|Home | About | Journals | Submit | Contact Us | Français|
Primary open-angle glaucoma (POAG) is a genetically complex neuropathy that affects retinal ganglion cells and is a leading cause of blindness worldwide. WDR36, a gene of unknown function, was recently identified as causative for POAG at locus GLC1G. Subsequent studies found disease-associated variants in control populations, leaving the role of WDR36 in this disease unclear. To address this issue, we determined the function of WDR36. We studied Wdr36 in zebrafish and found it is the functional homolog of yeast Utp21. Utp21 is cell essential and functions in the nucleolar processing of 18S rRNA, which is required for ribosome biogenesis. Evidence for functional homology comes from sequence alignment, ubiquitous expression, sub-cellular localization to the nucleolus and loss-of-function phenotypes that include defects in 18S rRNA processing and abnormal nucleolar morphology. Additionally, we show that loss of Wdr36 function leads to an activation of the p53 stress–response pathway, suggesting that co-inheritance of defects in p53 pathway genes may influence the impact of WDR36 variants on POAG. Although these results overall do not provide evidence for or against a role of WDR36 in POAG, they do provide important baseline information for future studies.
Glaucomas are a heterogeneous group of blinding neuropathies that result in retinal ganglion cell (RGC) death, visual field deficits and optic nerve atrophy (1). They are currently a leading cause of blindness worldwide, and estimates predict 60 million people will be diagnosed with the disease by 2010 (2). The most prevalent form of glaucoma is primary open-angle glaucoma (POAG), the incidence of which increases with advancing age (3). POAG is a complex disease, with multiple genetic and environmental factors interacting to cause disease (1,4,5). High intraocular pressure (IOP) is the greatest risk factor for developing glaucoma, and is thought to result from defects in aqueous humor drainage within the anterior segment of the eye. It is unknown exactly how elevated IOP impacts POAG. Currently, the most accepted hypothesis is that elevated pressure in the anterior chamber of the eye is translated to the posterior chamber, creating stress at the optic nerve head and, ultimately, resulting in death of RGCs (6–11). Though high IOP is the most common risk factor, it is not required for the development of POAG. About one-third of cases occur in normal tension patients (IOP < 22 mmHg) (12). Together, this suggests that glaucoma can result from insults to the drainage structures in the anterior chamber and/or from intrinsic defects in RGCs (1).
Determining the genetic risk factors of POAG has been an area of intense study, but the complex nature of the disease has made the search very difficult (13). Through human genetic screens, at least 20 loci have been found that link to POAG, but causative genes have only been reported for three (13). These genes are MYOCILIN, OPTINEURIN and WDR36. Genetically, both MYOCILIN and OPTINEURIN have been well established in influencing POAG, but the endogenous functions of these proteins as well as the exact mechanism by which they cause disease are unknown (1). MYOCILIN is the causative gene at locus GLC1A, and is associated with high IOP cases of POAG (13,14). MYOCILIN is expressed at high levels in trabecular meshwork cells, and it is thought that disease causing mutations result in cellular dysfunction as a result of blocked secretion and ectopic peroxisome targeting (15–20). Compromised trabecular meshwork cells result in raised IOP, which then contributes to glaucoma. OPTINEURIN is the causative gene at locus GLC1E and was found to cause normal-tension POAG (21). Genetic studies in a variety of populations revealed that some initially described disease-causing mutations are also found in normal individuals (22). This indicates that OPTINEURIN causes POAG when in permissive backgrounds, and acts as a key modifier in others (1). In cultured cells, OPTINEURIN can act to regulate apoptosis through interactions with tumor necrosis factor-alpha, but it is unknown how this occurs and whether this contributes to the apoptotic RGC cell loss found in POAG (23–26).
