Introduction

The Monkeypox virus (MPXV) outbreaks have spread to 117 nations [1], creating a significant burden on human health and society. MPXV strains are broadly categorized into two genetic clades: the Congo Basin clade, associated with higher virulence but reduced transmissibility, and the West African clade, characterized by lower virulence yet higher prevalence [2,3,4]. Hence, distinguishing between these two clades at point-of-care is of significance for implementing clinical interventions.

Nucleic acid detection, a commonly used method for virus detection, can effectively differentiate MPXV from other orthopoxviruses by targeting specific gene fragments with distinct signature [5]. Notably, the specific complement binding protein genes, C3L (pertinent to the former Congo Basin clade) and G2R_WA (specific to the West African clade), have proven valuable for discerning these two clades [6]. Laboratory-based polymerase chain reaction (PCR) technique has shown high sensitivity and accuracy in detecting such two clades [7, 8]. Nonetheless, the necessity for multiple rounds of thermal cycling renders such method intricate, cost-intensive, and time-consuming. As a remedy, many isothermal amplification methods (e.g., recombinase polymerase amplification [9]) have been developed for MPXV detection. By eliminating the thermal cycling requirement, these methods significantly diminish the detection expenses and operational complexity [10]. However, due to the affection of reaction time and inhibition effect to amplification degree, these strategies suffer from the limitations in accuracy and quantification [11].

Lateral flow assay (LFA) has been widely developed as a rapid, cost-effective, and user-friendly detection platform for recognizing viruses, including COVID-19, influenza, Ebola, Zika, and dengue viruses [12,13,14,15]. However, the conventional color reporters used in LFAs, such as AuNPs [16] and AgNPs [17], often yield weak signals, leading to insufficient sensitivity in nucleic acid testing for distinguishing viruses. The advancement in novel labeling nanomaterials, particularly fluorescent nanoparticles like fluorescent dyes [18], fluorescent microspheres [19], quantum dots [20], and upconversion nanoparticles (UCNPs) [14], has marked a breakthrough in this field [21]. These materials offer stronger and more quantifiable detection signals, greatly improving the detection sensitivity and accuracy of LFAs. Notably, the utilization of UCNPs with their unique anti-Stokes properties offers a solution to effectively avoid background fluorescence, enabling them an excellent nanomaterial for labeling and detecting [22]. Numerous LFAs developed using UCNPs exhibit prolonged fluorescence stability and an elevated signal-to-noise ratio [23], which have been effectively employed for detection of various targets with robust detection capability, such as bacteria [23], biotoxin [24], and metal ions [25]. Furthermore, when coupled with optimization strategies, such as regulation of fluorescence intensity and surface probe density [26], the detection performance of UCNPs-based LFA can be further improved to expand their utilization.

This study uses UCNPs with high fluorescent intensity to create a high-performance LFA with double T-lines and a shared C-line, which is capable of simultaneously detecting two clades of MPXV. The resulting platform enables both convenient at-home virus screening using a smartphone and highly sensitive laboratory detection utilizing a benchtop fluorescence analyzer. This dual-mode detection offers the benefits of cost-effectiveness, user-friendliness, and broad applicability.

Materials and methods

Materials

Y(CF3COO)3, Yb(CF3COO)3, and Er(CF3COO)3 were prepared by Suzhou University. Sodium trifluoroacetate, 1-Octadecene (ODE), oleic acid (OA), and N-Hydroxysulfosuccinimide sodium salt (Sulfo-NHS) were purchased from Aladdin. N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride crystalline (EDC), Trizma base (Tris), and Proclin 300 were obtained from Sigma Aldrich. Ethanol absolute, cyclohexane, chloroform, dimethyl sulfoxide (DMSO), and glycine were obtained from Sinopharm Group Chemical Reagent Co., Ltd. HEPES, MES, and BSA were obtained from Solarbio. DSPE-PEG-COOH and Streptavidin were obtained from Xi’an Ruixi Biotechnology Company and Shanghai Maclin Biochemical Technology Co., Ltd, respectively. Nucleic acid sequences of MPXV and their probe sequences were synthesized by Sangong Bioengineering (Shanghai) Co., Ltd (Table 1). All reagents are analytical grade. The components of LFA, including sample pad (SB08), absorbent pad (H-1), immersing pad (8965), nitrocellulose (NC) membrane (Sartorius CN 95), and PVC backing pad, were bought from Shanghai Jening Biotechnology Co., Ltd.

