Optimization of Acceptor to Donor Ratio and Biotinylated Anitbody Concentration

We next determined the S/B at a greater range of untagged recombinant NP (20 000–81.9 pg/mL) using donor/SA acceptor ratios of 10–40 μg/mL and 20–20 μg/mL and different concentrations of biotinylated antibody from 0.5 to 5.0 nM (Figure S-2). We selected a donor/SA ratio of 10–40 μg/mL for both pairs and 1.0 and 5.0 nM biotinylated antibody for Pairs 1 + 4 and 11 + 4, respectively, based on the S/B calculations (Figure S-2). The standard curves were generated using recombinant NP for these conditions, and no plateau was observed at the higher concentrations, indicating that the upper limit of detection was more than 20 000 pg/mL (20 ng/mL) (Figure S-3).

Lysis Buffer Selection and Viral Lysate Testing

To determine the limits of detection and dynamic range of the assay, we next generated standard curves using concentrations of NP from 200 000 to 0.02 ng/mL for both pairs (Figure ​Figure22A,C; Table S-1). We observed a significant hook effect at concentrations of recombinant NP above 320 ng/mL for pair 1 + 4 and 1600 ng/mL for pair 11 + 4 (Figure ​Figure22; Table S-1). The hook effect indicates the concentration of analyte at which an overabundance of analyte titrates the donor and acceptor away from each other, and this hooking dictates the upper limit of detection before the signal will start to decrease. Two different lysis buffers were used to optimize any lysis buffer effects: The NCATS lysis buffer consisted of 0.5% Triton-X 100 with protease inhibitor; the other was the AlphaLISA lysis buffer from the manufacturer. For pair 1 + 4, both lysis buffers performed equally, while the AlphaLISA lysis buffer performed better for pair 11 + 4 in generating the standard curve. The overall S/B was higher for pair 11 + 4 at higher concentrations of NP, but pair 1 + 4 had a higher sensitivity as seen by the larger S/B values at lower concentrations (Figure ​Figure22B,D; Table S-1).

Open in a separate window

Figure 2

SARS-CoV-2 NP detection in virus-infected cell lysates and tissue culture supernatants. (A) Standard curve for Pair 1 + 4 demonstrating hook effect using AlphaLISA or NCATS lysis buffer. (B) (Left) AlphaLISA signal (counts) for 24 and 48 h virus-infected cell lysates and 0, 24, and 48 h tissue culture supernatants at various dilutions. (Right) Same as left panel with a narrower dilution range. (C) Standard curve for Pair 11 + 4 demonstrating hook effect using AlphaLisA or NCATS lysis buffer. (D) (Left) AlphaLISA signal (counts) for 24 and 48 h virus-infected cell lysates and 0, 24, and 48 h tissue culture supernatants at various dilutions. (Right) Same as left panel with a narrower dilution range. Data points beyond the hook effect are shown but not included in the range for the nonlinear regression curve fit. N = triplicate wells. Error bars indicate SD.

To determine whether the detection system could detect native NP in SARS-CoV-2-infected lysates, TCS collected at 0, 24, and 48 h along with cell lysates collected at 24 and 48 h postinfection were tested at multiple dilutions. Pair 11 + 4 produced absolute AlphaLISA signal (counts) than greater that of pair 1 + 4 at the highest concentrations, but both pairs exhibited a hook effect at 1/15 lysate dilutions (Figure ​Figure22D). NP was better detected in TCS by pair 11 + 4 than by pair 1 + 4. The concentrations of NP in ng/mL were interpolated from the standard curves. For both pairs, the concentration of NP was approximately 9-fold greater in 48 h TCS than in 24 h TCS. For pair 11 + 4, the 24 and 48 h TCS concentrations were 80.3 and 708 ng/mL, respectively. The 24 and 48 h lysate NP concentrations were significantly higher at 7380 and 30 200 ng/mL, respectively. These concentrations were comparable, although not exactly the same, when comparing pairs 1 + 4 and 11 + 4.

