Fig 1.
Ybx1 is highly expressed in the progenitor cells of earlier embryonic mouse cortex and conditional ablation of Ybx1 causes impaired cortical development.
(A) In situ hybridization using DIG-labeled RNA probes (antisense and sense) targeting Ybx1 was performed on coronal sections of mouse brains at E11.5 to P0. Representative cortical regions were shown as magnified images. Scale bars, 500 μm. (B) Immunostaining for Ybx1 protein was done on coronal sections of mouse brains at E11.5, E13.5, E15.5, E17.5, and P0. Representative cortical regions were shown. Scale bars, 100 μm. (C) Representative images of P0 neonatal mice from the same litter with different genotypes were shown at 2 and 12 h after birth, respectively. (D) Quantitative analysis of neonatal body weight at P0 was shown as box and whisker plots. Ybx1fl/fl, Nestin-Cre, Ybx1+/fl, and Nestin-Cre, Ybx1fl/fl were designated as control (Ctrl), heterozygous (Het), and cKO, respectively. Ctrl (n = 39 pups) vs. Het (n = 26 pups), p = 0.37; Ctrl vs. cKO (n = 28 pups), **p = 0.0014; Het vs. cKO, *p = 0.021; ns, not significant; all by one-way ANOVA followed by Tukey’s multiple comparison test. (E) Representative images of the P0 brains from cKO (#1), Het (#2), and Ctrl (#3) pups shown in (D). Black arrowed lines indicate the cortical length and width. (F) Quantitative analysis of cortical length and width at P0 was shown as box and whisker plots. n = 21 pups for Ctrl, n = 15 pups for Het, n = 16 pups for cKO: p = 0.65 for width, p = 0.80 for length; ns, not significant; all by one-way ANOVA followed by Tukey’s multiple comparison test. (G) Representative images of P0 coronal brain sections stained with an antibody against Ybx1 protein demonstrated efficient ablation of Ybx1 in the cKO brain. The asterisk indicates the enlarged ventricle in the cKO brain. Scale bar, 1,000 μm. (H) Quantification of lateral ventricle area of P0 coronal brain sections was presented as box and whisker plots: Ctrl (n = 21 confocal fields) vs. Het (n = 22 confocal fields), *p = 0.035; Ctrl vs. cKO (n = 21 confocal fields), ****p = 2.03E−11; Het vs. cKO, ****p = 2.03E−011; all by one-way ANOVA followed by Tukey’s multiple comparison test. (I) Quantification of P0 cortex thickness was presented as box and whisker plots: Ctrl (n = 20 confocal fields) vs. cKO (n = 21 confocal fields), ****p = 6.73-05; Het (n = 19 confocal fields) vs. cKO, **p = 0.0014; ns, not significant; all by one-way ANOVA followed by Tukey’s multiple comparison test. The data underlying all the graphs shown in the figure are included in S1 Data.
Fig 2.
Conditional ablation of Ybx1 leads to a reduction in cortical progenitor cell pool and a decrease in cortical thickness.
