Role of DENN Domain-Containing Protein 5b (dennd5b) during early embryonic development of zebrafish

Abstract

Prior studies on non-canonical Wnt signaling have established Dishevelled Associated Activator of Morphogenesis 1 (Daam1) as a crucial link in cell movements by cytoskeletal rearrangement. Overexpression or depletion of Daam1 blocks gastrulation in Xenopus embryos and results in phenotype characteristic of spina bifida. A yeast two-hybrid screen has been performed to further identify factors required downstream of Daam1. Among many, DENN (Differentially Expressed in Normal versus Neoplastic cells) domain-containing protein 5 (Dennd5A) is identified as a binding partner of Daam1. In zebrafish, dennd5a and dennd5b are human orthologues of DENND5A and DENND5B. Until now, no data on zebrafish dennd5b’s expression or function is available. This current study elucidates the expression and function of the dennd5b during the early embryonic development of zebrafish. dennd5b shows 68.18% sequence similarity with human DENND5B and 68.56% with zebrafish dennd5a. Semi-quantitative RT-PCR showed maternal deposition of dennd5b at 0 h post-fertilization (hpf), continued expression through gastrulation, somite formation, and persistence into the larval stage. Spatial analysis demonstrated ubiquitous expression during cleavage and gastrulation, followed by restriction to the brain and neural tube during early somite stages, with brain-specific expression maintained through late embryogenesis and larval stage. Functional studies of dennd5b revealed compressed head and tail deformity. Both loss-of-function and gain-of-function perturbations disrupted convergence and extension movements, affecting rhombomere patterning, neural plate morphology, somite organization, notochord structure, and prechordal plate formation. Together, these findings establish dennd5b as essential for zebrafish embryogenesis, particularly in neural development, highlighting a conserved role downstream of Daam1 in non-canonical Wnt-mediated morphogenesis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11033-025-11023-y.

Introduction

Wnt signaling is one of the key regulatory signaling pathways that play a vital role in tissue homeostasis, stem cell renewal, proliferation, motility of cells, the establishment of the primary axis, and head formation during vertebrate development [1–3]. Wnt/β-catenin signaling specifies posterior and ventral fates during gastrulation [4, 5]. Importantly, perturbation of the Wnt signaling pathway causes pathological conditions such as birth defects, tumorigenesis, and other diseases [2]. A wide array of research suggests that misregulated Wnt signaling can cause human cancers in the colon [6], breast [7], lung [8], and skin [9]. More recent works have implicated Wnt signaling in neurodegenerative diseases such as Alzheimer’s [10] and Parkinson’s disease [11].

Wnt molecules, that act as the ligand in this pathway are conserved secreted glycoproteins that bind to a group of membrane-bound receptors Frizzled (Fz) and co-receptors, Low-density lipoprotein related receptor protein (LRP 5/6) gene families [1, 12]. Further downstream of this pathway, the Wnt signal is transduced into the cytoplasmic phosphoprotein Dishevelled (Dvl). Dvl further branches the signal into two distinct pathways, canonical/β-catenin dependent pathway, and non-canonical/β-catenin independent pathway. The Wnt/β-catenin dependent, canonical signaling pathway is the most well-studied Wnt signaling pathway, which is involved in the primary axes, anterior-posterior, dorsal-ventral, and left-right formation in vertebrates [13].

The non-canonical Wnt pathway/β-catenin independent plays an important role in convergent extension and gastrulation cell movement by cytoskeletal rearrangement [14]. In a noncanonical pathway, the Wnt ligand binds with Frizzled or other receptors, such as Ryk and Ror [15, 16], and transduces the signals to Dvl. Prior studies on non-canonical Wnt signaling have established that Daam1 provides a crucial link between Dvl and cytoskeletal rearrangement [17, 18]. Daam1 binds with Dvl and imparts signaling into the Rho GTPase in mediating cytoskeletal changes necessary for gastrulation cell movements, but how Daam1 accomplishes these effects is still unknown.

To further identify the factors required downstream of Daam1 for cytoskeletal changes, a yeast two-hybrid screen was performed using mouse c-terminal Daam1 as bait, and several new proteins were isolated [19, 20]. A cDNA of Dennd5A (R6IP1A or R6IP1) is among the identified binding partners of Daam1. Dennd5B, also referred to as Rab6-interacting Protein 1B-like protein (R6IP1B or R6IP1-like), is one of the two identified members of the DENND5 protein subfamily [21, 22].

The DENN domain is the evolutionarily conserved tripartite module found in species as diverse as humans, worms, flowering plants, and yeasts [23]. DENN domains have three regions, all of which have specific variations of sequence homology [24]. DENN domain module mainly exerts GEF activity on Rab proteins. The uDENN module helps interaction within DENND5A and Rab11 protein [22, 23, 25]. Other domains related to the DENN tripartite domain include RUN1, RUN2 domains, and a PLAT domain that separates the two RUN domains. The PLAT, Polycystin-1 Lipoxygenase Alpha-Toxin domain, also referred to as LH2 (Lipoxygenase homology) domain, is found in a variety of membrane or lipid-associated proteins [26]. The RUN domain is present in one or two copies within various proteins, particularly those associated with the functions of guanosine triphosphatases (GTPases) in the Rap and Rab families. The RUN domain is considered to serve as a specific effector for small GTPases [27]. However, the specific partners and roles of RUN domains have yet to be elucidated.

At the cellular level, DENND5B functions as a guanine nucleotide exchange factor (GEF), which facilitates the conversion of guanosine diphosphate (GDP) to guanosine triphosphate (GTP), thus converting inactive GDP-bound Rab proteins into their active GTP-bound forms [25]. The human DENN domain-containing family of proteins contains 18 members whose physiological functions remain largely uncharacterized [23]. Most of these proteins are classified as small GTPases and play a crucial role in regulating membrane trafficking through the coordination of biogenesis, transport, and fusion processes involving intracellular organelles and vesicles [28, 29].

