Introduction

In recent years, the CRISPR/Cas (clustered regularly interspaced short palindromic repeats/CRISPR-associated) system has emerged as a powerful tool for nucleic acid detection due to its high specificity and programmability1,2. Among the various CRISPR/Cas systems, Cas13a has garnered significant attention for its RNA-targeting capabilities3,4,5. Cas13a, an RNA-guided ribonuclease, forms a ribonucleoprotein complex (RNP) with CRISPR RNA (crRNA)6,7. Upon binding to complementary target RNA, Cas13a activates its HEPN (higher eukaryotes and prokaryotes nucleotide-binding domain) motifs, leading to indiscriminate cleavage of surrounding single-stranded RNAs (ssRNAs)8. This collateral cleavage activity, coupled with a fluorescent-labeled RNA reporter, enables sensitive and specific molecular diagnosis6,9.

Cas13a has been reported to detect RNA with attomolar sensitivity when used in conjunction with nucleic acid preamplification strategies, including reverse-transcription recombinase polymerase amplification (RT-RPA) and reverse-transcription loop-mediated isothermal amplification (RT-LAMP)10,11. However, incorporating preamplification steps into CRISPR-based detection requires multiple manual interventions and can increase detection costs12. To address these challenges, researchers are exploring amplification-free CRISPR-based molecular diagnostic technologies. These approaches include digital microfluidic technology and the application of multiple guide RNAs13,14,15,16. Another strategy involves using Cas tandem systems (e.g., Cas13–Cas12, Cas13–Cas14, and Cas13–Csm6), in which one Cas protein is responsible for recognizing the target, while a second Cas protein amplifies the signal through a cascade mechanism17,18,19. Compared to single Cas13a reaction systems, which rely on a single target to activate a single Cas13a RNP, Cas tandem systems can achieve higher sensitivity. However, using multiple Cas enzymes introduces challenges related to optimization, scalability, and buffer compatibility, as well as increased costs and processing time.

Recent functional studies of Cas13a have revealed critical features governing its target recognition and specificity. The crRNA spacer, which guides target RNA binding, plays a pivotal role in initiating the allosteric response20. Within the spacer, the seed region (nucleotides 9–14) requires perfect base pairing for stable target binding, while the adjacent switch region (nucleotides 5–8) induces activation of the catalytic cleft upon target RNA binding21. Moreover, extending complementarity between the 3’-flank of the crRNA spacer (referred to as the tag) and the target RNA (anti-tag) can inhibit type VI-A Cas13’s nuclease activity8,22. Studies have shown that an 8-nt complementary anti-tag sequence inhibits both cis- and trans-RNase activities of Cas13a8,22. This extended pairing enhances specificity by distinguishing self from non-self targets, a feature critical for Cas13-based diagnostics20. Insights into seed, switch, and tag: anti-tag pairing mechanisms provide a deeper understanding of Cas13a’s regulatory architecture, offering valuable perspectives for precise RNA targeting and control.

Here, we investigate the allosteric effects of target RNA secondary structure and the influence of anti-tag sequences on RNA detection using Leptotrichia wadei (Lwa)Cas13a. We discover that the presence of anti-tag sequences and target RNA secondary structures inhibits the trans-cleavage activity of Cas13a. Building on this finding, we design a specific CRISPR anti-tag hairpin containing secondary structure and anti-tag sequences. Using this design, we develop a simple and sensitive CRISPR anti-tag mediated room temperature RNA detection (CARRD) assay with a single LwaCas13a/crRNA complex. In the presence of target RNA, the CRISPR anti-tag hairpin mediates CRISPR cascade signal amplification, enabling highly sensitive, one-pot, amplification-free RNA detection at room temperature. Using this assay, we successfully detect target viral RNAs, including HIV and HCV. Furthermore, we clinically validate the CARRD assay by detecting HIV viral RNA extracted from clinical plasma samples, demonstrating its potential for simple and accurate viral RNA detection in resource-limited settings.

