A program for ligation at threonine sites: application to the controlled total synthesis of glycopeptides

Abstract

A method by which to accomplish formal threonine ligation has been developed. The method accomplishes ligations of two peptide domains. We have also demonstrated the ability to successfully ligate two independent glycopeptide domains.

1. Introduction

Glycoproteins are a class of naturally occurring biomacromolecules, which are biosynthesized through post-translational protein glycosylation. A great deal of effort has been directed toward the examination of the role that glycosylation plays in various critical protein functions, such as protein folding, proteolytic stability, and intercellular communication.¹ Of particular interest is the fact that many glycoproteins possess exploitable therapeutic activity, and may serve as promising candidates in the development of vaccines,² diagnostic techniques,³ and therapeutic agents.⁴ Prominent examples of therapeutically valuable glycoproteins include the erythropoietic agent, erythropoietin (EPO)^4a,5 and the fertility agent, human follicle stimulating hormone (hFSH).⁶ Despite considerable interest in this class of biomacromolecules, the field of glycobiology faces a significant obstacle to the rigorous evaluation of glycoproteins: the isolation of significant quantities of homogeneous glycoproteins from natural sources is often prohibitively difficult, due to the fact that most naturally occurring glycoproteins are biosynthesized as heterogeneous mixtures of glycoforms.

Given the great difficulties associated with the isolation of homogeneous glycoproteins, we recognized that an opportunity for chemical synthesis might lay in the challenge of using total chemical synthesis to gain access to homogeneous glycoprotein samples.⁷ The biology of such agents could then be studied in further detail. Moreover, through chemical synthesis, it would be possible to gain access to fully synthetic analog glycoproteins, possessing targeted modifications in the carbohydrate or protein domains. Of course, such targets could not necessarily be obtained through purely biologic means, since they would lack the viable biosynthetic pathways for reaching the desired structural types. In the context of our glycoprotein synthesis program, we have targeted for total synthesis the clinically relevant glycoproteins, EPO and hFSH.⁸ Several years ago, at the outset of our EPO total synthesis endeavor, we identified a number of significant methodological shortcomings that would need to be addressed before a viable total synthesis effort could be launched.

In particular, we were concerned with the dearth of general techniques and protocols available for the merging of two glycopeptide fragments. To address this issue, we developed an efficient method for the coupling of two differentially glycosylated peptide fragments. This advance was predicated on the landmark break-through of Kent and co-workers in the field of peptide synthesis, termed native chemical ligation (NCL).⁹ NCL is a widely used technique that allows for the merger of large peptide fragments, one of which presents a C-terminal thioester, and the other an N-terminal cysteine residue (Fig. 1a). Direct extension of the NCL method to the realm of glycopeptide synthesis is complicated by the difficulties associated with synthesizing pre-formed glycopeptide thioester. The solution developed in our laboratories (outlined in Fig. 1b) involves a relatively stable C-terminal ortho-thiophenolic ester, presented on one of the glycopeptides, and a protected N-terminal cysteine residue incorporated on the other fragment.¹⁰ Upon simultaneous reduction of the two disulfides, the phenol moiety undergoes intramolecular O→S migration, providing an intermediate thioester, which undergoes thioester exchange with the free cysteine of the second glycopeptide. The intermediate then suffers spontaneous intramolecular transfer to yield the bidomainal glycopeptide adduct, incorporating two differentiated sites of glycosylation. The general logic of this NCL method has more recently been extended to a direct oxo-ester variant, in which a phenolic ester equipped with para-NO₂ or para-CN substitution, is sufficiently activated to undergo direct cysteine ligation without the need for the intermediacy of a thioester species.¹¹

Figure 1.

(a) Native chemical ligation; (b) Glycopeptide NCL; (c) Free-radical-mediated reduction of Cys/Ala; (d) Proposed native chemical ligation at threonine.

These glycopeptide ligation protocols do suffer from an important limitation in terms of generalizability, in that they require the presence of a cysteine residue at the ligation site. In fact, there is often a paucity of cysteine sites in naturally occurring proteins and glycoproteins. The need for a menu of efficient cysteine-free ligation methods thus remains quite high.

