14-3-3 binding motif phosphorylation disrupts Hdac4-organized condensates to stimulate cardiac reprogramming

SUMMARY

Cell fate conversion is associated with extensive post-translational modifications (PTMs) and architectural changes of sub-organelles, yet how these events are interconnected remains unknown. We report here the identification of a phosphorylation code in 14-3-3 binding motifs (PC14-3-3) that greatly stimulates induced cardiomyocyte (iCM) formation from fibroblasts. PC14-3-3 is identified in pivotal functional proteins for iCM reprogramming, including transcription factors and chromatin modifiers. Akt1 kinase and protein phosphatase 2A are the key writer and key eraser of the PC14-3-3 code, respectively. PC14-3-3 activation induces iCM formation with the presence of only Tbx5. In contrast, PC14-3-3 inhibition by mutagenesis or inhibitor-mediated code removal abolishes reprogramming. We discover that key PC14-3-3-embedded factors, such as histone deacetylase 4 (Hdac4), Mef2c, and Foxo1, form Hdac4-organized inhibitory nuclear condensates. PC14-3-3 activation disrupts Hdac4 condensates to promote cardiac gene expression. Our study suggests that sub-organelle dynamics regulated by a PTM code could be a general mechanism for stimulating cell reprogramming.

Graphical Abstract

In brief

Liu et al. identify that PC14-3-3 activation in pivotal functional proteins, such as Hdac4 and Mef2c, is essential in cardiac reprogramming. PC14-3-3 activation disrupts Hdac4-organized repressive condensates leading to the release of multiple 14-3-3-motif-embedded proteins from the condensates, which turns on cardiac gene expression and stimulates cardiac reprogramming.

INTRODUCTION

Heart disease is the leading cause of death in the United States and around the world. In heart injuries, such as myocardial infarction, millions of cardiomyocytes undergo irreversible necrosis and fibrosis. For patients who survive the myocardial infarction, the injured muscle tissues in the infarcted heart area are replaced with non-functional fibrotic scar tissues, which leads to severely weakened pumping function of the heart. There is a critical need to develop effective therapeutic strategies to preserve the pumping function after myocardial infarction. Direct reprogramming of fibroblasts into cardiomyocyte (CM)-like cells (called induced CMs [iCMs]) by introducing three cardiac transcription factors, Gata4, Mef2C, and Tbx5 (G, M, and T), has emerged as an attractive strategy to repair damaged hearts.^1-3 These iCMs, when generated in situ in an infarcted heart, integrate electrically and mechanically with the surrounding myocardium, leading to a reduction in scar size and an improvement in heart function. However, the low conversion rate, poor purity, and lack of precise conversion of iCMs remain significant challenges. Limited understanding of the molecular mechanisms of iCM reprogramming is a key obstacle preventing its effective clinical applications.

Cardiac reprogramming is a cell fate conversion process involving extensive post-translational modifications (PTMs).^4-9 PTMs of histones have been extensively studied as a coding system to guide alterations in cell plasticity and cell fate, such as differentiation/programming and dedifferentiation/reprogramming. Coordinated PTMs on chromatin function as a histone code, and this code is added, read, and erased by chromatin modifiers.¹⁰ In parallel with the PTM coding system of histones, PTMs of free proteins are also essential for organ development and cell differentiation. Due to the complexity of PTMs on free proteins, most studies focus on specific PTM sites in a particular protein and generally define those PTMs as a feature belonging to the individual protein. However, the functional output of similar PTMs on a broad range of motif-embedded proteins as a coding system has not been explored. We propose that the histone code can be logically extended to include transcription factors and chromatin modifiers called the “PTM code.” This PTM code will be a complementary coding system to the genetic code to guide cell fate conversion, such as iCM reprogramming.

Cell fate conversion is also associated with extensive architectural changes of sub-organelles and organelles, yet how these organelle/sub-organelle dynamics are regulated and contribute to cell fate change is largely unknown. Biomolecular condensates (phase separation of proteins) are non-membrane-bound sub-organelles composed of heterogeneous or homogeneous proteins.^11,12 Recent studies indicate that chromatin modifiers and transcription factors can aggregate into nuclear condensates and synergistically regulate gene expression.^13-15 These condensates could play an essential role in cell fate change. Condensate formation typically requires a phase separation region within the intrinsically disordered regions (IDRs) that does not form a fixed three-dimensional structure.¹⁶ Notably, phosphorylation can alter the overall charge of IDRs in the protein or alter the protein-protein interaction to regulate the condensate formation.¹⁷

In this study, we aim to characterize the roles of PTMs in the molecular mechanisms of iCM reprogramming. We report our identification of a phosphorylation code in 14-3-3 binding motif proteins (PC14-3-3) that guide cardiac reprogramming. Our initial screen has identified the presence of PC14-3-3 in pivotal functional proteins that stimulate iCM formation. We next show that PC14-3-3 code removal abolishes cardiac reprogramming. Furthermore, we also identified that Akt1 kinase and protein phosphatase 2A (PP2A) were a key writer and a key eraser of the PC14-3-3 code, respectively. Activation of the PC14-3-3 code by Akt1 expression and PP2A inhibition greatly enhanced reprogramming, inducing iCM formation with the presence of only Tbx5. Mechanistically, PC14-3-3 activation disrupts nuclear inhibitory histone deacetylase 4 (Hdac4) condensates that trap other PC14-3-3-embedded factors, which in turn release transcription factors to promote cardiac gene expression. Our research reveals PC14-3-3 as an essential phosphorylation code that regulates sub-organelle dynamics to stimulate cardiac reprogramming, which may have general significance in cell fate determination and cell fate change.

RESULTS

Phosphorylation in 14-3-3 binding motifs is a phosphorylation code embedded in diverse classes of proteins that regulates cardiac reprogramming

To determine the impact of phosphorylation on cardiac reprogramming factors in iCM induction, we first constructed a phosphorylation-deficient mutant library for cardiac reprogramming factors. We picked the three primary factors MGT (Mef2c, Gata4, and Tbx5) as well as three other transcription factors (TFs) (Hand2, Mesp1, and Nkx2-5) that play important functions in cardiac reprogramming. We attempted to include all the possible phosphorylation sites on those TFs, including the known (from PhosphoSitePlus)^18,19 and predicted ones (Table S1). Three types of software, PhosphoSVM,²⁰ NetPhos,²¹ and Musite,²² for phosphorylation prediction were used to identify 132 sites on all the TFs with a high confidence rate (Table S1). Therefore, 132 known or predicted phosphorylation sites on these proteins were mutated into alanine to mimic dephosphorylated states.

Primary factors MGT or MGT plus either Hand2, Mesp1, or Nkx2-5 was introduced into mouse embryonic fibroblasts (MEFs) to induce cardiac reprogramming. TFs with dephosphorylation-mimicking mutations were tested side by side with the wild type (WT) counterparts for their ability to affect iCM induction following the established iCM evaluation protocol (Figure 1A). 10 dephosphorylation-mimicking modification sites were observed to affect reprogramming based on the expression of Myh6 (Figure 1B). Key mutants that dramatically affect reprogramming were also validated by measuring the CM structure gene cardiac troponin T (cTnt) protein expression with immunocytochemistry (Figures 1C, 1D, and S1A). These results indicate that phosphorylation modifications of TFs are critical for cardiac reprogramming.

Figure 1. Phosphorylation in 14-3-3 binding motifs is a phosphorylation code embedded in diverse classes of proteins that can regulate cardiac reprogramming.