Recently, the functional deficiency of Rab proteins has been implicated in a diverse range of neurodevelopmental disorders and neurodegenerative conditions in humans [30]. A study on knockout mice has demonstrated that Dennd5b plays an essential role in the post-Golgi secretion of chylomicrons, influencing intestinal triglyceride absorption, body composition, and peripheral lipoprotein metabolism [31]. Another recent study showed that Dennd5b-deficient mice lead to differential expression of key genes involved in hepatic lipid metabolism and lipid storage. These mice are also resistant to Hypercholesterolemia and diet-induced Hepatic Steatosis [32]. However, the role of DENND5B in early embryonic development remains largely unexplored.

In our current study, we sought to explore the expression and function of human DENND5B’s orthologue, zebrafish dennd5b. To our knowledge, no data has been reported on the developmental expression and function of zebrafish dennd5b. Although humans may appear to be different from zebrafish, about 70% of human genes are found in zebrafish [33]. A significant number of genes and essential biological pathways are highly conserved between humans and zebrafish. Most of the diseases that affect humans could be theoretically modeled in zebrafish. Together with the expression and functional analysis, this study provides us with a better understanding of how dennd5b is expressed and functions during the early embryonic development of vertebrates.

Materials and methods

Fish

All experiments conducted in this project are approved by IACUC protocol no. 2022 − 1173 by Sam Houston State University Institutional Animal Care and Use Committee (IACUC) (PI: Sharmin Hasan).

Nomenclature

All orthologous genes mentioned in this current study are written in italics (e.g., DENND5B, Dennd5b, dennd5b) [34]. Human genes are written in uppercase italics (DENND5B), whereas zebrafish genes are written in lowercase italics (dennd5b). Mouse genes are written in first letter capital, italicized (Dennd5b, Daam1). Protein symbols are the same as gene names but not italicized (e.g., DENND5B, dennd5b, Dennd5b).

Sequence conservation studies

The sequences for human and chimpanzee DENND5B, mouse Dennd5b, and zebrafish dennd5b, are obtained from the National Center for Biotechnology Information (NCBI) databases (https://www.ncbi.nlm.nih.gov). The nucleotide sequence identity comparison is shown in supplemental table S1. Zebrafish dennd5b domain information is obtained from UniProt (https://www.uniprot.org/uniprot/F1QQ55). To assess the evolutionary conservation of these sequences, we performed multiple sequence alignments using Clustal Omega. Following the alignment process, we calculated the identity percentages to quantify the similarity between the proteins. To visually represent the conservation of these sequences, we employed the Sequence Manipulation Suite’s Color Align Conservation tool.

Total RNA extraction and purification

Total RNA is extracted using TRIzol Reagent (Sigma Aldrich). For DNase treatment, 1 µL of Turbo DNase 2U/µL (Thermo Fisher Scientific) is used for every 10 µg of total RNA. RNA purification is carried out using the phenol-chloroform ethanol precipitation method. The total RNA concentration is checked by a Nanodrop spectrophotometer (Thermo Fisher Scientific).

cDNA synthesis

SuperScript III First-Strand Synthesis System (Thermo Fisher Scientific) is used to synthesize first-strand cDNA from total RNA using oligo(dT) primers.

Designing primers

Zebrafish actb2 is used as the internal reference gene. Zebrafish dennd5b and actb2 mRNA sequence were obtained from the NCBI database. A forward and a reverse RT-PCR primer (dennd5b RT F and dennd5b RT R) are designed to target 637 base pairs (bp) of the dennd5b coding sequence and partial 3’ untranslated region (UTR) which appeared to be the highly specific region of the dennd5b mRNA. A forward and a reverse primer (actb2 F, actb2 R) is designed to target a partial region of the Open Reading Frame (ORF) of the zebrafish actb2 gene. For RNA probe synthesis, two non-overlapping cDNA fragments of dennd5b are PCR amplified with dennd5b F1 and dennd5b R1 and dennd5b F2 and dennd5b R2 primer sets. All primers sequence used in this current study are listed in supplemental table S2.

Semi-quantitative RT-PCR

To investigate the temporal expression of dennd5b, cDNA was synthesized (Thermo Fisher Scientific) from total RNA collected from zebrafish embryos at 0, 2, 4, 6, 8, 24, 48, and 72 hpf. PCR amplification was performed for dennd5b and the reference gene actb2 using 30 cycles. The initial denaturation step was carried out at 98 °C for 3 min, followed by 30 cycles of denaturation at 98 °C for 30 s, annealing at 55 °C for 30 s, and extension at 68 °C for 30 s. A final extension step was performed at 68 °C for 5 min. Each PCR reaction (10 µL) contained 1 µL each of forward and reverse primers (10 µM), 1 µL of cDNA template, 5 µL of KOD One™ PCR Master Mix (Millipore Sigma), and 2 µL of nuclease-free water. PCR products were resolved on 1% agarose gels and visualized under UV illumination (Supplemental Fig. S1). For semi-quantitative analysis, three independent RT-PCR replicates were performed. Gel images were processed in ImageJ (NIH, https://imagej.nih.gov/ij/). Lanes were selected using the rectangular tool, and band intensities were quantified using the Plot Lanes and wand tool functions. Intensity values were exported to Microsoft Excel, and relative expression levels were calculated as the ratio of dennd5b to actb2 within the same lane. Expression values were normalized to the developmental stage with the highest expression level. Final graphs were generated in Microsoft Excel.

PCR purification

A crisp single band of 255 bp PCR product resulting from dennd5b F2 and dennd5b R2 primer sets was selected for purification to be further cloned into a pGEM-T vector. A Qiagen PCR purification kit is used for the purification of the PCR products. The purified PCR product concentration is measured using Nanodrop (Thermo Fisher Scientific) and is stored at −20 °C.

