Histone methyltransferase Dot1L is a coactivator for thyroid hormone receptor during Xenopus development

Abstract

Histone modifications are associated with transcriptional regulation by diverse transcription factors. Genome-wide correlation studies have revealed that histone activation marks and repression marks are associated with activated and repressed gene expression, respectively. Among the histone activation marks is histone H3 K79 methylation, which is carried out by only a single methyltransferase, disruptor of telomeric silencing-1–like (DOT1L). We have been studying thyroid hormone (T3)-dependent amphibian metamorphosis in two highly related species, the pseudo-tetraploid Xenopus laevis and diploid Xenopus tropicalis, as a model for postembryonic development, a period around birth in mammals that is difficult to study. We previously showed that H3K79 methylation levels are induced at T3 target genes during natural and T3-induced metamorphosis and that Dot1L is itself a T3 target gene. These suggest that T3 induces Dot1L expression, and Dot1L in turn functions as a T3 receptor (TR) coactivator to promote vertebrate development. We show here that in cotransfection studies or in the reconstituted frog oocyte in vivo transcription system, overexpression of Dot1L enhances gene activation by TR in the presence of T3. Furthermore, making use of the ability to carry out transgenesis in X. laevis and gene knockdown in X. tropicalis, we demonstrate that endogenous Dot1L is critical for T3-induced activation of endogenous TR target genes while transgenic Dot1L enhances endogenous TR function in premetamorphic tadpoles in the presence of T3. Our studies thus for the first time provide complementary gain- and loss-of functional evidence in vivo for a cofactor, Dot1L, in gene activation by TR during vertebrate development.—Wen, L., Fu, L., Shi, Y.-B. Histone methyltransferase Dot1L is a coactivator for thyroid hormone receptor during Xenopus development.

Epigenetic modifications of histones in the eukaryotic chromatin have been implicated in playing important roles in many cellular processes, including transcription, DNA repair, and cell cycle regulation (1–5). These modifications include posttranslational acetylation, methylation, phosphorylation, and ubiquitylation, and they have a significant influence on chromatin structure and on interactions between chromatin and chromatin-interacting proteins, thus affecting the chromatin function (1–6). On the basis of their associations with steady-state genome-wide mRNA levels, different histone modifications have been grouped into so-called activation or repression marks (2–4, 7–11).

Diverse families of epigenetic enzymes participate in the modification and demodification of histones, at least in vitro. Often, multiple enzymes can modify the same histone residues. An exception is the methylation of histone H3 lysine 79 (H3K79), an activation histone mark. Biochemical studies have shown that only a single known methyltransferase is capable of methylating H3K79 in vitro (1). This enzyme, disruptor of telomeric silencing-1–like (Dot1L), initially identified as a Dot1 in Saccharomyces cerevisiae (12), belongs to the lysine methyltransferase (KMT) family (1, 13–15). Unlike other KMTs, which contain a SET {Su(var)3–9, enhancer of zeste [E(Z)], and trithorax (trx)} domain, Dot1L is the only known non-SET-domain-containing KMT (1, 13, 16). Consistent with the in vitro findings, Dot1/Dot1L deletion in yeast, Drosophila, and mice leads to a complete loss of H3K79 methylation (17–19). Similarly, in vivo studies in the anuran amphibian Xenopus tropicalis suggest that Dot1L is the only functional H3K79 methyltransferase during Xenopus embryogenesis and tadpole growth period (20).

We are interested to determine the function of Dot1L during vertebrate development and its associated molecular mechanisms. We have chosen Xenopus metamorphosis as a model system. Amphibian metamorphosis resembles mammalian postembryonic development, a period around birth when plasma thyroid hormone (T3) concentration is high (21, 22). Unlike postembryonic development in mammals, which takes place mostly within the uterus and thus is difficult to manipulate, amphibian metamorphosis can be easily manipulated by controlling the availability of T3 to the animals. This makes it a good model to study how T3 regulates vertebrate development in vivo (21–24).

Earlier studies have shown that T3 regulates metamorphosis by regulating target gene transcription through T3 receptors (TRs) (25–32). For genes that are induced by T3, TRs function as heterodimers with 9-cis retinoic acid receptors (RXRs). These heterodimers bind to T3 response element (TRE) in or around the promoters of target genes and activate or repress their transcription by recruiting coactivator or corepressor complexes in the presence or absence of T3, respectively. Indeed, it has been shown that a number of T3 receptor (TR)-associated cofactors identified from in vitro and cell-culture studies are recruited by TR to endogenous target genes in a T3-dependent manner to regulate target gene transcription and Xenopus development (33–42). Many of these cofactors and associated complexes are capable of modifying histones, including acetylation/deacetylation and methylation. T3 has consistently been shown to induce histone acetylation and methylation, as well as to induce the removal of nucleosomes at endogenous target genes during metamorphosis (43, 44).

