Members of the enterovirus genus are promising oncolytic agents. Their morphogenesis involves the generation of both genome-packed infectious capsids and empty capsids. The latter are typically considered as an impurity in need of removal from the final product.
Trang 1Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/chroma
Spyridon Konstantinidisa, ∗, Murphy R Poplyka, Andrew R Swartza, Richard R Rustandib,
Rachel Thompsonb, Sheng-Ching Wanga
a Vaccine Process Research and Development, Merck & Co., Inc., Rahway, NJ, USA
b Analytical Research and Development, Merck & Co., Inc., Rahway, NJ, USA
a r t i c l e i n f o
Article history:
Received 7 April 2022
Revised 15 June 2022
Accepted 16 June 2022
Available online 17 June 2022
Keywords:
Enterovirus
Empty capsids
Cation exchange chromatography
High throughput
Oncolytic virus
a b s t r a c t
Membersoftheenterovirusgenusarepromisingoncolyticagents.Theirmorphogenesisinvolvesthe gen-erationofbothgenome-packedinfectiouscapsidsandemptycapsids.Thelatteraretypicallyconsidered
asanimpurityinneedofremovalfromthefinalproduct.Theseparationofemptyandfullcapsidscan takeplace withcentrifugation methods,whichareoflow throughputandpoorly scalable,orscalable chromatographicprocesses,whichtypicallyrequirepeakcuttingandasignificanttrade-off between pu-rityandyield.Herewedemonstratetheapplicationofpackedbedcationexchange(CEX)column chro-matographyforthe separationofemptycapsids frominfectiousvirionsfor aprototype strain of Cox-sackievirusA21.Thisseparation wasdevelopedusinghighthroughputchromatographytechniquesand scaledupasabindandelutepolishingstep.Theseparationwasrobustoverawiderangeofoperating conditionsandreturnedhighlyresolvedemptyandfullcapsids.TheCEXstepcouldbeoperatedinbind andeluteorflowthroughmodewithsimilarselectivityandreturnedyieldsgreaterthan70%forfull ma-ture virusparticles Similarperformance wasalsoachievedusingaselectionofotherbeadbasedCEX chromatographymedia,demonstratinggeneralapplicabilityofthistypeofchromatographyfor Coxsack-ievirusA21purification.TheseresultshighlightthewideapplicabilityandexcellentperformanceofCEX chromatographyforthepurificationofenteroviruses,suchasCoxsackievirusA21
© 2022MerckSharp&DohmeCorp.,asubsidiaryMerck&Co.,Inc.,Kenilworth,NJ,USA.Publishedby
ElsevierB.V ThisisanopenaccessarticleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/4.0/)
1 Introduction
Enteroviruses belong to the picornavirus family and are non-
enveloped viruses with a diameter of 27–30 nm They have a pos-
itive sense single stranded RNA genome of ∼7.4 kb long encoding
structural and non-structural proteins necessary for their replica-
tion [ 1, 2] Such viruses are attractive oncolytic agents employed
in cancer treatment [2–4] Recently, a prototype strain of Coxsack-
Abbreviations: Abs., Absorbance; AC, Affinity capture; BSA, Bovine serum albu-
min; C salt , Salt concentration in gradient; C salt,o , Starting salt concentration in gradi-
ent; CCCH, Clarified cell culture harvest; CEX, Cation exchange; CV, Column volume;
CV elution , Column volume number in the elution phase of a column; CVA, Coxsack-
ievirus; E1 – E3 , Elution pools 1 – 3; ED, Effective dose; FT1 – FT5 , Flowthrough
pools 1 – 5; GSH, Glutathione; HCP, Host cell protein; HT, High throughput; IEX,
Ion exchange; PAGE, Polyacrylamide gel electrophoresis; SDS, Sodium dodecyl sul-
fate; S , Strip pool; sd, Standard deviation; VP0 – 4, Viral polypeptide 0 - 4; W , Wash
pool
∗ Corresponding author
E-mail address: spyridon.konstantinidis@merck.com (S Konstantinidis)
ievirus A21 (CVA21) was also demonstrated as a potentially novel therapeutic for bladder cancer [5]in addition to other cancers in- volving tumors overexpressing the cell surface receptor intercellu- lar adhesion molecule 1 [6] The capsids of full mature enterovirus virions are composed of 60 copies of four viral proteins (VP) VP1– VP4 arranged in a shell that packages the RNA genome The gen- eration of such full mature virus virions is the result of a com- plex morphogenesis comprised of multiple steps [ 7, 8] Upon re- ceptor binding and delivery of the RNA genome, the nascently expressed P1 polyprotein is cleaved by virally encoded proteases into VP0, VP1 and VP3 These associate to form protomers ([(VP0, VP1, VP3) 1]) which are then assembled into pentamers ([(VP0, VP1, VP3) 5]) Pentamers can assemble into either empty procap- sids ([(VP0, VP1, VP3) 5] 12) or encapsidate the replicated genome to form provirions ([(VP0, VP1, VP3) 5] 12RNA)) Provirions undergo an autocatalytic cleavage of VP0 into VP2 and VP4, and the accrued re-arrangement of capsid proteins results into stable icosahedral full mature virions ([(VP1, VP2, VP3, VP4) 5] 12RNA)
https://doi.org/10.1016/j.chroma.2022.463259
0021-9673/© 2022 Merck Sharp & Dohme Corp., a subsidiary Merck & Co., Inc., Kenilworth, NJ, USA Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
Trang 2Virus maturation can be affected by a plethora of factors [8]and
in vitro cell culture production of enteroviruses, such as CVA21,
may not lead exclusively to the assembly of full mature virions,
which are the infectious particles displaying the desired oncolytic
activity; instead non-infectious particles may be assembled, such
as empty procapsids The latter can potentially elicit undesired
immune responses and are often the subject of scrutiny from
regulatory authorities [9] The separation of full mature virions
from empty procapsids, or full particles/capsids from empty par-
ticles/capsids, is a challenging task typically achieved by differ-
ential centrifugation which exploits buoyancy density differences
between the particles [ 10, 11] Such separations are, however, not
desirable for large scale bioprocessing [ 12, 13] Hence, alternative
routes for purifying full mature virions from empty procapsids, in
addition to process (e.g., host cell proteins) and product related im-
purities, are sought Empty procapsids are routinely encountered
during the processing of adeno-associated virus vectors and here
their separation from full particles has been achieved by employing
ion exchange chromatography (e.g., [14–16]) However, this separa-
tion method results in closely-eluting full and empty particle peaks
and requires peak cutting which is challenging to implement at
manufacturing scale and can result in yield losses in favor of purity
[17] Similarly, chromatography-based purifications of enteroviruses
return low product yields and focus predominantly on the reduc-
tion of process related impurities [ 18, 19]
Recently, a novel glutathione affinity chromatography (GSH AC)
capture step has been shown to purify CVA21 from clarified cell
culture harvests [20] However, upstream process conditions at in-