Cloning of dennd5b

The purified PCR products of 255 bp of dennd5b are cloned into the multiple cloning sites of the pGEM-T vector system. (Promega). The purified plasmids are sent for sequencing with dennd5b specific primers dennd5b F2 or dennd5b R2 (Genewiz). Sequencing data successfully confirmed the presence of 255 bp dennd5b, which showed 100% sequence similarity with the zebrafish dennd5b database sequence (https://useast.ensembl.org/index.html).

RNA probe synthesis

To enhance the sensitivity and accuracy of our in situ hybridization (ISH) experiments, we incorporated multiple specificity controls into the standard protocol. These included two non-overlapping antisense probes for dennd5b and three additional controls: a corresponding sense probe of dennd5b, a positive control probe (myod1), and a negative control (no probe) to assess nonspecific binding. The use of multiple non-overlapping antisense probes together with sense, positive, and negative controls provides high confidence in the specificity of ISH signals [35–39].

For RNA probe synthesis, the 255 bp fragment of dennd5b was cloned into the pGEM-T plasmid. After sequence verification, the plasmid was linearized and used to generate antisense and sense RNA probes using SP6 and T7 RNA polymerases, respectively, with the DIG RNA Labeling Kit (Thermo Fisher Scientific) following the manufacturer’s instructions. To generate the second non-overlapping antisense probe, a 1.3 kb region of dennd5b cDNA was amplified by PCR with dennd5b F1 and T7 promoter incorporated reverse primer dennd5b R1. The size of the product was verified in 1% Agarose gel electrophoresis and purified using Gene Jet PCR purification kit (Thermo Fisher Scientific). The purified PCR product was then used as a template for in vitro transcription with the DIG RNA Labeling Kit (Thermo Fisher Scientific) to synthesize anti-sense probe. All synthesized RNA probes were quantified using a Nanodrop spectrophotometer (Thermo Fisher Scientific), verified by 1% agarose gel electrophoresis, and stored at −80 °C until use. Additional RNA probes for non-canonical Wnt signaling marker genes krox20, myod1, tbxta, dlx3b, and ctslb were provided by Dr. Margot L. K. Williams (Baylor College of Medicine).

Dechorionation of eggs

The embryos are collected at different embryonic (2 hpf-48 hpf) and early larval stage to larval stage (72 hpf-96 hpf). Early stage embryos are dechorionated by using sharp metal forceps and or Pronase reagent (Sigma Aldrich) treatment [35, 36].

Whole-mount in situ hybridization

Whole-mount in situ hybridization of zebrafish embryos was performed following established protocols [35, 36]. For imaging, embryos were mounted in 50% glycerol, and images were captured using a KEYENCE VHX digital microscope. After imaging, embryos were stored at 4 °C.

Knockdown study

We employed a translation-blocking dennd5b morpholino (MO) (Gene Tools Inc.) to knockdown its expression at the early stages of development. Morpholinos are typically made up of 25 morpholine bases oligonucleotides that target the mRNA of interest by complementary base pairing. Translational blocking MO bind to the 5′ untranslated region (UTR) near the translational start site (ATG), inhibiting ribosomal assembly [40]. In parallel, we used a zebrafish-specific standard control MO (STD-MO, Gene Tools Inc.) as a negative control. Zebrafish dennd5b MO oligo sequence is 5′-GCTCATCACGGCTACAGATGATCCG-3′, and zebrafish specific standard control MO is 5′- CCTCTTACCTCAGTTACAATTTATA-3′. We also used a zebrafish p53 MO to verify the off-target effect exerted by the MO (Gene Tools Inc.). Morpholinos are sold in lipophilic powder. Upon receiving, we added nuclease-free water to make a concentration of 1 µg/µL (1 ng/nL), 3 µg/µL (3 ng/nL), 6 µg/µL (6 ng/nL), and 9 µg/µL (9 ng/nL).

Overexpression study

The full-length open reading frame of zebrafish dennd5b is amplified using gene-specific primers is cloned into the pCS2-Flag-N3 vector (Received from Raymond Habas, Temple University, PA, USA). The purified plasmid is sequenced (Eurofins Genomics) to verify the sequence integrity of dennd5b. The sequence-verified dennd5b-pCS2-Flag-N3 and pCS2-Flag-N3 (control) plasmids are linearized with NotI HF (NEB) following the manufacturer’s protocol. After linearization, linearized dennd5b-pCS2-Flag-N3 and pCS2-Flag-N3 are subjected to 1% agarose gel electrophoresis and purified using the Gene Jet Gel Extraction Kit (Thermo Fisher Scientific). The linearized and purified 500 ng of dennd5b-pCS2-Flag-N3 DNA and an empty pCS2-Flag-N3 plasmid DNA is used for in vitro transcription using SP6 mMessage mMachine in vitro Transcription Kit (Thermo Fisher Scientific) protocol. The in vitro transcribed mRNAs of dennd5b and pCS2-Flag-mRNA are purified using a MEGAclear Purification Kit (Thermo Fisher Scientific).

Rescue study

For the rescue experiment, we used an in vitro transcribed full-length open reading frame (ORF) of zebrafish dennd5b that lacked the morpholino binding site. The morpholino was designed to target the 5′ UTR, including the start codon and the immediate downstream codon. Only 6 out of the 25 nucleotides in the morpholino matched the in vitro transcribed dennd5b mRNA sequence, ensuring minimal interference.

Zebrafish breeding and microinjection

Zebrafish embryos are obtained from natural crosses in our zebrafish facility. Eggs are collected from the breeding cages with a strainer and rinsed with egg water (1.5 mL stock salts added to 1 L distilled water = 60 µg/mL final concentration). A 1.0 mm OD glass capillary (World Precision Instruments) was pulled into two needles using a micropipette puller Model P97 (Sutter Instruments) using the setup (Heat 645, Pull 60, Velocity 80, Time 100). A micro-injector with a foot pedal, PLI-97 (Harvard Apparatus) is used to inject the morpholino into the yolk of the zebrafish embryo. The injected healthy eggs are moved into a clean Petri dish using a gentle stream of egg water.

