Two CRISPR chd2 Knockout Zebrafish Strains Show Distinct Morphology and Behavior

Scientists have created the first two chd2-knockout zebrafish lines, and the pair of mutant strains is already revealing that different disruptions of the same gene can push developing animals in opposite directions. The gene, chromodomain helicase DNA binding domain 2, encodes an ATPase belonging to the SNF2-like family of helicase-related enzymes, and it carries two chromatin organization modifier domains, a SNF2-related helicase/ATPase domain and a DNA binding domain. In humans, CHD2 participates in remodeling chromatin structure, affects histone acetylation and regulates the transcription of multiple genes. It is expressed ubiquitously and plays a critical role in neurogenesis and neuronal differentiation, and it is an important gene associated with childhood epileptic encephalopathy. Patients with CHD2 mutations typically manifest myoclonic seizures and photosensitivity, and some present with other neuropsychiatric disorders including autism spectrum disorders, intellectual disability and developmental delay.

The new study, published in Development Genes and Evolution, used CRISPR/Cas9 genome editing to generate two zebrafish strains carrying different chd2 mutations, named fdu901 and fdu902. The fdu901 allele carries a four-base deletion (11,979-11982delGGGT) in exon 2, producing a frameshift that results in 64 new amino acids before a termination codon and prevents all functional domains of the protein from being translated. The fdu902 allele carries a single-base deletion (27350delG) in exon 13, also causing a frameshift, but this one results in only 3 new amino acids before termination and retains the whole chromo domain and part of the helicase ATP-binding domain. That structural difference allowed the researchers to conduct a comparative phenotypic study of total versus partial loss of chd2 function in a living vertebrate.

The choice of zebrafish was deliberate. The human CHD2 protein has high homology to those of mouse (96.5 percent) and zebrafish (73.0 percent), making both animals suitable candidate models for CHD2 loss. In mice, homozygous mutation of Chd2 leads to general intrauterine growth delay, and the pups cannot survive after birth, while heterozygous mutation leads to reduced viability in newborns, retarded embryonic development and abnormal spinal morphological development. Chd2 heterozygous mice display aberrant cortical rhythmogenesis and severe deficits in long-term memory, and the gene plays an essential role in the maintenance of radial glial cell self-renewal and neuronal output during cerebral cortical development. Earlier zebrafish work relied on morpholino antisense knockdown, which produces seizure-like behavior and photosensitivity in larvae and developmental defects such as stunted growth, microcephaly, malformed pectoral fins and body curvature, but knockdown techniques can carry off-target effects that true genetic knockouts avoid.

To build the strains, the team injected Cas9 mRNA and guide RNAs targeting exon 2 or exon 13 into one-cell zebrafish embryos, identified two founders, and verified the mutant sites by sequencing against wild-type DNA. The chimeric founders were then mated with fish from the Tu background for two generations to obtain heterozygous carriers, and heterozygous males and females of each line were crossed to produce homozygous mutants, heterozygotes and wild-type siblings for all experiments. Crucially, unlike the situation in mice, both homozygous chd2 knockout zebrafish survived and the mutations could be stably inherited, giving researchers a model that can be studied across the postnatal period, something the lethal mouse knockout cannot offer.

The morphological consequences were substantial and strain-specific. When the researchers compared offspring of heterozygote-by-heterozygote crosses for each allele with wild-type crosses, a greater proportion of fdu901 animals died at 24 hours post fertilization: mortality was 27.7 percent in the fdu901 cross and 23.5 percent in the fdu902 cross, against 7.8 percent in the wild type. At 4 days post fertilization, the deformity rate showed the reverse pattern, with 16.0 percent of the fdu902 cross affected versus 11.7 percent of the fdu901 cross and 1.8 percent of the wild type. The deformities seen after chd2 knockout mainly included pericardial edema, ocular edema, stunted growth, microcephaly, malformed pectoral fins and body curvature. Embryo mortality was higher in fdu901 than in fdu902, and developmental delay was more obvious in fdu901, while fdu902 carried the heavier deformity burden among survivors.

