The serine-threonine protein kinase PAK4 is dispensable in zebrafish: identification of a morpholino-generated pseudophenotype

Abstract

TALEN-based inactivation of the zebrafish pak4 gene resulted in embryos and adult fish that appear normal and fertile. This is in contrast to our previously published studies which were based on the use of antisense morpholino oligonucleotides (MOs). We have excluded potential explanations such as gene duplication, alternate splicing, cryptic initiation of translation, and translation-independent RNA function. Our conclusion is that pak4 is dispensable in zebrafish, and that even when corroborated by robust controls, such as RNA rescue, MOs may elicit misleading pseudophenotypes that do not correspond to results obtained by genetic mutations, and should thus be used with caution.

Figure 1. TALEN design and results of indel mutations in zebrafish pak4.

(A) Alignment of wild type pak4 sequence with mutated PCR amplicons. A pair of TALEN was designed overlapping (left arm) and downstream (right arm) the start ATG codon. The target sequences of the TAL Effector DNA-binding domains are highlighted in red. The start codon and two polymorphic nucleotides found within the right arm target are underlined. wt, wild type EK. Δ2, Δ4 and Δ5, alleles with 2-, 4- and 5-nucleotide deletion, respectively. Ins6, allele with 6-nucleotide insertion. (B) Results of immunoblotting for pak4 protein. Total protein from 100 embryos at shield stage was collected, immunoprecipitated and analyzed using a custom-made zebrafish pak4 anti-peptide antibody. wt, wild type EK. Δ2/Δ4, F₃ mutants of mixed genotypes pak4Δ2/Δ2, pak4Δ4/Δ4 and pak4Δ2/Δ4. pak4 protein with a predicted molecular weight of 72 kDa was detected in the wild type embryos but not in the mutants. Duplicate membrane probed with antibody to α-tubulin was used as loading control.

Figure 2. Morphological phenotype of live pak4Δ2/Δ4 mutants.

(A) wild type and F₃ mutant fish at 2 dpf. (B) wild type and F₃ mutant fish at 6 months. Fish were first photographed for documentation followed by fin biopsy genotyping for confirmation. (C) Whole mount in situ hybridization of wild type and mutant embryos at 7-somite for scl. Lateral (left) and dorsal (right) views are shown. Red and blue arrowheads indicate anterior and posterior blood domains, respectively. (D) Whole mount in situ hybridization of wild type and mutant F₃ embryos at 30 hpf for mpo. wt, wild type EK. In panels A, C and D, Δ2/Δ4 refers to the mixed population of homozygous mutant and trans-heterozygous mutant embryos. Panel B shows individuals of each defined genotype, all of which were normal and fertile.

Figure 3. Testing for other possible forms of expressed pak4 by RT-PCR.

(A) Positions of amplicons a–d corresponding to pak4 cDNA. a: 5′-UTR and 5′-end of coding region; primers F17 and R3. b: region amplified by primer pair (MO1F and MO3R) annealing to the target sites for the translation-blocking MO and splice MO1. c: region encoding kinase domain; primers F7 and R2. d: complete open reading frame; primers F2 and R2. TALEN mutation site and epitope site were indicated. (B) Agarose gel electrophoresis of amplicons a–d using cDNAs extracted from wild type (wt) and F₃ pak4 mixed null (Δ2/Δ4) embryos at shield stage as template. β-actin was used as control.

Figure 4. Testing for possible alternative translational initiation downstream from the mutation site.

(A) Construction of the PAK4-GFP fusions. The 5′-UTR and open reading frame of pak4 were cloned upstream of an eGFP cassette. The TALEN mutation site is indicated by the red arrow. (B) Expression of the fusions in zebrafish embryos. PAK4-eGFP RNAs were synthesized by in vitro transcription and injected into one-cell stage embryos. Green fluorescence was monitored at 24 hpf. Only the wild type construct exhibited fluorescence. (C) Western blot analysis of PAK4-eGFP RNA-injected embryos. Embryos at 24 hpf were dechorionated and used for total protein extraction and Western analysis with a GFP-specific antibody. eGFP RNA was injected into embryos as a positive control. UI, uninjected wild type EK. wt, wild type PAK4-eGFP. Δ2, PAK4Δ2-eGFP. Δ4, PAK4Δ4-eGFP. Ponceau Red staining of the Western blot membrane is shown as a loading control.

Figure 5. Testing the ability of mutant RNAs to rescue the MO knockdown phenotype.

(A) Whole mount in situ hybridization of wild type and mutant embryos at 30 hpf for mpo. (B) MZpak4 knockdown-rescue data. Percentages of embryos showing undetectable, reduced or normal staining as compared to control for mpo expression at 30 hpf in (A). UI, uninjected wild type EK. MO, 6 ng pak4 MO cocktail. MO+wt, 6 ng pak4 MO cocktail plus 800 pg wild type PAK4-eGFP RNA. MO+Δ2, 6 ng pak4 MO cocktail plus 800 pg PAK4Δ2-eGFP RNA. MO+Δ4, 6 ng pak4 MO cocktail plus 800 pg PAK4Δ4-eGFP RNA. Only the wild type pak4 RNA resulted in significant rescue of mpo expression.

Figure 6. Testing for stress contribution to the pak4 MO phenotype.

(A) Injection of standard control MO (6 ng) into wild type and pak4 mutant embryos. Live embryos at 3 dpf were anesthetized and photographed. Images shown are representatives of 92 out of 98 for wild type and 17 out of 17 for mutant. (B) Heat exposure of embryos. Wild type and F₃ mutants were exposed to 35°C from blastula stage to 24 hpf. Control samples (28.5°C) were incubated for 5 hours longer than the heated embryos in order to compensate for different developmental rates.