WDR36 encodes a protein of unknown function that was recently described as a causative gene for adult-onset POAG (27). Studies mapping POAG loci in affected families identified and narrowed a genetic linkage to 5q22.1 (GLC1G) (27–29). In one of these families, a coding variation in WDR36 was found to segregate with all affected family members and not with any unaffected individuals (27). Further analysis of 17 unrelated patients with either high- or low-tension POAG revealed a total of four amino acid changes that were not found in 200 control individuals (27). When other groups extended the study of WDR36 to different populations, the relationship between WDR36 and glaucoma appeared more complex. Originally described disease-causing variants have been found in control individuals with an equal frequency as patients with POAG (30–33). Although these data indicate that WDR36 is not causative for POAG in all populations, recent studies suggest that WDR36 may act as a modifier of the disease. Hauser et al. (30) found correlations between WDR36 variants and POAG severity in POAG pedigrees. Additional genetic association studies have also found a higher frequency of rare WDR36 variants in glaucoma patients compared with controls, and have suggested that WDR36 variants may contribute to a small number of unrelated glaucoma cases (32,34,35). Taken together, the current genetic data suggest that WDR36 may act as a modifier of POAG, and/or as a causative gene for POAG in certain populations. Alternatively, WDR36 variants may simply mark certain disease haplotypes and not be directly involved in POAG pathology. To better understand the role of WDR36 in POAG, determining its function and manipulating the gene in animal models are crucial. Although the function of WDR36 is unclear, it is known that it is a 100 kDa protein containing four conserved protein domains: a guanine nucleotide-binding protein (G-beta) WD40 repeat, an AMP-dependant synthetase and ligase, a cytochrome cd1—nitrite reductase-like domain and a putative Utp21-specific WD40-associated domain (27). In human and mouse tissues, WDR36 is widely expressed, and it was also found to be upregulated during human T-cell proliferation (27,36).
In this study, we use zebrafish to determine the function of Wdr36, the homolog of human WDR36, and investigate the affected cellular signaling pathways for its involvement in POAG. We show that Wdr36 is functionally homologous to the yeast Utp21, a component of the rRNA processome, and is involved in 18S rRNA processing and nucleolar homeostasis. We further show that wdr36 loss of function results in activation of the p53 stress–response pathway, consistent with disrupted nucleolar function.
To gain insight into the cellular function of vertebrate WDR36, BLASTp analysis using human WDR36 was carried out against Drosophila melanogaster (Dm), Caenorhabditis elegans (Ce) and Sacchaormyces cerevisiae (Sc)—organisms with extensive protein function data. The closest related proteins in each species were: Dm, AAK93538 (NP_650284); Ce, Y45F10D.7 (NP_502661); Sc, Utp21 (NP_013513) (37,38). Of the homologous proteins identified, only the putative yeast ortholog, Utp21, was studied. It was found that this protein is part of the yeast rRNA small-subunit (SSU) processome and is essential for 18S rRNA maturation (38,39). Human WDR36 shares 24% identity with yeast Utp21. Subsequent protein alignments revealed that Utp21 and zebrafish Wdr36 share 25% identity, whereas human WDR36 and zebrafish Wdr36 share 64% identity. To test the hypothesis that vertebrate Wdr36 is the functional homolog of yeast Utp21, we explored the expression and cellular consequences of loss of Wdr36 function in zebrafish.
The expression of wdr36 was examined by RT–PCR and in situ hybridization. Similar to that reported for human and mouse, RT–PCR analysis suggested that zebrafish wdr36 is ubiquitous as transcripts were detected in all embryonic and adult tissues assayed (Fig. 1) (27). It also indicated that wdr36 is maternally contributed (Fig. 1). Zebrafish embryos do not begin zygotic transcription until the 512-cell stage, and instead rely on a large supply of maternal protein and transcripts for early development. Ubiquitous wdr36 expression was also shown by in situ hybridization when the alkaline phosphatase reporter reaction was allowed to develop for longer periods of time (data not shown). Interestingly, shorter reporter reaction times showed enrichment of wdr36 transcripts within tissues undergoing high levels of proliferation (Fig. 1). Tissues with the greatest levels included the developing eye cup, proliferative zones within the CNS and cells throughout the gut (Fig. 1). Within the eye, as differentiation progressed, high levels of gene expression were maintained in the lens and retinal periphery where growth is ongoing. These results indicate that wdr36 is ubiquitously expressed, but is enriched in proliferative cells during development.
To investigate the sub-cellular localization of Wdr36, we expressed various epitope-tagged versions of Wdr36 within multiple cell types in zebrafish embryos. Fusion protein expression was driven by an inducible heat shock promoter (hsp70) in an attempt to control the protein levels. Following heat shock induction, epitope-tagged Wdr36 localized to both the cytoplasm and nucleoli (Fig. 2). Multiple epitope tags (N- and C-terminus 6X-myc or YFP) all showed equivalent localizations. Nucleolar localization was confirmed by co-expressing with either Histone2A fusions, which localize to chromatin but not the nucleolus, or with B23, which specifically labels the nucleolus (Fig. 2). In general, Wrd36 fusions first localized to nucleoli and then, as protein levels rose, became more enriched in the cytoplasm.
