1. Introduction
As science advances, innovative biotechnologies continuously enhance food production, quality, and animal and human welfare. The clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9), a prototype in genome engineering, has unlocked new possibilities in transgenesis and breeding. It operates by triggering two DNA repair mechanisms-non-homologous end joining (NHEJ) and homology-directed repair (HDR)-when it induces double-strand breaks [
1]. Both mechanisms have been employed in aquaculture to improve consumer-valued traits within genetic breeding programs. NHEJ is used to knock out (KO) or disrupt functional genes, while HDR is utilized to precisely knock in (KI) exogenous genes of interest, thereby improving target traits.
Recently, various CRISPR/Cas9 systems have emerged to improve the target-editing efficiency for KI via the HDR pathway. Successful applications have been observed in model animals using single-stranded oligodeoxynucleotides (ssODNs) as repair templates, which enables the insertion of small DNA fragments [
2⇓-
4]. Yoshimi et al. [
5] extended the ssODN-mediated approach to KI larger sequences with two 80-bp ssODNs combined with CRISPR/Cas9, achieving an approximately
integration rate in rat zygotes [
5]. Later, using the CRISPR/Cas9-ssODNs-mediated KI system, a
KI efficiency in sheep zygotes was obtained [
6]. However, erroneous repair was found to be more likely to occur when ssODNs were used as repair templates in zebrafish (Danio rerio) [
7]. Alternatively, modified donor plasmids with homologous arms (HAs) flanked by two single guide RNA (sgRNA) targeted sequences offer high site-specific KI rates [
8,
9], and this HA-mediated KI has been adapted to zebrafish and medaka (Oryzias latipes) [
9,
10]. Theoretically, directly offering linear double-stranded DNA (dsDNA) flanked by two HAs derived from the
- and
- ends of the targeted site while ignoring the difference in stability between circular DNA and dsDNA donors will cause the KI efficiency to increase. In addition to the type of donor, the use of an appropriate concentration of each component of the CRISPR/Cas9 system has a strong positive impact on KI by reducing off-target events and embryo lethality. In this regard, choosing the right delivery system and component dosages holds promise for achieving highly efficient KIs in non-model fish.
Transgenesis and CRISPR/Cas9-mediated genome editing have transformed aquaculture breeding, yielding commercial products such as transgenic AquAdvantage salmon [
11,
12], gene-edited tiger puffer fish, and red sea bream [
13]. While NHEJ dominates for consumer-focused traits such as growth, coloration, and reproduction, HDR-based KIs are effective for improving omega-3 fatty acid content and disease resistance [
14⇓-
16]. However, the integration of foreign genes via the HDR pathway raises concerns about low KI efficiency and introgression, the latter of which directly impact consumer acceptance of gene-inserted fish [
17]. To address this issue, strategies have been developed to make genetically modified fish reproductively sterile using NHEJ-mediated gene disruption, thereby reducing the environmental impacts [
17⇓-
19]. For example, the luteinizing hormone(LHorlh)gene regulates gametogenesis and gestation by binding the receptor [
20,
21]. LH-deficient female zebrafish are infertile, whereas the mutant males are fertile, indicating that the
gene facilitates fish oocyte maturation and triggers ovulation [
22]. However, interruption of the
gene in channel catfish (Ictalurus punctatus) can result in the production of both sterile males and females [
23].
Large-scale disease outbreaks are inevitable, necessitating improved disease-control methods. Antimicrobial peptides (AMPs) are polypeptides that serve as substitutes for antibiotics in a variety of species’ initial line of defense against microbial invasions with reduced antibiotic residues [
24,
25]. AMPs and antimicrobial peptide genes (AMGs) including cecropin, hepcidin, piscidin, epinecidin-1, lysozyme, and lactoferrin have been used for decades to improve disease resistance in a variety of aquatic animals, as feed supplements or transgenes [
14,
26]. Cathelicidins are a particularly important AMP family that share a common cathelin-like domain [
27] and exhibit broad-spectrum antimicrobial and immune-modulating activities [
28]. Recent investigations have shown that alligator-derived cathelicidin inhibits fish pathogens both in vivo and in vitro [
29⇓-
31]. Therefore, integrating AMGs into the genomic DNA has broad prospects for establishing novel disease-resistant fish lines.
Fish transgenic for AMGs could provide a significant option to address disease problems; however, an additional goal would be to prevent the possibility of the breeding of escapees with wild populations. Hypothetically, a reproductive gene such as that is responsible for gametogenesis and gestation could be knocked out at the DNA level with the replacement of a cathelicidin gene, leading to sterile fish with heightened disease resistance. The gene-edited sterilized fish from this approach could have their fertility temporarily restored with hormone therapy used for artificial fish spawning, making it achievable to produce environmentally compatible and disease-resistant fish lines. In this study, two CRISPR/Cas9 delivery systems-HA- and ssODN-mediated KI-were employed to insert the Alligator sinensis cathelicidin (As-Cath) gene at the channel catfish locus to develop a reversibly sterile and disease-resistant line. We compared the KI efficiency, hatchability, and fry survival from various systems, and then restored the fertility of As-Cath-integrated sterile founders through hormone therapy. In addition, the bacterial resistance of and their offspring individuals from the new fish line was further evaluated.
2. Materials and methods
2.1. Ethical approval
The care and use of animals followed the applicable guidelines from expert training courses. Experimental protocols in the current study were approved by the Auburn University Institutional Animal Care and Use Committee (AU-IACUC; Protocol Review Number: 2021-4003). All fish studies were conducted in compliance with the procedures and standards established by the Association for Assessment and Accreditation of Laboratory Animal Care(AAALAC).
2.2. Target locus for gene insertion
As the target integration site, we selected the
gene, which is widely expressed in the theca cells of the ovary and aids in egg maturation and ovulation during gonadal development [
22]. Based on the published genome of channel catfish [
32], the chosen lh site for sgRNA targeting is located in the middle of exon 2 (
Fig. 1). The inserted segment is derived from the coding sequence (CDS) of the cathelicidin gene of Alligator sinensis (As-Cath, GeneBank accession number XM_006037211.3) [
29].