(A) Immunostaining of coronal brain sections at P0 using antibodies against radial glial progenitor marker Pax6 and intermediate progenitor marker Tbr2. Representative cortical regions were shown. White dotted lines mark the boundaries. Scale bar, 100 μm. (B-F) Quantification of Pax6+ (B) and Tbr2+ (C) cell numbers, and cortical layer thicknesses (D–F) shown in (A). All statistical data are presented as box and whisker plots. For Pax6+ cell numbers, Ctrl (n = 26 confocal fields) vs. cKO (n = 24 confocal fields), ****p = 1.42E−05; Het (n = 23 confocal fields) vs. cKO, ****p = 9.54E−05. For Tbr2+ cell numbers, Ctrl (n = 26 confocal fields) vs. cKO (n = 24 confocal fields), ****p = 1.19E−05; Het (n = 23 confocal fields) vs. cKO, ***p = 1.85E−04. For VZ thickness, Ctrl (n = 26 confocal fields) vs. cKO (n = 24 confocal fields), ****p = 9.26E−05; Het (n = 23 confocal fields) vs. cKO, ****p = 1.33E−05. For SVZ thickness, Ctrl (n = 26 confocal fields) vs. cKO (n = 24 confocal fields), ***p = 9.49E−04; Het (n = 23 confocal fields) vs. cKO, **p = 0.0077. For whole cortical thickness, Ctrl (n = 26 confocal fields) vs. cKO (n = 24 confocal fields), ****p = 8.47E−06; Het (n = 23 confocal fields) vs. cKO, ****p = 2.61E−05. (G) Immunostaining of coronal brain sections at P0 for layer V marker Ctip2 and layer VI marker Tbr1. Representative cortical regions were shown. White dotted lines mark the boundaries of each layer. Scale bar, 100 μm. (H–M) Quantification of Ctip2+ (H) and Tbr1+ (I) neuron numbers, and the thickness of WM layer (J), layer VI (K), layer V (L), and layers I–IV (M) shown in (G). All statistical data are presented as box and whisker plots. For Ctip2+ neuron numbers, Ctrl (n = 22 confocal fields) vs. cKO (n = 20 confocal fields), ****p = 1.89E−06; Het (n = 20 confocal fields) vs. cKO, ***p = 1.62E−04. For Tbr1+ neuron numbers, Ctrl (n = 19 confocal fields) vs. cKO (n = 21 confocal fields), ****P = 4.62E−08; Het (n = 21 confocal fields) vs. cKO, ****P = 5.50E−05. For WM layer thickness, Ctrl (n = 22 confocal fields) vs. cKO (n = 20 confocal fields), ***p = 3.50E−04; Het (n = 20 confocal fields) vs. cKO, **p = 0.0038. For layer VI thickness, Ctrl (n = 22 confocal fields) vs. cKO (n = 20 confocal fields), ****p = 1.43E−05; Het (n = 20 confocal fields) vs. cKO, ***p = 2.56E−04. For layer V thickness, Ctrl (n = 22 confocal fields) vs. cKO (n = 20 confocal fields), ***p = 5.16E−04; Het (n = 20 confocal fields) vs. cKO, ***p = 5.71E−04. For thickness of layers, I–IV, Ctrl (n = 22 confocal fields) vs. cKO (n = 20 confocal fields), **p = 0.0093; Het (n = 20 confocal fields) vs. cKO, *p = 0.032. (N) A stacked bar chart summarizing the data in (D–F, J–M) shows the distribution of each layer in P0 cortex. At least 3 mice were analyzed for each genotype. All analyses were performed by one-way ANOVA followed by Tukey’s multiple comparison test. ns, not significant. The data underlying all the graphs shown in the figure are included in S1 Data.
Fig 3.
Ybx1 is required for maintenance of the cortical progenitor pool from E13.5.
