The ease of generating transgenic worms has been an important advantage in C. elegans for decades. Prior to the use of fluorophores as co-injection markers, two popular co-injection strategies took advantage of whole-body phenotypes induced by single point mutations. One strategy used wild-type genes to rescue visible loss-of-function phenotypes. Two examples of this rescuing strategy include pha-1 (wt) rescue of pha-1 (e2123ts) (Granato et al., 1994) and unc-119 (wt) rescue of unc-119 ( ed3 ) (Maduro and Pilgrim, 1995; Praitis et al., 2001). A second strategy used gain-of-function alleles that induced visible phenotypes injected into wild-type animals. One example of this second strategy is rol-6 ( su1006 ) to induce a roller phenotype (Kramer et al., 1990; Mello et al., 1991). These strategies either require injecting into mutant animals, which can be challenging, or generating mutant animals, which can affect worm health or the ability to visualize specific phenotypes. Many of these transgenic lines were converted into integrated strains that remain useful today, particularly if the co-injection, whole-animal phenotypes could be removed.

The rol-6 ( su1006 ) gain-of-function allele, as its name suggests, induces worms to roll (Cox et al., 1980; Park and Horvitz, 1986; Chen et al., 2003; Thein et al., 2003). The rol-6 ( su1006 ) mutation also induces a developmental delay (Sparling et al., 2023) and lower mating efficiency in males (Evans, 2006). Larval lineage tracing is more difficult in rol-6 ( su1006 ) larva due to their twisted cuticle (Evans, 2006). In addition, for experiments that involve visualizing the worm nervous system, the ventral nerve cord becomes more difficult to visualize in rol worms than in wild-type animals.

The advent of CRISPR-Cas9 gene editing now makes it possible to engineer single point mutations with surgical precision. CRISPR-Cas9 gene editing has been extensively used in C. elegans to edit single loci in animals, with a guide RNA targeting a single locus in the worm genome (Cho et al., 2013; Arribere et al., 2014; Zhao et al., 2014; Paix et al., 2015; Prior et al., 2017; Dokshin et al., 2018). In addition, the SKI LODGE system developed by the Mair lab uses the popular dpy-10 ( cn64 ) guide RNA to target the endogenous copy of dpy-10 as the co-injection marker and an inserted copy of the dpy-10 guide sequence to edit the SKI LODGE cassette inserted at a safe harbor locus (Silva-García et al., 2019). Thus, efficient guide RNAs can simultaneously target more than one locus in a genome.

We asked whether we could use CRISPR-Cas9 to “un-roll” roller worms made with traditional microinjection techniques. Traditional microinjection techniques generate extrachromosomal arrays that contain many copies of both the experimental plasmid and the co-injection plasmid. Thus, worms with integrated rol-6 (gf ) arrays contain an unknown, but likely high, number of rol-6 (gf) copies that would need to be reverted to wild-type during “un-rolling.” CRISPR/Cas9 broadly involves two steps: a guide RNA targets the Cas9 enzyme to a specific location to induce a double-strand DNA break and an exogenous repair template provides the sequence to the endogenous repair machinery to generate the desired edit via homologous recombination  (Arribere et al., 2014; Zhao et al., 2014; Paix et al., 2015; Prior et al., 2017; Dokshin et al., 2018). We designed a guide RNA targeted to the rol-6 ( su1006 ) allele and a repair template to replace rol-6 (gf) with wild-type rol-6 . Importantly, in the unlikely event that the endogenous rol-6 locus were targeted, the repair template should have replaced that edit with wild-type rol-6 ( Figure 1A ).

We used ribonucleoprotein-complex CRISPR-Cas9 following established methods to introduce the RNA and Cas9 protein via microinjection into adult roller BW1932 (genotype ctIs39 ) worms (Fay et al., 1999). We chose the healthiest line of normally-moving worms and assigned this gene-edited strain a new genotype: znyIs15 (* ctIs39 ). We quantified the number of worms that displayed wild-type versus roller phenotypes in both the ctIs39 and znyIs15 worms. We found that while 100% of ctIs39 worms exhibited roller movement, the unrolled znyIs15 worms all displayed wild-type, sinusoidal movement ( Figure 1B, 1D-E). Finally, we genotyped znyIs15 worms for rol-6 ( su1006 ) to determine if all rol-6 array copies were reverted to wild-type. We were surprised to find that our unrolled znyIs15 DNA only displayed a faint WT band ( Figure 1J ). Perplexed, we then sequenced the znyIs15 rol-6 genotyping PCR product. We discovered that znyIs15 worms harbored a single-nucleotide insertion near the target edit site, inducing V69D and a frame shift, resulting in an early stop codon ( Figure 1K ). The sequencing results suggest that for successfully unrolled worms, the repair machinery ignored the repair template and instead randomly inserted a nucleotide during repair of the majority of array copies of rol-6 . Further, it appears that this insertion occurred in multiple copies of rol-6 (gf) within the array and likely inactivates most of the array copies of rol-6 . These inactivated copies rol-6 , combined with a few reverted WT rol-6 edits, resulted in unrolled worms.

