To construct the LbCas12a expression plasmid pCDF-PylT-LbCas12a, the LbCas12a gene was amplified from plasmid pET28a-LbCas12a (ToloBio) with the primers of LbCas12a-F/R with high-fidelity DNA polymerase, and the amplicons were purified and digested with BamHI and SalI before being inserted into the same sites in pCDF-PylT (Neumann et al., 2009) via T4 DNA ligation.

The purification procedures of the recombinant proteins were the same, as previously described (Kang et al., 2022). Briefly, to obtain acetylated LbCas12a, the plasmids of pCDF-PylT-LbCas12a and pET28a-AcrVA5 (Dong et al., 2019) were co-transformed into Escherichia coli BL21 (DE3), and the transformants were induced by the addition of IPTG at a final concentration of 0.2 mM. Acetylated LbCas12a proteins were purified following the steps of the Ni-NTA column (HisTap^TM HP, GE Healthcare), ion exchange chromatography (HiTrap^TM SP HP, GE Healthcare), and molecular sieve (HiLoad^TM 16/600, Superdex^TM 200 pg, GE Healthcare). Proteins were quantitated using the Bradford method and stored at −80 °C in the storage buffer containing 50 mM Tris-HCl, 600 mM NaCl, 2 mM DTT, 0.2 mM EDTA, and 50% glycerol. The purification of CobB was the same as what was described previously (Kang et al., 2022).

To create short dsDNA, two complementary oligonucleotides of both forward and reverse were synthesized (Supplementary Table S2) (Sangon, China) and annealed at a ratio of 5:1 in 1× Taq buffer to form dsDNA, which was then stored at −20°C. The generated dsDNA was diluted to 100 nM with nuclease-free water before use.

To generate FAM-labeled double-stranded DNA targets, the target DNA fragment was amplified from plasmid pClone007s-COL1A2 (Linxian et al., 2019) with the primers of M13-F and M13-R-FAM with a high-fidelity DNA polymerase. Then, purified amplicons were quantitated with the NanoDrop 2000 instrument (Thermo Fisher Scientific) and stored at −20 °C.

To perform the deacetylation reaction, 1 μM acLbCas12a was treated with 1 μM CobB with the addition of 1 mM NAD⁺ in 1× NEB buffer 3.1 in a 20-μL reaction system. The reaction was carried out at 37 °C for 60 min, and the acetylation status of acLbCas12a was then analyzed by Western blot.

For the in vitro cis-cleavage assay, CobB-treated acLbCas12a was incubated in a 20-μL reaction system containing 150-ng target dsDNA, 1 μM acetylated LbCas12a, 1 μM CobB, 1 μM crRNA, and 1 mM NAD⁺. The reaction was performed at 37°C for 60 min in 1× NEB buffer 3.1 and then stopped by heating at 80°C for 15 min, followed by immediately chilling on ice before being further analyzed by PAGE.

For the in vitro trans-cleavage assay, CobB-treated acLbCas12a was incubated in a 20-μL reaction system containing 0.5 μM acetylated LbCas12a, 1 μM CobB, 0.5 nM target dsDNA, 12.5 nM crRNA, 0.5 μM 8C-FQ-reporter, and 0.5 μM NAD⁺. The reaction was performed at 37 °C in 1× NEB buffer 3.1, and the results were detected in a real-time PCR machine (Applied Biosystems QuantStudio 3) for 30 min with the fluorescence signal collected every 15 s (λex: 488 nm and λem: 535 nm).

The cell lines of HeLa and 97H were maintained in DMEM/high glucose (HyClone) supplemented with 10% FBS (Gibco) and 1× penicillin–streptomycin (Gibco). The cell lines of HCT116 were maintained in McCoy’s 5A medium (Sigma) supplemented with 10% FBS (Gibco) and 1× penicillin–streptomycin (Gibco). For pharmacological compound treatment, cells at ∼60% confluence were pre-incubated with FK886 (Sigma-Aldrich) at a final concentration of 100 nM and continually cultured for another 24 h before being counted and collected. The cell lines present in this study were obtained from Xiamen Immocell Biotechnology Co., Ltd.

