Neuronal expression of ndst3 in early zebrafish development is responsive to Wnt signaling manipulation

Abstract

Heparan sulfate proteoglycans (HSPGs) are constituents of the cell surface and extracellular matrix and are vital for various activities within the cell. The N-deacetylase/N-sulfotransferase (heparin glucosaminyl) family of enzymes, or NDST, modifies heparan sulfate (HS) by catalyzing both the N-deacetylation and the N-sulfation of N-acetylglucosamine residues. In zebrafish, a single ndst3 gene is an orthologue of both mammalian NDST3 and NDST4 genes. The role of ndst3 in zebrafish development has not been investigated and such study may provide insight into the role(s) of both mammalian orthologues. Here, we characterized expression of ndst3 during early development in zebrafish and found it to be predominately neuronal. We found that expression of ndst3 is sensitive to Wnt signaling manipulation, with stimulation of the Wnt pathway resulting in robust expansion of ndst3 expression domains. Finally, using CRISPR/Cas9 genome editing, we mutagenized the ndst3 gene and isolated an allele, ndst3^nu20, resulting in a frameshift and premature protein truncation. We discovered Ndst3 is not essential for zebrafish survival as ndst3^nu20 homozygous mutants are viable and fertile.

1. Introduction

Heperan sulfate proteoglycans (HSPGs) are found at the cell surface and within the extracellular matrix (ECM) where they are involved in numerous cellular activities, such as morphogen gradient formation and cell signaling (1). HSPGs consist of a core protein to which heperan sulfate (HS) glycosaminoglycan side chains are attached. These HS side chains, composed of repeating residues of D-glucuronic acid and N-acetyl-D-glucosamine, undergo modification by various enzymes to create specific types of sulfanation patterns which are key in determining the tissue-specific interactions and functions of HSPGs. The enzyme N-deacetylase/N-sulfotransferase (heparin glucosaminyl), or NDST, is a bifunctional enzyme important to HS modification that acts as the first step in HS sulfanation by catalyzing both the N-deacetylation and the N-sulfation of N-acetylglucosamine residues (1) (Fig. 1).

Fig. 1. Enzymatic action of Ndst3.

Heparan sulfate proteoglycans (HSPGs) are made up of a core protein to which heparan sulfate glycosaminoglycan (GAG) side chains are attached. The ndst3 gene encodes the bifunctional enzyme N-deacetlyase/N-sulfotransferase (heparin glucosaminyl) which is involved in the first step of heparan sulfate modification. The Ndst3 enzyme catalyzes both the N-deacetylase and N-sulfanation of the N-acetyl glucosamine in the GAG side chains.

The NDST enzymes are divided into three main groups: NDST1, NDST2, and a combined group of NDST3 and NDST4 (Fig. 2). Of the four NDST genes found in humans, expression of NDST1 and NDST2 is detected in most tissue types, whereas expression of NDST3 and NDST4 is limited primarily to brain tissue (2-4) (Fig. 3). Expression of NDST1 is highest in the brain, spinal cord, spleen, liver, and pancreas while expression of NDST2 is highest in bone marrow and lymphoid tissue (2-4) (Fig. 3). Expression of NDST3 is found predominantly in the brain and eye, with some expression also detected in the thymus, spleen, and liver; NDST4 expression is almost exclusively limited to the brain (2-4) (Fig. 3). Homozygous missense mutations and compound heterozygous missense mutations in NDST1 have been associated with intellectual disability (ID) (5, 6) and ID with respiratory problems (7).

Fig. 2. Phylogenetic analysis of NDST enzymes.

The NDST proteins can be divided into three branches: NDST1 (yellow), NDST2 (green), NDST3 and NDST4 (blue). All proteins were retrieved from the National Center for Biotechnology Information (NCBI) database (35). The confidence level of each node was calculated with 100 boostrap replicates. Details of phylogenetic reconstruction described in Methods.

Fig. 3. Expression of the NDST genes in human.

