Scholarly article on topic 'Quantification and characterization of virus-like particles by size-exclusion chromatography and nanoparticle tracking analysis'

Quantification and characterization of virus-like particles by size-exclusion chromatography and nanoparticle tracking analysis Academic research paper on "Chemical sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Journal of Chromatography A
OECD Field of science
Keywords
{Vaccine / HPLC / "Enveloped VLP" / MALS}

Abstract of research paper on Chemical sciences, author of scientific article — Petra Steppert, Daniel Burgstaller, Miriam Klausberger, Andres Tover, Eva Berger, et al.

Abstract The rapid quantification of enveloped virus-like particles (VLPs) requires orthogonal methods to obtain reliable results. Three methods—nanoparticle tracking analysis (NTA), size-exclusion HPLC (SE-HPLC) with UV detection, and detection with multi-angle light scattering (MALS)—for quantification of enveloped VLPs have been compared, and the lower and upper limits of detection and quantification have been evaluated. NTA directly counts the enveloped VLPs, and a particle number is obtained with a lower limit of detection (LLOD) of 1.7×107 part/mL and lower limit of quantification (LLOQ) of 3.4×108 part/mL. SE-HPLC with UV detection was calibrated with standards characterized by NTA, and a LLOD of 6.9×109 part/mL and LLOQ of 2.1×1010 part/mL were found. SE-HPLC with MALS does not require a pre-calibrated sample because with a spherical model based on the Rayleigh-Gans-Debye approximation, the particle concentration can be directly deduced from the scattered light. A LLOD of 4.8×108 part/mL and LLOQ of 2.1×109 part/mL were measured and substantially lower compared to the UV method. The absolute particle concentration measured by SE-HPLC-MALS is one order of magnitude lower compared to measurement by NTA, which is explained by the wide size distribution of an enveloped VLP suspension. The model used for evaluation of light scattering data assumes monodisperse, homogeneous, and spherical particles.

Academic research paper on topic "Quantification and characterization of virus-like particles by size-exclusion chromatography and nanoparticle tracking analysis"

Accepted Manuscript

Title: Quantification and characterization of virus-like particles by size-exclusion chromatography and nanoparticle tracking analysis

Author: <ce:author id="aut0005" author-id="S0021967316317514-515f654a16bbaaea063aeeb54afabb8b"> Petra Steppert<ce:author id="aut0010" author-id="S0021967316317514-974f1ccd8ea4af83fe94dee6ea437732"> Daniel Burgstaller<ce:author id="aut0015" author-id="S0021967316317514-534beb075333198537d00cb3903bcf00"> Miriam Klausberger<ce:author id="aut0020" author-id="S0021967316317514-4dfb70d446810f3888f98232ebc4ad05"> Andres Tover<ce:author id="aut0025" author-id="S0021967316317514-8d1419894c8fec8ebaf0bf74531436d6"> Eva Berger<ce:author id="aut0030" author-id="S0021967316317514-634e62e16957c990b6957e17044f21b1"> Alois Jungbauer

PII: DOI:

Reference:

S0021-9673(16)31751-4

http://dx.doi.org/doi:10.1016/j.chroma.2016.12.085 CHROMA 358179

To appear in:

Journal of Chromatography A

Received date: Revised date: Accepted date:

14-10-2016 12-12-2016 31-12-2016

Please cite this article as: Petra Steppert, Daniel Burgstaller, Miriam Klausberger, Andres Tover, Eva Berger, Alois Jungbauer, Quantification and characterization of virus-like particles by size-exclusion chromatography and nanoparticle tracking analysis, Journal of Chromatography A http://dx.doi.org/10.10167j.chroma.2016.12.085

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Quantification and characterization of virus-like particles by size-exclusion chromatography and nanoparticle tracking analysis

Petra Steppert1, Daniel Burgstaller1, Miriam Klausberger1, Andres Tover2, Eva Berger3 and Alois Jungbauer13*

1 Department of Biotechnology, University of Natural Resources and Life Sciences Vienna, Vienna, Austria

2 Icosagen AS, Tartumaa, Estonia

3 ACIB GmbH, Muthgasse 11, Vienna, Austria

*Corresponding author Muthgasse 18 1190 Vienna Austria

Tel: +4314765479083 Fax: +4314765479009 e.mail: alois.jungbauer@boku.ac.at

Highlights

• Orthogonal quantification methods for enveloped virus-like particles

• Multiple options for virus-like particle detection

• HPLC-method for accurate quantification

• HPLC method correlates with single particle tracking analysis

Abstract

The rapid quantification of enveloped virus-like particles (VLPs) requires orthogonal methods to obtain reliable results. Three methods—nanoparticle tracking analysis (NTA), size-exclusion HPLC (SE-HPLC) with UV detection, and detection with multi-angle light scattering (MALS)—for quantification of enveloped VLPs have been compared, and the lower and upper limits of detection and quantification have been evaluated. NTA directly counts the enveloped VLPs, and a particle number is obtained with a lower limit of detection (LLOD) of 1.7*107 part/mL and lower limit of quantification (LLOQ) of 3.4*108 part/mL. SE-HPLC with UV detection was calibrated with standards characterized by NTA, and a LLOD of 6.9*109 part/mL and LLOQ of 2.1x1010 part/mL were found. SE-HPLC with MALS does not require a pre-calibrated sample because with a spherical model based on the Rayleigh-Gans-Debye approximation, the particle concentration can be directly deduced from the scattered light. A LLOD of 4.8x108 part/mL and LLOQ of 2.1 x109 part/mL were measured and substantially lower compared to the UV method. The absolute particle concentration measured by SE-HPLC-MALS is one order of magnitude lower compared to measurement by NTA, which is explained by the wide size distribution of an enveloped VLP suspension. The model used for evaluation of light scattering data assumes monodisperse, homogeneous, and spherical particles.

Keywords: vaccine; HPLC; enveloped VLP; MALS

1. Introduction

The emerging demand and interest in enveloped virus-like particles (VLPs) for vaccination or gene therapy applications has increased the need for robust particle quantification methods [1 -4]. The availability of rapid in-process control methods would considerably improve development of the research field and production processes.

Because VLPs are non-infective, they cannot be quantified by well-established virus quantification techniques, such as TCID50 or the plaque assay [5]. Alternative approaches for particle quantification rely on detection of nucleic acids or specific epitopes. Moreover, these techniques are not feasible when particles lack genomic material and specific epitopes are not presented on the particle outer surface. Therefore, enveloped VLPs are frequently characterized using a combination of multiple biochemical and biophysical methods. Biochemical methods, such as Western blot analysis or mass spectrometry, are used to identify specific structural proteins. These methods cannot distinguish between structures embedded in particles or present free in solution, however. Biophysical characterization methods measure particle morphology, size, and size distribution but often cannot discriminate among structural diversities of particles [3, 6].

Traditionally, transmission electron microscopy (TEM) is used to visualize VLPs and measure particle size. However, with TEM only dried specimens can be visualized. The sample preparation often introduces artefacts or particle deformation which can lead to misinterpretation of particle concentration and particle size [3, 7-9].