(A) A schematic diagram of the phosphorylation screen strategy. Mutated individual genes were transduced into MEFs derived from a transgenic α-major histocompatibility complex (αMHC)-GFP reporter mouse with retroviruses expression. Medium was changed every 2 days. Reprogramming efficiency was evaluated 14 days after transduction by qPCR.

(B) Myh6 expression measured after MEF induction with dephosphorylation mutation library.

(C and D) Immunocytochemistry (ICC) of cardiac markers cTnT of S82A mutation construct or WT-transduced cells by fluorescence microscopy (100×). Mef2c-GT, Mef2c+Gata4+Tbx5. *p < 0.05, **p < 0.01, and ***p < 0.001 vs. MGT. Fold change (FC) of GFP-positive or cTnT-positive cells counted in 100× power field (PF). n = 10.

(E) Mutations of 14-3-3 binding motif phosphorylation disrupted 14-3-3 binding to Mef2c. HEK293 cells were transfected with 14-3-3 together with either WT or S82A or T20A mutation of Mef2c. Mef2c was immunoprecipitated with anti-Mef2c antibody followed by western blotting and detection of 14-3-3 isoform YWHAB-FLAG by FLAG antibody.

(F) A schematic diagram of the 14-3-3 binding motif dephosphorylation-mimicking screen strategy. Mutated individual genes were transduced into MEFs derived from a transgenic αMHC-GFP reporter mouse with retroviruses expressing MGT. Reprogramming efficiency was evaluated 14 days after transduction by qPCR as shown in (A).

(G) Relative expression of Myh6 in iCM reprogramming with WT and mutant PC14-3-3-containing proteins 14 days after MGT transduction. **p < 0.01 vs. WT. n = 3.

(H) Relative expression of Myh6 in iCM reprogramming with WT and mutant PC14-3-3-containing proteins 14 days after MGT transduction. Data are normalized to the MGT+empty plasmid group. *p < 0.05, **p < 0.01, and ***p < 0.001 vs. WT. FC, fold change.

Data represent three independent experiments and are presented as mean ± SD. Each independent experiment consists of at least three technical replicates. One-way ANOVA.

Next, we examined how these phosphorylation sites affected reprogramming. Since phosphorylation is well known for regulating the adjacent motif function, we mapped the phosphorylation site onto predicted motifs using the eukaryotic linear motif (ELM) resource.²³ Intriguingly, 7 out of 10 phosphorylation sites that affected reprogramming are within the 14-3-3 binding motif predicted by ELM (Figure S1B). Further validation of these sites using 14-3-3 motif predictors,^24,25 especially 14-3-3-Pred, revealed that 5 out of the 10 sites (Mef2c S82, Mef2c T20, Gata4 S261, Mesp1 S82, Hand2 T103) were predicted to be 14-3-3 binding motifs (Figure S1B). We further validated that the Mef2c S82A mutant diminished 14-3-3’s interaction with the mutant, as shown by our immunoprecipitation (IP) assays (Figure 1E). Consequently, at least, the phosphorylation of 14-3-3 binding motifs within Mef2c could affect reprogramming.

We then examined the pathways related to the 14-3-3-binding-motif-containing proteins during cardiac reprogramming with Gene Ontology (GO) analysis. Our analysis indicated that 14-3-3-binding-motif-containing proteins were highly related to multiple pathways involved in cardiac reprogramming (Figure S1C). Many TFs (Yap1, Foxo1), chromatin modifiers (Hdac4), and histone proteins (H3) related to heart development, muscle cell differentiation, and muscle cell apoptosis were also identified. To determine if phosphorylation of 14-3-3 binding motifs embedded in heart development or muscle cell differentiation proteins could contribute to regulating cardiac reprogramming, we next selected 19 mutations in 18 key TFs, chromatin modifiers, and histone proteins carrying known 14-3-3 binding motifs to validate the possible function of their 14-3-3 binding motifs in cardiac reprogramming (Table S2). We created a 14-3-3 binding motif dephosphorylation-mimicking mutant protein as a control for each WT protein (Figure 1F). Primary factors MGT plus WT or 14-3-3 binding motif mutant protein was introduced into MEFs to induce cardiac reprogramming. The effects of mutant proteins on iCM induction were examined by the established iCM evaluation protocol (Figure 1A). 7 out of 18 tested protein mutants, including those of Hdac4, Foxo1, Nrip1, and Bad, significantly inhibited cardiac reprogramming based on Myh6 and Actc1 expression (Figures 1G and S1D). More importantly, compared to the WT protein, a general decrease in cardiac gene Myh6 and Actc1 was observed from the mutants (Figures 1H and S1E). These observations indicate that phosphorylation of 14-3-3 binding motifs in these factors is the key phosphorylation event for cardiac reprogramming.

14-3-3 binding motif phosphorylation is required for and positively associated with cell fate conversion from fibroblasts to iCMs

Next, we determined the changes of 14-3-3 binding motif phosphorylation during reprogramming. We established an inducible reprogramming cell line using cardiac-specific Myh6 promoter-driven fluorescence marker gene GFP to evaluate cardiac reprogramming (Figure 2A). The polycistronic MGT and a transactivator to control the tetracycline-responsive promoter were introduced into this inducible cell line (see STAR Methods for more details). Successfully reprogrammed Myh6− GFP+ cells can be separated from the unsuccessfully reprogrammed Myh6− GFP− cells by flow cytometry sorting (Figure 2B). Since 14-3-3 binding motifs are broadly present in many proteins, we first examined the phosphorylation levels of all the 14-3-3 binding motifs during reprogramming using the phospho-(Ser) 14-3-3 binding motif antibody (CST 9601). Indeed, overall 14-3-3 binding motif phosphorylation is much higher in successfully reprogrammed CMs than nonreprogrammed cells (Figure 2C). In order to study 14-3-3 binding motif phosphorylation dynamics during the reprogramming process, proteins from MGT- and control-vector-infected MEF cells were harvested every 2 days up to day 14. Overall 14-3-3 binding motif phosphorylation increased during reprogramming in MGT-infected cells, especially from days 4 to 12 (Figure S2A). To determine whether a high level of phosphorylation in 14-3-3 binding motifs is specific to CMs among heart cells, neonatal CMs and cardiac fibroblasts (CFs) were also purified, and the phosphorylation of 14-3-3 binding motifs was compared (Figure S2B). Much higher phosphorylation in 14-3-3 binding motifs was observed in CMs than in CFs.

Figure 2. Phosphorylation in 14-3-3 binding motifs is required and positively associated with conversion from fibroblasts to iCMs.

(A) A schematic diagram of establishing an inducible cardiac reprogramming cell line with cardiac-specific Myh6 promoter-driven marker gene GFP.

(B) Flow cytometry sorting of reprogrammed cells. The inducible cardiac reprogramming cell line was treated with doxycycline for 2 days. Successfully reprogrammed Myh6− GFP+ cells and unsuccessfully reprogrammed Myh6− GFP− cells were separated by flow cytometry.

(C) Detection of 14-3-3 binding motif phosphorylation by western blot. Protein lysates from Myh6− GFP+ cells and Myh6− GFP− cells isolated from (B) were used for the assay.

(D and E) 14-3-3 inhibitor (14-3-3 antagonist I, 2-5) (D) and direct shRNA knockdown of 14-3-3 proteins (E) significantly decreased reprogramming efficiency by qPCR. n = 3.