Expression of NDST1 and NDST2 are found in many tissues (A-B), with the highest levels of NDST1 expression detected in the brain, spleen, liver, pancreas, and skin (A) while robust NDST2 expression is found mainly in bone marrow and lymphoid tissue (B). Expression levels of NDST3 and NDST4 are lower and restricted to specific tissue types (C-D). Expression of NDST3 is found mostly in the brain and eye, with some expression detected in lymphoid tissue and liver (C) while expression levels of NDST4 are limited almost exclusively to the brain (D). X-axis data color-coded by organ system: blue = brain; turquois = eye; red = bone marrow and lymphoid tissues; orange = liver and gallbladder; green = pancreas; pink = breast and female reproductive system; black = skin; purple = endocrine tissue. nTPM = normalized protein-coding transcripts per million. Details of graph construction described in Methods.

In mouse, expression of Ndst1 and Ndst2 has similarly been detected in most tissues, while expression of Ndst3 and Ndst4 appears more restricted to nervous tissue (8, 9). Mice lacking Ndst1 display perinatal lethality, with defects consisting of respiratory failure (10, 11), brain abnormalities (12), ocular deformities (13), craniofacial defects (12), and delayed endochondral bone formation (14). Interestingly, Purkinje cell-specific deletion of Ndst1 in Ndst1^f/f;L7-Cre⁺ mice has no effect on brain morphology, behavior, or locomotion (15). Mice deficient in Ndst2 are viable and fertile but display connective tissue-type mast cell defects (16, 17). Female compound Ndst1^f/f;Ndst2^−/−;L7-Cre⁺ mice exhibit strongly impaired reproductive ability, but display no other discernable defects (15). Compound null Ndst1/Ndst2 mice die during early development (18). Null Ndst3 mice are viable and fertile and show only subtle behavioral abnormalities consisting of a reduced anxiety-related behavior (9). Compound null Ndst2/Ndst3 mice appear healthy, while compound null Ndst1/Ndst3 mice die upon birth and display strong brain and craniofacial defects (9). These defects often exceed in severity and frequency to those seen in Ndst1 null mice, indicating Ndst1 and Ndst3 may have partially redundant functions related to craniofacial and brain development. Mice lacking Ndst4 are viable and fertile and show only a subtle defect in colonic epithelial homeostatis (19). The double Ndst3/Ndst4 mouse has not been described to date.

Less is known of the roles ndst genes play in zebrafish development. Zebrafish have five ndst genes: ndst1a, ndst1b, ndst2a, ndst2b, and ndst3. Alignment of zebrafish Ndst proteins show a high level of conservation (Fig. 4). Expression of the zebrafish ndst genes during early development is distinct with partially overlapping expression patterns (20). Morpholino-induced knockdown of ndst1b results in impaired vasculogenesis (21), malformed cranial facial cartilage, and shortened pectoral fins (20). To date, no mutants in any of the zebrafish ndst genes have been described.

Fig. 4. The Ndst proteins in zebrafish.

Multalin comparison of the five zebrafish Ndst proteins (Ndst1a, Ndst1b, Ndst2a, Ndst2b, and Ndst3). The five Ndst proteins in zebrafish are highly conserved, differing mainly in their N-terminal region. High amino acid consensus values of 90% or greater are marked in red. Low amino acid consensus values of 50% or greater are marked in blue.

The genomic region surrounding ndst3 in zebrafish shares synteny with mammals, avians, and amphibians (Fig. 5), as well as reptiles and other teleosts (20). The orientation of the genes within this region also remains conserved (Fig. 5). Interestingly, there appears to have been a local duplication of Ndst3 in mammals, avians, amphibians, and reptiles (20) that gave rise to the Ndst4 gene (Fig. 5). While zebrafish have paralogues of ndst1 and ndst2, only one homologue of ndst3 exists. Zebrafish ndst3 is orthologous to Ndst3 and Ndst4, and zebrafish Ndst3 shares a greater than 70% amino acid sequence homology to tetrapod NDST3 and NDST4 (20).