fection can challenge this step due to an undesired co-elution of
both CVA21 full mature virions and empty procapsids Here, we
report the deployment of cation exchange (CEX) chromatography
as a polishing step for the purification of CVA21 The virus is pro-
duced in adherent cell culture and purified in a three-column pro-
cess, which employs the GSH AC step, an intermediate ion ex-
change (IEX) chromatography step, and the CEX polishing step
High throughput chromatography techniques were employed to
develop the CEX-based polishing step and to generate information
regarding its wide applicability It is demonstrated that CEX chro-
matography can be deployed robustly in either bind and elute or
flowthrough mode, returning in both cases mature virions in high
yields while eliminating empty procapsids from the resultant prod-
uct pool When deployed in bind and elute mode, the polishing
step eluted full mature virions in a concentrated form and the sep-
aration displayed baseline resolution from empty procapsids The
scalability of the CEX polishing step was also demonstrated and it
was furthermore shown that efficient and effective purification of
the CVA21 full particles from empty procapsids could be obtained
using a diverse selection of cation exchange resins These results
serve to demonstrate the value of cation exchange chromatography
in the purification of enteroviruses, such as CVA21
2 Materials and methods
In this work, unless specified otherwise, chemicals and buffers
were from Sigma-Aldrich® (MO, USA), chromatography resins,
columns, stations, and 96 well PreDictor TM plates were from Cy-
tiva (Uppsala, Sweden), robotic stations and pertinent components
were from Tecan Group Ltd (Männedorf, Switzerland), and virus
stocks were acquired from ATCC (VA, USA)
2.1 Material generation
2.1.1 Generation of CVA21 enterovirus
CVA21 was produced using 3 L BioFlo® vessels operated
through a BioFlo 320 bioprocess control station (Eppendorf, NY,
USA) Human lung fibrobast MRC-5 cells (cat 05072101; ECACC,
Salisbury, UK) were cultured on 1 g L −1 Cytodex-1 microcarriers (Cytiva) using GIBCO TM William’s E Medium (Thermo Fisher Sci- entific Inc., MA, USA), supplemented with 10% (v/v) HyClone TM Bovine Calf Serum (Cytiva), 0.1% P188, 4 mM L-glutamine, and 20
mM glucose Cell culture took place at controlled conditions of 37
°C and pH 7.2 At 3 days post bioreactor batching, and once the cells had reached >90% confluency, the bioreactor underwent a 80% media exchange into serum-free cell culture media The tem- perature of the reactor was then controlled to either 37 °C (process A) or 34 °C (process B) Upon media exchange, the cells were in- fected with a multiplicity of infection of 0.05 using a CVA21 virus stock (cat VR-860) At 4 days post infection, and at an observed cytopathic effect of >90%, the reactor was harvested The collected viral fluid was filtered with a Sartopure® GF + depth filter (Sarto- rius, Göttingen, Germany) and clarified with a Sartoclean® CA 3 μm|0.8 μm filter (Sartorius) before it was processed further Alter- natively, the clarified harvest was stored at 4 °C or -70 °C for short and long term storage, respectively, until further testing Results presented hereafter employ CVA21 material generated from pro- cess B, unless stated otherwise
2.2 Large scale GSH affinity column chromatography
GSH affinity chromatography (GSH AC) was performed as de- scribed in [20]using a GSH Sepharose® 4 FF column Briefly, the column was typically loaded with 200 column volumes (CVs) of clarified cell culture harvest (CCCH) and eluted in 5 CVs with a
15 mM Tris, pH 8.0, 100 mM NaCl, 1 mM Dithiothreitol, 1 mM GSH buffer At large scale, the collected elution pool (i.e., GSH AC product) was purified further with the application of an interme- diate IEX step, and a bind and elute CEX step Conversely, at high throughput (HT) scale, the GSH AC product was employed directly for the development of the CEX step Purified intermediates were stored at 4 °C or –70 °C for short and long term storage, respec- tively All buffers contained 0.005% polysorbate-80 (PS80) The GSH
AC capture of CVA21 produced from upstream process B, led to a co-elution of full and empty CVA21 particles Application of this step to CVA21 material generated from upstream process A led to reduced empty procapsids in the elution product pool
2.3 High throughput chromatography 2.3.1 RoboColumn chromatography
Miniature column HT chromatography experiments employed Opus® RoboColumns® (Repligen, MA, USA) on a Tecan EVO® 150 robotic station (base unit), which was equipped with an 8–channel liquid handling arm and an eccentric robot manipulator arm and operated by EVOware® v2.8 The station was fitted with short stainless-steel tips and integrated with Te-Chrom TM, Te-Shuttle TM, and Infinite® M10 0 0pro reader devices The described configu- ration allowed for up to eight chromatographic separations to
be executed in parallel in a process described in [21] Here, all buffers and solutions contained 0.005% PS80, and all chromatog- raphy phases were run at a residence time of 2 min
A total of 16 RoboColumn-based separations were performed using GSH AC product as feed ( Table 1) Separations #1–14 and
#16 employed 200 μL RoboColumns and they aimed to evaluate the separation of full mature virions from empty procapsids on a selection of ion exchange resins Separation #15 employed 600 μL RoboColumns and sought to evaluate the separation of full mature virions from process related impurities Furthermore, separations
#1–6 and #10–16 were run in bind and elute mode, whereas sep- arations #7–9 were run in flowthrough mode
For all separations, fractions were collected every 200 μL in full- area UV transparent 96 well microplates (Corning Life Sciences) and read on the M10 0 0pro reader at 260 nm, 900 nm and 990
Trang 3Table 1
Details of high throughput RoboColumn chromatography separations carried out to screen the polishing purification of Coxsackievirus A21 via
bind and elute and flowthrough mode cation exchange chromatography and anion exchange chromatography as a function of resin, mobile phase
conditions, phase duration, in column volumes (CVs), and gradient slope In all separations the pH value was kept constant across all phases other
than the stripping of the columns [NaCl] depicts the NaCl concentration in the equilibration (Equil.), load, and wash buffer in each separation
This concentration was also the starting concentration in the elution gradient where applicable Separations #1–6, #10–14, and #16, employed
200 μL RoboColumns and clarified cell culture harvest (CCCH) from upstream process B Separation #15 employed 600 μL RoboColumns and
CCCH from upstream process A In all separations, the collected fractions had a nominal volume of 200 μL
