Development of Cas12a-Based Cell-Free Small-Molecule Biosensors via Allosteric Regulation of CRISPR Array Expression

Abstract

Cell-free biosensors can detect various molecules, thus promising to transform the landscape of diagnostics. Here, we developed a simple, rapid, sensitive, and field-deployable small-molecule detection platform based on allosteric transcription factor (aTF)-regulated expression of a clustered regularly interspaced short palindromic repeats (CRISPR) array coupled to Cas12a activity. To this end, we engineered an expression cassette harboring a T7 promoter, an aTF binding sequence, a Cas12a CRISPR array, and protospacer adjacent motif-flanked Cas12a target sequences. In the presence of the ligand, dissociation of the aTF allows transcription of the CRISPR array; this leads to activation of Cas12a collateral activity, which cleaves a single-stranded DNA linker to free a quenched fluorophore, resulting in a rapid, significant increase of fluorescence. As a proof of concept, we used TetR as the aTF to detect different tetracycline antibiotics with high sensitivity and specificity and a simple, hand-held visualizer to develop a fluorescence-based visual readout. We also adapted a mobile phone application to further simplify the interpretation of the results. Finally, we showed that the reagents could be lyophilized to facilitate storage and distribution. This detection platform represents a valuable addition to the toolbox of cell-free, CRISPR-based biosensors, with great potential for in-field deployment to detect non-nucleic acid small molecules.

Figure 1

Schematic of the cell-free biosensor. (a) The transcription template contains a T7 promoter sequence, an operator sequence, a Cas12a CRISPR array that expresses poly pre-crRNA (4 crRNAs), followed by a T7 terminator sequence, and target sequences for Cas12a crRNA flanked by the Cas12a PAM sequence (TTTN). In the absence of its cognate ligand, the aTF recognizes and binds to the operator sequence located downstream of the T7 promoter, thus inhibiting the in vitro transcription of the Cas12a array as well as the activity of Cas12a. (b) In the presence of the aTF-cognate ligand, the ligand/aTF complex dissociates from the operator sequence, allowing the activity of the T7 RNA polymerase and the expression of the Cas12a array. The expressed pre-crRNAs are then processed by Cas12a into mature crRNAs, which bind with Cas12a to form active RNPs. The produced crRNAs guide the Cas12a enzyme to the PAM-flanked targets present on the same expression template. (c) The recognition and binding to the PAM-flanked target sequences by the active RNPs initiate the cleavage of the dsDNA target sequences by Cas12a and subsequent (collateral) non-specific cleavage of the surrounding fluorescently labeled ssDNA reporter, generating a detectable fluorescent signal.

Figure 2

Assessment of aTF-regulated CRISPR array expression and Cas12a activity. (a) Establishment of the in vitro transcription reaction and Cas12a activity. The in vitro transcription reactions were carried out with a transcription template expressing Cas12a pre-crRNAs (2.5 nM DNA) or in the absence of a transcription template in the presence of the Cas12a enzyme and ssDNA reporters. NTC: no template control. The values are shown as mean ± SD (n = 3). (b) Assessment of the regulation activity of TetR on the Cas12a-based sensing system. Purified TetR protein was added to the in vitro transcription reaction at two concentrations (2.5 and 10 μM). The values are shown as mean ± SD (n = 3). (c) Schematic of the effect of heparin on the in vitro transcription reaction. In the absence of heparin, the T7 RNA polymerase enables multiple turnover transcription events, generating a large number of CRISPR arrays. The addition of heparin inhibits the multiple turnover transcription events, thus controlling the production of CRISPR arrays. (d) Suppression activity of TetR and heparin on the in vitro transcription reaction. The expression system was tested with 2.5 μM TetR, 66 μg/mL heparin, both TetR and heparin, or no TetR and no heparin control. The values are shown as mean ± SD (n = 3).

Figure 3

Establishment of a Cas12a-based tetracycline biosensor. (a) Schematic of the Cas12a-based tetracycline biosensor. (b) Assessment of the Cas12a-based tetracycline sensing system. The responsiveness of the system was tested using 5 μM tetracycline with the addition of TetR and heparin. The values are shown as mean ± SD (n = 3). (c) Dose–response of the biosensor with tetracycline. The values represent three independent replicates shown as points. Data were measured as end-point detections at 120 min. Significant differences in fluorescent signal between the no-ligand control and other tetracycline concentrations were determined using one-way ANOVA with Dunnett’s multiple comparison test (****P = 0.0005; ***P < 0.0001). (d) Detection of other tetracycline antibiotics and assessment of the biosensor’s specificity. Different tetracycline-related (doxycycline, oxytetracycline) and non-tetracycline (kanamycin, ampicillin, erythromycin) antibiotics were tested together with tetracycline and the no-ligand control. Each of the antibiotics was used at a 10 μM concentration. Data were measured at 120 min. The values are shown as mean ± SD (n = 3). (e) Comparison between the ROSALIND system and the Cas12a-based system. The ROSALIND detection reactions were run following the previous protocol. The FAM ssDNA reporter was used in the Cas12a-based system, and the two systems were detected at an excitation wavelength of 486 nm and an emission wavelength of 510 nm. The ROSALIND detection reaction showed similar kinetics in response to 12.5 μM tetracycline as in the previous report (right panel). The values are shown as mean ± SD (n = 3).

Figure 4

Visual readout, mobile phone application, and freeze-drying for field-deployable applications. (a) Schematic of visual detection of Cas12a-based sensing using the P51 visualizer. After incubation at 37 °C for 1–2 h, the detection reaction tubes can be directly placed in the P51 visualizer, and the detection results can be observed in the dark with the naked eye. (b) Visual detection of the dose–response with tetracycline as shown in Figure 3c. (c) Visual detection of other tetracycline antibiotics as shown in Figure S3. NL: no ligand. (d) Schematic of the mobile phone application for result interpretation. The application can interpret detection results from P51 visualization by capturing pictures of detection reactions in P51 directly or uploading an already captured image. The tubes identified as negative are shown in red boxes, while the positive ones are shown in green boxes. Confidence scores are shown on top of the detection reactions. (e) Example of an app-processed visual detection result. The reactions containing ≥2 μM tetracycline were identified as positive. (f) Assessment of lyophilized detection reactions stored at different temperatures. The detection reactions without the ligand were lyophilized and kept at RT, 4, or −20 °C for 2 days, and then, 5 μM tetracycline was added in a 20 μL reaction. Values are shown as mean ± SD (n = 3). (g) Assessment of detection reactions with environmental samples. Left panel, schematic of collection and spiking environmental water samples. Middle panel, endpoint fluorescence readouts of tetracycline detection in spiked environmental samples (5 μM). Data were measured at 120 min. The values are shown as mean ± SD (n = 3). Right panel, representative end-point visual detection of the reactions in the middle panel.