Polymer Microarrays Rapidly Identify Competitive Adsorbents of Virus-like Particles (VLPs)

Here we show that an adaption of polymer micro array method for screening bacterial-surface interactions allows for screening of polymers for desirable material-viron interactions. Non-pathogenic virus like particles including fluorophores are exposed to the arrays in aqueous buffer as a simple model of virons carried to the surface in saliva/sputum. Competitive binding of Lassa and Rubella particles is measured to probe the relative binding properties of a selection of copolymers. This provides the first step in the development of a method for discovery of novel materials with promise for viral binding, with the next being development of this method to assess absolute viral adsorption and assessment of the attenuation of the activity of live virus which we propose would be part of a material scale up step carried out in biological laboratory safety level 3 facilities and the use of more complex media to represent biological fluids. Abstract The emergence of SARS-CoV-2 highlights the global need for platform technologies to enable rapid development of diagnostics, vaccines, treatments, and personal protective equipment (PPE). However, many current technologies require the detailed mechanistic knowledge of specific material-virion interactions before they can be employed, for example to aid in the purification of vaccine components, or in design of more effective PPE. Here we show that an adaption of polymer micro array method for screening bacterial-surface interactions allows for screening of polymers for desirable material-viron interactions. Non-pathogenic virus like particles including fluorophores are exposed to the arrays in aqueous buffer as a simple model of virons carried to the surface in saliva/sputum. Competitive binding of Lassa and Rubella particles is measured to probe the relative binding properties of a selection of copolymers. This provides the first step in the development of a method for discovery of novel materials with promise for viral binding, with the next being development of this method to assess absolute viral adsorption and assessment of the attenuation of the activity of live virus which we propose would be part of a material scale up step carried out in biological interactions. Determination of absolute load, more models of real and measurement of of live virus Abstract The emergence of SARS-CoV-2 highlights the global need for platform technologies to enable rapid development of diagnostics, vaccines, treatments, and personal protective equipment (PPE). However, many current technologies require the detailed mechanistic knowledge of specific material-virion interactions before they can be employed, for example to aid in the purification of vaccine components, or in design of more effective PPE. Here we show that an adaption of polymer micro array method for screening bacterial-surface interactions allows for screening of polymers for desirable material-viron interactions. Non-pathogenic virus like particles including fluorophores are exposed to the arrays in aqueous buffer as a simple model of virons carried to the surface in saliva/sputum. Competitive binding of Lassa and Rubella particles is measured to probe the relative binding properties of a selection of copolymers. This provides the first step in the development of a method for discovery of novel materials with promise for viral binding, with the next being development of this method to assess absolute viral adsorption and assessment of the attenuation of the activity of live virus which we propose would be part of a material scale up step carried out in

Here we show that an adaption of polymer micro array method for screening bacterial-surface interactions allows for screening of polymers for desirable material-viron interactions. Non-pathogenic virus like particles including fluorophores are exposed to the arrays in aqueous buffer as a simple model of virons carried to the surface in saliva/sputum. Competitive binding of Lassa and Rubella particles is measured to probe the relative binding properties of a selection of copolymers. This provides the first step in the development of a method for discovery of novel materials with promise for viral binding, with the next being development of this method to assess absolute viral adsorption and assessment of the attenuation of the activity of live virus which we propose would be part of a material scale up step carried out in biological laboratory safety level 3 facilities and the use of more complex media to represent biological fluids.
