Study on the biological significance of YAP gene alternative splicing

概要

PTPN11 (Protein Tyrosine Phosphatase, Non-Receptor Type 11), which is also referred to as SHP2, encodes a ubiquitously expressed protein phosphatase. The function of SHP2 is best defined by its positive role in the RAS-ERK signaling pathway in the cytoplasm. On the other hand, SHP2 is reported to activate the transcription of Wnt/β-catenin target genes by tyrosine-dephosphorylating parafibromin, a core component of the nuclear PAF1 complex, in the nucleus. SHP2 distributes in both cytoplasm and nucleus in low cell density but is excluded from the nucleus in high cell density, indicating SHP2 nuclear function is regulated in a cell-density-dependent manner. Previously, Tsutsumi et al. reported that YAP/TAZ, promote nuclear accumulation of SHP2 at low cell density via physical interaction. However, immunoblotting of mouse NIH3T3 cell lysate using anti-YAP/TAZ antibody identifies two discrete bands for YAP protein, which is due to the differential splicing of exon 6 in mouse Yap gene. Similarly, human YAP 2α and YAP 2γ isoforms, which are considered as the conserved isoforms from mouse Yap-short isoform and the Yap-long isoform, are also due to the differential splicing of exon 6 in human YAP gene. However, to date no research has been conducted to elucidate the biological significance of the two conserved YAP isoforms nor the mechanism of splicing regulation.

To investigate whether human YAP 2α and YAP 2γ isoforms also form a complex with SHP2 in an isoform-specific manner. We constructed the expression vectors of all 8 human YAP isoforms from AGS human gastric cancer epithelial cells. Flag-tagged SHP2 was respectively co-expressed with the 8 HA-tagged human YAP isoforms in HEK 293T (Human Embryonic Kidney cells 293 with SV40 large T-antigen) cells. The cell lysates were subjected to immunoprecipitation with anti-Flag. We observed that exon 6 excluding isoforms (YAP 1α, 1β, 2α and 2β) could form complex with SHP2, whereas, exon 6 including isoforms YAP 1γ, 1δ, 2γ and 2β) could not. Exon 6 encodes a γ-segment which is inserted into and then disrupts the leucine-zipper motif (coiled-coli) in YAP/Yap. To verify the necessity of the leucine-zipper motif in the YAP-SHP2 complex formation, YAP 2α leucine-zipper mutant （YAP 2αLZM）was constructed and co-expressed with Flag-SHP2 in HEK293T cells. YAP 2αLZM fails to interact with SHP2. These findings suggest that human YAP interacts with SHP2, which requires functional leucine-zipper motif, is in an isoform-specific manner.

Since SHP2 protein doesn’t contain leucine-zipper motif, we hypothesized that TAZ, a homolog of YAP, mediated YAP-SHP2 interaction. We observed YAP 2α-TAZ heterodimerization by co-expressing Flag-YAP 2α and HA-TAZ in HEK293T cells. The interaction is diminished by using YAP 2αLZM or TAZLZM. Furthermore, YAP-SHP2 interaction is abrogated in the absence of TAZ in HEK293T cells. The observations suggest that YAP-SHP2 interaction requires leucin-zipper-mediated YAP-TAZ heterodimerization.

Next, we confirmed that SHP2 localized in both nucleus and cytoplasm under low cell density condition, whereas it is excluded from nucleus under high cell condition. To investigate whether SHP2 subcellular distribution is regulated by YAP in an isoform-specific manner, we constructed and overexpressed YAPS127A mutants, which localize in the nucleus even though under high cell density, in AGS cells. As a result, YAP 2αS127A strictly co-localizes with SHP2 in the nucleus whereas YAP 2γS127A does not, under high cell density. The result suggests SHP2 subcellular distribution is regulated in a YAP isoform-specific manner.

Given that YAP 2α and YAP 2γ differentially regulate SHP2 subcellular distribution, they might possess different biological significance. To test this idea, YAPKO AGS cells were established. Either YAP 2α or YAP 2γ was then reintroduced back to the YAPKO cells to establish isoform solely expressing AGS cells (referred to as YAPKO-2α and YAPKO-2γ). We compared the cell proliferation and motility of the obtained cells. Consequently, both YAPKO- 2α and YAPKO-2γ cells display enhanced proliferation and motility compared with parental YAPKO AGS cells. Markedly, YAPKO-2γ cells exhibit more proliferative and motility potential than YAPKO-2α cells. Moreover, we observed that YAP 2α-expressing NIH3T3 cells exhibit weak and slight colony formation, whereas YAP 2γ-expressing NIH3T3 cells remarkably form larger and greater number of colonies than YAP 2α-cells do in soft agar. The results indicate YAP 2γ potentiates the malignant transformation of normal NIH3T3 fibroblasts more powerful than YAP 2α dose under in vitro culture condition.

