Integrated autolysis, DNA hydrolysis and precipitation enables an improved bioprocess for Q-Griffithsin, a broad-spectrum antiviral and clinical-stage anti-COVID-19 candidate

Abstract

Across the biomanufacturing industry, innovations are needed to improve efficiency and flexibility, especially in the face of challenges such as the COVID-19 pandemic. Here we report an improved bioprocess for Q-Griffithsin, a broad-spectrum antiviral currently in clinical trials for COVID-19. Q-Griffithsin is produced at high titer in E. coli and purified to anticipated clinical grade without conventional chromatography or the need for any fixed downstream equipment. The process is thus both low-cost and highly flexible, facilitating low sales prices and agile modifications of production capacity, two key features for pandemic response. The simplicity of this process is enabled by a novel unit operation that integrates cellular autolysis, autohydrolysis of nucleic acids, and contaminant precipitation, giving essentially complete removal of host cell DNA as well as reducing host cell proteins and endotoxin by 3.6 and 2.4 log₁₀ units, respectively. This unit operation can be performed rapidly and in the fermentation vessel, such that Q-GRFT is obtained with 100% yield and > 99.9% purity immediately after fermentation and requires only a flow-through membrane chromatography step for further contaminant removal. Using this operation or variations of it may enable improved bioprocesses for a range of other high-value proteins in E. coli.

Keywords: Biologics manufacturing; Broad spectrum antiviral; COVID-19; Downstream recovery; Griffithsin.

Fig. 1

High-titer expression of Q-GRFT by two-stage fermentation. A) Relative expression of Q-GRFT and GRFT was compared by SDS-PAGE densitometry in microfermentations of an autolysis E. coli strain (DLF_R004) vs. two non-autolysis controls (DLF_R003 and DLF_25). N = 3 for each group. B) Q-GRFT and biomass accumulation were monitored during instrumented 1 L fermentations of DLF_R004. Solid line and shaded area: gDCW/L, mean ± SD, N = 6. Circles: Q-GRFT g/L, mean ± SD, N = 1–3 per timepoint. Red dashed line: mid-exponential growth.

Fig. 2

Characterization of autolysis effects on endotoxin release and development of a heat shock-triggered autolysis method. A) EU per mg total soluble protein following lysis of DLF_R004 by either sonication or autolysis using the freeze-thaw method . B) muGFP fluorescence relative to sonication-lysed controls for DLF_R004 with or without 0.1% Triton X-100 at various temperatures. N = 3. C) muGFP fluorescence relative to sonication-lysed controls for DLF_R004 vs. DLF_R003 at 60 °C in the absence of Triton X-100. N = 3.

Fig. 3

In-fermenter autolysis and precipitation for rapid release and purification of Q-GRFT. A) SDS-PAGE showing accumulation of Q-GRFT throughout a 1 L fermentation and purified Q-GRFT harvested from the fermenter after autolysis and precipitation. Lanes: a = Mk 12 unstained standard (Thermo Fisher Scientific, Waltham, MA); b-f = His-GRFT standard, 1 μg to 62.5 ng in 2-fold series; g-n, 1.5 μg total protein from fermentation samples at 20, 22, 25, 28, 38, 46, 49, and 64 hrs after inoculation; o, 1 μg total soluble protein following autolysis and precipitation (67 hrs after inoculation). B) Temperature and pH profile inside the fermenter during the autolysis and precipitation operation. The upward inflection of pH at approximately 90 mins. marks the addition of ammonium sulfate.

Fig. 4

Yield of Q-GRFT and levels of key impurities after each stage of the bioprocess. At the fermentation stage, impurity levels are estimates based on the literature and the titer of Q-GRFT as determined by ELISA. At other process stages, impurity and Q-GRFT abundances were determined by the assays described in Materials and Methods. * : indicates that the corresponding value was not determined at a given process stage. * *: indicates that the value shown is an upper bound (DNA was not detectable by qPCR in the STIC-purified samples, LOQ = 0.02 pg/μL). Columns show the grand mean of 3 bioprocess replicates (except for fermentation, N = 6 bioprocess replicates) with technical replicates of N = 3, 4, 3, or 2 for yield, endotoxin, HCP and DNA measurements, respectively. Error bars show SD of the 3 bioprocess replicates. Yield at the fermentation stage is 100% by definition.

Fig. 5

Overview of the improved Q-GRFT bioprocess. A) Fermentation. B) An integrated procedure for cellular autolysis, auto-hydrolysis of DNA and RNA, and contaminant precipitation, which can be performed within the fermenter. C) and D) Lysate clarification by centrifugation and/or dead-end filtration. E) Buffer exchange by tangential flow filtration. F) Flow-through membrane chromatography. G) Bulk drug substance.

Fig. 6

Economic overview of the improved Q-GRFT bioprocess vs. a process using two chromatography columns . A) Total COGS and throughput for both processes at four scales. B) Cost structure for the processes and scales shown in A. U = upstream, R = primary recovery, P = purification, F = fill/finish. M = materials, Fa = facility, C = consumables, L = labor and QC. Color bar = COGS ($/g).