A number of useful solutions to the cysteine ligation problem have been developed, many of which depend on the use of cysteine or a thiol-containing amino acid surrogate in the ligation step. Following thiol-mediated ligation, the cysteine or surrogate residue is converted to the desired amino acid. In this way, methionine ligation has been achieved through homo-cysteine coupling, followed by post-ligational methylation.¹² Similarly, serine ligation has been accomplished through NCL followed by conversion of cysteine to serine.¹³ A number of cysteine-free ligation methods make use of a two-step ligation–desulfurization sequence. For instance, metal-based post-ligational thiol reduction has been used to accomplish formal alanine¹⁴ and phenylalanine¹⁵ ligations. Despite the appeal of such a strategy, we noted that traditional desulfurization methods suffer from a lack of substrate generality, due to the susceptibility of many common functional groups (particularly thiol moieties) to the standard harsh reduction conditions. Based on the disclosure by Hoffmann and Walling in the 1950s,¹⁶ we recently developed a very mild and chemoselective free-radical-mediated desulfurization strategy, which allows for the post-translational conversion of cysteine to alanine in complex glycopeptide and peptide settings (Fig. 1c).¹⁷ This method has proven to be tolerant of a wide range of functionalities, including sulfur-containing groups—such as Thz, Cys(Acm), biotin, and thioesters —as well as amino acids, including methionine, and even complex carbohydrate moieties. This overall alanine ligation strategy has now been successfully applied in the context of the synthesis of a complex glycopeptide fragment of erythropoietin (Ala¹-Gly²⁸).^8c In addition, our mild sulfur reduction protocol has been employed to accomplish a formal valine ligation, through the coupling and post-ligational reduction of unnatural amino acid surrogates (γ-thiol valine¹⁸ or β-thiol valine^18,19) to valine residues. Moreover, a dual native chemical ligation at lysine has also been achieved recently. A γ-thiol group on the N-terminal lysine mediates double chemical ligation at both α and ε amines, followed by a free-radical desulfurization.²⁰

The development of a variety of methods by which to formally accomplish NCL at a diverse range of amino acid residues could be of substantial value. As noted above, due to its low frequency in nature, direct ligation at cysteine itself is of limited practical value. Furthermore, in developing a synthetic strategy toward a glycoprotein or glycopeptide target, it is often desirable to merge two pieces of relatively equal size. As the menu of amino acid ligation options expands, greater flexibility may be brought to the design of synthetic routes. Finally, the two-step ligation/reduction strategy provides the opportunity to explore the consequences of protein engineering with site-specifically modified glycoproteins.

In the hopes of broadening the range of options for amino acid ligation, we sought to extend our two-step ligation/reduction protocol to the development of a formal threonine ligation. Threonine was selected for its relative abundance in nature, particularly in comparison with cysteine. We describe herein the discovery and formulation of a mild and efficient two-step formal threonine ligation protocol.

The central idea is outlined in Figure 1d. Thus, as in the case of the valine ligation,¹⁸ a thiol-containing threonine surrogate would be incorporated at the N-terminus of Peptide 2. One could envision two possible sites at which to install a thiol group onto the threonine residue: at the β position or the γ position. We expected that the γ-thiol threonine would serve as the more productive surrogate in establishing the initial acylation event. Peptide 2, incorporating the γ-thiol threonine, was expected to undergo trans-thioester-ification with Peptide 1 (presented as either a thioester or activated oxo-ester), to generate a thioester-linked intermediate, which would then undergo spontaneous intramolecular acyl transfer, generating the new amide bond. Subsequent radical-based desulfurization would serve to remove the thiol, providing the target peptide with threonine at the ligation site.

2. Results and discussion

Our first objective was to synthesize a γ-thiol threonine amino acid surrogate. To accomplish this, we drew on the earlier work of Rapoport and co-workers,²¹ in which allo-threonine was obtained from D-vinylglycine. Through modification of the Rapoport route, we sought to accomplish the diastereoselective syntheses of the γ-thiol threonine derivatives, 5 and 6 (Scheme 1). Compound 5 was to be used directly for single amino acid extension studies (see Table 1), while compound 6 would be elaborated to peptide 37, which would serve as the substrate in the threonine ligation studies at the peptide level (see Table 2). Thus, as shown in Scheme 1, vinylglycine 1²² was epoxidized with an excess of mCPBA to afford a 5:1 ratio of syn and anti epoxides, 2a and 2b, which could be separated by chromatography. The major diastereomer, 2a, has the desired syn configuration, perhaps as a consequence of hydrogen bonding of mCPBA to the nitrogen functionality.²³ Upon exposure to the sodium salt of thioacetic acid, epoxide 2a was opened to provide the acetylated thiol, 3. The latter was transformed, in a straightforward fashion, to the target compounds, 5 and 6.

Scheme 1.

Synthesis of g-thiol threonine.