Interestingly, H3K79 methylation levels at TR target genes are increased both during natural metamorphosis, when endogenous T3 levels are high, or when premetamorphic tadpoles are treated with exogenous T3 (43). Furthermore, Dot1L is activated by T3 at the transcription level during Xenopus metamorphosis (45). Because Dot1L is the only enzyme known to be capable of methylation H3K79 in vitro, these findings suggest that Dot1L is transcriptionally activated by T3 to further enhances histone methylation, chromatin remodeling, and gene activation by TR during metamorphosis.

To test the idea that Dot1L functions as a TR coactivator during Xenopus development, we took advantage of the fact that the development, including metamorphosis, in Xenopus laevis and the related species X. tropicalis is highly conserved (44–52). This allowed us to make use of the ability to easily carry out transgenesis in the pseudo-tetraploid X. laevis (53) and to knock out the endogenous Dot1L gene in the diploid X. tropicalis (54–58) to increase or reduce Dot1L level in vivo, respectively. We report here our findings demonstrating that Dot1L is an important TR coactivator during Xenopus development.

MATERIALS AND METHODS

Animal rearing, staging, and treatment

Animal care and treatment were provided as approved by the Animal Use and Care Committee of Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health. Adult X. laevis and X. tropicalis were purchased from Nasco (Fort Atkinson, WI, USA). Embryos and tadpoles were staged according to Nieuwkoop and Faber (59).

Premetamorphic tadpoles were treated with or without 10 nM T3 for 18 h at 18 to 20°C (X. laevis) and at 25°C (X. tropicalis), and were humanely killed for gene expression analyses.

Plasmid constructs and restriction enzyme–mediated insertion transgenesis

The X. tropicalis Dot1L full coding region was cloned by using PCR with high-fidelity DNA polymerase PrimerStar (Takara, Kyoto, Japan) into pCR-TOPO and confirmed by sequencing. After digesting the resulting plasmid with restriction enzymes SacII and StuI, the Dot1L coding region was subcloned into the in vitro transcription vector pYSLZ60, which also added 3 copies of the FLAG tag to the N terminus of the Dot1L (pYSLZ60-Dot1L). The green fluorescent protein (GFP) coding sequences were amplified with primers 5′-GACAAGCTTCTCGAGATGGGTAAAGGAGAAGAACTTTTC-3′ (upstream primer, XhoI site in italics) and 5′-GAAGCTTGAGCTCGAGATCTGAGTC-3′ (downstream primer, XhoI site in italics) from pS65T-C1 (Clontech Laboratories, Mountain View, CA, USA), digested with XhoI, and inserted into XhoI-digested pYSLZ60 (calf intestinal alkaline phosphatase treated). The colonies with the correct orientation were identified by colony-PCR with T7 universal primer and the downstream primer used for the cloning. The final clone was sequenced to confirm the presence of the 3 copies of the FLAG tag at the N terminus of the GFP (pYSLZ60-GFP). The coding region for the FLAG-tagged Dot1L was cut out from pYSLZ60-Dot1L with restriction enzymes AgeI and EcoRI, and subcloned into a heat shock promoter-driven transgenesis vector (53) to produce pDP-HS-FLAG-Dot1L. For overexpression in culture cells and transgenesis, the heat shock promoter was replaced with the cytomegalovirus (CMV) promoter by using restriction enzymes KpnI and AgeI to produce the construct pDP-CMV-FLAG-Dot1L. Restriction enzyme–mediated insertion transgenesis was carried out as previously described (53, 60).

Cell-culture transfection and transcription assay

The vector pGL4.10 (Promega, Madison, WI, USA) was used to make the reporter construct with the firefly luciferase gene under the control of the T3-inducible X. tropicalis TH/bZIP. Briefly, the TH/bZIP promoter was PCR amplified from X. tropicalis genomic DNA with a primer set 5′-GAGCTCGGTACCGGGCTGTGCAGTCAGTACATATC-3′ (KpnI site underlined) and 5′-AAGCTTAGATCTCCTGAATGGGCAGCAGAACTTCC-3′ (BglII site underlined), and the high-fidelity DNA polymerase PrimerStar (Takara) and cloned into pCR Blunt II TOPO for sequencing. The TH/bZIP promoter fragment was excised from the TOPO vector by double digestion with SacI and HindIII, gel purified, and ligated into predigested pGL4.10 vector bearing the same restriction ends to generate the TH/bZIP promoter reporter plasmid.

