Affinity Sedimentation and Magnetic Separation With Plant-Made Immunosorbent Nanoparticles for Therapeutic Protein Purification

The virus-based immunosorbent nanoparticle is a nascent technology being developed to serve as a simple and efficacious agent in biosensing and therapeutic antibody purification. There has been particular emphasis on the use of plant virions as immunosorbent nanoparticle chassis for their diverse morphologies and accessible, high yield manufacturing via plant cultivation. To date, studies in this area have focused on proof-of-concept immunosorbent functionality in biosensing and purification contexts. Here we consolidate a previously reported pro-vector system into a single Agrobacterium tumefaciens vector to investigate and expand the utility of virus-based immunosorbent nanoparticle technology for therapeutic protein purification. We demonstrate the use of this technology for Fc-fusion protein purification, characterize key nanomaterial properties including binding capacity, stability, reusability, and particle integrity, and present an optimized processing scheme with reduced complexity and increased purity. Furthermore, we present a coupling of virus-based immunosorbent nanoparticles with magnetic particles as a strategy to overcome limitations of the immunosorbent nanoparticle sedimentation-based affinity capture methodology. We report magnetic separation results which exceed the binding capacity reported for current industry standards by an order of magnitude.

The concept of a plant virus-based immunosorbent nanoparticle (VIN), a plant virus or virus-like particle displaying antibody-binding proteins, has been proposed to capture antibodies for biosensing (Uhde-Holzem et al., 2016;Kuo et al., 2018) and therapeutic antibody purification (Werner et al., 2006). This nascent technology is one approach to address the need to reduce capital intensity for equitable and accessible antibody-related healthcare solutions, which could be harnessed to treat more prevalent diseases with availability of inexpensive and adequate production and purification capacity (Buyel et al., 2017). Purification can cost up to 80% of the total manufacturing expenses for antibody and other biopharmaceutical products (Yang et al., 2020). The need for lower cost and limited resource biomanufacturing technology is an important step in moving towards a more distributed form of production that benefits global health (Gomez-Marquez and Hamad-Schifferli, 2019). It may also be that these lower cost and limited resource biomanufacturing technologies transcend even global needs, as humankind considers extended duration space exploration and is faced with stringent life support system requirements in perhaps the most limited resource environment that humans will face (Menezes et al., 2015;Aglietti, 2020). Recent literature highlights the potential of plant-based manufacturing to close human health risk gaps for manned exploration missions (McNulty et al., 2021b).
Initial VIN research has primarily focused on nanomaterial design, considering three plant virion chassis [potato virus X (Uhde-Holzem et al., 2016), bamboo mosaic virus (Kuo et al., 2018), turnip vein clearing virus (TVCV) (Werner et al., 2006)] and several ligand display strategies, including multiple fusion sites on the coat protein, linker inclusions, modulations of ligand display density, and two different immunosorbent ligands (both based on functional fragments of Staphylococcus aureus Protein A). Additional research is needed to evaluate broader functionalities of VIN technology, characteristics for reliable biomanufacturing, and compatibility with advanced multimaterial configurations.
There have been multiple approaches to engineering virusbased nanomaterials into multi-material configurations including layer-by-layer assembled thin biofilms (Tiu et al., 2017), electrospun nanofibers (Shin et al., 2014), and biofunctionalized magnetic particles, to name a few. Within these approaches, bio-functionalized magnetic particles have been distinguished at large as an important platform within biosensing (Zhang and Zhou, 2014), and protein purification (Schwaminger et al., 2019). However, the virus-based nanomaterial research exploring bio-functionalized magnetic particles to date has been limited to gene therapy (Chan et al., 2005;Majidi et al., 2015) and molecular imaging (Huang et al., 2011;Shukla and Steinmetz, 2015). Given the demonstrated ability of virus-based nanomaterials as reagents to enhance target binding and sensitivity over traditional ligands (Sapsford et al., 2006;Soto et al., 2009;Koch et al., 2015) we perceive a general synergy and advantage in developing virus-functionalized magnetic particles for sensing and protein purification.
In this study we present a new vector for production of VINs, develop an optimized purification process for VINs that is generalizable to other plant virus-based nanomaterials, characterize key functional VIN properties, and in the process, identify potential limitations of the VIN methods used to date. In response to identification of these limitations, we present a novel VIN-magnetic particle coupled system to overcome these limitations.
Preliminary results suggest enhanced immunosorbent characteristics as compared to commercial immunosorbent magnetic particle standards and also provides new perspectives for utilization of plant virus-based nanomaterials.

