Designed for life: biocompatible de novo designed proteins and components

Abstract

A principal goal of synthetic biology is the de novo design or redesign of biomolecular components. In addition to revealing fundamentally important information regarding natural biomolecular engineering and biochemistry, functional building blocks will ultimately be provided for applications including the manufacture of valuable products and therapeutics. To fully realize this ambitious goal, the designed components must be biocompatible, working in concert with natural biochemical processes and pathways, while not adversely affecting cellular function. For example, de novo protein design has provided us with a wide repertoire of structures and functions, including those that can be assembled and function in vivo Here we discuss such biocompatible designs, as well as others that have the potential to become biocompatible, including non-protein molecules, and routes to achieving full biological integration.

Keywords: biocompatibility; de novo protein design; synthetic biology.

Figure 1.

The diversity of de novo designed protein structures. (a) Pizza variant, nvPizza2-S16H58, which coordinates a CdCl₂ nanocrystal [26]. PDB: 5CHB. (b) De novo designed reaction centre with heme B, synthetic Zn porphyrin and Zn(II) cations [27]. PDB: 5VJS. (c) Catalytic helical barrel, CC-Hept-I18C-L22H-I25E. Catalytic triad residues are shown [28]. PDB: 5EZC. (d) DFsc-Zn(II)₂ used by Ulas et al. [29] for semiquinone radical stabilization. PDB: 2LFD. (e) Designed beta solenoid proteins, SynRFR24.1 (red, PDB: 4YC5) and SynRFR24.t1428 (blue, PDB: 5DNS) [30]. (f) sTIM-11 [31]. PDB: 5BVL.

Figure 2.

A selection of catalytic de novo proteins. (a) Representation of the structure of a de novo protein which performs carbonic anhydrase activity. The solution nuclear magnetic resonance structure of the α3D scaffold (PDB: 2A3D [53]), was modified to bind zinc (grey) and hydrate CO₂ [54]. (b) Molecular dynamics simulation model of C45, which can catalyse the oxidation of a range of small molecules, including 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) [12]. (c) The de novo protein Dft2 was modified to bind manganese and perform catalase activity [55,56]. The crystal structure shown is that of the variant, P0 (PDB: 5C39); variants with two and three manganese-binding sites exhibit higher activity. (d) A 4-helix bundle library protein, Syn-F4, which performs enantioselective hydrolysis of ferric enterobactin (FeEnt) [57]. As no structure is available of this protein, the structure shown is a representative 4-helix bindle from the Hecht lab (PDB: 2JUA). (e) Crystal structure of the heptameric coiled-coil CC-Hept-I18C-L22H-I25E with hydrolase activity towards p-nitrophenyl acetate (pNPA). Catalytic triad residues are shown. (PDB: 5EZC) [28].

Figure 3.

Crystal structures of de novo inhibitors binding to their targets. Crystal structures of inhibitor complexes. (a) Inhibitor peptide αMCL1 (red) binds the human BCL2 homologue, Mcl-1 (blue), with picomolar affinity [60]. PDB: 5JSB. (b) Inhibitor peptide HB1.6928.2.3 (red), which can bind influenza haemagglutinin [59]. PDB: 5VLI.

Figure 4.

De novo transmembrane components for signalling. (a) A synthetic GPCR mimic [136]. The synthetic receptor consists of a ligand-binding pocket featuring a cationic metal complex (red), an Aib oligomer (grey) and a pair of pyrene molecules attached to a chiral diamine (purple and blue). This complex adopts one of two mirror image conformational states on complexation with a chiral ligand. The binding of a chiral ligand (magenta) to one end of an Aib oligomer propagates its conformational influence along the entire length. The signal is output by the conformationally responsive fluorophore (purple and blue). Thus, the binding of the cofactor perturbs the global conformation, which is reported by the fluorophore component. (b) A translocatable sensor [137] in which two head groups are coupled to a steroid spacer (grey). The external sensor is a protonated morpholine (red or blue), while the second head group is a neutral pyridineoxime ‘pro catalyst’ (magenta or green). When the head groups are polar (red or green), they prefer to sit in the aqueous phase; when non-polar (blue or magenta), they prefer to sit in the membrane. Binding of a zinc cofactor from within the vesicle pulls the pro-catalyst head group into the aqueous phase on the interior of the vesicle. This allows the hydrolysis of the substrate within the vesicle, generating the output signal.