Abnormal expression pattern of knock-in marker in Eimeria tenella using CRISPR/Cas9

Abstract

Gene editing technology has been widely applied in the genetic manipulation of many organisms and is increasingly being utilized in eukaryotic pathogens. However, its efficiency often requires improvement. In our study using CRISPR/Cas9 to genetically manipulate Eimeria tenella, we aimed to insert a tag into a target gene locus via homologous recombination, but observed outcomes inconsistent with expectations. Whole-genome sequencing analysis of the integration sites revealed that the transgenic E. tenella did not exhibit correct targeted integration. These results indicate that creating double-strand breaks (DSB) at specific genomic sites to trigger homology-directed repair (HDR) for gene modification can lead to mislocalized expression. This study provides insights for utilizing CRISPR/Cas9 technology in genetic editing, particularly in E. tenella, and offers suggestions for improving strategies that employ the co-transfection of multiple plasmids, such as Cas9-gRNA and donor plasmids.

Introduction

Among the seven main Eimeria species that infect poultry, E. tenella is regarded as the most important pathogen because it causes the most acute and fatal form of cecal coccidiosis, leading to bloody feces and high mortality, which results in the most severe economic losses to the poultry industry(Attree et al., 2021; Blake et al., 2020). Despite nearly three decades of research and notable progress in genetic manipulation techniques for Eimeria, further advancement has been seriously hindered by four major bottlenecks: the lack of an efficient in vitro culture system(Feix et al., 2021), low transfection efficiency(Clark et al., 2008; Liu et al., 2008), difficulties in obtaining clonal populations(Liu et al., 2013; Tang et al., 2016; Yan et al., 2009), and a scarcity of selectable markers(Sun et al., 2023a; Yu et al., 2023; Zhao et al., 2024). The CRISPR/Cas9-based gene editing platform is driving a transformative leap in Eimeria research, not only paving the way for novel coccidiosis control agents but also establishing it as a powerful model for studying fundamental biological processes such as sexual reproduction.

The development of vaccines against chicken coccidiosis has transitioned through three key phases: starting with live wild-type oocyst vaccines, advancing to attenuated live vaccines generated by serial passage, and evolving into novel platforms like genetically engineered and subunit vaccines(Blake et al., 2017; Liu et al., 2023; Wang and Suo, 2020). Subunit and genetically engineered vaccines, recognized for their superior safety and immunogenicity, represent a major future direction for coccidiosis control and are essential for achieving sustainable poultry production. Although numerous key antigens—such as microneme, rhoptry, gametocyte, and refractile body proteins—have been explored in various platforms, including DNA, recombinant protein, viral-vectored, and nanoparticle vaccines, the level of immune protection they induce has yet to meet the requirements for commercial application(Blake et al., 2017; Zhang et al., 2024). Breakthroughs in vaccine development hinge on three focal areas: antigen design and screening, delivery system optimization, and the establishment of a scientific evaluation framework(Wang and Suo, 2020). The discovery and optimization of Eimeria antigens form the fundamental premise for novel vaccines. This research plan aims to utilize CRISPR/Cas9-based gene editing to identify a panel of secreted antigens.

But an unexpected phenotype was observed in a CRISPR/Cas9 knock-in experiment targeting the Green Fluorescent Protein (GFP) gene into the E. tenella genome. The GFP construct lacked an exogenous promoter, yet GFP expression was detected, seemingly driven by the homologous arms used for targeting. Below are the possible explanations for this phenomenon:(i) Endogenous Promoter Activity from the Target Locus. If the GFP gene was inserted downstream of an actively transcribed gene, native upstream promoter or 5′ UTR elements could drive GFP expression. This is often referred to as a promoter trap effect. (ii)Cryptic or Unannotated Promoter Elements in the Homologous Arms. The homology arms might contain cryptic or unannotated promoter elements capable of initiating transcription. (iii)Readthrough Transcription from an Upstream Gene. If GFP were inserted downstream of a gene without a strong terminator, transcriptional readthrough may lead to GFP expression. (iv)Partial Integration of a Promoter from Plasmid Backbone. Unintended integration of vector backbone sequences containing promoters could result in unexpected GFP expression. (v)Bidirectional or Enhancer Activity Nearby. Insertion near bidirectional promoters or strong enhancers might stimulate GFP transcription. Resolving this ambiguity is critical for accurately assessing gene-editing efficiency and avoiding misinterpretation of functional studies.

