Hydrophobin-Protein a Fusion Protein Produced in Plants Efficiently

International Journal of Molecular Sciences

Article Hydrophobin-Protein A Fusion Protein Produced in Plants Eﬃciently Puriﬁed an Anti-West Nile Virus Monoclonal Antibody from Plant Extracts via Aqueous Two-Phase Separation

Collin Jugler 1, Jussi Joensuu 2 and Qiang Chen 1,*

1 The Biodesign Institute and School of Life Sciences, Arizona State University, Mail Zone 5401, 1001 S. McAllister Avenue, Tempe, AZ 85287, USA; [email protected] 2 VTT Technical Research Centre of Finland Ltd, Espoo, Finland; jussi.joensuu@iki.ﬁ * Correspondence: [email protected]; Tel.: +(480)-239-7802

Received: 2 March 2020; Accepted: 18 March 2020; Published: 20 March 2020

Abstract: The development of monoclonal antibodies (mAbs) has provided vast opportunities to treat a wide range of diseases from cancer to viral infections. While plant-based production of mAbs has effectively lowered the upstream cost of mAb production compared to mammalian cell cultures, further optimization of downstream processing, especially in extending the longevity of Protein A resin by an effective bulk separation step, will further reduce the overall prohibitive cost of mAb production. In this study, we explored the feasibility of using aqueous two-phase separation (ATPS) in capturing and separating plant-made mAbs from host proteins. Our results demonstrated that an anti-West Nile virus mAb (E16) was efficiently separated from most plant host proteins by a single ATPS step, comprising the mixing of plant extracts containing Hydrophobin-Protein A fusion protein (HPA) and E16 and the subsequent incubation with an inexpensive detergent. This simple ATPS step yielded a highly enriched E16 mAb preparation with a recovery rate comparable to that of Protein A chromatography. The ATPS-enriched E16 retained its structural integrity and was fully functional in binding its target antigen. Notably, HPA-based ATPS was also effective in enriching E16 from plant host proteins when both HPA and E16 were produced in the same leaves, supporting the potential of further streamlining the downstream purification process. Thus, ATPS based on plant-produced HPA in unpurified extract is a cost-effective yet efficient initial capture step for purifying plant-made mAbs, which may significantly impact the approach of mAb purification.

Keywords: monoclonal antibody (mAb); aqueous two-phase separation (ATPS); hydrophobin; West Nile virus; plant-made antibody

1. Introduction Monoclonal antibodies (mAbs) have dramatically transformed how we approach the treatment of a myriad of diseases [1]. Examples of mAb-therapeutics include cetuximab to treat colorectal cancer [2], avelumab to treat a variety of tumors [3], and the combination of three anti-Ebola mAbs, known as ZMapp, to treat Ebola virus infection [4]. However, the high production cost of mAbs using mammalian cell-based platforms translates into a high market price for approved treatments. Plant-made mAbs have shown great potential in reducing the upstream costs of producing these valuable biologics by eliminating the need for expensive facilities, bioreactors, and culture media required for mammalian cell-based protein production [5]. Although the upstream processing costs can be addressed through plant-based expression platforms [6], the downstream processing remains essentially equivalent to the

Int. J. Mol. Sci. 2020, 21, 2140; doi:10.3390/ijms21062140 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, 2140 2 of 12 mammalian cell-produced counterparts [7] and improvements or alternatives to the current methods could help to further reduce the overall production costs. Although Protein A affinity chromatography is a universally accepted standard purification step in mAb production, the resins for this chromatographic step are expensive and contribute significantly to the overall cost-of-goods of mAb-based drugs [8]. As a result, it is common practice to recycle and reuse the Protein A resin for mAb purification which can reduce column efficiency, as well as result in a reduction in the purity and overall recovery of the antibody [9,10]. The reduction of column efficiency is often attributed to ligand leaching, impurity build-up, restriction of pore access, and blocking access to the Protein A ligand itself [11]. Particularly relevant to plant-made mAbs, studies have shown that the presence of native plant proteins and small-molecule compounds, such as phenolics and alkaloids, in the raw plant extract can impact mAb binding to Protein A in addition to contributing to resin fouling [12–14]. Although the presence of these contaminating proteins and compounds can be reduced or eliminated by combining multiple and varying chromatography steps, prolonging the longevity of the Protein A resin of the initial capture step remains a challenge. This calls for the development of alternative strategies that can effectively enrich mAbs from plant extracts at a low cost. Aqueous two-phase separation (ATPS) has been a useful method for protein purification by taking advantage of unique characteristics of structurally different polymers to separate biomaterials [15]. In particular, thermo-separating surfactants have been used with hydrophobin tags to purify target proteins from fungi, plants, and insect cells [16–19]. Recently, it was demonstrated that a novel bifunctional fusion protein could be utilized to capture antibodies in-solution by way of a water-surfactant two-phase system. The bifunctional protein was a genetic fusion of the immunoglobulin-binding domain of Protein A from Staphylococcus aureus with a class II hydrophobin, HFBI, from Trichoderma reesei [20]. When the bifunctional fusion protein (HPA) was mixed with a mAb in solution, the mAb-HPA complex partitioned to the surfactant-rich phase due to the surfactant-like behavior induced by the hydrophobin moiety of the fusion protein. Upon removal of the residual aqueous phase, where the majority of the production host proteins reside, and after the addition of a low pH buffer to dissociate the mAb from the Protein A binding domain, the mAb was enriched and efficiently separated from most production host proteins [20]. While this strategy was successful in purifying mAbs from media of hybridoma culture, its feasibility in enriching mAbs from systems with more complex mixtures of proteins and other compounds, such as plant extract, has not been demonstrated. In the current study, we utilized the same bifunctional protein in a water-surfactant two-phase system to serve as an initial capture step in purifying plant-made mAbs. To achieve this, we transiently expressed two protein molecules separately in Nicotiana benthamiana: an anti-West Nile virus (WNV) mAb, E16, and the bifunctional fusion protein (HPA) described above, containing the antibody-binding domain of Protein A and the hydrophobin tag. When plant extracts containing E16 and HPA were mixed, E16 was efficiently separated from the majority of plant host proteins by a single ATPS step, yielding a highly enriched mAb preparation with the expected E16 function. Additionally, we explored the feasibility of co-expressing E16 and HPA in the same plant to further streamline the initial capture step. Co-expression of three polypeptides of E16 light chain (LC), heavy chain (HC) and HPA in the same plant produced two properly folded, functional proteins that interact with each other to achieve efficient separation of E16 from the host molecules. ATPS based on plant-produced HPA in unpurified extract has potential to change the way that plant-made mAb purification is approached.

2. Results and Discussion

2.1. Transient Expression of HPA and E16 mAb in N. benthamiana Transient expression of the HPA fusion protein in N. benthamiana plants was assessed with an enzyme-linked immunosorbent (ELISA) assay. The temporal expression pattern of HPA is similar to that of other proteins using MagnICON vectors [21], increasing up until 8–10 days post infiltration (DPI), where the protein level peaked at 218.7–272.9 µg/g FLW (Figure1). Previous work with a Int. J. Mol. Sci. 2020, 21, 2140 3 of 12 Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 3 of 12 diproteinfferent protein construct expression [20]. In vector addition reported to using significantly a different higher ex levelspression of expression vector, othe of ther factors same proteinsuch as constructdifferences [20 ].in In plant addition growth to usingand agroinfiltration a different expression conditions vector, may other also factorscontribute such to as the di ffdifferenceerences in in plantHPA growth accumulation and agroinfiltration in leaves. While conditions optimizing may also the contributeexpression to level the diofff thiserence protein in HPA in the accumulation MagnICON insystem leaves. is While required optimizing in future the work, expression our current level of expression this protein system in the did MagnICON provide systema sufficient is required amount in of futureHPA work,fusion our protein current for expression analyzing system its efficiency did provide in apurifying sufficient a amount plant-made of HPA mAb. fusion E16 protein was foralso analyzingtransiently its eexpressedfficiency in in purifying leaves of a N. plant-made benthamiana mAb. and E16 its was expression also transiently levels were expressed quantitated in leaves as ofwe N.reported benthamiana previouslyand its [22]. expression levels were quantitated as we reported previously [22].