Separation Resin pH [NaCl] (mM) Equil CVs Load CVs Elution CVs Strip CVs Gradient slope (mM CV −1 )
a, load prepared by diluting glutathione affinity chromatography elution (GSH AC) product 3-fold into concentrated equilibration buffer; b load
prepared by adjusting GSH AC product to desired conditions with small additions of 1 M acetic acid and 4 M NaCl stocks; c, prepared as in (a)
with the addition of bovine serum albumin and λ-DNA spikes; d, [NaCl] at end of gradient was 1500 mM; e, [NaCl] at end of gradient was 10 0 0
mM; f, gradient followed by a 3 CV step elution at 10 0 0 mM [NaCl]; g, Not applicable (NA); h, stripped with a pH 7.0, 100 mM Tris, 10 0 0 mM
NaCl buffer; i, stripped with a pH 7.5, 100 mM Tris, 10 0 0 mM NaCl buffer; j, stripped with a pH 6.0, 50 mM citrate, 10 0 0 mM NaCl buffer
nm The latter two wavelengths were used for pathlength correc-
tion purposes [22] The made measurements were employed to
construct chromatographic traces These were used to design the
pooling of the collected fractions and to identify fractions in need
of further analysis Here, the fractions were pooled in a fashion
yielding up to five pools containing flowthrough fractions ( FT1 –
FT5 ), one pool containing wash fractions ( W ), and one pool con-
taining strip fractions ( S ) The fractions collected during the elu-
tion of the RoboColumns were typically pooled in up to three dif-
ferent ways (i.e., E1 –E3 ), unless stated otherwise Pools E1 and E2
contained the fractions collected in approximately the first and
second half of the main elution peak, respectively Pool E3 con-
tained all fractions included in pools E1 and E2 in addition to
a few fractions collected after the last fraction included in pool
E2 Pooling was carried out on a separate Tecan EVO 200 robotic
station, operated by EVOware v2.8, which was equipped with an
8-channel disposable tip liquid handling arm Here, pools were
generated at a desired volume by mixing equal volumes of frac-
tions of interest, per RoboColumn, in separate wells of Thermo
Scientific TM Armadillo PCR 96-well plates (Thermo Fisher Scien-
tific Inc.) The fractions included in each pool are detailed in Table
S1 (Supporting information) Generated pools and fractions were
either analyzed immediately or stored at 4 °C or –70 °C until
their analysis Plates containing fractions and pools were sealed
with Thermo Scientific TM Nunc TM sealing tape (Thermo Fisher
Scientific Inc.)
For bind and elute chromatography, the RoboColumns were
eluted in NaCl gradients with slopes of ∼55–63 mM CV −1 follow-
ing their washing At the end of a gradient, the RoboColumns were
stripped for 5 CVs with a 100 mM Tris, pH 7.0, 10 0 0 mM NaCl
buffer unless stated otherwise The same buffer was also used in
flowthrough chromatography based separations to strip the Robo-
Columns at the end of their wash, with the exception of separation
#9 which employed a 100 mM Tris, pH 7.5, 1500 mM NaCl buffer
( Table 1) Linear elution salt gradients were simulated by multi-
step gradients wherein each step had a size of 1 CV and a salt level
( C salt) determined by the equation C salt= C salt,o+ gradient slope ×
C V elution Here, C salt,ois the salt level in the employed equilibration
and wash buffers or the load, and CV elutioncorresponds to the num- ber of CVs for which a RoboColumn was eluted for The steps in the gradient were generated by mixing low and high NaCl concen- tration buffers, per pH, in Axygen® 2.2 mL 96-well deep square well plates (Corning Life Sciences) at different ratios to obtain the desired C salt
Separation of full mature CVA21 virions from empty procapsids The full/empty CVA21 particle separation was tested in a range of mo- bile phase conditions for CEX resin Poros TM HS 50 (Thermo Fisher Scientific Inc.) Additional CEX resins Capto TM S ImpAct, Capto SP ImpRes, Capto S, Nuvia TM HR-S (Bio-Rad, CA, USA), and Nuvia S (Bio-Rad) were evaluated, along with the AEX resin Nuvia HP-Q (Bio-Rad) The CEX-based separations (i.e., #1–14) employed a 50
mM citrate buffer system at a pH range of 3.8–6.0, whereas the AEX-based separation (#16) employed a 50–70 mM Tris, pH 9.0 buffer system In both cases, the employed mobile phases included NaCl concentrations ([NaCl]) of 50 mM–1500 mM ( Table 1) The equilibration, loading, and wash chromatography phases were car- ried out at pH, buffer, and [NaCl] conditions matching those of the equilibration buffer
For separations #1–6, #10–14, and #16, the GSH AC prod- uct (feed) was diluted 3-fold in concentrated buffers to provide the load to the RoboColumns; this was loaded for 30–60 CVs ( Table1) The concentrated buffers were prepared at compositions (pH, buffer, and [NaCl]) matching those of the equilibration buffers post the 3-fold dilution of the GSH AC product For the AEX-based separation (#16), the Tris concentration was increased to 70 mM in the load compared to 50 mM in the equilibration buffer For sepa- rations #7–9, the GSH AC product was adjusted to match the equi- libration buffer with the addition of small amounts of 1 M citrate,
pH 4.0 and 5 M NaCl solutions it was loaded to the RoboColumns for 20 CVs ( Table1)
Separation of full mature CVA21 virions from process related impu- rities The described CEX bind and elute RoboColumn methodol- ogy was also employed to perform column challenge experiments These were carried out by increasing the levels of impurities pre-
Trang 4sented to the resin and determining their impact on the separation
of full mature virus particles from impurities, such as host cell DNA
and bovine serum albumin (BSA) For this purpose, 600 μL Poros
50 HS RoboColumns were employed in separation #15 ( Table 1)
They were equilibrated for 5 CVs before they were loaded for 20
CVs and washed for 5 CVs with equilibration buffer The Robo-
Columns were then eluted for 13 CVs in a multi-step NaCl gra-
dient with a slope of 57.7 mM CV −1 and stripped for 5 CVs The
mobile phases employed during the equilibration and wash of the
RoboColumns were comprised of a 50 mM citrate, pH 4.0, 250 mM
NaCl, 0.005% PS80 buffer system The strip employed a 50 mM cit-
rate, pH 6.0, 10 0 0 mM NaCl, 0.0 05% PS80 buffer The generation of
elution buffers took place as described in Section 2.3.1 to return
1/3 CV steps, each at an increasing NaCl level, while using a 50
mM citrate, pH 4.0, 10 0 0 mM NaCl, 0.0 05% PS80 buffer In these
experiments, the GSH AC product was diluted 3-fold in concen-
trated buffers to match the equilibration buffer composition post
dilution, as described in Section 2.3.1.1 The final material loaded
to the RoboColumns was then generated by spiking the diluted
GSH AC product with small volumes of concentrated BSA and λ