Results

Working principle of the CARRD assay

In a conventional CRISPR/Cas13a activation reaction, the Cas13a endonuclease and crRNA form a functional RNP complex that binds RNA targets with sequences complementary to the crRNA spacer region. Upon binding to a target RNA, the Cas13a RNP undergoes a conformational change that activates its collateral cleavage activity23. The activated Cas13a RNP indiscriminately cleaves ssRNA fluorescent reporters, amplifying the detection signal and enabling RNA detection (Fig. 1a, i). However, in the presence of target RNA containing an anti-tag sequence, this conformational change is inhibited, preventing the activation of the trans-cleavage reaction of CRISPR/Cas13a (Fig. 1a, ii).

Fig. 1: Schematic of CARRD: a CRISPR anti-tag mediated room temperature RNA detection assay.

a Schematic of (i) conventional Cas13a/crRNA activation on target RNA (red) without an anti-tag sequence and (ii) Cas13a/crRNA inactivation on target RNA (red) with an anti-tag (green). b Design of the CRISPR anti-tag hairpin. This hairpin-structured mediator contains an anti-tag sequence (5’ - rGrUrUrUrUrArGrU − 3’) in the loop region (pink) and a chimeric double-stranded DNA (blue) and RNA (red) with asymmetric lengths in the stem region. The RNA sequence is complementary to the crRNA in the Cas13a/crRNA RNP complex. c Schematic illustration of the one-pot, one-step CARRD assay at room temperature. The reaction comprises three main components: the Cas13a/crRNA complex, the CRISPR anti-tag hairpin, and the FQ reporter. In the presence of target RNA, Cas13a/crRNA is activated, initiating a trans-cleavage reaction and simultaneously cleaving the CRISPR anti-tag hairpin’s loop region. Subsequently, more Cas13a/crRNA complexes are activated, further amplifying the fluorescence signal and enabling highly sensitive and specific RNA detection. The reactions can be carried out at room temperature. Created in BioRender. Moon, J. (2025) https://BioRender.com/r57b155.

Inspired by the inhibitory role of the anti-tag sequence in Cas13a activation, we designed a specific hairpin-structured target RNA, termed the “CRISPR anti-tag hairpin,” and developed a one-pot, one-enzyme, amplification-free CARRD assay for RNA detection at room temperature. The CRISPR anti-tag hairpin consists of an anti-tag sequence and a double-stranded chimeric region of DNA and target RNA with asymmetric lengths, where the target RNA region is fully complementary to the crRNA (Fig. 1b). Due to its hairpin structure and anti-tag sequence, the CRISPR anti-tag hairpin cannot initially activate the trans-cleavage reaction of Cas13a in the absence of target RNA, even if it contains the target RNA sequence. In contrast, in the presence of target RNA, Cas13a is activated, initiating a trans-cleavage reaction that results in cleavage of the anti-tag sequence (loop region) of the CRISPR anti-tag hairpin. This cleavage exposes the target RNA sequence within the hairpin, allowing it to be recognized by the Cas13a/crRNA complex and further activating Cas13a’s trans-cleavage activity. This results in a cascade signal amplification, enabling sensitive RNA detection (Fig. 1c). Thus, by leveraging the CRISPR anti-tag hairpin, we established a one-pot, amplification-free, sensitive, and specific CRISPR cascade signal amplification method using a single Cas13a enzyme.

Development and optimization of the CARRD assay

Although CRISPR/Cas13a reactions are typically performed at 37 °C in most Cas13a-based diagnostic studies, we conducted in vitro trans-cleavage assays at various reaction temperatures (25, 30, 33, and 37 °C). Surprisingly, we found that the CRISPR/Cas13a reaction targeting ssRNA (Fig. 2a, target RNA) performed more effectively at 25 °C than at 37°C (Fig. 2b, c). This result aligns with a recent study showing that the CRISPR/Cas13a system operates efficiently across a broad temperature range (7–37 °C) while maintaining rapid trans-cleavage kinetics once activated24. This discrepancy from earlier reports suggesting 37 °C as the optimal temperature may result from differences in RNA secondary structure, which influence the efficiency of ternary complex formation. Efficient formation of the Cas13a ternary complex at lower temperatures has been identified as a key factor enabling robust cleavage activity even at room temperature24. This finding significantly simplifies the equipment requirements (e.g., eliminating the need for a heater) and is ideal for simple, minimally instrumented point-of-care diagnostic applications. Thus, we conducted subsequent Cas13a experiments at 25 °C.

