Introduction

Dengue fever (DF), dengue hemorrhagic fever (DHF), or dengue shock syndrome (DSS) are viral infections transmitted by mosquitoes that have emerged as significant public health issues in tropical and subtropical areas across the globe1. Around 3.9 billion individuals in 128 nations are susceptible to contracting dengue, with 70% of documented cases primarily found in Asia2, with developing countries also facing a substantial economic burden from these diseases3. Recent research in Pakistan has revealed that the dengue virus is circulating extensively4. Since July 2024, reports regarding dengue disease have been resurgent in Iran5. Four serotypes (DEN-1, DEN-2, DEN-3, and DEN-4) can currently be distinguished, which are genetically similar but have different antigenic properties. These serotypes can circulate simultaneously in the same geographical area, creating a complicated obstacle for efforts to prevent and manage them. The incidence of recorded fatalities due to dengue has experienced a substantial increase over the years, escalating from 960 in the year 2000 to 4.032 in the year 20156. According to the World Health Organization (WHO), there are an estimated 250,000-500,000 cases of dengue hemorrhagic fever (DHF) and DSS worldwide each year, associated with ≈ 25,000 deaths. Multiple diagnostic tests are available to detect dengue virus infection. Serological assays identify dengue virus-specific IgM antibodies, which usually emerge in the patient’s blood after about a week of infection. Confirmation of dengue infection can also be achieved by detecting the presence of the NS1 protein. The virus, classified within the Flavivirus family and the Flavivirus genus, contains three structural proteins (capsid, membrane, and envelope) and seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5)7. The non-structural glycoprotein NS1, with a molecular weight of 48–50 kilodaltons, is considered vital for the replication of the virus and a useful biomarker for detecting dengue. Enzyme-Linked Immunosorbent Assay (ELISA) and Rapid Diagnostic Tests (RDTs) are the most common methods for routine clinical diagnostics due to their ease of use and reliability. PCR is also widely used for viral confirmation, particularly in cases of co-infection or when more detailed viral characterization is required8. The concentration of NS1 varies during the infection, with initially modest concentration levels during the first two days but increasing rapidly, reaching their highest point around day four9. Progress has centered on developing cheap, simple, yet sensitive diagnostic alternatives, such as lateral flow immunoassays (LFIAs)10. LFAs commonly employ cellulose-based paper, notably nitrocellulose (NC) membrane, modified with analyte-specific capturing antibodies, with colloidal metal nanoparticles employed as detection labels. Existing colorimetric lateral flow immunoassays (LFIAs) for DF detection encounter drawbacks such as reduced sensitivity, specificity, and elevated limits of detection (LOD)11.

Aptamers are short chains of nucleotides that possess a distinct shape and selectively attach to particular targets. Aptamers are selected from a large pool of 10¹³ to 10¹⁶ randomly generated candidates in a laboratory setting. They offer advantages such as being cost-effective, easily produced, and having minimal variation across batches compared to antibodies12. Aptamers also possess increased stability compared to antibodies and can be easily modified chemically, and numerous aptasensors have been created13. Indeed, LFA-based aptamer-antibody diagnostics are effective in identifying viral proteins from Salmonella Typhimurium14. Additionally, LFA-based aptamer-antibodies can detect thrombin15, interleukin-616, and CXCL917, exhibiting exceptional accuracy and outstanding detection thresholds.

This article describes the cloning of the NS1 protein into the pET-23b vector and its expression in E. coli. A diagnostic test was developed using a novel aptamer-antibody sandwich assay targeting NS1 utilizing the recombinant protein. This test was also successfully applied to real serum samples. LOD determined as 5.2 ng/mL (95% CI: 4.6–5.8) in buffer and 6.1 ng/mL (95% CI: 5.5–6.8) in serum, based on ImageJ quantification and logistic regression analysis.