To study the function of Wdr36 in vivo, a zebrafish line with a viral insert in the wdr36 gene was utilized (wdr36hi3630aTg). This recessive mutant was identified in a forward genetic screen for essential genes (40). The viral insert in the wdr36 mutant was found within the first exon, about 40 nucleotide basepairs upstream of the start codon (Fig. 3A). Embryos heterozygous for the insert showed no phenotypic differences from wild-type (WT) siblings during development in any assays performed. Homozygous viral insert mutants (wdr36 mutants) showed a reduction in wdr36 transcript, but not a total loss, as assessed by semi-quantitative RT–PCR (Fig. 3B). We were not able to determine protein levels as our attempts to generate specific antibodies failed. wdr36 mutants do not show visible phenotypes until ~3.5 days post-fertilization (dpf) at which point affected embryos are characterized by smaller heads and eyes. These phenotypes progress during development and by 6 dpf the small head and eye phenotypes are obvious compared with WT siblings (Fig. 3C). The other profound ocular phenotype is a lens opacity that presents with variable penetrance (Fig. 3C). Additionally, an obvious gut phenotype, including liver necrosis and absence of swim bladder inflation, develops by 6 dpf (Fig. 3C). Histological analysis of the eye at 5 dpf showed only subtle phenotypes. Although cellular differentiation and lamination of the retina appeared normal, a thickening of the lens epithelium was found and presumably accounts for the lens opacity. The proliferative margin of the retina also appeared to be reduced in the wdr36 mutant embryos. Other than these subtle defects and smaller size, no other ocular abnormalities were found. Consistent with gross inspection of the embryo, histology of gut structures showed severe differentiation defects in the liver, pancreas and intestine (Fig. 6A, data not shown). All phenotypes progressively worsen and the embryos die by 11 dpf. Lethality is most likely due to defects in the gut organs. At 10 dpf, the eyes of mutants are very small compared with WTs, and a pin-hole pupil is present. Histology revealed progressive and severe lens degeneration, accounting for the dysmorphic pupil (Fig. 3D). Additionally, at this late stage of pathogenesis, the retina showed signs of degeneration, with thinning of all retinal layers and presence of dying cells with pyknotic nuclei (Fig. 3D).
To confirm the specificity of the phenotypes observed in the wdr36 mutant embryos and to provide additional tools for analysis, antisense morpholinos (MOs) were designed against wdr36. Two non-overlapping MOs were designed; one targeting the start ATG codon to disrupt translation and one targeting the junction of exon 2:intron 2 to disrupt splicing. Both MOs gave identical phenotypes, and RT–PCR indicated that splicing of wdr36 was disrupted by the second MO (Fig. 4A). Morphant embryos were also very similar to wdr36 genetic mutants, confirming that the collective phenotypes result from loss of wdr36 (Fig. 4B). The only clear difference observed between wrd36 morphants and wdr36 mutants was that the phenotypes showed slightly earlier in development when using MOs. With ATG MO injection, the small eye and head phenotypes were detected by 36 hpf. At 48 hpf, a delay in retinal lamination was observed by histology, along with phenotypes similar to wdr36 mutant embryos (Fig. 4C). For example, the later onset lens degeneration was also observed at 8 dpf in wdr36 morphants (Fig. 4C).
The WDR36 variants associated with disease are inherited in a heterozygous manner, opening the possibility that disease may be caused by haploinsufficiency (27,30). To begin to address this possibility, we characterized eyes from adult heterozygous wdr36–insert fish. Overall appearance of heterozygous eyes, or the fish in general, did not differ from WT. Similarly, there were no changes in retinal histology between adult heterozygotes versus WT eyes (Fig. 3E). To explore whether more subtle defects occurred in the eyes of wdr36–insert heterozygotes, we measured gap43 and c1q levels by real-time RT–PCR. Transcription of gap43 is upregulated in zebrafish RGCs following axotomy or optic nerve crush, and c1qc is upregulated in optic neuropathies including glaucoma (41–44). Both genes were found to be expressed at low levels and no change in expression was found between wdr36–insert heterozygote and WT eyes for either gene (data not shown). Finally, no indications of increased IOP were noted in wdr36–insert heterozygotes, as judged by measuring corneal thickness and eye size to body length ratio, which can be affected with changes in IOP.