This research addresses the issue of unintended GFP expression by systematically characterizing its molecular mechanisms and pinpointing its origin. We will evaluate whether such expression originates from non-integrated donor plasmids, sustained on-target activity, or off-target insertions. Our findings will contribute to developing definitive diagnostic standards, thereby increasing the precision and credibility of CRISPR-based applications in E. tenella.

Materials and methods

Ethics statement

The animal study was approved by the Beijing Administration Committee of Laboratory Animals. All procedures were conducted in strict accordance with the guidelines of the Institutional Animal Care and Use Committee of the Institute of Animal Sciences, Chinese Academy of Agricultural Sciences (IAS2024-25), and followed the International Guiding Principles for Biomedical Research Involving Animals.

Parasites and animals

In this study, experiments were conducted using the Beijing strain of E. tenella, which was provided by Professor Suo Xun from the National Animal Protozoa Laboratory at China Agricultural University. The virulence of the strain was maintained through successive passages in specific pathogen-free (SPF) chickens. Oocysts were propagated in vitro and subjected to sporulation culture, after which they were stored in a 2.5% potassium dichromate solution (Zhangjiakou Zhongyuan Chemical Co., Ltd., Zhangjiakou, China) at 4°C. Prior to animal infection, all oocysts were thoroughly washed to remove the potassium dichromate and accurately enumerated using a McMaster Egg Slide Counting Chamber to ensure precise infection doses.

Arbor Acres (AA) broilers used for the passage of the Beijing strain of E. tenella and the transgenic recombinant strain were purchased from Beijing Arbor Acres Poultry Breeding Company Limited. (Beijing, China). The birds were housed in coccidia-free isolators and provided with feed and water that had been strictly sterilized at high temperatures (80°C for over 2 hours) to prevent coccidial contamination. During the passage of the transgenic recombinant strain, prolyl-tRNA synthetase (PRS)(Sun et al., 2023b) was supplemented into the drinking water at a concentration of 0.8 g/2 L. Oocysts were collected from the feces of infected birds between 5.5 and 9 days post-infection (dpi).

Genomic DNA extraction from coccidia

Oocysts (2 × 10⁷) purified with sodium hypochlorite were subjected to disruption using 1 mm glass beads via vortexing for 10 minutes to thoroughly break the oocyst walls, sporocysts, and sporozoites. The resulting homogenate was transferred to a clean 1.5 mL centrifuge tube, and genomic DNA was extracted following the instructions of the Blood & Tissue Genomic DNA Extraction Kit (TIANGEN Biotech, Beijing, China). The concentration and purity of the extracted DNA were measured using a spectrophotometer, and the DNA was stored at –20°C for future use.

Plasmid design

The gene sequences of Mic2 (ETH2_1343100) and Rop35 (ETH2_0408400) from E. tenella were retrieved from the ToxoDB database (https://toxodb.org/toxo/). Based on these sequences, 5′ homology arms of 1500–2000 bp and 3′ homology arms of 1000–1200 bp (located outside the gene coding regions) were defined. Specific primers were then designed to amplify these corresponding regions from the genomic DNA of E. tenella. The GFP reporter gene was cloned downstream of the homology arms, and the drug-selection gene (PRS) was included as a separate expression cassette within the same plasmid. For the construction of pCRISPR-Cas9-Mic2 and pCRISPR-Cas9-Rop35 plasmids, the upstream and downstream fragments containing the gRNA sequences were amplified with four distinct primer pairs, respectively. These fragments were then seamlessly inserted into the AgeI and NcoI sites of the pHis4-Cas9-U6gRNA (UPRT) vector. The assembly of all verified PCR fragments was performed sequentially with the pEASY-Basic Seamless Cloning and Assembly Kit (TransGen Biotech Co., Ltd., Beijing, China) as per the manufacturer's instructions, with the corresponding primers listed in Table 1.

Table 1.

Primers used for constructing the two transfection plasmids.

