MYCL promotes iPSC-like colony formation via MYC Box 0 and 2 domains

Human induced pluripotent stem cells (hiPSCs) can differentiate into cells of the three germ layers and are promising cell sources for regenerative medicine therapies. However, current protocols generate hiPSCs with low efficiency, and the generated iPSCs have variable differentiation capacity among different clones. Our previous study reported that MYC proteins (c-MYC and MYCL) are essential for reprogramming and germline transmission but that MYCL can generate hiPSC colonies more efficiently than c-MYC. The molecular underpinnings for the different reprogramming efficiencies between c-MYC and MYCL, however, are unknown. In this study, we found that MYC Box 0 (MB0) and MB2, two functional domains conserved in the MYC protein family, contribute to the phenotypic differences and promote hiPSC generation in MYCL-induced reprogramming. Proteome analyses suggested that in MYCL-induced reprogramming, cell adhesion-related cytoskeletal proteins are regulated by the MB0 domain, while the MB2 domain regulates RNA processes. These findings provide a molecular explanation for why MYCL has higher reprogramming efficiency than c-MYC.

. HDFs were transduced with SeV carrying KLF4-OCT3/4-SOX2 (KOS), KLF4 (K), and c-MYC or MYCL on day 0. We used an MOI (multiplicity of infection) of 4.3 for each virus. StemFit AK03N without bFGF was used during the transduction and subsequent induction of iPSC-like colonies. We performed immunostaining of the reprogramming HDFs 1 to 7 days after the transduction and analyzed the results using ArrayScan. (B) Representative immunostaining images of reprogramming HDFs stained by anti-TRA-1-60 antibody (green) and Hoechst (blue) 7 days after the transduction. Scale bar, 300 μm. Ph, phase contrast. (C) Proliferation and expression of TRA-1-60 ( +) cells during reprogramming. HDFs were transduced with SeV, including c-MYC or MYCL, and immunostaining was performed from days 1 to 7. The number of total cells was counted as Hoechst-positive cells. Mean ± SD values are shown. n = 3, *p < 0.05 by paired t-test. (D) Schematic representation of HDF reprogramming with episomal plasmid vector (EpiP). HDFs were transduced with EpiP carrying SOX2, KLF4, OCT3/4-shp53, LIN28A, EBNA1, and c-MYC or MYCL. StemFit AK03N without bFGF was used during the transfection and subsequent induction of iPSC-like colonies. We performed flow cytometry of the reprogramming HDFs every three days from 1 to 19 days plus day 21 after the transduction. (E) Representative immunostaining images of reprogramming HDFs stained by anti-TRA-1-60 antibody (green) and Hoechst (blue) 21 days after the transduction. Scale bar, 300 μm. Ph, phase contrast. (F) Proliferation and expression of TRA-1-60 ( +) cells during reprogramming were analyzed by flow cytometry. HDFs were transduced with EpiP, including c-MYC or MYCL. Flow cytometry was performed every three days from days 1 to 19 days plus day 21. Mean ± SD for n = 3, *p < 0.05 and **p < 0.01 by paired t-test. (G) The number of iPSC-like and non-iPSC-like colonies derived from 1 × 10 5 HDFs transduced with EpiP including c-MYC or MYCL on day 21. Mean ± SD values are shown. n = 3, **p < 0.01 by unpaired t-test. (H) Percentage of CD13 ( +) cells during EpiP reprogramming determined by flow cytometry. Mean ± SD values are shown. n = 3, *p < 0.05 and **p < 0.01 by paired t-test.  www.nature.com/scientificreports/ the transfected factors, cell toxicity, and the time required for the iPSC-like colonies to appear: the SeV system requires about 7 days, but the EpiP system needs about 21 days based on our observations. We found two types of colonies: "iPSC-like" and "non-iPSC-like" colonies. The iPSC-like colonies produced by MYCL were more flattened and showed a monolayered colony morphology, with each cell tightly packed and expressing TRA-1-60. The non-iPSC-like colonies produced by c-MYC showed a cell aggregation-like morphology, in which individual cells were irregularly aggregated and did not express TRA-1-60. We counted the number of iPSC-like and non-iPSC-like colonies on day 21 and found that c-MYC induced iPSC-like colonies as well as many non-iPSC-like colonies, but MYCL induced almost only iPSC-like colonies and more of them than c-MYC ( Fig. 1G and Supplementary Fig. S3).