To test whether wdr36 is functionally homologous to yeast utp21, we investigated the processing of 18S rRNA following loss of Wdr36. rRNA is transcribed as a large 35S pre-rRNA transcript. The transcript undergoes a series of enzymatic cleavages within the nucleolus by large nucleo-protein complexes resulting in mature 18S, 5.8S and 25S rRNAs (45). 18S rRNA is utilized for the production of the small ribosomal subunit, and 5.8S and 25S rRNAs are used for the large subunit. The first cleavage steps in 35S rRNA processing separate the immature 18S rRNA from the 5.8 and 25S rRNAs. Distinct processomes for large and small subunit rRNAs then complete the maturation steps. The SSU processome, of which Utp21 is a part, is required specifically for small subunit (18S rRNA) maturation. After production in the nucleolus, mature subunits are transported to the cytoplasm where they are utilized for ribosome assembly. This process has been well studied in yeast, but in vertebrates the exact order of cleavage events has been found to vary across species. An outline of 18S maturation in the zebrafish is shown in Figure 5A (46). Northern blot analysis, using a probe against 18S rRNA, was carried out to investigate rRNA processing in wdr36 mutants and morphants. This probe recognizes both the mature and immature transcripts, as well as the intermediate forms of 18S rRNA. The presence of the precursor, intermediate and mature forms was observed in WT controls (Fig. 5B). In contrast, high levels of the immature forms of 18S rRNA was observed in wdr36 mutant and morphant embryos, but products from later maturation steps and the mature 18S rRNA were greatly reduced (Fig. 5B). This result shows a defect at one or more steps in the maturation of 18S rRNA following loss of wdr36, consistent with Wdr36 being the functional homolog of Utp21.
To address whether the observed defects in 18S rRNA processing had an effect on the nucleolus, histological sections of retinal neurons, intestinal epithelia and liver cells from wdr36 mutants and morphants were analyzed at high magnifications. In these tissues, the nucleolus appeared large and dysmorphic with loss of Wdr36 compared with controls (Fig. 6A). To confirm these findings, nucleolar morphology was analyzed by transmission electron microscopy. Again, Wdr36-deficient nucleoli were enlarged, dysmorphic and often less electron dense compared with WT control cells (Fig. 6B).
It is well documented that disruption of nucleolar morphology and ribosome biogenesis results in activation of the p53 stress–response pathway (47). When the nucleolus is disrupted, activated ADP-ribosylating factors (Arfs) and free ribosomal proteins are released from the nucleus into the cytoplasm. Once in the cytoplasm, these proteins interact with Mdm2. The normal function of Mdm2 is to bind, ubiquinate and maintain p53 in an inactive state. When Mdm2 interacts with free nucleolar proteins released following disruption of the nucleolus, Mdm2 no longer inhibits p53. The resultant activation of p53 can then promote transcription of itself along with pro-apoptotic and/or cell cycle arrest genes. To address whether a p53 stress–response was activated with loss of wdr36, we first investigated whether expression of p53 and its target genes were altered following loss of Wdr36. Semi-quantitative and real-time quantitative RT–PCR showed that while p53 transcripts were present, the levels were not different between control and Wdr36-deficient embryos (Fig. 7A and B). However, a truncated isoform of p53, Δ113p53, was significantly upregulated in wdr36 mutants and morphants as shown by semi-quantitative and quantitative RT–PCR (Fig. 7A and B). The Δ113p53 isoform results from an alternate start site in the fourth intron and lacks part of the DNA-binding domain (48,49). The corresponding human isoform, Δ133p53, causes upregulation of cell cycle arrest genes but acts in a dominant-negative fashion to block induction of pro-apoptotic transcripts (49). This was previously confirmed in the zebrafish def mutant, which displays hypoplastic digestive organs (48). In def mutants, markers of cell cycle arrest were transcriptionally upregulated, but markers of apoptosis were not. We found similar results after loss of Wdr36. The cyclin-dependent kinase inhibitor p21 was significantly upregulated, whereas the pro-apoptotic gene bax was not altered with the onset of Wdr36-deficient phenotypes (Fig. 7A). Overall, these experiments suggest that the p53 stress–response pathway is activated with loss of wdr36.