2.3. Design of donor DNA, sgRNA, and CRISPR/Cas9 system
Gene-targeted KI can be engineered via HDR using dsDNAs or ssODNs as donor templates. In the current study, we employed two CRISPR/Cas9-mediated systems to conduct targeted KI of the As-Cath fragment at the lh locus. For the first system, the CDS of the As-Cath gene was cloned into the pUC57_mini vector at the EcoRV enzyme digestion site to create the ssODN1_As-Cath_- ssODN2 construct as a plasmid donor. Two sgRNAs (sgRNA1 and sgRNA2) were co-injected to operate as "scissors", cutting the lh gene and linearizing the plasmid donor, respectively. Two short ssODNs were provided to ligate the ends of both cut sites, labeled as the two-hit two-oligo (2H2OP) system (
Fig. 1(a)). ssODN1 consists of
, of which the upstream
are derived from the part of exon 2 of the
gene, and the remaining 40 bp are homologous to the pUC57_mini backbone. For ssODN2, the upstream 40 bp are from the pUC57_mini backbone, while the downstream 40 bps come from a portion of exon 2 of the
gene. The dsDNA donor was created by constructing the As-Cath CDS sequence flanked with two HAs of 300 bp derived from the
gene of channel catfish on either side of the insert DNA; we tagged the second construct as HA1_As-Cath_HA2. More specifically, 163 bp of HA1 (the left homology arm) were derived from the upstream of exon 2, 136 bp are identical to intron 1, and 1 bp originated from exon 1. HA2 (the right homology arm) contains 21 bps from the downstream of exon 2,85 bps from intron 2, and 194 bps from upstream of exon 3 (Appendix A). Here, we used one sgRNA (sgRNA1) to cut the
site in the channel catfish genomic DNA and provided a linear dsDNA as the donor template; this system was labeled as dsDNA (
Fig. 1(b)). For both constructs, the expression of the As-Cath gene was driven by the zebrafish ubiquitin (UBI) promoter [
33]. The linear dsDNA, circular plasmid, and ssODNs were synthesized by Genewiz (Azenta Life Sciences, USA).
The sgRNAs were selected via the CRISPR design online tool (CRISPR Guide RNA Design Tool, Benching
(
https://zlab.bio/guide-design-resources.)), targeting the
gene of channel catfish and the donor plasmid. Candidate sgRNA sequences were compared to the whole genome of channel catfish via the Basic Local Alignment Search Tool to avoid cleavage of off-target sites. In addition, putative off-target sites were excluded using the online tool Cas-OFFinder
(
‡ http://www.rgenome.net/cas-offinder/.)[
34]. Eventually, sgRNA1 for the
locus and sgRNA2 for the donor plasmid were obtained. The Maxis-cript T7 kit (Thermo Fisher Scientific, USA) was used to generate sgRNAs in vitro, according to the instructions. Next, purified sgRNAs were prepared using the RNA Clean and Concentrator Kit (Zymo Research. The concentration and quality of the sgRNAs were detected with a Nanodrop 2000 spectrophotometer (Thermo Fisher Scientific, USA) and
agarose gel with
tris-borate-EDTA (TBE) buffer, respectively. The synthetic sgRNAs were diluted to a concentration of about
and then divided into polymerase chain reaction (PCR) tubes (2
per tube) and stored at
until use. The Cas9 protein powder was purchased from PNA BIO Inc. (USA); it was then diluted with DNase/RNase-free water to
and kept at
until use. The sgRNA and universal primer used in this study are listed in Table S1 (Appendix A). Two different dosages of the donor DNA template and two control groups were set up-50
, sham-injected control (iCT, only the
phenol red solution was injected), and non-injected control (nCT, no injection)-for each KI system.
2.4. Transgenic fish production and rearing
Mature channel catfish females and males were collected for artificial spawning according to the guidelines provided by Elaswad et al. [
35] with some modifications. In brief, we selected individuals weighing more than
for spawning. Females were anesthetized with
tricaine methanesulfonate (MS222; Hardover, USA) and implanted with
of luteinizing hormone-releasing hormone analog (LHRHa) to induce ovulation. The eggs were then gently stripped in a
greased spawning pan. Mature males were euthanized with 300 ppm MS222, and then the testes were collected, rinsed, weighed, and crushed; sperm were prepared in
saline solution
. Two milliliters of sperm solution was added to approximately 300 eggs and gently mixed. After
of mixing, sufficient pond water was added to the eggs to activate the sperm, then the sperm/egg mixture was gently swirled for
. More water was added, and the embryos were kept in a single layer in the pan; the embryos were then allowed to harden for
before microinjection.
The CRISPR/Cas9 system used for KI microinjections was combined with Cas9 protein, sgRNA, and donor template in a ratio of 2:1:1, including one component of phenol red as an indicator. For the ssODN1_As-Cath_ssODN2 construct (the 2H2OP system),
of Cas9 protein
of sgRNA1/sgRNA2
of donor plasmid (50 and
),
of ssODN1/ssODN2 (50 and
), and
of phenol red solution were mixed for microinjection. With respect to the HA1_As-Cath_HA2 construct (the dsDNA system),
of Cas9 protein
of sgRNA1
of donor dsDNA
and
of phenol red, and
of DNase-free water were mixed to bring the volume to
in total. For each mixture of the CRISPR/Cas9 system, we mixed Cas9 protein and sgRNA first and incubated them on ice for
; the donor templates were then supplemented. For the iCT group, we only injected phenol red (diluted with
saline). The mixed solution for each treatment was microinjected into one-cell stage embryos, as previously described [
36]. Every
of the mixture was loaded into a
outside diameter of borosilicate glass capillary that was pulled into a needle by means of a vertical needle puller (David Kopf Instruments, USA) and injected into 600 embryos. We injected 1000 embryos, dividing them into five random replicates for each treatment; another 200 embryos with three replicates were prepared for each control group. All embryos were from the same parents, and the microinjection was terminated after 90 min post-fertilization.
