Deploying a RAS pipeline against the SARS-CoV-2 pandemic

July 30, 2020, by Dom Esposito

Late on the evening of March 18, 2020, I was working on contingency plans for staffing the RAS Reagents Core (RRC) during the SARS-CoV-2 pandemic, when the first indication that we would become involved in studying the pandemic arrived. It came in the form of an email with the subject line “Help”. The sender was Dr. Matt Hall, a collaborator from his days in Michael Gottesman’s lab and now a Branch Chief at the National Center for Advancing Translational Sciences. Matt is part of an urgent project to develop nanobodies that could bind to the spike proteins of SARS-CoV-2 and potentially prevent the virus from binding to its ACE2 receptor on human cells. Matt asked for our protein purification help because the expression of the nanobodies was too low to supply sufficient proteins for assay development and structural studies.

We obtained approval from NCI to reallocate some resources to tackle this urgent need, but within 48 hours our planning had to be altered again to accommodate new NIH staffing, work-from-home, and social distancing requirements. Thus our work for NCATS required us to minimize personal interactions and lab and office co-occupancy, which arrangements were facilitated by round-the-clock, seven day work schedules. Technologies from the RAS Initiative that proved especially important included our electronic lab record system to coordinate all phases of the work, and an alternative bacterial host, Vibrio natriegens, that was essential when E. coli expression was inadequate. Over the course of several weeks we developed a high-yield production protocol for the NCATS nanobodies that enabled assay development and CryoEM structural determinations [Fu, Hall, et al., manuscript in preparation].

Seven days after Matt Hall’s email a second request for pandemic-related assistance arrived, this one from Dr. Kaitlyn Sadtler, an Earl Statdman Investigator at the National Institute for Biomedical Imaging and Bioengineering (NIBIB). Kaitlyn was working with Matt Hall on a cross-institute project to develop a quantitative ELISA-based serology assay for CoV-2 antibodies compatible with population surveys of 10,000 patients or more. The initial concept was that we would purify proteins just for assay development. However, a flurry of virtual meetings with Kaitlyn, Matt, and the third project lead, Dr. Matt Memoli, a clinician from the National Institute of Allergy and Infectious Diseases (NIAID), quickly resulted in escalation of the scale of the work to support deployment of the assays to sample human populations. In addition, it was highly desirable to detect antibodies to both the full-length spike protein of the virus, as well as the receptor-binding domain (RBD) of spike, so that the final assays would have maximum sensitivity (few false negatives) and specificity (few false positives).

There were multiple versions of these proteins circulating—spike protein variants from the NIAID Vaccine Research Center (generated initially in the lab of Dr. Jason McLellan at U-Texas at Austin) and Dr. Florian Krammer’s lab at Mt. Sinai, and RBD variants from the Krammer lab as well as from Dr. Aaron Schmidt’s lab at the Ragon Institute. Armed with the four DNA constructs obtained from these different sources, our lab was asked to rapidly generate a series of proteins which we had never seen before, and for which published protocols were lacking in detail. In addition, spike had gained notoriety for being difficult to produce, with best case yields around 1-2 mg per liter of cell culture. Additionally, proteins were to be generated using secretion from mammalian cells —a process which the RRC had used infrequently and generally only in small scale. 

Around the same time, word came from Dr. Doug Lowy, NCI Principal Deputy Director, that NCI was initiating a large CoV-2 serology effort centered in Dr. Ligia Pinto’s HPV Serology Group, one of our neighbors in the NCI’s Advanced Technology Research Facility. Ligia’s group would be developing assays and standardized reagents for a centralized serology hub as part of a new NCI SeroNet initiative, and she asked if we could generate proteins for their assay optimization work. Fortunately there was significant overlap in the proteins required for these two projects, so the team was able to develop a plan for large-scale production of these essential reagents. 

As a consequence, while the bulk of RRC staff were transitioning to telework status we recalled a dozen personnel to begin work on production of the needed CoV-2 proteins. The team quickly tested expression of the four proteins and confirmed that yields of spike proteins were indeed problematic. Using the existing low-yield protocols, RRC eukaryotic expression team members implemented round-the-clock operations to continuously feed, transfect, and harvest cells so as to provide material to the purification groups. Nearly every piece of our cell culture equipment was in service. Once the immediate needs of both the SeroNet and cross-institute groups were satisfied, RRC turned to optimizing production of the proteins for the projected higher levels of protein required. Protein purifiers worked 12+ hour shifts to process large culture volumes (over 180 liters total) and test multiple expression and purification conditions.

In the meantime, Kaitlyn’s team at NIBIB was able to demonstrate that both spike proteins produced similar ELISA results, and that the RBD format from the Ragon Institute gave better assay results, thus contracting our protein needs for that project from four to two. The RRC team was able to improve production of both spike and RBD proteins dramatically—increasing spike yields by nearly 10-fold over published data and identifying the vital variables in expression and purification that led to the improvements. In keeping with the urgency of the public health crisis, this work published less than 2 months after receipt of the first SARS-CoV2 plasmid in the laboratory. More importantly, our optimized processes (since published) were used to produce single large batches of spike and RBD protein to support the NIH cross-institute serosurvey initiative. Currently, the project has enrolled nearly all of the target 10,000 subjects and more than a quarter of the samples have already been assayed. When completed, this study will be one of the most diverse and quantitative serosurveys for IgG, IgA, and IgM antibodies to SARS-CoV-2 antigens, sampling people across the country and a variety of demographics to give a more complete understanding of the course of CoV-2 infection.

In yet another application of the SARS-CoV-2 antigens produced by the RRC team, we received an urgent request from the NIAID Vaccine Research Center for spike and RBD proteins to help support serology work on the phase 1 trial of the joint Moderna/NIAID mRNA vaccine. We delivered the proteins from our frozen stocks in less than 48 hours, allowing the VRC to complete serological assays which were recently published in the New England Journal of Medicine.

I should emphasize that producing these reagent-grade proteins in a matter of weeks would not have been possible without the NCI’s consistent investment in and support of the RAS Initiative. The urgency of the effort was new, but our staff benefitted from the efficient and smoothly-operating pipeline that we have evolved to deliver hundreds of RAS pathway proteins to our RAS Initiative colleagues. Research on RAS-driven cancers has been disrupted by the pandemic, but perhaps our assistance in understanding SARS-CoV-2 will help us get back to the work of defeating RAS more quickly.