To probe these effects deeper, we analyzed apoptosis directly by scoring pyknotic nuclei and the expression of activated caspase-3 immunoreactivity. No change in either indicator of apoptosis was found at the onset of overt phenotypes (data not shown). At 10 dpf, in concordance with the observed trend toward an increase in p53, evidence of apoptosis was found by both measures (data not shown). In contrast, defects in cell cycle regulation were found in wdr36 MOs at the onset of overt phenotypes. Propidium iodide fluorescent-activated cell sorting (FACS) (DNA content), BrdU labeling (S-phase) and phospho-Histone H3 [late G2/M-phase] analyses were used to probe changes in the cell cycle. At 36 hpf, FACS analysis showed that Wdr36-deficient embryos had disproportionately greater percentage of cells in S-phase (Fig. 7C). Furthermore, long-pulse BrdU-labeling confirmed the FACS results and suggested that in addition to changes in cell cycle kinetics, cell cycle exit was also delayed within the retina. Altered cell cycle exit was indicated by the presence of proliferating cells in the wdr36 morphant central retina at 58 hpf (Fig. 7D). At this time in WT retinas, proliferation was restricted to the ciliary marginal zone (CMZ; Fig. 7D). A delay in cell cycle withdrawal was confirmed by observing increased M-phase cells by phospho-Histone H3 analysis (Fig. 7E).
Finally, to address the dependence of these stress responses on p53 function, we measured p53, Δ113p53, p21, and bax expression by quantitative RT-PCR in wdr36 morphants that also were depleted of either p53 isoform. This was accomplished by co-injecting one-cell stage embryos with wdr36 MO and one of two previously described p53 MOs; one that targets the ATG start codon (p53 ATG MO), which will only disrupt the translation of full-length p53 protein, and a second splice-disrupting MO targeting the intron 4–exon 5 boundary (p53EI5), which will affect both full-length and Δ113p53 proteins (48,50). In these experiments, we found that the induction of both Δ113p53 and p21 caused by loss of Wdr36 was dependent on p53 function. Knockdown of full-length p53 alone partially rescued these stress–response genes to control levels, while co-depletion of both p53 isoforms completely prevented their induction (Fig. 7F).
In this study, a function for the POAG gene identified at locus GLC1G, WDR36, was determined. This is the first POAG gene currently identified for which a clear function has been described. The two previously discovered genes, MYOCILIN and OPTINEURIN, have both been extensively studied, but no endogenous function has yet been determined.
We have provided evidence that WDR36 functions in the processing of 18S rRNA. BLAST studies and protein alignments have shown that WDR36 is structurally homologous to yeast Utp21, which is part of the SSU processome. In yeast, this processome is essential for the maturation of 18S rRNA, and thus ribosome biogenesis (39). Evidence for Wdr36 as the functional homolog of Utp21 comes from data showing tissue distribution, sub-cellular localization and loss-of-function phenotypes.
Ribosome biogenesis is an essential process, so ubiquitous expression of 18S rRNA processing genes is expected. Previous evidence of widespread expression in mouse and human tissues (27), along with data presented here of expression patterns by both RT–PCR and in situ hybridization, shows that wdr36 is ubiquitously expressed. Further, enrichment in highly proliferative tissues was observed, consistent with an increased demand for ribosome biogenesis in proliferating cells.
At the sub-cellular level, Utp21 is localized to the nucleolus, the site of rRNA processing in the cell (38). Fusion proteins showed that Wdr36 also localizes to the nucleolus in vertebrate cells. YFP-tagged Wdr36 was found in this sub-compartment of the nucleus, as demonstrated by co-expression studies with the chromatin maker Histone2A, and a known nucleolar component, B23. Supporting these findings, mass spectrometry analysis of nucleoli isolated from cultured human cells showed WDR36 to be a component of this structure (http://www.lamondlab.com/NOPdb/) (51). Along with nucleolar localization, the epitope-tagged Wdr36 fusion proteins also showed strong cytoplasmic expression. Potentially, this may be a consequence of over-expression or epitope tagging. Alternatively, endogenous Wdr36 may also reside within the cytoplasm, similar to other nucleolar proteins (52).
Overall, our studies also do not rule out the possibility of additional functions for Wdr36. Some nucleolar proteins are known to have functions outside of ribosome biogenesis and these proteins also localize in the cytoplasm (47). For example, there is evidence that nucleolar components can directly influence the cell cycle and that proteins such as B23 and rRNA processing factors, re-localize to structures within the cytoplasm during mitosis (47).
