Skip to main content
NCBI home page
Genetics logoLink to Genetics
. 2021 Nov 15;220(2):iyab206. doi: 10.1093/genetics/iyab206

phiC31 integrase for recombination-mediated single-copy insertion and genome manipulation in Caenorhabditis elegans

Fang-Jung Yang 1,†, Chiao-Nung Chen 2,†, Tiffany Chang 1,†, Ting-Wei Cheng 3, Ni-Chen Chang 1,‡, Chia-Yi Kao 1,§, Chih-Chi Lee 1,††, Yu-Ching Huang 1,‡‡, Jung-Chen Hsu 1, Jengyi Li 1, Meiyeh J Lu 1, Shih-Peng Chan 2,3,✉, John Wang 1,✉
Editor: B Goldstein
PMCID: PMC9208643  PMID: 34791215

Abstract

Caenorhabditis elegans benefits from a large set of tools for genome manipulation. Yet, the precise single-copy insertion of very large DNA constructs (>10 kb) and the generation of inversions are still challenging. Here, we adapted the phiC31 integrase system for C. elegans. We generated an integrated phiC31 integrase expressing strain flanked by attP sites that serves as a landing pad for integration of transgenes by recombination-mediated cassette exchange (RCME). This strain is unc-119(−) so RMCE integrants can be produced simply by injection of a plasmid carrying attB sites flanking unc-119(+) and the gene(s) of interest. Additionally, phiC31 integrase is removed concomitantly with integration, eliminating the need to outcross away the integrase. Integrations were obtained for insert sizes up to ∼33.4 kb. Taking advantage of this integration method we establish a dual-color fluorescent operon reporter system able to study post-transcriptional regulation of mRNA. Last, we show that large chromosomal segments can be inverted using phiC31 integrase. Thus, the phiC31 integrase system should be a useful addition to the C. elegans toolkit.

Keywords: phiC31 integrase, recombination-mediated cassette exchange, integration, dual-fluorescent operon reporter, inversion

Introduction

Genomic manipulations are a cornerstone of modern experimental biology. Molecular genetic analysis relies on the ability to create precise mutations, transgenes, and genomic rearrangements such as inversions or translocations. Caenorhabditis elegans benefits from a large set of genetic tools to do this (Stinchcomb et al. 1985; Mello et al. 1991; Way et al. 1991; Praitis et al. 2001; Frøkjaer-Jensen et al. 2008, 2014;Dickinson and Goldstein 2016; Iwata et al. 2016; Dejima et al. 2018; Nance and Frøkjaer-Jensen 2019; Nonet 2020).

In some cases, precise single-copy insertions, possibly of large size, are necessary or desired in C. elegans. Single- or low-copy insertions can be created using biolistic bombardment (Praitis et al. 2001) and miniMos (Frøkjaer-Jensen et al. 2014), but the insertion sites for both methods cannot be controlled. MosSCI was the first efficient method for achieving precise single-copy insertions (Frøkjaer-Jensen et al. 2008), and later CRISPR/Cas9-based methods have been shown to be quite effective (Dickinson et al. 2013, 2015). These methods, along with recent advancements now enable the straightforward insertion of transgenes up to about 10 kb (Philip et al. 2019; Fan et al. 2020). Longer insertions of 12 kb (Das et al. 2015) and 16 kb (Frøkjaer-Jensen et al. 2012) have been reported, but less frequently. Recently, an Flp/FRT system has been developed that can insert transgenes of sizes up to ∼14 kb range relatively easily through recombination-mediated cassette exchange (RMCE) (Nonet 2020). While inserts of greater sizes may be possible, this has not been tested. Additionally, because Flp recombinase recombines two identical FRT sites, the products are also FRT sites. Consequently, an inserted sequence may be excised. phiC31 integrase is a phage-derived site-specific recombinase that has been used for genomic manipulations in E. coli (Merrick et al. 2018), yeast (Xu and Brown 2016), plants (Thomson et al. 2010; Bernabe-Orts et al. 2020; Cody et al. 2020), flies (Groth et al. 2004; Bateman et al. 2006; Venken et al. 2006), fish (Hu et al. 2011; Lu et al. 2011; Kirchmaier et al. 2013), mouse (Belteki et al. 2003), and humans (Groth et al. 2000). phiC31 integrase recombines an attB and an attP site to create an attR and an attL site. Recombination is a unidirectional reaction, so excision of integrated DNA is not a concern. phiC31 integrase has been used for “simple” insertion of donor constructs, for example by recombination of a single attB site on a plasmid with a single attP landing site in the genome, and for RCME between a pair of oppositely oriented attB sites on a plasmid with a pair of oppositely oriented attP sites in the genome. In addition, phiC31 can be used to delete DNA segments that are flanked by an attB and an attP site in the same orientation or invert DNA segments if the two sites are in an opposite orientation. Despite its potential utility, limited tests to establish a phiC31 integrase system have been unsuccessful or partially successful in C. elegans (Vargas 2012; Frøkjaer-Jensen et al. 2014).

We had several motivations for developing the phiC31 integrase system in C. elegans. First, we were having difficulties inserting moderately large sequences (∼5–15 kb) with the available techniques. Larger constructs also allow for the conducting of more complex transgenic experiments, such as multigene constructs, especially those with longer genes or longer regulatory regions. Another reason was to be able to create precise very large insertions (>10 kb). This capability would allow better characterization of a case of non-Mendelian inheritance of chromosomes in C. elegans (and Caenorhabditis species), whereby fathers that are heterozygous for chromosomes with a length difference, transmit longer chromosomes preferentially to sons (Wang et al. 2010; Le et al. 2017). Additionally, an efficient phiC31 integrase system could allow the testing of multiple transgene variants at the same genomic context faster. A final motivation is synthetic biology. phiC31 and other site-specific recombinases have been used to invert or delete DNA segments as both components and outputs in synthetic recorders and logic circuits (Merrick et al. 2018). The vast majority of synthetic biology experiments have been done in single cellular organisms, but could be applied to animals.

In this study, we adapted the phiC31 integrase system for C. elegans. We first used MosSCI to insert an RMCE “landing pad” strain that also expresses phiC31 integrase and demonstrated that we could generate precise RMCE insertions. Then, we improved the landing pad strain such that RMCE integrants can be screened by phenotypic rescue of Unc into wild-type (WT) worms. Using this improved strain, we were able to create insertions of very large constructs, including one of ∼33.4 kb. Taking advantage of this we establish a dual-color fluorescent operon reporter system able to study post-transcriptional regulation of mRNA. Last, we show that large chromosomal segments can be inverted using phiC31 integrase. Thus, the phiC31 integrase system should be a useful addition to the C. elegans toolkit.

Materials and methods

Additional details for some methods are in Supplementary File S1.

Caenorhabditis elegans strains and maintenance

Supplementary Table S1 lists the strains used and generated in this study. All worms were grown on NGM plates seeded with Escherichia coli strain OP50. Strains were maintained at 20°C in an incubator or at room temperature (∼22–23°C), except for strains with let-7 mutations, which were grown at 15°C.

Generation of the phiC31 integrase expressing landing pad strain

A plasmid containing the landing pad construct [fosmid6-1(−)+unc+phiC31_43.1+delete-rol6_122GFP; Supplementary File S2] was cloned using the 4 fragment MultiSite Gateway Pro Plus system (ThermoFisher, cat no. 12537100). The first donor fragment consists of a core 40-bp attP site (within a 221-bp segment) and Caenorhabditis briggsae unc-119(+). The second fragment contained the germline glh-2 promoter, phiC31 integrase, and the glh-2 3′-UTR. phiC31 was codon optimized (https://worm.mpi-cbg.de/codons/cgi-bin/optimize.py) (Redemann et al. 2011) and has three artificial introns and an SV40 NLS on the C-terminus. The third fragment is composed of a partial rol-6 promoter. Originally the dominant rol-6(su1006) allele was intended both as a marker for MosSCI integration and as a negative selection marker for successful RMCE. However, we failed to obtain such a Rol strain through MosSCI, so a truncated donor fragment was made instead to be compatible with the 4-fragment Gateway cloning. The fourth fragment contained myo-2p::GFP and an attP site (same as above). The attP sites are arranged in “opposite” orientation in the landing pad construct.