Nanoparticle tracking analysis (NTA) is a non-invasive method that visualizes particles in suspension by a focused laser beam. A video of the illuminated particles is captured and subsequently analyzed. The movement of each single particle, caused by Brownian motion, is tracked frame by frame to obtain an average mean squared displacement, which is converted into the particle diffusion coefficient; applying the Stokes-Einstein equation (Eq. 1) gives the hydrodynamic diameter of the particle:

r2 _ 4TKBt

(x,y)2=Zi^i Eq. 1

where (x, y)2 is the particle's mean squared displacement, T is the temperature, Kb the Boltzmann's constant, t the time period, n the dynamic viscosity, and dh the hydrodynamic diameter. The particle concentration is obtained when the tracked particles are counted and related to the sample volume [10]. The selectivity of tracking

and counting particles depends on the particles' light scattering intensity in contrast to the background signals [11]. This signal-to-noise ratio is highly influenced by several parameters that are adjustable by the user during capture and analysis of the video [12].

Size-exclusion chromatography (SEC) is widely used in a preparative scale for polishing viruses [14-16] and VLPs [17, 18] as well as for analytical purposes [19-21]. The size of enveloped VLPs implies that particles are usually eluted within the void fraction and that the resolution is too low to discriminate between structural diversities of particles. Traditional detection of UV absorbance requires a calibration curve relating the measured response signal to a VLP concentration, initially quantified by an orthogonal method. A substantial drawback of this method is that a particle size cannot be obtained. SEC coupled with a multi-angle light scattering (MALS) detector enables direct quantification and characterization of VLPs without preparation of a calibration curve. The MALS detector measures the excess Rayleigh ratios at the corresponding angles, which expresses the light scattering intensity of the sample in relation to solution [22, 23]. Here the particle size and number are calculated based on the particles' differential scattering intensity, which is a function of their refractive index and the radius of the scattered spherical particles (Eq. 2):

where N is the number of particles, k an optical constant dependent on the refractive index of the solution, R(Q) the Raleigh ratio at the scattering angle (Q), and ii(d) the single particle differential scattering intensity [23, 24]. To obtain the single particle differential scattering intensity, the Rayleigh-Gans-Debye (RGD) approximation can be used when the total phase shift of the incident light wave as it passes through the molecule is negligible and the refractive index of the particles is very close to the refractive index of the surrounding solution [22, 25]. Another, emerging separation method frequently coupled to MALS detector for analytical applications, is asymmetric flow field-flow fractionation (AFFF). AFFF-MALS was successfully used for separation and quantification of viruses [7, 24, 26] and VLPs [27, 28]. Comparison of particle size distribution and particle concentration measured for influenza viruses proved well correlated results between SEC-MALS and AFFF-MALS [7]. In AFFF particles are separated according to their hydrodynamic size in the

absence of a stationary phase using a separation channel with a semipermeable membrane at the bottom. The separation is achieved by flow velocity gradient of the longitudinal flow and by the cross flow towards the membrane. Compared to SEC, AFFF provides a broad, universal separation range accessible by a single channel, enabling improved resolution and separation of macromolecules or aggregates. Furthermore, large sample volumes can be applied to the AFFF channel without causing band broadening, which might be important for low concentration samples [22]. Although it is reported that unspecific sample-surface interactions are reduced compared to SEC, this might not be true for all applications. Several interacting flow parameters have to be carefully optimized to avoid unspecific interaction of samples with the membrane or sample-sample interactions, potentially promoting sample aggregation [7, 29]. Nevertheless, SEC is the more well-established and characterized method in the field of biotechnology and can be easily scaled up to an industrial scale for preparative applications. The idea behind our work was to develop SE-HPLC methods, using UV or MALS detection, for quantification of enveloped VLPs useful for in-process control during production and downstream processing. As an appropriate model, HIV-1 gag VLPs were selected [30]. They are spherical particles with diameters of about 100 to 150 nm. They consist of about 2500 HIV-1 gag monomers [31], arranged on the inner surface of the particles surrounded by the host cell-derived lipid membrane. They are produced up to high titers, and we have already developed and established purification strategies in our previous work [32]. Finally, we compared the SE-HPLC methods with NTA and examined the advantages and limitations of each.

2. Material and methods

2.1 Chemicals and buffers

D(+) Sucrose was obtained from Acros Organics (Glee, Belgium) and L-Arginine from Pierce Thermo Fisher Scientific (Waltham, MA, USA). All other chemicals were purchased from Sigma Aldrich (St. Louis, MO, USA) or Merck (Darmstadt, Germany), if not otherwise noted. All buffers were prepared with ultra-pure water and 0.1 ^m filtered using a polyether sulfone membrane (Pall Corporation, Ann Arbor, MI, USA). Sodium-phosphate buffers were prepared by mixing equimolar Na2HPO4 and NaH2PO4 to the required pH.

2.2 Chromatographic instruments

Preparative chromatography for preparation of HIV-1 gag VLP standard material was performed on an ÄKTA explorer 100 system equipped with a P-960 sample pump and a Frac-950 fraction collector (GE Healthcare, Uppsala, Sweden) and controlled by Unicorn software 10.1.

SE-HPLC experiments were performed on an Agilent Series 1100 System (Agilent, Waldbronn, Germany) consisting of a well plate automatic liquid sampler (WP ALS) for injection, a degasser, and a quaternary pump. UV absorbance (280 nm) was detected by a diode array detector. ChemStation for LC 3D systems (Rev. B. 04.03) software was used to control the HPLC system and for analysis of UV absorbance data. MALS signals were detected by a DAWN HELEOS 18-angle detector (Wyatt, Santa Barbara, CA, USA). The MALS detector was coupled to the differential refractive index detector Optilab rEX (Wyatt, Santa Barbara, CA, USA) [33-35]. MALS data were acquired by Astra 5.3.4 software, and data analysis was performed using Astra 6.1.2. For evaluation of particle size and number, the sphere model fit and number density evaluation procedure were used. A particle's refractive index of 1.46 was used for calculation [36]. To obtain the particle concentration, the particle number was related to the injection volume.