Primer Name	Sequence (5′→ 3′)	Purpose
Mic2ID-F1	ATGACCATGTACGTACAGTGTGCAAAACGGATGC	Amplification of the Mic2-5′ gene
Mic2ID-R1	CGCCACCGGATGCCCACATCTCTGATTGTTC
Mic2ID-F2	AGATGTGGGCATCCGGTGGCGGAGGTTCTATGGTGAGCAAGGGCGAGGAGC	Amplification of the GFP gene
Mic2ID-R2	TGTCTTTGGTACCAAGCTTCTTGTACAGCTCGTCCATGCCGAGAGTGATC
Mic2ID-F3	CAAGAAGCTTGGTACCAAAGACAATACTGTGCCTCTG	Amplification of the turboID gene
Mic2ID-R31	GCTGAGCCAGATCAGTCCAATGGATCCTGATTAGTATG
Mic2ID-F41	GGATCCATTGGACTGATCTGGCTCAGCCGCTGCAGTAG	Amplification of the 3′UTR-5′actin-PRS gene
Mic2ID-R4	GCGGCGACGTTTCCATTCACTTGTACAGCTCGTCCATGCCGC
Mic2D-F5	GCTGTACAAGTGAATGGAAACGTCGCCGCCGCGGGTTTC	Amplification of the Mic2-3′ gene
Mic2ID-R5	GACTCACTATAGGGCTACGTAGCAAAAAGAGAGCTCACACCCCTCCA
Mic2ID-F6	TTGCTACGTAGCCCTATAGTGAGTCGTATTACAATTC	Amplification of the Cloning vector pSimple-tub+eGFP gene
Mic2ID-R6	GTTTTGCACACTGTACGTACATGGTCATAGCTGTTTCCTG
Rop35ID-F1	AGCTATGACCATGACCGGTATGAGCCTGCGCGCCACTCTT	Amplification of the Rop35-5′ gene
Rop35ID-R1	ATAGAACCTCCGCCACCCAGGAACTTGTAGGGAGTCTG
Rop35ID-F2	CAAGTTCCTGGGTGGCGGAGGTTCTATGGTGAGCAA	Amplification of the GFP-turboID gene
Rop35ID-R2	GCTGAGCCAGATCAGTCCAATGGATCCTGATTAGTATG
Rop35ID-F3	GGATCCATTGGACTGATCTGGCTCAGCCGCTGCAGTAG	Amplification of the 3′UTR-5′actin-PRS gene
Rop35ID-R3	AGGGCCGGGTCACTTGTACAGCTCGTCCATGCCGC
Rop35ID-F4	ACGAGCTGTACAAGTGACCCGGCCCTTGGCTTTTC	Amplification of the Rop35-3′ gene
Rop35ID-R4	GAAGCTGCAGCAACAGGAG
Rop35ID-F5	CTCCACCGGTGCCCTATAGTGAGTCGTATTACAATTCAC	Amplification of the Cloning vector pSimple-tub+eGFP gene
Rop35ID-R5	AGGCTCATACCGGTCATGGTCATAGCTGTTTCCTGTGTGA
Mic2ID-JDF1	CGACAACCACTACCTGAGCTACCAGTC	Identify the turboID sequence
Mic2ID-JDR1	CCGCTATTCCACAGGAAAGCTTCTCAC
Mic2ID-JDF2	GCCTACAACGTCAACATCAAGTTGGAC	Identify the Mic2-3′/Rop35-3′ sequence
Mic2ID-JDR2	CTTCGCTATTACGCCAGCTGGCGAAAG
Mic2ID-JDF3	GCAGTGAGCGCAACGCAATTAATGTGAG	Identify the Mic2-5′/Rop35-5′ sequence
Mic2ID-JDR3	CACGCTGAACTTGTGGCCGTTTACGTC

Design and selection of guide RNA (gRNA)

Genomic sequences of Mic2 (ETH2_1343100) and Rop35 (ETH2_0408400) were retrieved from the ToxoDB database (https://toxodb.org/toxo/). The gRNA target sites were designed to be located approximately 20 bp upstream of the stop codon. Specifically, the EtMic2 gRNA target site is situated near the 5′ end of the gene with the sequence 5′-ATTCGGTCCACTCACTGAAG-3′, and the EtRop35 gRNA target site is also at the 5′ end with the sequence 5′-AGACTACGTGCGGACTTTGC-3′. Both sites were selected based on the following criteria: (1) perfect match to the on-target genomic sequence; (2) zero predicted off-target hits; and (3) local GC content between 45% and 65%.

