Author contributions: C.H.: conceived the project and analyzed the data; performed all the experiments with the help of J.W. and W.L.; H.G.: performed all the experiments with the help of J.W. and W.L.; Y.M. and M.W.: conceived the project and analyzed the data; wrote the manuscript with the help of C.H. All the authors discussed the results and commented on the manuscript. C.H. and H.G. contributed equally to this work.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS January 25, 2013.
Increasing evidence suggests that metabolic remodeling plays an important role in the regulation of somatic cell reprogramming. Threonine catabolism mediated by L-threonine dehydrogenase (TDH) has been recognized as a specific metabolic trait of mouse embryonic stem cells. However, it remains unknown whether TDH-mediated threonine catabolism could regulate reprogramming. Here, we report TDH as a novel regulator of somatic cell reprogramming. Knockdown of TDH inhibits, whereas induction of TDH enhances reprogramming efficiency. Moreover, microRNA-9 post-transcriptionally regulates the expression of TDH and thereby inhibits reprogramming efficiency. Furthermore, protein arginine methyltransferase (PRMT5) interacts with TDH and mediates its post-translational arginine methylation. PRMT5 appears to regulate TDH enzyme activity through both methyltransferase-dependent and -independent mechanisms. Functionally, TDH-facilitated reprogramming efficiency is further enhanced by PRMT5. These results suggest that TDH-mediated threonine catabolism controls somatic cell reprogramming and indicate the importance of post-transcriptional and post-translational regulation of TDH. STEM CELLS 2013;31:953–965
The reprogramming of somatic cells into induced pluripotent stem (iPS) cells by the defined transcriptional factors has provided great potential for regenerative medicine and biomedical research [1–4]. However, the reprogramming efficiency is extremely low. Over the years, tremendous efforts have been directed toward understanding the underlying mechanisms of reprogramming process and improving reprogramming efficiency. This has led to the identification of dozens of factors with abilities to enhance reprogramming efficiency, including transcriptional factors, miRNAs, inhibitors of specific signaling pathways, and chemical compounds [5–8]. However, the detailed mechanisms whereby somatic cells are reprogrammed into iPS cells still remain obscure.
Metabolism is either directly or indirectly involved in almost every aspect of cellular functions. Increasing evidence suggests that somatic cells and pluripotent stem cells exhibit different metabolic profiles. For example, it has been recognized that differentiated somatic cells mainly use mitochondrial oxidative phosphorylation for their energy production, whereas pluripotent stem cells rely on glycolysis [9–11]. Therefore, generation of iPS cells appears to involve metabolic reprogramming from mitochondrial oxidative phosphorylation to glycolysis. In supporting of this idea, it has been shown that stimulation of glycolysis enhances, whereas inhibition of glycolysis reduces reprogramming efficiency [12–14]. In addition, hypoxic environments, which have been linked to promoting glycolysis through transactivtion of various glycolytic enzyme genes, improve reprogramming efficiency . Furthermore, the reprogramming factor c-Myc is well documented for its function in promoting glycolysis and enhancing reprogramming efficiency. These combined findings strongly suggest that metabolic reconstruction plays an important role in somatic cell reprogramming.
Intriguingly, by evaluating the common metabolites in embryonic stem (ES) cells and differentiated embryoid bodies, it has been revealed that mouse ES cells exist in an unusual state associated with one carbon metabolism . Further analyses suggest that specific expression of L-threonine dehydrogenase (TDH) is responsible for this unique metabolic state of ES cells. TDH is recognized as the first and rate-limiting mitochondrial enzyme that hydrolyzes threonine into glycine and acetyl-CoA, with glycine facilitating one-carbon metabolism and acetyl-CoA feeding the tricarboxylic acid (TCA) cycle. Plenty of TDH facilitates consumption of threonine as a metabolic fuel to support rapid growth of mouse ES cells [16, 17]. Therefore, mouse ES cells are critically dependent on one amino acid-threonine, but not other amino acids . Furthermore, inhibition of TDH enzyme activity by the use of its inhibitors leads to death of mouse ES cells . These findings demonstrate the critical role of TDH in maintaining mouse ES cell growth. However, it still remains unknown whether TDH-mediated threonine catabolism is involved in the regulation of somatic cell reprogramming, and if this is the case, how is TDH regulated?
Here, we report TDH as a novel positive regulator of somatic cell reprogramming. Ectopic expression of TDH markedly enhances reprogramming efficiency, whereas TDH knockdown greatly inhibits it. In addition, miR-9 is shown to post-transcriptionally regulate tdh gene expression, by which miR-9 exerts its inhibitory effect on reprogramming. We also demonstrate that protein arginine methyltransferase 5 (PRMT5) directly binds to TDH. Via this association, PRMT5 mediates specific arginine (R180) methylation of TDH and regulates its enzyme activity. Functionally, PRTM5 enhances the promoting effect of TDH on reprogramming. Taken together, these results provide evidences to demonstrate the important role of TDH-mediated threonine catabolism in promoting somatic cell reprogramming, and uncover the molecular mechanisms by which TDH is intricately regulated both at post-transcriptional and post-translational levels.