Table 1.

Threonine extension by coupling and subsequent free-radical desulfurization

Entry	Peptide	Coupling Product/Yield/Time	Desulfurization Product/Yield/Time
1			FmocRNEDLSGAT—OMe 9 / 92 % / 2 h
2			FmocRLGDSTAGQT—OMe 12 / 81% / 2 h
3			FmocRLGDSTAGYT—OMe 15 / 89% / 2 h
4			FmocRLGDSTAGWT—OMe 18 / 90% / 2 h
5			FmocRTGDSAGTT—OMe 21 / 84% / 2 h
6			FmocRLGDSTAGLT—OMe 24 / 80% / 2 h
7			FmocKYDSRGFT—OMe 27 / 80% / 2 h
8			FmocRTGDSAGVT—OMe 30 / 95% / 2 h
9			FmocRTGDSAGIT—OMe 33 / 82% / 2 h
10			FmocVRSYTAGPT—OMe 36 / 83% / 2 h

Key: (a) buffer at pH 6.5 (6.0 M Gn·HCl, 188.8 mM Na₂HPO₄), TCEP, rt (b) TCEP, VA-044, tBuSH, 37 °C. VA-044=2,2′-azobis-[2-(2-imidazolini-2-yl)propane] dihydrochloride.

TCEP=tris(2-carboxyethyl)phosphine.

Table 2.

NCL at threonine through ligation and subsequent free-radical desulfurization

Entry	Peptide A	Ligation Product/Yield/Time	Desulfurization Product/Yield/Time
1			FmocRLGDSTAGYTGAPRHSWG—OMe 39 / 98% / 2 h
2			FmocRLGDSTAGWTGAPRHSWG—OMe 41 / 85% / 2 h
3			FmocRTGDSAGITGAPRHSWG—OMe 43 / 96% / 2 h
4

Key: (a) buffer at pH 6.5 (6.0 M Gn·HCl, 188.8 mM Na₂HPO₄), TCEP, RT; (b) TCEP, VA-044, tBuSH, 37 °C.

With γ-thiol threonine derivative 5 in hand, we next sought to evaluate the feasibility of the proposed threonine ligation method. We first examined the protocol in the context of a single amino acid extension of a variety of peptide substrates. As shown in Table 1, when a relatively sterically less demanding amino acid (Ala, Gln, Tyr, Trp, Phe) was presented at the C-terminus of the peptide, the amino acid extension generally proceeded very quickly and with good yield (see entries 1–4 and 7). Even when the β-branched amino acid, threonine, was incorporated at the C-terminus, the coupling was complete within 1 h (entry 5). As expected, as the C-terminus became more sterically hindered (Leu, Val, Ile, Pro), the reaction rate suffered. However, coupling could still be accomplished within a reasonable time frame, and in moderate to good yields (entries 6, 8–10). As shown in Table 1, a variety of different C-terminal esters participated successfully in the extension protocol, including thiophenyl ester, ortho-thiophenolic ester, and para-nitrophenyl ester. The diversity of C-terminal esters amenable to this protocol is of significance, as the ortho-thiophenolic ester is compatible with glycopeptide ligation, while para-nitrophenyl ester is particularly efficient at promoting coupling at sterically hindered ligation sites. We also note that the threonine ligation protocol is able to accommodate the presence of an unprotected lysine in the substrate peptide (25, entry 7). All of the coupling products obtained in Table 1 were subsequently subjected to our standard radical-based desulfurization conditions to provide the desired threonine extension products in very good yields.

We now sought to examine the two-step ligation/reduction protocol in the context of a peptide-peptide coupling. Peptide 37, possessing the γ-thiol threonine surrogate at its N-terminus, was prepared from compound 6 (see Supplementary data for details). According to our general procedure, 37 and the peptide coupling partner were dissolved in a guanidine buffer solution. Upon addition of TCEP, the disulfide moieties were cleaved to presumably give rise to the free thiol functionalities, which then underwent the anticipated ligation reaction. As shown in Table 2, we found the ligation rate to be dependent on the nature of the C-terminal amino acid. Thus, when a less sterically demanding amino acid—such as Tyr (13) or Trp (16)—was present at the C-terminus, ligation was complete within 1 h (entries 1 and 2). However, in the case of the more hindered amino acid, Ile (31), the reaction took up to 7 h to reach completion (entry 3). Once more, in each case, subsequent free-radical-based desulfurization was readily achieved in high yields through use of our mild reduction method. Finally, as shown in Table 2, entry 4, our new protocol could be readily extended to a glycopeptide ligation setting. Thus, peptide 44, presenting an N-linked hexasaccharide domain, underwent ligation with peptide 37 to provide glycopeptide 45 in 82% yield. Upon exposure to our previously described reduction conditions, glycopeptide 46, incorporating threonine at the ligation site, was obtained in 96% yield.