Fig. 2: Cas13a activity across temperatures and anti-tag substrate designs.

a The panel of nucleic acid substrates investigated with the LwaCas13a system in this study. Substrate elements are colored as follows: RNA, red; DNA, blue; anti-tag sequence; pink, rU; green. The 8-nucleotide ‘anti-tag’ contains 5’ – rGrUrUrUrUrArGrU-3’. b The trans-cleavage reaction of the LwaCas13a system operating at different temperature conditions (25, 30, 33, and 37 °C) in real-time and c fluorescence change fold (F/FNTC) measured at 60 min. Raw fluorescence units, RFU; No target control, NTC. [Target ssRNA] = 100 pM. (n = 3 and data represent mean ± s.d. of three technical replicates). d LwaCas13a reaction with nucleic acid substrate elements of Fig. 2a at 60 min. [Nucleic acid substrate elements] = 1 nM (n = 3 and data represent mean ± s.d. of three technical replicates). Statistical analysis for n = 3 samples comparing 5’-RNA tag HP and 5’-RNA rU8 HP was performed using a two-tailed t-test (p = 1.655 × 10−6), indicating a statistically significant difference. Source data are provided as a Source Data file. Created in BioRender. Moon, J. (2025) https://BioRender.com/r57b155.

Recent in vivo studies have demonstrated that extended complementarity between the target RNA and the 3’-flank of the crRNA (tag:anti-tag pairing) can significantly reduce RNA cleavage by the type VI-A Cas13a system8,22,25. Here, we investigated the cleavage activity of LwaCas13a, focusing on various RNA structures and the effects of anti-tag sequences (Fig. 2a). First, we compared the trans-cleavage activity of LwaCas13a on target RNA with and without an anti-tag sequence (5’–rGrUrUrUrUrArGrU–3’). The results showed reduced trans-cleavage activity on the target RNA containing the anti-tag (anti-tag target in Fig. 2a) compared to the target RNA without the anti-tag sequence, consistent with previous findings using LseCas13a, LbuCas13a, and LshCas13a8,22 (Fig. 2d). Moreover, when the target RNA was paired with a complementary DNA sequence to form a double-stranded structure (dsTarget in Fig. 2a), the trans-cleavage activity further decreased compared to the single-stranded RNA target (Target RNA in Fig. 2a, d). This signal reduction effect was even more pronounced when the RNA containing the anti-tag sequence formed a double-stranded structure (dsAnti-tag Target in Fig. 2a, d).

Next, we built a hairpin structure by incorporating the anti-tag sequence into the loop region to compare its trans-cleavage activity. In this configuration, the anti-tag sequence was positioned differently at the 3’ and 5’ ends of the target RNA. We designed two constructs: 5’-DNA tag HP (5’-DNA-8nt anti-tag segment-target RNA) and 5’-RNA tag HP (5’-target RNA-8nt anti-tag segment-DNA) (Fig. 2a). The 5’-DNA tag HP showed a similar result to the double-stranded target RNA:DNA (dsTarget), indicating that the loop structure does not significantly affect its performance (Fig. 2d). In contrast, the 5’-RNA tag HP showed a markedly reduced signal, even lower than that of the double-stranded anti-tag target RNA:DNA (dsAnti-tag Target). This result suggests that the loop structure at the 3’ end of the RNA has a pronounced inhibitory effect on Cas13a’s trans-cleavage activity (Fig. 2d). These findings are consistent with a previous study reporting that the 3’ region of the protospacer exerts an allosteric inhibitory effect[26](https://www.nature.com/articles/s41467-025-64205-4#ref-CR26 “Kimchi, O., Larsen, B. B., Dunkley, O. R. S., Te Velthuis, A. J. W. & Myhrvold, C. RNA structure modulates Cas13 activity and enables mismatch detection. Preprint at bioRxiv https://doi.org/10.1101/2023.10.05.560533

(2023).“).