The loss-of-function phenotypes described in this study provide evidence that Wdr36 is the functional homolog of Utp21. A wdr36 viral insert mutant shows defects in highly proliferative tissues; the eye, CNS and gut. These defects are consistent with the enriched expression in proliferating tissues observed by in situ hybridization and a higher need for ribosome biogenesis with cell expansion. MO knockdown of Wdr36 shows equivalent phenotypes as the mutants, confirming that these defects are due to loss of Wdr36. Informatively, both mutants and morphants showed disrupted 18S rRNA processing. Northern blot analysis indicated a strong reduction in processed 18S rRNA products and accumulation of larger immature pre-RNAs. In addition, both wdr36 mutant and morphant embryos show disrupted nucleolar morphology. These data suggest that the accumulation of unprocessed rRNAs within the nucleolus causes the enlarged nucleolar phenotype.
The results we have presented here are consistent with WDR36 being the functional homolog of Utp21, a component of the SSU processome and a protein involved in 18S rRNA processing. While an essential role for human WDR36 in rRNA processing is implicated by our experiments, the exact function that WDR36 plays within the processome is not clear. WDR36 contains nine WD40 repeat domains, which commonly facilitate protein–protein interactions. WDR36 may be necessary for structural integrity of the processome complex and/or have a necessary enzymatic function. The yeast work done to date on Utp21 does not offer any further insight. Utp21 was identified as part of the SSU processome by showing physical interaction with components of the complex, but its specific function was not probed (38).
It is surprising that disrupting a component of such a cell essential process does not result in earlier, more severe phenotypes. The maternal expression of wdr36 provides one possibility for the mild early phenotypes. WT wdr36 transcripts and perhaps protein derived from the WT allele of the wdr36 heterozygotic parent likely enables the mutant embryo to survive early development. It is not until after the WT maternal supplies are lost by degradation or diluted during cell divisions that phenotypes can be observed. This interpretation is suggested from observations within the developing retina, where earlier-born cells within the central retina show a more normal nucleolar morphology than later-born neurons at the periphery or in actively proliferating cells of the ciliary margin. This also explains the observed earlier phenotypes found with the ATG MO, which will block translation of both maternal and zygotic mRNA, although it still will not affect maternal protein.
Activation of the p53 stress–response pathway after disruptions in nucleolar morphology is well documented (47). We show that in wdr36 mutants and morphants, the p53 stress–response pathway is also activated. With loss of Wdr36 during embryogenesis, the disruption in nucleolar homeostasis results in significant upregulation of the truncated form of p53, Δ113p53. It has previously been reported that this isoform results in an arrest of the cell cycle, and may also act to block p53-mediated apoptosis (48,49). Consistently, the loss of Wdr36 initially results in delayed cell cycle progression, without cell death. This response, versus one including apoptosis, may be advantageous during development by giving embryos a chance to recover from a transient stress. With loss of Wdr36, however, stress caused by nucleolar defects is not transient, and the classic p53 stress–response, including apoptosis, appears to occur. Evidence for a direct role of the p53 stress pathways in gene expression caused by Wdr36 deficiencies was found by blocking p53 using MOs. While complete molecular recovery of the phenotypes was found, the gross appearance of the wdr36 morphant embryos was not significantly affected by co-injection of either p53 MO. To address whether maternal stores of p53 protein—which would not be affected by the MOs—were preventing phenotype rescue, wdr36 MO was also injected into embryos from an incross of p53M214K/M214K zebrafish (53). These fish are homozygous for a non-functional p53 allele and are thus not able to maternally contribute WT p53. Again, no change in phenotype versus WT embryos injected with wdr36 MO was noted. Together, this suggests that either alternate stress pathways are activated, or simply the inability to establish functional protein translation machinery itself causes the observed gross phenotypes.