Transfection and selection

Sporozoites were isolated from freshly collected sporulated oocysts of E. tenella and enumerated. The plasmid was linearized with the restriction enzyme AgeI and subsequently introduced into the sporozoites via restriction enzyme-mediated integration (REMI). Each transfection reaction consisted of 1 × 10⁷ sporozoites, Cas9-gRNA: Donor fragment (1:5), and 5 μL of AgeI. Transfection was performed using the LONZA Nucleofector 2D (AMAXA, Switzerland) with program U-033. Following transfection, the sporozoites were inoculated via the cloacal route into 7-day-old chickens, with four chickens per group, at a dose of 1 × 10⁶ sporozoites per bird. Oocysts were recovered from feces between 5 and 9 days post-infection. The presence of GFP- and mCherry-fluorescent oocysts in the population was confirmed by fluorescence microscopy. Following sporulation, the total oocysts were subjected to fluorescence-activated cell sorting (FACS) (MoFlo, Dako Cytomation, Fort Collins, CO, USA). A total of 2,000 fluorescent oocysts were sorted and passaged in chickens. Throughout the rearing period, PRS was administered at 0.8 g/2 L in the drinking water until the luminescence rate reached approximately 85%.

Observation of endogenous development of transgenic E. tenella

Parasite strains with a luminescence rate exceeding 90% were inoculated into one-week-old SPF chickens. The inoculation doses were administered as follows: 1 × 10⁷ sporozoites per chicken for ceca collected at 96 and 120 hours post-infection (hpi), 5 × 10⁶ at 144 hpi, 1 × 10⁶ at 168 hpi, and 1 × 10⁵ at 192 hpi. Freshly collected ceca were processed by removing their contents and rinsing thoroughly with 1 × PBS buffer. Surface moisture was then blotted dry with absorbent paper. Each cleaned cecum was spread on a glass slide, and the mucosal surface was gently scraped. The collected material was transferred to another slide, flattened under a coverslip, and compressed to maximize transparency. The resulting intestinal smears were directly examined under a fluorescence microscope (TCS SP8, Leica, Germany) to detect fluorescence signals from invading sporozoites, second-generation merozoites, and unsporulated oocysts. Sporulated oocysts were obtained through purification from fecal samples.

Analysis and identification of integration sites in E. tenella using third-generation sequencing

One-month-old Arbor Acres (AA) broiler chickens were inoculated with 2 × 10⁵ sporozoites each of the EtMic2-TurboID and EtRop35-TurboID strains. At 120 hours post-inoculation, the ceca were collected, opened longitudinally, and gently cleared of contents, followed by thorough washing with 1 × PBS buffer. The mucosal surface of the ceca was scraped using a glass slide, and the tissue scrapings were transferred to a beaker containing digestion solution (0.5% sodium taurocholate and 0.25% trypsin). Digestion was performed for 30 minutes at 42°C under constant stirring using a thermostatic shaker. The digested mixture was then filtered through gauze and centrifuged at 3600 rpm for 6 minutes. The resulting pellet was transferred to a 1.5 mL microcentrifuge tube and further centrifuged at 3000 rpm until a clean white pellet of merozoites was obtained. Genomic DNA was extracted from approximately 5 × 10⁸ merozoites. Whole-genome resequencing was conducted by Novogene Co., Ltd. (Beijing, China).

Results

Construction and selection of transgenic E. tenella

Parasite invasion is a rapid process characterized by the secretion of numerous proteins into the host cell. Among these, rhoptry proteins trigger immune responses and cellular stress pathways, representing the host's earliest reactions to invasion. Based on gene sequences of two representative secretory proteins in EtMic2 (ETH2_1343100) and EtRop35 (ETH2_0408400) retrieved from the ToxoDB database (https://toxodb.org/toxo/), primers were designed to amplify approximately 1500 bp upstream (promoter) and 1000 bp downstream of each target gene. The regulatory sequences of these two selected secretory proteins were used to drive GFP reporter gene expression, while the drug selection marker—a mutated prolyl-tRNA synthetase (PRS) gene—was constructed as an independent expression cassette co-expressed with mCherry (Fig. 1A). The completed plasmids were linearized with restriction enzymes and co-transfected with the CRISPR/Cas9 system into E. tenella. Stable positive parasite strains (luminescence rate >85%) were obtained through continuous passaging under PRS selection and fluorescence‐activated cell sorting (FACS). Microscopic observation and flow cytometric analysis revealed GFP expression in only about 0.02% of the first generation (data not shown). However, the fluorescence rate increased to approximately 20% in the second generation and further rose to 85% by the third generation. These results indicate that both GFP and mCherry were effectively expressed and exhibited high functionality and efficiency in E. tenella (Fig. 1B, C).

