Protein is a sequence-defined polyamide molecule consisting of 20 proteinogenic amino acids, and is exclusively produced by the cellular translation system. On the other hand, a chemical approach known as “chemical protein synthesis (CPS)” has recently emerged as an alternative to protein production. While in cells, amino acid monomers are coupled one-by-one by the ribosome, in CPS, peptide chains are elongated by an amide-bond forming reaction generally on solid supports. Solid-phase peptide synthesis (SPPS) allows for the coupling of a wide variety of amino acids, whereas the available monomers are normally limited to the 20 proteinogenic amino acids in the cellular translation system. Therefore, using CPS, peptide chemists have synthesized various modified or artificial proteins such as proteins with posttranslational modifications including methylation, acetylation, phosphorylation and ubiquitination, fluorophore-labeled proteins, mirror-image proteins, and so on.^1–7 

CPS consists of two major steps: the first step is SPPS, and the second step is peptide ligation. SPPS was originally developed by the Merrifield group in 1963⁸  and revolutionized chemical peptide synthesis. Thanks to the immobilized reactive sites on the solid support, it is possible to efficiently repeat the amino-acid coupling reaction and subsequent deprotection reaction via simple separation of waste solutions after each reaction. More recently, this process has been further streamlined by the advancement of automated peptide synthesizers. However, SPPS has an intrinsic difficulty in synthesis of long-chain peptides (e.g., more than 50 amino acid residues). One possible cause for the decreased efficiency in peptide elongation is the formation of secondary and tertiary structure of intermediate peptide chain under synthesis. As the peptide chain becomes longer, the exposure of the N-terminal reactive site on the solid support would decrease due to the formed structures, thereby reducing the accessibility of coupling or deprotection reagents.^9,10  Given that most proteins are composed of more than 100 amino acid residues, chemical synthesis of target proteins solely through SPPS has been impractical, while the state-of-the-art peptide synthesizers are pushing the limit of single-shot SPPS.¹¹  Currently, therefore, ligation chemistries for assembling peptide segments are still essential to build full-length proteins of interest. In this chapter, we describe the basic peptide ligation chemistries and its application especially focusing on recent strategies of one-pot multiple peptide ligation to streamline CPS procedures, although another good review paper summarizing one-pot peptide ligation is available.¹² 

The most widely used reaction for peptide ligation is native chemical ligation (NCL),¹³  in which one C-terminal peptide thioester and another N-terminal cysteinyl peptide are coupled to form a native peptide bond (Fig. 1A). Due to the high reactivity and chemoselectivity in aqueous solution, NCL does not require protecting groups on the side chains of peptide, making it versatile and easily accessible. In fact, the majority of previously synthesized proteins have been synthesized by using NCL and its extended methods.¹⁴  However, conventional NCL demands a Cys residue at the ligation site, which is a relatively rare amino acid among the 20 proteinogenic amino acids.¹⁵  To synthesize target proteins containing no Cys residues or proteins containing Cys residues only at terminal regions, various chemical tools beyond NCL have been established so far.

One reliable method is desulfurization reaction, in which Cys residues are converted to Ala residues via sulfur atom abstraction reaction (Fig. 1A). Since Ala is a relatively abundant amino acid,¹⁵  the applicable scope of NCL-based CPS dramatically increased by combining the desulfurization reaction. Based on Hoffmann’s early work of desulfurization of mercaptan in 1956,¹⁶  the first demonstration of desulfurization-mediated Cys-to-Ala conversion after peptide ligation with NCL was reported by using metal catalysts in 2001.¹⁷  However, this method requires excess amount of metals,¹⁸  and sometimes causes undesired side-reactions such as racemization of secondary alcohol.¹⁹  In 2007, metal-free desulfurization employing radical initiator, external thiol compound and phosphine²⁰  was developed and now becomes the gold standard in CPS. More recently, other desulfurization strategies such as photoredox desulfurization,²¹  P–B desulfurization,²²  tetraethylborate-mediated desulfurization,²³  UV-irradiation mediated desulfurization,²⁴  phosphine-only photodesulfurization²⁵  and so on, have been also established as other options for Cys-to-Ala conversion.

Desulfurization also triggered the development of side chain thiolated amino-acids,⁶  which enable peptide ligation via NCL-like mechanism (i.e. thioester exchange and subsequent S-to-N acyl shift) (Fig. 1B). A phenylalanine derivative tethering β-thiol group was first synthesized and successfully applied to NCL followed by desulfurization to generate natural amino-acid sequences.²⁶  After this report, thiol-containing unnatural amino acids that can be desulfurized to recover natural amino acids such as Val, Leu, Pro, Glu, Lys, and so on, have been synthesized and successfully applied to SPPS and NCL.⁶ 

As another means to extend the reaction scope of NCL beyond N-terminal cysteinyl peptide, removable ligation auxiliary groups have been intensely studied so far^27,28  after the concept of auxiliary group-mediate ligation was first reported in 1996.²⁹  Many of the auxiliary groups tethering NCL-active thiol groups are introduced on N-terminal amino group (Fig. 1C). The thiol group can work as a peptide thioester capturing element via thiol-thioester exchange with counterpart peptide, and then, a new amide bond forms via intramolecular S-to-N acyl transfer. Since the auxiliaries can be introduced regardless of N-terminal amino acid residue, any ligation junction could be chosen in theory. However, in practice, this method was almost limited to Gly junction because of significantly slow intramolecular S-to-N acyl transfer. To address this issue, recently, auxiliaries beyond Gly junctions have been developed by using α-unsubstituted β-mercaptoethyl scaffolds.^30–32  The latest auxiliary of -mercapto-β-(4-methoxy-2-pyridinyl)-ethyl (MMPyE) auxiliary (Fig. 1C) with efficient attachment/detachment features enabled NCL at sterically challenging junctions such as Leu–Val.³³ 

