A Novel Purification Procedure for Active Recombinant Human DPP4 and the Inability of DPP4 to Bind SARS-CoV-2

Abstract

Proteases catalyse irreversible posttranslational modifications that often alter a biological function of the substrate. The protease dipeptidyl peptidase 4 (DPP4) is a pharmacological target in type 2 diabetes therapy primarily because it inactivates glucagon-like protein-1. DPP4 also has roles in steatosis, insulin resistance, cancers and inflammatory and fibrotic diseases. In addition, DPP4 binds to the spike protein of the MERS virus, causing it to be the human cell surface receptor for that virus. DPP4 has been identified as a potential binding target of SARS-CoV-2 spike protein, so this question requires experimental investigation. Understanding protein structure and function requires reliable protocols for production and purification. We developed such strategies for baculovirus generated soluble recombinant human DPP4 (residues 29-766) produced in insect cells. Purification used differential ammonium sulphate precipitation, hydrophobic interaction chromatography, dye affinity chromatography in series with immobilised metal affinity chromatography, and ion-exchange chromatography. The binding affinities of DPP4 to the SARS-CoV-2 full-length spike protein and its receptor-binding domain (RBD) were measured using surface plasmon resonance and ELISA. This optimised DPP4 purification procedure yielded 1 to 1.8 mg of pure fully active soluble DPP4 protein per litre of insect cell culture with specific activity >30 U/mg, indicative of high purity. No specific binding between DPP4 and CoV-2 spike protein was detected by surface plasmon resonance or ELISA. In summary, a procedure for high purity high yield soluble human DPP4 was achieved and used to show that, unlike MERS, SARS-CoV-2 does not bind human DPP4.

Keywords: Covid-19; DPP4; protease; recombinant protein.

Figure 1

Overview of DPP4 purification workflow.

Figure 2

Elution profiles of DPP4 chromatography. (A) Chromatogram from Phenyl Sepharose that was equilibrated with 12% ammonium sulphate (AS) in 10 mM Tris-HCl pH 7.6 and eluted with 0% AS in 10 mM Tris-HCl pH 7.6 buffer. Inset: Fibroblast activation protein (FAP) activity. (B) Chromatogram from Nickel Sepharose, which was equilibrated with 20 mM imidazole in 200 mM NaCl, 10 mM Tris-HCl, pH 7.6 and eluted with an increasing concentration gradient of imidazole at 30 mM (a), 100 mM (b), 500 mM (c), 1000 mM (d) in 10 mM Tris-HCl pH 7.6 buffer. Inset: FAP activity. (C) Chromatogram from DEAE Sepharose that was equilibrated with 10 mM Tris-HCl pH 7.6 and eluted with 200 mM NaCl in 10 mM Tris-HCl pH 7.6 buffer. Inset: sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS–PAGE; 4–12% Bis-Tris gel) of the resulting purified soluble DPP4, stained with Sypro ruby Protein was measured by optical density at 280 nm in these chromatograms.

Figure 3

ELISA and virus neutralisation assays. (A) Purified DPP4 protein was used to capture MERS-CoV spike clamp, SARS-CoV-2 spike clamp or a control clamped protein. Clamp-stabilised proteins were detected using a clamp-specific mAb HIV1281. Data shown represent means of duplicate values. (B) DPP4, ACE2, a MERS-specific mAb m336 or a non-specific mAb C05 were used at 20 µg/mL and incubated with MERS-CoV pseudovirus. Percent MERS-CoV inhibition is the percentage reduction in luciferase signal (RLU) compared to virus-only control. N.D. indicates no inhibition detected. Data shown are the mean of duplicate values with error bars representing SD.

Figure 4

Surface plasmon resonance assays. Purified soluble human ACE2 (A) and DPP4 (B) were exposed to CM5 chips that had been coated with SARS-CoV-2 RBD or spike protein, or were not coated. Experimental data are shown in red. Calculated data fit using a 1:1 binding model are shown in black. Ligands were injected at increasing concentrations of (A) ACE2 at 0.50 nM, 2.5 nM, 12 nM, 62 nM and 310 nM and (B) DPP4 at 1.6 nM, 8.0 nM, 40 nM, 200 nM and 1000 nM.

Figure 5

Protein structures. (A) DPP4 monomer (PDB ID 1W1I) [46]. (B) Adenosine Deaminase (ADA) (PDB ID 1W1I) [46]. (C) MERS-CoV spike RBD (PDB ID 4L72) [45], and (D,E) SARS-CoV-2 spike RBD (PDB ID 6M0J) [48] shown with space-filling surfaces with similar orientation and scale. The observed and predicted interfacial binding sites are highlighted in green, with the charged residues highlighted in red (negative) and blue (positive), and the hydrophobic residues are highlighted in orange. In particular, compared to the other molecules SARS-CoV-2 spike RBD has an extensive hydrophobic surface, and fewer charged residues at the predicted DPP4 binding site that is in a similar location to MERS-CoV spike RBD (D) [28], but has more charges surfaces on the opposite side of SARS-CoV-2 spike RBD, which is a potential alternative binding site (E) [28]. The N-glycosylated residues in DPP4 are highlighted in yellow, and are sufficiently distant from the binding interface to expect that they do not bind to ADA [20].