To further evaluate the anti-tag sequence’s effect, we designed a hairpin structure where the loop region contained an 8nt rU sequence (used as a cleavage substrate for Cas13a) instead of the anti-tag sequence (named 5’-RNA rU8 HP, Fig. 2a). The results showed that trans-cleavage activity noticeably decreased in the hairpin structure containing the anti-tag sequence compared to the hairpin structure containing the 8nt rU sequence (Fig. 2d). These findings further confirm that the trans-cleavage activity of Cas13a can be effectively inhibited by the cooperative effects of the hairpin structure and the anti-tag sequence. Based on these results, we obtained a comprehensive understanding of the factors affecting Cas13a activation, such as reaction temperature, target RNA secondary structure, and anti-tag sequences. Leveraging these insights, we used a specific CRISPR anti-tag hairpin as the mediator of the cascade signal amplification reaction and developed a simple, sensitive, amplification-free CARRD assay that operates at room temperature.

Cascade signal amplification mechanism induced by the CRISPR anti-tag hairpin

We hypothesized that the RNA sequence within the hairpin containing the anti-tag would not activate the Cas13a reaction (e.g., 5’-RNA tag HP, Fig. 2d). However, after the anti-tag is cleaved, the RNA sequence would activate Cas13a (e.g., dsTarget or dsAnti-tag Target in Fig. 2d), leading to further signal amplification and serving as a cascade signal amplifier. Since the anti-tag sequence contains rU residues, we anticipated these residues would be susceptible to cleavage by the trans-cleavage activity of Cas13a activated by the target RNA27.

To test this hypothesis, we designed a sequential reaction to verify whether the anti-tag cleavage occurs and whether the trans-cleavage reaction of Cas13a can subsequently be activated by the RNA sequence within the CRISPR anti-tag hairpin. In step 1, we conducted the Cas13a/M crRNA reaction in the presence and absence of target M. Instead of using a fluorescent-quencher (FQ) reporter for fluorescence measurement, we added CRISPR anti-tag hairpins containing the anti-tag and HIV RNA:DNA sequences and incubated the reaction at room temperature (25 °C) for 20 min. In step 2, we added the reactant from step 1 as a target to a solution containing Cas13a/HIV crRNA and FQ reporters, and subsequently measured the fluorescence signal (Fig. 3a).

Fig. 3: Two-step CRISPR/Cas13a activation using anti-tag hairpins with variable DNA blockers.

a Schematic illustration of the sequential Cas13a reaction using a CRISPR anti-tag hairpin (HP). Cas13a/M crRNA and the CRISPR anti-tag HP were incubated with or without target M at 25 °C. The step 1 reaction product was used as a target for the step 2 Cas13a/HIV crRNA reaction and diluted tenfold. b CRISPR anti-tag HP panel used in Fig. 3a. Each HP (14D, 16D, 18D, and 20D HP) contains an 8 nt anti-tag sequence in the loop region and a 24-nt HIV RNA sequence in the stem region. Each HP has a DNA blocker of varying lengths (14, 16, 18, and 20 bp) complementary to the HIV RNA. Nucleotide, nt. c Fluorescence fold normalized to no target control (NTC) at 1 hr of trans-cleavage assay with LwaCas13a and different HPs. [Cas13a/M crRNA and Cas13a/HIV crRNA] = 10 nM, [Target M] = 10 pM, [CRISPR anti-tag HP] = 10 nM, [FQ reporter] = 500 nM. (n = 3 and data represent mean ± s.d. of three technical replicates). d Step 1: Analysis of the trans-cleavage effect of Cas13a/M crRNA on 14D HP by native PAGE gel. Step 2: Analysis of the Cas13a/HIV crRNA reaction on the reactant from step 1 (reactant 1–4) using the FQ reporter. [Cas13a/M crRNA and Cas13a/HIV crRNA] = 10 nM, [Target M] = 100 pM, [CRISPR anti-tag HP] = 100 nM, [FQ reporter] = 500 nM. The experiment was independently repeated twice with similar results. Source data are provided as a Source Data file. Created in BioRender. Moon, J. (2025) https://BioRender.com/r57b155.