The X. laevis A6 cells were cultured in the culture medium containing 60% L-15, 10% charcoal-stripped fetal bovine serum, and 1% streptomycin and ampicillin at 25°C in a humid incubator with an open water container inside. The day before transfection, 1.0 × 10⁴ A6 cells per well were seeded into a 96-well plate with culture medium but without antibiotics. Transfection was carried out by following the Lipofectamine 2000 manual (Thermo Fisher Scientific, Waltham, MA, USA). Briefly, for each well of the 96-well plate, 150 ng of pDP-CMV-FLAG-Dot1L or pDP-CMV-EGFP combined with 50 ng of TH/bZIP promoter reporter plasmid or Dot1L promoter plasmids (45) and 0.5 ng of phRG-TK Renilla luciferase plasmids were diluted in 25 μl Opti-MEM, and at the same time, 0.2 μl Lipofectamine 2000 was diluted in 25 μl of Opti-MEM. Then the Opti-MEM-diluted plasmids and Lipofectamine 2000 were combined and incubated for another 15 min. Next, 50 μl of the DNA Lipofectamine 2000 mixture was added to each well in the 96-well plate. One day after transfection, the medium was changed to culture medium containing 1% antibiotics. For T3 treatment, 20 nM T3 was added to the culture medium. Dual luciferase assays were carried out 3 d after transfection by following the dual luciferase assay manual (Promega, Madison, MI, USA).

Oocyte injection and transcription assay

An adult X. laevis female was humanely killed and dissected to obtain the ovary. The ovary was treated with 1 mg/ml collagenase (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at room temperature. Stage VI oocytes were selected for injection. X. tropicalis TRα and RXRα mRNA were in vitro transcribed as previously described (48). Dot1L and GFP mRNAs were in vitro transcribed from plasmid pYSLZ60-Dot1L and pYSLZ60-GFP, respectively, after linearization with NotI. Dot1L mRNA (4.6 ng/oocyte) or GFP mRNA (4.6 ng/oocyte) were coinjected with TRα/RXRα mRNA (230 pg/oocyte) into oocyte cytoplasm. After allowing the oocytes to recover for at least 2 h, TH/bZIP promoter plasmid (69 pg/oocyte) and phRG-TK Renilla luciferase plasmid (6.9 pg/oocyte) were coinjected in the nucleus of the oocytes. The injected oocytes were incubated at 18°C with or without treatment with 100 nM T3 for 18 h. At least 5 oocytes were pooled together and homogenized with 1× lysis buffer (20 μl/oocyte) for each independent sample. Three samples were collected for dual luciferase assay according to the manufacturer’s instructions.

RNA extraction and real-time quantitative PCR

Whole animals or the intestine from at least 3 tadpoles were homogenized together with Trizol reagent (Thermo Fisher Scientific, Waltham, MA, USA). Total RNA was extracted with Direct-zol RNA MiniPrep (Zymogen Energy, Shasta Lake, CA, USA) according to the manual. The RNA concentration was measured with NanoDrop (Thermo Fisher Scientific). The RNA was reverse transcribed with the QuantiTect reverse transcription kit (Qiagen, Germantown, MD, USA). Briefly, 1 μg total RNA was treated with genome DNA Wipe-out to remove genomic DNA and then reverse transcribed. qPCR was carried out via the SYBR Green method with 2× SYBR Green PCR mixture (Thermo Fisher Scientific). The primers used for real-time quantitative PCR (qPCR) analysis of gene expression are listed in Table 1.

TABLE 1.

Oligonucleotide primers used for qPCR and amplicon size

Gene	Primer sequence, 5′–3′		Size, bp
	Forward	Reverse
X. tropicalis
TH/bZIP	CCAAGGGAAACGGGTGGCTT	GTGCCACCTCTGCGGAAAGT	172
ST3	GGTTATGTGTGGCGCCTTCG	AATGGGAAAGGGCCCAGAGG	192
Klf9	GGCACAGGTGTCCTTATGCT	AAGGGCGTTCACCTGTATGG	93
X. laevis
TH/bZIP	CCACCTCCACAGAATCAGCAG	CAAGCAGAGAACGAGCAAGG	91
ST3	GGGGGTTCCATGAACCGTGA	GAGCTCACTCAGCCCCTCTG	147
Klf9	GCCCCAGTCAGGTCAACCAAT	CTTAAGGCAGTCAGGCCACG	150
X. tropicalis and X. laevis
EF1α	CGGAACTACCCTGCTGGAAG	GGCAAAGGTAACCACCATGC	172