3. Conclusion

In conclusion, we have described the development of a useful new entry in the field of native chemical ligation. Through an efficient two-step ligation/reduction protocol, it is now possible to formally achieve NCL at threonine sites, in both peptide and gly-copeptide settings. This methodological advance, taken in concert with the previous entries of our group and others, has served to significantly expand the NCL menu, which was originally restricted to cysteine-based ligations. Further applications and extensions of this method to the synthesis of important biologic level agents are underway in our laboratory.

4. Experimental section

4.1. General

Anhydrous THF, diethyl ether, CH₂Cl₂, toluene, and benzene were obtained from a dry solvent system (passed through column of alumina) and used without further drying. NMR spectra (¹H and ¹³C) were recorded on a Bruker Advance DRX-500 MHz, or a Bruker DRX-600 MHz. Low-resolution mass spectral analyses were performed with a JOEL JMS-DX-303-HF mass spectrometer or Waters Micromass ZQ mass spectrometer.

HPLC: All separations of peptides and glycopeptides involved a mobile phase of 0.05% TFA (v/v) in water (solvent A)/0.04% TFA in acetonitrile (solvent B). Preparative and analytical HPLC separations were performed using a Rainin HPXL solvent delivery system equipped with a Rainin UV-1 detector. LC–MS chromatographic separations were performed using a Waters 2695 Separations Module and a Waters 996 Photodiode Array Detector equipped with X-Bridge™ C18 column (5.0 μm, 2.1×150 mm), X-Terra™ MS C18 column (3.5 μm, 2.1× 100.0 mm) or Varian Microsorb C18 column (2×150mm) at a flow rate of 0.2 mL/min. HPLC separations were performed using: X-Bridge™ Prep C18 column OBD™ (5.0 mm, 19×150 mm), a flow rate of 16 mL/min. Microsorb 100-5 C18 column at a flow rate of 16.0 mL/min or Microsorb 300-5 C4 column at a flow rate of 16.0 mL/min.

Automated peptide synthesis was performed on an Applied Biosystems Pioneer continuous flow peptide synthesizer. Peptides were synthesized under standard automated Fmoc protocols. The deblock mixture was a mixture of 100/5/5 of DMF/piperidine/DBU. Upon completion of automated synthesis on a 0.05 mmol scale, the peptide resin was washed into a peptide synthesis vessel with CH₂Cl₂. The resin cleavage was effected by treatment with AcOH/ TFE/CH₂Cl₂ (2:2:6) for 2 ×1 h to yield peptidyl acids in good yield. The peptidyl acids were modified on C-terminus and/or N-terminus. The resulting peptides were subjected to a deprotection cocktail (60.0 mg of phenol, 0.2 ml of water, 0.15 ml of triisopro-pylsilane, and 3.0 ml TFA) for 2.0 h. TFA was removed by N₂. The oily residue was triturated with diethyl ether to give a white suspension, which was centrifuged and the ether subsequently decanted. The resulting solid was ready for HPLC purification.

4.2. Preparation and characterization of compounds 2a–6

4.2.1. (S)-Methyl 2-(((9H-fluoren-9-yl)methoxy)carbonylamino)-2-((S)-oxiran-2-yl) acetate (2a)

To a stirred solution of 1 (200 mg, 0.593 mmol) in CH₂Cl₂ (6 mL) at 0 °C was added m-chloroperbenzoic acid (1.022 g, 5.93 mmol) and the reaction was warmed to rt. After 24 h, the mixture was filtered through a glass filter, the solids were extracted with CH₂Cl₂, and the combined organic phase was washed with 10% NaHCO₃, water, dried, and evaporated. The two diastereomers were then separated by chromatography (CH₂Cl₂/ MeOH=300:1) to give 2a (138 mg, 67% yield) as a white solid. ¹HNMR (500 MHz, CDCl₃) δ 2.59 (m,1H), 2.76 (m,1H), 3.46 (s, 1H), 3.81 (s, 3H), 4.20 (t, J=6.8 Hz, 1H), 4.40 (d, J=6.9 Hz, 2H), 4.70 (dd, J= 1.5 Hz, 8.9 Hz, 1H), 5.23 (d, J=8.8 Hz, 1H), 7.29 (m, 2H), 7.32 (t, J=6.3 Hz, 2H), 7.56 (t, J=6.0 Hz, 2H), 7.76 (d, J=6.3 Hz, 2H); ¹³C NMR (125 MHz, CDCl₃) δ 43.9, 47.1, 51.2, 53.0, 53.1, 67.2, 120.1, 125.1, 127.1, 127.1, 127.8, 141.3, 141.4, 143.6, 143.8, 156.2, 170.2; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ −5.21 (c 1.08, CHCl₃); IR (liquid film) (v_max/cm⁻¹): 3342, 3066, 3019, 2953, 2848, 1724; HRMS: m/e calcd for C₂₀H₁₉NO₅Na⁺: 376.1161, found: 376.1153.