First, we compared the trans-cleavage reactions for three target HIV RNA lengths (20, 22, and 24 nt) to investigate the effect of RNA length on incorporation into the hairpin (HP). At a target RNA concentration of 10 pM, the 22- and 24-nt target RNAs exhibited similar signals, whereas the 20-nt target RNA produced a weak fluorescence signal (Supplementary Fig. 1). Based on this result, we selected a 22-nt HIV RNA sequence and designed CRISPR anti-tag HPs with various complementary DNA blocker lengths (14-, 16-, 18-, and 20-bp) to compare the sequential reactions using these four HP types (14D-, 16D-, 18D-, and 20D-HP).

In the sequential reactions using these four HP types, no fluorescence signal was observed in the absence of HPs, regardless of the presence of target M (Fig. 3c and Supplementary Fig. 2). This indicates that the reaction between Cas13a/M crRNA and target M in step 1 does not affect fluorescence signal generation in step 2. Similarly, when the DNA blocker complementary to the HIV sequence was relatively long (20 bp; 20D HP), the fluorescence signal increase was negligible regardless of the presence of target M (Fig. 3c and Supplementary Fig. 2). This finding suggests that when the DNA blocker is long, the target RNA within the hairpin is not effectively recognized by Cas13a/HIV crRNA, regardless of whether the loop region (anti-tag sequence) is cleaved.

However, when hairpins containing 14, 16, or 18 bp DNA blockers were used, a notable increase in fluorescence signal was detected in the presence of target M. Unlike the 20D HP, when the DNA blocker complementary to the target RNA was shorter, the HIV target RNA within HP could activate Cas13a/HIV crRNA in step 2 after the loop (anti-tag sequence) was cleaved by the activated Cas13a/M crRNA in step 1. In the absence of target M, a weak fluorescence signal was generated, which indicates the background signals resulting from the interactions between HIV target RNA within the HP and Cas13a/HIV crRNA; However, these results demonstrated that signals undetectable in the absence of the hairpin were amplified to observable levels when the hairpin is used (Fig. 3c and Supplementary Fig. 2). In other words, initially undetectable signals from low concentrations of target M were amplified to observable levels using the HP. The optimal DNA blocker length was 14 bp, which effectively blocked the target RNA sequence within the HP before loop cleavage. After loop cleavage, however, Cas13a can effectively recognize the target RNA within the HP and further induce trans-cleavage (Fig. 3c).

To further confirm the hairpin cleavage, we conducted a CRISPR sequential reaction and analyzed the reaction products using PAGE gel analysis. In the presence of target M, two bands corresponding to the 14D HP (hairpin containing a 14-bp DNA blocker) were observed (Fig. 3d), indicating that the HP was cleaved by the trans-cleavage reaction of Cas13a/M crRNA (Fig. 3d, step 1, lane 4). Additionally, FQ reporter cleavage was predominantly observed in the step 1 reactant containing target M and 14D HP; however, the cleavage reaction diminished in the absence of target M (Fig. 3d, step 2, lanes 1 and 2). This result confirms that the HIV sequence in 14D HP is detected by Cas13a/HIV crRNA in step 2 only after 14D HP cleavage. When 14D HP was absent during step 1, FQ reporter cleavage by Cas13a/M crRNA activated by target M was barely detectable due to the tenfold dilution as the reaction proceeded from step 1 to step 2 (Fig. 3d, step 2, lanes 3 and 4). Based on the results of these CRISPR sequential reactions, we concluded that following the cleavage of the anti-tag sequence in the loop region of the hairpin mediator, the target sequence in the stem region is recognized by the Cas13a/crRNA complex. This recognition triggers a trans-cleavage reaction and enables CRISPR cascade signal amplification.

Sensitivity and specificity of the CARRD assay

Next, we investigated whether the CRISPR anti-tag HP, which includes the target RNA sequence, could improve target RNA detection efficiency under a one-step reaction condition using a single Cas13a/crRNA complex. To test this, we prepared solutions containing Cas13a/HIV crRNA, the CRISPR anti-tag hairpin, and the FQ reporter, and measured the fluorescence signal in the presence or absence of the HIV target RNA. We used the same CRISPR anti-tag HPs (14-, 16-, 18-, and 20D HP) containing the HIV target RNA sequence as those used in Fig. 3 and compared their results. Additionally, we evaluated the performance of the CARRD assay against the conventional Cas13a/crRNA reaction without the CRISPR anti-tag HP.