All injected and control embryos were transferred into 10-L tubs filled with
of Holtfreter solution (59 mmol NaCl,2.4 mmol
, and
) [
37] and 10-12 ppm doxycycline for hatching immediately after microinjection. All tubs were placed in the same flow-through hatching trough, and a heater was put upstream of the trough to ensure that the water temperature was
; the dissolved oxygen levels were maintained at
via continuous aeration. Holtfreter solution was replaced twice per day, and dead embryos/ fry were collected and recorded daily during hatching for analysis. The hatched fry were transferred to a Holtfreter solution without doxycycline and fed with live Artemia nauplii four times per day. After 1 week of culture in tubs, all fry were stocked separately into a 60-L aquaria (120 fish per tank) in a recirculating system for growth. The feed pellet size was adjusted according to the size of the fish’s mouth as the fish grew. In detail, the fry in tanks were fed with Purina AquaMax powdered feed (50% crude protein, 17% crude fat,
crude fiber, and
ash) four times per day for 2 months. Then, the fingerlings were fed with Aquaxcel WW Fish Starter 4512 (45% crude protein, 12% crude fat, 3% crude fiber, and
phosphorus) twice a day for 2 months. Juvenile fish were fed with WW 4010 Transition feed (40% crude protein, 10% crude fat,
crude fiber, and
phosphorus) once a day [
15]. All fish were fed to satiation.
2.5. Integration analysis and mutation detection
After a 4-month culture, all fingerlings(20 - 40g)were pit-tagged (Biomark Inc., USA) to distinguish each individual. The fish from different treatments were then mixed together and randomly dispersed into two circular tanks (1200-L volume filled with of water) with the same density (120 fish per tank) for monthly growth comparison. Meanwhile, the pelvic fin clip and barbel were taken from anesthetized fish for DNA extraction and genotypic identification. During this phase, all fish received WW 4010 Transition feed once a day to satiation. Different genotyping strategies were involved for these two constructs. For the ssODN1_As-Cath_ssODN2 construct, the CDS region of As-Cath was amplified to confirm gene insertion using the primers Cath1- (forward and reverse), and the promoter region was amplified via the primers Prom1-F/R. As for the junctions, the ssODN1 and ssODN2 regions were amplified using the primers ssODN1-F/R and ssODN2-F/R to determine whether it was an on-target insertion. For the HA1_As-Cath_HA2 construct, the As-Cath and promoter regions were detected using the primers Cath2-F/R and Prom2-F/R, respectively. Then, the left HA and right HA junctions were amplified via the primers HA1-F/R and HA2-F/R. The primers were designed using the online software Primer3Plus and are listed in Table S2 (Appendix A). PCR was performed in a 10-μL system, and the PCR products were visualized by running agarose gel with tris-acetate-EDTA (TAE) buffer. A bright band of each region with the corresponding length indicated an on-target positive _As-Cath . Here, if we could determine that some individuals had been inserted with the As-Cath transgene but could not detect the junctional regions (HA- or ssODN-region), we identified them as potential off-target positives .
With respect to the
_As-Cath
(
http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi.) fish, we selected 60 individuals to be tested for
mutations. In this case, PCR was performed in a
volume system using an Expand High Fidelity
PCR System (Roche Diagnostics, USA) according to Elaswad et al. [
35], and LH-F/R primers were used in both constructs. Next, a surveyor mutation detection assay was performed via a Surveyor Mutation Detection Kit (IDT, USA) according to the detailed instructions [
38]. A negative control reaction was included in the assay by using genomic DNA from the nCT group. Surveyor-digested DNA samples were electrophoresed for
in
agarose gel using
TBE buffer and compared with wild-type (WT) samples.
2.6. DNA sequencing
For the integrated As-Cath, promoter, and junction sequences, PCR of the positive samples was performed in a volume system. The PCR products were then purified using a QIAquick PCR Product Purification Kit (QIAGEN, Germany) according to the manufacturer’s instructions. Before sequencing, all purified DNA samples were quantitated and identified using Nanodrop and by running electrophoresis using agarose gel. The primers Cath1-F/Cath2-F and Prom1-F/Prom-2F were used for the sequencing of As-Cath and promoter regions for HA1_As-Cath_HA2 and ssODN1_As-Cath_ssODN2 constructs, respectively; the primers HA1-F/HA2-F and ssODN1-F/ssODN2-F were used for the sequencing of junctional regions for these two constructs, respectively.
Regarding
mutations, we cloned the PCR products of putative mutant individuals using a TOPO TA Cloning Kit (Invitrogen, USA), following the instructions with some modifications. In brief, PCR was performed on each mutant individual that was previously identified with a Surveyor assay using the primers LH-F/R for the next cloning steps. In addition, the DNA of three WT individuals from the nCT group was prepared using the same primers and procedures, and then combined into one reaction and cloned as a WT control for sequencing. After cloning, we transformed the pCR
-TOPO vector containing the PCR products into One Shot TOP10 Electrocomp
Escherichia coli (E. coli) (Invitrogen, USA), as previously described [
35]. Then, 15 single colonies were randomly picked up to perform Colony PCR, and LH-F primer was used for the sequencing of
mutant samples.
2.7. Determination of mosaicism and transgene expression
Five 12-month-old on-target positive fish and five sham-injected control fish were randomly chosen and sacrificed. Fourteen tissues, including the skin, liver, kidney, spleen, blood, intestine, gill, stomach, fin, barbel, muscle, eye, brain, and gonad of each individual, were collected in 1.5-mL tubes and immediately transferred into liquid nitrogen for DNA and RNA isolation. PCR and quantitative real-time PCR (qRT-PCR) were conducted to detect the As-Cath gene’s potential mosaicism and mRNA level. The total RNAs were isolated from various tissues using TRIzol reagent (Thermo Fisher Scientific, USA) and were reverse-transcribed to cDNA using an iScript Synthesis Kit (Bio-Rad, USA), following the manufacturer’s protocols.
qRT-PCR was performed on a C1000 Thermal Cycler using a Sso-Fast
EvaGreen Supermix kit (Bio-Rad, USA), according to the instructions. Concentrations of the cDNA products were diluted to
, and a
template was used in a
PCR reaction volume. The mRNA level of
rRNA was used as an internal control, and the detailed qRT-PCR procedure was set up according to Coogan et al. [
39]. The primers (Cath_RT-F and Cath_RT-R) used for qRT-PCR are listed in Table S2. CFX Manager Software (version 1.6, Bio-Rad) was used to collect the raw crossing-point
values. The expression level of a target gene to the
rRNA gene from transgenic fish against non-transgenic sibling fish was converted to fold differences. Each sample was analyzed in triplicate using the formula
, which sets the zero expression of the non-transgenic full-siblings to
for comparison.