Instead of Cys residue, the use of selenocysteine (Sec) at the N-terminal peptide, which generally exists as diselenide-linked dimer, has become prevalent in the field of CPS³⁴  because of its higher nucleophilicity than that of Cys.³⁵  Meanwhile, peptide selenoester has also been recognized as a good alternative to thioester because of the higher reactivity as an acyl donor.³⁶  By combining N-terminal Sec peptide and C-terminal peptide selenoester, diselenide–selenoester ligation (DSL) was developed as significantly efficient peptide ligation in an additive-free manner (Fig. 1D).³⁷  DSL was further improved by setting reducing condition in the presence of diphenyldiselenide as a radical scavenger to prevent undesired deselenization, which was named reductive diselenide–selenoester ligation (rDSL).³⁸  It is noteworthy that another advantage to use Sec is the selective deselenization of Sec to Ala in the presence of free Cys residues, which enables ligation at both Cys and Ala sites in the same target protein sequence (Fig. 1E).³⁹ 

Besides NCL, there have been various efforts to develop chemoselective amide bond-forming reaction between two peptide segments. Notably, peptide thioesters had been used in peptide ligation prior to NCL, where two peptide segments with minimal protection (i.e. protection of Lys side-chain amino groups) were regioselectively coupled between the C-terminal thioester and the N-terminal main chain amino group (Fig. 2A).⁴⁰  This reaction proceeds via silver ion-mediated thioester activation followed by active oxyester formation with alcohol additives to generate a new amide bond. As another type of peptide ligation, traceless Staudinger ligation was reported (Fig. 2B),^41,42  where an intermediate of Staudinger reaction,⁴³  iminophosphorane generated from intermolecular coupling of azide and phosphine, was utilized as an acyl acceptor to attack intramolecular thioester and produce a new peptide bond after leaving hydrolyzed phosphine. Although this strategy overcame the limitation of Cys requirement involved in NCL, the handling of phosphine-containing peptide could include a certain difficulty due to its susceptibility to oxidation.⁴⁴ 

In 2006, Bode et al. reported another type of chemoselective peptide ligation, named α-ketoacid–hydroxylamine (KAHA) ligation (Fig. 2C), between side chain unprotected peptide segments.⁴⁵  After many trials and optimization of the N-terminal hydroxylamine structure,⁴⁶  5-oxaproline was found to be stable, easily prepared, incorporated into N-terminus of peptides in SPPS, and show clean ligation in a mixture of water/organic solvents (e.g. DMSO and NMP) with oxalic acid (Fig. 2C).⁴⁷  Unfortunately, the use of 5-oxaproline in the peptide ligation affords a homoserine residue at the ligation junction, which is a non-proteinogenic amino-acid. To address this issue, N-terminal oxazetidine instead of 5-oxaproline was successfully adopted to generate a proteinogenic Ser and Thr residue after ligation (Fig. 2C).^48,49 

As another powerful Cys-independent ligation strategy, serine/threonine ligation (STL) was reported by Li et al. in 2010 (Fig. 2D).⁵⁰  Between N-terminal Ser or Thr peptides and C-terminal peptide salicylaldehyde ester, STL proceeds via oxazolidine formation, O–N acyl transfer, and acidic hydrolysis of benzylidene acetal intermediate to retain native peptide sequences. Notably, the aldehyde capture–rearrangement process (i.e. oxazolidine formation followed by O–N acyl transfer process) was previously achieved by Tam’s pseudoproline ligation utilizing C-terminal peptide glycoaldehyde ester,⁵¹  although this strategy could not recover native amino acid structure after ligation due to the stable acetal formation. STL has been applied to synthesize naturally occuring proteins⁵²  such as human erythrocyte acylphosphatase,⁵³  MUC1 glycopeptides⁵⁴  and adiponectin collagenous domains.⁵⁵  STL mechanism was expanded to the use of N-terminal Cys or penicillamine (Pen) peptides instead of Ser/Thr in 2020 (Fig. 2D).⁵⁶  This cysteine/penicillamine ligation (CPL) enables to select not only Cys but also Ala and Val at the ligation junction by combining desulfurization reaction. Recently, C-terminal peptide salicylaldehyde ester has been synthesized through Boc-SPPS, leading to the incorporation of base-labile posttranslational modifications such as O-acetylation and S-palmytoylation.⁵⁷  Furthermore, N,O/S-benzylidene acetals, which are intermediates of STL and CPL, have been utilized as aggregation disruptors for the synthesis of aggregation-prone peptides and proteins.^58,59 

When the size of target protein is ∼100 amino acids or less, it is often the case that single ligation reaction between two peptide segments is sufficient. However, when the target protein becomes larger, two or more ligation steps with three or more peptide segments become necessary. To assemble three or more peptide segments, “middle segment(s)” arises besides N- and C-terminal segments. The N- or C-terminal reactivity of the middle segments should be intendedly controlled to suppress competing side reactions such as intramolecular cyclization and intermolecular self-ligation (Fig. 3A). For example, when three peptide segments are assembled by NCL, either the N-terminal Cys residue or C-terminal thioester in the middle segment should be inactivated in the first ligation step. If the three peptides are sequentially ligated in C-to-N direction, N-terminal Cys reactivity of the middle segment should be lower than that of C-terminal segment. This would be achieved for example through introduction of protecting groups on the Cys residue (Fig. 3B). On the other hand, in the case of N-to-C ligation, C-terminal thioester reactivity of the middle segment should be low compared to that of N-terminal segment. For example, the less reactive thioesters or thioester precursors are available for this purpose. After the first ligation, therefore, activation steps of peptide terminal functional groups are often required in the assembly of three or more segments (Fig. 3C).