The results showed that fluorescence signal intensity was significantly higher when CRISPR anti-tag HPs were applied compared to reactions without the HP (Supplementary Fig. 3a). Among the HPs tested, 14D HP exhibited the strongest fluorescence intensity enhancement. Further optimization of the 14D HP concentration revealed that 5 nM of 14D HP produced the greatest fluorescence difference between the presence and absence of the target RNA (Supplementary Fig. 3b). Under optimized conditions, we monitored real-time fluorescence signals for HIV RNA detection across concentrations ranging from 10 aM to 1 pM. As shown in Fig. 4a and Supplementary Fig. 4, the fluorescence signal intensity was greatly enhanced with the addition of HIV 14D HP (CARRD assay), significantly improving detection sensitivity compared to the conventional Cas13a/HIV crRNA reaction without 14D HP. The limit of detection (LOD), calculated as the mean blank + 3σ of the blank, was determined to be 10 aM for the CARRD assay—10,000 times more sensitive than the conventional Cas13a/crRNA reaction (LOD: 100 fM). Additionally, time-dependent fluorescence signal changes for detecting 100 fM of HIV RNA demonstrated that the fluorescence intensity at 1 h with the CARRD assay was 22.9-fold higher than that of the conventional reaction (Fig. 4c).

Fig. 4: Sensitive and specific detection of HIV and HCV RNA using the CARRD assay.

Comparison of a HIV detection and b HCV detection by the CARRD assay with the HP and conventional CRISPR/Cas13a assay without the HP. The ΔFluorescence intensity (FTarget RNA – FControl) of the CARRD assay (orange) and conventional CRISPR/Cas13a (green) at t = 60 min for various concentrations of the a HIV RNA and b HCV RNA targets (n = 3 and data represent mean ± s.d. of three technical replicates). The CARRD assay displayed superior sensitivity compared to the conventional Cas13a/crRNA reaction. c Time-dependent fluorescence signal changes of the CARRD assay (red) and conventional Cas13a/HIV crRNA (gray). The target HIV RNA concentration is 100 fM. Each curve was adjusted by subtracting the control signal (*n *= 3 and data represent mean ± s.d). d Specificity test using Cas13a/HIV crRNA and HIV 14D HP (black), and Cas13a/HCV crRNA and HCV 14D HP (orange) by the CARRD assay. Fluorescence intensity was measured at t = 60 min. For the Cas13a/HIV crRNA reaction; [HIV RNA] = 1 pM, [HCV RNA] = 10 pM, [HIV 14D HP] = 5 nM. For the Cas13a/HCV crRNA reaction; [HCV RNA] = 1 pM, [HIV RNA] = 10 pM, [HCV 14D HP] = 1 nM (n = 3 and data represent mean ± s.d. of three technical replicates). Source data are provided as a Source Data file.

To evaluate the assay’s versatility, we applied the CARRD assay to detect hepatitis C virus (HCV) RNA (Fig. 4b). For HCV detection, the CRISPR anti-tag HP was designed with a double-stranded chimeric region composed of the HCV target RNA sequence, complementary DNA, and an anti-tag sequence in the loop region. We optimized the complementary DNA lengths to 14, 15, and 16 nt and compared the results. Among the HPs tested, the 14D HP showed the strongest fluorescence intensity (Supplementary Fig. 5a). Further optimization confirmed that the detection efficiency was optimal at a 14D HP concentration of 1 nM (Supplementary Fig. 5b). Using these optimized conditions, we measured real-time fluorescence signals for HCV RNA detection across a concentration range of 10 aM to 1 pM (Fig. 5b and Supplementary Fig. 6). Similar to the HIV detection results, the CARRD assay with the HCV 14D HP achieved a detection limit of 10 aM for HCV RNA.

Fig. 5: Comparison of crRNA variants in CARRD and conventional assays.

a Schematic of the crRNAs targeting different sites in the HIV gag gene and sequence information of the crRNAs for HIV gene detection. b Comparison of ΔFluorescence intensities (FTarget – FControl) using various crRNAs (crRNA #1, #2, and #3) in a conventional Cas13a/crRNA reaction (w/o HP) and the CARRD assay (with HP) (n = 3 and data represent mean ± of s.d. of three technical replicates). c Time-dependent fluorescence signal changes of CARRD assay (orange) and conventional Cas13a/HIV crRNA (green) using different types of crRNAs. The target HIV RNA concentration is 100 fM. Each curve was adjusted by subtracting the control signal (n = 3 and data represent mean ± s.d. of three technical replicates). Source data are provided as a Source Data file.