Western blot analysis

Tadpoles were homogenized together with RIPA lysis buffer (20 mM Tris-HCl, pH 8; 2 mM EDTA, pH 8; 0.5% NP-40; 25 mM β-glycerophosphate; 100 mM NaF; 100 mM PMSF, proteinase inhibitor cocktail tablet). Western blot was carried out as previously described (61) by using anti-FLAG antibody (1:1000 dilution; Sigma-Aldrich), anti-β-actin antibody (1:3000 dilution; Abcam, Cambridge, MA, USA), anti-H3 antibody (1:5000 dilution; Abcam), or anti-H3K79me2 (1:1000 dilution; Abcam). The X-films were digitally scanned into a computer with a scanner (Epson Perfection V600 Photo; Epson, Long Beach, CA, USA) and analyzed with ImageJ software National Institutes of Health, Bethesda, MD, USA; https://imagej.nih.gov/ij/index.html).

Chromatin immunoprecipitation assay

Chromatin immunoprecipitation (ChIP) assay on oocyte samples was done essentially as previously described (37, 62, 63). Briefly, the cytoplasm of stage IV oocytes from X. laevis was injected with TR/RXR mRNAs (2.3 ng each/oocyte) along with FLAG-Dot1L mRNA (4.6 ng/oocyte) or FLAG-GFP mRNA (4.6 ng/oocyte). TH/bZIP promoter plasmid (57.5 pg/oocyte) and phRG-TK Renilla luciferase plasmid (11.5 pg/oocyte) were coinjected in the nucleus of the oocytes after the mRNA injection. The injected oocytes were incubated at 18°C with or without treatment with 100 nM T3 for 18 h and homogenized together to prepare extracts for ChIP assay. An aliquot of the extract equivalent to 2 to 5 oocytes was used for each ChIP assay, with 2 repeats. Each repeat had an input control, which had an aliquot of the extract equivalent to one fifth of that used for the ChIP assay. Anti-TR (new PB) (48) and anti-FLAG (Sigma-Aldrich) antibodies were used for ChIP of TR and FLAG-Dot1L, respectively, and salmon sperm DNA–protein G–agarose (Pierce, Rockford, IL, USA) was used to pull down the antibody-bound chromatins. After reversal of the DNA–protein cross-link, the immunoprecipitated DNA was purified with the MiniPrep PCR Purification kit (Qiagen, Germantown, MD, USA) and subjected to qPCR with a TaqMan primer/probe set targeting the thyroid hormone response element (TRE) within the TH/bZIP promoter (48). qPCR analysis was performed on a StepOne Plus Real-Time PCR System (Thermo Fisher Scientific).

Statistical analysis

All statistical analyses were carried out by 2-tailed Student’s t tests.

RESULTS

Dot1L overexpression increases gene activation by T3

To investigate whether Dot1L can coactivate transcription by TR in the presence of T3, we first analyzed its effect on two known TR target promoters in the A6 Xenopus cell line by cotransfection. Firefly luciferase reporters driven by the promoters of Xenopus T3 target genes Dot1L (45) or TH/bZIP (64) were cotransfected with a plasmid expressing Xenopus Dot1L or the control plasmid expressing GFP, both under the control of the CMV promoter. As shown in Fig. 1A, both promoters were induced when the cells were treated with T3. Overexpression of Dot1L led to increased transcriptional activation of both promoters, suggesting that Dot1L can function as a coactivator for endogenous TR.

Figure 1.

