Chapter 1
Chemical Peptide Ligations

Yihui Cao and Xuechen Li

Department of Chemistry, the University of Hong Kong, Hong Kong SAR, P. R. China

1.1 Introduction

Proteins play crucial roles in basic physiological processes and are responsible for a variety of biochemical functions, including signaling transduction, energy utilization, and immune response. Correlating protein structure with function has always been a charming topic among researchers. Although recombinant expression from bacteria or cell lines is a convenient means to produce proteins, it is still difficult to control specific post-translational modifications such as glycosylations, incorporate any uncanonical amino acid, or introduce unnatural reporters such as fluorescent tags, using the natural cellular machinery [1]. Herein, chemical protein synthesis that assembles protein sequence through atom-by-atom control provides a solution for generating site-specific natural or unnatural modification(s), and mirror-image proteins.

The solid-phase peptide synthesis (SPPS) by Merrifield provides an efficient peptide synthesis approach [2]. Utilization of SPPS methodology, along with the condensation of protected peptide fragments, has significantly expanded the range of polypeptide lengths that can be achieved via chemical synthesis. However, SPPS is limited by the peptide length. Due to the statistical reasons for linear stepwise coupling, each step during SPPS is incomplete, which causes byproducts to accumulate with peptide chain elongation. The peptide length from SPPS mostly remains within 50 amino acids to maintain good synthesis quality. Besides, the limited solubility of protected peptides in organic solvents hampers the ability of this method to meet the increasing synthesis demand for complex protein structures [3]. Consequently, novel synthetic approaches that can be conducted in aqueous buffers for handling unprotected peptide segments are strongly demanded.

The concept of peptide ligation, which allows condensation of unprotected peptide segments, was first proposed in the 1980s [4,5]. It involves a weakly activated peptide C-terminus to chemoselectively react with the N-terminus of the second peptide, resulting in a ligation intermediate that links the two fragments together, followed by an irreversible rearrangement step to form a natural peptide linkage (Figure 1.1). In this chapter, the driving forces of ligation, chemoselectivity details, and their applications in protein synthesis are discussed.

Figure 1.1 Generic chemical peptide ligation.

1.2 Ligation Driven by Trans-esterification

1.2.1 Native Chemical Ligation

Native chemical ligation (NCL) (Figure 1.2a), developed by Kent et al. in 1994, is the most widely applied ligation method [6]. NCL requires one peptide with a C-terminal thioester and the second peptide with an N-terminal cysteine (Cys). Ligation occurs when two fragments are mixed in a neutral or slightly basic aqueous buffer. The thiol group of N-terminal Cys undergoes reversible trans-thioesterification, replacing the thioester at the C-terminal of the first fragment. After that, a rapid [1,4] -to- acyl transfer converts the thioester intermediate into a native Xaa-Cys peptide (Xaa represents any amino acid) and generates the desired ligation product. Chemoselectivity can be considered to originate from "soft base-soft acid" interaction between the thioester (soft acid) and the thiol group (soft base) from free Cys. Other nucleophiles present on unprotected peptides, such as amines, are "hard" bases that do not have the same reactivity as the thiol group. Next, the irreversible and rapid acyl transfer drives the equilibrium. Even though the internal Cys could be involved in the reversible trans-thioesterification, the reaction equilibrium cannot move forward and does not produce a stable product. The ligating C-terminal residue and the thioester type highly affect the NCL reaction rate. For instance, -branched amino acids (Val, Ile, Thr) significantly decrease ligation rates due to their bulky side chains [7]. Additionally, because of carbonyl oxygen interference, proline carbonyl is less electrophilic, which restricts trans-esterification [6]. For the formation of intermolecular thioesters, 4-mercaptophenylacetic acid (MPAA) is widely used as an additive due to its low pKa, good water solubility in NCL buffer, and odorless nature. Other thiols with higher pKa values, such as trifluoroethanethiol (TFET), can be also used [8].

Figure 1.2 Ligation through trans-esterification.