The evidence that we have presented, that WDR36 is important for 18S rRNA processing, leads to the question whether WDR36 is the POAG causative (or modifying) gene at locus GLC1G? The possibility of a gene encoding an essential and ubiquitous protein, yet underlying a very specific pathogenesis-like POAG, is surprising, but is not without precedence. For example, four pre-mRNA splicing factors have recently been identified as causative for certain types of autosomal dominant retinitis pigmentosa (RP), another retina-specific blinding disorder in humans (54–57). In addition, mutations in essential nucleolar, 18S rRNA processing genes, such as RPS19 and dyskerin, have been shown to cause tissue-specific human disease (58–60). If WDR36 does play a role in POAG, it likely occurs in a permissive genetic context, much like the case with OPTENEURIN. This is not surprising given the complex nature of POAG. In line with this possibility, we have shown that Wdr36 interacts with p53; a key regulator of the apoptotic pathway. In POAG, it is established that RGCs are lost because of apoptosis (61,62). Genetic studies have been conducted to assess whether a direct relationship between p53 genetic variants and POAG exists, but, like other genes implicated in POAG, different studies have yielded somewhat contradictory results (63–66). Our data showing an interaction of p53 and WDR36 suggest that variants in these genes could potentially synergize in POAG pathogenesis. It may be that genetic changes in each gene alone do not cause disease, but when co-inherited, RGCs are at a higher risk for apoptosis, particularly when stressed by elevated IOP or other factors. Consistent with this concept, WDR36 has already been suggested to affect the severity of MYOCILIN-based disease (30). Overall, the functional evidence we have presented lays essential groundwork to address these issues, but does not answer whether WDR36 is involved in POAG. Genetic studies have shown that associated variants are present in a heterozygous state, suggesting disease may be caused by haploinsufficiency, dominant negative activities or possibly through a gain of function mechanism (27,30). Our initial results with heterozygous wdr36 mutants suggest that haploinsufficiency of WDR36 alone is not sufficient for disease. Further studies determining the effects of disease-associated variants on the normal function of WDR36 will be necessary to address how and if WDR36 may influence POAG.
In conclusion, we have shown that WDR36 functions in 18S rRNA processing. Loss of Wdr36 in zebrafish leads to proliferation defects in the eye and other tissues, as well as severe dysgenesis of lens. At the cellular level, loss of Wdr36 disrupts 18S rRNA maturation and leads to nucleolar morphology defects resulting in activation of the p53 stress–response pathway. The results from this study provide basis for further research to determine whether WDR36 variants are in fact involved in POAG.
A 563 bp fragment of the zebrafish wdr36 gene was amplified by PCR with the following primers: F: 5′-TGCAGACTATGAGCCGACTG-3′ and R: 5′-TAATACGACTCACTATAGGGCGACTTGGGCCAGGTCAAAATCT-3′. The R primer includes a T7 RNA polymerase promoter (shown in italics). This PCR product was used as a template for cRNA probe generation. cRNA probes were produced and whole-mount in situ hybridization was conducted as described previously (67,68). Ten to fifteen embryos, treated with 0.003% phenylthiourea (PTU), were analyzed in each in situ experiment, PTU inhibits melanin synthesis. For post-hybridization sectioning, embryos were fixed in 4% paraformaldehyde/PBS and infiltrated with 15% sucrose, 30% sucrose and then 100% Tissue-Tek OCT (Miles Inc., Elkhart, IN, USA). Embryos were oriented in a freezing mold and 10 µm sections were cut on a cryostat and mounted on gelatin-coated glass slides.
Total RNA was extracted from either single zebrafish embryos using 100 µl TRIZOL reagent (Invitrogen, Carlsbad, CA, USA) and 1 µl 10 mm glycogen (Roche, Indianapolis, IN, USA) or from pools of 10–50 embryos using the RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA). cDNA was produced using 0.25–1.0 µg of RNA with the SuperScript III Reverse Transcriptase enzyme (Invitrogen). Semi-quantitative PCR was done using Amplitaq Gold (Roche). qPCRs were carried out using the iQ SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) on the iCycler iQ Real-Time Detection System (Bio-Rad). At least three-independent cDNA samples were assayed in triplicate for each primer set. ef1α was used for normalization. Data were analyzed using the ΔΔCt method (69). Primers are listed in Table 1.
C-terminally tagged versions of Wdr36, B23 and H2A were generated using the Gateway Clonase II system (Invitrogen) and tol2-based vectors were modified for Gateway recombinations (70). Constructs were generated with primers and vectors found in Table 1. The final constructs were co-injected either with a plasmid containing Gal4 under the control of the ef1α promoter or into the HSP70–Gal4 transgenic zebrafish line. Constructs were injected into 1–4-cell embryos as described above. Injected HSP70–Gal4 tg embryos were subjected to a 30 min incubation at 37°C at 24 hpf to activate Gal4 expression and consequently the UAS promoter to drive fusion protein expression. Embryos were imaged from 24 to 48 hpf by confocal microscopy.