(F and G) ICC and quantification of cardiac marker cTnT of MGT-transduced cells with 14-3-3 inhibitor or shRNA knockdown of 14-3-3 proteins. n = 10. *p < 0.05, **p < 0.01, and ***p < 0.001 vs. MGT.

Data represent three independent experiments and are presented as mean ± SD. Each independent experiment consists of at least three technical replicates. One-way ANOVA.

A major function of the phosphorylated 14-3-3 binding motifs is to engage the interaction of the motif-containing proteins with its reader: 14-3-3.^26,27 To further determine the role of phosphorylated 14-3-3 binding motifs in iCM formation, we applied 14-3-3 inhibitor (14-3-3 antagonist I, 2-5) (Figure 2D) and short hairpin RNA (shRNA) against all isoforms of 14-3-3 proteins (Figures 2E and S2C) to disrupt the interaction of PC14-3-3 with 14-3-3s. Both the 14-3-3 inhibitor and shRNAs against 14-3-3 significantly decrease reprogramming efficiency (Figures 2F and 2G), indicating that the interaction of PC14-3-3-containing proteins with its reader 14-3-3 is essential for cardiac reprogramming. We also further validated the function of 14-3-3 in reprogramming by expressing a dominant-negative K49E mutation in YWHAE^28,29 during the reprogramming process. The results show that disrupting 14-3-3 function by K49E mutation also reduced cardiac reprogramming efficiency in the reprogramming cell line (Figure S2D). In contrast, overexpression of 14-3-3 proteins did not induce much change in cardiac reprogramming efficiency (Figure S2E). These results indicate that the phosphorylation in 14-3-3 binding motifs, but not 14-3-3 protein, is key to cardiac reprogramming.

Based on these results, we hypothesize that the phosphorylation of 14-3-3 binding motifs in many important functional proteins could be coordinated during cardiac reprogramming, and this phosphorylation stimulation is essential for cardiac reprogramming. We define the phosphorylation acquired on those 14-3-3 binding motifs as a phosphorylation code of 14-3-3 binding motifs (PC14-3-3).

PP2A inhibitor okadaic acid (OA) and Akt1 treatment leads to PC14-3-3 activation and drastically enhances reprogramming

Identifying the PC14-3-3 code prompted us to identify the key kinases (writers) and phosphatases (erasers) of the code to regulate reprogramming. Intriguingly, 6 out of those 8 14-3-3-binding-motif-embedded proteins identified in Figure 1G that affected reprogramming are regulated by OA or PP2A,^30-39 and 3 out of the 8 have been reported to be regulated by Akt1 (Table S3).^33,40-43 Moreover, 7 out of 8 are regulated by either PP2A or Akt1. Importantly, Akt1 has been reported to enhance reprogramming.⁴³ Therefore, we focused on the potential of PP2A and Akt1 as a key eraser and a key writer, respectively.

We first examined PP2A and its inhibitor OA on PC14-3-3 activation and cardiac reprogramming. The addition of pharmaceutical PP2A inhibitor OA significantly activated the PC14-3-3 code (Figures 3A and S3A). To test whether there is a correlation between PC14-3-3 stimulation and reprogramming efficiency, we next examined the effect of OA on reprogramming. OA treatment promoted higher cTnT and α-actinin expression with up-regulation of cardiac genes (Figures S3B-S3D). PP2A inhibition also promoted higher spontaneous beating and calcium transient activities of iCMs derived from fibroblasts (Figures S3E and S3F). Consistent with these observations, PP2A knockdown by shRNA increased cardiac reprogramming efficiency (Figures S3G and S3H). Therefore, PP2A inhibition enhanced reprogramming mainly by activating PC14-3-3.

Figure 3. PP2A inhibitor okadaic acid (OA) and Akt1 treatment leads to PC14-3-3 activation and drastically enhances reprogramming.

(A) PP2A inhibitor OA activated PC14-3-3. Phosphorylation of 14-3-3 binding motifs was detected by western blot of protein lysates from MEF cells with OA treatment.

(B) Relative expression of CM marker genes of iCM reprogramming with Akt1 and OA treatment 14 days after MGT transduction with or without shRNA YWHAZ/E. Statistical analyses were performed between MEFs with Akt1 and OA or DMSO-treated groups. *p < 0.05 vs. MGT. #p < 0.05 and ##p < 0.01 vs. MGT+Akt1+OA. n = 3.

(C and D) ICC and quantification of cardiac markers TnT and Myh6-GFP of MGT-transduced cells with or without Akt1 and OA treatment with or without shRNA YWHAZ/E by fluorescence microscopy (100×). ICC results indicated increased iCM maturation with the treatment of MGT+Akt+OA. ***p < 0.001 vs. MGT. ###p < 0.001 vs. MGT+Akt1+OA. n = 10.

(E) Quantification of spontaneous Ca²+ oscillation cells per field with indicated viral infection treatment for 1–4 weeks (n = 50 from 10 wells). ***p < 0.001 vs. MGT+DMSO. ###p < 0.001 vs. MGT+Akt1+OA.

(F) Quantification of beating iCMs loci with indicated viral infection for 1–4 weeks (n = 10). ***p < 0.001 vs. MGT+DMSO. ###p < 0.001 vs. MGT+Akt1+OA.

(G) Phosphorylation of 14-3-3 binding motifs was detected by western blot of protein lysates from MEF cells with or without MK-2206 (Akt1 inhibitor) + DT-061 (PP2A activator) treatment.

(H) Successfully reprogrammed Myh6− GFP+ cells were detected by flow cytometry with or without MK-2206 (Akt1 inhibitor) + DT-061 (PP2A activator) treatment in iCM cell line.

(I) Relative expression of CM marker genes in iCM reprogramming with Akt1 and OA treatment 14 days after Tbx5 transduction. *p < 0.05, **p < 0.01, and ***p < 0.001 vs. NC.

(J and K) ICC and Quantification of cardiac markers α-actinin of Tbx5 or control-transduced cells with or without Akt1 and OA treatment by fluorescence microscopy (400×). ***p < 0.001, vs. NC.

(L and M) Quantification of spontaneous Ca²+ oscillation cells per field (L) or beating iCM loci with Tbx5 or control-transduced cells (M) with or without Akt1 and OA treatment for 1–4 weeks (n = 10). *p < 0.05, **p < 0.01, ***p < 0.001, #p < 0.05, ##p < 0.01, and ###p < 0.001 vs. NC.

Data represent three independent experiments and are presented as mean ± SD. Each independent experiment consists of at least three technical replicates. One-way ANOVA.

We then examined the combinatorial effect of OA and Akt1 on PC14-3-3 activation and reprogramming. OA and Akt1 combination induced much higher PC14-3-3 activation than any individual treatment (Figure S3A), and higher PC14-3-3 activation also led to higher reprogramming efficiency. Particularly, the sarcomeric, contractility and ion channel related gene expression (Actn2, Actc1, Ryr2, and Myh6) indicated that reprogramming efficiency increased dramatically with OA and Akt1 treatment compared to the standard MGT procedure, represented by a 75-fold increase of Myh6 expression (Figure 3B). The PC14-3-3 activation by OA and Akt1 for 2 weeks also promoted iCM maturation as indicated by immunostaining against cTnT, the specific components of the sarcomere in CMs (Figures 3C and 3D). Moreover, spontaneously contracting cells were apparent 1 week after PC14-3-3 activation treatment (Figure 3E). Approximately 60 beating loci per well were identified in the PC14-3-3-activated group 4 weeks after transduction, 10-fold more than in the MGT control group (Figure 3E; Videos S1, S2, S3, S4, S5, and S6). Consistent with more beating cells, PC14-3-3-activated iCMs showed significantly higher spontaneous Ca²⁺ oscillations with various frequencies (Figure 3F; Videos S1, S2, S3, S4, S5, and S6). More importantly, YWHAZ/E knockdown significantly decreased Akt1+OA-induced reprogramming (Figures 3B-3F). All our data and analyses indicated that although there could be diverse kinases and phosphatases for PC14-3-3 code, most of the motifs that have a significant effect on reprogramming can be phosphorylated by Akt1 and dephosphorylated by PP2A. Consequently, PP2A inhibitor OA and Akt1 synergistically led to PC14-3-3 activation and drastically enhanced iCM reprogramming. In contrast, PP2A activator DT-061 and/or Akt1 inhibitor MK-2206 treatment downregulated PC14-3-3 (Figure 3G) and reprogramming efficiency in an inducible reprogramming cell line (Figure 3H).