4.2.2. (2S,3S)-Methyl 2-(((9H-fluoren-9-yl)methoxy)carbonylamino)-4-(acetylthio)-3-hydroxybutanoate (3)

Compound 2a was dissolved in toluene (3.2 mL) and treated with a solution of sodium acetate (88 mg, 1.077 mmol) and thioacetic acid (80 μL, 1.077 mmol) in DMF (1.6 mL). The reaction mixture was stirred at rt for 2 h. The mixture was concentrated under N₂. The concentrate was dissolved in EtOAc and washed with NH₄Cl, water, and brine. The organic layer was dried, concentrated, and purified by chromatography (Hexane/EtOAc=2:1) to give 3 (72 mg, 80%) as a light yellow foam. ¹H NMR (500 MHz, CDCl₃) δ 2.35 (s, 3H), 3.01 (m, 3H), 3.75 (s, 3H), 4.23 (m, 2H), 4.41 (d, J=7.0 Hz, 2H), 4.51 (d, J=9.3 Hz, 1H), 5.69 (d, J9.4 Hz, 1H), 7.31 (m, 2H), 7.39 (m, 2H), 7.61 (m, 2H), 7.75 (m, 2H); ¹³C NMR (125 MHz, CDCl₃) δ 30.6, 32.8, 47.2, 52.8, 57.3, 67.3, 71.5, 120.0, 120.0, 125.1, 125.2, 127.1, 127.8, 141.3, 141.4, 143.6, 143.8, 156.7, 170.9, 196.6; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ 11.69 (c 0.68, CHCl₃); IR (liquid film) (v_max/cm⁻¹): 3373, 3065, 3019, 2952, 2847, 1747, 1722, 1697; HRMS: m/e calcd for C₂₂H₂₃NO₆SNa⁺: 452.1144, found: 452.1155.

4.2.3. (2S,3S)-Methyl 2-(((9H-fluoren-9-yl)methoxy)carbonylamino)-3-hydroxy-4-(methyldisulfanyl) butanoate (4)

Compound 3 (47 mg, 0.11 mmol) was dissolved in MeOH (1.87 mL) and treated with 0.2 N NaOH solution (1.87 mL) at 0 °C for 20 min. The reaction mixture was carefully neutralized by the addition of 1 N HCl at 0 °C, diluted with EtOAc and washed with water and brine. The organic layer was concentrated and dried in vacuo, generating 43 mg of crude compound (2S,3S)-methyl 2-(((9H-fluoren-9-yl)methoxy)-carbonylamino)-3-hydroxy-4-mercaptobutanoate, which was directly used in the next step.

S-Methyl methanethiolsulfonate (36 μL, 0.38 mmol) and DIEA (12 mL, 0.11 mmol) were added to CH₂Cl₂ (0.5 mL). The crude residue of (2S,3S)-methyl 2-(((9H-fluoren-9-yl)methoxy)carbonyl-amino)-3-hydroxy-4-mercaptobutanoate in CH₂Cl₂ (0.8 mL) was added dropwise to the above solution and stirred at rt for 2 h. The reaction mixture was concentrated and purified by chromatography (Hexane/EtOAc=2:1), to provide 4 (33 mg, 70% in two steps) as a light yellow foam. ¹H NMR (500 MHz, CDCl₃) δ 2.42 (s, 3H), 2.65 (dd, J=9.2 Hz 14.2 Hz, 1H), 2.71 (d, J=2.5 Hz, 1H), 2.85 (dd, J=3.8 Hz, 14.0 Hz, 1H), 3.79 (s, 3H), 4.22 (t, J=6.8 Hz, 1H), 4.43 (m, 3H), 4.51 (d, J=9.4 Hz, 1H), 5.58 (d, J=9.4 1H), 7.29 (m, 2H), 7.37 (t, J=7.1 Hz, 2H), 7.60 (t, J=7.1 Hz, 2H), 7.75 (m, 2H); ¹³C NMR (125 MHz, CDCl₃) δ 22.9, 41.4, 47.2, 52.9, 57.1, 67.2, 69.9, 120.0, 120.1, 125.1, 127.1, 127.8, 141.3, 141.4, 143.6, 143.8, 156.7, 170.9; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ 48.90 (c 0.27, CHCl₃); IR (liquid film) (v_max/cm⁻¹): 3406, 3019, 2952, 2918, 2850, 1724; HRMS: m/e calcd for C₂₁H₂₃NO₅S₂Na⁺: 456.0915, found: 456.0909.