Table 1 The sequences of two MPXV clades and their corresponding detection probe
Full size table

Preparation and modification of NaYF4: Yb3+, Er3+

The UCNPs (NaYF4: Yb3+, Er3+) were synthesized using thermal decomposition method, according to our last paper [26]. Details were shown in the Supplementary materials.

Nucleic acid probe modification on UCNP surface

The two kinds of recognition probes were designed according to the specific nucleic acid fragments of two clades of MPXV (C3L and G2R_WA) reported in the literature, and the sequences were shown in Table 1. The amino modified G2R_WA and C3L recognition probes were conjugated to water-soluble UCNPs by a condensation reaction to form UCNPs-probes. Specifically, the as-prepared PEG@UCNPs (40 μL) were mixed with MES buffer (20 mM, pH = 6.0) in a ratio of 1:2 (v/v). EDC and Sulfo-NHS (5 μL, 24 mg/mL) were added for carboxyl group activation. The mixture was placed in a constant temperature shaker (37 °C, 45 min, 220 rpm). After resuspension in 0.6 mL MES buffer, recognition probes (120 µL, 100 nM) modified with amino group were added to continue the reaction in the shaker for 2 h. Furthermore, 18 µL sealing buffer (15% Glycine, 1% Proclin 300, pH = 11) was added to complete site sealing. The products (UCNPs-probes) were washed and resuspended in 0.6 mL probe-diluent (0.5% v/v Tween 20, 4% (w/v) BSA and SSC (2 ×), pH = 7.0).

Construction of UCNPs-based LFA

The UCNPs-based LFA was prepared according to the literature [27]. For single-target detection, UCNPs-probes for C3L clade or G2R_WA clade were evenly sprayed (5 µL/cm) on the immersing pad. To form T-lines and C-line, biotin-modified capture probes and control probes were fixed on NC membrane using streptavidin as intermediate. The corresponding capture probes and control probes were sprayed on the NC membrane with a dosage of 1 µL/cm. Each component was dried for 2 h at 37 °C. Then, the NC membrane, sample pad, immersing pad, and absorbent pad were pasted onto a PVC backing pad with an overlap of 2 mm. The pasted pads were cut into strips with a 3.9 mm width by programmable shear. For dual-target detection, the approach is to spray G2R_WA and C3L UCNPs-probes mixed on immersing pad, build two T-lines with two clades of capture probes, and mix the control probes of two clades at 0.5 times the original concentration to build a mixed C-line.

Detection of single MPXV

To evaluate the detection capabilities of the UCNPs-based LFA, we prepared standard samples of MPXV (G2R-WA and C3L) with gradient concentrations (0, 0.001, 0.01, 0.1, 0.5, 5, 25, 50, 100 nM). The procedure involved mixing 50 μL sample with 50 μL optimized sample-diluent, which was then applied to the designated sample area. After an 8-min reaction, the fluorescence signals on both T-line and C-line were read and analyzed. For reading results via a smartphone camera under darkroom conditions, specific parameters including a shutter speed of 1/200 and ISO 2600 were applied. Similarly, the benchtop fluorescence analyzer we developed could quantitatively assess the fluorescence intensity of each dot within the detection area and translate them into a peak graph. Each target concentration was repeated 3 times, while the blank sample underwent ten repetitions. Employing the mean value and standard deviation, a standard curve was fitted, and the detection limit was calculated according to the formula (S/N = 3).