Materials and methods

Materials

TSC (sodium citrate tribasic), HAuCl4 (gold chloride), tris(2-carboxyethyl) phosphine TCEP, NaCl, Tween20, PBS Buffer, streptavidin, and bovine serum albumin (BSA), all from Sigma Aldrich; a sample pad (Millipore); a conjugate pad (Millipore); a nitrocellulose membrane (NC 45, Schleicher & Schuell); an adsorption pad (Millipore); NS1 Antibody (Orient Gene, China); and restriction endonucleases and modifying enzymes were purchased from New England Biolabs (Mississauga, Canada). The anti-His6 MAb was purchased from Abcam (Shanghai, China). Glutathione (reduced and oxidized), sodium deoxycholate, L-arginine, GAM-HRPO, urea, and other general molecular biology grade reagents were purchased from Sigma–Aldrich (Steinheim, Germany). Ni–NTA agarose, plasmid DNA isolation, and gel extraction kits were obtained from Qiagen (Mississauga, Canada). Low-range protein molecular weight markers, 40% acrylamide: bisacrylamide, glycine, and protein assay reagents were purchased from Bio-Rad (Mississauga, Canada). Thioled-aptamer Sequence 5’ SH/CCGCACCGGGCAGGACGTCCGGGGTCCTCGGGGGGCGGG CCCCCCCCCCCCCCCCCCCC-3’18, the underlined part is also added to the sequence in our study. 5’-biotin-GGGGGGGGGGGGGGGGGGGG‐3’, were synthesized for the control zone by Sinacolon in Tehran, Iran.

Synthesis of gold nanoparticles (Au NPs) and Aptamer-Au NPs

The Synthesis of gold nanoparticles (Au NPs) was done according to Frens et al.19, with certain modifications. To a solution of 0.01% HAuCl4, TSC (1%, 1 mL) was added and heated to boiling. After 2–5 min, the initial yellow solution changed and turned colorless, subsequently transitioning to a bluish-gray hue. After an additional 5 min, the solution changed to reddish-purple, which signified the successful synthesis of Au NPs. The solution was agitated for a further 10 min and thereafter cooled to ambient temperature. To obtain gold nanoparticles (AuNPs) of varying sizes, varying concentrations, and volumes of trisodium citrate (TSC) are added to gold chloride (HAuCl4). To prepare the aptamer-conjugated Au NPs, the disulfide-modified oligonucleotide (50 µM) was treated with TCEP (100 µM) for 1 h at room temperature. Subsequently, the solution was combined with 1 mL of a 20 nm Au NP solution (1 nM). The resulting mixture was then incubated at room temperature for 12 h. NaCl (3 M, 20 µL) was gradually introduced into the mixture and subjected to sonication for 10 s. This step was done 5 times with a time interval of 1 h. The final solution was kept at room temperature for 24 h. The solution underwent centrifugation at a speed of 13,500 revolutions per minute for 30 min to separate the AuNPs from the unreacted reagents. The aptamer-Au NPs were subsequently rinsed thrice with a 1/2 × PBS buffer (pH 7.4) solution containing 0.01% Tween 20. They were then resuspended in a PBS buffer20.

Formation of aptamer-streptavidin conjugates

A biotinylated aptamer (42 µL) solution with an optical density of 1.0, measured at 260 nm using a Shimadzu UV-2600 spectrophotometer, was combined with 250 µL of streptavidin (2 mg mL−1). The optical density was measured following standard nucleic acid quantification protocols, and the Shimadzu UV-2600 instrument specifications confirm measurement reliability. Following a 1-hour incubation at room temperature, PBS1 × (508 µL) was introduced into the mixture. The surplus aptamer was eliminated using centrifugation for 30 min using a centrifugal filter (with a cutoff of 30,000, Millipore) at a speed of 6000 revolutions per minute. The conjugate underwent two washes using PBS (500 µL) in the same centrifugal filter. The residual solution in the filter was gathered and subsequently diluted to a volume of 500 µL by adding PBS. Streptavidin-biotinylated DNA was applied onto the nitrocellulose membrane in the control zone using the BioJet (BJQ300) dispenser.

Isolation of NS1

Construction of plasmid

The NS1 full-length nucleotide sequence of dengue (DEN-2) was codon optimized for E. coli expression and chemically synthesized by Metabion International AG (Metabion, Planegg, Germany). The optimized codon NS1 gene was obtained from GenScript Codon Optimization. The NS1 sequence was cloned into the expression vector pET23b (Novagen, USA) with NotI and EcoRI. The sequence was suggested in our data by Sanger sequencing. These observations should be validated with larger-scale clinical and field testing.