1.2.1.1 Desulfurization

Even though NCL is a powerful technique, the demand for one N-terminal Cys restricts its potential applications in protein chemical synthesis due to its low natural abundance (1.8%). The situation has changed after the invention of post-ligation desulfurization. Desulfurization, first reported in 2001 by Dawson, took advantage of metal-catalyzed reduction under a hydrogen atmosphere [9]. Nevertheless, the requirement of excess metal could occasionally cause side reactions and result in low yields. In addition, utilization of metal catalysts potentially induced epimerization of secondary alcohols and caused reduction of thiols and thioesters [10]. Later, the establishment of free radical-based desulfurization by Danishefsky et al. provided a milder and more reliable means for chemoselective peptide desulfurization (Figure 1.2b) [10]. Radical-induced desulfurization requires a radical initiator, phosphine compound, and hydrogen source. The reaction is initiated by water-soluble radical initiator 2,2-azobis[2-(2-imidazolin-2-yl)propane]dihydrochloride (VA-044) at 37 °C. The initiator radical attacks the Cys on the peptide, generating a thiyl radical. Subsequently, the thiyl radical rapidly reacts with tris(2-carboxyethyl)phosphine (TCEP) to generate the phosphonium radical. After that, the phosphine sulfide is cleaved from the complex, driving the alkyl radical formation. The desulfurized product is eventually generated after the application of hydrogen atom transfer (HAT) from thiol additives such as tert-butylthiol (tBuSH) or glutathione (GSH). Other thiyl radicals generated continue the radical chain reaction until its full conversion.

Based on the radical desulfurization strategy, photon-induced radical initiation has been developed. Pyane's group designed a flow chemistry system allowing NCL and UV-induced desulfurization to take place sequentially [11]. Later, visible light-induced metal complexes [12] and peroxide [13] radical initiator were reported, expanding the scope of radical initiator. Lately, Li et al. applied a novel radical generator, tetra-organylborate, to peptide desulfurization [14]. This strategy significantly increases the desulfurization reaction rate. It can be accomplished through a simple add-and-done procedure to finish within 30 seconds. Sodium tetraethylborate effectively serves as a radical initiator in the presence of atmospheric oxygen, inducing peptide desulfurization. Besides, the byproduct triethylborane from the initiation step serves as a hydrogen donor, demonstrating comparable efficiency to thiol additives such as tBuSH. Therefore, odorous thiol additives are not necessary for this strategy. In addition, the expeditious production of radicals can surpass the inhibitory impact of traces of MPAA residue , which allows NCL and desulfurization to perform in one pot. Furthermore, the mild conditions are compatible with some reductive functional groups, such as serotonylated substrates.

With the effective desulfurization technique established, the development of thiolated amino acids has drawn wide attention. The first thiolated amino acid to be used in NCL-desulfurization was -thiolated phenylalanine [15]. NCL-desulfurization via both the nickel method and the free radical VA-044 method has been successfully applied to it [16,17]. Remarkably, commercially available penicillamine (Pen) has been used as the valine surrogate to expand the ligation site to one of the most abundant amino acids (6.8%) [18]. So far, a number of research groups have contributed to developing 13 Fmoc-SPPS-compatible thiolated amino acids, including -thiolated, -thiolated, and -thiolated amino acids [1,19,20]. Although desulfurization may be difficult to some steric hindrance residue [18], NCL surrogate has been significantly expanded.

1.2.1.2 Auxiliary-assisted NCL

In parallel with the development of desulfurization, another strategy to overcome the low Cys abundance problem is the auxiliary-assisted NCL (Figure 1.2c). It introduces a thiol-containing handle at or near the peptide N-terminus, which mediates trans-thioesterification and -to- acyl transfer in a similar manner as Cys. After ligation, the auxiliary thiol hand can be removed. For glycopeptides, the branching sugar can serve as the handle through a thiol substitution. Its ligation site can be 1-6 amino acids away from the sugar-substituted amino acid [21,22]. In such cases, the acyl transfer via a large ring transition state significantly affects the reaction rate and efficiency, as compared to NCL. On the other hand, a variety of substitutions, including -2-mer-captoethyl type and -2-mercaptobenzyl type [23], have been developed. They are designed to allow for acyl transfer through a five- or six-member ring for a rapid...