It has been reported that before the increase in the expression of TRA-1-60, a decrease in the expression of CD13, a marker of fibroblasts 18 , is observed in somatic cell reprogramming. Therefore, we confirmed the expression of CD13 during reprogramming. The percentage of CD13 ( +) cells decreased daily in HDFs transduced with c-MYC or MYCL, but the number of CD13 (-) cells rapidly increased in c-MYC compared to MYCL ( Fig. 1H and Supplementary Fig. S4). In particular, the CD13 (-) TRA-1-60 (-) population was larger on day 10 with c-MYC reprogramming than MYCL reprogramming, but the CD13 (-) TRA-1-60 ( +) population from days 16 to 21 was larger with MYCL reprogramming (Supplementary Fig. S4). These results suggested that MYCL promotes TRA-1-60 ( +) cells more than c-MYC, but c-MYC suppresses CD13 expression more than MYCL.
MYC Box 0 and 2 domains are crucial for colony formation during reprogramming. Next, we prepared domain deletion mutants to identify which domains in the N-terminus of MYC proteins influence reprogramming ( Fig. 2A and Supplementary Fig. S5). We previously showed that a c-MYC mutant lacking transformation activity enhances the formation of iPSC-like colonies. This mutant has a point mutation in the transactivation domain of the N-terminal region, W135E ( Fig. 2A and Supplementary Fig. S5B), but can bind to genomic DNA 7 . On the other hand, the bHLHLZ domain in the C-terminus region is a well-known binding domain of MAX 19 . Mutants in the C-terminus region prevent MYC proteins from binding to DNA and thus reprogramming 7 . Finally, we tested the reprogramming activities of these mutants using the EpiP reprogramming system because, as explained above, this method provided a clearer phenotype and was easier to manipulate than the SeV method.
The EpiP mutants were transfected into HDFs with other reprogramming factors, and the number of iPSClike and non-iPSC-like colonies was counted ( Fig. 2B and Supplementary Fig. S6). c-MYC-ΔMB0 promoted the formation of iPSC-like colonies and inhibited the formation of non-iPSC-like colonies compared to c-MYC-WT. In contrast, MYCL-ΔMB0 showed almost no ability to form iPSC-like colonies (Fig. 2B). We confirmed that the protein expression of each domain deletion mutant by western blotting showed no difference compared with c-MYC-or MYCL-WT (Supplementary Fig. S7 and S8). These results demonstrate that the MB0 domain has different functions in c-MYC and MYCL for reprogramming and that c-MYC-ΔMB0 has a similar function as MYCL-WT. Figure 2B shows that c-MYC-ΔMB1 promoted iPSC-like colony formation like c-MYC-ΔMB0, but it also led to the formation of non-iPSC-like colonies. The formation of iPSC-like colonies by MYCL-ΔMB1 was about a quarter that by MYCL-WT. Unlike c-MYC-WT, c-MYC-ΔMB2 did not induce non-iPSC-like colonies, but it did induce a rate of iPSC-like colonies similar to c-MYC-WT. MYCL-ΔMB2 showed little ability to form iPSC-like colonies, resembling MYCL-ΔMB0. c-MYC-ΔMB3a, -ΔMB3b, and -ΔMB4 had similar colony-forming activities as c-MYC-WT. MYCL-ΔMB3b showed the same reprogramming efficiency as MYCL-WT, but MYCL-ΔMB4 formed about the same small number of iPSC-like colonies as MYCL-ΔMB1. The ΔbHLHLZ mutants of both c-MYC and MYCL failed to induce colonies and were therefore considered to have lost MYC function completely. Thus, the results indicate that in c-MYC, the MB0 and MB2 domains are repressive for iPSC-like colony formation, but in MYCL, they are promotive. Other domains also influenced the colony formation efficiency, but the effect was small.