2.3 HIV-1 gag VLPs

CHO cell culture supernatant (CHOEBNALT85 cell line) containing HIV-1 gag VLPs was provided by Icosagen (Tartumaa, Estonia) [32]. The HIV-1 gag VLPs were purified by either sucrose density gradient centrifugation (20% to 60% (w/v) sucrose) or preparative anion-exchange chromatography using a 1 mL CIMmultus QA monolith (BIA Separations, Ajdovscina, Slovenia) as described in detail previously [32]. Briefly, for the purification of HIV-1 gag VLPs by density gradient centrifugation, HIV-1 gag VLPs were pelleted through 20% (w/v) sucrose at 77,100 xg for 2.5 h at 4°C (Beckmann L8-80M ultracentrifuge (Indianapolis, IN, USA) equipped with a SW41Ti rotor (Brea, CA, USA), then re-suspended in phosphate-buffered saline (PBS) pH 7.4, loaded onto the sucrose gradient, and centrifuged at 93,500 xg for 17.5 h at 4°C. The band containing the purified HIV-1 gag VLPs was collected between densities of 1.16 and 1.18 g/cm3 (measured at 25°C) and yielded in 9.1x1010 part/mL when analyzed by NTA. Preparative anion-exchange chromatography, using a 1 mL CIMmultus QA monolith, was performed on an ÄKTA explorer 100 controlled by Unicorn software

5.10 (GE Healthcare, Uppsala, Sweden). The system was equipped with a P-960 sample-pump and a Frac-950 fraction collector. The experiments were performed at a flow rate of 1 mL/min using 50 mM HEPES, 350 mM NaCl pH 7.5 as mobile phase A and 50 mM HEPES, 2 M NaCl pH 7.5 as mobile phase B. The column was equilibrated with 15 CV of mobile phase A, before 50 mL of 0.8 ^m filtered (Millex AA filter, Millipore, Bedford, MA, USA) CHO cell culture supernatant was loaded. Elution was achieved after a wash step (mobile phase A for 15 bed volumes) by 0-25-45% B steps with a hold of 15 bed volumes each. Regeneration was performed with 100% B for 30 bed volumes followed by sanitization with 1 M NaOH for 30 bed volumes. 1 mL fractions were collected by the fraction collector and pooled according to the chromatogram the purified HIV-1 gag VLPs (9.8*1010 part/mL measured by NTA)were collected during elution within the 45% B step.

2.4 SE-HPLC

A TSKgel G5000PWXl column (300.0 mmx7.8 mm i.d.) in combination with a TSKgel PWXL guard column (40.0 mm*6.0 mm i.d.) (Tosoh Bioscience, Stuttgart, Germany) was used. For preliminary chromatographic experiments, 100 ^L of samples was injected, PBS pH 7.4 was used for isocratic elution, and a flow rate of 0.2 mL/min was applied. Elution fractions were collected manually according to the chromatogram. During method development, buffer screening experiments were performed using 12 buffers summarized in Table 1. The elution buffer (l) was selected, and the flow rate was adjusted to 0.3 mL/min for calibration and evaluation of the process parameters of the method. The HIV-1 gag VLP standard material or samples were optionally diluted in the elution buffer, and 25 or 50 ^L was injected. All experiments were performed at 25°C.

2.5 Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis

Electrophoresis (200 V, 400 mA for 50 min) was performed under reduced MES-SDS running conditions using NuPAGE® Bis/Tris Mini gels 4-12% (Invitrogen, Carlsbad, CA, USA) and an X-cell SureLock® Mini-Cell (Invitrogen, Carlsbad, CA, USA) electrophoresis chamber. Samples were prepared with NuPAGE® LDS sample buffer (Invitrogen, Carlsbad, CA, USA), reduced with 100 mM dithiothreitol, and heated for 10 min at 95°C, and 10 ^L of each sample was loaded per band. SeeBlue® Plus2

Pre-stained Protein Standard (Invitrogen, Carlsbad, CA, USA) was used as the protein molecular weight ladder. Protein bands were stained using a silver staining procedure as described by Heukeshoven et al. [37] or transferred (40 V, 200 mA for 2 h) onto a 0.2 mm nitrocellulose membrane (Whatman, Dassel, Germany) for Western blot analysis, using 50 mM sodium borate, 0.1% (w/v) SDS, and 20% (v/v) methanol as transfer buffer. The membrane was blocked overnight with 3% bovine serum albumin (BSA) in PBS-T (0.1% (w/v) Tween-20 in PBS) at 4°C. The membrane was incubated with monoclonal mouse anti-HIV-1 p24 (Icosagen AS, Tartumaa, Estonia) at a 1:1000 dilution in 1% BSA in T-PBS for 2 h. Secondary antibody incubation was performed with anti-mouse IgG conjugated with alkaline phosphatase (Sigma Aldrich, St. Louis, MO, USA) at a 1:1000 dilution in PBS-T with 1% BSA for 1 h. Visualization of bands was carried out by Lumi PhosTM (Thermo Fisher Scientific, Waltham, MA, USA) on the Lumi Imager (Boehringer Ingelheim, Ingelheim, Germany).

2.6 Total protein content and dsDNA

The total protein content was determined by the Bradford assay in a 96-well plate format using a Coomassie blue G-250-based protein dye reagent (Bio-Rad Laboratories, Hercules, CA, USA) in accordance with the manufacturer's instructions. Calibration curves were obtained using BSA standards diluted in TE-Buffer. The dsDNA content was determined with the Quant-iTTM PicoGreen® dsDNA kit (Invitrogen, Carlsbad, CA, USA) in a 96-well plate format according to the manufacturer's instructions. Signals for both methods were measured using a Genius Pro plate reader (Tecan, Mannedorf, Switzerland).

2.7 NTA

Particle concentration of HIV-1 gag VLP standard material and samples was measured by NTA using a NanoSight LM-10 instrument equipped with a blue laser (405 nm) and NTA 2.0 software (Malvern Instruments Ltd., Worcestershire, UK). For each sample, three serial dilutions were prepared in particle-free water to achieve an average between 20 and 60 particles/frame. The camera level was adjusted manually, and 60 s videos were captured at room temperature. Each video was analyzed three times for particles with diameters between 100 to 200 nm using three sets of optimized analysis parameters, which were kept constant during the measurements.

2.8 Endonuclease treatment

The endonuclease treatment protocol was developed in-house based on the manufacturer's recommendations. Samples were optionally spiked with 1 pg lambda DNA (Invitrogen, Carlsbad, CA, USA), diluted 1:20 in 1.08 M Tris-HCl, 1.08 mM MgCl2, and 80 mM NaCl, pH 8.0, and treated for 1 h with 150 U/mL Benzonase purity grade II (Merck KgA, Darmstadt Germany) at 37°C. Blank samples were diluted in the same amount of buffer and incubated for 1 h at 37°C, but no Benzonase was added. Digestion of samples and blank samples was stopped by 5 mM EDTA.

2.9 TEM

The samples were incubated for 1 min on 400-mesh copper grids (Agar Scientific Ltd, Stansted, UK) coated with Pioloform film and shaded with carbon. Sample fixation was performed for 15 min with 2.5% glutaraldehyde solution in 100 mM cacodylate buffer, pH 7.0. After three wash steps with water, samples were negatively stained with 1% aqueous uranyl acetate solution for 30 s [38]. The air-dried specimens were analyzed in a Tecnai G2 200 kV transmission electron microscope (FEI, Eindhoven, Netherlands), operating at 80 keV.

3 Results and discussion

We aimed to develop and characterize HPLC-based analytical methods for quantification of HIV-1 gag VLPs from CHO cell culture supernatant and purified samples. After separation of HIV-1 gag VLPs from impurities by SEC, HIV-1 gag VLPs were detected and quantified using UV absorbance signals measured at 280 nm or MALS signals.