Simultaneous detection of two clades of MPXV

A dual-target LFA was developed by integrating two single-target UCNPs-based LFAs. This involved mixing the G2R_WA and C3L UCNPs-probes and applying them to the binding pad. Two separate T-lines were created using their respective capture probes, and a new C-line was formed by mixing the two control probes at half concentration. The performance of this dual-target LFA was evaluated using standard samples of G2R_WA and C3L targets. The detection process remained consistent with the previously described method.

Detection of spiked sample

Spiked serum samples of G2R-WA and C3L targets at different concentrations (0, 0.1, 0.5, 2.5, 5, 10, 25 nM) were introduced to test the anti-interference capability of the developed UCNPs-based LFA. The 50 μL as-prepared spiked serum samples were mixed with sample-diluent in a ratio of 1:1 and then detected 3 times using the developed LFA.

Furthermore, plasmids representing MXPV were designed by Basic Local Alignment Search Tool (BLAST). The forward and reverse primers and plasmids were synthesized by Sangong Bioengineering (Shanghai) Co., Ltd. The plasmids with different concentrations (103, 104, 105 copies/μL) were amplified using qPCR (Jena, qTOWER3). The reaction conditions were as follows: pre-denaturation at 95 °C for 5 min, followed by 45 cycles of 20 s denaturation at 95 °C and 30 s annealing at 60 °C. An asymmetric amplification model was implemented with a 10 × Primer Mix and an 8:1 ratio of forward primers to reverse primers to generate more free single strands. The resulting amplified products, mixed with 80 μL sample dilution, were detected using the developed LFA 3 times.

Results and discussion

Development of high-performance POCT platform for MPXV

The detection principle of MPXV using a UCNPs-based LFA was shown in Fig. 1. Utilizing UCNPs with a larger particle size can effectively enhance the fluorescence signal intensity in detection due to their increased fluorescence intensity and greater surface modification capability. As a result, UCNPs with a diameter size of 500 nm were adopted to construct high-performance LFA. The developed UCNPs-based LFA operates on a sandwich strategy. As depicted in Fig. 1A, the presence of the target triggers the corresponding T-line to emit fluorescence signals, which can be read and analyzed through a dual-mode approach. The fluorescence images, achieved using a smartphone camera in conjunction with a 980-nm laser, exhibit clear bands that can be easily discerned with the naked eyes, enabling straightforward and portable at-home self-testing. In addition, the fluorescence signal on the LFA can be scanned point-by-point using a custom benchtop fluorescence analyzer to form a fluorescence peak graph, facilitating precise and sensitive quantitative analysis. Each peak fluorescence intensities correspond to detection signals of each T-line or C-line (Fig. 1B).

Fig. 1
figure 1

The principle of the UCNPs-based LFA for dual MPXV detection. A The structure and detection principle of UCNPs-based LFA. B The dual-mode reading using a smartphone and a fluorescence analyzer

Full size image

Characterization of UCNPs

The sensitivity and stability of UCNPs-based LFA are closely related to the fluorescence intensity, size uniformity, and conjugation capability of UCNPs. NaYF4 was selected as the matrix material due to the low phonon energy and high chemical stability, and Yb3+ and Er3+ were chosen as activators and sensitizers with molar concentration ratios of 18% and 2%. To obtain large-size UCNPs, the heating time at 320 °C was extended to 110 min during the synthesis process. This adjustment reduced the surface quenching and greatly improved the fluorescence intensity. In particular, UCNPs with a large size (500 nm) exhibited superior detection performance owing to their larger surface area and stronger fluorescence intensity, as indicated in our prior publication [26]. To validate the characteristics of the synthesized UCNPs, we conducted analyses on their composition, morphology, size, and fluorescence intensity. XRD patterns (Fig. 2A) confirm that the diffraction peaks of the synthesized UCNPs match those of the standard card (β-NaYF4, JCPDS no. 16–0334). Morphological insights from TEM (Fig. 2B) illustrate that the prepared UCNPs exhibit uniform hexagonal prism shapes with an average size of 500 nm. The UCNPs emit a vibrant green fluorescence under 980 nm excitation light (inset of Fig. 2C). The fluorescence spectra consistent with the standard emission spectra of NaYF4: Yb, Er, which reveal emission peaks at 540 nm and 650 nm (Fig. 2C). Due to the stronger intensity, the 540 nm emission peak serves as the primary analysis signal in this developed platform. In addition, a recognition probe concentration of 0.36 nM was chosen for UCNP conjugation.