Figure 1. Temporal expression pattern of Hydrogen-Protein A (HPA) fusion protein in N. benthamiana plants.Figure Leaves 1. Temporal were infiltrated expression with pattern the HPA of Hydrog constructen-Protein in the TMV-based A (HPA) fusion MagnICON protein systemin N. benthamiana and leaf materialplants. wasLeaves collected were infiltrated 5–12 days with post the infiltration HPA cons (DPI).truct Leafin the protein TMV-based extracts MagnICON were analyzed system with and an leaf ELISAmaterial assay was to collected detect HPA. 5–12 At days least post three infiltration independent (DPI). experiments Leaf protein are extracts represented were analyzed and error with bars an representELISA assay SEM. to detect HPA. At least three independent experiments are represented and error bars 2.2. Purificationrepresent ofSEM. mAb from Plant Extract via ATPS

2.2.We Purification evaluated of themAb eff frectivenessom Plant Extract of ATPS via inATPS purifying plant-made mAb with plant extracts from leaves harvested at 7 DPI. This sampling point was chosen because the levels of expression were high for We evaluated the effectiveness of ATPS in purifying plant-made mAb with plant extracts from both HPA (155.2 µg/g FLW) and E16 (339.9 µg/g FLW). A clarified extract from HPA and E16-expressing leaves harvested at 7 DPI. This sampling point was chosen because the levels of expression were high leaves was mixed volumetrically 1:1 and processed through the ATPS steps. As a reference, an aliquot for both HPA (155.2 μg/g FLW) and E16 (339.9 μg/g FLW). A clarified extract from HPA and E16- of the same E16 extract was processed in parallel in Protein A affinity chromatography. As shown expressing leaves was mixed volumetrically 1:1 and processed through the ATPS steps. As a in Figure2, ATPS significantly enriched E16, with HC and LC being the prominent protein bands reference, an aliquot of the same E16 extract was processed in parallel in Protein A affinity on the SDS-PAGE. Compared to Protein A chromatography-purified samples, a few more minor chromatography. As shown in Figure 2, ATPS significantly enriched E16, with HC and LC being the non-mAb bands were observed in the lane containing the ATPS-enriched E16 sample (Figure2, Lane 3 prominent protein bands on the SDS-PAGE. Compared to Protein A chromatography-purified compared to Lane 6). Specifically, densitometry analysis showed that the purity of ATPS-enriched E16 samples, a few more minor non-mAb bands were observed in the lane containing the ATPS-enriched is ~24% less than that of E16 purified by a two-step process of low pH precipitation and Protein A E16 sample (Figure 2, Lane 3 compared to Lane 6). Specifically, densitometry analysis showed that chromatography [22]. Nevertheless, this single ATPS step did remove most of the plant host proteins the purity of ATPS-enriched E16 is ~24% less than that of E16 purified by a two-step process of low from the E16 purification feedstream. The recovery rate of E16 through ATPS was 47.7% 5.2%, pH precipitation and Protein A chromatography [22]. Nevertheless, this single ATPS step did± remove as measured by ELISA, comparable to that of Protein A chromatography [22] but higher than that most of the plant host proteins from the E16 purification feedstream. The recovery rate of E16 through reported for ATPS purification of mAbs from hybridoma supernatants (24%) [20]. According to ATPS was 47.7% ± 5.2%, as measured by ELISA, comparable to that of Protein A chromatography our expression data, the molar ratio of HPA:mAb used in ATPS to achieve this recovery rate was [22] but higher than that reported for ATPS purification of mAbs from hybridoma supernatants (24%) approximately 2:1 (1:2 mass ratio). Thus, this favorable ratio [20] may have contributed to the high [20]. According to our expression data, the molar ratio of HPA:mAb used in ATPS to achieve this recovery of E16. Compared with Protein A chromatography, ATPS is simpler and is less expensive. recovery rate was approximately 2:1 (1:2 mass ratio). Thus, this favorable ratio [20] may have For example, the crude clarified leaf extract in Lane 2 (Figure2) can be directly used in ATPS for E16 contributed to the high recovery of E16. Compared with Protein A chromatography, ATPS is simpler purification. In contrast, the crude extract had to go through a bulk separation step (overnight low-pH and is less expensive. For example, the crude clarified leaf extract in Lane 2 (Figure 2) can be directly precipitation) to remove host proteins (i.e., Rubisco) and other contaminants to yield the starting used in ATPS for E16 purification. In contrast, the crude extract had to go through a bulk separation extract for Protein A chromatography (Figure2, Lane 4). This bulk separation step is necessary as the step (overnight low-pH precipitation) to remove host proteins (i.e., Rubisco) and other contaminants removal of most abundant plant host protein (Rubisco) (evidenced by the reduction of the ~ 50 kDa to yield the starting extract for Protein A chromatography (Figure 2, Lane 4). This bulk separation band intensity from Lane 1 to Lane 4) and other impurities are essential to maintain the efficiency step is necessary as the removal of most abundant plant host protein (Rubisco) (evidenced by the reduction of the ~ 50 kDa band intensity from Lane 1 to Lane 4) and other impurities are essential to

Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 4 of 12 Int.Int. J. J. Mol. Mol. Sci. Sci. 2019 2019, ,20 20, ,x x FOR FOR PEER PEER REVIEW REVIEW 44 of of 12 12 maintain the efficiency of Protein A chromatography [23]. By eliminating the bulk separation step, maintainmaintain thethe efficiencyefficiency ofof ProteinProtein AA chromatographychromatography [23].[23]. ByBy eliminatingeliminating thethe bulkbulk separationseparationATPS step,step, saves time and resources, and potentially decreases the loss of mAb due to proteolytic ATPSATPS savessaves timetime andand resources,resources, andand potentiallypotentially decreasesdecreases thethe lossloss ofof mAbmAb duedue toto proteolyticproteolyticdegradation by reducing the time it is exposed to plant host proteolytic enzymes upon degradationdegradation byby reducingreducing thethe timetime itit isis exposedexposed toto plantplant hosthost proteolyticproteolytic enzymesenzymeshomogenization uponupon [24]. The ATPS strategy reported here also has advantages over other ATPS methods homogenizationhomogenizationInt. J. Mol. Sci. 2020 [24]. [24]., 21 The ,The 2140 ATPS ATPS strategy strategy reported reported here here also also has has advantages advantages over over other other ATPS ATPS methods methodsreported4 of previously 12 [20]. Instead of using purified HPA, the crude plant extract containing HPA was reportedreported previouslypreviously [20].[20]. InsteadInstead ofof usingusing purifiedpurified HPA,HPA, thethe crudecrude plantplant extractextract containingcontaining HPAHPAdirectly waswas mixed with E16-containing extract in our process, eliminating extra steps for HPA Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 4 ofpurification. 12 This time and cost-saving improvement did not compromise the purity of the resulted directlydirectlyof Protein mixedmixed A with chromatographywith E16-containingE16-containing [23]. extractextract By eliminating inin ourour process, theprocess, bulk eliminat separationeliminatinging step, extraextra ATPS stepssteps saves forfor timeHPAHPA and purification.purification. ThisThis timetime andand cost-savingcost-saving improvemenimprovementt diddid notnot compromisecompromise thethe puritypurity ofof thethe resultedmAb,resulted but actually yielded better recovery of mAb (47.7%) from the more complex protein containing resources,maintain and potentially the efficiency decreases of Protein the A chromatography loss of mAb due [23]. toBy proteolyticeliminating the degradation bulk separation by step, reducing the plant crude extract than the more cumbersome method (24%) from the hybridoma supernatant [20]. mAb,mAb,time butbut it actuallyactually is exposedATPS yieldedyieldedsaves to planttime betterbetter and host recoveryresources,recovery proteolytic ofandof mAbmAb potentially enzymes (4(47.7%)7.7%) decreases upon fromfrom homogenization thethethe moremoreloss of complex complexmAb due [24 protein protein].to Theproteolytic ATPScontainingcontaining strategy The efficient enrichment and the remarkable rate of recovery supports our ATPS method as a simple, plantplantreported crudecrude extract hereextractdegradation also thanthan has thebythe advantages morereducingmore cumbersomecumbersome the over time other it meis ATPSme exposedthodthod methods (24%)(24%) to plant fromfrom reported host thethe hybridomaproteolytichybridoma previously enzymes [supernatantsupernatant20]. Instead upon [20].of[20]. using TheThe efficientefficient enrichment enrichmenthomogenization andand [24]. thethe The remarkableremarkable ATPS strategy raterate reported ofof recoveryrecovery here also supportshassupports advantages ourour over ATAT PSotherPS methodmethod ATPS methods asas a acost simple,simple, effective, yet efficient bulk separation step to enrich plant-made mAbs from most of the host purified HPA,reported the previously crude plant [20]. extract Instead of containing using purified HPA HPA, was the directlycrude plant mixed extract with containing E16-containing HPA was extract costcost effective,effective, yetyet efficientefficient bulkbulk separationseparation stepstep toto enrichenrich plant-madeplant-made mAbsmAbs fromfrom mostmost ofof theplantthe hosthost proteins. in our process,directly eliminating mixed with extraE16-containing steps for extract HPA purification.in our process, This eliminat timeing and extra cost-saving steps for HPA improvement plantplantdid proteins.proteins. not compromisepurification. This the time purity and cost-saving of the resulted improvemen mAb,t did but not actually compromise yielded the purity better of the recovery resulted of mAb (47.7%) frommAb, the but more actually complex yielded better protein recovery containing of mAb (47.7%) plant from crude the more extract complex than protein the more containing cumbersome plant crude extract than the more cumbersome method (24%) from the hybridoma supernatant [20]. method (24%)The efficient from the enrichment hybridoma and the supernatant remarkable rate [20 of]. recovery The effi supportscient enrichment our ATPS method and theas a simple, remarkable rate of recoverycost supports effective, our yet efficient ATPS method bulk separation as a simple, step to costenrich e ffplant-madeective, yet mAbs effi cientfrom most bulk of separation the host step to enrich plant-madeplant proteins. mAbs from most of the host plant proteins.