3.1 NTA and characterization of HIV-1 gag VLP standard material

For development of a SEC-HPLC method, standard material purified by density gradient centrifugation or anion-exchange chromatography was used. The standard material was characterized and analyzed by TEM, NTA and the total protein concentration was determined by Bradford assay. An average particle size of 140±23 nm (n=79) was measured from representative TEM pictures, which clearly demonstrated the presence of spherical VLP-like structures (Fig. 1A). The measured sizes are comparable to previously reported values [30, 39]. NTA of purified materials

resulted in an average hydrodynamic diameter of 176±66 nm (n=26). Particle sizes determined by TEM are usually underestimated compared to NTA because of shrinkage of specimens during sample fixation and staining [12, 40]. Particle concentration was counted by NTA. The chromatographic purified standard material consisted of 9.8*1010 part/mL, equivalent to a total protein concentration of 230.6 pg/mL resulting in an average protein content of 2.4 pg/109 particles. The particle concentration measured in the material purified by density gradient centrifugation was 9.1*1010 part/mL, equivalent to a total protein concentration of 91.1 pg/mL resulting in a protein content of 1.0 pg/109 particles.

Prior to calibration of HPLC, the limits of detection and quantification of NTA were evaluated because the concentrations of particles in the standard and in samples were counted by NTA. Based on recommendations in the literature [11, 41], the lower limit of detection (LLOD) was calculated assuming that only one single particle is present per video frame. A minimum of 20 particles for the lower limit of quantification (LLOQ) and a maximum of 60 particles per video frame for the upper limit of quantification (ULOQ) were assumed (Table 2). The number of particles/frame must be between 20 and 60 to enable reliable particle quantification by NTA.

3.2 Method development

Preliminary chromatographic experiments were performed by injecting 100 pL of HIV-1 gag VLPs present in crude CHO supernatant or purified by density gradient centrifugation (Fig. 2A). HIV-1 gag VLPs could be eluted within the void fraction of the column, during the first peak at 6.1 mL (Fig. 2A). The presence of HIV-1 gag VLPs in this fraction was confirmed by SDS-PAGE (Fig. 2B), Western blot (Fig. 2C), and TEM (Fig. 1B, C). SDS-PAGE (Fig. 2B) clearly demonstrates the presence of bands at 55 kDa, which were identified as gag polyprotein by Western blot analysis using anti-p24 for detection (Fig. 2C). HIV-1 gag VLPs present in CHO supernatant could be clearly resolved from other impurities, which eluted after the VLPs. Separation of endonuclease-treated CHO supernatant indicated that the majority of DNA impurities eluted between 8 and 11.5 mL because decreased signals were detected in this region compared to non-endonuclease-treated CHO supernatant (Fig. 2A). The TEM pictures from elution fractions (Fig. 1B, C) confirmed that intact spherical particles eluted from the SEC column, and an average particle diameter of 140±23 nm (n=13) was estimated from these pictures. These values are similar to those obtained from

measurements of HIV-1 gag VLPs present in the standard material (Fig. 1A). The SEC elution fractions could not be quantified by NTA because the sample volume was insufficient to perform accurate particle concentration measurements based on three serial dilutions.

3.2.1 Buffer screening

For method optimization, the flow rate was increased to 0.3 mL/min. Furthermore, buffer screening experiments were performed to avoid adsorption caused by electrostatic or hydrophobic interaction of HIV-gag-1 VLPs with the stationary phase. These secondary effects are an important issue when membrane proteins, lipids, or enveloped viruses are purified or analyzed by SEC [42-45]. Partial adsorption and increased retention were observed when HIV-1 gag VLP standard material purified by AIEX chromatography was injected in the presence of PBS buffer (data not shown). We screened different pH values from pH 6.0 to pH 9.0 (Fig. 3A), evaluated the impact of different additives (Fig. 3B), and analyzed the effect of 250 mM NaCl compared to 500 mM NaCl (Fig. 3C). The UV absorbance signals of HIV-1 gag VLPs were analyzed in CHO supernatant or in purified material and were stable over the entire pH range in the presence of 500 mM NaCl (Fig. 3A). The minor deviations of the measured response signals were within the expected experimental error (8.7% relative standard deviation (RSD) for injections of CHO supernatant and 11.5% RSD for injections of purified HIV-1 gag VLPs). Because high ionic strength might promote hydrophobic interactions with the hydrophobic stationary phase (methacrylate), mobile phase modifiers and surfactants (i-propanol, Tween 80, arginine, and urea), which potentially prevent these interactions, were tested. Additionally, the effects of additives stabilizing the HIV-1 gag VLP structure, such as EDTA or sucrose, were investigated. Addition of i-propanol to the buffer reduced the UV response signal drastically, indicating that the lipid membrane of HIV-1 gag VLPs might be affected, destroyed, or precipitated by this solvent. The presence of other additives did not have a substantial impact on the UV absorbance signals, however (Fig. 3B). This result suggests that electrostatic interactions were the main reason for the observed adsorption effects. We concluded that 250 mM was the minimal NaCl concentration needed to avoid electrostatic interactions because about 140 mM NaCl, as present in PBS, was insufficient to suppress adsorption, and further increasing NaCl concentrations to 500 mM did not further improve suppression of the interaction

(Fig. 3C). Based on these results, 25 mM Na-phosphate supplemented with 250 mM NaCl, pH 8.0, was selected as elution buffer.

3.3 SEC-UV method

HIV-1 gag VLP standard material purified by AIEX chromatography was used for preparation of calibration curves and to determine the performance parameters of the method. Two individual calibration curves for injection volumes of 25 or 50 pL were prepared. For each injection volume, three individual calibration curves were prepared on different days using seven different HIV-1 gag VLP standards ranging from 1.4x1010 to 6.5*1010 part/mL (Fig. 4A). The run time of 90 min per sample did not allow measurement of several replicates within one day. The peak area under the UV absorbance curve at 280 nm eluting at 6.1 mL was numerically integrated. When 25 pL was injected, linearity was confirmed by F-test statistics [46] at the 95% confidence level from 1.4x1010 to 6.5x1010 part/mL, with an average slope of 7.6x10-9, an average intercept of -55.4, and an average correlation coefficient of 0.9907 (Fig. 4B). For 50 pL injections, linearity was confirmed by F-test statistics at the 95% confidence level only between 1.4x1010 to 4.9x1010 part/mL, with an average slope of 1.4x10-8, an average intercept of -33.3, and an average correlation coefficient of 0.9864 (Fig. 4B). The residuals were equally distributed, and the RSD was lower than 15% in both cases (data not shown). The LLOD and LLOQ for the SEC-UV method were calculated according to the recommendations made by ICH Q2 [47] using the approach based on the standard deviation of the response signal and the slope of the calibration curve [47]; the limits are expressed as LLOD=3a/S and LLOQ=10ct/S, with o as the residual standard deviation and S as the slope of the calibration curve. The ULOQs were defined from the highest concentration reproducibly analyzed (RSD <15%) within the linear range of the method. An injection volume of 50 pL enabled detection and quantification in a lower concentration range (2.1x1010 to 4.9x1010 part/mL), whereas an injection volume of 25 pL allowed quantification of HIV-1 gag VLPs up to 6.5x1010 part/mL (Table 2). The absolute measurement range of the SEC-UV (2.1x1010 to 6.5x1010 part/mL) method was in a concentration range about 100 times higher than that of NTA (3.4x108 to 1.0x109 part/mL). While the measurement concentration range of NTA was about 1.5 times wider (Table 2).