MosSCI (Frøkjaer-Jensen et al. 2008) was used to integrate the phiC31 integrase construct at the ttTi5605 site on LG II. After screening >10,000 P0 equivalents, we obtained one integrant [BRC0546(antIs30)], which we verified by a combination of PCR assays and sequencing of amplification products. Expression and splicing of phiC31 were also confirmed by RT-PCR.

To obtain an Unc-119 strain, the Cbr-unc-119(+) gene within the antIs30 landing pad was mutated using CRISPR (Cho et al. 2013; Paix et al. 2015). Three Unc-119 lines were obtained from individuals injected with Cas9 protein and using a pool of three unc-119 sgRNAs. One allele had a 691-bp deletion [Cbr-unc-119(ant40)]. This mutated landing pad transgene was named antIs31 and then backcrossed to N2 4 times, each time retaining Unc-119 individuals, to yield BRC0566, the main landing pad strain used for injections.

RMCE plasmids

Plasmid names, injection concentrations, and descriptions are in Supplementary Table S2. For the initial proof-of-principle RMCE experiments plasmid pBRC_double_attB_GFP, containing sur-5p::GFP flanked by oppositely oriented attB sites, was constructed using conventional cloning. To facilitate the RMCE experiments using the Unc-119 strain (BRC0566), two general plasmids were constructed. The first plasmid, pBRC_double_attB_GFP_donor, contains a GFP marker to facilitate visually following the insertion; it carries oppositely oriented attB sites flanking Cbr-unc-119(+), sur-5p::GFP and the R1-R2 Gateway cloning sites. The second plasmid, pCG150_double_phiC31_attB, lacks fluorescent markers and carries oppositely oriented attB sites flanking a C. elegans unc-119 minimal rescue fragment and the R3-R4 Gateway cloning sites (Supplementary Figure S1). Conventional restriction site-mediated cloning is also possible for both plasmids. Plasmid pBRC_double_attB_GFP_mCherry was generated by Gateway cloning into pBRC_double_attB_GFP_donor. The operon reporters (Supplementary Figure S2) carrying two fluorescence colors were constructed by conventional cloning into pCG150_double_phiC31_attB.

For the bacterial artificial chromosome (BAC) insertion experiments, the fire ant BAC 47D10 was first retrofitted with a cassette containing Cbr-unc-119(+), the kanamycin resistance gene, and two “outward” facing attB sites flanking sur-5p::GFP, to yield attB_retrofitted_Sinv47D10 using recombineering (Quick & Easy BAC Modification Kit, Gene Bridges). RMCE of this BAC would insert ∼137 kb (∼125.4 kb of fire ant sequence and ∼11.8 kb of the retrofitted sequence and vector) into the phiC31 landing pad. A shortened BAC containing ∼20.7 kb of fire ant sequence, attB_retrofitted_Sinv47D10_short, was generated by recombineering-mediated deletion. RMCE of this shortened BAC would insert ∼33.4 kb. Additional details in Supplementary File S1.

Plasmids were isolated using the Mini Plus Plasmid DNA Extraction System (Viogene, cat no. GF2002); BACs were isolated using the Qiagen Genomic-tip 20/G (Qiagen, cat no. 10223).

RMCE experiments

For the initial landing pad strain BRC0546 (antIs30), adult (non-Unc) hermaphrodites were injected using standard methods (Evans 2006) with the donor cassette (sur-5p::GFP) and the coinjection marker (myo-2p::RFP) plasmids. Then, transgenic F1 were singled out based on GFP fluorescence. After, F2 and subsequent progeny were screened for Unc, Green, and non-Red individuals, who were putatively homozygous for the sur-5p::GFP RMCE insert.

For the improved landing pad strain BRC0566 (antIs31), adult (Unc) hermaphrodites were injected as above. Then, transgenic non-Unc F1s were singled out. Next, F2 and subsequent progenies were screened for non-Unc individuals that segregated all non-Unc individuals, and hence were putatively homozygous lines. Screening for homozygous lines was attempted up to the F6 generation if approximately three-quarters of the progeny were non-Unc at each generation. Otherwise, the transgene was presumed to be an extrachromosomal array and screening was stopped. For candidate homozygous lines, the GFP and mCherry fluorescence patterns were used to help ascertain if the RMCE insertion was likely correct. Specifically, there should be a loss of the myo-2p::GFP within the landing pad and the coinjection marker mCherry as well as gain of any donor construct GFPs and/or mCherry.

PCR and sequencing assays were conducted to validate the RMCE insertions for all potentially correct insertion strains. First, a set of four “junction” PCR assays using primers flanking the landing pad and primers on the construct near the edge of insert were used to confirm both integration and orientation. Positive PCR products were also sequenced to verify that recombination produced two proper attR sites. Second, for the initial landing pad strain, two PCR assays were used to test if the Cbr-unc-119 and phiC31 integrase genes were lost. Third, for the dual-cassette experiment, a series of PCR assays were designed spanning various subparts across the whole insert. Finally, in some cases, long range PCR assays across the entire insert were conducted. All primers used are listed in Supplementary Table S3.

Dual-color operon reporter experiments

Four operon reporters carrying two fluorescence colors were injected into BRC0566 to obtain single-copy integrants (Supplementary Figure S2). The seam cell-specific operon reporter plasmids, SCMp::GFP(PEST)::H2B::lin-41_3′UTR::SL2::mCherry::H2B::unc-54_3′UTR and SCMp::GFP(PEST)::H2B::lin-41_3′UTR_ΔLCS::SL2::mCherry::H2B::unc-54_3′UTR, contain the seam cell-specific promoter expressing an operon containing GFP and mCherry, both as histone chimeras. GFP is also fused to a PEST degron and contains the lin-41 3′-UTR or lin-41 3′-UTRΔLCS (LCS, let-7 complementary site) (Rausch et al. 2015). An SL2 sequence, containing the gpd-2/gpd-3 intergenic region (Merritt et al. 2008), precedes mCherry which contains the unc-54 3′-UTR. The ubiquitously expressed reporter plasmids, dpy-30p::GFP(PEST)::H2B::lin-41_3′UTR::SL2::mCherry::H2B::unc-54_3′UTR and dpy-30p::GFP(PEST)::H2B::lin-41_3′UTR_ΔLCS::SL2::mCherry::H2B::unc-54_3′UTR, were similarly constructed except with the dpy-30 promoter. These operon reporters were cloned into pCG150_double_phiC31_attB for injection into BRC0566 with coinjection marker egl-20p::mCherry. The seam cell-specific construct was also crossed into the let-7(n2853) hypomorphic mutant to test the post-transcriptional regulatory role of let-7 on the lin-41 3′-UTR. For validation of SCMp::GFP(PEST)::H2B::lin-41_3′UTR::SL2::mCherry::H2B::unc-54_3′UTR and dpy-30p::GFP(PEST)::H2B::lin-41_3′UTR::SL2::mCherry::H2B::unc-54_3′UTR, we conducted a double “PCR walking” assay: one primer was fixed to the left- or right-flanking region of ttTi5605 and other primers were staggered approximately every 1.5 kb along the transgene (Supplementary Figure S3). For the ΔLCS transgenes, we verified using long range and junction PCR assays as well as by sequencing the PCR product that spanned the ΔLCS region of the 3′-UTR. After validation of the insertion lines, images were taken of the worms and then GFP and mCherry expression levels were quantified.

For microscopy assays, synchronized arrested L1 larvae of animals carrying a dual-color reporter in the WT or let-7(n2853) background were plated on NGM agar plates with OP50 bacteria and incubated at 20°C and then imaged at the L2 and mid L4 stages (18 and 40 h after plating). Images were acquired in the green, red, and transmitted light channels (with differential interference contrast) using a 63× Zeiss oil immersion objective on a Zeiss fluorescence microscope (Axio Imager.M2) and Axiovision 4.8 software. For data analysis, worms were selected based on visual inspection of gonad length and vulva morphology (Mok et al. 2015). Seam cells were selected using the Fiji imaging platform (Schindelin et al., 2012). Images were denoised using a Gaussian blurred algorithm and segmented using a Yen global threshold. Signal intensity in the green channel was divided by that in the red channel for each cell; relative signal intensities were then averaged for each worm. Two to 10 seam cells in 11–14 worms per genotype were quantified. Median signal intensity and frequency of the signal intensity were calculated and graphed using GraphPad Prism software.

