Structural Studies of Inhibitors with Clinically Relevant Influenza Endonuclease Variants

Vital to the treatment of influenza is the use of antivirals such as Oseltamivir (Tamiflu) and Zanamivir (Relenza); however, antiviral resistance is becoming an increasing problem for these therapeutics. The RNA-dependent RNA polymerase acidic N-terminal (PAN) endonuclease, a critical component of influenza viral replication machinery, is an antiviral target that was recently validated with the approval of Baloxavir Marboxil (BXM). Despite its clinical success, BXM has demonstrated susceptibility to resistance mutations, specifically the I38T, E23K, and A36 V mutants of PAN. To better understand the effects of these mutations on BXM resistance and improve the design of more robust therapeutics, this study examines key differences in protein–inhibitor interactions with two inhibitors and the I38T, E23K, and A36 V mutants. Differences in inhibitor binding were evaluated by measuring changes in binding to PAN using two biophysical methods. The binding mode of two distinct inhibitors was determined crystallographically with both wild-type and mutant forms of PAN. Collectively, these studies give some insight into the mechanism of antiviral resistance of these mutants.


Mutation Generation
Point mutations were generated in the PAN endonuclease pET 28a plasmid by QuikChange mutagenesis (Agilent).Each PCR reaction of 50 μl contained 50 ng of template, 125 ng primer pair, 200 μM dNTPs and 3 units of Pfu DNA polymerase.The PCR cycles were initiated at 95°C for 1 min to denature the template DNA, followed by 12 amplification cycles.Each amplification cycle consisted of 95 °C for 50 sec, 60 °C for 1 min, and 68 °C for 6 min.The PCR cycles were finished an extension step at 68 °C for 7 min.The PCR products were treated with 5 units of DpnI at 37°C for 1 h and transformed into ultra competent cells and plasmid extracted.All mutations were first verified by Sanger Sequencing (Eton Biosciences) and then full, Non-Sanger plasmid sequencing (Primordium Labs) to ensure no other mutations occurred during the QuikChange prior to expression and purification.

Protein Expression and Purification
Expression and purification of PA N endonuclease was performed as reported previously (J. Med. Chem. 2019, 62, 9438-9449).The pandemic isolate A/California/04/2009(H1N1) Nterminal PA (PA N ) endonuclease Δ52-64:Gly truncated construct was expressed from a pET-28a parent vector containing a kanamycin-resistance reporter gene with expression inducible by the Lac1 operon.PA N endonuclease was expressed as an 8-histidine tagged fusion protein cleavable by TEV protease.The transformation protocol was adapted from pET system manual (Novagen) using single competent BL21 cells.Briefly, 1 μL of 25 ng/μL recombinant plasmid was used for transformation.Cells were mixed with the plasmid and were heat shocked at 42 °C for 30 sec followed by incubation on ice for 3 min.Outgrowth was plated on LB agarose plates containing 50 μg/mL kanamycin and was incubated overnight at 37 °C.One colony was scraped from the LB plate and added to 5 mL of SOC broth containing 50 μg/mL kanamycin and was incubated overnight at 37 °C with shaking at 125 rpm.SOC media (100 mL) containing 50 μg/mL kanamycin was combined with the 5 mL overnight growth and was incubated with shaking at 200 rpm at 37 °C until the OD 600 of this starter culture reached >2 (3-4 h).The culture was equally divided into 6, 2L flasks containing 1L of expression media (TB media with added 0.2% dextrose, 0.1 mM MnCl 2 , and 0.1 mM MgSO 4 , 50 μg/mL kanamycin).
Cells were grown to and OD 600 between 0.4-0.6 at room temperature with shaking at 200 rpm (3-4 h).Expression was then induced by addition of IPTG to a final concentration of 0.1 mM.
The cultures were grown with vigorous shaking (250 rpm) overnight at room temperature.
After ~18 h the cells were harvested by centrifuging at 2000g for 30 min at 4 °C.The resulting paste was stored at -80 °C prior to lysis.
The supernatant was decanted from the pellet, and a HisTrap FF (Cytiva) column was utilized to isolate His-tagged fusion protein from the cell lysates according to the manufacturer's recommendations at 4 °C.Briefly, cell-free lysates were loaded on 5 mL column that had previously been charged with Ni ions.The column was then washed with binding buffer (20 mM Na 2 PO 4 , 500 mM NaCl, 25 mM imidazole, pH 7.4) until fraction absorbance reached a steady baseline.The protein was then eluted over a gradient from 0-100% elution buffer (20 mM Na 2 PO 4 , 500 mM NaCl, 500 mM imidazole, pH 7.4) at a flow rate of 4 mL/min.PA N endonuclease eluted between 40-60% elution buffer.SDS-PAGE analysis showed a band corresponding to PA N endonuclease running at ~23 kDa with several small impurities.
Fractions containing PA N endonuclease were combined in a 10K MWCO dialysis bag with 1000 units of TEV protease and were dialyzed against dialysis buffer (100 mM NaCl, 1 mM dithiothreitol, 1 mM MnCl 2 , 20 mM Tris, 5% glycerol, pH 8.0) overnight with three buffer exchanges.The proteolytic cleavage of the fusion protein is slow and greatly benefits from the addition of excess TEV protease.A white precipitate forms over time.The solution was run through the Histrap FF column equilibrated with the same binding buffer as before.The resulting flow through contained His-cleaved PA N endonuclease, which was then concentrated to 5-10 mg/mL using a pressurized Amicon and/or spin Amicon concentrator.The concentrated protein was then purified on a gel-permeation size exclusion column (GE Superdex 75, 10/300 GL) according to manufacturer recommendations in buffer (150 mM NaCl, 2 mM MgCl 2 , 2 mM MnCl 2 , 20 mM HEPES, pH 7.5).A large peak corresponding to the cleaved PA N endonuclease eluted at ~12 mL eluent.A small shoulder before the main peak was occasionally observed, which contained primarily uncleaved and/or unfolded PA N endonuclease construct.
Fractions containing pure cleaved PA N endonuclease were combined and concentrated to 2-5 mg/mL.Stored protein was flash-frozen in liquid nitrogen and was kept at -80 °C.This protein was suitable for use in enzyme or thermal shift assays or for protein crystallography.