In the case of four or more segment ligation, convergent synthetic route can be chosen in addition to the linear synthetic route where the peptide segments are sequentially connected from one end to the other. In the convergent route, SPPS-prepared peptides are separately coupled in the first ligation and each ligated segment is assembled together at the end (Fig. 3D). In principle, the convergent route is preferred in terms of synthetic efficiency because it involves fewer steps to reach the final product. Moreover, the purity of the final product obtained from convergent route tends to be high. This is because separation of the product obtained with two similar-sized peptides could be easier than that obtained between large-sized peptide and small-sized peptide (i.e. product obtained in one-directional sequential ligation), due to the change of physicochemical properties such as hydrophobicity and net charge. It is also noteworthy that the order of peptide segment ligation is crucial in CPS. When the intermediate segments are not soluble in the ligation solution or soluble-but-unreactive due to unknown reasons (many of them could have structural problems such as aggregation), the synthesis fails at this point. Therefore, it is essential not only to optimize the order of peptide ligation but also to establish chemical tools and methods for ligating peptides in flexible orders.

As the target proteins of interest become larger and the number of peptide segments to be assembled increases, the yield of the final product becomes low due to the repeated reaction and purification process. HPLC purification is commonly used for isolating target peptides, and typically conducted after each ligation reaction. Since the purification step always includes the risk of sample loss mainly caused by the poor physical properties of target peptides such as aggregation and adsorption, increased number of purification often results in decreased yields. To address this issue, the concept and strategies of “one-pot multiple peptide ligation” have been developed so far, where purification and isolation of intermediate peptides are skipped and two or more ligation reactions are conducted in one-pot manner. Since the number of purification in one-pot ligation is smaller than that in conventional method, an improvement in the synthetic yield of the final product is expected. One-pot peptide ligation is typically conducted with one-directional sequential coupling of peptide segments in either C-to-N or N-to-C direction. Strategies and mechanisms enabling one-pot sequential ligation are introduced below divided by the direction of ligation.

In the C-to-N NCL-mediated one-pot ligation, N-terminal Cys residue of the middle segment needs to be dormant during the first ligation and to be activated after the ligation in the reaction mixture. In 2004, Kent et al. established the first C-to-N three segment one-pot ligation by using thiazolidine (Thz) as an inactivated Cys residue (Fig. 4A).⁶⁰  The N-terminal Thz moiety was cleanly converted to Cys residue by methoxyamine treatment at mild acidic pH (∼4) after the first NCL in one-pot manner. This conversion proceeds quantitatively because the thermodynamically stable oxime formation between methoxyamine and formaldehyde generated by hydrolysis of Thz push the equilibrium reaction to the naked Cys exposure. After the Thz-to-Cys conversion, the second NCL proceeded smoothly in the presence of large excess of methoxyamine to produce the target Crambin protein. This one-pot three segment ligation strategy became popular and has been widely applied to the synthesis of various proteins such as glycosylated ribonuclease C,⁶¹  l- and d-kaliotoxin,⁶²  VEGF,⁶³  human α-synuclein,⁶⁴  Chimadanin,^65,66  hCCL5,⁶⁷  CCL14 glycoprotein,⁶⁸  and ESAT6 protein fragment.⁶⁹  Using the same reaction mechanism, one-pot four segment ligation was also achieved to synthesize ribonuclease A (N-acetylglucosamine) and histone H2B bearing ubiquitin and N-acetylglucosamine.⁷⁰  However, the excess amount of methoxyamine in the reaction mixture is often problematic as a certain amount of peptide thioester could be decomposed.^18,60,70  Recently, we have developed a Thz-mediated one-pot NCL strategy by using 2-aminobenzamide-based aldehyde scavenger, 2-amino-5-methoxy-N′,N′-dimethylbenzohydrazide (AMDBH).⁷¹  AMDBH efficiently converted Thz into Cys at pH ∼ 4 but not at pH ∼ 7 of the NCL condition, and thioester degradation caused by AMDBH was negligible compared to methoxyamine. Accordingly, a pH-controlled one-pot four segments ligation was established to synthesize Arabidopsis ubiquitinated histone H2A.Z (209 amino-acids) in good yield. As another mechanism of Thz-to-Cys conversion for one-pot peptide ligation, metal-mediated hydrolysis of Thz was demonstrated by using Pd or Cu salts as a Lewis-acid. In 2018, Brik et al. reported that allylpalladium(ii) chloride dimer, [Pd(allyl)Cl]₂, could be utilized for Thz-mediated three segment ligation for the synthesis of activity-based probe composed of ubiquitinated histone H2A.⁷²  Cupric sulfate, CuSO₄, was also adopted in the Thz-mediated one-pot three segments NCL by Otaka group.⁷³ 

Selenazolidine (Sez), which is selenium substituent of Thz, was used in CPS as a protected form of Sec to synthesize human phosphohistidine phosphatase 1 in non-one-pot fashion.⁷⁴  Later, Sez-to-Sec conversion by cupric chloride, CuCl₂, was successfully applied to one-pot three segment ligation to prepare human thiosulfate:glutathione sulfurtransferase protein.⁷⁵  More recently, similar one-pot three segment ligation for synthesizing the thioredoxin-like domain of SELENOF has been achieved by using Fmoc-masked Sez moiety to prevent undesired deselenization.⁷⁶ 