Fig. 5. Genomic synteny of ndst3 region.

Schematic illustrating the chromosomal location of the ndst3 gene and surrounding region in various vertebrate species. Arrowed boxes represent the orientation of the genes. Numbers represent the distance in mega base pairs between genes. METTL14 (methyltransferase 14, N6-adenosine-methyltransferase subunit); PRSS12 (serine protease 12); NDST3 (N-deacetylase/N-sulfotransferase 3), NDST4 (N-deacetylase/N-sulfotransferase 4); UGT8 (UDP glycosyltransferase 8).

We reasoned that a mutation affecting ndst3 function in zebrafish would provide not only new insight into the role of ndst3 in zebrafish, but also provide important information on the roles of Ndst3 and Ndst4 in mammals. Given the importance of HSPGs in Wnt signaling, we investigated the relationship between ndst3 expression and Wnt signaling. In this study, we characterize the expression of ndst3 in zebrafish during early development and find that expression of ndst3 is responsive to stimulation of the Wnt signaling pathway. In addition, to our knowledge, we describe the first-reported knockout of ndst3 in zebrafish.

2. Results

2.1. Strong neuronal expression pattern of ndst3 observed during early development

In order to characterize ndst3 expression during early zebrafish development, we used whole mount in situ hybridization (WISH). We detected no maternal contribution of ndst3 at the 8-cell stage (Fig. 6A), and found ndst3 to be expressed in a predominately neuronal pattern during early development (Figs 6B-G). At 8 somites (s), ndst3 expression was observed in the forebrain, midbrain, rhombomeres and neural tube (Fig. 6B). At 17 s, the strong neuronal expression pattern intensified (Fig. 6C). In addition, ndst3 expression was observed in the forming cloaca at 17 s (Fig. 6C). At 24 hours post fertilization (hpf), strong expression continues in the spinal cord (Figs 6D, E), as well as specific regions of the brain, such as the telencephalon/diencephalon boundary and tegmentum of the midbrain (Figs 6D, F, G). Faint expression was seen in the lens (Fig. 6F).

Fig. 6. Predominately neuronal ndst3 expression pattern is observed during early development.

Maternal expression of ndst3 is not observed at 8-cell (A). Expression of ndst3 is seen in the forebrain (fb), midbrain (mb), rhombomeres (rh), and neural tube (nt) at 8 s (B), 17 s (C), and 24 hpf (D-G). Expression is observed in the cloaca (cl) at 17 s (C). At 24 hpf, expression can be seen in specific regions of the brain, such as the telencephalon/diencephalon (tl/dn) boundary and tegmentum (tg) of the midbrain (D, F, G). Anterior to the left. Lateral views (A, B, D, E, G), dorsal view (F), dorsal view of head with lateral view of trunk (C). Scale bar: 100μm. nc = notochord, ov = otic vesicle

2.2. Expression of ndst3 at 48 hpf is sensitive to Wnt signaling manipulation

The importance of HSPGs in signaling and development is well known (1). The role for HS in development was first demonstrated by the Drosophila mutant sulfateless (sfl), encoding the only known Ndst in fly, characterized by perturbed Wingless (Wg), hedgehog (Hh), Decapentaplegic (Dpp) and fibroblast growth factor (Fgf) signaling (22). Given that Wnt signaling plays a crucial role in neurodevelopment (23), and we observed strong expression of ndst3 in the zebrafish brain (Figs. 6B-D, F, G), we investigated what affect manipulation of Wnt signaling may have on ndst3 expression.

We characterized ndst3 expression in the zebrafish at 48 hpf using WISH and found strong expression in the rhombomeres and within the proliferating zones of the brain (Figs 7A, C, E). Brief treatment of embryos at 48 hpf with lithium chloride (LiCl), which activates Wnt signaling through inhibition of GSK3β (23, 24), resulted in robust expansion of ndst3 expression domains, particularly in the tectum (Figs 7B, D, F). These results indicate that expression of ndst3 in the zebrafish at 48 hpf is responsive to stimulation of the Wnt signaling pathway.

