Last update: Sep 2020


Cas12b, formerly known as C2c1, is an alternative to Cas9 and Cas12a as an RNA-guided endonuclease in animal and plant genome engineering. CRISPR-Cas12b is a class 2 type V-B CRISPR system.

Cas12b was first identified in 2015 by the research team led by Feng Zhang (Broad Institute of MIT and Harvard), who also contributed to the development of CRISPR-Cas9 and CRISPR-Cas13 systems. Cas12b was isolated from thermophilic bacteria usually found in geysers, hot springs, volcanoes, et cetera. Thus, it usually works at higher temperatures compared to other Cas nucleases.

Similarly to Cas12a, Cas12b has a single RuvC domain and recognizes T-rich PAMs (VTTV for AaCas12b and AacCas12b, ATTN for BthCas12b; V is A, C, or G). It cleaves non-target and target strands respectively 14-17 and 23-24 bases downstream of the PAM sequence. The double strand break (DSB) results in staggered 6-8 base 5'-overhangs. Similarly to Cas9, Cas12b uses both crRNA and tracrRNA that can be fused into sgRNA.

Figure 1. Crystal structure of Cas12b and its gRNA bound to DNA

Cas12b nucleases are generally smaller than Cas9 and Cas12a nucleases. For example, BhCas12b consists of 1108 amino acids and AacCas12b of 1129. Because of this small size, Cas12b can be potentially easier to deliver into cells via adeno-associated viruses (AAVs), which are widely used for in vivo delivery of gene therapies, e.g., for treating lipoprotein lipase deficiency in humans.

Cas12b has been recognized to have very high target specificity (low off-target editing), as shown by the example of AaCas12b, which has low tolerance even for single base pair mismatches in mouse and human cells. This characteristic makes Cas12b a strong candidate for therapeutic genome editing.

One huge disadvantage of Cas12b nucleases is the high temperature required for their operation. Commonly researched AacCas12b isolated from Alicyclobacillus acidoterrestris displays optimum activity at 48 °C, which makes it nearly impossible to use it in mammalian and plant cells. Fortunately, scientists have found several other Cas12b variants which cleave DNA at lower temperatures.

Cas12b in Humans

In 2019, Feng Zhang and his associates published a paper in Nature Communications, demonstrating the applications of Cas12b in human genome editing. In their study, they focused predominantly on BhCas12b isolated from Bacillus hisashii. The goal of the study was to identify Cas12b nucleases that can operate at lower temperatures than previously studied AacCas12b so that the system can be employed for human genome editing.

First, the team conducted a BLAST search of the sequence databases, finding 27 members of the Cas12b family. Of those, they chose 14, and confirmed 4 to have some cleavage ability: AkCas12b, BhCas12b, EbCas12b, and LsCas12b. Only AkCas12b and BhCas12b could make a viable number of DSBs at 37 °C (human body temperature), so the researchers focused on those variants. That said, percentage indels generated by those Cas12b nucleases was below 1%, which meant that the researchers had to modify the nucleases or their sgRNAs in order to increase genome editing efficiencies to acceptable levels. While altering AkCas12b did not result in an increased number of indels, 5-nt truncation on the 5′ end of the BhCas12b sgRNA contributed to up to a 30-fold efficiency increase.

One problem that the researchers encountered was the frequent failure of Cas12b nucleases to generate DSBs. Instead, Cas12b made single-strand breaks, i.e., nicks in the DNA sequence, usually on the non-target strand. This behavior was especially prevalent at lower temperatures. Therefore, scientists decided to engineer Cas12b in order to create a better functioning nuclease.

Researchers noted that many enzymes that are active at lower temperatures have glycine residues on their surface. Those residues contribute to increased protein flexibility and enzymatic activity. Thus, the team generated glycine substitutions at 66 residues on the surface of Cas12b nuclease, progressively refining their methods until they developed BhCas12b v4 (variant 4), which exhibited high activity for a range of target sequences. BhCas12b v4 also displayed higher dsDNA cleavage activity at 37 °C and reduced number of DNA nicking in vivo (in human cells).

Next, Zhang and his associates tested BhCas12b v4-mediated genome editing at 56 targets within 5 genes in 293T (human) cells, identifying efficient cleavage at ATTN PAMs, and less efficient, but still adequate cleavage at TTTN and GTTN PAMs. The positive control used for the experiments was AsCas12a with the TTTV PAM sequence. As you can see in Fig.2.a, on average, BhCas12b v4 targeting ATTN PAM generated more indels than AsCas12a for the sites tested in this experiment.