To ensure that this strategy is possible across roller strains developed by different labs, we next unrolled an independent roller strain ( OH6020 ; genotype otIs185 ) (Sarin et al., 2007). Following the same protocol, we were again able to isolate unrolled worms after CRISPR-Cas9 gene editing, assigning this gene-edited strain the genotype znyIs16 (* otIs185 ). Similar to our previous quantifications, we found that the unrolled znyIs16 worms completely lost the roller phenotype of otIs185 worms ( Figure 1C, 1F-G). We also genotyped znyIs16 animals for rol-6 ( su1006 ) , again finding that unrolled znyIs16 DNA only displayed a faint WT band and a larger-than-expected PCR product ( Figure 1J ). Similar to our znyIs15 sequencing results, we identified a 79-base pair insertion near the edit site in znyIs16 worms. This insertion induced R68L and a frame shift ( Figure 1K ), again generating an early stop codon and likely inactivating most of the array copies of rol-6 . As with znyIs15, a small fraction of array copies of rol-6 in znyIs16 do appear to have been reverted to WT. Together, our data indicate that the guide RNA successfully targeted Cas9 to rol-6 (gf) loci, but that we chose a suboptimal repair strategy. These results indicate that we can use CRISPR-Cas9 to unroll worms with roller phenotypes induced by the rol-6 ( su1006 ) gain-of-function allele, but that unrolling worms appears to be more successful by inactivating the array copies of rol-6 (gf) than by reverting them to wild type rol-6 .

We next validated the utility of unrolling worms by visualizing the ventral nerve cord in unrolled worms by generating transgenic lines labeling the nervous system with p rab-3 ::mCherry. Indeed, unrolled worms displayed the typical, wild type untwisted ventral nerve cords that were easier to visualize with confocal microscopy than their roller counterparts ( Figure 1H- I). These data indicate that we can unroll worm strains initially generated with roller co-injection markers instead of laboriously remaking integrated strains with different co-injection markers.

Our data show that CRISPR-Cas9 can be successfully used to edit multiple copies of rol-6 ( su1006 ) in C. elegans , including many copies of rol-6 (gf) in multicopy arrays following microinjection and integration. Co-injection markers serve an important purpose in many transgenic strains: they identify the desired transgenic array during crossing and maintenance. Some transgenic arrays are only identifiable via their co-injection markers with a conventional, dissection microscope. Thus, some strains harboring rol-6 (gf) co-injection markers may not be easy to work with once unrolled. This is strain-dependent, as we found that we can visualize the GFP signal of ctIs39 /znyIs15 with a fluorescent dissection microscope. One way to address this limitation would be to intentionally insert a cassette into the repair template that could be identified by PCR-based genotyping, similar to the cassette we unintentionally inserted to yield znyIs16. Together, our results identify a new CRISPR-based method to easily revert phenotype-causing rol-6 (gf) point mutations from existing transgenic worm strains. This theory may also be applied to other gain-of-function co-injection markers.

Genome editing

The rol-6 ( su1006 ) guide RNA (crRNA) and repair template (HDR, ssODN) were designed using Integrated DNA Technologies (IDT) Alt-R™ HDR Design Tool ( https://www.idtdna.com/site/order/designtool/index/HDRDESIGN ).

znyIs15 and znyIs16 were generated using an adapted CRISPR protocol (Ghanta and Mello, 2020). Briefly, the rol-6 ( su1006 ) crRNA, tracrRNA, and Cas9 protein were mixed and injected into BW1932 and OH6020 adult worms. All CRISPR reagents were obtained from IDT.

Injected worms were singled onto plates and allowed to recover and lay eggs for three days. Then, plates were examined for normal-moving progeny. All normal-moving progeny were singled onto new plates. After another three days, plates were examined for F2 progeny that exhibited normal sinusoidal movement. The healthiest line for each “un-rolling” was selected for subsequent experiments.

Quantifying rol phenotype

We counted L4-adult worms freely moving on plates with OP50 lawns, assaying worms for the roller phenotype versus sinusoidal movement. Approximately 100 worms were quantified per genotype per session. Worms were quantified on three separate sessions on different days.

Transgenesis

Transgenic worms were generated following established techniques. pwAS70 [ rab-3 p::mCh] and pwAS1 [ unc-122 p::gfp] were co-injected at 5ng/uL and 30 ng/uL, respectively, into N2 or VOE643 worms to generate znyEx214 and znyEx220, respectively. znyEx214 was crossed to otIs185 to generate VOE689 .

Microscopy

For white-light images, worms were left on agar with OP50 . Worms were imaged on a dissection microscope at 25x magnification. Images were captured with an iPhone. Images were cropped and rotated in FIJI.

For fluorescent images, worms were mounted on 2% agarose pads and immobilized in 10 mM levamisole (Sigma 31742-250MG). Worms were imaged on a spinning disk confocal microscope (Nikon Ti2 Inverted Confocal with Yokogawa W1 Spinning Disk Package, Nikon Instruments) with a CFI super fluor 40x, 1.3 NA oil immersion objective (Nikon Instruments). Z-stack digital micrographs (step size: 0.5 microns) were acquired with a back-illuminated cCMOS camera (Teledyne Photometrics) using Nikon Elements software (Nikon Instruments).

Statistics

We performed two-tailed Fisher's exact tests with Prism 10 (GraphPad).

Strain table

Plasmids

Genotyping rol-6

rol-6 ( su1006 ) Fwd: CTG AAA ATT TCC AGA TGA CCC TAA CTA CGG

rol-6 ( su1006 ) Rev: GAA TGG ACC ATC TGG GAA TCC ACC

AatII (New England Biolabs # R3162S) cuts N2 (421 and 31 bp) and rol-6 ( su1006 ) (242, 179 and 31 bp).

1kb plus DNA ladder (New England Biolabs).

Undigested PCR product was sent for Sanger sequencing with Quintara Biosciences.

CRISPR-Cas9 reagents

Software

ImageJ (NIH) (Schindelin et al., 2012); Prism 10 (Graphpad). Figure was assembled using Adobe Illustrator.