A 23.4-kDa ADP-ribosylation factor-like 6 interacting protein (Arl6ip) was first named in mice (Pettersson et al.,2000). It possesses four transmembrane domains, as predicted by Kyte-Doolittle analysis. Mouse Arl6ip is located in the intracytoplasmic membrane and functions in protein transport, membrane trafficking, and cell signaling. It is expressed in all hematopoietic cell lineages, especially in early myeloid progenitor cells. In addition, arl6ip mRNA is transcribed abundantly in the brain, bone marrow, thymus, and lung by Northern blot analysis (Pettersson et al.,2000). Mouse arl6ip cDNA (GenBank Accession No. AF223953) is identical to mouse TBX2 (Accession No. AF172088) isolated from the thymic stromal cell line (Pettersson et al.,2000) and mouse Aip-1 (Accession No. AF133669) (Ingley et al.,1999). Importantly, it also shares 97% identity with human ARMER/KIAA0069 (Accession No. D31885) isolated from myeloblast cell line KG-1 (Nomura et al.,1994; Pettersson et al.,2000; Lui et al.,2003). EST clones of partial arl6ip cDNA from nematode, zebrafish, rat, and porcine exhibit a high level of homology. Therefore, arl6ip is believed to be a highly evolutionarily conserved gene from invertebrates to vertebrates (Pettersson et al.,2000).
Meanwhile, using yeast two-hybrid screening, mouse Arl6ip was confirmed to physically interact with mouse Arl6 (Ingley et al.,1999), which belongs to the ARF/ARL family in the Ras superfamily (Clark et al.,1993). At the same time, a mutation of Arl6 has been reported to potentially cause Bardet-Biedl syndrome (BBS) in humans (Chiang et al.,2004; Fan et al.,2004) who suffer from obesity, blindness, polydactyly, renal abnormalities, and cognitive impairment (Takada et al.,2005). Under these conditions, a key question arises. What would happen if, during embryogenesis, Arl6ip expression is either absent or present, but becomes dysfunctional? To address this question, we first hypothesized that Arl6ip might have a functional role during development since arl6ip is expressed more broadly than arl6 in developing embryos. Mouse arl6, for instance, is reported to be expressed in node, ectoderm, and neural plate during the embryonic stage (Takada et al.,2005). This gives further support to the importance of Arl6ip expressions and functions during embryogenesis. As a consequence, we further hypothesized that abnormal function of Arl6ip could cause human disease in a manner similar to the association between Arl6 and BBS, as described above. Therefore, we examined the expression pattern and functions of arl6ip transcripts to clarify the role of Arl6ip during embryogenesis. Zebrafish (Danio rerio) is an excellent model vertebrate for the study of developmental biology and genetic analysis (Bilotta and Saszik,2001). It was therefore used as a model system in this study, which involved cloning zebrafish arl6ip cDNA and then characterizing the expression pattern and biological function of Arl6ip during embryonic development.
RESULTS AND DISCUSSION
Dynamic Expression Patterns of arl6ip Transcripts
To study the spatiotemporal expression patterns of arl6ip in zebrafish embryos, we detected arl6ip transcripts by reverse transcription polymerase chain reaction (RT-PCR) and whole-mount in situ hybridization. Results showed that arl6ip was expressed at the 1-cell stage in blastodisc (Fig. 1A,C), indicating that arl6ip is a maternally expressed gene. The arl6ip mRNA was also expressed at the 8-cell stage, 24, 48, 72, and 96 hr post-fertilization (hpf), and even in adulthood (Fig. 1A).
In order to detect endogenous Arl6ip in the early embryonic stages, we performed Western blot analysis to detect the endogenous Arl6ip protein at the 1-cell stage, 2 1/4 hpf (128 cell stage), 3 1/3 hpf, 5 1/4 hpf and 10 hpf of zebrafish embryos. Results showed that Arl6ip was present at the 2 1/4, 3 1/3, 5 1/4, and 10 hpf (Fig. 1B), but that Arl6ip could not be detected at the one-cell stage (data not shown). Although the level of Arl6ip protein expression may not be detectable at the one-cell stage, its presence at the 128-cell stage is sufficient evidence that Arl6ip may be a maternally translated protein.
