Description

The Glow Worms research stream is a course-based undergraduate research education (CURE) laboratory that focuses on making fluorescent protein knock-in strains at genetic loci of interest to the research community. To facilitate making strains, we aimed to determine whether the self-excising cassette (SEC) selection method for CRISPR knock-in could be improved. Indeed, we recently reported that using high-quality DNA and reducing Cas9/sgRNA concentration facilitates recovery of homology directed repair (HDR) knock-ins, requiring microinjection of only 10 P0 worms (Huang et al. 2021). Here, we show that cloning of SEC-based CRISPR plasmids can also be highly efficient and reproducible when optimized based on quantitative considerations.

Construction of the Cas9/sgRNA plasmid requires inserting a unique gRNA sequence into plasmid pDD162 (Dickinson et al. 2013) or one of its derivatives (Ward 2015; Aljohani et al. 2020) via site-directed mutagenesis. This involves a PCR step, followed by a KLD ligation step. Using a standard NEB® Q5 PCR protocol of 30 cycles (M0491) and a standard NEB® KLD reaction protocol (M0554), we found that 64.1% of clones included a correct gRNA insertion (Fig. 1A, B). Incorrect clones could be categorized into one of three groups: (1) those lacking the gRNA insertion (i.e. the pDD162 PCR template was transformed instead); (2) those lacking a base or two at the 5’ end of the gRNA (i.e. presumably the result of oligos in the primer pool with incomplete synthesis errors); or (3) those where a reproducible recombination of pDD162 had occurred (Fig. 1B). Based on the types of incorrect clones observed, we hypothesized that we were both using too much DNA in the KLD reaction and that the 5 minute room temperature incubation was insufficient for DpnI to digest the methylated pDD162 template prior to transformation. Indeed, limiting the PCR to 15 cycles and inclusion of a 20 minute 37°C incubation step in the KLD ligation protocol resulted in 90% of clones being correct Cas9/sgRNA plasmids (Fig. 1A, B). This is highly cost effective, as we now routinely transform with as little as 10 µL of competent cells and only need to screen one clone.

Construction of the FP/SEC plasmid requires insertion of two PCR-amplified homology arms into a pre-digested plasmid via Gibson assembly (Dickinson et al. 2015). Using the published protocol, which specifies a 4:1 volume ratio of purified homology arms to purified vector DNA (Dickinson et al. 2015), we found that only 35% of clones were correct assemblies, and the percentage of correct clones was highly variable among unique assembly reactions (Fig. 1D). Indeed, sometimes we would need to screen as many as 10 clones to identify a correctly assembled plasmid. This low efficiency was most likely due to both student inexperience with cloning and inconsistencies in the amount of DNA being used in the assemblies. Thus, we decided to take a more quantitative approach, following the recommendations of the NEBuilder® HiFi DNA Assembly protocol for a four fragment assembly, which specifies a 1:1 ratio of homology arms and vector DNA at 0.05 pmol each for all fragments. We quantified the DNA using ethidium bromide staining, an NEB® DNA ladder standard (N3200S), and Bio-Rad Image Lab 6.0 software (Fig. 1C). This definitely resulted in improved efficiency, yielding 58% of clones with correct assemblies (Fig. 1D). Nevertheless, the frequency of assembly errors remained high (42%), so we wondered whether we were again using too much DNA. We reduced the total amount of DNA 16-fold and modified the protocol to a 2:1 ratio of homology arms to vector DNA, which we based on our typical yield of homology arm and vector DNA following spin column clean-up, as well as limiting the total reaction volume to 10 µL (see Methods). Remarkably, this resulted in a highly reproducible success rate, with an average of 89% of clones being correct for 14 unique assemblies representing 14 different gene targets (Fig 1D). Thus, we now routinely need to screen just one or two clones per assembly, and oftentimes 100% of the clones are correct when screening three or more.

Overall, in combination with our previous report of highly improved knock-in efficiency via the SEC selection method (Huang et al. 2021), our new finding that SEC-based plasmids can be reproducibly cloned at 90% efficiency has essentially resolved the two major drawbacks of plasmid-based CRISPR knock-in. As such, plasmid-based approaches are equally attractive relative to recent approaches utilizing commercially available Cas9/RNP products and linear DNA repair templates that do not require any cloning (Dokshin et al. 2018; Ghanta and Mello 2020). As an undergraduate CURE, plasmid-based approaches are more cost effective, much more scalable to a large number of students, an excellent educational opportunity for molecular cloning, and plasmids are a permanent archive of sequence-verified clones that can be relatively easily repurposed for alternative tags such as other fluorescent markers or auxin-inducible degron (AID) (Ashley et al. 2021). We anticipate that our improvements will make plasmid-based CRISPR knock-in accessible for a wider range of researchers.