-DNA (Thermo Fisher Scientific Inc.) stocks to final concentrations
of 0.1 g L −1 and 200 ng mL −1, respectively These corresponded to
loading 1.2 mg BSA and 2.4 μg λ-DNA to the Poros 50 HS Robo-
Columns which represented a > 100-fold increase of such impuri-
ties in a typical GSH AC product Absorbance measurements of the
collected 200 μL fractions and their pooling took place as described
in Section2.3.1
2.4 Stability of CVA21 in chromatography mobile phase conditions
The impact of three factors on the stability of CVA21 was in-
vestigated: pH, [NaCl], and time For this purpose, GSH AC product
was used, and it was diluted 3-fold in concentrated buffers to yield
a final composition of 50 mM citrate buffers at 18 combinations of
pH (3.8, 3.9, 4.0, 4.1, 4.2, 4.5) and [NaCl] (100 mM, 400 mM, 700
mM) conditions The starting GSH AC product was also included in
this study as a control The 18 conditions and control were pre-
pared in triplicate in separate wells of an Axygen 2.2 mL 96-well
deep square well plate and upon their preparation the plate was
sealed with Thermo Scientific Nunc sealing tape and shaken at
1100 rpm for 1.5 h at room temperature At the end of the incuba-
tion period, an aliquot was taken from each well of the plate and
added to 0.5 mL Matrix TM 2D barcoded tubes (Thermo Fisher Sci-
entific Inc.) which were stored at -70 °C until their analysis The
plate was then sealed and left at room temperature for one day,
under shaking, until a new aliquot was transferred to a second set
of 0.5 mL Matrix 2D barcoded tubes, also stored at -70 °C until
their analysis All used buffers contained 0.005% PS80
2.5 Large scale cation exchange column chromatography
The Poros 50 HS CEX microscale purification method was
scaled-up large scale using a 10 cm bed height column with a bed
volume of 200 mL which was connected to an ÄKTA chromatog-
raphy station, controlled by UNICORN TM v7 The column was first
equilibrated for 4 CVs using a 50 mM citrate, pH 4.0, 400 mM NaCl,
0.005% PS80 buffer, followed up by its loading for 27 CVs The col-
umn was then washed for 4 CVs with a 25 mM citrate, pH 4.0,
50 0 mM NaCl, 0.0 05% PS80 buffer before it was eluted for 4 CVs
with a 25 mM citrate, pH 4.0, 800 mM NaCl, 0.005% PS80 buffer
Finally, the column was stripped for 4 CVs using a PBS buffer at
pH 7.0, 10 0 0 mM NaCl and 0.005% PS80 Here, the load to the col-
umn was a process intermediate obtained by purifying clarified cell
culture harvest with the GSH AC and intermediate IEX steps The
intermediate IEX product was adjusted to match the composition
of the equilibration buffer for the CEX step with the addition of
small amounts of 1 M citrate, pH 4.0 and 5 M NaCl solutions The residence time across all steps was set to 3 min, instead of 2 min used in HT scale
2.6 Sucrose density gradient centrifugation analysis of CVA21 process intermediates
The presence of CVA21 empty procapsids and full mature viri- ons in CEX chromatography loads and fractions from a large scale purification was verified via sucrose density gradient centrifugation performed Continuous sucrose gradients were prepared at 11 mL
in Polyclear ultracentrifuge tubes (Seton Scientific, CA, USA) using
15 mM Tris, pH 8.0, 150 mM NaCl, 0.005% PS80 buffers contain- ing sucrose at 15% (w/v) and 45% (w/v) Upon application of 1 mL samples to the top of the tubes, the gradients were centrifuged
at 360 0 0 rpm for 10 0 min at 4 °C using an Optima TM-SE Ultra- centrifuge (Beckman Coulter, CA, USA) Twelve fractions of equal volumes were then collected from the top of the gradients using
a piston gradient fractionator (Biocomp Instruments, Canada) and stored at 4 °C until further analysis
2.7 Analytical methods 2.7.1 Quantitative western blotting
CVA21 full mature virion (VP4) and empty procapsid (VP0) con- tents in samples were determined via quantitative western blot- ting using a Sally Sue TM system and a 12–230 kDa Sally Sue TM Separation Module kit (Protein Simple, CA, USA) Here, it needs to
be emphasized that while VP0 is included in both provirions and empty procapsids, the presence of the former in the purified CCCH samples is expected to be negligible [20] Therefore the VP0 mea- surements were indicative of the presence of empty propcapsids in tested samples Samples were prepared using an Anti–Rabbit De- tection Module (Protein Simple) according to the manufacturer’s protocol and were denatured in a Mastercycler® Gradient (Eppen- dorf) for 5 min at 95 °C For their analysis, an anti–VP4 rabbit pAb (Lifetein LLC, NJ, USA), diluted to 20 μg mL −1 in Antibody Diluent
2 (Protein Simple), was used Upon their preparation, the samples were loaded to the capillaries for 9 sec, separated for 40 min at
250 V, and immobilized for 250 sec This was followed by their exposure to antibody diluent for 23 min, to anti–VP4 rabbit pri- mary antibody for 30 min, and to the anti–rabbit secondary anti- body for 30 min The capillaries were then imaged with the chemi- luminescence detection settings and a 8 exposure time setting Peaks were integrated with a dropped lines method All samples were diluted with a concentrated Tris, pH 7.5 buffer, 0.005% PS80
to a final composition of ∼150 mM Tris, pH 7.5, 0.005% PS80 prior
to their analysis Assay results were employed to determine yields via mass balancing
2.7.2 Infectivity assay
An automated, high-throughput viral imaging infectivity assay was used to measure CVA21 potency Briefly, in this assay, the tested samples and a CVA21 positive reference control were used
to infect confluent 384 well tissue culture cell plates, which were planted with SK-MEL-28 cells (cat HTB-72; ATCC) Upon their in- fection and incubation, the plates were fixed, permeabilized, and stained with Hoechst 33342 (nuclei stain) (cat H3570; Thermo Fisher Scientific Inc.) Subsequently, the cells in the plates were immunostained with purified rabbit anti-CVA21 pAb (National Bi- ologics Laboratory) and labeled with Alexa Fluor® 488 AffiniPure donkey anti-rabbit IgG (cat 711-545-152; Jackson ImmunoResearch Inc, PA, USA) The plates were then imaged for the stained nuclei and the fluorescently tagged viral protein on a BioTek Cytation TM3 reader (Agilent Technologies, CA, USA) The images were analyzed
on the reader’s software to count total (nuclei stain) and infected
Trang 5(tagged viral protein) cells and these counts were used to calculate
the percentage of infected cells in each well of a tested plate A
dose response curve was then generated from the estimated per-
centage of infected cells for each sample and for the CVA21 ref-
erence standard in order to calculate the associated effective dose
(ED) 50 Finally, the relative potency for each test sample was de-