To further determine the key role of PC14-3-3 in cardiac reprogramming, we tested whether PC14-3-3 activation could replace any of the primary cardiac reprogramming factors, Mef2c, Gata4, or Tbx5. We found that Tbx5 was the only TF required to achieve cardiac reprogramming when PC14-3-3 was activated by Akt1+OA when using neonatal CFs. The Tbx5, Akt1, and OA combination activated cardiac genes Actn2, Actc1, Ryr2, and Myh6 from CFs, similar to the MGT combination. Furthermore, immunostaining against cTnT, a specific component of the sarcomere in CMs, clearly identified the sarcomeric structure (Figures 3J and 3K), indicating that Akt1+OA+TBX5 was able to promote iCM maturation. OA+Akt1+Tbx5 also induced beating iCMs (Figure 3L) with spontaneous Ca²⁺ oscillations (Figure 3M) from neonatal CFs (Videos S7, S8, S9, and S10). Together, these data indicate that PC14-3-3 renders two reprogramming factors, Mef2c and Gata4, dispensable in cardiac reprogramming.

PC14-3-3 activation disrupts Hdac4-organized condensates containing numerous 14-3-3 binding motif proteins

Our early investigation (Figure 1) revealed that protein mutants Hdac4 and Hdac5, deficient in phosphorylation of 14-3-3 binding motifs, significantly inhibited cardiac reprogramming (Figure 1). Hdac4 and Hdac5 belong to the class IIa HDAC family and possess identical 14-3-3 binding motifs. We therefore examined the dynamics of class IIa HDACs (Hdac4+5+9) proteins during cardiac reprogramming. Surprisingly, under iCM-inducing conditions (DMEM and M199 medium in a 4:1 ratio with 10% FBS), endogenous class IIa HDACs form condensates in the nucleus (Figures 4A, 4B, and S4D). More importantly, PC14-3-3 activation by Akt1+OA decreased the number of class IIa HDAC condensates, whereas PC14-3-3 repression by Akt1 inhibitor MK-2206 and PP2A activator DT-061 increased it.

Figure 4. PC14-3-3 activation regulates Hdac4-organized condensates with multiple 14-3-3-motif-embedded proteins.

(A and B) ICC of endogenous Hdac4 in reprogramming cells treated with OA+Akt1 or MK-2206 (Akt1 inhibitor) + DT-061 (PP2A activator) by fluorescence microscopy. Magnification image showing the Hdac4+5+9 condensates (green dots) in a single reprogramming cell. Sample size: MK+DT (n = 321), DMSO (n = 344), and Akt1+OA (n = 287).

(C and D) A schematic diagram showing Hdac4 14-3-3 binding motifs, intrinsically disordered regions (IDRs), potential phase separation regions, and the construction strategy for optoIDR assay.

(E) Time-lapse images of the HE293T cell expressing Hdac4205-646aa-optoIDR with laser excitation. A droplet fusion event occurs in the Hdac4-3SA205-646aa optoIDR group but not in the Hdac4205-646aa optoIDR group.

(F–H) Co-overexpression of Hdac4-3SA with PC14-3-3-embedded factors Mef2c (F), Foxo1 (G), and Nrip (H) showing that Hdac4 co-localized with these factors in the same condensate.

(I–K) Co-overexpression of Hdac4-3SA with PP2A subunits showing that Hdac4 co-localizes with PP2A subunits B56alpha (I), B56delta (J), and B56epsilon (K) in the same condensate.

Data represent three independent experiments and are presented as mean ± SD. Each independent experiment consists of at least three technical replicates. One-way ANOVA.

Between Hdac4 and Hdac5, the most effective regulator was Hdac4 in cardiac reprogramming, as determined by qPCR of CM markers (Figure S1D). Compared to WT Hdac4, Hdac4 mutants (Hdac4 -3SA) with depleted phosphorylation of 14-3-3 binding motifs (serine 246, 467, and 632 [S246, S467, and S632] to alanine mutations) also largely abolished reprogramming as revealed by immunostaining of cTnT and Myh6 (Figures S4A-S4C). We therefore focused on Hdac4 and its PC14-3-3 motifs in HDAC condensate dynamics for PC14-3-3-mediated reprogramming.

We first applied an optoIDR assay to test PC14-3-3 function in Hdac4 condensate formation.⁴⁴ The photoinducible, self-associating Cry2 protein was labeled with mCherry and fused to the IDR of WT Hdac4 or mutant Hdac4 (Hdac4-3SA) (Figures 4C and 4D), which contains the 14-3-3 binding motifs around S246, S467, and S632 (Figure 4C). The fusion of all three 14-3-3 binding motif (S246, S467, and S632) mutant Hdac4-3SA IDRs to Cry2-mCherry facilitated the rapid formation of micron-sized spherical droplets upon blue light stimulation in HEK293 cells (Figure 4E). In contrast, WT Hdac4-IDR-Cry2-mCherry did not form condensates with blue light stimulation (Figure 4E). Given that Hdac4 contains three PC14-3-3 binding motifs, we conducted additional tests to determine which specific motif could regulate Hdac4 condensate formation. By deleting both the 14-3-3 binding motif and the adjacent IDR individually, we found that 14-3-3 binding motif and adjacent region IDRs around the 14-3-3 binding motif phosphorylation sites S246 and S632 were required for Hdac4 condensate formation (Figure S4E). In particular, the S246 site in Hdac4 is a nuclear localization signal, and previous reports have indicated that 14-3-3 binding to the S246 region prevents Hdac4 from translocating into the nucleus. Notably, S246 did not form condensates, implying that nuclear localization is a prerequisite for condensate formation. Together, these data indicate that Hdac4 forms condensates and that both S246 and S632 are crucial for condensate formation, with S246 being particularly significant, as condensate formation may require nuclear localization. More importantly, the Hdac4 condensates are regulated by PC14-3-3.

We next investigated the potential components within Hdac4 condensates. GO analysis of 14-3-3 binding protein showed that there is a cluster of proteins that are related to the nuclear body or granule (Figure S4F). Intriguingly, optoIDR assay indicated that not only Hdac4 but also other PC14-3-3-embedded factors (Foxo1 and Mef2c) can regulate both reprogramming and condensate formation through PC14-3-3 activation (Figures S4G-S4J). Most importantly, co-overexpression of PC14-3-3 mutant versions of Hdac4 (Hdac4-3SA) with PC14-3-3-embedded proteins in HEK293 cells show that Mef2c, Foxo1, or Nrip1 could co-localize with Hdac4 in the same puncta (Figures 4F-4H).