Figure 2. Aqueous two-phase separation (ATPS)-based purification of E16 mAb from N. benthamiana

FigureFigureFigure 2.2. AqueousAqueous 2. Aqueous two-phasetwo-phase two-phase separationseparation separation (ATPS)-b(ATPS)-b (ATPS)-basedasedased purificationpurification purification ofof E16E16 of E16mAbmAb mAb fromfrom from N.N. benthamianabenthamianaN. benthamiana leaves. Total soluble proteins were extracted from plants and E16 was purified by ATPS (Lanes 1–3) leaves.leaves.leaves. TotalTotal Total solublesolubleFigure soluble 2. proteinsproteins Aqueous proteins two-phasewerewere were extractedextracted separation extracted fromfrom (ATPS)-b from plantsplantsased plants andand purification andE16E16 E16waswas of was purifiE16purifi mAb purifiededed from byby ATPSN.ATPS by benthamiana ATPS (Lanes(Lanes (Lanes 1–3)1–3) 1–3) or Protein A affinity chromatograph (Lanes 4-6), analyzed on a 12% SDS-PAGE gel under reducing leaves. Total soluble proteins were extracted from plants and E16 was purified by ATPS (Lanes 1–3) oror ProteinProteinor Protein AA affinityaffinity A affinity chromatographchromatograph chromatograph (Lanes(Lanes (Lanes 4-6),4-6), 4-6), analyzedanalyzed analyzed onon aa on 12%12% a 12% SDS-PAGESDS-PAGE SDS-PAGE gelgel under gelunder under reducingreducing reducing conditions, and visualized with Coomassie stain. Lane 1, total leaf soluble proteins; Lane 2, proteins or Protein A affinity chromatograph (Lanes 4-6), analyzed on a 12% SDS-PAGE gel under reducing conditions, and visualized with Coomassie stain. Lane 1, total leaf soluble proteins; Lane 2, proteinsin in the upper aqueous phase of ATPS; Lane 3, ATPS-enriched E16; Lane 4, leaf soluble proteins after conditions,conditions, andandconditions, visualizedvisualized and withvisualizedwith CoomassieCoomassie with Coomassie stain.stain. stain. LaneLane Lane 1,1, total total1, total leafleaf leaf soluble solublesoluble proteins; proteins;proteins; Lane LaneLane 2, proteins 2,2, proteinsproteins inin thethethe upperupper upper aqueousaqueous aqueousin the upper phasephase phase aqueous ofof of ATPS; ATPS;phase ATPS; of LaneLane LaneATPS; 3,3, Lane ATPS-e ATPS-enrichedATPS-e 3, ATPS-enrichednrichednriched E16; E16; E16; Lane LaneLane 4,4, leafleaf leaf soluble solublesoluble proteins proteinsproteins after after afterafter low low pH precipitation; Lane 5, Protein A flow through; Lane 6, Protein A column-purified E16. : HC, lowlow pHpH pH precipitation; precipitation;low pH precipitation; Lane Lane 5,5, Protein Protein Lane 5, A AProtein flow flowflow Athro through;thro flowugh;ugh; thro ugh;Lane Lane Lane 6, 6, Protein Protein6, Protein A A co co column-purifiedlumn-purifiedlumn-purifiedlumn-purified E16. E16. E16. : HC, : ::HC, HC, HC, :: LC, : minor impurity, non-mAb bands. : LC, : minor impurity, non-mAb bands. :: LC,LC,LC, :: : minor minorminor impurity, impurity,impurity, non-mAbnon-mAb bands.bands. 2.3. Characterization of ATPS-Purified E16 mAb 2.3. Characterization2.3. Characterization of ATPS-Purified of ATPS-Purified E16 E16 mAb mAb 2.3.2.3. CharacterizationCharacterization ofof ATPS-PurifiedATPS-Purified E16E16 mAbmAb Western blot analysis was used to characterize E16 recovered from ATPS. When the blot was Western blot analysis was used to characterize E16 recovered from ATPS. When the blot was Western blot analysis was used to characterize E16 recovered from ATPS. When the blot was WesternWestern blotprobedblot analysisanalysis with an anti-human waswas usedused gamma-chain toto characterizecharacterize antibody, E16E16 we recoveredrecovered observed the fromfrom expected ATPS.ATPS. size When Whenof HC (50 thethe kDa) blotblotprobed waswas with an anti-human gamma-chain antibody, we observed the expected size of HC (50 kDa) probed withsimilar an anti-humanto that of Protein-A-purified gamma-chain E16 under antibody, reducing we conditions observed (Figure the expected 3a), confirming size that of HCthe (50 kDa) probedprobed withwith anan anti-humananti-human gamma-chaingamma-chain antibody,antibody, wewe observedobserved thethe expectedexpected sizesize ofof HCHC (50similar(50 kDa)kDa) to that of Protein-A-purified E16 under reducing conditions (Figure 3a), confirming that the similar to that50 kDa of band Protein-A-purified on SDS-PAGE (Figure E16 2) under is indeed reducing the E16 conditionsHC and not (Figurethe Rubisco3a), large confirming subunit. that the similarsimilar toto thatthat ofof Protein-A-purifiedProtein-A-purified E16E16 underunder reducingreducing conditionsconditions (Figure(Figure 3a),3a), confirmingconfirming 50 thatthat kDa thethe band on SDS-PAGE (Figure 2) is indeed the E16 HC and not the Rubisco large subunit. 50 kDa bandSimilarly, on SDS-PAGE the expected (Figure 25 kDa2 )LC is band indeed was the observ E16ed HC under and reducing not the conditions Rubisco when large the subunit. blot was Similarly, 5050 kDakDa bandband incubatedonon SDS-PAGESDS-PAGE with an (Figureanti-human(Figure 2)2) kappa-chain isis indeedindeed antibodythethe E16E16 (Figure HCHC andand 3b). notnotTo ensure thethe RubiscoRubisco that E16 largewaslarge fully subunit.Similarly,subunit. the expected 25 kDa LC band was observed under reducing conditions when the blot was the expected 25 kDa LC band was observed under reducing conditions when the blot was incubated Similarly,Similarly, thethe assembledexpectedexpected and 2525 kDathekDa integrity LCLC bandband was wasmaintainedwas observobserv througheded underunder ATPS, reducingreducing samples were conditionsconditions also analyzed whenwhen under thethe non- blotincubatedblot waswas with an anti-human kappa-chain antibody (Figure 3b). To ensure that E16 was fully with an anti-humanreducing conditions. kappa-chain As shown antibody in Figure (Figure 3c, both 3theb). Protein To ensure A column that and E16 ATPS-enriched was fully assembled E16 mAbs and the incubatedincubated withwith anan anti-humananti-human kappa-chainkappa-chain antibodyantibody (Figure(Figure 3b).3b). ToTo ensureensure thatthat E16E16 waswasassembled fullyfully and the integrity was maintained through ATPS, samples were also analyzed under non- integrity waswere maintained observed as a through ~179 KDa ATPS,band. This samples represents were the also fully analyzedassembled and under glycosylated non-reducing E16 mAb conditions. assembledassembled andand thethe integrityintegrity waswas maintainedmaintained throughthrough ATPS,ATPS, samplessamples werewere alsoalso analyzedanalyzed underunderreducing non-non- conditions. As shown in Figure 3c, both the Protein A column and ATPS-enriched E16 mAbs As shownconsisting in Figure of3 c,a heterotetramer both the Protein of HC Aand column LC as we and observed ATPS-enriched previously [22] E16. Overall, mAbs the were enriched observed as reducingreducing conditions. conditions.mAb is fullyAs As shown shownassembled in in FigureFigureand retains 3c, 3c, both itsboth integrity the the Protein Protein after subjection A A column column to ATPS, and and ATPS-enriched ATPS-enrichedfurther supporting E16 E16 ourwere mAbsmAbs observed as a ~179 KDa band. This represents the fully assembled and glycosylated E16 mAb a ~179 KDa band. This represents the fully assembled and glycosylated E16 mAb consisting of a werewere observedobservedmethod asas aa ~179 ~179as an KDaeffectiveKDa bandband capture.. ThisThis step representsrepresents for purifying the theplant-made fullyfully assembledassembled mAbs. andand glycosylatedglycosylated E16E16consisting mAbmAb of a heterotetramer of HC and LC as we observed previously [22]. Overall, the enriched heterotetramer of HC and LC as we observed previously [22]. Overall, the enriched mAb is fully consistingconsisting ofof aa heterotetramerheterotetramer ofof HCHC andand LCLC asas wewe observedobserved previouslypreviously [22][22].. Overall,Overall, thethe enrichedenrichedmAb is fully assembled and retains its integrity after subjection to ATPS, further supporting our assembled and retains its integrity after subjection to ATPS, further supporting our method as an mAbmAb isis fullyfully assembledassembled andand retainsretains itsits integrityintegrity afterafter subjectionsubjection toto ATPS,ATPS, furtherfurther supportingsupportingmethod ourour as an effective capture step for purifying plant-made mAbs. effective capture step for purifying plant-made mAbs. methodmethod asas anan effectiveeffective capturecapture ststepep forfor purifyingpurifying plant-madeplant-made mAbs.mAbs. 2.4. Antigen Specificity of ATPS-Purified E16 E16 recognizes a conformational epitope consisting of four discontinuous polypeptides on the surface of the domain III (DIII) of the WNV E protein [25]. Therefore, the binding between E16 and DIII is highly specific and contingent upon the structural integrity of both proteins. To verify that ATPS-enriched E16 retains recognition for its epitope, a binding assay utilizing a yeast display system was performed [22]. Flow cytometric analysis revealed that the incubation of ATPS-purified E16 with Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 5 of 12