MATERIALS AND METHODS
Cell Culture, Viral production, and Induction of Pluripotent Stem Cells
Oct4-EGFP MEF cells were derived from the mouse strain B6; 129S4-Pou5f1tm1Jae/J (Jackson Laboratory, stock no. 008204, Bar Harbor, Maine). MEF cells were isolated from E13.5 embryos and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 100 μg/ml streptomycin. The isolated MEF cells in early passages (up to passage 2) were used for mouse iPS cells generation.
Induction of pluripotent stem cells from Oct4-EGFP MEF cells was carried out as described previously . Briefly, plat-E packaging cells were transfected with pMXs plasmids containing mouse Oct4, Sox2, Klf4, c-Myc, or TDH. Viral supernatants were collected 36 and 60 hours post-transfection, and two rounds of infection were performed. The culture medium containing retrovirus particles was filtered through a 0.45 μm polyvinylidene difluoride filter (Millipore, Billerica, Massachusetts). Oct4-EGFP MEF cells were seeded at a density of 30,000 cells per well in a six-well plate 1-day before infection. Equal amounts of supernatants containing each of the four retroviruses were mixed and incubated with MEF cells supplemented with 4 μg/ml polybrene (Sigma, St.louis, Missour) for 24 hours. Three or four days after transduction, MEF cells were harvested and plated on the mitomycin-C-treated MEF feeder cells, and the culture medium was changed to mouse embryonic stem cell medium. GFP-positive clones were finally counted under a fluorescence microscope.
Reagents and Antibodies
The following reagents used for this study were purchased from the indicated sources: antibodies against glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (Santa Cruz, California; 1:1,000), actin (Cell Signaling, Danvers and Beverly, Massachusetts; 1:1000), TDH (Millipore, 1:1,000), PRMT5 (Santa Cruz, 1:1,000; Cell Signaling, 1:1,000), cytochrome c (Santa Cruz, 1:1,000), GFP (MBL 1:4,000), Flag (Sigma 1:4,000); AP-conjugated secondary antibodies against mouse and rabbit IgG (Promega, Madison,Wisconsin); 3H-SAM (PE, NET155); threonine (Sigma, T8441); and 3-HNV (sigma, H4002).
Primers used to generate plasmids encoding full-length TDH, full-length PRMT5 and their truncated deletion mutants were listed in supporting information Table 1. The TDH cDNA and PRMT5 cDNA were amplified by RT-PCR using total RNA from R1-ES cells.
Dual-Luciferase Reporter Assay
TDH-3′UTR and TDH-3′UTR-mut were constructed into pSI-CHECK2-report plasmid (Promega). The indicated plasmids and miR-9 mimics were cotransfected into R1-ES cells using oligofectamine (Invitrogen, Carlsbad, California). Firefly and Renilla luciferase activities were analyzed by Dual-Luciferase Reporter Assay system according to the manufacturer's instruction (Promega). The relative luciferase activities were calculated by normalizing the firely luciferase activity to Renilla luciferase activity. The represented data were mean ± SD from at least three independent experiments.
Introduction of miRNA Mimics and Inhibitors
Mimics and inhibitors of mmu-miRNAs were synthesized by Genepharma Company (Shanghai, People's Republic of China). The sequences for miRNA mimics and inhibitors were as follows: miR-9 mimics, sense 5′-UCUUUGGUUAUCUAGCUGUAUGA-3′ and antisense 5′-AUACAGCUAGAUAACCAAAGAUU-3′; miR-31 mimics, sense 5′-AGGCAAGAUGCUGGCAUAGCUG-3′ and antisense 5′-UUUCCGUUCUACGACCGUAUCG-3′; miR-149 mimics, sense 5′-UCUGGCUCCGUGUCUUCACUCCC-3′ and antisense 5′-UUAGACCGAGGCACAGAAGUGAG-3′; miR-218 mimics, sense 5′-UUGUGCUUGAUCUAACCAUGU-3′; and antisense 5′-UUAACACGAACUAGAUUGGUA-3′; mimics control, sense 5′-UUCUCCGAACGUGUCACGUTT-3′ and antisense 5′-TTAAGAGGCUUGCACAGUGCA-3′; and miR-9 inhibitors, 5′-AGAAACCAAUAGAUCGACAUACU-3′.
For each transfection in a six-well plate, 100 nM miRNA mimics or inhibitors were used. Transfection of MEF cells by oligofectamine (Invitrogen) was performed according to the manufacturer's instruction.
Total RNA was isolated using Trizol (Ambion). One microgram of total RNA was used to synthesize cDNA using PrimeScriptTM RT reagent kit (Takara, DRR037A) according to the manufacturer's instruction. The reverse transcription primers were as follows: U6, 5′-CGCTTCACGAATTTGCGTGTCAT-3′; mmu-miRNA-9, 5′-GTCGTATCCAGTGCGTGTCGTGGAGTCGGCAATTGCACTGGATACGACTCATACAGCT-3′; mmu-miR-31, 5′-GTCGTATCCAGTGCGTGTCGTGGAGTCGGCAATTGCACTGGATACGACCAGCTATG-3′; mmu-miR-149, 5′-GTCGTATCC AGTGCGTGTCGTGGAGTCGGCAATTGCACTGGATACGACGGGAGTGAA-3′; mmu-miR-218, 5′-GTCGTATCCAGT GCGTGTCGTGGAGTCGGCAATTGCACTGGATACGACACATGGT-3′.