The distribution of arl6ip mRNA was ubiquitous in whole embryos at the 1-, 4-, and 8-cell stages and 13 hpf (Fig. 1C–F). Furthermore, arl6ip also appeared in the brain, optic primordia, trunk, hypochord, spinal cord, myotome, fin-bud, heart, kidney, and retina when embryos developed after 24 hpf (Fig. 1G–L). In addition, the expression level of arl6ip mRNA varied among different tissues at 24 hpf, with weak expression in the trunk, but high expression in the head. At 72 hpf, arl6ip is weakly expressed in the trunk, but it is strongly expressed in the brain, eyes, and kidney. While arl6ip was dominantly expressed in the amacrine cells of the inner nuclear layer (INL), it was weakly expressed in the ganglion cell layer (GCL), or ciliary marginal zone (cmz), at 96 hpf by cross-sectioning observation (Fig. 1L). However, arl6ip was hardly seen in either the outer nuclear layer or photoreceptors at 96 hpf. Based on these lines of evidence, this study demonstrates that arl6ip expression during the early stages of zebrafish embryogenesis is dynamic, suggesting, in turn, that the expressions of arl6ip transcripts is stage-dependent.
Translation Inhibitor Specifically Blocks the Synthesis of Arl6ip
To understand Arl6ip function during embryogenesis, we designed two types of arl6ip-morpholino (MO), arl6ip-MO1 and arl6ip-MO2, to specifically knock down the translation of endogenous arl6ip mRNA. The arl6ip-MO1 located at the coding region of arl6ip mRNA, whereas the arl6ip-MO2 targeted the 5′ untranslated region (5′ UTR) of arl6ip mRNA (Fig. 2A). In order to demonstrate that translation of endogenous arl6ip mRNA was effectively inhibited by the injected arl6ip-MO1, we extracted and analyzed the total proteins from 18-hpf embryos after injection of arl6ip-MO1. Western blot analysis was used to clearly demonstrate that the endogenous Arl6ip protein was totally absent in the arl6ip-morphants (Fig. 2B). This result was consistent with embryos injected with arl6ip-MO2 (data not shown). We also designed a control MO, arl6ip-control-MO, which was a complementary strand of arl6ip-MO1, but in the same direction (Fig. 2A). Results showed that the arl6ip-control-MO did not affect arl6ip expression in developing zebrafish embryos, even after injection as high as 8 ng/embryo. This indicates that the phenotypes induced by arl6ip-MO are specific and that this result was caused by the effective inhibition of the translation of endogenous arl6ip mRNA.
Defective Phenotypes of arl6ip-Morphants
After injecting 4 ng/embryo of arl6ip-MO1, we observed defective phenotypes of embryos from 18 to 96 hpf, including those displaying microphthalmia (small eyes), flat head, pericardial edema, deficient pectoral fins, loss of organized pigment pattern, and abnormal trunk (Fig. 2H′–J′). Moreover, when we performed a histological microsection and stained with hematoxylin and eosin to observe the defective internal organs or tissues, we found that the pneumatic duct was abnormal and the tectum tissue in head was defective in the arl6ip-morphants at 96 hpf (Fig. 2R,S). Identical deformities occurred in embryos when the 8 ng/embryo of arl6ip-MO2 was injected. In contrast, when compared to wild-type, the development of embryos injected with 8 ng/embryo of arl6ip-control-MO was correspondingly normal (Fig. 2G). However, when embryos injected with arl6ip-MO1 (Fig. 2H–L) were compared to the wild-type (Fig. 2C–G), symptomatic microphthalmia was a noticeable defect, indicating that ocular development is arrested.