Among a variety of Cys thiol protecting groups developed so far,⁷⁷  acetamidomethyl (Acm) group⁷⁸  has been commonly used for Cys protection in CPS because of its compatibility with SPPS procedures (e.g. stability in both acidic and basic conditions) and its orthogonality of deprotection⁷⁹  (Fig. 4B). However, the orthogonal and harsh deprotection conditions such as the use of heavy metals (Hg²⁺ or Ag⁺) have prevented Acm from application to one-pot peptide ligation because of the necessity of purification step after the removal. To address this issue, in 2016, Brik et al. developed a novel Acm removal method using palladium chloride, PdCl₂, and achieved C-to-N three segment peptide ligation with N-terminal Acm-protected middle segment in one-pot manner.⁸⁰  As a derivative of Acm group, trifluoroacetamidomethyl (Tfacm) group was developed by Liu et al. for another type of Cys thiol protecting group and successfully applied to C-to-N one-pot ligation (Fig. 4B).⁸¹  Notably, Tfacm is efficiently removed by adjusting the pH value to 11.5 after NCL-mediated peptide ligation. Iterative pH adjustment between 6.5 and 11.5 enabled one-pot four segment ligation to synthesize crambin and hCCL21 protein. More recently, Tfacm-mediated one-pot three segment ligation was applied to prepare K11-linked diubiquitin.⁸² 

Protection of N-terminal amino group of Cys in the middle segments is also possible to conduct one-pot C-to-N ligation although free thiol group of Cys may participate in the equilibrated thiol-thioester exchange reaction during NCL. The most popular choice of primary amine protection is carbamates such as Boc, Fmoc and Cbz, which have been successfully applied to peptide synthesis.⁸³  The carbamate-type protecting group first applied to one-pot peptide ligation was p-dihydroxyborylbenzyloxycarbonyl (Dobz), which can be removed by H₂O₂ treatment (Fig. 4C).⁸⁴  One-pot three segment C-to-N ligation with a middle segment containing Dobz-protected N-terminal Cys was conducted by adding H₂O₂ after first NCL to synthesize methylated histone H3 and acetylated histone H4.⁸⁵  Importantly, the addition of an appropriate amount of H₂O₂ to the reaction mixture is necessary for efficient deprotection and prevention of Met oxidation.

In 2018, we have reported a C-to-N one-pot sequential NCL strategy using allyloxycarbonyl (Alloc) groups in the N-terminus of middle segments (Fig. 4C).⁸⁶  Quantitative and rapid removal of Alloc group was demonstrated by adding a palladium(0) complex, Pd/TPPTS complex, with the aid of 4-mercaptophenylacetic acid (MPAA), which is commonly used as a thiol catalyst to activate peptide thioester and accelerate NCL.⁸⁷  This deprotection proceeds through nucleophilic attack of MPAA on π-allyl palladium complex to reproduce active palladium(0) complexes in the NCL solution. MPAA also contributes to quenching of the deprotection agent (i.e. Pd complex), which is an essential step for the second and following NCL to prevent undesired deprotection during ligation. By utilizing this unique mechanism, one-pot C-to-N five segment ligation was achieved for the first time to afford histone H2A.X in high isolated yield. We also confirmed that the Alloc deprotection reaction by Pd/TPPTS complex was compatible with Acm deprotection triggered by PdCl₂, enabling the one-pot ligation to assemble Cys-containing protein.⁸⁸  One inherent problem associated with the use of palladium is the significant affinity of Pd against thiols abundant in a NCL reaction mixture, leading to the requirement of excess amount of Pd agents. Therefore, we embarked on the exploration of the other Alloc deprotection agents. Accordingly, organoruthenium complex was found to show more than 50-fold activity over Pd/TPPTS complex and catalytic Alloc removal was achieved even in the presence of excess thiol moieties.⁸⁹  Thanks to the efficient and repetitive Alloc deprotection strategy, linker histone H1.2 and heterochromatin protein 1α (HP1α), bearing site-specific PTMs including phosphorylation, ubiquitination, citrullination, and acetylation have been synthesized and analyzed in biochemical assays.

As another carbamate type protecting group, Fmoc group was also applied to the one-pot peptide ligation in 2020.⁹⁰  NCL and Fmoc removal were repeated in the same reaction mixture through iterative pH adjustments (pH 7 for NCL and pH 11 for Fmoc removal) in the presence of 20% piperidine to achieve one-pot C-to-N four segment ligation for the synthesis of 86-residue polypeptide segment of Plasmodium falciparum protein Pf-AMA1 and human lysozyme.

Photoirradiation-triggered activation of N-terminal Cys would become an ideal one-pot ligation strategy because operation of adding some deprotection reagent is not required and the progress of the deprotection reaction can be controlled simply by turning ON/OFF the irradiation. The first use of a photolabile protecting group in one-pot peptide ligation is reported by Fujii and Otaka et al., where 4-(dimethylamino)phenacyloxycarbonyl (Mapoc) moiety as a protecting group of N-terminal amine was introduced (Fig. 4D).⁹¹  Mapoc group exhibited adequate stability against TFA cleavage condition commonly used in Fmoc-SPPS and was successfully applied to one-pot C-to-N three segment ligation without readjustment of ligation conditions such as pH to synthesize human-brain natriuretic peptide. In 2016, another photolabile protecting group of 6-nitroveratryloxycarbonyl (Nvoc), which is one of the most frequently used photolabile groups for masking amino group⁹²  was employed (Fig. 4D). Although extractive removal of thiophenol, a thiol catalyst in NCL solution, was required for the efficient Nvoc deprotection, one-pot C-to-N three segment ligation was achieved to obtain SNX-482.⁹³ 