2.8. Reproductive evaluation and restoration of parental KI fish
All
fish were stocked into a 0.04-ha earthen pond at Fish Genetics at Auburn University for growth and maturation. At the age of two years, some
individuals are expected to reach sexual maturity [
40]. To evaluate the reproduction of 2-year-old KI founders, on-target positive
, off-target positive
_As-Cath
, and WT fish were selected to conduct a three-round mating experiment. Firstly, three pairs of WT, six pairs of
_As-Cath
, and four pairs of
_As-Cath
mature parents were randomly placed into 13 tanks
for two weeks of natural spawning to evaluate the spawnability of each genotype, and egg masses were collected from the spawnable parents. We then primed the males with a
LHRHa implant and
human chorionic gonadotropin (HCG) in the unspawned groups with a 1-week observation period to determine whether the
_As-Cath
females were fertile. After this period, we recruited six more pairs of
_As-Cath
fish to assess the effects of hormone therapy using three dosages of a combination of HCG and LHRHa implant (1200 IU.
HCG +
LHRHa,
LHRHa,2000 IU
LHRHa) and a control group injected with
NaCl. A 30-g egg mass for each genotype, with three replicates, was collected to calculate the fecundity (eggs
body weight (BW)). The masses were then transferred into tubs for hatchability and fry-survival determination. Fish were fed ad libitum throughout the experiment.
2.9. Generation and genotype analysis for fish
All the fry were separated into 60-L tanks by different genotypes. After four months of culture, fin clips and barbels were collected for DNA extraction from individuals of each genotype, except the control groups. The same culture and genotyping procedures as described above were applied to the generation.
2.10. Experimental challenge with Flavobacterium covae (F. covae) and Edwardsiella ictaluri (E. ictaluri)
Gene-edited channel catfish were cultured in 60-L aquariums in the greenhouse of the Fish Genetics Laboratory at Auburn University (approved by AU-IACUC). To determine the resistance against pathogens, both and fish were challenged by . covae and . ictaluri.
2.10.1. F. covae challenge
Healthy fingerlings with a BW of (mean standard error), including four genotypes (15 fish per genotype)-that is, _As-Cath _As-Cath , negative _As- (negative fish without As-Cath insertion and lh mutation), and WT-were mixed and acclimated in one hatching trough for five days and then transferred to a tank in the challenge room for acclimation for another prior to bacterial infections. All fish were randomly/equally separated into two 60-L buckets (30 L of water). In brief, a revived F. covae isolate (strain ALG-00- 530) on modified Shieh agar (MSA) was inoculated into multiple cultures of of modified Shieh broth (MSB) in 50-mL sterile flasks and grown in a shaker incubator at for at . These cultures were then expanded into cultures (5-mL additions) in 500-mL flasks and grown for another 12 h. The optical density (OD) was adjusted to ; then, spread plate dilutions were performed to determine the final inoculum concentration. One hundred microliters of each inoculum was serially diluted and spread onto MSA agar plates in duplicate and incubated at for to quantify the concentration of the inoculum. Two flasks containing of inocula were immediately added to two 60-L buckets with fish following preparation, respectively. The fish were then immersed statically in buckets for at approximately (immersion dose: ); afterward, all fish were gently moved back into the 1800-L tank containing 1000 L of water, and water flow was resumed. Meanwhile, a mock-challenged tank was used as the control, incorporating another 40 fish in of rearing water for with sterile MSB (325mL)instead of the bacterial culture. With respect to the challenge of the fry , four families of fry fish per family)-namely, _As-Cath _As-Cath _As-Cath , and WT-were selected, and each family was randomly divided into three replicates with 15 fish per basket. The same challenge procedure and strain of . covae with a dose of (immersion dose: ) were implanted for the generation.
2.10.2. E. ictaluri challenge
Sixty
fish
, including the above four genotypes, were prepared for the E. ictaluri challenge. E. ictaluri (S97- 773) was provided by the USDA-ARS, Aquatic Animal Health Research Unit, USA. The detailed procedures of the E. ictaluri challenge were performed according to Simora et al. [
30], with some modifications. In brief,
of frozen glycerol stock of
. ictaluri was inoculated into
of brain-heart infusion broth (BHIB; Hardy Diagnostics) at
in a shaker incubator at
for
. Bacteria were then subcultured into
of BHIB for another
under the same conditions until the cell density reached about
, based on the
value. All
individuals were transferred into one 1800-L tank for the challenge. Before starting E. ictaluri infection, the water was lowered to a total of
; then,
of
. ictaluri suspension containing 3.
cells was added to the tank, resulting in a final immersion dose of
. The fish were immersed statically for
with aeration
, and then the water was restored. In addition to the infected groups, one control tank containing 30 fish received only BHIB as a mock-challenged group. With respect to the challenge of the
fingerlings
, a total of four genotypes containing 60 fish were selected, and the same challenge procedure and strain of
. ictaluri, with a dose of
(immersion dose:
), were implanted for the
generation.
During the first of the experiment, we checked for mortality every and then three times daily. Challenged fish were continuously monitored for ten days for external clinical signs of . covae/E. ictaluri and confirmation of bacteria colony growth was obtained by isolating bacteria from the kidney and liver to determine the cause of death. Dead individuals were recorded and collected over time.