Real-time PCR was performed using SYBR premix EX Taq (TaKaRa) and ROX and analyzed with Stratagene Mx3000p (Agilent Technologies). Real-time PCR primer sequences were as follows: mmu-miRNA-9, 5′-GCCTCTTTGGTTATCTAG-3′ and 5′-CAGTGCGTGTCGTGGAGT-3; mmu-miR-31, 5′-GCCTGCAAGATGCTGG-3′ and 5′-CAGTGCGTGTCGTGGAGT-3′; mmu-miR-149, 5′-GCCTCTGGCTCCGTGTCT-3′ and 5′-CAGTGCGTGTCGTGGAGT-3′; mmu-miR-218, 5′-GCCTTGTGCTTGATCT-3′ and 5′-CAGTGCGTGTCGTGGAGT-3′; TDH, 5′-CCTGGAGGAGGAACAACTGACTA-3′ and 5′-ACTCGAATGTGCCGTTCTTTG-3′; actin, 5′-GACCTGACTGACTACCTCATGAAGAT-3′ and 5-'GTCACACTTCATGATGGAGTTGAAGG-3′. The signal intensities of miRNA and TDH mRNA were normalized to intensities of U6 snRNA and actin, respectively. The data were represented as mean ± SD from three independent experiments.
Western Blot Analysis and Coimmunoprecipitation
Western blot analysis was performed as previously described . For coimmunoprecipitation, cells were lysed in IP lysis buffer (0.5% NP-40, 150 mM NaCl, 20 mM HEPES, pH 7.5, 2 mM EDTA, 1.5 mM MgCl2) supplemented with protease inhibitor cocktail for 1 hour on ice. Cell lysates were incubated with Protein A/G-Sepharose beads coated with the indicated antibodies at 4°C for 4 hours. The immunoprecipitates were subjected to Western blot analysis.
iPS cells were seeded on 0.1% gelatin-coated plate. Three days after seeding, cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton-X-100 in phosphate buffered saline (PBS), and blocked with 1% Bovine serum albumin in PBS. Cells were then incubated at room temperature with anti-Oct4 (Santa Cruz, sc5279), anti-Sox2 (Millipore, ab5603), anti-Ssea1 (Santa Cruz, sc21702), or anti-Nanog (Millipore, ab5731) antibody for 2 hours and with rhodamine-conjugated anti-rabbit IgG secondary antibody for an additional 1 hour. The cells were washed twice with PBS containing 0.1% Tween 20 and stained with Hoechst 33342 (Sigma). Images were captured with an Olympus fluorescence microscope.
To generate lentiviruses expressing TDH, PRMT5, or control shRNAs, HEK 293T cells grown on a 6-cm dish were transfected with 2 μg of TDH, PRMT5 shRNAs (cloned in PLKO.1) or control vector, 2 μg of pREV, 2 μg of pGag/Pol/PRE, and 1 μg of pVSVG. Twenty-four hours after transfection, cells were cultured with DMEM containing 20% FBS for an additional 24 hours. The culture medium containing lentivirus particles was centrifugated at 1,000 × g for 5 minutes. Viruses in the supernatant were used for infection. The knockdown efficiency was evaluated by Western blot analysis and real-time RT-PCR analysis. The shRNA sequence targeting TDH is 5′-GCCTTTGTCCTGTGCCTAAAT-3′. PRMT5 targeting shRNA sequences are 5′-CTCTTGTGAATGCGTCTCTT-3′ (PRMT5-1), and 5′-AGCTCTGAGTTCTCTTCCTA-3′ (PRMT5-2).
Measurement of TDH Enzyme Activity
TDH enzyme activity was measured similarly as previously described . To extract mitochondrial protein, cells were homogenized in the buffer containing 10 mM Tris-HCl (pH 7.5), 250 mM sucrose, and 2 mM EDTA on ice with a glass pestle douncer. Mitochondria were isolated from homogenates through a two-step differential centrifugation procedure at 600 × g and 11,000 × g for 10 minutes each at 4°C. Mitochondrial protein was then extracted in 100 mM potassium phosphate buffer (pH 7.4) containing 0.1% NP-40, 10 mM dithiothreitol (DTT) and 1× protease inhibitor cocktail. TDH enzyme activity was determined by measuring the rate of formation of NADH at 25°C. The assay mixture contained 100 mM Tris-HCl (pH 8.0), 25 mM L-threonine, 5 mM NAD+, and 25 mM NaCl. The reaction was initiated by the addition of mitochondrial protein extract, and absorbance of the reaction mixture at 340 nM was recorded continuously on a spectrophotometer (Beckman). Protein concentrations were determined by a BCA assay kit (Pierce), and enzyme activities were normalized by protein concentrations.
TDH (34-373) was expressed as a GST fusion protein in Escherichia coli strain Rosetta. The cells were cultured at 37°C until the A600 nm reached 0.6 and were then induced with 0.2 mM isopropyl β-D-thiogalactoside (Promega) for 16 hours at 20°C. The cells were suspended in 50 mM Tris-HCl (pH 8.0) containing 50 mM NaCl, 1 mM DTT (Promega), and 1 mg/ml lysozyme, incubated on ice for 30 minutes, and sonicated. After spinning at 13,000 × g for 15 minutes at 4°C, the supernatant was incubated with glutathione Sepharose beads for 2 hours. The beads-bounded proteins were eluted using the elution buffer (50 mM Tris-HCl and 20 mM glutathione). In this study, GST-TDH (34-373) was used instead of full length TDH (1-373). The reason is that GST tagged full-length TDH protein was expressed in the presence of inclusion bodies in E. coli cells.