File list (4)  interact with viral surface components via charge-mediated association, and incorporated additional hydrophobic and hydrophilic co-monomers to tune relative binding affinities. The method was exemplified using non-replicating virus-like particles (VLPs), from Rubella and Lassa viruses, as structural mimics of infectious disease-causing pathogens, but without the full genome encoding for infectivity. These are ideally suited to probing viral binding outside the high level of biosafety restrictions required for live-virus work. Various approaches used to fabricate polymer microarrays with in situ polymerisation onto hydrogel coated class slides have been previously described [14][15][16][17] . The polymer microarray in this study involved the contact printing and subsequent in situ photopolymerisation of an array of 300 monomer mixtures (14 homopolymers, alongside 182 binary, 72 ternary and 32 quaternary copolymers) onto an epoxy functionalised, poly(2-hydroxyethylmethacrylate) (pHEMA) coated glass slide in triplicate. Further details of the slide preparation and details on the identity of the 300 copolymers (Error! Reference source not found., Error! Reference source not found.) are included in the supplementary information. In brief, the 300 copolymers were formed from the 14 monomers depicted in Figure 1 b. The charged monomers were selected to provide electrostatic interactions with charged amino acid residues in the surface-displayed VLP proteins. The other monomers (hydrophilic and hydrophobic) were selected to broaden the chemical diversity of the array system and introduce further selectivity via H-bonding interactions at, for example, serine and threonine residues or solvophobic association at leucine, isoleucine and aromatic rich regions of the proteins and at viral lipid membranes. After printing, the fabricated slides were imaged using phase contrast microscopy (Error! Reference source not found.a) and the chemical identities of the polymer spots were analysed using time of flight-secondary ion mass spectrometry (ToF-SIMS), representative ion image presented in Error! Reference source not found.a. Both techniques showed good polymer spot localisation, and evidence from ToF-SIMS indicated no significant carryover between the printing of different monomer solutions after appropriate optimisation of the process parameters.
We then assessed the short-term adsorption, as an accessible measure for binding interactions, of each VLP to the library in aqueous media as a simplified version of biological fluids. A solution containing both AlexaFluor-488 tagged Rubella VLPs (2.5 g/mL) and AlexaFluor-555 tagged Lassa VLPs (2.5 g/mL) in Dulbecco's Phosphate Buffered Saline (DPBS) were prepared and incubated with the polymer arrays (details of the VLP preparation are included in the SI). The concentrations chosen were arbitrary and it is anticipated that to fully understand binding a range of concentration would be employed in future work. The microarray was washed (4 x 100 mL Milli-Q water, 1x100 mL DPBS) and then placed immediately into the incubation solution, rocked in the dark gently at ambient temperature to achieve uniform exposure of the spots to the VLPs. After four hours, found to be optimal for measuring equilibrium adsorption, the slide was washed again (1 x 100 mL DPBS, 4 x 100 mL Milli-Q water) and allowed to air-dry in a dark cupboard. Fluorescence images of the array both before and after incubation were acquired using an automated microscope (IMSTAR) and processed using Image J software. Each spot of a composite image (autofluorescence was subtracted from the final result) was cropped using a circle to the border of the spots in order to determine the fluorescence intensity per pixel from each spot due to VLP binding.
After incubation with dye-labelled VLPs there was measurable fluorescence intensity (at least 3 x background) observed on the array slide, localised on some of the printed spots, indicating adsorption ( Figure 1c, Error! Reference source not found. in the SI shows there was little autofluorescence). The results were broadly reproducible even with non-optimised incubation protocols, with the three technical repeats for both types of VLPs showing very similar fluorescence profiles. It was also apparent that preferential adsorption of one labelled VLP compared to the other occurred at some polymer spots. To gain a measure of this selectivity, a binding selectivity index was calculated using the Lassa/Rubella ratio in Figure 2. This approach enabled the rapid identification of co-polymer materials that were capable of discriminating between different VLPs based in competitive adsorption. Furthermore, it was possible to separate the selectivity ratios into statistically significant groups (p < 0.05) for high, intermediate and low VLP-binding polymers ( Figure 2, Error! Reference source not found.,Error! Reference source not found.). These data provide insight on the chemistries of co-polymers able to sequester Lassa and/or Rubella VLPs from aqueous suspensions.