Sequential Two- versus Three-Step Assay Protocol Optimization

We next optimized the sequential addition of assay reagents including lysis buffer, biotinylated antibody, acceptor beads, and donor beads in the two- versus three-step assay protocol (Figure S-4). For pairs 1 + 4 and 11 + 4, the two-step assay produced a greater dynamic range for recombinant NP detection than did the three-step assay without exhibiting a hook effect (Figure S-4) as determined by the S/B ratios. NCATS lysis buffer performed slightly better than AlphaLISA lysis buffer. We further optimized the two- and three-step assays for recombinant NP spiked into Vero-E6 cells grown in 384-well plates (Figure S-5). In most cases, the two-step assays performed better with 20 000 cells/well as measured by S/B ratios. The effect of media and cell density was also tested for assay interference. Media affected pair 1 + 4 to a greater extent than pair 11 + 4, while cell density did not significantly affect S/B ratios. The lower limit of detection for both pairs was 0.22 ng/mL with an upper limit at or beyond 400 ng/mL (Figure S-6).

Assay Volume Optimization

Next, we optimized the total assay volume in a 384-well plate format using recombinant NP in AlphaLISA LB and found that higher total assay volumes produced lower absolute AlphaLISA signal (counts), but 100 uL total volume performed better than 50 uL total volume (Figure ​Figure33A,B; Table S-2). Again, pair 11 + 4 exhibited a higher limit of detection than pair 1 + 4 (Figure ​Figure33B,D). In this assay, the 20 μL condition utilized the 384-well Proxiplate, while the 50 and 100 uL conditions utilized the 384-well CulturPlate. The main difference between the two is the shallower circle wells in the Proxiplate compared to the deeper square wells in the CulturPlate, suggesting the plate type and well dimensions change the S/B ratios.

Open in a separate window

Figure 3

Optimization of assay volumes for 384-well plate using recombinant NP. (A) Detection of titrated NP in wells with different final volumes of 20, 50, and 100 μL using Pair 1 + 4. (B) Detection of titrated NP in wells with different starting volumes 20, 50, and 100 μL using Pair 11 + 4. N = triplicate wells. Error bars indicate SD.

We further optimized media conditions and found that cell culture media itself had a profound effect on the S/B ratios, while the concentration of FBS had little to no impact on the assay performance (Figure S-7).

Batch to Batch Reagent Reproducibility

A second batch of reagents was produced, and the concentration of biotinylated antibody was matched to generate the same performance as the first batch for pair 1 + 4 and pair 11 + 4 (Figure S-8). We tested the second batch of reagents with the recombinant NP spiked into Vero-E6 cell culture in 384-well plates and found better performance with pair 11 + 4 when cells were present in the wells (Figure ​Figure44A–D, Table S-3). The lower limit of detection was 0.2 ng/mL, and the upper limit was greater than 500 ng/mL NP. Altogether, pair 11 + 4 performed better than did pair 1 + 4 and was selected as the final pair for future assays. To confirm that the second batch of reagents performed well, we tested the same conditions for optimization as before with cell densities ranging from 20 000 to 0 cells/well, along with a no media control in PBS.

Open in a separate window

Figure 4

Optimization of 384-well plate cell number in simulated infection using second preparation of reagents. (A) Detection of titrated NP in wells culturing 20 000, 10 000, 5000, or 0 cells/well along with a no-media control in Vero-E6 cells using Pair 1 + 4 (second batch). (B) Detection of titrated NP in wells culturing 20 000, 10 000, 5000, or 0 cells/well along with a no-media control in Vero-E6 cells using Pair 11 + 4 (second batch). N = triplicate wells. Error bars indicate SD.

