Investigating Embryonic Expression Patterns and Evolution of AHI1 and CEP290 Genes, Implicated in Joubert Syndrome

Joubert syndrome and related diseases (JSRD) are developmental cerebello-oculo-renal syndromes with phenotypes including cerebellar hypoplasia, retinal dystrophy and nephronophthisis (a cystic kidney disease). We have utilised the MRC-Wellcome Trust Human Developmental Biology Resource (HDBR), to perform in-situ hybridisation studies on embryonic tissues, revealing an early onset neuronal, retinal and renal expression pattern for AHI1. An almost identical pattern of expression is seen with CEP290 in human embryonic and fetal tissue. A novel finding is that both AHI1 and CEP290 demonstrate strong expression within the developing choroid plexus, a ciliated structure important for central nervous system development. To test if AHI1 and CEP290 may have co-evolved, we carried out a genomic survey of a large group of organisms across eukaryotic evolution. We found that, in animals, ahi1 and cep290 are almost always found together; however in other organisms either one may be found independent of the other. Finally, we tested in murine epithelial cells if Ahi1 was required for recruitment of Cep290 to the centrosome. We found no obvious differences in Cep290 localisation in the presence or absence of Ahi1, suggesting that, while Ahi1 and Cep290 may function together in the whole organism, they are not interdependent for localisation within a single cell. Taken together these data support a role for AHI1 and CEP290 in multiple organs throughout development and we suggest that this accounts for the wide phenotypic spectrum of AHI1 and CEP290 mutations in man.


Introduction
Joubert syndrome and related diseases (JSRD) are a group of inherited ciliopathies, characterised by a cerebello-retinal-renal phenotype. The brain phenotype is a developmental midbrain malformation leading to cerebellar vermis hypoplasia or aplasia, and seen in brain MRI imaging as a ''molar tooth sign'' [1]. Other structural brain defects including hypoplasia of the corpus callosum and occipital meningoencephalocele have been reported [2,3].
Retinal dysplasia and degeneration occur in a proportion of patients and may lead to progressive blindness [1,4,5]. Ocular coloboma may also be a feature [6]. Renal disease is variable, with nephronophthisis [7] and multicystic dysplasia [8] as reported phenotypes, which may lead to end stage renal failure. Consistent with JSRD as a ciliopathy are the findings of polydactaly and liver fibrosis [9,10].
Mutations in AHI1 and CEP290 are both a frequent causes of Joubert syndrome and genetic variants in both genes may act as modifier alleles, especially in regard to a retinal and CNS phenotype [27,28]. Mutations in AHI1 (Abelson-helper integration site-1) are the most common genetic cause of JSRD, accounting for 12% of cases and 20% of individuals with Joubert syndrome and Leber's congenital amaurosis [13,14,29,30,31].
AHI1 is highly conserved throughout evolution and encodes the Ahi1 protein (also known as Jouberin). We have previously demonstrated that Ahi1 localises to centrosomes/basal bodies of renal epithelial cells, and that it interacts with the nephrocystin-1 protein [32]. At a genomic level, there is evidence for AHI1 mutations/polymorphisms having an oligogenic effect and modulating the phenotype. For example Tory et al. describe 5 patients with homozygous NPHP1 mutations together with a R830W mutation in AHI1, which leads to a more severe central nervous system defect [27]. Similarly, the relative risk (RR) of a retinal defect associated with nephronophthisis was significantly increased (RR = 7.5 (95% CI 4.0-11.2)) in the presence of the R830W AHI1 mutation [31].
Mutations in CEP290 may cause a wide spectrum of human disease ranging from isolated Leber's congenital amaurosis to JSRD, Meckel syndrome (MKS) and Bardet-Biedl syndrome (BBS) [33,34,35]. Renal disease is common in patients with CEP290 mutations, and may be secondary to NPHP and renal cortical cysts [3,36].