Int. J. Mol. Sci. 2020, 21, 2140 5 of 12

WNV DIII-displaying yeast cells resulted in an E16-specific peak that shared a similar percentage of positive yeast cells and fluorescent intensity with that of the Protein A column-purified E16 (Figure4). This providesInt. J. Mol. additional Sci. 2019, 20, xevidence FOR PEER REVIEW that ATPS enrichment does not alter the mAb’s specificity5 of 12 to its antigen, furtherFigure supporting 3. Western theblot useanalysis of thisof ATPS-purified method for E16. purifying ATPS- or Protein mAbs A column-purified that require E16 functional was integrity. separated on 12% SDS-PAGE gels under reducing (a and b) conditions or on a 4–20% SDS-PAGE gel under non-reducing (c) conditions and blotted onto polyvinylidene fluoride (PVDF) membranes. A goat anti-human gamma chain antibody (a and c) or a goat anti-human kappa chain antibody (b) were incubated with the membranes to detect heavy chain or light chain, respectively. Lane 1, ATPS-

enriched E16; Lane 2, Protein A column-purified E16. HC: heavy chain, LC: light chain, (HL)2: assembled mAb with two light and heavy chains.

separatedwas performed on 12% SDS-PAGE [22]. Flow cytometric gels under analysis reducing reveal (aedand thatb )the conditions incubation or of onATPS-purified a 4–20% SDS-PAGE E16 with gel WNV DIII-displaying yeast cells resulted in an E16-specific peak that shared a similar percentage of under non-reducingFigure 3. Western (c) conditions blot analysis andof ATPS-purified blotted onto E16. polyvinylidene ATPS- or Protein A ﬂuoride column-purified (PVDF) E16 membranes. was A positive yeast cells and fluorescent intensity with that of the Protein A column-purified E16 (Figure goat anti-humanseparated gamma on 12% SDS-PAGE chain antibody gels under (a andreducingc) or (a a and goat b) conditions anti-human or on kappa a 4–20% chain SDS-PAGE antibody gel (b) were 4). Thisunder provides non-reducing additional (c) conditions evidence and that blotted ATPS onto enrichment polyviny lidenedoes not fluoride alter (PVDF)the mAb’s membranes. specificity A to its incubated with the membranes to detect heavy chain or light chain, respectively. Lane 1, ATPS-enriched antigen,goat anti-humanfurther supporting gamma chain the antibody use of (athis and methc) or aod goat for anti-human purifying kappa mAbs chain th antibodyat require (b ) functionalwere E16; Laneintegrity.incubated 2, Protein with A column-puriﬁedthe membranes to detect E16. HC:heavy heavy chain or chain, light LC:chain, light respectively. chain,(HL) Lane2 1,: assembledATPS- mAb with two lightenriched and E16; heavy Lane chains. 2, Protein A column-purified E16. HC: heavy chain, LC: light chain, (HL)2: assembled mAb with two light and heavy chains.

2.4. Antigen Specificity of ATPS-Purified E16 E16 recognizes a conformational epitope consisting of four discontinuous polypeptides on the surface of the domain III (DIII) of the WNV E protein [25]. Therefore, the binding between E16 and DIII is highly specific and contingent upon the structural integrity of both proteins. To verify that ATPS-enriched E16 retains recognition for its epitope, a binding assay utilizing a yeast display system was performed [22]. Flow cytometric analysis revealed that the incubation of ATPS-purified E16 with WNV DIII-displaying yeast cells resulted in an E16-specific peak that shared a similar percentage of positive yeast cells and fluorescent intensity with that of the Protein A column-purified E16 (Figure 4). This provides additional evidence that ATPS enrichment does not alter the mAb’s specificity to its Figureantigen, 4. Binding further of ATPS-enrichedsupporting the use E16 of to this domain meth IIIod offor WNV purifying E displayed mAbs th onat therequire surface functional of yeast cell.

Yeast cellsintegrity. displaying domain III of WNV E protein were incubated with ATPS-enriched E16, Protein A column-purifiedFigure 4. Binding E16 (positive of ATPS-enriched control), E16 or to a domain generic III human of WNV IgGE displayed (hIgG, on negative the surface control), of yeast and then stained withcell. a Yeast Fluorescein cells displaying isothiocyanate domain III (FITC)-conjugatedof WNV E protein were antibody. incubated Yeast with ATPS-enriched cells were then E16, processed by flow cytometryProtein A tocolumn-purified monitor the E16 specific (positive binding control), of or E16 a generic to WNV human E domainIgG (hIgG, III negative (DIII). control), and then stained with a Fluorescein isothiocyanate (FITC)-conjugated antibody. Yeast cells were then 2.5. ATPS withprocessed Co-expression by flow cytometry of E16 and to monitor HPA inthe Plantsspecific binding of E16 to WNV E domain III (DIII).

Purification of plant-made mAbs could be further streamlined by expressing both the mAb and its purifying reagent molecule, such as HPA, in the same plant. Thus, we investigated the feasibility of co-expressing E16 and HPA together in the same plant and using the co-expressed proteins to purify E16 via ATPS. As the production of incompletely assembled antibodies would be deleterious to the immune response, the in vivo assembly of the antibody is a tightly controlled process involving the sequential interaction between the HC, LC, and associated chaperons [26]. Specifically, the early formation of HC dimers is crucial for the subsequent binding of LCs and any interference of this event may affectFigure the 4. overallBinding ofassembly ATPS-enriched of E16 the to full domain antibody III of WNV [26 E, displayed27]. Our on previous the surface of results yeast have shown cell. Yeast cells displaying domain III of WNV E protein were incubated with ATPS-enriched E16, that the accumulationProtein A column-purified levels of a mAb E16 (positive in plants control), is greatlyor a generic reduced human IgG when (hIgG, expressed negative control), in the format of a recombinant immuneand then stained complex with a Fluorescein (RIC) compared isothiocyanate to (FITC)-conjugated its parent IgG antibody. [28,29 Yeast]. These cells were results then support the hypothesis thatprocessed the in by planta flow cytometryinteraction to monitor of mAbthe specific HC-antigen binding of E16 fusion to WNV and E domain its antigen III (DIII). binding domain (Fab) between RICs interferes with HC assembly. To avoid the potential interaction of E16 HC and HPA in planta which may also interfere with E16 HC and LC assembly, we took advantage of the phenomenon of “competing replicons” of the MagnICON vectors [30–32]. By cloning the E16 HC and HPA genes into a vector that is based on the same TMV backbone, only the HC or HPA, but not both would be preferentially expressed in any given cell to avoid their interaction. We anticipated that Int. J. Mol. Sci. 2019, 20, x FOR PEER REVIEW 6 of 12