Figure 2. Human genome editing using BhCas12b v4

Looking at the number of ATTN sites within the human genome (representing the number of sites that can be cleaved), researchers concluded that BhCas12b v4 has a comparable number of targetable sites to Cas12a, and thus could be a viable alternative to the latter. Also, as shown in Fig.21.b, where y-axis stands for a fraction of indels per respective indel lengths (x-axis), BhCas12b was also found to generate large deletions of 5–15 bp, bigger than the ones made by SpCas9 and AsCas12a.

To further investigate the genome editing of human cells using BhCas12b v4, the team tested the performance of Cas12b ribonucleoproteins (RNPs) in editing primary human T cells. The team created BhCas12b v4-sgRNA (RNP) complexes and delivered them into human CD4+ T cells using electroporation, which yielded 32–49% indels for 3 target genes tested. This series of experiments shows that BhCas12 v4 could be used in human genome editing employing multiple methods.

Lastly, the team studied the targeting specificity of BhCas12b in 293T cells. They chose 9 target sites that had roughly similar indel activities for all Cas nucleases tested, as shown in Fig.3.a. AsCas12a was only tested on the last 3 sites. The researchers then performed Guide-Seq analysis, with results presented in Fig.3.b. The circle graphs show the number (wedges), and the relative proportion of DSB sites for each nuclease tested. The purple and dark/light blue wedges show on-target cleavage proportion for each nuclease, while the light-colored wedges represent off-target DSBs. The number under the circle stands for a fraction of on-target cuts. As you can see, all detected cleavage sites were on-target in the case of BhCas12b v4 and AsCas12a. On the contrary, the SpCas9 displayed way poorer performance, with a high number of off-target cleavage. Notably, SpCas9 knocked out the DNMT1 gene only in 10% of cases, making DSBs at 101 other sites, while BhCas12b v4 was perfectly on-target. The team ran an additional Guide-Seq analysis for the sites that could not be targeted by both BhCas12b v4 and SpCas9 and therefore were tested only by BhCas12b and found off-target cleavage at only 2 out of 14 tested sites.

Figure 3. Testing the specificity of BhCas12b v4 nuclease

Furthermore, as shown in Fig.3.c, indel activity decreased markedly when single base mismatches between the guide (sg)RNA and target DNA were introduced into the sgRNA at the respective positions (x-axis). The single mismatch tolerance was especially low for the first 9 and 15th-19th nucleotides (nt). On the right, you can see that the indel activity was even lower when double mismatches (2 consecutive nt) were present in sgRNA until the 20th nt. These results demonstrate the high specificity of BhCas12b, consistent with the similar characteristic displayed by AacCas12b in previous studies.

Overall, Zhang and his collaborators engineered an improved Cas12b nuclease, creating BhCas12b v4, and investigated its properties, showing that it can be employed for high specificity genome editing in human cells. Because of its small size and low off-target editing, it could be applied in in vivo genome editing in the near future.

Cas12b in Plants

In 2020, Ming et al. attempted to develop a CRISPR-Cas12b system that could be applied in plant genome engineering. They tested the performance of 4 orthologs of Cas12b: AacCas12b, AaCas12b, BthCas12b, and BhCas12b in rice genome engineering, in order to find the most promising candidate. The team found that AaCas12b from Alicyclobacillus acidiphilus offers the comparatively highest efficiency in targeted mutagenesis.

In their first set of experiments, the team compared the genome editing efficiency of AacCas12b, AaCas12b, and BthCas12b in OsEPFL9 (GTTG PAM) and OsGS3 (ATTC PAM) genes of rice protoplasts. As shown in Fig.4, AaCas12b had the highest (above 10%) editing efficiency of the three, while BthCas12b displayed the poorest performance. All 3 orthologs created 4–14 bp deletions (markedly higher than those generated by Cas9) 12–24 nucleotides (nt) away from the PAM site.