Next, we analyzed the effect of the MYC-deletion mutants on the expression of TRA-1-60 and CD13 by flow cytometry 16 days after the start of reprogramming (Fig. 2C). Mutants that increased the number of iPSC-like colonies also increased the expression of TRA-1-60, while those that reduced the number of iPSC-like colonies lowered the TRA-1-60 expression ( Fig. 2C and Supplementary Fig. S9). c-MYC-WT showed little TRA-1-60  boxes show important domains for MYC function,  including MB0, 1, 2, 3a (c-MYC only), 3b, 4, and basic-helix-loop-helix leucine zipper motif (bHLHLZ). The percentage of common amino acids in each MYC box domain between MYCL and c-MYC is shown (Identity%). The numbers on the right indicate amino acid lengths. (B) Number of iPSC-like and non-iPSClike colonies transduced with EpiP including c-MYC-WT/mutants (left) or MYCL-WT/mutants (right) on day 21. Mean ± SD values are shown. n = 3, **p < 0.01, ***p < 0.001 and ****p < 0.0001 by ordinary one-way ANOVA and Dunnett's test vs. WT. (C) Expression of TRA-1-60 ( +) HDFs and CD13 ( +) HDFs transduced with EpiP including c-MYC-WT/mutants (left) or MYCL-WT/mutants (right) on day 16. Mean ± SD values are shown. n = 3, *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 by ordinary one-way ANOVA and Dunnett's test vs. WT. (D) Representative flow cytometry images for TRA-1-60 and CD13 for HDFs transduced with EpiP including c-MYC-WT/ΔMB0 or MYCL-WT/ΔMB0 10 and 21 days after the transduction. Numbers indicate the expression percentage of each quadrant. (E) Proliferation of HDFs transduced with EpiP including c-MYC-WT/ΔMB0 or MYCL-WT/ΔMB0 10 and 21 days later. Mean ± SD values are shown. n = 3, ***p < 0.001 by ordinary one-way ANOVA and Dunnett's test vs. c-MYC-WT. The number of cells was counted using a Cell Counter model R1 (OLYMPUS). www.nature.com/scientificreports/ expression, whereas c-MYC-ΔMB0 upregulated the expression. MYCL-ΔMB0, unlike MYCL-WT, failed to upregulate the expression of TRA-1-60. The CD13 expression was also correlated with colony formation. In c-MYC, a significant decrease in CD13 expression was observed for mutants that promoted non-iPSC-like colony formation. As for MYCL, only a slight decrease in CD13 expression was observed for mutants that promoted iPSC-like colony formation. From these results, we concluded that the MB0 domain is essential for the function of MYC in reprogramming but functions differently between c-MYC and MYCL.
To analyze the function of the MB0 domain in more detail, we analyzed the expression of TRA-1-60 and CD13 10 and 21 days after the start of reprogramming by flow cytometry (Fig. 2D). In the case of c-MYC-WT, there was a strong decrease in CD13 expression on day 10, and most cells were CD13 negative on day 21. In the cases of c-MYC-ΔMB0 and MYCL-WT, there was a slight decrease in CD13 expression on day 10, and more than half of cells were expressing TRA-1-60 on day 21. Finally, in the case of MYCL-ΔMB0, there was no change in CD13 or TRA-1-60 expression. More study is needed to determine how CD13 is regulated by c-MYC and MYCL.
Additionally, c-MYC-WT showed higher cell proliferation on day 10, but c-MYC-ΔMB0 resulted in a lower cell proliferation comparable more with MYCL-WT than with c-MYC-WT on day 10 (Fig. 2E). We attributed this effect to the lost transformation activity of c-MYC-ΔMB0. From days 10 to 21, the cell proliferation increased significantly in c-MYC-ΔMB0 and MYCL-WT, and a concomitant increase in the CD13 (-) TRA-1-60 ( +) population was observed (Fig. 2D, E). These observations suggest that the number of cells that were reprogrammed increased rapidly with c-MYC-ΔMB0 and MYCL-WT. With c-MYC-WT, the cell proliferation continued until day 21. However, the CD13 (-) TRA-1-60 ( +) population hardly increased (Fig. 2D), indicating that these cells were not reprogramming but changing to other highly proliferative cell types. From these results, we concluded that the MB0 domain functions negatively in c-MYC and positively in MYCL for reprogramming.

MYCL regulates cytoskeleton-and cell adhesion-related proteins during reprogramming via the MB0 domain.
