3.4. Purification and characterization of the yellow fever
VLPs
In parallel to the investigation of perfusion processes discussed above,
the VLPs produced by stable cell pool 2 were purified and characterized
in order to allow evaluation of product attributes and to establish
hallmarks for future integration of upstream and downstream processing.
Steric exclusion chromatography (SXC) was selected for single-step
purification of YF VLPs due to the good results reported for this
technique for concentration and purification of large biomolecules, such
as nucleic acids, large proteins and viruses (Lee et al., 2012;
Marichal-Gallardo et al., 2017; Levanova & Poranen, 2018). SXC works by
capturing the large biomolecules at a non-reactive hydrophilic surface
by their mutual steric exclusion of polyethylene glycol (PEG) (Lee et
al., 2012). Since the contaminants in the cell culture supernatant are
much smaller than the VLPs, they are eliminated during feed in the
flowthrough, whereas VLPs are subsequently eluted by reducing PEG
concentration.
Figure 5A shows a typical chromatogram obtained for a single-step
purification from cell culture supernatant using SXC with 8% (m/v) of
PEG 6,000. The chemiluminescent immunoblot of the fractions collected
during chromatographic steps shows that most VLPs were recovered as a
concentrated eluate fraction and that no product was lost either in the
feed flowthrough or in the wash step. Some product was detected in the
fraction of the regeneration with 1M NaCl, indicating that there was
some degree of adsorption of VLPs to the membrane used as stationary
phase. Although adsorption is not the intended mechanism of SXC, the low
intensity of the signal in the regeneration shows that it is a minor
part of the VLPs that is lost in this fraction.
Panels C, D and E (Figure 5) show SDS-PAGE, Western blot and SEC-HPLC
analyses performed to evaluate purity, identity and multimeric state of
the VLPs, respectively. As shown in Figure 5C, sample complexity was
significantly decreased when the cell culture supernatant is compared to
the SXC eluate. The Western blot in Figure 5D shows that two protein
bands in the supernatant, in the SXC eluate (containing 8% PEG) and in
a concentrated/diafiltered sample of the eluate are recognized by the
pan-flavivirus 4G2 antibody. The presence of these two bands having
molecular mass (MM) around that of the envelope protein (50 kDa)
correspond to two different glycoforms of the E protein, as we have
shown in a previous work (Lima et al., 2019). In the SXC eluate the
presence of PEG changed protein migration, but upon diafiltration the
pattern is similar to the one observed in the supernatant sample.
SEC-HPLC analysis of SXC eluate (Fig. 5D) further shows that
contaminants, such as other proteins or eventually monomeric E protein
not assembled in VLPs, were successfully removed in the purification
step.
Concentrations of VLP, total proteins and DNA in the cell culture
supernatant and in the SXC eluate were quantified in order to have an
overview of SXC performance as a purification step for YF VLPs (Table
1). VLP recovery and DNA removal were 72.3% and 60.3% respectively,
which are lower levels than usually reported for SXC, but in a range
comparable to other types of chromatography (Pato et al., 2019). On the
other hand, there was 99.9% total protein removal, which is comparable
to values reported by other authors for SXC (Lee et al., 2012;
Marichal-Gallardo et al., 2017; Levanova & Poranen, 2018).
The SXC-purified VLPs were further analyzed by transmission electron
microscopy (TEM) and analytical size exclusion chromatography (A-SEC) to
further investigate VLP structure and size. Negative-stained TEM
confirmed the 3D structure of VLPs (Fig. 6A), whereas A-SEC runs
revealed a non-symmetric and broader peak when compared to the protein
standards used (Fig. 6B), which goes in line with the variability in
sizes observed in the panel of micrographs (Fig. 6A).