2.11. Statistical analysis
Spawnability, hatchability, fecundity, fry survival rate, and growth data were analyzed using one-way analysis of variance (ANOVA)/Tukey’s multiple comparisons test to determine the mean differences among treatments. To compare the KI efficiency of different groups, one-way ANOVA/Tukey’s multiple comparisons and odds ratio (OR) (Table S3 in Appendix A) were adopted. The survival curves of the challenge experiments were determined using Kaplan-Meier plots followed by a log-rank (Mantel-Cox) test. All statistical analysis was achieved via GraphPad Prism 9.4.1 (GraphPad Software, LLC). Gene expression between transgenic and non-transgenic fish was analyzed with an unpaired Student’s two-sample -test. Statistical significance was set at , and all data were presented as the mean standard error of the mean(SEM).
3. Results
3.1. Targeted KI of As-Cath gene into the lh locus
Both the 2H2OP and dsDNA systems were able to induce As-Cath-integrated catfish lines with high integrated rates, but the 2H2OP system had significant off-target effects (
Figs. 2(a) and
(b), Figs. S1-S4 in Appendix A). More specifically, the 2H2OP system containing
of donors (2H2OP50) showed the highest KI efficiency, at
, followed by the groups 2H2OP100 (17.76%, 27/152), dsDNA50 (12.21%, 26/213), and dsDNA100 (10.25%, 25/244) (Table S4 in Appendix B). Although the 2H2OP50 group was able to introduce the highest KI efficiency
(
Fig. 3(a)), the
system or
of donors brought a significantly higher KI efficiency than the dsDNA method
or
of donors
(
Figs. 3(b) and
(c)). However, the dsDNA with
donors demonstrated the highest on-target KI efficiency
compared with the other treatments
(
Fig. 3(d)). In contrast, only one on-target KI case was observed in the 2H2OP system, which was significantly lower than in the dsDNA
(
Fig. 3(e)). Although different dosages of donors exhibited a significant effect on the total KI efficiency, our results indicated that this difference was not significant in the on-target KI
(
Fig. 3(f)).
According to the OR, the 2H2OP system and a low dosage tended to bear a higher total integrated rate, which was 2.30 and 1.47 times greater than that of the dsDNA (OR = 2.30 for 2H2OP vs dsDNA) and high dosage for , respectively. Nonetheless, dsDNA showed an overwhelming superiority in on-target integration, which was more than 20 times greater than that in the 2H2OP system (TableS3). Taken together, the findings showed that the dsDNA system accompanied by a dosage of of donors yielded the highest on-target KI efficiency in our current study.
Given the non-As-Cath-integrated fish, we did detect individuals with only the mutation. More specifically, , and of fish with deficiency in the 2H2OP50, 2H2OP100, dsDNA50, and dsDNA100 groups, respectively, were detected by the Surveyor mutation test (Table S4 in Appendix A). The sequencing results revealed2,2,1, and 3 types of mutations in -mutant individuals from the 2H2OP100 group (Fig. S5 in Appendix A).
3.2. Effects of the dosage and CRISPR/Cas9 system
Different donor dosages and CRISPR/Cas9-mediated systems exhibited toxicity to the fish embryos, decreasing the hatchability and fry survival rate. Although there were no significant differences in hatching rates among these four CRISPR/Cas9-mediated injected groups compared with the iCT group
, the hatching rate was lower than that of the nCT group
(
Fig. 3(g)). Moreover, the lethality of embryos was consistent across different donor dosages
or CRISPR/Cas9-mediated systems (2H2OP vs dsDNA)
, and was significantly higher than that in the nCT group (
Figs. 3(h) and
(i)). Regarding fry survival, high survival rates of over
were observed in all groups (P=0.0747;
Fig.3(j)) with-) out significant effects from the CRISPR/Cas9 system (2H2OP vs dsDNA,
;
Fig. 3(k)) and dosages (50 vs
,
;
Fig. 3(l)).
3.3. Mosaicism and As-Cath expression
PCR and qRT-PCR were used to detect the As-Cath transgene and its expression in the different tissues of on-target positive fish. The results revealed that three of the five
_As-Cath
fish expressed the As-Cath in all 14 sampled tissues (skin, liver, kidney, spleen, blood, intestine, gill, stomach, fin, barbel, muscle, eye, brain, and gonad) (
Figs. 4(a) and
(b)), but one of them had expression in 11 tissues (except barbel, muscle, and gill) and another in eight tissues (skin, liver, blood, intestine, gill, barbel, muscle, and gonad) (Fig. S6 in Appendix A), suggesting mosaicism in the on-target positive individuals. We found that the expression of As-Cath was detected even without pathogenic infections for the three on-target positive individuals. The three highest mRNA levels were determined in the kidney (28.91-fold changed), skin (24.30-fold), and gill (8.45-fold), followed by the muscle (7.43-fold), spleen (6.05-fold), and barbel (4.81-fold). However, the eye (1.33-fold), intestine (1.59-fold), and fin (1.61-fold) had the lowest expression compared with other tissues (
Fig. 4(c)).
In addition, compared with the WT individuals, the mRNA level of
in gonads was down-regulated in
_As-Cath
females at the age of one year
, but there was no significant difference in that of males
(
Fig. 4(d)). LH levels in the body are often regulated by feedback mechanisms involving sex hormones such as estrogen and testosterone. Females typically have higher estrogen levels, which can affect LH secretion. Any disruptions to these hormonal feedback loops may impact LH mRNA levels more prominently in females.
3.4. Reproductive sterility and restoration of reproduction
A three-round mating experiment determined the potential for complete control of channel catfish reproduction (
Fig. 5(a)). Our outcomes revealed that three pairs of WT
eggs
BW) and two pairs of
_As-Cath
fish
eggs
BW) were respectively spawned during the first 2-week natural mating, but no spawn was observed in the
_As-Cath
pairs
. Compared with the
_As-Cath
pairs, the WT and
_As-
fish had higher spawnability under natural pairing conditions
. In addition, the
_As-Cath
pairs did not show a significant difference in spawnability compared with the WT pairs
(
Fig. 5(b)).