To purify recombinant PRMT5 protein, Flag-PRMT5 was transfected into HEK 293T cells. Twenty-four hours after transfection, cell lysates were immunoprecipitated with anti-Flag M2 affinity beads (Sigma). The beads were washed extensively and proteins were eluted with 3× Flag peptide (Sigma). The eluted protein concentration was determined by a BCA assay kit.
In Vitro and In Vivo Methylation Assays
To perform an in vitro methylation assay, 1 μg of Flag-PRMT5 protein was mixed with 5 μg of GST-TDH protein in the presence of 3H-SAM (2.75 μCi). The mixture was incubated at room temperature for 1 hour, stopped by adding SDS sample buffer, and boiled for 5 minutes at 100°C. The mixture was then resolved by SDS-PAGE, followed by Coomassie Brilliant Blue (CBB) staining. 3H-methyl-TDH was detected by autoradiography after 50 days exposure. For in vivo methylation assay, Flag-TDH, Flag-TDH/R180K, Flag-TDH/R190K, or Flag-TDH/R192K was transfected into HEK 293T cells together with Flag-PRMT5. Twenty-four hours after transfection, HEK 293T cells were fed with 3H-SAM (0.55 μCi/ml) for 1 hour at 37°C. Cell lysates were immunoprecipitated with anti-Flag M2 affinity beads. The immunoprecipitates were subjected to Western blot analysis. 3H-methyl-TDH was detected by autoradiography after 50 days exposure.
The iPS cells were harvested by collagenase IV treatment, centrifuged, and the cell pellets were suspended in PBS. We injected 200 μl of the cell suspension (1 × 106) into the dorsal flank of nude mice. Two months after injection, tumors were dissected and fixed in 4% paraformaldehyde. Paraffin-embedded tissue was sliced and stained with hematoxylin and eosin.
Statistical analysis was carried out using Microsoft Excel software and GraphPad Prism to assess differences between experimental groups. Statistical significance was analyzed by Student's t test and expressed as a p value. p values lower than .05 were considered to be statistical significance.
TDH Improves Reprogramming Efficiency of Mouse Embryonic Fibroblast Cells
TDH is recently reported to be copiously expressed in mouse ES cells. TDH-mediated threonine catabolism is required for the rapid growth of mouse ES cells. To address whether TDH regulates somatic cell reprogramming, we first compared the levels of TDH in mouse embryonic fibroblast (MEF) cells, R1 ES cells, and iPS cells by Western blot and real-time reverse transcriptase-polymerase chain reaction (RT-PCR) analyses. Consistent with the previous report , TDH was highly expressed in ES cells and iPS cells but was expressed at extremely low levels in MEF cells (Fig. 1A, 1B). Moreover, both mRNA levels and the enzyme activity of TDH were increased during Oct4, Sox2, c-Myc, and Klf4 (OSCK)-induced MEF reprogramming (Fig. 1C, 1D), indicating that TDH may facilitate reprogramming. We next introduced TDH into MEF cells, which carry an Oct4-enhanced green fluorescent protein (EGFP) reporter that activates GFP expression upon the induction of pluripotency, along with OSCK transduction. The cells were maintained in the culture medium containing normal concentration of threonine (0.8 mg/ml, Thr−) or high concentration of threonine (2 mg/ml, Thr+). As shown in Figure 1E and 1F, ectopic expression of TDH resulted in an increase in the GFP-positive colonies in reprogramming with OSCK under both normal and high concentration of threonine culture conditions. By contrast, knockdown of TDH inhibited OSCK-induced reprogramming efficiency (Fig. 1G). The efficiency of shRNA-mediated TDH knockdown was confirmed in R1-ES cells by both Western blot and real-time RT-PCR analyses (supporting information Fig. S1). These results demonstrate that TDH promotes reprogramming of MEF cells.
To determine whether TDH-derived iPS cells still retain pluripotency, GFP-positive colonies derived from TDH plus OSCK-transduced MEF cells were expanded and analyzed for markers of pluripotency. Immunofluorescence analysis showed that these cells were positive for mouse ES cell markers, such as Oct4, Sox2, Nanog, and Ssea1 (Fig. 1H). In addition, semiquantitative RT-PCR analysis revealed that all the examined iPS clones expressed ES cell marker genes, including Oct4, Sox2, Nanog, Rex1, Fbx15, and Esg1, at the levels comparable to those in R1 mouse ES cells (Fig. 1I). To further test pluripotency of these iPS cells, multiple iPS clones were separately injected subcutaneously into dorsal flanks of nude mice. Six weeks after injection, the excised teratomas were analyzed. Histological examination showed that the teratomas contained various tissues representing all three germ layers, including intestinal epithelium from endoderm, muscle and skin from mesoderm, and nerve tissue from ectoderm (Fig. 1J). These results suggest that the iPS cells generated via TDH induction are pluripotent.
