Fluorescence-Linked Aptamer Assay for SARS-CoV-2 Spike-Protein: A Step-by-Step Performance Analysis in Clinical Samples

Abstract

The COVID-19 pandemic has been a main concern over the last two years and has become one of the most important crises in the history of human health. Today, there is still a need for affordable and reliable diagnostic tests for massive disease monitoring. Previously, a set of highly specific DNA-aptamers (C7/C9) binding to the SARS-CoV-2 Spike (S) protein were isolated but its performance in clinical samples remained to be tested. Here, 242 samples were collected through three different methods and subjected to florescence-linked aptamer assays (FLAA) based on C7/C9 aptamers through two readout protocols. Then, a step-by-step statistical approach which included agreement tests, proportion comparisons and binomial and multinomial logistic regressions was used to predict optimal conditions for the novel C7/C9 FLAA test. RTqPCR threshold cycles, symptoms onset and processing time were influential factors on FLAA test results. Naturally occurring mutations on S were also detected and analyzed. Aminoacidic substitutions D614G and T732A appeared relevant for aptamer recognition although further studies are necessary. The methodology presented here is the first step to determine the performance and diagnosis across a range of clinical contexts and it might serve as a base for a complete analysis applicable to other designs of new diagnostic tests.

Keywords: COVID-19; FLAA; SARS-CoV-2; aptamer; diagnostic test; spike protein.

Scheme 1

Tridimensional (PDB: 6ZGG) and linear representations of SARS-CoV-2 spike protein. Mutations T478K, T732A, D614G and P681H are presented as red, blue and green volumes. Red dots in linear drawing of S-protein represent sites of reported mutations for Omicron variant. Black circles are detected mutations shared with Omicron variant. Yellow rectangle corresponds to characteristic mutations of delta variant.

Figure 1

Design of the FLAA assay for SARS-CoV-2 S-protein detection.

Figure 2

Distribution of fluorescence readings in tested samples. (a) Scatter-plot colored for direct and indirect protocols; (b) Scatter-plot colored by sample type.

Figure 3

Binomial test for RTqPCR and FLAA positive and negative proportions. No discriminative probability was 0.5, p< 0.05 were considered significant.

Figure 4

Frequency plots for positive and negative samples percentage divided by protocol and sample types. Yellow, red and purple rectangles represent proportions that should be the same as RTqPCR. ** represent most varying frequency ratios.

Figure 5

ROC curves for (a) overall samples; (b) protocol; and (c) sample type. (d) Fagan nomogram for post-test probability of positive (red line) and negative outcomes (blue line). Yellow squares delimit cutoff zones with high sensibility and specificity.

Figure 6

Estimated marginal means plots for (a) time from symptoms onset and (b) time for sample processing.

Figure 7

Recalculated ROC curve using all cut-off points found for experimental and time variables. The numbers in parentheses correspond to (Sensitivity, Specificity, Youden index, AUC). Red line: ROC curve, Black line: non-discrimination line.

Figure 8

Recalculated ROC curves categorized by nCtOrf1ab ranges. (a) Scheme of Ct ranges used for curves adjustment, Red arrow: Ct scale, for independent, green bar: <20;0–20, pink bar: 20–25 and blue bar: 25–30; for overlapped, pink bar: <25, blue bar: <30. (b) ROC curve for Ct = 20–25, (c) ROC curve for Ct = 25–30 (), (d) ROC curve for Ct < 20, (e) ROC curve for Ct < 25 and (f) ROC curve for Ct < 30. Red lines: ROC curves, Black lines: non-discrimination lines.

Figure 9

Frequency plot and dot-plots for means of normalized fluorescence and CtOrf1ab. * = moderate proportion variability; ** = high proportion variability and *** = highest variability in positive:negative FLAA test results. Pink zone shows the means below negative samples.

Figure 10

Estimated marginal means plots for (a) normalized fluorescence; (b) time for Symptoms onset and (c) time for sample processing.