Unprotected amino acids are completely insoluble in organic solvents, limiting their use in organic synthesis. To address this issue, we report a Ta-catalyzed strategy for the elongation of unprotected amino acids using N-trimethylsilylimidazole. More specifically, silylation of the unprotected amino acids was carried out at both termini to generate a linear intermediate that was soluble in organic solvents, thereby permitting the efficient introduction of amino acid esters at the C-terminus. Furthermore, the in situ silicon protection approach employed herein was applicable to varying lengths of amino acid chains, leading to the generation of N-unprotected β-, γ-, and δ-peptides without any loss of enantio- or diastereopurity. These innovative results have facilitated the elimination of several steps in the preparation of desired peptides and have simplified the synthesis of peptides containing long-chain amino acids and branched peptides.

Unprotected amino acids, widely abundant in nature, are readily accessible aminocarboxylic compounds with a wide range of bioactivities owing to their structurally diverse side chains. However, the intrinsic bifunctionality of unprotected amino acids, possessing nucleophilic and electrophilic moieties, has restricted their use primarily to self-condensation protocols. Consequently, sequence- and length-controlled elongation reactions could not be achieved.^1–3) As the sequence-defined peptides have significantly enhanced biological activities in relation to the peptides composed of single amino acids, it is essential to develop cross-condensation procedures enabling selective peptide bond formation between different amino acids. To meet these demands, the introduction of protection/deprotection steps for each elongation step at the N- and/or C-termini of amino acids has been widely investigated since the first report of such an approach in 1932,⁴⁾ and is now fully rooted. As these methods have become more prevalent, significant progress has been made in the development and application of available protecting groups over the decades.

However, the functional groups that represent the most efficient protecting groups possess high molecular weights, resulting in the production of waste in amounts comparable to or exceeding that of the desired product.⁵⁾ Furthermore, the required repeated protection and deprotection steps ultimately reduce the final overall yield. It is therefore necessary to urgently establish an efficient, selective, and general method for peptide bond formation using unprotected amino acids. From these long-term dilemmas, several protocols to accomplish peptide bond formation using unprotected amino acids have emerged.^6–13) Notably, our research group proposed condensation reactions of unprotected amino acids that employed silicon- and aluminum-based reagents, which facilitated the synthesis of oligopeptides via sequential elongation reactions.^14,15) Furthermore, we recently realized chemoselective peptide bond formation between different types of unprotected amino acids.¹⁶⁾ A key feature in realizing these reactions was the design of a 5-membered ring, which acts as a competent electrophile, from unprotected amino acids and reagents. This 5-membered ring served dual roles in situ, involving the preparation of an active ester and suppressing over-reaction by protecting the amino group and reacting smoothly with an amino acid ester (Chart 1a). However, despite these advances, significant room for improvement remains in the development of general and practical protocols. For example, further optimization is required for preparing key reagents and overcoming substrate limitations caused by the stability profiles of the intermediates such as β-amino acids, which cannot form 5-membered rings. Peptides containing long-chain amino acids such as β-amino acids have methylene introduced into the amino and carboxyl groups, allowing free rotation and stabilizing the secondary structures like β-turns, β-sheets, and α-helices. From these structural advantages, they have attracted strong interest in the pharmaceutical and foldamer fields (Chart 1b). Therefore, there is a strong need to develop a robust method capable of incorporating long-chain β-, γ-, and δ-amino acids.

Chart 1. Backgrounds

Previously, our group discovered that the use of appropriate silylating reagents provided regioselective silicon protection to the N- and/or C-termini of unprotected amino acids.¹⁶⁾ Against these backgrounds, we hypothesized that the incorporation of N-trimethylsilylimidazole (TMS-IM) could form a linear multi-silylating intermediate to overcome the abovementioned issues, whilst also allowing the subsequent reaction with the amino acid ester to proceed efficiently.