3.4 SEC-MALS method

The second quantification method was detecting SEC eluates by the MALS. The light scattering detector response was converted into a geometric radius, and the particle number was calculated using the sphere model fit and the number density procedure based on the RGD approximation. The key prerequisite for applying RGD approximation—that the relative refractive index of the scattering particles is close to unity—was tested by comparing the change in the differential refractive index between elution buffer and HIV-1 gag VLPs. The measured value was very close to unity (lower than 1.0001), and the prerequisite was fulfilled. The MALS signals were proportional to the injected amount of particles, and the average diameter of HIV-1 gag VLPs eluted at 6.1 mL was 186±22 nm (n=7, Fig. 5A), which is comparable to the particle sizes measured by NTA (176±66 nm). The narrower particle size distribution measured by MALS might be ascribed to the separation of impurities by SEC before particle quantification. Although the MALS detection directly provides the information about the particle number eluting within the designated peak, values were always about 1 log smaller than particle concentrations obtained from NTA measurements (Fig. 5B).

The difference in the values obtained by NTA and MALS will be discussed in detail in section 3.6 Comparison of SEC-UV, SEC-MALS, and NTA. However, to relate the particle concentrations detected by MALS to the values measured by NTA, a linear correlation between both particle measurement methods was made (Fig. 5B). The validity of the linear model was confirmed by F-test statistics at the 95% confidence level between HIV-1 gag standard concentrations of 1.4*1010 to 6.5*101° part/mL, with an average slope of 0.14, an average intercept of 3.3*108, and an average correlation coefficient of 0.9947 (Fig. 5B). However, when 25 pL of 1.4x1010 part/mL HIV-1 gag VLPs (lowest standard concentration) or 50 pL of 6.5*1010 part/mL HIV-1 gag VLPs (highest standard concentration) were injected, MALS signals could not be detected precisely anymore (RSD >20%). Therefore, we concluded that these values were below the LLOQ or above the ULOQ (Table 2) and were not included in the regression model. The residuals for all other injections were equally distributed, and the RSD was lower than 10% (data not shown). Based on the slope of the correlation curve and the standard deviation of the MALS response signals, the LLOD and LLOQ were determined in a way similar to that already described for the SEC-UV method in section 3.3 SEC-UV method. The SEC-MALS method demonstrated a lower sensitivity

compared to the SEC-UV method, and a precise quantification of HIV-1 gag VLPs was possible up to 4.9x1010 part/mL (equivalent to 9.0x109 part/mL directly measured by MALS) (Table 2) when sample volumes in the range of 25 to 50 pL were injected.

3.5 Quantification of HIV-1 gag VLPs in various samples

The applicability of both SE-HPLC methods for quantification of HIV-1 gag VLPs was demonstrated by analyzing various samples with different sample matrixes and impurity contents (Table 3). The samples were collected during preparative purification of HIV-1 gag VLPs using AIEX monoliths [32] (Fig. 6A, B). Including samples from the loading material, flow through (FT), wash fraction and two fractions containing HIV-1 gag VLPs collected during the first (P1, 25% B) and the second elution step (P2, 45% B). Additionally, the impact of different dsDNA concentrations of the samples was investigated using samples optionally spiked with 3.3 pg/mL DNA and/or treated with 150 U/mL Benzonase for 1 h (Fig. 6B). Analysis of samples collected during AIEX purification resulted in similar particle concentrations measured by the different methods (Fig. 6A). Furthermore, the presence or absence of dsDNA impurities in the sample material did not influence any of the results (Fig. 6B). When samples in a low particle concentration range were analyzed, such as P2 Blank, P2 Spike, or P2 Spike Benz in Fig. 6B, NTA proved to be the most sensitive method.

3.6 Comparison of SEC-UV, SEC-MALS, and NTA

In general, each type of detection described here measures different attributes that are used for quantification of HIV-1 gag VLPs and therefore carry different limitations and advantages. When HIV-1 gag VLPs are analyzed by NTA, the hydrodynamic diameter of the particles is calculated, and the particle concentration is obtained when the tracked particles are counted and related to the sample volume. NTA tends to overestimate the particle concentration, especially of crude materials containing background particles such as host cell debris, other extracellular particles, aggregated proteins, or dsDNA complexes [13]. Therefore, both SE-HPLC methods, independent from the detection mode, benefit from the separation of HIV-1 gag VLPs from impurities before quantification, although separation of impurities that are similar in size to HIV-1 gag VLPs cannot be achieved by the SEC column. When the UV signals are detected, the evaluated particle concentration is based on the calibration curve using HIV-1 gag VLP standards, initially quantified by NTA. However, no information

about the particle size is provided by this detection method, and a possible overestimated particle concentration of the HIV-1 gag VLP standard is propagated during analysis of unknown samples. Otherwise, HIV-1 gag VLPs can be directly quantified by the MALS detector.

MALS is an absolute detection method, similarly to NTA, and can provide information about the particle concentration without preparation of a calibration curve. Here the particle number is calculated based on the differential scattering intensity of the particle, which depends on the size of the scattered particle and the particle's refractive index. This model results in a particle size described by a geometric radius of a sphere measured in solution. Given that TEM pictures proved that HIV-1 gag VLPs are spherical, it is reasonable to compare the hydrodynamic diameter measured by NTA to the geometric diameter measured by MALS. However, the model is based on the assumption that monodisperse, homogeneous, spherical particles are eluted and analyzed [24], [23], which might not be true for the HIV-1 gag VLPs eluting in the void fraction of the SEC column. Furthermore, aggregation might be induced by interaction of the VLPs with the stationary phase. This factor was assumed to be the reason for a 0.5 log decreased particle concentration measured by SEC-MALS compared to field flow fractionation coupled to MALS in an analysis of influenza viruses by Wei et al. [7]. Partially adsorbed or retained HIV-1 gag VLPs by the stationary phase could be another reason for underestimation of particle concentrations measured by SEC-MALS. Furthermore, the particle's refractive index has the largest impact on determination of a precise particle number [23]. For our calculations, we assumed a refractive index based on a value reported in the literature for a VLP with a 100-nm diameter surrounded by a 10-nm lipid bilayer [36].

