One-pot iterative native chemical ligation–desulfurisation chemistry leveraging a coumarin-based photolabile protecting group for cysteine

Introduction

The development of chemoselective peptide ligation methodologies has revolutionised chemical protein synthesis, providing rapid access to structurally complex polypeptide and protein molecules.^1–4 Native chemical ligation (NCL),^1,5 which enables fusion of peptide fragments by an amide bond-forming reaction between a peptide bearing an N-terminal cysteine (Cys) residue and a peptide with a C-terminal thioester functionality, remains the gold standard method for protein synthesis (Fig. 1a).⁶ Although this method initially relied on the availability of a suitably placed Cys (1.8% abundance) at the ligation junction, subsequent introduction of desulfurisation chemistry, which converts Cys to alanine (Ala) at the ligation junction,^7–9 along with the development of thiolated analogues of other amino acids, has greatly expanded the scope of NCL.^4,10–19

Fig. 1.

(a) Mechanism of an NCL reaction (b) Synthesis of Fmoc-Cys(DEAMC)-OH building block 1: (a) 2 (prepared as per published protocol³¹), Fmoc-Cys-OH, DIPEA, THF, rt, 16 h, 65%.

With these impressive advances, large proteins can now be assembled successfully by sequential ligations of multiple fragments. A popular ligation approach involves the ligation of three or more fragments in a bi-directional manner that benefits from the use of bifunctional fragments that feature a thioester at the C-terminus and an N-terminal Cys residue. For these approaches, masking of the Cys residue or thioester during an initial NCL reaction is essential to prevent undesired side reactions, including peptide cyclisation or self-ligation. Several strategies have been developed for performing multi-fragment ligation–desulfurisation protocols in a ‘one-pot’ manner,²⁰ including the use of Cys protecting groups; thiazolidine (Thz),^21–24 acetamidomethyl (Acm)^25,26 and trifluoromethylacetamidomethyl (TfAcm).²⁷ However, depending on the peptide sequence and modifications, the conditions required for deprotection may be incompatible or necessitate intermediate purification, significantly hampering the speed, efficiency and overall yield of chemical protein synthesis.²⁸

Photolabile protecting groups have emerged as attractive alternatives as they provide the ability to perform iterative ligations in a one-pot manner due to the reagent-free nature of the deprotection steps. However, notable examples, which include nitroveratryl and quinoline-based protecting groups,^29,30 require harsh UV irradiation for deprotection, potentially leading to unwanted side reactions on peptide and protein molecules. Capitalising on our previous work on the use of the 7-diethylamino-3-methyl coumarin (DEAMC) protecting group on selenocysteine (Sec) to facilitate one-pot diselenide-selenoester ligation chemistry,³¹ herein we validate the DEAMC moiety as a suitable protecting group for Cys, which can be cleanly removed under mild conditions by photoirradiation using 450-nm visible light. Importantly, this method enables chemoselective desulfurisation of unprotected Cys residues in the presence of DEAMC-protected Cys residues, thus streamlining the one-pot NCL–desulfurisation protocol.

Results and discussion

The proposed DEAMC-protected cysteine was synthesised in a single step by alkylating Fmoc-Cys-OH with our previously reported DEAMC-Br reagent 2 in the presence of N,N-diisopropylethylamine (DIPEA), furnishing the target building block 1 in 65% yield (Fig. 1b).

In order to assess the photodeprotection efficiency of DEAMC along with its stability during ligation, a model peptide H₂N-C(DEAMC)SPGYS-NH₂ (3) was synthesised by Fmoc-strategy solid-phase peptide synthesis (Fmoc SPPS) and purified by reverse-phase HPLC (see ESI for synthetic details). The purified peptide 3 was dissolved in a denaturing buffer comprising 6 M of guanidine·HCl and irradiated with blue light (λ = 450 nm, LED) to test the deprotection of the DEAMC group from the side chain of Cys. Pleasingly, clean and quantitative deprotection was observed within 30 s, affording the corresponding cysteinyl peptide (4) with an unprotected N-terminal cysteine residue (see ESI for details). Next, model peptide 3 was dissolved in the same denaturing buffer, this time with the inclusion of tris(2-carboxyethyl)phosphine (TCEP) and 4-mercaptophenylacetic acid (MPAA), additives typically used for NCL, and incubated at 37°C for 16 h. Importantly, the DEAMC remained intact under these conditions, demonstrating its compatibility under NCL conditions (see the ‘Stability and deprotection of C(DEAMC)SPGYS 3’ section in the Supplementary material for details). These control experiments of model peptide 3 provided confidence in the proposed use of DEAMC to mask N-terminal Cys-containing peptide thioesters for use in iterative ligation-based assembly of large peptides and proteins.