Furthermore, the 1-week hormone priming
LHRHa
) of the males did not stimulate the
_As-
females to give eggs, indicating that the
-deficient females blocked oocyte maturation and ovulation. However, our results indicated that a combination of LHRHa and HCG effectively induced spawning for the
_As-Cath
females when both males and females were primed. More specifically, two, two, and one female gave eggs 24-48 h post-hormone injection from the 1200 IU (6213 eggs
BW),1600 IU (5514 eggs
BW), and
group combined with
LHRHa, respectively. These three treatments significantly improved the fecundity compared with
injection
. The fecundity decreased with increasing hormone dosage, but the difference among these three hormone dosages was not significant
. Nevertheless, the fecundity was restored to a normal level when
or1600
combined with
LHRHa was adopted (
Fig. 5(c)). Compared with the WT and the other hormonal-therapy groups, the
HCG group significantly reduced the fecundity
and hatchability
(
Fig. 5(d)). Although different hormone treatments had varying effects on fecundity and hatchability, they had no effects on fry survival at the early stage
(
Fig. 5(e)).
3.5. genotyping, growth comparison in and
As mentioned above, three WT, two _As-Cath , and five _As-Cath families were generated from our three-round mating experiment. However, genotype analysis determined that only one family of the _As-Cath line integrated rate in the offspring) and two families of the _As-Cath line integrated rate in the progeny of family 1 and integrated rate in the offspring of family 2) had the As-Cath gene detectable in the generation. These results further confirmed the existence of the mosaic phenomenon in the founders.
To determine the pleiotropic effects of disruption and As-Cath integration on fish growth, we compared the BW of the founders and the progeny, respectively, over time. The growth data suggested that the _As-Cath individuals did not show superiority in terms of growth in the first nine months in the generation. Nonetheless, _As-Cath fish exhibited the largest body gain (36.35g)compared with other genotypes(25g). Furthermore, significantly faster growth was demonstrated in the generation of _As-Cath after a 3-month culture (Table S5 in Appendix A).
3.6. Enhanced resistance against fish pathogens
Dead fish were diagnosed and confirmed with bacterial infection (F. covae or E. ictaluri) based on clinical symptoms and bacterial cultures from their liver and kidneys. Enhanced resistance against bacteria in As-Cath-integrated fish was observed compared with WT/negative individuals from our challenge experiments in both
and
generations. According to the
. covae challenge observation, most of the infected WT fish showed visible clinical signs: saddleback, fin erosion, and yellow skin discoloration on the body. However, some As-Cath transgenic fish showed mild clinical signs, with frayed fins, but survived the infection. There was no significant difference in survival rate between the two types of controls (WT and
_As-Cath
) in both
vs
,
and
generation (26.67% vs
). However, the
_As-Cath
and
_As-Cath
fish exhibited significantly improved survival post-F. covae infection compared with the WT group in both
founders
_As-Cath
vs WT:
vs
_As-Cath
vs WT:
vs
,
and
progeny
_As-Cath
vs WT:
vs
_As-Cath
vs WT:
vs
,
). In addition, the on-target insertion of the
gene resulted in improved resistance against
. covae in comparison with the off-target positives, without statistically differing in both generations
vs
for
, and
vs
for
). Furthermore, our findings revealed that the
progeny was more resistant to
. covae than its
parent (
Figs. 6(a) and
(b)).
The typical "hole-in-the-head" clinical sign was observed in dead WT channel catfish, accompanied by an accumulation of ascites, a hemorrhage, and exophthalmia. However, some As-Cath transgenic fish showed skin ulcers and swollen bellies but survived the infection. Increased resistance to E. ictaluri was also observed in the
vs WT:
vs
_As-
vs WT:
vs
) and
generations
WT:
vs
), with results similar to those of the F. covae challenge. Overall, As-Cath-integrated individuals showed a significant improvement in survival rate compared with WT fish
vs
for
vs
for
). Nevertheless, there was no significant difference in
.As-
and
_As-Cath
fish
vs
for
;
vs
for
) (
Figs. 6(c) and
(d)).
4. Discussion
In contrast to previous gene-editing oriented exclusively to improving the desired traits, the present study took into account ways to decrease the potential impact of transgenic fish on ecosystems and genetic biodiversity. More specifically, we successfully integrated an AMG into a reproduction-associated locus using CRISPR/Cas9-mediated engineering. We identified a suitable KI system for channel catfish to achieve boosted resistance against fish pathogens and reproductive control, with the aim of reducing the reliance on antibiotics and anti-parasitics in aquaculture. The HA-mediated CRISPR/Cas9 system displayed a high integrated rate, low off-target events, and low toxicity. In addition, reproduction was found to be entirely controllable and could effectively be restored with hormone therapy in the new fish line. In general, the insertion of the cathelicidin transgene at the locus for enhanced resistance against infectious diseases and comprehensive reproductive confinement to improve consumer-valued qualities and to promote the environmental sustainability of transgenic fish appears promising.
Several obstacles appear in the CRISPR/Cas9-mediated KI system when it is used in the embryos of non-model animals. In the history of genome editing, the initial CRISPR/Cas9 systems were proposed based on mammalian cells or embryos of the model animals. When moving from model to non-model animals, there are several uncertainties (e.g., embryo size, developmental period, and sensitivity to Cas9 protein) for which researchers must optimize a fitted system when starting a new species’ genome editing. Yoshimi et al. [
5] demonstrated that the ssODN-mediated end-joining approach induced a high integrated rate of
in rats when a short ssODN template was provided. Conversely, recent works have indicated that ssODN-mediated KI can induce a high percentage (17.8%) of indel mutations in sheep [
41]. In the current study, we used CRISPR/Cas9 systems assisted by ssODNs and HAs to create on-target KIs of the As-Cath gene at the lh locus. Although a high KI efficiency of 22.38% (64/286) was detected in the ssODN-mediated system, it caused a high off-target frequency
in the channel catfish. Our results agree with findings in zebrafish, which have illustrated that erroneous ssODN integration occurred when various template lengths were adopted [
7]. These studies suggest that ssODN-mediated KI efficiency in fish models relies heavily on ssODN templates [
42], and caution is warranted when employing ssODNs to create KI models.