Differential scanning fluorometry (DSF)
Each well of a 96-well 0.   For a step-by-step explanation of the process used here please refer to Bio-protocol 2020, 10, e3574-e3574.

Protein Crystallography
Purified protein for crystallization was stored at 2.2-4.3 mg/mL at -80 °C after flash freezing in buffer consisting of 150 mM sodium chloride, 20 mM HEPES (pH 7.5), 2 mM MgCl 2 , and 2 mM MnCl 2 .Co-crystallization and crystal soaking methods were used to obtain co-crystal structures of inhibitors bound to PA N endonuclease.BXA was purchased from Fisher and used without further purification.Compound 23 was previously synthesized according to a prior procedure (J.Med.Chem. 2019, 62, 9438-9449).For co-crystallization, protein was incubated with 0.5 mM inhibitor for 1 hour on ice prior to setting the crystallization drops.For crystal soaking, fully formed holo crystals were transferred to a new drop containing 5 µL of reservoir solution and 1 µL of 50 mM DMSO inhibitor stock solution (final concentration 8.3 mM).
Crystals were left undisturbed overnight and either stored in liquid nitrogen or collected on an in-house X-ray diffractometer the following day.In both crystallization methods, crystals were    The electron density of BXA is displayed as a gray mesh.Mesh is 2Fo-Fc contoured at 1.8σ.
plate wells were sealed prior to analysis, and the plate was then heated in a thermocycler from 25 to 99 °C at a ramp rate of 0.05 °C/sec.Fluorescence was read using the ROX filter channel (λex = 580 nm; λem = 623 nm), and the fluorescence signal was fitted to a first derivative curve to identify TM.Native WT PAN endonuclease was generally observed to melt with a TM= 58−59 °C, I38T TM= 58-59 °C, E23K TM= 60-61 °C, and A36V TM= 51-52 °C.

Figure S1 .
Figure S1.Representative thermal shift traces for I38T mutant with BXA.Fluorescent melting curve trace (top) and first derivative of the fluorescent trace (bottom).The maximum of the derivative peak for each trace is the melting point of the protein.Red traces are WT with BXA melting curve replicates, blue traces are I38T mutant with BXA melting curve replicates.A total of eight independent replicates were performed for each experiment.

Figure S2 .
Figure S2.Left: Crystal structure of WT PAN (gray) overlayed with I38T mutant (cyan).Right: crystal structure of WT PAN (gray) overlayed with E23K mutant (pink).The protein backbone (WT gray, I38T cyan, and E23K pink) is shown as a cartoon and Mn 2+ ions are shown as purple spheres.

Figure S5 .
Figure S5.Co-crystal structure of 23 bound to WT PAN (gray) overlayed with 23 bound to the

Table of Contents
2 mL optical MicroAmp (ThermoFisher) thermocycler plate contained a volume of 20 μL containing final concentrations of 1 μg PAN endonuclease, 200 μM or 1 mM inhibitor, and 1×SYPRO orange Thermal Shift dye in buffer (150 mM NaCl, 2 mM MnCl2, 20 mM HEPES pH 7.5) with 4% DMSO.A master mix containing PAN endonuclease, 1×SYPRO orange Thermal Shift dye, and buffer was made and 16 μL added to each well.Either 1mM inhibitor stocks or 20% DMSO stocks were made, to which 4 μL was added to the plate to make a total of 20 μL final volume.Each well was mixed thoroughly by pipetting up and down, with care to prevent air bubbles from forming.The presence of this small concentration of DMSO was found to have a negligible effect on ΔTM values of native PAN endonuclease.Thermocycler

Table S1 .
X-ray crystallographic data collection and refinement statistics for the WT, I38T, and E23K apo structures and E23K with 23 bound.

Table S2 .
X-ray crystallographic data collection and refinement statistics for WT and I38T with 23 bound, and I38T and E23K with BXA bound.

Table S3 .
X-ray crystallographic data collection and refinement statistics for A36V with