Transformation and expression of the NS1 protein

The vectors carrying the NS1 sequence were introduced into the BL21 (DE3) strain of E. coli using a combination of CaCl2 and heat shock. Positive clones were obtained by isolating them on an LB-agar medium with 100 µg/ml ampicillin and incubating them at 37 °C for one night. A single, separate colony with a positive result was introduced into 10 mL of LB medium enhanced with 100 µg/ml of ampicillin. The culture was thereafter agitated at 190 revolutions per minute (rpm) and kept at 37 °C for the night. The culture that had been developed overnight was combined with a fresh medium containing antibiotics in a ratio of 1:50, resulting in a final volume of 1 L. The mixture was then incubated at 37 °C until the optical density (O.D. 600 nm) reached around 0.6–0.8. The protein expression was stimulated by employing 0.5 mM IPTG, followed by a 4-hour expression duration. Subsequently, the cells were harvested using centrifugation at a speed of 4000 revolutions per minute for 20 min while maintaining a temperature of 4 °C. Subsequently, the lysate, which consisted of all the cellular components, was generated. The pellet was placed in a solution of 25 mL of Buffer A (50 mM Tris–HCl pH 8, 100 mM NaCl, and 1 mM β-Mercaptoethanol) that included 1 mM PMSF. Later, a solution containing 5 mg/ml of lysozyme, 10 µg/ml of DNase A, and 5 mM of MgCl2 was supplied. The solution was placed in a temperature-controlled environment at 4 °C and stirred for 30 min. The complete cellular lysate was sonicated using five cycles, each consisting of 20 s of sonication followed by 2 min of rest, at an amplitude of 40%. The inclusion bodies were purified using centrifugation at a speed of 14,000 revolutions per minute for 30 min. Finally, the inclusion bodies were restored in Buffer B (Buffer A and 8 M urea) overnight at a temperature of 4 °C.

Purification of NS1 via affinity chromatography from inclusion bodies

To isolate the protein from the inclusion bodies, a technique called Ni2+NTA affinity chromatography was used. This technique utilizes a column called HisTrap HP 5 mL, which Sigma-Aldrich manufactures in the United States. The reason for using this technique is that the recombinant NS1 protein (rNS1) has a 6× histidine tag attached to its C-terminal end. After attaining column equilibrium by running 10 column volumes (CV) of Buffer B through it, the protein was injected into the column at a flow rate of 0.5 ml/min. Subsequently, a wash was performed with Buffer B, using a volume equal to 10 times the volume of the column, to remove any proteins that were bound non-specifically. The elution process was conducted using a gradient of Buffer B and Buffer BI (Buffer B and 0.5 M imidazole) at a flow rate of 1 ml/min. The obtained samples were analyzed using 12% SDS-PAGE. The tubes containing the rNS1 were consolidated for later refolding.

rNS1 refolding

The rNS1 protein was refolded using the rNS1 refolding technique. In this protocol, 10 mg of rNS1 that had undergone denaturation with urea was diluted to a concentration of 100 mg/ml before the refolding step. The rNS1 protein underwent a 20-fold dilution using Buffer ASB (Buffer A and 0.1% (w/v) ASB-14) and was subsequently subjected to dialysis in a solution of 50 mM Tris–HCl (pH 8), 100 mM NaCl, 1 mM DTT, and 0.05% ASB-14. All of these investigations were carried out at a temperature of 4 °C.

Characterization

UV-Vis spectroscopy

The synthesis of spherical gold nanoparticles (AuNPs) and DNA-conjugated AuNPs was suggested in our data using a Shimadzu UV-2600 UV-Vis spectrophotometer (Shimadzu Corporation, Kyoto, Japan) in the 400 to 700 nm wavelength range. UV-Vis spectroscopy is a reliable technique for verifying the formation of spherical AuNPs, as their absorption spectra in the visible range reflect a uniform response to light due to the even distribution of conduction band electrons across their surface. These observations should be validated with larger-scale clinical and field testing.

Dynamic light scattering (DLS)

Gold nanoparticles (AuNPs) were characterized using dynamic light scattering (DLS) with the SZ-100z instrument (Horiba Jobin Yvon). A sample volume of 2 mL was required for the DLS analysis. DLS is used to confirm the synthesis of spherical AuNPs by measuring their average size and polydispersity index (PDI). The PDI, a unitless value ranging from 0 to 1, quantifies the degree of size variation within the nanoparticle population. A PDI value close to 1 suggests a broad size distribution, while a value near 0 reflects uniform size and successful synthesis. These observations should be validated with larger-scale clinical and field testing.