Enzyme Activity of TDH is Required for Its Promoting Effect on Reprogramming
To determine whether TDH enzyme activity is essential for its promoting effect on reprogramming, we first performed the domain analysis of TDH using NCBI conserved domain prediction software. A potential enzymatic active domain was found in the region spanning amino acids 150-200 of TDH (Fig. 2A). To confirm this prediction, wild-type TDH and its deletion mutants TDHΔ150-200, TDHΔ170-200, and TDHΔ190-200 were assayed for the enzyme activity. As shown in Figure 2B, compared with wild-type TDH, all the examined TDH mutants almost completely lost their enzyme activity. We next evaluated the effects of wild-type TDH and enzymatic inactive TDH mutants TDHΔ150-200, TDHΔ170-200, and TDHΔ190-200 on OSCK-induced reprogramming. As expected, wild-type TDH indeed greatly enhanced the reprogramming efficiency, whereas these enzymatic inactive TDH mutants failed to show any stimulatory effect on OSCK-induced reprogramming (Fig. 2C), indicating that the enzyme activity of TDH is indispensable to enhance the reprogramming efficiency.
TDH is known to convert threonine into two products glycine and acetyl-CoA, with glycine facilitating one-carbon metabolism and acetyl-CoA feeding the TCA cycle. To determine which product of threonine catabolism mediates the effect of TDH on reprogramming, we used 3-hydroxynorvaline (3-HNV), a synthetic variant of threonine containing an extra carbon atom. The TDH can hydrolyze this threonine analog, producing glycine and propionyl-CoA. However, the latter is unable to be used by TCA cycle. MEF cells were transduced with TDH plus OSCK and maintained in the culture medium containing the increasing amounts of 3-HNV. As shown in Figure 2D, 3-HNV inhibited TDH-facilitated reprogramming efficiency in a dose-dependent manner. In addition, when the concentration of 3-HNV was added up to 2 mM, it almost completely reversed the promoting effect of TDH on OSCK-induced reprogramming (Fig. 2E, lanes 1, 3, and 4). These results indicate that acetyl-CoA produced from TDH-mediated threonine catabolism is important for TDH in promoting reprogramming. We next evaluated the effect of glycine on OSCK-induced reprogramming. The culture medium was supplemented with 4 mM glycine, instead of the normal concentration of 0.4 mM, for maintaining OSCK-transduced MEF cells. We did not observe any effect of supplemented glycine on OSCK-induced reprogramming (data not shown).
To further determine whether TDH-mediated threonine catabolism is required for reprogramming, we used the culture medium deprived threonine, glycine, tryptophan, or valine. Intriguingly, deprivation of threonine resulted in a futile induction of iPS cells by either OSCK or OSCK plus TDH (Fig. 2F). In contrast, lacking of glycine, tryptophan, or valine did not show a significant effect on OSCK- or OSCK plus TDH-induced reprogramming (Fig. 2F and supporting information Fig. S2). These data indicate the specific effect of threonine on reprogramming. Since mouse ES cells are critically dependent on threonine for their growth, we sought to exclude the possibility that the failure of OSCK or OSCK plus TDH to induce iPS cell generation in the absence of threonine is due to the lack of nutrition support for iPS cell growth. To this end, MEF cells were preincubated with the threonine-depleted culture medium for the increasing periods of time and then cultured with the complete medium until the GFP-positive colonies were counted after a total of 16 days induction. As shown in Figure 2G, 6 days preincubation of MEF cells with the threonine-depleted culture medium greatly reduced the reprogramming efficiency by either OSCK or OSCK plus TDH inductions. Moreover, preincubation of MEF cells with the threonine-depleted culture medium for either 10 or 16 days resulted in a complete failure of iPS cell generation (Fig. 2G). The subsequent real-time RT-PCR and cell proliferation analyses showed that incubation with the threonine-depleted culture medium did not affect TDH mRNA levels and cell growth rate of MEF cells 6 days after OSCK induction (Fig. 2H, 2I). Taken together, these results strongly suggest that TDH-mediated threonine catabolism is essential for the reprogramming progress.
TDH is Post-Transcriptionally Regulated by MiR-9
MicroRNA is known to regulate gene expression at the post-transcriptional level. Increasing evidence suggests that miRNA plays an important role in almost all cellular processes. Therefore, we sought to investigate whether TDH expression could be regulated by miRNA. We searched the TargetScan database for the potential miRNAs that could target the 3′UTR of TDH. Through this bioinformatic analysis, four miRNA candidates were identified (supporting information Fig. S3A). However, only miR-9 showed the inhibitory effect on TDH 3′UTR-luciferase activity and TDH protein expression (supporting information Fig. S3B-S3D). Therefore, we focused on the study of microRNA-9. The 3′UTR of TDH contains a putative region (nucleotides 278-285) that matched to the miR-9 seed region (Fig. 3A). The subsequent real-time RT-PCR analysis showed that miR-9 had an expression pattern opposite to that of TDH, and miR-9 levels were much lower in iPS and ES cells than those in MEF cells (Figs. 1A, 3B), further supporting that TDH may be a real target of miR-9. We then introduced miR-9 mimics into MEF cells together with luciferase reporter plasmids containing wild-type or mutant 3′UTR of TDH (Fig. 3A). The reporter activity was noticeably suppressed by the presence of wild-type 3′UTR of TDH, however, this activity was reversed when the 3′UTR was mutated (Fig. 3C, supporting information Fig. S4). In addition, the TDH 3′UTR-luciferase activity was increased by treatment with miR-9 inhibitors (supporting information Fig. S4). These data suggest that 3′UTR of TDH was inhibited by miR-9. In supporting of this, treatment of miR-9 inhibitors resulted in the elevated protein levels of TDH in R1 ES cells, whereas miR-9 mimics showed the opposite effect (Fig. 3D). However, neither miRNA-9 inhibitors nor miR-9 mimics caused any change in TDH mRNA levels (Fig. 3E). Correlated with the effects of miR-9 inhibitors and mimics on TDH protein levels, introduction of miR-9 inhibitors into R1 ES cells increased TDH enzyme activity, whereas addition of miR-9 mimics decreased it (Fig. 3F). Taken together, these results suggest that TDH is post-transcriptionally regulated by miR-9.