Lysis Buffer and Incubation Time Optimization, And Determination of Cell Lysate Interference

The AlphaLISA buffer and NCATS lysis buffers (LB) were compared in plates incubated for 30 min versus 1 h and for the AlphaLISA buffer there was a slight increase in signal after 1 h incubation (Figure S-9). The incubation time had little effect on the S/B for the NCATS LB condition. NP detection in the presence of lysed 10 000 and 5000 cells/well performed similarly to NP detection in media and better than in 20 000 cells/well. A concentration of 20 000 cells/well would be considered over-confluent, beyond 100%. The no-media PBS control had a lower S/B ratio overall. The main difference between the LBs was that a signal plateau was observed for the NCATS LB indicating an upper limit of detection of approximately 1500 ng/mL, whereas the AlphaLISA buffer could potentially detect higher concentrations. Importantly, the sensitivity of the assay with NCATS LB reveals a similar lower limit of detection of 0.23 ng/mL as in previous experiments, whereas the AlphaLISA buffer was observed be less sensitive, for example, with a lower limit of detection of 2.06 ng/mL for 10 000 cells/well.

Estimation of NP Concentration in Cell Lysates Using Standard Curve Interpolation

Finally, we estimated the concentration of NP in TCS and lysate from cells infected with live SARS-CoV-2 for 24 and 48 h compared to the noninfected (0 h) controls using interpolation from the 10 000 cells/well standard curve (Figure S-10). Dilutions of TCS and lysates from 15-fold to over 1900-fold were created, and NCATS LB was added to the samples. The linear portion of the curves was used for interpolation. The concentrations of NP in lysates was determined to be 7300 and 34 092 ng/mL for 24 and 48 h lysates, respectively (Figure S-10). In correspondence with the NP concentration, the viral titers were calculated to be 8.2 × 10⁰⁶ and 1.1 × 10⁰⁹ focus forming units (FFU)/mL for 24 and 48 h TCS, respectively, indicating increasing amounts of viral particles with longer infection times. This data indicated that the NP in viral lysates could be quantified using the AlphaLISA NP detection reagents and standard curve interpolation.

Preparation of Antibody Pair Matrixing

A matrix with all antibodies provided by PerkinElmer was tested for the ability to detect NP.

The untagged NP and His-tagged NP were diluted in AlphaLISA lysis buffer at 10 000, 1000, and 100 pg/mL. For initial testing, 20 μg/mL AlphaLISA Acceptor (final), 1 nM (final) biotinylated antibody, and 20 μg/mL donor (final) were used. A two-step assay was performed (data points in duplicate) using the following protocol:

• 1.

  Dispense 5 μL of recombinant protein.

• 2.

  Dispense 10 μL of mix of acceptor beads and biotinylated antibody.

• 3.

  Incubate 60 min at room temperature (RT).

• 4.

  Dispense 5 μL of SA-Donor Beads.

• 5.

  Incubate 30 min at room temperature.

• 6.

  Read plates on EnVision (PerkinElmer).

Antibody Concentration Optimizations

Best pairs 1 + 4 and 11 + 4 were optimized using the following parameters: Biotinylated antibody concentrations tested were 0.5, 1, 2, and 5 nM. Acceptor bead–SA donor bead concentrations tested were at 20–20 μg/mL and 10–40 μg/mL. The concentration of untagged NP recombinant protein started at 20 000 pg/mL followed by 2.5-fold dilutions. The dispensing protocol was the same as for the antibody pair matrixing.

Vero-E6 Cell Culture

Vero-E6 (grown in EMEM, 10% FBS, and 1% penicillin/streptomycin), were cultured in T175 flasks and passaged at 95% confluency. Briefly, cells were washed once with PBS and dissociated from the flask using 0.25% Trypsin. Cells were counted prior to seeding.