Fig. 1.

Establishment and selection of the stably transfected E. tenella cell line. A. Schematic representation of the expression vector. In the designed construct, the GFP reporter gene is driven by the approximately 1.5 kb upstream promoter region of the E. tenella microneme protein gene Mic2 or the rhoptry protein gene Rop35, while the downstream ∼1 kb region serves as the 3′ homologous arm. The drug selection marker PRS and mCherry are co-expressed under the control of an actin promoter. B, C. Fluorescence analysis of stably transfected parasites. Oocysts were collected from feces during 5.5–9 days post-infection with either the Et-Mic2(B) or Et-Rop35(C) strain. After purification, sporulated oocysts were observed under fluorescence microscopy to examine GFP and mCherry expression. Scale bar = 100 μm.

Stable transfected E. tenella and GFP development in the whole life cycle

To investigate GFP expression throughout the life cycle of E. tenella, we observed the endogenous and exogenous developmental stages of the transgenic parasite by microscopy. Following infection in the ceca of chickens, transgenic E. tenella exhibited typical developmental characteristics during its endogenous stages: sporozoites were mainly located within the cecal epithelial cells, and between 2 and 4 hours post-infection, they were observed penetrating goblet cells and intestinal epithelial cells. Fluorescence was detected in sporozoites at 1 day post-infection (dpi), and in tissue smear samples collected from 2 to 5 dpi, fluorescent sporozoites, merozoites, and secondary merozoites were also observed. Interestingly, both sporulated and unsporulated oocysts showed fluorescence, and sporozoite nuclei exhibited particularly strong fluorescence (Fig. 2A, B).

Fig. 2.

Observation of the endogenous development stages of transgenic E. tenella expressing fluorescent protein in chickens. A. SPF chickens were infected at 7 days of age with an E. tenella Mic2-positive population. Cecal samples were collected every 24 hours and examined microscopically to observe the following parasite developmental stages: sporozoites at 1 dpi, trophozoites at 2 dpi and 3 dpi, first-generation schizonts at 4 dpi, second-generation merozoites at 5 dpi, and mature oocysts at 6 dpi. Bar=25μ m (2 dpi and 4 dpi) , Bar=50μm(1 dpi,3 dpi,5 dpi, and 6 dpi) . B. Seven-day-old SPF chickens were infected with an E. tenella Rop35-positive population. Cecal samples were collected every 24 hours for microscopic examination of the parasitic stages: sporozoites at 1 day post-infection (dpi), trophozoites at 2 and 3 dpi, first-generation schizonts at 4 dpi, second-generation merozoites at 5 dpi, and mature oocysts at 6 dpi. Bar=25μ m (1 dpi,2dpi and 3 dpi) , Bar=50μm(4 dpi,5 dpi and 6 dpi).

However, upon querying the ToxoDB database for RNA-Seq data of Mic2 (ETH2_1343100), we found that the transcript abundance (in TPM) was nearly zero in the unsporulated oocyst stage. This pattern does not align with our fluorescence expression observations. As for Rop35 (ETH2_0408400), although its expression was relatively low in both unsporulated and sporulated oocysts, it was not close to zero. We therefore speculate that this discrepancy may be due to unexpected expression driven by the targeted homologous arm.

Identification and analysis of the integration sites in E. tenella-Mic2 and E. tenella-Rop35