When the quantification methods are compared in terms of feasibility, SE-HPLC methods show big advantages compared to NTA. NTA is very user sensitive, and several parameters are adjustable during video capture and analysis. To minimize user influence, measurement of serial dilutions and triplicate analysis with optimized analysis parameters were performed. This measurement procedure, using the LM-10 equipment, is not automatable and takes a minimum of 30 to 45 min per sample. This time is even more prolonged due to sample preparation and dilution to meet the narrow measurement range covering very low particle concentrations (Table 2). The SE-HPLC methods are effectively automated and therefore suitable for high-throughput analysis of samples. Although no appropriate experiments were performed, it is

inferred that the analysis time can be reduced to at least 60 min per sample. The SE-HPLC methods allow direct measurements of samples with relatively high particle concentrations and thus reduce effort for sample preparation and dilution compared to NTA. SE-HPLC also provides information about the purity of the sample and is very robust. However, NTA is the most sensitive method, and samples with very low particle concentrations can be quantified using it.

Conclusion

We developed and demonstrated very useful and reliable tools based on SEC for rapid and automated quantification of HIV-1 gag VLPs. The methods allow quantification of enveloped VLPs in various sample matrixes at different purity stages and are operated in a 100-fold increased concentration range compared to NTA, reducing effort in sample preparation before analysis. The detection of particles using MALS allows direct quantification of particles and omits the mandatory preparation of a calibration curve before measurement. However, by implementation of a correlation curve, particle concentration measured by MALS could be related to values measured by NTA.

Acknowledgements

The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2007-2013) under grant agreement n° 312004. This work also has been partly supported by the Federal Ministry of Science, Research and Economy (BMWFW), the Federal Ministry of Traffic, Innovation and Technology (bmvit), the Styrian Business Promotion Agency SFG, and the Standortagentur Tirol and ZIT - Technology Agency of the City of Vienna through the COMET-Funding Program managed by the Austrian Research Promotion Agency FFG. We thank Gerhard Sekot for assisting with TEM pictures.

References

[1] P. Pushko, P. Pumpens, E. Grens, Development of Virus-Like Particle Technology from Small Highly Symmetric to Large Complex Virus-Like Particle Structures, Intervirology, 56 (2013) 141-165.

[2] C.M. Thompson, E. Petiot, A. Lennaertz, O. Henry, A.A. Kamen, Analytical technologies for influenza virus-like particle candidate vaccines: challenges and emerging approaches, Virology Journal, 10 (2013) 1-14.

[3] L.H.L. Lua, N.K. Connors, F. Sainsbury, Y.P. Chuan, N. Wibowo, A.P.J. Middelberg, Bioengineering virus-like particles as vaccines, Biotechnology and Bioengineering, 111 (2014) 425-440.

[4] L.K. Pattenden, A.P.J. Middelberg, M. Niebert, D.I. Lipin, Towards the preparative and large-scale precision manufacture of virus-like particles, Trends in Biotechnology, 23 (2005) 523-529.

[5] A.J. Darling, J.A. Boose, J. Spaltro, Virus assay methods: Accuracy and validation, Biologicals, 26 (1998) 105-110.

[6] A. Zeltins, Construction and characterization of virus-like particles: A review, Molecular Biotechnology, 53 (2013) 92-107.

[7] Z. Wei, M. McEvoy, V. Razinkov, A. Polozova, E. Li, J. Casas-Finet, G.I. Tous, P. Balu, A.A. Pan, H. Mehta, M.A. Schenerman, Biophysical characterization of influenza virus subpopulations using field flow fractionation and multiangle light scattering: correlation of particle counts, size distribution and infectivity, J Virol Methods, 144 (2007) 122-132.

[8] S. Gutiérrez-Granados, L. Cervera, F. Godia, J. Carrillo, M.M. Segura, Development and validation of a quantitation assay for fluorescently tagged HIV-1 virus-like particles, Journal of Virological Methods, 193 (2013) 85-95.

[9] H. Malenovska, Virus quantitation by transmission electron microscopy, TCID50, and the role of timing virus harvesting: A case study of three animal viruses, Journal of Virological Methods, 191 (2013) 136-140.

[10] D. Griffiths, P. Hole, J. Smith, A. Malloy, B. Carr, Counting and sizing of virus and protein aggregates by Nanoparticle Tracking Analysis (NTA), in: Nanotechnology 2010: Bio Sensors, Instruments, Medical, Environment and Energy - Technical Proceedings of the 2010 NSTI Nanotechnology Conference and Expo, NSTI-Nanotech 2010, 2010, pp. 176-179.

[11] S. Heider, C. Metzner, Quantitative real-time single particle analysis of virions, Virology, 462-463 (2014) 199-206.

[12] R A. Dragovic, C. Gardiner, A.S. Brooks, D.S. Tannetta, D.J.P. Ferguson, P. Hole, B. Carr, C.W.G. Redman, A.L. Harris, P.J. Dobson, P. Harrison, I.L. Sargent, Sizing and phenotyping of cellular vesicles using Nanoparticle Tracking Analysis, Nanomedicine: Nanotechnology, Biology, and Medicine, 7 (2011) 780-788.

[13] G.G. Mironov, A V. Chechik, R. Ozer, J.C. Bell, M.V. Berezovski, Viral Quantitative Capillary Electrophoresis for Counting Intact Viruses, Analytical Chemistry, 83 (2011) 5431 -5435.

[14] B. Kalbfuss, M. Wolff, R. Morenweiser, U. Reichl, Purification of cell culture-derived human influenza A virus by size-exclusion and anion-exchange chromatography, Biotechnology and Bioengineering, 96 (2007) 932-944.

[15] P. Nestola, R.J.S. Silva, C. Peixoto, P.M. Alves, M.J.T. Carrondo, J.P.B. Mota, Robust design of adenovirus purification by two-column, simulated moving-bed, size-exclusion chromatography, Journal of Biotechnology, 213 (2015) 109-119.

[16] T. Weigel, T. Solomaier, A. Peuker, T. Pathapati, M.W. Wolff, U. Reichl, A flow-through chromatography process for influenza A and B virus purification, Journal of Virological Methods, 207 (2014) 45-53.

[17] C. Peixoto, M.F.Q. Sousa, A.C. Silva, M.J.T. Carrondo, P.M. Alves, Downstream processing of triple layered rotavirus like particles, Journal of Biotechnology, 127 (2007) 452-461.

[18] P. Lagoutte, C. Mignon, S. Donnat, G. Stadthagen, J. Mast, R. Sodoyer, A. Lugari, B. Werle, Scalable chromatography-based purification of virus-like particle carrier for epitope based influenza A vaccine produced in Escherichia coli, Journal of Virological Methods, 232 (2016) 8-11.

[19] J. Transfiguracion, A.P. Manceur, E. Petiot, C.M. Thompson, A.A. Kamen, Particle quantification of influenza viruses by high performance liquid chromatography, Vaccine, 33 (2015) 78-84.

[20] Y. Yang, H. Li, Z. Li, Y. Zhang, S. Zhang, Y. Chen, M. Yu, G. Ma, Z. Su, Size-exclusion HPLC provides a simple, rapid, and versatile alternative method for quality control of vaccines by characterizing the assembly of antigens, Vaccine, 33 (2015) 1143-1150.

[21] C. Ladd Effio, S.A. Oelmeier, J. Hubbuch, High-throughput characterization of virus-like particles by interlaced size-exclusion chromatography, Vaccine, 34 (2016) 1259-1267.