Production of a Plant Virus-Based Immunosorbent Nanoparticle
Intact VINs consisting of an assembled tobamovirus, TVCV, presenting a C-terminal coat protein fusion to a flexible linker domain (GGGGS) 3 coupled to a S. aureus Protein A fragment (domains D and E) were successfully produced in Nicotiana benthamiana plants via agroinfiltration and subsequently purified to a moderate extent ( Figures 1A-D). An illustration of the construct schematic and results of the PCR and DNA sequence verification of the transformation are included in Supplementary Figures S1, S2 and Supplementary Table S1. The vector used here simplifies previously published A. tumefaciens vectors (Werner et al., 2006) by combining multiple provectors into a single vector capable of producing intact VINs.
Agroinfiltrated N. benthamiana plants showed signs of viral infection typical of tobamoviruses (yellowing of leaves, stunted growth; data not shown). The coat protein fusion was expressed at high levels (~0.3 g VIN/kg leaf fresh weight, per total soluble protein results of purified VIN) in N. benthamiana plants collected 6-14 days post-infiltration. Furthermore, transmission electron microscope (TEM) images show that fully assembled virion particles were formed ( Figures 1B,C).
SDS-PAGE results confirm that there is a band at the expected size of the VIN coat protein Fc-affinity ligand fusion (CP-FcAL) (~33.5 kD) and western blot results confirm that it is an identity match for the expected CP-FcAL (via anti-protein A antibody). We did not observe bands corresponding to unfused Fc-affinity ligand on SDS PAGE gels or in western blots, although SDS-PAGE results do present the possibility of a minor presence of CP-FcAL degradation products. The CP-FcAL protein identity was also confirmed using mass spectrometry (Supplementary Figure S3).
In addition to agroinfiltration, we also demonstrated that mechanical transmission using the VINs generated by agroinfiltration is a viable route for production of fully assembled and functional plant virus-based immunosorbent particles.

Fc-Protein Capture and Elution
Protein A is well known to bind strongly with the conserved fragment crystallizable (Fc) region of many species and subclass variants of IgG. We show that VINs retain general immunosorbence for several species and subclasses of IgG (Supplementary Figure S4).
Next, we demonstrate that VINs are capable of capturing and then eluting human immunoglobulin G (hIgG) using a low pH elution mechanism (0.1 M glycine-HCl pH 2.5) (Figures 2A,B). VINs produced using mechanical transmission were also shown to retain immunosorbent functionality (Supplementary Figure S5). Tests using bovine serum albumin (BSA) as the target capture protein confirm that sedimentation of the target protein, and largely that of the VIN, required specific binding interactions ( Figure 2C). It was also observed that VINs would sediment in the absence of binding target proteins at centrifugation of 20,000 x g for 90 min (Supplementary Figure S6).
VINs are also capable of capturing and eluting Fc-fusion proteins ( Figures 2D,E). We successfully tested two prepurified plant-expressed Fc-fusion proteins with VIN-based capture and elution: recombinant capillary morphogenesis protein Fc-fusion (rCMG2-Fc) and recombinant parathyroid hormone Fc-fusion (rPTH-Fc). We observed that the biophysical characteristics (e.g., molecular mass, Svedberg coefficient) of the domain fused to the Fc region is critical to the sedimentation step (denoted III in Figure 2A) performance.
The hIgG sedimentation conditions (12,000 x g, 10 min) were not adequate for rCMG2-Fc (100 kDa) and rPTH-Fc (55 kDa). We determined that 20,000 x g for 20 min was adequate for sedimentation of the rPTH-Fc when bound to VINs (Supplementary Figure S7) and accordingly used this as an adequate centrifugation condition during operation with the larger rCMG2-Fc. Similarly, the smaller sizes of the Fcfusion proteins as compared to the hIgG required a higher PEG concentration (25% w/v) for the PEG-based buffer exchange step (screening data not shown). Further optimization is required to remove residual PEG in this higher concentration method, as can be observed by the PEG interference of electrophoresis (lanes VI in Figures 2D,E), although it has been shown that the presence of PEG does not impede performance of subsequent downstream processing operations including ion exchange and affinity chromatography (Roe, 2001).