termined by taking the ratio between a sample’s ED50 to the refer-
ence control’s ED50 and reporting it as a percentage (%Response)
2.7.3 SDS-PAGE
Samples were analyzed via gel electrophoresis using NuPAGE TM
12% Bis-Tris 1.0 mm gels (Invitrogen, CA, USA) to track CVA21
empty procapsids and full mature virus particles (VP0 and VP2, re-
spectively; VP4 could not be reliably tracked due to its molecular
weight being close to the low limit of the gel) and proteinaceous
impurities For this purpose, 700 μL of denaturing buffer was pre-
pared by mixing 200 μL of NuPAGE Sample Reducing Agent (10X)
(Invitrogen) and 500 μL of NuPAGE LDS Sample Buffer (4X) (Invit-
rogen) 14 μL and 26 μL of denaturing buffer and sample, respec-
tively, were mixed in wells of a Thermo Scientific Armadillo PCR
96-well plate This was sealed with a Thermo Scientific Nunc seal-
ing tape and centrifuged briefly at 30 0 0 rpm on a Sorvall Legend
XTR centrifuge (Thermo Fisher Scientific Inc.) The PCR plate was
then denatured in a Mastercycler Gradient (Eppendorf) for 10 min
at 70 °C Following denaturation, up to 25 μL of sample and 2 μL of
Mark12 Unstained Standard (Invitrogen) were loaded into separate
lanes of a gel The prepared gels were electrophoresed for 50 min
at 200 V in a 1X MOPS running buffer, prepared from NuPAGE
MOPS SDS Running Buffer (20X) (Invitrogen), and stained with a
Pierce TMSilver Stain Kit (Thermo Fisher Scientific Inc.) according to
the manufacturer’s protocol, with a 2 min development time Gel
images were generated on a Gel Doc TM EZ System (Bio-Rad) with
a Silver Stain autoexposure scan protocol VP0–VP4 were identified
based on their expected molecular weight and annotated where
possible by arrows
2.7.4 Total protein, DNA, and bovine serum albumin analytics
Quant-iT TM PicoGreen TM dsDNA (Invitrogen, CA, USA) and
Pierce TM Coomassie Plus (Bradford) (Thermo Fisher Scientific Inc.)
assays were deployed as per the manufacturer’s instructions BSA
quantitative western blotting analysis was performed as described
in [23]
3 Results and discussion
3.1 Identification of cation exchange chromatography for separation
of CVA21 full mature virions and empty procapsids
RoboColumn resin screening of GSH AC product from an early
static culture virus production process supported the application of
cation exchange chromatography for separating CVA21 empty pro-
capsids from full mature virions, as opposed to anion exchange
and hydrophobic interaction chromatography (data not shown)
These early results were further corroborated via screening HT
batch chromatography experiments ( Fig 1A–F) Full mature viri-
ons bound to cation exchangers Capto S ImpAct and SP ImpRes
only at pH 4.0, 420 mM NaCl ( Fig.1A and B, respectively), whereas
binding of empty procapsids to these resins was stronger across
a wider range of tested conditions ( Fig 1D and E, respectively)
Conversely, the multimodal resin Capto MMC ImpRes bound more
strongly both types of particles across a wider range of test condi-
tions ( Fig.1C and F)
The binding differences between the two particle types outlined
a pH and NaCl concentration operating space resulting in nearly
complete separation of CVA21 full mature virions from empty pro-
capsids on each resin For example, employing a binding condition
of pH 4.0, 540 mM NaCl on Capto SP ImpRes led to ∼100% and
∼13% flowthrough yields for full ( Fig.1B) and empty ( Fig.1E) par- ticles, respectively Likewise, the same condition on Capto S ImpAct led to flowthrough yields of ∼100% and ∼0% for full ( Fig.1A) and empty ( Fig.1D) particles, respectively The aforementioned operat- ing space became narrower with increased pH values suggesting that an optimal separation would need to employ mobile phases with a low pH While enteroviruses can be stable across a wide range of conditions [24], the employment of an acidic condition for separating full mature virus particles and empty procapsids, via CEX chromatography, led to concerns over potential infectivity losses for CVA21 These were addressed by the execution of a sta- bility study at room temperature which evaluated the relationship between CVA21 infectivity and factors including liquid conditions (pH and NaCl concentration) and hold duration (two time points)
3.1.1 Impact of acidic conditions on CVA21 infectivity
The performed stability study indicated an average decrease in CVA21 infectivity of 15.2% ± 10.6% between the two time points across the 18 liquid conditions tested ( Fig.1G vs H) The pH and [NaCl] effects on CVA21 infectivity were investigated based on re- gression analysis (Table S2) and they were found to differ between the two time points; after the 1.5 h hold ( Fig.1G), only NaCl con- centration had a significant and positive effect, whereas after the
28 h hold ( Fig.1H), both pH and NaCl concentration had positive and almost equal effects on CVA21 infectivity Moreover, the rela- tionship between infectivity and these two factors was stronger at the second time point compared to the first one (i.e., %R 2of ∼17% and ∼49% at the first and second time points, respectively, in Ta- ble S2) This implied that pH and [NaCl] affected CVA21 infectivity more prominently at increased hold times at room temperature The loss of infectivity as a function of time was also observed in the GSH AC product control sample which was buffered at pH 8.0 ( Fig.1G and H) The employment of one-way analysis of variance
to compare between the measured infectivities of the control sam- ple and of the tested 18 samples, at the first time point, showed no significant difference between the 19 samples (Table S3) Hence, the time dependent infectivity losses in Fig.1G and H were not specific to the tested acidic conditions alone; instead they also in- cluded inherent, short term infectivity losses for CVA21 at room temperature Based on these results, the application of CEX chro- matography to purify full mature virions form empty procapsids at acidic conditions was deemed to be a viable approach for CVA21 purification since no significant infectivity losses are expected over the short duration of the CEX step ( ∼5 h at large scale)
3.2 Poros 50 HS chromatography for separation of CVA21 full mature virions and empty procapsids
The early chromatography screening and stability testing results were followed by the characterization of the CEX-based CVA21 pu- rification using RoboColumns Here, Poros 50 HS resin was em- ployed in bind and elute mode (separations #1–6 in Table 1) due to its large pore size characteristics [25], rendering it bet- ter suited to adsorb large solutes, such as CVA21 particles The chromatograms from separations #1–6 ( Figs 2A and S1) demon- strated the excellent repeatability of the RoboColumn technique and yielded valuable information At pH 5.0 and 6.0 (separations