4.2.4. (2S,3S)-Methyl 2-amino-3-hydroxy-4-(methyldisulfanyl)butanoate (5)

To a stirred solution of 4 (57 mg, 0.131 mmol) in DMF (4.3 mL) was added diethylamine (1.4 mL). The reaction mixture was stirred at rt for 2 h and the solvent was evaporated. The residue was purified by chromatography (CH₂Cl₂/MeOH 40:1) to give 5 (25 mg, 89% yield) as a clear oil. ¹H NMR (500 MHz, CDCl₃) δ 2.43 (s, 3H), 2.84 (dd, J=6.0 Hz, 13.9 Hz, 1H), 2.92 (dd, J=7.0 Hz, 13.9 Hz, 1H), 3.66 (d, J=3.1 Hz, 1H), 3.76 (s, 3H), 4.17 (m, 1H); ¹³C NMR (125 MHz, CDCl₃) d 23.1, 41.7, 52.5, 56.2, 70.6, 174.3; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ 76.7 (c 0.3, CHCl₃); IR (liquid film) (v_max/cm⁻¹): 3366, 3303, 2952, 2918, 1738, 1437, 1245, 1174, 1022; HRMS: m/e calcd for C₆H₁₃NO₃S₂H⁺: 212.0415, found: 212.0413.

4.2.5. (2S,3S)-2-(tert-Butoxycarbonylamino)-3-hydroxy-4-(methyldisulfanyl) butanoic acid (6)

To a solution of 5 (28 mg, 0.133 mmol) and Boc₂O (57.8 mg, 0.265 mmol) in THF (0.7 mL) and MeOH (0.5 mL) was added TEA (0.056 mL, 0.399 mmol). The mixture was stirred for 2 h at room temperature. The reaction mixture was extracted with EtOAc three times. Combined organic layers were dried and purified by chromatography (Hex/EtOAc 2:1) to give (2S,3S)-methyl 2-(tert-butoxycarbonylamino)-3-hydroxy-4-(methyl-disulfanyl) butanoate (30 mg, 75% yield) as a clear oil. ¹H NMR (500 MHz, CDCl₃) d 1.43 (s, 9H), 2.42 (s, 3H), 2.71 (m, 1H), 2.88 (m, 1H), 3.78 (s, 3H), 4.42 (d, J=8.8 Hz, 2H), 5.32 (d, J=8.4 Hz, 1H); ¹³C NMR (125 MHz, CDCl₃) δ 22.9, 28.3, 41.6, 52.8, 56.7, 70.0, 80.4, 156.1, 171.2; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ 59.84 (c 0.37, CHCl₃); IR (liquid film) (v_max/cm⁻¹): 3407, 2979, 2918, 1752, 1716; HRMS: m/e calcd for C₁₁H₂₁NO₅S₂Na⁺: 334.0759, found: 334.0749.

(2S,3S)-methyl 2-(tert-butoxycarbonylamino)-3-hydroxy-4-(methyldisulfanyl) butanoate (20 mg, 0.064 mmol) was dissolved in THF (1 mL) and treated with 1 N NaOH solution (1 mL). The mixture was stirred for 2 h at rt. The reaction mixture was washed with Et₂O and the aqueous layer was adjusted to pH 3 using 1 N HCl, and extracted with Et₂O three times. The combined organic layer was dried and purified by chromatography (CH₂Cl₂/ MeOH=20:1 to 15:1) to give 6 (14 mg, 74% yield).*¹H NMR (500 MHz, CDCl₃) δd 1.39 (s, 9H), 2.38 (s, 3H), 2,70 (m, 1H), 2.84 (m, 1H), 4.44 (m, 2H), 5.45 (d, J=8.8 Hz, 1H); ¹³C NMR (125 MHz, CDCl₃) d 23.0, 28.3, 41.4, 56.6, 70.0, 80.7, 156.4, 174.7; ${\left[\alpha \right]}_{\mathrm{D}}^{20}$ 37.56 (c 0.28, CHCl₃); IR (liquid film) (v_max/cm⁻¹): 3393, 2979, 2929, 2853, 1710, 1691; HRMS: m/e calcd for C₁₀H₁₉NO₅S₂Na⁺: 320.0602, found: 320.0602.