[22] S. Podzimek, Light Scattering, Size Exclusion Chromatography and Asymmetric Flow Field Flow Fractionation Powerful Tools for the Characterization of Polymers, Proteins and Nanoparticles, John Wiley & Sons, Inc., Hoboken, NJ, USA, 2011.

[23] P.J. Wyatt, M.J. Weida, Method and apparatus for determining absolute number densities of particles in suspension, in, 2004.

[24] M. McEvoy, V. Razinkov, Z. Wei, J R. Casas-Finet, G.I. Tous, M.A. Schenerman, Improved particle counting and size distribution determination of aggregated virus populations by asymmetric flow field-flow fractionation and multiangle light scattering techniques, Biotechnol Prog, 27 (2011) 547-554.

[25] P.J. Wyatt, Measurement of special nanoparticle structures by light scattering, Analytical Chemistry, 86 (2014) 7171-7183.

[26] T. Bousse, D.A. Shore, C.S. Goldsmith, M.J. Hossain, Y. Jang, C.T. Davis, R.O. Donis, J. Stevens, Quantitation of influenza virus using field flow fractionation and multi-angle light scattering for quantifying influenza A particles, J Virol Methods, 193 (2013) 589-596.

[27] D.I. Lipin, Y.P. Chuan, L.H.L. Lua, A.P.J. Middelberg, Encapsulation of DNA and non-viral protein changes the structure of murine polyomavirus virus-like particles, Archives of Virology, 153 (2008) 2027.

[28] L.F. Pease, D.I. Lipin, D.H. Tsai, M.R. Zachariah, L.H.L. Lua, M.J. Tarlov, A.P.J. Middelberg, Quantitative characterization of virus-like particles by asymmetrical flow field flow fractionation, electrospray differential mobility analysis, and transmission electron microscopy, Biotechnol Bioeng, 102 (2009).

[29] Y.P. Chuan, Y.Y. Fan, L. Lua, A.P.J. Middelberg, Quantitative analysis of viruslike particle size and distribution by field-flow fractionation, Biotechnology and Bioengineering, 99 (2008) 1425-1433.

[30] L. Deml, C. Speth, M.P. Dietrich, H. Wolf, R. Wagner, Recombinant HIV-1 Pr55gag virus-like particles: potent stimulators of innate and acquired immune responses, Mol Immunol, 42 (2005).

[31] N.L. Goicochea, S.A.K. Datta, M. Ayaluru, C. Kao, A. Rein, B. Dragnea, Structure and stoichiometry of template-directed recombinant HIV-1 Gag particles, Journal of Molecular Biology, 410 (2011) 667-680.

[32] P. Steppert, D. Burgstaller, M. Klausberger, E. Berger, P.P. Aguilar, T.A. Schneider, P. Kramberger, A. Tover, K. Nobauer, E. Razzazi-Fazeli, A. Jungbauer, Purification of HIV-1 gag virus-like particles and separation of other extracellular particles, Journal of Chromatography A, (2016).

[33] J. Mogridge, Using Light Scattering to Determine the Stoichiometry of Protein Complexes, in: H. Fu (Ed.) Protein-Protein Interactions: Methods and Applications, Humana Press, Totowa, NJ, 2004, pp. 113-118.

[34] M.P. Tarazona, E. Saiz, Combination of SEC/MALS experimental procedures and theoretical analysis for studying the solution properties of macromolecules, Journal of Biochemical and Biophysical Methods, 56 (2003) 95-116.

[35] J. Wen, T. Arakawa, J.S. Philo, Size-Exclusion Chromatography with On-Line Light-Scattering, Absorbance, and Refractive Index Detectors for Studying Proteins and Their Interactions, Analytical Biochemistry, 240 (1996) 155-166.

[36] Y. Hu, X. Cheng, H. Daniel Ou-Yang, Enumerating virus-like particles in an optically concentrated suspension by fluorescence correlation spectroscopy, Biomedical Optics Express, 4 (2013) 1646-1653.

[37] J. Heukeshoven, R. Dernick, Simplified method for silver staining of proteins in polyacrylamide gels and the mechanism of silver staining, Electrophoresis, 6 (1985) 103-112.

[38] M.A. Hayat, S.E. Miller, Negative Staining, McGraw-Hill, New York, 1990.

[39] A. Lynch, A.E. Meyers, A.-L. Williamson, E.P. Rybicki, Stability studies of HIV-1 Pr55gagvirus-like particles made in insect cells after storage in various formulation media, Virology Journal, 9 (2012) 1-5.

[40] L.H. Lua, N.K. Connors, F. Sainsbury, Y.P. Chuan, N. Wibowo, A.P. Middelberg, Bioengineering virus-like particles as vaccines, Biotechnol Bioeng, 111 (2014) 425440.

[41] P. Kramberger, M. Ciringer, A. Strancar, M. Peterka, Evaluation of nanoparticle tracking analysis for total virus particle determination, Virol J, 9 (2012) 265.

[42] J. Transfiguracion, H. Jorio, J. Meghrous, D. Jacob, A. Kamen, High yield purification of functional baculovirus vectors by size exclusion chromatography, Journal of Virological Methods, 142 (2007) 21-28.

[43] T. Arakawa, D. Ejima, T. Li, J.S. Philo, The critical role of mobile phase composition in size exclusion chromatography of protein pharmaceuticals, Journal of Pharmaceutical Sciences, 99 (2010) 1674-1692.

[44] P. Hong, S. Koza, E.S.P. Bouvier, A review size-exclusion chromatography for the analysis of protein biotherapeutics and their aggregates, Journal of Liquid Chromatography and Related Technologies, 35 (2012) 2923-2950.

[45] R.D. Ricker, L.A. Sandoval, 15th International Symposium on Protein, Peptide and Polynucleotide AnalysisFast, reproducible size-exclusion chromatography of biological macromolecules, Journal of Chromatography A, 743 (1996) 43-50.

[46] P. Araujo, Key aspects of analytical method validation and linearity evaluation, Journal of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences, 877 (2009) 2224-2234.

[47] Validation of Analytical Procedures: Text and Methodology Q2(R1),, in, International Conference on Harmonization (ICH), http://www.ich.org, 1996.

Figure legends

Fig. 1 TEM pictures of negatively stained HIV-1 gag VLPs. Sample (A) was purified by density gradient centrifugation and then applied to SEC. (B) and (C) show images of pooled collected SEC elution fractions after injection of the density gradient purified standard material (pool of fractions 1-3 in Fig. 2). The scale bars correspond to 100 nm in (A, C) and 200 nm in (B).

Fig. 2 (A) Overlay of SEC profiles obtained from injection of HIV-1 gag VLPs purified by density gradient centrifugation, CHO supernatant (Supernatant), or endonuclease-treated CHO supernatant (Supernatant Benz). Injection volume was 100 pL, and elution buffer was PBS pH 7.4 at a flow rate of 0.2 mL/min. Fractions (1-3) were collected after injection of density gradient purified standard material and analyzed by (B) SDS-PAGE using silver stain and (C) Western blot analysis detecting HIV-1 p24. M: molecular mass marker; 1-3: elution fractions as indicated in (A).