The copolymers were chosen to be combinations of monomers with ionisable functionality, to interact with proteins on an electrostatic basis, and with monomers containing non-charged hydrophilic and hydrophobic side-chains, to modulate H-bonding and hydrophobic associations at non-charged regions of VLP surfaces. The highest selectivity ratio of Lassa over Rubella was shown for a copolymer produced from a monomer mixture 66:17:17 tBAEMA:HEMA:MAAcid %w/w (see Figure 2 for structures, Error! Reference source not found.,Error! Reference source not found.). All of the top Lassa/Rubella VLP binders contained significant amounts (>30%) of monomers expected to be positively charged under the incubation conditions (DPBS buffer). The Lassa VLPs contain a tripartite spike complex derived from a single precursor glycoprotein, 18 which computational studies have predicted to display an isoelectric point of 7.54 19 . The protein part of the VLP would thus have only a slight negative charge at the incubation conditions, and would not be expected to bind preferentially to only positive charged monomers. However, charge heterogeneities on the VLPs could nevertheless result in regions of more dense charge or of H-bond acceptor/donor interactions, which could still allow spatially-matched charge-charge and H-bonding interactions with an appropriately matched surface. The highest selectivity index for Rubella over Lassa was found to be from the 70:30 pEGDA:CEA comonomer combination (although it should be noted that the standard deviation from the three repeats is relatively high) ( Figure 2, Error! Reference source not found.,Error! Reference source not found.). The top Rubella binders typically include more acidic and neutral polymers rather than the strongly cationic polymers which were observed to be more selective for Lassa VLPs. It is expected that the E1 protein would play the largest role in surface charge of Rubella VLP and this has a pI of 6.5 20 . Thus it was apparent that simple charge-charge interactions were not predominant for VLP binding at surfaces, which may have arisen due to the surface presentation of the viral spike glycoproteins constrained at the VLP surface compared to the recombinantly expressed and purified proteins themselves.
Further experiments are required to understand fully the mechanisms by which the VLPs interacted with the different polymer surfaces, yet it is clear here that the developed method is able to identify multifaceted candidate materials that interact with VLPs in a rapid and experimentally simple fashion The ability to screen multiple combinations simultaneously for their binding affinity therefore has the potential to identify new and perhaps unanticipated candidate materials suitable for more selective binding and inactivation of viruses. Assessment of absolute binding affinity would require the construction of adsorption isotherms using a range of viron concentrations, ideally complemented with the absolute quantification of viral load. Figure 2: Binding selectivity indices for Lassa/Rubella for a selection of VLP binding copolymers grouped by statistically significant selectivity for Lassa VLP over Rubella (p < 0.05). Data was calculated from the fluorescence images depicted in Figure 1. Background was subtracted from the intensity values (per pixel) before the values were ratioed. Error given is the standard deviation from three technical replicates on one array slide. Groupings depicted are taken from Tukey HSD test carried out at 95% confidence interval and show representative groups from the dataset.
In summary, we have reported the development of a high-throughput copolymer microarray system which enabled rapid identification of copolymers with competitive adsorption of fluorescently tagged VLPs. Whilst more investigation is required to understand fully the mode of binding of the VLPs to the different polymer surfaces, there were clear differences in selectivity observed during the comparative binding experiments which could be exploited when developing novel materials. As the method was exemplified using virus-like particles, which mimic the structure of the native virus but without the viral genome, this study marks a starting point to study material-virion interactions. Determination of absolute viral load, more complex media models of real situations and measurement of inactivation of live virus using scaled up analysis in biological safety level 3 facilities are the next steps. Novel materials identified in such an approach could be exploited in applications including the downstream purification of vaccine components or the more clinically informed design of PPE.

