Statistical analysis
All data are expressed as mean ± SEM. Statistical analyses were performed using IBM SPSS Statistics 20. Normal distribution was assessed by Shapiro-Wilk test. Statistical significance was determined according to sample distribution and homogeneity of variance. Statistical comparisons between two groups were determined by two-tailed Student’s t test and Mann-Whitney U test. One-way ANOVA following by Bonferroni’s post hoc test, two-way ANOVA with simple effect test and Kruskal-Wallis test with Dunn’s post hoc test were performed for comparisons among multiple groups. P < 0.05 was considered statistically significant. The group sizes of the individual experiments were chosen based on our preliminary experiments. It was shown that the significance was reached when the number in each group was 3 or 4. In our study, n =5-8 fulfilled the statistical requirements. The exact statistical method and exact group size for the individual experiments are shown in the corresponding figure legends. No samples were excluded from the analyses.