Fig 1.
Structural features of CnaB domains and derived split-protein systems.
(A) Topology diagram of the CnaB fold. The β-strands are depicted as arrows and a black bar marks the locations of the isopeptide bond. The color coding indicates the design concept for CnaB-based split-protein systems. In these systems, the yellow sequence stretch is removed, resulting in a catcher (blue) and tag (red) moiety. The isopeptide bond is retained in this split-protein system and forms after association of the tag and catcher parts. (B) Structure of the CnaB domain from the S. pyogenes fibronectin binding FbaB (pdb code: 2x5p[54]) that served as template for the design of the SpyTag/SpyCatcher system (color coding as in (A)). The Asp and Lys that form the isopeptide bond are shown as sticks and colored in red and cyan, respectively.
Fig 2.
Structures of the CnaB-type domains that served as starting point for the design of the tag/catcher systems.
The systems are named according to the PDB entry they were derived from (see Table 1 for details) (A) 3phs; (B) 4oq1; (C) 3kptN; (D) 3kptC. Residues of the tag, catcher, and deleted regions are colored red, blue, and yellow, respectively (see Table 1 for details on the domain boundaries); active-site residues Asn, Lys and Glu are shown as sticks.
Table 1.
Overview of tag/catcher systems designed in the present study.
The names of the systems were derived from the PDB code of the respective crystal structure. 3kptN and 3kptC indicate that these constructs have been derived from the N- and C-terminal domain of PDB entry 3KPT, respectively.
Fig 3.
Structure-based multiple sequence alignment of the original Spy-system and the four constructs selected for the present study.
The active site residues are shown in bold. Residues of the tag, catcher, and deleted regions are marked in red, blue, and yellow, respectively. A “*” marks a strictly conserved sequence position; “:” and “.” denote decreasing degrees of sequence similarity.
Fig 4.
Conformational stability of the parent domains (violet) and tag/catcher (orange) constructs deduced from the molecular dynamics simulations.
(A) 3phs, (B) 4oq1, (C) 3kptN, and (D) 3kptC. The RMSD was calculated for the backbone of the residues forming the core of the domains (see Methods for a definition of the core residues).
Fig 5.
Evolution of van-der-Waals interaction energy of the tag in parent domains (violet) and tag/catcher (orange) constructs.
(A) 3phs, (B) 4oq1, (C) 3kptN, and (D) 3kptC.
Fig 6.
Hydrogen bonds between the tag and catcher region compared for domain (violet) and tag/catcher systems (orange).
The structure presentations show the spatial arrangement of the hydrogen bond network and the diagrams show the mean distance observed over the simulation time between the participating residues for (A) 3phs, (B) 4oq1, (C) 3kptN, and (D) 3kptC. Residues located C-terminally of the reactive Asn position did not form β-sheet hydrogen bonds and are therefore not shown.
Fig 7.
Covalent intermolecular bond formation assays for 4oq1C / 4oq1T-MBP (A) and 3kptCC / 3kptCT-MBP (B).
(A) Purified 4oq1T-MBP and 4oq1C were mixed each at 15 μM (final concentration) for 24h at 25°C with shaking at 500 rpm before boiling (10min, 95°C) and SDS-PAGE with Coomassie staining. 4oq1T-MBP and 4oq1C reacted spontaneously to form a covalent product increasing over time (lane 1: catcher input (30μM), lane 2: tag input (30μM), lane 3: 0h, lane 4: 1h, lane 5: 2h, lane 6: 3h, lane 7: 4h, lane 8: 24h). Same volume of samples were loaded. MW stands for molecular weight (kDa). (B) Bond formation assay monitoring the reaction of 3kptCT-MBP and 3kptCC. All experimental settings as described in (A).
Fig 8.
Covalent intermolecular bond formation assay between tag-MBP and respective mCherry-catcher proteins.
Purified tag-MBP and mCherry-catcher proteins were mixed each at 15 μM (final concentration) for 24h at 25°C with shaking at 500 rpm before boiling (10min, 95°C) and SDS-PAGE with Coomassie staining. (A) Time course of 4oq1T-MBP and mCherry-4oq1C reaction. (B) Time course of 3kptCT-MBP and mCherry-3kptCC reaction. Spontaneous isopeptide bond formation was detected between the protein partners used. Covalent product increased over time. Lane 1: mCherry-catcher input (30μM), lane 2: tag input (30μM), lane 3: 0h, lane 4: 1h, lane 5: 2h, lane 6: 3h, lane 7: 4h, lane 8: 24h. Same volume of samples were loaded. MW stands for molecular weight (kDa).