Because the PP2A complex plays a pivotal role in the regulation of PC14-3-3, as we have demonstrated in previous section (Figure 3), and dephosphorylation of Hdac4 is crucial for condensation formation (Figure 4E), we next examined whether PP2A, or its subunits, co-localized with Hdac4 condensates to maintain Hdac4 in a non-phosphorylated state. Indeed, we observed that PP2A subunits B56alpha, B56delta, and B56epsilon co-localized within Hdac4 condensates (Figures 4I-4K), suggesting that PC14-3-3 heavily regulates Hdac4-organized condensates. These results imply that PC14-3-3 activation serves as a general mechanism to regulate the formation and dynamics of 14-3-3-motif-embedded protein condensates, which are organized by Hdac4 and contain Mef2c, Foxo1, and Nrip1.

Disruption of Hdac4 condensates is responsible for PC14-3-3 stimulated cardiac reprogramming

To determine if Hdac4 condensate disruption is responsible for PC14-3-3-stimulated cardiac reprogramming, we first generated several 14-3-3 permanent binding forms of Hdac4. R18 is an unphosphorylated peptide that binds to 14-3-3 proteins with high affinity.^45,46 Therefore, replacing the 14-3-3 binding motif with the R18 sequence should induce permanent binding of 14-3-3 to the R18 sequence (see STAR Methods). We reason that if the effect of the Hdac4-3SA mutant on inhibiting cardiac reprogramming is due to the disrupted binding of Hdac4 to 14-3-3, then replacing the mutant 14-3-3 binding motifs with R18 that binds strongly to 14-3-3 proteins will largely abolish the inhibitory effect. To determine if this is indeed the case, the three 14-3-3 mutant binding motifs of Hdac4-3SA were replaced with an R18 peptide (Figures 5A and S5A).

Figure 5. PC14-3-3 dissociates Hdac4 condensates to facilitate cardiac reprogramming.

(A) A schematic diagram of Hdac4 mutants in 14-3-3 binding motifs combined with replacement of a permanent binding peptide R18.

(B) Immunostaining of specific mutant Hdac4 proteins in reprogramming cells.

(C) Co-overexpression of Hdac4 mutants with PC14-3-3-embedded factor Mef2c showing that R18 dissociates Hdac4 co-localized with MEF2C in the same condensate.

(D) Relative expression of CM marker genes during iCM reprogramming with expression of Hdac4 mutants 3R18 and 3SA. *p < 0.05, **p < 0.01, and ***p < 0.001, WT vs. 3SA. #p < 0.05, ##p < 0.01, and ###p < 0.001, 3SA vs. 3R18. n = 3.

(E) ICC of Hdac4 in iCM cells with Hdac4 overexpression and Hdac4 inhibitor treatment. Among Hdac4 inhibitors, TMP269, but not MC1568 and BML210, dissociates Hdac4 condensates.

(F and G) ICC by fluorescence microscopy (F) and quantification (G) of cardiac markers cTnT and Myh6-GFP of MGT-transduced cells with Hdac4 inhibitor TMP269 treatment (400×).

(H) Heatmap showing the effect of Hdac4 inhibitor TMP269 on representative cardiac-related gene expression (GO: 0048738) among MGT-transduced cells. n = 3.

(I) Chromatin immunoprecipitation (ChIP)-qPCR of Hdac4-GFP at cardiac gene loci from iCM cells treated with OA (PC14-3-3 activation) or MK+DT (PC14-3-3 inhibition). *p < 0.05 vs. IGG. n = 3.

Data represent three independent experiments and are presented as mean ± SD. Each independent experiment consists of at least three technical replicates. One-way ANOVA.

The combined three binding motif mutant construct Hdac4-3R18 was introduced into MEF cells with MGT for condensates and reprogramming analyses. For control groups, WT and PC14-3-3 mutant forms of Hdac4 in the MEF cells show an obvious difference in cellular distribution with immunostaining (Figures 5B and 5C). Unlike WT Hdac4, which was distributed evenly and had no condensates in the cytosol, mutant Hdac4 only existed in the nucleus and formed obvious biomolecular condensate-like puncta in the nucleus. Importantly, drastic dissociation of Hdac4 condensates was observed with the overexpression of Hdac4-3R18, indicating that 14-3-3 binding motifs were involved in the regulation of Hdac4 condensates (Figures 5B, 5C, and S5B). Moreover, Hdac4-3R18 indeed abolished the inhibitory effect of Hdac4 in cardiac reprogramming (Figure 5D). Consistent with the fact that PC14-3-3 activation disrupted Hdac4 condensates, other proteins within the condensates also lost their localization with Hdac4 condensates, such as Mef2c (Figures 5C and S5C). Taken together, these results indicate that PC14-3-3 directly regulates Hdac4 condensate formation and is a key target of PC14-3-3 to stimulate cardiac reprogramming.

It is also possible that PC14-3-3-stimulated cardiac reprogramming may be due to its effect on Hdac4 subcellular localization, but not condensate regulation, because phosphorylation of S246 on Hdac4 can induce Hdac4 translocation to the cytosol.^47,48 We constructed permanent nucleus-localized Hdac4 forms to differentiate between these two possibilities by deleting the C-terminal nuclear export signal (NES)⁴⁹ on Hdac4 (Figure S5A). The Hdac4-3R18-NES expression in the nucleus did not lead to condensate formation in MEFs (Figure S5B). Moreover, Hdac4-3R18 expression indeed abolished the inhibitory effect of Hdac4 in cardiac reprogramming (Figure S5D). This result reveals that the effect of PC14-3-3 on cardiac reprogramming is due to it regulating Hdac4 condensate dynamics but not Hdac4 subcellular localization. In addition, among the single 14-3-3 binding motif mutations of Hdac4, we observed significant changes in cardiac gene expression with Hdac4-1R18-246-2SA and Hdac4-1R18-632-2SA, but not Hdac4-1R18-467-2SA, implying that PC14-3-3 on S246 and S632 is more important for cardiac reprogramming (Figures S5E and S5F), which is consistent with the role of these two sites in condensate formation (Figure S4E).

To independently determine the role of Hdac4 condensate dissociation in reprogramming, we screened and tested various Hdac4 inhibitors in condensate formation and cardiac reprogramming. TMP269 and MC1568 are selective class II (IIa) HDAC (HDAC II, including Hdac4) inhibitors, and BML210 is a potent HDAC inhibitor that can inhibit Hdac4. Among the Hdac4 inhibitors tested, TMP269, but neither MC1568 nor BML210, dissociated Hdac4 condensates in iCM cells (Figure 5E). In parallel, TMP269 enhanced cardiac reprogramming, whereas MC 1568 and BML210 modestly inhibited it (Figure S5G). Hdac4 inhibitor TMP269 improved MGT-mediated iCM formation by enhancing higher Myh6-GFP and cTnT expression (Figures 5F and 5G). TMP269 also promoted higher calcium transient activities in iCMs derived from neonatal CFs (Figure S5H). Consistent with these observations, TMP269 treatment also promoted cardiac gene activation and downregulation of fibroblast genes at the genome level (Figures 5H and S5I). These results further support the conclusion that PC14-3-3 activation stimulates MGT-mediated iCM formation through Hdac4 condensate disruption.

