Fig 2: LIN28A-associated proteins involved in nuclear processes: Chromatin organization and splicing(A) String network of BAF/PBAF complex subunits (Hodges et al., 2016) merged with our LIN28A IP-MS data.(B) Immunoblotting of SMARCB1 and SMARCA4 co-immunoprecipitated with LIN28A from NPC lysates in the presence of RNase A. A different antibody (ab63740) was used for IP from the one used for IP-MS (sc-374460) for cross-validation, and normal rabbit IgG was used for the control IP. LIN28A enrichment was so high that LIN28A from the input lysate could not be detected in the same blot with the immunoprecipitated LIN28A.(C) U1 snRNP (GO:0005685), U2 snRNP (GO:0005686), and U4/U6.U5 tri-snRNP (GO:0046540) complex subunits and their interactions, merged with our LIN28A IP-MS data. See also Figures S4 and S5.

Fig 3: LIN28A impacts subcellular distribution of EIF3D(A) Venn diagram showing the overlaps between the LIN28A-associated proteins annotated in translational initiation (GO:0006413), SRP-dependent cotranslational targeting to membrane (GO:0006614), and ribosome (GO:0005840).(B) Relative abundance of eIF3 subunit proteins co-immunoprecipitated with LIN28A in the presence of RNase A (red bar) and those in the control IP (gray). EIF3G, H, I, K data are omitted as they were detected only once in all the samples. Data are represented as mean ± SEM. **FDR (false discovery rate) < 0.01, ****FDR < 0.0001 by SAINT analysis.(C and E) Representative immunofluorescence images of HEK293 cells transfected with LIN28A-expressing plasmids. Scale bar, 10 µm. Arrows indicate cells with intense nuclear LIN28A signals. Arrowheads indicate cells with higher LIN28A signals outside the nucleus.(D) Correlation of nuclear/cytosolic mean intensity ratio of LIN28A and EIF3D. Pearson r = 0.7234, p < 0.0001.(F) Correlation of nuclear/cytosolic mean intensity ratio of LIN28A and SF3B3. Pearson r = -0.5010, p = 0.048. See also Figure S6.

Fig 4: Immunoprecipitation-mass spectrometry (IP-MS) analysis to identify LIN28A interactome(A) Overall workflow of IP-MS used in the present study to identify significant interactors of endogenous LIN28A in human iPS cell-derived NPCs.(B) Western blot assessment of specificity of the LIN28A antibody used in IP-MS. The lysate from LIN28A over-expressing (OE) NPC was loaded next to the same amount of control lysate. The antibody specifically detected LIN28A protein.(C) Volcano plot of the proteins identified in LIN28A IP-MS to compare spectral counts (SpC) in the presence versus absence of RNase A. The y axis represents -log10(t test p value).(D) Venn diagram showing the overlap of significant LIN28A interactors identified in this study and those identified in previous studies. Pie chart below represents the number of interactions identified in certain number (1–5) of previous studies. Most of the interactions were identified in only one or two of 23 curated studies (Table S1).(E) Frequency distribution of percentage CRAPome studies that detected the LIN28A interactors in negative controls. LIN28A interactors identified in this study and those in previous studies are represented by interleaved bars in each bin. Recently published large-scale interactome study (Huttlin et al.), in which 273 LIN28A interactors were identified, is separately represented. See also Figure S1.

Fig 5: LIN28A impairs stress granule (SG) formation upon stress(A) Intersection of SG proteins from previous studies and RNase A-resistant LIN28A interactors in our study.(B) Immunofluorescence of G3BP1 in infrared fluorescence protein (iRFP) or LIN28A-overexpressing cells with sodium arsenite treatment (0.3 mM, 30 min). mEmerald was expressed as a co-transfection marker. Scale bar, 10 µm.(C) Quantification of percentage SG-positive cells among mEmerald-positive cells (n = 3 plates for each group). At least 35 transfected cells in four images per plate were assessed. Data are represented as mean ± SEM. ***p < 0.001, unpaired t test.