N-glycosylation and ubiquitinylation of PD-L1 do not restrict interaction with BMS-202: A molecular modeling study. (October 2020)

• Record Type: Journal Article
• Authors: Bailly, Christian; Vergoten, Gérard
• Is Part Of: Computational biology and chemistry. Volume 88(2020)
• Publication Date: 2020-10
• Subjects: PD-L1 Programmed cell death ligand 1 -- PTM post-translational modification -- Ub ubiquitin; PD-L1 -- BMS-202 -- Protein glycosylation -- Ubiquitin -- Drug-protein binding -- Molecular modelling; Chemistry -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; Biochemistry -- Data processing; Biology -- Data processing; Molecular biology -- Data processing; Periodicals; Electronic journals; 542.85
• DOI: 10.1016/j.compbiolchem.2020.107362
• Languages: English
• ISSNs: 1476-9271

Abstract

Graphical abstract: Highlights: Proteins PD-1 and PD-L1 (Programmed cell Death protein-1/Ligand 1) are major glycoprotein targets in oncology. We designed two complementary models of protein PD-L1 with full N-glycosylation and ubiquitination. N-glycosylation and ubiquitination do not impede the stabilization of the protein dimer by the anticancer drug BMS-202. The drug can target the different glycoforms of the protein. Abstract: The Programmed cell Death protein-1/Ligand 1 (PD-1/L1) checkpoint is a major target in oncology. Monoclonal antibodies targeting PD-1 or PD-L1 are used to treat different types of solid tumors and lymphoma. PD-L1-binding small molecules are also actively searched. The lead compound is the biphenyl drug BMS-202 which stabilizes PD-L1 protein dimers and displays a potent antitumor activity in experimental models. Here we have investigated the effect of N-glycosylation (at N35, N192, N200 and N219) and mono-ubiquitination (at K178) of PD-L1 on the interaction with BMS-202 by molecular modeling. Two complementary tridimensional models of PD-L1, based on available crystallographic structures, were constructed with BMS-202 bound. The structures were glycosylated, with a fucosylated bi-antennary N-glycan and ubiquitinated. Model 1 refers to glycoPD-L1 bearing 16 N-glycans, with or without 4 ubiquitin residues. Model 2 presents 8 N-glycans and 2 ubiquitin residues. In both cases, BMS-202 was bound to the protein interface, stabilizing a PD-L1 dimer. The incorporation of the N-glycans or the ubiquitins did not significantly alter the drug-protein recognition. The interface of the drug-stabilized protein dimer is unaffected by the glycosylation or ubiquitination. Calculations of the binding energies indicated that the glycosylation slightly reduces the stability of the drug-protein complexes but does not prevent the drug binding process. Our modeling study suggests that the drug can target efficiently the different forms of PD-L1 in cells, glycosylated, ubiquitinated or not. These models of N-glycosylated and ubiquitinated PD-L1 will be useful to study other PD-L1 protein complexes.