Having optimised conditions for the introduction and deprotection of a DEAMC-protected Cys residue, attention was turned to evaluating its utility in protein assembly using an iterative NCL–desulfurisation strategy. Towards this end, a 60-residue unglycosylated analogue (5) comprising three copies of the extracellular variable number tandem repeat (VNTR) of transmembrane glycoprotein mucin-1 (MUC1) was selected as a target peptide (Fig. 2a). Specifically, a A21C mutant of this 60-residue peptide was selected to showcase selective desulfurisation chemistry of a DEAMC-protected Cys at the ligation junction before carrying out subsequent photodeprotection and ligation steps. Based on the retrosynthetic analysis, this strategy necessitated the synthesis of three differentially functionalised peptides: fragment 6, a peptide bearing an N-terminal cysteine; fragment 7, a bifunctional peptide featuring an N-terminal DEAMC-protected cysteine and a C-terminal alkyl thioester; and fragment 8, a peptide with a C-terminal alkyl thioester. All three peptides were synthesised using Fmoc-SPPS on a hyper acid-labile 2-chlorotrityl chloride (2-CTC) resin. In the case of fragments 7 and 8, the fully elongated peptide was cleaved from resin under mild acidic conditions and subjected to C-terminal thioesterification using 7-azabenzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate (PyAOP), DIPEA and the alkyl thiol 3-mercaptoethylpropionate (see Fig. 2b for the synthesis of bifunctional peptide fragment 7, including the coupling of the N-terminal DEAMC-protected Cys building block 1).³² Finally, global acidolytic cleavage and purification by reverse-phase HPLC furnished the suitably functionalised fragments ready for assembly by ligation.

Fig. 2.

(a) Sequence of target A21C MUC1 60mer 5 (ligation junctions indicated by dotted lines). (b) General Fmoc-SPPS (see ESI), Coupling: Fmoc-Cys(DEAMC)-OH 1 (1.2 eq.), HOAt (2.4 eq.), DIC (2.4 eq.), DMF, rt, 3 h; Boc-protection: (a) General Fmoc deprotection (see ESI), (b) Boc-anhydride (5 eq.), DIPEA (15 eq.), DMF, rt, 1 h; Resin Cleavage: 1,1,1,3,3,3-hexafluoroisopropanol (HFIP)/DCM (30% v/v, 5 mL), rt, 1 h; Thioesterification: ethyl 3-mercaptopropionate (30 eq.), PyAOP (5 eq.), DIPEA (5 eq.), −30°C, 3 h; Acidolytic Cleavage: TFA, iPr₃SiH, H₂O (18 : 1 : 1 v/v/v), rt, 2 h. Analytical HPLC and inlay of ESI-MS of purified MUC1 fragment 7 (c) Synthesis of MUC1 60mer 5 by one-pot iterative NCL reactions. (d) HPLC trace of crude reaction mixture after one-pot four-step ligation assembly. (e) Analytical HPLC and inlay of ESI-MS of purified MUC1 60mer 5.

With peptides 6–8 in hand, the synthesis of the A21C MUC1 polypeptide (5) was initiated by reacting equimolar amounts of cysteinyl fragment 6 and bifunctional fragment 7 in the ligation buffer (6 M of guanidine·HCl, 1 M of HEPES, 200 mM of TCEP, 20 mM of MPAA, pH 7.0) at 5-mM peptide concentration (see Fig. 2c). Upon completion of NCL in 3 h (based on UPLC-MS analysis), the reaction mixture was acidified to pH 3.0 with 5 M of HCl, and MPAA was extracted with ethyl acetate.³³ This step is crucial since MPAA acts as a radical quencher, blocking the desulfurisation pathway. Although photodesulfurisation chemistry has been previously reported,^9,34 it was not pursued here due to concerns regarding the potential instability of the DEAMC group under UV irradiation. Instead, we investigated whether this desulfurisation step could be performed using 2,2′-azobis[2-(2-imidazolin-2-yl)propane] dihydrochloride (VA-044) as a radical initiator. To attempt this, the pH of the reaction mixture was re-adjusted to 7.0 using 5 M of NaOH and it was treated with VA-044 (15 mM) at 37°C for 16 h. Pleasingly, complete desulfurisation of Cys at the ligation junction was observed, whereas the DEAMC-protected Cys residue at the N-terminus remained unaffected. The reaction mixture was centrifuged to remove precipitated TCEP oxide and sulfide byproducts. The pellet was washed and re-suspended in an equivalent volume of ligation buffer (6 M of guanidine·HCl, 1 M of HEPES, 200 mM of TCEP, 200 mM of MPAA, pH 7.0) and recentrifuged. Supernatants were combined and irradiated with 450-nm light for 15 min, leading to the quantitative deprotection of the DEAMC group to afford 9 (now bearing a free N-terminal cysteine) as the major product. Without intermediate purification, 9 was subjected to another NCL reaction with thioester fragment 8 (4.0 eq.), which proceeded to completion within 3 h. After reverse-phase HPLC purification and lyophilisation, the target 60-residue C21A MUC1 polypeptide 5 was isolated in 34% yield over four steps (76% per step), underscoring the utility of this one-pot multi-step reaction manifold with the DEAMC-protected Cys as a key linchpin.

Conclusions

The DEAMC photolabile protecting group was successfully installed on the side chain of Cys, enabling selective desulfurisation of unprotected Cys residues, followed by photodeprotection by LED irradiation at λ = 450 nm. This mild, reagent-free step is fully compatible with an iterative NCL assembly strategy, allowing efficient one-pot synthesis of larger polypeptides and proteins without intermediary purification steps. The utility of this methodology was highlighted through the successful synthesis of a 60-residue A21C MUC1 polypeptide. Future work will involve broader application of DEAMC-protected cysteine in the ligation-based assembly of challenging protein targets, particularly those requiring selective desulfurisation of Cys/thiolated amino acid at the ligation junction in the presence of native cysteine residues.

Experimental section

Supplementary material

Supplementary material is available online.