4.3. General procedure for native chemical ligation at threonine site

To a solution of Peptide A (1.0 equiv, 4 mmol) and Peptide B (1.5 equiv, 6 μmol) in 0.5 mL of Guanidine buffer^‡ was added 0.5 M bond-breaker® TCEP solution (Pierce) (5.0 equiv, 20 mmol). The reaction mixture was stirred at room temperature. The reactions were monitored by LC–MS and purified directly by HPLC upon consumption of the starting material.

4.4. General procedure for desulfurization

To a solution of the peptide (0.6 mM) in 200.0 mL of water (or buffer) and 100.0 mLofCH₃CN were added 200.0 μL of 0.5 M bond-breaker® TCEP solution (Pierce), 20.0 μL of 2-methyl-2-propanethiol and 10.0 mL of radical initiator (0.1 M in water). The reaction mixture was stirred at 37 °C. The reactions were monitored by LC– MS and purified directly by HPLC upon consumption of the starting material.

4.4.1. MS characterization of peptides 7–46

• Compound 7. ESIMS calcd for C₅₆H₇₄N₁₂O₁₇S₂ [M+H]⁺ m/z=1251.48, found: 1251.60.

• Compound 8. ESIMS calcd for C₅₃H₇₅N₁₃O₁₉S [M+H]⁺ m/z=1230.51, found: 1231.12.

• Compound 9. ESIMS calcd for C₅₃H₇₅N₁₃O₁₉ [M+H]⁺ m/z=1198.54, found: 1198.88.

• Compound 10. ESIMS calcd for C₅₈H₇₉N₁₃O₁₇S₂ [M+H]⁺ m/z=1294.53, found: 1294.79.

• Compound 11. ESIMS calcd for C₅₅H₇₉N₁₄O₁₉S [M+H]⁺ m/z=1273.55, found: 1274.08.

• Compound 12. ESIMS calcd for C₅₅H₈₀N₁₄O₁₉ [M+H]⁺ m/z=1241.58, found: 1241.91.

• Compound 13. ESIMS calcd for C₆₀H₇₆N₁₂O₁₆S [M+H]⁺ m/z=1253.53, found: 1254.20.

• Compound 14. ESIMS calcd for C₅₉H₈₁N₁₃O₁₉S [M+H]⁺ m/z=1308.56, found: 1308.92.

• Compound 15. ESIMS calcd for C₅₉H₈₁N₁₃O₁₉ [M+H]⁺ m/z=1276.59, found: 1277.07.

• Compound 16. ESIMS calcd for C₆₂H₇₇N₁₃O₁₅S [M+H]⁺ m/z=1276.55, found: 1276.81.

• Compound 17. ESIMS calcd for C₆₁H₈₂N₁₄O₁₈S [M+H]⁺ m/z=1331.58, found: 1331.93.

• Compound 18. ESIMS calcd for C₆₁H₈₂N₁₄O₁₈ [M+H]⁺ m/z=1299.60, found: 1300.02.

• Compound 19. ESIMS calcd for C₄₉H₆₃N₁₁O₁₅S [M+H]⁺ m/z=1078.43 [M+2H]²⁺ m/z=539.72, found: 1078.58, 540.12.

• Compound 20. ESIMS calcd for C₄₈H₆₈N₁₂O₁₈S [M+H]⁺ m/z=1133.46, found: 1134.01.

• Compound 21. ESIMS calcd for C₄₈H₆₈N₁₂O₁₈ [M+H]⁺ m/z=1101.49, found: 1101.83.

• Compound 22. ESIMS calcd for C₅₇H₇₈N₁₂O₁₅S [M+H]⁺ m/z=1203.55, found: 1204.02.

• Compound 23. ESIMS calcd for C₅₆H₈₃N₁₃O₁₈S [M+H]⁺ m/z=1258.58, found: 1259.00.

• Compound 24. ESIMS calcd for C₅₆H₈₃N₁₃O₁₈ [M+H]⁺ m/z=1226.61, found: 1227.09.