CEP290 is highly conserved throughout evolution and encodes the Cep290 protein (also known as nephrocystin-6) [3]. We have previously demonstrated that Cep290 localises to centrosomes/ basal bodies of renal epithelial cells and that it interacts with ATF4 [3]. Additional CEP290/Cep290 interactions include a genetic interaction with CC2D2A [18] and protein complex interactions with CP110 and Rab8a [42]. In a similar manner to patients with NPHP1 mutations, there is some evidence that alleles of AHI1 may also modify the neurological phenotype of patients with CEP290 mutations [28].
Given this evidence implicating an important role for CEP290 and AHI1 in modulating brain, eye and other phenotypes in JSRD, we sought to identify the pattern of both AHI1 and CEP290 expression in kidney and central nervous system during development. Using human embryonic and fetal tissue we describe the expression pattern of AHI1 and CEP290 during stages of human renal and brain development (Table 1).
To determine whether AHI1 and CEP290 may have co-evolved, we performed a genomic survey of a large group of organisms across eukaryotic evolution. We also explored whether Ahi1 was required for recruitment of Cep290 to the centrosome.
Our novel data emphasises the important and extensive role of AHI1 and CEP290 in vertebrate development, from essential cellular signalling organelles, such as the primary cilium, to complex organ systems such as the brain and kidney.

Ethics statement
This study was conducted with full ethical approval. For human embryonic and fetal tissue samples, the samples were collected with appropriate maternal consents and ethical approval by the Newcastle and North Tyneside 1 Research Ethics Committee, UK.
In situ hybridisation probes for human expression studies of AHI1 were generated using PCR and subcloning into pGEM-Teasy vectors. All clones were sequenced to confirm orientation with respect to T7 and SP6 promoters. Both sense and antisense RNA probes were produced by in vitro transcription from the T7 and SP6 RNA polymerase sites respectively of the pGEM-T Easy vector. Probes were generated from two regions of the AHI1 gene (BC094800), nucleotides 1 to 276 (in 59UTR) and nucleotides 2105 to 2635 (in 39UTR) using the ''Riboprobe'' system (Promega). Both probe sets gave similar patterns however the 59UTR probe set gave a higher background in preliminary experiments (data not shown). The 39UTR probe set was used for all the experiments reported here.
In situ hybridisation probes for human expression studies of CEP290 were generated from two regions of CEP290 (NM_025114) using the ''Riboprobe'' system (Promega) as above: nucleotides 675 to 1381 and 6265 to 6924 (both in the coding sequence). Although a similar pattern was detected with both antisense probes, there was a higher background with the 675 to 1381 nucleotide sense probe (data not shown). The 6265 to 6924 nucleotide probe set was used for all the experiments reported here.
Antisense probes for sonic hedgehog (SHH, NM_000193) were used as a positive control. The SHH probe, which contains nucleotides from 460 to 950, showed the expected specific staining in the spinal cord and notochord (data not shown). Antisense RNA probe of developing brain homeobox 1 (DBX1, NM_001029865) transcribed from nucleotides 308 to 830 which shows specific staining at the sulcus limitans of spinal cord was also used as an experimental control (data not shown). Expression patterns were analysed using the Axioplan 2 imaging system (Zeiss).

Sequence analysis
Putative Ahi1 and Cep290 orthologues were identified using a combination of reciprocal best BLASTP and PSI-BLAST, with human Ahi1 (isoform a, NP_001128302.1) and Cep290 (NP_079390.3) as the query sequences [45,46]. The initial Ahi1 search generated a very large number of false positives, which were only similar in terms of WD40 repeats. To overcome this problem, we performed further searches using human Ahi1 amino acids 1-545. These protein sequences were used to query the nonredundant predicted proteomes of 44 organisms (33 flagellate, 11 non-flagellate) chosen to represent a wide evolutionary spread of eukaryotes. Searches were carried out using the NCBI stand-alone BLAST+ application (version 2.2.25+, [47]) against BLAST-formatted genomes from a variety of sources (see Supplementary Figure S1). We used the following criteria to define a putative Ahi1 or Cep290 orthologue: 1) An e-value less than 1e-05 (BLASTp) or 1e-10 (PSI-BLAST), 2) A reciprocal BLAST which returns the original sequence, and 3) Manual examination of sequences and alignments constructed using MAFFT [48].