RMCE with a BAC

The shorter BAC attB_retrofitted_Sinv47D10_short was injected as above for the plasmids. For the longer BAC attB_retrofitted_Sinv47D10, approximately three-quarters of the P0s (n = 319) were injected, as above, while in the remaining injections (n = 110), 70 µM spermidine and 30 µM spermine were added and the coinjection marker was omitted. In addition to the junction, Cbr-unc-119, and phiC31 integrase PCR assays, screening of BAC integrants also included PCR assays for the gain of the bacterial kanamycin resistance (kanR) gene, the loss of myo-2p::GFP, and five (attB_retrofitted_Sinv47D10_short) or seven (attB_retrofitted_Sinv47D10) ∼500-bp fragments that were spaced roughly evenly across the BACs. Most PCR assays were also conducted on the known incorrect BAC insertion lines, i.e., those with GFP and/or mCherry expression. Southern blot analysis was conducted according to standard protocol with random primer labeling using 32P-dCTP. Primers used for the BAC validation and Southern blot assays are in Supplementary Table S3. Oxford Nanopore Technologies (ONT) sequencing was used to test for the proper insertions for one candidate strain for each of the BACs (details in Supplementary File S1).

Generation of chromosomal inversion on LG IV

Co-CRISPR (Arribere et al. 2014; Paix et al. 2015) was used to generate one strain carrying a 40-bp core attP site (plus flanking regions; 221 bp total) inserted into the second intron of dpy-13 and another carrying a minimal 38-bp attB site in the fourth intron of unc-30. The sgRNAs, repair templates, and PCR primers used for validation along with additional sequence information are listed in Supplementary Table S3. The insertion alleles obtained were antSi50 (attB, BRC0609) in dpy-13 and antSi51 (attP, BRC0669) in unc-30; these strains are WT and have been backcrossed 3× to N2. Recombination was used to place the attB and attP insertions onto the same chromosome to generate BRC0671 (antSi50 antSi51).

The chromosomal inversion was generated as follows. First BRC0566 [antIs31; unc-119(ed9)] was outcrossed to N2 to remove unc-119(ed9) to yield BRC0664. Next, N2 males were crossed to BRC0664 and then heterozygous F1 males were crossed to BRC0671 to produce antIs31/+; antSi50 antSi51/+ + or +/+; antSi50 antSi51/+ + F2 cross progeny. myo-2p::GFP is difficult to detect in adults, so it was scored in the F3 progeny. F3 and F4 progeny were screened for Dpy Unc individuals, indicative of an inversion mediated by phiC31 integrase. Inversions were confirmed in the independently derived Dpy Unc lines by PCR and sequencing of the junctions.

Recombination rates between dpy-13 and unc-30 in the inversion allele, antIn1, and in the WT orientation [dpy-13(e184) unc-30(ok613)] were determined from the recombinant offspring of the dpy-13 unc-30/+ + heterozygotes. The recombination rate was calculated based on the fraction of non-Dpy non-Unc F2 that segregated non-Dpy Unc or Dpy non-Unc and the Punnett Square expectations for recombinants (see Supplementary File S1).

Results

RMCE proof-of-principle

To adapt the phiC31 recombination system for C. elegans, we created a strain possessing a landing pad for RMCE consisting of inverted attP sites that surround a GFP marker and a germline expressed phiC31 integrase (Figure 1A). This strategy was chosen for the following reasons. First, the subsequent injections of the desired sequences (or constructs) for RMCE would be simplified because coinjection of phiC31 integrase (as DNA, mRNA, or protein) would not be needed. Second, because phiC31 integrase protein would already be present in the strain, integration may be more efficient. Third, by placing the phiC31 integrase gene between the attP sites, RMCE would swap out the phiC31 integrase from the chromosome, eliminating the need to conduct crosses to remove the integrase. Finally, the GFP marker would permit visually tracking the presence of the landing pad in crosses or its absence after RMCE.

Figure 1.

Figure 1

Open in a new tab

phiC31 integrase-mediated RMCE in C. elegans. (A) Schematic of single-copy transgene integration via RMCE into the antIs31 landing pad. (B) Strains used for RMCE. Both BRC0546 and BRC0566 express phiC31 integrase in the germline. BRC0566 is derived from BRC0546 after mutating the Cbr-unc-119 gene. GFP expression in the pharynx from the myo-2 promoter is evident; weak expression in other parts of the body is also seen. (C) Images of two representative strains after successful RMCE. BRC0580 carries a single-copy integrant of sur-5p::GFP; BRC0790 carries a single-copy integrant of sur-5p::GFP and myo-2p::mCherry. Locations of GFP and mCherry expression match the expected pattern. (D) Validation of proper integrations with PCR assays for four representative strains. Orange PCR assays indicate test for proper recombination junctions and orientation. Integrants can also be in the reverse orientation, and those PCR assays are not depicted here. Blue indicates long PCR assays across the entire insert and demonstrate single-copy insertion. Arrows next to the gels indicate the correct long PCR product; dots are PCR artifacts caused by probable intramolecular annealing of the attR sites.

We generated this phiC31 integrase expressing landing pad strain using the MosSCI technique to target a construct containing the desired elements and Cbr-unc-119(+) into the ttTi5605 Mos1 insertion site on LG II (Frøkjaer-Jensen et al. 2008). This site was chosen because it is permissive for germline expression of transgenes (Frøkjaer-Jensen et al. 2008). After screening the progeny from approximately 10,000 young adult worms subjected to MosSCI, we succeeded in obtaining one correct integrant [Figure 1B; BRC0546; antIs30 [attP-phiC31-f Cbr-unc-119(+) glh-2p::phiC31 myo-2p::GFP attP-phiC31-r]; unc-119(ed9)]. We further verified that phiC31 integrase was expressed and properly spliced (Supplementary Figure S4).

Next, we tested whether the antIs30 phiC31 integrase expressing landing pad could be used for RMCE. We created a donor plasmid containing a globally expressed green fluorescent protein gene (sur-5p::GFP) that is flanked with inverted attB sites; this sequence to be inserted is 2559 bp. Recombination between both pairs of attP/attB sites would swap in sur-5p::GFP and swap out phiC31 integrase, Cbr-unc-119(+), and myo-2p::GFP. Compared with the original strain, homozygous recombinant individuals would have a switch in GFP expression location from the pharynx to the whole body and also from WT to the Unc phenotype.

We injected this donor plasmid and a coinjection marker (myo-2p::mCherry) into 26 antIs30; unc-119(ed9) P0s and then singled out all 28 WT F1 individuals that were transgenic for the donor plasmid based on their expression of sur-5p::GFP. Three F1s continued segregating F2 individuals expressing sur-5p::GFP (Table 1). In one case the F2 brood was composed of mostly WT animals and many Unc animals putatively homozygous for the donor construct and lacking mCherry expression (i.e., Unc, Sur-5p::GFP, and non-Myo2p::GFP), consistent with the F1 mother being heterozygous for the donor construct. In the second case, all the F2 were WT, with some expressing sur-5p::GFP and others not. We singled out nine sur-5p::GFP F2 individuals and found that all segregated many Unc, Sur-5p::GFP, and non-Myo2p::GFP F3 progeny, indicating that these F2 were likely heterozygous for the donor construct. The last F1 individual segregated F2 and subsequent F3 in a pattern that appeared consistent with only an extrachromosomal array (i.e., no integration) and was not examined further.

Table 1.

