Fig. 3

MILIP interacts with YBX1 and promotes translational activation of Snai1 a RNA pulldown followed by mass spectrometry analysis identified that YBX1 was the most abundant protein co-pulled down with MILIP antisense probes. S: scramble; AS: antisense. n = 1. b YBX1, but not p53, was co-pulled down with MILIP as shown in RNA pull-down (RPD) assays. β-Actin was included as a negative control. c MILIP was co-precipitated with YBX1 as detected by RNA immunoprecipitation (RIP) assays. LncRNA PLANE was used as a negative control. d In vitro-transcribed full-length (FL) MILIP and MILIP deletion mutants MILIP-ΔE1, MILIP-Δ1488-1895, MILIP-Δ990-1895 but not MILIP-ΔE2 were co-precipitated with GFP-YBX1 as shown in RIP assays. MILIP-ΔE1: a MILIP mutant with its exon1 deleted; MILIP-ΔE2: a MILIP mutant with its exon2 deleted. e MILIP was co-precipitated with full-length (FL) YBX1, YBX1 Δ Ala/Pro-rich N-terminal domain (AP) and YBX1 Δ C-terminal domain (CTD) but not YBX1 Δ cold shock domain (CSD) as shown in RIP assays. f SiRNA knockdown of MILIP reduced Snai1 protein levels but not its mRNA levels. One-way ANOVA followed by Tukey’s multiple comparisons test. g, h SiRNA knockdown of MILIP did not impinge on the half-life of Snai1 protein (g). Relative abundance of Snai1 protein levels normalized to respective β-Actin levels were quantitated using the ImageJ software (h). CHX, cycloheximide: 20 μg/ml. One-way ANOVA followed by Tukey’s multiple comparisons test. i, j The increased cell migration and invasion caused by MILIP overexpression was diminished by co-knockdown of Snai1 (i). Relative area of migrated and invasive cells was quantitated using the ImageJ software (j). One-way ANOVA followed by Tukey’s multiple comparisons test. Data are representatives or mean ± s.d.; n = 3 independent experiments