Array Fabrication
Standard glass microscope slides (Corning, 75 x 25 mm, 0.96 to 1.06 mm thick) were activated using an oxygen plasma (Diener Nano System, 40 KHz, 1000 W, 5 min, O2 gas at 0.09 mbar). The activated slides were then immediately immersed in a solution of 3-(glycidyloxypropyl)trimethoxysilane (10 mL) in toluene (500 mL, dried over 3Å molecular sieves) at 50°C under an argon atmosphere. After a minimum of 3 hours, the slides were removed from the solution, cooled to room temperature and sonicated with acetone (2x100mL). After being stored at <50 mTorr for a minimum of 12 hours, slides were dip-coated four times (using a Holmarc Opto-Mechatronics Pvt. Ltd Dip Coating Unit -Model HO-TH-01, dip speed 9 mm/s, return speed 2 mm/s) in a 4% w/v poly(2-hydroxyethyl methacrylate (poly(HEMA), suitable for cell culture) in ethanol (95% v/v in water) solution. Slides were then stored at <50 mTorr for a minimum of 12 hours before printing. Monomer printing and photo polymerisation was performed using 946MP6B pins (ArrayIt) that have a 220 m diameter and a Biodot XYZ3200 contact printer. The printing conditions were O2 <2000 ppm (argon atmosphere), 25°C and 30% relative humidity. Commercially available monomers were spotted onto the above modified microscope slides in triplicate, such that there were 3 array repeats on each slide. Monomer solutions were prepared as 50% v/v or w/v solutions in N,N-dimethylformamide (DMF). The required amount of monomer to produce the required copolymer was dispensed into an individual well of a 384 polypropylene well plate, such that the final volume was 20 L. Subsequently the radical photoinitiator (2,2-dimethoxy-2-phenylacetophenone) was added as a 10 L 3% w/v solution in DMF to each well, giving a final initiator concentration of 1% w/v. Monomer spots on the slides were polymerised by exposure to shortwave 365 nm UV light (density -30 mV/cm 2 ) after each round of printing of thirty slides and then all slides were exposed to shortwave 365 nm UV light for a further 10 minutes after all polymer spots were printed. The slides were then dried at <50 mTorr for at least 7 days before use.
Monomers were selected based on their previous interaction with proteins and based on their hydrophilicity, hydrophobicity and/or acidity in order to tune the copolymers produced to the differing regions of a proteins surface. The structures of the monomers are featured in Error: Reference source not found of the main paper and are reproduced again below in Figure S1. All 14 were used to synthesise their corresponded homopolymers, whilst all were also mixed in a 70:30 v/v to create the binary polymers. Ternary and quaternary polymers were prepared using the comonomer ratios in Table S1 and Table S2.

Time-of-Flight Secondary-Ion Mass Spectrometry (ToF-SIMS) Analysis
ToF-SIMS analysis was carried out using a ToF-SIMS IV (IONTOF GmbH) instrument operated using a 25 kV Bi3 + primary ion source exhibiting a pulsed target current of ∼1 pA. Samples were scanned at a pixel density of 100 pixels per mm, with 8 shots per pixel over a given area. An ion dose of 2.45 × 10 11 ions per cm 2 was applied to each sample area ensuring static conditions were maintained throughout. Both positive and negative secondary ion spectra were collected. Owing to the nonconductive nature of the samples, a low energy (20 eV) electron flood gun was applied to provide charge compensation.

Virus-Like Particle Preparation
Rubella VLPs were fabricated in 293 freestyle cells which were transiently transfected with a single mammalian expression plasmid encoding the three structural proteins Capsid, E2 and E1. 21,22 The particles were subsequently secreted into the cell culture media and then harvested 72 hours after transfection. Purification was performed using the sucrose cushion method, with the supernatant overlayed onto 20% sucrose and then spun at 117,734g at 4°C for 1.5 hours The subsequent pellet was resuspended in PBS buffer. Lassa VLPs were produced using a previously published method 18 . In brief, MDCK II cells stably expressing Lassa glycoprotein were grown to confluency and the supernatant harvested after 96 hrs. This was then ultracentrifuged at 110,000g at 4°C for 1.5 hours through a 20% sucrose cushion and then resuspended in PBS. Rubella VLPs were then tagged using an AlexaFluor-488 labelling kit (at a concentration of 53 g/mL), whilst Lassa VLPs were tagged using an AlexaFluor-555 labelling kit (at a concentration of 179 g/mL).