The PacBio CCS sequencing data for the E. tenella-Mic2 locus were of high quality, exhibiting a dominant read length of 20,000-25,000 bp with sufficient sequencing depth (Fig. 3A). In contrast, the effective reads for E. tenella-Rop35 were predominantly shorter than 5,000 bp. This distinct length profile likely reflects inherent biological differences between the two loci rather than data quality issues, and the high depth of these shorter reads was deemed appropriate for precise integration site analysis (Fig. 3B). Our analysis confirmed that the exogenous sequence Donor-Mic2 successfully integrated into multiple sites within the target genome, primarily localized to chromosomes HG994973.1, HG994968.1, and HG994970.1. High-fidelity integration of the exogenous sequences was robustly supported by the alignment data. For Donor-Mic2, three large segments (1603 bp, 1084 bp, and 1001 bp) on HG994973.1 showed perfect identity (100%), with most other aligned fragments also exhibiting near-perfect sequence identity. The exceptionally low mismatch, gap-open counts, and expected values (mostly 0) confirm the high reliability of these alignments, ruling out random matches (Table 2). Similarly, Donor-Rop35 integrated at multiple loci, with two large efficient integrations on chromosome HG994964.1 (2077 bp and 1139 bp, both near 100% identity) and multiple other perfect alignments on HG994973.1, HG994968.1, and HG994970.1. The high bitscores and negligible E-values further validate these alignments, collectively proving precise, multi-locus insertion (Table 3). Furthermore, the whole-genome mutation landscape revealed conspicuous mutation hotspots and structural variations on chromosomes HG994973.1 (harboring the target gene Mic2) and HG994964.1 (harboring Rop35) (Fig. 3C). This suggests that the experimental procedure may have induced widespread genomic instability or off-target effects. qseqid: Name of the exogenous sequence; qseqid: The start position of the exogenous sequence alignment to the genome; qend: The end position of the exogenous sequence alignment to the genome; sseqid: Chromosome name on the target genome; sstart: Start position on the target genome; ssend: End position on the target genome; pident: Identity value of the exogenous sequence alignment to the genome; length: Region length of the exogenous sequence alignment to the genome; mismatch: Number of mismatches in the alignment region of the exogenous sequence to the genome; gapopen: Number of gaps in the alignment region of the exogenous sequence to the genome; evalue: E-value of the alignment of the exogenous sequence to the genome region; bitscore: Alignment score; qlen: Length of the exogenous sequence; slen: Length of the genome chromosome; qcovs: Coverage of the exogenous alignment length relative to the genome chromosome qseqid: The start position of the exogenous sequence alignment to the genome; qend: The end position of the exogenous sequence alignment to the genome; sseqid: Chromosome name on the target genome; sstart: Start position on the target genome; ssend: End position on the target genome; pident: Identity value of the exogenous sequence alignment to the genome; length: Region length of the exogenous sequence alignment to the genome; mismatch: Number of mismatches in the alignment region of the exogenous sequence to the genome; gapopen: Number of gaps in the alignment region of the exogenous sequence to the genome; evalue: E-value of the alignment of the exogenous sequence to the genome region; bitscore: Alignment score; qlen: Length of the exogenous sequence; slen: Length of the genome chromosome; qcovs: Coverage of the exogenous alignment length relative to the genome chromosome

Fig. 3.

Analysis and identification of the integration sites of E. tenella-Mic2 and E. tenella-Rop35. A. Subsequent to the infection of one-month-old chickens with 2 × 10⁵ E. tenella-Mic2-positive parasites, merozoites were isolated from the ceca at day 5 post-infection for DNA extraction. The size profile of the genomic DNA demonstrated a prominent peak between 15,000 and 20,000 base pairs, which is consistent with the robust amplification and structural intactness of the inserted exogenous fragment. B. Following the isolation of merozoites from the ceca of one-month-old chickens infected with 2 × 10⁵ E. tenella-Rop35-positive parasites at day 5 post-infection, genomic DNA was extracted and analyzed. The analysis revealed a size distribution concentrated in the 15-20 kb range, indicating successful capture and good integrity of the target sequence without large-scale degradation. C. The whole-genome mutation landscape, generated using Circos software, visualizes the genomic variations present in the sample. This circular diagram displays the density distributions of single-nucleotide polymorphisms (SNPs), small insertions and deletions (indels), and structural variations (SVs) from the outer to the inner rings, with the outermost ring providing the genomic coordinates.

Table 2.