Serial concentrations of MO were next performed to assess the relationship between the MO concentration and degree of defect. Defects were categorized as mild and severe phenotypes at 24 hpf. Mild defects were tail in up position, flat head, irregular boundary between somites, and blurred eyes outline (Fig. 2N). Severe defects also consisted of flat head and irregular boundary between somites, but also shorter tail and absence of eyes (Fig. 2M). The percentages of mild and severe defects occurring after injection of arl6ip-MO1 and arl6ip-MO2 are shown in Table 1. Because we found that the degree of severity observed in defective phenotypes increased in proportion to the increase of MO concentration, it was concluded that the phenotypes induced by both arl6ip-MO1 and arl6ip-MO2 were dosage-dependent.
Table 1. The Occurrence Rates of Defective Zebrafish Embryos Derived From Fertilized Eggs Injected With Different Materialsa
Concentration (ng) (per embryo)
Survival rate (%)
Wild type-like rate (%)
Mild defect rate (%)
Severe defect rate (%)
The phenotypic defects of arl6ip-morphants were rescued by co-injection of arl6ip mRNA for arl6ip-MO1 or arl6ip-MO2. The definitions of arl6ip-MO1, arl6ip-MO2, and arl6ip-control-MO are provided in the Experimental Procedures section. The definitions of mild defect or severe defect are provided in Figure 2.
arl6ip mRNA + arl6ip-MO1
200 pg + 4 ng
arl6ip mRNA + arl6ip-MO2
200 pg + 4 ng
Arl6ip Is Involved in Heart Development
In the wild-type, the arl6ip transcripts were detected in the heart-tube of embryos at 48 hpf (Fig. 1I), but they were absent at 72 hpf (Fig. 1J). However, the arl6ip-morphants displayed the defective phenotypes such as pericardial edema (Fig. 2H–L) and looping failure when embryos were observed at 48 hpf (see Supp. Fig. S1, which is available online). Furthermore, we used some pre-migratory cardiac neural crest cell markers, such as wnt3a (Dorsky et al.,1998; Sato and Yost,2003), to examine the cardiac neural crest cells by whole-mount in situ hybridization. Results showed that wnt3a was not normally expressed in arl6ip-morphants, indicating that the pre-migratory cardiac neural crest cells were defective (data not shown). When the cardiac neural crest cells were ablated by laser, it was shown that the migration of neural crest cells to the heart was lost, resulting in the failure of the heart tube to undergo looping (Li et al.,2003). Therefore, we speculate that the loss of Arl6ip function in embryos may lead to defective cardiac neural crest cells, which, in turn, may lead to heart defects. However, some other possibilities should be considered before making a firm conclusion. For example, a loss of Wnt signaling may prevent cardiac mesoderm cells from proper migration to the heart field prior to neural crest migration (Marvin et al.,2001). Secondly, a number of other Wnt ligands has been shown to regulate neural crest migration (Yanfeng et al.,2003); therefore, the loss of Wnt3a expression itself may not necessarily equate to a loss of cardiac cell migration. Thirdly, the overall heart defect we observed in the arl6ip-morphants might be the result of specific effects in cardiac mesoderm cells in the early stages of heart morphogenesis as well as secondary effects in cardiac neural crest cell migration.
Fin-Bud Development Is Defective in arl6ip-Morphants
The arl6ip gene was expressed in zebrafish fin-bud at 48 hpf by whole-mount in situ hybridization (Fig. 1H). When we injected the arl6ip-MO1 into embryo to knock down Arl6ip, the pectoral fin did not develop normally (Fig. 2K), suggesting that arl6ip must play an important role in the development of zebrafish fin buds. Furthermore, we injected arl6ip-MO1 into the embryos derived from transgenic line Tg(shh::GFP), in which GFP is driven by the sonic hedgehog (shh) gene promoter and GFP should be expressed in fin-bud due to the fact that the Shh protein is involved in fin-bud development (Neumann et al.,1999). Results showed that the GFP was not displayed in the fin-bud of arl6ip-MO1-injected embryos (n=10) when we observed the embryos at 48 hpf by confocal fluorescence microscope (Supp. Fig. S2). We speculate that shh expression might be affected in fin-bud development when Arl6ip is absent. Therefore, our hypothesis that expression of Arl6ip is required for fin-bud development during zebrafish embryogenesis gains further support.