• Compound 25. ESIMS calcd for C₆₀H₇₀N₁₂O₁₆ [M+H]⁺ m/z=1215.51, [M+2H]²⁺ m/z=608.26; found: 1216.04, 608.71.

• Compound 26. ESIMS calcd for C₅₉H₇₆N₁₂O₁₆S [M+H]⁺ m/z=1241.52, [M+2H]²⁺ m/z=621.27; found: 1241.94, 621.63.

• Compound 27. ESIMS calcd for C₅₉H₇₆N₁₂O₁₆ [M+H]⁺ m/z=1209.56, [M+2H]²⁺ m/z=605.29; found: 1210.14, 605.69.

• Compound 28. ESIMS calcd for C₅₀H₆₄N₁₂O₁₇ [M+H]⁺ m/z=1105.46, found: 1105.80.

• Compound 29. ESIMS calcd for C₄₉H₇₀N₁₂O₁₇S [M+H]⁺ m/z=1131.48, [M+2H]²⁺ m/z=566.25; found: 1131.87, 566.81.

• Compound 30. ESIMS calcd for C₄₉H₇₀N₁₂O₁₇ [M+H]⁺ m/z=1099.51, found: 1099.86.

• Compound 31. ESIMS calcd for C₅₁H₆₆N₁₂O₁₇ [M+H]⁺ m/z=1119.48, found: 1120.10.

• Compound 32. ESIMS calcd for C₅₀H₇₂N₁₂O₁₇S [M+H]⁺ m/z=1145.50, [M+2H]²⁺ m/z=573.26; found: 1145.93, 573.68.

• Compound 33. ESIMS calcd for C₅₀H₇₂N₁₂O₁₇ [M+H]⁺ m/z=1113.52, found: 1113.87.

• Compound 34. ESIMS calcd for C₅₈H₇₂N₁₂O₁₆ [M+H]⁺m/z=1193.53, found: 1193.59.

• Compound 35. ESIMS calcd for C₅₇H₇₈N₁₂O₁₆S [M+H]⁺ m/z=1219.55, found: 1220.01.

• Compound 36. ESIMS calcd for C₅₇H₇₈N₁₂O₁₆ [M+H]⁺ m/z=1187.57, found: 1188.02.

• Compound 37. ESIMS calcd for C₄₄H₆₅N₁₅O₁₂S₂ [M+H]⁺ m/z=1060.45, found: 1060.44.

• Compound 38. ESIMS calcd for C₉₇H₁₃₃N₂₇O₂₈S [M+2H]²⁺ m/z=1078.99, [M+3H]³⁺ m/z=719.66, found: 1079.34, 719.96.

• Compound 39. ESIMS calcd for C₉₇H₁₃₃N₂₇O₂₈ [M+2H]²⁺ m/z=1063.00, [M+3H]³⁺ m/z=709.00, found: 1063.35, 709.23.

• Compound 40. ESIMS calcd for C₉₉H₁₃₄N₂₈O₂₇S[M+2H]²⁺ m/z=1090.50, [M+3H]³⁺ m/z=727.33, found: 1090.91, 727.57.

• Compound 41. ESIMS calcd for C₉₉H₁₃₄N₂₈O₂₇ [M+2H]²⁺ m/z=1074.50, [M+3H]³⁺ m/z=716.68, found: 1074.79, 716.97.

• Compound 42. ESIMS calcd for C₈₈H₁₂₄N₂₆O₂₆S [M+2H]²⁺ m/z=997.46, [M+3H]³⁺ m/z=665.31, found: 997.82, 665.60.

• Compound 43. ESIMS calcd for C₈₈H₁₂₄N₂₆O₂₆ [M+2H]²⁺ m/z=981.47, [M+3H]³⁺ m/z=654.65, found: 981.72, 654.97.

• Compound 44. ESIMS calcd for C₉₈H₁₄₆N₁₆O₄₆S₂ [M+2H] ⁺ m/z=1174.46, found: 1175.42.

• Compound 45. ESIMS calcd for C₁₃₃H₁₉₉N₃₁O₅₇S [M+2H]²⁺ m/z=1588.18, [M+3H]³⁺ m/z=1059.12, found: 1588.27, 1059.25.

• Compound 46. ESIMS calcd for C1₃₃H₁₉₉N₃₁O₅₇ [M+2H]²⁺ m/z=1572.19, [M+3H]³⁺ m/z=1048.46, found: 1572.66, 1048.74.