SAXS data based glycosylated models of human alpha-1-acid glycorprotein, a key player in health, disease and drug circulation.

Plasma Alpha-1-glycoprotein (AGP) binds diverse drugs, its isoforms and their levels vary significantly in acute phases of health. Relative binding pattern of drugs to AGP and albumin has been used to model their release profiles, and structural insights on glycosylated form of AGP will improve predictions. Main challenge is the heavy and heterogeneous glycosylation of AGP molecules. Our small angle X-ray scattering (SAXS) data on plasma extracted AGP showed interparticulate effect from 283 to 313 K which disappeared irreversibly upon further heating to 343K. Using ALPHAFOLD2 server, the protein only portion could be modelled but as expected its theoretical SAXS profile did not match acquired experimental data. Using mass spectra-based information, we attached representative glycan motifs at known sites to compute four models of fully glycosylated AGP. Importantly, calculated SAXS profiles of these models agreed with our experimental data. These representative glycosylated models were further analyzed for molecular motions using Normal Mode Analysis and all-atom Molecular Dynamics simulations in reference to SAXS data. Overall, we show that SAXS data-based models of glycoprotein are better representation of this biopharmaceutical molecule and provide them for structure-based drug profile estimations.