To evaluate the specificity of the CARRD assay, we first tested HIV detection using Cas13a/HIV crRNA and HIV 14D CRISPR anti-tag HP. A strong fluorescence increase was observed with the HIV RNA target, while the HCV RNA did not enhance the fluorescence signal (Fig. 4d, black column). Similarly, for HCV detection using Cas13a/HCV crRNA and HCV 14D CRISPR anti-tag HP, a strong fluorescence signal was observed with the HCV target RNA, while the HIV RNA produced no significant fluorescence signal (Fig. 4d, orange column). These results demonstrate the high specificity of the CARRD assay.

Evaluation of crRNA efficiency in HIV detection using the CARRD assay

It is well-known that the binding site and sequence of crRNA can significantly affect detection efficiency13,28. Therefore, to optimize the CARRD assay for HIV detection, we designed and tested different crRNAs. We designed three crRNAs along the HIV-1 genome (Fig. 5a), and we compared fluorescence signals using the conventional Cas13a/crRNA reaction and the CARRD assay. Figure 5b, c shows the differences in fluorescence signals in the presence and absence of target HIV RNA for the conventional Cas13a/crRNA reaction (without CRISPR anti-tag HP) and the CARRD assay (with CRISPR anti-tag HP) using the three crRNAs (crRNA #1, #2, and #3).

The increase in fluorescence intensity varied depending on the type of crRNA, despite using the same target RNA concentration. When the target RNA was present, crRNA #1 exhibited the highest trans-cleavage activity among the three crRNAs, regardless of the presence or absence of the HP. In contrast, with the conventional Cas13a/crRNA reaction using crRNA #3, no significant fluorescence signal increase was observed, even in the presence of the target, likely due to low or no trans-cleavage activity13,29. Interestingly, when the CRISPR anti-tag hairpin was applied, clear fluorescence differences were observed between the target and control. An additional increase in fluorescence signal was observed for all three crRNAs in the CARRD assay compared to the conventional Cas13a reaction (Fig. 5b, c). The signal intensity followed the order crRNA #1 > #2 > #3, consistent with the results of the conventional Cas13a/crRNA reaction. Based on these findings, we selected crRNA #1 for subsequent testing of HIV clinical plasma samples.

Clinical validation with HIV clinical plasma samples

We applied the CARRD assay to analyze HIV clinical plasma samples and validate its clinical feasibility. We obtained 30 deidentified clinical plasma specimens and extracted RNA from 140 µL of plasma using commercial nucleic acid extraction kits (Fig. 6a). Before testing these RNA samples with the CARRD assay, we performed real-time reverse transcription-quantitative polymerase chain reaction (RT-qPCR) to confirm the presence of nine HIV-positive samples and 21 HIV-negative samples (Fig. 6b, c). The fluorescence intensity results analyzed using the CARRD assay are shown in Fig. 6d. The fluorescence threshold was determined using background-subtracted fluorescence from negative control samples and set to the median of the negative samples (mean + 3σ), indicated by the blue line (Fig. 6d)4. In RT-qPCR, a Cq value of 40 or higher was considered negative (Supplementary Table 1). Comparing the results of RT-qPCR and the CARRD assay revealed that all results matched except for samples 24 and 25 (Fig. 6b, c). Based on the confusion matrix in Fig. 6b, the assay demonstrated a sensitivity of 77.8%, a specificity of 100%, and an overall accuracy of 93.3%. The positive predictive value (PPV) and negative predictive value (NPV) were calculated to be 100 and 91.3%, respectively. These results confirm the assay’s high precision and specificity, demonstrating its robust ability to differentiate true positives from true negatives in clinical plasma samples. The scatter plot in Fig. 6e compares the fluorescence intensity of samples from two groups: positive samples (n = 9) and negative samples (n = 21). Gray dots represent negative samples, showing generally lower fluorescence intensity, while red dots represent positive samples, showing higher fluorescence intensity. The black dot with error bars highlights a significant difference in the mean fluorescence values between the two groups (Fig. 6e). Additionally, we generated a receiver operating characteristic (ROC) curve to evaluate the diagnostic performance of the CARRD assay, using RT-qPCR as the reference standard (Fig. 6f). The area under the curve (AUC) was 0.9709, demonstrating an excellent diagnostic performance.