To verify the abovementioned hypothesis, we 1st investigated the effect of TMS-IM silylation on peptide bond formation using H-l-Trp(Boc)-OH and H-l-Ala-Ot-Bu as model substrates. Upon heating the unprotected amino acid (1), 2 equivalent (equiv.) of the nucleophile (2), and 4 equiv. of TMS-IM at 50°C for 24 h, a trace of the corresponding dipeptide was observed. MS results supported the detection of di-silylation of unprotected amino acids when the nucleophilic amino acid ester and TMS-IM were stirred together. This finding indicates that the silylation of the amino group is stabilized by weak coordination to the C-terminal carbonyl group, resulting in selective peptide bond formation (see Supplementary Materials). Considering our previous report that the addition of a catalytic amount of Ta promotes the condensation reaction between amino acid esters and amino acid silyl esters,¹⁷⁾ we subsequently applied identical conditions to the current system, resulting in a dramatic improvement in the yield of 3a (Chart 2). Although the direct reason why Ta is more effective in peptide bond formations than other metal alkoxides¹⁸⁾ is unclear, its size, affinity for carbonyl or amino groups, and stability may have contributed significantly to activation for amino acid silyl esters. For the model reaction, the reaction was nearly complete after 12 h. However, a reaction time of 24 h was used to obtain all dipeptides in good yield. Based on these results, we explored the scope of this N-unprotected dipeptide synthesis beginning with coupling partner of H-l-Ala-Ot-Bu (3a–3n) and H-l-Trp(Boc)-OH (3o–3x) (Chart 2). Consequently, peptide bond formation proceeded in good to excellent yields, independent of the electronic environment defining the aromatic ring (3b–3f and 3p), with negligible impact on the other active functional groups (3g–3m and 3o–3u). Bulky unnatural amino acids containing quaternary carbon centers were tolerated under these reaction conditions (3n and 3v). However, when the amino acid methyl esters were used as nucleophiles, self-condensation followed by cyclization occurred, resulting in the formation of diketopiperazine. As is often the case in the Ta-catalyzed peptide synthesis, the catalyst coordinated with the methyl ester and acted as an electrophile.¹⁷⁾ Meanwhile, the amino acid ethyl and i-Pr esters were compatible with the reaction of unprotected amino acids to afford the corresponding dipeptides in good yields (3w and 3x). In addition, successful tripeptide synthesis was achieved using 3b as the nucleophile (3y).

Chart 2. Elongation of the Unprotected Amino Acids Using TMS-IM^a

Encouraged by these positive results, we turned toward examining the substrate scope of unnatural long-chain amino acids as β-, γ-, and δ-amino acids. As reported previously, incorporating such long-chain peptides enhances stability through improved van der Waals interactions, folding, and steric effects, leading to their incorporation in various pharmaceuticals.^19–25) As mentioned above, a previous route to peptide synthesis using unprotected amino acids involved the formation of a 5-membered intermediate, which rendered the application of β- (or longer) amino acids extremely challenging. However, it was assumed that the current approach, which leads to the in situ protection of both amino acid termini, would proceed regardless of the amino acid length to enable the efficient preparation of high-value β-, γ-, and δ-peptides. Therefore, various unnatural amino acids were investigated by applying the reaction conditions outlined in Chart 1, revealing that the amino acid length had no effect on peptide bond formation (6a, 6h, and 6i). Efficient reactions were observed even when amino acid esters bearing bulky (6b, 6c, 6f, and 6g) or active functional groups (6d and 6e) were used (Chart 3).

Chart 3. Elongation Reaction of the Unprotected β-, γ-, and δ-Amino Acids^a

This length-independent condensation of amino acids was subsequently considered in the context of performing multi-peptide bond formation reactions. Upon applying optimized reaction conditions to amino acids with carboxyl groups on their side chains (e.g., aspartic acid [Asp] and glutamic acid [Glu]), peptide bond formation was observed not only at the primary chain but also at the side chain. As the residual unreacted substrates were observed under the proposed reaction conditions, the reaction yields were enhanced upon increasing the amount of TMS-IM, and the corresponding tri-amino acid complexes (7a and 7b), which are motifs of branched peptides, were obtained in high yields (Chart 4). In addition, we investigated further elongation reactions using the compounds prepared by these results in anticipation of the next stage of development. Consequently, tetra- (7c), penta- (7d), and hepta-amino acid complexes (7e) were successfully synthesized using 7a and 3a in moderate to good yields, respectively.

Chart 4. Multiple Elongations of Unprotected Asp and Glu^a

We developed an efficient method for forming peptide bonds between unprotected amino acids and amino acid esters using commercially available reagents. In a previously reported elongation reaction of unprotected amino acids, self-condensation was suppressed by the formation of a relatively stable 5-membered ring intermediate, but it was difficult to apply to long-chain amino acids. On the other hand, the current method proceeds with linear multi-silyl protection in situ, allowing the synthesis of a peptide containing long-chain amino acids with enhanced bioactivity in relation to that of peptides composed of natural amino acids. Based on these advantages, multi-peptide bond formation was achieved throughout the simultaneous reaction with amino acid esters using unprotected Asp and Glu. This success would lead to the development of new branching oligopeptide synthesis with Asp and Glu as junctions. Furthermore, these reactions proceeded without any loss of enantio- or diastereopurity, demonstrating the application toward the synthesis of complex bioactive peptides.