Photolabile protecting group on the thiol group of N-terminal Cys residue can also work. A 6-nitroveratryl (NV) type protecting group was introduced through late-stage conjugation (Fig. 4D), in which rac-2-Br-2-(4,5-dimethoxy-2-nitrophenyl)-acetic acid (rac-2-Br-DMNPA) derivative was attached on the thiol of N-terminal Cys peptide in solution phase.⁹⁴  In the one-pot C-to-N three segment ligation, the removal of this photolabile protecting group was done in the presence of MPAA and semicarbazide as a scavenger to produce a mini-protein chlorotoxin. The selenol moiety of N-terminal selenocysteine was also masked by a photolabile protecting group, named 7-diethylamino-3-methyl coumarin (DEAMC) for one-pot ligation strategy (Fig. 4D).⁹⁵  Removal of DEAMC group proceeded in a mild and reagent-free manner by irradiation of visible light (λ = 450 nm) without undesired deselenization. By utilizing DSL and rDSL strategy, DEAMC-protected peptide segments were successfully applied to one-pot C-to-N three or four segment ligation to synthesize fragments of mucin-1 and apolipoprotein CIII. Importantly, the use of β-selenoleucine⁹⁶  and photodeselenization chemistry enabled the ligation at Ser–Leu junction. More recently, to overcome the limitations of conventional nitro-benzyl type protecting groups such as incompatibility with desulfurization, a quinoline-based photolabile protecting group, named 7-(piperazin-1-yl)-2-(methyl)quinolinyl (PPZQ) was developed⁹⁷  (Fig. 4D). PPZQ can be introduced not only by Fmoc SPPS as a PPZQ-protected cysteine, but also by late-stage modification through Cys conjugation chemistry. In addition, PPZQ can be efficiently removed by photolysis in NCL reaction mixture and is compatible with desulfurization. One-pot C-to-N four segment ligation was conducted by repeating NCL at neutral pH and uncaging of PPZQ at pH 3 to synthesize human γ-synuclein protein. Notably, metal-free desulfurization reaction was also conducted in one-pot manner after removing MPAA via ultrafiltration.

Photo-uncaging-mediated C-to-N one-pot ligation was also achieved with KAHA ligation.⁹⁸  5-Oxaproline protected tethering o-nitrobenzyl type photoprotecting group was synthesized, coupled on N-terminal of middle segment bearing C-terminal α-ketoacid and applied to one-pot C-to-N three segment ligation to afford NEDD8 protein.

Although NCL-based ligation is commonly conducted in the reducing condition to keep Cys residues as the reduced form (i.e. prevent disulfide formation), reduction switch-mediated one-pot C-to-N ligation is possible when the first ligation is performed in the absence of strong reducing agent such as TCEP. Azide group was introduced instead of amino group of N-terminal Cys residue in the middle segment and one-pot C-to-N three segment ligation by switching the reducing condition was demonstrated⁹⁹  (Fig. 4E). After the completion of the first ligation containing MPAA that work as a weak reducing agent, TCEP-mediated Staudinger reaction to generate N-terminal amino group for the second NCL. As a result, protein toxin Mambalgin-1 was successfully synthesized. Another reduction switch utilizing intramolecular Se–S bond of N-selenoethyl cysteine (SetCys) was reported in 2020 (Fig. 4E).¹⁰⁰  Upon TCEP-mediated reduction of the Se–S bond, the selenoethyl moiety can be removed under mild condition. One-pot C-to-N three segment ligation was conducted with TCEP-free first NCL and subsequent TCEP-mediated second NCL to produce K1 polypeptide. Interestingly, this one-pot reaction can be performed even in the presence of all three peptide segments in the first ligation.

One-pot C-to-N three segment ligation without N-terminal Cys activation was reported,¹⁰¹  where the reactivity difference between N-terminal cysteine of a C-terminal segment and sterically hindered β-thioleucine of a middle segment was exploited to realize regioselectivity in the first ligation. After the regioselective first ligation between C-terminal oxyester-based thioester precursor of the middle segment and the Cys of C-terminal segment, β-thioleucine-mediated second ligation occured with N-terminal segment bearing the same thioester precursor to synthesize a partial peptide of human erythropoietin. This mechanism is called kinetically controlled ligation (KCL), which is commonly employed in one-pot N-to-C ligation as described in the sections below.

When attempting one-pot ligation with N-to-C direction, there is a need to keep the peptide C-terminal reactivity of the middle segment suppressed during the first ligation. However, it is challenging to introduce protecting groups directly onto the C-terminal acyl donor, such as thioesters. The concept of KCL has surmounted this problem by simply exploiting multiple acyl donors with different reactivity derived from different leaving group ability (Fig. 5A). Therefore, protecting groups to suppress the peptide C-terminal reactivity are not necessary. In 2006, the first KCL strategy for one-pot N-to-C three segment ligation was reported by Kent et al.,¹⁰²  where the NCL reactivity difference between aryl thioester and less reactive alkyl thioester was employed (Fig. 5B). Since the first ligation between pepitde aryl thioester and cysteine residue was sufficiently faster than possible side reactions such as self-cyclization and polymerization of the middle segment, the KCL concept has proven useful. By combining Thz-mediated one-pot C-to-N three segment ligation, they achieved convergent synthesis of crambin protein from six peptide segments. Importantly, the difference of reactivity in KCL can be accelerated by the difference of C-terminal amino acid residues neighboring thioester. Therefore, it is recommended to choose both less reactive amino acids such as Pro, Ile, Val, and Thr at C-terminus of the middle segments and relatively reactive amino acids such as Gly and Ala at C-terminus of N-terminal segments¹⁰³  (Fig. 5A). Generally, to keep the original thioesters intact, thiol catalysts are not used in the first ligation and then added in the second ligation to accelerate the ligation. As an extended strategy of this KCL, by employing N-terminal selenocysteine instead of cysteine, deselenization step was inserted between the first and second ligation to afford mono-selenocysteine substituted analogue of human thiosulfate:glutathione sulfurtransferase (TSTD1).⁷⁵ 

