AsiaChem | Chemistry in Japan | December 2021 Volume 2 Issue 1

www.asiachem.news December 2021 | 15 85% adduct in 3 hrs with a rate constant of 1.1 M-1S-1. Interestingly, it was found, unlike the reports of Heinis and Winter,12 that DCO modification did not result in losing phage infectivity and more than 80% of phage remained viable after modification. This suggests that crosslinking of phage coat protein is negligible with DCO. Amide bond formation In one of the first reports, Robert et al. reported a general route for post-translational cyclization of mRNA display libraries by treating translated peptide with disuccinimidyl glutarate (DSG) at pH 8.25 DSG reacted near-quantitively with N-terminal amine and an internal Lys ε-amino group crosslinked via two amide bonds. The same group then demonstrated mRNA display of DSG-linked library against Gαi1, successfully discovering a strong cyclic peptide binder with Kd = 2.1 nM. 26 Disufide-rich loop formation Disulfide bond formation was one of the first approaches developed to cyclize linear peptides displayed on phage but due to the instability of disulfide bond in reducing cellular environment, this approach finds little practical value for in vivo applications. However, plant-based cyclotides are a unique class of peptides having multiple loops in the form of cysteine knots. Their remarkable thermal and proteolytic stability and a wide range of biological activities make them ideal macrocycles to be screened as ligands for target proteins. There are several reports of selection of cyclotides with novel function using in vitro displays.27-30 As a recent demonstration, Wenyu et al. reported mRNA-display of a cyclotide library derived from Momordica cochinchinensis trypsin inhibitor-II (MCoTI-II), in which two loops, 1 and 5, were randomized. The selection campaign against human Factor XIIa (hFXIIa) successfully yielded an extraordinary potent and selective variant, referred to as MCoFx1, giving Ki of 0.37 nM to hFXIIa that is greater than three orders of magnitude selective over trypsin and other related proteases.31 Cyclization using genetic code reprogramming Genetic code reprogramming is a powerful technique which enables incorporation of nonproteinogenic amino acids in translated polypeptides via codon reassignment32 or expansion.33, 34 The technique has evolved and matured over the years (for recent reviews see these references35, 36) in which task of reprogramming is achieved through a combination of an Escherichia coli reconstituted cell-free translation system and pre-aminoacylated tRNA with various nonproteinogenic amino acids facilitated by flexizymes. This system, referred to as FIT (Flexible In-vitro Translation), enables for devising many unique methods for macrocyclization of peptides discussed in the following sections. Thioether Bond Formation Thioether bond formation by nucleophilic substitution Unlike the aforementioned strategy of adding an external organic moiety with multiple halogens, this strategy results in the formation of one thioether bond per cycle. The halo part is incorporated at the initiator position or at a suitable side chain through genetic code reprogramming.37 An intramolecular substitution reaction by a downstream Cys thiol results in the formation of a physiologically stable thioether linkage. Suga group has explored, evolved and exploited this technique thoroughly, resulting in a number of interesting macrocyclic libraries and successful selections against various targets (for recent representative examples see references37-45). In 2008, Goto et al. have used a methioninedepleted FIT system where the initiation codon AUG becomes vacant, and engineered the initiation event. To this sytem is added an aminoacyltRNAfMet CAU charged with N-chloroacetylated amino acid, such as tryptophan (ClAc-Trp) or tyrosine (ClAc-Tyr), prepared by a flexizyme (eFx).37 The ClAc-Trp-tRNAfMet CAU was set as an initiator, for example, for the peptide expression, ribosome elongates amino acids starting from the ClAc-Trp according to mRNA template sequence, followed by a Cys residue at a downstream position. When the peptide synthesis is completed, the Cys thiol spontaneously reacts with the ClAc group to yield a thioether linked macrocyclic peptide (Figure 3A). It should be noted that other haloAc group, such as BrAc and IAc, yielded many byproducts originating from adducts of thiols present in the translation system, e.g. mercaptoethanol, DTT, and Cys. Thus, the ClAc group was the perfect reactivity toward the Cys thiol in peptide chain that effectively promotes the desired intramolecular reaction over undesired intermolecular reaction. This strategy has been applied to constructing mass libraries (over trillion members) of thioether macrocycles in combination with genetic code reprogramming for the incorporation of exotic amino acids46-48 including N-methyl-L-amino acids49,50, D-amino acids51-53, and β-amino acids54,55, etc. Suga group has integrated this strategy with mRNA display, referred to as RaPID (Random nonstandard Peptides Integrated Discovery) system, and enabled the ‘rapid’ discovery of various potent macrocycles56 against extracellular and intracellular proteins and has reported more than 35 successful selection outcomes with a range of low nM to pM KD values in the period of a decade.57-84 D W C C L/D W HS N H Cl O Pu mRNA•cDNA C Pu mRNA•cDNA N H S O L/D W HS N H Cl O C C C HS SH C C C C HS D W N H S O S S S Br Br Br A B C fM P D V C Cab F W K Y NH Cl O HS T P D V C F W K Y C S S E Urotensin-II fM P D V C Cab F W K Y NH O S Urotensin-II analog D Y HS O Br C D C GryA Y D C O S Figure 3. Ribosomal synthesis of macrocycles closed by a thioether bond via nucleophilic substitution. (A) N-terminal ClAc-Trp installed by the genetic code reprogramming reacts with a downstream Cys. (B) Tricyclic peptide synthesis of the intramolecular N-terminal ClAc with the second downstream Cys in the combination with TBMB that crosslinks three remaining Cys residues. (C) The ClAc group on the sidechain of Cab installed by the genetic code reprogramming reacts with a downstream Cys. The sequence represents a sequence of human urotensin II. (D) Macrocyclization using thioether bond formation by intramolecular reaction between non-proteinogenic amino acid O2beY and Cys inside living bacterial cells via intein-based protein splicing.

RkJQdWJsaXNoZXIy NDU2MA==