Process Characterization
We evaluated process performance from the perspective of nanomaterial stability, capture and elution functionality, and particle integrity. VINs are stable throughout the freeze-thaw process for up to 12 cycles without noticeable degradation of CP-FcAL when stored at either −20°C or −80°C. Long-term stability of VINs was evaluated over a series of timepoints (2, 4, 8 weeks), temperatures (−20°C, 4°C, 20°C), and protease inhibitors [none, 2 mM ethylenediaminetetraacetic acid (EDTA) and 1 mM phenylmethylsulfonyl fluoride (PMSF)] (Supplementary Figure S8). VIN CP-FcAL were intact over the duration evaluated at −20 and 4°C, while the addition of protease inhibitors was shown to prolong stability at 20°C, with no discernable degradation observed for up to 2 weeks (Supplementary Figure S8). Next, we demonstrated that the VIN functionality is retained when using samples of hIgG spiked into wild-type N. benthamiana plant extract ( Figure 3A). Comparable performance was observed when crude N. benthamiana extracts of VIN were used in conjunction with antibodies spiked into crude N. benthamiana extracts (Supplementary Figure S9). Furthermore, we were able to demonstrate that the VINs recovered from a single capture and elution cycle can be reused for an additional cycle ( Figure 3B). A minor fraction of hIgG was recovered with the VIN in both cycles, suggesting that hIgG recovery could be improved by optimization of the elution step (e.g., CP-FcAL binding affinity, buffer composition). The VIN recovered from the second use cycle could not be used for a third cycle with the established sedimentation conditions. Accordingly, VIN particle integrity was investigated to probe limitations of the sedimentation method. We first compared VINs generated by two different methods of plant extraction, a blender or liquid nitrogen-assisted mortar and pestle ( Figures  3C,D). We observed a statistically significant difference in the mean particle length between the two extraction methods (p <0.001), with blender-based extraction resulting in a shorter mean VIN length (blender: X = 217 nm, σ = 84 nm, N = 428; mortar and pestle: X = 239 nm, σ = 80 nm, N = 558). We then investigated the VIN particle lengths at steps throughout the capture and elution with initial VIN generated using liquid nitrogen-assisted mortar and pestle extraction, focusing on the VINs lost in the supernatant during the VIN-hIgG complex sedimentation step ( Figure 3E) and the final  Frontiers in Bioengineering and Biotechnology | www.frontiersin.org April 2022 | Volume 10 | Article 865481 5 recovered VINs ( Figure 3F), denoted III and IV in Figure 2A, respectively. There is an observed statistically significant difference in the mean particle length for the initial VINs, VINs lost in the supernatant ( X = 241 nm, σ = 94 nm, N = 337), and final recovered VINs ( X = 196 nm, σ = 93 nm, N = 486) (I and III, p = 0.039; I and IV, p < 0.001; III and IV, p < 0.001).

Process Development
The appreciable level of impurities present in the purified VIN solutions, as well as an interest in improving scalability of the processing by removing the chloroform-based liquid-liquid extraction step, motivated an investigation into process development of the VIN purification. We performed a 2-factor 2-level process optimization of the extraction step (buffer composition-50 mM sodium acetate 86 mM NaCl pH 5.0, 100 mM potassium phosphate pH 7.0; protease inhibitors-none, 2 mM EDTA + 1 mM PMSF) followed by an addition of a heat hold step that was investigated with a temperature screening (30-70°C) after each processing operation ( Figure 4A).
Results of the process optimization conditions tested, including verification of bind-and-elute functionality of VIN produced from the different purification schemes, can be found in Supplementary Figures S10-S12.
The chloroform-based liquid-liquid extraction step was removed from the processing scheme with comparable or improved VIN recovery and purity upon inclusion of a low pH extraction and 60°C heat hold post-PEG precipitation ( Figure 4B). Interestingly, the addition of protease inhibitors to the extraction buffer reduced VIN recovery and increased the presence of what appears to be degradation products. Similarly distinct from wild-type virion processing, the heat hold resulted in significant VIN loss when introduced at processing steps prior to PEG precipitation. This behavior can be attributed to the FcAL presentation, as wild-type tobacco mosaic (wt-TMV) is routinely processed with early-stage processing heat holds (Smith et al., 2006). Nano differential scanning fluorimetry results (Figures 4C-E) indicate that the VINs prepared according to either protocol detailed in this study exhibit a melting temperature of~82°C, supporting that the improved protocol does not introduce discernible differences in VIN CP-FcAL stability. There are multiple distinct conformational shifts within the crude solution consistent with the heterogeneity of solution. Figure 5A illustrates the basic concept and utility of the VINcoupled magnetic particles (VIN-MPs) generated in this study and Figures 5B,C displays TEM images of the intact VIN-MPs complex. A magnetic separation-based capture and elution method was developed with hIgG spiked into phosphate buffered saline (PBS) (not shown) and crude N. benthamiana plant extract that confirms the VIN immunosorbent functionality in this novel configuration ( Figure 5D). A faint presence in the elution that may indicate that some minor amount of VIN is recovered in addition to the target hIgG.