To confirm which genes are regulated by the MYCL MB0 domain in reprogramming, we analyzed protein expressions during reprogramming because it was reported that gene expressions do not correlate well with protein expressions 20 . We performed a comprehensive analysis of expressed proteins during reprogramming induced by c-MYC and MYCL WT and ΔMB0 mutants. We used SeV-reprogramming HDFs on days 3, 5, and 7 days and EpiP-reprogramming HDFs on day 10 as samples for mass spectrometry (MS) (Fig. 3A) because the percentage of TRA-1-60 ( +) cells was much higher with SeV than with EpiP for observations up to day 7 (Fig. 1C, F). There was more than a two-fold increase in the expression of 520 (SeV) and 128 (EpiP) proteins with MYCL-WT reprogramming compared to c-MYC-WT reprogramming (Fig. 3B, groups (i) and (ii), respectively) and 183 (EpiP) proteins with c-MYC-ΔMB0 reprogramming compared to c-MYC-WT reprogramming (Fig. 3B, group (iii)). Overall, we identified 18 proteins common to the three groups (Fig. 3B, group (iv)). Then, we applied a Gene Ontology (GO) analysis using DAVID and detected enriched terms during reprogramming 21,22 (Fig. 3C, D, and Table 1), finding cytoskeleton-and cell adhesion-related proteins are involved in the promotion of reprogramming by MYCL-WT. The same analysis was performed to identify proteins whose expression was upregulated by c-MYC-WT compared with MYCL-WT and c-MYC-ΔMB0 (Supplementary Fig. S10A and Table 2). These proteins were associated with the proliferation of non-iPSC-like colonies. We found that c-MYC-WT regulates proteins involved in cell proliferation, such as the cell cycle and DNA replication. To understand the function of the MB0 domain in reprogramming, MS analysis was applied to HDF samples transfected with MYCL-WT, MYCL-ΔMB0, or c-MYC-ΔMB0 (Supplementary Fig. S10B and Table 3). GO analysis indicated that these proteins were associated with cell adhesion and RNA processing.
We also compared phosphorylated proteins during SeV reprogramming with MYCL and c-MYC. In total, there was more than a two-fold relative increase of 17 phosphorylated proteins with MYCL-WT and 132 phosphorylated proteins with c-MYC-WT. The GO analysis indicated that the phosphorylated proteins increased by MYCL included cytoskeleton-related proteins and those increased by c-MYC included transcription-related proteins ( Supplementary Fig. S11).
MYCL regulates RNA processing-related proteins during reprogramming via the MB2 domain. Our analysis also revealed that, along with the MYCL MB0 domain, the MYCL MB2 domain is important for reprogramming (Fig. 2B). It has been reported that the c-MYC MB2 domain is involved in transformation activity, and tryptophan 135 within the MB2 domain is necessary for this activity 10 . MYCL also has a tryptophan residue within its MB2 domain but little transformation activity 23 . We hypothesized that this domain in MYCL has reprogramming function. We therefore produced a mutant in which tryptophan 96 was substituted with glutamate (W96E). This tryptophan is equivalent to tryptophan 135 in c-MYC ( Fig. 4A and Supplementary  Fig. S5B). We confirmed the expression of MYCL-W96E by western blotting (Supplementary Fig. S12). Next, we examined the effect of MYCL-W96E for reprogramming. HDFs were transfected with reprogramming factors including MYCL-WT or -W96E. MYCL-W96E could not induce iPSC-like colonies, suggesting tryptophan 96 is crucial for reprogramming (Fig. 4B, C). We thus hypothesized that the residue might be important for MYCL to bind to other proteins. To identify the binding proteins, we produced GST-fusion recombinant proteins of the MYCL MB2 domain (Fig. 4A). GST-MYCL-MB2-WT or -W96E proteins were immobilized on glutathione Sepharose, and affinity columns were prepared. Cell lysates were applied to the column, and, after washing, the bound proteins were eluted. We used the cell lysates from reprogramming HDFs, but since it was difficult to collect a large amount, we also used cell lysates from hiPSCs. The reason for using the hiPSC lysates is that many of the proteins expressed in reprogramming HDFs are highly expressed in hiPSCs as well 16,[24][25][26][27] .