#5 and #6, respectively), only a single peak was observed in the elution gradient, whereas at pH 3.8–4.5 (separations #1–4, re- spectively) two peaks were observed; one in the gradient and one in the strip For each separation, the fractions in each peak were pooled into elution ( E3 ) and strip ( S ) pools (Table S1) The areas of the elution and strip peaks in the chromatograms in- creased and decreased, respectively, with increasing pH This indi- cated the presence of two solute populations with their retention
Trang 6Fig 1 Preliminary high throughput cation exchange batch chromatography screening results for Coxsackievirus A21 and its infectivity dependence on liquid conditions and
time: (A)–(C) Flowthrough yields for full mature virions, as a function of binding pH and NaCl concentration ([NaCl]), for resins Capto S ImpAct, Capto SP ImpRes, Capto MMC ImpRes, respectively; (D)–(F) Flowthrough yields for empty procapsids, as a function of binding pH and NaCl concentration ([NaCl]), for resins Capto S ImpAct, Capto SP ImpRes, Capto MMC ImpRes, respectively; (G) and (H) %Response, depicting CVA21 infectivity based on the deployed viral imaging infectivity assay, as a function of pH for
a 1.5 h and 28 h hold, respectively, at room temperature In (A)–(F) the yields (z-axis) are averages of duplicates The colorbar in (C) denotes the color scale across (A)–(C) The colorbar in (F) denotes the color scale across (D)–(F) In (G) and (H), symbols (o), ( ), and ( ♦) correspond to NaCl concentrations of 100 mM, 400 mM, and 700 mM, respectively, and symbol ( ) corresponds to a non-acidic control sample Error bars correspond to ±1 standard deviation (sd)
being strongly affected by pH; a weaker binding one, eluting in
the salt gradient, and a stronger binding one, eluting in the strip
The composition of the two populations was determined via SDS-
PAGE analysis of pools E3 and S ( Fig 3A) Here, it is noted that
this analysis did not employ concentration normalizations during
gel loading and hence its results also reflected volumetric concen-
trations as depicted by the generation of pools from the collected
fractions (Table S1) Furthermore, silver stain also stains single and
double stranded DNA and RNA (e.g., [26]) and for samples which
were rich in full mature virions this resulted to the observation
of a band at the top of the loaded gel lanes which was attributed
to genomic RNA of CVA21 For separations #1–4, E3 contained pri-
marily full mature CVA21 particles (abundant VP2 band and little
to no presence of VP0 band), whereas S contained empty procap-
sids and small amounts of full mature virions (abundant VP0 band
and presence of VP2 band) ( Fig 3A) In contrast, for separations
#5–6, E3 contained both full mature CVA21 particles and empty
procapsids, and S contained neither of the two ( Fig.3A)
The composition of these two populations was also investigated
by quantitative western blotting analyses of the FT, W, E3 and S
pools ( Fig 2B) For all six separations, the FT and W pools con-
tained ∼0% of the full and empty particles included in the loaded
GSH AC product Hence, all binding conditions, spanning a pH
range of 3.8–6.0, resulted in high binding of CVA21 particles to
Poros 50 HS Elution yields, as depicted by the E3 pool yields, var- ied between ∼75%–∼100% and increased with increasing pH Strip yields, as depicted by the S pool yields, decreased with increas- ing pH ( Fig.2B) These led to mass balance closures well in excess
of 80% for full mature CVA21 virions Hence, the elution yields of CVA21 full particles were high and varied within a narrow range However, for separations #1–6, the E3 pool yields for the empty procapsids varied between ∼0% and 60% and increased rapidly with increasing pH ( Fig.2B) Mass balance closures for these par- ticles ( ∼30%–∼70%) were poorer than those observed for the full CVA21 particles and the unaccounted empty particles were consid- ered to be irreversibly bound to the resin While this would have
a negative effect on the re-use of a column, unless the empty pro- capsids were removed through a cleaning in place strategy, here their irreversible binding was considered to be a desirable feature
of this step Hence, a separation between CVA21 full mature virions and empty procapsids was also achieved using packed bed column chromatography, and its resolution depended on pH This agreed with the early batch-based chromatography HT screens ( Fig.1A–F) The results generated from separations #1–6 provided sufficient information to derive the conclusion that attempting to purify full mature CVA21 particles from empty particles at less acidic pH con- ditions led to their co-elution in the salt gradient This was pri- marily due to pH effects on the retention of the empty procap-
Trang 7Fig 2 Bind and elute Poros 50 HS high throughput RoboColumn chromatography results for the separation of Coxsackievirus A21 full mature virions from empty procapsids:
(A) 3D plot of chromatographic traces of recorded absorbance at 260 nm as a fraction number (y-axis) from six separations (#1–6), each at a different pH (x-axis) Lines (-) and (—) denote absorbances (Abs.) and salt levels, respectively, normalized by their maximum, and symbols ( ◦) and ( ) denote duplicated experiments (R1 and R2) The z-axis is a normalized scale from 0 to 1 where 1 denotes the maximum; (B) Bar plot of elution and strip yields and mass balances for full mature virions and empty procapsids (each bar corresponds to a different pH/separation) Error bars correspond to ±1 standard deviation (sd); (C) Elution salt levels of main elution peak as a function
of pH The salt level was determined by identifying the fraction associated to the beginning of the elution peak
sids which decreased more as a function of the pH compared to
the retention of the full particles This suggested that the encap-
sidation of genome in the full particles led to an increase in their
negative net charge which led to electrostatic repulsions with CEX
resins and reduced retention compared to the empty particles This
agrees with earlier studies for a selection of picornaviruses, such
as enterovirus 71, which observed the presence of fewer negatively
charged surface patches for empty particles compared to full par-
ticles [27] The difference in the retention between the two CVA21
particle types on Poros 50 HS was exploited to define operating
windows for establishing the polishing purification step for CVA21
and to deploy it at large scale
3.2.1 Operating windows for separating CVA21 full mature virions
and empty procapsids via bind and elute chromatography
Retention trends were generated (Fig S2) using the bind and