Fabrication of immunochromatographic strips

The immunochromatographic strip consisted of four components: a sample pad, a conjugate pad, a nitrocellulose membrane, and an adsorption pad. The sample pad was soaked in PBS (10 mM pH 7.2) containing 0.5% Tween 20 and 5% sucrose for thirty minutes and then air-dried at 37 °C for another 30 min after rinsing. The nitrocellulose membrane was submerged in a solution of PBS (10 mM, pH 7.2) containing 1% Tween 20 and 1% BSA for thirty minutes. Afterward, the membrane was allowed to air dry at room temperature, and 1 µL of NS1 Antibody was put into the nitrocellulose membrane as test lines, using PBS buffer. It was applied onto the test zone with the BJQ300 and allowed to dry at ambient temperature. The conjugate pad was submerged in a solution of PBS (10 mM, pH 7.2) containing 0.5% Tween 20 for thirty minutes. Afterward, it was allowed to air dry at room temperature. A total of 7 µl of 30-nanometer nanoparticles, which were linked to aptamer molecules, were deposited onto a conjugate pad and allowed to dry at room temperature in the absence of light. The size and form of nanoparticles are critical parameters that determine the detection sensitivity. Studies have indicated that gold nanoparticles (AuNPs) ranging in size from 20 to 40 nm are commonly employed for detection in lateral flow assays (LFAs)21.

Sensitivity and specificity assay

Sensitivity assays for LFA involved measuring serially diluted immunoaffinity-purified DENV2 NS1 proteins (1.25, 20.5, 80, and 320 ng/ml) for chromatographic presentation on the LFIA. The detection process was completed within 10 min, and the results were inspected visually. Specificity was checked with the NS1 recombinant protein of DENV1, 3, and 4, ZIKV, and YFV virus (Orient Gene Company, Zhejiang, China).

Statistical analysis

Peak area intensities for the test lines were extracted from standardized regions of raw strip images using ImageJ (v1.53). All measurements were converted to floating-point values. Missing or non-visible signals within the analysis region were replaced by zero intensities to maintain dataset completeness for statistical processing. For Limit of Detection (LOD) estimation in buffer and serum matrices, binary classification (positive/negative) was generated using a dynamic threshold equal to the median peak area value of all replicates within the respective dataset. Logistic regression models were fitted separately for buffer and serum test data, using NS1 concentration as the predictor and detection probability as the response. The LOD was defined as the concentration corresponding to 50% detection probability, with 95% confidence intervals (CI) computed by non-parametric bootstrapping (1000 iterations). Reported LODs were 5.2 ng/mL (95% CI: 4.6–5.8) for buffer and 6.1 ng/mL (95% CI: 5.5–6.8) for serum (see Figs. 5B and 6B; Supplementary Table S1). For group comparisons among virus categories (DENV2, other Flavivirus NS1 proteins), Kruskal–Wallis nonparametric testing was used due to non-normal intensity distributions. Significant omnibus test results (p < 0.01) were followed by Bonferroni-adjusted pairwise comparisons to identify specific differences. Receiver Operating Characteristic (ROC) curves were generated for buffer and serum datasets, with Area Under the Curve (AUC) calculated to quantify discrimination ability: AUC = 0.97 for buffer and AUC = 0.94 for serum. Principal Component Analysis (PCA) was performed on normalized (scaled to zero mean, unit variance) peak area intensities from both matrix types combined, to visualize the separation of virus serotypes and matrix effects. The first two principal components explained 82% (PC1) and 12% (PC2) of the variance, respectively, consistent with clustering patterns shown in Supplementary Table S1. All statistical analyses and plotting were conducted in Python (v3.10) using numpy, pandas, scipy, statsmodels, and matplotlib packages, with reproducibility ensured via fixed random seeds.

Results and discussion

Development of aptamer-antibody lateral flow assay

The NS1-targeting aptamer sequence was selected based on a study conducted in 2015 by Chen et al. The aptamer sequence is modified by adding a poly C sequence to its 3’ end. This modified sequence is then attached to synthesized AuNPs and placed on a conjugated pad. A short poly G oligonucleotide sequence complementary to the 3’ aptamer deposited on the nitrocellulose membrane is used for the control zone. In addition, an antibody from OrientGene (Zhejiang, China) specifically targeting the NS1 protein was immobilized on the nitrocellulose membrane as test lines. These analyses further validated the accuracy of the protein. The refined protein was employed as the detection target for our aptamer antibody test. The presence of both test and control lines indicated positive instances. However, the reactivity of the control line alone demonstrated a negative outcome, affirming the devices’ reliability.

