A surface plasmon resonance assay to determine the effect of influenza neuraminidase mutations on its affinity with antiviral drugs.
Thesis DisciplineChemical Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
The outbreak of pandemic influenza and its ability to spread rapidly makes it a severe threat to public health. Antiviral drugs such as oseltamivir (Roche’s Tamiflu™) and zanamivir (GlaxoSmithKline’s Relenza™) are neuraminidase (NA) inhibitors (NI), which bind more tightly to NA than its natural substrate, sialic acid. However, the virus can acquire resistance to antiviral drugs by developing single point mutations (such as H274Y) in the target protein. Thus in some cases the drugs may not be as effective as expected. The high level of inconsistency exhibited by fluorometric assays and the short half-life of the chemiluminescent assay for monitoring drug resistance lead to the need for a simple, label-free, reliable assay. To address this problem, this work focused on three main objectives: 1) to determine the binding affinities of two common anti-viral drugs (oseltamivir and zanamivir) against the influenza NA wild type and drug resistant mutants using bioinformatics software Schrodinger Suite™ 2010. 2) To develop a reliable label-free, real-time, surface plasmon resonance (SPR) assay to measure the binding affinity between influenza viral coat protein neuraminidase (wild type and mutant) and anti-viral drugs. 3) To develop an SPR inhibition assay to quantitatively compare the interactions of sialic acid, zanamivir and oseltamivir with the viral coat protein neuraminidase (wild type and mutant).
The entire docking process was carried out using Schrödinger Suite™ 2010. The 2009 pandemic H1N1 neuraminidase (PDB: 3NSS) was used throughout the docking studies as the wild type structure. Five mutants (H274Y, N294S, H274N, A346N and I222V) and three ligands (sialic acid, oseltamivir and zanamivir) were built using the maestro module. The grid-based ligand docking with energetics (GLIDE) module and induced fit docking (IFD) module were used for docking studies. The binding affinities, Gibbs free energy change (∆G) and molecular mechanics-generalized born energy/ solvent accessible area (MM-GB/SA) values for wild-type NA interactions show that both the antiviral drugs studied interact strongly with the wild-type protein. The ∆G values for all antiviral interactions with mutant NA forms were reduced in magnitude, thereby indicating that they are less favourable than interactions with the wild-type protein. A similar trend was observed with MM-GB/SA results. Amongst all of the computed values, MM-GB/SA was the closest to the experimental data. In several cases, the interactions between the anti-viral drugs and NA mutants were markedly less favourable than those between sialic acid and the same mutants, indicating that these mutations could confer anti-viral resistance.
Influenza NA wild-type and H274Y mutant were expressed in baculovirus expression system (BVES) in insect cells. The expressed proteins were partially purified using the standard purification techniques of anion exchange and size exclusion chromatography (SEC). A fluorometric activity assay was performed on the recombinant proteins. Both the wild type and the mutant showed similar level of activities. In addition, the recombinant NAs were used in an inhibition assay. Oseltamivir was found to be sensitive to wild type protein (IC50 = 0.59 nM) and resistant to the H274Y mutant protein (IC50 = 349.43 nM). On the other hand, zanamivir was sensitive to both wild type (IC50 = 0.26 nM) and the H274Y mutant (IC50 = 0.44 nM). This indicated that zanamivir was a more potent inhibitor than oseltamivir. These findings were in good agreement with the literature.
An SPR assay for accurate monitoring of influenza antiviral drug resistance was developed. A spacer molecule (1, 6- hexanediamine) was site-specifically tethered to the inert 7-hydroxyl group of zanamivir. The tethered zanamivir was immobilized onto an SPR GLC chip to obtain a final immobilization response of 431 response units (RU). The reference subtracted binding responses obtained for NA wild-type and H274Y mutant were analysed using the ProteOn Manager™ Software tools. The SPR curves were fitted to a simple Langmuir 1:1 model with drift to obtain association rate constant (ka) and dissociation rate constants (kd). The relative binding values obtained from literature and the current SPR assay (1.9 and 1.7 respectively) suggested that the current SPR assay yielded similar results to the existing labelled enzymatic assay. In addition, an SPR inhibition assay was developed. The calculated IC50-spr values were compared and it was observed that oseltamivir was sensitive to wild type protein (IC50-spr = 7.7 nM) and resistant to the H274Y mutant protein (IC50-spr = 256 nM). On the other hand, zanamivir was sensitive to both wild type (IC50-spr = 2.16 nM) and the H274Y mutant (IC50-spr = 2.4 nM). Sialic acid was also found to be sensitive to both wild type (IC50-spr = 5.5 nM) and H274Y mutant (IC50-spr = 3.25 nM). In the cases studied, the viral proteins remained sensitive to sialic acid, consistent with retention of virulence of these mutant strains. It was concluded that zanamivir is a more potent inhibitor than oseltamivir for treating the H274Y mutant. Comparison of the SPR inhibition results with the docking results revealed a similar trend. The wild-type NA and H27Y mutant retained binding affinity for sialic acid and zanamivir. Oseltamivir showed a significant decrease in binding affinity for the H274Y mutant compared with the wild-type. This was because of the disruption of the salt bridge formation within NA that was vital for oseltamivir activity.
To my knowledge, this is the first SPR biosensor assay developed to monitor influenza antiviral drug resistance. There is a tremendous scope to extend this study to more mutants and new antiviral drugs. This could pave the way for a reliable SPR biosensor assay to replace low consistency labelled enzymatic assays.