A role for Biofoundries in rapid development and validation of automated SARS-CoV-2 clinical diagnostics

Abstract

The SARS-CoV-2 pandemic has shown how a rapid rise in demand for patient and community sample testing can quickly overwhelm testing capability globally. With most diagnostic infrastructure dependent on specialized instruments, their exclusive reagent supplies quickly become bottlenecks, creating an urgent need for approaches to boost testing capacity. We address this challenge by refocusing the London Biofoundry onto the development of alternative testing pipelines. Here, we present a reagent-agnostic automated SARS-CoV-2 testing platform that can be quickly deployed and scaled. Using an in-house-generated, open-source, MS2-virus-like particle (VLP) SARS-CoV-2 standard, we validate RNA extraction and RT-qPCR workflows as well as two detection assays based on CRISPR-Cas13a and RT-loop-mediated isothermal amplification (RT-LAMP). In collaboration with an NHS diagnostic testing lab, we report the performance of the overall workflow and detection of SARS-CoV-2 in patient samples using RT-qPCR, CRISPR-Cas13a, and RT-LAMP. The validated RNA extraction and RT-qPCR platform has been installed in NHS diagnostic labs, increasing testing capacity by 1000 samples per day.

Fig. 1. MS2-SARS-CoV-2 VLP production and characterization.

a Schematic of the genetic construct of the engineered MS2-SARS-CoV-2 N-gene VLP encompassing the MS2 maturation protein (MP) in green and coat proteins (CP) in orange, linked via His-tag (yellow), under transcriptional control of the T7 promoter and T7 terminator sequences. The SARS-CoV-2 N protein RNA is packaged using a downstream pac site. Schematic created using DNAplotlib. b The VLP constructs were expressed in E. coli and purified using HiTrap® TALON® Crude and HiTrap® Heparin columns. SDS-PAGE analysis of the purification steps includes a protein marker (M) followed by pellet (P) and soluble fraction (S) of the cell lysate, followed by the column flow through (FT) and protein elution fractions, where the CP-His-CP dimer (~28 kDa) is indicated by an arrow. c Purified VLPs were analysed by dynamic light scattering (DLS), which showed a uniform particle population of ~27 nm. Error bars represent the SD of three technical replicates. d Reverse-transcriptase droplet digital PCR (RT-ddPCR) was performed for absolute quantification of the purified VLPs. Serial dilutions of 1, 10, and 100 thousand-fold of the purified VLPs in the presence and absence of a reverse transcription (RT) enzyme were analysed. Droplets were clustered using a threshold determined using a python implementation of an online tool (http://definetherain.org.uk). Dotted lines represent the cut offs for the positive and negative clusters. Any data points between the two dotted lines are considered droplet “rain”. Source data are available in the Source Data file.

Fig. 2. MS2-SARS-CoV-2 VLP detection with multiple target primer–probe sets and qPCR master mixes.

a VLP dilutions of 2.5, 25, and 250 copies per reaction were analysed by one-step RT-qPCR using the CDC primer–probe sets N1, N2, and N3 with the TaqPath master mix (Thermo Fisher Scientific) and reported as Ct values. b A Ct value standard curve for VLP concentrations of 1, 5, 10, 15, 20, 25, 50, 100, 150, 200, and 250 VLP copies per reaction was determined using the N1 primer–probe set and the TaqPath master mix. c VLP dilutions of 250, 25, and 2.5 copies per reaction were analysed using the TaqPath, Luna Universal (NEB), and Fast Virus (Thermo Fisher Scientific) RT-qPCR master mixes with the N1, N2, and N3 CDC primer–probe sets. All measurements in a and c are reported as mean ± SE of n = 3 independent experiments with three technical replicates. Measurements in b are reported as mean ± SD and are representative of n = 2 independent experiments with three technical replicates. Statistical difference between the TaqPath and Luna, as well as TaqPath and Fast Virus master mix Ct values was analysed using an unpaired two-sided t-test with (black star) indicating p < 0.05 and (double black star) p < 0.01. Source data are available in the Source Data file.