Preparation of Viral Lysate and Tissue Culture Supernatant Production

Vero-E6 cells were plated in 12-well plates at 450 000 cells/well in 1.25 mL of growth media. Cells were incubated for 24 h at 37 °C. A 250 μL aliquot of SARS-CoV-2 [USA-WA1/2020 strain (Gen Bank: MN985325.1)] was used at an MOI of 0.05. Cells were inoculated for 45 min at 37 °C. Supernatant was collected and pooled for 0 h TCS. A final concentration of 0.5% Triton-X 100 and 1× protease inhibitor was used for all samples. TCS samples were stored at −20 °C until needed. Next, 1.5 mL of fresh media was added to each well, and cells were incubated for 24 and 48 h at 37 °C. Six wells were harvested at each time point. Supernatant was collected from 6 wells each at 24 and 48 h. Samples were pooled for 24 or 48 h TCS. For viral titer calculations, 200 μL of TCS was removed and diluted with equal volume of lysis buffer. Samples were stored at −20 °C until needed. Lysate was collected from 6 wells each at 24 and 48 h. Cells were rinsed once with ice-cold PBS, and 250 μL/well of cell lysis buffer + PI was added to lyse cells. Protease inhibitor was added to the lysis buffer before lysing cells. Cells were scraped to the bottom of the well, and pooled lysates were collected into 1.5 mL tubes on ice. Tubes were vortexed for 10 s and placed on ice for 10 min. This was repeated three times. Lysates were spun down for 30 min at 13.2 000 rpm at 4 °C and stored at −20 °C until further use.

Optimization of Assay Using Viral Lysates and Tissue Culture Supernatant

The best conditions of pairs 1 + 4 and 11 + 4 were used to expand the standard curve and test TCS and lysate samples: 1 + 4: 10–40 μg/mL with 1 nM biotinylated antibody, 11 + 4: 10–40 μg/mL with 5 nM biotinylated antibody. An 11-point curve of recombinant untagged NP started at 200 000 ng/mL, followed by 5-fold dilutions. Lysates and TCS were diluted in PBS + Triton-X 100 + protease inhibitors 1:15 (1×) followed by 3-fold dilutions. The dispensing protocol was the same as above.

Simulation of Viral Infection Using Recombinant NP

The best conditions of pairs 1 + 4 and 11 + 4 were used to test the assay in a cell-based format: 1 + 4: 10–40 μg/mL with 1 nM biotinylated antibody and 11 + 4: 10–40 μg/mL with 5 nM biotinylated antibody. Vero E6 cells were plated at 20 000 and 50 000 cells/well in 384-well plates (TC) and incubated overnight. An 11-point curve of recombinant NP started at 4000 ng/mL initial (400 ng/mL final) followed by 3.5-fold dilutions in AlphaLISA lysis buffer.

For second batch preparation, the best conditions of pairs 1 + 4 and 11 + 4 were used to test the assay in a cell-based format (newly conjugated and biotinylated Ab were used and optimized to reproduce previous antibodies): 1 + 4: 10–40 μg/mL with 0.5 nM Biotinylated antibody, 11 + 4: 10–40 μg/mL with 1 nM Biotinylated antibody. Vero E6 cells were plated at 20 000, 10 000, and 5000 cells/well in 384-well plate (TC) and incubated overnight. The assay was also tested without cells and without media as a control. An 11-point curve of recombinant NP started at 2000 ng/mL initial (500 ng/mL final) followed by 3.5-fold dilutions in media.

384-Well Plate Assay Statistics

Whole-plate statistics were calculated using recombinant NP at a concentration of 500 ng/mL in media using a scaled down protocol to conserve reagents in a white small-volume 384-well plate (Greiner 784075). Reagents were added using a fully automated BioRAPTR FRD Workstation (Beckman Coulter). Briefly, 8 uL of cell culture media with or without NP was added to columns 3–24 and columns 1–2, respectively. Next, 2 uL of 5X AlphaLISA Lysis buffer was then dispensed to wells and agitated for 30 min. Then, 4 uL of mixed biotinylated anti-NP antibody and AlphaLISA acceptor beads were dispensed into each well and incubated for 60 min. Last, 2 uL of streptavidin donor beads were dispensed and incubated for 30 min at room temperature. The plate was then read on the Pherastar with an AlphaLISA module from BMG Labtech. Z-factor, CV, and S/B was calculated for the entire plate.

This research was supported in part by the Intramural Research Program of the National Center for Advancing Translational Sciences, NIH. We thank Dr. Arturo Gonzalez-Moya and his team at PerkinElmer for assay development support.