Fig. 3 The effect of the elution buffer composition on SEC response obtained after injection of 50 pL purified HIV-1 gag VLPs, optionally diluted in the appropriate buffer or present in crude CHO supernatant. (A) pH values from pH 6.0 to pH 9.0 were screened using 25 mM Na-phosphate, 500 mM NaCl pH 6.0/ pH 7.0/ pH 7.5/ pH 8.0, and 50 mM Bis-Tris, 500 mM NaCl pH 9.0. (B) 25 mM Na-phosphate, 500 mM NaCl pH 8.0 (d) was modified by addition of (f) 20% (v/v) i-propanol, (g) 200 mM arginine, (h) 0.1% (w/v) Tween 80, (i) 1 mM EDTA, (j) 250 mM urea, or (k) 250 mM sucrose. (C) Comparison of 25 mM Na-phosphate, 250 mM pH 8.0 and 25 mM Na-phosphate, 500 mM NaCl pH 8.0. The normalized response was obtained from the integrated peak areas related to injections (A) at pH 7.5, (B) in the presence of buffer (h) and (C) samples diluted 1:5.

Fig. 4 (A) SEC elution peaks detected by UV absorbance at 280 nm, obtained after 25 pL injection of HIV-1 gag VLP concentrations, ranging from 1.4*1010 to 6.5*101° part/mL. (B) The calibration curves of the SEC-UV method for injection of 25 or 50 pL of HIV-1 gag VLPs.

Fig. 5 (A) SEC elution peaks detected by the 90° light scattering signal, obtained after 25 |jL injection of HIV-1 gag VLP concentrations, ranging from 1.4*1010 to 6.5*101° part/mL. The geometric radius obtained from all measurements is overlaid. (B) The calibration curve of the SEC-MALS method, correlating particle concentrations measured by SEC-MALS to particle concentrations measured by NTA.

Fig. 6 Quantification of HIV-1 gag VLPs by NTA, SEC-UV, and SEC-MALS in various samples. The samples were obtained during preparative purification using an AIEX monolith (A) and (B), optionally spiked with 0.3 jg/mL DNA (Spike) and endonuclease treated with 150 U/mL Benzonase (Benz). Blank samples were diluted with appropriate buffers, but no DNA or Benzonase was added. A detailed summary of individual sample composition is reported in Table 3. Samples marked with * were below LLOQ; + were below LLOD.

Fig. 1

»0 nrn '."y 200 nm 100 nm

Fig. 2

Fig. 3 (A)

CL <0 0)

"I 0.8

z 0.6 A

--0 — 500 mM NaCI

--v — 250 mM NaCI

/ / / /

^ V. / .....

o—■

Fig. 4

Geometric radius (nm)

Fig. 6

- I-1 NTA

Tables

Table 1 Summary of elution buffers tested during method development Buffer

(a) 25 mM Na-phosphate, 500 mM NaCI, pH 6.0

(b) 25 mM Na-phosphate, 500 mM NaCI, pH 7.0

(c) 25 mM Na-phosphate, 500 mM NaCI, pH 7.5

(d) 25 mM Na-phosphate, 500 mM NaCI, pH 8.0

(e) 50 mM Bis-Tris, 500 mM NaCI, pH 9.0

(f) 25 mM Na-phosphate, 500 mM NaCI, 20% (v/v) i-propanoI, pH 8.0

(g) 25 mM Na-phosphate, 500 mM NaCI, 200 mM arginine, pH 8.0 25 mM Na-phosphate, 500 mM NaCI, 0.1% (w/v) Tween 80, pH

(h) 8.0

(i) 25 mM Na-phosphate, 500 mM NaCI, 1 mM EDTA, pH 7.0 (j) 25 mM Na-phosphate, 500 mM NaCI, 250 mM urea, pH 8.0 (k) 25 mM Na-phosphate, 500 mM NaCI, 250 mM sucrose, pH 8.0 (I) 25 mM Na-phosphate, 250 mM NaCI, pH 8.0

Table 2. LLOD, LLOQ, and ULOQ for quantification of HIV-1 gag VLPs by NTA, SEC-UV, and SEC-MALS. Particle concentrations measured by SEC-MALS were obtained directly from MALS or were correlated with NTA particle concentrations by the linear correlation function displayed in Fig. 5B.

Detection method Injection volume (ML) LLOD (part/mL) LLOQ (part/mL) ULOQ (part/mL)

NTA 300 1.7E+07 3.4E+08 1.0E+09

SEC-UV 25 8.0E+09 2.4E+10 6.5E+10

50 6.9E+09 2.1E+10 4.9E+10

SEC-MALS 5.6E+09 1.7E+10

25-50 6.5E+10

correlated

SEC-MALS directly 25-50 4.8E+08 2.1E+09 9.0E+09

Table 3. Composition of samples analyzed by SEC-UV, SEC-MALS, and NTA.

dsDNA (ng/mL) Total

Sample Sample matrix protein (Mg/mL)

Load ** Cell culture medium, pH 7.4 10 349

AIEX purification FT Cell culture medium, pH 7.4 40 171

Wash 50 mM HEPES, 350 mM NaCl, pH 7.2 65 371

P1 50 mM HEPES, 762 mM NaCl, pH 7.2 24652 175

P2 50 mM HEPES, 1.093 M NaCl, pH 7.2 602 118

P1 Blank 50 mM HEPES, 762 mM NaCl, 5 mM EDTA, pH n.d. 19430 204

P1 Benz t 50 mM HEPES, 766 mM NaCl, 54 mM Tris-HCl, 54 mM MgCl2, 5 mM EDTA, pH n.d. 5665 204

P1 Spike * 50 mM HEPES, 766 mM NaCl, 54 mM Tris-HCl, 54 mM MgCl2, 5 mM EDTA, pH n.d. 21109 196

Endonuclease P1 Spike 50 mM HEPES, 766 mM NaCl, 54 mM Tris-HCl, 54 mM 5031 196

treatment of Benz *t MgCl2, 5 mM EDTA, pH n.d.

AIEX elution P2 Blank 50 mM HEPES, 1.093 mM NaCl, 5mM EDTA, pH n.d. 237 62

fractions P2 Benz t 50 mM HEPES, 1.097 mM NaCl, 54 mM Tris-HCl, 54 mM 160 62

MgCl2, 5 mM EDTA, pH n.d.

P2 Spike * 50 mM HEPES, 1.097 mM NaCl, 54 mM Tris-HCl, 54 mM MgCl2, 5 mM EDTA, pH n.d. 3245 59

P2 Spike 50 mM HEPES, 1.097 mM NaCl, 54 mM Tris-HCl, 54 mM 158 59

Benz *t MgCl2, 5 mM EDTA, pH n.d.

*Samples were spiked with 3.3 ^g/mL lambda DNA.

t Samples were treated with 150 U/mL Benzonase.

** Samples were 0.8 ^m

filtered.

n.d. - not

determined