All injected embryos, whether by 2, 4, or 8 ng/embryo of arl6ip-MO1, had expired by the time development had reached 120 hpf. However, when a combination of 200 pg/embryo of synthetic arl6ip mRNA and 4 ng of arl6ip-MO1 was injected, about 44% of embryos were rescued. Meanwhile, about 54% of embryos (Fig. 2P) became wild-type (Fig. 2O) by injecting 200 pg of arl6ip mRNA combined with 4 ng/embryo arl6ip-MO2 (Table 1), indicating that the defects (Fig. 2Q) caused by arl6ip-MO were specific. The reason why arl6ip-MO2, as opposed to arl6ip-MO1, was more effective in rescuing the defects may have been the result of the structural interaction between MO and the synthetic mRNA we injected. The arl6ip-MO1 was designed to hybridize a 25-nucleotide sequence (a one-nucleotide before AUG and a 21-nucleotide after AUG) located at the coding region of arl6ip mRNA. However, since the synthetic arl6ip mRNA was designed from AUG to UAA, the synthetic arl6ip may have been blocked by arl6ip-MO1 during rescue.
Arl6ip Is Required for Retina Development
We had originally hypothesized that Arl6ip might play important roles during development, and we further hypothesized that abnormal function of Arl6ip could cause human disease. Significantly, among these malformed arl6ip-morphants, microphthalmia was the predominant phenotypic defect induced by arl6ip-MO1. Therefore, it was important to gain a better understanding of the role that Arl6ip plays in ocular development.
The molecular markers of eye specification start to be expressed in the anterior neural pate during gastrulation (60–80% epiboly stages) (Chow and Lang,2001). Moreover, the optic primordium starts to evaginate from the lateral wall of the diencephalon (Sidman and Rakic,1982). In the retina, the composition of cell types is different. To compare the embryos between 24 and 36 hpf, almost all retinal cells are mitotic progenitor cells at 24 hpf, but many postmitotic neurons are generated at 36 hpf (Schmitt and Dowling,1994; Hu and Easter,1999). When we performed histochemical examination of 72- and 96-hpf wild-type embryos, we found that the optic nerve, outer plexiform layer, inner plexiform layer, ganglion cell layer, inner nuclear layer, and outer nuclear layer of the retina all differentiated normally (Fig. 3A,C). However, in the arl6ip-morphants, these tissue layers did not differentiate normally; rather, they entirely lost their organization and decreased in size (Fig. 3B,D). This ocular defect did not recover, even after further development, when we observed it again at 96 hpf (Fig. 3D). Furthermore, we used a retinal axon marker, Zn8 antibody, to carry out immunological examination. The Zn8 antibody recognizes DM-GRASP and labels retinal axons (Fashena and Westerfield,1999). The level of Zn8 was markedly reduced in the retina of morphants, and only a small number of cells at the ventronasal retina expressed Zn8. This evidence suggests that the RGC differentiation is severely compromised in the arl6ip-morphants, resulting in defective retinal axons (Fig. 3E,F). These data should strengthen the conclusion obtained from laser microscope observation that the retinal axon is defective in the arl6ip-morphants.
When the embryos lost the Arl6ip function, we found that the retinal size was smaller than that of wild-type at 72 hpf, even though they developed until 96 hpf (Fig. 3B,D). The retinal morphology of the arl6ip-morphants at 96 hpf appeared to be similar to that of the wild-type at 36 hpf reported by Li et al. (2000). The normal differentiation of retinal cells did not occur in the arl6ip-morphants. We suggested that the differentiation of retinal cells might delay severely in arl6ip-morphants. Therefore, we used bromodeoxyuridine (BrdU) labeling to detect cell-cycle progression and found that the BrdU-positive signals were highly expressed in the arl6ip-morphants at 72 hpf (Fig. 3G–J). It is highly possible that the absence of Arl6ip might influence cell-cycle progression, resulting in the inability of retinal mitotic cells to enter the differentiation process.
