Novel alternatively spliced variant form of human CDK5RAP2

MORC2 regulates RBM39-mediated CDK5RAP2 alternative splicing to promote EMT and metastasis in colon cancer

Colorectal carcinogenesis and progression are associated with aberrant alternative splicing, yet its molecular mechanisms remain largely unexplored. Here, we find that Microrchidia family CW-type zinc finger 2 (MORC2) binds to RRM1 domain of RNA binding motif protein 39 (RBM39), and RBM39 interacts with site 1 of pre-CDK5RAP2 exon 32 via its UHM domain, resulting in a splicing switch of cyclin-dependent kinase 5 regulatory subunit associated protein 2 (CDK5RAP2) L to CDK5RAP2 S. CDK5RAP2 S promotes invasion of colorectal cancer cells in vitro and metastasis in vivo. Mechanistically, CDK5RAP2 S specifically recruits the PHD finger protein 8 to promote Slug transcription by removing repressive histone marks at the Slug promoter. Moreover, CDK5RAP2 S, but not CDK5RAP2 L, is essential for the promotion of epithelial-mesenchymal transition induced by MORC2 or RBM39. Importantly, high protein levels of MORC2, RBM39 and Slug are strongly associated with metastasis and poor clinical outcomes of colorectal cancer patients. Taken together, our findings uncover a novel mechanism by which MORC2 promotes colorectal cancer metastasis, through RBM39-mediated pre-CDK5RAP2 alternative splicing and highlight the MORC2/RBM39/CDK5RAP2 axis as a potential therapeutic target for colorectal cancer.

An alternative splice isoform of mouse CDK5RAP2 induced cytoplasmic microtubule nucleation

The centrosome lacks microtubule (MT)-nucleation activity in differentiated neurons. We have previously demonstrated that MTs were nucleated at the cytoplasm of mouse neurons. They are supposed to serve seeds for MTs required for dendrite growth. However, the factors that activate the cytoplasmic γ-tubulin ring complex (γTuRC) are unknown. Here we report an alternative splicing isoform of cyclin-dependent kinase 5 regulatory subunit-associated protein 2 (CKD5RAP2) as a candidate for the cytoplasmic γTuRC activator. This isoform lacked exon 17 and was expressed predominantly in the brain and testis. The expression was transient during the development of cortical neurons, which period coincided with the period we reported cytoplasmic MT nucleation. This isoform resulted in a frameshift and generated truncated protein without a centrosomal localization signal. When this isoform was expressed in cells, it localized diffusely in the cytoplasm. It was co-immunoprecipitated with γ-tubulin and MOZART2, suggesting that it can activate cytosolic γTuRCs. After cold-nocodazole depolymerization of MTs and subsequent washout, we observed numerous short MTs in the cytoplasm of cells transfected with the cDNA of this isoform. The isoform-overexpressing cells exhibited an increased amount of MTs and a decreased ratio of acetylated tubulin, suggesting that MT generation and turnover were enhanced by the isoform. Our data suggest the possibility that alternative splicing of CDK5RAP2 induces cytoplasmic nucleation of MTs in developing neurons.

Centromeric chromatin integrity is compromised by loss of Cdk5rap2, a transcriptional activator of CENP-A

Centromeres are chromosomal loci where kinetochores assemble to ensure faithful chromosome segregation during mitosis. CENP-A defines the loci by serving as an epigenetic marker that recruits other centromere components for a functional structure. However, the mechanism that controls CENP-A regulation of centromeric chromatin integrity remains to be explored. Separate studies have shown that loss of CENP-A or the Cdk5 regulatory subunit associated protein 2 (Cdk5rap2), a key player in mitotic progression, triggers the occurrence of lagging chromosomes. This prompted us to investigate a potential link between CENP-A and Cdk5rap2 in the maintenance of centromeric chromatin integrity. Here, we demonstrate that loss of Cdk5rap2 causes reduced CENP-A expression while exogenous Cdk5rap2 expression in cells depleted of endogenous Cdk5rap2 restores CENP-A expression. Indeed, we show that Cdk5rap2 is a nuclear protein that acts as a positive transcriptional regulator of CENP-A. Cdk5rap2 interacts with the CENP-A promoter and upregulates CENP-A transcription. Accordingly, loss of Cdk5rap2 causes reduced level of centromeric CENP-A. Exogenous CENP-A expression partially inhibits the occurrence of lagging chromosomes in Cdk5rap2 knockdown cells, indicating that lagging chromosomes induced by loss of Cdk5rap2 is due, in part, to loss of CENP-A. Aside from manifesting lagging chromosomes, cells depleted of Cdk5rap2, and thus CENP-A, show increased micronuclei and chromatin bridge formation. Altogether, our findings indicate that Cdk5rap2 serves to maintain centromeric chromatin integrity partly through CENP-A.

Activation of an exonic splice-donor site in exon 30 of CDK5RAP2 in a patient with severe microcephaly and pigmentary abnormalities

This report constitutes the first report of a cryptic exonic splice‐donor site in CDK5RAP2, highlights the importance of evaluating novel splice mutations, and suggests that the phenotypic range associated with CDK5RAP2 mutations may include skin pigmentary abnormalities.

Species-Specific Expression of Full-Length and Alternatively Spliced Variant Forms of CDK5RAP2

CDK5RAP2 is one of the primary microcephaly genes that are associated with reduced brain size and mental retardation. We have previously shown that human CDK5RAP2 exists as a full-length form (hCDK5RAP2) or an alternatively spliced variant form (hCDK5RAP2-V1) that is lacking exon 32. The equivalent of hCDK5RAP2-V1 has been reported in rat and mouse but the presence of full-length equivalent hCDK5RAP2 in rat and mouse has not been examined. Here, we demonstrate that rat expresses both a full length and an alternatively spliced variant form of CDK5RAP2 that are equivalent to our previously reported hCDK5RAP2 and hCDK5RAP2-V1, repectively. However, mouse expresses only one form of CDK5RAP2 that is equivalent to the human and rat alternatively spliced variant forms. Knowledge of this expression of different forms of CDK5RAP2 in human, rat and mouse is essential in selecting the appropriate model for studies of CDK5RAP2 and primary microcephaly but our findings further indicate the evolutionary divergence of mouse from the human and rat species.

Subdiffraction imaging of centrosomes reveals higher-order organizational features of pericentriolar material

The centrosome is the main microtubule organization centre of animal cells. It is composed of a centriole pair surrounded by pericentriolar material (PCM). Traditionally described as amorphous, the architecture of the PCM is not known, although its intricate mode of assembly alludes to the presence of a functional, hierarchical structure. Here we used subdiffraction imaging to reveal organizational features of the PCM. Interphase PCM components adopt a concentric toroidal distribution of discrete diameter around centrioles. Positional mapping of multiple non-overlapping epitopes revealed that pericentrin (PCNT) is an elongated molecule extending away from the centriole. We find that PCM components occupy separable spatial domains within mitotic PCM that are maintained in the absence of microtubule nucleation complexes and further implicate PCNT and CDK5RAP2 in the organization and assembly of PCM. Globally, this work highlights the role of higher-order PCM organization in the regulation of centrosome assembly and function.