To extend this in vitro observation, we performed in vivo tumor formation assay in nude mice. The same number of HA-tagged YAP 2α or YAP 2γ ectopically expressing NIH3T3 cells were subcutaneously injected into both flanks of the nude mice. Consequently, both YAP 2α and YAP 2γ cells developed tumors obviously. However, the size of the tumors obtained from YAP 2α cells, were remarkably larger than that from the YAP 2γ cells.

The opposite observations therefore suggest that the different microenvironment for YAP 2α and YAP 2γ cells between in vitro culture condition and in vivo condition might be responsible for this differential outcome. We hypothesized that immune response from macrophages between YAP2α and YAP 2γ in the nude mice might be the reason for the opposite outcomes. To test this idea, F4/80, and CD11b was used as a pan-macrophage marker for IHC staining. Significantly increasing infiltration of macrophages presented in YAP 2γ tumors but not 2α tumors. Given that YAP 2γ cells form smaller tumors in nude mice as mentioned above, the infiltrated macrophages is considered to suppress the YAP 2γ tumor growth. To further confirm this idea, CD86 were used as M1 macrophage marker and CD163 were used as M2 macrophage marker for IHC staining. We observed a remarkably stronger signal by CD86 antibody, whereas slightly enhanced signal by CD163 antibody in YAP 2γ tumors than YAP 2α tumors. Hence, we conclude that overexpression of YAP 2γ in NIH3T3 cells results in the infiltration of macrophage (mainly M1), whereas overexpression of YAP 2α does not.

To gain the molecular insight for the observation and further explore the specific downstream target genes of YAP 2α and YAP 2γ respectively, we carried out the Mouse Cancer Targeted RNA Panel, which profiles the expression of 395 cancer-associated genes (including pre-defined housekeeping genes as internal control). As a result, Monocyte chemoattractant protein 1 (Mcp-1 also known as chemokine (C-C motif) ligand 2, Ccl2) exhibits higher expression with more than 2-fold in YAP 2γ cells than that in YAP 2α cells, furthermore the result shows a good consistency among the 3 sample pairs.
Moreover, we confirmed that the expression of YAP 2γ promoted mRNA expression of Ccl2. In contrast, the expression of YAP 2α suppressed mRNA expression of Ccl2.

Given that YAP 2α isoform which binds with and therefore promotes SHP2 nuclear translocalization, suppresses the mRNA expression of Ccl2, SHP2 might affect Ccl2 expression at transcriptional level via complex formation with YAP 2α-TEAD in Ccl2 gene promoter region. To test this idea, we constructed 3 luciferase reporters by fusing murine Ccl2 promoter regions (-1200bp/+85bp, -600bp/+85bp, -300bp/+85bp) with a firefly luciferase cDNA in pGL3 plasmid because there is a TEAD binding element locates 983bp~989bp upstream of the Ccl2 transcription start site. Overexpression of SHP2 significantly reduces the Ccl2-1200bp/+85bp reporter activity, while Ccl2-600bp/+85bp and Ccl2-300bp/+85bp reporter show no markedly difference. Similarly, YAP 2γ promotes Ccl2-1200bp/+85bp activity significantly more potentially than YAP 2α. No significant difference has been observed in Ccl2-600bp/+85bp and Ccl2-300bp/+85bp reporter between YAP 2α and YAP 2γ. SHP2-PA, which is a YAP/TAZ binding deficient mutant, failed to suppress the Ccl2-1200bp/+85bp reporter activity, suggesting the transcriptional inhibition of Ccl2 is due to the nuclear SHP2. Furthermore, we confirmed SHP2 localized in Ccl2-1200bp/+85bp promoter region by chromatin IP using anti-SHP2 antibody. Taken together, these observations suggest that YAP 2α promote the nuclear translocalization of SHP2, which then suppress Ccl2 transcriptional activity through in Ccl2- 1200bp/-600bp region.

Furthermore, we identified SRSF3 as the splicing regulator for YAP exon 6 alternative splicing. Loss of SRSF3 expression decreased the exon 6 inclusion rate in YAP. Additionally, activation of K-RAS signaling pathway suppressed SRSF3 expression, then resulted in the reduction of the exon 6 inclusion rate.

Collectively, our study revealed a mechanism underlying the regulation of SHP2 nuclear translocalization, which is mediated by YAP/TAZ heterodimerization in an isoform-specific manner, and we also characterized the different biological significance of YAP 2α and YAP 2γ isoforms through differentially regulating SHP2 subcellular distribution. Furthermore, we have investigated the critical switch for exon 6 splicing: Exon 6 inclusion is upregulated by SRSF3. In contrast, suppress the expression of SRSF3 by aberrantly activation of K-RAS results in exon 6 skipping.