Moreover, the photoreceptor cell layer was not formed in the arl6ip-morphants. We continued to examine whether rhodopsin mRNA and red opsin mRNA were transcribed normally in the arl6ip-morphants. After the total RNA was extracted from wild-type embryos and arl6ip-morphants at 48, 72, and 96 hpf, RT-PCR was used to detect the presence of the transcripts of rhodopsin and red opsin genes, which are expressed exclusively in rods and red cones, respectively. Zebrafish arl6ip begins to express at about 50 hpf (Raymond et al.,1995), which is specific for photoreceptor cells and agrees with our inability to detect the expression of either rhodopsin or red opsin in wild-type or arl6ip-morphants at 48 hpf (Fig. 3K). However, after starting transcription, rhodopsin mRNA and red opsin mRNA were detected later in the wild-type embryos at 72 and 96 hpf. While the rhodopsin and red opsin genes were transcribed normally in the wild-type embryos, we found that neither rhodopsin mRNA nor red opsin mRNA was detected in the arl6ip-morphants at 72 and 96 hpf (Fig. 3K), indicating that loss-of-function of Arl6ip causes failure of retinal ganglion cells and photoreceptors to differentiate in the arl6ip-morphants.
In summary, we found that the arl6ip gene is expressed in the brain, optic primordia, hypochord, spinal cord, myotome, heart, fin-bud, kidney, trunk, and retina of zebrafish embryos. Loss of Arl6ip function induces some obvious morphological defects, such as microphthalmia, heart defect, and fin-bud deficiency, and some internal organ defects, such as malformed kidney and tectum, strongly suggesting that Arl6ip is required for vertebrate development during zebrafish embryogenesis. It is, therefore, worthwhile to further determine whether Arl6ip is involved in early embryonic development, which will lead us closer to characterizing the underlying mechanisms and links between Arl6 mutation and Arl6/Arl6ip interaction in human disease.
Culturing of Zebrafish
Zebrafish wild-type AB strain and transgenic lines of Tg(cmlc2::EGFP) (Huang et al.,2003) and Tg(shh:EGFP) (2.2shh::gfp:ABC#15) (Shkumatava et al.,2004) were cultured at approximately 28°C, fed two times daily, and maintained under a 14-hr day and 10-hr night photoperiod. Embryonic development was observed with a stereo dissecting microscope, MZFL III (Leica Microscope System Ltd., Germany), to determine the stage based on Kimmel et al. (1995) or Schmitt and Dowling (1994,1999). Fertilized eggs were collected and cultured in a 90×15-mm petri dish. The developmental stages were distinguished by hpf as in Kimmel et al. (1995).
RT-PCR was performed according to the procedures described by Chen et al. (2001), except that the sequences of PCR primers were as follows: arl6ip forward (5′-ATGGCTGAGGGCGACAATAAAAGTGCAAAT-3′) and reverse (5′-TTACTCGTTCTTCTTCTCCTTCTGTTTAAGAA-3′) primers; rhodopsin forward (5′-GAAGCTGCGCACACCCCTCAACTACATCCT-3′) and reverse (5′-GACCATGATGATGACCA- TGCGGGTGAC-3′) primers; red opsin forward (5′-TCAAGAAGCTCCGTCACCCTCTC-3′) and reverse (5′-AACTGTCGGTTCATGAAGACATA-3′) primers; and tubulin forward (5′-CCCTTCCCTCGTCTCCAC-3′) and reverse (5′-GCCAGTGTACCAGTGAAGGGA-3′) primers. PCR amplification was performed by Taq DNA polymerase (Viogene) for 25 cycles. Each cycle consisted of denaturation for 30 sec at 94°C, annealing for 30 sec at 52°C, and extension for 30 sec at 72°C. The last extension step was extended for 5 min at 72°C.
Whole-Mount In Situ Hybridization
cDNAs were subcloned into the pGEM-T easy vector (Promega). All riboprobes were labeled with digoxigenin (DIG) by using the DIG RNA labeling kit (Roche). The procedures were described by Lee et al. (2006) and Chen et al. (2001), except for embryos no later than 24 hpf whose chorion was removed by using pronase (10 μg/ml) (Sigma). Cryostat cross-sectioning was performed according to standard protocols (Westerfield,1995).