Fig. 2
figure 2

Characterization of UCNPs. A XRD patterns, B TEM photo, and C fluorescence spectra of UCNPs

Full size image

Detection validation of the UCNPs-based LFA

The real product of constructed LFA is shown in Fig. 3A. After being dropped on sample area, a 100 μL of sample flows through the detection area together with UCNPs-probes on the immersing pad. This process leaves behind discernible signals on T-line and C-line after undergoing corresponding reactions. Under the 980-nm excitation light, the detection area displays prominent green fluorescent bands that are markedly distinguished from the background, proving a high signal-to-noise ratio. Evidently, with an increase in the target concentration, the T-line signal rises while the C-line signal decreases. This occurs because more UCNPs are captured at the T-line, and fewer flow through the C-line. In addition, the fluorescence signal of the detection can be read by a custom benchtop fluorescence analyzer (Fig. 3B). The detail structure of the benchtop fluorescence analyzer was shown in our previous work [26]. The relationship curve of fluorescence intensity and scanning position was constructed to yield quantitative analytical outcomes (Fig. 3C), with an analysis procedure outlined in the Supplementary materials. Finally, the quantized T/C value (integral area ratio of T-line and C-line peak) serves as the detection signal to shape the concentration-signal graph (Fig. 3D), which shows a good linear upward trend and differentiation degree in the detection area. It is noteworthy that since the C-line signal also varies with the target concentration, utilizing the T/C value as the analysis result can effectively enhance the stability and signal differentiation of the system.

Fig. 3
figure 3

UCNPs-based LFA platform for MPXV detection. A Real product of LFA and its fluorescence photos. B Real product of benchtop fluorescence analyzer. C Scanned by benchtop fluorescence analyzer, the relationship graph of fluorescence intensity, and scanning position with different target concentrations. D The relationship graph of quantized T/C value and target concentration

Full size image

Development of a high-performance LFA platform

To enable the developed LFA platform to achieve better detection performance, variety parameters were optimized. Firstly, a precisely tailored sample dilution effectively mitigates false positives, prevents non-specific binding on NC membrane, and significantly improves the signal-to-noise ratio for both MPXV clades. The optimized sample dilution recipe consists of HEPES (2.383%, w/v, pH = 8), NaCl (1.58%, w/v), BSA (5%, w/v), Tween 20 (5%, v/v), EDTA (0.0465%, w/v), and Glycine (15%, w/v). Furthermore, the concentration of probes modified on the T-line and C-line also profoundly impacts the strength of the fluorescence signal, thereby influencing the accuracy of result interpretation. Our experimental findings determined the optimal T-line and C-line dosages is 100 μM and 75 μM, respectively. A reduced probe concentration on C-line can result in a more pronounced change in the C-line signal with respect to the target. The concentration of UCNPs-probes exerts an impact on their binding capability to the target, which consequently influences the signal-to-noise ratio and the sensitivity of LFA. To further optimize the performance of the developed platform, the concentration of UCNPs-probes was regulated (Fig. S1). It is evident that G2R_WA and C3L both exhibit an optimum detection range and LOD at 2 × UCNPs-probes concentration (1 × concentration: 3.3 mg/mL).