(A) E13.5 coronal brain sections were immunostained with antibodies against the radial glial cell marker Pax6 and the intermediate progenitor marker Tbr2. Representative cortical regions were shown. White dotted lines mark the boundaries. Scale bar: 50 μm. (B–F) Quantification of Pax6+ (B) and Tbr2+ (C) cell numbers, and cortical layer thicknesses (D to F) shown in (A). All statistical data are presented as box and whisker plots. For Pax6, Ctrl (n = 21 confocal fields) vs. cKO (n = 22 confocal fields), **p = 0.0018; Het (n = 18 confocal fields) vs. cKO, *p = 0.019. For Tbr2, Ctrl (n = 21 confocal fields) vs. cKO (n = 22 confocal fields), *p = 0.025; Het (n = 18 confocal fields) vs. cKO, *p = 0.014. For whole cortical thickness, Ctrl (n = 21 confocal fields) vs. cKO (n = 22 confocal fields), **p = 0.0018; Het (n = 18 confocal fields) vs. cKO, *p = 0.016. For VZ thickness, Ctrl (n = 21 confocal fields) vs. cKO (n = 22 confocal fields), **p = 0.0023; Het (n = 18 confocal fields) vs. cKO, *p = 0.026. For SVZ thickness, Ctrl (n = 21 confocal fields) vs. cKO (n = 22 confocal fields), **p = 0.0055; Het (n = 18 confocal fields) vs. cKO, *p = 0.047. (G) A stacked bar chart summarizing the data in (D–F) shows the distribution of each layer in E13.5 cortex. (H) E15.5 coronal brain sections were immunostained with antibodies against the radial glial cell marker Pax6 and the intermediate progenitor marker Tbr2. Representative cortical regions were shown. White dotted lines mark the boundaries. Scale bar: 100 μm. (I–O) Quantification of Pax6+ (I) and Tbr2+ (J) cell numbers, and cortical layer thicknesses (K–O) shown in (H). All statistical data are presented as box and whisker plots. For Pax6, Ctrl (n = 24 confocal fields) vs. cKO (n = 22 confocal fields), ***p = 2.93E−04; Het (n = 21 confocal fields) vs. cKO, **p = 0.0071. For Tbr2, Ctrl (n = 24 confocal fields) vs. cKO (n = 22 confocal fields), ***p = 8.58E−04; Het (n = 21 confocal fields) vs. cKO, **p = 0.0042. For VZ thickness, Ctrl (n = 24 confocal fields) vs. cKO (n = 22 confocal fields), ****p = 5.60E−05; Het (n = 21 confocal fields) vs. cKO, ***p = 2.72E−04. For SVZ thickness, Ctrl (n = 24 confocal fields) vs. cKO (n = 22 confocal fields), ***p = 2.75E−04; Het (n = 21 confocal fields) vs. cKO, ***p = 8.94E−04. For CP thickness, Ctrl (n = 24 confocal fields) vs. cKO (n = 22 confocal fields), **p = 0.0062; Het (n = 21 confocal fields) vs. cKO, *p = 0.034. For IZ thickness, Ctrl (n = 24 confocal fields) vs. cKO (n = 22 confocal fields), ***p = 1.99E−04; Het (n = 24 confocal fields) vs. cKO, ***p = 5.31E−04. For whole cortical thickness, Ctrl (n = 24 confocal fields) vs. cKO (n = 22 confocal fields), ****p = 1.73E−05; Het (n = 21 confocal fields) vs. cKO, ***p = 1.68E−04. (P) A stacked bar chart summarizing the data in (K–O) shows the distribution of each layer in E15.5 cortex. At least 3 embryos were analyzed for each genotype. All analyses were performed by one-way ANOVA followed by Tukey’s multiple comparison test. ns, not significant. The data underlying all the graphs shown in the figure are included in S1 Data.
Fig 4.
Ybx1 regulates proliferation and differentiation of cortical progenitor cells.