One-pot three segment KCL employing two different alkyl thioester reactivities were also developed.⁶⁵  In the first ligation, N-terminal peptide bearing alkyl thioester derived from trifluoroethanethiol (TFET) with low pK _a of ∼7.3 was ligated with N-terminal Cys middle segment bearing less reactive alkyl thioester (Fig. 5C). After the completion of the first ligation, C-terminal peptide was added along with TFET as an external thiol additive in the second ligation. In this strategy, thanks to the property of the alkyl thioester, desulfurization was also conducted in one-pot fashion after the second NCL whereas aryl thioesters are known to inhibit desulfurization.¹⁰⁴  This KCL method successfully produced site-specifically sulfated madanin-1⁶⁶  and anopheline.¹⁰⁵ 

In the conventional three segment KCL, the selection of Val thioester in the first ligation was not favored because of the relatively slow kinetics in NCL. However, Wang et al. addressed this issue by the introduction of β-mercaptan on the C-terminal valine residue (i.e. the introduction of penicillamine residue instead of Val), which enables the formation of highly reactive strained cyclic thioester, β-thiolactone, as a C-terminal valyl thioester surrogate (Fig. 5D).¹⁰⁶  Consequently, the first ligation between N-terminal peptides bearing the β-thiolactone moiety and middle segments bearing C-terminal less reactive alkyl thioesters was chemoselectively coupled in the absence of thiol additives. After the second ligation with conventional NCL conditions containing MPAA, two different peptidic hormone, LPH and hPTH were synthesized via desulfurization of isolated full-length thiolated mutants.

Remarkably high reactivity of DSL compared to NCL was also utilized for one-pot three segment KCL (Fig. 6A). Cysteine-free 94 residue early secretory antigenic protein-6 was synthesized via tandem 2 step ligation consisting of the first additive free DSL and the second TFET-mediated NCL followed by desulfurization in one-pot manner.³⁷  This strategy was expanded to use β-selenoleucine instead of selenocysteine in the first DSL and β-thioaspartic acid in the second NCL for the synthesis of UL22A protein.¹⁰⁷ 

The different reactivity between aryl thioester and alkyl thioester was also employed in direct coupling method between peptide thioester and N-terminal amino group to achieve one-pot N-to-C three segment ligation.¹⁰⁸  The key for the regioselective peptide bond formation was addition of silver ion. N-terminal amino group of the middle segment selectively reacted with aryl thioester of N-terminal segment in the absence of Ag⁺, but not with the alkyl thioester of the middle segment itself (Fig. 6B). Then, Ag⁺-assisted second ligation was conducted between the alkyl thioester and N-terminal amino group of C-terminal segment that include O-acylisopeptide structure¹⁰⁹  to increase its solubility. Importantly, this one-pot ligation included Gly–Lys and Gly–Asp junction, indicating that steric hindrance was not affected for the regioselectivitiy of the first ligation. It is also notable that primary amines of Lys side chain and the O-acylisopeptide moiety were protected by isonicotinyloxycarbonyl group¹¹⁰  to retain the hydrophilicity of the peptide. As a result, Tim-3 Ig domain carrying GlcNAc was successfully obtained. Similar one-pot N-to-C three segment ligation was applied to the synthesis of Sec-substituted ferredoxin¹¹¹  and membrane protein caveolin-1.¹¹²  This thioester–amine direct coupling method was further expanded to one-pot four segment ligation, where aryl selenoester was revealed to selectively react with N-terminal amino group of a middle segment bearing C-terminal aryl thioester in the presence of N,N-diisopropylethylamine and dipyridyldisulfide in DMSO solution.¹¹³  By combining the aforementioned three segment ligation strategy, four peptide segments were successfully assembled in one-pot manner to provide 153 amino acid superoxide dismutase.

Combinatorial use of different ligation mechanism can also provide another type of one-pot peptide ligation. In 2016, Li et al. reported one-pot N-to-C three segment ligation with STL followed by NCL (Fig. 6C).¹¹⁴  A middle segment bearing N-terminal Thr and C-terminal aryl thioester was ligated with N-terminal segment with salicylaldehyde ester in the first STL, and after acidic acetal ring opening reaction and solvent exchange from pyridine/acetic acid to NCL buffer, the second NCL was conducted with N-terminal Cys segment to successfully synthesize a 66 amino acid partial segment of interleukin-25.

Another category of one-pot N-to-C ligation has emerged by utilizing C-terminal acyl donor precursors. In NCL-mediated one-pot N-to-C ligation, thioester precursors were employed to control peptide C-terminal reactivity. Originally, C-terminal thioester precursors have been developed to address the fact that the thioester moiety is labile in the basic condition used in Fmoc-based SPPS.¹⁴  One-pot N-to-C peptide ligation can be achieved through activating latent thioester precursors of the middle segments after the first ligation and conducting the second ligation in the same ligation mixture. Unlike KCL, in which two reactive acyl donors exists in the first ligation, this strategy should effectively suppress the side reactions such as self-cyclization and oligomerization of a middle segment because of the ignorable reactivity of the acyl donor precursors. Therefore, ligation junction would be selected in more flexible fashion regardless of the properties of C-terminal amino acids.