Dot1L enhances T3-activation of target genes by liganded TR. A) Dot1L overexpression increases T3 induction in cell cultures. Firefly luciferase driven by promoters of Xenopus T3 target genes Dot1L or TH/bZIP was cotransfected into A6 Xenopus cell line with plasmids pDP-CMV-FLAG-Dot1L, which expresses FLAG-tagged Dot1L (Dot11L) under CMV promoter, or control plasmids pDP-CMV-GFP (control), which expresses GFP under CMV promoter, respectively. All transfections also included phRG-TK to express Renilla luciferase under control of TK promoter as internal control. Transfected cells were treated with 20 nM T3 for 24 h before being collected for luciferase assay; ratio of firefly luciferase activity to Renilla luciferase activity (in arbitrary units) was plotted. T3 induced transcription from both promoters via endogenous TRs in cultured cells, and Dot1L overexpression enhanced transcription. pGL4-Basic was control reporter lacking functional promoter. B) Dot1L increases T3-induced transcription from TH/bZIP promoter in vivo in Xenopus oocytes. Dot1L mRNA or control GFP mRNA was coinjected with TR/RXR mRNAs into oocyte cytoplasm, followed by nuclear injection of reporter firefly luciferase plasmid driven by T3-dependent TH/bZIP promoter and internal control Renilla luciferase plasmid. After incubating overnight at 18°C in presence or absence of 100 nM T3, oocytes were collected for luciferase activity, with ratio of firefly luciferase activity to Renilla luciferase activity (in arbitrary units) as measure of TH/bZIP promoter activity. There was little effect by T3 or Dot1L in the absence of overexpressed TR/RXR, and in the presence of TR/RXR, Dot1L enhanced induction of TH/bZIP promoter by T3. *P < 0.05, **P < 0.01.

To further test the role of Dot1L in transcriptional activation by TR, we next investigated Dot1L function in the reconstituted Xenopus oocyte system, where we can study transcriptional regulation by TR in the context of chromatin, thus mimicking the promoter context in native chromatin (65, 66). To ensure response to T3, we expressed TR and RXR by microinjecting TR/RXR mRNAs into the oocyte cytoplasm because oocytes have little endogenous TR/RXR, followed by microinjecting the reporter DNA, firefly luciferase plasmid driven by the T3-dependent TH/bZIP promoter, and the internal control Renilla luciferase DNA into the oocyte nucleus. As shown in Fig. 1B, when Dot1L mRNA or the control GFP mRNA was coinjected with TR/RXR mRNAs into the oocyte cytoplasm, the reporter activity in the presence of T3 was strongly enhanced by the expression of Dot1L compared to GFP. In the absence of T3, there was little effect by Dot1L. In addition, the promoter had little activity in the absence of TR/RXR mRNA injection regardless of the presence or absence of T3 or Dot1L or GFP. Thus, Dot1L functions as a TR coactivator in vivo.

Knocking down endogenous Dot1L reduces T3 activation of endogenous target genes in premetamorphic X. tropicalis tadpoles

To investigate whether Dot1L is involved in gene activation by TR in vivo, we generated a transcription activator–like effector nuclease (TALEN) against X. tropicalis Dot1L (20). As reported previously, this TALEN, consisting of 2 arms targeting Dot1L exon 5 (Fig. 2A), could efficiently knock down the endogenous Dot1L (with about 90% of the Dot1L locus mutated) when the mRNAs encoding the arms were microinjected into fertilized eggs, leading to reduced histone H3K79 methylation (Fig. 2B, C) (20). The animals derived from the TALEN-injected fertilized eggs (the F₀ generation animals) developed normally to feeding-stage tadpoles (3 d old, stage 45) but began to die after 9 d of age, before the onset of metamorphosis (stage 54) (20). This makes it impossible to study the role of Dot1L in gene regulation by TR during natural metamorphosis (stages 54–66).

Figure 2.

TALEN against Dot1L can effectively knock down Dot1L activity in X. tropicalis tadpoles. A) Schematic diagram of Dot1L TALEN and its target sequence in X. tropicalis Dot1L. Four types of repeat variable diresidues recognizing nucleotide A, G, T, or C are shown in different-color boxes, respectively. Left arm contained FokI-ELD nuclease and right arm contained FokI-KKR nuclease, which, when both arms bind to their respective binding sites, heterodimerize to form functional nuclease to make double-stranded break in intervening sequence. B, C) Dot1L knockdown significantly suppresses H3K79me2 level. B) mRNAs encoding left and right arms of Dot1L TALEN were microinjected into fertilized eggs. Two weeks later, when animals reached premetamorphic stages 48–49, several tadpoles were humanely killed, and proteins were isolated from whole animals minus digestive tract for Western blot analysis. Film was digitally scanned and analyzed with ImageJ software. C) Band density of methylated H3K79 was normalized against that of actin (H3K79/Actin) or H3 (H3K79/H3), respectively, with ratios for control set as 1. Result from normalization against either actin or H3 showed significant reduction in H3K79me2 level in TALEN mRNA-injected tadpoles; apparent 2 bands for H3 were due to slight shift when putting down film for exposure.