(A) Coronal brain sections at P0 were stained with antibodies recognizing BrdU and Ki67. Pregnant mothers received a BrdU pulse 24 h before pup dissection at P0. Representative cortical regions were shown. Scale bar: 100 μm. (B and C) Quantification of BrdU+ cell numbers (B) and the percentage of cells exiting the cell cycle (C) at P0 shown in (A). All statistical data are presented as box and whisker plots. For BrdU+ cell numbers, Ctrl (n = 27 confocal fields) vs. cKO (n = 24 confocal fields), ***p = 2.11E−04; Het (n = 23 confocal fields) vs. cKO, ***p = 4.24E−04. For Ki67−BrdU+/BrdU+, Ctrl (n = 27 confocal fields) vs. cKO (n = 24 confocal fields), **p = 0.0015; Het (n = 23 confocal fields) vs. cKO, **p = 0.0051. (D) Coronal brain sections at E13.5 were stained with antibodies recognizing BrdU and Ki67. Pregnant mothers received a BrdU pulse 24 h before embryo collection. Representative cortical regions were shown. Scale bar, 50 μm. (E and F) Quantification of BrdU+ cell numbers (E) and the percentage of cells exiting the cell cycle (F) at E13.5 shown in (D). All statistical data are presented as box and whisker plots. For BrdU+ cell numbers, Ctrl (n = 22 confocal fields) vs. cKO (n = 25 confocal fields), **p = 0.0052; Het (n = 23 confocal fields) vs. cKO, *p = 0.012. For Ki67−BrdU+/BrdU+, Ctrl (n = 22 confocal fields) vs. cKO (n = 25 confocal fields), **p = 0.0028; Het (n = 23 confocal fields) vs. cKO, *p = 0.027. (G) Representative images of primary and secondary neurospheres formed by NSCs isolated at E13.5. Scale bar, 50 μm. (H–K) Quantification of the sizes (H, J) and numbers (I, K) of the primary (H, I) and secondary (J, K) neurospheres show in (G). All statistical data are presented as box and whisker plots. In H, Ctrl (n = 42 neurospheres) vs. cKO (n = 38 neurospheres), ****p = 6.65E−05; Het (n = 46 neurospheres) vs. cKO, ****p = 5.22E−05. In J, Ctrl (n = 42 neurospheres) vs. cKO (n = 38 neurospheres), ****p = 2.76E−05; Het (n = 46 neurospheres) vs. cKO, ****p = 2.26E−05. (L) Schematic drawings of the IUE experiment. Pregnant mice were injected at E14.5 with plasmids. Subsequently, a single EdU pulse was administrated at E15.5, and embryos were dissected at E16.5 for analysis. (M) Confirmation of Ybx1 knockdown in the cortex after IUE of shYbx1. Immunostaining for GFP and Ybx1 on coronal sections of E16.5 cortex. Representative cortical regions were shown. Scale bar, 50 μm. (N) Immunostaining for GFP, EdU, and Ki67 on coronal sections of E16.5 mouse cortex after knockdown of Ybx1 using IUE. Representative cortical regions were shown. Scale bar, 50 μm. (O and P) Quantification of percentage of EdU+GFP+/GFP+ (O) and Ki67-EdU+/EdU+GFP+ (P) shown in (N). Data are presented as box and whisker plots: in O, shCtrl (n = 22 confocal fields) vs. shYbx1 (n = 19 confocal fields), ***p = 1.38E−04; in P, shCtrl (n = 22 confocal fields) vs. shYbx1 (n = 19 confocal fields), ***p = 8.30E−04. At least 3 pups or embryos were analyzed for each genotype or condition. Analyses were performed by one-way ANOVA followed by Tukey’s multiple comparison test (B, C; E, F; H–K), or by unpaired Student t test (O, P). ns, not significant. The data underlying all the graphs shown in the figure are included in S1 Data.
Fig 5.
Ybx1 controls the stability of its m5C-modified target transcripts in neural progenitor cells.
