the γ,δ‐enamide aldehyde (3  a) (Figure S7).

In summary, we have disclosed the enantioselective addition of branched aldehydes to allenamides, catalyzed by a synergistic combination of a peptide and a gold catalyst. The reaction furnishes α‐alkylated aldehydes bearing either a quaternary or a fully substituted carbon stereogenic center in good to very good enantioselectivities and proceeds under mild conditions. The aldehyde‐enamide products can be elaborated into a range of different chiral building blocks including diols, lactones and piperidines. Acid and base additives proved important to control the formation of the electrophilic gold intermediate and its reaction with the enamine intermediate rather than undesired formation of allenamide dimers. Structural modifications as well as in‐solution NMR spectroscopic studies showed that the conformational properties of the peptidic catalyst are important for the catalytic efficiency and the selectivity of the reaction. These data demonstrate the robustness and chemoselectivity of peptide catalysts of the type H‐Pro‐Pro‐Xaa, even in increasingly complex environments with metal catalysts. They lay the foundation for the use of peptides in even more complex synergistic catalytic processes, reminiscent of catalytic cascades used by nature.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.