In a glove box, a mixture of unprotected amino acid (0.250 mmol), amino acid tert-butyl ester (0.500 mmol), 1-(trimethylsilyl)imidazole (146 µL, 1.00 mmol), Ta(OEt)₅ (6.50 µL, 0.0250 mmol), and dry dichloromethane (1 mL) was placed in a flame-dried 20 mL test tube equipped with a magnetic stirring bar and stirred vigorously at 50°C for 24 h under Ar atmosphere. The reaction mixture was then diluted with CHCl₃ (4.50 mL) and transferred onto a SiO₂ column. The reaction mixture was purified by flash column chromatography to provide the corresponding peptide as a white solid.

This work was partially supported by a Grant-in-Aid for Specially Promoted Research (23H05407) from the Japan Society for the Promotion of Science (JSPS).

The authors declare no conflict of interest.

This article contains supplementary materials.

• 1)  Greene T. W., Wuts P. G. M., “Protective Groups in Organic Synthesis,” John Wiley & Sons, New York, 1999.
• 2)  Kocienski P. J., “Protecting Groups,” Georg Thieme Verlag, Stuttgart, New York, 2004.
• 3)  Isidro-Llobet A., Álvarez M., Albericio F., Chem. Rev., 109, 2455–2504 (2009).
• 4)  Bergmann M., Zervas L., Ber. Dtsch. Chem. Ges. B, 65, 1192–1201 (1932).
• 5)  Fuse S., Masuda K., Otake Y., Nakamura H., Chem. Eur. J., 25, 15091–15097 (2019).
• 6)  van Leeuwen S. H., Quaedflieg P. J. L. M., Broxterman Q. B., Liskamp R. M. J., Tetrahedron Lett., 43, 9203–9207 (2002).
• 7)  van Leeuwen S. H., Quaedflieg P. J. L. M., Broxterman Q. B., Milhajlovic Y., Liskamp R. M. J., Tetrahedron Lett., 46, 653–656 (2005).
• 8)  Lanigan R. M., Karaluka V., Sabatini M. T., Starkov P., Badland M., Boulton L., Sheppard T. D., Chem. Commun. (Camb.), 52, 8846–8849 (2016).
• 9)  Sabatini M. T., Boulton L. T., Sheppard T. D., Sci. Adv., 3, e1701028 (2017).
• 10)  Sabatini M. T., Karaluka V., Lanigan R. M., Boulton L. T., Badland M., Sheppard T. D., Chem. Eur. J., 24, 7033–7043 (2018).
• 11)  Li Z., Fu H., Gong H., Zhao Y., Bioorg. Chem., 32, 170–177 (2004).
• 12)  Spengler J., Böttcher C., Albericio F., Burger K., Chem. Rev., 106, 4728–4746 (2006).
• 13)  Nobuta T., Tsuchiya N., Suto Y., Yamagiwa N., Org. Biomol. Chem., 22, 703–707 (2024).
• 14)  Hattori T., Yamamoto H., J. Am. Chem. Soc., 144, 1758–1765 (2022).
• 15)  Hattori T., Yamamoto H., Chem. Sci. (Camb.), 14, 5795–5801 (2023).
• 16)  Hattori T., Yamamoto H., J. Am. Chem. Soc., 146, 25738–25744 (2024).
• 17)  Muramatsu W., Hattori T., Yamamoto H., J. Am. Chem. Soc., 141, 12288–12295 (2019).
• 18)  Tsuji H., Yamamoto H., J. Am. Chem. Soc., 138, 14218–14221 (2016).
• 19)  Ordóñez M., Cativiela C., Tetrahedron Asymmetry, 18, 3–99 (2007).
• 20)  Guo L., Almeida A. M., Zhang W., Reidenbach A. G., Choi S. H., Guzei I. A., Gellman S. H., J. Am. Chem. Soc., 132, 7868–7869 (2010).
• 21)  Kudo F., Miyanaga A., Eguchi T., Nat. Prod. Rep., 31, 1056–1073 (2014).
• 22)  Cabrele C., Martinek T. A., Reiser O., Berlicki Ł., J. Med. Chem., 57, 9718–9739 (2014).
• 23)  Legrand B., Maillard L. T., ChemPlusChem, 86, 629–645 (2021).
• 24)  Tanaka T., Oba M., Ichiki T., Suemune H., Chem. Pharm. Bull., 49, 1178–1181 (2001).
• 25)  Honda T., Wakabayashi H., Kanai K., Chem. Pharm. Bull., 50, 307–308 (2002).