Summary of phiC31-mediated RMCE experiments

Experimenta Transgene insert size (kb) P0 injected Transgenic F1 Candidate linesb Correct RMCE linesc Success rated
pBRC_double_attB_GFPe,f 2.6 26 28 2 2 7.1%
pBRC_double_attB_GFP_donorf 6.4 165 104 7 5 4.8%
pBRC_double_attB_GFP_mCherryg (total) 7.4 238 207 26 24 11.6%
 Trial 1 180 80 4 2 2.5%
 Trial 2 47 112 19 19 17.0%
 Trial 3 11 15 3 3 20.0%
pDonor_double_phiC31_attB_SCMp::GFP(PEST)::H2B::lin-41_3′UTR:: SL2::mCherry::H2B::unc-54_3′UTRh 14.9 21 19 6j 6 ≥31.6%
pDonor_double_phiC31_attB_SCMp::GFP(PEST)::H2B::lin-41_3′UTRΔLCS::SL2::mCherry::H2B::unc-54_3′UTRh 14.8 13 30i 26 26 86.7%
pDonor_double_phiC31_attB_dpy-30p::GFP(PEST)::H2B::lin-41_3′UTR::SL2::mCherry::H2B::unc-54_3′UTRh 8.9 15 30i 9j 9 ≥30%
pDonor_double_phiC31_attB_dpy-30p::GFP(PEST)::H2B::lin-41_3′UTRΔLCS::SL2::mCherry::H2B::unc-54_3′UTRh 8.8 36 51 9j 9 ≥17.6%
Subtotal plasmids 514 469 85 81 17.3%
attB_retrofitted_Si47D10_short (fire ant BAC)g 33.4 111 87 6 1k,l (4k) 1.1% (4.6%)
attB_retrofitted_Si47D10 (fire ant BAC)m 137.2 429 138 3 0k,l 0%
Subtotal BACs 540 225 9 1 (4) 0.4% (1.8%)
Open in a new tab
a

Plasmids injected into BRC0566 (antIs31 [attP-f Cbr-unc-119(ant40) glh-2p::phiC31 myo-2p::GFP attP-r] II; unc-119(ed9) III) resulting in rescue of unc-119 mutants into WT, unless indicated otherwise.

b

Based on expected GFP and mCherry phenotypes.

c

Based on PCR for the recombination junctions, unless indicated otherwise.

d

Percent of transgenic F1s that yielded correct insertions.

e

In this experiment, the plasmid was injected into BRC0546 (antIs30 [attP-f Cbr-unc-119(+) glh-2p::phiC31 myo-2p::GFP attP-r] II; unc-119(ed9) III) and homozygous RMCE integrants would have an Unc phenotype.

f

Coinjection marker: plasmid carrying myo-2p::mCherry.

g

Coinjection marker: plasmid carrying myo-3p::mCherry.

h

Coinjection marker: plasmid carrying egl-20p::mCherry.

i

First 30 non-Unc F1s were picked and the rest were ignored.

j

Number of integrated lines followed, although more were present.

k

Based also on PCR assays distributed across inserted BAC sequence.

l

For 1 candidate, also Oxford Nanopore sequencing and Southern blot assay.

m

Coinjection marker: plasmid carrying myo-3p::mCherry or none (see Materials and methods).

To confirm that the Unc worms were the product of phiC31-mediated RMCE at the landing pad, we conducted molecular validation experiments on the two independent strains. PCR assays for the two recombination junctions and the presence of phiC31 integrase as well as with flanking primers across the landing pad were consistent with the donor sur-5p::GFP construct replacing the phiC31 integrase at the landing pad (Supplementary Figure S5). Additionally, sequencing of the junction PCR products verified that both recombination junctions were attR sites, the outcome of attP × attB recombination. Together these results demonstrate that phiC31-mediated RMCE is possible in C. elegans, can occur at relatively high frequency [two out of two integrated transgenic lines; 7% (2 of 28) of transgenic F1], and integration can occur in the F1 or the F2 generations.

Improving and facilitating RMCE

Although the strain BRC0546 [antIs30; unc-119(ed9)] can be used for RMCE, identifying recombinants is demanding because it requires screening for Unc individuals in the context of WT individuals. Screening for WT individuals rescued for the Unc phenotype would be easier. Therefore, we used CRISPR/Cas9 to mutate the Cbr-unc-119 gene within the landing pad, yielding antIs31[attp-phiC31-f; Cbr-unc-119(ant40); glh-2p::phiC31; myo-2p::GFP; attp-phiC31-r] (Figure 1B). We then backcrossed this strain four times to N2 to produce BRC0566 (antIs31 ant40 II; unc-119(ed9) III).

To facilitate generating donor constructs for RMCE, we developed two plasmids (Supplementary Table S2). The first one (pBRC_double_attB_GFP_donor) has sur-5p::GFP, which permits visual following of the recombinant transgene in subsequent crosses, while the second one (pCG150_double_phiC31_attB) does not, which permits experiments where tracking of custom fluorescent signals is desired. For both plasmids, the gene of interest can be added into these plasmids by standard or Gateway cloning. Finally, both plasmids have unc-119(+) (C. briggsae or C. elegans) for screening.

We next used these plasmids to confirm that RMCE in BRC0566 works (Table 1). We injected the pBRC_double_attB_GFP_donor plasmid, which would insert ∼6.4 kb at the landing pad, along with an mCherry coinjection marker into BRC0566. We obtained 104 transgenic F1 that gave rise to seven putatively integrated lines having the predicted recombinant phenotype (WT, Sur-5p::GFP, and non-mCherry; Figure 1C). Of these, five were validated to be correct (Figure 1D) corresponding to a success rate of 4.8% (based on F1s picked). Because the pairs of attP sites in the genome and of attB sites on the donor construct are identical, the donor construct can integrate in two orientations, which we call orientation 1 and 2. We obtained two and three such lines, respectively. We also injected a plasmid where Gateway cloning was used to insert myo-2p::mCherry into pBRC_double_attB_GFP_donor, creating a construct with an insert size of ∼7.4 kb. In a series of three experiments, we obtained 24 successfully integrated lines with success rates ranging from 2.5% to 20% (Figure 1, C and D; Table 1). Finally, we used conventional cloning to insert various two gene operons into pCG150_double_phiC31_attB to create donor inserts of ∼9 to ∼15 kb. These operon constructs were designed as a proof-of-concept platform for dual expression (more below). Again, we easily obtained proper insertions that were molecularly validated (Figure 2 and Table 1; Supplementary Figure S3). Together these results demonstrated that phiC31-mediated RCME can be used in this improved strain and that larger insertions up to ∼15 kb can be generated relatively easily.

Figure 2.

Figure 2

Open in a new tab

Measuring the activity of let-7 using a dual-color reporter. (A, H) Schematic of the dual-color reporters used to monitor miRNA activity in C. elegans. Transcription of the single-copy integrated reporter is driven by (A) a seam cell specific promoter (SCMp) or (H) the ubiquitously active dpy-30 promoter. The GFP reporter is fused to the lin-41 3′-UTR (as shown in Supplementary Figure S2), which contains two conserved let-7 complementary sites (LCSs) that are regulated by the let-7 miRNA, while mCherry is fused to the unc-54 3′-UTR and serves as an internal control. (B, I) The expression of dual-color reporters with (B) the SCMp or (I) the dpy-30 promoter at the mid-L4 stage in WT C. elegans. (C, D) Repression of an SCMp driven let-7 target reporter (i.e., intact lin-41 3′-UTR; chsIs001) in mid L4 worms depends on let-7(+) (SPN369) and is lost in let-7(n2853) mutants (SPN391). (E, F) an SCMp driven lin-41 3′-UTR reporter without the let-7 binding sites (i.e., ΔLCS; chsIs002) in mid-L4 worms is regulated neither in let-7 WT (SPN377) nor in let-7(n2853) mutant (SPN393) individuals. (C–F) L2-staged animals serve as let-7 negative controls. Seam cells are marked with arrowheads and hyp7 nuclei are marked with arrows. (G) Quantification of reporter expression in worms. Violin plots show the GFP signal intensity divided by the mCherry intensity for each worm per condition. Two to 10 seam cells were quantified per worm. The numbers of worms examined for each condition are indicated on top. Horizontal line and dotted lines indicate median value and first/third quartiles, respectively. *P < 0.05 and ****P < 0.0001, two-tailed unpaired t-test.

Single-copy dual-fluorescent operon reporter system

A two-color fluorescent reporter system has been developed to more accurately quantify post-transcriptional control by microRNAs (miRNA) at the single cell level (Ecsedi et al. 2015). In this system, two reporters are driven by the same promoter but have different 3′-UTRs. A test 3′-UTR is fused to GFP while a control, unregulated 3′-UTR is fused to mCherry. Thus, regulation of the 3′-UTR such as via miRNAs should be reflected by changes in the GFP/mCherry signal ratio. To maintain robustness of the signal, the GFP and mCherry reporters are integrated site-specifically in single copy. However, in this design two integration events, onto separate chromosomes, are required.