Fig. 7. Neuronal expression of ndst3 at 48 hpf is sensitive to Wnt manipulation.

Non-manipulated embryos at 48 hpf show expression of ndst3 in the proliferating zones of the brain. Embryos treated for eight minutes with 0.3M LiCl at 48 hpf and collected after a two-hour incubation show an expanded ndst3 expression domain, with strong ndst3 expression observed in the tectum (tc) (B, D, F). Anterior to the left. Lateral views (A, B). Dorsal views (C-F)

2.3. CRISPR/Cas9-induced mutagenesis of ndst3 results in viable and fertile mutants

Using CRISPR/Cas9 genome editing, we mutagenized the ndst3 gene and isolated an allele, ndst3^nu20, which results in a 10 bp deletion in the first coding exon (Fig. 8A), resulting in a frameshift and premature protein truncation (Fig. 8A-C). The ndst3^nu20 mutant protein is 273 amino acids (aa), opposed to wildtype Ndst3 which is 874 aa (Fig. 8C), and is truncated well before the N-deacetylase active site near amino acid 477 (Fig. 8B, C). Interestingly, homozygous ndst3^nu20 fish are phenotypically normal and fertile.

Fig. 8. Creation of ndst3 knockout zebrafish.

(A) The ndst3^nu20 allele is the result of a 10 bp deletion, resulting in a frameshift and protein truncation, I247AfsX28. (B) Schematic of NDST3 illustrating protein domains and enzyme active sites. The deacetylase and sulfotransferase sites are marked with red asterisks. (C) The Multalin comparison of human NDST3 (hNDST3), mouse NDST3 (mNDST3), human NDST4 (hNDST4), mouse NDST4 (mNDST4), and zebrafish Ndst3 (zNdst3). The N-deacetylase domain is marked in tan. The N-sulfotransferase domain is marked in purple. The yellow highlighted region represents the ndst3 CRISPR target site. High amino acid consensus values of 90% or greater are marked in red. Low amino acid consensus values of 50% or greater are marked in blue. The region containing the N-deacetylase active site is marked with an arrow.

3. Discussion

3.1. Expression of ndst3 is predominately neuronal in zebrafish during early development

Similar to what was previously reported by Filipek-Gornoick et al, we found ndst3 to be expressed within the central nervous system (CNS) at 24 hpf (20). Although Filipek-Gorniok et al reported maternal expression of ndst3 at the 2-cell stage, we detected no maternal contribution of ndst3 at the 8-cell stage (Fig 6A). Consistent with our results, RNA-seq data indicates an extremely low level of maternally deposited ndst3, with only 1 transcript per million (TPM) reported at the 2-cell stage and 2 TPM seen at both the 128-cell and 1K-cell stage (25). The strong expression we observed of ndst3 in the developing brain is consistent with the proposed role of NDST3 in psychiatric disease etiology (see below).

3.2. Ndst3 activity is not essential for zebrafish survival

In the NDST proteins, the domain responsible for N-deacetylase activity is predicted to be in the N-terminal half of the protein (20, 26), with the active site known to be located near the cysteine at position 486 (20, 27, 28) (Fig. 8B, C). The domain responsible for N-sulfotransferase activity is known to be located in the carboxyl half of the protein (20, 29) (Fig. 8B, C). The ndst3^nu20 mutant protein is truncated well before the N-deacetylase active site (Fig. 8B, C), indicating that the mutant protein should lack both N-deacetlylase and N-sulfurtransferase activity and most likely be non-functional. This suggests that Ndst3 activity is not essential for zebrafish survival under laboratory conditions.

