Genetic conversion of a split-drive into a full-drive element

Abstract

The core components of CRISPR-based gene drives, Cas9 and guide RNA (gRNA), either can be linked within a self-contained single cassette (full gene-drive, fGD) or be provided in two separate elements (split gene-drive, sGD), the latter offering greater control options. We previously engineered split systems that could be converted genetically into autonomous full drives. Here, we examine such dual systems inserted at the spo11 locus that are recoded to restore gene function and thus organismic fertility. Despite minimal differences in transmission efficiency of the sGD or fGD drive elements in single generation crosses, the reconstituted spo11 fGD cassette surprisingly exhibits slower initial drive kinetics than the unlinked sGD element in multigenerational cage studies, but then eventually catches up to achieve a similar level of final introduction. These unexpected kinetic behaviors most likely reflect differing transient fitness costs associated with individuals co-inheriting Cas9 and gRNA transgenes during the drive process.

Fig. 1. Super-Mendelian performance is comparable between split and full gene-drive elements in single-generation crosses.

a Schematic of the genetic constructs engineered and tested in the study for sGD-to-fGD conversion. A tdTom-expressing split gene-drive cassette (sGD) that contains a hackPAM in between its markers is genetically paired with a specific EGFP-expressing Cas9, which contains a gRNA (that does not home) targeting the sequence next to the hackPAM in sGD, that drives itself into the sGD locus and forms an autonomous full gene-drive cassette (fGD) using marker sequences as homology arms. The resulting cassette expresses both tdTom and EGFP markers. b Upon successful hacking of the spo11 sGD, vasa and nos-driven fGDs (vfGD and nfGD, respectively) were assessed for F₁ germline conversion capacities. The panel follows the layout described in (a), except for the sGD marker being tdTomato and represented in orange and purple for sGD and fGD configurations, respectively. Sex of the parental (F₁) trans-heterozygote is indicated in the X axis under the Cas9 line used, as well as by circles (female) or triangles (male), used to show the data of each individual cross. Error bars represent mean values ± SEM. Stars represent statistical significance (****P < 0.0001) on differences in conversion efficiency between sGD and fGD (black, two-sided t test) and Cas9 deriving from Mendelian frequencies in fGD (green, χ²). For vfGD, n = 2854 (N = 42) for female crosses and n = 1827 (N = 36) for male crosses. For nfGD, n = 1015 (N = 23) and n = 990 (N = 21), respectively. Source Data are provided as a Source Data file. sGD data, which are reproduced from a previous study,  serve as a comparison for the fGD. c Regions surrounding the spo11 gRNA target site were amplified from a small subset (32 for sGD, 18 for fGD) of single non-fluorescent F₂ individuals generated in (b), sequenced by Sanger and analyzed. A bar depicts the % of GD⁺ (purple) and % of non-fluorescent (GD⁻, gray) flies. The prevalence of NHEJ genotypes is shown in blue or red, describing the kind of NHEJ allele that is formed and % among the total tested GD⁻ (NHEJ/WT) heterozygotes. We note that, even though they are likely to very rare, large deletions affecting primer binding (>100 bp deletions) would be classified as WT due to the detection of only the parental WT allele.

Fig. 2. spo11 fGD multigenerational cage trials and NHEJ profile assessment.

a Setup of fGD cage trials followed previous experiments with the sGD-Cas9 configuration. Virgin heterozygote fGD/+ and WT (+/+) flies were seeded at 1:3 ratio in the initial generation and allowed to mate at random at each generation (G_n). Flies in a cage were counted and scored for presence or absence of the phenotypic markers and randomly passed on to the following generation (G_n+1). Gray traces depict the predicted fGD performance. Orange and green traces depict sGD cage experiments where transgene and Cas9 are unlinked and sort independently of one another. sGD data reproduced from a previous study serves as a comparison for the fGD. Purple traces depict fGD cage trials, where transgene and Cas9 are linked as one genomic unit. Source Data for fGD are provided as a Source Data file. b NHEJ cage trial data were obtained by deep-sequencing the target site region of pooled non-fluorescent individuals at specific generations. At each generation, NHEJ alleles are shown to represent their distribution among the total population (fGD⁺ and fGD⁻, left) or only for non-fluorescent fGD⁻ flies (right). Purple bars show the fGD⁺ population percentage. At G₂ and G₄, there are many transient alleles that disappear by G₈, when the large array of NHEJs is simplified to a subset of, most likely, non-detrimental alleles that is maintained by G₂₀. Sequences of the most prevalent NHEJ alleles at G₂₀ are shown next to the graph. c Mathematical models on fGD cage trials were run using fitted parameter values and 100 stochastic simulations were plotted for the spo11 fGD (purple). Thicker lines depict the mean of the 100 simulations.