Chemicals. All chemicals used were supplied by Sigma-Aldrich (USA) or New England Biolabs (USA) if not specified explicitly.
Gene construction. The C1 domain of Cpe0147 from Clostridium perfringens (GenBank accession no. EDT23863.1), the Staphylococcus epidermidis SdrG N2 and N3 domain genes, as well as GB1 were synthesized codon-optimized for expression in Escherichia Coli as linear DNA fragments with suitable overhangs (GenScript, China). Genes were cloned into pET22b or pQE80L Vectors. The T11A mutant and circular permutation were introduced by blunt end ligation cloning using T4 Ligase. Final open reading frames of all constructs were checked by DNA sequencing (GenScript, China). The complete sequences of all protein constructs used are listed in Supplementary Information.
Protein expression and purification. The genes encoding the proteins used in this article were synthesized and codon-optimized for expression in Escherichia coli (BL 21) cells. Precultures of 5 mL in LB medium (BD Difco, USA) containing 100 µ g/mL ampicillin (Macklin, China), grown overnight at 37 °C, were inoculated in 500 ml LB medium containing 100 µg/mL ampicillin and grown for 3 h at 37 °C and then induced with 0.4 mM of IPTG (Aladdin, China) overnight at 25 °C. Bacteria were harvested by centrifugation at 10000 g, and pellets were stored frozen at -80 °C until purification.
All purification steps were performed at 4 °C to 8 °C. The bacterial pellet was resuspended in PBS, the cells were then lysed through mechanically sonication (ATPIO, China) followed by centrifugation at 11000 g for 1 h. The His6-tagged proteins were purified by affinity chromatography using a Talon column (Sigma-Aldrich, USA). The supernatant was washed extensively and then eluted in the same buffer supplemented with 200 mM imidazole. Protein containing fractions were concentrated in the centrifugal filters, exchanged into measurement buffer by desalting columns (Sigma-Aldrich, USA), and frozen in aliquots in liquid nitrogen to be stored at -80 °C until used in experiments. Protein concentrations were measured by spectrophotometry at 280 nm with typical final concentrations of ~ 100 mM (Thermo Nanodrop 2000, USA).
AFM sample preparation. More detailed AFM-based single molecule force spectroscopy protocol have been published previously28,39. In brief, AFM cantilevers (Bruker, MLCT, USA) and glass surfaces (SAIL BRAND, China) were modified with aminosilane.
Glass surface: Glass substrates were cut into 1 × 1 cm2 slides, soaked in a chromic mixture overnight, thoroughly washed with deionized (DI) water, ethanol and acetone successively, and then dried under a steam of nitrogen to produce surfaces with exposed hydroxyl groups. These substrates were immersed in an anhydrous toluene solution containing 1% (v/v) APTES (Merck, USA) at room temperature (R.T.) for 1 h for amination. Then, they were washed with toluene and ethanol, dried under a nitrogen flow. Finally, surfaces were baked at 90 °C for 30 min. Glass substrates were stored in a desiccator under Argon and typically used within half month.
Cantilevers: Silicon nitride (Si3N4) cantilevers (MLCT-D, Bruker) were first cleaned with Milli-Q water, and then placed in a chromic mixture (chromic acid) at 80 °C for 30 min. After that, the cantilevers were washed with water, then ethanol, and dried under a steam of nitrogen. Then the hydroxylated cantilevers were immersed in 1% (v/v) APTES in toluene for 1 h. After that, they were rinsed with toluene, then ethanol, dried under a nitrogen stream, and incubated at 80 °C for 45 min. Finally, they were stored overnight under Argon and used in the following steps the next day.
Both glass substrates and cantilevers were immersed in DMSO containing 0.2 mM Mal-PEG-NHS (MW: 5000 Da, Nanocs, USA) for 1 h. After being washed with DMSO, ethanol and dried under a nitrogen stream, the resulting Mal-coated glass substrates and cantilevers were kept dry at -20 °C for use in the following protein modification steps in single-molecule experiments.
AFM-based single molecule force spectroscopy. The force spectroscopy experiments were carried out using a commercial JPK ForceRobot 300 AFM system (JPK Instruments AG, Germany). Experiments were conducted at room temperature (22 °C) and performed in 10 mM PBS buffer with or without 10 mM TCEP if needed. Soft silicon nitride MLCT-D cantilevers of typical spring constant of 30 ~ 40 pN nm− 1 were used for all experiments and calibrated using the thermal tune method after allowing the cantilever to equilibrate in solution for at least 30 min. Cantilevers were briefly and gently (~ 300 pN) brought in contact with the functionalized surface and held at the surface for 0.5 s, then retracted at constant velocity of 1.6 µm s− 1. The force-extension curves were recorded using JPK data processing software and were further analyzed by a custom-written procedure in Igor 6.12 (Wavemetric, Inc).
Molecular Dynamics Simulations. The molecular dynamics simulations were conducted by the GROMACS software52 with the ff14SB force field53 and TIP3P water54. To model the ester bond, we introduced a covalent bond between the Thr-11 (Oγ1) and Gln-141 (Cδ). The extra atoms (Hγ1 of Thr-11; Nε2, Hε21, and Hε22 of Gln-141) were removed, with their partial charges being integrated into the nearby heavy atoms. The atomic coordinates of the C1WT were taken from the Protein Data Bank (entry 4MKM)16. The five N-terminal residues lacking structural information were not included. In constructing the C1CP, a ELP linker with the length of 20 amino acids was added between the two termini of the C1, and the residues Leu-125 and Asp-126 were used as the new termini. The three-dimensional structure of the linker was modelled by the ModLoop55,56.
The C1WT and C1CP were solvated in the rectangular water boxes with the dimensions of ~ 241 Å×84 Å×85 Å (with 48891 water molecules) and 1017 Å×74 Å×74 Å (with 159145 water molecules), respectively. Sodium ions were added to neutralize the systems. The LINCS algorithm was used to restrain the covalent bond involving hydrogen atoms57. After a 50000-step minimization using the steepest descent method, each system was equilibrated for 0.1 ns in the NVT ensemble and another 0.1 ns in the NPT ensemble. The temperature and pressure were controlled at 298.0 K and 1.0 atm, respectively. The heavy atoms of the proteins were restrained to their original positions by a harmonic potential during the minimization and equilibrium steps. Starting from the equilibrated structures, we performed steered MD simulations by applying pulling force between the termini residues along the x-axis in NVT ensemble. For the C1WT, the constant pulling force were applied to the termini residues with the strength of 1500 pN and the simulations were lasted for 10 ns. For the C1CP, we firstly conducted a constant velocity pulling simulations with the pulling speed of 49 Å/ns, such that the ester bond stars to sustain pulling force. Then we performed the same constant pulling force simulations as that in the C1WT. Two independent simulations were conducted for each system. To increase the statistics, we also performed another two (one) constant pulling force simulations with the length of 35 ns for the C1WT (C1CP). The snapshots sampled during the constant pulling force simulations were used for analysis. The software PyMOL was used for the structure visualization.