As pioneering research on Fmoc-compatible thioester precursors, in 2008, Blanco-Canosa and Dawson reported N-acyl-benzimidazolinone (Nbz) that can be synthesized from 4-diaminobenzamide (Dbz) moiety on solid support and readily used in NCL without a solution-phase activation step.¹¹⁵  To overcome synthetic issues of Nbz, its structure was improved to create N-acyl-N′-methyl-benzimidazolinone (MeNbz) in 2015.¹¹⁶  In 2018, Blanco-Canosa et al. reported in situ formation of MeNbz in the ligation mixture via intramolecular cyclization of p-cyanophenyloxycarbonyl o-amino(methyl)aniline (pCN-Phoc-MeDbz), which is a isolable intermediate in the conversopm from MeDbz to MeNbz (Fig. 7A).¹¹⁷  pCN-Phoc-MeDbz efficiently converts into MeNbz at pH 7.4–7.6, and MeNbz is subjected to NCL in a one-pot manner after thioesterification with a thiol catalyst such as 4-mercaptophenol. Using this activation method, one-pot three segment N-to-C ligation was performed with model peptides by adjusting pH 7.5 during pCN-Phoc-MeDbz-mdeiated second ligation after the completion of the first NCL at 6.8. A similar method using p-fluorophenyloxycarbonyl MeDbz, which is synthesized from readily available p-fluorophenyl chloroformate was also reported and applied to one-pot three segment N-to-C ligation.¹¹⁸ 

Among various thioester precursors, peptide hydrazide has become the most prevalent thioester precursor mainly because of the good availability and selective and efficient conversion to thioester. Liu et al. reported NaNO₂-mediated conversion of hydrazide to acyl azide and subsequent thioesterification with an external thiol in 2011.¹¹⁹  In 2018, another efficient thioesterification strategy based on peptide hydrazide using Knorr pyrazole synthesis¹²⁰  was also reported by Dawson et al. In this strategy, stoichiometric acetylacetone (acac) reacts with C-terminal hydrazide to form N-acyl pyrazole, and in situ thioesterification with external aryl thiol such as MPAA occurs under the aqueous acidic conditions around pH 3 (Fig. 7B). This thioesterification method was successfully applied for one-pot N-to-C three segment ligation with model peptides by repeatedly conducting thioesterification with stoichiometric acac at pH 3 and NCL at 7. This in situ hydrazide-to-pyrazole conversion was employed to form intramolecular thiolactone of a middle segment bearing C-terminal 4S-mercaptoproline in one-pot three segment N-to-C ligation.¹²¹  After desulfurization of the ligated product, natural Pro-containing variable number tandem repeat region of mucin 1 trimer was synthesized. Notably, by utilizing the high reactivity of the strained thiolactone, efficient one-pot ligation was realized at otherwise low reactive Pro junctions in both the first and second ligation.

Thioester precursors activated by N,S-acyl-shift have also been used for one-pot N-to-C ligation. In 2012, Melnyk et al. achieved one-pot N-to-C three segment ligation using the cyclic disulfide form of bis(2-sulfanylethyl) amide (SEA) as a latent thioester of the internal segment (Fig. 7C),¹²²  which was originally developed as an N,S-acyl-shift-mediated thioester surrogate.¹²³  This intramolecular disulfide bond is stable against moderate reducing reagents such as MPAA, while it can be cleaved by strong reducing reagents such as TCEP or DTT. Therefore, to achieve one-pot N-to-C ligation, the first ligation was performed with N-terminal segment bearing conventional alkl thioester and a cysteinyl middle segment baring C-terminal SEA in the presence of MPAA, and subsequently, TCEP was added to the ligation solution for the activation of SEA in the second ligation in a one-pot manner. K1 domain of the hepatocyte growth factor (HGF),¹²²  Human Defensin DEFB133,¹²⁴  and small-ubiquitin-like modifier (SUMO) 2/3 dimers^125,126  were successfully synthesized through this strategy. As the selenium analog of SEA, bis(2-selenylethyl)amide (SeEA), which is an N,Se-acyl-shift-mediated selenoester precursor, was also developed and applied to one-pot N-to-C three and four segment ligation (Fig. 7C).¹²⁷  As the diselenide form of SeEA is unreactive with DTT but activated by TCEP, SEA can be selectively used as an acyl donor in the presence of SeEA. By combining the previously developed one-pot three segment ligation explained above, one-pot four segment ligation was demonstrated to synthesize a 36 amino acid branched peptide. Interestingly, SeEA peptide can be applied to another type of one-pot N-to-C ligation because of the faster kinetics in NCL than SEA. In the presence of TCEP, N-terminal peptide bearing SeEA was selectively reacted with N-terminal Cys of a middle segment bearing SEA because of the faster kinetics, although SEA was also activated in the condition. This KCL-type one-pot N-to-C three segment ligation was successfully employed for the synthesis of biotinylated NK1 domain of HGF.¹²⁸ 

The use of two orthogonally activatable thioester precursors, cysteinylprolyl ester (CPE)¹²⁹  and N-alkylcysteine (NAC)¹³⁰  enabled one-pot N-to-C three segment ligation. CPE converts into thioester via N,S-acyl shift and subsequent 2,5-diketopiperazine formation at weakly basic conditions (pH > 7.8), while N,S-acyl shift of NAC is accelerated under slightly acidic conditions (pH 4–6) (Fig. 7D). By leveraging this orthogonality, three peptide segments of histone H4 were assembled in a one-pot manner through CPE-mediated first ligation at pH 7.8 and following NAC-mediated second ligation at pH 5.5.¹³¹  This strategy was extended to one-pot four segment ligation by combining MPOH thioester in the first ligation with the aforementioned pair of thioester precursors for the synthesis of O-glycosylated histone H2A at the 40th serine residue.¹³² 