Magnetic Separation With Virus-Based Immunosorbent Nanoparticle
The VIN-MP were either coupled with amine-terminated (VIN-MP-N) or carboxyl-terminated (VIN-MP-C) superparamagnetic particles. The VIN-MP-N coupled at significantly higher densities (>0.2 mg VIN/mg MP) than the VIN-MP-C (≤0.1 mg VIN/mg MP) and resulted in higher hIgG capture. Therefore, VIN-MP-N were selected as the basis for additional study.
Two different coupling agents were tested: glutaraldehyde (VIN-MP-NG) and 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) (VIN-MP-NE). Furthermore, we screened a range of EDC concentrations for use in the VIN-MP-NE synthesis reaction. We observed that the coupling density of VINs to MPs could be tuned with the concentration of EDC used in the covalent coupling reaction ( Figure 6A). We also observed that the VIN concentration in the reaction medium could be used to tune coupling density. For example, we found that increasing the VIN concentration from 1.42 to 2.10 mg/ml resulted in an increase in the coupling density from 0.274 to 0.352 mg VIN/mg MP when using an EDC concentration of 0.5 mg/ml. We observed a non-monotonic relationship between coupling density and hIgG binding/elution load per mass of MP ( Figure 6B). An optimal coupling density of 0.3-0.4 mg VIN/mg MP was identified. The hIgG binding/elution load per mass of VIN was consistent at approximately 0.3 mg hIgG/mg VIN below a coupling density of 0.4 mg VIN/mg MP, above which a negative correlation between coupling density and  hIgG binding/elution load per mass of VIN is observed ( Figure 6C). This suggests that higher coupling densities may provide less available hIgG binding sites due to steric hinderances or electrostatic interactions. A comparison of VIN capture and elution performance with hIgG in PBS is shown in Figures 6D,E for magnetic separation using VIN-MP and sedimentation separation using uncoupled VIN over a range of hIgG loading densities. The VIN-MP hIgG elution recovery normalized by mass of VIN is approximately 35-55% of that of the uncoupled VIN.
Sedimentation and magnetic separation-based operations were further compared in purification of plant-expressed monoclonal antibody (mAb) A20 from crude N. benthamiana plant extract (Figure 7). Non-reducing condition SDS-PAGE results indicate that fully assembled mAb A20 is produced and recovered by VIN in elution. The relative extents of mAb A20 elution recovery for uncoupled VIN and VIN-MP are consistent with the hIgG in PBS results.

Biotic Purification Technologies
Virus-based nanomaterials present promising characteristics as an alternative technological platform to traditional chemical methods of biopharmaceutical purification in their inexpensive and scalable production coupled with their high replication fidelity, biophysical properties, stability, and accessible modifications leading to wide-ranging functionality. There are safety and regulations concerns of commercializing selfreplicating technology, but this barrier has been addressed by researchers by removing the requisite replication machinery, as is done to generate virus-like particle technology (Zeltins, 2013;Marsian and Lomonossoff, 2016), using a plant virus to avoid human infection (Balke and Zeltins, 2019), designing the virusbased nanomaterial to rapidly shed the transgenic gene inserts (Torti et al., 2021), and/or inactivating the virus (Koudelka et al., 2015). As these methodologies are well established, we do not address virus nanoparticle containment strategies within the scope of VIN process development.
A range of biotic technologies beyond virus-based nanomaterials have been developed and studied for biopharmaceutical purification (Dias and Roque, 2016;Mahmoodi et al., 2019). These can be classified by utility as fusion tags [e.g., inteins (Wood et al., 1999), carbohydrate binding modules (Shoseyov et al., 2006)], thermo-responsive biopolymers [e.g., elastin-like polypeptides (Sheth et al., 2014a)], and hydrophobic nanoparticles [e.g., polyhydroxyalkanoates (Banki et al., 2005), oleosins (Bhatla et al., 2010), hydrophobins (Jugler et al., 2020)]. Fusion tags have by and large been the most widely adopted biotic purification technology for the accessibility they present to early-stage research labs. However, they are limited as a platform technology by the influence of product-specific characteristics, complications of tag cleavage, and generally unfavorable commercial-scale economics (Fc-fusion tags being a notable exception to most of these limitations) (Bell et al., 2013). There has been some adoption and maturation of the other biotic technologies that overcome these limitations, with an observed emphasis on elastin-like polypeptides (ELPs) (Sheth et al., 2014a;Sheth et al., 2014b) and oleosins (McLean et al., 2012). VINs represent another contender within this group of technology, albeit at a more nascent stage of development. Advantages of VIN technology include the simplicity of production (ELP: culturebased system; VIN: plant-based system) and within that the projected high yield per hectare (oleosin: <1 kg/ha; VIN: 200 + kg per hectare) (Werner et al., 2006) that position VINs as inexpensive purification reagents.
Mechanical transmission-based production of VINs, as we have demonstrated, could be considered for economical manufacturing for its enhanced simplicity over agrobacteriumbased methods and its reliability and stability over transcriptbased methods. A main barrier to this strategy is the variability and escape of the VIN functionality over the course of multiple virion replication and plant passage cycles. One strategy to alleviate these concerns would be to embed selective pressures into the processing procedure, although this would introduce yet unsolved barriers within quality assurance and quality control that would hinder commercialization. Studies have shown the use of selective pressures during production can cause viruses to sacrifice reproductive fitness for selected characteristics (e.g., thermal and structural stability) (Dessau et al., 2012) and that genetic stability of virus-based nanomaterials can be achieved (Le Nouën et al., 2017).