AHI1 expression during human brain and kidney development
There has been no previous documentation of AHI1 spatial expression during development in human tissue. Previous characterisation has been limited to hybridisation to RNA northern blot showing expression in fetal brain, kidney, liver and lung [13,14] and Western blotting, demonstrating Ahi1 protein expression in fetal cerebellum and spinal cord [50].
Using specific probes targeted towards nucleotides 2105 to 2635 of AHI1 mRNA we observed a strong and specific pattern of expression in the brain and spinal cord from CS16 to CS23 ( Figure 1; see also Table 1 for developmental overview). AHI1 expression was abundant in the developing telencephalon ( Figure 1B), especially the neuroepithelium, and in the neural retina ( Figure 1C, D respectively). Strikingly, we observed expression within the choroid plexus ( Figure 1I, J). Prominent expression was also seen in the mesencephalon and throughout the developing hindbrain (metencephalon and myelencephalon, Figure 1F-H) including the rhombic lip of the developing cerebellum ( Figure 1H). There was particularly prominent staining within the alar plate of the spinal cord as well as in the spinal and other ganglia of the peripheral nervous system ( Figure 1K, L). SHH and DBX1 antisense probes were used as positive controls and the expected patterns of expression were seen in the developing spinal cord (data not shown).
AHI1 expression during human nephrogenesis was examined from stages CS14 through to 9 PCW (Figure 2). These studies revealed that AHI1 transcripts are abundant in both the developing mesonephros and metanephros. Extra renal sites of expression included prominent staining of the embryonic liver and embryonic gonad ( Figure 2D, F). Detailed studies through mesonephric development showed discrete AHI1 expression in the mesonephric excretory unit at CS12 (data not shown) and CS14-CS22 ( Figure 2G). By CS23, AHI1 expression was visible in degenerating glomeruli of the mesonephros and at 9 PCW, expression was detectable in the mesonephric tubule, ducts and paramesonephric duct, structures that will form parts of the future reproductive system ( Figure 2G/CS23 and 2G/9PCW respectively). AHI1 expression was seen in early (permanent) metanephric kidney ( Figure 2H) with intense staining in the metanephric cap and ureteric bud at CS14 and CS16. By CS22 and later, developing glomeruli, tubules and collecting ducts strongly expressed AHI1 ( Figure 2H). More detailed analysis of AHI1 expression at 9 PCW revealed intense staining in both tubular epithelium and glomerular structures ( Figure 2H/9 PCW). Expression was seen in cortical and medullary epithelium, including the epithelium lining of the renal calyces. Glomerular staining was intense in a subset of cells whilst there was an absence of expression in the renal interstitium, indicating low levels of AHI1 transcript in interstitial cells and vascular endothelium.
Thus, in human tissue AHI1 is strongly expressed in all developing renal tissues from the mesonephros to the precursors of the adult kidney, with a broad expression throughout the nephron and collecting ducts during renal development.

CEP290 expression during human brain and kidney development
The CEP290 spatial expression during development in human tissue has not previously been reported. Previous characterisation has been limited to hybridisation to RNA northern blot of adult tissues showing expression of a major CEP290 transcript of approximately 8 kb that is expressed strongly in placenta and weakly in adult brain [3].
Using specific probes targeted towards nucleotides 6265 to 6924 of CEP290 mRNA we observed a strong and specific pattern of expression in the brain and spinal cord from CS16 to CS22  Figure 3). CEP290 expression was seen in the developing telencephalon ( Figure 3B), especially the neuroepithelium, and in the neural retina ( Figure 3C, D respectively). Prominent expression was seen in the choroid plexus ( Figure 3E, F) and throughout the developing hindbrain ( Figure 3H). Expression of CEP290 is also seen in the rhombic lip of the developing cerebellum ( Figure 3H) and the mesencephalic and cerebellar neuroepithelium ( Figure 3I, J). In the spinal cord there was particularly prominent staining within the alar plate and some staining of the spinal and sympathetic ganglia ( Figure 3K, L). This expression pattern seen in the developing human brain with CEP290 mirrors the pattern seen for AHI1 (compare Figure 1 with Figure 3).