One potential reason ndst3^nu20 mutants are viable may be due to compensatory effects provided by other ndst genes. In the mouse, Ndst1 and Ndst3 are indicated to have partially redundant functions (9). Therefore, it is possible that ndst1a and/or ndst1b may provide compensatory function for ndst3 in zebrafish development; creation and analysis of compound mutant lines would help address this issue. Another reason ndst3^nu20 mutants are viable and do not display an obvious phenotype may be due to genetic compensation provided by an upregulation of genes encoding enzymes with similar activity (30).

3.2. Ndst3 activity may play a role Wnt signaling and neurodevelopment

The results found in ndst3^nu20 zebrafish mutants are similar to those found in Ndst3-null and Ndst4-null mouse mutants, both of which are viable and fertile with only subtle defects observed in behavior and the colon, respectively (9, 19). It is intriguing that Ndst3-null mouse mutants are viable and fertile, but display behavioral abnormalities (9). In large-scale genome-wide association studies (GWAS), NDST3 has been linked to both schizophrenia (SZ) and bipolar disorder (BD) in Ashkenazi Jewish (31) and Han Chinese populations (32). Although ndst3^nu20 zebrafish mutants do not display obvious morphological defects, it is feasible that they may too possess behavioral abnormalities. As many assays and tracking tools exist to study behavioral responses in the zebrafish (33), investigating the behavior of ndst3^nu20 mutants may be an attractive area for future research. Our results indicate that ndst3 expression within the zebrafish brain is responsive to stimulation of the Wnt signaling pathway. The link between Wnt signaling and heparan sulfate modifications is well established (34-36). Observed rapid upregulation of ndst3 expression may be part of the Wnt regulatory loop or change the neuronal precursors sensitivity to other HS dependent cytokines. Interestingly, altered Wnt signaling has been found in patients suffering from SZ and BD (37, 38). Given the potential role of Wnt signaling in the pathogenesis of SZ (39) and BD (40) and the proposed link of NDST3 with SZ (31, 32) and BD (32), further investigation into the role Ndst3 may play in Wnt signaling and neurodevelopment is warranted.

4. Experimental Procedures

4.1. Phylogenetic analysis

NDST protein sequences from the following species were collected from the National Center for Biotechnology Information (NCBI) database (41): human (NP_001534.1, NP_003626.1, NP_004775.1, NP_072091.1); mouse (NP_001335029.1, NP_034941.2, NP_112463.2, NP_072087.2), chicken (XP_004945002.1, XP_4216132.2, XP_004941301.4, XP_015131840.30), western clawed frog (NP_001011159.1, NP_001136290.1, NP_001120519.1, XP_031758874.1), and zebrafish (NP_001311428.1, NP_001311384.1, XP_005156429.1, XP_009305632.2, XP_009305879.1).

The online software platform Phylogeny.fr (42) was used for analysis of the NDST proteins. Amino acid sequences were aligned using MUSCLE v. 3.8.31 (43) and configured for highest accuracy. The Gblocks v.0.91b (44) program was used to remove poorly aligned and/or divergent regions. A maximum-likelihood phylogenetic tree was made using the PhyML v 3.1 (45) using the default amino acid substitution model (WAG). Bootstrap replicates of N=100 were performed to assess reliability of branch topology. The final phylogenetic tree was graphed using TreeDyn v198.3 (46).

4.2. Chromosomal synteny analysis

The Ensembl database (release 107) (47) was used to collect data, including chromosomal location and orientation, of NDST3, NDST4, UGT8 (UDP glycosyltransferase 8), PRSS12 (serine protease 12), and METTL14 (methyltransferase 14, N6-adenosine-methyltransferase subunit) for human, mouse, chicken, western clawed frog, and zebrafish.

4.3. Human NDST Expression Data Analysis

The Human Protein Atlas database (www.proteinatlas.org Version 21.1) was used to collect RNA expression data for the NDST genes (2-4). The top eight data points displaying the highest expression levels for each NDST gene were collected from the consensus dataset. Using this selected group of tissues samples, the level of each NDST gene for each selected tissue type was collected from the consensus dataset. The consensus dataset, representing normalized RNA expression levels of 55 tissue types, was created by combining the Human Protein Atlas (HPA) RNA-seq and Genotype-Tissue Expression (GTEx) RNA-seq datasets (2-4).