2.5. ATPS with Co-expression of E16 and HPA in Plants Purification of plant-made mAbs could be further streamlined by expressing both the mAb and its purifying reagent molecule, such as HPA, in the same plant. Thus, we investigated the feasibility of co-expressing E16 and HPA together in the same plant and using the co-expressed proteins to purify E16 via ATPS. As the production of incompletely assembled antibodies would be deleterious to the immune response, the in vivo assembly of the antibody is a tightly controlled process involving the sequential interaction between the HC, LC, and associated chaperons [26]. Specifically, the early formation of HC dimers is crucial for the subsequent binding of LCs and any interference of this event may affect the overall assembly of the full antibody [26,27]. Our previous results have shown that the accumulation levels of a mAb in plants is greatly reduced when expressed in Int.Int.the J. J. Mol.format Mol. Sci. Sci. 2019of 2019 a, ,20 20, ,x x FOR FOR PEER PEER REVIEW REVIEW 44 of of 12 12 Int. J. Mol. Sci. 2019, 20recombinant, x FOR PEER immune REVIEW complex (RIC) compared to its parent IgG [28,29]. These results4 support of 12 the hypothesis that the in planta interaction of mAb HC-antigen fusion and its antigenmaintain maintainbinding domain thethe efficiencyefficiency ofof ProteinProtein AA chromatographychromatography [23].[23]. ByBy eliminatingeliminating thethe bulkbulk separationseparation step,step, maintain the efficiency(Fab) between of Protein RICs Ainterferes chromatography with HC assembly. [23]. By To eliminating avoid the potential the bulk interaction separationATPSATPS of E16step, savessaves HC and timetime andand resources,resources, andand potentiallypotentially decreasesdecreases thethe lossloss ofof mAbmAb duedue toto proteolyticproteolytic Int. J. Mol. Sci. 2020, 21, 2140 6 of 12 ATPS saves timeHPA and in plantaresources, which and may potentiallyalso interfere decreases with E16 HCthe and loss LC of assembly, mAb due we totook proteolytic degradationadvantagedegradation of the byby reducingreducing thethe timetime itit isis exposedexposed toto plantplant hosthost proteolyticproteolytic enzymesenzymes uponupon degradation byphenomenon reducing theof “competing time it replicons”is exposed of the to Magn plantICON host vectors proteolytic [30–32]. By enzymes cloninghomogenizationhomogenization the upon E16 HC and [24]. [24]. The The ATPS ATPS strategy strategy reported reported here here also also has has advantages advantages over over other other ATPS ATPS methods methods HPA genes into a vector that is based on the same TMV backbone, only the HC or HPA, but not both homogenizationthe LC, [24]. being The delivered ATPS strategy in a non-competing reported here also PVX-based has advantages vector, wouldover other be expressed ATPSreported methodsreported in all previously previously infiltrated [20]. [20]. Instead Instead of of using using purified purified HPA, HPA, the the crude crude plant plant extract extract containing containing HPA HPA was was would be preferentially expressed in any given cell to avoid their interaction. We anticipated that the reported previously [20]. Instead of using purified HPA, the crude plant extract containing HPAdirectly was mixed with E16-containing extract in our process, eliminating extra steps for HPA cells.LC, Only being in the delivered cells that in a preferentially non-competing expressed PVX-based the vect HCor, would over HPA be expressed would a in fully alldirectly infiltrated assembled mixed cells. E16 with be E16-containing extract in our process, eliminating extra steps for HPA purification. This time and cost-saving improvement did not compromise the purity of the resulted directly mixedproduced. Onlywith in UponE16-containing the cells homogenization that preferentially extract of in the expressedour leaf process, material, the HC eliminatthe over fully HPAing assembled wouldextra asteps fully E16 assembled wouldforpurification. HPA bind E16 to Thisbe HPA, time and cost-saving improvement did not compromise the purity of the resulted purification.while Thisproduced. any time unassembled and Upon cost-saving homogenization LC would improvemen not of associatethet leaf did material, not with compromise any the HPA, fully thusas thesembled facilitatingpurity E16 of wouldthe anmAb, mAb,resulted initial bind but but to capture actuallyHPA,actually and yielded yielded better better recovery recovery of of mAb mAb (4 (47.7%)7.7%) from from the the more more complex complex protein protein containing containing mAb, but actuallyenrichmentwhile yielded any of assembledunassembled better recovery E16 LC would via of mAb ATPS not (4 associate from7.7%) plant from with crude the any more HPA, extract. complexthus Indeed,facilitating protein our an resultscontaininginitialplantplant capture crudeindicatedcrude and extractextract that thanthan thethe moremore cumbersomecumbersome memethodthod (24%)(24%) fromfrom thethe hybridomahybridoma supernatantsupernatant [20].[20]. plant crudewe extract wereenrichment than able the to of enrich more assembled cumbersome E16 that E16 wasvia ATPS me co-expressedthod from (24%) plant withfrom crude HPAthe extract. hybridoma in N. Indeed, benthamiana supernatantour resultsleaves,TheThe indicated efficient efficient[20]. albeit that enrichment enrichment with a and and the the remarkable remarkable rate rate of of recovery recovery supports supports our our AT ATPSPS method method as as a a simple, simple, The efficientlower enrichmentwe purity were (9.4%ableand tothe less enrich remarkable pure E16 by that densitometry rate was of co-expressed recovery analysis) supports with thanHPA our in when ATN. benthamianaPS they method were leaves,as expressed costacost simple, albeit effective,effective, in with separate ayetyet efficientefficient bulkbulk separationseparation stepstep toto enrichenrich plant-madeplant-made mAbsmAbs fromfrom mostmost ofof thethe hosthost cost effective,plants yetlower (Figureefficient purity5, bulk Lane (9.4% separation 3). less The pure yield by step densitometry of to E16 enrich is lower, analysplant-made asis) onlythan mAbs awhen percentage theyfrom were most of expressed infiltrated ofplant plantthe host inproteins. proteins. cellsseparate would plants (Figure 5, Lane 3). The yield of E16 is lower, as only a percentage of infiltrated cells would plant proteins.express E16 HC due to the competition from HPA. The low yield of E16 may also have altered the express E16 HC due to the competition from HPA. The low yield of E16 may also have altered the ratio of E16 to HPA, contributing to lower ATPS efficiency and lower E16 purity. Therefore, further ratio of E16 to HPA, contributing to lower ATPS efficiency and lower E16 purity. Therefore, further optimizationoptimization to better to better balance balance the the expression expression of of each each moleculemolecule inin planta planta is isneeded needed to improve to improve the yield the yield and purityand purity of mAbs of mAbs by thisby this approach. approach. Nevertheless, Nevertheless, toto our knowledge, knowledge, this this is the is the first first demonstration demonstration of co-expressionof co-expression of a of mAb a mAb and and a bifunctionala bifunctional protein protein capable of of binding binding the the mAb mAb in the in thesame same plant, plant, and furthermore,and furthermore, the successfulthe successful partitioning partitioning and and enrichmentenrichment of of the the mAb mAb by by a capturing a capturing molecule molecule in in a simplea simple two-phase two-phase extraction extraction system. system. This This provides provides evidence for for potentially potentially further further enhancing enhancing the the efficiencyefficiency of the of purification the purification scheme scheme proposed proposed in in this this study.study.

FigureFigure 2. 2. Aqueous Aqueous two-phase two-phase separation separation (ATPS)-b (ATPS)-basedased purification purification of of E16 E16 mAb mAb from from N. N. benthamiana benthamiana Figure 2. Aqueous two-phase separation (ATPS)-based purification of E16 mAb from N. benthamiana leaves.leaves. Total Total soluble soluble proteins proteins were were extracted extracted from from plantsplants and and E16 E16 was was purifi purifieded by by ATPS ATPS (Lanes (Lanes 1–3) 1–3)