CEP290 expression during human nephrogenesis was examined from CS12 through to 9 PCW (Figure 4). CEP290 transcripts are abundant in both the developing mesonephros and metanephros ( Figure 4E, F). Specific areas of CEP290 expression are seen in the mesonephric excretory unit at CS12-CS23 ( Figure 4E). By 9 PCW, CEP290 expression was seen in early (permanent) metanephric kidney with intense staining in the developing glomeruli, tubules and collecting ducts ( Figure 4F, 4E/9PCW). Thus, like AHI1, CEP290 is strongly expressed in the mesonephros and metanephros precursors of the adult kidney. Similar to AHI1, expression of CEP290 was low/absent in the renal interstitium.
AHI1 and CEP290 are highly evolutionarily conserved and are generally present together in animal genomes This great similarity of expression in humans prompted us to ask if AHI1 and CEP290 might have co-evolved. We have previously shown using a reciprocal BLASTp approach that AHI1 is highly evolutionarily conserved, and is restricted to the genomes of ciliated organisms [49]. Given the strong overlap in expression between AHI1 and CEP290, we extended this analysis using PSI-BLAST and reciprocal BLASTp to identify additional homologues. We searched the predicted proteomes of 44 eukaryotic organisms chosen to represent a wide evolutionary spread of organisms, both ciliated (33 organisms) and non-ciliated (11 organisms). Both ahi1 and cep290 evolved prior to the evolution of multicellularity, are conserved from ciliated protozoa to man, and are restricted to the genomes of organisms that build cilia or flagella (Supplementary Figure S1). Putative orthologues of both AHI1 and CEP290 were present in the genomes of organisms that build only sensory cilia (Daphnia, Tribolium) as well as in the genomes of organisms that build either motile cilia/flagella only, or both motile and sensory cilia, implying that both ahi1 and cep290 are required for basal body/cilium function in cells with motile cilia and cells with sensory cilia. Building on our previous observation that ahi1 was only present in the genomes of organisms that build a canonical nine-triplet centriole [49], both ahi1 and cep290 were absent from the genome of Toxoplasma gondii and the two nematode genomes in our study, which all build specialised, non-canonical centrioles. Therefore, ahi1 and cep290 may both be required for triplet centriole function. In animals, ahi1 and cep290 were both found in all organisms except in Dipthera (Anopheles and Drosophila) and Gallus, indicating a possible common function in metazoa that is consistent with the expression data. However, while there was a general trend for ahi1 and cep290 to both be present in a given genome, this was not invariable. We found only a predicted AHI1 orthologue in 5 genomes (Aureococcus anophagefferens, Naegleria gruberi,, Trichomonas vaginalis, Trypanosoma brucei) and only a predicted CEP290 orthologue in a further 5 genomes (Anopheles gambiae, Drosophila melanogaster, Micromonas pulsilla, Monosiga brevicollis, Gallus gallus) suggesting that, while ahi1 and cep290 might function together in animals, they are not co-dependent in all organisms.

Cep290 does not require Ahi1 for localisation
Cep290 is a 290 kDa protein that is localised at the centrosome [3,36,42,51]. To examine whether Ahi1 is necessary for correct recruitment of Cep290 to centrosomes [52,53], we ablated Ahi1 expression using siRNA in IMCD3 renal epithelial cells [49] and visualised Cep290 and centrosomes using anti-Cep290 and antigamma tubulin immunofluorescence. We examined the appearance of the Cep290 signal at centrosomes throughout the cell. There was no difference in the distribution of centrosomes between Ahi1-silenced cells and controls ( Figure 5A), indicating that the centrosome is still able to re-orient to the apical cell surface in the absence of Ahi1. We found no obvious differences in the immunofluorescence pattern between control and Ahi1 siRNAtreated cells ( Figure 5B and C), regardless of the centrosome position within the cell, indicating that Ahi1 and Cep290 are not interdependent for localisation to centrosomes.