Expression of NS1 protein

To examine our lateral flow test strip, the NS1 protein of the dengue (DEN-2) virus was expressed and purified from bacterial cell lysates. Bacterial cells containing the NS1 protein were cultured overnight at 37 °C while induced with 0.5 mM IPTG. The recombinant protein was purified by Ni2+NTA affinity chromatography and refolded by urea. The SDS-PAGE analysis was conducted to assess the purity of the isolated protein. SDS-PAGE analysis indicated that the rNS1 protein was successfully purified using IMAC (Fig. 1). A single band was observed at the expected molecular weight of rNS1 protein in lanes 4 and 5, corresponding to the eluted fractions in which the rNS1 protein was purified to homogeneity. The purified rNS1 protein was used for LFA in the present study.

Fig. 1
figure 1

IMAC purification. SDS–PAGE analysis of IMAC-purified NS1 protein; Lane M: standard protein molecular weight markers; Lane 1: unbound protein; Lane 2, Lane 3: washes; Lane 4 and Lane 5: elutions.

Full size image

Characterization of gold nanoparticles (AuNPs)

UV-Vis spectroscopy

A reliable method for confirming the formation of spherical gold nanoparticles (AuNPs) is through the analysis of their absorption spectra in the visible range using UV-Vis spectroscopy. Spherical AuNPs exhibit a uniform response to visible light due to the even distribution of conduction band electrons across their surface. Specifically, 30 nm AuNPs show a distinct absorption peak at approximately 525 nm, corresponding to their characteristic surface plasmon resonance (SPR) (Fig. 2A). Upon binding to DNA, the nanoparticles form localized surface plasmon resonance (LSPR) bands, characterized by a reduction in intensity and a shift toward longer wavelengths. In this study, the AuNPs exhibited a strong resonance peak at 525 nm (λmax). When linked to aptamers, the nanoparticles demonstrated a minor shift of 2–3 nm (Fig. 2B), indicating successful aptamer attachment. This result is consistent with previous findings that show the interaction between oligonucleotide sequences and gold nanoparticles.

Fig. 2
figure 2

(A) UV-Vis for 30 nm gold nanoparticles synthesized with sodium citrate indicated in our tests that AuNPs exhibited a strong resonance peak at 525 nm (λmax). (B) SPR band of nanoparticles stabilized with sodium citrate in the absence and presence of aptamer. A shift of 2–3 nm suggests a relatively small change in the nanoparticle environment, which suggests that the aptamer is interacting with AuNPs. These observations should be validated with larger-scale clinical and field testing.

Full size image

Dynamic light scattering (DLS)

Dynamic light scattering (DLS) is a method employed to verify the synthesis of AuNPs. Nevertheless, this method is exclusively appropriate for examining spherical AuNPs. It is deduced. The average size of gold nanoparticles is determined to be 30 nm. Additionally, dynamic light scattering (DLS) can be utilized to measure both the average size of the gold nanoparticles (AuNPs) and their polydispersity index (PDI). The PDI, or Polydispersity Index, is a unitless value ranging from 0 to 1 that quantifies the degree of variation in nanoparticle sizes. In this study, the spherical nanoparticle PDI (polydispersity index) is 0.224, indicating a good distribution coefficient. The results of the DLS measurements are shown in Fig. 3, which illustrates the size and distribution of the nanoparticles.

Fig. 3
figure 3

The size distribution of gold nanoparticles was analyzed using Dynamic Light Scattering (DLS). (A) The green curve suggests that most of the particle volume is concentrated around a peak diameter, (B) while the red line shows the cumulative volume contribution, which increases with particle size. The cumulative distribution reaches 100% as it includes all particle sizes. The particle size distribution by intensity reveals a predominant peak around 30 nm with a good distribution coefficient (PDI: 0.224). (C) The graph shows that the highest volume contribution comes from particles with diameters around 30 nm, while smaller and larger particles contribute less. These observations should be validated with larger-scale clinical and field testing.

Full size image

Test procedure LFA

The LFA described here deviates from the conventional use of paired antibodies. Instead, it is designed as a sandwich assay, where a capture antibody is immobilized on a membrane in the test zone, and a detecting aptamer coupled to gold nanoparticles is immobilized in the control zone using a short oligonucleotide sequence that is complementary to the aptamer. The NS1 present in the sample forms a binding interaction with the aptamer. This complex is then immobilized in the test zone by attaching to the capture antibody. The marker line becomes crimson and perceptible. In the absence of the target NS1, no marker line was observed. Upon entering the control zone, the oligonucleotide complex establishes a control line, thereby confirming the functionality of the test (Fig. 4).