Briefly, the HOLMES(NAD⁺) system comprises acetylated Cas12a, CobB, crRNA, target dsDNA, and single-stranded DNA (ssDNA) FQ-reporters. As acetylated Cas12a loses trans-cleavage activities with target dsDNA as the template, ssDNA FQ-reporters will not be cleaved by acetylated Cas12a, and no fluorescence signal could be detected. Only in the presence of NAD⁺, CobB then removes the acetyl unit from acetylated Cas12a and restores its trans-cleavage activities, cleaving the ssDNA FQ-reporters and illuminating fluorescence signals (Figure 1). Considering the strong relationship between NAD⁺ and the fluorescence signals in HOLMES(NAD⁺), the system can theoretically be employed for NAD⁺ detection.

To test the hypothesis, we first need to produce acetylated Cas12a and establish the HOLMES(NAD⁺) system. Usually, the acetylation process could be performed by the in vitro incubation of purified recombinant Cas12a together with AcrVA5, as previously described, where AcrVA5 efficiently acetylates the key lysine residue in Cas12a at the presence of acetyl-CoA (Kang et al., 2022). However, a further step is required to purify the acetylated Cas12a after the in vitro reaction, where the procedure is time-consuming and could be harmful to both the activity and stability of Cas12a (Supplementary Figures S1–S3). Therefore, we used an in vivo acetylation strategy by purification of acetylated LbCas12a from E. coli BL21 (DE3) co-expressing both LbCas12a and AcrVA5. Western blotting results showed that the purified LbCas12a protein was efficiently acetylated (Supplementary Figure S4), and the acetylated LbCas12a (acLbCas12a) lost both cis- and trans-cleavage activities with target dsDNA (Supplementary Figures S5–S7), which was consistent with our previous study (Kang et al., 2022) and demonstrated the practicability of the in vivo acetylation strategy. We further validated that acLbCas12a could be efficiently deacetylated by CobB (Supplementary Figure S8) and recovered its cis- and trans-cleavage activities when the NAD⁺ cofactor was added. The restored cis- and trans-cleavage enzymatic activities were comparable to those of the control group using unacetylated Cas12a (Supplementary Figures S5–S7). Notably, the reaction system with acLbCas12a, CobB, crRNA, target dsDNA, and ssDNA FQ-reporters but without NAD⁺ is named the HOLMES(NAD⁺) system.

Then, HOLMES(NAD⁺) stability was analyzed, and no significant difference was found in fluorescence intensities after 20 times of repeated freeze–thaw treatment of the enzyme mixture of Cas12a and CobB (Supplementary Figure S9). When HOLMES(NAD⁺) reactions were tested under temperatures ranging from 30°C to 45°C, the most optimal temperature was determined to be 37°C (Supplementary Figure S10). In addition, we found that HOLMES(NAD⁺) was of both high repeatability (Supplementary Figure S5) and high precision (Supplementary Figure S6). Therefore, the above results clearly showed that HOLMES(NAD⁺) could be triggered by the NAD⁺ cofactor to illuminate fluorescence signals, demonstrating that the system could be employed for stable and accurate NAD⁺ detection.

The limit of detection (LOD) of HOLMES(NAD⁺) was further determined to be 22.5 nM, using serially diluted NAD⁺ solutions ranging from 1 mM to 7.8 nM. We found that the illuminated fluorescence signals gradually increased with the increase in the NAD⁺ concentration (Figures 2A,B), and the value of the coefficient of determination (R ²) was 0.997 between the fluorescence intensity and NAD⁺ concentrations (from 1 mM to 31.3 nM) (Figure 2C). The high R ² value not only reflected the linear relationship between the added NAD⁺ concentration and the activated LbCas12a trans-cleavage activities but also indicated that HOLMES(NAD⁺) could be employed for the quantitative determination of the NAD⁺ concentration.