Affinity Sedimentation Processing
Affinity sedimentation shares several characteristics with affinity precipitation, which provides potential benefits of low cost and buffer usage (Hilbrig and Freitag, 2003), ability to achieve high concentration factors (Low et al., 2007), high throughput (Shukla and Thömmes, 2010), and minimal concerns of fouling at the expense of higher recovery and selectivity generally achieved by a chromatographic counterpart (Mondal et al., 2006), which can consist of as much as 50% of the total pharmaceutical manufacturing costs (Kelley, 2009). Affinity sedimentation and precipitation methods also exhibit generally higher tolerance to variation in feed streams, as we have shown with VINs through processing of crude plant extract, making them well-suited to early-stage downstream processing.
Biotic purification technologies have been applied in a diverse range of processing strategies including liquid chromatography, inverse transition cycling, aqueous two-phase partitioning, and affinity precipitation. The uncoupled VIN application methodology presented here is based on what we are terming as pseudo-secondary effect affinity sedimentation, in which the affinity interaction and sedimentation mechanisms are partially coupled with a dependence of the sedimentation on affinity interaction (i.e., VIN-target protein complex characteristics influence sedimentation velocity). We derive this terminology from affinity precipitation processing which has distinguished methodologies as either primary effect (coupled affinity/ precipitation) or secondary effect (independent affinity/ precipitation) (Hilbrig and Freitag, 2003).
We performed initial work to identify centrifugation conditions as a function of target protein characteristics and Frontiers in Bioengineering and Biotechnology | www.frontiersin.org April 2022 | Volume 10 | Article 865481 loading (e.g., antibody versus Fc-fusion protein), but it may be valuable in future works to develop a model to understand this relationship more deeply between the VIN-target protein complex morphology and sedimentation velocity. It was qualitatively observed that Fc-fusion protein recovery was lower than that of hIgG regardless of centrifugation conditions. We hypothesize that lower binding affinities and differing biophysical characteristics of the VIN-Fc-fusion protein complex are contributing to these observed differences. However, future work is required to elucidate these underlying mechanisms.
While protein solids, such as those formed in precipitation and sedimentation, have been shown to be stable in long term storage (Harrison et al., 2015), here we show that sedimentation impacts VIN structural integrity and presumably contributes to the unstable performance over multiple reuse cycles. From these results, one may infer the importance of fully intact VIN on sedimentation characteristics and thus centrifugal recovery. The mechanism of particle breakage is suspected to be mechanically induced during pellet resuspension and not a direct result of the protein pellet formation from sedimentation. The mechanical properties of viruses have been studied extensively using computational and physical methods (Mateu, 2012;Buzón et al., 2020). TMV, in the same genus as the TVCV used for the VIN in this study, has been attributed a Young's modulus of 6 ± 3 GPa (Schmatulla et al., 2007), although mixed reports suggest the value could be lower (Falvo et al., 1997). Furthermore, there are reports on icosahedral virions that demonstrate minor changes in coat protein composition resulting in significant modulation of mechanical properties (Medrano et al., 2019).
Thus, we hypothesize that the presumably stiff TVCV rod-like particle basis combined with increased drag by the loose affinity ligand display sheath (using the largest genetically inserted virion coat protein presentation to date) surrounding the VIN enhances its vulnerability to applied shear stress during pellet resuspension via repeated pipette tip aspiration. It may be that a smaller affinity ligand presentation, such as an affibody (Frejd and Kim, 2017) or synthetic peptide (Lund et al., 2012), would decrease particle breakage (but also necessitate more aggressive centrifugation conditions with the smaller virion size). Gentler resuspension methods with lower shear would also be worth investigating to decrease breakage. Similar shear sensitivity in pipette-based resuspension has been shown for larger biomolecule systems such as with cell-based pelleting, wherein higher pellet compaction and tip velocities were shown to result in cell losses (Delahaye et al., 2015). Pellets formed during VIN processing and use were highly compacted and required considerable pipetting for complete resuspension, suggesting that pipette-based resuspension could have also played a role of particle degradation in this system. Gentler resuspension techniques should be explored in the future to improve particle integrity during operation. However, the simple and inexpensive production of VIN technology is amenable to single-use application, in which the VIN reagent is disposed after only a single cycle of operation. In this case, the question of how to improve particle integrity during operation beyond ensuring purity of the final recovered antibody product may be lower priority.
In this study, we investigated stability of the VINs over long term storage, multiple freeze-thaw cycles, and at elevated temperatures, primarily focusing on coat protein fusion primary structure. Additional stability concerns include impact to protein secondary structure, virion particle structure, and nucleic acid integrity. Exposure to multiple freeze-thaw cycles has been shown to degrade virion nucleic acid and infectivity (Krajden et al., 1999). Thermostability was confirmed at the level of secondary protein structure via performance check of bindand-elute functionality, which led to the integration of a high heat hold into the VIN purification scheme.