Genomic mapping of E. tenella-Mic2 integration sites.

qstart	qend	sseqid	sstart	ssend	pident	length	mismatch	gapopen	evalue	bitscore	qlen	slen	qcovs
232	1834	HG994973.1	2750733	2752335	100	1603	0	0	0	2961	12947	4564562	28
4316	5399	HG994973.1	949812	948729	100	1084	0	0	0	2002	12947	4564562	28
8372	9372	HG994973.1	2752338	2753338	100	1001	0	0	0	1849	12947	4564562	28
3673	4318	HG994968.1	2157682	2157037	100	646	0	0	0	1194	12947	3810913	5
7139	7443	HG994970.1	1437282	1436978	99.672	305	1	0	4.16e-157	558	12947	4007228	17
6370	6574	HG994970.1	1439328	1439124	100	205	0	0	3.48e-103	379	12947	4007228	17
5406	5595	HG994970.1	1441763	1441574	100	190	0	0	7.58e-95	351	12947	4007228	17
6701	6891	HG994970.1	1438601	1438410	99.479	192	0	1	9.81e-94	348	12947	4007228	17
6027	6213	HG994970.1	1440198	1440012	100	187	0	0	3.53e-93	346	12947	4007228	17
6207	6372	HG994970.1	1439698	1439533	100	166	0	0	1.66e-81	307	12947	4007228	17
6890	7053	HG994970.1	1437988	1437825	99.39	164	1	0	1e-78	298	12947	4007228	17
5867	6027	HG994970.1	1440709	1440549	100	161	0	0	1e-78	298	12947	4007228	17
7502	7646	HG994970.1	1436534	1436390	100	145	0	0	7.86e-70	268	12947	4007228	17
5727	5867	HG994970.1	1441127	1440987	100	141	0	0	1.31e-67	261	12947	4007228	17
5595	5727	HG994970.1	1441453	1441321	100	133	0	0	3.68e-63	246	12947	4007228	17
6572	6703	HG994970.1	1438910	1438779	100	132	0	0	1.32e-62	244	12947	4007228	17
7050	7139	HG994970.1	1437641	1437552	100	90	0	0	2.95e-39	167	12947	4007228	17
7441	7506	HG994970.1	1436737	1436672	100	66	0	0	6.47e-26	122	12947	4007228	17

Table 3.

Genomic mapping of E. tenella-Rop35 integration sites.

qstart	qend	sseqid	sstart	ssend	pident	length	mismatch	gapopen	evalue	bitscore	qlen	slen	qcovs
226	2302	HG994964.1	992292	994368	99.904	2077	2	0	0	3825	13550	1948711	24
8837	9975	HG994964.1	994368	995506	100	1139	0	0	0	2104	13550	1948711	24
795	856	HG994964.1	992783	992844	100	62	0	0	1.13e-23	115	13550	1948711	24
4784	5867	HG994973.1	949812	948729	100	1084	0	0	0	2002	13550	4564562	8
4141	4786	HG994968.1	2157682	2157037	100	646	0	0	0	1194	13550	3810913	5
7607	7911	HG994970.1	1437282	1436978	99.672	305	1	0	4.36e-157	558	13550	4007228	17
6838	7042	HG994970.1	1439328	1439124	100	205	0	0	3.64e-103	379	13550	4007228	17
5874	6063	HG994970.1	1441763	1441574	100	190	0	0	7.94e-95	351	13550	4007228	17
7169	7359	HG994970.1	1438601	1438410	99.479	192	0	1	1.03e-93	348	13550	4007228	17
6495	6681	HG994970.1	1440198	1440012	100	187	0	0	3.69e-93	346	13550	4007228	17
6675	6840	HG994970.1	1439698	1439533	100	166	0	0	1.74e-81	307	13550	4007228	17
7358	7521	HG994970.1	1437988	1437825	99.39	164	1	0	1.05e-78	298	13550	4007228	17
6335	6495	HG994970.1	1440709	1440549	100	161	0	0	1.05e-78	298	13550	4007228	17
7970	8114	HG994970.1	1436534	1436390	100	145	0	0	8.22e-70	268	13550	4007228	17
6195	6335	HG994970.1	1441127	1440987	100	141	0	0	1.38e-67	261	13550	4007228	17
6063	6195	HG994970.1	1441453	1441321	100	133	0	0	3.85e-63	246	13550	4007228	17
7040	7171	HG994970.1	1438910	1438779	100	132	0	0	1.39e-62	244	13550	4007228	17
7518	7607	HG994970.1	1437641	1437552	100	90	0	0	3.09e-39	167	13550	4007228	17
7909	7974	HG994970.1	1436737	1436672	100	66	0	0	6.77e-26	122	13550	4007228	17
1293	1320	HG994969.1	3433271	3433244	100	28	0	0	0.0000902	52.8	13550	3854180	0