Making use of the ability of tadpoles as early as the onset of feeding (stage 45) to respond to exogenous T3, we investigated whether the Dot1L knockdown affected the response of premetamorphic tadpoles to T3. We microinjected the mRNAs encoding the left and right arms of the Dot1L TALEN into fertilized eggs, and when the animals reached stage 48/49 (2 wk old), they were treated or not treated with 10 nM T3 for 18 h. The animals were then humanely killed for RNA extraction, and qPCR analysis was performed of several well-known T3 target genes. These included transcription factor TH/bZIP (64), TRβ (67), and matrix metalloproteinase stromelysin 3 (ST3) (68), all of which are known to be induced by T3 in all organs analyzed so far in X. laevis and/or X. tropicalis. As shown in Fig. 3, all 3 genes were indeed induced in wild-type (WT) animals after T3 treatment. In the Dot1L-knockdown siblings, the induction of all 3 genes were significantly reduced, supporting the notion of an important role of endogenous Dot1L in gene regulation by T3 in tadpoles.

Figure 3.

Dot1L knockdown reduces T3-induced target gene expression in X. tropicalis tadpoles. WT and Dot1L-knockdown tadpoles (as in Fig. 2) were treated with or without 10 nM T3 for 18 h, and several tadpoles were homogenized together for RNA extraction and qPCR analysis of expression of 3 well-known T3 target genes: TH/bZIP (A), TRβ (B), and ST3 (C). **P < 0.001 vs. WT.

Dot1L overexpression leads to increased H3K79 methylation and enhanced T3 induction of target genes in premetamorphic X. laevis tadpoles

To complement the above knockdown studies, we next investigated the effect of overexpressing Dot1L in premetamorphic tadpoles. To this end, we carried out transgenesis by using a double promoter construct approach in X. laevis (53). We used the CMV promoter to drive ubiquitous expression of FLAG-tagged Dot1L and lens-specific γ-crystallin promoter to drive the eye-specific expression of the marker GFP (Fig. 4A), which allowed easy identification of transgenic (green eyes) vs. WT (no green eyes) siblings under a fluorescent dissection microscope (Fig. 4B). We used the sperm nuclei–mediated transgenic procedure to generate transgenic animals and used the resulting F₀ animals to obtain F₁ generation transgenic tadpoles and WT siblings (Fig. 4B). Some stage 40 F₁ premetamorphic tadpoles were humanely killed, and proteins were isolated from the whole animals for Western blot analysis of the transgenic Dot1L (FLAG tagged). As shown in Fig. 4C, transgenic animals expressing FLAG-tagged Dot1L had higher levels of H3K79 methylation, demonstrating increased Dot1L function in the tadpoles.

Figure 4.

Transgenic overexpression of Dot1L increases H3K79 methylation in premetamorphic X. laevis tadpoles. A) Schematic diagram of double-promoter transgenic construct. FLAG-tagged Dot1L was driven by CMV promoter, while marker GFP was driven by lens-specific γ-crystallin promoter. B) Representative transgenic and WT F₁ generation tadpoles in bright and fluorescent view. Transgenic tadpoles had green fluorescent lens but otherwise were morphologically similar to their WT siblings. C) Overexpression of Dot1L increases H3K79me2 in tadpoles. Several tadpoles at stage 40 were humanely killed together, and proteins were isolated from whole animals for Western blot analysis of transgenic Dot1L (Flag), control actin (Actin), total histone H3 (H3), and dimethylated H3 K79 (H3K79me2). Significant increase in H3K79me2 level in transgenic tadpoles was observed.

We next treated stage 45 transgenic and WT tadpoles with or without 10 nM T3 for 18 h. Total RNA was isolated from whole animals and used for qPCR analysis of the expression of well-known T3 target genes TH/bZIP (64), Klf9 (69), and ST3 (68). These genes were found to be induced by T3 in WT animals, and their induction was significantly increased by the transgenic overexpression of Dot1L (Fig. 5). Thus, Dot1L can function as a coactivator for TR in tadpoles.

Figure 5.

Transgenic overexpression of Dot1L enhances T3-induced target gene expression in premetamorphic tadpoles. Stage 45 transgenic and WT X. laevis tadpoles were treated with or without 10 nM T3 for 18 h at 20°C. Total RNA was isolated from whole animals and used for qPCR analysis of expression of 3 known T3 target genes: Klf9 (A), TH/bZIP (B), and ST3 (C). Overexpression of Dot1L significantly enhanced T3-induced expression of all target genes. **P < 0.001 vs. WT.