The antisense polynucleotide MO that specifically inhibits the translation of arl6ip mRNA was designed and synthesized by Gene Tools. The sequences of arl6ip-MOs were designed as follows: arl6ip-MO1, 5′-ACTTTTATTGTCGCCCTCAGCCATG-3′, antisense nucleotides 122 to 146 of zebrafish arl6ip cDNA (Accession No. AY398312); arl6ip-MO2, 5′-GATGTTACTTGAGAGTTTAGGTTCC-3′, antisense nucleotides 89 to 113 of zebrafish arl6ip cDNA; and arl6ip-control-MO, 5′-GTACCGACTCCCGCTGTTATTTTCA-3′, an inverted sequence of arl6ip-MO1. All arl6ip-MOs were prepared at a stock concentration of 8.361 μg/μl, stored at −20°C, and diluted to the desired concentration for microinjection into zebrafish embryos at the 1-cell stage.
The coding region of arl6ip was subcloned into the pCS2+ vector. The mRNA of arl6ip was capped according to the protocol of the manufacturer (Epicentre). The synthesized mRNA for arl6ip was diluted to 50–200 ng/μl.
Western Blot Analysis
About 500 wild-type embryos at the 1-cell stage, 2 1/4, 3 1/3, 5 1/4, 10 and 18 hpf, and 500 arl6ip-MO1-injected-embryos at 18 hpf were collected and homogenized with 250 μl lysis buffer (Lee et al.,2007; Lui et al.,2003). The rabbit polyclonal antibody against ARMER (Lui et al.,2003; Abcam, Cat#ab24228) was used at a dilution of 1:500. The secondary antibody, goat polyclonal antibody against rabbit IgG (HRP) (Abcam, Cat#ab6721), or goat monoclonal antibody against rabbit IgG-AP (Santa Cruz Biotechnology, Cat#SC-2034), was used at a dilution of 1:5,000. The loading control, mouse-anti-β-tubulin (Abcam, Cat#ab7287) was used at a dilution of 1:500. The secondary antibody, goat antibody against mouse IgG-HRP (Santa Cruz Biotechnology, Cat#sc-2055), or goat antibody against mouse-AP (Abcam, Cat#ab6790), was used at a dilution of 1:5,000.
To identify the differentiation of retina cells, we immersed the 72- and 96-hpf larvae of wild-type and arl6ip-morphants in 4% paraformaldehyde overnight at 4°C and embedded them in paraffin. The embryos were sectioned at 10-μm intervals.
The protocol of immunofluorescent staining was performed according to standard protocols (Shen and Raymond,2004). The Zn8 antibody (Development Studies Hybridoma Bank) was used at a dilution of 1:200. The secondary antibody, Cy3 goat anti-mouse immunoglobulin G (H+L) (Chemicon International, Cat#AP181C), was used at a dilution of 1:200. The embryos were observed by confocal fluorescence microscope (TCS SP5, Leica).
The experimental procedure of BrdU labeling followed the procedures described by Hu and Easter (1999) with some modifications. Dechorionated embryos were kept on ice for 15 min and incubated in a cold 15-mM BrdU/15%DMSO medium for 30 min. Then, embryos were transferred to an embryonic medium at 28.5°C for 40 min, fixed in 4% PFA at room temperature for 2 hr, and transferred to methanol at −20°C overnight. The BrdU antibody (Abcam, Cat#Ab7796) was used at a dilution of 1:100. The nuclei were labeled with 45 ng/ml DAPI (Sigma) at room temperature for 10 to 30 min.
We are grateful to Miss Shuan Tseng (Veterinary Hospital, NTU) for helping with the paraffin sectioning, to Dr. Carl J. Neumann (European Molecular Biology Laboratory, Heidelberg, Germany) for providing transgenic zebrafish Tg(shh::EGFP), to Miss Yi-Hua Chiu (Instrument Center, TMO, Taipei Medical University) for assisting with the fluorescence confocal microscopy, and to Dr. Richard I. Dorsky for providing Wnt3a cDNA. This project was partially supported by a grant from the National Science Council, ROC.