We took advantage of the large cargo size of the phiC31 integrase system to streamline the dual-fluorescent system and put both reporters at one genomic location. We designed an operon reporter that permits examining the heterochronic downregulation of the lin-41 3′-UTR by the let-7 miRNA in the seam cells using quantitative fluorescence microscopy. This operon reporter consisted of a seam cell specific promoter (SCMp) driving GFP(PEST)::H2B::lin-41 3′UTR and mCherry::H2B::unc-54 3′UTR (Supplementary Figure S2). The gpd-2/gpd-3 intergenic region connects the two markers and provides a SL2-specific trans-splice site for the mCherry (Huang et al. 2001). We cloned this operon reporter into plasmid pCG150_double_phiC31_attB, injected the donor construct into the Unc landing pad strain (BRC0566), and obtained at least 6 WT candidate integrants. One of them was verified for proper RMCE insertions by double “PCR walking” (see Materials and methods; Supplementary Figure S3A).

We next used one of the integrated alleles, chsIs001, to establish that this operon correctly reports let-7/lin-41 3′-UTR regulation. let-7 is upregulated during the L4 stage. Thus, we followed GFP/mCherry expression ratios during development with the expectation that the expression ratio at the L4 stage should be less than at the L2 stage. As predicted, we observed reduced GFP/mCherry expression ratios in the seam cells (Figure 2C). To confirm that let-7 is necessary for the decrease in GFP expression, we crossed the dual reporter strain into the let-7(n2853) hypomorphic mutant and found that the GFP signal at the L4 stage were not down-regulated as in WT (Figure 2D). Finally, to test the requirement of the let-7 complementary target sequences on the lin-41 3′-UTR, we generated a new phiC31-mediated RMCE operon integrant (chsIs002) with the target site deleted. Examination of this integrant revealed no reduction of the GFP/mCherry ratio at the L4 stage in both the WT or let-7(n2853) backgrounds (Figure 2, E and F). Quantitation of the GFP/mCherry ratios are shown in Figure 2G.

To expand the utility of the reporter to the whole animal, we replaced the SCM promoter with a ubiquitously expressing dpy-30 promoter (Figure 2H). We used phiC31-mediated RMCE of this construct to obtain worm lines stably expressing GFP and mCherry in almost all tissues, including the gonad (chsIs004, Figure 2I; Supplementary Figure S3C). Interestingly, the GFP/mCherry expression ratio varied along the gonadal axis, likely reflecting spatial-temporal regulation of the LIN-41 (or possibly of the UNC-54) 3′-UTR in this organ. In addition, the GFP/mCherry expression ratio differed among cell types, and was easily noticed in the intestines and some neuron cells, further suggesting cell-type specific regulation of the LIN-41 3′-UTR.

BAC integration

The ability to generate very large single-copy transgenes (10–30 kb, or even 100+ kb) at specific locations would facilitate more complex experiments. To this end, we tried to integrate a shorter (attB_retrofitted_Sinv47D10_short) and a longer (attB_retrofitted_Sinv47D10) BAC that upon RMCE would insert about 33.4 and 137.2 kb, respectively (Figure 3; Supplementary Figures S6A and S7A). These BACs contain a retrofitted cassette that carries Cbr-unc-119(+), sur-5p::GFP, and two attB sites. The attB sites are oriented such that the entire BAC (including vector sequence), except for sur-5p::GFP, would be inserted (Figure 3; Supplementary Figures S6A and S7A).

Figure 3.

Figure 3

Open in a new tab

phiC31 integrase mediated insertion of a ∼33.4-kb single-copy transgene. (A) Top, schematic for RMCE of attB_retrofitted_Sinv47D10_short BAC (BAC 47D10 retrofitted with an attB cassette and then shortened) into the antIs31 landing pad. Proper insertion would insert ∼33.4 kb, comprising ∼20.7 kb of fire ant sequence, ∼0.9 kb of ampR, and ∼11.8 kb of retrofitted vector. Relative locations of the probe (blue circle with #2 and bar) and the relevant EcoRI restriction sites for the Southern blot assay are shown. Schematic not to scale. (B) Southern blot validation. Genomic DNA from N2, BRC0566 (antIs31), and the insertion strain BRC0906 (antIs32) as well as attB_retrofitted_Sinv47D10_short BAC DNA was digested with EcoRI, and subjected to Southern blot assay. Probe #2 hybridized to single fragments of ∼4 kb on the BAC and the insertion strain; the expected fragment size is 4016 bp. This is consistent with a single-copy insertion or a low copy number insertion with no rearrangement in the probed region. (C) IGV plot of Oxford Nanopore (ONT) sequence reads indicates that antIs32 carries a correct single-copy insertion. ① Blue bar shows the extent of the ideal single-copy insertion. ② Subsections of the ideal insertion as labeled. ③ Gray is coverage depth which is similar across the insert and the flanking region, consistent with a single-copy insertion. Colored vertical bars are potential mutational differences between the predicted BRC0906 genome and the sequence reads. The sequence for the attB_retrofitted_Sinv47D10_short BAC was assembled and polished from ONT reads only, so some potential mutations could be assembly errors in the BAC. ④ Map positions of individual reads. Reads cross junctions and the insert cleanly without internal breaks in alignment. This is another indication of a correct insertion. See Supplementary Figure S6 for additional PCR validation results and Southern blot assays.

To test RMCE of the shorter BAC, we injected it along with an mCherry coinjection marker into the landing pad strain and obtained 87 F1 non-Unc individuals. Of these, six gave rise to potentially correct integrated lines based on segregating exclusively non-Unc, non-GFP, and non-mCherry progeny (Table 1). Based on PCR assays, four had proper junctions as well as the predicted pattern of presence and absence of markers on the BAC or landing pad, consistent with phiC31-mediated RMCE (Supplementary Figure S6, A and B). To further confirm proper integration, we conducted Southern analysis on one of the insertion alleles, antIs32. All three probes yielded results consistent with single-copy insertion (Figure 3; Supplementary Figure S6C). Nevertheless, the Southern blot assay cannot exclude low copy insertions with no rearrangements in the probed region, therefore we sequenced this strain to about 60× coverage using the ONT platform. Inspection of the reads revealed uniform mapping coverage across the antIs32 insertion allele. Additionally, the reads crossed the attR junctions cleanly and those within the insertion mapped without internal breaks in alignment. Together, the sequencing results clearly show a single-copy correct insertion of the BAC and demonstrate that phiC31 can be used to cleanly integrate constructs up to ∼33.4 kb.

We next attempted to integrate ∼137.2 kb using the longer BAC. After injection of this BAC with and without an mCherry coinjection marker into the landing pad strain, we obtained 138 F1 non-Unc individuals that produced 22 integrated lines. However, only three lines were non-Unc, non-GFP, and non-mCherry, which would be expected of potential RMCE integrants (additional details in Supplementary File S1). Moreover, none were proper, clean single-copy insertions based on additional PCR assays, and for one strain, Southern analysis and ONT sequencing (Supplementary Figure S7).

Generating large inversions

phiC31 integrase can be used to invert sequences that are flanked by oppositely oriented attP and attB sites (Merrick et al. 2018). To test if it is possible to generate a very large inversion in C. elegans we first created a strain carrying the attP and attB sites in opposing orientations ∼9 Mb apart on LG IV. The attP (antSi50) and attB (antSi51) sites were inserted into an intron of the dpy-13 and unc-30 genes, respectively (Figure 4A). Because the insertions are in introns, both insertion alleles are WT. However, an inversion would truncate both genes and produce a Dpy Unc phenotype.

Figure 4.

Figure 4

Open in a new tab

Generation of a large balancer inversion. (A) Schematic for generating an inversion spanning the dpy-13 to unc-30 interval on LG IV. An attB and an attP site were inserted into an intron of the dpy-13 and unc-30 genes, as indicated. Worms carrying this chromosome were crossed to a phiC31 integrase expressing strain and then Dpy Unc strains carrying homozygous inversions were isolated. Red and green arrows represent primers used in PCR assays testing for the original or inverted orientations. Intron and exon structure are simplified and are illustrative only; gene sizes and distances are not to scale. (B) PCR assays for representative Dpy Unc strains that carry the phiC31 integrase induced inversion. Seven strains from four independent lines (first digit of lane IDs) are shown. PCR arrays for the recombinant junctions are positive while those for the original orientations are negative, except for N2. Minus sign (−) is water. (C) The inversion, antIn1, suppresses recombination in the dpy-13 to unc-30 interval.