We finally examined the function of Hdac4 condensates in regulating chromatin and cardiac gene expression. We examined if PC14-3-3 directly regulated Hdac4 at cardiac gene promoters during iCM formation. Cells overexpressing GFPHdac4-NES deletion were used for chromatin IP assays in an iCM cell line using GFP antibodies (Figure S5J). OA or MK2206+DT-061 were used to activate or repress PC14-3-3 activation. Hdac4 was localized to cardiac gene promoters including Myh6, Tnnt2, Actc1, and Myocd, and PC14-3-3 inactivation by MK2206+DT-061 enhanced Hdac4 binding (Figure 5I). In contrast, PC14-3-3 activation with OA reduced Hdac4 binding. We also tested H3K27ac on those promoters as a marker for gene activation and observed that H3K27ac was negatively associated with Hdac4 recruitment (Figure S5K). These results show that Hdac4 binding to the cardiac promoter within the Hdac4 condensates induces a repressive effect for cardiac gene expression and PC14-3-3 activation disrupts the repressive effect.

Together, these data illustrate that without PC14-3-3 activation, Hdac4 inhibitory condensates trap other 14-3-3-binding-motif-carrying proteins inside the Hdac4 condensates and repress gene expression. In contrast, PC14-3-3 activation disrupts the repressive Hdac4 condensates and releases transcription/reprogramming factors, which may work synergistically to induce cardiac gene expression for cardiac reprogramming.

DISCUSSION

The key discovery of this research is the identification of PC14-3-3 and its critical role in cardiac reprogramming. Inspired by the histone code concept, PC14-3-3 is defined as the coordinated phosphorylation modification at the 14-3-3 binding motifs within 14-3-3 binding proteins. The major characteristics of the PC14-3-3 code are as follows: (1) the code is embedded within the 14-3-3 binding motifs in a large array of diverse proteins, which include TFs, chromatin modifiers, and various pathway factors. The key factors carrying the PC14-3-3 that play a critical role in cardiac reprogramming are Hdac4, Mef2c, Foxo1, and Nrip. (2) The code is recognized and engaged by the reader 14-3-3 proteins. (3) The code is added or removed by multiple kinases and phosphatases, with Akt1 and PP2A being the major identified writer and eraser in cardiac reprogramming. (4) A key function of the PC14-3-3 code is to induce a global phase change of 14-3-3-motif-embedded proteins like Hdac4, Mef2c, Nrip1, and Foxo1 from a condensate phase to a free phase, that is, the activation of PC14-3-3 disrupts Hdac4 condensates and releases Mef2c, Foxo1, and Nrip1 for proper cardiac gene expression.

In this study, we identify Akt1 kinase and PP2A phosphatase as key writers and erasers of the PC14-3-3 code, respectively, with 14-3-3 proteins as readers. Among the initial seven PC14-3-3-containing proteins affecting reprogramming, five are regulated by OA or PP2A and two are regulated by Akt, and 6/7 are regulated by either PP2A or Akt1. Moreover, PP2A inhibitor OA and Akt1 kinase synergistically lead to PC14-3-3 activation, drastically enhance iCM reprogramming, and render Mef2c and Gata4 dispensable for reprogramming. In contrast, shRNA knockdown of readers 14-3-3ε (YWHAE) and 14-3-3ζ (YWHAZ) largely abolishes Akt1- and OA-stimulated reprogramming. Furthermore, PP2A spatially co-localizes with PC14-3-3-regulated Hdac4 condensates. Nevertheless, aside from Akt1 and PP2A, we expect that there will be numerous writers and erases of the PC14-3-3 code, similar to those of the histone code.²⁷

An important discovery is that PC14-3-3 very likely regulates cardiac reprogramming at a sub-organelle level (condensates) via spatial protein interactions. Our investigation identifies the presence of Hdac4 condensates and their dynamic changes during cardiac reprogramming. PC14-3-3 activation decreases the size and number of Hdac4 condensates, whereas PC14-3-3 inhibition increases them. Hdac4 14-3-3 binding motifs are localized in the adjacent condensate regulation IDR. Mutations in the Hdac4 14-3-3 binding motifs greatly affect Hdac4 condensate dynamics. Our studies reveal that Hdac4 inhibits transcription and Hdac4 condensates are associated with transcriptionally inactive chromatin areas. Importantly, mutations in Hdac4 14-3-3 binding motifs show strong correlation between Hdac4 condensate formation and inhibition of cardiac reprogramming. Inducing PC14-3-3 by 14-3-3 permanent binding forms of Hdac4 also further shows strong correlation between inhibition of Hdac4 condensate formation and cardiac reprogramming. Finally, we have identified that among Hdac4 inhibitors, TMP269, but neither MC1568 nor BML210, dissociates Hdac4 condensates. In parallel, TMP269 enhances cardiac reprogramming, whereas MC 1568 and BML210 inhibit it. These results suggest that Hdac4 condensate dissociation is responsible for PC14-3-3-stimulated reprogramming.

Interestingly, numerous factors, such as Mef2c, Nrip1, and Foxo1, co-localize within the Hdac4 condensates. This finding is consistent with the notion that Hdac4/5 can physically interact with Mef2c,^50,51 Nrip1,⁵² and Foxo1.^53,54 Also, PC14-3-3 appears in the adjacent condensate regulation IDR of those factors. Our studies imply that PC14-3-3-code-embedded proteins may be involved in a distinct layer for chromatin and gene regulation. PC14-3-3-code-embedded proteins spatially present in the same biomolecular condensates could provide rationality why seemingly discrete modifications or signals can have a unified function in cell fate change such as cardiac reprogramming. Our study indicates that PC14-3-3 activation induces a global phase change of 14-3-3-motif-embedded proteins like Hdac4, Mef2c, Nrip1, and Foxo1, and by regulating the PC1-3-3 activation of different proteins, it likely induces a multilevel regulation of this condensate and its function. PC14-3-3 activation serves as a general mechanism to regulate the composition or composition activation in the condensates. It is also worth noting the presence of numerous TF and regulators within the PC14-3-3 condensates. The disruption of Hdac4 condensates may not only diminish the Hdac4 activity but may also increase the activities of Mef2c, Nrip1, and Foxo1. Moreover, the presence of Mef2c in the condensates also explains at least partially why PC14-3-3 activation replaces Mef2c and Gata4 overexpression for cardiac reprogramming. Overall, the identification of the composition and dynamics of Hdac4 condensates during cardiac reprogramming and their regulation by PC14-3-3 has provided mechanistic insights into cardiac reprogramming.

This study implies that coding systems based on motifs and PTM modifications, such as PC14-3-3, very likely serve as a general mechanism in the biological system. We predict that a PTM code has the following characteristics: (1) the coding information depends on motif sequences and PTM modifications. (2) This type of code possesses a collaborative function for a specific biological process by connecting the diversity of its embedded proteins. (3) Naturally, the code’s collaborative function is executed at multiple levels but is very likely spatiotemporally connected. In this article, we identify that PC14-3-3 function depends on the phosphorylation of 14-3-3 binding motifs, PC14-3-3 activation plays a key role in cardiac reprogramming, and its function is executed by disrupting Hdac4 condensates, which connect Hdac4, numerous TFs, and possibly chromatin. These studies imply that PC14-3-3 activation represents a general mechanism in cell fate changes beyond CM reprogramming. In support of this notion, the SARS2 nucleoprotein serves as a protein that regulates liquid-liquid phase separation, interacting with 14-3-3 in a phosphorylation-dependent manner.^55,56 We are optimistic that PTM code studies within free proteins will provide insights into biological processes, particularly cell fate establishment and cell fate change, which may reveal targets for organ disease and regenerative medicine.

