At-line process analytical technology (PAT) for more efficient scale up of biopharmaceutical microfiltration unit operations.

Tangential flow microfiltration (MF) is a cost-effective and robust bioprocess separation technique, but successful full scale implementation is hindered by the empirical, trial-and-error nature of scale-up. We present an integrated approach leveraging at-line process analytical technology (PAT) and mass balance based modeling to de-risk MF scale-up. Chromatography-based PAT was employed to improve the consistency of an MF step that had been a bottleneck in the process used to manufacture a therapeutic protein. A 10-min reverse phase ultra high performance liquid chromatography (RP-UPLC) assay was developed to provide at-line monitoring of protein concentration. The method was successfully validated and method performance was comparable to previously validated methods. The PAT tool revealed areas of divergence from a mass balance-based model, highlighting specific opportunities for process improvement. Adjustment of appropriate process controls led to improved operability and significantly increased yield, providing a successful example of PAT deployment in the downstream purification of a therapeutic protein. The general approach presented here should be broadly applicable to reduce risk during scale-up of filtration processes and should be suitable for feed-forward and feed-back process control.

Targeted purification development enabled by computational biophysical modeling.

Chromatographic and non-chromatographic purification of biopharmaceuticals depend on the interactions between protein molecules and a solid-liquid interface. These interactions are dominated by the protein-surface properties, which are a function of protein sequence, structure, and dynamics. In addition, protein-surface properties are critical for in vivo recognition and activation, thus, purification strategies should strive to preserve structural integrity and retain desired pharmacological efficacy. Other factors such as surface diffusion, pore diffusion, and film mass transfer can impact chromatographic separation and resin design. The key factors that impact non-chromatographic separations (e.g., solubility, ligand affinity, charges and hydrophobic clusters, and molecular dynamics) are readily amenable to computational modeling and can enhance the understanding of protein chromatographic. Previously published studies have used computational methods such as quantitative structure-activity relationship (QSAR) or quantitative structure-property relationship (QSPR) to identify and rank order affinity ligands based on their potential to effectively bind and separate a desired biopharmaceutical from host cell protein (HCP) and other impurities. The challenge in the application of such an approach is to discern key yet subtle differences in ligands and proteins that influence biologics purification. Using a relatively small molecular weight protein (insulin), this research overcame limitations of previous modeling efforts by utilizing atomic level detail for the modeling of protein-ligand interactions, effectively leveraging and extending previous research on drug target discovery. These principles were applied to the purification of different commercially available insulin variants. The ability of these computational models to correlate directionally with empirical observation is demonstrated for several insulin systems over a range of purification challenges including resolution of subtle product variants (amino acid misincorporations). Broader application of this methodology in bioprocess development may enhance and speed the development of a robust purification platform.

Determination of the in vivo stoichiometry of tyrosyl radical per betabeta' in Saccharomyces cerevisiae ribonucleotide reductase.

The class I ribonucleotide reductases catalyze the conversion of nucleotides to deoxynucleotides and are composed of two subunits: R1 and R2. R1 contains the site for nucleotide reduction and the sites that control substrate specificity and the rate of reduction. R2 houses the essential diferric-tyrosyl radical (Y(*)) cofactor. In Saccharomyces cerevisiae, two R1s, alpha(n) and , have been identified, while R2 is a heterodimer (betabeta'). beta' cannot bind iron and generate the Y(*); consequently, the maximum amount of Y(*) per betabeta' is 1. To determine the cofactor stoichiometry in vivo, a FLAG-tagged beta ((FLAG)beta) was constructed and integrated into the genome of Y300 (MHY343). This strain facilitated the rapid isolation of endogenous levels of (FLAG)betabeta' by immunoaffinity chromatography, which was found to have 0.45 +/- 0.08 Y(*)/(FLAG)betabeta' and a specific activity of 2.3 +/- 0.5 micromol min(-1) mg(-1). (FLAG)betabeta' isolated from MMS-treated MHY343 cells or cells containing a deletion of the transcriptional repressor gene CRT1 also gave a Y(*)/(FLAG)betabeta' ratio of 0.5. To determine the Y(*)/betabeta' ratio without R2 isolation, whole cell EPR and quantitative Western blots of beta were performed using different strains and growth conditions. The wild-type (wt) strains gave a Y(*)/betabeta' ratio of 0.83-0.89. The same strains either treated with MMS or containing a crt1Delta gave ratios between 0.49 and 0.72. Nucleotide reduction assays and quantitative Western blots from the same strains provided an independent measure and confirmation of the Y(*)/betabeta' ratios. Thus, under normal growth conditions, the cell assembles stoichiometric amounts of Y(*) and modulation of Y(*) concentration is not involved in the regulation of RNR activity.