Discussion
We have demonstrated early embryonic expression of AHI1 and CEP290 in human embryonic tissues, with a remarkable similarity in expression patterns. Both AHI1 and CEP290 are widely expressed in multiple tissues, including developing brain, eye, spinal cord and renal tissues. Previous expression studies in murine tissues showed Ahi1 expression in neurons of the developing hindbrain/midbrain/forebrain, pituitary, testis and kidney [50]. Ahi1 expression in murine brain was reported during early embryonic life (day 10.5) and persisted into adult life. Interestingly, in Doering's studies, mouse Ahi1 protein expression was not seen in the developing or mature cerebellum although AHI1 was present in the human cerebellum, suggesting a species difference [50]. Indeed, these studies commented that the highest AHI1 expression in adult brain tissue was in the cerebellar vermis and cerebellar peduncles [50]. A functional role for AHI1 in brain development is suggested by the numerous structural defects associated with Joubert syndrome [13,14].
We were able to demonstrate prominent AHI1 expression in the developing retina. It is now well established that the light sensitive photoreceptor cells of the retina are specialised sensory cilia. Inherited forms of retinal degeneration, where these cells are affected are now a hallmark of diseases termed ciliopathies [54,55]. Numerous gene defects, encoding proteins expressed in cilia, may be associated with retinal degeneration. NPHP5 mutations are a prominent cause of retinal degeneration [56]. Ahi1 2/2 mutant mice fail to develop photoreceptor outer segments, leading to retinal degeneration [31,57].
Ahi1 was seen to be expressed at the transition zone of photoreceptor cells of wild-type mice, colocalising with the transition zone marker Rpgrip1 [57]. The Ahi12/2 murine model mimics closely the defect seen in JSRD patients with AHI1 mutations [7,9,30]. Our data of human AHI1 expression in retinal development confirms this requirement of AHI1 for normal photoreceptor development.
Consistent with the renal phenotype seen in some JSRD patients [29], we demonstrate for the first time AHI1 and CEP290 expression in the human embryonic kidney. Indeed, murine data showing detailed kidney expression of these genes is also lacking. The pattern of expression during development of the mesonephros and metanephros points to a role in tubulogenesis, epithelial cell polarity and maintenance of orientated cell division, as has been described for other ciliary proteins [58].
In human adult kidney, we have previously demonstrated that Ahi1 protein expression is limited to the distal tubule and collecting duct nephron segments [32], whilst murine kidney tissue from a 5 month week old animal revealed Ahi1 expression at the corticomedullary junction [59]. A postnatal increase in Ahi1 protein expression has also been described [59]. This suggests that renal Ahi1 expression must be regulated throughout development and into adult life, and that a role for maintenance of distal tubular function continues into adult life.
There have been no previous studies documenting the renal expression of CEP290 in human kidney. Previous studies have documented expression of murine Cep290, with RNA blot analysis revealing expression at embryonic day 7 to E17 in whole embryos, and kidney and brain expression at P0. [36]. In situ hybridisation studies in murine cerebellum have demonstrated expression in the external (proliferative) and internal granule layer in the midline at E18 and the latter expression persists into adulthood [36]. At the earlier human stages studied here, CEP290 expression is detected in cerebellar neuroepithelium, also a proliferative layer, as well as mesencephalic and telencephalic neuroepithelium.
The Cep290 protein has also been localized within photoreceptor cells, to the connecting cilium adjoining the outer segment to the photoreceptor cell body [3,60].
CEP290 mutations are the commonest monogenic form of Leber's congenital amaurosis (LCA) [61]. A murine model of Cep290 deficiency reveals early post-natal photoreceptor degeneration [60] whilst human studies in patients with CEP290 mutations suggest that whilst there is initial development of rod photoreceptors during early development, the distal segments of photoreceptors are structurally abnormal after maturation, causing early and severe loss of visual function typically within the first decade of life [61]. The retinal expression of CEP290 we have shown during early development in humans is consistent with a role for CEP290 in photoreceptor cell development.