4.4. Whole Mount in situ hybridization

The 469 bp template for the ndst3 anti-sense RNA probe was amplified from wildtype RNA using a primer tailed with the T7 RNA polymerase promoter (5’-CCCTAAAGAACTAGACAAAAGCA-3’ and 5’-TAATACGACTCACTATAGGGCCAGTCGATGCCTTTATGGT-3’). Whole-mount in situ hybridization was performed as previously described (48), using low stringency conditions.

4.5. LiCl experiment

The LiCl experiments were performed as previously described (49). Wildtype clutches of embryos were divided into two groups: treated and untreated. The treated embryo groups were submerged in 0.3M LiCl egg water at room temperature for eight minutes, while untreated embryos were submerged in room temperature egg water. After eight minutes, treated and untreated embryos were transferred to separate clean egg waters and incubated at 37°C for two hours. After a two-hour incubation, all embryos were fixed in 4% paraformaldehyde.

4.6. CRISPR/Cas9 genome editing

Genome editing was performed as previously described (50, 51). We used linearized plasmid pCS2-nCas9n (Addgene plasmid #47929) (51) as a template for Cas9 mRNA synthesis (mMessage mMachine ThermoFisher Scientific). The gRNA template was made by PCR and RNA was synthesized using the MEGAshortscript T7 kit (ThermoFisher Scientific). 30-40 pg of gRNA was co-injected with 150 pg of Cas9 mRNA per embryo at one cell stage. The ndst3^nu20 mutant line was created by targeting sequence (5’- GGGGCAGCCCAGATCATCCC -3’) in exon 2 of ndst3 (Fig 3A). Indels in the ndst3 gene were identified by a loss of MboI restriction enzyme site. The G0 founders positive for mutation were outcrossed with WT fish to produce F1 heterozygotes, which were then analyzed using sequencing to identify the mutation I247AfsX28. Individuals carrying desirable, identified mutations were outcrossed with wildtype to generate F2 heterozygous family.

4.7. Genotyping

The ndst3^nu20 genotyping protocol is based on the elimination of an MboI site in the mutant ndst3^nu20 allele. Primers 5’- TATTCTCAGCTTGGACAGGACA -3’ and 5’- AGCTTTCTGCCTGTTAAAAACG -3 were used to amplify a 632bp segment of ndst3^nu20 prior to diagnostic MboI restriction enzyme digestion. Resulting fragments of 480bp and 152 bp were separated using a 2.0% agarose gel.

4.8. Zebrafish lines and Maintenance

All fish used in this study where raised and cared for in accordance with approved protocol by Ann & Robert H. Lurie Children’s Hospital of Chicago Institutional Animal Care and Use Committee (IACUC# 13-036) and complied with NIH standards provided in the “Guide for the Care and Use of Laboratory Animals”. The following wild type and mutant fish lines were used: AB (ZFIN ID: ZDB-GENO-960809-7), TU (ZFIN ID: ZDB-GENO-990623-3), and ndst3^nu20.

4.9. Multalin Analysis

All protein sequences were obtained from the National Center for Biotechnology Information (NCBI) database (41): zebrafish - Ndst1a (NP_001311428.1), Ndst1b (NP_001311384.1), Ndst2a (XP_009305632.2), Ndst2b (XP_005156429.1), and Ndst3 (XP_009305879.1); human – NDST3 (NP_004775.1) and NDST4 (NP_072091.1); mouse – NDST3 (NP_112463.2) and NDST4 (NP_072087.2). These sequences were then aligned using MultAlin v 5.4.1 (52), using an absolute scoring method and Blosum62 symbol comparison table with a gap opening default value of 12 and gap extension default value of 2.