leaves. TotalFigure soluble 5. proteinsATPS purification were extracted of E16 from with plants HPA co-expressedand E16 was purifi in theed same by ATPSN. benthamiana (Lanes 1–3)leaf. oror Total ProteinProtein leaf AA affinityaffinity chromatographchromatograph (Lanes(Lanes 4-6),4-6), analyzedanalyzed onon aa 12%12% SDS-PAGESDS-PAGE gelgel underunder reducingreducing or Protein Asoluble affinityFigure proteins chromatograph 5. ATPS were purification extracted (Lanes of from4-6), E16 with leavesanalyzed HPA co-expressing co-expressed on a 12% SDS-PAGE in both the E16same andgel N. benthamiana HPA.under Thereducing plantleaf. Total conditions, extractconditions, leaf was and and visualized visualized with with Coomassie Coomassie stain. stain. Lane Lane 1, 1, total total leaf leaf soluble soluble proteins; proteins; Lane Lane 2, 2, proteins proteins conditions, andthen visualized processedsoluble proteins with through Coomassiewere ATPS extracted to stain. enrich from Lane leaves E16. 1, ATPSco-etotalxpressing leaf fractions soluble both were proteins;E16 analyzedand HP LaneA. onThe 2, a proteinsplant 12% SDS-PAGEextract inin the thewas upperupper gel aqueousaqueous phasephase ofof ATPS;ATPS; LaneLane 3,3, ATPS-eATPS-enrichednriched E16;E16; LaneLane 4,4, leafleaf solublesoluble proteinsproteins afterafter in the upperunder aqueousthen reducing processedphase conditionsof ATPS;through Lane ATPS and 3, visualized to ATPS-e enrich nrichedE16. with ATPS Coomassie E16; fractions Lane stain.were4, leaf analyzed Lanesoluble 1, on totalproteins a 12% leaf SDS-PAGE solubleafter lowlow proteins; pHgel pH precipitation; precipitation; Lane Lane 5, 5, Protein Protein A A flow flow thro through;ugh; Lane Lane 6, 6, Protein Protein A A co column-purifiedlumn-purified E16. E16. : :HC, HC, under reducing conditions and visualized with Coomassie stain. Lane 1, total leaf soluble proteins; low pH precipitation;Lane 2, proteins Lane 5, in Protein the upper A flow aqueous through; phase Lane of ATPS;6, Protein Lane A 3,co ATPS-enrichedlumn-purified E16. :: HC, HC, : ::LC, LC, LC, :: :minor minor impurity, impurity, non-mAb non-mAb bands. bands. Lane 2, proteins in the upper aqueous phase of ATPS; Lane 3, ATPS-enriched E16. : HC, : LC, : LC, : minorminor impurity,impurity, non-mAbnon-mAb bands.bands. : minor impurity, non-mAb bands. 2.3.2.3. Characterization Characterization of of ATPS-Purified ATPS-Purified E16 E16 mAb mAb Scalability is one of the deciding factors for the adaptation of ATPS in the commercial setting 2.3. Characterization of ATPS-Purified E16 mAb Western blot analysis was used to characterize E16 recovered from ATPS. When the blot was for mAb production. The scalability of ATPS has been demonstrated in processing variousWestern protein blot analysis was used to characterize E16 recovered from ATPS. When the blot was probed with an anti-human gamma-chain antibody, we observed the expected size of HC (50 kDa) Westernproducts blot analysis with high was purity used andto characterize recovery [33 ].E16 Moreover, recovered it hasfrom been ATPS. shown When that the ATPS probedblot can was bewith integrated an anti-human gamma-chain antibody, we observed the expected size of HC (50 kDa) similar to that of Protein-A-purified E16 under reducing conditions (Figure 3a), confirming that the probed withseamlessly an anti-human with other gamma-chain downstream antibody, processing we observed techniques the [expected34]. In a size recent of study,HC similar(50 the kDa) scalabilityto that of Protein-A-purified of E16 under reducing conditions (Figure 3a), confirming that the 50 kDa band on SDS-PAGE (Figure 2) is indeed the E16 HC and not the Rubisco large subunit. similar to thatHPA-based of Protein-A-purified ATPS for mAb E16 production under reducing has been conditions demonstrated (Figure [20]. 3a), This confirming further suggests50 that kDa the theband potential on SDS-PAGE (Figure 2) is indeed the E16 HC and not the Rubisco large subunit. Similarly, the expected 25 kDa LC band was observed under reducing conditions when the blot was 50 kDa bandapplication on SDS-PAGE of using (Figure HPA-containing 2) is indeed plant-extract the E16 HC in purifyingand not the plant-made Rubisco mAbs.largeSimilarly, subunit. the expected 25 kDa LC band was observed under reducing conditions when the blot was Similarly, the expected 25 kDa LC band was observed under reducing conditions when theincubated incubatedblot was withwith anan anti-humananti-human kappa-chainkappa-chain antibodyantibody (Figure(Figure 3b).3b). ToTo ensureensure thatthat E16E16 waswas fullyfully incubated 3.with Materials an anti-human and Methods kappa-chain antibody (Figure 3b). To ensure that E16 wasassembledassembled fully and and the the integrity integrity was was maintained maintained through through ATPS, ATPS, samples samples were were also also analyzed analyzed under under non- non- assembled and the integrity was maintained through ATPS, samples were also analyzed underreducingreducing non- conditions. conditions. As As shown shown in in Figure Figure 3c, 3c, both both the the Protein Protein A A column column and and ATPS-enriched ATPS-enriched E16 E16 mAbs mAbs reducing conditions.3.1. Materials As shown in Figure 3c, both the Protein A column and ATPS-enriched E16werewere mAbs observedobserved asas aa ~179~179 KDaKDa bandband. . ThisThis representsrepresents thethe fullyfully assembledassembled andand glycosylatedglycosylated E16E16 mAbmAb were observed asThe a ~179 salts KDa for phosphateband. This burepresentsffered saline, the fully sodium assembled L-ascorbate, and glycosylated carbonate, E16consistingconsisting bicarbonate, mAb ofof aa citric heterotetramerheterotetramer ofof HCHC andand LCLC asas wewe observedobserved previouslypreviously [22][22]. . Overall,Overall, thethe enrichedenriched consisting acid,of a heterotetramer yeast nitrogen of base HC and without LC as amino we observed acids, yeast previously synthetic [22]. drop-out Overall, the medium mAbenrichedmAb isis supplements, fullyfully assembledassembled andand retainsretains itsits integrityintegrity afterafter subjectionsubjection toto ATPS,ATPS, furtherfurther supportingsupporting ourour mAb is fullyammonium assembled persulfate, and retains tetramethylethlyenediamine its integrity after subjection (TEMED),to ATPS, further glucose supporting andmethodmethod galactose our as as an werean effective effective all capture capture st stepep for for purifying purifying plant-made plant-made mAbs. mAbs. method as purchasedan effective fromcapture Sigma-Aldrich step for purifying (St. plant-made Louis, MO, mAbs. USA). Coomassie Brilliant Blue R-250 and OmniPur®phenylmethylsulfonyl fluoride (PMSF) were purchased from EMD Chemicals (Gibbstown, NJ, USA). Ethylenediaminetetraacetic acid (EDTA), Tween-20, Tris-base and sodium dodecyl sulfate (SDS) were purchased from Fisher Scientific (Fair Lawn, NJ, USA). All detection antibodies used for ELISAs and Western blots were purchased from SouthernBiotech (Homewood, AL, USA). Plates used for Int. J. Mol. Sci. 2020, 21, 2140 7 of 12

ELISAs were purchased from Corning (Kennebunk, ME, USA). The KPL 3, 30, 5, 50–tetramethylbenzidine (TMB) peroxidase substrate kit for ELISA was purchased from SeraCare (Milford, MA, USA). Proteomics grade Triton X-114 was purchased from Amresco®Chemicals (Solon, OH, USA). Electrophoresis grade glycine was purchased from MP Biomedicals LLC (Solon, OH, USA). 2-methylpropan-1-ol (isobutanol) was purchased from Acros Organics (Fair Lawn, NJ, USA). Centrifugal filter units were purchased from Millipore (Billerica, MA, USA). Pierce detergent removal spin columns, Millipore Express PLUS 0.22µm vacuum membranes and Pierce ECL Western blotting substrate were all purchased from Thermo Scientific (Rockford, IL, USA). MabSelect resin was purchased from GE Healthcare (Upsala, Sweden). 30% (29:1) acrylamide/bis solution and polyvinylidene fluoride (PVDF) membrane were purchased from Bio-Rad (Hercules, CA, USA).

3.2. Bacterial Strains and Growth The coding sequence of HPA [20] was cloned into the MagnIcon vector pICH11599 and transformed into Agrobacterium tumefaciens GV3101 cells as described previously [35].

3.3. Agroinfiltration N. benthamiana plants were grown and infiltrated with target gene-harboring A. tumefaciens GV3101 strains as previously described [36]. Specifically, A. tumefaciens strains containing the E16 HC 30 module, E16 LC 30 module [22], their corresponding 50 modules, and the integrase construct [37] in the MagnICON system were co-infiltrated into leaves of 5-7-week old N. benthamiana with a combined OD600 of 0.74. GV3101 strains containing the HPA 30 module, the 50 module, and the integrase module were co-infiltrated into leaves with a final OD600 = 0.9. For co-expression of E16 and HPA in the same leaves, the six strains of GV3101 required for E16 HC, LC and HPA expression were co-infiltrated into the leaves N. benthamiana plants with a final OD600 of 0.98 as described previously [38].

3.4. HPA ELISA HPA-expressing leaves were homogenized in 1 PBS (8 mM Na HPO , 2 mM KH PO , 15 mM × 2 4 2 4 NaCl, 2.5 mM KCl) pH 7.4, supplemented with 10 mg/mL of sodium L-ascorbate, 2 mM PMSF, and 1mM EDTA in a fresh leaf mass to buffer volume ratio of 1:1.5 [39]. The plant extract was then clarified by centrifugation twice at 16,873 g for 10 min at 4 C. Samples were then diluted in 100 mM × ◦ bicarbonate buffer (pH 9.2) along with the HPA standard, ranging from 50 ng/mL to 0.78 ng/mL in a 96-well plate. The plate was then incubated at 37 ◦C for 2 h. After blocking with 5% skim milk in PBST (1 PBS with 0.05% Tween-20), generic human IgG was added to the plate to interact with × the antibody-binding domain of HPA coated on the plate. The plate was incubated at 37 ◦C for 1 h, and a horseradish peroxidase-conjugated goat anti-human kappa chain was added as the detection antibody. The plate was developed with TMB and then read at 450nm as previously described [40]. Protein expression level was then calculated as micrograms of recombinant protein per gram of fresh leaf weight (FLW).

3.5. Aqueous Two-Phase Separation Leaves expressing each recombinant protein (HPA or E16) were harvested at 7 days post infiltration (DPI) and homogenized in extraction buffer separately in a blender as described [41]. After clarification by centrifugation, equal volumes of E16 and HPA infiltrated leaf extract were mixed and incubated at 4 ◦C for 1 h. The mixed extract was then diluted 1:10 in PBS pH 7.4 and incubated in a 26 ◦C water bath for 5 min. Triton X-114 was added to 1mL of the pre-warmed mixed extract at a concentration of 4% w/v and mixed well. Phase separation was achieved by a 5-min incubation in at 26 ◦C. The upper aqueous phase was removed and 0.5 mL of 100 mM glycine-HCl (pH 2.0) was added to the lower detergent-rich phase and thoroughly mixed to dissociate E16 from HPA. After phase separation as described above, the upper phase containing the enriched E16 was transferred to another tube and pH adjusted to Int. J. Mol. Sci. 2020, 21, 2140 8 of 12 pH 7.0 with 1M Tris-base. The HPA fusion protein was recovered by adding isobutanol to the lower detergent-rich phase as described previously [20]. The same ATPS procedure was followed when HPA and E16 were co-expressed in the same leaf, except there was no separate homogenization and mixing of leaf extracts. When necessary, the enriched E16 fractions were concentrated with a 100K MWCO centrifugal filter unit and passed through a detergent removal spin column before further analysis.

