Crystallization of a nonreplicating rotavirus vaccine candidate

Abstract Nonreplicating rotavirus vaccine (NRRV) candidates are being developed with the aim of serving the needs of developing countries. A significant proportion of the cost of manufacturing such vaccines is the purification in multiple chromatography steps. Crystallization has the potential to reduce purification costs and provide new product storage modality, improved operational flexibility, and reduced facility footprints. This communication describes a systematic approach for the design of the crystallization of an NRRV candidate, VP8 subunit proteins fused to the P2 epitope of tetanus toxin, using first‐principles models and preliminary experimental data. The first‐principles models are applied to literature data to obtain feasible crystallization conditions and lower bounds for nucleation and growth rates. Crystallization is then performed in a hanging‐drop vapor diffusion system, resulting in the nucleation and growth of NRRV crystals. The crystals obtained in a scaled‐up evaporative crystallization contain proteins truncated in the P2 region, but have no significant differences with the original samples in terms of antibody binding and overall conformational stability. These results demonstrate the promise of evaporative crystallization of the NRRV.

crystallization scale sublinearly (Hong et al., 2018). Although the purification of some therapeutic proteins such as insulin have used crystallization, crystallization technology effective for large-molecule therapeutic proteins is still lacking and needs to be developed.
Technology for the design and control of the crystallization of proteins is much less mature than for small molecules. Methods for inducing supersaturation are also more limited due to the need of maintaining protein stability and quality. Many proteins are easily denatured by changes in temperature and pH, addition of precipitants, and agitation. Proteins have complex thermodynamics, slow kinetics, large uncertainties, and potential for protein aggregation that greatly restrict allowable paths through the phase diagram, which is equivalent to threading an unknown narrow winding path through an uncertain high-dimensional space.
This communication describes a systematic approach to the design of the NRRV crystallization by a combination of firstprinciples models and preliminary experimental data. Literaturereported results for truncated VP8 subunit proteins of rotaviruses (Dormitzer et al., 2002;Kraschnefski et al., 2008Kraschnefski et al., , 2005Scott et al., 2005;Yu et al., 2008;Zhang et al., 2007) are analyzed to obtain feasible crystallization conditions and lower bounds on the crystal nucleation and growth rates. Proof-of-concept crystallization experiments are performed for validation of the analysis and characterization of the crystals.

| Estimation of crystallization rates
Preliminary well-or vial-based experimental data can provide lower bounds on crystallization rates. Most experimental studies apply screening methods such as hanging-or sitting-drop vapor diffusion systems (McPherson, 2004). Vapor diffusion systems place a droplet containing protein, buffer, and precipitant in vapor equilibrium with a reservoir containing higher concentration of buffer and precipitant.
Water evaporates, which increases the concentration of protein and precipitant, until the droplet reaches equilibrium with the reservoir. This process produces a gradual increase of supersaturation, resulting in nucleation and growth of crystals.
Nucleation within such drops is describable by the stochastic model (Goh et al., 2010): is the nucleation rate which is a function of states that change with time t (more details below), P t ( ) 0 is the time evolution of the probability that the droplet contains no crystals, and V t ( ) is the volume of the droplet. The analytical solution of Equation (1) is The induction time t ind is the time when at least one crystal has nucleated. The cumulative distribution function (CDF) for the induction time and the corresponding probability distribution func- (4) and the mean induction time is The nucleation rate in the droplet is modeled by the classical homogeneous nucleation expression (Nielsen, 1964), where A and B are nucleation parameters, is the concentration of protein, and C t ( ) P, sat is the solubility. The most rigorous definition for the supersaturation is in terms of chemical potentials but S t ( ) is nearly always written in terms of concentrations to avoid the time and expense of computing the chemical potential of the solution phase. Although there is substantial evidence that not all primary nucleation is described by classical nucleation theory (Erdemir et al., 2009), the above expression has been shown to correlate well with experimental data for most solute-solvent systems while having only two fitting parameters (Kim & Mersmann, 2001).
Since the amount of protein in the droplet C t V t ( ) ( ) where subscript "e" refers to the conditions in the droplet at the equilibrium volume before nucleation. This expression can be rearranged to provide a lower bound for the nucleation rate of where subscript "lb" indicates a lower bound. Some publications directly report an induction time assuming that the time for a nucleus to grow large enough to be observable is negligible . Other publications report only the total time for nucleation and growth (Dormitzer et al., 2002;Kraschnefski et al., 2008Kraschnefski et al., , 2005Yu et al., 2008;Zhang et al., 2007). The above lower bound remains valid, although less tight, for the induction times reported in these publications.
HONG ET AL.