Results

Zebrafish dennd5b nucleotide and protein sequence show similarity among different species

The zebrafish dennd5b gene is located on chromosome 4 and encodes a protein comprising 1,325 amino acids. In contrast, the human orthologue of DENND5B is located on chromosome 12 and contains 1,274 amino acids. Genes that are pivotal in biological processes often show a remarkable degree of conservation across a wide variety of organisms. This conservation is reflected in the stability of their sequences over extended evolutionary timescales. Therefore, we examined the functional characteristics of orthologous genes by conducting a thorough analysis of their coding region and protein sequences. A comparative analysis of human DENND5B’s sequence similarity with closely related vertebrate organisms and zebrafish reveals notable nucleotide sequence similarities among these proteins: 68.15% with zebrafish, 99.84% with chimpanzees, and 90.66% with mice (Supplemental Table S1). DENN domains have three regions: downstream of DENN (dDENN), the main DENN domain, and upstream of DENN domain (uDENN). Other domains related to the DENN domain include RUN1, PLAT, and RUN2 domain (Fig. 1a). The RUN domain is named after RPIP8 Rap2 interacting protein 8 (RPIP8), UNC-14, and new molecule containing SH3 at the carboxyl terminus (NESC) [27]. Further examination of the amino acid sequences of DENND5B across humans, chimpanzees, mice, and zebrafish uncovered that the second RUN domain displayed a significantly higher level of conservation when compared to the other domains (Fig. 1b). The protein sequence alignment between dennd5b with its paralogue dennd5a also indicated diminished conservation levels overall, except for the second RUN domain (Fig. 1c). Moreover, by coding sequence comparison, between the dennd5a and dennd5b genes there exists 68.56% sequence similarity.

Fig. 1.

Multiple sequence alignments of dennd5b amino acid sequences. a The protein domains of zebrafish Dennd5b, including uDENN, cDENN, dDENN, RUN1, PLAT, and RUN2, are illustrated in a schematic representation. The gray line beneath the schematic denotes the positions of the amino acid residues, while the lighter gray line above the schematic marks the specific locations of the individual domains b The amino acid sequence alignment shows the high level of conservation of DENND5B between humans, chimpanzees, mice, and zebrafish. Amino acids highlighted in black shade indicate conservation of the residue between all proteins, those in grey reflect highly similar residues between proteins. Below the amino acid sequences, the blue line represents the position of the uDENN domain, while the light blue lines indicate the central DENN domain. The light pink line corresponds to the dDENN domain. The dark yellow lines denote the locations of the first RUN domain (RUN1), and the light yellow lines mark the second RUN domain (RUN2). Additionally, the green lines represent the PLAT domain, which separates the two RUN domains within DENND5B c Amino acid sequence alignments between zebrafish dennd5a and dennd5b, and light-yellow lines indicate the locations of the RUN2 domain in zebrafish dennd5b

dennd5b is expressed throughout zebrafish embryonic development

To investigate the temporal expression of dennd5b, we performed semi-quantitative RT-PCR at 0, 2, 4, 6, 8, 24, 48, and 72 hpf using actb2 as an internal control for normalization. Following the zebrafish developmental framework of Kimmel et al. (1995) [41], zebrafish development at 28.5 °C can be divided into three phases: maternal expression, embryonic development, and larval development. The maternal phase encompasses the zygote and early cleavage periods (0–¾ hpf), during which development relies on maternally deposited RNAs and proteins prior to zygotic genome activation. Subsequent embryonic stages include the blastula (2¼ hpf), characterized by the midblastula transition (MBT) and onset of epiboly. During the gastrulation (5¼ hpf) stage, involution, convergence, and extension movements establish germ layers and body axes. During the segmentation (10 hpf) stage, somites, neuromeres, and organ primordia formation begins. By the pharyngula stage (24 hpf), axis straightening, circulation, pigmentation, and fin development are evident. In the hatching period (48 hpf), organogenesis is largely complete, and cartilage develops in the head and pectoral fins, followed by asynchronous hatching. By the early larval stage (72 hpf), larvae inflate their swim bladders and initiate feeding and avoidance behaviors.

Our semi-quantitative RT-PCR results demonstrated that dennd5b is maternally expressed from the 1-cell stage (0 hpf) and persists throughout early development up to 72 hpf with no significant changes of expression (Fig. 2a, b). Strong constitutive expression of actb2 was detected across all stages, confirming semi-quantitative RT-PCR assay reliability (Fig. 2a). To assess spatial expression, whole-mount in situ hybridization was performed from blastula (2 hpf) to larval stages (96 hpf). dennd5b transcripts were detected ubiquitously during the blastula (2 hpf) (Fig. 2c), at shield (6 hpf) (Fig. 2d), and at 80% epiboly (8 hpf) (Fig. 2e). By the end of gastrulation (6 hpf-10 hpf), at 100% epiboly stage (10 hpf) dennd5b expression is detected in the neural region (Fig. 2f, g,h). At early somitogenesis (12 hpf), dennd5b became enriched in the developing brain (Fig. 2i, j,k). By 16 hpf, dennd5b transcripts localized strongly to the brain, neural tube and tail bud regions (Fig. 2l). At 22–24 hpf, strong expression was detected in the forebrain, midbrain, midbrain–hindbrain boundary, hindbrain, neural tube and tail regions (Fig. 2m, n,o). In 48 hpf, 72 hpf and at 96 hpf, expression was largely restricted to the brain (Fig. 2p, q,r, s,t, u).

Fig. 2.