(A-C) Gene Ontology (GO) analysis of transcripts with altered expression levels in E13.5 Ybx1 cKO cortex (A), E18.5 Ybx1 cKO cortex (B), and E13.5 Ybx1 KD cortical progenitors (C). GO terms in biological processes were shown. (D) GO analysis of Ybx1 target transcripts identified by anti-Ybx1 RIP-seq in the E14.5 mouse cortex. (E) Venn diagram showing the overlap of target mRNAs identified by three cKO or KD RNA-seq and anti-Ybx1 RIP-seq. GO terms in biological processes were shown. (F) Validation of m5C modification on Ybx1 target mRNAs by anti-m5C immunoprecipitation combined with RT-qPCR. In vitro transcribed GFP mRNA which has no m5C modification was used as a negative control. Data represent mean ± SD (n = 3 replicates): ***p = 5.99E−04 for Ccnd2; ***p = 1.37E−04 for eEF1g; **p = 0.0045 for Rpsa; ***p = 8.89E−04 for Rps5; ***p = 8.06E−04 for Uhmk1; not significant (ns) for GFP; by unpaired Student t test. (G) Ybx1 target transcripts exhibit accelerated degradation in the Ybx1 cKO cortex. Radial glial cells (RGCs) dissected from E14.5 Ybx1 cKO and control embryos were cultured, treated with actinomycin D (ActD), and collected at different time points. Ybx1 target mRNA levels were measured by RT-qPCR while β-actin mRNA was used a non-target control. Data represent mean ± SEM (n = 3 replicates): for Ccnd2, **p = 0.0017 (2.5 h), ***p = 7.50E−04 (5 h), ***p = 2.32E−04 (7.5 h), **p = 0.0018 (10 h); for eEF1g, **p = 0.0014 (2.5 h), ***p = 3.98E−04 (5 h), ***p = 3.10E−04 (7.5 h), **p = 0.0022 (10 h); for Rpsa, **p = 0.0096 (2.5 h), **p = 0.0012 (5 h), ***p = 4.09E−04 (7.5 h), ***p = 5.08E−04 (10 h); for Rps5, ***p = 5.67E−04 (2.5 h), ***p = 1.33E−04 (5 h), ***p = 1.65E−04 (7.5 h), ***p = 6.41E−04 (10 h); for Uhmk1, **p = 0.0013 (2.5 h), **p = 0.0025 (5 h), **p = 0.0011 (7.5 h), ***p = 7.91E−04 (10 h); for β-actin, ns, not significant; by unpaired Student t test. The half-lives for each mRNA in Ctrl and cKO were calculated (? indicates that the half-lives cannot be calculated). The data underlying all the graphs shown in the figure are included in S1 Data.
Fig 6.
Knockdown of Ybx1 target mRNAs impairs the proliferation and differentiation of cortical progenitor cells.
(A) Representative images of neurospheres formed after siRNA-mediated knockdown of Ccnd2, eEF1g, Rpsa, Rps5, or Uhmk1 in cultured E13.5 neural stem cells (NSCs). Scale bar, 50 μm. (B and C) Quantification of the sizes (B) and numbers (C) of the neurospheres after siRNA-mediated knockdown of Ybx1 target mRNAs show in (A). At least 3 replicates were performed. All statistical data are presented as box and whisker plots: in B, siCcnd2 (n = 55 neurospheres) vs. siCtrl (n = 51 neurospheres), ****p = 5.87E−06; sieEF1g (n = 52 neurospheres) vs. siCtrl, ****p = 5.46E−07; siRpsa (n = 57 neurospheres) vs. siCtrl, ****p = 1.57E−07; siRps5 (n = 49 neurospheres) vs. siCtrl, ****p = 6.12E−07; siUhmk1 (n = 45 neurospheres) vs. siCtrl, ****p = 9.87E−08; by one-way ANOVA followed by Tukey’s multiple comparison test; ns, not significant. (D) Immunostaining for GFP, EdU, and Ki67 on coronal sections of E16.5 mouse cortex after knockdown of Ybx1 targets in the cortex using IUE of a cocktail shRNA against all targets. Scale bar, 50 μm. (E and F) Quantification of percentage of EdU+GFP+/GFP+ (E) and Ki67-EdU+/EdU+GFP+ (F) shown in (D). At least 3 embryos were analyzed for each condition. Data are presented as box and whisker plots: in E, shCtrl (n = 23 confocal fields) vs. shCocktail (n = 27 confocal fields), ***p = 2.10E−04; in F, shCtrl (n = 23 confocal fields) vs. shCocktail (n = 27 confocal fields), ***p = 6.86E−04; by unpaired Student t test (E, F). The data underlying all the graphs shown in the figure are included in S1 Data.
Fig 7.
Overexpression of Ybx1 targets rescues the cortical development defects caused by Ybx1 ablation.
