In recent years, the detailed molecular functions of DYX1C1 and DCDC2 have been emerging. Both genes have been assigned roles in developmental dyslexia, neuronal migration and cilia. Given these overlapping functions, the question arises whether there is a functional relationship between these two genes.
Here, we show that DYX1C1 and DCDC2 proteins both interact with CPAP. We validate the previously suggested protein-protein interaction between CPAP and DYX1C1 by co-immunoprecipitation and identify the p23 domain as the interaction-mediating domain. This result is consistent with a previous study showing protein-protein interactions of DYX1C1 mainly via p23 (7). P23 domains are known to be important for protein-protein interactions, for example with the chaperone Hsp90 (42–44). In addition, it is known that DYX1C1 is interacting with Hsp70 and Hsp90 and has a chaperone function (16, 51, 52). Interestingly, TPR domain containing proteins (TTC proteins) have been reported to have important roles in cilia functioning in IFT and dynein assembly (53). A ciliary function of the CCT/TRiC chaperone complex has previously been described affecting the BBSome(54). Interestingly, it has been shown that Dyx1c1 interacts with the chaperones Cct3, Cct4, Cct5 and Cct8 while CCT4, CCT5 and CCT8 are reportedly localizing to the centrosome (16, 54). It is conceivable that DYX1C1 acts as a chaperone for DCDC2 and CPAP. It remains to be determined whether the chaperone-functions and cilia-related functions of DYX1C1 are disparate or functionally related. Next to its role in cilia, implicated functions of DYX1C1 include chaperone, ubiquitin system, estrogen signalling and autophagy roles (51, 52, 55, 56). DCDC2 has been proposed to function in microtubule stabilization and ciliary signalling (57, 58).
Both Dyx1c1 and Dcdc2 have been shown to produce neuronal migration defects when knocked down in rats (3, 6). Similarly, CPAP has been shown to regulate neuronal migration independently of its function in centrosome, via its microtubule-destabilizing domain PN2-3 (40). CPAP controls, via its PN2-3 domain, ciliary length and centriolar and ciliary tubulin assembly and disassembly (39, 59–61). CPAP promotes cilia disassembly and thereby keeping a neural stem cell pool (35). Ascl1— a proneural transcription factor known for its role in neurogenesis — is a transcriptional regulator of Cpap, while DYX1C1 perturbation affects ASCL1 expression (7, 40, 62). This indicates a complex regulatory mechanism of CPAP by DYX1C1 and ASCL1. FOXP2, a gene often associated with speech and language disorders, affects DCDC2 expression in SH-SY5Y cells providing a further link between language-related gene expression and DCDC2 (63). The complex DYX1C1/CPAP/DCDC2 might act in conjunction in migrating neurons, possibly independently of cilium and centrosome. Interestingly, DYX1C1 is highly upregulated in differentiating human neurons (32). Further studies in human stem cell derived neurons might shed more light on the role of these genes in neuronal migration.
Both DYX1C1 and DCDC2 have been reported as cytoskeletal interactors (7, 57). DYX1C1 associates with microtubule proteins and DCDC2 has a role in microtubule stabilization (7, 57). DCDC2 localizes to primary cilia and is involved in ciliary signaling (18, 58). Some studies report a physical localization of dyx1c1 at or around the basal body (6, 7, 31, 45). It might be a dynamic process, shuttling between the cytoplasm and temporal accumulation around the basal body. DCDC2 and DYX1C1 do not co-occur at the cilium and our immunoprecipitations were carried out on non-starved, cycling cells both suggesting that their combined action may take place in the cytoplasm, consistent with in situ proximity ligation assay (PLA) data (7). Future studies using PLA might pinpoint the subcellular localization of their interaction with CPAP.
Genetic interaction studies have been successfully used to study cellular pathways in model systems, for example zebrafish (64). They can provide useful information about protein-protein complexes, downstream pathways and parallel pathways. This approach has been used to dissect interactions of ciliary genes – for example, the genetic interactions of BBS and PCD genes have been modelled in zebrafish (46–49). Exacerbation of a ciliary phenotype can take place among physically interacting and among non-interacting proteins. For example, concomitant loss of bbs7 and bbs1 – members of the BBsome protein complex – and of dnah6 with dnai1 and dnah5 - where no physical interaction takes place – produces an aggravated phenotype (47, 48). Previous studies have reported a ciliary phenotype in dyx1c1 and dcdc2 morphant zebrafish (18, 25). Here, we show an exacerbated phenotype in double dyx1c1 and dcdc2b morpholino injected zebrafish. The tissue expression specificity as well as the subcellular localization might overlap only partially raising the question whether dyx1c1 and dcdc2 might act in parallel pathways thereby aggravating the ciliary phenotype. Future studies should address whether dyx1c1 morphants can be rescued by dcdc2b overexpression and vice-versa.
We show a mutual transcriptional regulation between DYX1C1 and DCDC2 in the human ciliated RPE1 cell line, but not in the morphant zebrafish. This might be due to dilution of the effect in the whole organism as compared to an isolated cell type. Transcriptional control likely is indirect as there is no evidence that either proteins would act as transcription factors. The regulation of DCDC2 by DYX1C1 has not been reported previously(7, 58, 65). This might point to a cell-specific or cilia-specific effect, which could be further studied using starved and non-starved cells as well as different non-ciliated and ciliated cell lines.