Separate detection of two clades of MPXV

To verify the detection performance of UCNPs-based LFA, detections were performed using G2R_WA and C3L standard samples with gradient concentrations. Each target concentration was tested 3 times, while the blank sample underwent ten repetitions. As the fluorescence signals stabilized after an 8-min reaction, the detection signals at 8 min were chosen for analysis (Fig. S2). Taking G2R_WA detection as an example, as depicted in the fluorescence photos (Fig. 4A), within the target concentration range of 0.01–50 nM, the T-line fluorescence signals initially increased and then attenuated with rising target concentration. In contrast, the C-line fluorescence signal gradually weakened, and notable background fluorescence was absent. This smartphone-based reading approach is straightforward to execute and equipment-free, enabling qualitative detection within a certain range.

Fig. 4
figure 4

Detection of two clades of MPXV separately. A Fluorescence photos and B peak patterns of G2R_WA with gradient concentration. The standard curve of C G2R_WA clade and D C3L clade. The fitted standard curves are y = 0.53x + 0.075 (G2R_WA) and y = 0.28x + 0.127 (C3L), respectively

Full size image

Moreover, the benchtop fluorescence analyzer permits quantification of the fluorescence signals emanating from both T-line and C-line, forming distinctive peak patterns (Fig. 4B). Through mathematical fitting, these quantified fluorescence signal values (T/C value) can be correlated with the target concentration to achieve accurate quantitative detection. Especially, the fluorescence signal, which may not be discernible in fluorescence photos, become readily measurable on the peak pattern, substantiating the fluorescence reader’s capability to capture lower-level detection signals. The standard curve for both clades (Fig. 4C–D) exhibits a strong linear correlation (R2 = 0.999). Repeated experiments revealed an average coefficient of variation (CV) of less than 15% for both clades, meeting the stability criteria. The LOD for both qualitative or quantitative analysis was calculated using the formula S/N = 3 (defined as blank + 3 std dev of the blank), yielding an LOD of 0.6 pM for G2R_WA and an LOD of 1 pM for C3L, respectively. For qualitative analysis, signal is taken from the average grey-value in green channel of the T-line region in the fluorescence photograph. For quantitative analysis, the quantified fluorescence intensity can be read directly by the fluorescence analyzer. In detail, the obtained signal is greater than 3 times the standard deviation of blank can be recognized as coming from the target rather than the systematic error. This represents a substantial enhancement in sensitivity compared to relevant POCT nucleic acid detection methods operating at the nanomolar (nM) level [28,29,30]. To assess the stability of the developed UCNPs-based LFA, LFAs from the same batch were stored at room temperature (25 °C). Furthermore, the stability of the developed UCNPs-based LFA has been demonstrated over a 20-day period under room temperature storage conditions (Fig. S3).

Integration for both two clades of MPXV detection

The above results serve to validate the constructed UCNPs-based LFA platform’s notable sensitivity and stability in independently detecting the two MPXV clades. To further substantiate the dual-target detection capability of developed LFA, the detection of both G2R_WA and C3L was integrated into a single LFA setup. The results from the dual-target detection are shown in Fig. 5A. No peaks are observable in T-line 1 and T-line 2 in the control group, indicating the absence of false positives in the dual-target detection. Upon the addition of a target, the corresponding T-line manifests a peak value that varies with concentration, while the other T-line remains devoid of any signal, affirming the absence of cross-reaction within the dual-target detection system.