Limitations of the study

We recognize that our studies have not encompassed all 14-3-3 binding motifs (PC14-3-3), a limitation inherent to the scope of our research. Consequently, future investigations into PC14-3-3 will incorporate more comprehensive statistical and computational methodologies, including computational identification of 14-3-3 binding motifs and prediction of condensation proteins. Additionally, we acknowledge that while we have established a connection between PC14-3-3 and condensation in the cardiac reprogramming system, further research is required to determine if this relationship extends to other systems.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Zhong Wang (zhongw@med.umich.edu).

Materials availability

No materials have been generated in this study.

Data and code availability

• All raw RNA sequencing data from this study have been submitted to the NCBI GEO database under accession number GSE255806. All other data reported in this paper will be shared by the lead contact upon request.

• No original code was generated for this study. All code used for analysis was properly cited in the STAR Methods.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Mouse lines

The α-MHC-GFP transgenic mice were used to derive MEFs and NCFs.^58-60 αMHC-Cre/Rosa26A-Flox-Stop-Flox-GCaMP3 mice was purchased from Jackson Laboratories (011038, 014538). MEFs and NCFs isolation were performed with 4–6 months old female mice. All animal related procedures were approved by the Institutional Animal Care and Use Committee of the University of Michigan and are consistent with the National Institutes of Health Guide for Use and Care of Animals.

METHOD DETAILS

Plasmids

The pMXs based retroviral polycistronic vector encoding Mef2c, Gata4, Tbx5 were provided by Dr. Li Qian’s lab.⁵⁷ Mutation constructions of cardiac genes Mef2c, Gata4, Tbx5, Hand2 and Nkx2-5 were constructed by overlap PCR strategy. The main primers used for the cloning were listed in Table S5. Most of the genes related with 14-3-3 were acquired from Addgene and cloned into pMX plasmids. The plasmids purchase from Addgene what listed in Table S4. The 14-3-3 permanent bind forms of Hdac4 fragments of Hdac4-3R18 were synthesized by Genscript. The shRNA lentivirus vector for 14-3-3 isoforms, Ppp2cb were from Vector Core of University of Michigan. The R18 sequence used to replace 14-3-3 binding motif in Hdac4 are as follows: “PHCVPRDLSWLDLEANMCLP”.

Inducible fibroblast cell line construction

The inducible fibroblast cell line construction was generated from α-MHC-GFP transgenic mice.⁶¹ All the Lentiviral Transfer Plasmids was revised from pTripz backbone. The polycistronic MGT construct under the control of a tetracycline responsive promoter for temporal control of MGT reprogramming factor expression and constitutive expression neomycin resistance gene (neo) was cloned. The second plasmid could express transactivator (reverse tetracycline-controlled transactivator) to induce tetracycline responsive promoter activation and constitutive expression puromycin resistance gene (puro) expression plasmids. The third plasmid was revised from retrovirus pMX to express Hdac4 and hygromycin resistance gene. Targeted plasmids was packaged into lentivirus or retrovirus, selected with G418 (750 μg/mL, invitrogen), puromycin (4ug/ml, invitrogen) or hygromycin B (1000 μg/mL, invitrogen) to generate MEF cell line (icMEF) that can be reprogrammed simply by the addition of doxycycline to the culture media.

Inhibitor information

Okadaic Acid (10011490), TMP269 (17738), MC 1568 (16265), BML-210 (10005019), MK-2206(11593) were purchased from Cayman. DT-061 (S8774) was purchased from SELLECK CHEMICALS. 14-3-3 Antagonist I, 2–5 (100081) was purchased from Sigma.

Retrovirus and lentivirus preparation

75% confluent 10 cm dish of Platinum-E (Plat-E) Retroviral Packaging Cell Line (cell biolabs) were transfected with totally 10 μg retrovirus vectors using Lipofectamine 2000 (Thermo Fisher Scientific) in 1.5 mL Opti-MEM (Thermo Fisher Scientific). After 24 hours, the medium was changed with 10 mL fresh DMEM medium with 10% FBS. Viral medium was collected twice 48 hours and 72 hours after transfection. The Viral medium was filtered through a 0.45-mmcellulose filter and was added 1/5 vol of 40% PEG8000 solution. The mixture was kept at 4 °C overnight and spun at 3000 g, 4 °C, 30 minutes to concentrate. The virus was resuspended by fresh MEFs medium with 8 μg/mL polybrene (Sigma).

Similarly, 10 μg lentiviral vectors with 6 μg psPAX2 and 4 μg pMD2.G were packaged into a 60% confluent 10 cm plate of HEK293T cells (ATCC) with 0.75 mL Opti-MEM by using Lipofectamine 2000 (Invitrogen) with another 0.75 mL Opti-MEM with DNA mixture. 4–6 h later, Opti-MEM was changed by 10 mL fresh MEFs medium. Then, the virus was collected.

Direct reprogramming of fibroblasts to iCMs

Preparation of MEFs (isolated at E13.5) was described here.⁵⁹ Briefly, embryos were harvested from transgenic mice of α-MHC-GFP at 13.5 days postcoitum followed by decapitation and removal of internal organs, including the heart. The tissue was minced and digested with TrypLE Express Enzyme (Thermo Fisher Scientific, Waltham, MA, USA). Cells were resuspended in MEFs medium (10% FBS and 2 mM L-glutamine contained DMEM medium) and plated onto one 10 cm dish per embryo. Cells were regularly passaged at 1:3 (passage 1). Passage 3 MEFs were used for reprogramming.

NCFs were isolated from P2-P3 α-MHC-GFP transgenic or αMHC-Cre/Rosa26A-Flox-Stop-Flox-GCaMP3 mice.^58,60 Briefly, heart tissue was isolated, minced, and digested with 0.05% trypsin-EDTA. Then NCFs were collected with type II collagenase (0.5 mg/mL) in Hanks’ Balanced salt solution. After washing and resuspending in Dulbecco’s PBS with 2.5 g BSA and 0.5 m EDTA, cells were incubated with CD90.2 microbeads (Miltenyi Biotec, MidiMACS Starting Kit) at 4°C for 30min. Positive cells were isolated by magnetic-activated cell sorting and plated onto a 10-cm dish with FB medium (Iscove’s modified Dulbecco’s medium with 20% FBS and 1% penicillin/streptomycin) for future use. Isolated fibroblasts were routinely examined under fluorescence microscope or with GFP immunostaining to determine potential CM contamination.

The optimized protocol of direct cardiac reprogramming was described here.⁶² Briefly, fresh fibroblasts were seeded on tissue culture dishes at a density of 10,000 cells/cm2 before virus infection. Fibroblasts were infected with fresh viral mixture containing 8 μg/mL polybrene (Sigma) 24 h after seeding. Twenty-four hours later, the viral medium was replaced with induction medium composed of DMEM/199 (4:1) (Gibco, Thermo Fisher Scientific) containing 10% FBS, 1% penicillin/streptomycin and 10 μL/mL GlutaMAX (Gibco, Thermo Fisher Scientific). Medium was changed every 2–3 days with or without indicated chemicals for two weeks before cells were examined. 1 μg/mL puromycin (SIGP8833-25MG, Sigma) was added into the medium 3 days after infection to eliminate cells without infection.

