Challenges in the use of sortase and other peptide ligases for site-specific protein modification

Abstract

Site-specific protein modification is a widely-used biochemical tool. However, there are many challenges associated with the development of protein modification techniques, in particular, achieving site-specificity, reaction efficiency and versatility. The engineering of peptide ligases and their substrates has been used to address these challenges. This review will focus on sortase, peptidyl asparaginyl ligases (PALs) and variants of subtilisin; detailing how their inherent specificity has been utilised for site-specific protein modification. The review will explore how the engineering of these enzymes and substrates has led to increased reaction efficiency mainly due to enhanced catalytic activity and reduction of reversibility. It will also describe how engineering peptide ligases to broaden their substrate scope is opening up new opportunities to expand the biochemical toolkit, particularly through the development of techniques to conjugate multiple substrates site-specifically onto a protein using orthogonal peptide ligases.

Scheme 1. Summary of strategies used to optimise reactions of peptide ligase to enable complex protein modification reactions including both substrate and protein engineering described in this review.

Scheme 2. Catalytic mechanism of (A) peptide ligases, (B) Sortase A on the surface of Gram-positive bacteria, (D) peptide asparaginyl ligases (PALs) to produce cyclic peptides in plants, (F) subtilisin in Bacillus amyloliquefaciens. Substrate specificity of (C) Sa Sortase A, (E) PALs and (G) subtiligase.

Fig. 1. Exemplar yeast and phage constructs used for directed evolution of sortases. In both cases, sortases are encoded by phage or yeast cells and the activity of the encoded sortase is probed by addition of a biotinylated sortase substrate (e.g. Biotinyl-LPETGG) which enables isolation of phage of yeast encoding active sortases. (A) Aga1p–Aga2p strategy used by Chen et al. i to increase sortase activity. (B) M13 Phage strategy used by Piotukh et al to identify sortases with altered specificity.

Fig. 2. Location of mutations in sortase variants mapped onto the crystal structure of Sortase A. (A) Apo-crystal structure of WT sortase A determined by Zong et al. (1t2p) (B) structure of a LPETG peptide bound to Sortase A. (1t2w) (C) location of mutations observed in eSrtA(2A-9) shown in purple. Active site cysteine yellow. (D) Location of mutations observed in eSrtA(4S-9) shown in red. (E) Location of mutations observed in SrtA(5M) (orange) and SrtA(7M) (orange and blue). (F) Location of mutations found in SrtAβ (dark green).

Scheme 3. Mechanisms of competing transpeptidation and hydrolysis reaction. In the absence of an acyl acceptor substrate, the by-product peptide reversibly forms the starting material competitively inhibiting the hydrolysis reaction.

Scheme 4. Substrate engineering strategies employed to enhance product yields with SrtA, Butelase and OaAEP1. (A) Formation of a β-hairpin prevents binding of SrtA to the reaction product. (B) Hydroxyacetamide products are not substrates for the reverse reaction. (C) Cyclisation of the diglycyl motif with loss of serine generates a diketopiperazine. (D) A GlyGlyHis motif is a ligand for Ni₂₊ in solution which sequesters the product peptide as an inactive complex. (E) β-Thioacetamide products are not substrates for the reverse reaction. (F) Enzyme selectivity is exploited: while OaAEP1 can act on a NGL sequence to form an NGV product, the NGV sequence is a poor substrate. (G) The product peptide with an N-terminal cysteine is sequestered by formation of a complex.

Scheme 5. Examples of application of (A) strategy for the dual labelling of both termini of the same protein using SpSrtA and SaSrtA. Adapted from Antos et al. (B) Strategy for the triple labelling of distinct capsid proteins in a M13 bacteriophage particle. Adapted from Hess et al.

Scheme 6. Recent examples of expansion of the substrates for peptide ligases to enable segment assembly and the generation of complex assemblies such as triubiquitins. (A) The use of tertbutylthiol cysteine disulfides as leucine isosteres enables the generation of sortase substrates which can then be deactivated by reduction and desulfurisation. (B) Incorporation of azidoacetyl glycyl lysine into proteins enables subsequent reduction using 2-diphenylphosphinobenzoic acid (2DPBA) and labelling using sortases. (C) Extension of this approach to applications with multiple orthogonal sortases enables the synthesis of specific triubiquitin and diubiquitylated SUMO constructs using both internal and N-terminal labelling.

Scheme 7. Examples of the combined application of SrtA and butelase-1 to enable double labelling of proteins and formation of protein fusions. Orthogonal labelling combining SaSrtA and butelase-1. (A) dual labelling of ubiquitin via a three-step tandem ligation with native chemical ligation. (B) One-pot conjugation of two nanobodies via their C-termini to produce C-to-C protein conjugates. This was done with a PEG linker and oligonucleotide linker. (C) One-pot conjugation at the C-terminus of the light chain and heavy chain of an antibody.

Scheme 8. Examples of the use of P (A) orthogonal labelling combining butelase-1 and VyPAL2 to prepare a cycloprotein-drug conjugate. (B) pH controlled orthogonal ligation with VyPAL2 to produce a fluorescein-drug-labelled affibody. (C) Substrate controlled orthogonal labelling of an anti-UBC6e nanobody via OaAEP1 (D) Use of C-terminal 2-aminoethylamides to enable C–C tail-to-tail protein dimerisation using OaAEP1. General structure of peptide substrates for homodimerisation and strategy to enable heterodimerisation via use of C-terminal protein thioesters.

Holly Morgan

Bruce Turnbull

Michael Webb