Comprehensive proteomic analysis suggested that the MYCL MB0 domain influences the expression of cell adhesion-related proteins, and MYCL shows an up-regulation of phosphorylated cytoskeletal proteins (Fig. 3C, D, and Supplementary Fig. S11A). Cell adhesion is mediated by adhesion molecules, such as integrins and cadherins, which function in the extracellular matrix (ECM) and cell-cell adhesion and are important for cell  CRABP2  EXOC5  TRIP12  SLC25A24  PAIP1  NEK7   NOL11  DNAJC9  GADD45GIP1 SF3A3  ARL1  MICAL1   POLR2L  CRIP1  GTF2I  AGK  PODXL  S100A13   TOMM20  CCDC43  AHNAK  PPIL3  REXO2  STT3A   CFL2  SRBD1  ENDOD1  IDI1  FNTA  H3-3A,H3-3B   POTEF  SUGP2  PDHA1  ELN  FBLN2  SCPEP1   TSPYL5  MTA1  HOOK3  COL4A2  NDRG3  TXLNG   SGTA  TMSB4X  HLA-H  MAP2  EXOC2  MPI   REEP5  ACTG1  PUM1  ATP5MG  SOX2  HMGN1   SGCD  HINT1  KPNA4  PIK3R4  DBI  UBQLN2   HMGB3  PTGR1  METTL26  COX7C  TUBB2A  NDUFB9   SYAP1  EMC2  B4GALT4  NCKAP1  SRP9 Table 3. MS analysis of identified proteins in cells reprogrammed with MYCL-WT and c-MYC-ΔMB0 compared with MYCL-ΔMB0. Three groups are described: (i) proteins whose peptide counts increased more than two-fold in MYCL-WT compared to MYCL-ΔMB0; (ii) proteins whose peptide counts increased more than two-fold in c-MYC-ΔMB0 compared to MYCL-ΔMB0; and (iii) commonly identified proteins. n = 3.  www.nature.com/scientificreports/ communication and the regulation of fundamental physiological processes such as tissue development and maintenance 36,37 . Human iPSCs and hESCs have unique focal adhesion localization, and appropriate adhesion to the ECM is required to regulate reprogramming via MET and maintain pluripotency [38][39][40] . Accordingly, our study supports MYCL regulating cell-substrate adhesion through its MB0 domain to promote reprogramming. In other words, MYCL might regulate proteins involved in cell adhesion and the cytoskeleton directly or indirectly to cause MET and promote reprogramming. In c-MYC, loss of the MB0 domain positively affects iPSC-like colony formation, suggesting that this domain has a different function compared to MYCL. This functional difference is somewhat surprising since the domain is well conserved (Supplementary Fig. S5B). We would like to clarify this point in the future. We also found that the MB2 domain has an important function in MYCL-reprogramming (Fig. 2B,C). Deleting the MB2 domain completely compromised the reprogramming ability of MYCL. In c-MYC, the MB2 domain has an important function in transformation activity 14 , and tryptophan 135 in the MB2 domain is essential for this activity. The equivalent tryptophan residue in MYCL is tryptophan 96. MYCL has little transformation activity, but we showed that the mutation of tryptophan 96 completely lost the reprogramming ability of MYCL. To further investigate the function, we sought interacting proteins by affinity column chromatography. We found 31 proteins, including 26 RBPs, that interact with the MYCL MB2 domain (Table 4, genes written in blue). A GO analysis suggested that some of the 31 proteins are involved in RNA processing (Table 4). It has been reported that altered RNA processing affects somatic cell reprogramming 41 . Therefore, we hypothesize that MYCL also promotes MET in reprogramming by regulating RNA processing via interactions with RBPs at its MB2 domain. An illustrative summary of how MYCL regulates cell reprogramming through these two domains is shown in Fig. 5.
To conclude, we have demonstrated that MYCL promotes more efficient reprogramming than c-MYC, regulates the expression of cell adhesion and cytoskeletal proteins, and is involved in RNA processing via a single tryptophan residue in the MB2 domain. Following these findings, we propose that MYCL causes MET by regulating the expression of proteins involved in the promotion of reprogramming from the RNA-processing stage. . We speculate that MYCL promotes reprogramming through the synergistic effects of these two mechanisms.