Statistical Analysis
Statistical analysis and graphical plots were carried out in R version 4.0.0 using RStudio version 1.2.5033 as integrated development environment (IDE). Tukey HSD tests were carried out using the HSD.test function from the agricolae package version 1.3-2, with p-value < 0.05 used to group means. Boxplots were plotted using the ggplot + geom_boxplot function from the ggplot2 package version 3.3.0. Figure S1: Chemical structure of monomers used in the fabricated copolymer array. Note that in the two bottom groups, only the group on the left was used for the synthesis of ternary and quaternary copolymers. Table S1: Ternary and quaternary polymers were made in the following ratios on %v/v basis. One of the monomers from the top group was taken and then mixed with different amounts of HA, HEMA or MAAcid to make the second, third and further copolymer components. % of each material added to create ternary and quaternary polymers, where Group A is any of the monomers featured in the top of Figure S1. Group A HA  HEMA  MAAcid  66  17  17  0  66  17  0  17  66  0  17  17  17  66  0  17  17  66  17  0  0  66  17  17  17  0  66  17  17  17  66  0  0  17  66  17  17  17  0  66  17  0  17  66  0  17  17  66  70  10  10  10  51  17  17  17   40   20  20  20  25  25  25 Table S2: Identity of each copolymer spot in a repeat. Labelling as in Figure S1 A B C    Table S3 for full list of group memberships) (d) FITC/Rhodamine (Rubella/Lassa) ratio (colour used for illustration of Tukey HSD groupings, p < 0.05, see Table S3 for full list of group memberships)

Abstract
The emergence of SARS-CoV-2 highlights the global need for platform technologies to enable rapid development of diagnostics, vaccines, treatments, and personal protective equipment (PPE). However, many current technologies require the detailed mechanistic knowledge of specific materialvirion interactions before they can be employed, for example to aid in the purification of vaccine components, or in design of more effective PPE. Here we show that an adaption of polymer micro array method for screening bacterial-surface interactions allows for screening of polymers for desirable material-viron interactions. Non-pathogenic virus like particles including fluorophores are exposed to the arrays in aqueous buffer as a simple model of virons carried to the surface in saliva/sputum. Competitive binding of Lassa and Rubella particles is measured to probe the relative binding properties of a selection of copolymers. This provides the first step in the development of a method for discovery of novel materials with promise for viral binding, with the next being development of this method to assess absolute viral adsorption and assessment of the attenuation of the activity of live virus which we propose would be part of a material scale up step carried out in biological laboratory safety level 3 facilities and the use of more complex media to represent biological fluids.

Main Text
Common strategies for selective biomolecular recognition in diagnostics typically utilise antigenantibody interactions, such as in common ELISA immunoassays. 1, 2 Whilst these assays typically allow high selectivity to be obtained, there are a number of drawbacks which limit their more widespread usage, including the cost of manufacture (each antigen needs a specific antibody to be developed) and the storage and transport of what are typically thermally sensitive reagents. These disadvantages become more important when the target application requires interaction with classes of related biomolecules rather than specific individual analytes. Prior studies have used low-cost polymers to modify nanocrystals 3 and chromatographic materials 4, 5 with the aim of introducing broad-spectrum binding affinity towards viral targets. However, the myriad of putative copolymer structures derived from even a small number of monomers, means that to date only a fraction of the chemical space available for polymeric affinity agents and biomolecular sequestrants has been explored.