Fig. 5
figure 5

Integration for both two clades of MPXV detection. A Fluorescence photos and peak patterns of dual-target detection. B Standard curves of dual-target detection

Full size image

Differing from single-target detection, the C-line signal in the dual-target scenario is influenced by both clade-specific fluorescence probes. Even at higher target concentrations, the C-line retains elevated signals, resulting in a small T/C value. Therefore, new standard curves were established for the dual-target detection through four-parameter fitting (Fig. 5B). According to the new standard curves, the LOD for dual-target detection is 0.9 pM (G2R_WA) and 1.2 pM (C3L), with upper limits reaching 25 nM for both G2R_WA and C3L. Remarkably, these values are not significantly different from those achieved in single-target detection. These findings collectively demonstrate that the dual-target LFA can accurately and effectively discern the presence of MPXV infection and identify the specific clades, offering a heightened detection performance when compared to existing POCT methods. Through comparison with other LFA-based methods for MPXV nucleic acid detection (Table S1), although our method has the advantage of a rapid readout, the LOD is not comparable to other LFAs with nucleic acid amplification pretreatment. To our knowledge, this is the first POCT method for simultaneous detection of G2R and C3L on a single LFA. Additionally, compared to other optical LFAs developed for nucleic acid detection without amplification (Table S2), our method has considerable competitiveness in terms of LOD and detection time.

Practical application validation

To evaluate the anti-interference ability of the developed UCNPs-based LFA in real application, we introduced a series of gradient concentrations of G2R-WA and C3L into human serum as spiked clinical samples. By establishing the correlation between fluorescence intensity and target concentration, the detection results were compared with the actual results (Fig. 6). The findings demonstrated that all detection outcomes fell within the range defined by the two dashed lines (± 15%), which is indicated to meet the common requirement of POCT. This suggests that the established UCNPs-based LFA platform is suitable for application in serum samples detection.

Fig. 6
figure 6

Practical application validation of UCNPs-based LFA detection in spiked serum sample. The area between the two dashed lines indicates the range where the deviation is less than 15%

Full size image

To further validate the practical detection capabilities of the developed LFA, we designed plasmids to replace MPXV as clinical samples using the Basic Local Alignment Search Tool (BLAST). Plasmid samples with different concentrations (103, 104, 105 copies/μL) were amplified using qPCR for 45 cycles. An asymmetric amplification model was performed with 10 × primer Mix and an 8:1 ratio of forward primers to reverse primers to produce more free single strands. The obtained results (Table 2) show that the detection outcomes of the UCNPs-based LFA can still align with the individual actual concentrations, with a deviation of less than ± 15%. This capability enables the detection of low concentrations, down to 103 copies/μL. Collectively, these findings substantiate the effective applicability of the developed UCNPs-based LFA platform for practical applications.

Table 2 Clinical validation of UCNPs-based LFA for detection in MPXV plasmid samples
Full size table

In summary, the developed platform demonstrates greater suitability for point-of-care testing scenarios by effectively eliminating biological sample interference, reducing detection time, and enhancing detection sensitivity through cost-effective methods. However, it remains crucial to integrate amplification technology to fully meet the demands of detecting low-abundance clinical samples. This highlights a prospective research direction for us: incorporating targeted enrichment methods or cyclic reaction systems to genuinely achieve clinical applications without amplification.

Conclusion

Utilizing UCNPs with enhanced fluorescence intensity and greater probe modification capacity, this study has developed a high-performance LFA with double T-lines and a shared C-line, offering both qualitative and quantitative dual-mode detection options through smartphone-based reading and a benchtop fluorescence analyzer. The resulting UCNPs-based LFA demonstrates remarkable sensitivity, rapidity, and stability, which can achieve an LOD of 0.9 pM (G2R_WA) and 1.2 pM (C3L) of two MPXV clades within 8 min. In addition, the spiked sample detection proved that the developed detection platform has excellent anti-interference and can realize the MPXV detection. This validates the proposed UCNPs-based LFA’s potential for rapid identification of the presence and specific clade of MPXV infection. Although the strategy showcases considerable potential in the field of nucleic acid detection, it is still necessary to combine amplification techniques to meet the detection needs of real clinical samples. This points out the future research direction in this area. Enrichment or cyclic reactions may be used to further optimize sensitivity, and really achieve non-amplification detection, and eliminate the impact of primer design, amplification procedures, and equipment.