Temporal and spatial expression of dennd5b during zebrafish embryogenesis. a-b Semi-quantitative RT-PCR analysis of dennd5b expression at 0, 2, 4, 6, 8, 24, 48, and 72 hpf developmental stages. Zebrafish actb2 was used as an internal control for normalization. b Densitometry-based quantification of dennd5b expression levels relative to actb2. Data are normalized to the 24 hpf timepoint and expressed as mean ± SEM from n = 3 replicates. c- v′ Whole-mount in situ hybridization showing dennd5b mRNA expression at various developmental stages. c-e Early development (2–8 hpf). f-h 10 hpf. i-k 7–8 somite stage (12 hpf), expression in developing brain region. l 16 hpf, expression in brain regions including forebrain (FB), midbrain (MB), hindbrain (HB), and developing neural tube and tail bud. m 22 hpf, expression in FB, MB, HB and neural tube. n-o 24 hpf, strong expression in brain regions (HB, MB, FB) with developing neural tube and tail. p-q 48 hpf, prominent brain expression with developing body structures. r-s 72 hpf, continued brain expression in developing larva. t-u 96 hpf, expression maintained in brain regions of free-swimming larva. In situ hybridization of dennd5b with sense probe, which shows no staining. a′ 2 hpf, b′ 6 hpf, c′ 8 hpf, d′-f′ 10 hpf, g′-i′ 12 hpf, j′ 16 hpf, k′ 22 hpf, l′-m′ 24 hpf, n′-o′ 48 hpf, p′-q′ 72 hpf, r′-s′ 96 hpf. t′-v′ Expression of positive control gene myod1 in the somite at 12 hpf, 16 hpf and 24 hpf, respectively. Scale bar 500 μm

Our negative control probe produced no detectable staining in the corresponding stages of zebrafish embryos and larvae (Fig. 2a′-s′). As a positive control, myod1 expression was analyzed which showed strong staining in muscle precursors at 12, 16, and 24 hpf (Fig. 2t′,u′,v′). Specificity of dennd5b ISH signals was further confirmed using a short non-overlapping antisense probe of dennd5b (255 bp) at multiple stages, which yielded consistent staining patterns (Supplementary Fig. 2a–i), while a sense probe of dennd5b produced no detectable signal in the corresponding stages of zebrafish embryos (Supplementary Fig. 2j–r).

MO knockdown of dennd5b indicates its critical role in early embryo development

To assess the optimal dosage of the dennd5b translation blocker MO, we first conducted microinjections aimed at identifying an optimum dose of the synthetic blocker oligonucleotide, morpholino that would minimize off-target effects and toxicity. We classified the resulting phenotypes based on the severity of morphological alterations in the embryos. A “no phenotype” classification was assigned to embryos that presented with no visible defects, mirroring the appearance of control, uninjected embryos (Fig. 3a). The zebrafish standard control MO-injected embryos (n = 216), which are injected with an identical highest dose (9ng) as dennd5b MO, appeared normal confirming the phenotype observed in dennd5b morphant embryos being specific (Fig. 3b).

Fig. 3.

MO knockdown of dennd5b indicates its critical role in embryo development. a An uninjected zebrafish embryo at 48 hpf exhibits no observable phenotype. b Zebrafish embryos injected with the standard control morpholino show no signs of physical deformity at 48 hpf. c–e A mild phenotype, f–h a moderate phenotype, and i–k a severe phenotype was observed following dennd5b MO knockdown, underscoring the critical role of dennd5b in early zebrafish embryonic development. The scale bar indicates 890 μm. l A graphical representation illustrates the effects of three different dosages of dennd5b MO injection, demonstrating a dose-dependent increase in phenotypic severity, including a compressed head and deformed tail, as compared to embryos injected with the standard control MO

At 24 hpf, typical zebrafish embryos exhibit an elongated trunk and tail that grow in a straight and parallel alignment to the yolk tube structure (black arrowhead in Fig. 3a, b). We categorized the severity of defects in morphants based on the development of both the yolk tube and the tail along the anteroposterior axis. In mild phenotypes, the yolk tube is fully extended but shows a noticeably reduced diameter, suggesting possible growth abnormalities. The tail, in this case, is slightly curved at the tip, indicating deviations from typical morphology, while the head displays minimal compression with subtle deformities. These characteristics help clarify the presentation of the mild phenotype seen in morphants (black arrow in Fig. 3c, d,e). The moderate phenotype also shows an extended yolk tube with decreased diameter; however, the tail curves significantly, failing to extend beyond the yolk tube, while the head displays moderate compression (black arrow in Fig. 3f, g,h). Severe phenotypes present with a grossly compressed head and either a completely deformed or absent tail (black arrow Fig. 3i, j,k). Our results spanned across multiple rounds of microinjection, are graphically represented in Fig. 3l.

At 3 ng dose, the resulting embryos showed 53.4% as no phenotype, 37.77% as mild, 14.91% as moderate, and 5.14% as severe phenotype. For the 6 ng, the observed phenotypes comprised 48.54% no phenotype, 26.90% mild, 17.54% moderate, and 7.01% severe. The highest dosage, 9 ng dennd5b morpholino injection resulted in 46.28% no phenotype, 25.42% mild, 14.87% moderate, and 13.43% severe. Previous studies have documented that a higher dose, exceeding 9 ng per embryo, can lead to off-target effects [42]. To mitigate this risk, we co-injected a zebrafish p53 MO with the dennd5b MO and zebrafish standard control MO. Notably, the p53 MO injection did not alter the phenotypes noted from the dennd5b MO knockdown or zebrafish standard control morpholino knockdown. We also injected 1 ng p53 MO alone to observe its impact in zebrafish embryos which did not produce any phenotype. Additionally, several uninjected embryos (n = 562) were kept aside under identical conditions to rule out potential human error. Remarkably, none of the uninjected control embryos (100%) or those injected with the zebrafish standard control morpholino presented any significant observable phenotypes (Fig. 3a, b).

dennd5b overexpression causes compressed head and tail defects in a dose-dependent manner