The expression of both AHI1 and CEP290 in the choroid plexus is a novel and interesting finding. The choroid plexus produces cerebrospinal fluid and is circulated through the ventricles of the brain and the subarachnoid space. The ependymal cells lining the ventricular space are known to possess motile cilia, and that defects in these motile cilia can contribute to an obstruction of CSF flow leading to hydrocephalus [62,63,64,65]. In contrast, the murine ciliopathy model Tg737 orpk demonstrated hydrocephalus before formation of motile cilia in the ependymal cells [66], implicating defects in choroid plexus function (such as increased CSF production) as a primary event in these animals. Primary cilia are intimately related to the regulation of CSF production. Indeed clusters of primary cilia on choroid plexus epithelia cells have been shown to act as chemosensors to regulate CSF production [67]. Upon ahi1 knockdown in zebrafish, we have previously shown hydrocephalus, alongside other defects associated with a ciliopathy [49]. A similar appearance, with prominent hydrocephalus was seen with cep290 (alias nphp6) knockdown in zebrafish [68]. To date, murine models of Ahi1 and Cep290 knockdown have not demonstrated gross hydrocephalus. The brain defect that these mice have revealed includes cerebellar hypoplasia with a vermis/ midline fusion defect early in development [69].
In humans with JSRD, hydrocephalus is a rarer finding, however recently mutations in the CC2D2A gene (whose protein product interacts with CEP290 [18]) have been associated with ventriculomegaly and hydrocephalus [70]. Our experiments suggest that a functional role for AHI1 and CEP290 is likely within the developing brain, including the choroid plexus.
A proteomics and interaction mapping project has recently been published which details a number of important interactions for both Ahi1 and Cep290 [71]. Mks1, Mks6, and Tectonic1 bind to Ahi1 protein, which also co-purified with inversin, whilst Cep290 bound strongly with nephrocystin-5. Depletion of nephrocystin-5 had no effect on Cep290 localization, whilst nephrocystin-5 failed to localize to the centrosome in the absence of Cep290 [71]. The studies defined 3 protein ''connecting modules'' where nephrocystin proteins 21, 24 and 28 function together at the apical surface of the cell, nephrocystin-5 and -6 function at the centrosome and the MKS group of proteins (Mks1, Mks6 and Tctn2) function in a hedgehog signalling pathway [71]. Ahi1 was seen as a linking protein between MKS and centrosomal modules. In an attempt to determine whether Ahi1 is necessary for correct recruitment of Cep290 to the area around centrosomes, we found no loss of localisation, which is consistent with Sang's recently published localisation data set [71].
In conclusion, we have demonstrated in human development an expression pattern for AHI1 and CEP290 which includes neuronal, retinal and renal structures. AHI1 and CEP290 are highly conserved throughout evolution, but can be expressed and function independently.

Supporting Information
Supplementary Figure S1 Distribution of ahi1 and cep290 proteins and cilium and centriole architecture across eukaryotes. Putative ahi1 and cep290 homologues are found in organisms that build both motile and sensory cilia. Both are present in most animals; however ahi1 and cep290 can also occur individually in organisms. Ultrastructural information was not available for all organisms included in this study. A ? denotes an unknown architecture or one where conflicting data have been reported. Architectures are in accordance with those described in Simms et al. [49]. (TIF) Figure 5. Cep290 does not require Ahi1 for localisation to centrosomes. Immunofluorescence of IMCD3 monolayers co-transfected with siGLO and either negative-control siRNA (A and B), or siRNA against Ahi1 (A and C). (A) Graph to show centrosome position (gamma tubulin immunofluorescence, graph expressed as mean +/2 s.e.m.) from apical to basal in control and Ahi1-silenced cells. Apical refers to the upper 1.5-2 mm of the cell (the distance between the nuclear envelope and the plasma membrane in these cells; this is unchanged following silencing of Ahi1, data not shown). Arrows denote the apical and mid-cell positions used in B and C. (B and C) Confocal maximum intensity projections of control (B) and Ahi1-silenced (C) cells showing Cep290 (red), gamma tubulin (green) and DNA (blue). Transfected cells were identified using siGLO (yellow). In each case, the top panel represents an apically-oriented centrosome (position 1 on the graph, A) and the bottom panel represents a centrosome from the mid-cell region (position 2 on the graph). Scale bar: 5 mm. doi:10.1371/journal.pone.0044975.g005