To induce the inversion, we crossed the phiC31 integrase expressing transgene, antIs31, into the attP and attB bearing strain to generate antIs31/+; antSi50 antSi51/+ + heterozygotes. From 15 F1 heterozygotes, we obtained eight independent Dpy Unc lines, via individuals found in the F2 or F3 generations. We then used PCR assays to show that the Dpy and Unc phenotypes were consistent with an inversion within the dpy-13 and unc-30 genes (Figure 4B). This result demonstrates that phiC31 integrase can mediate precise inversions in C. elegans and such inversions can be multimegabase in size.

Inversions should suppress recombination. To determine the recombination rate in the dpy-13 unc-30 interval for the phiC31-mediated inversion, we crossed N2 males to one of the inversion alleles, (antIn1[dpy-13(ant41) unc-30(ant42)]), obtained F1 heterozygotes (dpy-13 unc-30/+ +), singled out non-Dpy non-Unc F2 individuals, and then scored for F2 heterozygous recombinants (reflecting recombination events in the F1 mother) by examining the F3 segregation pattern. None of the 117 singled out F2 individuals segregated any recombinant genotypes (Figure 4C); this proportion is different from that expected for the 8.22 cM genetic distance between the two genes (P = 2.0 × 10e−6, expect 13.0, exact binomial test). Control crosses using the double mutant dpy-13(e184) unc-30(ok613) that is in the WT orientation yielded 14/143 recombinants (Figure 4C), which corresponded to 7.27 cM, and was not different from expected (P = 0.79, expect 15.9, exact binomial test). These results demonstrate that antIn1 strongly suppresses recombination in the dpy-13 unc-30 interval.

Discussion

We have demonstrated that the phiC31 integrase system can be used in C. elegans to insert transgenes through RMCE. phiC31 RMCE insertion rates were reasonably high for insertions up to ∼14.9 kb, and we were able to obtain an integrant of ∼33.4 kb. Using phiC31-mediated RMCE, we developed a system able to study post-transcriptional gene regulation by integration of a dual-fluorescence operon reporter. Additionally, we showed that phiC31 integrase can be used to generate inversions, and such inversions can be large.

The main advantage of the phiC31 integrase system is the ability to insert very large transgenes. We have demonstrated a precise, single-copy insertion of a ∼33.4-kb fragment. This size is in the range of typical fosmids and is more than double that of other reported methods. In comparison, MosSCI (Frøkjaer-Jensen et al. 2008, 2012), CRISPR/Cas9 (Dickinson et al. 2013), and recent variants, such as the use of SECs (Das et al. 2015; Dickinson et al. 2015), mmCRISPi (Philip et al. 2019), the adaptation of SapTrap for assembly of MosSCI targeting vectors (Fan et al. 2020), and split-selection safe harbor approaches (Stevenson et al. 2020), have made large insertions of 5–12 kb fairly straightforward, and insertions of ∼16 kb have been reported (Frøkjaer-Jensen et al. 2012). Similarly, the largest reported insertion in the Flp/FRT system was ∼14 kb prior to removal of the self-excising cassette (SEC) (Nonet 2020). Larger insertions were not attempted for the Flp/FRT system, but since it also uses recombinational insertion rather than homology directed repair, it may have a similar capacity as phiC31. Biolistic bombardment (Praitis et al. 2001) and miniMos (Frøkjaer-Jensen et al. 2014) are alternatives for generating single or low copy insertions of very large constructs, but insertion sites for both methods are random.

phiC31-mediated RMCE is simple, requiring only the injection of two plasmids, a donor construct and a coinjection marker, followed by screening for WT integrants. No extra manipulations or crosses are required to remove the phiC31 integrase, which makes obtaining insertion strains fairly fast. We can typically obtain homozygous integrants at the F3 generation and occasionally already by the F2 generation. However, it should be noted that comparable speeds can be achieved for direct MosSCI insertions and for the initial integration when using SECs in the CRISPR/Cas9 and Flp/FRT systems (Frøkjaer-Jensen et al. 2008, 2012; Dickinson et al. 2015; Nonet 2020). We also obtained integrants fairly easily. Our average success rate (based on the number of transgenic F1s) was 17.3% for plasmids up to ∼14.9 kb. Given the caveat that we did not conduct direct side-by-side comparisons, this efficiency seems similar to that of the other integration methods (Frøkjaer-Jensen et al. 2008, 2012; Dickinson et al. 2015; Nonet 2020). Overall, this permits rapid tests of genetic constructs especially when single-copy is important.

Another advantage is that phiC31 integrase mediates unidirectional recombination between attP and attB sites, which are dissimilar, to generate attR and attL sites, which are also dissimilar. Thus, insertions cannot be excised. In contrast, in the Flp based system, the FRT recombination sites are the same before and after recombination, so an insertion could be subject to immediate excision. Unidirectional recombination also permits additional rounds of phiC31 RMCE into strains already containing RMCE transgenes. The Flp system may result in the excision of prior integrated transgenes.

Our current design does have several limitations. First, we currently have only one landing pad for integration. A future aim would be to develop additional landing pads on different chromosomes, which would offer more experimental flexibility, especially for conducting downstream crosses or for testing different local genomic environments. Second, integrations into the landing pad can occur in two orientations because the two attP sites are identical in sequence. This could be resolved by placing different attP variants on the “left” and “right” end of the landing pad (with matching changes for the attB sites on the donor construct) to force insertion in a specific orientation, similar to what has been done for the Flp/FRT system (Dickinson et al. 2015; Nonet 2020). A third issue is that the effect of constitutive germline expression of phiC31 integrase is unknown. Constitutive expression of phiC31 integrase in the D. melanogaster germline seems to have little deleterious effect (Bateman et al. 2006). Similarly, antIs31 does not exhibit any overt negative phenotypes. Nevertheless, we did lose the transgene once after passaging for over 1 year. Thus, continuous expression of phiC31 integrase might be slightly deleterious, which could be minimized by building a strain with conditional phiC31 expression. Another potential shortcoming is that the donor construct contains unc-119(+), which may not be desirable. To enable markerless insertion, the incorporation of self-excision cassettes commonly used in CRISPR/Cas9 and Flp/FRT systems may be useful (Dickinson et al. 2015; Nonet 2020). Last, the landing pad is marked with a myo-2p::GFP which is barely detectible under a dissecting microscope. Consequently, scoring for its disappearance after RMCE is occasionally difficult. This could be solved with a brighter fluorescent marker or a stronger promoter.

Dual-fluorescence operon reporter system

With the ability to generate very large insertions (>10 kb), the phiC31 integrase system permits more complex transgenic experiments. As a proof-of-concept, we used RMCE to integrate dual-fluorescence operon reporters for post-transcriptional gene regulation. Our experiments confirmed let-7 regulation of the lin-41 3′-UTR in the seam cells and also revealed other potential regulation of the lin-41 3′-UTR (or possibly of the unc-54 3′-UTR) in other tissues.

Quantitative analysis of fluorescence protein amounts has long been employed to monitor gene expression and regulation in C. elegans (Breimann 2019). While initial approaches were largely limited to detecting the presence or absence of expression by location or time (Merritt et al. 2008), modern analysis techniques have become more sophisticated. Short half-life reporters (e.g., ∼2 h for GFP/PEST) has allowed better quantification of temporal GFP expression in living worms (Frand et al. 2005). Another improvement has come with the advent of single-copy insertion [e.g., using MosSCI (Frøkjaer-Jensen et al. 2008) and CRISPR/Cas9 (Dickinson and Goldstein 2016)]. For example, a quantitative two-color fluorescent reporter system was developed to more accurately measure in vivo images with one reporter containing a test 3′-UTR and the other containing the unc-54 3′-UTR that serves as the calibration control (Ecsedi et al. 2015). Although powerful, this design required placing two reporters driven by the same promoter, both in single copy, into two MoscSCI landing sites on different chromosomes, which is somewhat difficult (because of the construct size) and labor intensive. The phiC31 approach, which allows RMCE with a larger DNA fragment, reduces the time and labor bottleneck while also inserting both reporters into the same genomic site, which would remove any position effects. Thus, single-copy integration by the phiC31 system should be of high value because it accelerates the examination of gene regulation, especially for long constructs such as dual-fluorescence operon reporters.

Insertion size limitations

Of the two BACs tested, we only succeeded using the shorter one to integrate ∼33.4 kb. The success rate for this shorter BAC clone was only 1.1%. However, because we only verified one of the four candidate lines with ONT sequencing, the success rate could be as high as 4.6%. Regardless, 1.1% (or 4.6%) is much lower than the 17.3% average success rate for the insertions of <15 kb. While we only tested one BAC of ∼30 kb, this suggests that integration of transgenes ∼30 kb or larger by RMCE may be challenging.

