Abstract
Self-labelling protein tags (SLPs) are resourceful tools that revolutionized sensor imaging, having the versatile ability of being genetically fused with any protein of interest and undergoing activation with alternative probes specifically designed for each variant (namely, SNAP-tag, CLIP-tag and Halo-tag). Commercially available SLPs are highly useful in studying molecular aspects of mesophilic organisms, while they fail in characterizing model organisms that thrive in harsh conditions. By applying an integrated computational and structural approach, we designed a engineered variant of the alkylguanine-DNA-alkyl-transferase (OGT) from the hyper-thermophilic archaeon Saccharolobus solfataricus (SsOGT), with no DNA-binding activity, able to covalently react with O6 -benzyl-cytosine (BC-) derivatives, obtaining the first thermostable CLIP-tag, named SsOGT-MC8 . The presented construct is able to recognize and to covalently bind BC- substrates with a marked specificity, displaying a very low activity on orthogonal benzyl-guanine (BG-) substrate and showing a remarkable thermal stability that broadens the applicability of SLPs. The rational mutagenesis that, starting from SsOGT, led to the production of SsOGT-MC8 was first evaluated by structural predictions to precisely design the chimeric construct, by mutating specific residues involved in protein stability and substrate recognition. The final construct was further validated by biochemical characterization and X-ray crystallography, allowing us to present here the first structural model of a CLIP-tag establishing the molecular determinants of its activity, as well as proposing a general approach for the rational engineering of any O6 -alkylguanine-DNA-alkyl-transferase turning it into a SNAP- and a CLIP-tag variant.