Description

Transgenic animals are powerful tools in the study of basic biological processes using C. elegans. The recent development of recombination-mediated cassette exchange (RMCE) integration approaches in worms provides a relatively rapid efficient method to create single copy transgenes with inserts up to at least 12 kb (Nonet, 2020,Yang et al., 2021). RMCE is similar to MosSCI (Frokjaer-Jensen et al., 2008) in that it depends on integrating at landing sites that have been engineered. In the case of MosSCI, a transposon is excised, and the double stranded break is repaired by a template often using synthesis-dependent strand annealing (SDSA) which is error prone (Frokjaer-Jensen et al., 2008). Similar approaches mediated by cas9 cleavage also have high error rates, approaching 65% in some studies (Au et al., 2019). By contrast, RMCE uses recombination to insert the template into the genome which rarely yield erroneous inserts (Nonet, 2020).

The RMCE approach I developed takes advantage of two distinct recombinases. A plasmid template delivered into the gonad of young adult animals is first integrated into the genome using FLP recombinase. Both the template and the landing site contain two distinct FLP integration sites, FRT and FRT3. Recombination between the FRT and FRT3 sites in the plasmid and the landing site yields replacement of the genomic FRT FRT3 interval with plasmid FRT FRT3 interval. This recombination likely occurs in two steps; first a loop in by recombination at one of the sites, followed by excision by recombination at the other. These recombination events typically occur in the F1 germline, though occasionally it occurs in the P0 animal (Nonet, 2020).

Two methods have been developed for performing RMCE. The first method uses a landing site and an unlinked source of FLP recombinase (usually bqSi711). Injection of the plasmid leads to integration of the plasmid at the landing site which is identified as a Rol (or Hyg^R) animal. After the initial insertion is made homozygous, the self-excising marker cassette (SEC) is then excised using a heat shock Cre protocol, leading to the final insertion. The insertion is then outcrossed from bqSi711 using simple crosses. The second method utilizes a landing site which contains a germline FLP expressing transcription unit contained within the landing site. In this case, the FLP expression element is excised using a heat shock Cre protocol. One limitation of the RMCE approach is the lack of landing sites for integration. Here, I describe several new landing sites created using a CRISPR integration approach.

I integrated a two-component landing site using CRISPR/cas9 just adjacent to the position of the ttTi5605 Mos1 insertion that has been widely used for creating single copy insertions using MosSCI (Frokjaer-Jensen et al., 2008). I also integrated a single-component landing site at same position on Chr II and another at a site adjacent to cxTi10882, another commonly used Mos1 insertion site on Chr IV. In addition, I modified the previously described two-component landing site jsTi1453 I, deleting the left miniMosarm from that landing site (Fig. 1A, B). I first characterized these novel landing sites by integrating the identical mec-4p GFP-C1 tbb-2 3’ construct at each site and comparing the expression level of insertions at the new landing sites to identical insertions at the previously described landing sites (Nonet, 2020). The expression of GFP-C1 in touch receptor neurons (TRN) was easily detected in all the new landing site transgenes, though the expression level was slightly lower than that observed in integrations at the four previously described sites (Fig. 1C).

Recently developed bipartite reporters including a tetO/tetR Tet OFF system exhibit background expression in both the pharynx and the rectal gland cells (Nonet, 2020). Because a similar rectal gland background signal was observed at several different landing sites using distinct bipartite systems, I speculated that the miniMos transposon arm might be contributing to the background. Comparison of the identical tetO 7X ∆mec-7p GFP-C1 reporter integrated at both jsTi1453 and the js1570 ∆mosL arm derivative confirmed this was the case as expression in the rectal gland cells was undetectable in the js1570 derived transgene (Fig 1D). Despite the reduction in background, the tetO reporter still robustly expressed GFP in TRNs, when driven by the identical mec-4 promoter tet OFF driver (Fig. 1C).

In developing a new recombination-mediated homolog exchange technique (https://sites.wustl.edu/nonetlab/rmhe/), I have used these new landing sites to create addition RMCE insertions. I collated the insertion frequency data from a set of over 90 injection sessions in which I counted the number of Roller F1 animals obtained, and the number of insertions obtained (Fig. 1E). These data demonstrate that js1570, jsSi1579 and jsSi1669 all behave comparably to previously described landing sites, yielding insertions at a rate of approximately 1 per 3 injected animals. However, the jsSi1691 single component site yielded insertions at a lower frequency (1 per 10 injected animals). I speculate this is due to lower expression of FLP from the jsSi1691 landing site since integration at the same position using jsSi1579 and bqSi711 as a source of FLP yields normal integration frequencies. A recent study indicated that for CRISPR/cas9 genome modifications one could obtain a much higher frequency of integration events using specifically treated miniprep DNA (Huang et al., 2021). A preliminary set of 7 injections performed while this manuscript was under review and presented in Figure 1E suggest that similar benefits are seen for RMCE, raising the insertion frequency to above 1 integration per P0 in this admittedly small sample size.

I previously demonstrated that RMCE yields insertions of expected structure in greater than 95% of cases (Nonet, 2020). I have also characterized most of the transgenes obtained at these additional landing sites either by confirming the presence of an expected fluorescence pattern, recombinase activity, tetO reporter activity, or tetR driver activity. In some cases, long range PCR combined with restriction digests and/or sequencing was also performed to confirm the structure of insertions. All of the insertions at js1570 and jsSi1669 have been verified. The analysis of insertions on Chr II is ongoing, but to date only one of over 50 well-characterized insertions is incorrect with that insertion containing a distinct region of Chr II inserted adjacent to the loxP site. Three other cases of unexpected outcomes are also worth mentioning. In one case, I obtained a homozygous Rol insertion which still expressed the rpl-28p GFP-his-58 marker from the landing cassette. In addition, in two cases, I was unable to excise the SEC by heat shock. However, in all three cases, an independent sister insertion from the same injection session was used for the excision step to isolate the final insertion. I have not attempted to determine the molecular structure of these unusual ‘faulty’ insertions since they are easily identified and discarded.

The new landing site are now available at the CGC and should provide additional flexibility in creating RMCE-based transgenic animals. I also plan to create additional landing sites at well-characterized high expressing genomic positions on the remaining chromosomes that currently do not contain landing sites.