For spontaneous beating and calcium transient assessment experiments, induction medium was replaced every 2–3 days by mature medium containing StemPro-34 SF medium (Gibco, Thermo Fisher Scientific), GlutaMAX (10 μL/mL, Gibco, Thermo Fisher Scientific), ascorbic acid (50 μg/mL, Sigma), recombinant human VEGF165 (5 ng/mL, R&D Systems), recombinant human FGF10 (25 ng/mL, R&D Systems), and recombinant human FGF basic146 aa (10 ng/mL, R&D Systems) for another 2 weeks.⁶³

Western blot

Proteins were extracted from cells by adding lysis buffer and centrifuged at 4°C for 15 min at 12,000 rpm. The membranes were blocked with 4% BSA for 1 h at room temperature and then incubated with primary antibodies over night at 4°C. After 3 washings with TBST, the membranes were incubated with appropriate secondary antibody in TBST solution for another 1 h at room temperature. After 3 washings, the membranes were scanned and quantified by Odyssey CLx Imaging System (LI-COR Biosciences, USA).

Immunocytochemistry

Anti-Hdac4 + 5 + 9 antibody (Abcam, ab131524) is the antibody used to detect the endogenous condensate formation. This antibody directed target the 14-3-3 bind motif of Hdac4 around phosphorylated site S246. Flag antibody (Sigma F1804), H3K27ac (Abcam ab245911), V5 (Abcam ab9116), Mef2c (CST 5030), HA (ab9110), cTnT (Thermo Fisher Scientific), and GFP (Thermo Fisher Scientific) Antibodies were used. Cells were first fixed with 4% formaldehyde for 15 min. Then they were permeabilized with 0.1% Triton X-100 in PBS for another 15 min at room temperature. Cells were blocked with 4% horse serum in PBS for 1 h and then incubated with primary antibodies overnight at 4°C followed by incubation with appropriate fluorogenic secondary antibodies (Thermo Fisher Scientific) at room temperature for 1 h. The images were captured using the Keyence All-in-One Fluorescence Microscope BZ-X810. cTnT and α-actinin positive cells were manually quantified by single-blind method from randomly HPFs within each well.⁶⁴

Quantitative real-time PCR (qPCR)

Total RNAs from all cells were extracted using Trizol Reagent (Thermo Fisher Scientific) following the manufacturer’s instructions. RNA integrity was determined using formaldehyde denaturalization agarose gel electrophoresis. RNA concentrations were measured with Nanodrop spectrophotometer (Thermo Fisher Scientific). RNA was reverse transcribed by using iScript cDNA Synthesis Kit (BioRad). qPCR was performed using StepOne Real-Time PCR System (Thermo Fisher Scientific). Primer oligonucleotides were synthesized by Sigma and are listed in Table S5.

Spontaneous beating and calcium transient assessment

Spontaneous beating assessment was performed by light microscopy at room temperature after indicated treatment of MEFs or NCFs at several time points. The beating cell number was manually quantified by single-blind method in each well of 48-well plate. αMHC-Cre/Rosa26A- Flox-Stop-Flox-Gcamp3 NCFs were used for calcium transient assessment that was performed by fluorescence microscopy at room temperature after indicated reprogramming treatments. Three HPFs of view were randomly selected within each well.

RNA-sequencing

Total RNAs from all groups of cells were isolated using TRIzol following the provider’s instructions. RNA (RIN >8.5) was used for RNA-seq library preparation by using NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (E7760S). The libraries were sequenced using Hiseq 4000 by the University of Michigan Sequencing Core. The quantification of RNA expression was estimated by Kallisto.⁴⁶ The transcript abundance was imported using tximport followed by DESeq2 analysis. The abundance of genes was used to calculate fold change and p values. Differentially expressed genes were identified as a fold change greater than 2 and padj values less than 0.05.

Flow cytometry

Fluorescence flow cytometry data were from 10,000 single-cell events which were collected using a standard MoFloAstrios flow cytometer (Immunocytometry Systems; BD Biosciences). Data were analyzed using Summit (BD Biosciences).

Condensation and optoIDR assay

The condensate regulation of IDRs of Hdac4, MEF2C and Foxo1 were predicted by dSCOPE, which is a web server developed for detecting sequences critical for phase separation related proteins.^65,66 Predicted IDRs on Hdac4 were deleted respectively, and each truncated Hdac4 constructs were inserted into mammalian expression vector pMX. Each pMX plasmid was transformed into HEK293T cells to test the condensate formation of each truncated Hdac4 construct.

The photoactivatable, self-associating Cry2 protein was labeled with mCherry and fused to the intrinsically disordered regions (IDR) Hdac4, MEF2C and Foxo1. The fusion of the IDR of Hdac4, MEF2C and Foxo1 to Cry2-mCherry were tested for the rapid formation of micron-sized spherical droplets upon blue light stimulation.⁴⁴ WT and mutated fusion protein were compared for their condensate formation ability by blue light stimulation within 10 min. The assay is performed using Keyence All-in-One Fluorescence Microscope BZ-X810.

QUANTIFICATION AND STATISTICAL ANALYSIS

Results were presented as mean ± sd. Statistical difference between groups was analyzed by one-way ANOVA followed by the Student–Newman–Keuls multiple comparisons tests. A P- value < 0.05 was regarded as significant. All experiments were conducted with a minimum of biological triplicates, and the corresponding sample sizes are indicated in the figure legends.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-Hdac4 + 5 + 9	Abcam	Cat# ab131524, RRID:AB_11156240
flag	Sigma	Cat# F1804, RRID:AB_262044
H3K27ac	Abcam	Cat# ab245911
V5	Abcam	Cat# ab9116, RRID:AB_307024
Mef2c	CST	Cat# 5030, RRID: AB_10548759
HA	Abcam	Cat# ab9110, RRID:AB_307019
cTnT	Thermo Fisher Scientific	Cat# MS295P, RRID:AB_61806
GFP	Thermo Fisher Scientific	Cat# a11122, RRID:AB_221569
Biological samples
pMX-MGT-puro	(Wang et al.)57	N/A
See Table S4 for a list of plasmids from Addgene.	This paper	N/A
See Table S4 for a list of plasmids from Sigma.	This paper	N/A
14-3-3 permanent bind forms of Hdac4 fragments of Hdac4-3R18	Genscript	this paper
Chemicals, peptides, and recombinant proteins
Okadaic Acid	Cayman	Cat# 10011490
TMP269	Cayman	Cat# 17738
MC 1568	Cayman	Cat# 16265
BML-210	Cayman	Cat# 10005019
MK-2206	Cayman	Cat# 11593
DT-061	SELLECK CHEMICALS	Cat# S8774
14-3-3 Antagonist I, 2-5	Sigma	Cat# 100081
Deposited data
Bulk RNA-seq data	This paper	NCBI GEO: GSE255806
Experimental models: Cell lines
inducible fibroblast cell line for cardiac reprogramming (icMEF)	this paper	this paper
Platinum-E (Plat-E) Retroviral Packaging Cell Line	cell biolabs	RV-101
Experimental models: Organisms/strains
α-MHC-GFP transgenic mice	(Qian et al.)2	N/A
αMHC-Cre transgenic mice	Jackson Laboratories	Cat# 011038
Rosa26A-Flox-Stop-Flox-GCaMP3	Jackson Laboratories	Cat# 014538
Oligonucleotides
See Table S5 for a list of oligonucleotides.	This paper	N/A

Highlights.

• A phosphorylation code in 14-3-3 binding motifs (PC14-3-3) is identified

• PC14-3-3 activation via writer Akt1 and eraser PP2A stimulates cardiac reprogramming

• PC14-3-3 code activation replaces Mef2c and Gata4 in cardiac reprogramming

• PC14-3-3 activation disrupts Hdac4 condensates to stimulate cardiac reprogramming