Compared with the ssODN-mediated system, HA-assisted KI can achieve a
HDR-mediated KI in human cells with various homogeneous sequences [
9,
43]. In addition, Simora et al. [
44] determined that HA-mediated CRISPR/Cas9 provided with a linear dsDNA donor displayed a total integration rate of
at the non-coding region of the channel catfish genome, which is drastically higher than that of this work (29% vs 11.16% (51/457)). We believe this difference in integration rate is due to the different sample sizes, unknown functions in the target regions (non-coding vs
locus), efficiency of sgRNA and HA, and unpredictable genetic interaction; the larger sample size from our study could give more robust conclusions. These findings reveal that the HA-mediated system is more effective in the catfish species than the ssODN. The KI efficiency of HDR-induced CRISPR/Cas9 has been at a low level, including in cell lines and model animals [
5,
7,
9]. Fortunately, novel CRISPR/Cas-mediated techniques are constantly being developed. For example, the CRISPR/Cas12i-mediated system shows promise in multiplexed genome editing with high mutation rates in human T cells [
45]. Moreover, using homology-independent targeted integration (HITI) vectors, Kelly et al. [
46] established a CRISPR/Cas9 HITI system for the insertion of large DNA donors, with a high integrated efficiency of
in human 293T cells. Recently, a novel study demonstrated that long sequences can be integrated in mammalian cells using the single-strand annealing proteins (SSAPs) with catalytically inactive dCas9 (dCas-SSAP) system with a high on-target KI efficiency
[
47]. These new tools or systems hold promise for application from model to non-model animals and could improve genome-editing efficiency.
Although we predicted and avoided possible off-target events using well-acknowledged software, the actual integration results showed the existence of off-target activities. This was mainly due to the failure of in silico prediction to predict bona fide off-target sites in vivo [
48,
49]. Furthermore, the frequency of off-target events is higher in animal experiments in vivo than in cellular experiments in vitro [
50]. Most publications contend that the observed unintended mutations/insertions are a major concern in the application of the CRISPR/Cas9 system and could confound the interpretation of findings [
49,
51,
52]. However, although some reports claim that no detectable undesirable mutations/insertions from the genotypes or phenotypes have been revealed in mice and fish [
44,
53,
54], the following underlying potential can be noted: ① Unaltered phenotypes may be observed, since off-target cleavage can occur in a non-coding region [
55]. ② The researchers tend to focus on the
founders with intended insertions rather than those harboring possible off-target mutations [
56-
57]. ③ Most studies using animal models do not use genome-wide methodologies for detecting off-target cases, which could conceal some infrequent off-target editing sites [
50]. In the same case, except for
mutations, we did not conduct a thorough detection of all off-target individuals, due to it being time-consuming and expensive. Nevertheless, this does not preclude us from keeping the non-analyzed off-target individuals, as we will eventually genotype them in a genome-wide and unbiased way.
The pivotal role of
in regulating reproduction in channel catfish is evident from our gene expression study. LH plays a crucial role in the final oocyte maturation and ovulation processes in females, as supported by significantly lower
expression in
- knockout catfish females when compared with controls. Conversely, we observed no substantial alterations in
expression in lh-knockout males, underscoring the comparatively lesser role of
in male reproduction. Our observations resonate with similar studies conducted in other species, such as zebrafish [
22]. In lhb-mutant female zebrafish, sterility is a recurring outcome, further reinforcing the critical involvement of
in female reproduction. However, lhb-mutant female zebrafish were fertile, as evidenced in the reduced Gonadosomatic Index (GSI) observed through morphological and histological analyses in
mutant males, and there were no discernible differences in sperm motility and testicular structure when comparing
mutants with control male zebrafish [
22]. Therefore, our findings affirm the pivotal role of
in female channel catfish for oocyte maturation and ovulation, with its impact on male fertility being less pronounced, but still resulting in sterility. Notably, our data, distinct from zebrafish, underscore that testicular development and gene expression in channel catfish do not necessarily equate to fertility, a notable divergence from the observed patterns in zebrafish, and illustrates that great diversity of reproductive strategy and genetics across teleost species.
Genetic mosaicism has been and remains another small obstacle. In this study, we failed to effectively obtain
of individuals without mosaics. In essence, mosaicism from CRISPR/Cas9- genome-edited organisms is common in the case of fertilized-egg-based editing, and mosaic animals have been observed in mice [
58,
59], rats [
57], and zebrafish [
60,
61], with a variety of frequencies, especially in P1 transgenic fish. CRISPR/Cas9 engineered mosaicisms bring undesired consequences, hindering the generation of homozygous positive offspring and prolonging the generation of homozygotes. We evaluated the expression of As-Cath transgene from five on-target positive
founders and found that one individual had no expression in the gonad. In our study, several mosaic events were determined in the germline, resulting in the inability to transfer the As-Cath gene to the offspring. Although early sperm/testis or egg/ovary genotyping can be effective in avoiding the creation of undesirable offspring, it is challenging to access the germline DNA without sacrificing the parents. However, homozygosity should be achieved in the
and
progeny, which may be appropriate for further study. New strategies-that is, Easi-CRISPR, C-CRISPR [
6], CRISPR/Cas9 HITI [
46], and dCas9-SSAP [
47]-might be used to reduce the induction of mosaic animals.
Regardless of the type of CRISPR/Cas9-mediated genome editing, microinjection always has irreversible effects on embryos; in our current study, these were increased mortality and decreased hatchability. High embryonic deaths were observed from sham-and CRISPR/Cas9-mediated-microinjection in our study, revealing that major mortality occurs due to the injection of the yolk, while fewer impacts stem from the DNA donors and reagents [
44]. Although a high dosage resulted in a high embryonic mortality and lower hatching rate, it did not significantly reduce the fry survival rate compared with the WT group, which is in agreement with the findings by Elaswad et al. [
35]. This may be because microinjection only has a detrimental effect on the yolk of the embryo. Still, this effect no longer affects the fry once the fertilized eggs have successfully hatched. Given the unavoidable physical lethality of embryos, off-target effects, and mosaicisms, we recommend the microinjection of about 3000 fertilized eggs for non-model fish species in order to afford enough gene-edited fish for subsequent validation experiments.