Fig 9.
Candidate sites for modification of 4oq1T and 3kptCT (catcher region: blue; tag region: red; linking region: yellow).
(A) 4oq1 crystal structure with residues H243-L245 of the linker sequence shown in yellow sticks, interacting residues in the catcher region within 4 Å are shown in blue sticks. (B) Enlargement showing the interacting residues in detail. Circles denote specific interactions. (C) Overlay of 11 tag peptide structures from the 4oq1 MD simulation; for the sake of clarity, only one catcher structure is shown. (D) 3kptC crystal structure with residues N502-Q504 of the linker sequence shown in yellow sticks, interacting residues in the catcher region within 4 Å are shown in blue sticks. (E) Enlargement showing the interacting residues in detail. Circles denote specific interactions. (F) Overlay of 11 tag peptide structures from the 3kptC MD simulation; for the sake of clarity, only one catcher structure is shown.
Fig 10.
Comparative analysis of different 4oq1T-MBP variants.
(A) Sequence alignment of different 4oq1T-MBP variants (I: 4oq1T wildtype (residues V246-N259), II: N-terminal L245 extension of 4oq1T, III: N-terminal H243-Q244-L245 extension of 4oq1T, IV: C-terminal R257-G258-N259 truncation of 4oq1T). (B) Comparative covalent intermolecular bond formation assay between different 4oq1T-MBP variants and mCherry-4oq1C (0h-24h). Purified 4oq1T variants and mCherry-4oq1C proteins were mixed each at 15 μM (final conc.) for 24h at 25°C with shaking at 500 rpm before boiling (10min, 95°C) and SDS-PAGE with Coomassie staining. Interaction I: mCherry-4oq1C + 4oq1T (wildtype), interaction II: mCherry-4oq1C + 4oq1T (L), interaction III: mCherry-4oq1C + 4oq1T (HQL), interaction IV: mCherry-4oq1C + 4oq1T (ΔRGN). (lane 1: mCherry-catcher input (30μM), lane 2: tag input (30μM), lane 3: 0h, lane 4: 1h, lane 5: 2h, lane 6: 3h, lane 7: 4h, lane 8: 24h). Same volume of samples were loaded. MW stands for molecular weight (kDa).
Fig 11.
Comparative analysis of different 3kptCT-MBP variants.
(A) Sequence alignment of different 3kptCT-MBP variants (I: 3kptCT wildtype (residues T505-K518), II: N-terminal N506-Q507 extension of 3kptCT, III: C-terminal GWI instead of PTK in 3kptCT). (B) Comparative covalent intermolecular bond formation assay between different 3kptCT-MBP variants and mCherry-3kptCC. Purified 3kptCT variants and mCherry-3kptCC proteins were mixed each at 15 μM (final conc.) for 24h at 25°C with shaking at 500 rpm before boiling (10min, 95°C) and SDS-PAGE with Coomassie staining. Interaction I: mCherry-3kptCC + 3kptCT (wildtype), interaction II: mCherry-3kptCC + 3kptCT (NQ), interaction III: mCherry-3kptCC + 3kptCT (GWI). (lane 1: mCherry-catcher input (30μM), lane 2: tag input (30μM), lane 3: 0h, lane 4: 1h, lane 5: 2h, lane 6: 3h, lane 7: 4h, lane 8: 24h). Same volume of samples were loaded. MW stands for molecular weight (kDa).
Fig 12.
Bar diagrams summarizing the quantified reconstitution rates for (A) 4oq1 variants and (B) 3kptC variants.
For each time point, the reconstitution of the wiltype-tag and different tag-variants is compared (please refer to Fig 11 for the exact sequence of the different tag variants). Each value represents the average over 3–4 independent experiments. o.n. = over night (24 h). Purified mCherry-catcher (15μM final conc.) and respective purified tag-MBP (15μM final conc.) were mixed at 25°C with shaking (500rpm) and a time course of covalent bond formation was performed (0h, 1h, 2h, 3h, 4h, o.n. = over night). Samples were boiled (10min, 95°C) prior to gel loading. Four independent assays per catcher-tag pair were performed (triplicate for mCherry-3kptCC-3kptCT(NQ)) and analyzed via densitometry (ImageJ). Based on the band intensities in the gel mean values and standard deviation were calculated.