Fig. 3. SARS-CoV-2 N-gene target RNA detection by CRISPR-Cas13a and RT-LAMP diagnostic systems.

a Schematic of the CRISPR-Cas13a nucleic acid detection workflow from patient samples. Viral RNA is amplified using target-specific primers and detected with target-specific crRNA activating Cas13a to collaterally cleave a fluorescent probe. b CDC N1, N2, and N3 primer sets were employed to amplify the N-gene RNA released from MS2-SARS-CoV-2 VLPs at 2.5, 25, and 250 copies per reaction. The CRISPR-Cas13a detection time course was analysed using a fluorescence microplate reader. Error bars represent the SEM of n = 3 biologically independent amplification reactions and four CRISPR detection technical replicates. c Schematic of the RT-loop-mediated isothermal amplification (RT-LAMP) diagnostic workflow using target-specific LAMP primers. The isothermal amplification of a target results in the acidification of the RT-LAMP master mix and a subsequent pH-associated colour change that is detected using a microplate reader. d Time-course detection of three replicate RT-LAMP reactions using the MS2-SARS-CoV-2 VLPs at 0, 5, 10, 20, 30, and 40 copies per reaction, performed at 65 °C using the BMG CLARIOstar Plus microplate reader. Source data are available in the Source Data file.

Fig. 4. SARS-CoV-2 platform development using the MS2-SARS-CoV-2 VLP standard.

a Schematic of the modular platform for the detection of SARS-CoV-2. Viral RNA is isolated using the Analytik Jena FeliX liquid handler with the Analytik Jena or Promega RNA extraction kits. Sample RNA is transferred and detection reactions are set up using the Labcyte Echo platform. These include qPCR, validated for the TaqPath (Thermo Fisher Scientific), Luna (NEB), and Fast Virus (Thermo Fisher Scientific) RT-qPCR master mix options, RT-LAMP nucleic acid detection, and the CRISPR-Cas13a diagnostic workflow. b Automated RNA extraction was developed using VLP dilutions of 10³ and 10⁴ copies/mL for the Analytik Jena and Promega RNA extraction kits. Efficiency of the extractions using both kits was analysed by RT-qPCR with the CDC N1 primer–probe set using the TaqPath master mix. c Dilutions of VLPs used for RT-qPCR in b were analysed using the CRISPR NAT to demonstrate the use of this workflow as an alternative diagnostic option. Error bars in b represent mean ± SE of n = 3 biologically independent samples with three technical replicates. Error bars in c represent mean ± SE of n = 3 independent amplification replicates and four technical replicates for CRISPR detection. Source data are available in the Source Data file.

Fig. 5. SARS-CoV-2 platform validation using patient samples.

a Schematic of a typical workflow tracking viral copy number from a hypothetical patient sample where an input of 200 µL with a minimum 1 particle/µL results in 3–4 copies of extracted RNA per microliter resulting in a range of 15–20 RNA copies per qPCR reaction. b Comparison of RT-qPCR Ct value results for the same 173 patient samples obtained by the North West London Pathology (NWLP) laboratory and the London Biofoundry (LBF), based on the VLP workflow using the Analytik Jena RNA extraction kit and CDC N1 primer–probe set. c Validation of the Promega RNA extraction kit using 65 patient samples with the Analytik Jena RNA extraction workflow previously validated in b. Error bands in b and c represent the 95% confidence interval. d Validation of the CRISPR NAT using patient samples with Ct values determined via RT-qPCR. Error bars represent the mean ± SE of n = 3 independent amplification replicates and four technical replicates for CRISPR detection. e Validation of the RT-LAMP colorimetric NAT using patient samples with Ct values determined via RT-qPCR. Error bars represent the mean ± SE of n = 3 independent amplification reactions. Source data are available in the Source Data file.