(A) Overexpression of Ybx1 targets (OE-Cocktail) in Ybx1 cKO mouse cortex using IUE, followed by immunostaining for GFP, Pax6, and Tbr2 on coronal sections at E16.5. Representative cortical regions were shown. Scale bar, 50 μm. (B and C) Quantification of percentage of Pax6+GFP+/GFP+ (B) and Tbr2+GFP+/GFP+ (C) shown in (A). Data are presented as box and whisker plots. In B, “Ctrl + OE-Ctrl” (n = 20 confocal fields) vs. “cKO + OE-Ctrl” (n = 18 confocal fields), ****p = 4.26E−05; “cKO + OE-Ctrl” vs. “cKO + OE-Cocktail” (n = 21 confocal fields), **p = 0.0040; “Ctrl + OE-Ctrl” vs. “cKO + OE-Cocktail”, p = 0.10. In C, “Ctrl + OE-Ctrl” (n = 20 confocal fields) vs. “cKO + OE-Ctrl” (n = 18 confocal fields), ***p = 5.69E-04; “cKO + OE-Ctrl” vs. “cKO + OE-Cocktail” (n = 21 confocal fields), *p = 0.043; “Ctrl + OE-Ctrl” vs. “cKO + OE-Cocktail”, *p = 0.015. (D) Overexpression of Ybx1 targets (OE-Cocktail) in control and Ybx1 cKO mouse cortex using IUE, followed by immunostaining for GFP, EdU, and Ki67 on coronal sections at E16.5. Representative cortical regions were shown. Scale bar, 50 μm. (E and F) Quantification of percentage of EdU+GFP+/GFP+ (E) and Ki67-EdU+/EdU+GFP+ (F) shown in (D). Data are presented as box and whisker plots. In E, “Ctrl + OE-Ctrl” (n = 27 confocal fields) vs. “Ctrl + OE-Cocktail” (n = 25 confocal fields), **p = 0.0099; “Ctrl + OE-Ctrl” vs. “cKO + OE-Ctrl” (n = 24 confocal fields), ****p = 3.08E−10; “Ctrl + OE-Ctrl” vs. “cKO + OE-Cocktail” (n = 26 confocal fields), p = 0.19. “Ctrl + OE-Cocktail” vs. “cKO + OE-Cocktail”, ****p = 8.25E−6. “cKO + OE-Ctrl” vs. “cKO + OE-Cocktail”, ****p = 3.08E−10. In F, “Ctrl + OE-Ctrl” (n = 26 confocal fields) vs. “Ctrl + OE-Cocktail” (n = 28 confocal fields), *p = 0.015; “Ctrl + OE-Ctrl” (n = 26 confocal fields) vs. “cKO + OE-Ctrl” (n = 24 confocal fields), ****p = 3.08E−10; “Ctrl + OE-Ctrl” vs. “cKO + OE-Cocktail” (n = 24 confocal fields), p = 0.51. “Ctrl + OE-Cocktail” vs. “cKO + OE-Cocktail”, ***p = 0.00015. “cKO + OE-Ctrl” vs. “cKO +OE-Cocktail”, ****p = 3.08E−10. At least 3 embryos were analyzed for each genotype or condition. Analyses were performed by one-way ANOVA followed by Tukey’s multiple comparison test. ns, not significant. The data underlying all the graphs shown in the figure are included in S1 Data.
Fig 8.
Ybx1 regulates the G1–S phase transition of cortical progenitor cell cycle.
(A) Schematic drawings showing Ybx1 target mRNAs-encoding proteins are involved in regulating normal cell cycle progression. (B and C) Cell cycle flow cytometry analysis determining the percentage of cells in G0/G1, S, and G2/M phases in cortical progenitor cells from E13.5 (B) and E15.5 (C) Ybx1 cKO and littermate control embryos. Data are presented as mean ± SEM: n = 3 replicates; in B, *p = 0.031 for G0/G1 phase, *p = 0.044 for S phase, **p = 0.0012 for G2/M phase; in C, ***p = 4.33E−04 for G0/G1 phase, ***p = 9.75E−04 for S phase, ***p = 3.40E−04 for G2/M phase; by unpaired Student t test. The data underlying all the graphs shown in the figure are included in S1 Data.