Polymer microarrays have been developed to facilitate simultaneous investigation of many thousands of chemically unique materials for biologic-material affinity on a single surface [6][7][8][9][10][11][12][13] . This high-throughput approach has now been used to identify materials for a range of biomedical applications, such as the inhibition of bacterial biofilm formation 13 and the growth of stem cells with controllable behaviour 8 . Polymer microarrays can be easily fabricated using inkjet or contact printing, coupled with in situ polymerisation from low quantities of commercial photocurable monomers. 6 In this work we present a method based on a polymer microarray platform to rapidly identify materials derived from commercially available monomers capable of differential adsorption of virus-like particles (VLPs) in competitive binding experiments. The fabricated array contained monomer units expected to interact with viral surface components via charge-mediated association, and incorporated additional hydrophobic and hydrophilic co-monomers to tune relative binding affinities. The method was exemplified using non-replicating virus-like particles (VLPs), from Rubella and Lassa viruses, as structural mimics of infectious disease-causing pathogens, but without the full genome encoding for infectivity. These are ideally suited to probing viral binding outside the high level of biosafety restrictions required for live-virus work. Various approaches used to fabricate polymer microarrays with in situ polymerisation onto hydrogel coated class slides have been previously described [14][15][16][17] . The polymer microarray in this study involved the contact printing and subsequent in situ photopolymerisation of an array of 300 monomer mixtures (14 homopolymers, alongside 182 binary, 72 ternary and 32 quaternary copolymers) onto an epoxy functionalised, poly(2-hydroxyethylmethacrylate) (pHEMA) coated glass slide in triplicate. Further details of the slide preparation and details on the identity of the 300 copolymers ( Error: Reference source not found, Error: Reference source not found) are included in the supplementary information.
In brief, the 300 copolymers were formed from the 14 monomers depicted in Figure 1b. The charged monomers were selected to provide electrostatic interactions with charged amino acid residues in the surface-displayed VLP proteins. The other monomers (hydrophilic and hydrophobic) were selected to broaden the chemical diversity of the array system and introduce further selectivity via H-bonding interactions at, for example, serine and threonine residues or solvophobic association at leucine, isoleucine and aromatic rich regions of the proteins and at viral lipid membranes. After printing, the fabricated slides were imaged using phase contrast microscopy (Error: Reference source not founda) and the chemical identities of the polymer spots were analysed using time of flight-secondary ion mass spectrometry (ToF-SIMS), representative ion image presented in Error: Reference source not founda. Both techniques showed good polymer spot localisation, and evidence from ToF-SIMS indicated no significant carryover between the printing of different monomer solutions after appropriate optimisation of the process parameters.
We then assessed the short-term adsorption, as an accessible measure for binding interactions, of each VLP to the library in aqueous media as a simplified version of biological fluids. A solution containing both AlexaFluor-488 tagged Rubella VLPs (2.5 g/mL) and AlexaFluor-555 tagged Lassa VLPs (2.5 g/mL) in Dulbecco's Phosphate Buffered Saline (DPBS) were prepared and incubated with the polymer arrays (details of the VLP preparation are included in the SI). The concentrations chosen were arbitrary and it is anticipated that to fully understand binding a range of concentration would be employed in future work. The microarray was washed (4 x 100 mL Milli-Q water, 1x100 mL DPBS) and then placed immediately into the incubation solution, rocked in the dark gently at ambient temperature to achieve uniform exposure of the spots to the VLPs. After four hours, found to be optimal for measuring equilibrium adsorption, the slide was washed again (1 x 100 mL DPBS, 4 x 100 mL Milli-Q water) and allowed to air-dry in a dark cupboard. Fluorescence images of the array both before and after incubation were acquired using an automated microscope (IMSTAR) and processed using Image J software. Each spot of a composite image (autofluorescence was subtracted from the final result) was cropped using a circle to the border of the spots in order to determine the fluorescence intensity per pixel from each spot due to VLP binding.
After incubation with dye-labelled VLPs there was measurable fluorescence intensity (at least 3 x background) observed on the array slide, localised on some of the printed spots, indicating adsorption ( Figure 1c, Error: Reference source not found in the SI shows there was little autofluorescence). The results were broadly reproducible even with non-optimised incubation protocols, with the three technical repeats for both types of VLPs showing very similar fluorescence profiles. It was also apparent that preferential adsorption of one labelled VLP compared to the other occurred at some polymer spots. To gain a measure of this selectivity, a binding selectivity index was calculated using the Lassa/Rubella ratio in Figure 2. This approach enabled the rapid identification of co-polymer materials that were capable of discriminating between different VLPs based in competitive adsorption. Furthermore, it was possible to separate the selectivity ratios into statistically significant groups (p < 0.05) for high, intermediate and low VLP-binding polymers ( Figure 2, Error: Reference source not found,Error: Reference source not found). These data provide insight on the chemistries of copolymers able to sequester Lassa and/or Rubella VLPs from aqueous suspensions.