The active form of the Saccharomyces cerevisiae ribonucleotide reductase small subunit is a heterodimer in vitro and in vivo.

The class I ribonucleotide reductases (RNRs) are composed of two homodimeric subunits: R1 and R2. R2 houses a diferric-tyrosyl radical (Y*) cofactor. Saccharomyces cerevisiae has two R2s: Y2 (beta2) and Y4 (beta'2). Y4 is an unusual R2 because three residues required for iron binding have been mutated. While the heterodimer (betabeta') is thought to be the active form, several rnr4delta strains are viable. To resolve this paradox, N-terminally epitope-tagged beta and beta' were expressed in E. coli or integrated into the yeast genome. In vitro exchange studies reveal that when apo-(His6)-beta2 ((His)beta2) is mixed with beta'2, apo-(His)betabeta' forms quantitatively within 2 min. In contrast, holo-betabeta' fails to exchange with apo-(His)beta2 to form holo-(His)betabeta and beta'2. Isolation of genomically encoded tagged beta or beta' from yeast extracts gave a 1:1 complex of beta and beta', suggesting that betabeta' is the active form. The catalytic activity, protein concentrations, and Y* content of the rnr4delta and wild type (wt) strains were compared to clarify the role of beta' in vivo. The Y* content of rnr4delta is 15-fold less than that of wt, consistent with the observed low activity of rnr4delta extracts (<0.01 nmol min(-1) mg(-1)) versus wt (0.06 +/- 0.01 nmol min(-1) mg(-1)). (FLAG)beta2 isolated from the rnr4delta strain has a specific activity of 2 nmol min(-1) mg(-1), similar to that of reconstituted apo-(His)beta2 (10 nmol min(-1) mg(-1)), but significantly less than holo-(His)betabeta' (approximately 2000 nmol min(-1) mg(-1)). These studies together demonstrate that beta' plays a crucial role in cluster assembly in vitro and in vivo and that the active form of the yeast R2 is betabeta'.

Disentangling the web of allosteric communication in a homotetramer: heterotropic inhibition of phosphofructokinase from Bacillus stearothermophilus.

A strategy for isolating each of the four potentially unique heterotropic pairwise allosteric interactions that exist in the homotetramer phosphofructokinase from Bacillus stearothermophilus is described. The strategy involves the construction of hybrid tetramers containing one wild-type subunit and three mutant subunits that have been modified to block binding of both the substrate, fructose 6-phosphate (Fru-6-P), and the allosteric inhibitor, phospho(enol)pyruvate (PEP). Each type of binding site occurs at a subunit interface, and mutations on either side of the interface have been identified that will greatly diminish binding at the respective site. Consequently, four different types of mutant subunits have been created, each containing a different active site and allosteric site modification. The corresponding 1:3 hybrids isolate a different pair of unmodified substrate and allosteric sites with a unique structural disposition located 22, 30, 32, and 45 A apart, respectively. The allosteric inhibition exhibited by the unmodified sites in each of these four hybrids has been quantitatively evaluated in terms of a coupling free energy. Each of the coupling free energies is unique in magnitude, and their relative magnitudes vary with pH. Importantly, the sum of these coupling free energies at each pH is equal to the total heterotropic coupling free energy associated with the tetrameric enzyme. The latter quantity was assessed from the overall inhibition of a control hybrid that removed the homotropic interactions in PEP binding. The results do not agree with either the concerted or sequential models that are often invoked to explain allosteric behavior in oligomeric enzymes.