One striking phenotype appeared only in fdu902. Decreased pigmentation in the fdu902 cross began to appear at 2 days post fertilization, and a specific scoliosis emerged at 3 days post fertilization and gradually worsened at 4 days, after which the affected larvae died successively from 5 days onward. Of 27 randomly selected scoliotic larvae from the fdu902 cross that were genotyped by sequencing, 14 were heterozygous and 13 were homozygous for the mutation. In fdu901, by contrast, no signs of this specific scoliosis with decreased pigmentation appeared at any point during the entire 1-to-7-day observation period, although body curvature did occur in that strain. The authors suggest that partial preservation of the chd2 protein structure may affect a pathway of pigmentation formation, a phenomenon they flag as deserving more in-depth study, and they note that in clinical practice mutations at different CHD2 sites are also known to lead to different phenotypes.

Behavioral testing added another layer of contrast. The team tested 7-day-post-fertilization larvae in 24-well plates at 28.5 degrees Celsius, recording activity with a ViewPoint system at 25 frames per second pooled into 1-minute bins, across five separate days of mating batches, and genotyped each larva by PCR and sequencing after testing. Only morphologically normal larvae were used. During 30 minutes of light exposure, the locomotor activity of fdu901 heterozygotes and homozygotes was lower than that of wild type, while the locomotor activity of fdu902 heterozygotes and homozygotes was higher than wild type, although no significant differences were detected between mutant and wild-type groups within each strain’s comparison. In light-to-dark transition tests, both strains showed photosensitive reactions, with sudden increases in total distance traveled at the transitions, but again without statistical differences from wild type.

Further behavioral measures largely came back negative. The frequency of six activity intensity levels (30, 40, 50, 60, 70 and 80-120) measured in the first 25 minutes showed no significant difference between groups, and the same held for the 30-to-60-minute window, where fdu901 tended toward lower activity and fdu902 toward higher activity than wild type but without statistical significance within strains. A center/periphery test, which assessed the percentage of time spent and distance traveled in the peripheral zone of a divided tank as a proxy for danger awareness, likewise showed no significant differences between or within the fdu901 and fdu902 groups. Taken together, the authors summarize, both knockout strains displayed no significant behavioral differences from wild type overall, but the two strains showed different alterations from each other, with the excitability of fdu902 higher than that of fdu901.

Why might the two strains behave so differently? The authors point to the position of each mutation within the protein. The fdu902 knockout terminates translation within the ATP helicase domain, and a previous study suggested that helicase dysfunction in humans might specifically result in neuronal hyperexcitability in the absence of syndromic or dysmorphic features. Clinical cases have documented that patients with various CHD2 mutations exhibit different behavioral phenotypes accompanied by seizures, gross motor delays, anxiety or emotional withdrawal, supporting the idea that mutations at different sites can produce contrasting phenotypes. The morphological divergence may similarly reflect complete loss of function in fdu901 versus some remaining function in fdu902, and the authors add that higher mortality in fdu901 might result in a lower deformity rate among its surviving larvae.

The knockout deformity rates were also notably lower than the roughly 50 percent deformity rate reported for morpholino chd2 knockdown, a difference the authors attribute partly to the off-target effects morpholinos can elicit, primarily mediated through p53 activation, and partly to possible genetic compensation, in which a deleterious mutation leads to transcriptional upregulation of related genes that assume the mutated gene’s function, a mechanism absent in gene knockdown. The authors caution that additional kinds of behavioral detection, or observation of these zebrafish at older ages, may be needed to probe the relationship between CHD2 and neuropsychiatric disorders more deeply. Even so, the two survivable, stably inherited lines, one deleting all functional domains and one retaining part of the protein, now stand as tools for dissecting the functions of chd2’s different domains in a vertebrate that, unlike the mouse, can carry the homozygous loss through development.

Subject of Research: Generation and phenotypic comparison of two CRISPR/Cas9 chd2-knockout zebrafish strains

Article Title: Effects of two chd2-knockout strains on the morphology and behavior in zebrafish

Article References: Li, T., Li, W., Li, F., Lin, J., Zhang, Y., Zhang, Q., Sun, Y., Chen, X., Zhou, S., & Li, Q. (2024). Effects of two chd2-knockout strains on the morphology and behavior in zebrafish. Development Genes and Evolution, 234(2), 173-180. https://doi.org/10.1007/s00427-024-00721-5

Image Credits: AI Generated

DOI: 10.1007/s00427-024-00721-5

Keywords: chd2, zebrafish, CRISPR/Cas9, gene knockout, chromatin remodeling, neurodevelopment, epileptic encephalopathy, developmental deformity, larval behavior, locomotor activity, genetic compensation, Development Genes and Evolution