br Fig circFAT e sequestered miR g
Fig. 4. circFAT1(e2) sequestered miR-548g to stabilize RUNX1. (A) Relative distribution of circFAT1(e2) in MGC-803 cells. (B) The putative sequence of miR-548g and circFAT1(e2) with five binding sites. (C and D) Interaction between miR-548g and circFAT1(e2) was determined by luciferase reporter assay in MGC803 and MKN-28 cells. (E) RIP assay was carried out to determine the interaction between circFAT1(e2) and miR-548g using a specific miR-548g probe or scramble in MGC-803 cells, P < 0.05. Relative expression of miR-548g was assessed in (F) circFAT1(e2) overexpressed and (G) circFAT1(e2) knockdown MGC-803 Fluxametamide by RT-PCR; miR-1 expression was used as negative control, P < 0.05. (H) RIP assay was used to further determine the interaction between miR-548g and circFAT1(e2) in the cytoplasm and nucleus of MGC-803 cells via a specific miR-548g probe, P < 0.05. (I) Correlation between circFAT1(e2) and miR-548g in six GC cell lines and one normal human gastric epithelial mucosa cell line.
Results from above study revealed that circFAT1(e2) are distributed in the cytoplasm and nucleus of GC cells, and cytoplasmic circFAT1(e2) was demonstrated to play a role as a GC inhibitor by interacting with miR-548g. However, whether nuclear circFAT1(e2) participates in the tumorigenesis of GC remains undetermined. Using the online catRAPID algorithm, we found multiple proteins that may be bound by circFAT1(e2) (Table S3). First, among these proteins, YBX1 got the highest score. Second, YBX1 plays important roles in GC with multi-ncRNAs, so it was chosen for further exploration. CircFAT1(e2) was artificially divided into four parts (a, 1-1000; b, 1001-2000; c, 2001-3000; d, 3001-3283). These four parts were subjected to catRAPID to analyze the interaction with YBX1 protein; part a (1-1000) and part d (3001-3283) showed higher scores (Fig. 6A). We then conducted RNA immunoprecipitation (RIP) in MGC-803 cells to pull down circFAT1(e2) using an anti-YBX1 antibody, followed by RT-PCR analysis and Western
blot assay for circFAT1(e2) expression. The results showed that cir-cFAT1(e2) expression was significantly higher in the anti-YBX1 group than the IgG group (P < 0.05, Fig. 6B). Moreover, the four parts of circFAT1(e2) were transcribed in vitro and then served as anchors to pull down YBX1 in cell nucleus lysate, respectively. Western blot assay showed a specific enrichment of YBX1 in part a and part d groups (Fig. 6C). These results showed that circFAT1(e2) specifically interacted with YBX1 in GC cells.
3.7. circFAT1(e2) inhibited GC cell tumorigenesis through YBX1
RT-PCR analysis demonstrated that overexpression of circFAT1(e2) significantly downregulated the expression of three targeted genes of YBX1 (EGFP, c-Met, and CDC25A). However, YBX1 could abolish this eﬀect induced by circFAT1(e2) in MGC-803 and MKN-28 cells (P < 0.05, Fig. 7A and B). Cell growth curve analysis showed that overexpression of circFAT1(e2) in MGC-803 and MKN-28 cells could
Fig. 6. circFAT1(e2) directly bind to YBX1. (A) The catRAPID algorithm was applied to predict the RNA-protein interaction of circFAT1(e2) and YBX1. (B and C) The interaction between circFAT1(e2) and YBX1 was validated by RNA immunoprecipitation (RIP) in MGC-803 cells.
significantly reduce the cell number, however, YBX1 could abolish the reduction of cells induced by circFAT1(e2) overexpression (P < 0.05, Fig. 7C and D). Colony formation assays indicated that overexpression of circFAT1(e2) in MGC-803 and MKN-28 cells could significantly re-duce the colony number, however, YBX1 could abolish the reduction of colonies induced by circFAT1(e2) overexpression (Fig. 7E and F). These findings demonstrated that circFAT1(e2) could inhibit GC cell pro-liferation through YBX1.