To determine the functional consequence of miR-9-mediated TDH regulation, the effect of miR-9 on reprogramming was evaluated. We introduced the mimics or inhibitors of miR-9 into MEF cells together with OSCK transduction. As shown in Figure 3G and 3H, miR-9 inhibitors enhanced OSCK-induced reprogramming efficiency, whereas miR-9 mimics inhibited it. To further examine whether miR-9 regulates reprogramming by targeting TDH, miR-9 inhibitors were introduced into MEF cells expressing TDH shRNA or control shRNA together with OSCK transduction. As shown in Figure 3I, the promoting effect of miR-9 inhibitors on reprogramming was almost completely reversed by TDH knockdown. These results suggest the importance of the miR-9-TDH axis in regulating somatic cell reprogramming. Since TDH was specifically expressed in ES cells and was downregulated upon leukemia inhibitory factor (LIF) withdrawal-induced differentiation (Fig. 3J), we sought to determine whether miR-9 is involved in LIF withdrawal-induced downregulation of TDH. The real-time RT-PCR analysis showed that levels of miR-9 were strongly increased in ES cells after deprivation of LIF (Fig. 3K). Furthermore, treatment of miR-9 inhibitors attenuated the LIF withdrawal-induced downregulation of TDH in ES cells (Fig. 3L). These data indicate that miR-9 also plays an important role in regulating TDH expression during ES cell differentiation.
TDH Physically Interacts with PRMT5
To understand the mechanisms whereby TDH promotes reprogramming, we sought to identify new TDH-interacting partners. Using an affinity purification approach, a protein with a relative molecular weight of approximately 70 kDa that specifically bound to the recombinant TDH protein was identified and mass spectrometry analysis revealed it as the PRMT5 protein (Fig. 4A). The subsequent subcellular fractionation analysis revealed the mitochondrial colocalization of TDH and PRMT5 (Fig. 4B). To verify the interaction of TDH with PRMT5, we expressed GFP-PRMT5 together with Flag-TDH in HEK 293T cells. Reciprocal coimmunoprecipitation experiments using either anti-GFP or anti-Flag antibody revealed the interaction between exogenously expressed TDH and PRMT5 (Fig. 4C, 4D). The interaction of endogenous TDH with PRMT5 was also confirmed by the coimmunoprecipitation assay using anti-PRMT5 antibody (Fig. 4E). In an in vitro GST-pull down assay with recombinant glutatione S-trnasferase (GST)-TDH and Flag-PRMT5 proteins, these two proteins bind directly to each other (Fig. 4F). The interaction between TDH and PRTM5 takes place in the mitochondria, as manifested by the presence of TDH in the anti-PRMT5 immunoprecipitates from the mitochondrial fraction, but not from the cytoplasmic fraction (Fig. 4G). Together, these results indicate that PRMT5 is a new binding partner of TDH.
To delineate the TDH-binding domain in PRMT5, we generated a panel of PRMT5 deletion mutants (supporting information Fig. S5A, S5C). The immunoprecipitation analyses demonstrated that C-terminal 457 amino acids (amino acids 181-637) of PRMT5 were involved in the association with TDH (supporting information Fig. S5B). Interestingly, PRMT5Δ, a methyltransferase inactive mutant of PRMT5 that lacks amino acids 360-372 , exhibited an even stronger binding affinity toward TDH than wild-type PRMT5 (supporting information Fig. S5D), we currently do not know the reason for this discrepancy. We also generated a panel of TDH deletion mutants including N-terminal region, the central enzymatic active domain, and C-terminal region (supporting information Fig. S5E). The C-terminal 173 amino acids (amino acids 200-373) of TDH were shown to be responsible for its binding to PRMT5 (supporting information Fig. S5F).
PRMT5 Regulates TDH Enzyme Activity in Both Methyltransferase-Dependent and -Independent Manners
Given the direct interaction of PRMT5 and TDH, we wondered if PRMT5 could regulate the enzyme activity of TDH. When PRMT5 was knocked down by two different sets of shRNAs in ES cells, the enzyme activity of TDH was greatly decreased (Fig. 5A). In addition, in MEF cells that lack endogenous TDH expression, knockdown of PRMT5 resulted in a decrease in the enzyme activity of exogenously expressed TDH (Fig. 5B). The subcellular fractionation experiments indicated that the mitochondrial localization of TDH was not affected by PRMT5 knockdown (Fig. 5C). These results suggest that the enzyme activity of TDH is indeed regulated by PRMT5.