Next, the team studied the PAM requirements for AacCas12b and AaCas12b, which were the 2 most promising orthologs in the genome editing of rice protoplasts. They targeted several VTTV (V is A, C, or G) PAM sites in rice protoplasts. They found that both orthologs are often able to recognize VTTV PAMs, especially ATTV and GTTG PAMs, which is mostly consistent with previous reports on PAM sites recognized by Cas12b orthologs used in human cells.

Figure 4. Comparison of editing efficiency of 3 Cas12b nucleases

Since AaCas12b displayed the highest mutation rates, it was deemed the most appropriate for targeted mutagenesis in rice, and Ming et al. used this ortholog only for the next set of experiments. The team tested the targeting specificity of AaCas12b by assessing the mutation frequency (representing genome editing efficiency) when double mismatches (in 2 consecutive nucleotides) are introduced into 20 nt long crRNA protospacer sequences at 2 sites (OsEPFL9-sgRNA02 and Os12g24050-sgRNA01). The results are presented in Fig.5. MM stands for a mismatch. As you can see, compared to the on-target control (no mismatches, 2nd bar from the bottom), mutation frequencies in the presence of double mismatches were extremely low for all mismatch sites and were not exceeding 3%. This data suggests that AaCas12b is a highly specific nuclease when employed in rice cells.

Figure 5. Off-targeting analysis with double mismatches in sgRNA

Moreover, the team showed that AaCas12b could not perform genome editing when its protospacers were shortened to 18 nucleotides or below (Fig.6). Though further studies are needed to confirm this behavior, these experimental results suggest that AaCas12b may be very different from Cas9 and Cas12a, which can often still make DSBs when equipped with 17–18 nt protospacers.

Figure 6. Genome editing with truncated protospacers

In the 2nd part of the paper, Ming et al. assessed the ability of AacCas12b and AaCas12b to generate rice mutants by targeting the OsEPFL9-sgRNA02 site, transforming the constructs into rice calli by Agrobacterium. 36.4% of the 22 T0 stable transgenic lines generated by AacCas12b had monoallelic mutations at the target site, and none had biallelic mutations. In the case of 24 T0 lines generated by AacCas12b, 54.2% had monoallelic, and 25% had biallelic mutations. This data shows that AacCas12b and AaCas12b can generate stable mutants in rice.

In the next set of experiments, the team attempted Cas12b-mediated multiplexed genome editing by targeting 3 genes with 3 sgRNAs (OsROC5-sgRNA02, OsEPFL9-sgRNA02, and OsGS3-sgRNA02) using AacCas12b and AaCas12b, studying 24 independent T0 lines per each nuclease. As shown in Fig.7, AacCas12b generated monoallelic mutation in 4.2%, 50%, and 25% of T0 lines in the respective 3 genes, and biallelic mutation in 12.5% of T0 lines in the OsEPFL9 gene. Of those, 4 lines were double mutants.

Figure 7. Multiplex genome editing using AacCas12b

As shown in Fig.8, AaCas12b generated a monoallelic mutation in 0%, 66.7%, and 70.8% of T0 lines in the respective 3 genes. Also, it generated a biallelic mutation in 29.2% and 45.8% of T0 lines in the OsEPFL9 and OsGS3 genes. Overall, 16 lines were double mutants, and 7 were biallelic double mutants. 0% mutation rate in the OsROC5-sgRNA02 site is consistent with the low editing activity measured for this sgRNA in protoplasts in the previous experiments. No off-target effects were detected in the randomly selected double mutants generated by AacCas12b and AaCas12b. This experimental data suggests that AacCas12b and AaCas12b might be potentially used in multiplexed genome editing, though further studies editing more genes are certainly needed.

Figure 8. Multiplex genome editing using AaCas12b

Next, the team employed three Cas12b orthologs in CRISPRi in order to control gene transcription levels. First, they introduced mutations into RuvC domains of AacCas12b, AaCas12b, and BthCas12b to render them DNase dead (dCas12b). They used those dCas12b nucleases on their own (without fusion to transcription repressors) to target the Os04g39780 gene. The team created 27 CRISPRi constructs (9 constructs per dCas12b ortholog) targeting 3 different PAMs, using 3 sgRNAs targeting either the promoter or the coding sequence, and assessed target gene expression by quantitative PCR with reverse transcription (qRT–PCR). While dBthCas12b mediated poor transcriptional repression, dAacCas12b and dAaCas12b were able to reduce gene expression to 25–75% of the original value for most target sites, both at the promoter region and the coding sequence. When fused to three SRDX repressor domain, dCas12b proteins mediated comparable levels of transcription repression to when they were acting on their own. This result hints towards CRISPRi effects being the result of transcription interference through dCas12b binding, and not chromatin modifications caused by the SRDX repressor.