As another example of the use of N,S-acyl-shift-mediated thioester precursor, Otaka et al. have employed N-sulfanylethylanilide (SEAlide)¹³³  for one-pot N-to-C peptide ligation. SEAlide possesses a unique property; phosphate accelerates N,S-acyl shift of SEAlide as acid–base catalysis to form thioester and therefore can start peptide ligation (Fig. 7E).^134,135  Using this activation mechanism, a middle segment bearing C-terminal SEAlide was first ligated with N-terminal segment in the absence of phosphate and then, the second ligation was conducted in the presence of phosphate to complete one-pot N-to-C three segment ligation to afford human atrial natriuretic peptide.¹³⁶  This strategy was extended to one-pot four segment ligation by combining KCL mechanism including the different reactivity between arylthioester and alkylthioester to synthesize a model peptide. This SEAlide-mediated three segment ligation was also effectively applied for the synthesis of monoglycosylated human GM2-activator protein¹³⁷  and CXCL14.¹³⁸  Interestingly, this N-to-C three segment ligation proceeded even in the presence of C-terminal peptide in the first ligation.¹³⁹  Furthermore, Otaka et al. also reported photo-caged SEAlide for the one-pot N-to-C ligation.⁹³  N,S-Acyl shift is completely blocked by introducing NV protecting group on thiol moiety of SEAlide, and the NV group is readily removed upon UV light irradiation. To suppress the undesired byproduct generated by thiyl radical derived from thiophenol, extraction of thiophenol with diethyl ether is necessary before the photolysis reaction. The synthesis of reduced form SNX-482 was achieved via one-pot N-to-C four segment ligation by repeating NCL and UV irradiation. Notably, two-pot four segment convergent ligation was also accomplished for the synthesis of the same target using N-to-C ligation with photo-caged SEAlide peptide and C-to-N ligation with NVOC-protected N-terminal cysteine peptide.

KAHA ligation was also applied to one-pot N-to-C ligation by introducing photolabile protecting group into α-ketoacid.⁹⁸  For the protection of C-terminal α-ketoacid, 6-membered-ring acetal protecting group bearing 2-nitrophenyl was designed and synthesized. A middle segment with N-terminal 5-oxaproline and C-terminal photo-protected α-ketoacid was first ligated with N-terminal segment bearing non-protected α-ketoacid, and after removal of the photolabile protecting group, C-terminal segment was ligated to produce NEDD8 protein. Moreover, NEDD8 protein was also synthesized via two-pot four segment convergent ligation by integrating this photolysis-mediated one-pot N-to-C strategy and a C-to-N strategy mentioned in the Section 3.2.4.

In this review, recent advances of CPS via one-pot ligation strategies are summarized. The method of one-pot multiple peptide ligation allows us to not only save time by skipping laborious purification steps but also increase the yield of target proteins by avoiding undesired aggregation and adsorption of intermediate peptide segments in purification and isolation process. Currently, the range of target proteins that can be prepared through CPS is continuously expanding by virtue of repeatedly improved and sophisticated peptide chemistries including solid-phase synthetic chemistry, thioester synthesis, peptide solubilization strategy, and so on. Among these essential techniques, one-pot ligation chemistries are also indispensable for efficient production of sufficient amount of proteins.

While various one-pot ligation methods developed so far have already streamlined the procedures of CPS, there are still several issues that are important but not yet achieved: (1) one-pot ligation of six or more segments is not achieved. (2) The maximum size of one-pot ligated peptide is around 200 amino acids. (3) The most of ligation are conducted in mM concentration. (4) The direction of one-pot ligation is limited to one way in a sequential fashion (C-to-N or N-to-C). To address these issues, successive evolution of peptide chemistry is absolutely essential. Continuous development of peptide terminal activity controlling strategies such as novel ligation-compatible protecting groups and acyl donor precursors would be of importance as the employment of the difference of reaction kinetics seems to include difficulty in one-pot five or more segment ligation. Development of a pair of protecting groups that are orthogonally removable in ligation mixture could switch the ligation direction in one-pot manner. This methodology would be precious because the flexibility of the ligation order of peptide segments can avoid the undesired aggregation or precipitation of intermediate peptides to reach the full-length target proteins. To synthesize aggregation- or precipitation-prone proteins such as membrane proteins and large proteins, fast and efficient ligation in more diluted conditions will be demanded. To this end, ligation accelerating auxiliary devices would be a one promising methodology, in addition to further development of ongoing Se-based strategies. Peptide aggregation-disrupting chemistries will be also required efforts. One-pot convergent peptide ligation is still challenging although “two-pot convergent ligation” where the first two ligations (e.g. segment 1–2 and 3–4) are conducted in the separate reaction solutions and then, the second ligation is performed by mixing these solutions, are reported.^93,98  If the site-selective two ligations of segment 1–2 and 3–4 are realized in one-pot by using, for example, a proximity effect of nucleic acid hybridization, one-pot convergent ligation will be demonstrated. Furthermore, simultaneous ligation of three or more peptide segments in one-pot manner could be the most ideal way to assemble multiple peptide segments, possibly by using multiple ligation mechanisms that proceeds in the same reaction conditions or strictly controlled proximity effect-mediated ligation. It is also noteworthy that since the HPLC-based purification and isolation of large proteins are often troublesome, some innovative technology for separating proteins of interest from by-products is awaited. In future where the more sophisticated CPS technologies than current ones will be available, cutting-edge one-pot ligation strategies will not only be utilized for accessing to new target proteins in academic filed, but also impact on general society with the aim of broader applications such as drug discovery and material science by associating process chemistry area to produce large amount of synthetic proteins in industrial settings.