The in vitro transcribed and purified mRNAs of dennd5b-Flag-pCS2-mRNA and Flag-pCS2-mRNAs are injected with different dosages to determine the dose-dependent impact of overexpression of mRNAs. The dennd5b-Flag-pCS2-mRNA injected into the blastomere of one-cell zebrafish embryos dose-dependently produced embryos with a compressed head and defects in posterior body formation, particularly in the tail. We classified the changes in zebrafish embryos caused by overexpression based on severity. The “no phenotype” category is for embryos that look normal, and at 24 h after fertilization, they are healthy like uninjected control embryos (refer to Fig. 3a). For the overexpression study control group, zebrafish Flag-pCS2-mRNA injected embryos have a normal head, straight trunk, and tail that grow parallel to the yolk tube (black arrow Fig. 4a). Defects resulting from mRNA injections are categorized by the head structure, yolk tube, and tail’s extent. For the mild phenotype, the yolk tube is fully extended but smaller, with a slightly curved tail and a minimally compressed head (Fig. 4b, c,d). In the moderate phenotype, the yolk tube is also fully extended and smaller, but the tail curves and doesn’t extend beyond the yolk tube, with a more compressed head (Fig. 4e, f,g). For the severe phenotype, the head is severely compressed, and the tail is severely curled upward or absent (Fig. 4h, i,j).

Fig. 4.

Zebrafish dennd5b overexpression resulted in a compressed head and a deformed tail phenotype. a Zebrafish standard control RNA injected embryo showed no major signs of physical deformity (no phenotype) at 48 hpf. b-d A mild phenotype, e-g A moderate phenotype, and h-j A severe phenotype emerged due to dennd5b mRNA injection suggesting its critical role in zebrafish early embryonic development. The scale bar represents 890 μm. k Graphical representation of different dosages of mRNA injection resulted in notable physical deformities of a compressed head and a deformed tail phenotype, which occur in a mRNA dose-dependent manner at 48 hpf

We compiled data from multiple separate microinjections experiments to create a comprehensive graphical representation (Fig. 4k). In the case of overexpressing mRNAs in zebrafish embryos, injecting them into the blastomeres or the space between the cytoplasm and the blastomere can be quite time-consuming. As a result, we had fewer embryos injected with mRNA than those injected with morpholino (MO) injected into the yolk. At the 1–2 cell stage, we injected 130 pg of dennd5b-Flag-pCS2-mRNA either into the blastomeres or at the junction of the blastomere and yolk. The resulting phenotypes were classified as follows: no phenotype (60%), mild phenotype (26.92%), moderate phenotype (6.41%), and severe phenotype (6.41%). For the 260 pg dennd5b-Flag-pCS2-mRNA injections, the observed phenotypes were recorded as, no phenotype (62.22%), mild phenotype (13.33%), moderate phenotype (13.33%), and severe phenotype (11.11%). In our overexpression study, we found that 390 pg of dennd5b-Flag-pCS2-mRNA was the maximum dose that could be used, as exceeding this amount led to significant embryonic lethality. The phenotypes observed with this dosage were categorized as, no phenotype (16.94%), mild phenotype (5.08%), moderate phenotype (16.95%), and severe phenotype (61.01%). To further corroborate our findings on dennd5b-Flag-pCS2-mRNA overexpression, we injected 125 pg, 250 pg, and 350 pg of pCS2-Flag mRNA alongside the dennd5b-Flag-pCS2-mRNA. In comparison to the gene-specific mRNA injections, increasing the doses of pCS2-Flag mRNA did not result in any significant phenotypic changes, confirming the specificity of the effects attributed to dennd5b mRNA on embryonic development. For every round of injections, we kept a separate set of uninjected eggs (n = 106) to rule out human error in microinjection, assess egg quality, and ensure the viability of the eggs.

dennd5b knockdown phenotype can be partially rescued by dennd5b mRNA

To further confirm that the structural phenotype observed in zebrafish embryos results specifically from the loss of dennd5b function, we conducted rescue experiments. We co-injected the knockdown MO targeting the 5′ untranslated region (UTR) of dennd5b including the start codon and the adjacent downstream codon along with an in vitro transcribed dennd5b mRNA lacking the MO binding site (19 base-mismatched) into one-cell stage embryos. This design ensured that the rescue mRNA would not be affected by the MO, allowing for the assessment of dennd5b’s specific role in development.​ Zebrafish embryos injected with 7.2 ng of morpholino alone, 23.8% showed no phenotype, 47.6% displayed mild defects, and 19.0% had moderate defects. Co-injection of 7.2 ng MO with 90 pg of dennd5b mRNA, which lacks the morpholino binding site, resulted in an improved phenotype distribution with 42.1% of embryos exhibited no phenotype, 31.6% were mild, 21.1% moderate, and only 5.3% were severe supporting the specificity of the knockdown and confirming that the observed defects are due to loss of dennd5b function (Fig. 5). When the MO dose was increased to 10.8 ng, 50.0% of embryos showed no phenotype, while 31.6% exhibited mild and 15.8% moderate phenotypes. In the group injected with 10.8 ng MO and 90 pg mRNA, 20.8% of embryos showed no phenotype, 54.7% were mild, 13.2% moderate, and 11.3% severe. As expected, all embryos in the uninjected control group developed normally, with 100% showing no phenotype. These results also indicate that co-injection with dennd5b mRNA partially rescues the MO-induced phenotype with a lower dose of MO (7.2 ng) than a higher dose (10.8 ng) (Fig. 5).

Fig. 5.