Virus-Based Immunosorbent Nanoparticle Process Development
The process development investigation presented here provides insights on the differences between wild-type plant virions and protein display plant virions, which is relevant technology for a host of biomedical applications. The improved process can serve as a roadmap for future virus-based nanomaterial purification. We observed that the protein presentation confers additional processing sensitivities to the virion (which is otherwise described as a glassy surface), likely due to the interactions between the presented ligand and N. benthamiana plant host cell impurities, noticeable in the heat hold of crude solution and the inclusion of protease inhibitors to extraction.
We observed that the two effective unit procedure modifications, low pH extraction and heat hold post-PEG precipitation, possess low orthogonality in impurity clearance mechanisms for the starting stream used although there is still discernable benefit in combining the methods as observed by the improvement in SDS-PAGE band purity.
The process development in this work focused on addition and removal of unit procedures at a high-level to inform process design. There is value in future research performing parameter optimization with an emphasis on maximizing recovery and purity with a fixed purification scheme. More rigorous analytical investigation of binding affinity (e.g., using isothermal calorimetry) and solution purity (e.g., using highperformance liquid chromatography) are necessary for this future optimization and technological maturation.

Magnetic Separation
The VIN-MP results presented in this study, representing the first virus-based nanomaterial system coupled with MPs for protein purification, reflect a greater than 25x increase in binding capacity compared to current industry standards for affinity protein capture with magnetic particles-Pierce ™ Protein A Magnetic Beads (≥40 µg rabbit IgG/mg MP) (ThermoFisher Scientific), SureBeads ™ Protein A Magnetic Beads (≥6 µg IgG/mg MP) (Bio-Rad Laboratories), VIN-MP (>1,000 µg hIgG/mg MP).
These exciting results provide strong evidence for the sensitivity-enhancing properties of virus-based nanomaterials and their usefulness as ligand scaffolding in biotechnological applications.
Frontiers in Bioengineering and Biotechnology | www.frontiersin.org April 2022 | Volume 10 | Article 865481 Furthermore, the VIN-MP system served to decouple the affinity and separation mechanisms of processing, as compared to the partially coupled behavior in VIN sedimentation operation, thereby increasing the process robustness to changes in the sample solution, including the diversity and concentration of the target protein. Additional investigation is required to experimentally assess this capability in a larger set of processing conditions. Investigation of the reusability of VIN-MP and for magnetic separation and particle integrity over operation is also of importance for future testing. Preliminary results suggest there may be a minor presence of VIN in the eluate. Two possible means of explaining this observation are proteolytic cleavage along a covalently bonded FcAL resulting in detachment of the VIN from MP or minor particle breakage from resuspension of the VIN-MP after magnetic separation resulting in the presence of VIN fragments.
The uncoupled VIN sedimentation operation demonstrated 2-3 times higher capture capacity per VIN mass than the VIN-MP system. The uncoupled VIN operation also resulted in hIgG recoveries as high as 95% of the feed, whereas VIN-MP operation was maximal at 42% recovery of the feed hIgG ( Figure 6). It will be important to identify the cause for the lower recovery in future development by additional screening of lower hIgG concentrations and optimization of the capture and elution methodology.
Higher capture capacity may be particularly advantageous for large-scale protein purification-an area for which affinity precipitation (Swartz et al., 2018) and magnetic separation (Schwaminger et al., 2019) are receiving growing interest. However, given the nature of VINs as inexpensive and simply produced reagents, the value of maximizing Fcprotein binding per VIN in a small-scale commercial application (as is the current niche of magnetic particle purification) is most likely weighted less than factors such as Fc-protein recovery, process duration, labor time, amenability to automation, and equipment costs.