Fig. 4
figure 4

Schematic diagram of NS1 detection using a lateral-flow strip assay-based aptamer-antibody sandwich.

Full size image

Analytical sensitivity and specificity

Analytical sensitivity was evaluated using immunoaffinity-purified DENV NS1 proteins of DENV2 at concentrations of 1.25, 5, 20, 80, and 320 ng/mL in buffer and matched ranges in spiked human serum (Figs. 5 and 6). Raw lateral flow images were quantified by ImageJ, peak areas cleaned by replacing missing values with zero, and converted to floats for analysis. To ensure robust binary class separation in logistic regression modeling, dynamic detection thresholds equal to the median peak area for each matrix were applied. Logistic regression analysis indicated that, in buffer, the concentration corresponding to a 50% probability of detection (LOD) was 5.2 ng/mL (95% CI: 4.6–5.8 ng/mL), and in serum it was 6.1 ng/mL (95% CI: 5.5–6.8 ng/mL). These values closely match the visual LODs determined in the original trial but now include statistically rigorous CIs derived from quantitative signal intensity data. Updated Figs. 5B and 6B display fitted logistic curves with LOD values and 95% CI shading. Kruskal–Wallis tests across virus groups (DENV1–4, ZIKV, YFV) for both buffer and serum datasets suggested in our data significant differences in median peak area (p < 0.01). Post-hoc pairwise rank-based comparisons with Bonferroni adjustment indicated in our tests that DENV2 produced appeared to be significantly higher signals at lower NS1 concentrations than all other flaviviruses in both matrices (p < 0.05). In addition, one-way ANOVA with Tukey’s post hoc test was applied to datasets in Fig. 7. Results from Fig. 7A confirm that in spiked human serum (6–10 ng/mL NS1), DENV2 showed significantly higher peak areas than other conditions (p < 0.01). This statistical output reinforces the visual observations in Figs. 5A and 6A, demonstrating strong assay specificity. ROC curve analysis using quantitative peak area data yielded AUC values of 0.97 for buffer and 0.94 for serum, indicating excellent discrimination performance even under the protein-rich serum matrix that can promote nanoparticle aggregation. The improved ROC performance in buffer reflects reduced background and matrix interference. Principal component analysis (PCA) performed on normalized peak areas revealed that PC1 accounted for 82% and PC2 for 12% of total variance. PCA score plots displayed clear separation between buffer and serum datasets, as well as distinct clustering of DENV2 from other viruses, consistent with high specificity and limited cross-reactivity. The aptamer–antibody LFA achieved performance comparable to sensitive ELISA platforms, while maintaining rapid turnaround (~ 10 min) and low complexity suitable for point-of-care deployment. For example, Prommool et al.’s modified ELISA achieved 0.2–0.02 ng/mL LOD depending on serotype but required ≥ 4 h incubation and trained personnel. Thio-NAD cycling ELISA (Chen et al.) achieved 1.152 pg/mL LOD at the cost of specialized reagents and setup. Other rapid formats such as the paper-based microfluidic assay of Prabowo et al. reached 0.78 ng/mL but with higher per-test cost. In contrast, antibody-only LFAs (Lai et al.) reported 15.625–31.5 ng/mL LOD, underscoring the sensitivity gain from aptamer integration. The nucleic acid aptamers likely improve target capture by recognizing larger or additional epitopes compared to antibodies, while being thermally stable and low-cost to produce. Given the high structural similarity of NS1 proteins among flaviviruses, cross-reactivity is a critical concern. Our assays demonstrated no detectable cross-reactivity with NS1 from ZIKV, YFV, DENV1, DENV3, and DENV4 at tested concentrations, confirming matrix-robust specificity for DENV2 NS1. Supporting this observation, Fig. 7B shows that across 5, 20, and 80 ng/mL in buffer, DENV2 generated significantly higher peak area intensities than other flaviviruses (p < 0.01, p < 0.001, p < 0.0001), consistent with specificity trends in Fig. 5B. These results, derived from both visual inspection and statistical quantitation, indicate that the aptamer–antibody LFA can sensitively and specifically detect DENV2 NS1 in both buffer and serum matrices, meeting criteria for field-ready, low-resource diagnostic deployment. Comparative performance metrics are summarized in Supplementary Table S1, which includes updated LOD values with confidence intervals, statistical analysis results, and comparison to other NS1 detection platforms These observations should be validated with larger-scale clinical and field testing.