3.6. Protein A Affinity Chromatography of E16 For use as a reference standard, plant-expressed E16 was also purified by Protein A affinity chromatography as described previously [42]. Briefly, control leaves infiltrated with the E16 construct were homogenized as described above, with the exception that the extraction buffer pH was 5.2. After clarification by centrifugation, the plant extract was re-adjusted to pH 5.2 and further incubated overnight at 4 ◦C. After removal of the precipitated plant host proteins by centrifugation, the extract was adjusted to pH 7.2, filtered through a Millipore Express PLUS 0.22µm membrane. The filtered extract was then loaded onto a Protein A column with MabSelect resin at a rate of 0.8-1.2mL/minute. E16 was eluted from the Protein A column with 100mM citric acid, which was then neutralized with 1M Tris-base.

3.7. SDS-PAGE and Western Blotting Fractions of the starting extract and the flow through (Protein A column) or upper phase (ATPS), along with equivalent amounts of purified E16 from both methods were subjected to reducing SDS-PAGE on a 12% gel and stained with Coomassie Brilliant Blue R-250. For co-expression, only the ATPS samples were included. The purity of E16 mAb was determined by imaging and quantitating Coomassie blue-stained protein bands on SDSPAGE using a densitometer as described previously [43]. For Western blot analysis, ATPS and column-purified E16 samples were separated via SDS-PAGE on 12% or 4–20% gradient gels under reducing or non-reducing conditions, respectively, and then transferred to a PVDF membrane. After blocking with 5% milk in PBST, the membranes were probed with either goat anti-human gamma chain-HRP or kappa chain-HRP antibody diluted in 1% milk in PBST. After washing with PBST, Pierce ECL Western blotting substrate was mixed according to the manufacturer’s instructions and used to cover the membrane. The ImageQuant LAS 4000 imager and associated software was then used to image the blots.

3.8. E16 ELISA for ATPS E16 Recovery Recovery of E16 purified from ATPS was quantified by an ELISA as previously described [42]. Briefly, 96-well plates were coated with the WNV domain III (DIII) of the E protein [41]. After blocking with 5% milk in PBST, fractions from the various ATPS purification steps were added to the wells, along with mammalian cell-produced E16 as a standard [37]. A goat anti-human kappa chain-HRP was used as the detection antibody. The plate was then developed with TMB and read with a plate reader at 450nm. E16 recovery by ATPS was calculated as the percentage of purified E16 obtained from the total E16 in the mixed plant extract.

3.9. Yeast Display Assay The yeast display assay was adapted from our published procedure [22]. Saccharomyces cerevisiae EBY100 cells harboring the gene for WNV DIII in the yeast surface display vector, pYD1, were grown in synthetic yeast medium with 2% glucose at 30 ◦C until log phase growth was reached (OD600 2-5). Cells were then grown an additional 36 h at room temperature in synthetic yeast medium with 2% galactose to induce expression and display of DIII on the yeast surface. Cells were collected, washed, and incubated with ATPS-enriched E16, Protein-A-puriﬁed E16 (positive control) or a generic human IgG (negative control). Cells were then stained with goat anti-human IgG-FITC and counted with a Gallios ﬂow cytometer. Int. J. Mol. Sci. 2020, 21, 2140 9 of 12

4. Conclusions The use of plants as a production platform for mAb therapeutics has great potential to not only reduce the upstream cost of production [6], but also provide opportunities to improve efficacy and safety [40,44,45]. However, the downstream processing costs between plant and cell culture-based platforms remain essentially equivalent, highlighting the need for alternatives or improvements to the current methods for mAb purification [46]. Removing as many contaminating molecules as possible by introducing an efficient yet inexpensive initial capture step, as described here, will prolong the longevity of the Protein A resins in the affinity chromatography step that are the current gold standard for mAb purification. In this study, we built upon the novel concept of utilizing a bifunctional fusion protein to capture a mAb in an ATPS to reduce downstream costs of purifying mAbs from plants. By expressing both the E16 mAb and the HPA fusion protein in plants, directly mixing plant extracts in a controlled ratio, and enriching the mAb through the facile ATPS method described here, we were able to produce a highly enriched E16 mAb with a recovery rate comparable to that obtained from the conventional Protein A affinity chromatography. Furthermore, the recovered mAb was fully assembled and fully functional by retaining recognition of its target antigen. This highly enriched E16 preparation not only can be easily purified to homogeneity by the Protein A affinity chromatography step but may also increase the longevity of the Protein A resin due to the reduction of host proteins in the feedstream of the affinity chromatography. Of note is the demonstration of a mAb and the HPA fusion protein being co-expressed within the same plant and the subsequent enrichment of the mAb via ATPS with an inexpensive detergent. These steps towards improving the purification methods of plant-made mAbs are promising and warrant further optimization and studies on a larger scale to validate the utility of this system.

Author Contributions: Q.C. conceptualized the research, C.J. performed the research; C.J. and Q.C. analyzed the data; J.J. provided research material resource; C.J. and Q.C. wrote the paper with J.J. providing critical comments. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported in part by a grant from National Institute of Allergy and Infectious Diseases (NIAID) # R33AI101329 to QC. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Ecker, D.M.; Jones, S.D.; Levine, H.L. The therapeutic monoclonal antibody market. mAbs 2014, 7, 9–14. [CrossRef][PubMed] 2. Jonker, D.J.; O’Callaghan, C.J.; Karapetis, C.S.; Zalcberg, J.R.; Tu, D.; Au, H.-J.; Berry, S.R.; Krahn, M.; Price, T.; Simes, R.J.; et al. Cetuximab for the Treatment of Colorectal Cancer. New Engl. J. Med. 2007, 357, 2040–2048. [CrossRef][PubMed] 3. Kim, E.S. Avelumab: First Global Approval. Drugs 2017, 77, 929–937. [CrossRef][PubMed] 4. Qiu, X.; Wong, G.; Audet, J.; Bello, A.; Fernando, L.; Alimonti, J.B.; Fausther-Bovendo, H.; Wei, H.; Aviles, J.; Hiatt, E.; et al. Reversion of advanced Ebola virus disease in nonhuman primates with ZMapp. Nature 2014, 514, 47–53. [CrossRef] 5. Chen, Q.; Davis, K. The potential of plants as a system for the development and production of human biologics. F1000Research 2016, 5.[CrossRef] 6. Nandi, S.; Kwong, A.T.; Holtz, B.R.; Erwin, R.L.; Marcel, S.; McDonald, K.A. Techno-economic analysis of a transient plant-based platform for monoclonal antibody production. mAbs 2016, 8, 1456–1466. [CrossRef] 7. Ma, J.K.C.; Drossard, J.; Lewis, D.; Altmann, F.; Boyle, J.; Christou, P.; Cole, T.; Dale, P.; van Dolleweerd, C.J.; Isitt, V.; et al. Regulatory approval and a ﬁrst-in-human phase I clinical trial of a monoclonal antibody produced in transgenic tobacco plants. Plant Biotechnol. J. 2015, 13, 1106–1120. [CrossRef] 8. Ghose, S.; Hubbard, B.; Cramer, S.M. Binding capacity diﬀerences for antibodies and Fc-fusion proteins on protein A chromatographic materials. Biotechnol. Bioeng. 2007, 96, 768–779. [CrossRef] 9. Jiang, C.; Liu, J.; Rubacha, M.; Shukla, A.A. A mechanistic study of Protein A chromatography resin lifetime. J. Chromatogr. A 2009, 1216, 5849–5855. [CrossRef] Int. J. Mol. Sci. 2020, 21, 2140 10 of 12