We suspect the main issue could be technical. Both purification and injection of large BACs may cause shearing of the DNA. Broken DNA in the C. elegans germline seems to undergo intermolecular recombination easily, presumably at microhomology regions, to form extrachromosomal arrays (Stinchcomb et al. 1985). In the Flp/FRT study, the model for RMCE is that in most cases, injected DNA forms an extrachromosomal array intermediate prior to integration (Nonet 2020). This model is compatible with the 100% incorrect integrants that we obtained for the larger ∼137-kb BAC. The PCR assays for integrants showed internal deletions, evidence of insertions with intermolecular recombination, or both (Supplementary Table S4). While C. elegans biology may impose an upper limit, we hope that refinement of the injection protocol will allow inserting larger transgenes.

Inversions

In addition to RMCE, we also used phiC31 integrase to create a large inversion on LG IV. The ability to generate precise inversions is useful in genetics and synthetic biology. In genetic analysis, large inversions can be used as genetic balancers, such as for maintaining strains carrying lethal mutations. Similarly, genetic balancers could be used to keep certain allele combinations together, as occurs naturally for selfish gene complexes, e.g., the pha-1/sup-35 maternal selfish element (Ben-David et al. 2017). Recently, a CRISPR/Cas9 approach was used to produce a powerful set of genetic balancers that spans about 89% of the C. elegans genome (Iwata et al. 2016; Dejima et al. 2018). By suitable placement of attP and attB sites, the phiC31 integrase system could be an alternative or additional approach for obtaining balancer for the remaining regions or for generating balancers over subregions.

In synthetic biology, recombinases have been used to control inversion switches in state machines or genome recorders (Friedland et al. 2009; Roquet et al. 2016). The detection of events in C. elegans has been limited to event reporting, such as by the induction of a fluorescent marker gene expression upon exposure to stresses (Link et al. 1999; Anbalagan et al. 2012). This approach is useful in situations when the signal, e.g., heavy metal presence, is occurring at the same time as, or possibly just prior to, observation. However, transient events that happened too far in the past cannot be recorded. Inversion recorders would fix the information into the genome such that prior exposure would be decoded later, even in subsequent generations. It would be interesting to utilize serine recombinases to record events in an animal, as has been done in microbial systems.

In summary, we have established phiC31 integrase RMCE as another option for precise single-copy insertion of transgenes in C. elegans. Integrations are straightforward to obtain for insert sizes up to ∼15 kb. Although, our system is currently limited to one landing pad, this site is suitable for quickly testing many constructs in the same genomic context and for inserting very large (up to ∼33 kb), complex genetic constructs. As an example, we have used the greater capacity to develop a dual-operon reporter system able to study post-transcriptional gene regulation. Finally, we have shown that the phiC31 integrase can be used to invert DNA sequences in C. elegans. Thus, the phiC31 integrase system should be a useful tool for genetic engineering, genetic analysis, and synthetic biology.

Data availability

The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article. Strains BRC0546 and BRC0566 are available from the Caenorhabditis Genetics Center (CGC). Plasmids pBRC_double_attB_GFP_donor (Addgene ID 169692) and pCG150_double_phiC31_attB (Addgene ID 169693) are available from Addgene. Other strains and plasmids are available upon request. Raw Oxford Nanopore sequence data are available at NCBI SRA under accession numbers: SRR16511097, SRR16511098, SRR16511099, SRR16519834, and SRR16519835 (Project ID: PRJNA773226). The GenBank accession numbers for the assembled BAC sequences are OK668367 and OK668368. Supplemental materials consisting of supplemental methods, supplemental figures, supplemental sequences, supplemental tables, and a detailed phiC31 integrase-mediated RMCE protocol have been deposited at figshare (https://doi.org/10.25386/genetics.16908922).

Acknowledgments

We thank Dr Helge Grosshans, Dr Hitoshi Sawa, Dr Chun-Liang Pan, Dr Yi-Chun Wu, the Caenorhabditis Genetics Center (CGC) for providing C. elegans strains and plasmids, and the C. elegans Core Facility of the National Core Facility for Biopharmaceuticals, Ministry of Science and Technology, Taiwan. We also thank two anonymous reviewers for comments that helped to improve the manuscript.

Funding

This work was supported by grants from the Ministry of Science and Technology, Taiwan (MOST 108-2311-B-002-012), National Taiwan University Hospital (UN108-040), National Taiwan University Hospital Yunlin Branch (NTUHYL110.I003), and National Taiwan University (109L7224 and 110L7205) to S.P.C. and MOST (MOST 100-2311-B-001-015-MY3, MOST 104-2314-B-001-009-MY5) and Academia Sinica (AS-103-CDA-L01, AS-GC-109-09) to J.W.

Conflicts of interest

The authors declare that there is no conflict of interest.