The copolymers were chosen to be combinations of monomers with ionisable functionality, to interact with proteins on an electrostatic basis, and with monomers containing non-charged hydrophilic and hydrophobic side-chains, to modulate H-bonding and hydrophobic associations at non-charged regions of VLP surfaces. The highest selectivity ratio of Lassa over Rubella was shown for a copolymer produced from a monomer mixture 66:17:17 tBAEMA:HEMA:MAAcid %w/w (see Figure 2 for structures, Error: Reference source not found,Error: Reference source not found). All of the top Lassa/Rubella VLP binders contained significant amounts (>30%) of monomers expected to be positively charged under the incubation conditions (DPBS buffer). The Lassa VLPs contain a tripartite spike complex derived from a single precursor glycoprotein, 18 which computational studies have predicted to display an isoelectric point of 7.54 19 . The protein part of the VLP would thus have only a slight negative charge at the incubation conditions, and would not be expected to bind preferentially to only positive charged monomers. However, charge heterogeneities on the VLPs could nevertheless result in regions of more dense charge or of H-bond acceptor/donor interactions, which could still allow spatially-matched charge-charge and H-bonding interactions with an appropriately matched surface. The highest selectivity index for Rubella over Lassa was found to be from the 70:30 pEGDA:CEA comonomer combination (although it should be noted that the standard deviation from the three repeats is relatively high) (Figure 2, Error: Reference source not found,Error: Reference source not found). The top Rubella binders typically include more acidic and neutral polymers rather than the strongly cationic polymers which were observed to be more selective for Lassa VLPs. It is expected that the E1 protein would play the largest role in surface charge of Rubella VLP and this has a pI of 6.5 20 . Thus it was apparent that simple charge-charge interactions were not predominant for VLP binding at surfaces, which may have arisen due to the surface presentation of the viral spike glycoproteins constrained at the VLP surface compared to the recombinantly expressed and purified proteins themselves.
Further experiments are required to understand fully the mechanisms by which the VLPs interacted with the different polymer surfaces, yet it is clear here that the developed method is able to identify multifaceted candidate materials that interact with VLPs in a rapid and experimentally simple fashion The ability to screen multiple combinations simultaneously for their binding affinity therefore has the potential to identify new and perhaps unanticipated candidate materials suitable for more selective binding and inactivation of viruses. Assessment of absolute binding affinity would require the construction of adsorption isotherms using a range of viron concentrations, ideally complemented with the absolute quantification of viral load. Figure 2: Binding selectivity indices for Lassa/Rubella for a selection of VLP binding copolymers grouped by statistically significant selectivity for Lassa VLP over Rubella (p < 0.05). Data was calculated from the fluorescence images depicted in Figure 1. Background was subtracted from the intensity values (per pixel) before the values were ratioed. Error given is the standard deviation from three technical replicates on one array slide. Groupings depicted are taken from Tukey HSD test carried out at 95% confidence interval and show representative groups from the dataset.
In summary, we have reported the development of a high-throughput copolymer microarray system which enabled rapid identification of copolymers with competitive adsorption of fluorescently tagged VLPs. Whilst more investigation is required to understand fully the mode of binding of the VLPs to the different polymer surfaces, there were clear differences in selectivity observed during the comparative binding experiments which could be exploited when developing novel materials. As the method was exemplified using virus-like particles, which mimic the structure of the native virus but without the viral genome, this study marks a starting point to study material-virion interactions. Determination of absolute viral load, more complex media models of real situations and measurement of inactivation of live virus using scaled up analysis in biological safety level 3 facilities are the next steps. Novel materials identified in such an approach could be exploited in applications including the downstream purification of vaccine components or the more clinically informed design of PPE.