10. Chen, Q. Expression and Purification of Pharmaceutical Proteins in Plants. Biol. Eng. 2008, 1, 291–321. [CrossRef] 11. Pathak, M.; Rathore, A.S. Mechanistic understanding of fouling of protein A chromatography resin. J. Chromatogr. A 2016, 1459, 78–88. [CrossRef] 12. Valdés, R.; Gómez, L.; Padilla, S.; Brito, J.; Reyes, B.; Álvarez, T.; Mendoza, O.; Herrera, O.; Ferro, W.; Pujol, M.; et al. Large-scale purification of an antibody directed against hepatitis B surface antigen from transgenic tobacco plants. Biochem. Biophys. Res. Commun. 2003, 308, 94–100. [CrossRef] 13. Hey, C.; Zhang, C. Process Development for Antibody Purification from Tobacco by Protein A Affinity Chromatography. Chem. Eng. Technol. 2012, 35, 142–148. [CrossRef] 14. Bai, Y.; Glatz, C.E. Capture of a recombinant protein from unclarified canola extract using streamline expanded bed anion exchange. Biotechnol. Bioeng. 2003, 81, 855–864. [CrossRef][PubMed] 15. Walter, H.; Johansson, G.; Brooks, D.E. Partitioning in aqueous two-phase systems: Recent results. Anal. Biochem. 1991, 197, 1–18. [CrossRef] 16. Linder, M.B.; Qiao, M.; Laumen, F.; Selber, K.; Hyytiä, T.; Nakari-Setälä, T.; Penttilä, M.E. Efficient Purification of Recombinant Proteins Using Hydrophobins as Tags in Surfactant-Based Two-Phase Systems. Biochemistry 2004, 43, 11873–11882. [CrossRef] 17. Lahtinen, T.; Linder, M.B.; Nakari-Setälä, T.; Oker-Blom, C. Hydrophobin (HFBI): A potential fusion partner for one-step purification of recombinant proteins from insect cells. Protein Expr. Purif. 2008, 59, 18–24. [CrossRef] 18. Joensuu, J.J.; Conley, A.J.; Lienemann, M.; Brandle, J.E.; Linder, M.B.; Menassa, R. Hydrophobin Fusions for High-Level Transient Protein Expression and Purification in Nicotiana benthamiana. Plant Physiol. 2010, 152, 622. [CrossRef] 19. Reuter, L.J.; Shahbazi, M.-A.; Mäkilä, E.M.; Salonen, J.J.; Saberianfar, R.; Menassa, R.; Santos, H.A.; Joensuu, J.J.; Ritala, A. Coating Nanoparticles with Plant-Produced Transferrin–Hydrophobin Fusion Protein Enhances Their Uptake in Cancer Cells. Bioconjug. Chem. 2017, 28, 1639–1648. [CrossRef] 20. Kurppa, K.; Reuter, L.J.; Ritala, A.; Linder, M.B.; Joensuu, J.J. In-solution antibody harvesting with a plant-produced hydrophobin–Protein A fusion. Plant Biotechnol. J. 2018, 16, 404–414. [CrossRef] 21. Lai, H.; He, J.; Hurtado, J.; Stahnke, J.; Fuchs, A.; Mehlhop, E.; Gorlatov, S.; Loos, A.; Diamond, M.S.; Chen, Q. Structural and functional characterization of an anti-West Nile virus monoclonal antibody and its single-chain variant produced in glycoengineered plants. Plant Biotechnol. J. 2014, 12, 1098–1107. [CrossRef] 22. Lai, H.; Engle, M.; Fuchs, A.; Keller, T.; Johnson, S.; Gorlatov, S.; Diamond, M.S.; Chen, Q. Monoclonal antibody produced in plants efficiently treats West Nile virus infection in mice. Proc. Natl. Acad. Sci. USA 2010, 107, 2419–2424. [CrossRef] 23. Chen, Q. Expression and manufacture of pharmaceutical proteins in genetically engineered horticultural plants. In Transgenic Horticultural Crops: Challenges and Opportunities—Essays by Experts; Mou, B., Scorza, R., Eds.; Taylor & Francis Boca Raton: Abingdon, UK, 2011; pp. 83–124. ISBN 978-1-4200-9379-7. 24. De Muynck, B.; Navarre, C.; Boutry, M. Production of antibodies in plants: Status after twenty years. Plant Biotechnol. J. 2010, 8, 529–563. [CrossRef] 25. Lai, H.; Paul, A.M.; Sun, H.; He, J.; Yang, M.; Bai, F.; Chen, Q. A plant-produced vaccine protects mice against lethal West Nile virus infection without enhancing Zika or dengue virus infectivity. Vaccine 2018, 36, 1846–1852. [CrossRef] 26. Feige, M.J.; Hendershot, L.M.; Buchner, J. How antibodies fold. Trends Biochem. Sci. 2010, 35, 189–198. [CrossRef] 27. Baumal, R.; Potter, M.; Scharff, M.D. Synthesis, assembly, and secretion of gamma globulin by mouse myeloma cells. 3. Assembly of the three subclasses of IgG. J. Exp. Med. 1971, 134, 1316–1334. [CrossRef] 28. Huang, Z.; Chen, Q.; Hjelm, B.; Arntzen, C.; Mason, H. A DNA replicon system for rapid high-level production of virus-like particles in plants. Biotechnol. Bioeng. 2009, 103, 706–714. [CrossRef] 29. Phoolcharoen, W.; Bhoo, S.H.; Lai, H.; Ma, J.; Arntzen, C.J.; Chen, Q.; Mason, H.S. Expression of an immunogenic Ebola immune complex in Nicotiana benthamiana. Plant Biotechnol. J. 2011, 9, 807–816. [CrossRef][PubMed] 30. Dietrich, C.; Maiss, E. Fluorescent labelling reveals spatial separation of potyvirus populations in mixed infected Nicotiana benthamiana plants. J. Gen. Virol. 2003, 84, 2871–2876. [CrossRef][PubMed] Int. J. Mol. Sci. 2020, 21, 2140 11 of 12

31. Chen, Q.; Lai, H. Gene delivery into plant cells for recombinant protein production. Biomed. Res. Int. 2015, 2015, 932161. [CrossRef][PubMed] 32. Lico, C.; Capuano, F.; Renzone, G.; Donini, M.; Marusic, C.; Scaloni, A.; Benvenuto, E.; Baschieri, S. Peptide display on Potato virus X: Molecular features of the coat protein-fused peptide affecting cell-to-cell and phloem movement of chimeric virus particles. J. Gen. Virol. 2006, 87, 3103–3112. [CrossRef][PubMed] 33. Iqbal, M.; Tao, Y.; Xie, S.; Zhu, Y.; Chen, D.; Wang, X.; Huang, L.; Peng, D.; Sattar, A.; Shabbir, M.A.B.; et al. Aqueous two-phase system (ATPS): An overview and advances in its applications. Biol. Proced. Online 2016, 18, 18. [CrossRef][PubMed] 34. Show, P.-L.; Ling, T.-C.; Lan, J.C.-W.; Tey, B.-T.; Ramanan, R.N.; Yong, S.-T.; Ooi, C.-W. Review of Microbial Lipase Purification Using Aqueous Two-phase Systems. Curr. Org. Chem. 2015, 19, 19–29. [CrossRef] 35. Hurtado, J.; Acharya, D.; Lai, H.; Sun, H.; Kallolimath, S.; Steinkellner, H.; Bai, F.; Chen, Q. In vitro and in vivo efficacy of anti-chikungunya virus monoclonal antibodies produced in wild-type and glycoengineered Nicotiana benthamiana plants. Plant Biotechnol. J. 2019.[CrossRef][PubMed] 36. Yang, M.; Sun, H.; Lai, H.; Hurtado, J.; Chen, Q. Plant-produced Zika virus envelope protein elicits neutralizing immune responses that correlate with protective immunity against Zika virus in mice. Plant Biotechnol. J. 2018, 16, 572–580. [CrossRef] 37. He, J.; Lai, H.; Engle, M.; Gorlatov, S.; Gruber, C.; Steinkellner, H.; Diamond, M.S.; Chen, Q. Generation and Analysis of Novel Plant-Derived Antibody-Based Therapeutic Molecules against West Nile Virus. PLoS ONE 2014, 9, e93541. [CrossRef] 38. Leuzinger, K.; Dent, M.; Hurtado, J.; Stahnke, J.; Lai, H.; Zhou, X.; Chen, Q. Efficient Agroinfiltration of Plants for High-level Transient Expression of Recombinant Proteins. J. Vis. Exp. 2013.[CrossRef] 39. Yang, M.; Lai, H.; Sun, H.; Chen, Q. Virus-like particles that display Zika virus envelope protein domain III induce potent neutralizing immune responses in mice. Sci. Rep. 2017, 7, 7679. [CrossRef] 40. Dent, M.; Hurtado, J.; Paul, A.M.; Sun, H.; Lai, H.; Yang, M.; Esqueda, A.; Bai, F.; Steinkellner, H.; Chen, Q. Plant-produced anti-dengue virus monoclonal antibodies exhibit reduced antibody-dependent enhancement of infection activity. J. Gen. Virol. 2016, 97, 3280–3290. [CrossRef] 41. He, J.; Peng, L.; Lai, H.; Hurtado, J.; Stahnke, J.; Chen, Q. A Plant-Produced Antigen Elicits Potent Immune Responses against West Nile Virus in Mice. Biomed. Res. Int. 2014, 2014, 10. [CrossRef] 42. Lai, H.; He, J.; Engle, M.; Diamond, M.S.; Chen, Q. Robust production of virus-like particles and monoclonal antibodies with geminiviral replicon vectors in lettuce. Plant Biotechnol. J. 2012, 10, 95–104. [CrossRef] [PubMed] 43. Huang, Z.; Phoolcharoen, W.; Lai, H.; Piensook, K.; Cardineau, G.; Zeitlin, L.; Whaley, K.; Arntzen, C.J.; Mason, H.; Chen, Q. High-level rapid production of full-size monoclonal antibodies in plants by a single-vector DNA replicon system. Biotechnol. Bioeng. 2010, 106, 9–17. [CrossRef][PubMed] 44. Chen, Q. Glycoengineering of plants yields glycoproteins with polysialylation and other defined N-glycoforms. Proc. Natl. Acad. Sci. USA 2016, 113, 9404–9406. [CrossRef][PubMed] 45. Hiatt, A.; Bohorova, N.; Bohorov, O.; Goodman, C.; Kim, D.; Pauly, M.H.; Velasco, J.; Whaley, K.J.; Piedra, P.A.; Gilbert, B.E.; et al. Glycan variants of a respiratory syncytial virus antibody with enhanced effector function and in vivo efficacy. Proc. Natl. Acad. Sci. USA 2014, 111, 5992–5997. [CrossRef][PubMed] 46. Fulton, A.; Lai, H.; Chen, Q.; Zhang, C. Purification of monoclonal antibody against Ebola GP1 protein expressed in Nicotiana benthamiana. J. Chromatogr. A 2015, 1389, 128–132. [CrossRef][PubMed]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).