Fig. 5
figure 5

(A) Visual LOD for NS1 in buffer: DENV2 detected at 5 ng/mL, while other flaviviruses required ≥ 20 ng/mL for visible signals using the aptamer–antibody LFA strip. (B) ImageJ quantification (0–80 ng/mL) with logistic regression gave LOD = 5.2 ng/mL (95% CI: 4.6–5.8). Scatter plots show fitted curves with CI shading. Non-parametric analysis (Kruskal–Wallis with Bonferroni correction) indicated significantly higher DENV2 signals at low concentrations (p < 0.05), consistent with assay specificity observed across the tested flaviviruses.

Full size image
Fig. 6
figure 6

(A) Visual LOD for NS1 detection in 10‑fold diluted human serum to minimize AuNP aggregation; DENV2 visually detected at 5 ng/mL, with diminished or absent signals for other flaviviruses below 20 ng/mL. (B) Quantitative ImageJ analysis (5–10 ng/mL range) gave LOD = 6.1 ng/mL (95% CI: 5.5–6.8). ROC analysis indicated AUC = 0.94. PCA plots from combined buffer/serum datasets revealed clear matrix-driven clustering and virus separation, consistent with ANOVA findings in Fig. 7A showing significantly higher DENV2 signal intensities (p < 0.01).

Full size image
Fig. 7
figure 7

Statistical comparison of NS1 peak area intensities across matrices and viral targets. (A) In serum (6–10 ng/mL), DENV2 showed significantly higher signals than other flaviviruses, confirmed by one‑way ANOVA with Tukey’s (p < 0.01, p < 0.001; n = 5). (B) In buffer (5, 20, 80 ng/mL), DENV2 signals remained significantly greater across all concentrations (p < 0.01 to p < 0.0001; ns = not significant). Results align with LOD, ROC, and PCA trends, reinforcing assay specificity for DENV2 in both serum and buffer matrices.

Full size image

Conclusion

A new and simple lateral flow strip assay was successfully developed using an aptamer–antibody sandwich with gold nanoparticles for DENV2 detection. The detection limits were determined as 5.2 ng/mL (95% CI: 4.6–5.8) in buffer and 6.1 ng/mL (95% CI: 5.5–6.8) in spiked human serum, based on ImageJ quantification and logistic regression modeling. Quantitative performance analysis indicated in our tests ROC AUC = 0.97 for buffer and 0.94 for serum detection, with Kruskal–Wallis testing (p < 0.01) and Bonferroni-adjusted pairwise comparisons confirming DENV2 specificity. Principal component analysis (PC1 = 82%, PC2 = 12%) revealed clear separation between matrices and virus serotypes. These observations should be validated with larger-scale clinical and field testing.

This study presents a highly sensitive, selective, rapid, cost-effective lateral flow aptamer–antibody sandwich for detecting NS1 in DENV2. The assay maintains sub‑10 ng/mL sensitivity in complex biological matrices like serum, demonstrating robust matrix tolerance and strong real‑world applicability in point‑of‑care diagnostics. Therefore, the aptamer–antibody sandwich‑based lateral flow strip assay for DENV2 detection developed in this study has great potential for reliable on‑site DENV2 detection.

Future perspectives

Building on the current aptamer–antibody LFA platform, several avenues can be pursued to enhance clinical applicability and diagnostic coverage. First, lowering the limit of detection through optimization of aptamer sequences, nanoparticle surface chemistry, and signal amplification strategies could enable earlier detection in febrile patients and asymptomatic carriers. Second, expanding serotype coverage beyond DENV2 by incorporating multi-aptamer panels or multiplexed test lines may improve utility in regions with concurrent circulation of multiple Flavivirus serotypes. Third, integrating the platform into portable, battery-operated readers capable of quantitative image analysis could support results standardization in field conditions. Furthermore, validation with large-scale, diverse clinical sample sets—including different disease stages, co-infections, and varying matrix complexities—will be essential for regulatory approval. Finally, exploring stability under high humidity and temperature, as well as long-term shelf life studies, will help adapt the device to tropical and resource-limited settings28. These improvements, alongside ongoing surveillance and diagnostic integration strategies, could transform the presented device into a robust tool for public health interventions against dengue and related arboviruses29.