Literature cited

  1. Anbalagan C, Lafayette I, Antoniou-Kourounioti M, Haque M, King J, et al. 2012. Transgenic nematodes as biosensors for metal stress in soil pore water samples. Ecotoxicology. 21:439–455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arribere JA, Bell RT, Fu BX, Artiles KL, Hartman PS, et al. 2014. Efficient marker-free recovery of custom genetic modifications with CRISPR/Cas9 in Caenorhabditis elegans. Genetics. 198:837–846. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bateman JR, Lee AM, Wu CT.. 2006. Site-specific transformation of Drosophila via phiC31 integrase-mediated cassette exchange. Genetics. 173:769–777. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Belteki G, Gertsenstein M, Ow DW, Nagy A.. 2003. Site-specific cassette exchange and germline transmission with mouse ES cells expressing phiC31 integrase. Nat Biotechnol. 21:321–324. [DOI] [PubMed] [Google Scholar]
  5. Ben-David E, Burga A, Kruglyak L.. 2017. A maternal-effect selfish genetic element in Caenorhabditis elegans. Science. 356:1051–1055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bernabe-Orts JM, Quijano-Rubio A, Vazquez-Vilar M, Mancheno-Bonillo J, Moles-Casas V, Selma S, et al. 2020. A memory switch for plant synthetic biology based on the phage ΦC31 integration system. Nucleic Acids Res. 48:3379–3394. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Breimann L, Preusser F, Preibisch S.. 2019. Light-microscopy methods in C. elegans research. Curr Opin Syst Biol. 13:82–92. [Google Scholar]
  8. Cho SW, Lee J, Carroll D, Kim JS, Lee J.. 2013. Heritable gene knockout in Caenorhabditis elegans by direct injection of Cas9-sgRNA ribonucleoproteins. Genetics. 195:1177–1180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cody JP, Graham ND, Zhao C, Swyers NC, Birchler JA.. 2020. Site-specific recombinase genome engineering toolkit in maize. Plant Direct. 4:e00209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Das A, Dickinson DJ, Wood CC, Goldstein B, Slep KC.. 2015. Crescerin uses a TOG domain array to regulate microtubules in the primary cilium. Mol Biol Cell. 26:4248–4264. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dejima K, Hori S, Iwata S, Suehiro Y, Yoshina S, et al. 2018. An aneuploidy-free and structurally defined balancer chromosome toolkit for Caenorhabditis elegans. Cell Rep. 22:232–241. [DOI] [PubMed] [Google Scholar]
  12. Dickinson DJ, Goldstein B.. 2016. CRISPR-based methods for Caenorhabditis elegans genome engineering. Genetics. 202:885–901. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dickinson DJ, Pani AM, Heppert JK, Higgins CD, Goldstein B.. 2015. Streamlined genome engineering with a self-excising drug selection cassette. Genetics. 200:1035–1049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Dickinson DJ, Ward JD, Reiner DJ, Goldstein B.. 2013. Engineering the Caenorhabditis elegans genome using Cas9-triggered homologous recombination. Nat Methods. 10:1028–1034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ecsedi M, Rausch M, Großhans H.. 2015. The let-7 microRNA directs vulval development through a single target. Dev Cell. 32:335–344. [DOI] [PubMed] [Google Scholar]
  16. Evans TC, editor. 2006. Transformation and microinjection. In: The C. elegans Research Community, editor. Wormbook. [Google Scholar]
  17. Fan X, De Henau S, Feinstein J, Miller SI, Han B, et al. 2020. SapTrap assembly of Caenorhabditis elegans MosSCI transgene vectors. G3 (Bethesda). 10:635–644. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Frand AR, Russel S, Ruvkun G.. 2005. Functional genomic analysis of C. elegans molting. PLoS Biol. 3:e312. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Friedland AE, Lu TK, Wang X, Shi D, Church G, et al. 2009. Synthetic gene networks that count. Science. 324:1199–1202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Frøkjaer-Jensen C, Davis MW, Ailion M, Jorgensen EM.. 2012. Improved Mos1-mediated transgenesis in C. elegans. Nat Methods. 9:117–118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Frøkjaer-Jensen C, Davis MW, Hopkins CE, Newman BJ, Thummel JM, et al. 2008. Single-copy insertion of transgenes in Caenorhabditis elegans. Nat Genet. 40:1375–1383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Frøkjaer-Jensen C, Davis MW, Sarov M, Taylor J, Flibotte S, et al. 2014. Random and targeted transgene insertion in Caenorhabditis elegans using a modified Mos1 transposon. Nat Methods. 11:529–534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Groth AC, Fish M, Nusse R, Calos MP.. 2004. Construction of transgenic Drosophila by using the site-specific integrase from phage phiC31. Genetics. 166:1775–1782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Groth AC, Olivares EC, Thyagarajan B, Calos MP.. 2000. A phage integrase directs efficient site-specific integration in human cells. Proc Natl Acad Sci USA. 97:5995–6000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hu G, Goll MG, Fisher S.. 2011. PhiC31 integrase mediates efficient cassette exchange in the zebrafish germline. Dev Dyn. 240:2101–2107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Huang T, Kuersten S, Deshpande AM, Spieth J, MacMorris M, et al. 2001. Intercistronic region required for polycistronic pre-mRNA processing in Caenorhabditis elegans. Mol Cell Biol. 21:1111–1120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Iwata S, Yoshina S, Suehiro Y, Hori S, Mitani S.. 2016. Engineering new balancer chromosomes in C. elegans via CRISPR/Cas9. Sci Rep. 6:33840. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kirchmaier S, Hockendorf B, Moller EK, Bornhorst D, Spitz F, et al. 2013. Efficient site-specific transgenesis and enhancer activity tests in medaka using phiC31 integrase. Development. 140:4287–4295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Le TS, Yang FJ, Lo YH, Chang TC, Hsu JC, et al. 2017. Non-Mendelian assortment of homologous autosomes of different sizes in males is the ancestral state in the Caenorhabditis lineage. Sci Rep. 7:12819. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Link CD, Cypser JR, Johnson CJ, Johnson TE.. 1999. Direct observation of stress response in Caenorhabditis elegans using a reporter transgene. Cell Stress Chaper. 4:235–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lu J, Maddison LA, Chen W.. 2011. PhiC31 integrase induces efficient site-specific excision in zebrafish. Transgenic Res. 20:183–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mello CC, Kramer JM, Stinchcomb D, Ambros V.. 1991. Efficient gene transfer in C. elegans: extrachromosomal maintenance and integration of transforming sequences. EMBO J. 10:3959–3970. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Merrick CA, Zhao J, Rosser SJ.. 2018. Serine integrases: advancing synthetic biology. ACS Synth Biol. 7:299–310. [DOI] [PubMed] [Google Scholar]
  34. Merritt C, Rasoloson D, Ko D, Seydoux G.. 2008. 3' UTRs are the primary regulators of gene expression in the C. elegans germline. Curr Biol. 18:1476–1482. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Mok DZ, Sternberg PW, Inoue T.. 2015. Morphologically defined sub-stages of C. elegans vulval development in the fourth larval stage. BMC Dev Biol. 15:26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Nance J, Frøkjaer-Jensen C.. 2019. The Caenorhabditis elegans transgenic toolbox. Genetics. 212:959–990. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Nonet ML. 2020. Efficient transgenesis in Caenorhabditis elegans using Flp recombinase-mediated cassette exchange. Genetics. 215:903–921. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Paix A, Folkmann A, Rasoloson D, Seydoux G.. 2015. High efficiency, homology-directed genome editing in Caenorhabditis elegans using CRISPR-Cas9 ribonucleoprotein complexes. Genetics. 201:47–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Philip NS, Escobedo F, Bahr LL, Berry BJ, Wojtovich AP.. 2019. Mos1 element-mediated CRISPR integration of transgenes in Caenorhabditis elegans. G3 (Bethesda). 9:2629–2635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Praitis V, Casey E, Collar D, Austin J.. 2001. Creation of low-copy integrated transgenic lines in Caenorhabditis elegans. Genetics. 157:1217–1226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Rausch M, Ecsedi M, Bartake H, Müllner A, Großhans H.. 2015. A genetic interactome of the let-7 microRNA in C. elegans. Dev Biol. 401:276–286. [DOI] [PubMed] [Google Scholar]
  42. Redemann S, Schloissnig S, Ernst S, Pozniakowsky A, Ayloo S, Hyman AA, et al. 2011. Codon adaptation-based control of protein expression in C. elegans. Nat Methods. 8:250–252. [DOI] [PubMed] [Google Scholar]
  43. Roquet N, Soleimany AP, Ferris AC, Aaronson S, Lu TK.. 2016. Synthetic recombinase-based state machines in living cells. Science. 353:aad8559. [DOI] [PubMed] [Google Scholar]
  44. Schindelin J, , Arganda-CarrerasI, , FriseE, , KaynigV, , Longair M, . et al. 2012. Fiji: an open-source platform for biological-image analysis. Nat Methods. 9:676–682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Stevenson ZC, Moerdyk-Schauwecker MJ, Jamison B, Phillips PC.. 2020. Rapid self-selecting and clone-free integration of transgenes into engineered CRISPR safe harbor locations in Caenorhabditis elegans. G3 (Bethesda). 10:3775–3782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Stinchcomb DT, Shaw JE, Carr SH, Hirsh D.. 1985. Extrachromosomal DNA transformation of Caenorhabditis elegans. Mol Cell Biol. 5:3484–3496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Thomson JG, Chan R, Thilmony R, Yau YY, Ow DW.. 2010. PhiC31 recombination system demonstrates heritable germinal transmission of site-specific excision from the Arabidopsis genome. BMC Biotechnol. 10:17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Vargas ML. 2012. Studies of a site-specific recombination system and analysis of new modulators of Notch signaling in C. elegans. Columbia University.
  49. Venken KJ, He Y, Hoskins RA, Bellen HJ.. 2006. P[acman]: a BAC transgenic platform for targeted insertion of large DNA fragments in D. melanogaster. Science. 314:1747–1751. [DOI] [PubMed] [Google Scholar]
  50. Wang J, Chen PJ, Wang GJ, Keller L.. 2010. Chromosome size differences may affect meiosis and genome size. Science. 329:293. [DOI] [PubMed] [Google Scholar]
  51. Way JC, Wang L, Run JQ, Wang A.. 1991. The mec-3 gene contains cis-acting elements mediating positive and negative regulation in cells produced by asymmetric cell division in Caenorhabditis elegans. Genes Dev. 5:2199–2211. [DOI] [PubMed] [Google Scholar]
  52. Xu Z, Brown WR.. 2016. Comparison and optimization of ten phage encoded serine integrases for genome engineering in Saccharomyces cerevisiae. BMC Biotechnol. 16:13. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article. Strains BRC0546 and BRC0566 are available from the Caenorhabditis Genetics Center (CGC). Plasmids pBRC_double_attB_GFP_donor (Addgene ID 169692) and pCG150_double_phiC31_attB (Addgene ID 169693) are available from Addgene. Other strains and plasmids are available upon request. Raw Oxford Nanopore sequence data are available at NCBI SRA under accession numbers: SRR16511097, SRR16511098, SRR16511099, SRR16519834, and SRR16519835 (Project ID: PRJNA773226). The GenBank accession numbers for the assembled BAC sequences are OK668367 and OK668368. Supplemental materials consisting of supplemental methods, supplemental figures, supplemental sequences, supplemental tables, and a detailed phiC31 integrase-mediated RMCE protocol have been deposited at figshare (https://doi.org/10.25386/genetics.16908922).

Articles from Genetics are provided here courtesy of Oxford University Press

RESOURCES