We next determined whether PRMT5 methyltransferase activity is critical for regulating TDH enzyme activity. Wild-type PRMT5 or its methyltransferase inactive mutant PRMT5Δ was individually introduced into HEK 293T cells together with Flag-TDH, and to our surprise, PRMT5Δ noticeably increased TDH enzyme activity, although to a lesser extent compared with wild-type PRMT5 (Fig. 5D), indicating that PRMT5 may regulate TDH enzyme activity through both methyltransferase-dependent and -independent mechanisms. To further confirm this, we performed an in vitro assay with purified GST-TDH plus either Flag-PRMT5 or Flag-PRMT5Δ. The results showed that both PRMT5 and PRMT5Δ increased the enzyme activity of TDH even in the absence of S-adenosyl-L-methionine (SAM) (Fig. 5E, lanes 1-3), implying the existence of methyltransferase-independent effect of PRMT5 on TDH enzyme activity. Intriguingly, in the presence of SAM, the promoting effect of PRMT5 on TDH activity was further enhanced (Fig. 5E, lane 2 vs. 5). However, PRMT5Δ-mediated enhancement of TDH enzyme activity showed no obvious change in the absence and presence of SAM (Fig. 5E, lane 3 vs. 6). Taken together, these results suggest that PRMT5 could regulate TDH enzyme activity via both methyltransferase-dependent and -independent mechanisms.
Arg180 in TDH is Methylated by PRMT5
We next investigated whether TDH is subjected for methylation by PRMT5. To this end, an in vitro methylation assay was performed using purified GST-TDH plus either Flag-PRMT5 or Flag-PRMT5Δ in the presence of 3H-SAM. 3H-methyl-TDH was readily detected after incubation of TDH with PRMT5, but not with PRMT5Δ (Fig. 6A). This result suggests that TDH can be directly methylated by PRMT5. To identify the potential methylated arginine residue(s) in TDH mediated by PRMT5, we first evaluated the conserved arginine residues in TDH across different species, including mouse, zebrafish, Drosophila melanogaster, and Caenorhabditis elegans. Three arginine residues Arg180, Arg190, and Arg192 in TDH, which located in its enzymatic active domain, were found to be highly conserved (Fig. 6B). To determine whether these conserved arginine residues could be direct targets for methylation by PRMT5, the TDH mutant derivatives were prepared in which Arg180, Arg190, and Arg192 were individually altered to the similarly charged lysine residue, and the ability of TDH/R180K, TDH/R190K, or TDH/R192K to act as a substrate for PRMT5 was measured in vitro. Intriguingly, in contrast to TDH, TDH/R190K and TDH/R192K, which were efficiently methylated, TDH/R180K was unable to be methylated by PRMT5 (Fig. 6C). The subsequent in vivo methylation assay further confirmed the specific methylation of Arg180 in TDH by PRMT5 (Fig. 6D). Taken together, these results suggest that Arg180 in TDH is physiologically methylated by PRMT5.
PRMT5 Enhances the Promoting Effect of TDH on Reprogramming
To investigate the physiological significance of PRMT5-mediated TDH regulation, TDH was introduced into MEF cells expressing PRMT5 shRNA or control shRNA together with OSCK. When PRMT5 was knocked down, the promoting effect of TDH on iPS cell generation was inhibited (Fig. 7A, lane 2 vs. 4). In contrast, induction of PRMT5 increased TDH-induced reprogramming efficiency (Fig. 7B, lane 4 vs. 5). The mutant PRMT5Δ also exhibited a promoting effect on TDH-induced reprogramming, although to a lesser extent compared to wild-type PRMT5 (Fig. 7B, lane 5 vs. 6). This correlated well with the increased enzyme activity of TDH by PRMT5Δ (Fig. 5D). Together, these results indicate that PRMT5 promotes TDH-induced reprogramming by regulating TDH enzyme activity.
Given that Arg180 in TDH is physiologically methylated by PRMT5 (Fig. 6C, 6D), we sought to determine the importance of Arg180 methylation of TDH in mediating its enhancing effect on reprogramming. We first examined the enzyme activity of TDH/R180K. When Arg180 was mutated to lysine, the enzyme activity of TDH was decreased (Fig. 7C, lane 1 vs. 2). In addition, in contrast to wild-type TDH, the enzyme activity of TDH/R180K was no longer regulated by PRMT5 (Fig. 7C, lane 2 vs. 4 and lane 1 vs. 3). The residual activity of TDH/R180K (Fig. 7C, lane 2) indicates that TDH may be regulated by other factor(s) in addition to PRMT5. We next evaluated the effect of TDH/R180K on reprogramming. TDH/R180K exhibited a much weaker ability to enhance OSCK-induced reprogramming efficiency compared to wild-type TDH (Fig. 7D, lanes 1-3). Furthermore, although PRMT5 consistently enhanced TDH-induced reprogramming efficiency (Fig. 7D, lane 2 vs. 5), it failed to show any effect on TDH/R180K-induced reprogramming (Fig. 7D, lane 3 vs. 6). Therefore, these results suggest that PRMT5-mediated methylation of TDH at arginine residue (R180) plays an important role in regulating somatic cell reprogramming.