Histological specimens for light microscopy were processed as described previously (71). In brief, embryos were fixed in primary fixative [2% paraformaldehyde, 2.5% glutaraldehyde, 3% sucrose, 0.06% phosphate buffer (pH 7.4)] at 4°C for 24 h and then washed in 0.1 m PBS, dehydrated through an ethanol series and propylene oxide and then infiltrated with EMbed-812/Araldyte resin mixture. Semi-thin plastic sections were cut with a glass knife on a JB4 microtome and stained with 1% toluidine blue in 1% borax. For transmission electron microscopy, an additional fixation in 1.0% osmium was included, followed by dehydration in MeOH/Aralhydrate. Embryos were then embedded in EMBed-812/DER 736. Ultrathin sections (60–70 nm) were collected on coated grids and stained with uranyl acetate and lead citrate for contrast. Images were captured using Hitachi H600 transmission electron microscopy.
Two MOs were designed against wdr36: one to block mRNA splicing and one to block translation. The wdr36 MOs and previously described MOs against p53 isoforms were ordered from Gene Tools (Corvallis, OR, USA) (48,50). A list of MOs used is given in Table 1. MOs were resuspended in water and injected into one- to two-cell stage embryos using a Nanoject II injector (Drummond Scientific, Broomall,PA, USA).
Both eyes from three adult wdr36 heterozygous and three WT siblings were dissected and used for expression analysis by qRT–PCR as described earlier. Primers used are listed in Table 1. To determine the body length to eye ratio, the lengths of four adult fish from each group were measured from tip of mouth to beginning of the tail by ruler and eye size was determined using morphometric analysis (MetaMorph Software, Molecular Devices, Sunnyvale, CA, USA) of dorsal view images, using the same ruler for calibration. These measures were then divided to obtain the ratio of eye size: body length and groups were compared by Student’s t-test. Corneal thickness was determined by using morphometric analysis (MetaMorph) to measure the width of corneas from histological sections of three eyes from each group. For each fish, five different sections were measured and averaged before comparison was made across fish. Groups were then compared by Student’s t-test.
Total RNA was isolated as described earlier. One microgram of total RNA was separated on 1% agarose–formaldehyde gels and capillary transferred with 20× SSC onto positively charged nylon membranes (Roche). Blots were analyzed using the DIG Northern Starter Kit following the manufacturer's protocol (Roche). A plasmid previously described containing the 18S target sequence was generously provided by the Dawid Lab (NICHD, NIH) and used to make DIG-labeled probes against 18S rRNA (46). An 815 bp fragment of ef1a was PCR amplified, TA cloned into the pCRII-TOPO vector (Invitrogen) and used to make the DIG-labeled control probe. Probes were produced as per manufactures recommendation with 1 µg of linearized plasmid (Roche, DIG Northern Starter Kit). Primers to amplify ef1α: 5′-TGATCTACAAATGCGGTGGA-3′; 5′-GAGACTCGTGGTGCATCTCA-3′.
Cell cycle/DNA content was analyzed as described previously (72). Approximately 30–40 whole embryos at specified stages were pooled, washed with E3 solution (5 mm NaCl, 0.17 mm KCl, 0.33 mm CaCl2, 0,33 mm MgSO4), placed into DMEM+10% serum and disaggregated on ice. Cells were serially passed through 105 and 40 µm filters, washed with 1X PBS, stained with 0.05 mg/ml propidium iodide and RNase treated. Three pools of embryos for each condition and time point were assayed independently. FACS analysis was done on a FACS Calibur (Becton-Dickinson, San Jose, CA, USA) and analyzed using Modfit LT software (Verity Software House, Topsham, ME, USA).
National Institutes for Health (R01EY16060 to B.A.L., F30AG029763 to J.M.S., and T32 EY014537—Research Training Program in Vision Science to J.M.S.).
We thank Michael Walter, PhD for informative and helpful discussions; Igor Dawid, PhD and Adam Amsterdam, PhD for reagents; Pat Cliff and Melissa Reske for technical support; the MCW Flow Cytometry Core Facility and Clive Wells, PhD of the MCW EM Core Facility.
Conflict of Interest statement. None declared.