To assess the pleiotropic effects, we compared the growth performance of the on-target/off-target As-Cath-integrated fish line with the WT population. Our findings demonstrated that off-target insertions did not exhibit significant growth depression or improvement in various families of
founder. Nonetheless, the preliminary data revealed that the
_As-Cath
fish had a greater gain in BW compared with the WT individuals after a 3-month culture in the tank, and the growth differences are emerging in the
progeny. Our hypothesis is that the AMP transgenics should be healthier, allowing faster growth under certain culture conditions or life stages. cfGnRH-deficient channel catfish did not show significant effects in growth and survival throughout a 4-year culture compared with WT fish [
19]. However, potential pleiotropic effects could exist when the
gene is replaced by As-Cath, as in our case. Therefore,
mosaic founders carrying the As-Cath transgene should be used to produce homozygous families; then, comparisons of the growth, survival rate, seinability, and carcass traits could be performed to make the enhanced performance of the
_As-Cath
fish line more transparent to farmers and the public in the future.
HDR-mediated KI is rarely applied in aquaculture due to the very low integration efficiency, but most of the traits have been achieved via NHEJ-mediated KO [
18]. In addition, few studies have proved that gene-mutants can induce disease-resistant fish lines via KO to date [
14]. In contrast, the integration of AMG is encouraging as a means of improving resistance against pathogens in fish [
14,
17]. Consumers generally have relatively little understanding of transgenesis and have more negative attitudes toward trans-genic organisms than gene-edited organisms [
62]; however, there is still some public pushback against transgenic/gene-edited animals that is hindering them from reaching the market. Here, we reasonably contend that cathelicidin transgenic catfish would not pose a threat to food safety, since ① the gut will digest most proteins and inactivate them even when consumed raw; ② since cooking denatures proteins and the proteins are further digested in the stomach, amino acids rather than active proteins are absorbed by humans; and ③ even though gene sequence is ever-changing in various beings, there are only 20 different types of encoded amino acids that are frequently consumed by humans. In this vein, we are raising attention to the potential benefits and risks of our As-Cath transgenic catfish by making them transparent to the public.
Nonetheless, scientists and breeders need to be aware of the possible damage that genetically modified fish could cause to the environment and ecosystems [
17]. Reproductive sterility via genome editing has been attracting research attention and offers opportunities to reduce environmental risks in aquaculture [
62]. Moreover, representative examples have illustrated that reproductive confinement is promising in model and cultured fish by knocking out/disrupting reproduction-related genes [
23,
63⇓⇓-
66]. Recently, Qin et al. [
19] demonstrated that the reproduction-blocked channel catfish are sterile, and this reproductive confinement can be lifted through hormone therapy with LHRHa. In this study, the dose of
HCG coupled with
LHRHa restored fecundity at the highest level in comparison with other hormone treatments; however, this improvement was not significantly different from the improvement observed with a dose of
. It’s important to note that while the mechanism of LHRHa in synchronizing spawning is well established, its specific necessity for synchronizing spawning in our P1 transgenic fish, similar to wild-type fish, warrants further investigation.
Both on-target and off-target As-Cath transgenic catfish were established in the current study. It should be noted that biosafety is crucial when using gene editing/transgenesis in fish aquaculture. Key safety concerns include the comprehensive analysis of off-target effects through genome-wide assessments using advanced sequencing. In addition, assessing the stability of the genetic modification over generations is vital, achieved by tracking progeny to ensure that the intended changes persist and do not lead to unintended physiological or biochemical consequences due to alterations at non-target loci. Furthermore, since AMG transgenic fish produce AMPs as part of their genetic manipulation, the impact on the health of the host fish should be fully investigated. This includes assessing whether the peptides affect the fish’s growth, immune system, and overall well-being. If these fish are intended for human consumption in the future, thorough safety assessments will be required. These will involve evaluating whether the AMPs or any other components of the transgenic fish pose a health risk to consumers. Special attention should be paid to allergenic potential and possible effects on the human gut microbiome.
5. Conclusions
We established catfish line that confers enhanced resistance to fish pathogens by expressing the cathelicidin protein, accompanied by sterility in both males and females by replacing the lh locus with the cathelicidin transgene. Our study has demonstrated that the insertion of the cathelicidin gene at the locus by harnessing the HA- or ssODN-mediated CRISPR/Cas9 system can be a robust approach to produce sterile and environmentally sound fish lines with enhanced disease resistance. Encouragingly, the CRISPR/ Cas9-mediated KI of AMGs at the reproduction-related loci coupled with hormone therapy could be applied in other commercial fish to increase profits and lower the environmental dangers posed by escaped genetically modified individuals. We contend that the genome-editing tool should be used as a complement to existing breeding techniques, rather than as a replacement for them. Hence, a combination of genome editing and conventional selective breeding is required to maximize the benefits of CRISPR/Cas9 tools in aquatic applications and to hasten the breeding process. In conclusion, this study showed the potential of overexpressing a disease-resistant peptide inserted at a reproduction-related gene using CRISPR/Cas9 engineering in catfish, which may provide a strategy for decreasing bacterial disease problems in aquaculture while simultaneously reducing environmental risks. Reproduction was successful but reduced compared with the WT; thus, hormone therapy to restore fertility requires further research and improvement.
Acknowledgments
We thank Dr. Eric Peatman for providing the CFX96 Real-Time System. This project was partially supported by USDA Grant (2018-33522-28769), and Alabama Agricultural Experiment Station grant (AAES-AIR). Jinhai Wang was supported by a scholarship from the China Scholarship Council. Rhoda Mae C. Simora is based at the College of Fisheries and Ocean Sciences, University of the Philippines Visayas Miagao, Philippines.
Compliance with ethics guidelines
Jinhai Wang, Baofeng Su, De Xing, Timothy J. Bruce, Shangjia Li, Logan Bern, Mei Shang, Andrew Johnson, Rhoda Mae C. Simora, Michael Coogan, Darshika U. Hettiarachchi, Wenwen Wang, Tasnuba Hasin, Jacob Al-Armanazi, Cuiyu Lu, and Rex A. Dunham declare that they have no conflict of interest or financial conflicts to disclose.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.eng.2023.12.005.
PDF
(2550KB)