Discussion

In this study, we employed the CRISPR/Cas9 system to generate a GFP reporter knock-in strain in E. tenella. Unexpectedly, GFP was stably expressed throughout the parasite's entire life cycle, including in unsporulated oocysts, despite the absence of an exogenous promoter. Based on whole-genome sequencing data, the observed integration appears to be random. The underlying cause of this outcome is not fully understood, but we hypothesize that while the designed ∼1500 bp homologous arms were intended to facilitate homologous recombination, they might inadvertently contain sequence motifs with promoter activity, potentially driving direct gene expression. Furthermore, it remains uncertain whether the specific loci of random integration themselves possess regulatory elements that enhance expression. The genetic manipulation system for E. tenella is not yet fully mature, and the optimal conditions for achieving precise gene knockout using CRISPR/Cas9 in this parasite remain to be systematically established.

Gene editing technology utilizes “molecular scissors”—nucleases—to create site-specific double-strand breaks (DSB) in the genome. Cells then repair these breaks either through homology-directed repair (HDR) or non-homologous end joining (NHEJ), enabling precise modification of target genes. Similar to Toxoplasma gondii, E. tenella possesses a Ku80-dependent DNA repair system, with NHEJ being the predominant repair pathway(Fox et al., 2009; Huynh and Carruthers, 2009). This characteristic makes targeted integration of foreign genes via homologous recombination particularly challenging. Studies have shown that exogenous DNA fragments carrying homology arms of about 1000 bp tend to integrate nonspecifically at multiple genomic loci during recombination in E. tenella(Tang et al., 2016; Yan et al., 2009). This may explain the observed tendency for random integration of exogenous DNA in the present study.

From the first successful transient transfection of E. tenella sporozoites expressing β-galactosidase in 1998(Kelleher and Tomley, 1998) to the achievement of efficient genetic editing using the CRISPR/Cas9 system in 2020(Tang et al., 2020), the development of recombinant Eimeria vectors has completed a leap from “impossible” to “feasible”. Between 2007, when fluorescent proteins such as YFP and RFP were introduced as reporter markers(Hao et al., 2007), and the later identification of drug-selectable markers like prolyl-tRNA synthetase (PRS)(Sun et al., 2023b), the screening efficiency for E. tenella has been significantly enhanced. Early methods relied solely on fluorescent reporters and required multiple rounds of FACS sorting, which was time-consuming and labor-intensive(Huang et al., 2011; Yan et al., 2009). Subsequently, the introduction of drug-selection systems based on markers such as the dihydrofolate reductase–thymidylate synthase (DHFR-TSm2m3, TgDHFR) gene and prolyl-tRNA synthetase (PRS) — combined with fluorescent reporters in a “drug-FACS” dual-selection strategy — shortened the screening cycle by 2- to 3-fold and yielded more stable transgenic populations(Clark et al., 2008; Di Cristina et al., 1999). Notably, a single-vector approach fusing the fluorescent protein with the drug-resistance marker proved more efficient than co-transfection with two separate plasmids(Clark et al., 2008; Hanig et al., 2012; Sun et al., 2023b). These advances have significantly accelerated the development of genetic manipulation platforms for Eimeria. Building on this foundation, the present study employs two fluorescent proteins (GFP and mCherry) along with one drug-selectable marker, enabling the enrichment of a high proportion of positive parasites within three passages.

In conclusion, our findings underscore that in apicomplexan parasites, reporter gene “expression” does not necessarily equate to “correct expression”. Only through the dual validation of integration site and transcriptional regulation can the accuracy and reproducibility of CRISPR knock-in experiments be ensured.

Funding

This study was supported by the Youth Innovation Program of Chinese Academy of Agricultural Sciences (Y2023QC10), the National Natural Science Foundation of China (32273035) and the Agricultural Science and Technology Innovation Program of China (ASTIP-IAS15).

CRediT authorship contribution statement

Yanzhen Liao: Writing – original draft, Formal analysis, Data curation, Conceptualization. Lin Liang: Data curation. Ruiying Liang: Writing – review & editing. Jiabo Ding: Writing – review & editing. Dandan Hu: Writing – review & editing. Hongbin Si: Writing – review & editing. Xingju Song: Writing – review & editing. Xinming Tang: Writing – review & editing, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition.

Disclosures

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.