Graphical representation shows partial rescue of MO dose dependent phenotype by injecting a rescue RNA. A bar graph showing the lower dose of MO and rescue RNA injection leads to a reduction in physical deformities specifically mild to moderate phenotypes as observed at 48 hpf

dennd5b plays a role in convergence and extension (CE) movements of zebrafish embryos

Body axis extension defects can be quantified at the early somite stage (ss) by measuring gap angles defined as the angle formed between the head, mid-yolk, and tail of embryos [43] as denoted in white dashed lines (Fig. 6a). Whereas control embryos displayed a normal gap angle, both dennd5b morphants (Fig. 6b) and overexpressed embryos (Fig. 6c) exhibited increased gap angles. To assess whether dennd5b knockdown or overexpression disrupted hypoblast cell migration, we performed whole-mount in situ hybridization for tbxta and myod1 at early somitogenesis (7-8ss). While expression of both markers remained robust, their spatial organization was altered. Compared to controls (green arrow, Fig. 6d), axial tissues were significantly broader in both dennd5b morphants (Fig. 6e) and overexpressed embryos (Fig. 6f). In addition, somite morphology was disrupted, with somites extending less posteriorly and adopting a more lateral orientation, resulting in reduced spacing between somites (Fig. 6e, f). Compared to the control embryos (Fig. 6g), notochord cells expressing tbxta also fail to intercalate into a compact midline, leading to a broadened notochord horizontally and shortened vertically (Fig. 6h, i). Together, these results suggest that dennd5b function influences mesendoderm migration by the end of gastrulation.

Fig. 6.

dennd5b plays a role in CE movement. Zebrafish embryos at the one-cell stage were injected with dennd5b MO, and dennd5b mRNA. These embryos were either viewed from the dorsal side, ventral side and or laterally. a-c The angle between the anterior-most and the posterior-most ends of the body axis was measured (scale bar, 420 μm). d-a′ Knockdown or overexpression of dennd5b affects the expression of CE movement-related genes as determined by WISH. The expression pattern of the neuroectoderm marker dlx3b, notochord marker tbxta, somite marker myod1 and rhombomere 3,5 of the hindbrain marker krox20 were determined at 7–8 somite stage (12 hpf). The expression pattern of the prechordal plate marker ctslb was determined at 1–3 somite stage (10–11 hpf) scale bar 500 μm

To further assess CE of neuroectodermal cells, we analyzed neural plate formation using dlx3b in situ hybridization at the early somite stage (7-8ss). In control embryos, posterior views showed neural plate boundaries (red arrows) running parallel to the notochord (Fig. 6j), and anterior views revealed a continuous neural plate boundary encircling the embryo (Fig. 6m). In contrast, both dennd5b morphants and overexpressed embryos displayed disrupted neural plate morphology, with posterior boundaries misaligned (Fig. 6k, l) and anterior boundaries reduced/disrupted (Fig. 6n, o).

We next examined rhombomere patterning using krox20, a marker of rhombomeres 3 and 5. Although expression was detectable in both dennd5b knockdown and overexpressed embryos, compared to the control embryos (Fig. 6p, s), the rhombomeres domains were reduced (double arrows, Fig. 6q, r) and an abnormal expansion of these domains are evident in morphant and overexpressed embryos (Fig. 6t, u), suggesting impaired CE within the hindbrain. Finally, we assessed the prechordal plate formation using ctslb. Compared to controls (Fig. 6v, y), dennd5b morphants exhibited a shortened prechordal plate along the mediolateral axis (Fig. 6w, z), whereas overexpressed embryos displayed an extended prechordal plate (Fig. 6x, a′). Taken together, these data demonstrate that both knockdown and overexpression of dennd5b disrupt convergence and extension movements in multiple germ layers, resulting in broad defects in axis elongation and tissue morphogenesis.

Discussion

DENND5 was initially identified through a yeast two-hybrid screen as an interacting partner of the Rab6 GTPase, implicating it in intracellular trafficking [44]. Structurally, the protein comprises N-terminal DENN domains, which serve as GDP-GTP exchange factors (GEFs) for Rab proteins such as Rab39 [25], and C-terminal RUN domains (RUN1 and RUN2), separated by a PLAT domain. RUN domains are evolutionarily conserved modules involved in mediating protein-protein interactions, particularly with cytoskeletal elements and small GTPases [27, 45, 46]. Sequence comparison of DENND5 orthologs in vertebrates, such as human, chimpanzee, mouse, and zebrafish revealed strong conservation in the RUN2 domain, suggesting its central role in DENND5 function (Fig. 1b). Intriguingly, despite the paralogous relationship between dennd5a and dennd5b in zebrafish, these genes share only ~ 68.5% sequence identity, pointing to potential subfunctionalization or neofunctionalization. However, their RUN2 domains are nearly identical (Fig. 1c), indicating that this region likely retains core functionality conserved across vertebrate evolution. Prior studies have shown that RUN2 of DENND5 binds Sorting Nexin 1 (SNX1), forming transient complexes that mediate vesicle trafficking from early endosomes to the trans-Golgi network [47, 48]. This reinforces the hypothesis that DENND5 plays critical, possibly non-redundant roles in intracellular transport during embryogenesis.

Our temporal and spatial expression analyses demonstrate that dennd5b is expressed continuously throughout zebrafish embryogenesis, from the earliest cleavage stages through larval development. Semi-quantitative RT-PCR revealed maternal deposition of dennd5b transcripts and persistent expression up to at least 72 hpf without significant fluctuations (Fig. 2a, b). Whole-mount ISH further revealed a broad distribution of dennd5b transcripts during cleavage, blastula, and gastrulation, consistent with maternal expression. However, beginning at somitogenesis, expression became progressively enriched in the developing nervous system, with strong localization to the brain, neural tube, tail bud and later to distinct forebrain, midbrain, and hindbrain locations. This transition from ubiquitous to tissue-specific expression suggests that dennd5b functions in both early general cellular processes and later in neural development (Fig. 2c-u). These findings are consistent with RNA-seq profiles from zebrafish embryos and larvae which showed a high expression during blastula to 75% epiboly and at prim-5 stage (24 hpf) [49]. These suggest an essential role for dennd5b in early developmental events, possibly related to cell polarity and cytoskeletal organization which are critical during gastrulation cell movements. The spatial expression pattern parallels key signaling pathway known to regulate brain development, including non-canonical Wnt signaling [50, 51].

