ACAM, a novel member of the neural IgCAM family, mediates anterior neural tube closure in a primitive chordate

Abstract

The neural IgCAM family of cell adhesion molecules, which includes NCAM and related molecules, has evolved via gene duplication and alternative splicing to allow for a wide range of isoforms with distinct functions and homophilic binding properties. A search for neural IgCAMs in ascidians (Ciona intestinalis, Ciona savignyi, and Phallusia mammillata) has identified a novel set of truncated family members that, unlike the known members, lack fibronectin III domains and consist of only repeated Ig domains. Within the tunicates this form appears to be unique to the ascidians, and it was designated ACAM, for Ascidian Cell Adhesion Molecule. In C. intestinalis ACAM is expressed in the developing neural plate and neural tube, with strongest expression in the anterior sensory vesicle precursor. Unlike the two other conventional neural IgCAMs in C. intestinalis, which are expressed maternally and throughout the morula and blastula stages, ACAM expression initiates at the gastrula stage. Moreover, C. intestinalis ACAM is a target of the homeodomain transcription factor OTX, which plays an essential role in the development of the anterior central nervous system. Morpholino (MO) knockdown shows that ACAM is required for neural tube closure. In MO-injected embryos neural tube closure was normal caudally, but the anterior neuropore remained open. A similar phenotype was seen with overexpression of a secreted version of ACAM. The presence of ACAM in ascidians highlights the diversity of this gene family in morphogenesis and neurodevelopment.

Keywords: Ascidian; IgCAM; Neural cell adhesion molecules; Neural tube defect; OTX.

Figure 1

Ciona intestinalis has three homologs of neural Ig-containing cell adhesion molecules (IgCAMs) A. Diagram of the neural homophilic binding IgCAMs in humans compared to Ciona intestinalis, Ciona savignyi, Phallusia mammillata and Oikopleura dioica. Alternatively spliced variants of the human ortholog are also shown. B. Neighbor joining analysis of amino acid sequences of the first 4 (or less) Ig domains of IgCAM members from Anopheles gambiae (Ag), O. doioica (oiko), Hydra vulgaris (Hv), Danio rerio (Dr), Homo sapiens (Hs), Mus musculus (Mm), Strongylocentrotus purpuratus (Sp), Branchiostoma floridae (Bf), P. mammillata (Pm), Caenorhabditis elegans (Ce), C. intestinalis (Ci) C. savignyi (Cs) and Drosophila melanogaster (Dm). The numbers indicate bootstrap values. The dashed boxes indicate clustering of ascidian cell adhesion molecules within each subfamily. The red circles indicate high bootstrap values that support our conclusions.

Figure 2

Ciona intestinalis IgCAMs are differentially expressed. A. and B. In situ hybridization of C. intestinalis ACAM at gastrula neurula stages, respectively. Neural plate diagram (B′) based on (Hudson et al., 2007). Yellow arrowheads indicate expression in neural precursor cells; red arrowheads indicate expression in muscle precursor cells. C and D. ACAM expression by ISH at tailbud (C) and tadpole (D) stages, the arrows indicate the strong expression in the anterior brain. E and F. In situ hybridization for NCAM1/2 at gastrula (E) stage and neurula (F) stage. Yellow arrowheads indicate expression in neural precursor cells; orange arrowheads indicate expression in mesenchyme cells. G and H. In situ hybridization for NrCAM at neurula stage (G) and mid tailbud (H) stage. At neurula expression is seen in muscle precursors (red arrowheads), while at mid tailbud expression is seen primarily in the palps (green arrowheads); fainter expression was seen in the sensory vesicle (white arrowhead). I. RT-PCR analysis of ACAM, NCAM1/2 and NrCAM at the indicated stages of C. intestinalis embryonic development (E=early; L=late).

Figure 3

eGFP expression driven by ACAM cis-regulatory domain (expression construct ACAM>H2B∷eGFP). A and B. Brightfield (A) and fluorescent (B) image of a neurula stage embryo expressing ACAM>H2B∷eGFP. Brackets indicate eGFP expression in the neural plate. C and D. Brightfield (C) and fluorescent (D) images of tailbud embryo co-expressing ACAM>H2B∷eGFP and pan neural marker ETR>H2B∷RFP. Arrowhead indicates high eGFP expression in the forebrain. (Sc=spinal cord, M= muscle, SV=sensory vesicle).

Figure 4. ACAM is a target of the homeobox transcription factor OTX

A and B. RT-PCR (30 cycles) analysis of ACAM expression in embryos electroporated with ETR>H2BeGFP (A), or ETR>OTX∷EnR (B) in gastrula (G), neurula (N) and initial tailbud (ITB) stages. C. qRT-PCR analysis of ACAM expression from embryos as in A and B, normalized to actin expression. D and E. Brightfield and fluorescent images of tadpoles co-electroporated with ACAM>H2B∷eGFP and either control plasmid, ETR>H2B∷eGFP (D), or ETR>Otx∷EnR (E). F. Brightfield and fluorescent images of a tailbud embryo co- electroporated with ETR>OTX∷EnR and ACAM>H2B∷GFP. G. Diagram of the OTX binding sites upstream of the ACAM cis-regulatory domain, X marks the mutated sites in the indicated construct. H. Percent of embryos with eGFP in the indicated categories (full expression, ∼10-100 cells, and 0-10 cells) electroporated with the four plasmid constructs indicated in G. For cells in the ∼10-100 cells, and 0-10 cells categories the percentage showing expression only in the muscle cells is indicated (striped bars). The number of embryos scored for each plasmid is indicated in parentheses. I-K. Representative fluorescent images of embryos for the categories quantified in H. Expression from the ACAM expression plasmids is in green, and expression from a control plasmid (Bra>H2B∷RFP) is in red.

Figure 5. ACAM knockdown and steric inhibition results in open anterior neural tube

A. Single embryo RT-PCR for ACAM transcript levels (top panel), and MO-induced mis-splicing at the intron 4-exon 4 junction of the ACAM transcript (bottom panel). Embryos were injected with either control (C-MO), or a mixture of a splice- and translation-blocking ACAM morpholino oligonucleotides (ACAM-MOs). PCR against actin shows RNA present in all samples. B. Percent of embryos with neural tube closure defects in C-MO and ACAM-MO injected cohorts. Results from two separate experiments are shown (◆n= 6 ●n=9 n=20 n=24) C. Brightfield, anti-CRALBP, and anti-Arrestin fluorescent images of tadpoles after C-MO or ACAM-MOs injection. Arrowheads indicate exposed sensory vesicle. D. Brightfield, anti-CRALBP, and anti-HA staining of tadpoles electroporated with either control plasmid ACAM>Venus or ACAM>secACAM∷HA plasmid. E. Percent of embryos with open neural tubes in control (ACAM>Venus) and ACAM>secACAM∷HA electroporated cohorts. Averages of a total of 4 experiments are shown (n for ACAM>Venus=77, 110, 76, 73; n for ACAM>secACAM∷HA=68, 158, 118, 70) *p=0.0283 Student T test. F. Single plane confocal image of late tailbud embryos electroporated with control (ACAM>Venus) or ACAM>secACAM∷Venus. Embryos were immunostained for Venus (green), and counter-stained with rhodamine labeled phalloidin (red). Arrowheads indicate anterior neuropore.