For perspective on these other factors, consider that our recently published study on evaluating the costs of the affinity capture step of mAb purification (McNulty et al., 2021a) yielded results that the unit costs of magnetic separation (modeled using an industry standard technology with a comparable capture and elution protocol) were lower than uncoupled VIN sedimentation in process duration (73% reduction), labor time (30% reduction), and equipment mass (71% reduction) when processing a single lab-scale sample (2 ml volume tube). These results support the favorable position of VIN-MP in comparison to uncoupled VIN sedimentation for lab-scale applications.

Summary and Future Directions
Virus-based nanomaterials provide a highly diverse and tunable technology that can be adapted to overcome the limitations of the application methodology. For example, the length of a rod-like plant virus such as the one used in this study, TVCV, is proportional to the length of the viral genomic information and, as such, the length can be modulated through the addition or subtraction of genomic information (e.g., addition of non-functional genomic information can be used to increase virion particle length) (Saunders and Lomonossoff, 2017). Increasing VIN length in this manner is one approach to investigate for increasing the binding site occupation in the VIN-MP system. Other techniques useful for VIN performance optimization include density modulation of the protein display (Cruz et al., 1996) and a multi-ligand protein display (Werner et al., 2006). Last not least, use of affinity ligands other than Protein A domains, in particular, affibodies evolved from Protein A/Z domains (Nord et al., 1997), should expand the usability of the technology beyond monoclonal antibody binding/ capture (Ståhl et al., 2017).
In this study, we have presented the development of a virusbased nanomaterial to serve as a protein purification reagent, characterized performance using a pseudo-secondary effect affinity sedimentation bind-and-elute protocol, expanded functionality to Fc-fusion proteins, identified limitations of the technology operated in that procedure, and developed a magnetic particle coupled system for magnetic separation to improve processing. This provides further evidence supporting virusbased nanomaterials as simple and inexpensive reagents for protein purification and suggests a path forward for technological development.

Gene Constructs
The viral expression vector used in this study is based on previously reported TVCV-based vectors (Werner et al., 2006). The viral expression vector used here (pICH25892; plasmid kindly provided by Nomad Biosciences GmbH) is an assembly of the previously reported 5′ provector containing the TVCV coat protein (minus the stop codon) fused to a C-terminal glycine-rich flexible linker (pICH20701) and the 3' provector containing the D and E antibody-binding domains from S. aureus protein A with short flanking sequences (amino acids 29-161; GenBank accession no. J01786) (pICH21767).

Production of Virus-Based Immunosorbent Nanoparticle
VINs were primarily produced via whole plant agroinfiltration using A. tumefaciens containing viral expression vector pICH25892 according to a previously reported method with minor modifications (Xiong et al., 2018). A final cell density of OD 600 = 0.2 was used for agroinfiltration. Post-infiltration plants were cultivated at 60% relative humidity with a 16 h photoperiod, 23°C/20°C temperature regime, and a photosynthetic photon flux density of 425 µmol/(m 2 ·s) derived from a combination of high-pressure sodium, high-pressure metal halide, and incandescent lights for a duration of 6-12 days post-inoculation.
VINs were also produced via direct mechanical transmission of intact VINs. A total volume of 300 µl of purified VIN solution (~0.1 mg/ml) was applied per plant in Frontiers in Bioengineering and Biotechnology | www.frontiersin.org April 2022 | Volume 10 | Article 865481 aliquots of 100 µl for each of three middling leaves. An abrasive powder (Celite) was lightly sprinkled on each leaf and each leaf was gently rubbed by hand. The surfaces of the leaves were rinsed with water at 20 min post-inoculation to remove excess inoculation reagents.