Dot1L enhances TR/RXR binding to TRE in vivo

To investigate how Dot1L enhances T3 induction of target gene expression, we again used the oocyte transcription system. We injected mRNAs encoding FLAG-Dot1L or FLAG-GFP (as a negative control) along with TR/RXR mRNAs into the cytoplasm of X. laevis oocytes. Subsequently, a mixture of TH/bZIP promoter luciferase plasmid and phRG-TK Renilla luciferase plasmid were injected into the oocyte nuclei. After incubation at 18°C for 18 h in the presence or absence of 100 nM T3, the oocytes were collected for ChIP assay with anti-TR antibody or anti-FLAG antibody (against FLAG-Dot1L or FLAG-GFP), and the presence of the TH/bZIP promoter region in the precipitated DNA was analyzed by qPCR. The results revealed a significant increase in TR/RXR binding to the TRE in the presence or absence of T3 when Dot1L was expressed (Fig. 6). Interestingly, anti-FLAG antibody failed to detect any difference between the FLAG-Dot1L- and FLAG-GFP-expressed samples in the presence or absence of T3, suggesting that the association of Dot1L, if any, with the TRE region was either too weak or too transient to be detected by the ChIP assay under our conditions. Regardless, the results suggest that Dot1L enhances gene activation by T3 via enhanced TR/RXR binding to the TRE, possibly via modifying histone H3.

Figure 6.

Dot1L enhances TR/RXR binding to TRE in vivo. TR/RXR mRNAs were injected into cytoplasm of X. laevis oocytes along with mRNA encoding FLAG-Dot1L or FLAG-GFP, respectively, followed by injection of TH/bZIP promoter plasmid and phRG-TK Renilla luciferase plasmid into nuclei. After incubation at 18°C for 18 h in presence or absence of 100 nM T3, oocytes were collected for ChIP assay with anti-TR antibody or anti-FLAG antibody (for FLAG-Dot1L or FLAG-GFP). DNAs precipitated by antibodies were purified with PCR purification kit and quantitated by qPCR analysis along with their input controls. Percentage of ChIP DNA relative to input was calculated. *P < 0.05.

DISCUSSION

Thyroid hormone affects diverse developmental, physiologic, and pathologic processes, among which anuran metamorphosis is the most dramatic. The total dependence of amphibian metamorphosis on T3 has enabled extensive studies to be conducted on the in vivo mechanisms of T3 action. Earlier studies have shown that TR is both necessary and sufficient to mediate the metamorphic effects of T3 during Xenopus metamorphosis (25–32). This makes metamorphosis a unique model to study how TR regulates gene expression and vertebrate development in vivo. Using this model, we have discovered here that the histone lysine methyltransferase Dot1L functions as a novel TR coactivator during development in vivo.

In vitro and cell-culture studies have long demonstrated that TR functions by recruiting cofactors to regulate target gene transcription. For T3-inducible genes, TR activates or represses their transcription by recruiting coactivator or corepressor complexes in the presence or absence of T3, respectively. Many TR cofactors have been identified through biochemical and molecular approaches. In the absence of T3, TR recruits corepressor complexes to the T3-inducible promoters (70–77). This leads to local histone deacetylation and transcriptional inhibition. T3 binding to TR leads to the release of corepressor complexes and recruitment of coactivator complexes to activate gene transcription. Such coactivator complexes are capable of disrupting or remodeling chromatin and/or modifying histones (41, 42, 45, 65, 76, 78–83).

Many transcription cofactors encoding histone modification enzymes have been identified. Their corresponding histone modifications have been classified as activation or repression histone marks on the basis of genome-wide association with gene expression levels in cultured cells (2, 4–11). Often histone modifications by enzymes that are transcriptional coactivators are histone activation marks, while those by corepressor enzymes are repression marks. Thus, histone acetyltransferases SRC1-3 and CBP/p300 and methyltransferase PRMT1 are transcription coactivators and the corresponding histone modifications are activation marks; furthermore, these coactivators have been implicated as TR coactivators in vitro, as well as during Xenopus metamorphosis (33, 34, 36, 42, 81, 84–86).

Histone H3K79 methylation is a well-known activation histone mark. Interestingly, H3K79 is one of the few histone marks located in the globular domain of histone H3 instead of the N or C terminus and is methylated in vitro by only a single known methyltransferase, Dot1L (1, 13, 16). Consistent with the in vitro findings, knocking down Dot1L in Xenopus tadpoles leads to a proportional decrease in the levels of methylated H3K79 (Fig. 2) (20), and Dot1 deletion in yeast, Drosophila, and mice leads to a complete loss of H3K79 methylation (17–19). Thus, our earlier observation that H3K79 methylation levels at T3 target genes are increased during natural and T3-induced metamorphosis suggests that Dot1L may function as a TR coactivator during Xenopus development.