Emerging evidence suggests that the metabolic transition from oxidative phosphorylation to glycolysis facilitates somatic cell reprogramming [12–14]. Our study for the first time demonstrates that TDH-mediated threonine catabolism promotes reprogramming efficiency, which provides another solid evidence to support the important function of metabolic remodeling in regulating somatic cell reprogramming. TDH-mediated hydrolysis of threonine produces glycine and acetyl-CoA, which provide fuel for nucleotide synthesis and ATP generation, respectively. The observation that 3-HNV, a threonine analog, inhibits TDH-induced reprogramming efficiency (Fig. 2 and E) suggests that acetyl-CoA mediates the downstream effect of TDH on reprogramming, since the propionyl-CoA, generated by TDH-meditated hydrolyzation of 3-HNV, cannot be used by TCA cycle and ATP generation. In contrast to 3-HNV, supplemented glycine fails to show any effect on reprogramming efficiency. This indicates that glycine may not be required for TDH-promoted reprogramming. Nonetheless, we cannot exclude the possibility that cells are unable to efficiently deliver the supplemented glycine into mitochondria to feed the glycine cleavage system. The detailed mechanism of TDH-enhanced reprogramming remains to be further investigated.
The importance of miRNA-mediated protein post-transcriptional regulation has recently been recognized in almost all cellular processes. Here, we demonstrate that a microRNA miR-9 post-transcriptionally represses TDH expression by targeting its 3′UTR, through which miR-9 inhibits somatic cell reprogramming. Although miR-9 was recently shown to directly target the reprogramming factor LIN28 , the inhibitory effect of miR-9 on reprogramming appears to be dependent on TDH, as evidenced by the fact that inhibition of miR-9 fails to enhance reprogramming efficiency of TDH knockdown cells. In our study, we also notice that levels of miR-9 are decreasing during iPS cell induction (data not shown), correlating with increases in TDH protein expression and enzymatic activity during reprogramming (Fig. 1C, 1D). These results strongly suggest that miR-9 physiologically plays an important role in regulating TDH-facilitated reprogramming. miR-9 has been previously identified as a regulator of specific differentiation pathways. For instance, miR-9, expressed early during ES cell differentiation, participates in neuronal differentiation [21, 22]. In addition, inhibition of miR-9 prevents downregulation of SIRT1 during ES cell differentiation . Moreover, treatment of miR-9 inhibitors attenuates downregulation of TDH during ES cell differentiation initiated by LIF withdrawal (Fig. 3L). These combined findings suggest that miR-9 may regulate differentiation through modulating multiple cellular targets.
Arginine methylation is becoming increasingly known as an important type of protein modification, which has been implicated in the regulation of various cellular processes, including RNA processing, transcriptional regulation, and signal transduction [24, 25]. The methylation of arginine residues is catalyzed by PRMTs . PRMT5, a member of the PRMT family of enzymes, regulates diverse cellular processes by catalyzing arginine methylation of nonhistone proteins, as well as histones [24, 26]. For example, PRMT5 exerts as a transcriptional repressor by methylating histones H3 and H4 [27, 28]. PRMT5 is also shown to regulate the function of p53 and E2F1 by catalyzing their arginine methylation [19, 29]. Interestingly, a recent study has revealed that PRMT5-mediated H2A methylation in the cytoplasm of ES cells plays an important role in regulating pluripotency . Moreover, PRMT5 is also shown to be involved in the regulation of somatic cell reprogramming . In a search for potential TDH binding proteins using mass spectrometry, PRMT5 is identified as a novel interacting partner of TDH. Through direct interaction, PRMT5 mediates specific methylation at arginine 180 in TDH. This methylation modification increases the enzyme activity of TDH and enhances the promoting effect of TDH on reprogramming. Interestingly, PRMT5 also increases TDH enzyme activity and enhances TDH-induced reprogramming efficiency via a methyltransferase-independent mechanism, as manifested by the observation that PRMT5Δ, an enzymatic inactive mutant of PRMT5, exhibits the similar ability to increase TDH enzyme activity and reprogramming efficiency as wild-type PRMT5. In agreement with our findings, PRMT3 was recently shown to control ribosomal subunit homeostasis in a methyltransferase-independent manner . These results suggest that PRMT5 could regulate TDH enzyme activity and TDH-facilitated reprogramming via both methyltransferase-dependent and -independent mechanisms, despite the mechanism for the latter is still lacking.
Recent studies suggest that, in contrast to mouse, human lacks a functional TDH enzyme due to the tdh gene mutations [16, 33]. It appears that human primarily uses serine/threonine dehydratase pathway for threonine catabolism . Therefore, it would be interesting to know whether human somatic cell reprogramming could be facilitated by either accelerating serine/threonine dehydratase-mediated threonine catabolism or induction of mouse TDH. In summary, our data strongly suggest the involvement of threonine catabolism mediated by TDH in regulating somatic cell reprogramming and provide the detailed mechanisms whereby TDH undergoes post-transcriptional and post-translational regulation by miR-9 and PRMT5, respectively.
We thank Professors Shilai Bao and Yunyu Shi for kindly providing us with 3H-SAM. This work was supported by Grant (XDA01020104) from Chinese Academy of Sciences; Grants (2010CB912804 and 2011CB966302) from Ministry of Science and Technology of China; Grant (31030046) from National Natural Science Foundation of China; and the Fundamental Research Funds for Central Universities (USTC, WK2060190018).