Purification of Virus-Based Immunosorbent Nanoparticle
VIN-expressing N. benthamiana leaf tissue was stored at −80°C after harvest and processed with minor modifications to a previously reported protocol (Werner et al., 2006). A single round of PEG-assisted precipitation step was performed rather than two. Extraction was performed using either a blender (NutriBullet; NutriBullet, LLC, Pacoima, CA) or liquid nitrogen-assisted mortar and pestle with 0.1 M potassium phosphate pH 7.0 extraction buffer at a 3:1 buffer volume to biomass weight extraction ratio. In the case of the mortar and pestle method, the homogenized leaf powder was mixed with the buffer and nutated for 30 min at 4°C for extraction.

Binding and Elution of Fc-Proteins
Binding and elution of Fc-proteins was performed according to a previously reported protocol (Werner et al., 2006

Binding and Elution of Fc-Proteins With Virus-Based Immunosorbent Nanoparticle-Magnetic Particles
A VIN-MP solution was prepared by resuspending 2.5 mg of VIN-MP in 0.5 ml of 0.1 M sodium phosphate buffer pH 8.0 binding buffer. Further, the particles were magnetically separated; the supernatant was aspirated and discarded (repeated three times).
Crude protein extracts from N. benthamiana leaves (leaf juice press extraction followed by microfiltration with filter paper) in binding buffer were spiked with various amounts of hIgG. A volume of 100 µl hIgG-containing crude extract was added to the VIN-MP solution. The mixture was briefly vortexed and then incubated nutating at 4°C for 30 min. The incubated solution was then magnetically separated and washed three times with binding buffer. The binding buffer was removed after wash and 50 µl of 0.2 M glycine buffer pH 2.5 elution buffer was added. The solution was briefly vortexed to resuspend magnetic particles and further incubated nutating at 4°C for 5 min. Particles were again magnetically separated and the supernatant was collected as the eluate. The eluate was then pH neutralized with 13 µl of 1.5 M Tris-HCl buffer pH 8.8. The elution and neutralization steps were repeated three times and pooled together.

Protein Analysis
Protein concentration was measured using Bradford and Pierce Modified Lowry assays.
Sample protein compositions were analyzed by SDS-PAGE and western blot. SDS-PAGE samples were loaded using constant volume (30 µl). Western blot analysis was performed using a primary antibody of rabbit anti-protein A (1:25,000 dilution) (Sigma-Aldrich, St. Louis, MO) and a secondary antibody of goat anti-rabbit IgG-HRP (1:3,000 dilution) (Southern Biotech, Birmingham, AL) for detection of VIN CP-FcAL. A secondary antibody of goat anti-human IgG-HRP (1:2,500 dilution) was used to detect human IgG, rCMG2-Fc, and rPTH-Fc.

Electron Microscopy
Carbon film on 300 mesh copper discs (Ted Pella, Redding, CA, United States) were prepared for increased hydrophilicity by glow discharge at 30 mA for 30 s on a glass slide. 5 µl liquid VIN solution samples were loaded onto the prepared disc, incubated 30 s, and then blotted with filter paper. Negative stain was applied in five sequential rounds of 5 µl uranyl sulfate loading, 30 s incubation, and filter paper blotting. TEM was performed using a JEM-1230 transmission electron microscope (JEOL, Peabody, MA, United States).
The lengths of VIN particles imaged by TEM were manually measured using straight line analysis with ImageJ (National Institutes of Health, Bethesda, MD, United States). Statistically significant differences in mean particle lengths of different VIN solutions were determined by equal variance two sample t-test (α = 0.05). The equal variance assumption was evaluated by two sample F-test (α = 0.05).

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusion of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
MM, AS, AG, YG, and KM designed and executed the experiments. All authors were involved in analysis and interpretation of data. MM wrote the initial manuscript draft. KM edited the manuscript draft. AG, YG, SN, and KM supervised the project.

FUNDING
This material is based upon work supported by the NASA under grant or cooperative agreement award number NNX17AJ31G. This work was also supported by a NASA Space Technology Research Fellowship (NASA grant number 80NSSC18K1157). This work is supported by the Translational Research Institute through NASA Cooperative Agreement NNX16AO69A. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Aeronautics and Space Administration (NASA) or the Translational Research Institute for Space Health (TRISH).