| 1751
Numerous expressions are available for modeling the crystal growth rate G t ( ), which are all increasing functions of supersaturation. The supersaturation is maximum before nucleation because the supersaturation decreases after nucleation due to the crystal growth. Then the growth rate can be related to the mean where N t ( ) is the number of crystals, t n is the time when at least n crystals have nucleated, and the subscript "e" refers to the droplet conditions before nucleation. The first inequality is introduced because the times when the crystals have nucleated cannot be directly measured. This expression can be rearranged to provide a lower bound for the crystal growth rate, As before, when publications do not report the induction time directly, this lower bound remains valid while being less tight. Table 1 reports the lower bounds on the crystal nucleation and growth rates calculated for the literature-reported crystallization results for truncated VP8 subunit proteins (Dormitzer et al., 2002;Kraschnefski et al., 2008Kraschnefski et al., , 2005Scott et al., 2005;Yu et al., 2008;Zhang et al., 2007). Although the estimated crystallization rates are very slow, evaporation-based crystallization should be feasible for truncated VP8 subunit proteins by providing a method for controlling supersaturation to deal with the uncertainties in the crystallization kinetics. The slow primary nucleation rate is addressable by seeding, and the slow crystal growth rate is addressable by increasing the surface area of the crystals. These estimated crystallization rates can be applied for the preliminary design of evaporative crystallizers.

| Identification of feasible crystallization conditions
To identify feasible crystallization conditions for truncated VP8 subunit proteins, a mechanistic model developed for predicting the pH and ionic strength of cell culture media (Hong et al., 2021) was applied to the literature-reported crystallization media (Dormitzer et al., 2002;Kraschnefski et al., 2008Kraschnefski et al., , 2005Scott et al., 2005;Yu et al., 2008;Zhang et al., 2007) (Table 2)

| Fermentation and purification
The P2-VP8-P[8] sequence was modified to improve product quality and expression titer (Dalvie et al., 2020). The modified P2-VP8-P[8] was expressed and secreted from Pichia pastoris (Komagataella phaffii NRRL Y-11430). The fermentation and protein purification were carried out in an automated, benchtop, multiproduct manufacturing system, as previously reported (Crowell et al., 2018). Cells were grown with 4% glycerol for biomass accumulation and 1% methanol, supplemented with 67 g/L sorbitol, for production. The temperature, pH, and dissolved oxygen were maintained at 25°C, 6.5, and 25%, respectively.
Purified protein was concentrated approximately 10-fold using

kDa molecular weight cut off (MWCO) Amicon Ultra Centrifugal
Filter Units (Millipore Sigma) according to the manufacturer's recommended protocol. The concentrated protein was then dialyzed against 0.1 M PIPES, pH 6.5, using a 3.5 kDa MWCO Slide-A-Lyzer G2 dialysis cassette (Thermo Fisher Scientific) according to the manufacturer's recommended protocol.

| Crystallization experiments
Crystallization was performed using a hanging-drop T A B L E 2 Crystallization media for truncated VP8 subunit proteins of rotaviruses (TNE: 20 mM Tris-HCl pH 8.0, 100 mM NaCl, 1 mM EDTA)  (Wei et al., 2018). Total intensity data (the peak area under the curve for a waveform) at various temperatures were acquired from the plate reader and normalized using min-max normalization. Origin 9.4 software package was used to calculate the T m value by plotting first derivative of total intensity data against corresponding temperatures. Volkin.