elute chromatograms in Fig.2A to describe the interaction between
the GSH AC purified CVA21 particles and the Poros 50 HS resin
A quadratic relationship was, therefore, derived describing the de-
pendence of elution salt on pH for the main elution peak ( Fig.2C)
and binding at a salt level below the fitted line would result to the
binding of CVA21 particles, in the GSH AC product, to the Poros
50 HS resin Here, it needs be emphasized that in Fig.2C the salt
levels represent their concentration in elution buffers at the inlet
of a column Hence, a safety factor of at least ∼60 mM NaCl (i.e.,
approximately equal to the employed gradient slope in mM CV −1)
would need be considered when choosing the binding NaCl con-
centration based on these results
At a pH range of 3.8–5.0, the required elution salt levels var-
ied between ∼350 mM and ∼800 mM NaCl ( Fig.2C) These were
sufficiently high to enable the selection of conservative NaCl con- centrations in binding conditions (e.g., 200 mM–650 mM) to avoid any loss of CVA21 particles in the flowthrough, while allowing for the robust preparation of mobile phases and ease of implementa- tion The latter is particularly important when taking into consider- ation that at large scale processing the product of the GSH AC step, eluted in a 100 mM NaCl buffer, was loaded directly to the follow- ing IEX step, which was run in flowthrough mode Hence, adopting
a pH binding condition within 3.8–5.0 for the CEX polishing step would only require the pH adjustment of the intermediate product and not its dilution (typically undesired at large scale processing) While the Poros 50 HS operating pH range of 3.8–5.0 led to op- timal and robust conditions for binding of GSH AC purified CVA21 particles, the optimal elution pH range for separating full mature virions from empty procapsids was narrower Pool E3 contained
quantifiable amounts of empty procapsids ( Fig 2B) at a pH be- tween 4.2 and 5.0 ( ∼16%–∼60%) Conversely, elution pH values be- tween 3.8 ≤ pH< 4.2 led to the complete separation of CVA21 full particles from empty particles ( Figs.2B and 3A) This supported the selection of these pH values as the optimal elution pH range Such elution pH conditions had additional benefits that rendered them well suited to large scale processing SDS-PAGE analysis for sep- arations #1 ( Fig 3B) and #2 ( Fig.3C) supported that within this elution pH range, full mature CVA21 virions could be eluted in a concentrated form, and collected with robust collection windows, via the application of a single step elution method with a step at high salt level
Finally, the near optimal separation of CVA21 mature virions from empty procapsids at a pH range of 4.2–4.5 could also ren- der this pH range as a viable alternative for eluting CVA21 At such
Trang 8Fig 3 Bind and elute Poros 50 HS high throughput RoboColumn chromatography SDS-PAGE results for the separation of Coxsackievirus A21 full mature virions from empty
procapsids: (A) Gel images of GSH affinity chromatography product ( feed ), 3-fold diluted feed in concentrated equilibration buffer ( load ), elution pool 3 ( E3 ), and strip pool ( S ) for six separations (#1–6), each at a different pH; (B)–(D) Gel images of fractions comprising pool E3 for six separations (#1–6), each at a different pH, respectively In (A) and (B) text above each lane denotes the identity of the tested sample
pH conditions, pool E3 corresponded to empty procapsid elution
yields of ∼16% and ∼35%, respectively ( Fig.2B) However, for these
two conditions, the E3 pools were comprised of fractions 68–84
and 71–81, respectively (Table S1), which included fractions at the
tail of the corresponding elution peaks ( Figs.2A and S1) The late
eluting fractions contained decreasing and increasing amounts of
full mature virions and empty procapsids, respectively ( Fig.3D and
E as indicated by the intensity of the VP2 and VP0 bands) Exclud- ing a few fractions from the tail of the elution peaks (e.g., fractions
76, 77 and 79–81 for separations #3 and #4, respectively) would result in product pools free of empty procapsids, at the cost of a marginal reduction in the elution yields for full CVA21 particles; the majority of the full particles eluted in a small number of frac- tions at lower salt levels than the empty procapsids ( Fig.3D and
Trang 9Fig 4 Flowthrough mode Poros 50 HS high throughput RoboColumn chromatography SDS-PAGE results for the separation of Coxsackievirus A21 full mature virions from
empty procapsids: (A) Gel images of 3-fold diluted GSH affinity chromatography product adjusted to the desired pH and NaCl concentration conditions ( load ), flowthrough pool ( FT ), wash pool ( W ), and strip pool ( S ) for three separations (#7–9, respectively), each at a different binding condition (pH and NaCl concentration); (B) Gel images of
load and 13 fractions collected during the loading of the column for separation #9 employing a binding condition of pH 4.5 and 550 mM NaCl In (A) and (B) text above
each lane denotes the identity of the tested sample
E) Hence, purifying CVA21 full particles at a range of 4.2 ≤ pH
≤ 4.5 would be near optimal albeit with the requirement of more
stringent peak collection control to avoid co-purifying empty pro-
capsids in the CEX CVA21 product pool
3.2.2 Separation of CVA21 full mature virions from empty procapsids
via flowthrough chromatography
The stronger binding of CVA21 empty procapsids to the Poros
50 HS resin at acidic conditions, compared to its full mature
virions, made possible the purification of the GSH AC product
via flowthrough mode CEX chromatography Separations #7–9
( Table1) were carried out at pH conditions of 3.8, 4.0 and 4.5, re-
spectively (Fig S3) Here, the employed salt level was determined
based on the retention trends elucidated from the bind and elute
experiments ( Fig.2C) and the SDS-PAGE analysis results in Fig.3B,
C and E High flowthrough yields for full mature CVA21 particles
were obtained for separations #7–9 based on quantitative western
blotting analysis (i.e., 92.7% ± 2.9%, 91.7% ± 4.1% and 84.8% ± 8.0%,
respectively) The highly robust nature of CVA21 full and empty
particle separation at pH ≤ 4.0 was also supported by separations
#7 and #8; at this pH range, the tested flowthrough pools were
free of empty procapsids ( Fig.4A) Separation #9, carried out at a
pH of 4.5, led to the inclusion of a small amount of empty procap-
sids in the flowthrough product pool (13.6% ±7.3% and Fig.4A) This
corresponded to a ∼3-fold reduction in the co-purified empty pro-
capsids in the E3 pool of the bind and elute separation #4 ( Figs.2B
and 3A) The breakthrough of empty procapsids for separation #9
was tracked via SDS-PAGE analysis ( Fig 4B), which showed that