Since currently there are no efficient Cas12a transcriptional activation systems in plants, the team attempted to activate genes using Cas12b, this time also including BhCas12b-v4 (previously used in humans) in their experiments. They engineered sgRNA scaffolds to increase activation efficiency. Since AaCas12b showed the most promise in their early attempts, they focused on this ortholog. The most notable transcription activation for OsER1 and OsGW7 genes was achieved with dAa-Cas12b fused to TV transcriptional activator and coupled with Aa3.8 sgRNA scaffold (3-5 fold activation) and with dAaCas12b-TV and Aac.3 sgRNA scaffold-mediated recruitment of MS2-VPR (5-8 fold activation). Thus, the researchers managed to demonstrate an effective transcriptional activation system employing dAaCas12b.

In their study, Ming et al. established Crispr-Cas12b as another system that can be employed in plant genome engineering as an alternative to Cas9 and Cas12a. They showed it could be used for genome editing as well as CRISPRi. One of the challenges is to find the Cas12b variant that has high efficacy at lower temperatures since plants themselves usually grow in lower temperatures than mammals and their cells. The improved activity at lower temperatures could be achieved by protein and sgRNA engineering.


In 2019, Teng et al. developed a sensitive and accurate DNA detection method called CDetection, which takes advantage of the high specificity of Cas12b nuclease. It can detect even attomolar (aM, 10−18) or smaller concentrations of DNA in a sample. It takes advantage of non-canonical collateral (non-specific) single-strand DNA trans-cleavage performed by Cas12b when triggered by its binding to the target double-strand DNA.

CDetection exhibits higher sensitivity than the DETECTR system based on Cas12a. Thanks to the high specificity of Cas12b, CDetection can recognize even single-nucleotide differences when equipped with tuned guide RNA (tgRNA). It does not need an extra step required in Cas13-SHERLOCK- or Cas14-DETECTR-based dsDNA detection methods and instead can directly detect DNA. The researchers tested the new system using mock samples, demonstrating direct circulating free DNA (cfDNA) detection. However, they admit that CDetection might not work so effectively when studying actual clinical specimens of poor purity and integrity.

Nonetheless, Teng et al. theorize that CDetection could be used for fast and cheap molecular diagnostics and clinical research in the future. It might be able to detect point mutations related to over 20,000 known human diseases.



[1] Ming, M., Ren, Q., Pan, C. et al. CRISPR–Cas12b enables efficient plant genome engineering. Nat. Plants 6, 202–208 (2020).

[2] Strecker, J., Jones, S., Koopal, B. et al. Engineering of CRISPR-Cas12b for human genome editing. Nat Commun 10, 212 (2019).

[3] UniProt. 2019. CRISPR-associated endonuclease Cas12b. [Accessed 23 July 2020].

[4] Broad Institute of MIT and Harvard. 2019. Scientists engineer new CRISPR platform for DNA targeting. [Accessed 23 July 2020].

[5] Yan, F., Wang, W. & Zhang, J. CRISPR-Cas12 and Cas13: the lesser known siblings of CRISPR-Cas9. Cell Biol Toxicol 35, 489–492 (2019).

[6] Teng, F., Guo, L., Cui, T. et al. CDetection: CRISPR-Cas12b-based DNA detection with sub-attomolar sensitivity and single-base specificity. Genome Biol 20, 132 (2019).

[7] Wang, D., Tai, P.W.L. & Gao, G. Adeno-associated virus vector as a platform for gene therapy delivery. Nat Rev Drug Discov 18, 358–378 (2019).


Figure 1. Broad Institute of MIT and Harvard. 2019. Scientists engineer new CRISPR platform for DNA targeting. [Accessed 23 July 2020].

Figure 2&3. Strecker, J., Jones, S., Koopal, B. et al. Engineering of CRISPR-Cas12b for human genome editing. Nat Commun 10, 212 (2019).

Figure 4-8. Ming, M., Ren, Q., Pan, C. et al. CRISPR–Cas12b enables efficient plant genome engineering. Nat. Plants 6, 202–208 (2020).