Fig. 6: Detection of HIV clinical samples by the CARRD assay.

a Schematic of the CARRD assay using HIV clinical plasma samples. b,** c** Comparison of the CARRD assay and RT-qPCR between positive and negative samples. A total of 30 clinical samples were evaluated using RT-qPCR and the CARRD assay. When compared with RT-qPCR, the CARRD assay results demonstrated strong agreement. Samples with a Cq value of RT-qPCR over 40 were classified as negative samples. d The fluorescence signal of the CARRD assay obtained from 30 clinical plasma samples (n = 3 and data represent mean ± s.d. of three technical replicates). e Scatter diagram for HIV diagnosis using the CARRD assay (unpaired two-tailed Student’s *t-*test, p = 0.00076). Each dot represents the mean value from three technical replicates for a single sample (n = 3). Horizontal lines indicate the group mean for the negative and positive samples, respectively. f ROC curve of 30 clinical samples (AUC = 0.9709). ROC analysis was performed with the standard parameters in Prism 10.3.1, using the Wilson/Brown method for confidence interval (CI) calculation. All statistical analyses were performed at a 95% confidence level. Area under the curve, AUC. Source data are provided as a Source Data file. Created in BioRender. Moon, J. (2025) https://BioRender.com/r57b155.

Discussion

In this study, we investigated the effects of chimeric DNA/RNA and CRISPR anti-tag sequences on the allosteric regulation of Cas13a for RNA detection. Our results show that the presence of anti-tag sequences and secondary structures in target RNA significantly inhibits the trans-cleavage activity of Cas13a. Building on this discovery, we proposed a cascade amplification method using a CRISPR anti-tag hairpin mediator and developed a simple, sensitive, and affordable CARRD assay for RNA detection at room temperature with a LOD of 10 aM (10,000 times more sensitive than the conventional Cas13a/crRNA detection). This method employs a single CRISPR/Cas13a enzyme, eliminating the need for preamplification and heating equipment. By leveraging the inherent inhibitory effect of the anti-tag sequence, the CRISPR anti-tag hairpin initially prevents Cas13a’s cleavage activity, minimizing background noise and false-positive signals. In the presence of target RNA, the Cas13a enzyme is specifically activated, initiating the non-specific cleavage of ssRNA sequences, releasing the target RNA sequence from the CRISPR anti-tag hairpin. The released target RNA sequence further activates Cas13a’s trans-cleavage reactions, enabling cascade signal amplification. Unlike other approaches that typically require multiple types of enzymes for cascade amplification3,9,17,18,30,31,32,33,34, our study demonstrates that CRISPR cascade amplification can be achieved using a single Cas13a/crRNA complex, achieving an attomolar sensitivity for RNA detection. This significantly simplifies the assay and reduces costs (Supplementary Table 2).

The development of the CARRD assay addresses several key challenges in previous CRISPR/Cas13a-based diagnostic methods. By eliminating the need for pre-nucleic acid amplification and reverse transcription, the CARRD assay streamlines the detection process and lowers costs, making it more suitable for field testing and resource-limited settings. Operating effectively at room temperature, the assay reduces the complexity and time associated with multiple handling steps while eliminating the need for high-temperature incubators. This demonstrates the potential for simple, sensitive, and minimally instrumented detection of viral RNAs, such as HIV and HCV. Notably, the CRISPR anti-tag hairpin structure retains its signal amplification effect across diverse RNA targets, underscoring the versatility of the design. Consequently, the CARRD assay provides a promising solution for improving current molecular diagnostics and expanding the accessibility of RNA virus detection in diverse clinical and environmental contexts. Future research will focus on adapting the CARRD assay to alternative readout formats, such as paper-based detection or integration with portable, low-cost fluorescence readers to enhance its suitability for point-of-care use in low-resource settings.

Desp