empty procapsids were flowing through at low amounts during
early stages of column loading and their abundance increased with
increasing loading Hence, their further reduction would require
fine-tuning of both pH and [NaCl] in the binding conditions instead
of modulating the amount of the loaded GSH AC product alone
Despite this, those pH conditions that were found to be optimal
in a bind and elute based purification (i.e., pH < 4.2) were also
optimal when deployed in a flowthrough mode based purification
Purifying CVA21 GSH AC product with Poros 50 HS, with op-
timal bind and elute or flowthrough mode chromatography con-
ditions, led to product pools with high full mature CVA21 parti-
cle yields and free of empty procapsids Hence, both purification
modes were viable Considering large scale unit operations, imple-
menting a polishing step in flowthrough mode is easier than in
bind and step elution However, the latter offers the significant ad- vantage of product concentration via volumetric reduction, which
is desirable for subsequent downstream unit operations This con- tributed to the selection of bind and elute Poros 50 HS chromatog- raphy for the polishing of CVA21 at large scale
3.3 Large scale CVA21 full mature virion purification via Poros 50 HS bind and elute chromatography
The polishing of CVA21 via Poros 50 HS bind and elute chro- matography, at pH of 4.0, was verified at large scale using a 200
mL column ( Fig.5A) The large scale purification process deployed the CEX polishing step after the preceding IEX and GSH AC steps Bound CVA21 particles were step-eluted from the Poros 50 HS col- umn at 800 mM NaCl This led to the observation of a single peak containing concentrated full mature CVA21 particles and no empty particles The latter were recovered during the stripping of the col- umn with a neutral pH and high salt buffer This behavior, along with the absence of any particles in the collected flowthrough ( Fig.5B), agreed with the observation made from the HT scale ex- periments ( Figs.2B and 3A, C) Good agreement was also observed across scales for full mature virion yields with large scale yields of
∼91% and ∼84% yields, based on quantitative western blotting and infectivity assays, respectively
The high product volumes generated from the large scale run enabled the use of sucrose density gradient centrifugation analy- sis to verify the CEX-based separation of full CVA21 particles from empty particles The process intermediate, which was loaded to the 200 mL Poros 50 HS column, was shown to contain both par- ticle types ( Fig.6A), whereas the Poros 50 HS elution product was free of empty particles ( Fig 6B) These results demonstrated fur- ther the scalability of the HT scale results and provided additional confirmation for the performance of the selected conditions; they led to high elution yields and a complete separation of full CVA21 particles from empty ones in a robust and easy to implement pu- rification at large scale
3.4 Purification of CVA21 full mature virions from process related impurities via Poros 50 HS bind and elute chromatography
Apart from separating full mature virions from empty procap- sids, the CEX step was also determined to flow through small
Trang 10Fig 5 Bind and elute Poros 50 HS large scale chromatography results for the separation of Coxsackievirus A21 full mature virions from empty procapsids: (A) Chromato-
graphic trace at 280 nm on the left-hand side y-axis and conductivity, pH traces on the right-hand side y-axis The x-axis represents column volumes; (B) SDS-PAGE analysis
of fractions collected across the entire loading phase ( FT ), elution phase ( E ) and strip phase ( S ) of the chromatogram in (A) Text above each lane denotes the identity of the tested sample
Fig 6 SDS-PAGE analysis of fractions collected during sucrose density gradient centrifugation for: (A) Starting material (CEX Load) purified by the large scale bind and
elute Poros 50 HS polishing step; (B) Elution product pool (CEX Elution) from the polishing step In both (A) and (B) the second lane shows the sample that was analyzed
by sucrose density gradient centrifugation and lanes B1–B12 show the fractions collected during their sucrose density gradient centrifugation analysis from the top to the bottom collected layers In (A) and (B), text above each lane denotes the identity of the tested sample Boxes with dashed lines denote the location of empty procapsids and full mature virions in the collected fractions
amounts of persistent high molecular weight proteinaceous im-
purities, which were not removed by the preceding GSH AC and
intermediate IEX steps This was observed while purifying CCCH
from upstream process A and testing the generated products with
overexposed SDS-PAGE gels (Fig S4) This trend was also observed
when purifying CCCH from upstream process B; for separation #9
( Fig 4B) the collected 20 CV flowthrough pool contained both
CVA21 particles and faint bands of protein impurities at similar
molecular weights to those observed in Fig S4 This supported fur-
ther the operation of the Poros 50 HS step in bind and elute mode
rather than flowthrough mode for CVA21 purification
The observation that the bind and elute CEX step contributed
to a further reduction of process related impurities led to the ex-
ecution of studies aiming to challenge its purification potential
Here, large amounts of BSA and λ-DNA were spiked to GSH AC
product, as described in Section 2.3.1.2, and the CEX purification
was performed with a binding condition of pH 4.0, 250 mM NaCl
and a NaCl gradient of 58 mM CV −1 BSA represented a major
process related impurity to be removed by the downstream pro- cess, present due to the inclusion of bovine calf serum in the cell culture, whereas λ-DNA represented a molecularly-distinct type of contaminant (i.e., host cell DNA) in need of removal from the final purified product It needs be highlighted that the employed GSH
AC product purified in this study was generated by upstream pro- cess A and as shown in Fig S4 it included a low amount of empty procapsids
The overlay of chromatographic traces from the HT total pro- tein and DNA assays led to the observation of three peaks in the salt gradient and two peaks in the neutral pH column strip ( Fig.7) Their pooling ( Fig.7) was followed by analytical testing to deter- mine the presence of full CVA21 particles and BSA via quantitative western blotting The former were observed only in elution pool
E2 , leading to an elution yield of 96.7% ± 5.2%. BSA was observed
in pools E2 and E4 with increased presence at higher salt levels (i.e., 17.6 μg ± 1.0 μg and 370.0 μg ± 29.3 μg, respectively, based
on quantitative western blotting analysis) Pool E1 was not tested