The specificity of the HOLMES(NAD⁺) system was then tested with the employment of various small molecules, including similar nucleotides and precursors for NAD⁺ biosynthesis. As expected, although some small molecules, such as AMP, ADP, ATP, NMN, NADPH, and NADP⁺ may share similar molecular structures with NAD⁺, they are not the cofactors of CobB and are theoretically unable to restore the trans-cleavage activities of acLbCas12a, resulting in no fluorescence signal output. However, to our great surprise, in addition to NAD⁺, HOLMES(NAD⁺) also illuminated fluorescence signals when NADH was supplemented, which is inconsistent with the fact that NAD⁺ is the only cofactor of CobB (Figure 2D). Therefore, we hypothesized that the commercially available NADH was contaminated with NAD⁺, which resulted in the above unexpected signal output for NADH. To test this hypothesis, NADH was then measured by the mass spectrometry (MS) method, and the tested NADH solution was found to be a mixture of NADH and NAD⁺, which could therefore explain the above unexpected results for NADH. Moreover, the concentration of NAD⁺ in the NADH solution (500 nM) was 154.5 nM when measured using the quantitative MS method, which was consistent with the HOLMES(NAD⁺) method, that is, 140.8 nM (Supplementary Figures S16–S19). Taken together, the above findings not only proved the high specificity of the HOLMES(NAD⁺) method but also demonstrated its ability in target NAD⁺ quantitation.

In addition to pure NAD⁺ solutions, we also verified the HOLMES(NAD⁺) system with extracts from real biology samples. Total NAD⁺ was first extracted from different cancer cell lines by heating at 95°C and then neutralized before being detected. Using the HOLMES(NAD⁺) method, the cellular NAD⁺ levels were identified to be approximately 200–600 pmol cells. The samples were also measured with other methods in parallel, and the results were consistent with those of the MS and MTT methods (Figure 3A). Moreover, the HOLMES(NAD⁺) method was further used to quantitatively measure the changes in the cellular NAD⁺ concentrations, including 97H, HeLa, and HCT116. After being treated with FK866, an inhibitor of the nicotinamide phosphoribosyltransferase, the cellular NAD⁺ levels dropped in all tested cell lines (Figure 3B), which is consistent with the previous report (Pittelli et al., 2010).

Previous studies have uncovered the potential of CRISPR technologies in non-nucleic acid detection, and the targets include both proteins (e.g., microcystin-LR, thrombin, and IFN-γ) (Deng et al., 2021; Wang et al., 2021; Wu et al., 2021) and small molecules (e.g., ATP, antibiotics, uric acid, and melamine) (Liang et al., 2019; Qiao et al., 2021; Chen et al., 2022; Mahas et al., 2022). Here, with the use of NAD⁺-based posttranslational modification of Cas12a, we create an NAD⁺-specific CRISPR diagnostic system, namely, HOLMES(NAD⁺), further expanding the Cas12a application in non-nucleic acid detection scenarios. Through comparison of different NAD⁺ detection methods, we find that different methods show distinct advantages and disadvantages (Supplementary Table S7). For example, although the MS method requires expensive instruments and professional operations, it can measure more than one target molecules in one test and allow for organic extraction that may better preserve molecules, such as NADH and NADPH (Yamada et al., 2006). HOLMES(NAD⁺) and other methods, by contrast, detect merely the target NAD⁺ concentration, which could be less convenient when the detection of multiple molecules is needed. In addition, we also observed that the detection efficiencies of the enzyme-based HOLMES(NAD⁺) and MTT methods might be slightly affected by intracellular substances, i.e., the efficiencies are enhanced in the HCT116 and HeLa samples, but reduced in the 97H sample in comparison with the MS method. Overall, HOLMES(NAD⁺) shows apparent advantages in rapidness (half an hour), convenience, cost-effectiveness, and accuracy in the quantitative detection of NAD⁺ concentrations, which therefore shows great potential in both scientific research and industries.