In microbes, clustered regularly interspaced short palindromic repeats (CRISPR) systems provide adaptive immunity against foreign DNA by utilizing CRISPR nuclease in complex with a RNA (CRISPR-RNA) to target foreign DNA for their cleavage impairment. CRISPR-RNA targets DNA with sequences complementary to RNA with a requirement that they be adjacent to a special motif called protospacer adjacent motif (PAM). Upon binding, CRISPR-RNA unwinds dsDNA target and the RNA of CRISPR-RNA forms a heteroduplex with complementary strand of the unwound DNA target resulting in a three-stranded nucleic acid structure, also known as R-loop.

These nucleases, which can be programmed via RNA, to bind and cleave any DNA has been repurposed as revolutionary genome engineering tools, where both their DNA binding and DNA cleavage functions are being used for wide-ranging applications including gene editing, regulation, visualization. But resolving CRISPR-RNA’s off-target activity problems along with a better understanding of its molecular mechanisms are near prerequisites for its efficient and accurate use, rational engineering and more widespread adoption  especially for clinical applications. 

Figure 1. Different steps of DNA targeting by CRISPR-RNA

Single molecule imaging for molecular mechanism of CRISPR

There are many steps in DNA targeting by CRISPR-RNA, some of which are summarized in Figure 1. Understanding of CRISPR's molecular mechanism will improve with individually and combinatorially investigating these steps which has been my goal during my PhD. I have employed single molecule imaging (chiefly smFRET), complemented with biochemical assays, to investigate these steps. Single molecule techniques for ideal for such investigations because:

  • observe wide ranging events (transient to long-lived).
  • detect rare events.
  • identify distinct sub-steps (for e.g. distinct FRET values for distinct sub-steps)
  • evaluate intrinsic heterogeneity of a biological system by identifying molecular sub-populations
  • all the above in real-time which allows for a kinetic analysis of distinct sub-steps.

I have performed these relevant investigations with different DNA sequences i.e. with and without off targets. So, the changing nature of a given step with off-targets not only helps in aiding the molecular understanding of it but also its specificity. Judicious selection of fluorophore labeling positions can allow you to investigate any of the given steps either individually or combinatorially. Two such labeling geometry is shown below (Figure 2), which we employed to investigate how Cas9-RNA recognizes/rejects and unwinds different DNA sequences.

We have also investigated other important steps of CRISPR-RNA targeting using a similar approach and extended it to CRISPR-Cpf1 family and other CRISPR-Cas9 variants which are currently the most widely used genome engineering CRISPR tools. The obtained information can help us understand the mechanism of these genome engineering tools, design strategies to improve their efficiency and accuracy, perform rationale-guided engineering of new CRISPR tools and CRISPR inhibitors/activators.

Figure 2. smFRET assay to study DNA interrogation i.e.(recognition-rejection) by Cas9-RNA.
(a) Schematic of single-molecule FRET assay to study DNA interrogation by Cas9-RNA. Specific interaction between Donor (Cy3) labeled and surface immobilized DNA target and acceptor (Cy5) labeled Cas9-RNA in solution results in FRET, where different FRET values report also report on nature of these interactions. (b) DNA targets with different positions and extent of off-target bp (mismatches) were used in these experiments. (c) A representative smFRET time-trajectory showing the real-time interaction between Cas9-RNA and a DNA target molecule with 12 PAM-distal mismatches. DNA targets with < 12 PAM-distal mismatches were ultra-stably bound by Cas9-RNA. (d) These experiments with different DNA targets helped us uncover bimodal binding nature of Cas9-RNA i.e. Cas9-RNA binds DNA in predominantly two distint sub-steps, first being a PAM-surveillance step and if PAM is detected then the DNA unwinding ensues leading to second sub-step. (e) Schematic of single-molecule FRET assay to study Cas9-RNA induced DNA unwinding. The FRET pair were placed in the two opposite strands of the dsDNA target, the separation between which increase during unwinding leading to a lower FRET efficiency. FRET efficiency reports on the different extent of DNA unwinding. (f) DNA unwinding in real-time with mismatches and different Cas9 variants shows how both PAM-distal mismatches and mutations of engineered Cas9s (deCas9 & dCas9-HF1) can destablize a fully unwound DNA state. (g) This destabilization correlated with reduced rate of cleavage underlying the importance of DNA unwinding for cleavage action of Cas9.

Figure 3. Cpf1 (Cas12a) vs. Cas9.
Ribonucleoprotein complex of Cpf1, also known as Cas12a, with the guide-RNA (Cpf1-RNA) interrogates DNA in two primary modes. Mode I involves PAM-surveillance which is independent of DNA target sequence. If PAM is detected then the mode II ensues which involves DNA unwinding and R-loop formation. This dual mode of DNA interrogation is common between Cpf1-RNA and Cas9-RNA. Major differences are in the specificity of stable binding and cleavage. Cpf1-RNA requires ≥ 17 bp of matching base pairs (bp) for being near-permanently locked onto the DNA target and to carry out its cleavage. Whereas Cas9-RNA can be permanently locked onto the DNA with only ≥ 9 bp but carries out its cleavage only when matching bp ≥ 16 bp. Another difference is in the manner in which Cpf1 achieves near permanent/stable binding. Post the DNA cleavage, Cpf1 employs a septum that prevents re-hybridization between two strands of the DNA thus preventing any dissolution of the R-loop and Cpf1-RNA dissociation from the DNA target.


My 3 minute case for doing smFRET for CRISPR that I presented to a general non-specialist audience.