Our cotransfection studies showed that Dot1L overexpression enhanced TR function in cell cultures, and our analyses in the frog oocyte model system demonstrated that Dot1L could coactivate T3 target promoter in a context mimicking native chromatin in vivo. More importantly, taking advantage of the ease of carrying out transgenesis in X. laevis and TALEN-mediated gene knockdown in the diploid X. tropicalis, we provided complementary evidence to support a role of Dot1L in gene regulation by T3. First, TALEN-mediated knockdown of Dot1L reduced the levels of methylated H3K79 in tadpoles and the extent gene activation by T3, indicating that endogenous Dot1L is important for T3 induction of target genes during development. Second, transgenic overexpression of Dot1L increased the levels of methylated H3K79 in tadpoles as well as the expression levels of T3 target genes after T3 treatment of the transgenic premetamorphic tadpoles, demonstrating that the transgenic Dot1L enhances T3 induction of target genes by endogenous TR in tadpoles.

Our mechanistic studies in frog oocytes showed that Dot1L enhanced TR binding to the TRE in vivo. This may be responsible for the enhanced transcriptional activation of T3 target genes by T3 in the oocytes or during development. It remains unclear how Dot1L enhanced TR binding to the TRE. It is possible that the increased methylation of H3K79 may stabilize TR/RXR binding to the TRE in the chromatinized promoter context. Alternatively, Dot1L may be associated with TR bound to the promoter, which may stabilize TR/RXR binding in vivo. This association, if it exists, is weak or transient because our ChIP assay did not detect any association of Dot1L with the TRE region in vivo. On the other hand, our earlier observation that T3 treatment of premetamorphic tadpoles enhanced H3K79 methylation at the TRE regions in vivo supports the recruitment of Dot1L to the TRE regions. Regardless, our findings here indicate that Dot1L can function as a coactivator for TR, although the exact mechanism may need further investigation.

While numerous in vitro and cell-culture studies have provided detailed molecular information on the properties and potential functions of many cofactors in gene regulation by TR, relatively few experiments have been carried out to investigate the in vivo roles of the cofactors during vertebrate development. Our studies here provide for the first time complementary gain- and loss-of functional evidence for a cofactor, Dot1L, in gene activation by TR during vertebrate development. Our findings suggest that as a TR coactivator, Dot1L is likely important for amphibian metamorphosis. In addition, because Dot1L is the only methyltransferase for H3K79 and methylated H3K79 is associated with the expression and activation of diverse genes by diverse transcription factors (1, 2, 4–6, 43, 45, 76), Dot1L may affect metamorphosis by functioning as a coactivator for transcription factors other than TR during metamorphosis. Thus, it would be of interest to investigate whether Dot1L affects Xenopus metamorphosis primarily through gene regulation by TR. Our earlier studies have shown that Dot1L is critical for tadpole growth and development before metamorphosis, as Dot1L-knockdown tadpoles (with about a 90% mutation rate) die before the onset of metamorphosis (20). This makes it impossible to study natural metamorphosis of Dot1L-knockout or -knockdown animals. While it was possible for us to study gene regulation in F₀ generation knockdown tadpoles after T3 treatment, analyzing the biologic consequence of Dot1L deletion would require a larger number of animals of homogeneous background. This can be approached by generating Dot1L-knockout tadpoles and inducing them to undergo metamorphosis with exogenous T3. Alternatively, it may be possible in the future to make use of gene editing technologies (54, 55, 87, 88) to conditionally knock out Dot1L only during metamorphosis. Such studies should allow us to determine how Dot1L affects vertebrate development.

Glossary

ChIP

chromatin immunoprecipitation

CMV

cytomegalovirus

Dot1L

disruptor of telomeric silencing-1–like

GFP

green fluorescent protein

KMT

lysine methyltransferase

qPCR

quantitative PCR

RXR

retinoic acid receptor

ST3

stromelysin 3

T3

thyroid hormone

TALEN

transcription activator–like effector nuclease

TR

T3 receptor

TRE

T3 response element

WT

wild type

AUTHOR CONTRIBUTIONS

L. Wen and Y.-B. Shi designed the research; L. Wen and L. Fu performed the research; and L. Wen, L. Fu, and Y.-B. Shi analyzed the data and wrote the article.