Functional analysis of dennd5b using MO knockdown revealed dose-dependent phenotypes, including brain compression, and tail deformity, all indicative of disrupted neural development and mesoderm patterning. Consistent with its expression pattern, dennd5b was enriched in the developing brain and neural tube, and perturbation of its function disrupted both neurogenesis and tail formation. Overexpression produced even more pronounced cranial and axial deformities, underscoring the importance of maintaining tightly regulated dennd5b dosage during embryogenesis (Figs. 3 and 4). These reciprocal phenotypes suggest that both insufficient and excessive dennd5b activity can perturb critical developmental pathways.

Previous studies have shown that both gain-of-function and loss-of-function of non-canonical Wnt signaling components can lead to abnormal CE movements during zebrafish gastrulation [52–54]. During zebrafish gastrulation, CE movements occur primarily in the anterior and posterior mesendoderm and neuroectoderm cell layers [55]. The mesendoderm, also referred to as the hypoblast, comprises axial and paraxial cell populations [56]. To investigate whether dennd5b similarly regulates CE processes, we examined the expression of marker genes such as krox20, myod1, ctslb, dlx3b, and tbxta in dennd5b morphants and over-expressed embryos at the somite stages. Whole-mount in situ hybridization revealed that compared to controls, both knockdown and overexpression embryos exhibited altered rhombomere patterning (krox20), neural plate morphology (dlx3b), compressed and laterally expanded somites (myod1), undulated notochords (tbxta), and shortened or expanded prechordal plates (ctslb) (see Fig. 6).

These tissue-specific defects correspond closely with the endogenous expression domains of dennd5b. During early development, dennd5b is broadly expressed but becomes enriched in the brain, neural tube, and tail bud by the segmentation stages, and later localizes to distinct forebrain, midbrain, hindbrain regions. The overlap between expression domains and affected structures suggests that dennd5b plays a direct role in coordinating CE movements within both mesodermal and neuroectodermal lineages. Together, these findings indicate that precise regulation of dennd5b levels is essential for proper axis elongation and morphogenesis.

Unlike CRISPR/Cas9-induced mutants, MO knockdowns typically do not trigger compensatory mechanisms, such as transcriptional upregulation of paralogs or related pathways [57]. While MO technology is a powerful tool for transient gene knockdown in zebrafish, it is not without limitations. MOs may produce exaggerated or non-physiological phenotypes that differ from stable gene knockouts. To overcome this, in our study, we used a standard control zebrafish MO that did not exert a significant effect on the phenotype of the injected control groups. Additionally, our MO targeted the 5′ untranslated region (UTR) and start codon of dennd5b. The in vitro transcribed mRNA used for our rescue experiments lacked the MO-binding sequence, minimizing the likelihood of direct MO binding and ensuring specific rescue. MOs can also elicit off-target effects, often mediated by p53 activation. To overcome this, we used p53 MO co-injection controls to minimize nonspecific cell death, subtle phenotypes might still result from non-specific toxicity or interactions with unintended mRNAs.

Recent studies implicate DENND5B in neurological conditions, including epilepsy where DENND5B levels were reduced in both human patients with temporal lobe epilepsy and chronic epileptic mice [21]. Moreover, overexpression of DENND5B suppressed seizure activity, while knockdown exacerbated it. These findings align with our findings on dennd5b which is prominently expressed in the developing brain, and its overexpression leading to pronounced neural defects in zebrafish.

Interestingly, in humans, mutations in the paralogue DENND5A are associated with developmental epileptic encephalopathies [58], suggesting a broader role for the DENND5 gene family in neural development and function. A recent study reported that the De novo variants in DENND5B cause a neurodevelopmental disorder [59]. Our zebrafish model provides a useful in vivo system for mechanistic dissection of DENND5B function in neurodevelopment and may ultimately contribute to the development of therapeutic strategies targeting these pathways. This study identifies dennd5b as a critical regulator of early zebrafish embryogenesis. Our expression and functional data support dennd5b’s involvement in brain patterning and axial morphogenesis, likely mediated through its conserved RUN domains.

Our findings from zebrafish models provide a foundational platform to examine how mutations in dennd5b might contribute to these neurodevelopmental syndromes. This study provides a foundation for future research into the functional roles of dennd5b in neural development. In the future, we aim to further validate the role of dennd5b by CRISPR/Cas9-based knockout models to complement our MO knockdowns, to assess genetic compensation and the full range of phenotypic consequences. Given the conserved nature of DENND5B and its potential implications in neurodevelopmental diseases, investigating its function in zebrafish models could provide valuable insights into the pathophysiology of human epileptic encephalopathies.

Conclusion

In conclusion, our findings highlight the conservation of DENND5B across species and underscore its critical role in early development. The spatial and temporal expression patterns, along with functional data from MO knockdowns and overexpression studies, suggest that dennd5b is essential for proper embryogenesis in zebrafish. Furthermore, the parallels between zebrafish and human data point to the potential involvement of DENND5B in neurodevelopmental disorders, particularly epilepsy. Further functional and rescue experiments will be key to elucidating the precise mechanisms by which dennd5b contributes to neural development.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Sharmin Hasan conceived and designed the study, conducted the RT-PCR, in situ hybridization, microinjection experiments, and managed data acquisition and interpretation. Alicia Mendoza was responsible for the microinjection experiments, in situ hybridization, data acquisition, and analysis. Khaled Mohamed Nassar performed initial data collection, in situ hybridization experiments. Magdalen Marston carried out the cloning experiments. Andre Gil performed the semi-quantitative RT-PCR data analysis. All authors contributed to the experiments, data collection, analysis, and collaboratively wrote the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Conflict of interest

The authors declare no competing interests.